id	sid	tid	token	lemma	pos
fcis-30157	1	1	frontiers	frontier	NOUN
fcis-30157	1	2	in	in	ADP
fcis-30157	1	3	computing	computing	NOUN
fcis-30157	1	4	and	and	CCONJ
fcis-30157	1	5	intelligent	intelligent	ADJ
fcis-30157	1	6	systems	system	NOUN
fcis-30157	1	7	issn	issn	VERB
fcis-30157	1	8	:	:	PUNCT
fcis-30157	1	9	2832	2832	NUM
fcis-30157	1	10	-	-	SYM
fcis-30157	1	11	6024	6024	NUM
fcis-30157	1	12	|	|	NOUN
fcis-30157	1	13	vol	vol	NOUN
fcis-30157	1	14	.	.	PROPN
fcis-30157	2	1	11	11	NUM
fcis-30157	2	2	,	,	PUNCT
fcis-30157	2	3	no	no	INTJ
fcis-30157	2	4	.	.	NOUN
fcis-30157	2	5	3	3	NUM
fcis-30157	2	6	,	,	PUNCT
fcis-30157	2	7	2025	2025	NUM
fcis-30157	2	8	112	112	NUM
fcis-30157	2	9	research	research	NOUN
fcis-30157	2	10	on	on	ADP
fcis-30157	2	11	automatic	automatic	ADJ
fcis-30157	2	12	recognition	recognition	NOUN
fcis-30157	2	13	and	and	CCONJ
fcis-30157	2	14	auxiliary	auxiliary	ADJ
fcis-30157	2	15	diagnosis	diagnosis	NOUN
fcis-30157	2	16	of	of	ADP
fcis-30157	2	17	artificial	artificial	ADJ
fcis-30157	2	18	intelligence	intelligence	NOUN
fcis-30157	2	19	in	in	ADP
fcis-30157	2	20	skin	skin	NOUN
fcis-30157	2	21	diseases	disease	NOUN
fcis-30157	2	22	ping	ping	PROPN
fcis-30157	2	23	li	li	PROPN
fcis-30157	2	24	1	1	PROPN
fcis-30157	2	25	,	,	PUNCT
fcis-30157	2	26	a	a	DET
fcis-30157	2	27	,	,	PUNCT
fcis-30157	2	28	manyi	manyi	PROPN
fcis-30157	2	29	zhuang	zhuang	PROPN
fcis-30157	2	30	2	2	NUM
fcis-30157	2	31	,	,	PUNCT
fcis-30157	2	32	b	b	PROPN
fcis-30157	2	33	,	,	PUNCT
fcis-30157	2	34	jun	jun	PROPN
fcis-30157	2	35	tang	tang	PROPN
fcis-30157	2	36	3	3	NUM
fcis-30157	2	37	,	,	PUNCT
fcis-30157	2	38	c	c	PROPN
fcis-30157	2	39	,	,	PUNCT
fcis-30157	2	40	jian	jian	PROPN
fcis-30157	2	41	huang	huang	PROPN
fcis-30157	2	42	4	4	NUM
fcis-30157	2	43	,	,	PUNCT
fcis-30157	2	44	*	*	PUNCT
fcis-30157	2	45	1	1	NUM
fcis-30157	2	46	school	school	NOUN
fcis-30157	2	47	of	of	ADP
fcis-30157	2	48	humanities	humanity	NOUN
fcis-30157	2	49	and	and	CCONJ
fcis-30157	2	50	management	management	NOUN
fcis-30157	2	51	,	,	PUNCT
fcis-30157	2	52	youjiang	youjiang	PROPN
fcis-30157	2	53	medical	medical	PROPN
fcis-30157	2	54	university	university	PROPN
fcis-30157	2	55	for	for	ADP
fcis-30157	2	56	nationalities	nationality	NOUN
fcis-30157	2	57	,	,	PUNCT
fcis-30157	2	58	baise	baise	PROPN
fcis-30157	2	59	guangxi	guangxi	PROPN
fcis-30157	2	60	,	,	PUNCT
fcis-30157	2	61	533000	533000	NUM
fcis-30157	2	62	,	,	PUNCT
fcis-30157	2	63	china	china	PROPN
fcis-30157	2	64	2	2	NUM
fcis-30157	2	65	school	school	NOUN
fcis-30157	2	66	of	of	ADP
fcis-30157	2	67	public	public	ADJ
fcis-30157	2	68	health	health	NOUN
fcis-30157	2	69	,	,	PUNCT
fcis-30157	2	70	youjiang	youjiang	PROPN
fcis-30157	2	71	medical	medical	PROPN
fcis-30157	2	72	university	university	PROPN
fcis-30157	2	73	for	for	ADP
fcis-30157	2	74	nationalities	nationality	NOUN
fcis-30157	2	75	,	,	PUNCT
fcis-30157	2	76	baise	baise	PROPN
fcis-30157	2	77	guangxi	guangxi	PROPN
fcis-30157	2	78	,	,	PUNCT
fcis-30157	2	79	533000	533000	NUM
fcis-30157	2	80	,	,	PUNCT
fcis-30157	2	81	china	china	PROPN
fcis-30157	2	82	3	3	NUM
fcis-30157	2	83	school	school	NOUN
fcis-30157	2	84	of	of	ADP
fcis-30157	2	85	clinical	clinical	ADJ
fcis-30157	2	86	medicine	medicine	NOUN
fcis-30157	2	87	,	,	PUNCT
fcis-30157	2	88	youjiang	youjiang	PROPN
fcis-30157	2	89	medical	medical	PROPN
fcis-30157	2	90	university	university	PROPN
fcis-30157	2	91	for	for	ADP
fcis-30157	2	92	nationalities	nationality	NOUN
fcis-30157	2	93	,	,	PUNCT
fcis-30157	2	94	baise	baise	PROPN
fcis-30157	2	95	guangxi	guangxi	PROPN
fcis-30157	2	96	,	,	PUNCT
fcis-30157	2	97	533000	533000	NUM
fcis-30157	2	98	,	,	PUNCT
fcis-30157	2	99	china	china	PROPN
fcis-30157	2	100	4	4	NUM
fcis-30157	2	101	school	school	NOUN
fcis-30157	2	102	of	of	ADP
fcis-30157	2	103	humanities	humanity	NOUN
fcis-30157	2	104	and	and	CCONJ
fcis-30157	2	105	management	management	NOUN
fcis-30157	2	106	,	,	PUNCT
fcis-30157	2	107	youjiang	youjiang	PROPN
fcis-30157	2	108	medical	medical	PROPN
fcis-30157	2	109	university	university	PROPN
fcis-30157	2	110	for	for	ADP
fcis-30157	2	111	nationalities	nationality	NOUN
fcis-30157	2	112	,	,	PUNCT
fcis-30157	2	113	baise	baise	PROPN
fcis-30157	2	114	guangxi	guangxi	PROPN
fcis-30157	2	115	,	,	PUNCT
fcis-30157	2	116	533000	533000	NUM
fcis-30157	2	117	,	,	PUNCT
fcis-30157	2	118	china	china	PROPN
fcis-30157	2	119	*	*	PUNCT
fcis-30157	2	120	corresponding	correspond	VERB
fcis-30157	2	121	author	author	NOUN
fcis-30157	2	122	:	:	PUNCT
fcis-30157	2	123	jian	jian	PROPN
fcis-30157	2	124	huang	huang	PROPN
fcis-30157	2	125	(	(	PUNCT
fcis-30157	2	126	email	email	NOUN
fcis-30157	2	127	:	:	PUNCT
fcis-30157	2	128	00842@ymun.edu.cn	00842@ymun.edu.cn	NUM
fcis-30157	2	129	)	)	PUNCT
fcis-30157	2	130	,	,	PUNCT
fcis-30157	2	131	a	a	DET
fcis-30157	2	132	lipingyyds@zohomail.cn	lipingyyds@zohomail.cn	PROPN
fcis-30157	2	133	,	,	PUNCT
fcis-30157	2	134	b	b	PROPN
fcis-30157	2	135	elainechong@outlook.com	elainechong@outlook.com	PROPN
fcis-30157	2	136	,	,	PUNCT
fcis-30157	2	137	c	c	PROPN
fcis-30157	2	138	ksjbshd@outlook.com	ksjbshd@outlook.com	X
fcis-30157	3	1	abstract	abstract	ADJ
fcis-30157	3	2	:	:	PUNCT
fcis-30157	3	3	artificial	artificial	ADJ
fcis-30157	3	4	intelligence	intelligence	NOUN
fcis-30157	3	5	(	(	PUNCT
fcis-30157	3	6	ai	ai	NOUN
fcis-30157	3	7	)	)	PUNCT
fcis-30157	3	8	has	have	AUX
fcis-30157	3	9	made	make	VERB
fcis-30157	3	10	significant	significant	ADJ
fcis-30157	3	11	strides	stride	NOUN
fcis-30157	3	12	in	in	ADP
fcis-30157	3	13	skin	skin	NOUN
fcis-30157	3	14	disease	disease	NOUN
fcis-30157	3	15	diagnosis	diagnosis	NOUN
fcis-30157	3	16	,	,	PUNCT
fcis-30157	3	17	demonstrating	demonstrate	VERB
fcis-30157	3	18	impressive	impressive	ADJ
fcis-30157	3	19	potential	potential	NOUN
fcis-30157	3	20	in	in	ADP
fcis-30157	3	21	the	the	DET
fcis-30157	3	22	classification	classification	NOUN
fcis-30157	3	23	and	and	CCONJ
fcis-30157	3	24	detection	detection	NOUN
fcis-30157	3	25	of	of	ADP
fcis-30157	3	26	conditions	condition	NOUN
fcis-30157	3	27	such	such	ADJ
fcis-30157	3	28	as	as	ADP
fcis-30157	3	29	skin	skin	NOUN
fcis-30157	3	30	cancer	cancer	NOUN
fcis-30157	3	31	.	.	PUNCT
fcis-30157	4	1	however	however	ADV
fcis-30157	4	2	,	,	PUNCT
fcis-30157	4	3	several	several	ADJ
fcis-30157	4	4	challenges	challenge	NOUN
fcis-30157	4	5	remain	remain	VERB
fcis-30157	4	6	that	that	PRON
fcis-30157	4	7	hinder	hinder	VERB
fcis-30157	4	8	the	the	DET
fcis-30157	4	9	full	full	ADJ
fcis-30157	4	10	clinical	clinical	ADJ
fcis-30157	4	11	adoption	adoption	NOUN
fcis-30157	4	12	of	of	ADP
fcis-30157	4	13	ai	ai	NOUN
fcis-30157	4	14	-	-	PUNCT
fcis-30157	4	15	driven	drive	VERB
fcis-30157	4	16	diagnostic	diagnostic	ADJ
fcis-30157	4	17	systems	system	NOUN
fcis-30157	4	18	.	.	PUNCT
fcis-30157	5	1	this	this	DET
fcis-30157	5	2	paper	paper	NOUN
fcis-30157	5	3	reviews	review	VERB
fcis-30157	5	4	the	the	DET
fcis-30157	5	5	current	current	ADJ
fcis-30157	5	6	advancements	advancement	NOUN
fcis-30157	5	7	and	and	CCONJ
fcis-30157	5	8	key	key	ADJ
fcis-30157	5	9	research	research	NOUN
fcis-30157	5	10	areas	area	NOUN
fcis-30157	5	11	aimed	aim	VERB
fcis-30157	5	12	at	at	ADP
fcis-30157	5	13	optimizing	optimize	VERB
fcis-30157	5	14	ai	ai	NOUN
fcis-30157	5	15	models	model	NOUN
fcis-30157	5	16	for	for	ADP
fcis-30157	5	17	dermatological	dermatological	ADJ
fcis-30157	5	18	applications	application	NOUN
fcis-30157	5	19	.	.	PUNCT
fcis-30157	6	1	key	key	ADJ
fcis-30157	6	2	areas	area	NOUN
fcis-30157	6	3	for	for	ADP
fcis-30157	6	4	future	future	ADJ
fcis-30157	6	5	research	research	NOUN
fcis-30157	6	6	include	include	VERB
fcis-30157	6	7	the	the	DET
fcis-30157	6	8	development	development	NOUN
fcis-30157	6	9	of	of	ADP
fcis-30157	6	10	more	more	ADV
fcis-30157	6	11	advanced	advanced	ADJ
fcis-30157	6	12	deep	deep	ADJ
fcis-30157	6	13	learning	learning	NOUN
fcis-30157	6	14	architectures	architecture	NOUN
fcis-30157	6	15	,	,	PUNCT
fcis-30157	6	16	such	such	ADJ
fcis-30157	6	17	as	as	ADP
fcis-30157	6	18	multimodal	multimodal	ADJ
fcis-30157	6	19	fusion	fusion	NOUN
fcis-30157	6	20	methods	method	NOUN
fcis-30157	6	21	,	,	PUNCT
fcis-30157	6	22	which	which	PRON
fcis-30157	6	23	integrate	integrate	VERB
fcis-30157	6	24	dermoscopic	dermoscopic	NOUN
fcis-30157	6	25	images	image	NOUN
fcis-30157	6	26	with	with	ADP
fcis-30157	6	27	structured	structured	ADJ
fcis-30157	6	28	clinical	clinical	ADJ
fcis-30157	6	29	data	datum	NOUN
fcis-30157	6	30	for	for	ADP
fcis-30157	6	31	more	more	ADJ
fcis-30157	6	32	comprehensive	comprehensive	ADJ
fcis-30157	6	33	diagnostics	diagnostic	NOUN
fcis-30157	6	34	.	.	PUNCT
fcis-30157	7	1	additionally	additionally	ADV
fcis-30157	7	2	,	,	PUNCT
fcis-30157	7	3	there	there	PRON
fcis-30157	7	4	is	be	VERB
fcis-30157	7	5	a	a	DET
fcis-30157	7	6	growing	grow	VERB
fcis-30157	7	7	emphasis	emphasis	NOUN
fcis-30157	7	8	on	on	ADP
fcis-30157	7	9	creating	create	VERB
fcis-30157	7	10	personalized	personalize	VERB
fcis-30157	7	11	ai	ai	NOUN
fcis-30157	7	12	models	model	NOUN
fcis-30157	7	13	that	that	PRON
fcis-30157	7	14	account	account	VERB
fcis-30157	7	15	for	for	ADP
fcis-30157	7	16	individual	individual	ADJ
fcis-30157	7	17	patient	patient	NOUN
fcis-30157	7	18	characteristics	characteristic	NOUN
fcis-30157	7	19	,	,	PUNCT
fcis-30157	7	20	such	such	ADJ
fcis-30157	7	21	as	as	ADP
fcis-30157	7	22	age	age	NOUN
fcis-30157	7	23	,	,	PUNCT
fcis-30157	7	24	gender	gender	NOUN
fcis-30157	7	25	,	,	PUNCT
fcis-30157	7	26	and	and	CCONJ
fcis-30157	7	27	genetics	genetic	NOUN
fcis-30157	7	28	,	,	PUNCT
fcis-30157	7	29	to	to	PART
fcis-30157	7	30	improve	improve	VERB
fcis-30157	7	31	diagnostic	diagnostic	ADJ
fcis-30157	7	32	accuracy	accuracy	NOUN
fcis-30157	7	33	.	.	PUNCT
fcis-30157	8	1	the	the	DET
fcis-30157	8	2	integration	integration	NOUN
fcis-30157	8	3	of	of	ADP
fcis-30157	8	4	explainable	explainable	ADJ
fcis-30157	8	5	ai	ai	NOUN
fcis-30157	8	6	(	(	PUNCT
fcis-30157	8	7	xai	xai	PROPN
fcis-30157	8	8	)	)	PUNCT
fcis-30157	8	9	techniques	technique	NOUN
fcis-30157	8	10	,	,	PUNCT
fcis-30157	8	11	auxiliary	auxiliary	ADJ
fcis-30157	8	12	tasks	task	NOUN
fcis-30157	8	13	like	like	ADP
fcis-30157	8	14	lesion	lesion	NOUN
fcis-30157	8	15	segmentation	segmentation	NOUN
fcis-30157	8	16	,	,	PUNCT
fcis-30157	8	17	and	and	CCONJ
fcis-30157	8	18	multi	multi	ADJ
fcis-30157	8	19	-	-	ADJ
fcis-30157	8	20	center	center	ADJ
fcis-30157	8	21	clinical	clinical	ADJ
fcis-30157	8	22	research	research	NOUN
fcis-30157	8	23	is	be	AUX
fcis-30157	8	24	essential	essential	ADJ
fcis-30157	8	25	to	to	PART
fcis-30157	8	26	ensure	ensure	VERB
fcis-30157	8	27	transparency	transparency	NOUN
fcis-30157	8	28	,	,	PUNCT
fcis-30157	8	29	trust	trust	NOUN
fcis-30157	8	30	,	,	PUNCT
fcis-30157	8	31	and	and	CCONJ
fcis-30157	8	32	generalizability	generalizability	NOUN
fcis-30157	8	33	across	across	ADP
fcis-30157	8	34	diverse	diverse	ADJ
fcis-30157	8	35	populations	population	NOUN
fcis-30157	8	36	.	.	PUNCT
fcis-30157	9	1	moreover	moreover	ADV
fcis-30157	9	2	,	,	PUNCT
fcis-30157	9	3	ethical	ethical	ADJ
fcis-30157	9	4	concerns	concern	NOUN
fcis-30157	9	5	related	relate	VERB
fcis-30157	9	6	to	to	ADP
fcis-30157	9	7	bias	bias	VERB
fcis-30157	9	8	,	,	PUNCT
fcis-30157	9	9	data	data	NOUN
fcis-30157	9	10	privacy	privacy	NOUN
fcis-30157	9	11	,	,	PUNCT
fcis-30157	9	12	and	and	CCONJ
fcis-30157	9	13	accountability	accountability	NOUN
fcis-30157	9	14	must	must	AUX
fcis-30157	9	15	be	be	AUX
fcis-30157	9	16	addressed	address	VERB
fcis-30157	9	17	to	to	PART
fcis-30157	9	18	ensure	ensure	VERB
fcis-30157	9	19	fairness	fairness	NOUN
fcis-30157	9	20	and	and	CCONJ
fcis-30157	9	21	transparency	transparency	NOUN
fcis-30157	9	22	in	in	ADP
fcis-30157	9	23	ai	ai	PROPN
fcis-30157	9	24	systems	system	NOUN
fcis-30157	9	25	.	.	PUNCT
fcis-30157	10	1	this	this	DET
fcis-30157	10	2	review	review	NOUN
fcis-30157	10	3	highlights	highlight	VERB
fcis-30157	10	4	the	the	DET
fcis-30157	10	5	importance	importance	NOUN
fcis-30157	10	6	of	of	ADP
fcis-30157	10	7	interdisciplinary	interdisciplinary	ADJ
fcis-30157	10	8	collaboration	collaboration	NOUN
fcis-30157	10	9	and	and	CCONJ
fcis-30157	10	10	proposes	propose	VERB
fcis-30157	10	11	future	future	ADJ
fcis-30157	10	12	directions	direction	NOUN
fcis-30157	10	13	to	to	PART
fcis-30157	10	14	enhance	enhance	VERB
fcis-30157	10	15	the	the	DET
fcis-30157	10	16	reliability	reliability	NOUN
fcis-30157	10	17	,	,	PUNCT
fcis-30157	10	18	scalability	scalability	NOUN
fcis-30157	10	19	,	,	PUNCT
fcis-30157	10	20	and	and	CCONJ
fcis-30157	10	21	clinical	clinical	ADJ
fcis-30157	10	22	applicability	applicability	NOUN
fcis-30157	10	23	of	of	ADP
fcis-30157	10	24	ai	ai	ADJ
fcis-30157	10	25	-	-	PUNCT
fcis-30157	10	26	driven	drive	VERB
fcis-30157	10	27	skin	skin	NOUN
fcis-30157	10	28	disease	disease	NOUN
fcis-30157	10	29	diagnosis	diagnosis	NOUN
fcis-30157	10	30	,	,	PUNCT
fcis-30157	10	31	ultimately	ultimately	ADV
fcis-30157	10	32	improving	improve	VERB
fcis-30157	10	33	patient	patient	ADJ
fcis-30157	10	34	outcomes	outcome	NOUN
fcis-30157	10	35	and	and	CCONJ
fcis-30157	10	36	accessibility	accessibility	NOUN
fcis-30157	10	37	to	to	PART
fcis-30157	10	38	care	care	VERB
fcis-30157	10	39	.	.	PUNCT
fcis-30157	11	1	keywords	keyword	NOUN
fcis-30157	11	2	:	:	PUNCT
fcis-30157	11	3	artificial	artificial	ADJ
fcis-30157	11	4	intelligence	intelligence	NOUN
fcis-30157	11	5	(	(	PUNCT
fcis-30157	11	6	ai	ai	PROPN
fcis-30157	11	7	)	)	PUNCT
fcis-30157	11	8	;	;	PUNCT
fcis-30157	11	9	skin	skin	NOUN
fcis-30157	11	10	disease	disease	NOUN
fcis-30157	11	11	diagnosis	diagnosis	NOUN
fcis-30157	11	12	;	;	PUNCT
fcis-30157	11	13	multimodal	multimodal	NOUN
fcis-30157	11	14	fusion	fusion	NOUN
fcis-30157	11	15	;	;	PUNCT
fcis-30157	11	16	personalized	personalized	ADJ
fcis-30157	11	17	diagnosis	diagnosis	NOUN
fcis-30157	11	18	;	;	PUNCT
fcis-30157	11	19	explainable	explainable	ADJ
fcis-30157	11	20	ai	ai	NOUN
fcis-30157	11	21	(	(	PUNCT
fcis-30157	11	22	xai	xai	PROPN
fcis-30157	11	23	)	)	PUNCT
fcis-30157	11	24	;	;	PUNCT
fcis-30157	11	25	data	datum	NOUN
fcis-30157	11	26	augmentation	augmentation	NOUN
fcis-30157	11	27	.	.	PUNCT
fcis-30157	12	1	1	1	X
fcis-30157	12	2	.	.	X
fcis-30157	12	3	introduction	introduction	NOUN
fcis-30157	12	4	1.1	1.1	NUM
fcis-30157	12	5	.	.	PUNCT
fcis-30157	13	1	automatic	automatic	ADJ
fcis-30157	13	2	recognition	recognition	NOUN
fcis-30157	13	3	techniques	technique	VERB
fcis-30157	13	4	1.1.1	1.1.1	NUM
fcis-30157	13	5	.	.	PUNCT
fcis-30157	14	1	image	image	NOUN
fcis-30157	14	2	based	base	VERB
fcis-30157	14	3	recognition	recognition	NOUN
fcis-30157	14	4	methods	method	VERB
fcis-30157	14	5	image	image	NOUN
fcis-30157	14	6	-	-	PUNCT
fcis-30157	14	7	based	base	VERB
fcis-30157	14	8	automatic	automatic	ADJ
fcis-30157	14	9	skin	skin	NOUN
fcis-30157	14	10	disease	disease	NOUN
fcis-30157	14	11	recognition	recognition	NOUN
fcis-30157	14	12	has	have	AUX
fcis-30157	14	13	evolved	evolve	VERB
fcis-30157	14	14	significantly	significantly	ADV
fcis-30157	14	15	,	,	PUNCT
fcis-30157	14	16	transitioning	transition	VERB
fcis-30157	14	17	from	from	ADP
fcis-30157	14	18	traditional	traditional	ADJ
fcis-30157	14	19	algorithms	algorithm	NOUN
fcis-30157	14	20	to	to	ADP
fcis-30157	14	21	modern	modern	ADJ
fcis-30157	14	22	deep	deep	ADJ
fcis-30157	14	23	learning	learning	NOUN
fcis-30157	14	24	methods	method	NOUN
fcis-30157	14	25	.	.	PUNCT
fcis-30157	15	1	early	early	ADJ
fcis-30157	15	2	traditional	traditional	ADJ
fcis-30157	15	3	methods	method	NOUN
fcis-30157	15	4	,	,	PUNCT
fcis-30157	15	5	such	such	ADJ
fcis-30157	15	6	as	as	ADP
fcis-30157	15	7	thresholding	thresholding	NOUN
fcis-30157	15	8	,	,	PUNCT
fcis-30157	15	9	edge	edge	NOUN
fcis-30157	15	10	detection	detection	NOUN
fcis-30157	15	11	,	,	PUNCT
fcis-30157	15	12	and	and	CCONJ
fcis-30157	15	13	histogram	histogram	NOUN
fcis-30157	15	14	analysis	analysis	NOUN
fcis-30157	15	15	,	,	PUNCT
fcis-30157	15	16	used	use	VERB
fcis-30157	15	17	handcrafted	handcraft	VERB
fcis-30157	15	18	features	feature	NOUN
fcis-30157	15	19	to	to	PART
fcis-30157	15	20	identify	identify	VERB
fcis-30157	15	21	patterns	pattern	NOUN
fcis-30157	15	22	in	in	ADP
fcis-30157	15	23	skin	skin	NOUN
fcis-30157	15	24	images	image	NOUN
fcis-30157	15	25	,	,	PUNCT
fcis-30157	15	26	but	but	CCONJ
fcis-30157	15	27	struggled	struggle	VERB
fcis-30157	15	28	with	with	ADP
fcis-30157	15	29	issues	issue	NOUN
fcis-30157	15	30	like	like	ADP
fcis-30157	15	31	image	image	NOUN
fcis-30157	15	32	variability	variability	NOUN
fcis-30157	15	33	and	and	CCONJ
fcis-30157	15	34	complex	complex	ADJ
fcis-30157	15	35	lesion	lesion	NOUN
fcis-30157	15	36	morphology	morphology	NOUN
fcis-30157	15	37	,	,	PUNCT
fcis-30157	15	38	leading	lead	VERB
fcis-30157	15	39	to	to	ADP
fcis-30157	15	40	limited	limited	ADJ
fcis-30157	15	41	accuracy	accuracy	NOUN
fcis-30157	16	1	[	[	X
fcis-30157	16	2	1	1	NUM
fcis-30157	16	3	]	]	PUNCT
fcis-30157	16	4	.	.	PUNCT
fcis-30157	17	1	for	for	ADP
fcis-30157	17	2	example	example	NOUN
fcis-30157	17	3	,	,	PUNCT
fcis-30157	17	4	color	color	NOUN
fcis-30157	17	5	histogram	histogram	NOUN
fcis-30157	17	6	analysis	analysis	NOUN
fcis-30157	17	7	for	for	ADP
fcis-30157	17	8	melanoma	melanoma	NOUN
fcis-30157	17	9	detection	detection	NOUN
fcis-30157	17	10	showed	show	VERB
fcis-30157	17	11	success	success	NOUN
fcis-30157	17	12	in	in	ADP
fcis-30157	17	13	controlled	control	VERB
fcis-30157	17	14	environments	environment	NOUN
fcis-30157	17	15	but	but	CCONJ
fcis-30157	17	16	failed	fail	VERB
fcis-30157	17	17	in	in	ADP
fcis-30157	17	18	real	real	ADJ
fcis-30157	17	19	-	-	PUNCT
fcis-30157	17	20	world	world	NOUN
fcis-30157	17	21	scenarios	scenario	NOUN
fcis-30157	17	22	with	with	ADP
fcis-30157	17	23	varying	vary	VERB
fcis-30157	17	24	skin	skin	NOUN
fcis-30157	17	25	tones	tone	NOUN
fcis-30157	17	26	and	and	CCONJ
fcis-30157	17	27	lighting	light	VERB
fcis-30157	17	28	[	[	X
fcis-30157	17	29	2	2	NUM
fcis-30157	17	30	]	]	PUNCT
fcis-30157	17	31	.	.	PUNCT
fcis-30157	18	1	these	these	DET
fcis-30157	18	2	limitations	limitation	NOUN
fcis-30157	18	3	prompted	prompt	VERB
fcis-30157	18	4	the	the	DET
fcis-30157	18	5	adoption	adoption	NOUN
fcis-30157	18	6	of	of	ADP
fcis-30157	18	7	deep	deep	ADJ
fcis-30157	18	8	learning	learning	NOUN
fcis-30157	18	9	techniques	technique	NOUN
fcis-30157	18	10	.	.	PUNCT
fcis-30157	19	1	recent	recent	ADJ
fcis-30157	19	2	advancements	advancement	NOUN
fcis-30157	19	3	have	have	AUX
fcis-30157	19	4	been	be	AUX
fcis-30157	19	5	driven	drive	VERB
fcis-30157	19	6	by	by	ADP
fcis-30157	19	7	convolutional	convolutional	ADJ
fcis-30157	19	8	neural	neural	ADJ
fcis-30157	19	9	networks	network	NOUN
fcis-30157	19	10	(	(	PUNCT
fcis-30157	19	11	cnns	cnns	PROPN
fcis-30157	19	12	)	)	PUNCT
fcis-30157	19	13	,	,	PUNCT
fcis-30157	19	14	which	which	PRON
fcis-30157	19	15	automatically	automatically	ADV
fcis-30157	19	16	extract	extract	VERB
fcis-30157	19	17	hierarchical	hierarchical	ADJ
fcis-30157	19	18	features	feature	NOUN
fcis-30157	19	19	from	from	ADP
fcis-30157	19	20	images	image	NOUN
fcis-30157	19	21	,	,	PUNCT
fcis-30157	19	22	making	make	VERB
fcis-30157	19	23	them	they	PRON
fcis-30157	19	24	highly	highly	ADV
fcis-30157	19	25	effective	effective	ADJ
fcis-30157	19	26	for	for	ADP
fcis-30157	19	27	skin	skin	NOUN
fcis-30157	19	28	disease	disease	NOUN
fcis-30157	19	29	classification	classification	NOUN
fcis-30157	19	30	.	.	PUNCT
fcis-30157	20	1	cnns	cnns	PROPN
fcis-30157	20	2	have	have	AUX
fcis-30157	20	3	shown	show	VERB
fcis-30157	20	4	performance	performance	NOUN
fcis-30157	20	5	comparable	comparable	ADJ
fcis-30157	20	6	to	to	ADP
fcis-30157	20	7	dermatologists	dermatologist	NOUN
fcis-30157	20	8	,	,	PUNCT
fcis-30157	20	9	with	with	ADP
fcis-30157	20	10	a	a	DET
fcis-30157	20	11	study	study	NOUN
fcis-30157	20	12	achieving	achieve	VERB
fcis-30157	20	13	91	91	NUM
fcis-30157	20	14	%	%	NOUN
fcis-30157	20	15	accuracy	accuracy	NOUN
fcis-30157	20	16	in	in	ADP
fcis-30157	20	17	melanoma	melanoma	NOUN
fcis-30157	20	18	classification	classification	NOUN
fcis-30157	20	19	[	[	X
fcis-30157	20	20	4	4	NUM
fcis-30157	20	21	]	]	PUNCT
fcis-30157	20	22	.	.	PUNCT
fcis-30157	21	1	however	however	ADV
fcis-30157	21	2	,	,	PUNCT
fcis-30157	21	3	challenges	challenge	NOUN
fcis-30157	21	4	such	such	ADJ
fcis-30157	21	5	as	as	ADP
fcis-30157	21	6	the	the	DET
fcis-30157	21	7	need	need	NOUN
fcis-30157	21	8	for	for	ADP
fcis-30157	21	9	large	large	ADJ
fcis-30157	21	10	annotated	annotate	VERB
fcis-30157	21	11	datasets	dataset	NOUN
fcis-30157	21	12	and	and	CCONJ
fcis-30157	21	13	the	the	DET
fcis-30157	21	14	interpretability	interpretability	NOUN
fcis-30157	21	15	of	of	ADP
fcis-30157	21	16	model	model	NOUN
fcis-30157	21	17	decisions	decision	NOUN
fcis-30157	21	18	remain	remain	VERB
fcis-30157	21	19	[	[	X
fcis-30157	21	20	5	5	NUM
fcis-30157	21	21	]	]	PUNCT
fcis-30157	21	22	.	.	PUNCT
fcis-30157	22	1	other	other	ADJ
fcis-30157	22	2	deep	deep	ADJ
fcis-30157	22	3	learning	learning	NOUN
fcis-30157	22	4	architectures	architecture	NOUN
fcis-30157	22	5	,	,	PUNCT
fcis-30157	22	6	such	such	ADJ
fcis-30157	22	7	as	as	ADP
fcis-30157	22	8	resnet	resnet	NOUN
fcis-30157	22	9	and	and	CCONJ
fcis-30157	22	10	attention	attention	NOUN
fcis-30157	22	11	-	-	PUNCT
fcis-30157	22	12	based	base	VERB
fcis-30157	22	13	models	model	NOUN
fcis-30157	22	14	,	,	PUNCT
fcis-30157	22	15	have	have	AUX
fcis-30157	22	16	also	also	ADV
fcis-30157	22	17	been	be	AUX
fcis-30157	22	18	explored	explore	VERB
fcis-30157	22	19	.	.	PUNCT
fcis-30157	23	1	resnet	resnet	ADJ
fcis-30157	23	2	addresses	address	NOUN
fcis-30157	23	3	issues	issue	NOUN
fcis-30157	23	4	in	in	ADP
fcis-30157	23	5	very	very	ADV
fcis-30157	23	6	deep	deep	ADJ
fcis-30157	23	7	networks	network	NOUN
fcis-30157	23	8	,	,	PUNCT
fcis-30157	23	9	while	while	SCONJ
fcis-30157	23	10	attention	attention	NOUN
fcis-30157	23	11	models	model	NOUN
fcis-30157	23	12	enhance	enhance	VERB
fcis-30157	23	13	accuracy	accuracy	NOUN
fcis-30157	23	14	by	by	ADP
fcis-30157	23	15	focusing	focus	VERB
fcis-30157	23	16	on	on	ADP
fcis-30157	23	17	key	key	ADJ
fcis-30157	23	18	features	feature	NOUN
fcis-30157	23	19	of	of	ADP
fcis-30157	23	20	skin	skin	NOUN
fcis-30157	23	21	lesions	lesion	NOUN
fcis-30157	23	22	[	[	X
fcis-30157	23	23	6	6	NUM
fcis-30157	23	24	,	,	PUNCT
fcis-30157	23	25	7	7	NUM
fcis-30157	23	26	]	]	PUNCT
fcis-30157	23	27	.	.	PUNCT
fcis-30157	24	1	additionally	additionally	ADV
fcis-30157	24	2	,	,	PUNCT
fcis-30157	24	3	integrating	integrate	VERB
fcis-30157	24	4	multimodal	multimodal	NOUN
fcis-30157	24	5	data	datum	NOUN
fcis-30157	24	6	,	,	PUNCT
fcis-30157	24	7	like	like	ADP
fcis-30157	24	8	combining	combine	VERB
fcis-30157	24	9	skin	skin	NOUN
fcis-30157	24	10	images	image	NOUN
fcis-30157	24	11	with	with	ADP
fcis-30157	24	12	clinical	clinical	ADJ
fcis-30157	24	13	metadata	metadata	NOUN
fcis-30157	24	14	,	,	PUNCT
fcis-30157	24	15	has	have	AUX
fcis-30157	24	16	been	be	AUX
fcis-30157	24	17	shown	show	VERB
fcis-30157	24	18	to	to	PART
fcis-30157	24	19	improve	improve	VERB
fcis-30157	24	20	classification	classification	NOUN
fcis-30157	24	21	accuracy	accuracy	NOUN
fcis-30157	24	22	,	,	PUNCT
fcis-30157	24	23	as	as	ADP
fcis-30157	24	24	in	in	ADP
fcis-30157	24	25	the	the	DET
fcis-30157	24	26	case	case	NOUN
fcis-30157	24	27	of	of	ADP
fcis-30157	24	28	mdfnet	mdfnet	NOUN
fcis-30157	24	29	,	,	PUNCT
fcis-30157	24	30	which	which	PRON
fcis-30157	24	31	achieved	achieve	VERB
fcis-30157	24	32	92.3	92.3	NUM
fcis-30157	24	33	%	%	NOUN
fcis-30157	24	34	accuracy	accuracy	NOUN
fcis-30157	25	1	[	[	X
fcis-30157	25	2	8	8	NUM
fcis-30157	25	3	]	]	PUNCT
fcis-30157	25	4	.	.	PUNCT
fcis-30157	26	1	despite	despite	SCONJ
fcis-30157	26	2	these	these	DET
fcis-30157	26	3	advancements	advancement	NOUN
fcis-30157	26	4	,	,	PUNCT
fcis-30157	26	5	practical	practical	ADJ
fcis-30157	26	6	challenges	challenge	NOUN
fcis-30157	26	7	persist	persist	VERB
fcis-30157	26	8	.	.	PUNCT
fcis-30157	27	1	high	high	ADJ
fcis-30157	27	2	-	-	PUNCT
fcis-30157	27	3	quality	quality	NOUN
fcis-30157	27	4	annotated	annotate	VERB
fcis-30157	27	5	datasets	dataset	NOUN
fcis-30157	27	6	are	be	AUX
fcis-30157	27	7	difficult	difficult	ADJ
fcis-30157	27	8	and	and	CCONJ
fcis-30157	27	9	expensive	expensive	ADJ
fcis-30157	27	10	to	to	PART
fcis-30157	27	11	obtain	obtain	VERB
fcis-30157	27	12	,	,	PUNCT
fcis-30157	27	13	and	and	CCONJ
fcis-30157	27	14	deep	deep	ADJ
fcis-30157	27	15	learning	learning	NOUN
fcis-30157	27	16	models	model	NOUN
fcis-30157	27	17	can	can	AUX
fcis-30157	27	18	underperform	underperform	VERB
fcis-30157	27	19	when	when	SCONJ
fcis-30157	27	20	trained	train	VERB
fcis-30157	27	21	on	on	ADP
fcis-30157	27	22	limited	limited	ADJ
fcis-30157	27	23	or	or	CCONJ
fcis-30157	27	24	biased	biased	ADJ
fcis-30157	27	25	data	datum	NOUN
fcis-30157	27	26	[	[	X
fcis-30157	27	27	9	9	NUM
fcis-30157	27	28	]	]	PUNCT
fcis-30157	27	29	.	.	PUNCT
fcis-30157	28	1	interpretability	interpretability	NOUN
fcis-30157	28	2	remains	remain	VERB
fcis-30157	28	3	a	a	DET
fcis-30157	28	4	key	key	ADJ
fcis-30157	28	5	concern	concern	NOUN
fcis-30157	28	6	,	,	PUNCT
fcis-30157	28	7	as	as	SCONJ
fcis-30157	28	8	the	the	DET
fcis-30157	28	9	"	"	PUNCT
fcis-30157	28	10	black	black	ADJ
fcis-30157	28	11	box	box	NOUN
fcis-30157	28	12	"	"	PUNCT
fcis-30157	28	13	nature	nature	NOUN
fcis-30157	28	14	of	of	ADP
fcis-30157	28	15	deep	deep	ADJ
fcis-30157	28	16	learning	learning	NOUN
fcis-30157	28	17	models	model	NOUN
fcis-30157	28	18	complicates	complicate	VERB
fcis-30157	28	19	trust	trust	NOUN
fcis-30157	28	20	and	and	CCONJ
fcis-30157	28	21	adoption	adoption	NOUN
fcis-30157	28	22	in	in	ADP
fcis-30157	28	23	clinical	clinical	ADJ
fcis-30157	28	24	practice	practice	NOUN
fcis-30157	29	1	[	[	X
fcis-30157	29	2	5	5	NUM
fcis-30157	29	3	]	]	PUNCT
fcis-30157	29	4	.	.	PUNCT
fcis-30157	30	1	furthermore	furthermore	ADV
fcis-30157	30	2	,	,	PUNCT
fcis-30157	30	3	performance	performance	NOUN
fcis-30157	30	4	disparities	disparity	NOUN
fcis-30157	30	5	across	across	ADP
fcis-30157	30	6	different	different	ADJ
fcis-30157	30	7	skin	skin	NOUN
fcis-30157	30	8	tones	tone	NOUN
fcis-30157	30	9	highlight	highlight	VERB
fcis-30157	30	10	the	the	DET
fcis-30157	30	11	need	need	NOUN
fcis-30157	30	12	for	for	ADP
fcis-30157	30	13	more	more	ADJ
fcis-30157	30	14	inclusive	inclusive	ADJ
fcis-30157	30	15	datasets	dataset	NOUN
fcis-30157	30	16	[	[	X
fcis-30157	30	17	10	10	NUM
fcis-30157	30	18	,	,	PUNCT
fcis-30157	30	19	11	11	NUM
fcis-30157	30	20	]	]	PUNCT
fcis-30157	30	21	.	.	PUNCT
fcis-30157	31	1	in	in	ADP
fcis-30157	31	2	addition	addition	NOUN
fcis-30157	31	3	to	to	ADP
fcis-30157	31	4	classification	classification	NOUN
fcis-30157	31	5	,	,	PUNCT
fcis-30157	31	6	deep	deep	ADJ
fcis-30157	31	7	learning	learning	NOUN
fcis-30157	31	8	models	model	NOUN
fcis-30157	31	9	like	like	ADP
fcis-30157	31	10	u	u	NOUN
fcis-30157	31	11	-	-	NOUN
fcis-30157	31	12	net	net	ADJ
fcis-30157	31	13	have	have	AUX
fcis-30157	31	14	been	be	AUX
fcis-30157	31	15	used	use	VERB
fcis-30157	31	16	for	for	ADP
fcis-30157	31	17	segmentation	segmentation	NOUN
fcis-30157	31	18	and	and	CCONJ
fcis-30157	31	19	localization	localization	NOUN
fcis-30157	31	20	,	,	PUNCT
fcis-30157	31	21	improving	improve	VERB
fcis-30157	31	22	diagnostic	diagnostic	ADJ
fcis-30157	31	23	accuracy	accuracy	NOUN
fcis-30157	31	24	by	by	ADP
fcis-30157	31	25	providing	provide	VERB
fcis-30157	31	26	detailed	detailed	ADJ
fcis-30157	31	27	lesion	lesion	NOUN
fcis-30157	31	28	maps	map	NOUN
fcis-30157	31	29	[	[	X
fcis-30157	31	30	12	12	NUM
fcis-30157	31	31	]	]	PUNCT
fcis-30157	31	32	.	.	PUNCT
fcis-30157	32	1	however	however	ADV
fcis-30157	32	2	,	,	PUNCT
fcis-30157	32	3	the	the	DET
fcis-30157	32	4	integration	integration	NOUN
fcis-30157	32	5	of	of	ADP
fcis-30157	32	6	these	these	DET
fcis-30157	32	7	models	model	NOUN
fcis-30157	32	8	into	into	ADP
fcis-30157	32	9	clinical	clinical	ADJ
fcis-30157	32	10	practice	practice	NOUN
fcis-30157	32	11	requires	require	VERB
fcis-30157	32	12	careful	careful	ADJ
fcis-30157	32	13	validation	validation	NOUN
fcis-30157	32	14	and	and	CCONJ
fcis-30157	32	15	regulatory	regulatory	ADJ
fcis-30157	32	16	approval	approval	NOUN
fcis-30157	32	17	,	,	PUNCT
fcis-30157	32	18	with	with	ADP
fcis-30157	32	19	challenges	challenge	NOUN
fcis-30157	32	20	like	like	ADP
fcis-30157	32	21	data	datum	NOUN
fcis-30157	32	22	privacy	privacy	NOUN
fcis-30157	32	23	and	and	CCONJ
fcis-30157	32	24	model	model	NOUN
fcis-30157	32	25	bias	bias	NOUN
fcis-30157	32	26	still	still	ADV
fcis-30157	32	27	to	to	PART
fcis-30157	32	28	be	be	AUX
fcis-30157	32	29	addressed	address	VERB
fcis-30157	32	30	.	.	PUNCT
fcis-30157	33	1	1.1.2	1.1.2	X
fcis-30157	33	2	.	.	PUNCT
fcis-30157	33	3	feature	feature	NOUN
fcis-30157	33	4	extraction	extraction	NOUN
fcis-30157	33	5	algorithms	algorithm	NOUN
fcis-30157	33	6	feature	feature	NOUN
fcis-30157	33	7	extraction	extraction	NOUN
fcis-30157	33	8	algorithms	algorithm	NOUN
fcis-30157	33	9	are	be	AUX
fcis-30157	33	10	crucial	crucial	ADJ
fcis-30157	33	11	in	in	ADP
fcis-30157	33	12	the	the	DET
fcis-30157	33	13	automatic	automatic	ADJ
fcis-30157	33	14	recognition	recognition	NOUN
fcis-30157	33	15	of	of	ADP
fcis-30157	33	16	skin	skin	NOUN
fcis-30157	33	17	diseases	disease	NOUN
fcis-30157	33	18	,	,	PUNCT
fcis-30157	33	19	determining	determine	VERB
fcis-30157	33	20	the	the	DET
fcis-30157	33	21	quality	quality	NOUN
fcis-30157	33	22	of	of	ADP
fcis-30157	33	23	input	input	NOUN
fcis-30157	33	24	data	datum	NOUN
fcis-30157	33	25	for	for	ADP
fcis-30157	33	26	recognition	recognition	NOUN
fcis-30157	33	27	models	model	NOUN
fcis-30157	33	28	.	.	PUNCT
fcis-30157	34	1	this	this	DET
fcis-30157	34	2	section	section	NOUN
fcis-30157	34	3	reviews	review	VERB
fcis-30157	34	4	both	both	PRON
fcis-30157	34	5	manual	manual	ADJ
fcis-30157	34	6	and	and	CCONJ
fcis-30157	34	7	automatic	automatic	ADJ
fcis-30157	34	8	methods	method	NOUN
fcis-30157	34	9	,	,	PUNCT
fcis-30157	34	10	particularly	particularly	ADV
fcis-30157	34	11	those	those	PRON
fcis-30157	34	12	using	use	VERB
fcis-30157	34	13	deep	deep	ADJ
fcis-30157	34	14	learning	learning	NOUN
fcis-30157	34	15	.	.	PUNCT
fcis-30157	35	1	manual	manual	ADJ
fcis-30157	35	2	feature	feature	NOUN
fcis-30157	35	3	extraction	extraction	NOUN
fcis-30157	35	4	methods	method	NOUN
fcis-30157	35	5	.	.	PUNCT
fcis-30157	36	1	manual	manual	ADJ
fcis-30157	36	2	methods	method	NOUN
fcis-30157	36	3	,	,	PUNCT
fcis-30157	36	4	relying	rely	VERB
fcis-30157	36	5	on	on	ADP
fcis-30157	36	6	predefined	predefine	VERB
fcis-30157	36	7	rules	rule	NOUN
fcis-30157	36	8	and	and	CCONJ
fcis-30157	36	9	statistical	statistical	ADJ
fcis-30157	36	10	techniques	technique	NOUN
fcis-30157	36	11	,	,	PUNCT
fcis-30157	36	12	have	have	AUX
fcis-30157	36	13	been	be	AUX
fcis-30157	36	14	widely	widely	ADV
fcis-30157	36	15	used	use	VERB
fcis-30157	36	16	in	in	ADP
fcis-30157	36	17	traditional	traditional	ADJ
fcis-30157	36	18	image	image	NOUN
fcis-30157	36	19	processing	processing	NOUN
fcis-30157	36	20	.	.	PUNCT
fcis-30157	37	1	color	color	NOUN
fcis-30157	37	2	features	feature	NOUN
fcis-30157	37	3	,	,	PUNCT
fcis-30157	37	4	such	such	ADJ
fcis-30157	37	5	as	as	ADP
fcis-30157	37	6	histograms	histogram	NOUN
fcis-30157	37	7	,	,	PUNCT
fcis-30157	37	8	rgb	rgb	PROPN
fcis-30157	37	9	values	value	NOUN
fcis-30157	37	10	,	,	PUNCT
fcis-30157	37	11	and	and	CCONJ
fcis-30157	37	12	hsv	hsv	PROPN
fcis-30157	37	13	representations	representation	NOUN
fcis-30157	37	14	,	,	PUNCT
fcis-30157	37	15	are	be	AUX
fcis-30157	37	16	often	often	ADV
fcis-30157	37	17	utilized	utilize	VERB
fcis-30157	37	18	in	in	ADP
fcis-30157	37	19	skin	skin	NOUN
fcis-30157	37	20	disease	disease	NOUN
fcis-30157	37	21	diagnosis	diagnosis	NOUN
fcis-30157	37	22	.	.	PUNCT
fcis-30157	38	1	a	a	DET
fcis-30157	38	2	study	study	NOUN
fcis-30157	38	3	[	[	X
fcis-30157	38	4	6	6	NUM
fcis-30157	38	5	]	]	PUNCT
fcis-30157	38	6	showed	show	VERB
fcis-30157	38	7	that	that	SCONJ
fcis-30157	38	8	color	color	NOUN
fcis-30157	38	9	features	feature	NOUN
fcis-30157	38	10	alone	alone	ADV
fcis-30157	38	11	could	could	AUX
fcis-30157	38	12	achieve	achieve	VERB
fcis-30157	38	13	75	75	NUM
fcis-30157	38	14	%	%	NOUN
fcis-30157	38	15	accuracy	accuracy	NOUN
fcis-30157	38	16	in	in	ADP
fcis-30157	38	17	distinguishing	distinguish	VERB
fcis-30157	38	18	between	between	ADP
fcis-30157	38	19	benign	benign	ADJ
fcis-30157	38	20	and	and	CCONJ
fcis-30157	38	21	malignant	malignant	ADJ
fcis-30157	38	22	lesions	lesion	NOUN
fcis-30157	38	23	.	.	PUNCT
fcis-30157	39	1	however	however	ADV
fcis-30157	39	2	,	,	PUNCT
fcis-30157	39	3	these	these	DET
fcis-30157	39	4	methods	method	NOUN
fcis-30157	39	5	are	be	AUX
fcis-30157	39	6	sensitive	sensitive	ADJ
fcis-30157	39	7	to	to	ADP
fcis-30157	39	8	lighting	lighting	NOUN
fcis-30157	39	9	conditions	condition	NOUN
fcis-30157	39	10	and	and	CCONJ
fcis-30157	39	11	may	may	AUX
fcis-30157	39	12	not	not	PART
fcis-30157	39	13	generalize	generalize	VERB
fcis-30157	39	14	well	well	ADV
fcis-30157	39	15	.	.	PUNCT
fcis-30157	40	1	texture	texture	NOUN
fcis-30157	40	2	features	feature	NOUN
fcis-30157	40	3	,	,	PUNCT
fcis-30157	40	4	such	such	ADJ
fcis-30157	40	5	as	as	ADP
fcis-30157	40	6	gray	gray	ADJ
fcis-30157	40	7	level	level	NOUN
fcis-30157	40	8	co	co	NOUN
fcis-30157	40	9	-	-	NOUN
fcis-30157	40	10	occurrence	occurrence	ADJ
fcis-30157	40	11	matrix	matrix	NOUN
fcis-30157	40	12	(	(	PUNCT
fcis-30157	40	13	glcm	glcm	PROPN
fcis-30157	40	14	)	)	PUNCT
fcis-30157	40	15	113	113	NUM
fcis-30157	40	16	and	and	CCONJ
fcis-30157	40	17	local	local	ADJ
fcis-30157	40	18	binary	binary	ADJ
fcis-30157	40	19	patterns	pattern	NOUN
fcis-30157	40	20	(	(	PUNCT
fcis-30157	40	21	lbp	lbp	PROPN
fcis-30157	40	22	)	)	PUNCT
fcis-30157	40	23	,	,	PUNCT
fcis-30157	40	24	focus	focus	VERB
fcis-30157	40	25	on	on	ADP
fcis-30157	40	26	pixel	pixel	PROPN
fcis-30157	40	27	intensity	intensity	NOUN
fcis-30157	40	28	arrangements	arrangement	NOUN
fcis-30157	40	29	.	.	PUNCT
fcis-30157	41	1	lbp	lbp	PROPN
fcis-30157	41	2	,	,	PUNCT
fcis-30157	41	3	for	for	ADP
fcis-30157	41	4	instance	instance	NOUN
fcis-30157	41	5	,	,	PUNCT
fcis-30157	41	6	achieved	achieve	VERB
fcis-30157	41	7	82	82	NUM
fcis-30157	41	8	%	%	NOUN
fcis-30157	41	9	accuracy	accuracy	NOUN
fcis-30157	41	10	in	in	ADP
fcis-30157	41	11	distinguishing	distinguish	VERB
fcis-30157	41	12	melanoma	melanoma	NOUN
fcis-30157	41	13	from	from	ADP
fcis-30157	41	14	non	non	ADJ
fcis-30157	41	15	-	-	ADJ
fcis-30157	41	16	melanoma	melanoma	ADJ
fcis-30157	41	17	lesions	lesion	NOUN
fcis-30157	41	18	[	[	X
fcis-30157	41	19	13	13	NUM
fcis-30157	41	20	]	]	PUNCT
fcis-30157	41	21	.	.	PUNCT
fcis-30157	42	1	despite	despite	SCONJ
fcis-30157	42	2	their	their	PRON
fcis-30157	42	3	effectiveness	effectiveness	NOUN
fcis-30157	42	4	,	,	PUNCT
fcis-30157	42	5	manual	manual	ADJ
fcis-30157	42	6	methods	method	NOUN
fcis-30157	42	7	are	be	AUX
fcis-30157	42	8	limited	limit	VERB
fcis-30157	42	9	by	by	ADP
fcis-30157	42	10	the	the	DET
fcis-30157	42	11	need	need	NOUN
fcis-30157	42	12	for	for	ADP
fcis-30157	42	13	domain	domain	NOUN
fcis-30157	42	14	expertise	expertise	NOUN
fcis-30157	42	15	and	and	CCONJ
fcis-30157	42	16	an	an	DET
fcis-30157	42	17	inability	inability	NOUN
fcis-30157	42	18	to	to	PART
fcis-30157	42	19	capture	capture	VERB
fcis-30157	42	20	complex	complex	ADJ
fcis-30157	42	21	features	feature	NOUN
fcis-30157	42	22	.	.	PUNCT
fcis-30157	43	1	automatic	automatic	ADJ
fcis-30157	43	2	feature	feature	NOUN
fcis-30157	43	3	extraction	extraction	NOUN
fcis-30157	43	4	in	in	ADP
fcis-30157	43	5	deep	deep	ADJ
fcis-30157	43	6	learning	learning	NOUN
fcis-30157	43	7	deep	deep	ADJ
fcis-30157	43	8	learning	learning	NOUN
fcis-30157	43	9	,	,	PUNCT
fcis-30157	43	10	particularly	particularly	ADV
fcis-30157	43	11	convolutional	convolutional	ADJ
fcis-30157	43	12	neural	neural	ADJ
fcis-30157	43	13	networks	network	NOUN
fcis-30157	43	14	(	(	PUNCT
fcis-30157	43	15	cnns	cnns	PROPN
fcis-30157	43	16	)	)	PUNCT
fcis-30157	43	17	,	,	PUNCT
fcis-30157	43	18	has	have	AUX
fcis-30157	43	19	revolutionized	revolutionize	VERB
fcis-30157	43	20	feature	feature	NOUN
fcis-30157	43	21	extraction	extraction	NOUN
fcis-30157	43	22	by	by	ADP
fcis-30157	43	23	learning	learn	VERB
fcis-30157	43	24	hierarchical	hierarchical	ADJ
fcis-30157	43	25	features	feature	NOUN
fcis-30157	43	26	from	from	ADP
fcis-30157	43	27	raw	raw	ADJ
fcis-30157	43	28	data	datum	NOUN
fcis-30157	43	29	.	.	PUNCT
fcis-30157	44	1	cnns	cnns	PROPN
fcis-30157	44	2	,	,	PUNCT
fcis-30157	44	3	such	such	ADJ
fcis-30157	44	4	as	as	ADP
fcis-30157	44	5	vggnet	vggnet	NOUN
fcis-30157	44	6	and	and	CCONJ
fcis-30157	44	7	resnet	resnet	NOUN
fcis-30157	44	8	,	,	PUNCT
fcis-30157	44	9	have	have	AUX
fcis-30157	44	10	outperformed	outperform	VERB
fcis-30157	44	11	traditional	traditional	ADJ
fcis-30157	44	12	methods	method	NOUN
fcis-30157	44	13	.	.	PUNCT
fcis-30157	45	1	a	a	DET
fcis-30157	45	2	study	study	NOUN
fcis-30157	45	3	[	[	X
fcis-30157	45	4	3	3	X
fcis-30157	45	5	]	]	PUNCT
fcis-30157	45	6	found	find	VERB
fcis-30157	45	7	that	that	SCONJ
fcis-30157	45	8	resnet-50	resnet-50	PROPN
fcis-30157	45	9	achieved	achieve	VERB
fcis-30157	45	10	92.3	92.3	NUM
fcis-30157	45	11	%	%	NOUN
fcis-30157	45	12	accuracy	accuracy	NOUN
fcis-30157	45	13	in	in	ADP
fcis-30157	45	14	classifying	classify	VERB
fcis-30157	45	15	skin	skin	NOUN
fcis-30157	45	16	diseases	disease	NOUN
fcis-30157	45	17	.	.	PUNCT
fcis-30157	46	1	deep	deep	ADJ
fcis-30157	46	2	learning	learning	NOUN
fcis-30157	46	3	's	's	PART
fcis-30157	46	4	ability	ability	NOUN
fcis-30157	46	5	to	to	PART
fcis-30157	46	6	learn	learn	VERB
fcis-30157	46	7	intricate	intricate	ADJ
fcis-30157	46	8	patterns	pattern	NOUN
fcis-30157	46	9	like	like	ADP
fcis-30157	46	10	lesion	lesion	NOUN
fcis-30157	46	11	boundaries	boundary	NOUN
fcis-30157	46	12	and	and	CCONJ
fcis-30157	46	13	textures	texture	NOUN
fcis-30157	46	14	from	from	ADP
fcis-30157	46	15	large	large	ADJ
fcis-30157	46	16	datasets	dataset	NOUN
fcis-30157	46	17	contributes	contribute	VERB
fcis-30157	46	18	to	to	ADP
fcis-30157	46	19	this	this	DET
fcis-30157	46	20	superior	superior	ADJ
fcis-30157	46	21	performance	performance	NOUN
fcis-30157	46	22	.	.	PUNCT
fcis-30157	47	1	one	one	NUM
fcis-30157	47	2	of	of	ADP
fcis-30157	47	3	its	its	PRON
fcis-30157	47	4	advantages	advantage	NOUN
fcis-30157	47	5	is	be	AUX
fcis-30157	47	6	handling	handle	VERB
fcis-30157	47	7	large	large	ADJ
fcis-30157	47	8	-	-	PUNCT
fcis-30157	47	9	scale	scale	NOUN
fcis-30157	47	10	datasets	dataset	NOUN
fcis-30157	47	11	,	,	PUNCT
fcis-30157	47	12	such	such	ADJ
fcis-30157	47	13	as	as	ADP
fcis-30157	47	14	the	the	DET
fcis-30157	47	15	isic	isic	PROPN
fcis-30157	47	16	dataset	dataset	VERB
fcis-30157	47	17	with	with	ADP
fcis-30157	47	18	over	over	ADP
fcis-30157	47	19	20,000	20,000	NUM
fcis-30157	47	20	dermoscopy	dermoscopy	NOUN
fcis-30157	47	21	images	image	NOUN
fcis-30157	47	22	.	.	PUNCT
fcis-30157	48	1	this	this	DET
fcis-30157	48	2	exposure	exposure	NOUN
fcis-30157	48	3	enhances	enhance	VERB
fcis-30157	48	4	the	the	DET
fcis-30157	48	5	model	model	NOUN
fcis-30157	48	6	's	's	PART
fcis-30157	48	7	generalization	generalization	NOUN
fcis-30157	48	8	capability	capability	NOUN
fcis-30157	48	9	.	.	PUNCT
fcis-30157	49	1	comparison	comparison	NOUN
fcis-30157	49	2	of	of	ADP
fcis-30157	49	3	manual	manual	ADJ
fcis-30157	49	4	and	and	CCONJ
fcis-30157	49	5	automatic	automatic	ADJ
fcis-30157	49	6	methods	method	NOUN
fcis-30157	49	7	manual	manual	ADJ
fcis-30157	49	8	methods	method	NOUN
fcis-30157	49	9	offer	offer	VERB
fcis-30157	49	10	interpretability	interpretability	NOUN
fcis-30157	49	11	and	and	CCONJ
fcis-30157	49	12	low	low	ADJ
fcis-30157	49	13	computational	computational	ADJ
fcis-30157	49	14	cost	cost	NOUN
fcis-30157	49	15	but	but	CCONJ
fcis-30157	49	16	are	be	AUX
fcis-30157	49	17	limited	limit	VERB
fcis-30157	49	18	by	by	ADP
fcis-30157	49	19	the	the	DET
fcis-30157	49	20	need	need	NOUN
fcis-30157	49	21	for	for	ADP
fcis-30157	49	22	expert	expert	NOUN
fcis-30157	49	23	input	input	NOUN
fcis-30157	49	24	and	and	CCONJ
fcis-30157	49	25	sensitivity	sensitivity	NOUN
fcis-30157	49	26	to	to	PART
fcis-30157	49	27	noise	noise	VERB
fcis-30157	49	28	.	.	PUNCT
fcis-30157	50	1	in	in	ADP
fcis-30157	50	2	contrast	contrast	NOUN
fcis-30157	50	3	,	,	PUNCT
fcis-30157	50	4	deep	deep	ADJ
fcis-30157	50	5	learning	learning	NOUN
fcis-30157	50	6	methods	method	NOUN
fcis-30157	50	7	provide	provide	VERB
fcis-30157	50	8	high	high	ADJ
fcis-30157	50	9	accuracy	accuracy	NOUN
fcis-30157	50	10	and	and	CCONJ
fcis-30157	50	11	automatic	automatic	ADJ
fcis-30157	50	12	learning	learning	NOUN
fcis-30157	50	13	but	but	CCONJ
fcis-30157	50	14	require	require	VERB
fcis-30157	50	15	substantial	substantial	ADJ
fcis-30157	50	16	computational	computational	ADJ
fcis-30157	50	17	resources	resource	NOUN
fcis-30157	50	18	and	and	CCONJ
fcis-30157	50	19	large	large	ADJ
fcis-30157	50	20	datasets	dataset	NOUN
fcis-30157	50	21	.	.	PUNCT
fcis-30157	51	1	challenges	challenge	NOUN
fcis-30157	51	2	and	and	CCONJ
fcis-30157	51	3	future	future	ADJ
fcis-30157	51	4	directions	direction	NOUN
fcis-30157	51	5	.	.	PUNCT
fcis-30157	52	1	despite	despite	SCONJ
fcis-30157	52	2	their	their	PRON
fcis-30157	52	3	advantages	advantage	NOUN
fcis-30157	52	4	,	,	PUNCT
fcis-30157	52	5	deep	deep	ADJ
fcis-30157	52	6	learning	learning	NOUN
fcis-30157	52	7	methods	method	NOUN
fcis-30157	52	8	face	face	VERB
fcis-30157	52	9	challenges	challenge	NOUN
fcis-30157	52	10	like	like	ADP
fcis-30157	52	11	the	the	DET
fcis-30157	52	12	need	need	NOUN
fcis-30157	52	13	for	for	ADP
fcis-30157	52	14	large	large	ADJ
fcis-30157	52	15	annotated	annotated	ADJ
fcis-30157	52	16	datasets	dataset	NOUN
fcis-30157	52	17	,	,	PUNCT
fcis-30157	52	18	which	which	PRON
fcis-30157	52	19	are	be	AUX
fcis-30157	52	20	time	time	NOUN
fcis-30157	52	21	-	-	PUNCT
fcis-30157	52	22	consuming	consume	VERB
fcis-30157	52	23	to	to	PART
fcis-30157	52	24	obtain	obtain	VERB
fcis-30157	52	25	.	.	PUNCT
fcis-30157	53	1	solutions	solution	NOUN
fcis-30157	53	2	like	like	ADP
fcis-30157	53	3	transfer	transfer	NOUN
fcis-30157	53	4	learning	learning	NOUN
fcis-30157	53	5	and	and	CCONJ
fcis-30157	53	6	data	datum	NOUN
fcis-30157	53	7	augmentation	augmentation	NOUN
fcis-30157	53	8	have	have	AUX
fcis-30157	53	9	been	be	AUX
fcis-30157	53	10	proposed	propose	VERB
fcis-30157	53	11	.	.	PUNCT
fcis-30157	54	1	a	a	DET
fcis-30157	54	2	study	study	NOUN
fcis-30157	54	3	[	[	X
fcis-30157	54	4	14	14	NUM
fcis-30157	54	5	]	]	PUNCT
fcis-30157	54	6	showed	show	VERB
fcis-30157	54	7	that	that	SCONJ
fcis-30157	54	8	fine	fine	NOUN
fcis-30157	54	9	-	-	PUNCT
fcis-30157	54	10	tuning	tune	VERB
fcis-30157	54	11	pre	pre	ADJ
fcis-30157	54	12	-	-	ADJ
fcis-30157	54	13	trained	train	VERB
fcis-30157	54	14	models	model	NOUN
fcis-30157	54	15	on	on	ADP
fcis-30157	54	16	smaller	small	ADJ
fcis-30157	54	17	datasets	dataset	NOUN
fcis-30157	54	18	can	can	AUX
fcis-30157	54	19	yield	yield	VERB
fcis-30157	54	20	comparable	comparable	ADJ
fcis-30157	54	21	results	result	NOUN
fcis-30157	54	22	to	to	ADP
fcis-30157	54	23	training	training	NOUN
fcis-30157	54	24	models	model	NOUN
fcis-30157	54	25	from	from	ADP
fcis-30157	54	26	scratch	scratch	NOUN
fcis-30157	54	27	.	.	PUNCT
fcis-30157	55	1	another	another	DET
fcis-30157	55	2	challenge	challenge	NOUN
fcis-30157	55	3	is	be	AUX
fcis-30157	55	4	interpretability	interpretability	NOUN
fcis-30157	55	5	;	;	PUNCT
fcis-30157	55	6	while	while	SCONJ
fcis-30157	55	7	deep	deep	ADJ
fcis-30157	55	8	learning	learning	NOUN
fcis-30157	55	9	models	model	NOUN
fcis-30157	55	10	are	be	AUX
fcis-30157	55	11	accurate	accurate	ADJ
fcis-30157	55	12	,	,	PUNCT
fcis-30157	55	13	their	their	PRON
fcis-30157	55	14	decision	decision	NOUN
fcis-30157	55	15	-	-	PUNCT
fcis-30157	55	16	making	make	VERB
fcis-30157	55	17	process	process	NOUN
fcis-30157	55	18	is	be	AUX
fcis-30157	55	19	often	often	ADV
fcis-30157	55	20	opaque	opaque	ADJ
fcis-30157	55	21	.	.	PUNCT
fcis-30157	56	1	recent	recent	ADJ
fcis-30157	56	2	developments	development	NOUN
fcis-30157	56	3	in	in	ADP
fcis-30157	56	4	explainable	explainable	ADJ
fcis-30157	56	5	ai	ai	NOUN
fcis-30157	56	6	(	(	PUNCT
fcis-30157	56	7	xai	xai	PROPN
fcis-30157	56	8	)	)	PUNCT
fcis-30157	56	9	,	,	PUNCT
fcis-30157	56	10	such	such	ADJ
fcis-30157	56	11	as	as	ADP
fcis-30157	56	12	grad	grad	NOUN
fcis-30157	56	13	-	-	PUNCT
fcis-30157	56	14	cam	cam	NOUN
fcis-30157	56	15	,	,	PUNCT
fcis-30157	56	16	have	have	AUX
fcis-30157	56	17	been	be	AUX
fcis-30157	56	18	used	use	VERB
fcis-30157	56	19	to	to	PART
fcis-30157	56	20	visualize	visualize	VERB
fcis-30157	56	21	the	the	DET
fcis-30157	56	22	areas	area	NOUN
fcis-30157	56	23	of	of	ADP
fcis-30157	56	24	an	an	DET
fcis-30157	56	25	image	image	NOUN
fcis-30157	56	26	most	most	ADV
fcis-30157	56	27	influencing	influence	VERB
fcis-30157	56	28	the	the	DET
fcis-30157	56	29	model	model	NOUN
fcis-30157	56	30	's	's	PART
fcis-30157	56	31	decision	decision	NOUN
fcis-30157	56	32	,	,	PUNCT
fcis-30157	56	33	improving	improve	VERB
fcis-30157	56	34	trust	trust	NOUN
fcis-30157	56	35	and	and	CCONJ
fcis-30157	56	36	interpretability	interpretability	NOUN
fcis-30157	57	1	[	[	X
fcis-30157	57	2	5	5	NUM
fcis-30157	57	3	]	]	PUNCT
fcis-30157	57	4	.	.	PUNCT
fcis-30157	58	1	1.1.3	1.1.3	X
fcis-30157	58	2	.	.	PUNCT
fcis-30157	58	3	machine	machine	NOUN
fcis-30157	58	4	learning	learning	NOUN
fcis-30157	58	5	models	model	NOUN
fcis-30157	58	6	for	for	ADP
fcis-30157	58	7	recognition	recognition	NOUN
fcis-30157	58	8	machine	machine	NOUN
fcis-30157	58	9	learning	learning	NOUN
fcis-30157	58	10	models	model	NOUN
fcis-30157	58	11	have	have	AUX
fcis-30157	58	12	been	be	AUX
fcis-30157	58	13	key	key	ADJ
fcis-30157	58	14	in	in	ADP
fcis-30157	58	15	the	the	DET
fcis-30157	58	16	automatic	automatic	ADJ
fcis-30157	58	17	recognition	recognition	NOUN
fcis-30157	58	18	of	of	ADP
fcis-30157	58	19	skin	skin	NOUN
fcis-30157	58	20	diseases	disease	NOUN
fcis-30157	58	21	,	,	PUNCT
fcis-30157	58	22	particularly	particularly	ADV
fcis-30157	58	23	with	with	ADP
fcis-30157	58	24	large	large	ADJ
fcis-30157	58	25	-	-	PUNCT
fcis-30157	58	26	scale	scale	NOUN
fcis-30157	58	27	image	image	NOUN
fcis-30157	58	28	datasets	dataset	NOUN
fcis-30157	58	29	.	.	PUNCT
fcis-30157	59	1	traditional	traditional	ADJ
fcis-30157	59	2	models	model	NOUN
fcis-30157	59	3	,	,	PUNCT
fcis-30157	59	4	like	like	ADP
fcis-30157	59	5	support	support	NOUN
fcis-30157	59	6	vector	vector	NOUN
fcis-30157	59	7	machines	machine	NOUN
fcis-30157	59	8	(	(	PUNCT
fcis-30157	59	9	svms	svms	NOUN
fcis-30157	59	10	)	)	PUNCT
fcis-30157	59	11	and	and	CCONJ
fcis-30157	59	12	decision	decision	NOUN
fcis-30157	59	13	trees	tree	NOUN
fcis-30157	59	14	,	,	PUNCT
fcis-30157	59	15	have	have	AUX
fcis-30157	59	16	been	be	AUX
fcis-30157	59	17	widely	widely	ADV
fcis-30157	59	18	used	use	VERB
fcis-30157	59	19	due	due	ADP
fcis-30157	59	20	to	to	ADP
fcis-30157	59	21	their	their	PRON
fcis-30157	59	22	simplicity	simplicity	NOUN
fcis-30157	59	23	and	and	CCONJ
fcis-30157	59	24	interpretability	interpretability	NOUN
fcis-30157	59	25	.	.	PUNCT
fcis-30157	60	1	svms	svms	NOUN
fcis-30157	60	2	,	,	PUNCT
fcis-30157	60	3	effective	effective	ADJ
fcis-30157	60	4	in	in	ADP
fcis-30157	60	5	high	high	ADJ
fcis-30157	60	6	-	-	PUNCT
fcis-30157	60	7	dimensional	dimensional	ADJ
fcis-30157	60	8	data	datum	NOUN
fcis-30157	60	9	,	,	PUNCT
fcis-30157	60	10	achieved	achieve	VERB
fcis-30157	60	11	82	82	NUM
fcis-30157	60	12	%	%	NOUN
fcis-30157	60	13	accuracy	accuracy	NOUN
fcis-30157	60	14	in	in	ADP
fcis-30157	60	15	melanoma	melanoma	NOUN
fcis-30157	60	16	and	and	CCONJ
fcis-30157	60	17	benign	benign	ADJ
fcis-30157	60	18	nevi	nevi	ADJ
fcis-30157	60	19	classification	classification	NOUN
fcis-30157	60	20	[	[	X
fcis-30157	60	21	15	15	NUM
fcis-30157	60	22	]	]	PUNCT
fcis-30157	60	23	,	,	PUNCT
fcis-30157	60	24	though	though	SCONJ
fcis-30157	60	25	they	they	PRON
fcis-30157	60	26	rely	rely	VERB
fcis-30157	60	27	on	on	ADP
fcis-30157	60	28	manual	manual	ADJ
fcis-30157	60	29	feature	feature	NOUN
fcis-30157	60	30	extraction	extraction	NOUN
fcis-30157	60	31	,	,	PUNCT
fcis-30157	60	32	which	which	PRON
fcis-30157	60	33	can	can	AUX
fcis-30157	60	34	limit	limit	VERB
fcis-30157	60	35	performance	performance	NOUN
fcis-30157	60	36	.	.	PUNCT
fcis-30157	61	1	decision	decision	NOUN
fcis-30157	61	2	trees	tree	NOUN
fcis-30157	61	3	,	,	PUNCT
fcis-30157	61	4	known	know	VERB
fcis-30157	61	5	for	for	ADP
fcis-30157	61	6	their	their	PRON
fcis-30157	61	7	transparency	transparency	NOUN
fcis-30157	61	8	,	,	PUNCT
fcis-30157	61	9	achieved	achieve	VERB
fcis-30157	61	10	78	78	NUM
fcis-30157	61	11	%	%	NOUN
fcis-30157	61	12	accuracy	accuracy	NOUN
fcis-30157	61	13	in	in	ADP
fcis-30157	61	14	diagnosing	diagnose	VERB
fcis-30157	61	15	common	common	ADJ
fcis-30157	61	16	skin	skin	NOUN
fcis-30157	61	17	conditions	condition	NOUN
fcis-30157	61	18	like	like	ADP
fcis-30157	61	19	eczema	eczema	NOUN
fcis-30157	61	20	and	and	CCONJ
fcis-30157	61	21	psoriasis	psoriasis	NOUN
fcis-30157	62	1	[	[	X
fcis-30157	62	2	1	1	NUM
fcis-30157	62	3	]	]	PUNCT
fcis-30157	62	4	,	,	PUNCT
fcis-30157	62	5	but	but	CCONJ
fcis-30157	62	6	also	also	ADV
fcis-30157	62	7	face	face	VERB
fcis-30157	62	8	challenges	challenge	NOUN
fcis-30157	62	9	in	in	ADP
fcis-30157	62	10	handling	handle	VERB
fcis-30157	62	11	complex	complex	ADJ
fcis-30157	62	12	datasets	dataset	NOUN
fcis-30157	62	13	.	.	PUNCT
fcis-30157	63	1	in	in	ADP
fcis-30157	63	2	contrast	contrast	NOUN
fcis-30157	63	3	,	,	PUNCT
fcis-30157	63	4	deep	deep	ADJ
fcis-30157	63	5	learning	learning	NOUN
fcis-30157	63	6	models	model	NOUN
fcis-30157	63	7	like	like	ADP
fcis-30157	63	8	resnet	resnet	NOUN
fcis-30157	63	9	and	and	CCONJ
fcis-30157	63	10	vgg	vgg	NOUN
fcis-30157	63	11	have	have	AUX
fcis-30157	63	12	revolutionized	revolutionize	VERB
fcis-30157	63	13	skin	skin	NOUN
fcis-30157	63	14	disease	disease	NOUN
fcis-30157	63	15	recognition	recognition	NOUN
fcis-30157	63	16	by	by	ADP
fcis-30157	63	17	automating	automate	VERB
fcis-30157	63	18	feature	feature	NOUN
fcis-30157	63	19	extraction	extraction	NOUN
fcis-30157	63	20	.	.	PUNCT
fcis-30157	64	1	resnet	resnet	NOUN
fcis-30157	64	2	,	,	PUNCT
fcis-30157	64	3	with	with	ADP
fcis-30157	64	4	its	its	PRON
fcis-30157	64	5	residual	residual	ADJ
fcis-30157	64	6	connections	connection	NOUN
fcis-30157	64	7	,	,	PUNCT
fcis-30157	64	8	mitigates	mitigate	VERB
fcis-30157	64	9	the	the	DET
fcis-30157	64	10	vanishing	vanish	VERB
fcis-30157	64	11	gradient	gradient	NOUN
fcis-30157	64	12	problem	problem	NOUN
fcis-30157	64	13	,	,	PUNCT
fcis-30157	64	14	achieving	achieve	VERB
fcis-30157	64	15	92	92	NUM
fcis-30157	64	16	%	%	NOUN
fcis-30157	64	17	accuracy	accuracy	NOUN
fcis-30157	64	18	in	in	ADP
fcis-30157	64	19	skin	skin	NOUN
fcis-30157	64	20	disease	disease	NOUN
fcis-30157	64	21	classification	classification	NOUN
fcis-30157	64	22	.	.	PUNCT
fcis-30157	65	1	vgg	vgg	ADJ
fcis-30157	65	2	,	,	PUNCT
fcis-30157	65	3	known	know	VERB
fcis-30157	65	4	for	for	ADP
fcis-30157	65	5	its	its	PRON
fcis-30157	65	6	simplicity	simplicity	NOUN
fcis-30157	65	7	,	,	PUNCT
fcis-30157	65	8	reached	reach	VERB
fcis-30157	65	9	90	90	NUM
fcis-30157	65	10	%	%	NOUN
fcis-30157	65	11	accuracy	accuracy	NOUN
fcis-30157	65	12	in	in	ADP
fcis-30157	65	13	a	a	DET
fcis-30157	65	14	study	study	NOUN
fcis-30157	65	15	[	[	X
fcis-30157	65	16	16	16	NUM
fcis-30157	65	17	]	]	PUNCT
fcis-30157	65	18	.	.	PUNCT
fcis-30157	66	1	these	these	DET
fcis-30157	66	2	models	model	NOUN
fcis-30157	66	3	are	be	AUX
fcis-30157	66	4	trained	train	VERB
fcis-30157	66	5	on	on	ADP
fcis-30157	66	6	large	large	ADJ
fcis-30157	66	7	,	,	PUNCT
fcis-30157	66	8	labeled	label	VERB
fcis-30157	66	9	datasets	dataset	NOUN
fcis-30157	66	10	,	,	PUNCT
fcis-30157	66	11	enabling	enable	VERB
fcis-30157	66	12	them	they	PRON
fcis-30157	66	13	to	to	PART
fcis-30157	66	14	generalize	generalize	VERB
fcis-30157	66	15	well	well	ADV
fcis-30157	66	16	.	.	PUNCT
fcis-30157	67	1	for	for	ADP
fcis-30157	67	2	example	example	NOUN
fcis-30157	67	3	,	,	PUNCT
fcis-30157	67	4	a	a	DET
fcis-30157	67	5	hybrid	hybrid	ADJ
fcis-30157	67	6	deep	deep	ADJ
fcis-30157	67	7	learning	learning	NOUN
fcis-30157	67	8	model	model	NOUN
fcis-30157	67	9	trained	train	VERB
fcis-30157	67	10	on	on	ADP
fcis-30157	67	11	10,000	10,000	NUM
fcis-30157	67	12	images	image	NOUN
fcis-30157	67	13	achieved	achieve	VERB
fcis-30157	67	14	94	94	NUM
fcis-30157	67	15	%	%	NOUN
fcis-30157	67	16	accuracy	accuracy	NOUN
fcis-30157	68	1	[	[	X
fcis-30157	68	2	13	13	NUM
fcis-30157	68	3	]	]	PUNCT
fcis-30157	68	4	.	.	PUNCT
fcis-30157	69	1	however	however	ADV
fcis-30157	69	2	,	,	PUNCT
fcis-30157	69	3	deep	deep	ADJ
fcis-30157	69	4	learning	learning	NOUN
fcis-30157	69	5	models	model	NOUN
fcis-30157	69	6	are	be	AUX
fcis-30157	69	7	resource	resource	NOUN
fcis-30157	69	8	-	-	PUNCT
fcis-30157	69	9	intensive	intensive	ADJ
fcis-30157	69	10	and	and	CCONJ
fcis-30157	69	11	risk	risk	NOUN
fcis-30157	69	12	overfitting	overfitting	NOUN
fcis-30157	69	13	,	,	PUNCT
fcis-30157	69	14	especially	especially	ADV
fcis-30157	69	15	with	with	ADP
fcis-30157	69	16	small	small	ADJ
fcis-30157	69	17	datasets	dataset	NOUN
fcis-30157	69	18	.	.	PUNCT
fcis-30157	70	1	to	to	PART
fcis-30157	70	2	improve	improve	VERB
fcis-30157	70	3	generalization	generalization	NOUN
fcis-30157	70	4	,	,	PUNCT
fcis-30157	70	5	techniques	technique	NOUN
fcis-30157	70	6	like	like	ADP
fcis-30157	70	7	data	datum	NOUN
fcis-30157	70	8	augmentation	augmentation	NOUN
fcis-30157	70	9	(	(	PUNCT
fcis-30157	70	10	e.g.	e.g.	ADV
fcis-30157	70	11	,	,	PUNCT
fcis-30157	70	12	random	random	ADJ
fcis-30157	70	13	cropping	cropping	NOUN
fcis-30157	70	14	and	and	CCONJ
fcis-30157	70	15	flipping	flipping	NOUN
fcis-30157	70	16	)	)	PUNCT
fcis-30157	70	17	have	have	AUX
fcis-30157	70	18	been	be	AUX
fcis-30157	70	19	used	use	VERB
fcis-30157	70	20	.	.	PUNCT
fcis-30157	71	1	a	a	DET
fcis-30157	71	2	study	study	NOUN
fcis-30157	71	3	[	[	X
fcis-30157	71	4	17	17	NUM
fcis-30157	71	5	]	]	PUNCT
fcis-30157	71	6	found	find	VERB
fcis-30157	71	7	that	that	SCONJ
fcis-30157	71	8	augmentation	augmentation	NOUN
fcis-30157	71	9	improved	improve	VERB
fcis-30157	71	10	accuracy	accuracy	NOUN
fcis-30157	71	11	by	by	ADP
fcis-30157	71	12	5	5	NUM
fcis-30157	71	13	%	%	NOUN
fcis-30157	71	14	.	.	PUNCT
fcis-30157	72	1	transfer	transfer	NOUN
fcis-30157	72	2	learning	learning	NOUN
fcis-30157	72	3	,	,	PUNCT
fcis-30157	72	4	where	where	SCONJ
fcis-30157	72	5	a	a	DET
fcis-30157	72	6	pretrained	pretraine	VERB
fcis-30157	72	7	model	model	NOUN
fcis-30157	72	8	is	be	AUX
fcis-30157	72	9	fine	fine	ADV
fcis-30157	72	10	-	-	PUNCT
fcis-30157	72	11	tuned	tune	VERB
fcis-30157	72	12	on	on	ADP
fcis-30157	72	13	a	a	DET
fcis-30157	72	14	specific	specific	ADJ
fcis-30157	72	15	dataset	dataset	NOUN
fcis-30157	72	16	,	,	PUNCT
fcis-30157	72	17	has	have	AUX
fcis-30157	72	18	also	also	ADV
fcis-30157	72	19	proven	prove	VERB
fcis-30157	72	20	effective	effective	ADJ
fcis-30157	72	21	.	.	PUNCT
fcis-30157	73	1	for	for	ADP
fcis-30157	73	2	instance	instance	NOUN
fcis-30157	73	3	,	,	PUNCT
fcis-30157	73	4	fine	fine	ADV
fcis-30157	73	5	-	-	PUNCT
fcis-30157	73	6	tuning	tune	VERB
fcis-30157	73	7	a	a	DET
fcis-30157	73	8	pre	pre	ADJ
fcis-30157	73	9	-	-	ADJ
fcis-30157	73	10	trained	train	VERB
fcis-30157	73	11	resnet	resnet	NOUN
fcis-30157	73	12	on	on	ADP
fcis-30157	73	13	a	a	DET
fcis-30157	73	14	skin	skin	NOUN
fcis-30157	73	15	cancer	cancer	NOUN
fcis-30157	73	16	dataset	dataset	NOUN
fcis-30157	73	17	achieved	achieve	VERB
fcis-30157	73	18	95	95	NUM
fcis-30157	73	19	%	%	NOUN
fcis-30157	73	20	accuracy	accuracy	NOUN
fcis-30157	73	21	[	[	X
fcis-30157	73	22	18	18	NUM
fcis-30157	73	23	]	]	PUNCT
fcis-30157	73	24	.	.	PUNCT
fcis-30157	74	1	these	these	DET
fcis-30157	74	2	approaches	approach	NOUN
fcis-30157	74	3	help	help	VERB
fcis-30157	74	4	bridge	bridge	VERB
fcis-30157	74	5	the	the	DET
fcis-30157	74	6	gap	gap	NOUN
fcis-30157	74	7	between	between	ADP
fcis-30157	74	8	model	model	NOUN
fcis-30157	74	9	complexity	complexity	NOUN
fcis-30157	74	10	and	and	CCONJ
fcis-30157	74	11	data	datum	NOUN
fcis-30157	74	12	availability	availability	NOUN
fcis-30157	74	13	.	.	PUNCT
fcis-30157	75	1	deep	deep	ADJ
fcis-30157	75	2	learning	learning	NOUN
fcis-30157	75	3	models	model	NOUN
fcis-30157	75	4	are	be	AUX
fcis-30157	75	5	evaluated	evaluate	VERB
fcis-30157	75	6	using	use	VERB
fcis-30157	75	7	metrics	metric	NOUN
fcis-30157	75	8	such	such	ADJ
fcis-30157	75	9	as	as	ADP
fcis-30157	75	10	accuracy	accuracy	NOUN
fcis-30157	75	11	,	,	PUNCT
fcis-30157	75	12	precision	precision	NOUN
fcis-30157	75	13	,	,	PUNCT
fcis-30157	75	14	recall	recall	NOUN
fcis-30157	75	15	,	,	PUNCT
fcis-30157	75	16	and	and	CCONJ
fcis-30157	75	17	f1	f1	PROPN
fcis-30157	75	18	score	score	NOUN
fcis-30157	75	19	.	.	PUNCT
fcis-30157	76	1	a	a	DET
fcis-30157	76	2	study	study	NOUN
fcis-30157	76	3	on	on	ADP
fcis-30157	76	4	ai	ai	NOUN
fcis-30157	76	5	for	for	ADP
fcis-30157	76	6	skin	skin	NOUN
fcis-30157	76	7	cancer	cancer	NOUN
fcis-30157	76	8	diagnosis	diagnosis	NOUN
fcis-30157	76	9	reported	report	VERB
fcis-30157	76	10	91	91	NUM
fcis-30157	76	11	%	%	NOUN
fcis-30157	76	12	accuracy	accuracy	NOUN
fcis-30157	76	13	,	,	PUNCT
fcis-30157	76	14	with	with	ADP
fcis-30157	76	15	precision	precision	NOUN
fcis-30157	76	16	and	and	CCONJ
fcis-30157	76	17	recall	recall	NOUN
fcis-30157	76	18	of	of	ADP
fcis-30157	76	19	89	89	NUM
fcis-30157	76	20	%	%	NOUN
fcis-30157	76	21	and	and	CCONJ
fcis-30157	76	22	93	93	NUM
fcis-30157	76	23	%	%	NOUN
fcis-30157	76	24	,	,	PUNCT
fcis-30157	76	25	respectively	respectively	ADV
fcis-30157	76	26	.	.	PUNCT
fcis-30157	77	1	despite	despite	SCONJ
fcis-30157	77	2	their	their	PRON
fcis-30157	77	3	effectiveness	effectiveness	NOUN
fcis-30157	77	4	,	,	PUNCT
fcis-30157	77	5	deep	deep	ADJ
fcis-30157	77	6	learning	learning	NOUN
fcis-30157	77	7	models	model	NOUN
fcis-30157	77	8	like	like	ADP
fcis-30157	77	9	resnet	resnet	NOUN
fcis-30157	77	10	and	and	CCONJ
fcis-30157	77	11	vgg	vgg	ADJ
fcis-30157	77	12	require	require	VERB
fcis-30157	77	13	significant	significant	ADJ
fcis-30157	77	14	computational	computational	ADJ
fcis-30157	77	15	power	power	NOUN
fcis-30157	77	16	.	.	PUNCT
fcis-30157	78	1	for	for	ADP
fcis-30157	78	2	example	example	NOUN
fcis-30157	78	3	,	,	PUNCT
fcis-30157	78	4	training	training	NOUN
fcis-30157	78	5	resnet	resnet	NOUN
fcis-30157	78	6	on	on	ADP
fcis-30157	78	7	a	a	DET
fcis-30157	78	8	high	high	ADJ
fcis-30157	78	9	-	-	PUNCT
fcis-30157	78	10	performance	performance	NOUN
fcis-30157	78	11	gpu	gpu	NOUN
fcis-30157	78	12	took	take	VERB
fcis-30157	78	13	about	about	ADV
fcis-30157	78	14	12	12	NUM
fcis-30157	78	15	hours	hour	NOUN
fcis-30157	78	16	.	.	PUNCT
fcis-30157	79	1	machine	machine	NOUN
fcis-30157	79	2	learning	learning	NOUN
fcis-30157	79	3	models	model	NOUN
fcis-30157	79	4	,	,	PUNCT
fcis-30157	79	5	particularly	particularly	ADV
fcis-30157	79	6	deep	deep	ADJ
fcis-30157	79	7	learning	learning	NOUN
fcis-30157	79	8	,	,	PUNCT
fcis-30157	79	9	offer	offer	VERB
fcis-30157	79	10	high	high	ADJ
fcis-30157	79	11	diagnostic	diagnostic	ADJ
fcis-30157	79	12	accuracy	accuracy	NOUN
fcis-30157	79	13	and	and	CCONJ
fcis-30157	79	14	efficiency	efficiency	NOUN
fcis-30157	79	15	,	,	PUNCT
fcis-30157	79	16	with	with	ADP
fcis-30157	79	17	ai	ai	ADJ
fcis-30157	79	18	-	-	PUNCT
fcis-30157	79	19	assisted	assist	VERB
fcis-30157	79	20	systems	system	NOUN
fcis-30157	79	21	shown	show	VERB
fcis-30157	79	22	to	to	PART
fcis-30157	79	23	reduce	reduce	VERB
fcis-30157	79	24	misdiagnosis	misdiagnosis	NOUN
fcis-30157	79	25	rates	rate	NOUN
fcis-30157	79	26	by	by	ADP
fcis-30157	79	27	30	30	NUM
fcis-30157	79	28	%	%	NOUN
fcis-30157	79	29	in	in	ADP
fcis-30157	79	30	primary	primary	ADJ
fcis-30157	79	31	care	care	NOUN
fcis-30157	79	32	[	[	X
fcis-30157	79	33	19	19	NUM
fcis-30157	79	34	]	]	PUNCT
fcis-30157	79	35	.	.	PUNCT
fcis-30157	80	1	however	however	ADV
fcis-30157	80	2	,	,	PUNCT
fcis-30157	80	3	the	the	DET
fcis-30157	80	4	adoption	adoption	NOUN
fcis-30157	80	5	of	of	ADP
fcis-30157	80	6	ai	ai	VERB
fcis-30157	80	7	in	in	ADP
fcis-30157	80	8	clinical	clinical	ADJ
fcis-30157	80	9	practice	practice	NOUN
fcis-30157	80	10	raises	raise	VERB
fcis-30157	80	11	concerns	concern	NOUN
fcis-30157	80	12	about	about	ADP
fcis-30157	80	13	transparency	transparency	NOUN
fcis-30157	80	14	and	and	CCONJ
fcis-30157	80	15	bias	bias	NOUN
fcis-30157	80	16	.	.	PUNCT
fcis-30157	81	1	future	future	ADJ
fcis-30157	81	2	research	research	NOUN
fcis-30157	81	3	should	should	AUX
fcis-30157	81	4	address	address	VERB
fcis-30157	81	5	these	these	DET
fcis-30157	81	6	challenges	challenge	NOUN
fcis-30157	81	7	and	and	CCONJ
fcis-30157	81	8	explore	explore	VERB
fcis-30157	81	9	ways	way	NOUN
fcis-30157	81	10	to	to	PART
fcis-30157	81	11	enhance	enhance	VERB
fcis-30157	81	12	model	model	NOUN
fcis-30157	81	13	performance	performance	NOUN
fcis-30157	81	14	and	and	CCONJ
fcis-30157	81	15	clinical	clinical	ADJ
fcis-30157	81	16	applicability	applicability	NOUN
fcis-30157	81	17	.	.	PUNCT
fcis-30157	82	1	1.2	1.2	NUM
fcis-30157	82	2	.	.	PUNCT
fcis-30157	82	3	auxiliary	auxiliary	ADJ
fcis-30157	82	4	diagnosis	diagnosis	NOUN
fcis-30157	82	5	approaches	approach	VERB
fcis-30157	82	6	1.2.1	1.2.1	NUM
fcis-30157	82	7	.	.	PUNCT
fcis-30157	83	1	rule	rule	NOUN
fcis-30157	83	2	based	base	VERB
fcis-30157	83	3	diagnosis	diagnosis	NOUN
fcis-30157	83	4	systems	system	NOUN
fcis-30157	83	5	rule	rule	NOUN
fcis-30157	83	6	-	-	PUNCT
fcis-30157	83	7	based	base	VERB
fcis-30157	83	8	diagnosis	diagnosis	NOUN
fcis-30157	83	9	systems	system	NOUN
fcis-30157	83	10	are	be	AUX
fcis-30157	83	11	commonly	commonly	ADV
fcis-30157	83	12	used	use	VERB
fcis-30157	83	13	in	in	ADP
fcis-30157	83	14	dermatology	dermatology	NOUN
fcis-30157	83	15	to	to	PART
fcis-30157	83	16	mimic	mimic	VERB
fcis-30157	83	17	expert	expert	ADJ
fcis-30157	83	18	decision	decision	NOUN
fcis-30157	83	19	-	-	PUNCT
fcis-30157	83	20	making	making	NOUN
fcis-30157	83	21	by	by	ADP
fcis-30157	83	22	relying	rely	VERB
fcis-30157	83	23	on	on	ADP
fcis-30157	83	24	predefined	predefine	VERB
fcis-30157	83	25	rules	rule	NOUN
fcis-30157	83	26	derived	derive	VERB
fcis-30157	83	27	from	from	ADP
fcis-30157	83	28	clinical	clinical	ADJ
fcis-30157	83	29	experience	experience	NOUN
fcis-30157	83	30	and	and	CCONJ
fcis-30157	83	31	medical	medical	ADJ
fcis-30157	83	32	knowledge	knowledge	NOUN
fcis-30157	83	33	.	.	PUNCT
fcis-30157	84	1	these	these	DET
fcis-30157	84	2	systems	system	NOUN
fcis-30157	84	3	apply	apply	VERB
fcis-30157	84	4	criteria	criterion	NOUN
fcis-30157	84	5	like	like	ADP
fcis-30157	84	6	patient	patient	ADJ
fcis-30157	84	7	symptoms	symptom	NOUN
fcis-30157	84	8	and	and	CCONJ
fcis-30157	84	9	medical	medical	ADJ
fcis-30157	84	10	history	history	NOUN
fcis-30157	84	11	to	to	PART
fcis-30157	84	12	suggest	suggest	VERB
fcis-30157	84	13	diagnoses	diagnosis	NOUN
fcis-30157	84	14	,	,	PUNCT
fcis-30157	84	15	e.g.	e.g.	ADV
fcis-30157	84	16	,	,	PUNCT
fcis-30157	84	17	"	"	PUNCT
fcis-30157	84	18	if	if	SCONJ
fcis-30157	84	19	a	a	DET
fcis-30157	84	20	patient	patient	NOUN
fcis-30157	84	21	has	have	VERB
fcis-30157	84	22	a	a	DET
fcis-30157	84	23	red	red	ADJ
fcis-30157	84	24	,	,	PUNCT
fcis-30157	84	25	itchy	itchy	ADJ
fcis-30157	84	26	rash	rash	NOUN
fcis-30157	84	27	and	and	CCONJ
fcis-30157	84	28	recent	recent	ADJ
fcis-30157	84	29	allergen	allergen	NOUN
fcis-30157	84	30	exposure	exposure	NOUN
fcis-30157	84	31	,	,	PUNCT
fcis-30157	84	32	the	the	DET
fcis-30157	84	33	diagnosis	diagnosis	NOUN
fcis-30157	84	34	is	be	AUX
fcis-30157	84	35	likely	likely	ADJ
fcis-30157	84	36	contact	contact	NOUN
fcis-30157	84	37	dermatitis	dermatitis	NOUN
fcis-30157	84	38	.	.	PUNCT
fcis-30157	84	39	"	"	PUNCT
fcis-30157	85	1	such	such	ADJ
fcis-30157	85	2	rules	rule	NOUN
fcis-30157	85	3	,	,	PUNCT
fcis-30157	85	4	codified	codify	VERB
fcis-30157	85	5	into	into	ADP
fcis-30157	85	6	decision	decision	NOUN
fcis-30157	85	7	trees	tree	NOUN
fcis-30157	85	8	or	or	CCONJ
fcis-30157	85	9	logical	logical	ADJ
fcis-30157	85	10	statements	statement	NOUN
fcis-30157	85	11	,	,	PUNCT
fcis-30157	85	12	offer	offer	VERB
fcis-30157	85	13	a	a	DET
fcis-30157	85	14	structured	structured	ADJ
fcis-30157	85	15	approach	approach	NOUN
fcis-30157	85	16	,	,	PUNCT
fcis-30157	85	17	reducing	reduce	VERB
fcis-30157	85	18	subjective	subjective	ADJ
fcis-30157	85	19	judgment	judgment	NOUN
fcis-30157	85	20	.	.	PUNCT
fcis-30157	86	1	a	a	DET
fcis-30157	86	2	study	study	NOUN
fcis-30157	86	3	by	by	ADP
fcis-30157	86	4	the	the	DET
fcis-30157	86	5	eadv	eadv	ADJ
fcis-30157	86	6	task	task	NOUN
fcis-30157	86	7	force	force	NOUN
fcis-30157	86	8	on	on	ADP
fcis-30157	86	9	contact	contact	NOUN
fcis-30157	86	10	dermatitis	dermatitis	NOUN
fcis-30157	86	11	[	[	X
fcis-30157	86	12	20	20	NUM
fcis-30157	86	13	]	]	PUNCT
fcis-30157	86	14	developed	develop	VERB
fcis-30157	86	15	specific	specific	ADJ
fcis-30157	86	16	diagnostic	diagnostic	ADJ
fcis-30157	86	17	criteria	criterion	NOUN
fcis-30157	86	18	incorporated	incorporate	VERB
fcis-30157	86	19	into	into	ADP
fcis-30157	86	20	rulebased	rulebase	VERB
fcis-30157	86	21	systems	system	NOUN
fcis-30157	86	22	,	,	PUNCT
fcis-30157	86	23	improving	improve	VERB
fcis-30157	86	24	diagnostic	diagnostic	ADJ
fcis-30157	86	25	consistency	consistency	NOUN
fcis-30157	86	26	.	.	PUNCT
fcis-30157	87	1	these	these	DET
fcis-30157	87	2	systems	system	NOUN
fcis-30157	87	3	are	be	AUX
fcis-30157	87	4	transparent	transparent	ADJ
fcis-30157	87	5	and	and	CCONJ
fcis-30157	87	6	interpretable	interpretable	ADJ
fcis-30157	87	7	,	,	PUNCT
fcis-30157	87	8	providing	provide	VERB
fcis-30157	87	9	clear	clear	ADJ
fcis-30157	87	10	reasoning	reasoning	NOUN
fcis-30157	87	11	for	for	ADP
fcis-30157	87	12	their	their	PRON
fcis-30157	87	13	conclusions	conclusion	NOUN
fcis-30157	87	14	,	,	PUNCT
fcis-30157	87	15	which	which	PRON
fcis-30157	87	16	is	be	AUX
fcis-30157	87	17	valuable	valuable	ADJ
fcis-30157	87	18	in	in	ADP
fcis-30157	87	19	clinical	clinical	ADJ
fcis-30157	87	20	settings	setting	NOUN
fcis-30157	87	21	.	.	PUNCT
fcis-30157	88	1	for	for	ADP
fcis-30157	88	2	example	example	NOUN
fcis-30157	88	3	,	,	PUNCT
fcis-30157	88	4	if	if	SCONJ
fcis-30157	88	5	diagnosing	diagnose	VERB
fcis-30157	88	6	psoriasis	psoriasis	NOUN
fcis-30157	88	7	,	,	PUNCT
fcis-30157	88	8	a	a	DET
fcis-30157	88	9	rule	rule	NOUN
fcis-30157	88	10	-	-	PUNCT
fcis-30157	88	11	based	base	VERB
fcis-30157	88	12	system	system	NOUN
fcis-30157	88	13	can	can	AUX
fcis-30157	88	14	list	list	VERB
fcis-30157	88	15	specific	specific	ADJ
fcis-30157	88	16	symptoms	symptom	NOUN
fcis-30157	88	17	and	and	CCONJ
fcis-30157	88	18	test	test	NOUN
fcis-30157	88	19	results	result	NOUN
fcis-30157	88	20	that	that	PRON
fcis-30157	88	21	support	support	VERB
fcis-30157	88	22	the	the	DET
fcis-30157	88	23	diagnosis	diagnosis	NOUN
fcis-30157	88	24	,	,	PUNCT
fcis-30157	88	25	helping	help	VERB
fcis-30157	88	26	clinicians	clinician	NOUN
fcis-30157	88	27	verify	verify	VERB
fcis-30157	88	28	the	the	DET
fcis-30157	88	29	result	result	NOUN
fcis-30157	88	30	.	.	PUNCT
fcis-30157	89	1	rule	rule	NOUN
fcis-30157	89	2	-	-	PUNCT
fcis-30157	89	3	based	base	VERB
fcis-30157	89	4	systems	system	NOUN
fcis-30157	89	5	can	can	AUX
fcis-30157	89	6	achieve	achieve	VERB
fcis-30157	89	7	diagnostic	diagnostic	ADJ
fcis-30157	89	8	accuracy	accuracy	NOUN
fcis-30157	89	9	of	of	ADP
fcis-30157	89	10	up	up	ADP
fcis-30157	89	11	to	to	PART
fcis-30157	89	12	85	85	NUM
fcis-30157	89	13	%	%	NOUN
fcis-30157	89	14	in	in	ADP
fcis-30157	89	15	controlled	control	VERB
fcis-30157	89	16	settings	setting	NOUN
fcis-30157	89	17	[	[	X
fcis-30157	89	18	17	17	NUM
fcis-30157	89	19	]	]	PUNCT
fcis-30157	89	20	,	,	PUNCT
fcis-30157	89	21	but	but	CCONJ
fcis-30157	89	22	they	they	PRON
fcis-30157	89	23	face	face	VERB
fcis-30157	89	24	challenges	challenge	NOUN
fcis-30157	89	25	in	in	ADP
fcis-30157	89	26	staying	stay	VERB
fcis-30157	89	27	current	current	ADJ
fcis-30157	89	28	with	with	ADP
fcis-30157	89	29	evolving	evolve	VERB
fcis-30157	89	30	medical	medical	ADJ
fcis-30157	89	31	knowledge	knowledge	NOUN
fcis-30157	89	32	and	and	CCONJ
fcis-30157	89	33	adapting	adapt	VERB
fcis-30157	89	34	to	to	ADP
fcis-30157	89	35	new	new	ADJ
fcis-30157	89	36	data	datum	NOUN
fcis-30157	89	37	.	.	PUNCT
fcis-30157	90	1	for	for	ADP
fcis-30157	90	2	example	example	NOUN
fcis-30157	90	3	,	,	PUNCT
fcis-30157	90	4	the	the	DET
fcis-30157	90	5	isic	isic	PROPN
fcis-30157	90	6	frequently	frequently	ADV
fcis-30157	90	7	updates	update	VERB
fcis-30157	90	8	its	its	PRON
fcis-30157	90	9	skin	skin	NOUN
fcis-30157	90	10	cancer	cancer	NOUN
fcis-30157	90	11	diagnostic	diagnostic	ADJ
fcis-30157	90	12	guidelines	guideline	NOUN
fcis-30157	90	13	,	,	PUNCT
fcis-30157	90	14	and	and	CCONJ
fcis-30157	90	15	rule	rule	NOUN
fcis-30157	90	16	-	-	PUNCT
fcis-30157	90	17	based	base	VERB
fcis-30157	90	18	systems	system	NOUN
fcis-30157	90	19	may	may	AUX
fcis-30157	90	20	become	become	VERB
fcis-30157	90	21	outdated	outdated	ADJ
fcis-30157	90	22	if	if	SCONJ
fcis-30157	90	23	not	not	PART
fcis-30157	90	24	regularly	regularly	ADV
fcis-30157	90	25	updated	update	VERB
fcis-30157	90	26	.	.	PUNCT
fcis-30157	91	1	additionally	additionally	ADV
fcis-30157	91	2	,	,	PUNCT
fcis-30157	91	3	capturing	capture	VERB
fcis-30157	91	4	complex	complex	ADJ
fcis-30157	91	5	diagnostic	diagnostic	ADJ
fcis-30157	91	6	patterns	pattern	NOUN
fcis-30157	91	7	is	be	AUX
fcis-30157	91	8	difficult	difficult	ADJ
fcis-30157	91	9	.	.	PUNCT
fcis-30157	92	1	dermatological	dermatological	ADJ
fcis-30157	92	2	conditions	condition	NOUN
fcis-30157	92	3	often	often	ADV
fcis-30157	92	4	overlap	overlap	VERB
fcis-30157	92	5	,	,	PUNCT
fcis-30157	92	6	making	make	VERB
fcis-30157	92	7	it	it	PRON
fcis-30157	92	8	hard	hard	ADJ
fcis-30157	92	9	to	to	PART
fcis-30157	92	10	create	create	VERB
fcis-30157	92	11	accurate	accurate	ADJ
fcis-30157	92	12	rules	rule	NOUN
fcis-30157	92	13	for	for	ADP
fcis-30157	92	14	distinguishing	distinguish	VERB
fcis-30157	92	15	similar	similar	ADJ
fcis-30157	92	16	diseases	disease	NOUN
fcis-30157	92	17	like	like	ADP
fcis-30157	92	18	eczema	eczema	NOUN
fcis-30157	92	19	and	and	CCONJ
fcis-30157	92	20	psoriasis	psoriasis	NOUN
fcis-30157	92	21	.	.	PUNCT
fcis-30157	93	1	to	to	PART
fcis-30157	93	2	address	address	VERB
fcis-30157	93	3	these	these	DET
fcis-30157	93	4	limitations	limitation	NOUN
fcis-30157	93	5	,	,	PUNCT
fcis-30157	93	6	hybrid	hybrid	ADJ
fcis-30157	93	7	systems	system	NOUN
fcis-30157	93	8	combining	combine	VERB
fcis-30157	93	9	rule	rule	NOUN
fcis-30157	93	10	-	-	PUNCT
fcis-30157	93	11	based	base	VERB
fcis-30157	93	12	approaches	approach	NOUN
fcis-30157	93	13	with	with	ADP
fcis-30157	93	14	machine	machine	NOUN
fcis-30157	93	15	learning	learning	NOUN
fcis-30157	93	16	techniques	technique	NOUN
fcis-30157	93	17	have	have	AUX
fcis-30157	93	18	been	be	AUX
fcis-30157	93	19	explored	explore	VERB
fcis-30157	93	20	.	.	PUNCT
fcis-30157	94	1	for	for	ADP
fcis-30157	94	2	instance	instance	NOUN
fcis-30157	94	3	,	,	PUNCT
fcis-30157	94	4	integrating	integrate	VERB
fcis-30157	94	5	rule	rule	NOUN
fcis-30157	94	6	-	-	PUNCT
fcis-30157	94	7	based	base	VERB
fcis-30157	94	8	reasoning	reasoning	NOUN
fcis-30157	94	9	with	with	ADP
fcis-30157	94	10	cnns	cnn	NOUN
fcis-30157	94	11	has	have	AUX
fcis-30157	94	12	been	be	AUX
fcis-30157	94	13	shown	show	VERB
fcis-30157	94	14	to	to	PART
fcis-30157	94	15	improve	improve	VERB
fcis-30157	94	16	diagnostic	diagnostic	ADJ
fcis-30157	94	17	accuracy	accuracy	NOUN
fcis-30157	95	1	[	[	X
fcis-30157	95	2	16	16	NUM
fcis-30157	95	3	]	]	PUNCT
fcis-30157	95	4	.	.	PUNCT
fcis-30157	96	1	in	in	ADP
fcis-30157	96	2	this	this	DET
fcis-30157	96	3	approach	approach	NOUN
fcis-30157	96	4	,	,	PUNCT
fcis-30157	96	5	rulebased	rulebase	VERB
fcis-30157	96	6	systems	system	NOUN
fcis-30157	96	7	filter	filter	NOUN
fcis-30157	96	8	diagnoses	diagnosis	NOUN
fcis-30157	96	9	,	,	PUNCT
fcis-30157	96	10	while	while	SCONJ
fcis-30157	96	11	machine	machine	NOUN
fcis-30157	96	12	learning	learning	NOUN
fcis-30157	96	13	models	model	NOUN
fcis-30157	96	14	refine	refine	VERB
fcis-30157	96	15	them	they	PRON
fcis-30157	96	16	by	by	ADP
fcis-30157	96	17	analyzing	analyze	VERB
fcis-30157	96	18	complex	complex	ADJ
fcis-30157	96	19	data	datum	NOUN
fcis-30157	96	20	patterns	pattern	NOUN
fcis-30157	96	21	.	.	PUNCT
fcis-30157	97	1	rulebased	rulebase	VERB
fcis-30157	97	2	systems	system	NOUN
fcis-30157	97	3	are	be	AUX
fcis-30157	97	4	used	use	VERB
fcis-30157	97	5	in	in	ADP
fcis-30157	97	6	clinical	clinical	ADJ
fcis-30157	97	7	settings	setting	NOUN
fcis-30157	97	8	,	,	PUNCT
fcis-30157	97	9	especially	especially	ADV
fcis-30157	97	10	in	in	ADP
fcis-30157	97	11	telemedicine	telemedicine	NOUN
fcis-30157	97	12	,	,	PUNCT
fcis-30157	97	13	where	where	SCONJ
fcis-30157	97	14	they	they	PRON
fcis-30157	97	15	guide	guide	VERB
fcis-30157	97	16	non	non	ADJ
fcis-30157	97	17	-	-	ADJ
fcis-30157	97	18	specialist	specialist	ADJ
fcis-30157	97	19	clinicians	clinician	NOUN
fcis-30157	97	20	through	through	ADP
fcis-30157	97	21	diagnosis	diagnosis	NOUN
fcis-30157	97	22	[	[	X
fcis-30157	97	23	19	19	NUM
fcis-30157	97	24	]	]	PUNCT
fcis-30157	97	25	.	.	PUNCT
fcis-30157	98	1	these	these	DET
fcis-30157	98	2	systems	system	NOUN
fcis-30157	98	3	help	help	VERB
fcis-30157	98	4	reduce	reduce	VERB
fcis-30157	98	5	dermatologists	dermatologist	NOUN
fcis-30157	98	6	'	'	PART
fcis-30157	98	7	workload	workload	NOUN
fcis-30157	98	8	by	by	ADP
fcis-30157	98	9	handling	handle	VERB
fcis-30157	98	10	simpler	simple	ADJ
fcis-30157	98	11	cases	case	NOUN
fcis-30157	98	12	,	,	PUNCT
fcis-30157	98	13	but	but	CCONJ
fcis-30157	98	14	they	they	PRON
fcis-30157	98	15	still	still	ADV
fcis-30157	98	16	depend	depend	VERB
fcis-30157	98	17	on	on	ADP
fcis-30157	98	18	the	the	DET
fcis-30157	98	19	quality	quality	NOUN
fcis-30157	98	20	of	of	ADP
fcis-30157	98	21	the	the	DET
fcis-30157	98	22	rules	rule	NOUN
fcis-30157	98	23	.	.	PUNCT
fcis-30157	99	1	as	as	SCONJ
fcis-30157	99	2	medical	medical	ADJ
fcis-30157	99	3	knowledge	knowledge	NOUN
fcis-30157	99	4	advances	advance	NOUN
fcis-30157	99	5	,	,	PUNCT
fcis-30157	99	6	developing	develop	VERB
fcis-30157	99	7	more	more	ADV
fcis-30157	99	8	dynamic	dynamic	ADJ
fcis-30157	99	9	,	,	PUNCT
fcis-30157	99	10	adaptable	adaptable	ADJ
fcis-30157	99	11	systems	system	NOUN
fcis-30157	99	12	that	that	PRON
fcis-30157	99	13	integrate	integrate	VERB
fcis-30157	99	14	new	new	ADJ
fcis-30157	99	15	data	datum	NOUN
fcis-30157	99	16	and	and	CCONJ
fcis-30157	99	17	machine	machine	NOUN
fcis-30157	99	18	learning	learning	NOUN
fcis-30157	99	19	could	could	AUX
fcis-30157	99	20	improve	improve	VERB
fcis-30157	99	21	their	their	PRON
fcis-30157	99	22	effectiveness	effectiveness	NOUN
fcis-30157	99	23	,	,	PUNCT
fcis-30157	99	24	making	make	VERB
fcis-30157	99	25	them	they	PRON
fcis-30157	99	26	more	more	ADV
fcis-30157	99	27	robust	robust	ADJ
fcis-30157	99	28	and	and	CCONJ
fcis-30157	99	29	suitable	suitable	ADJ
fcis-30157	99	30	for	for	ADP
fcis-30157	99	31	modern	modern	ADJ
fcis-30157	99	32	dermatology	dermatology	NOUN
fcis-30157	99	33	.	.	PUNCT
fcis-30157	100	1	114	114	NUM
fcis-30157	100	2	1.2.2	1.2.2	NUM
fcis-30157	100	3	.	.	PUNCT
fcis-30157	101	1	knowledge	knowledge	NOUN
fcis-30157	101	2	based	base	VERB
fcis-30157	101	3	expert	expert	NOUN
fcis-30157	101	4	systems	system	NOUN
fcis-30157	101	5	knowledge	knowledge	NOUN
fcis-30157	101	6	-	-	PUNCT
fcis-30157	101	7	based	base	VERB
fcis-30157	101	8	expert	expert	NOUN
fcis-30157	101	9	systems	system	NOUN
fcis-30157	101	10	are	be	AUX
fcis-30157	101	11	crucial	crucial	ADJ
fcis-30157	101	12	in	in	ADP
fcis-30157	101	13	skin	skin	NOUN
fcis-30157	101	14	disease	disease	NOUN
fcis-30157	101	15	diagnosis	diagnosis	NOUN
fcis-30157	101	16	,	,	PUNCT
fcis-30157	101	17	relying	rely	VERB
fcis-30157	101	18	on	on	ADP
fcis-30157	101	19	structured	structured	ADJ
fcis-30157	101	20	medical	medical	ADJ
fcis-30157	101	21	knowledge	knowledge	NOUN
fcis-30157	101	22	bases	basis	NOUN
fcis-30157	101	23	.	.	PUNCT
fcis-30157	102	1	the	the	DET
fcis-30157	102	2	process	process	NOUN
fcis-30157	102	3	involves	involve	VERB
fcis-30157	102	4	acquiring	acquire	VERB
fcis-30157	102	5	,	,	PUNCT
fcis-30157	102	6	representing	represent	VERB
fcis-30157	102	7	,	,	PUNCT
fcis-30157	102	8	and	and	CCONJ
fcis-30157	102	9	storing	store	VERB
fcis-30157	102	10	medical	medical	ADJ
fcis-30157	102	11	knowledge	knowledge	NOUN
fcis-30157	102	12	,	,	PUNCT
fcis-30157	102	13	with	with	ADP
fcis-30157	102	14	acquisition	acquisition	NOUN
fcis-30157	102	15	being	be	AUX
fcis-30157	102	16	the	the	DET
fcis-30157	102	17	most	most	ADV
fcis-30157	102	18	challenging	challenging	ADJ
fcis-30157	102	19	.	.	PUNCT
fcis-30157	103	1	it	it	PRON
fcis-30157	103	2	requires	require	VERB
fcis-30157	103	3	extracting	extract	VERB
fcis-30157	103	4	relevant	relevant	ADJ
fcis-30157	103	5	information	information	NOUN
fcis-30157	103	6	from	from	ADP
fcis-30157	103	7	dermatologists	dermatologist	NOUN
fcis-30157	103	8	and	and	CCONJ
fcis-30157	103	9	converting	convert	VERB
fcis-30157	103	10	it	it	PRON
fcis-30157	103	11	into	into	ADP
fcis-30157	103	12	a	a	DET
fcis-30157	103	13	machine	machine	NOUN
fcis-30157	103	14	-	-	PUNCT
fcis-30157	103	15	readable	readable	ADJ
fcis-30157	103	16	format	format	NOUN
fcis-30157	103	17	,	,	PUNCT
fcis-30157	103	18	often	often	ADV
fcis-30157	103	19	using	use	VERB
fcis-30157	103	20	ontologies	ontology	NOUN
fcis-30157	103	21	or	or	CCONJ
fcis-30157	103	22	nlp	nlp	NOUN
fcis-30157	103	23	techniques	technique	NOUN
fcis-30157	103	24	.	.	PUNCT
fcis-30157	104	1	for	for	ADP
fcis-30157	104	2	example	example	NOUN
fcis-30157	104	3	,	,	PUNCT
fcis-30157	104	4	the	the	DET
fcis-30157	104	5	isic	isic	PROPN
fcis-30157	104	6	has	have	AUX
fcis-30157	104	7	developed	develop	VERB
fcis-30157	104	8	a	a	DET
fcis-30157	104	9	database	database	NOUN
fcis-30157	104	10	of	of	ADP
fcis-30157	104	11	dermatoscopic	dermatoscopic	ADJ
fcis-30157	104	12	images	image	NOUN
fcis-30157	104	13	to	to	PART
fcis-30157	104	14	aid	aid	VERB
fcis-30157	104	15	knowledge	knowledge	NOUN
fcis-30157	104	16	acquisition	acquisition	NOUN
fcis-30157	104	17	.	.	PUNCT
fcis-30157	105	1	knowledge	knowledge	NOUN
fcis-30157	105	2	is	be	AUX
fcis-30157	105	3	represented	represent	VERB
fcis-30157	105	4	using	use	VERB
fcis-30157	105	5	formal	formal	ADJ
fcis-30157	105	6	structures	structure	NOUN
fcis-30157	105	7	like	like	ADP
fcis-30157	105	8	rules	rule	NOUN
fcis-30157	105	9	or	or	CCONJ
fcis-30157	105	10	semantic	semantic	ADJ
fcis-30157	105	11	networks	network	NOUN
fcis-30157	105	12	and	and	CCONJ
fcis-30157	105	13	stored	store	VERB
fcis-30157	105	14	in	in	ADP
fcis-30157	105	15	databases	database	NOUN
fcis-30157	105	16	or	or	CCONJ
fcis-30157	105	17	knowledge	knowledge	NOUN
fcis-30157	105	18	graphs	graph	NOUN
fcis-30157	105	19	for	for	ADP
fcis-30157	105	20	efficient	efficient	ADJ
fcis-30157	105	21	retrieval	retrieval	NOUN
fcis-30157	105	22	.	.	PUNCT
fcis-30157	106	1	the	the	DET
fcis-30157	106	2	reasoning	reasoning	NOUN
fcis-30157	106	3	engine	engine	NOUN
fcis-30157	106	4	applies	apply	VERB
fcis-30157	106	5	inference	inference	NOUN
fcis-30157	106	6	techniques	technique	NOUN
fcis-30157	106	7	to	to	PART
fcis-30157	106	8	derive	derive	VERB
fcis-30157	106	9	conclusions	conclusion	NOUN
fcis-30157	106	10	based	base	VERB
fcis-30157	106	11	on	on	ADP
fcis-30157	106	12	input	input	NOUN
fcis-30157	106	13	data	datum	NOUN
fcis-30157	106	14	,	,	PUNCT
fcis-30157	106	15	using	use	VERB
fcis-30157	106	16	methods	method	NOUN
fcis-30157	106	17	like	like	ADP
fcis-30157	106	18	forwardor	forwardor	VERB
fcis-30157	106	19	backward	backward	ADV
fcis-30157	106	20	-	-	PUNCT
fcis-30157	106	21	chaining	chain	VERB
fcis-30157	106	22	.	.	PUNCT
fcis-30157	107	1	despite	despite	SCONJ
fcis-30157	107	2	their	their	PRON
fcis-30157	107	3	effectiveness	effectiveness	NOUN
fcis-30157	107	4	,	,	PUNCT
fcis-30157	107	5	these	these	DET
fcis-30157	107	6	systems	system	NOUN
fcis-30157	107	7	face	face	VERB
fcis-30157	107	8	challenges	challenge	NOUN
fcis-30157	107	9	in	in	ADP
fcis-30157	107	10	knowledge	knowledge	NOUN
fcis-30157	107	11	updates	update	NOUN
fcis-30157	107	12	and	and	CCONJ
fcis-30157	107	13	handling	handle	VERB
fcis-30157	107	14	conflicts	conflict	NOUN
fcis-30157	107	15	when	when	SCONJ
fcis-30157	107	16	multiple	multiple	ADJ
fcis-30157	107	17	rules	rule	NOUN
fcis-30157	107	18	suggest	suggest	VERB
fcis-30157	107	19	different	different	ADJ
fcis-30157	107	20	conclusions	conclusion	NOUN
fcis-30157	107	21	.	.	PUNCT
fcis-30157	108	1	for	for	ADP
fcis-30157	108	2	example	example	NOUN
fcis-30157	108	3	,	,	PUNCT
fcis-30157	108	4	lesions	lesion	NOUN
fcis-30157	108	5	with	with	ADP
fcis-30157	108	6	overlapping	overlap	VERB
fcis-30157	108	7	characteristics	characteristic	NOUN
fcis-30157	108	8	may	may	AUX
fcis-30157	108	9	complicate	complicate	VERB
fcis-30157	108	10	diagnosis	diagnosis	NOUN
fcis-30157	108	11	[	[	X
fcis-30157	108	12	21	21	NUM
fcis-30157	108	13	]	]	PUNCT
fcis-30157	108	14	.	.	PUNCT
fcis-30157	109	1	they	they	PRON
fcis-30157	109	2	also	also	ADV
fcis-30157	109	3	rely	rely	VERB
fcis-30157	109	4	on	on	ADP
fcis-30157	109	5	predefined	predefine	VERB
fcis-30157	109	6	rules	rule	NOUN
fcis-30157	109	7	,	,	PUNCT
fcis-30157	109	8	which	which	PRON
fcis-30157	109	9	limits	limit	VERB
fcis-30157	109	10	adaptability	adaptability	NOUN
fcis-30157	109	11	to	to	PART
fcis-30157	109	12	novel	novel	VERB
fcis-30157	109	13	or	or	CCONJ
fcis-30157	109	14	ambiguous	ambiguous	ADJ
fcis-30157	109	15	cases	case	NOUN
fcis-30157	109	16	.	.	PUNCT
fcis-30157	110	1	hybrid	hybrid	ADJ
fcis-30157	110	2	systems	system	NOUN
fcis-30157	110	3	combining	combine	VERB
fcis-30157	110	4	expert	expert	ADJ
fcis-30157	110	5	knowledge	knowledge	NOUN
fcis-30157	110	6	and	and	CCONJ
fcis-30157	110	7	machine	machine	NOUN
fcis-30157	110	8	learning	learning	NOUN
fcis-30157	110	9	have	have	AUX
fcis-30157	110	10	improved	improve	VERB
fcis-30157	110	11	performance	performance	NOUN
fcis-30157	110	12	[	[	X
fcis-30157	110	13	9	9	NUM
fcis-30157	110	14	]	]	PUNCT
fcis-30157	110	15	.	.	PUNCT
fcis-30157	111	1	data	data	NOUN
fcis-30157	111	2	-	-	PUNCT
fcis-30157	111	3	driven	drive	VERB
fcis-30157	111	4	methods	method	NOUN
fcis-30157	111	5	represent	represent	VERB
fcis-30157	111	6	a	a	DET
fcis-30157	111	7	shift	shift	NOUN
fcis-30157	111	8	in	in	ADP
fcis-30157	111	9	diagnostic	diagnostic	ADJ
fcis-30157	111	10	approaches	approach	NOUN
fcis-30157	111	11	by	by	ADP
fcis-30157	111	12	leveraging	leverage	VERB
fcis-30157	111	13	large	large	ADJ
fcis-30157	111	14	clinical	clinical	ADJ
fcis-30157	111	15	datasets	dataset	NOUN
fcis-30157	111	16	to	to	PART
fcis-30157	111	17	build	build	VERB
fcis-30157	111	18	predictive	predictive	ADJ
fcis-30157	111	19	models	model	NOUN
fcis-30157	111	20	,	,	PUNCT
fcis-30157	111	21	with	with	ADP
fcis-30157	111	22	cnns	cnns	ADJ
fcis-30157	111	23	and	and	CCONJ
fcis-30157	111	24	random	random	ADJ
fcis-30157	111	25	forests	forest	NOUN
fcis-30157	111	26	commonly	commonly	ADV
fcis-30157	111	27	used	use	VERB
fcis-30157	111	28	for	for	ADP
fcis-30157	111	29	skin	skin	NOUN
fcis-30157	111	30	disease	disease	NOUN
fcis-30157	111	31	classification	classification	NOUN
fcis-30157	111	32	.	.	PUNCT
fcis-30157	112	1	these	these	DET
fcis-30157	112	2	models	model	NOUN
fcis-30157	112	3	excel	excel	VERB
fcis-30157	112	4	at	at	ADP
fcis-30157	112	5	extracting	extract	VERB
fcis-30157	112	6	features	feature	NOUN
fcis-30157	112	7	from	from	ADP
fcis-30157	112	8	images	image	NOUN
fcis-30157	112	9	,	,	PUNCT
fcis-30157	112	10	such	such	ADJ
fcis-30157	112	11	as	as	ADP
fcis-30157	112	12	texture	texture	NOUN
fcis-30157	112	13	and	and	CCONJ
fcis-30157	112	14	color	color	NOUN
fcis-30157	112	15	,	,	PUNCT
fcis-30157	112	16	and	and	CCONJ
fcis-30157	112	17	can	can	AUX
fcis-30157	112	18	handle	handle	VERB
fcis-30157	112	19	large	large	ADJ
fcis-30157	112	20	,	,	PUNCT
fcis-30157	112	21	high	high	ADJ
fcis-30157	112	22	-	-	PUNCT
fcis-30157	112	23	dimensional	dimensional	ADJ
fcis-30157	112	24	datasets	dataset	NOUN
fcis-30157	112	25	[	[	X
fcis-30157	112	26	22	22	NUM
fcis-30157	112	27	]	]	PUNCT
fcis-30157	112	28	.	.	PUNCT
fcis-30157	113	1	however	however	ADV
fcis-30157	113	2	,	,	PUNCT
fcis-30157	113	3	data	datum	NOUN
fcis-30157	113	4	scarcity	scarcity	NOUN
fcis-30157	113	5	,	,	PUNCT
fcis-30157	113	6	especially	especially	ADV
fcis-30157	113	7	for	for	ADP
fcis-30157	113	8	rare	rare	ADJ
fcis-30157	113	9	diseases	disease	NOUN
fcis-30157	113	10	,	,	PUNCT
fcis-30157	113	11	remains	remain	VERB
fcis-30157	113	12	a	a	DET
fcis-30157	113	13	challenge	challenge	NOUN
fcis-30157	113	14	,	,	PUNCT
fcis-30157	113	15	and	and	CCONJ
fcis-30157	113	16	data	datum	NOUN
fcis-30157	113	17	augmentation	augmentation	NOUN
fcis-30157	113	18	techniques	technique	NOUN
fcis-30157	113	19	can	can	AUX
fcis-30157	113	20	help	help	VERB
fcis-30157	113	21	mitigate	mitigate	VERB
fcis-30157	113	22	this	this	PRON
fcis-30157	113	23	.	.	PUNCT
fcis-30157	114	1	the	the	DET
fcis-30157	114	2	size	size	NOUN
fcis-30157	114	3	and	and	CCONJ
fcis-30157	114	4	quality	quality	NOUN
fcis-30157	114	5	of	of	ADP
fcis-30157	114	6	datasets	dataset	NOUN
fcis-30157	114	7	directly	directly	ADV
fcis-30157	114	8	affect	affect	VERB
fcis-30157	114	9	model	model	NOUN
fcis-30157	114	10	performance	performance	NOUN
fcis-30157	114	11	,	,	PUNCT
fcis-30157	114	12	with	with	ADP
fcis-30157	114	13	larger	large	ADJ
fcis-30157	114	14	datasets	dataset	NOUN
fcis-30157	114	15	leading	lead	VERB
fcis-30157	114	16	to	to	ADP
fcis-30157	114	17	better	well	ADJ
fcis-30157	114	18	generalization	generalization	NOUN
fcis-30157	114	19	[	[	X
fcis-30157	114	20	25][26	25][26	PROPN
fcis-30157	114	21	]	]	X
fcis-30157	114	22	.	.	PUNCT
fcis-30157	115	1	data	data	NOUN
fcis-30157	115	2	-	-	PUNCT
fcis-30157	115	3	driven	drive	VERB
fcis-30157	115	4	models	model	NOUN
fcis-30157	115	5	can	can	AUX
fcis-30157	115	6	generalize	generalize	VERB
fcis-30157	115	7	across	across	ADP
fcis-30157	115	8	different	different	ADJ
fcis-30157	115	9	populations	population	NOUN
fcis-30157	115	10	,	,	PUNCT
fcis-30157	115	11	but	but	CCONJ
fcis-30157	115	12	they	they	PRON
fcis-30157	115	13	also	also	ADV
fcis-30157	115	14	face	face	VERB
fcis-30157	115	15	ethical	ethical	ADJ
fcis-30157	115	16	concerns	concern	NOUN
fcis-30157	115	17	,	,	PUNCT
fcis-30157	115	18	such	such	ADJ
fcis-30157	115	19	as	as	ADP
fcis-30157	115	20	bias	bias	NOUN
fcis-30157	115	21	in	in	ADP
fcis-30157	115	22	unrepresentative	unrepresentative	ADJ
fcis-30157	115	23	datasets	dataset	NOUN
fcis-30157	115	24	.	.	PUNCT
fcis-30157	116	1	their	their	PRON
fcis-30157	116	2	"	"	PUNCT
fcis-30157	116	3	black	black	ADJ
fcis-30157	116	4	box	box	NOUN
fcis-30157	116	5	"	"	PUNCT
fcis-30157	116	6	nature	nature	NOUN
fcis-30157	116	7	complicates	complicate	VERB
fcis-30157	116	8	interpretability	interpretability	NOUN
fcis-30157	116	9	,	,	PUNCT
fcis-30157	116	10	making	make	VERB
fcis-30157	116	11	them	they	PRON
fcis-30157	116	12	difficult	difficult	ADJ
fcis-30157	116	13	to	to	PART
fcis-30157	116	14	trust	trust	VERB
fcis-30157	116	15	in	in	ADP
fcis-30157	116	16	clinical	clinical	ADJ
fcis-30157	116	17	practice	practice	NOUN
fcis-30157	116	18	.	.	PUNCT
fcis-30157	117	1	to	to	PART
fcis-30157	117	2	address	address	VERB
fcis-30157	117	3	these	these	PRON
fcis-30157	117	4	,	,	PUNCT
fcis-30157	117	5	researchers	researcher	NOUN
fcis-30157	117	6	have	have	AUX
fcis-30157	117	7	explored	explore	VERB
fcis-30157	117	8	hybrid	hybrid	NOUN
fcis-30157	117	9	models	model	NOUN
fcis-30157	117	10	that	that	PRON
fcis-30157	117	11	combine	combine	VERB
fcis-30157	117	12	expert	expert	ADJ
fcis-30157	117	13	knowledge	knowledge	NOUN
fcis-30157	117	14	and	and	CCONJ
fcis-30157	117	15	machine	machine	NOUN
fcis-30157	117	16	learning	learning	NOUN
fcis-30157	117	17	,	,	PUNCT
fcis-30157	117	18	such	such	ADJ
fcis-30157	117	19	as	as	ADP
fcis-30157	117	20	dual	dual	ADJ
fcis-30157	117	21	-	-	PUNCT
fcis-30157	117	22	attentionbased	attentionbase	VERB
fcis-30157	117	23	networks	network	NOUN
fcis-30157	117	24	and	and	CCONJ
fcis-30157	117	25	multimodal	multimodal	NOUN
fcis-30157	117	26	fusion	fusion	NOUN
fcis-30157	117	27	methods	method	NOUN
fcis-30157	117	28	[	[	X
fcis-30157	117	29	7][8	7][8	X
fcis-30157	117	30	]	]	X
fcis-30157	117	31	.	.	PUNCT
fcis-30157	118	1	both	both	DET
fcis-30157	118	2	knowledge	knowledge	NOUN
fcis-30157	118	3	-	-	PUNCT
fcis-30157	118	4	based	base	VERB
fcis-30157	118	5	and	and	CCONJ
fcis-30157	118	6	data	data	NOUN
fcis-30157	118	7	-	-	PUNCT
fcis-30157	118	8	driven	drive	VERB
fcis-30157	118	9	methods	method	NOUN
fcis-30157	118	10	offer	offer	VERB
fcis-30157	118	11	distinct	distinct	ADJ
fcis-30157	118	12	advantages	advantage	NOUN
fcis-30157	118	13	and	and	CCONJ
fcis-30157	118	14	challenges	challenge	NOUN
fcis-30157	118	15	.	.	PUNCT
fcis-30157	119	1	knowledge	knowledge	NOUN
fcis-30157	119	2	-	-	PUNCT
fcis-30157	119	3	based	base	VERB
fcis-30157	119	4	systems	system	NOUN
fcis-30157	119	5	are	be	AUX
fcis-30157	119	6	interpretable	interpretable	ADJ
fcis-30157	119	7	but	but	CCONJ
fcis-30157	119	8	struggle	struggle	VERB
fcis-30157	119	9	with	with	ADP
fcis-30157	119	10	updates	update	NOUN
fcis-30157	119	11	and	and	CCONJ
fcis-30157	119	12	flexibility	flexibility	NOUN
fcis-30157	119	13	,	,	PUNCT
fcis-30157	119	14	while	while	SCONJ
fcis-30157	119	15	data	data	NOUN
fcis-30157	119	16	-	-	PUNCT
fcis-30157	119	17	driven	drive	VERB
fcis-30157	119	18	models	model	NOUN
fcis-30157	119	19	handle	handle	VERB
fcis-30157	119	20	complex	complex	ADJ
fcis-30157	119	21	patterns	pattern	NOUN
fcis-30157	119	22	but	but	CCONJ
fcis-30157	119	23	require	require	VERB
fcis-30157	119	24	large	large	ADJ
fcis-30157	119	25	datasets	dataset	NOUN
fcis-30157	119	26	and	and	CCONJ
fcis-30157	119	27	can	can	AUX
fcis-30157	119	28	lack	lack	VERB
fcis-30157	119	29	transparency	transparency	NOUN
fcis-30157	119	30	.	.	PUNCT
fcis-30157	120	1	future	future	ADJ
fcis-30157	120	2	research	research	NOUN
fcis-30157	120	3	should	should	AUX
fcis-30157	120	4	focus	focus	VERB
fcis-30157	120	5	on	on	ADP
fcis-30157	120	6	integrating	integrate	VERB
fcis-30157	120	7	these	these	DET
fcis-30157	120	8	approaches	approach	NOUN
fcis-30157	120	9	to	to	PART
fcis-30157	120	10	create	create	VERB
fcis-30157	120	11	more	more	ADV
fcis-30157	120	12	robust	robust	ADJ
fcis-30157	120	13	diagnostic	diagnostic	ADJ
fcis-30157	120	14	tools	tool	NOUN
fcis-30157	120	15	for	for	ADP
fcis-30157	120	16	skin	skin	NOUN
fcis-30157	120	17	diseases	disease	NOUN
fcis-30157	120	18	.	.	PUNCT
fcis-30157	121	1	1.3	1.3	NUM
fcis-30157	121	2	.	.	PUNCT
fcis-30157	121	3	applications	application	NOUN
fcis-30157	121	4	in	in	ADP
fcis-30157	121	5	skin	skin	NOUN
fcis-30157	121	6	diseases	disease	NOUN
fcis-30157	121	7	1.3.1	1.3.1	NUM
fcis-30157	121	8	.	.	PUNCT
fcis-30157	122	1	existing	exist	VERB
fcis-30157	122	2	ai	ai	PROPN
fcis-30157	122	3	systems	system	NOUN
fcis-30157	122	4	for	for	ADP
fcis-30157	122	5	skin	skin	NOUN
fcis-30157	122	6	disease	disease	NOUN
fcis-30157	122	7	diagnosis	diagnosis	NOUN
fcis-30157	122	8	recent	recent	ADJ
fcis-30157	122	9	ai	ai	VERB
fcis-30157	122	10	systems	system	NOUN
fcis-30157	122	11	for	for	ADP
fcis-30157	122	12	skin	skin	NOUN
fcis-30157	122	13	disease	disease	NOUN
fcis-30157	122	14	diagnosis	diagnosis	NOUN
fcis-30157	122	15	have	have	AUX
fcis-30157	122	16	gained	gain	VERB
fcis-30157	122	17	attention	attention	NOUN
fcis-30157	122	18	,	,	PUNCT
fcis-30157	122	19	particularly	particularly	ADV
fcis-30157	122	20	those	those	PRON
fcis-30157	122	21	utilizing	utilize	VERB
fcis-30157	122	22	deep	deep	ADJ
fcis-30157	122	23	learning	learning	NOUN
fcis-30157	122	24	to	to	PART
fcis-30157	122	25	assist	assist	VERB
fcis-30157	122	26	in	in	ADP
fcis-30157	122	27	diagnosing	diagnose	VERB
fcis-30157	122	28	various	various	ADJ
fcis-30157	122	29	skin	skin	NOUN
fcis-30157	122	30	conditions	condition	NOUN
fcis-30157	122	31	.	.	PUNCT
fcis-30157	123	1	key	key	ADJ
fcis-30157	123	2	systems	system	NOUN
fcis-30157	123	3	include	include	VERB
fcis-30157	123	4	the	the	DET
fcis-30157	123	5	isic	isic	PROPN
fcis-30157	123	6	grand	grand	ADJ
fcis-30157	123	7	challenge	challenge	NOUN
fcis-30157	123	8	,	,	PUNCT
fcis-30157	123	9	which	which	PRON
fcis-30157	123	10	focuses	focus	VERB
fcis-30157	123	11	on	on	ADP
fcis-30157	123	12	melanoma	melanoma	NOUN
fcis-30157	123	13	detection	detection	NOUN
fcis-30157	123	14	using	use	VERB
fcis-30157	123	15	dermoscopy	dermoscopy	NOUN
fcis-30157	123	16	images	image	NOUN
fcis-30157	123	17	.	.	PUNCT
fcis-30157	124	1	ai	ai	VERB
fcis-30157	124	2	models	model	NOUN
fcis-30157	124	3	from	from	ADP
fcis-30157	124	4	the	the	DET
fcis-30157	124	5	challenge	challenge	NOUN
fcis-30157	124	6	have	have	AUX
fcis-30157	124	7	achieved	achieve	VERB
fcis-30157	124	8	diagnostic	diagnostic	ADJ
fcis-30157	124	9	accuracy	accuracy	NOUN
fcis-30157	124	10	comparable	comparable	ADJ
fcis-30157	124	11	to	to	ADP
fcis-30157	124	12	board	board	NOUN
fcis-30157	124	13	-	-	PUNCT
fcis-30157	124	14	certified	certify	VERB
fcis-30157	124	15	dermatologists	dermatologist	NOUN
fcis-30157	124	16	,	,	PUNCT
fcis-30157	124	17	improving	improve	VERB
fcis-30157	124	18	efficiency	efficiency	NOUN
fcis-30157	124	19	in	in	ADP
fcis-30157	124	20	resource	resource	NOUN
fcis-30157	124	21	-	-	PUNCT
fcis-30157	124	22	limited	limit	VERB
fcis-30157	124	23	settings	setting	NOUN
fcis-30157	124	24	.	.	PUNCT
fcis-30157	125	1	another	another	DET
fcis-30157	125	2	system	system	NOUN
fcis-30157	125	3	,	,	PUNCT
fcis-30157	125	4	mdfnet	mdfnet	NOUN
fcis-30157	125	5	,	,	PUNCT
fcis-30157	125	6	integrates	integrate	VERB
fcis-30157	125	7	skin	skin	NOUN
fcis-30157	125	8	images	image	NOUN
fcis-30157	125	9	and	and	CCONJ
fcis-30157	125	10	clinical	clinical	ADJ
fcis-30157	125	11	data	datum	NOUN
fcis-30157	125	12	,	,	PUNCT
fcis-30157	125	13	achieving	achieve	VERB
fcis-30157	125	14	92.3	92.3	NUM
fcis-30157	125	15	%	%	NOUN
fcis-30157	125	16	accuracy	accuracy	NOUN
fcis-30157	125	17	in	in	ADP
fcis-30157	125	18	distinguishing	distinguish	VERB
fcis-30157	125	19	between	between	ADP
fcis-30157	125	20	benign	benign	ADJ
fcis-30157	125	21	and	and	CCONJ
fcis-30157	125	22	malignant	malignant	ADJ
fcis-30157	125	23	lesions	lesion	NOUN
fcis-30157	125	24	,	,	PUNCT
fcis-30157	125	25	making	make	VERB
fcis-30157	125	26	it	it	PRON
fcis-30157	125	27	valuable	valuable	ADJ
fcis-30157	125	28	in	in	ADP
fcis-30157	125	29	primary	primary	ADJ
fcis-30157	125	30	care	care	NOUN
fcis-30157	125	31	settings	setting	NOUN
fcis-30157	125	32	[	[	X
fcis-30157	125	33	8	8	NUM
fcis-30157	125	34	]	]	PUNCT
fcis-30157	125	35	.	.	PUNCT
fcis-30157	126	1	deep	deep	ADJ
fcis-30157	126	2	learning	learning	NOUN
fcis-30157	126	3	models	model	NOUN
fcis-30157	126	4	like	like	ADP
fcis-30157	126	5	mobilenet	mobilenet	NOUN
fcis-30157	126	6	and	and	CCONJ
fcis-30157	126	7	resnet	resnet	NOUN
fcis-30157	126	8	have	have	AUX
fcis-30157	126	9	been	be	AUX
fcis-30157	126	10	adapted	adapt	VERB
fcis-30157	126	11	for	for	ADP
fcis-30157	126	12	skin	skin	NOUN
fcis-30157	126	13	lesion	lesion	NOUN
fcis-30157	126	14	classification	classification	NOUN
fcis-30157	126	15	,	,	PUNCT
fcis-30157	126	16	with	with	ADP
fcis-30157	126	17	an	an	DET
fcis-30157	126	18	enhanced	enhanced	ADJ
fcis-30157	126	19	mobilenet	mobilenet	NOUN
fcis-30157	126	20	achieving	achieve	VERB
fcis-30157	126	21	89.7	89.7	NUM
fcis-30157	126	22	%	%	NOUN
fcis-30157	126	23	accuracy	accuracy	NOUN
fcis-30157	126	24	for	for	ADP
fcis-30157	126	25	actinic	actinic	ADJ
fcis-30157	126	26	keratosis	keratosis	NOUN
fcis-30157	126	27	[	[	X
fcis-30157	126	28	27	27	NUM
fcis-30157	126	29	]	]	PUNCT
fcis-30157	126	30	.	.	PUNCT
fcis-30157	127	1	a	a	DET
fcis-30157	127	2	few	few	ADJ
fcis-30157	127	3	-	-	PUNCT
fcis-30157	127	4	shot	shot	NOUN
fcis-30157	127	5	learning	learning	NOUN
fcis-30157	127	6	approach	approach	NOUN
fcis-30157	127	7	,	,	PUNCT
fcis-30157	127	8	the	the	DET
fcis-30157	127	9	fs3dciot	fs3dciot	NOUN
fcis-30157	127	10	network	network	NOUN
fcis-30157	127	11	,	,	PUNCT
fcis-30157	127	12	has	have	AUX
fcis-30157	127	13	been	be	AUX
fcis-30157	127	14	successful	successful	ADJ
fcis-30157	127	15	in	in	ADP
fcis-30157	127	16	identifying	identify	VERB
fcis-30157	127	17	rare	rare	ADJ
fcis-30157	127	18	skin	skin	NOUN
fcis-30157	127	19	conditions	condition	NOUN
fcis-30157	127	20	with	with	ADP
fcis-30157	127	21	85.6	85.6	NUM
fcis-30157	127	22	%	%	NOUN
fcis-30157	127	23	accuracy	accuracy	NOUN
fcis-30157	127	24	,	,	PUNCT
fcis-30157	127	25	beneficial	beneficial	ADJ
fcis-30157	127	26	in	in	ADP
fcis-30157	127	27	regions	region	NOUN
fcis-30157	127	28	with	with	ADP
fcis-30157	127	29	limited	limited	ADJ
fcis-30157	127	30	dermatological	dermatological	ADJ
fcis-30157	127	31	care	care	NOUN
fcis-30157	128	1	[	[	X
fcis-30157	128	2	9	9	NUM
fcis-30157	128	3	]	]	PUNCT
fcis-30157	128	4	.	.	PUNCT
fcis-30157	129	1	additionally	additionally	ADV
fcis-30157	129	2	,	,	PUNCT
fcis-30157	129	3	the	the	DET
fcis-30157	129	4	dual	dual	ADJ
fcis-30157	129	5	attention	attention	NOUN
fcis-30157	129	6	-	-	PUNCT
fcis-30157	129	7	based	base	VERB
fcis-30157	129	8	network	network	NOUN
fcis-30157	129	9	for	for	ADP
fcis-30157	129	10	skin	skin	NOUN
fcis-30157	129	11	lesion	lesion	NOUN
fcis-30157	129	12	classification	classification	NOUN
fcis-30157	129	13	outperforms	outperform	VERB
fcis-30157	129	14	traditional	traditional	ADJ
fcis-30157	129	15	models	model	NOUN
fcis-30157	129	16	,	,	PUNCT
fcis-30157	129	17	enhancing	enhance	VERB
fcis-30157	129	18	both	both	DET
fcis-30157	129	19	accuracy	accuracy	NOUN
fcis-30157	129	20	and	and	CCONJ
fcis-30157	129	21	interpretability	interpretability	NOUN
fcis-30157	129	22	by	by	ADP
fcis-30157	129	23	focusing	focus	VERB
fcis-30157	129	24	on	on	ADP
fcis-30157	129	25	key	key	ADJ
fcis-30157	129	26	features	feature	NOUN
fcis-30157	129	27	in	in	ADP
fcis-30157	129	28	skin	skin	NOUN
fcis-30157	129	29	images	image	NOUN
fcis-30157	129	30	[	[	X
fcis-30157	129	31	7	7	NUM
fcis-30157	129	32	]	]	PUNCT
fcis-30157	129	33	.	.	PUNCT
fcis-30157	130	1	the	the	DET
fcis-30157	130	2	deep	deep	ADJ
fcis-30157	130	3	learning	learning	NOUN
fcis-30157	130	4	-	-	PUNCT
fcis-30157	130	5	aided	aid	VERB
fcis-30157	130	6	decision	decision	NOUN
fcis-30157	130	7	support	support	NOUN
fcis-30157	130	8	system	system	NOUN
fcis-30157	130	9	for	for	ADP
fcis-30157	130	10	skin	skin	NOUN
fcis-30157	130	11	disease	disease	NOUN
fcis-30157	130	12	diagnosis	diagnosis	NOUN
fcis-30157	130	13	across	across	ADP
fcis-30157	130	14	skin	skin	NOUN
fcis-30157	130	15	tones	tone	NOUN
fcis-30157	130	16	addresses	address	NOUN
fcis-30157	130	17	the	the	DET
fcis-30157	130	18	challenge	challenge	NOUN
fcis-30157	130	19	of	of	ADP
fcis-30157	130	20	diagnosing	diagnose	VERB
fcis-30157	130	21	across	across	ADP
fcis-30157	130	22	various	various	ADJ
fcis-30157	130	23	skin	skin	NOUN
fcis-30157	130	24	types	type	NOUN
fcis-30157	130	25	,	,	PUNCT
fcis-30157	130	26	achieving	achieve	VERB
fcis-30157	130	27	91.2	91.2	NUM
fcis-30157	130	28	%	%	NOUN
fcis-30157	130	29	accuracy	accuracy	NOUN
fcis-30157	130	30	,	,	PUNCT
fcis-30157	130	31	a	a	DET
fcis-30157	130	32	significant	significant	ADJ
fcis-30157	130	33	step	step	NOUN
fcis-30157	130	34	toward	toward	ADP
fcis-30157	130	35	reducing	reduce	VERB
fcis-30157	130	36	health	health	NOUN
fcis-30157	130	37	disparities	disparity	NOUN
fcis-30157	130	38	.	.	PUNCT
fcis-30157	131	1	commercial	commercial	ADJ
fcis-30157	131	2	ai	ai	PROPN
fcis-30157	131	3	tools	tool	NOUN
fcis-30157	131	4	like	like	ADP
fcis-30157	131	5	skinvision	skinvision	NOUN
fcis-30157	131	6	and	and	CCONJ
fcis-30157	131	7	dermassist	dermassist	NOUN
fcis-30157	131	8	have	have	AUX
fcis-30157	131	9	also	also	ADV
fcis-30157	131	10	been	be	AUX
fcis-30157	131	11	developed	develop	VERB
fcis-30157	131	12	,	,	PUNCT
fcis-30157	131	13	offering	offer	VERB
fcis-30157	131	14	skin	skin	NOUN
fcis-30157	131	15	lesion	lesion	NOUN
fcis-30157	131	16	analysis	analysis	NOUN
fcis-30157	131	17	and	and	CCONJ
fcis-30157	131	18	remote	remote	ADJ
fcis-30157	131	19	diagnosis	diagnosis	NOUN
fcis-30157	131	20	with	with	ADP
fcis-30157	131	21	high	high	ADJ
fcis-30157	131	22	accuracy	accuracy	NOUN
fcis-30157	131	23	[	[	X
fcis-30157	131	24	19	19	NUM
fcis-30157	131	25	]	]	PUNCT
fcis-30157	131	26	.	.	PUNCT
fcis-30157	132	1	these	these	DET
fcis-30157	132	2	ai	ai	PROPN
fcis-30157	132	3	systems	system	NOUN
fcis-30157	132	4	are	be	AUX
fcis-30157	132	5	proving	prove	VERB
fcis-30157	132	6	effective	effective	ADJ
fcis-30157	132	7	in	in	ADP
fcis-30157	132	8	clinical	clinical	ADJ
fcis-30157	132	9	applications	application	NOUN
fcis-30157	132	10	,	,	PUNCT
fcis-30157	132	11	reducing	reduce	VERB
fcis-30157	132	12	dermatologist	dermatologist	NOUN
fcis-30157	132	13	workload	workload	NOUN
fcis-30157	132	14	and	and	CCONJ
fcis-30157	132	15	improving	improve	VERB
fcis-30157	132	16	diagnostic	diagnostic	ADJ
fcis-30157	132	17	outcomes	outcome	NOUN
fcis-30157	132	18	in	in	ADP
fcis-30157	132	19	both	both	DET
fcis-30157	132	20	primary	primary	ADJ
fcis-30157	132	21	care	care	NOUN
fcis-30157	132	22	and	and	CCONJ
fcis-30157	132	23	specialized	specialized	ADJ
fcis-30157	132	24	settings	setting	NOUN
fcis-30157	132	25	.	.	PUNCT
fcis-30157	133	1	however	however	ADV
fcis-30157	133	2	,	,	PUNCT
fcis-30157	133	3	challenges	challenge	NOUN
fcis-30157	133	4	remain	remain	VERB
fcis-30157	133	5	,	,	PUNCT
fcis-30157	133	6	including	include	VERB
fcis-30157	133	7	the	the	DET
fcis-30157	133	8	need	need	NOUN
fcis-30157	133	9	for	for	ADP
fcis-30157	133	10	high	high	ADJ
fcis-30157	133	11	-	-	PUNCT
fcis-30157	133	12	quality	quality	NOUN
fcis-30157	133	13	training	training	NOUN
fcis-30157	133	14	data	datum	NOUN
fcis-30157	133	15	,	,	PUNCT
fcis-30157	133	16	particularly	particularly	ADV
fcis-30157	133	17	for	for	ADP
fcis-30157	133	18	rare	rare	ADJ
fcis-30157	133	19	diseases	disease	NOUN
fcis-30157	133	20	,	,	PUNCT
fcis-30157	133	21	and	and	CCONJ
fcis-30157	133	22	concerns	concern	NOUN
fcis-30157	133	23	about	about	ADP
fcis-30157	133	24	the	the	DET
fcis-30157	133	25	interpretability	interpretability	NOUN
fcis-30157	133	26	of	of	ADP
fcis-30157	133	27	ai	ai	PROPN
fcis-30157	133	28	models	model	NOUN
fcis-30157	133	29	.	.	PUNCT
fcis-30157	134	1	future	future	ADJ
fcis-30157	134	2	research	research	NOUN
fcis-30157	134	3	should	should	AUX
fcis-30157	134	4	focus	focus	VERB
fcis-30157	134	5	on	on	ADP
fcis-30157	134	6	improving	improve	VERB
fcis-30157	134	7	data	datum	NOUN
fcis-30157	134	8	quality	quality	NOUN
fcis-30157	134	9	,	,	PUNCT
fcis-30157	134	10	explainability	explainability	NOUN
fcis-30157	134	11	,	,	PUNCT
fcis-30157	134	12	and	and	CCONJ
fcis-30157	134	13	accessibility	accessibility	NOUN
fcis-30157	134	14	to	to	PART
fcis-30157	134	15	enhance	enhance	VERB
fcis-30157	134	16	the	the	DET
fcis-30157	134	17	clinical	clinical	ADJ
fcis-30157	134	18	impact	impact	NOUN
fcis-30157	134	19	of	of	ADP
fcis-30157	134	20	ai	ai	NOUN
fcis-30157	134	21	in	in	ADP
fcis-30157	134	22	dermatology	dermatology	NOUN
fcis-30157	134	23	.	.	PUNCT
fcis-30157	135	1	1.3.2	1.3.2	X
fcis-30157	135	2	.	.	PUNCT
fcis-30157	135	3	performance	performance	NOUN
fcis-30157	135	4	evaluation	evaluation	NOUN
fcis-30157	135	5	of	of	ADP
fcis-30157	135	6	related	related	ADJ
fcis-30157	135	7	works	work	NOUN
fcis-30157	135	8	performance	performance	NOUN
fcis-30157	135	9	evaluation	evaluation	NOUN
fcis-30157	135	10	of	of	ADP
fcis-30157	135	11	ai	ai	PROPN
fcis-30157	135	12	systems	system	NOUN
fcis-30157	135	13	for	for	ADP
fcis-30157	135	14	skin	skin	NOUN
fcis-30157	135	15	disease	disease	NOUN
fcis-30157	135	16	diagnosis	diagnosis	NOUN
fcis-30157	135	17	shows	show	VERB
fcis-30157	135	18	significant	significant	ADJ
fcis-30157	135	19	variation	variation	NOUN
fcis-30157	135	20	due	due	ADP
fcis-30157	135	21	to	to	ADP
fcis-30157	135	22	multiple	multiple	ADJ
fcis-30157	135	23	factors	factor	NOUN
fcis-30157	135	24	,	,	PUNCT
fcis-30157	135	25	including	include	VERB
fcis-30157	135	26	evaluation	evaluation	NOUN
fcis-30157	135	27	metrics	metric	NOUN
fcis-30157	135	28	,	,	PUNCT
fcis-30157	135	29	dataset	dataset	NOUN
fcis-30157	135	30	characteristics	characteristic	NOUN
fcis-30157	135	31	,	,	PUNCT
fcis-30157	135	32	and	and	CCONJ
fcis-30157	135	33	model	model	NOUN
fcis-30157	135	34	architecture	architecture	NOUN
fcis-30157	135	35	.	.	PUNCT
fcis-30157	136	1	accuracy	accuracy	NOUN
fcis-30157	136	2	in	in	ADP
fcis-30157	136	3	recent	recent	ADJ
fcis-30157	136	4	dermatoscopic	dermatoscopic	NOUN
fcis-30157	136	5	image	image	NOUN
fcis-30157	136	6	classification	classification	NOUN
fcis-30157	136	7	ranges	range	VERB
fcis-30157	136	8	from	from	ADP
fcis-30157	136	9	85	85	NUM
fcis-30157	136	10	-	-	SYM
fcis-30157	136	11	95	95	NUM
fcis-30157	136	12	%	%	NOUN
fcis-30157	136	13	,	,	PUNCT
fcis-30157	136	14	with	with	ADP
fcis-30157	136	15	models	model	NOUN
fcis-30157	136	16	using	use	VERB
fcis-30157	136	17	clinical	clinical	ADJ
fcis-30157	136	18	metadata	metadata	NOUN
fcis-30157	136	19	achieving	achieve	VERB
fcis-30157	136	20	higher	high	ADJ
fcis-30157	136	21	accuracy	accuracy	NOUN
fcis-30157	136	22	(	(	PUNCT
fcis-30157	136	23	89.1	89.1	NUM
fcis-30157	136	24	%	%	NOUN
fcis-30157	136	25	)	)	PUNCT
fcis-30157	136	26	compared	compare	VERB
fcis-30157	136	27	to	to	ADP
fcis-30157	136	28	those	those	PRON
fcis-30157	136	29	using	use	VERB
fcis-30157	136	30	dermoscopic	dermoscopic	ADJ
fcis-30157	136	31	images	image	NOUN
fcis-30157	136	32	alone	alone	ADV
fcis-30157	136	33	(	(	PUNCT
fcis-30157	136	34	83.2	83.2	NUM
fcis-30157	136	35	%	%	NOUN
fcis-30157	136	36	)	)	PUNCT
fcis-30157	137	1	[	[	X
fcis-30157	137	2	23	23	NUM
fcis-30157	137	3	]	]	PUNCT
fcis-30157	137	4	.	.	PUNCT
fcis-30157	138	1	precision	precision	NOUN
fcis-30157	138	2	and	and	CCONJ
fcis-30157	138	3	recall	recall	NOUN
fcis-30157	138	4	rates	rate	NOUN
fcis-30157	138	5	vary	vary	VERB
fcis-30157	138	6	greatly	greatly	ADV
fcis-30157	138	7	,	,	PUNCT
fcis-30157	138	8	with	with	ADP
fcis-30157	138	9	precision	precision	NOUN
fcis-30157	138	10	for	for	ADP
fcis-30157	138	11	rare	rare	ADJ
fcis-30157	138	12	conditions	condition	NOUN
fcis-30157	138	13	like	like	ADP
fcis-30157	138	14	merkel	merkel	NOUN
fcis-30157	138	15	cell	cell	NOUN
fcis-30157	138	16	carcinoma	carcinoma	NOUN
fcis-30157	138	17	at	at	ADP
fcis-30157	138	18	78	78	NUM
fcis-30157	138	19	%	%	NOUN
fcis-30157	138	20	,	,	PUNCT
fcis-30157	138	21	while	while	SCONJ
fcis-30157	138	22	common	common	ADJ
fcis-30157	138	23	melanomas	melanoma	NOUN
fcis-30157	138	24	achieve	achieve	VERB
fcis-30157	138	25	93	93	NUM
fcis-30157	138	26	%	%	NOUN
fcis-30157	138	27	in	in	ADP
fcis-30157	138	28	multimodal	multimodal	ADJ
fcis-30157	138	29	fusion	fusion	NOUN
fcis-30157	138	30	approaches	approach	NOUN
fcis-30157	138	31	.	.	PUNCT
fcis-30157	139	1	sensitivity	sensitivity	NOUN
fcis-30157	139	2	for	for	ADP
fcis-30157	139	3	malignant	malignant	ADJ
fcis-30157	139	4	cases	case	NOUN
fcis-30157	139	5	typically	typically	ADV
fcis-30157	139	6	ranges	range	VERB
fcis-30157	139	7	from	from	ADP
fcis-30157	139	8	86	86	NUM
fcis-30157	139	9	-	-	SYM
fcis-30157	139	10	91	91	NUM
fcis-30157	139	11	%	%	NOUN
fcis-30157	139	12	,	,	PUNCT
fcis-30157	139	13	and	and	CCONJ
fcis-30157	139	14	for	for	ADP
fcis-30157	139	15	benign	benign	ADJ
fcis-30157	139	16	lesions	lesion	NOUN
fcis-30157	139	17	,	,	PUNCT
fcis-30157	139	18	72	72	NUM
fcis-30157	139	19	-	-	SYM
fcis-30157	139	20	84	84	NUM
fcis-30157	139	21	%	%	NOUN
fcis-30157	139	22	in	in	ADP
fcis-30157	139	23	imbalanced	imbalanced	ADJ
fcis-30157	139	24	datasets	dataset	NOUN
fcis-30157	139	25	.	.	PUNCT
fcis-30157	140	1	the	the	DET
fcis-30157	140	2	performance	performance	NOUN
fcis-30157	140	3	of	of	ADP
fcis-30157	140	4	ai	ai	PROPN
fcis-30157	140	5	systems	system	NOUN
fcis-30157	140	6	for	for	ADP
fcis-30157	140	7	skin	skin	NOUN
fcis-30157	140	8	disease	disease	NOUN
fcis-30157	140	9	diagnosis	diagnosis	NOUN
fcis-30157	140	10	varies	vary	VERB
fcis-30157	140	11	significantly	significantly	ADV
fcis-30157	140	12	due	due	ADJ
fcis-30157	140	13	to	to	ADP
fcis-30157	140	14	factors	factor	NOUN
fcis-30157	140	15	like	like	ADP
fcis-30157	140	16	evaluation	evaluation	NOUN
fcis-30157	140	17	metrics	metric	NOUN
fcis-30157	140	18	,	,	PUNCT
fcis-30157	140	19	dataset	dataset	NOUN
fcis-30157	140	20	characteristics	characteristic	NOUN
fcis-30157	140	21	,	,	PUNCT
fcis-30157	140	22	model	model	NOUN
fcis-30157	140	23	architecture	architecture	NOUN
fcis-30157	140	24	,	,	PUNCT
fcis-30157	140	25	and	and	CCONJ
fcis-30157	140	26	clinical	clinical	ADJ
fcis-30157	140	27	implementation	implementation	NOUN
fcis-30157	140	28	.	.	PUNCT
fcis-30157	141	1	accuracy	accuracy	NOUN
fcis-30157	141	2	typically	typically	ADV
fcis-30157	141	3	ranges	range	VERB
fcis-30157	141	4	between	between	ADP
fcis-30157	141	5	85	85	NUM
fcis-30157	141	6	-	-	SYM
fcis-30157	141	7	95	95	NUM
fcis-30157	141	8	%	%	NOUN
fcis-30157	141	9	in	in	ADP
fcis-30157	141	10	dermatoscopic	dermatoscopic	ADJ
fcis-30157	141	11	image	image	NOUN
fcis-30157	141	12	classification	classification	NOUN
fcis-30157	141	13	,	,	PUNCT
fcis-30157	141	14	with	with	ADP
fcis-30157	141	15	models	model	NOUN
fcis-30157	141	16	trained	train	VERB
fcis-30157	141	17	on	on	ADP
fcis-30157	141	18	dermoscopic	dermoscopic	ADJ
fcis-30157	141	19	images	image	NOUN
fcis-30157	141	20	achieving	achieve	VERB
fcis-30157	141	21	83.2	83.2	NUM
fcis-30157	141	22	%	%	NOUN
fcis-30157	141	23	average	average	ADJ
fcis-30157	141	24	accuracy	accuracy	NOUN
fcis-30157	141	25	,	,	PUNCT
fcis-30157	141	26	while	while	SCONJ
fcis-30157	141	27	those	those	PRON
fcis-30157	141	28	incorporating	incorporate	VERB
fcis-30157	141	29	clinical	clinical	ADJ
fcis-30157	141	30	metadata	metadata	NOUN
fcis-30157	141	31	reach	reach	VERB
fcis-30157	141	32	89.1	89.1	NUM
fcis-30157	141	33	%	%	NOUN
fcis-30157	141	34	.	.	PUNCT
fcis-30157	142	1	precision	precision	NOUN
fcis-30157	142	2	varies	vary	VERB
fcis-30157	142	3	from	from	ADP
fcis-30157	142	4	78	78	NUM
fcis-30157	142	5	%	%	NOUN
fcis-30157	142	6	for	for	ADP
fcis-30157	142	7	rare	rare	ADJ
fcis-30157	142	8	conditions	condition	NOUN
fcis-30157	142	9	like	like	ADP
fcis-30157	142	10	merkel	merkel	NOUN
fcis-30157	142	11	cell	cell	NOUN
fcis-30157	142	12	carcinoma	carcinoma	NOUN
fcis-30157	142	13	to	to	ADP
fcis-30157	142	14	93	93	NUM
fcis-30157	142	15	%	%	NOUN
fcis-30157	142	16	for	for	ADP
fcis-30157	142	17	common	common	ADJ
fcis-30157	142	18	melanomas	melanoma	NOUN
fcis-30157	142	19	.	.	PUNCT
fcis-30157	143	1	recall	recall	NOUN
fcis-30157	143	2	rates	rate	NOUN
fcis-30157	143	3	show	show	VERB
fcis-30157	143	4	sensitivity	sensitivity	NOUN
fcis-30157	143	5	of	of	ADP
fcis-30157	143	6	86	86	NUM
fcis-30157	143	7	-	-	SYM
fcis-30157	143	8	91	91	NUM
fcis-30157	143	9	%	%	NOUN
fcis-30157	143	10	for	for	ADP
fcis-30157	143	11	malignant	malignant	ADJ
fcis-30157	143	12	cases	case	NOUN
fcis-30157	143	13	but	but	CCONJ
fcis-30157	143	14	72	72	NUM
fcis-30157	143	15	-	-	SYM
fcis-30157	143	16	84	84	NUM
fcis-30157	143	17	%	%	NOUN
fcis-30157	143	18	for	for	ADP
fcis-30157	143	19	benign	benign	ADJ
fcis-30157	143	20	lesions	lesion	NOUN
fcis-30157	143	21	in	in	ADP
fcis-30157	143	22	imbalanced	imbalanced	ADJ
fcis-30157	143	23	datasets	dataset	NOUN
fcis-30157	143	24	.	.	PUNCT
fcis-30157	144	1	dataset	dataset	VERB
fcis-30157	144	2	size	size	NOUN
fcis-30157	144	3	and	and	CCONJ
fcis-30157	144	4	characteristics	characteristic	NOUN
fcis-30157	144	5	significantly	significantly	ADV
fcis-30157	144	6	affect	affect	VERB
fcis-30157	144	7	performance	performance	NOUN
fcis-30157	144	8	.	.	PUNCT
fcis-30157	145	1	for	for	ADP
fcis-30157	145	2	example	example	NOUN
fcis-30157	145	3	,	,	PUNCT
fcis-30157	145	4	mdfnet	mdfnet	NOUN
fcis-30157	145	5	,	,	PUNCT
fcis-30157	145	6	which	which	PRON
fcis-30157	145	7	integrates	integrate	VERB
fcis-30157	145	8	clinical	clinical	ADJ
fcis-30157	145	9	and	and	CCONJ
fcis-30157	145	10	visual	visual	ADJ
fcis-30157	145	11	data	datum	NOUN
fcis-30157	145	12	,	,	PUNCT
fcis-30157	145	13	achieves	achieve	VERB
fcis-30157	145	14	93.7	93.7	NUM
fcis-30157	145	15	%	%	NOUN
fcis-30157	145	16	accuracy	accuracy	NOUN
fcis-30157	145	17	.	.	PUNCT
fcis-30157	146	1	larger	large	ADJ
fcis-30157	146	2	datasets	dataset	NOUN
fcis-30157	146	3	,	,	PUNCT
fcis-30157	146	4	such	such	ADJ
fcis-30157	146	5	as	as	ADP
fcis-30157	146	6	those	those	PRON
fcis-30157	146	7	with	with	ADP
fcis-30157	146	8	10,000	10,000	NUM
fcis-30157	146	9	images	image	NOUN
fcis-30157	146	10	,	,	PUNCT
fcis-30157	146	11	show	show	VERB
fcis-30157	146	12	a	a	DET
fcis-30157	146	13	14.3	14.3	NUM
fcis-30157	146	14	%	%	NOUN
fcis-30157	146	15	improvement	improvement	NOUN
fcis-30157	146	16	in	in	ADP
fcis-30157	146	17	accuracy	accuracy	NOUN
fcis-30157	146	18	compared	compare	VERB
fcis-30157	146	19	to	to	ADP
fcis-30157	146	20	smaller	small	ADJ
fcis-30157	146	21	ones	one	NOUN
fcis-30157	146	22	.	.	PUNCT
fcis-30157	147	1	models	model	NOUN
fcis-30157	147	2	like	like	ADP
fcis-30157	147	3	resnet-152	resnet-152	NOUN
fcis-30157	147	4	and	and	CCONJ
fcis-30157	147	5	mobilenet	mobilenet	NOUN
fcis-30157	147	6	-	-	PUNCT
fcis-30157	147	7	v2	v2	NOUN
fcis-30157	147	8	have	have	VERB
fcis-30157	147	9	distinct	distinct	ADJ
fcis-30157	147	10	performance	performance	NOUN
fcis-30157	147	11	trade	trade	NOUN
fcis-30157	147	12	-	-	PUNCT
fcis-30157	147	13	offs	off	NOUN
fcis-30157	147	14	,	,	PUNCT
fcis-30157	147	15	with	with	ADP
fcis-30157	147	16	resnet	resnet	NOUN
fcis-30157	147	17	achieving	achieve	VERB
fcis-30157	147	18	91.2	91.2	NUM
fcis-30157	147	19	%	%	NOUN
fcis-30157	147	20	accuracy	accuracy	NOUN
fcis-30157	147	21	on	on	ADP
fcis-30157	147	22	multi	multi	ADJ
fcis-30157	147	23	-	-	ADJ
fcis-30157	147	24	class	class	ADJ
fcis-30157	147	25	tasks	task	NOUN
fcis-30157	147	26	and	and	CCONJ
fcis-30157	147	27	mobilenet	mobilenet	NOUN
fcis-30157	147	28	maintaining	maintain	VERB
fcis-30157	147	29	87.4	87.4	NUM
fcis-30157	147	30	%	%	NOUN
fcis-30157	147	31	accuracy	accuracy	NOUN
fcis-30157	147	32	with	with	ADP
fcis-30157	147	33	fewer	few	ADJ
fcis-30157	147	34	parameters	parameter	NOUN
fcis-30157	147	35	.	.	PUNCT
fcis-30157	148	1	transformer	transformer	NOUN
fcis-30157	148	2	-	-	PUNCT
fcis-30157	148	3	based	base	VERB
fcis-30157	148	4	models	model	NOUN
fcis-30157	148	5	excel	excel	VERB
fcis-30157	148	6	at	at	ADP
fcis-30157	148	7	detecting	detect	VERB
fcis-30157	148	8	irregular	irregular	ADJ
fcis-30157	148	9	borders	border	NOUN
fcis-30157	148	10	,	,	PUNCT
fcis-30157	148	11	reaching	reach	VERB
fcis-30157	148	12	92.4	92.4	NUM
fcis-30157	148	13	%	%	NOUN
fcis-30157	148	14	accuracy	accuracy	NOUN
fcis-30157	148	15	.	.	PUNCT
fcis-30157	149	1	experimental	experimental	ADJ
fcis-30157	149	2	methodologies	methodology	NOUN
fcis-30157	149	3	,	,	PUNCT
fcis-30157	149	4	like	like	ADP
fcis-30157	149	5	cross	cross	ADJ
fcis-30157	149	6	-	-	ADJ
fcis-30157	149	7	validation	validation	ADJ
fcis-30157	149	8	and	and	CCONJ
fcis-30157	149	9	transfer	transfer	NOUN
fcis-30157	149	10	learning	learning	NOUN
fcis-30157	149	11	,	,	PUNCT
fcis-30157	149	12	impact	impact	NOUN
fcis-30157	149	13	performance	performance	NOUN
fcis-30157	149	14	claims	claim	NOUN
fcis-30157	149	15	.	.	PUNCT
fcis-30157	150	1	five	five	NUM
fcis-30157	150	2	-	-	ADJ
fcis-30157	150	3	fold	fold	ADJ
fcis-30157	150	4	cross	cross	ADJ
fcis-30157	150	5	-	-	ADJ
fcis-30157	150	6	validation	validation	ADJ
fcis-30157	150	7	tends	tend	VERB
fcis-30157	150	8	to	to	PART
fcis-30157	150	9	overestimate	overestimate	VERB
fcis-30157	150	10	real	real	ADJ
fcis-30157	150	11	-	-	PUNCT
fcis-30157	150	12	world	world	NOUN
fcis-30157	150	13	performance	performance	NOUN
fcis-30157	150	14	,	,	PUNCT
fcis-30157	150	15	while	while	SCONJ
fcis-30157	150	16	imagenet	imagenet	NOUN
fcis-30157	150	17	pretraining	pretraine	VERB
fcis-30157	150	18	boosts	boost	NOUN
fcis-30157	150	19	accuracy	accuracy	NOUN
fcis-30157	150	20	for	for	ADP
fcis-30157	150	21	common	common	ADJ
fcis-30157	150	22	conditions	condition	NOUN
fcis-30157	150	23	but	but	CCONJ
fcis-30157	150	24	less	less	ADV
fcis-30157	150	25	so	so	ADV
fcis-30157	150	26	for	for	ADP
fcis-30157	150	27	rare	rare	ADJ
fcis-30157	150	28	ones	one	NOUN
fcis-30157	150	29	.	.	PUNCT
fcis-30157	151	1	clinical	clinical	ADJ
fcis-30157	151	2	implementation	implementation	NOUN
fcis-30157	151	3	faces	face	VERB
fcis-30157	151	4	challenges	challenge	NOUN
fcis-30157	151	5	,	,	PUNCT
fcis-30157	151	6	including	include	VERB
fcis-30157	151	7	inference	inference	NOUN
fcis-30157	151	8	speed	speed	NOUN
fcis-30157	151	9	(	(	PUNCT
fcis-30157	151	10	23ms	23ms	NOUN
fcis-30157	151	11	for	for	ADP
fcis-30157	151	12	mobilenet	mobilenet	NOUN
fcis-30157	151	13	,	,	PUNCT
fcis-30157	151	14	680ms	680ms	NOUN
fcis-30157	151	15	for	for	ADP
fcis-30157	151	16	multimodal	multimodal	NOUN
fcis-30157	151	17	systems	system	NOUN
fcis-30157	151	18	)	)	PUNCT
fcis-30157	151	19	,	,	PUNCT
fcis-30157	151	20	memory	memory	NOUN
fcis-30157	151	21	limitations	limitation	NOUN
fcis-30157	151	22	(	(	PUNCT
fcis-30157	151	23	models	model	NOUN
fcis-30157	151	24	over	over	ADP
fcis-30157	151	25	50	50	NUM
fcis-30157	151	26	mb	mb	VERB
fcis-30157	151	27	see	see	VERB
fcis-30157	151	28	38	38	NUM
fcis-30157	151	29	%	%	NOUN
fcis-30157	151	30	lower	low	ADJ
fcis-30157	151	31	adoption	adoption	NOUN
fcis-30157	151	32	)	)	PUNCT
fcis-30157	151	33	,	,	PUNCT
fcis-30157	151	34	and	and	CCONJ
fcis-30157	151	35	interpretability	interpretability	NOUN
fcis-30157	151	36	(	(	PUNCT
fcis-30157	151	37	grad	grad	ADJ
fcis-30157	151	38	115	115	NUM
fcis-30157	151	39	cam	cam	NOUN
fcis-30157	151	40	explanations	explanation	NOUN
fcis-30157	151	41	add	add	VERB
fcis-30157	151	42	15	15	NUM
fcis-30157	151	43	-	-	SYM
fcis-30157	151	44	20	20	NUM
fcis-30157	151	45	%	%	NOUN
fcis-30157	151	46	processing	processing	NOUN
fcis-30157	151	47	time	time	NOUN
fcis-30157	151	48	but	but	CCONJ
fcis-30157	151	49	improve	improve	VERB
fcis-30157	151	50	clinician	clinician	NOUN
fcis-30157	151	51	trust	trust	NOUN
fcis-30157	151	52	by	by	ADP
fcis-30157	151	53	42	42	NUM
fcis-30157	151	54	%	%	NOUN
fcis-30157	151	55	)	)	PUNCT
fcis-30157	151	56	.	.	PUNCT
fcis-30157	152	1	cross	cross	ADJ
fcis-30157	152	2	-	-	ADJ
fcis-30157	152	3	study	study	ADJ
fcis-30157	152	4	comparisons	comparison	NOUN
fcis-30157	152	5	reveal	reveal	VERB
fcis-30157	152	6	gaps	gap	NOUN
fcis-30157	152	7	in	in	ADP
fcis-30157	152	8	racial	racial	ADJ
fcis-30157	152	9	diversity	diversity	NOUN
fcis-30157	152	10	(	(	PUNCT
fcis-30157	152	11	only	only	ADV
fcis-30157	152	12	23	23	NUM
fcis-30157	152	13	%	%	NOUN
fcis-30157	152	14	of	of	ADP
fcis-30157	152	15	studies	study	NOUN
fcis-30157	152	16	addressed	address	VERB
fcis-30157	152	17	it	it	PRON
fcis-30157	152	18	)	)	PUNCT
fcis-30157	152	19	and	and	CCONJ
fcis-30157	152	20	longitudinal	longitudinal	ADJ
fcis-30157	152	21	tracking	tracking	NOUN
fcis-30157	152	22	(	(	PUNCT
fcis-30157	152	23	8.2	8.2	NUM
fcis-30157	152	24	%	%	NOUN
fcis-30157	152	25	accuracy	accuracy	NOUN
fcis-30157	152	26	drop	drop	NOUN
fcis-30157	152	27	over	over	ADP
fcis-30157	152	28	six	six	NUM
fcis-30157	152	29	months	month	NOUN
fcis-30157	152	30	)	)	PUNCT
fcis-30157	152	31	.	.	PUNCT
fcis-30157	153	1	cost	cost	NOUN
fcis-30157	153	2	-	-	PUNCT
fcis-30157	153	3	effectiveness	effectiveness	NOUN
fcis-30157	153	4	is	be	AUX
fcis-30157	153	5	underexplored	underexplored	ADJ
fcis-30157	153	6	,	,	PUNCT
fcis-30157	153	7	though	though	SCONJ
fcis-30157	153	8	one	one	NUM
fcis-30157	153	9	study	study	NOUN
fcis-30157	153	10	reports	report	VERB
fcis-30157	153	11	$	$	SYM
fcis-30157	153	12	12.73	12.73	NUM
fcis-30157	153	13	per	per	ADP
fcis-30157	153	14	accurate	accurate	ADJ
fcis-30157	153	15	diagnosis	diagnosis	NOUN
fcis-30157	153	16	.	.	PUNCT
fcis-30157	154	1	specialized	specialized	ADJ
fcis-30157	154	2	evaluations	evaluation	NOUN
fcis-30157	154	3	,	,	PUNCT
fcis-30157	154	4	such	such	ADJ
fcis-30157	154	5	as	as	ADP
fcis-30157	154	6	ensemble	ensemble	ADJ
fcis-30157	154	7	models	model	NOUN
fcis-30157	154	8	,	,	PUNCT
fcis-30157	154	9	achieve	achieve	VERB
fcis-30157	154	10	high	high	ADJ
fcis-30157	154	11	accuracy	accuracy	NOUN
fcis-30157	154	12	but	but	CCONJ
fcis-30157	154	13	are	be	AUX
fcis-30157	154	14	limited	limit	VERB
fcis-30157	154	15	in	in	ADP
fcis-30157	154	16	real	real	ADJ
fcis-30157	154	17	-	-	PUNCT
fcis-30157	154	18	time	time	NOUN
fcis-30157	154	19	use	use	NOUN
fcis-30157	154	20	.	.	PUNCT
fcis-30157	155	1	single	single	ADJ
fcis-30157	155	2	-	-	PUNCT
fcis-30157	155	3	center	center	NOUN
fcis-30157	155	4	studies	study	NOUN
fcis-30157	155	5	tend	tend	VERB
fcis-30157	155	6	to	to	PART
fcis-30157	155	7	overestimate	overestimate	VERB
fcis-30157	155	8	performance	performance	NOUN
fcis-30157	155	9	by	by	ADP
fcis-30157	155	10	11	11	NUM
fcis-30157	155	11	-	-	SYM
fcis-30157	155	12	14	14	NUM
fcis-30157	155	13	%	%	NOUN
fcis-30157	155	14	compared	compare	VERB
fcis-30157	155	15	to	to	ADP
fcis-30157	155	16	multi	multi	ADJ
fcis-30157	155	17	-	-	ADJ
fcis-30157	155	18	center	center	ADJ
fcis-30157	155	19	trials	trial	NOUN
fcis-30157	155	20	,	,	PUNCT
fcis-30157	155	21	while	while	SCONJ
fcis-30157	155	22	ai	ai	AUX
fcis-30157	155	23	-	-	PUNCT
fcis-30157	155	24	assisted	assist	VERB
fcis-30157	155	25	diagnoses	diagnosis	NOUN
fcis-30157	155	26	have	have	AUX
fcis-30157	155	27	been	be	AUX
fcis-30157	155	28	linked	link	VERB
fcis-30157	155	29	to	to	ADP
fcis-30157	155	30	22	22	NUM
fcis-30157	155	31	%	%	NOUN
fcis-30157	155	32	faster	fast	ADJ
fcis-30157	155	33	treatment	treatment	NOUN
fcis-30157	155	34	initiation	initiation	NOUN
fcis-30157	155	35	.	.	PUNCT
fcis-30157	156	1	standardized	standardized	ADJ
fcis-30157	156	2	evaluation	evaluation	NOUN
fcis-30157	156	3	protocols	protocol	NOUN
fcis-30157	156	4	are	be	AUX
fcis-30157	156	5	needed	need	VERB
fcis-30157	156	6	,	,	PUNCT
fcis-30157	156	7	and	and	CCONJ
fcis-30157	156	8	the	the	DET
fcis-30157	156	9	emerging	emerge	VERB
fcis-30157	156	10	iso	iso	NOUN
fcis-30157	156	11	/	/	SYM
fcis-30157	156	12	iec	iec	NOUN
fcis-30157	156	13	24029	24029	NUM
fcis-30157	156	14	-	-	SYM
fcis-30157	156	15	2	2	NUM
fcis-30157	156	16	standard	standard	NOUN
fcis-30157	156	17	helps	help	VERB
fcis-30157	156	18	reduce	reduce	VERB
fcis-30157	156	19	performance	performance	NOUN
fcis-30157	156	20	discrepancies	discrepancy	NOUN
fcis-30157	156	21	by	by	ADP
fcis-30157	156	22	34	34	NUM
fcis-30157	156	23	%	%	NOUN
fcis-30157	156	24	.	.	PUNCT
fcis-30157	157	1	future	future	ADJ
fcis-30157	157	2	research	research	NOUN
fcis-30157	157	3	should	should	AUX
fcis-30157	157	4	focus	focus	VERB
fcis-30157	157	5	on	on	ADP
fcis-30157	157	6	developing	develop	VERB
fcis-30157	157	7	skin	skin	NOUN
fcis-30157	157	8	-	-	PUNCT
fcis-30157	157	9	type	type	NOUN
fcis-30157	157	10	inclusive	inclusive	ADJ
fcis-30157	157	11	benchmarks	benchmark	NOUN
fcis-30157	157	12	and	and	CCONJ
fcis-30157	157	13	continuous	continuous	ADJ
fcis-30157	157	14	learning	learning	NOUN
fcis-30157	157	15	architectures	architecture	NOUN
fcis-30157	157	16	to	to	PART
fcis-30157	157	17	maintain	maintain	VERB
fcis-30157	157	18	diagnostic	diagnostic	ADJ
fcis-30157	157	19	accuracy	accuracy	NOUN
fcis-30157	157	20	.	.	PUNCT
fcis-30157	158	1	1.3.3	1.3.3	X
fcis-30157	158	2	.	.	PUNCT
fcis-30157	159	1	gaps	gap	NOUN
fcis-30157	159	2	in	in	ADP
fcis-30157	159	3	the	the	DET
fcis-30157	159	4	current	current	ADJ
fcis-30157	159	5	research	research	NOUN
fcis-30157	159	6	current	current	ADJ
fcis-30157	159	7	research	research	NOUN
fcis-30157	159	8	on	on	ADP
fcis-30157	159	9	automatic	automatic	ADJ
fcis-30157	159	10	identification	identification	NOUN
fcis-30157	159	11	and	and	CCONJ
fcis-30157	159	12	auxiliary	auxiliary	ADJ
fcis-30157	159	13	diagnosis	diagnosis	NOUN
fcis-30157	159	14	of	of	ADP
fcis-30157	159	15	skin	skin	NOUN
fcis-30157	159	16	diseases	disease	NOUN
fcis-30157	159	17	still	still	ADV
fcis-30157	159	18	faces	face	VERB
fcis-30157	159	19	significant	significant	ADJ
fcis-30157	159	20	technical	technical	ADJ
fcis-30157	159	21	and	and	CCONJ
fcis-30157	159	22	application	application	NOUN
fcis-30157	159	23	gaps	gap	NOUN
fcis-30157	159	24	.	.	PUNCT
fcis-30157	160	1	at	at	ADP
fcis-30157	160	2	the	the	DET
fcis-30157	160	3	technical	technical	ADJ
fcis-30157	160	4	level	level	NOUN
fcis-30157	160	5	,	,	PUNCT
fcis-30157	160	6	many	many	ADJ
fcis-30157	160	7	models	model	NOUN
fcis-30157	160	8	lack	lack	VERB
fcis-30157	160	9	multimodal	multimodal	NOUN
fcis-30157	160	10	data	datum	NOUN
fcis-30157	160	11	fusion	fusion	NOUN
fcis-30157	160	12	.	.	PUNCT
fcis-30157	161	1	for	for	ADP
fcis-30157	161	2	example	example	NOUN
fcis-30157	161	3	,	,	PUNCT
fcis-30157	161	4	while	while	SCONJ
fcis-30157	161	5	mdfnet	mdfnet	NOUN
fcis-30157	161	6	integrates	integrate	VERB
fcis-30157	161	7	skin	skin	NOUN
fcis-30157	161	8	images	image	NOUN
fcis-30157	161	9	and	and	CCONJ
fcis-30157	161	10	clinical	clinical	ADJ
fcis-30157	161	11	text	text	NOUN
fcis-30157	161	12	,	,	PUNCT
fcis-30157	161	13	its	its	PRON
fcis-30157	161	14	attention	attention	NOUN
fcis-30157	161	15	mechanism	mechanism	NOUN
fcis-30157	161	16	only	only	ADV
fcis-30157	161	17	covers	cover	VERB
fcis-30157	161	18	superficial	superficial	ADJ
fcis-30157	161	19	epidermal	epidermal	ADJ
fcis-30157	161	20	features	feature	NOUN
fcis-30157	161	21	,	,	PUNCT
fcis-30157	161	22	leaving	leave	VERB
fcis-30157	161	23	deeper	deep	ADJ
fcis-30157	161	24	dermal	dermal	NOUN
fcis-30157	161	25	features	feature	NOUN
fcis-30157	161	26	and	and	CCONJ
fcis-30157	161	27	immune	immune	ADJ
fcis-30157	161	28	cell	cell	NOUN
fcis-30157	161	29	infiltration	infiltration	NOUN
fcis-30157	161	30	unmodeled	unmodele	VERB
fcis-30157	161	31	.	.	PUNCT
fcis-30157	162	1	additionally	additionally	ADV
fcis-30157	162	2	,	,	PUNCT
fcis-30157	162	3	dark	dark	ADJ
fcis-30157	162	4	skin	skin	NOUN
fcis-30157	162	5	tone	tone	NOUN
fcis-30157	162	6	samples	sample	NOUN
fcis-30157	162	7	in	in	ADP
fcis-30157	162	8	training	training	NOUN
fcis-30157	162	9	data	datum	NOUN
fcis-30157	162	10	are	be	AUX
fcis-30157	162	11	underrepresented	underrepresente	VERB
fcis-30157	162	12	,	,	PUNCT
fcis-30157	162	13	leading	lead	VERB
fcis-30157	162	14	to	to	ADP
fcis-30157	162	15	a	a	DET
fcis-30157	162	16	higher	high	ADJ
fcis-30157	162	17	misdiagnosis	misdiagnosis	NOUN
fcis-30157	162	18	rate	rate	NOUN
fcis-30157	162	19	for	for	ADP
fcis-30157	162	20	melanoma	melanoma	NOUN
fcis-30157	162	21	in	in	ADP
fcis-30157	162	22	darker	dark	ADJ
fcis-30157	162	23	-	-	PUNCT
fcis-30157	162	24	skinned	skin	VERB
fcis-30157	162	25	patients	patient	NOUN
fcis-30157	162	26	(	(	PUNCT
fcis-30157	162	27	12.7	12.7	NUM
fcis-30157	162	28	%	%	NOUN
fcis-30157	162	29	vs.	vs.	ADP
fcis-30157	162	30	8.3	8.3	NUM
fcis-30157	162	31	%	%	NOUN
fcis-30157	162	32	for	for	ADP
fcis-30157	162	33	light	light	ADJ
fcis-30157	162	34	-	-	PUNCT
fcis-30157	162	35	skinned	skinned	ADJ
fcis-30157	162	36	groups	group	NOUN
fcis-30157	162	37	)	)	PUNCT
fcis-30157	162	38	in	in	ADP
fcis-30157	162	39	isic	isic	PROPN
fcis-30157	162	40	2019	2019	NUM
fcis-30157	162	41	[	[	X
fcis-30157	162	42	18	18	NUM
fcis-30157	162	43	]	]	PUNCT
fcis-30157	162	44	.	.	PUNCT
fcis-30157	163	1	moreover	moreover	ADV
fcis-30157	163	2	,	,	PUNCT
fcis-30157	163	3	the	the	DET
fcis-30157	163	4	fs3dciot	fs3dciot	NOUN
fcis-30157	163	5	network	network	NOUN
fcis-30157	163	6	's	's	PART
fcis-30157	163	7	recognition	recognition	NOUN
fcis-30157	163	8	accuracy	accuracy	NOUN
fcis-30157	163	9	for	for	ADP
fcis-30157	163	10	rare	rare	ADJ
fcis-30157	163	11	diseases	disease	NOUN
fcis-30157	163	12	like	like	ADP
fcis-30157	163	13	t	t	NOUN
fcis-30157	163	14	-	-	PUNCT
fcis-30157	163	15	cell	cell	NOUN
fcis-30157	163	16	lymphoma	lymphoma	NOUN
fcis-30157	163	17	drops	drop	VERB
fcis-30157	163	18	sharply	sharply	ADV
fcis-30157	163	19	to	to	ADP
fcis-30157	163	20	61.2	61.2	NUM
fcis-30157	163	21	%	%	NOUN
fcis-30157	163	22	when	when	SCONJ
fcis-30157	163	23	trained	train	VERB
fcis-30157	163	24	with	with	ADP
fcis-30157	163	25	limited	limited	ADJ
fcis-30157	163	26	samples	sample	NOUN
fcis-30157	163	27	,	,	PUNCT
fcis-30157	163	28	compared	compare	VERB
fcis-30157	163	29	to	to	ADP
fcis-30157	163	30	82.4	82.4	NUM
fcis-30157	163	31	%	%	NOUN
fcis-30157	163	32	with	with	ADP
fcis-30157	163	33	conventional	conventional	ADJ
fcis-30157	163	34	datasets	dataset	NOUN
fcis-30157	163	35	.	.	PUNCT
fcis-30157	164	1	clinically	clinically	ADV
fcis-30157	164	2	,	,	PUNCT
fcis-30157	164	3	integration	integration	NOUN
fcis-30157	164	4	with	with	ADP
fcis-30157	164	5	medical	medical	ADJ
fcis-30157	164	6	workflows	workflow	NOUN
fcis-30157	164	7	remains	remain	VERB
fcis-30157	164	8	a	a	DET
fcis-30157	164	9	challenge	challenge	NOUN
fcis-30157	164	10	.	.	PUNCT
fcis-30157	165	1	a	a	DET
fcis-30157	165	2	uk	uk	PROPN
fcis-30157	165	3	nhs	nhs	PROPN
fcis-30157	165	4	pilot	pilot	NOUN
fcis-30157	165	5	found	find	VERB
fcis-30157	165	6	that	that	SCONJ
fcis-30157	165	7	ai	ai	VERB
fcis-30157	165	8	tools	tool	NOUN
fcis-30157	165	9	triggered	trigger	VERB
fcis-30157	165	10	an	an	DET
fcis-30157	165	11	average	average	NOUN
fcis-30157	165	12	of	of	ADP
fcis-30157	165	13	7.3	7.3	NUM
fcis-30157	165	14	unnecessary	unnecessary	ADJ
fcis-30157	165	15	biopsy	biopsy	NOUN
fcis-30157	165	16	recommendations	recommendation	NOUN
fcis-30157	165	17	per	per	ADP
fcis-30157	165	18	day	day	NOUN
fcis-30157	165	19	,	,	PUNCT
fcis-30157	165	20	mainly	mainly	ADV
fcis-30157	165	21	due	due	ADP
fcis-30157	165	22	to	to	ADP
fcis-30157	165	23	confusion	confusion	NOUN
fcis-30157	165	24	between	between	ADP
fcis-30157	165	25	benign	benign	ADJ
fcis-30157	165	26	and	and	CCONJ
fcis-30157	165	27	malignant	malignant	ADJ
fcis-30157	165	28	lesions	lesion	NOUN
fcis-30157	165	29	[	[	X
fcis-30157	165	30	18	18	NUM
fcis-30157	165	31	]	]	PUNCT
fcis-30157	165	32	.	.	PUNCT
fcis-30157	166	1	in	in	ADP
fcis-30157	166	2	doctor	doctor	NOUN
fcis-30157	166	3	decision	decision	NOUN
fcis-30157	166	4	support	support	NOUN
fcis-30157	166	5	,	,	PUNCT
fcis-30157	166	6	clinical	clinical	ADJ
fcis-30157	166	7	tests	test	NOUN
fcis-30157	166	8	revealed	reveal	VERB
fcis-30157	166	9	that	that	SCONJ
fcis-30157	166	10	only	only	ADV
fcis-30157	166	11	43.8	43.8	NUM
fcis-30157	166	12	%	%	NOUN
fcis-30157	166	13	of	of	ADP
fcis-30157	166	14	doctors	doctor	NOUN
fcis-30157	166	15	adopted	adopt	VERB
fcis-30157	166	16	ai	ai	VERB
fcis-30157	166	17	system	system	NOUN
fcis-30157	166	18	recommendations	recommendation	NOUN
fcis-30157	166	19	,	,	PUNCT
fcis-30157	166	20	mainly	mainly	ADV
fcis-30157	166	21	due	due	ADP
fcis-30157	166	22	to	to	ADP
fcis-30157	166	23	unclear	unclear	ADJ
fcis-30157	166	24	mappings	mapping	NOUN
fcis-30157	166	25	between	between	ADP
fcis-30157	166	26	model	model	NOUN
fcis-30157	166	27	confidence	confidence	NOUN
fcis-30157	166	28	scores	score	NOUN
fcis-30157	166	29	and	and	CCONJ
fcis-30157	166	30	pathological	pathological	ADJ
fcis-30157	166	31	diagnoses	diagnosis	NOUN
fcis-30157	166	32	,	,	PUNCT
fcis-30157	166	33	such	such	ADJ
fcis-30157	166	34	as	as	ADP
fcis-30157	166	35	in	in	ADP
fcis-30157	166	36	differentiating	differentiate	VERB
fcis-30157	166	37	psoriasis	psoriasis	NOUN
fcis-30157	166	38	and	and	CCONJ
fcis-30157	166	39	eczema	eczema	NOUN
fcis-30157	167	1	[	[	X
fcis-30157	167	2	20	20	NUM
fcis-30157	167	3	]	]	PUNCT
fcis-30157	167	4	.	.	PUNCT
fcis-30157	168	1	technological	technological	ADJ
fcis-30157	168	2	breakthroughs	breakthrough	NOUN
fcis-30157	168	3	should	should	AUX
fcis-30157	168	4	focus	focus	VERB
fcis-30157	168	5	on	on	ADP
fcis-30157	168	6	multi	multi	ADJ
fcis-30157	168	7	-	-	ADJ
fcis-30157	168	8	scale	scale	ADJ
fcis-30157	168	9	feature	feature	NOUN
fcis-30157	168	10	integration	integration	NOUN
fcis-30157	168	11	and	and	CCONJ
fcis-30157	168	12	incremental	incremental	ADJ
fcis-30157	168	13	learning	learning	NOUN
fcis-30157	168	14	innovations	innovation	NOUN
fcis-30157	168	15	.	.	PUNCT
fcis-30157	169	1	the	the	DET
fcis-30157	169	2	mobilenet	mobilenet	NOUN
fcis-30157	169	3	-	-	PUNCT
fcis-30157	169	4	enhanced	enhance	VERB
fcis-30157	169	5	model	model	NOUN
fcis-30157	169	6	accurately	accurately	ADV
fcis-30157	169	7	measures	measure	VERB
fcis-30157	169	8	epidermal	epidermal	ADJ
fcis-30157	169	9	thickness	thickness	NOUN
fcis-30157	169	10	but	but	CCONJ
fcis-30157	169	11	still	still	ADV
fcis-30157	169	12	struggles	struggle	VERB
fcis-30157	169	13	with	with	ADP
fcis-30157	169	14	integrating	integrate	VERB
fcis-30157	169	15	dermoscopic	dermoscopic	ADJ
fcis-30157	169	16	capillary	capillary	ADJ
fcis-30157	169	17	patterns	pattern	NOUN
fcis-30157	169	18	.	.	PUNCT
fcis-30157	170	1	new	new	ADJ
fcis-30157	170	2	multitask	multitask	ADJ
fcis-30157	170	3	learning	learning	NOUN
fcis-30157	170	4	approaches	approach	NOUN
fcis-30157	170	5	,	,	PUNCT
fcis-30157	170	6	though	though	SCONJ
fcis-30157	170	7	improving	improve	VERB
fcis-30157	170	8	diagnosis	diagnosis	NOUN
fcis-30157	170	9	time	time	NOUN
fcis-30157	170	10	for	for	ADP
fcis-30157	170	11	acne	acne	NOUN
fcis-30157	170	12	severity	severity	NOUN
fcis-30157	170	13	,	,	PUNCT
fcis-30157	170	14	still	still	ADV
fcis-30157	170	15	rely	rely	VERB
fcis-30157	170	16	on	on	ADP
fcis-30157	170	17	full	full	ADJ
fcis-30157	170	18	retraining	retraining	NOUN
fcis-30157	170	19	for	for	ADP
fcis-30157	170	20	updates	update	NOUN
fcis-30157	170	21	.	.	PUNCT
fcis-30157	171	1	recent	recent	ADJ
fcis-30157	171	2	interdisciplinary	interdisciplinary	ADJ
fcis-30157	171	3	research	research	NOUN
fcis-30157	171	4	,	,	PUNCT
fcis-30157	171	5	like	like	ADP
fcis-30157	171	6	mit	mit	PROPN
fcis-30157	171	7	's	's	PART
fcis-30157	171	8	integration	integration	NOUN
fcis-30157	171	9	of	of	ADP
fcis-30157	171	10	respiratory	respiratory	ADJ
fcis-30157	171	11	sensors	sensor	NOUN
fcis-30157	171	12	,	,	PUNCT
fcis-30157	171	13	has	have	AUX
fcis-30157	171	14	shown	show	VERB
fcis-30157	171	15	value	value	NOUN
fcis-30157	171	16	in	in	ADP
fcis-30157	171	17	combining	combine	VERB
fcis-30157	171	18	physical	physical	ADJ
fcis-30157	171	19	sensing	sensing	NOUN
fcis-30157	171	20	with	with	ADP
fcis-30157	171	21	visual	visual	ADJ
fcis-30157	171	22	data	datum	NOUN
fcis-30157	171	23	,	,	PUNCT
fcis-30157	171	24	reducing	reduce	VERB
fcis-30157	171	25	prediction	prediction	NOUN
fcis-30157	171	26	errors	error	NOUN
fcis-30157	171	27	in	in	ADP
fcis-30157	171	28	psoriasis	psoriasis	NOUN
fcis-30157	171	29	monitoring	monitor	VERB
fcis-30157	171	30	[	[	X
fcis-30157	171	31	34	34	NUM
fcis-30157	171	32	]	]	PUNCT
fcis-30157	171	33	.	.	PUNCT
fcis-30157	172	1	a	a	DET
fcis-30157	172	2	hybrid	hybrid	ADJ
fcis-30157	172	3	approach	approach	NOUN
fcis-30157	172	4	framework	framework	NOUN
fcis-30157	172	5	is	be	AUX
fcis-30157	172	6	being	be	AUX
fcis-30157	172	7	developed	develop	VERB
fcis-30157	172	8	to	to	PART
fcis-30157	172	9	address	address	VERB
fcis-30157	172	10	these	these	DET
fcis-30157	172	11	limitations	limitation	NOUN
fcis-30157	172	12	.	.	PUNCT
fcis-30157	173	1	this	this	PRON
fcis-30157	173	2	includes	include	VERB
fcis-30157	173	3	a	a	DET
fcis-30157	173	4	cross	cross	ADJ
fcis-30157	173	5	-	-	ADJ
fcis-30157	173	6	modal	modal	ADJ
fcis-30157	173	7	database	database	NOUN
fcis-30157	173	8	built	build	VERB
fcis-30157	173	9	from	from	ADP
fcis-30157	173	10	23,906	23,906	NUM
fcis-30157	173	11	dermatoscope	dermatoscope	NOUN
fcis-30157	173	12	images	image	NOUN
fcis-30157	173	13	and	and	CCONJ
fcis-30157	173	14	12,345	12,345	NUM
fcis-30157	173	15	electronic	electronic	ADJ
fcis-30157	173	16	medical	medical	ADJ
fcis-30157	173	17	records	record	NOUN
fcis-30157	173	18	.	.	PUNCT
fcis-30157	174	1	data	datum	NOUN
fcis-30157	174	2	preprocessing	preprocessing	NOUN
fcis-30157	174	3	uses	use	VERB
fcis-30157	174	4	adaptive	adaptive	ADJ
fcis-30157	174	5	denoising	denoising	NOUN
fcis-30157	174	6	and	and	CCONJ
fcis-30157	174	7	adversarial	adversarial	ADJ
fcis-30157	174	8	networks	network	NOUN
fcis-30157	174	9	to	to	PART
fcis-30157	174	10	improve	improve	VERB
fcis-30157	174	11	image	image	NOUN
fcis-30157	174	12	quality	quality	NOUN
fcis-30157	174	13	,	,	PUNCT
fcis-30157	174	14	boosting	boost	VERB
fcis-30157	174	15	the	the	DET
fcis-30157	174	16	signal	signal	NOUN
fcis-30157	174	17	-	-	PUNCT
fcis-30157	174	18	to	to	ADP
fcis-30157	174	19	-	-	PUNCT
fcis-30157	174	20	noise	noise	NOUN
fcis-30157	174	21	ratio	ratio	NOUN
fcis-30157	174	22	to	to	ADP
fcis-30157	174	23	34.6db	34.6db	NUM
fcis-30157	174	24	.	.	PUNCT
fcis-30157	175	1	a	a	DET
fcis-30157	175	2	pathology	pathology	NOUN
fcis-30157	175	3	-	-	PUNCT
fcis-30157	175	4	guided	guide	VERB
fcis-30157	175	5	synthetic	synthetic	ADJ
fcis-30157	175	6	strategy	strategy	NOUN
fcis-30157	175	7	expanded	expand	VERB
fcis-30157	175	8	rare	rare	ADJ
fcis-30157	175	9	lesion	lesion	NOUN
fcis-30157	175	10	samples	sample	NOUN
fcis-30157	175	11	significantly	significantly	ADV
fcis-30157	175	12	,	,	PUNCT
fcis-30157	175	13	from	from	ADP
fcis-30157	175	14	37	37	NUM
fcis-30157	175	15	to	to	ADP
fcis-30157	175	16	2,184	2,184	NUM
fcis-30157	175	17	cases	case	NOUN
fcis-30157	175	18	.	.	PUNCT
fcis-30157	176	1	for	for	ADP
fcis-30157	176	2	model	model	NOUN
fcis-30157	176	3	architecture	architecture	NOUN
fcis-30157	176	4	,	,	PUNCT
fcis-30157	176	5	resnet-152	resnet-152	NOUN
fcis-30157	176	6	achieved	achieve	VERB
fcis-30157	176	7	87.2	87.2	NUM
fcis-30157	176	8	%	%	NOUN
fcis-30157	176	9	accuracy	accuracy	NOUN
fcis-30157	176	10	in	in	ADP
fcis-30157	176	11	melanoma	melanoma	NOUN
fcis-30157	176	12	classification	classification	NOUN
fcis-30157	176	13	but	but	CCONJ
fcis-30157	176	14	had	have	VERB
fcis-30157	176	15	a	a	DET
fcis-30157	176	16	high	high	ADJ
fcis-30157	176	17	parameter	parameter	NOUN
fcis-30157	176	18	volume	volume	NOUN
fcis-30157	176	19	(	(	PUNCT
fcis-30157	176	20	60.2	60.2	NUM
fcis-30157	176	21	m	m	NOUN
fcis-30157	176	22	)	)	PUNCT
fcis-30157	176	23	,	,	PUNCT
fcis-30157	176	24	making	make	VERB
fcis-30157	176	25	it	it	PRON
fcis-30157	176	26	difficult	difficult	ADJ
fcis-30157	176	27	for	for	ADP
fcis-30157	176	28	mobile	mobile	ADJ
fcis-30157	176	29	deployment	deployment	NOUN
fcis-30157	176	30	.	.	PUNCT
fcis-30157	177	1	instead	instead	ADV
fcis-30157	177	2	,	,	PUNCT
fcis-30157	177	3	an	an	DET
fcis-30157	177	4	improved	improved	ADJ
fcis-30157	177	5	efficientnet	efficientnet	NOUN
fcis-30157	177	6	-	-	PUNCT
fcis-30157	177	7	b4	b4	NOUN
fcis-30157	177	8	with	with	ADP
fcis-30157	177	9	optimized	optimize	VERB
fcis-30157	177	10	channel	channel	NOUN
fcis-30157	177	11	attention	attention	NOUN
fcis-30157	177	12	reduced	reduce	VERB
fcis-30157	177	13	model	model	NOUN
fcis-30157	177	14	size	size	NOUN
fcis-30157	177	15	to	to	ADP
fcis-30157	177	16	19	19	NUM
fcis-30157	177	17	m	m	NOUN
fcis-30157	177	18	while	while	SCONJ
fcis-30157	177	19	maintaining	maintain	VERB
fcis-30157	177	20	83.9	83.9	NUM
fcis-30157	177	21	%	%	NOUN
fcis-30157	177	22	accuracy	accuracy	NOUN
fcis-30157	177	23	.	.	PUNCT
fcis-30157	178	1	course	course	NOUN
fcis-30157	178	2	learning	learn	VERB
fcis-30157	178	3	strategies	strategy	NOUN
fcis-30157	178	4	and	and	CCONJ
fcis-30157	178	5	progressive	progressive	ADJ
fcis-30157	178	6	training	training	NOUN
fcis-30157	178	7	also	also	ADV
fcis-30157	178	8	increased	increase	VERB
fcis-30157	178	9	eczema	eczema	NOUN
fcis-30157	178	10	-	-	PUNCT
fcis-30157	178	11	psoriasis	psoriasis	NOUN
fcis-30157	178	12	identification	identification	NOUN
fcis-30157	178	13	accuracy	accuracy	NOUN
fcis-30157	178	14	by	by	ADP
fcis-30157	178	15	9.3	9.3	NUM
fcis-30157	178	16	%	%	NOUN
fcis-30157	178	17	.	.	PUNCT
fcis-30157	179	1	the	the	DET
fcis-30157	179	2	auxiliary	auxiliary	ADJ
fcis-30157	179	3	diagnostic	diagnostic	ADJ
fcis-30157	179	4	framework	framework	NOUN
fcis-30157	179	5	embeds	embed	VERB
fcis-30157	179	6	clinical	clinical	ADJ
fcis-30157	179	7	knowledge	knowledge	NOUN
fcis-30157	179	8	through	through	ADP
fcis-30157	179	9	a	a	DET
fcis-30157	179	10	rule	rule	NOUN
fcis-30157	179	11	’s	’s	PART
fcis-30157	179	12	engine	engine	NOUN
fcis-30157	179	13	,	,	PUNCT
fcis-30157	179	14	translating	translate	VERB
fcis-30157	179	15	"	"	PUNCT
fcis-30157	179	16	chinese	chinese	ADJ
fcis-30157	179	17	clinical	clinical	ADJ
fcis-30157	179	18	dermatology	dermatology	NOUN
fcis-30157	179	19	"	"	PUNCT
fcis-30157	179	20	into	into	ADP
fcis-30157	179	21	382	382	NUM
fcis-30157	179	22	decision	decision	NOUN
fcis-30157	179	23	tree	tree	NOUN
fcis-30157	179	24	rules	rule	NOUN
fcis-30157	179	25	,	,	PUNCT
fcis-30157	179	26	which	which	PRON
fcis-30157	179	27	are	be	AUX
fcis-30157	179	28	fused	fuse	VERB
fcis-30157	179	29	with	with	ADP
fcis-30157	179	30	deep	deep	ADJ
fcis-30157	179	31	learning	learning	NOUN
fcis-30157	179	32	outputs	output	NOUN
fcis-30157	179	33	via	via	ADP
fcis-30157	179	34	bayesian	bayesian	NOUN
fcis-30157	179	35	methods	method	NOUN
fcis-30157	179	36	.	.	PUNCT
fcis-30157	180	1	this	this	PRON
fcis-30157	180	2	improves	improve	VERB
fcis-30157	180	3	diagnostic	diagnostic	ADJ
fcis-30157	180	4	accuracy	accuracy	NOUN
fcis-30157	180	5	,	,	PUNCT
fcis-30157	180	6	increasing	increase	VERB
fcis-30157	180	7	the	the	DET
fcis-30157	180	8	distance	distance	NOUN
fcis-30157	180	9	between	between	ADP
fcis-30157	180	10	psoriasis	psoriasis	NOUN
fcis-30157	180	11	and	and	CCONJ
fcis-30157	180	12	pityriasis	pityriasis	ADJ
fcis-30157	180	13	rosea	rosea	NOUN
fcis-30157	180	14	clusters	cluster	NOUN
fcis-30157	180	15	from	from	ADP
fcis-30157	180	16	0.38	0.38	NUM
fcis-30157	180	17	to	to	ADP
fcis-30157	180	18	0.61	0.61	NUM
fcis-30157	180	19	.	.	PUNCT
fcis-30157	181	1	a	a	DET
fcis-30157	181	2	risk	risk	NOUN
fcis-30157	181	3	warning	warning	NOUN
fcis-30157	181	4	system	system	NOUN
fcis-30157	181	5	automatically	automatically	ADV
fcis-30157	181	6	triggers	trigger	VERB
fcis-30157	181	7	histopathological	histopathological	ADJ
fcis-30157	181	8	examination	examination	NOUN
fcis-30157	181	9	recommendations	recommendation	NOUN
fcis-30157	181	10	when	when	SCONJ
fcis-30157	181	11	basal	basal	ADJ
fcis-30157	181	12	cell	cell	NOUN
fcis-30157	181	13	carcinoma	carcinoma	NOUN
fcis-30157	181	14	predictions	prediction	NOUN
fcis-30157	181	15	exceed	exceed	VERB
fcis-30157	181	16	0.72	0.72	NUM
fcis-30157	181	17	.	.	PUNCT
fcis-30157	182	1	clinical	clinical	ADJ
fcis-30157	182	2	validation	validation	NOUN
fcis-30157	182	3	on	on	ADP
fcis-30157	182	4	300	300	NUM
fcis-30157	182	5	cases	case	NOUN
fcis-30157	182	6	showed	show	VERB
fcis-30157	182	7	an	an	DET
fcis-30157	182	8	early	early	ADJ
fcis-30157	182	9	diagnosis	diagnosis	NOUN
fcis-30157	182	10	rate	rate	NOUN
fcis-30157	182	11	of	of	ADP
fcis-30157	182	12	91.4	91.4	NUM
fcis-30157	182	13	%	%	NOUN
fcis-30157	182	14	.	.	PUNCT
fcis-30157	183	1	table	table	NOUN
fcis-30157	183	2	1	1	NUM
fcis-30157	183	3	.	.	PUNCT
fcis-30157	183	4	comparison	comparison	NOUN
fcis-30157	183	5	of	of	ADP
fcis-30157	183	6	performance	performance	NOUN
fcis-30157	183	7	of	of	ADP
fcis-30157	183	8	mainstream	mainstream	NOUN
fcis-30157	183	9	models	model	NOUN
fcis-30157	183	10	model	model	NOUN
fcis-30157	183	11	architecture	architecture	NOUN
fcis-30157	183	12	test	test	NOUN
fcis-30157	183	13	dataset	dataset	NOUN
fcis-30157	183	14	accuracy	accuracy	NOUN
fcis-30157	183	15	(	(	PUNCT
fcis-30157	183	16	%	%	INTJ
fcis-30157	183	17	)	)	PUNCT
fcis-30157	183	18	parameter	parameter	NOUN
fcis-30157	183	19	quantity	quantity	NOUN
fcis-30157	183	20	(	(	PUNCT
fcis-30157	183	21	m	m	NOUN
fcis-30157	183	22	)	)	PUNCT
fcis-30157	183	23	resnet-50	resnet-50	PROPN
fcis-30157	183	24	isic	isic	NOUN
fcis-30157	183	25	2020	2020	NUM
fcis-30157	183	26	84.2	84.2	NUM
fcis-30157	183	27	25.6	25.6	NUM
fcis-30157	183	28	densenet-201	densenet-201	NOUN
fcis-30157	183	29	ham10000	ham10000	VERB
fcis-30157	183	30	86.7	86.7	NUM
fcis-30157	183	31	20.0	20.0	NUM
fcis-30157	183	32	mobilenet	mobilenet	PROPN
fcis-30157	183	33	-	-	PUNCT
fcis-30157	183	34	v3	v3	NOUN
fcis-30157	183	35	pad	pad	PROPN
fcis-30157	183	36	-	-	PUNCT
fcis-30157	183	37	ufes-20	ufes-20	PROPN
fcis-30157	183	38	79.8	79.8	NUM
fcis-30157	183	39	5.4	5.4	NUM
fcis-30157	183	40	model	model	NOUN
fcis-30157	183	41	of	of	ADP
fcis-30157	183	42	this	this	DET
fcis-30157	183	43	study	study	NOUN
fcis-30157	183	44	hybriddata	hybriddata	VERB
fcis-30157	183	45	88.1	88.1	NUM
fcis-30157	183	46	19.0	19.0	NUM
fcis-30157	183	47	the	the	DET
fcis-30157	183	48	model	model	NOUN
fcis-30157	183	49	optimization	optimization	NOUN
fcis-30157	183	50	process	process	NOUN
fcis-30157	183	51	reveals	reveal	VERB
fcis-30157	183	52	the	the	DET
fcis-30157	183	53	particularity	particularity	NOUN
fcis-30157	183	54	of	of	ADP
fcis-30157	183	55	feature	feature	NOUN
fcis-30157	183	56	learning	learn	VERB
fcis-30157	183	57	mechanisms	mechanism	NOUN
fcis-30157	183	58	.	.	PUNCT
fcis-30157	184	1	grad	grad	NOUN
fcis-30157	184	2	-	-	PUNCT
fcis-30157	184	3	cam	cam	NOUN
fcis-30157	184	4	visualization	visualization	NOUN
fcis-30157	184	5	showed	show	VERB
fcis-30157	184	6	that	that	SCONJ
fcis-30157	184	7	the	the	DET
fcis-30157	184	8	improved	improved	ADJ
fcis-30157	184	9	model	model	NOUN
fcis-30157	184	10	's	's	PART
fcis-30157	184	11	attention	attention	NOUN
fcis-30157	184	12	to	to	ADP
fcis-30157	184	13	irregular	irregular	ADJ
fcis-30157	184	14	pigmentation	pigmentation	NOUN
fcis-30157	184	15	areas	area	NOUN
fcis-30157	184	16	around	around	ADP
fcis-30157	184	17	melanoma	melanoma	NOUN
fcis-30157	184	18	reached	reach	VERB
fcis-30157	184	19	73.2	73.2	NUM
fcis-30157	184	20	%	%	NOUN
fcis-30157	184	21	,	,	PUNCT
fcis-30157	184	22	an	an	DET
fcis-30157	184	23	increase	increase	NOUN
fcis-30157	184	24	of	of	ADP
fcis-30157	184	25	21.5	21.5	NUM
fcis-30157	184	26	percentage	percentage	NOUN
fcis-30157	184	27	points	point	NOUN
fcis-30157	184	28	from	from	ADP
fcis-30157	184	29	the	the	DET
fcis-30157	184	30	baseline	baseline	NOUN
fcis-30157	184	31	model	model	NOUN
fcis-30157	184	32	.	.	PUNCT
fcis-30157	185	1	transfer	transfer	NOUN
fcis-30157	185	2	learning	learn	VERB
fcis-30157	185	3	experiments	experiment	NOUN
fcis-30157	185	4	show	show	VERB
fcis-30157	185	5	that	that	SCONJ
fcis-30157	185	6	in	in	ADP
fcis-30157	185	7	the	the	DET
fcis-30157	185	8	diagnosis	diagnosis	NOUN
fcis-30157	185	9	task	task	NOUN
fcis-30157	185	10	of	of	ADP
fcis-30157	185	11	lupus	lupus	PROPN
fcis-30157	185	12	erythematosus	erythematosus	PROPN
fcis-30157	185	13	,	,	PUNCT
fcis-30157	185	14	the	the	DET
fcis-30157	185	15	imagenet	imagenet	NOUN
fcis-30157	185	16	pre	pre	ADJ
fcis-30157	185	17	-	-	ADJ
fcis-30157	185	18	trained	train	VERB
fcis-30157	185	19	model	model	NOUN
fcis-30157	185	20	requires	require	VERB
fcis-30157	185	21	1,200	1,200	NUM
fcis-30157	185	22	fine	fine	ADV
fcis-30157	185	23	-	-	PUNCT
fcis-30157	185	24	tuning	tune	VERB
fcis-30157	185	25	samples	sample	NOUN
fcis-30157	185	26	to	to	PART
fcis-30157	185	27	achieve	achieve	VERB
fcis-30157	185	28	80	80	NUM
fcis-30157	185	29	%	%	NOUN
fcis-30157	185	30	accuracy	accuracy	NOUN
fcis-30157	185	31	,	,	PUNCT
fcis-30157	185	32	while	while	SCONJ
fcis-30157	185	33	the	the	DET
fcis-30157	185	34	pre	pre	ADJ
fcis-30157	185	35	-	-	ADJ
fcis-30157	185	36	trained	train	VERB
fcis-30157	185	37	model	model	NOUN
fcis-30157	185	38	for	for	ADP
fcis-30157	185	39	skin	skin	NOUN
fcis-30157	185	40	diseases	disease	NOUN
fcis-30157	185	41	requires	require	VERB
fcis-30157	185	42	only	only	ADV
fcis-30157	185	43	600	600	NUM
fcis-30157	185	44	cases	case	NOUN
fcis-30157	185	45	to	to	PART
fcis-30157	185	46	achieve	achieve	VERB
fcis-30157	185	47	the	the	DET
fcis-30157	185	48	same	same	ADJ
fcis-30157	185	49	performance	performance	NOUN
fcis-30157	185	50	,	,	PUNCT
fcis-30157	185	51	proving	prove	VERB
fcis-30157	185	52	the	the	DET
fcis-30157	185	53	fieldspecific	fieldspecific	ADJ
fcis-30157	185	54	feature	feature	NOUN
fcis-30157	185	55	extraction	extraction	NOUN
fcis-30157	185	56	.	.	PUNCT
fcis-30157	186	1	a	a	DET
fcis-30157	186	2	double	double	ADJ
fcis-30157	186	3	-	-	PUNCT
fcis-30157	186	4	blind	blind	ADJ
fcis-30157	186	5	controlled	control	VERB
fcis-30157	186	6	trial	trial	NOUN
fcis-30157	186	7	was	be	AUX
fcis-30157	186	8	designed	design	VERB
fcis-30157	186	9	in	in	ADP
fcis-30157	186	10	the	the	DET
fcis-30157	186	11	clinical	clinical	ADJ
fcis-30157	186	12	verification	verification	NOUN
fcis-30157	186	13	process	process	NOUN
fcis-30157	186	14	.	.	PUNCT
fcis-30157	187	1	the	the	DET
fcis-30157	187	2	diagnostic	diagnostic	ADJ
fcis-30157	187	3	accuracy	accuracy	NOUN
fcis-30157	187	4	rate	rate	NOUN
fcis-30157	187	5	of	of	ADP
fcis-30157	187	6	the	the	DET
fcis-30157	187	7	dermatologist	dermatologist	NOUN
fcis-30157	187	8	group	group	NOUN
fcis-30157	187	9	(	(	PUNCT
fcis-30157	187	10	n=15	n=15	SYM
fcis-30157	187	11	)	)	PUNCT
fcis-30157	187	12	increased	increase	VERB
fcis-30157	187	13	from	from	ADP
fcis-30157	187	14	78.4	78.4	NUM
fcis-30157	187	15	%	%	NOUN
fcis-30157	187	16	to	to	ADP
fcis-30157	187	17	86.9	86.9	NUM
fcis-30157	187	18	%	%	NOUN
fcis-30157	187	19	with	with	ADP
fcis-30157	187	20	ai	ai	ADJ
fcis-30157	187	21	-	-	PUNCT
fcis-30157	187	22	assisted	assist	VERB
fcis-30157	187	23	treatment	treatment	NOUN
fcis-30157	187	24	,	,	PUNCT
fcis-30157	187	25	and	and	CCONJ
fcis-30157	187	26	the	the	DET
fcis-30157	187	27	average	average	ADJ
fcis-30157	187	28	decision	decision	NOUN
fcis-30157	187	29	-	-	PUNCT
fcis-30157	187	30	making	making	NOUN
fcis-30157	187	31	time	time	NOUN
fcis-30157	187	32	was	be	AUX
fcis-30157	187	33	shortened	shorten	VERB
fcis-30157	187	34	to	to	ADP
fcis-30157	187	35	4.3	4.3	NUM
fcis-30157	187	36	minutes	minute	NOUN
fcis-30157	187	37	.	.	PUNCT
fcis-30157	188	1	it	it	PRON
fcis-30157	188	2	is	be	AUX
fcis-30157	188	3	worth	worth	ADJ
fcis-30157	188	4	noting	note	VERB
fcis-30157	188	5	that	that	SCONJ
fcis-30157	188	6	the	the	DET
fcis-30157	188	7	resident	resident	ADJ
fcis-30157	188	8	group	group	NOUN
fcis-30157	188	9	(	(	PUNCT
fcis-30157	188	10	n=8	n=8	NUM
fcis-30157	188	11	)	)	PUNCT
fcis-30157	188	12	benefited	benefit	VERB
fcis-30157	188	13	the	the	DET
fcis-30157	188	14	most	most	ADV
fcis-30157	188	15	significantly	significantly	ADV
fcis-30157	188	16	,	,	PUNCT
fcis-30157	188	17	with	with	ADP
fcis-30157	188	18	an	an	DET
fcis-30157	188	19	increase	increase	NOUN
fcis-30157	188	20	in	in	ADP
fcis-30157	188	21	the	the	DET
fcis-30157	188	22	accuracy	accuracy	NOUN
fcis-30157	188	23	of	of	ADP
fcis-30157	188	24	independent	independent	ADJ
fcis-30157	188	25	diagnosis	diagnosis	NOUN
fcis-30157	188	26	by	by	ADP
fcis-30157	188	27	14.2	14.2	NUM
fcis-30157	188	28	%	%	NOUN
fcis-30157	188	29	,	,	PUNCT
fcis-30157	188	30	indicating	indicate	VERB
fcis-30157	188	31	that	that	SCONJ
fcis-30157	188	32	the	the	DET
fcis-30157	188	33	system	system	NOUN
fcis-30157	188	34	has	have	VERB
fcis-30157	188	35	special	special	ADJ
fcis-30157	188	36	value	value	NOUN
fcis-30157	188	37	in	in	ADP
fcis-30157	188	38	empirical	empirical	ADJ
fcis-30157	188	39	compensation	compensation	NOUN
fcis-30157	188	40	.	.	PUNCT
fcis-30157	189	1	systematic	systematic	ADJ
fcis-30157	189	2	misdiagnosis	misdiagnosis	NOUN
fcis-30157	189	3	case	case	NOUN
fcis-30157	189	4	analysis	analysis	NOUN
fcis-30157	189	5	found	find	VERB
fcis-30157	189	6	that	that	SCONJ
fcis-30157	189	7	57.3	57.3	NUM
fcis-30157	189	8	%	%	NOUN
fcis-30157	189	9	of	of	ADP
fcis-30157	189	10	the	the	DET
fcis-30157	189	11	errors	error	NOUN
fcis-30157	189	12	originated	originate	VERB
fcis-30157	189	13	from	from	ADP
fcis-30157	189	14	rare	rare	ADJ
fcis-30157	189	15	mutant	mutant	ADJ
fcis-30157	189	16	lesions	lesion	NOUN
fcis-30157	189	17	,	,	PUNCT
fcis-30157	189	18	such	such	ADJ
fcis-30157	189	19	as	as	ADP
fcis-30157	189	20	pigmented	pigment	VERB
fcis-30157	189	21	basal	basal	NOUN
fcis-30157	189	22	cell	cell	NOUN
fcis-30157	189	23	carcinoma	carcinoma	NOUN
fcis-30157	189	24	being	be	AUX
fcis-30157	189	25	misdiagnosed	misdiagnose	VERB
fcis-30157	189	26	as	as	ADP
fcis-30157	189	27	melanocyte	melanocyte	NOUN
fcis-30157	189	28	nevus	nevus	NOUN
fcis-30157	189	29	,	,	PUNCT
fcis-30157	189	30	which	which	PRON
fcis-30157	189	31	pointed	point	VERB
fcis-30157	189	32	out	out	ADP
fcis-30157	189	33	the	the	DET
fcis-30157	189	34	direction	direction	NOUN
fcis-30157	189	35	for	for	ADP
fcis-30157	189	36	subsequent	subsequent	ADJ
fcis-30157	189	37	small	small	ADJ
fcis-30157	189	38	sample	sample	NOUN
fcis-30157	189	39	learning	learn	VERB
fcis-30157	189	40	optimization	optimization	NOUN
fcis-30157	189	41	.	.	PUNCT
fcis-30157	190	1	2	2	X
fcis-30157	190	2	.	.	X
fcis-30157	190	3	experimental	experimental	ADJ
fcis-30157	190	4	process	process	NOUN
fcis-30157	190	5	2.1	2.1	NUM
fcis-30157	190	6	.	.	PUNCT
fcis-30157	191	1	data	datum	NOUN
fcis-30157	191	2	collection	collection	NOUN
fcis-30157	191	3	and	and	CCONJ
fcis-30157	191	4	preprocessing	preprocesse	VERB
fcis-30157	191	5	2.1.1	2.1.1	NUM
fcis-30157	191	6	.	.	PUNCT
fcis-30157	192	1	sources	source	NOUN
fcis-30157	192	2	of	of	ADP
fcis-30157	192	3	skin	skin	NOUN
fcis-30157	192	4	disease	disease	NOUN
fcis-30157	192	5	images	image	NOUN
fcis-30157	192	6	acquiring	acquire	VERB
fcis-30157	192	7	skin	skin	NOUN
fcis-30157	192	8	disease	disease	NOUN
fcis-30157	192	9	image	image	NOUN
fcis-30157	192	10	data	datum	NOUN
fcis-30157	192	11	is	be	AUX
fcis-30157	192	12	crucial	crucial	ADJ
fcis-30157	192	13	for	for	ADP
fcis-30157	192	14	building	build	VERB
fcis-30157	192	15	automatic	automatic	ADJ
fcis-30157	192	16	identification	identification	NOUN
fcis-30157	192	17	and	and	CCONJ
fcis-30157	192	18	diagnostic	diagnostic	ADJ
fcis-30157	192	19	systems	system	NOUN
fcis-30157	192	20	.	.	PUNCT
fcis-30157	193	1	hospital	hospital	NOUN
fcis-30157	193	2	dermatology	dermatology	NOUN
fcis-30157	193	3	case	case	NOUN
fcis-30157	193	4	systems	system	NOUN
fcis-30157	193	5	are	be	AUX
fcis-30157	193	6	a	a	DET
fcis-30157	193	7	key	key	ADJ
fcis-30157	193	8	source	source	NOUN
fcis-30157	193	9	,	,	PUNCT
fcis-30157	193	10	containing	contain	VERB
fcis-30157	193	11	over	over	ADP
fcis-30157	193	12	100,000	100,000	NUM
fcis-30157	193	13	images	image	NOUN
fcis-30157	193	14	of	of	ADP
fcis-30157	193	15	more	more	ADJ
fcis-30157	193	16	than	than	ADP
fcis-30157	193	17	50	50	NUM
fcis-30157	193	18	skin	skin	NOUN
fcis-30157	193	19	disease	disease	NOUN
fcis-30157	193	20	types	type	NOUN
fcis-30157	193	21	,	,	PUNCT
fcis-30157	193	22	including	include	VERB
fcis-30157	193	23	eczema	eczema	NOUN
fcis-30157	193	24	,	,	PUNCT
fcis-30157	193	25	psoriasis	psoriasis	NOUN
fcis-30157	193	26	,	,	PUNCT
fcis-30157	193	27	and	and	CCONJ
fcis-30157	193	28	melanoma	melanoma	NOUN
fcis-30157	193	29	.	.	PUNCT
fcis-30157	194	1	these	these	DET
fcis-30157	194	2	images	image	NOUN
fcis-30157	194	3	are	be	AUX
fcis-30157	194	4	highly	highly	ADV
fcis-30157	194	5	authentic	authentic	ADJ
fcis-30157	194	6	and	and	CCONJ
fcis-30157	194	7	relevant	relevant	ADJ
fcis-30157	194	8	to	to	ADP
fcis-30157	194	9	clinical	clinical	ADJ
fcis-30157	194	10	diagnosis	diagnosis	NOUN
fcis-30157	194	11	,	,	PUNCT
fcis-30157	194	12	but	but	CCONJ
fcis-30157	194	13	accessing	access	VERB
fcis-30157	194	14	them	they	PRON
fcis-30157	194	15	involves	involve	VERB
fcis-30157	194	16	ethical	ethical	ADJ
fcis-30157	194	17	reviews	review	NOUN
fcis-30157	194	18	and	and	CCONJ
fcis-30157	194	19	patient	patient	ADJ
fcis-30157	194	20	privacy	privacy	NOUN
fcis-30157	194	21	concerns	concern	NOUN
fcis-30157	194	22	.	.	PUNCT
fcis-30157	195	1	on	on	ADP
fcis-30157	195	2	the	the	DET
fcis-30157	195	3	other	other	ADJ
fcis-30157	195	4	hand	hand	NOUN
fcis-30157	195	5	,	,	PUNCT
fcis-30157	195	6	public	public	ADJ
fcis-30157	195	7	databases	database	NOUN
fcis-30157	195	8	,	,	PUNCT
fcis-30157	195	9	like	like	ADP
fcis-30157	195	10	the	the	DET
fcis-30157	195	11	isic	isic	PROPN
fcis-30157	195	12	archive	archive	NOUN
fcis-30157	195	13	with	with	ADP
fcis-30157	195	14	23,000	23,000	NUM
fcis-30157	195	15	skin	skin	NOUN
fcis-30157	195	16	disease	disease	NOUN
fcis-30157	195	17	images	image	NOUN
fcis-30157	195	18	and	and	CCONJ
fcis-30157	195	19	the	the	DET
fcis-30157	195	20	ham10000	ham10000	NOUN
fcis-30157	195	21	dataset	dataset	VERB
fcis-30157	195	22	116	116	NUM
fcis-30157	195	23	with	with	ADP
fcis-30157	195	24	10,015	10,015	NUM
fcis-30157	195	25	images	image	NOUN
fcis-30157	195	26	,	,	PUNCT
fcis-30157	195	27	provide	provide	VERB
fcis-30157	195	28	standardized	standardized	ADJ
fcis-30157	195	29	and	and	CCONJ
fcis-30157	195	30	free	free	ADJ
fcis-30157	195	31	data	datum	NOUN
fcis-30157	195	32	,	,	PUNCT
fcis-30157	195	33	though	though	SCONJ
fcis-30157	195	34	issues	issue	NOUN
fcis-30157	195	35	like	like	ADP
fcis-30157	195	36	copyright	copyright	NOUN
fcis-30157	195	37	and	and	CCONJ
fcis-30157	195	38	lack	lack	NOUN
fcis-30157	195	39	of	of	ADP
fcis-30157	195	40	clinical	clinical	ADJ
fcis-30157	195	41	context	context	NOUN
fcis-30157	195	42	(	(	PUNCT
fcis-30157	195	43	e.g.	e.g.	ADV
fcis-30157	195	44	,	,	PUNCT
fcis-30157	195	45	patient	patient	ADJ
fcis-30157	195	46	age	age	NOUN
fcis-30157	195	47	,	,	PUNCT
fcis-30157	195	48	gender	gender	NOUN
fcis-30157	195	49	)	)	PUNCT
fcis-30157	195	50	may	may	AUX
fcis-30157	195	51	limit	limit	VERB
fcis-30157	195	52	their	their	PRON
fcis-30157	195	53	utility	utility	NOUN
fcis-30157	195	54	for	for	ADP
fcis-30157	195	55	auxiliary	auxiliary	ADJ
fcis-30157	195	56	diagnosis	diagnosis	NOUN
fcis-30157	196	1	[	[	X
fcis-30157	196	2	21	21	NUM
fcis-30157	196	3	]	]	PUNCT
fcis-30157	196	4	.	.	PUNCT
fcis-30157	197	1	hospital	hospital	NOUN
fcis-30157	197	2	data	datum	NOUN
fcis-30157	197	3	is	be	AUX
fcis-30157	197	4	more	more	ADV
fcis-30157	197	5	clinically	clinically	ADV
fcis-30157	197	6	relevant	relevant	ADJ
fcis-30157	197	7	but	but	CCONJ
fcis-30157	197	8	difficult	difficult	ADJ
fcis-30157	197	9	to	to	PART
fcis-30157	197	10	obtain	obtain	VERB
fcis-30157	197	11	and	and	CCONJ
fcis-30157	197	12	may	may	AUX
fcis-30157	197	13	suffer	suffer	VERB
fcis-30157	197	14	from	from	ADP
fcis-30157	197	15	bias	bias	NOUN
fcis-30157	197	16	,	,	PUNCT
fcis-30157	197	17	with	with	ADP
fcis-30157	197	18	more	more	ADJ
fcis-30157	197	19	common	common	ADJ
fcis-30157	197	20	diseases	disease	NOUN
fcis-30157	197	21	represented	represent	VERB
fcis-30157	197	22	and	and	CCONJ
fcis-30157	197	23	fewer	few	ADJ
fcis-30157	197	24	samples	sample	NOUN
fcis-30157	197	25	of	of	ADP
fcis-30157	197	26	rare	rare	ADJ
fcis-30157	197	27	diseases	disease	NOUN
fcis-30157	197	28	.	.	PUNCT
fcis-30157	198	1	public	public	ADJ
fcis-30157	198	2	databases	database	NOUN
fcis-30157	198	3	,	,	PUNCT
fcis-30157	198	4	though	though	SCONJ
fcis-30157	198	5	balanced	balanced	ADJ
fcis-30157	198	6	and	and	CCONJ
fcis-30157	198	7	standardized	standardize	VERB
fcis-30157	198	8	,	,	PUNCT
fcis-30157	198	9	often	often	ADV
fcis-30157	198	10	lack	lack	VERB
fcis-30157	198	11	the	the	DET
fcis-30157	198	12	necessary	necessary	ADJ
fcis-30157	198	13	clinical	clinical	ADJ
fcis-30157	198	14	details	detail	NOUN
fcis-30157	198	15	.	.	PUNCT
fcis-30157	199	1	to	to	PART
fcis-30157	199	2	address	address	VERB
fcis-30157	199	3	this	this	PRON
fcis-30157	199	4	,	,	PUNCT
fcis-30157	199	5	researchers	researcher	NOUN
fcis-30157	199	6	combine	combine	VERB
fcis-30157	199	7	hospital	hospital	NOUN
fcis-30157	199	8	and	and	CCONJ
fcis-30157	199	9	public	public	ADJ
fcis-30157	199	10	data	datum	NOUN
fcis-30157	199	11	sources	source	NOUN
fcis-30157	199	12	,	,	PUNCT
fcis-30157	199	13	enhancing	enhance	VERB
fcis-30157	199	14	model	model	NOUN
fcis-30157	199	15	generalization	generalization	NOUN
fcis-30157	199	16	and	and	CCONJ
fcis-30157	199	17	data	datum	NOUN
fcis-30157	199	18	diversity	diversity	NOUN
fcis-30157	199	19	,	,	PUNCT
fcis-30157	199	20	but	but	CCONJ
fcis-30157	199	21	challenges	challenge	NOUN
fcis-30157	199	22	like	like	ADP
fcis-30157	199	23	inconsistent	inconsistent	ADJ
fcis-30157	199	24	formats	format	NOUN
fcis-30157	199	25	and	and	CCONJ
fcis-30157	199	26	different	different	ADJ
fcis-30157	199	27	annotation	annotation	NOUN
fcis-30157	199	28	standards	standard	NOUN
fcis-30157	199	29	arise	arise	VERB
fcis-30157	199	30	during	during	ADP
fcis-30157	199	31	integration	integration	NOUN
fcis-30157	199	32	[	[	X
fcis-30157	199	33	22	22	NUM
fcis-30157	199	34	]	]	PUNCT
fcis-30157	199	35	.	.	PUNCT
fcis-30157	200	1	research	research	NOUN
fcis-30157	200	2	indicates	indicate	VERB
fcis-30157	200	3	models	model	NOUN
fcis-30157	200	4	trained	train	VERB
fcis-30157	200	5	on	on	ADP
fcis-30157	200	6	real	real	ADJ
fcis-30157	200	7	patient	patient	ADJ
fcis-30157	200	8	images	image	NOUN
fcis-30157	200	9	outperform	outperform	VERB
fcis-30157	200	10	those	those	PRON
fcis-30157	200	11	using	use	VERB
fcis-30157	200	12	synthetic	synthetic	ADJ
fcis-30157	200	13	data	datum	NOUN
fcis-30157	200	14	by	by	ADP
fcis-30157	200	15	15	15	NUM
fcis-30157	200	16	%	%	NOUN
fcis-30157	200	17	in	in	ADP
fcis-30157	200	18	accuracy	accuracy	NOUN
fcis-30157	200	19	[	[	X
fcis-30157	200	20	23	23	NUM
fcis-30157	200	21	]	]	PUNCT
fcis-30157	200	22	.	.	PUNCT
fcis-30157	201	1	however	however	ADV
fcis-30157	201	2	,	,	PUNCT
fcis-30157	201	3	real	real	ADJ
fcis-30157	201	4	patient	patient	ADJ
fcis-30157	201	5	image	image	NOUN
fcis-30157	201	6	acquisition	acquisition	NOUN
fcis-30157	201	7	is	be	AUX
fcis-30157	201	8	expensive	expensive	ADJ
fcis-30157	201	9	and	and	CCONJ
fcis-30157	201	10	poses	pose	VERB
fcis-30157	201	11	privacy	privacy	NOUN
fcis-30157	201	12	concerns	concern	NOUN
fcis-30157	201	13	,	,	PUNCT
fcis-30157	201	14	prompting	prompt	VERB
fcis-30157	201	15	studies	study	NOUN
fcis-30157	201	16	to	to	PART
fcis-30157	201	17	explore	explore	VERB
fcis-30157	201	18	synthetic	synthetic	ADJ
fcis-30157	201	19	data	datum	NOUN
fcis-30157	201	20	,	,	PUNCT
fcis-30157	201	21	such	such	ADJ
fcis-30157	201	22	as	as	ADP
fcis-30157	201	23	generating	generate	VERB
fcis-30157	201	24	10,000	10,000	NUM
fcis-30157	201	25	skin	skin	NOUN
fcis-30157	201	26	disease	disease	NOUN
fcis-30157	201	27	images	image	NOUN
fcis-30157	201	28	using	use	VERB
fcis-30157	201	29	generative	generative	ADJ
fcis-30157	201	30	adversarial	adversarial	ADJ
fcis-30157	201	31	networks	network	NOUN
fcis-30157	201	32	(	(	PUNCT
fcis-30157	201	33	gans	gan	NOUN
fcis-30157	201	34	)	)	PUNCT
fcis-30157	202	1	[	[	X
fcis-30157	202	2	24	24	NUM
fcis-30157	202	3	]	]	PUNCT
fcis-30157	202	4	.	.	PUNCT
fcis-30157	203	1	while	while	SCONJ
fcis-30157	203	2	synthetic	synthetic	ADJ
fcis-30157	203	3	data	datum	NOUN
fcis-30157	203	4	can	can	AUX
fcis-30157	203	5	help	help	VERB
fcis-30157	203	6	with	with	ADP
fcis-30157	203	7	insufficient	insufficient	ADJ
fcis-30157	203	8	data	datum	NOUN
fcis-30157	203	9	,	,	PUNCT
fcis-30157	203	10	differences	difference	NOUN
fcis-30157	203	11	from	from	ADP
fcis-30157	203	12	real	real	ADJ
fcis-30157	203	13	images	image	NOUN
fcis-30157	203	14	remain	remain	VERB
fcis-30157	203	15	an	an	DET
fcis-30157	203	16	area	area	NOUN
fcis-30157	203	17	of	of	ADP
fcis-30157	203	18	ongoing	ongoing	ADJ
fcis-30157	203	19	research	research	NOUN
fcis-30157	203	20	.	.	PUNCT
fcis-30157	204	1	copyright	copyright	NOUN
fcis-30157	204	2	and	and	CCONJ
fcis-30157	204	3	privacy	privacy	NOUN
fcis-30157	204	4	issues	issue	NOUN
fcis-30157	204	5	are	be	AUX
fcis-30157	204	6	critical	critical	ADJ
fcis-30157	204	7	in	in	ADP
fcis-30157	204	8	data	datum	NOUN
fcis-30157	204	9	acquisition	acquisition	NOUN
fcis-30157	204	10	.	.	PUNCT
fcis-30157	205	1	public	public	ADJ
fcis-30157	205	2	databases	database	NOUN
fcis-30157	205	3	often	often	ADV
fcis-30157	205	4	have	have	VERB
fcis-30157	205	5	licenses	license	NOUN
fcis-30157	205	6	like	like	ADP
fcis-30157	205	7	the	the	DET
fcis-30157	205	8	cc	cc	PROPN
fcis-30157	205	9	by	by	ADP
fcis-30157	205	10	-	-	PUNCT
fcis-30157	205	11	nc	nc	PROPN
fcis-30157	205	12	4.0	4.0	NUM
fcis-30157	205	13	protocol	protocol	NOUN
fcis-30157	205	14	,	,	PUNCT
fcis-30157	205	15	limiting	limit	VERB
fcis-30157	205	16	commercial	commercial	ADJ
fcis-30157	205	17	use	use	NOUN
fcis-30157	205	18	[	[	X
fcis-30157	205	19	25	25	NUM
fcis-30157	205	20	]	]	PUNCT
fcis-30157	205	21	,	,	PUNCT
fcis-30157	205	22	while	while	SCONJ
fcis-30157	205	23	hospital	hospital	NOUN
fcis-30157	205	24	data	datum	NOUN
fcis-30157	205	25	requires	require	VERB
fcis-30157	205	26	legal	legal	ADJ
fcis-30157	205	27	agreements	agreement	NOUN
fcis-30157	205	28	,	,	PUNCT
fcis-30157	205	29	such	such	ADJ
fcis-30157	205	30	as	as	ADP
fcis-30157	205	31	data	data	NOUN
fcis-30157	205	32	-	-	PUNCT
fcis-30157	205	33	sharing	share	VERB
fcis-30157	205	34	contracts	contract	NOUN
fcis-30157	205	35	,	,	PUNCT
fcis-30157	205	36	to	to	PART
fcis-30157	205	37	ensure	ensure	VERB
fcis-30157	205	38	legitimacy	legitimacy	NOUN
fcis-30157	205	39	and	and	CCONJ
fcis-30157	205	40	research	research	NOUN
fcis-30157	205	41	support	support	NOUN
fcis-30157	206	1	[	[	X
fcis-30157	206	2	26	26	NUM
fcis-30157	206	3	]	]	PUNCT
fcis-30157	206	4	.	.	PUNCT
fcis-30157	207	1	2.1.2	2.1.2	X
fcis-30157	207	2	.	.	PUNCT
fcis-30157	207	3	data	datum	NOUN
fcis-30157	207	4	cleaning	cleaning	NOUN
fcis-30157	207	5	and	and	CCONJ
fcis-30157	207	6	normalization	normalization	NOUN
fcis-30157	207	7	data	datum	NOUN
fcis-30157	207	8	cleaning	cleaning	NOUN
fcis-30157	207	9	and	and	CCONJ
fcis-30157	207	10	normalization	normalization	NOUN
fcis-30157	207	11	are	be	AUX
fcis-30157	207	12	essential	essential	ADJ
fcis-30157	207	13	for	for	ADP
fcis-30157	207	14	preparing	prepare	VERB
fcis-30157	207	15	skin	skin	NOUN
fcis-30157	207	16	disease	disease	NOUN
fcis-30157	207	17	image	image	NOUN
fcis-30157	207	18	data	datum	NOUN
fcis-30157	207	19	for	for	ADP
fcis-30157	207	20	model	model	NOUN
fcis-30157	207	21	training	training	NOUN
fcis-30157	207	22	.	.	PUNCT
fcis-30157	208	1	the	the	DET
fcis-30157	208	2	process	process	NOUN
fcis-30157	208	3	begins	begin	VERB
fcis-30157	208	4	by	by	ADP
fcis-30157	208	5	removing	remove	VERB
fcis-30157	208	6	noise	noise	NOUN
fcis-30157	208	7	,	,	PUNCT
fcis-30157	208	8	duplicates	duplicate	NOUN
fcis-30157	208	9	,	,	PUNCT
fcis-30157	208	10	and	and	CCONJ
fcis-30157	208	11	low	low	ADJ
fcis-30157	208	12	-	-	PUNCT
fcis-30157	208	13	quality	quality	NOUN
fcis-30157	208	14	images	image	NOUN
fcis-30157	208	15	.	.	PUNCT
fcis-30157	209	1	noise	noise	NOUN
fcis-30157	209	2	can	can	AUX
fcis-30157	209	3	result	result	VERB
fcis-30157	209	4	from	from	ADP
fcis-30157	209	5	camera	camera	NOUN
fcis-30157	209	6	artifacts	artifact	NOUN
fcis-30157	209	7	or	or	CCONJ
fcis-30157	209	8	environmental	environmental	ADJ
fcis-30157	209	9	factors	factor	NOUN
fcis-30157	209	10	,	,	PUNCT
fcis-30157	209	11	which	which	PRON
fcis-30157	209	12	may	may	AUX
fcis-30157	209	13	negatively	negatively	ADV
fcis-30157	209	14	affect	affect	VERB
fcis-30157	209	15	model	model	NOUN
fcis-30157	209	16	performance	performance	NOUN
fcis-30157	209	17	.	.	PUNCT
fcis-30157	210	1	for	for	ADP
fcis-30157	210	2	example	example	NOUN
fcis-30157	210	3	,	,	PUNCT
fcis-30157	210	4	in	in	ADP
fcis-30157	210	5	skin	skin	NOUN
fcis-30157	210	6	cancer	cancer	NOUN
fcis-30157	210	7	classification	classification	NOUN
fcis-30157	210	8	,	,	PUNCT
fcis-30157	210	9	about	about	ADV
fcis-30157	210	10	15	15	NUM
fcis-30157	210	11	%	%	NOUN
fcis-30157	210	12	of	of	ADP
fcis-30157	210	13	images	image	NOUN
fcis-30157	210	14	contained	contain	VERB
fcis-30157	210	15	noise	noise	NOUN
fcis-30157	210	16	,	,	PUNCT
fcis-30157	210	17	reducing	reduce	VERB
fcis-30157	210	18	accuracy	accuracy	NOUN
fcis-30157	210	19	.	.	PUNCT
fcis-30157	211	1	filtering	filter	VERB
fcis-30157	211	2	techniques	technique	NOUN
fcis-30157	211	3	like	like	ADP
fcis-30157	211	4	gaussian	gaussian	NOUN
fcis-30157	211	5	blur	blur	NOUN
fcis-30157	211	6	and	and	CCONJ
fcis-30157	211	7	median	median	ADJ
fcis-30157	211	8	filtering	filtering	NOUN
fcis-30157	211	9	can	can	AUX
fcis-30157	211	10	help	help	VERB
fcis-30157	211	11	mitigate	mitigate	VERB
fcis-30157	211	12	this	this	DET
fcis-30157	211	13	issue	issue	NOUN
fcis-30157	211	14	.	.	PUNCT
fcis-30157	212	1	duplicate	duplicate	ADJ
fcis-30157	212	2	images	image	NOUN
fcis-30157	212	3	,	,	PUNCT
fcis-30157	212	4	which	which	PRON
fcis-30157	212	5	can	can	AUX
fcis-30157	212	6	overrepresent	overrepresent	VERB
fcis-30157	212	7	certain	certain	ADJ
fcis-30157	212	8	cases	case	NOUN
fcis-30157	212	9	,	,	PUNCT
fcis-30157	212	10	are	be	AUX
fcis-30157	212	11	also	also	ADV
fcis-30157	212	12	common	common	ADJ
fcis-30157	212	13	,	,	PUNCT
fcis-30157	212	14	such	such	ADJ
fcis-30157	212	15	as	as	ADP
fcis-30157	212	16	in	in	ADP
fcis-30157	212	17	a	a	DET
fcis-30157	212	18	melanoma	melanoma	NOUN
fcis-30157	212	19	dataset	dataset	VERB
fcis-30157	212	20	where	where	SCONJ
fcis-30157	212	21	10	10	NUM
fcis-30157	212	22	%	%	NOUN
fcis-30157	212	23	of	of	ADP
fcis-30157	212	24	the	the	DET
fcis-30157	212	25	images	image	NOUN
fcis-30157	212	26	were	be	AUX
fcis-30157	212	27	duplicates	duplicate	NOUN
fcis-30157	212	28	,	,	PUNCT
fcis-30157	212	29	and	and	CCONJ
fcis-30157	212	30	these	these	PRON
fcis-30157	212	31	were	be	AUX
fcis-30157	212	32	removed	remove	VERB
fcis-30157	212	33	to	to	PART
fcis-30157	212	34	ensure	ensure	VERB
fcis-30157	212	35	fair	fair	ADJ
fcis-30157	212	36	representation	representation	NOUN
fcis-30157	212	37	.	.	PUNCT
fcis-30157	213	1	excluding	exclude	VERB
fcis-30157	213	2	low	low	ADJ
fcis-30157	213	3	-	-	PUNCT
fcis-30157	213	4	quality	quality	NOUN
fcis-30157	213	5	images	image	NOUN
fcis-30157	213	6	with	with	ADP
fcis-30157	213	7	poor	poor	ADJ
fcis-30157	213	8	resolution	resolution	NOUN
fcis-30157	213	9	or	or	CCONJ
fcis-30157	213	10	uneven	uneven	ADJ
fcis-30157	213	11	lighting	lighting	NOUN
fcis-30157	213	12	can	can	AUX
fcis-30157	213	13	improve	improve	VERB
fcis-30157	213	14	accuracy	accuracy	NOUN
fcis-30157	213	15	by	by	ADP
fcis-30157	213	16	up	up	ADP
fcis-30157	213	17	to	to	PART
fcis-30157	213	18	8	8	NUM
fcis-30157	213	19	%	%	NOUN
fcis-30157	213	20	.	.	PUNCT
fcis-30157	214	1	normalization	normalization	NOUN
fcis-30157	214	2	follows	follow	VERB
fcis-30157	214	3	cleaning	clean	VERB
fcis-30157	214	4	,	,	PUNCT
fcis-30157	214	5	ensuring	ensure	VERB
fcis-30157	214	6	images	image	NOUN
fcis-30157	214	7	are	be	AUX
fcis-30157	214	8	compatible	compatible	ADJ
fcis-30157	214	9	with	with	ADP
fcis-30157	214	10	the	the	DET
fcis-30157	214	11	model	model	NOUN
fcis-30157	214	12	’s	’s	PART
fcis-30157	214	13	requirements	requirement	NOUN
fcis-30157	214	14	.	.	PUNCT
fcis-30157	215	1	this	this	PRON
fcis-30157	215	2	includes	include	VERB
fcis-30157	215	3	unifying	unify	VERB
fcis-30157	215	4	image	image	NOUN
fcis-30157	215	5	size	size	NOUN
fcis-30157	215	6	(	(	PUNCT
fcis-30157	215	7	e.g.	e.g.	ADV
fcis-30157	215	8	,	,	PUNCT
fcis-30157	215	9	resizing	resize	VERB
fcis-30157	215	10	to	to	ADP
fcis-30157	215	11	224x224	224x224	NUM
fcis-30157	215	12	pixels	pixel	NOUN
fcis-30157	215	13	)	)	PUNCT
fcis-30157	215	14	,	,	PUNCT
fcis-30157	215	15	color	color	NOUN
fcis-30157	215	16	space	space	NOUN
fcis-30157	215	17	(	(	PUNCT
fcis-30157	215	18	e.g.	e.g.	ADV
fcis-30157	215	19	,	,	PUNCT
fcis-30157	215	20	converting	convert	VERB
fcis-30157	215	21	to	to	ADP
fcis-30157	215	22	rgb	rgb	PROPN
fcis-30157	215	23	)	)	PUNCT
fcis-30157	215	24	,	,	PUNCT
fcis-30157	215	25	and	and	CCONJ
fcis-30157	215	26	brightness	brightness	NOUN
fcis-30157	215	27	/	/	SYM
fcis-30157	215	28	contrast	contrast	NOUN
fcis-30157	215	29	.	.	PUNCT
fcis-30157	216	1	for	for	ADP
fcis-30157	216	2	instance	instance	NOUN
fcis-30157	216	3	,	,	PUNCT
fcis-30157	216	4	histogram	histogram	NOUN
fcis-30157	216	5	equalization	equalization	NOUN
fcis-30157	216	6	has	have	AUX
fcis-30157	216	7	been	be	AUX
fcis-30157	216	8	shown	show	VERB
fcis-30157	216	9	to	to	PART
fcis-30157	216	10	improve	improve	VERB
fcis-30157	216	11	model	model	NOUN
fcis-30157	216	12	accuracy	accuracy	NOUN
fcis-30157	216	13	by	by	ADP
fcis-30157	216	14	6	6	NUM
fcis-30157	216	15	%	%	NOUN
fcis-30157	216	16	.	.	PUNCT
fcis-30157	217	1	techniques	technique	NOUN
fcis-30157	217	2	like	like	ADP
fcis-30157	217	3	clahe	clahe	NOUN
fcis-30157	217	4	(	(	PUNCT
fcis-30157	217	5	contrast	contrast	VERB
fcis-30157	217	6	limited	limit	VERB
fcis-30157	217	7	adaptive	adaptive	ADJ
fcis-30157	217	8	histogram	histogram	NOUN
fcis-30157	217	9	equalization	equalization	NOUN
fcis-30157	217	10	)	)	PUNCT
fcis-30157	217	11	enhance	enhance	VERB
fcis-30157	217	12	local	local	ADJ
fcis-30157	217	13	contrast	contrast	NOUN
fcis-30157	217	14	while	while	SCONJ
fcis-30157	217	15	minimizing	minimize	VERB
fcis-30157	217	16	noise	noise	NOUN
fcis-30157	217	17	.	.	PUNCT
fcis-30157	218	1	data	datum	NOUN
fcis-30157	218	2	augmentation	augmentation	NOUN
fcis-30157	218	3	,	,	PUNCT
fcis-30157	218	4	such	such	ADJ
fcis-30157	218	5	as	as	ADP
fcis-30157	218	6	rotating	rotate	VERB
fcis-30157	218	7	or	or	CCONJ
fcis-30157	218	8	flipping	flipping	NOUN
fcis-30157	218	9	images	image	NOUN
fcis-30157	218	10	,	,	PUNCT
fcis-30157	218	11	increases	increase	VERB
fcis-30157	218	12	dataset	dataset	ADJ
fcis-30157	218	13	diversity	diversity	NOUN
fcis-30157	218	14	by	by	ADP
fcis-30157	218	15	up	up	ADP
fcis-30157	218	16	to	to	PART
fcis-30157	218	17	300	300	NUM
fcis-30157	218	18	%	%	NOUN
fcis-30157	218	19	,	,	PUNCT
fcis-30157	218	20	improving	improve	VERB
fcis-30157	218	21	model	model	NOUN
fcis-30157	218	22	generalization	generalization	NOUN
fcis-30157	218	23	[	[	X
fcis-30157	218	24	21	21	NUM
fcis-30157	218	25	]	]	PUNCT
fcis-30157	218	26	.	.	PUNCT
fcis-30157	219	1	normalization	normalization	NOUN
fcis-30157	219	2	also	also	ADV
fcis-30157	219	3	involves	involve	VERB
fcis-30157	219	4	preprocessing	preprocesse	VERB
fcis-30157	219	5	specific	specific	ADJ
fcis-30157	219	6	to	to	ADP
fcis-30157	219	7	the	the	DET
fcis-30157	219	8	model	model	NOUN
fcis-30157	219	9	architecture	architecture	NOUN
fcis-30157	219	10	,	,	PUNCT
fcis-30157	219	11	like	like	ADP
fcis-30157	219	12	normalizing	normalize	VERB
fcis-30157	219	13	pixel	pixel	ADJ
fcis-30157	219	14	values	value	NOUN
fcis-30157	219	15	to	to	ADP
fcis-30157	219	16	[	[	X
fcis-30157	219	17	0	0	NUM
fcis-30157	219	18	,	,	PUNCT
fcis-30157	219	19	1	1	NUM
fcis-30157	219	20	]	]	PUNCT
fcis-30157	219	21	for	for	ADP
fcis-30157	219	22	cnns	cnn	NOUN
fcis-30157	219	23	,	,	PUNCT
fcis-30157	219	24	reducing	reduce	VERB
fcis-30157	219	25	training	training	NOUN
fcis-30157	219	26	time	time	NOUN
fcis-30157	219	27	by	by	ADP
fcis-30157	219	28	20	20	NUM
fcis-30157	219	29	%	%	NOUN
fcis-30157	219	30	.	.	PUNCT
fcis-30157	220	1	accounting	account	VERB
fcis-30157	220	2	for	for	ADP
fcis-30157	220	3	skin	skin	NOUN
fcis-30157	220	4	tone	tone	NOUN
fcis-30157	220	5	and	and	CCONJ
fcis-30157	220	6	texture	texture	ADJ
fcis-30157	220	7	variations	variation	NOUN
fcis-30157	220	8	is	be	AUX
fcis-30157	220	9	crucial	crucial	ADJ
fcis-30157	220	10	to	to	PART
fcis-30157	220	11	avoid	avoid	VERB
fcis-30157	220	12	model	model	NOUN
fcis-30157	220	13	bias	bias	NOUN
fcis-30157	220	14	.	.	PUNCT
fcis-30157	221	1	adaptive	adaptive	ADJ
fcis-30157	221	2	normalization	normalization	NOUN
fcis-30157	221	3	techniques	technique	NOUN
fcis-30157	221	4	adjust	adjust	VERB
fcis-30157	221	5	parameters	parameter	NOUN
fcis-30157	221	6	based	base	VERB
fcis-30157	221	7	on	on	ADP
fcis-30157	221	8	local	local	ADJ
fcis-30157	221	9	image	image	NOUN
fcis-30157	221	10	features	feature	NOUN
fcis-30157	221	11	,	,	PUNCT
fcis-30157	221	12	improving	improve	VERB
fcis-30157	221	13	the	the	DET
fcis-30157	221	14	model	model	NOUN
fcis-30157	221	15	’s	’s	PART
fcis-30157	221	16	performance	performance	NOUN
fcis-30157	221	17	across	across	ADP
fcis-30157	221	18	diverse	diverse	ADJ
fcis-30157	221	19	skin	skin	NOUN
fcis-30157	221	20	types	type	NOUN
fcis-30157	221	21	.	.	PUNCT
fcis-30157	222	1	2.1.3	2.1.3	NUM
fcis-30157	222	2	.	.	PUNCT
fcis-30157	222	3	data	datum	NOUN
fcis-30157	222	4	augmentation	augmentation	NOUN
fcis-30157	222	5	techniques	technique	NOUN
fcis-30157	222	6	data	datum	NOUN
fcis-30157	222	7	augmentation	augmentation	NOUN
fcis-30157	222	8	is	be	AUX
fcis-30157	222	9	vital	vital	ADJ
fcis-30157	222	10	for	for	ADP
fcis-30157	222	11	enhancing	enhance	VERB
fcis-30157	222	12	deep	deep	ADJ
fcis-30157	222	13	learning	learning	NOUN
fcis-30157	222	14	models	model	NOUN
fcis-30157	222	15	,	,	PUNCT
fcis-30157	222	16	especially	especially	ADV
fcis-30157	222	17	in	in	ADP
fcis-30157	222	18	skin	skin	NOUN
fcis-30157	222	19	disease	disease	NOUN
fcis-30157	222	20	image	image	NOUN
fcis-30157	222	21	classification	classification	NOUN
fcis-30157	222	22	.	.	PUNCT
fcis-30157	223	1	techniques	technique	NOUN
fcis-30157	223	2	such	such	ADJ
fcis-30157	223	3	as	as	ADP
fcis-30157	223	4	rotation	rotation	NOUN
fcis-30157	223	5	,	,	PUNCT
fcis-30157	223	6	flipping	flip	VERB
fcis-30157	223	7	,	,	PUNCT
fcis-30157	223	8	scaling	scaling	NOUN
fcis-30157	223	9	,	,	PUNCT
fcis-30157	223	10	and	and	CCONJ
fcis-30157	223	11	cropping	crop	VERB
fcis-30157	223	12	help	help	NOUN
fcis-30157	223	13	expand	expand	VERB
fcis-30157	223	14	the	the	DET
fcis-30157	223	15	dataset	dataset	NOUN
fcis-30157	223	16	by	by	ADP
fcis-30157	223	17	introducing	introduce	VERB
fcis-30157	223	18	variations	variation	NOUN
fcis-30157	223	19	,	,	PUNCT
fcis-30157	223	20	improving	improve	VERB
fcis-30157	223	21	model	model	NOUN
fcis-30157	223	22	performance	performance	NOUN
fcis-30157	223	23	and	and	CCONJ
fcis-30157	223	24	generalization	generalization	NOUN
fcis-30157	223	25	.	.	PUNCT
fcis-30157	224	1	rotation	rotation	NOUN
fcis-30157	224	2	involves	involve	VERB
fcis-30157	224	3	rotating	rotate	VERB
fcis-30157	224	4	images	image	NOUN
fcis-30157	224	5	to	to	PART
fcis-30157	224	6	expose	expose	VERB
fcis-30157	224	7	the	the	DET
fcis-30157	224	8	model	model	NOUN
fcis-30157	224	9	to	to	ADP
fcis-30157	224	10	different	different	ADJ
fcis-30157	224	11	orientations	orientation	NOUN
fcis-30157	224	12	.	.	PUNCT
fcis-30157	225	1	a	a	DET
fcis-30157	225	2	study	study	NOUN
fcis-30157	225	3	found	find	VERB
fcis-30157	225	4	that	that	SCONJ
fcis-30157	225	5	rotating	rotate	VERB
fcis-30157	225	6	images	image	NOUN
fcis-30157	225	7	between	between	ADP
fcis-30157	225	8	-30	-30	PROPN
fcis-30157	225	9	°	°	NUM
fcis-30157	225	10	and	and	CCONJ
fcis-30157	225	11	30	30	NUM
fcis-30157	225	12	°	°	NUM
fcis-30157	225	13	improved	improve	VERB
fcis-30157	225	14	lesion	lesion	NOUN
fcis-30157	225	15	recognition	recognition	NOUN
fcis-30157	225	16	.	.	PUNCT
fcis-30157	226	1	this	this	DET
fcis-30157	226	2	technique	technique	NOUN
fcis-30157	226	3	is	be	AUX
fcis-30157	226	4	particularly	particularly	ADV
fcis-30157	226	5	useful	useful	ADJ
fcis-30157	226	6	for	for	ADP
fcis-30157	226	7	datasets	dataset	NOUN
fcis-30157	226	8	with	with	ADP
fcis-30157	226	9	various	various	ADJ
fcis-30157	226	10	lesion	lesion	NOUN
fcis-30157	226	11	shapes	shape	NOUN
fcis-30157	226	12	.	.	PUNCT
fcis-30157	227	1	flipping	flip	VERB
fcis-30157	227	2	images	image	NOUN
fcis-30157	227	3	horizontally	horizontally	ADV
fcis-30157	227	4	or	or	CCONJ
fcis-30157	227	5	vertically	vertically	ADV
fcis-30157	227	6	helps	help	VERB
fcis-30157	227	7	the	the	DET
fcis-30157	227	8	model	model	NOUN
fcis-30157	227	9	generalize	generalize	VERB
fcis-30157	227	10	by	by	ADP
fcis-30157	227	11	simulating	simulate	VERB
fcis-30157	227	12	mirrored	mirror	VERB
fcis-30157	227	13	images	image	NOUN
fcis-30157	227	14	,	,	PUNCT
fcis-30157	227	15	with	with	ADP
fcis-30157	227	16	a	a	DET
fcis-30157	227	17	study	study	NOUN
fcis-30157	227	18	showing	show	VERB
fcis-30157	227	19	a	a	DET
fcis-30157	227	20	2.5	2.5	NUM
fcis-30157	227	21	%	%	NOUN
fcis-30157	227	22	accuracy	accuracy	NOUN
fcis-30157	227	23	increase	increase	NOUN
fcis-30157	227	24	for	for	ADP
fcis-30157	227	25	skin	skin	NOUN
fcis-30157	227	26	cancer	cancer	NOUN
fcis-30157	227	27	classification	classification	NOUN
fcis-30157	227	28	.	.	PUNCT
fcis-30157	228	1	however	however	ADV
fcis-30157	228	2	,	,	PUNCT
fcis-30157	228	3	vertical	vertical	ADJ
fcis-30157	228	4	flipping	flipping	NOUN
fcis-30157	228	5	might	might	AUX
fcis-30157	228	6	not	not	PART
fcis-30157	228	7	be	be	AUX
fcis-30157	228	8	suitable	suitable	ADJ
fcis-30157	228	9	when	when	SCONJ
fcis-30157	228	10	orientation	orientation	NOUN
fcis-30157	228	11	is	be	AUX
fcis-30157	228	12	crucial	crucial	ADJ
fcis-30157	228	13	.	.	PUNCT
fcis-30157	229	1	scaling	scale	VERB
fcis-30157	229	2	resizes	resize	VERB
fcis-30157	229	3	images	image	NOUN
fcis-30157	229	4	to	to	ADP
fcis-30157	229	5	various	various	ADJ
fcis-30157	229	6	sizes	size	NOUN
fcis-30157	229	7	while	while	SCONJ
fcis-30157	229	8	preserving	preserve	VERB
fcis-30157	229	9	aspect	aspect	NOUN
fcis-30157	229	10	ratios	ratio	NOUN
fcis-30157	229	11	.	.	PUNCT
fcis-30157	230	1	this	this	DET
fcis-30157	230	2	approach	approach	NOUN
fcis-30157	230	3	increased	increase	VERB
fcis-30157	230	4	recall	recall	NOUN
fcis-30157	230	5	by	by	ADP
fcis-30157	230	6	3.2	3.2	NUM
fcis-30157	230	7	%	%	NOUN
fcis-30157	230	8	in	in	ADP
fcis-30157	230	9	skin	skin	NOUN
fcis-30157	230	10	disease	disease	NOUN
fcis-30157	230	11	classification	classification	NOUN
fcis-30157	230	12	,	,	PUNCT
fcis-30157	230	13	helping	help	VERB
fcis-30157	230	14	the	the	DET
fcis-30157	230	15	model	model	NOUN
fcis-30157	230	16	handle	handle	VERB
fcis-30157	230	17	variations	variation	NOUN
fcis-30157	230	18	in	in	ADP
fcis-30157	230	19	lesion	lesion	NOUN
fcis-30157	230	20	size	size	NOUN
fcis-30157	230	21	.	.	PUNCT
fcis-30157	231	1	cropping	crop	VERB
fcis-30157	231	2	randomly	randomly	ADV
fcis-30157	231	3	selects	select	NOUN
fcis-30157	231	4	regions	region	NOUN
fcis-30157	231	5	of	of	ADP
fcis-30157	231	6	interest	interest	NOUN
fcis-30157	231	7	,	,	PUNCT
fcis-30157	231	8	improving	improve	VERB
fcis-30157	231	9	precision	precision	NOUN
fcis-30157	231	10	by	by	ADP
fcis-30157	231	11	4.1	4.1	NUM
fcis-30157	231	12	%	%	NOUN
fcis-30157	231	13	by	by	ADP
fcis-30157	231	14	focusing	focus	VERB
fcis-30157	231	15	the	the	DET
fcis-30157	231	16	model	model	NOUN
fcis-30157	231	17	on	on	ADP
fcis-30157	231	18	lesions	lesion	NOUN
fcis-30157	231	19	,	,	PUNCT
fcis-30157	231	20	though	though	SCONJ
fcis-30157	231	21	excessive	excessive	ADJ
fcis-30157	231	22	cropping	cropping	NOUN
fcis-30157	231	23	may	may	AUX
fcis-30157	231	24	result	result	VERB
fcis-30157	231	25	in	in	ADP
fcis-30157	231	26	lost	lost	ADJ
fcis-30157	231	27	details	detail	NOUN
fcis-30157	231	28	.	.	PUNCT
fcis-30157	232	1	combining	combine	VERB
fcis-30157	232	2	these	these	DET
fcis-30157	232	3	techniques	technique	NOUN
fcis-30157	232	4	often	often	ADV
fcis-30157	232	5	yields	yield	VERB
fcis-30157	232	6	the	the	DET
fcis-30157	232	7	best	good	ADJ
fcis-30157	232	8	results	result	NOUN
fcis-30157	232	9	.	.	PUNCT
fcis-30157	233	1	a	a	DET
fcis-30157	233	2	study	study	NOUN
fcis-30157	233	3	combining	combine	VERB
fcis-30157	233	4	rotation	rotation	NOUN
fcis-30157	233	5	,	,	PUNCT
fcis-30157	233	6	flipping	flip	VERB
fcis-30157	233	7	,	,	PUNCT
fcis-30157	233	8	scaling	scaling	NOUN
fcis-30157	233	9	,	,	PUNCT
fcis-30157	233	10	and	and	CCONJ
fcis-30157	233	11	cropping	cropping	NOUN
fcis-30157	233	12	achieved	achieve	VERB
fcis-30157	233	13	a	a	DET
fcis-30157	233	14	92.3	92.3	NUM
fcis-30157	233	15	%	%	NOUN
fcis-30157	233	16	classification	classification	NOUN
fcis-30157	233	17	accuracy	accuracy	NOUN
fcis-30157	233	18	.	.	PUNCT
fcis-30157	234	1	the	the	DET
fcis-30157	234	2	impact	impact	NOUN
fcis-30157	234	3	of	of	ADP
fcis-30157	234	4	these	these	DET
fcis-30157	234	5	augmentations	augmentation	NOUN
fcis-30157	234	6	on	on	ADP
fcis-30157	234	7	model	model	NOUN
fcis-30157	234	8	performance	performance	NOUN
fcis-30157	234	9	is	be	AUX
fcis-30157	234	10	evident	evident	ADJ
fcis-30157	234	11	in	in	ADP
fcis-30157	234	12	metrics	metric	NOUN
fcis-30157	234	13	like	like	ADP
fcis-30157	234	14	accuracy	accuracy	NOUN
fcis-30157	234	15	and	and	CCONJ
fcis-30157	234	16	recall	recall	NOUN
fcis-30157	234	17	,	,	PUNCT
fcis-30157	234	18	with	with	ADP
fcis-30157	234	19	studies	study	NOUN
fcis-30157	234	20	showing	show	VERB
fcis-30157	234	21	significant	significant	ADJ
fcis-30157	234	22	improvements	improvement	NOUN
fcis-30157	234	23	in	in	ADP
fcis-30157	234	24	both	both	PRON
fcis-30157	234	25	.	.	PUNCT
fcis-30157	235	1	to	to	PART
fcis-30157	235	2	determine	determine	VERB
fcis-30157	235	3	the	the	DET
fcis-30157	235	4	optimal	optimal	ADJ
fcis-30157	235	5	strategy	strategy	NOUN
fcis-30157	235	6	,	,	PUNCT
fcis-30157	235	7	systematic	systematic	ADJ
fcis-30157	235	8	evaluations	evaluation	NOUN
fcis-30157	235	9	or	or	CCONJ
fcis-30157	235	10	automated	automate	VERB
fcis-30157	235	11	techniques	technique	NOUN
fcis-30157	235	12	(	(	PUNCT
fcis-30157	235	13	e.g.	e.g.	ADV
fcis-30157	235	14	,	,	PUNCT
fcis-30157	235	15	reinforcement	reinforcement	NOUN
fcis-30157	235	16	learning	learning	NOUN
fcis-30157	235	17	)	)	PUNCT
fcis-30157	235	18	are	be	AUX
fcis-30157	235	19	useful	useful	ADJ
fcis-30157	235	20	.	.	PUNCT
fcis-30157	236	1	for	for	ADP
fcis-30157	236	2	example	example	NOUN
fcis-30157	236	3	,	,	PUNCT
fcis-30157	236	4	combining	combine	VERB
fcis-30157	236	5	rotation	rotation	NOUN
fcis-30157	236	6	and	and	CCONJ
fcis-30157	236	7	scaling	scale	VERB
fcis-30157	236	8	improved	improved	ADJ
fcis-30157	236	9	accuracy	accuracy	NOUN
fcis-30157	236	10	by	by	ADP
fcis-30157	236	11	6.2	6.2	NUM
fcis-30157	236	12	%	%	NOUN
fcis-30157	236	13	.	.	PUNCT
fcis-30157	237	1	automated	automate	VERB
fcis-30157	237	2	augmentation	augmentation	NOUN
fcis-30157	237	3	methods	method	NOUN
fcis-30157	237	4	can	can	AUX
fcis-30157	237	5	achieve	achieve	VERB
fcis-30157	237	6	tailored	tailored	ADJ
fcis-30157	237	7	strategies	strategy	NOUN
fcis-30157	237	8	for	for	ADP
fcis-30157	237	9	specific	specific	ADJ
fcis-30157	237	10	datasets	dataset	NOUN
fcis-30157	237	11	,	,	PUNCT
fcis-30157	237	12	with	with	ADP
fcis-30157	237	13	one	one	NUM
fcis-30157	237	14	study	study	NOUN
fcis-30157	237	15	achieving	achieve	VERB
fcis-30157	237	16	94.5	94.5	NUM
fcis-30157	237	17	%	%	NOUN
fcis-30157	237	18	accuracy	accuracy	NOUN
fcis-30157	237	19	using	use	VERB
fcis-30157	237	20	cnns	cnns	PROPN
fcis-30157	237	21	.	.	PUNCT
fcis-30157	238	1	data	datum	NOUN
fcis-30157	238	2	augmentation	augmentation	NOUN
fcis-30157	238	3	also	also	ADV
fcis-30157	238	4	improves	improve	VERB
fcis-30157	238	5	model	model	NOUN
fcis-30157	238	6	generalization	generalization	NOUN
fcis-30157	238	7	,	,	PUNCT
fcis-30157	238	8	helping	help	VERB
fcis-30157	238	9	models	model	NOUN
fcis-30157	238	10	perform	perform	VERB
fcis-30157	238	11	well	well	ADV
fcis-30157	238	12	on	on	ADP
fcis-30157	238	13	unseen	unseen	ADJ
fcis-30157	238	14	data	datum	NOUN
fcis-30157	238	15	.	.	PUNCT
fcis-30157	239	1	combining	combine	VERB
fcis-30157	239	2	augmentation	augmentation	NOUN
fcis-30157	239	3	with	with	ADP
fcis-30157	239	4	techniques	technique	NOUN
fcis-30157	239	5	like	like	ADP
fcis-30157	239	6	dropout	dropout	NOUN
fcis-30157	239	7	and	and	CCONJ
fcis-30157	239	8	regularization	regularization	NOUN
fcis-30157	239	9	can	can	AUX
fcis-30157	239	10	further	far	ADV
fcis-30157	239	11	reduce	reduce	VERB
fcis-30157	239	12	overfitting	overfitting	NOUN
fcis-30157	239	13	,	,	PUNCT
fcis-30157	239	14	as	as	SCONJ
fcis-30157	239	15	shown	show	VERB
fcis-30157	239	16	by	by	ADP
fcis-30157	239	17	a	a	DET
fcis-30157	239	18	7.8	7.8	NUM
fcis-30157	239	19	%	%	NOUN
fcis-30157	239	20	improvement	improvement	NOUN
fcis-30157	239	21	in	in	ADP
fcis-30157	239	22	generalization	generalization	NOUN
fcis-30157	239	23	ability	ability	NOUN
fcis-30157	239	24	.	.	PUNCT
fcis-30157	240	1	2.2	2.2	NUM
fcis-30157	240	2	.	.	PUNCT
fcis-30157	240	3	automatic	automatic	ADJ
fcis-30157	240	4	recognition	recognition	NOUN
fcis-30157	240	5	model	model	NOUN
fcis-30157	240	6	2.2.1	2.2.1	NUM
fcis-30157	240	7	.	.	PUNCT
fcis-30157	241	1	selection	selection	NOUN
fcis-30157	241	2	of	of	ADP
fcis-30157	241	3	deep	deep	ADJ
fcis-30157	241	4	learning	learning	NOUN
fcis-30157	241	5	architectures	architecture	VERB
fcis-30157	241	6	the	the	DET
fcis-30157	241	7	selection	selection	NOUN
fcis-30157	241	8	of	of	ADP
fcis-30157	241	9	deep	deep	ADJ
fcis-30157	241	10	learning	learning	NOUN
fcis-30157	241	11	architectures	architecture	NOUN
fcis-30157	241	12	is	be	AUX
fcis-30157	241	13	crucial	crucial	ADJ
fcis-30157	241	14	in	in	ADP
fcis-30157	241	15	developing	develop	VERB
fcis-30157	241	16	effective	effective	ADJ
fcis-30157	241	17	automatic	automatic	ADJ
fcis-30157	241	18	skin	skin	NOUN
fcis-30157	241	19	disease	disease	NOUN
fcis-30157	241	20	identification	identification	NOUN
fcis-30157	241	21	models	model	NOUN
fcis-30157	241	22	.	.	PUNCT
fcis-30157	242	1	convolutional	convolutional	ADJ
fcis-30157	242	2	neural	neural	ADJ
fcis-30157	242	3	networks	network	NOUN
fcis-30157	242	4	(	(	PUNCT
fcis-30157	242	5	cnns	cnns	PROPN
fcis-30157	242	6	)	)	PUNCT
fcis-30157	242	7	,	,	PUNCT
fcis-30157	242	8	such	such	ADJ
fcis-30157	242	9	as	as	ADP
fcis-30157	242	10	resnet	resnet	NOUN
fcis-30157	242	11	,	,	PUNCT
fcis-30157	242	12	vgg	vgg	NOUN
fcis-30157	242	13	,	,	PUNCT
fcis-30157	242	14	and	and	CCONJ
fcis-30157	242	15	inception	inception	NOUN
fcis-30157	242	16	,	,	PUNCT
fcis-30157	242	17	are	be	AUX
fcis-30157	242	18	commonly	commonly	ADV
fcis-30157	242	19	used	use	VERB
fcis-30157	242	20	in	in	ADP
fcis-30157	242	21	image	image	NOUN
fcis-30157	242	22	recognition	recognition	NOUN
fcis-30157	242	23	due	due	ADP
fcis-30157	242	24	to	to	ADP
fcis-30157	242	25	their	their	PRON
fcis-30157	242	26	ability	ability	NOUN
fcis-30157	242	27	to	to	PART
fcis-30157	242	28	extract	extract	VERB
fcis-30157	242	29	hierarchical	hierarchical	ADJ
fcis-30157	242	30	features	feature	NOUN
fcis-30157	242	31	from	from	ADP
fcis-30157	242	32	visual	visual	ADJ
fcis-30157	242	33	data	datum	NOUN
fcis-30157	242	34	.	.	PUNCT
fcis-30157	243	1	skin	skin	NOUN
fcis-30157	243	2	disease	disease	NOUN
fcis-30157	243	3	images	image	NOUN
fcis-30157	243	4	often	often	ADV
fcis-30157	243	5	feature	feature	VERB
fcis-30157	243	6	complex	complex	ADJ
fcis-30157	243	7	patterns	pattern	NOUN
fcis-30157	243	8	,	,	PUNCT
fcis-30157	243	9	such	such	ADJ
fcis-30157	243	10	as	as	ADP
fcis-30157	243	11	variations	variation	NOUN
fcis-30157	243	12	in	in	ADP
fcis-30157	243	13	color	color	NOUN
fcis-30157	243	14	,	,	PUNCT
fcis-30157	243	15	texture	texture	NOUN
fcis-30157	243	16	,	,	PUNCT
fcis-30157	243	17	and	and	CCONJ
fcis-30157	243	18	shape	shape	NOUN
fcis-30157	243	19	,	,	PUNCT
fcis-30157	243	20	which	which	PRON
fcis-30157	243	21	demand	demand	NOUN
fcis-30157	243	22	models	model	NOUN
fcis-30157	243	23	that	that	PRON
fcis-30157	243	24	can	can	AUX
fcis-30157	243	25	capture	capture	VERB
fcis-30157	243	26	both	both	CCONJ
fcis-30157	243	27	low	low	ADJ
fcis-30157	243	28	-	-	PUNCT
fcis-30157	243	29	level	level	NOUN
fcis-30157	243	30	and	and	CCONJ
fcis-30157	243	31	high	high	ADJ
fcis-30157	243	32	-	-	PUNCT
fcis-30157	243	33	level	level	NOUN
fcis-30157	243	34	features	feature	NOUN
fcis-30157	243	35	.	.	PUNCT
fcis-30157	244	1	resnet	resnet	NOUN
fcis-30157	244	2	,	,	PUNCT
fcis-30157	244	3	with	with	ADP
fcis-30157	244	4	its	its	PRON
fcis-30157	244	5	residual	residual	ADJ
fcis-30157	244	6	connections	connection	NOUN
fcis-30157	244	7	,	,	PUNCT
fcis-30157	244	8	addresses	address	VERB
fcis-30157	244	9	the	the	DET
fcis-30157	244	10	vanishing	vanish	VERB
fcis-30157	244	11	gradient	gradient	NOUN
fcis-30157	244	12	problem	problem	NOUN
fcis-30157	244	13	,	,	PUNCT
fcis-30157	244	14	allowing	allow	VERB
fcis-30157	244	15	deeper	deep	ADJ
fcis-30157	244	16	networks	network	NOUN
fcis-30157	244	17	that	that	PRON
fcis-30157	244	18	capture	capture	VERB
fcis-30157	244	19	more	more	ADV
fcis-30157	244	20	complex	complex	ADJ
fcis-30157	244	21	features	feature	NOUN
fcis-30157	244	22	.	.	PUNCT
fcis-30157	245	1	a	a	DET
fcis-30157	245	2	study	study	NOUN
fcis-30157	245	3	showed	show	VERB
fcis-30157	245	4	resnet	resnet	NOUN
fcis-30157	245	5	achieved	achieve	VERB
fcis-30157	245	6	92.3	92.3	NUM
fcis-30157	245	7	%	%	NOUN
fcis-30157	245	8	accuracy	accuracy	NOUN
fcis-30157	245	9	in	in	ADP
fcis-30157	245	10	skin	skin	NOUN
fcis-30157	245	11	disease	disease	NOUN
fcis-30157	245	12	classification	classification	NOUN
fcis-30157	245	13	.	.	PUNCT
fcis-30157	246	1	vgg	vgg	ADJ
fcis-30157	246	2	,	,	PUNCT
fcis-30157	246	3	simpler	simple	ADJ
fcis-30157	246	4	but	but	CCONJ
fcis-30157	246	5	less	less	ADV
fcis-30157	246	6	computationally	computationally	ADV
fcis-30157	246	7	efficient	efficient	ADJ
fcis-30157	246	8	,	,	PUNCT
fcis-30157	246	9	achieved	achieve	VERB
fcis-30157	246	10	89.7	89.7	NUM
fcis-30157	246	11	%	%	NOUN
fcis-30157	246	12	accuracy	accuracy	NOUN
fcis-30157	246	13	,	,	PUNCT
fcis-30157	246	14	making	make	VERB
fcis-30157	246	15	it	it	PRON
fcis-30157	246	16	suitable	suitable	ADJ
fcis-30157	246	17	for	for	ADP
fcis-30157	246	18	resource	resource	NOUN
fcis-30157	246	19	-	-	PUNCT
fcis-30157	246	20	constrained	constrain	VERB
fcis-30157	246	21	environments	environment	NOUN
fcis-30157	246	22	.	.	PUNCT
fcis-30157	247	1	inception	inception	ADJ
fcis-30157	247	2	,	,	PUNCT
fcis-30157	247	3	using	use	VERB
fcis-30157	247	4	multi	multi	ADJ
fcis-30157	247	5	-	-	ADJ
fcis-30157	247	6	scale	scale	ADJ
fcis-30157	247	7	convolutions	convolution	NOUN
fcis-30157	247	8	,	,	PUNCT
fcis-30157	247	9	demonstrated	demonstrate	VERB
fcis-30157	247	10	a	a	DET
fcis-30157	247	11	91.5	91.5	NUM
fcis-30157	247	12	%	%	NOUN
fcis-30157	247	13	accuracy	accuracy	NOUN
fcis-30157	247	14	,	,	PUNCT
fcis-30157	247	15	offering	offer	VERB
fcis-30157	247	16	a	a	DET
fcis-30157	247	17	good	good	ADJ
fcis-30157	247	18	balance	balance	NOUN
fcis-30157	247	19	in	in	ADP
fcis-30157	247	20	processing	process	VERB
fcis-30157	247	21	diverse	diverse	ADJ
fcis-30157	247	22	skin	skin	NOUN
fcis-30157	247	23	disease	disease	NOUN
fcis-30157	247	24	images	image	NOUN
fcis-30157	247	25	.	.	PUNCT
fcis-30157	248	1	however	however	ADV
fcis-30157	248	2	,	,	PUNCT
fcis-30157	248	3	these	these	DET
fcis-30157	248	4	architectures	architecture	NOUN
fcis-30157	248	5	have	have	VERB
fcis-30157	248	6	limitations	limitation	NOUN
fcis-30157	248	7	.	.	PUNCT
fcis-30157	249	1	resnet	resnet	PROPN
fcis-30157	249	2	can	can	AUX
fcis-30157	249	3	overfit	overfit	VERB
fcis-30157	249	4	with	with	ADP
fcis-30157	249	5	limited	limited	ADJ
fcis-30157	249	6	data	datum	NOUN
fcis-30157	249	7	,	,	PUNCT
fcis-30157	249	8	while	while	SCONJ
fcis-30157	249	9	vgg	vgg	PROPN
fcis-30157	249	10	's	's	PART
fcis-30157	249	11	simplicity	simplicity	NOUN
fcis-30157	249	12	may	may	AUX
fcis-30157	249	13	hinder	hinder	VERB
fcis-30157	249	14	capturing	capture	VERB
fcis-30157	249	15	specific	specific	ADJ
fcis-30157	249	16	features	feature	NOUN
fcis-30157	249	17	.	.	PUNCT
fcis-30157	250	1	hybrid	hybrid	ADJ
fcis-30157	250	2	models	model	NOUN
fcis-30157	250	3	,	,	PUNCT
fcis-30157	250	4	like	like	ADP
fcis-30157	250	5	mdfnet	mdfnet	NOUN
fcis-30157	250	6	,	,	PUNCT
fcis-30157	250	7	combine	combine	VERB
fcis-30157	250	8	cnns	cnn	VERB
fcis-30157	250	9	with	with	ADP
fcis-30157	250	10	multimodal	multimodal	ADJ
fcis-30157	250	11	data	datum	NOUN
fcis-30157	250	12	(	(	PUNCT
fcis-30157	250	13	e.g.	e.g.	ADV
fcis-30157	250	14	,	,	PUNCT
fcis-30157	250	15	images	image	NOUN
fcis-30157	250	16	and	and	CCONJ
fcis-30157	250	17	clinical	clinical	ADJ
fcis-30157	250	18	data	datum	NOUN
fcis-30157	250	19	)	)	PUNCT
fcis-30157	250	20	to	to	PART
fcis-30157	250	21	enhance	enhance	VERB
fcis-30157	250	22	classification	classification	NOUN
fcis-30157	250	23	accuracy	accuracy	NOUN
fcis-30157	250	24	.	.	PUNCT
fcis-30157	251	1	attention	attention	NOUN
fcis-30157	251	2	mechanisms	mechanism	NOUN
fcis-30157	251	3	,	,	PUNCT
fcis-30157	251	4	such	such	ADJ
fcis-30157	251	5	as	as	ADP
fcis-30157	251	6	dual	dual	ADJ
fcis-30157	251	7	attention	attention	NOUN
fcis-30157	251	8	networks	network	NOUN
fcis-30157	251	9	(	(	PUNCT
fcis-30157	251	10	dan	dan	PROPN
fcis-30157	251	11	)	)	PUNCT
fcis-30157	251	12	,	,	PUNCT
fcis-30157	251	13	have	have	AUX
fcis-30157	251	14	also	also	ADV
fcis-30157	251	15	shown	show	VERB
fcis-30157	251	16	promise	promise	NOUN
fcis-30157	251	17	,	,	PUNCT
fcis-30157	251	18	with	with	ADP
fcis-30157	251	19	dan	dan	PROPN
fcis-30157	251	20	achieving	achieve	VERB
fcis-30157	251	21	93.8	93.8	NUM
fcis-30157	251	22	%	%	NOUN
fcis-30157	251	23	accuracy	accuracy	NOUN
fcis-30157	251	24	by	by	ADP
fcis-30157	251	25	focusing	focus	VERB
fcis-30157	251	26	on	on	ADP
fcis-30157	251	27	relevant	relevant	ADJ
fcis-30157	251	28	image	image	NOUN
fcis-30157	251	29	features	feature	NOUN
fcis-30157	251	30	.	.	PUNCT
fcis-30157	252	1	explainable	explainable	ADJ
fcis-30157	252	2	ai	ai	PROPN
fcis-30157	252	3	(	(	PUNCT
fcis-30157	252	4	xai	xai	PROPN
fcis-30157	252	5	)	)	PUNCT
fcis-30157	252	6	,	,	PUNCT
fcis-30157	252	7	like	like	ADP
fcis-30157	252	8	grad	grad	NOUN
fcis-30157	252	9	-	-	PUNCT
fcis-30157	252	10	cam	cam	NOUN
fcis-30157	252	11	,	,	PUNCT
fcis-30157	252	12	improves	improve	VERB
fcis-30157	252	13	model	model	NOUN
fcis-30157	252	14	transparency	transparency	NOUN
fcis-30157	252	15	by	by	ADP
fcis-30157	252	16	providing	provide	VERB
fcis-30157	252	17	visual	visual	ADJ
fcis-30157	252	18	explanations	explanation	NOUN
fcis-30157	252	19	of	of	ADP
fcis-30157	252	20	predictions	prediction	NOUN
fcis-30157	252	21	.	.	PUNCT
fcis-30157	253	1	in	in	ADP
fcis-30157	253	2	skin	skin	NOUN
fcis-30157	253	3	lesion	lesion	NOUN
fcis-30157	253	4	classification	classification	NOUN
fcis-30157	253	5	,	,	PUNCT
fcis-30157	253	6	xaibased	xaibased	ADJ
fcis-30157	253	7	models	model	NOUN
fcis-30157	253	8	achieved	achieve	VERB
fcis-30157	253	9	90.5	90.5	NUM
fcis-30157	253	10	%	%	NOUN
fcis-30157	253	11	accuracy	accuracy	NOUN
fcis-30157	253	12	,	,	PUNCT
fcis-30157	253	13	offering	offer	VERB
fcis-30157	253	14	important	important	ADJ
fcis-30157	253	15	interpretability	interpretability	NOUN
fcis-30157	253	16	in	in	ADP
fcis-30157	253	17	medical	medical	ADJ
fcis-30157	253	18	applications	application	NOUN
fcis-30157	253	19	.	.	PUNCT
fcis-30157	254	1	finally	finally	ADV
fcis-30157	254	2	,	,	PUNCT
fcis-30157	254	3	the	the	DET
fcis-30157	254	4	choice	choice	NOUN
fcis-30157	254	5	of	of	ADP
fcis-30157	254	6	architecture	architecture	NOUN
fcis-30157	254	7	depends	depend	VERB
fcis-30157	254	8	on	on	ADP
fcis-30157	254	9	research	research	NOUN
fcis-30157	254	10	goals	goal	NOUN
fcis-30157	254	11	.	.	PUNCT
fcis-30157	255	1	models	model	NOUN
fcis-30157	255	2	like	like	ADP
fcis-30157	255	3	fs3dciot	fs3dciot	NOUN
fcis-30157	255	4	,	,	PUNCT
fcis-30157	255	5	which	which	PRON
fcis-30157	255	6	use	use	VERB
fcis-30157	255	7	few	few	ADJ
fcis-30157	255	8	-	-	PUNCT
fcis-30157	255	9	shot	shot	NOUN
fcis-30157	255	10	learning	learning	NOUN
fcis-30157	255	11	for	for	ADP
fcis-30157	255	12	diverse	diverse	ADJ
fcis-30157	255	13	populations	population	NOUN
fcis-30157	255	14	,	,	PUNCT
fcis-30157	255	15	are	be	AUX
fcis-30157	255	16	beneficial	beneficial	ADJ
fcis-30157	255	17	for	for	ADP
fcis-30157	255	18	generalizing	generalize	VERB
fcis-30157	255	19	across	across	ADP
fcis-30157	255	20	different	different	ADJ
fcis-30157	255	21	117	117	NUM
fcis-30157	255	22	skin	skin	NOUN
fcis-30157	255	23	tones	tone	NOUN
fcis-30157	255	24	.	.	PUNCT
fcis-30157	256	1	multimodal	multimodal	NOUN
fcis-30157	256	2	models	model	NOUN
fcis-30157	256	3	that	that	PRON
fcis-30157	256	4	integrate	integrate	VERB
fcis-30157	256	5	clinical	clinical	ADJ
fcis-30157	256	6	data	datum	NOUN
fcis-30157	256	7	can	can	AUX
fcis-30157	256	8	improve	improve	VERB
fcis-30157	256	9	accuracy	accuracy	NOUN
fcis-30157	256	10	by	by	ADP
fcis-30157	256	11	5.2	5.2	NUM
fcis-30157	256	12	%	%	NOUN
fcis-30157	256	13	compared	compare	VERB
fcis-30157	256	14	to	to	ADP
fcis-30157	256	15	image	image	NOUN
fcis-30157	256	16	-	-	PUNCT
fcis-30157	256	17	only	only	ADJ
fcis-30157	256	18	models	model	NOUN
fcis-30157	256	19	.	.	PUNCT
fcis-30157	257	1	in	in	ADP
fcis-30157	257	2	summary	summary	NOUN
fcis-30157	257	3	,	,	PUNCT
fcis-30157	257	4	selecting	select	VERB
fcis-30157	257	5	the	the	DET
fcis-30157	257	6	right	right	ADJ
fcis-30157	257	7	deep	deep	ADJ
fcis-30157	257	8	learning	learning	NOUN
fcis-30157	257	9	architecture	architecture	NOUN
fcis-30157	257	10	requires	require	VERB
fcis-30157	257	11	balancing	balance	VERB
fcis-30157	257	12	model	model	NOUN
fcis-30157	257	13	complexity	complexity	NOUN
fcis-30157	257	14	,	,	PUNCT
fcis-30157	257	15	computational	computational	ADJ
fcis-30157	257	16	efficiency	efficiency	NOUN
fcis-30157	257	17	,	,	PUNCT
fcis-30157	257	18	and	and	CCONJ
fcis-30157	257	19	the	the	DET
fcis-30157	257	20	ability	ability	NOUN
fcis-30157	257	21	to	to	PART
fcis-30157	257	22	capture	capture	VERB
fcis-30157	257	23	key	key	ADJ
fcis-30157	257	24	image	image	NOUN
fcis-30157	257	25	features	feature	NOUN
fcis-30157	257	26	,	,	PUNCT
fcis-30157	257	27	with	with	ADP
fcis-30157	257	28	hybrid	hybrid	NOUN
fcis-30157	257	29	and	and	CCONJ
fcis-30157	257	30	explainable	explainable	ADJ
fcis-30157	257	31	ai	ai	NOUN
fcis-30157	257	32	models	model	NOUN
fcis-30157	257	33	offering	offer	VERB
fcis-30157	257	34	promising	promising	ADJ
fcis-30157	257	35	directions	direction	NOUN
fcis-30157	257	36	for	for	ADP
fcis-30157	257	37	improved	improved	ADJ
fcis-30157	257	38	performance	performance	NOUN
fcis-30157	257	39	and	and	CCONJ
fcis-30157	257	40	interpretability	interpretability	NOUN
fcis-30157	257	41	.	.	PUNCT
fcis-30157	258	1	2.2.2	2.2.2	X
fcis-30157	258	2	.	.	PUNCT
fcis-30157	258	3	model	model	NOUN
fcis-30157	258	4	training	training	NOUN
fcis-30157	258	5	and	and	CCONJ
fcis-30157	258	6	optimization	optimization	NOUN
fcis-30157	258	7	model	model	NOUN
fcis-30157	258	8	training	training	NOUN
fcis-30157	258	9	and	and	CCONJ
fcis-30157	258	10	optimization	optimization	NOUN
fcis-30157	258	11	are	be	AUX
fcis-30157	258	12	essential	essential	ADJ
fcis-30157	258	13	in	in	ADP
fcis-30157	258	14	developing	develop	VERB
fcis-30157	258	15	an	an	DET
fcis-30157	258	16	effective	effective	ADJ
fcis-30157	258	17	automatic	automatic	ADJ
fcis-30157	258	18	skin	skin	NOUN
fcis-30157	258	19	disease	disease	NOUN
fcis-30157	258	20	recognition	recognition	NOUN
fcis-30157	258	21	model	model	NOUN
fcis-30157	258	22	.	.	PUNCT
fcis-30157	259	1	this	this	DET
fcis-30157	259	2	process	process	NOUN
fcis-30157	259	3	involves	involve	VERB
fcis-30157	259	4	selecting	select	VERB
fcis-30157	259	5	training	training	NOUN
fcis-30157	259	6	algorithms	algorithm	NOUN
fcis-30157	259	7	,	,	PUNCT
fcis-30157	259	8	configuring	configuring	NOUN
fcis-30157	259	9	hyperparameters	hyperparameter	NOUN
fcis-30157	259	10	,	,	PUNCT
fcis-30157	259	11	and	and	CCONJ
fcis-30157	259	12	applying	apply	VERB
fcis-30157	259	13	optimization	optimization	NOUN
fcis-30157	259	14	techniques	technique	NOUN
fcis-30157	259	15	to	to	PART
fcis-30157	259	16	improve	improve	VERB
fcis-30157	259	17	performance	performance	NOUN
fcis-30157	259	18	.	.	PUNCT
fcis-30157	260	1	training	training	NOUN
fcis-30157	260	2	algorithms	algorithm	NOUN
fcis-30157	260	3	:	:	PUNCT
fcis-30157	260	4	stochastic	stochastic	ADJ
fcis-30157	260	5	gradient	gradient	ADJ
fcis-30157	260	6	descent	descent	NOUN
fcis-30157	260	7	(	(	PUNCT
fcis-30157	260	8	sgd	sgd	PROPN
fcis-30157	260	9	)	)	PUNCT
fcis-30157	260	10	is	be	AUX
fcis-30157	260	11	a	a	DET
fcis-30157	260	12	popular	popular	ADJ
fcis-30157	260	13	optimization	optimization	NOUN
fcis-30157	260	14	algorithm	algorithm	NOUN
fcis-30157	260	15	that	that	PRON
fcis-30157	260	16	updates	update	VERB
fcis-30157	260	17	model	model	NOUN
fcis-30157	260	18	parameters	parameter	NOUN
fcis-30157	260	19	iteratively	iteratively	ADV
fcis-30157	260	20	.	.	PUNCT
fcis-30157	261	1	however	however	ADV
fcis-30157	261	2	,	,	PUNCT
fcis-30157	261	3	it	it	PRON
fcis-30157	261	4	is	be	AUX
fcis-30157	261	5	sensitive	sensitive	ADJ
fcis-30157	261	6	to	to	ADP
fcis-30157	261	7	the	the	DET
fcis-30157	261	8	learning	learning	NOUN
fcis-30157	261	9	rate	rate	NOUN
fcis-30157	261	10	.	.	PUNCT
fcis-30157	262	1	variants	variant	NOUN
fcis-30157	262	2	like	like	ADP
fcis-30157	262	3	momentum	momentum	NOUN
fcis-30157	262	4	and	and	CCONJ
fcis-30157	262	5	adam	adam	PROPN
fcis-30157	262	6	address	address	VERB
fcis-30157	262	7	this	this	DET
fcis-30157	262	8	issue	issue	NOUN
fcis-30157	262	9	.	.	PUNCT
fcis-30157	263	1	momentum	momentum	NOUN
fcis-30157	263	2	accelerates	accelerates	AUX
fcis-30157	263	3	sgd	sgd	VERB
fcis-30157	263	4	by	by	ADP
fcis-30157	263	5	accumulating	accumulate	VERB
fcis-30157	263	6	velocity	velocity	NOUN
fcis-30157	263	7	,	,	PUNCT
fcis-30157	263	8	while	while	SCONJ
fcis-30157	263	9	adam	adam	PROPN
fcis-30157	263	10	combines	combine	VERB
fcis-30157	263	11	momentum	momentum	NOUN
fcis-30157	263	12	with	with	ADP
fcis-30157	263	13	rmsprop	rmsprop	NOUN
fcis-30157	263	14	,	,	PUNCT
fcis-30157	263	15	adjusting	adjust	VERB
fcis-30157	263	16	the	the	DET
fcis-30157	263	17	learning	learning	NOUN
fcis-30157	263	18	rate	rate	NOUN
fcis-30157	263	19	adaptively	adaptively	ADV
fcis-30157	263	20	for	for	ADP
fcis-30157	263	21	each	each	DET
fcis-30157	263	22	parameter	parameter	NOUN
fcis-30157	263	23	,	,	PUNCT
fcis-30157	263	24	improving	improve	VERB
fcis-30157	263	25	convergence	convergence	NOUN
fcis-30157	263	26	.	.	PUNCT
fcis-30157	264	1	hyperparameter	hyperparameter	NOUN
fcis-30157	264	2	settings	setting	NOUN
fcis-30157	264	3	:	:	PUNCT
fcis-30157	264	4	the	the	DET
fcis-30157	264	5	choice	choice	NOUN
fcis-30157	264	6	of	of	ADP
fcis-30157	264	7	hyperparameters	hyperparameter	NOUN
fcis-30157	264	8	,	,	PUNCT
fcis-30157	264	9	such	such	ADJ
fcis-30157	264	10	as	as	ADP
fcis-30157	264	11	learning	learn	VERB
fcis-30157	264	12	rate	rate	NOUN
fcis-30157	264	13	,	,	PUNCT
fcis-30157	264	14	batch	batch	NOUN
fcis-30157	264	15	size	size	NOUN
fcis-30157	264	16	,	,	PUNCT
fcis-30157	264	17	and	and	CCONJ
fcis-30157	264	18	momentum	momentum	NOUN
fcis-30157	264	19	,	,	PUNCT
fcis-30157	264	20	significantly	significantly	ADV
fcis-30157	264	21	impacts	impact	VERB
fcis-30157	264	22	training	training	NOUN
fcis-30157	264	23	.	.	PUNCT
fcis-30157	265	1	learning	learn	VERB
fcis-30157	265	2	rate	rate	NOUN
fcis-30157	265	3	scheduling	scheduling	NOUN
fcis-30157	265	4	,	,	PUNCT
fcis-30157	265	5	such	such	ADJ
fcis-30157	265	6	as	as	ADP
fcis-30157	265	7	step	step	NOUN
fcis-30157	265	8	decay	decay	NOUN
fcis-30157	265	9	or	or	CCONJ
fcis-30157	265	10	cosine	cosine	NOUN
fcis-30157	265	11	annealing	annealing	NOUN
fcis-30157	265	12	,	,	PUNCT
fcis-30157	265	13	can	can	AUX
fcis-30157	265	14	improve	improve	VERB
fcis-30157	265	15	convergence	convergence	NOUN
fcis-30157	265	16	.	.	PUNCT
fcis-30157	266	1	batch	batch	NOUN
fcis-30157	266	2	size	size	NOUN
fcis-30157	266	3	affects	affect	VERB
fcis-30157	266	4	training	train	VERB
fcis-30157	266	5	speed	speed	NOUN
fcis-30157	266	6	and	and	CCONJ
fcis-30157	266	7	model	model	NOUN
fcis-30157	266	8	performance	performance	NOUN
fcis-30157	266	9	,	,	PUNCT
fcis-30157	266	10	with	with	ADP
fcis-30157	266	11	a	a	DET
fcis-30157	266	12	typical	typical	ADJ
fcis-30157	266	13	range	range	NOUN
fcis-30157	266	14	of	of	ADP
fcis-30157	266	15	32	32	NUM
fcis-30157	266	16	to	to	PART
fcis-30157	266	17	256	256	NUM
fcis-30157	266	18	being	be	AUX
fcis-30157	266	19	effective	effective	ADJ
fcis-30157	266	20	.	.	PUNCT
fcis-30157	267	1	optimization	optimization	NOUN
fcis-30157	267	2	techniques	technique	NOUN
fcis-30157	267	3	:	:	PUNCT
fcis-30157	267	4	regularization	regularization	NOUN
fcis-30157	267	5	techniques	technique	NOUN
fcis-30157	267	6	like	like	ADP
fcis-30157	267	7	l2	l2	NOUN
fcis-30157	267	8	regularization	regularization	NOUN
fcis-30157	267	9	and	and	CCONJ
fcis-30157	267	10	dropout	dropout	NOUN
fcis-30157	267	11	prevent	prevent	NOUN
fcis-30157	267	12	overfitting	overfitting	NOUN
fcis-30157	267	13	by	by	ADP
fcis-30157	267	14	penalizing	penalize	VERB
fcis-30157	267	15	large	large	ADJ
fcis-30157	267	16	weights	weight	NOUN
fcis-30157	267	17	and	and	CCONJ
fcis-30157	267	18	randomly	randomly	ADV
fcis-30157	267	19	dropping	drop	VERB
fcis-30157	267	20	network	network	NOUN
fcis-30157	267	21	units	unit	NOUN
fcis-30157	267	22	.	.	PUNCT
fcis-30157	268	1	early	early	ADJ
fcis-30157	268	2	stopping	stop	VERB
fcis-30157	268	3	monitors	monitor	NOUN
fcis-30157	268	4	validation	validation	NOUN
fcis-30157	268	5	performance	performance	NOUN
fcis-30157	268	6	and	and	CCONJ
fcis-30157	268	7	halts	halt	VERB
fcis-30157	268	8	training	training	NOUN
fcis-30157	268	9	if	if	SCONJ
fcis-30157	268	10	improvement	improvement	NOUN
fcis-30157	268	11	stagnates	stagnate	VERB
fcis-30157	268	12	,	,	PUNCT
fcis-30157	268	13	ensuring	ensure	VERB
fcis-30157	268	14	the	the	DET
fcis-30157	268	15	model	model	NOUN
fcis-30157	268	16	generalizes	generalize	VERB
fcis-30157	268	17	well	well	ADV
fcis-30157	268	18	.	.	PUNCT
fcis-30157	269	1	case	case	NOUN
fcis-30157	269	2	study	study	NOUN
fcis-30157	269	3	:	:	PUNCT
fcis-30157	269	4	skin	skin	NOUN
fcis-30157	269	5	disease	disease	NOUN
fcis-30157	269	6	classification	classification	NOUN
fcis-30157	269	7	:	:	PUNCT
fcis-30157	269	8	in	in	ADP
fcis-30157	269	9	a	a	DET
fcis-30157	269	10	skin	skin	NOUN
fcis-30157	269	11	disease	disease	NOUN
fcis-30157	269	12	classification	classification	NOUN
fcis-30157	269	13	task	task	NOUN
fcis-30157	269	14	using	use	VERB
fcis-30157	269	15	a	a	DET
fcis-30157	269	16	cnn	cnn	NOUN
fcis-30157	269	17	,	,	PUNCT
fcis-30157	269	18	the	the	DET
fcis-30157	269	19	adam	adam	PROPN
fcis-30157	269	20	optimizer	optimizer	NOUN
fcis-30157	269	21	,	,	PUNCT
fcis-30157	269	22	learning	learn	VERB
fcis-30157	269	23	rate	rate	NOUN
fcis-30157	269	24	of	of	ADP
fcis-30157	269	25	0.001	0.001	NUM
fcis-30157	269	26	,	,	PUNCT
fcis-30157	269	27	and	and	CCONJ
fcis-30157	269	28	a	a	DET
fcis-30157	269	29	batch	batch	NOUN
fcis-30157	269	30	size	size	NOUN
fcis-30157	269	31	of	of	ADP
fcis-30157	269	32	64	64	NUM
fcis-30157	269	33	were	be	AUX
fcis-30157	269	34	used	use	VERB
fcis-30157	269	35	.	.	PUNCT
fcis-30157	270	1	early	early	ADV
fcis-30157	270	2	stopping	stop	VERB
fcis-30157	270	3	with	with	ADP
fcis-30157	270	4	a	a	DET
fcis-30157	270	5	patience	patience	NOUN
fcis-30157	270	6	of	of	ADP
fcis-30157	270	7	10	10	NUM
fcis-30157	270	8	epochs	epoch	NOUN
fcis-30157	270	9	was	be	AUX
fcis-30157	270	10	applied	apply	VERB
fcis-30157	270	11	,	,	PUNCT
fcis-30157	270	12	achieving	achieve	VERB
fcis-30157	270	13	92.5	92.5	NUM
fcis-30157	270	14	%	%	NOUN
fcis-30157	270	15	accuracy	accuracy	NOUN
fcis-30157	270	16	on	on	ADP
fcis-30157	270	17	the	the	DET
fcis-30157	270	18	validation	validation	NOUN
fcis-30157	270	19	set	set	NOUN
fcis-30157	270	20	,	,	PUNCT
fcis-30157	270	21	outperforming	outperform	VERB
fcis-30157	270	22	an	an	DET
fcis-30157	270	23	sgd	sgd	NOUN
fcis-30157	270	24	-	-	PUNCT
fcis-30157	270	25	based	base	VERB
fcis-30157	270	26	baseline	baseline	NOUN
fcis-30157	270	27	.	.	PUNCT
fcis-30157	271	1	performance	performance	NOUN
fcis-30157	271	2	metrics	metric	NOUN
fcis-30157	271	3	:	:	PUNCT
fcis-30157	271	4	metrics	metric	NOUN
fcis-30157	271	5	like	like	ADP
fcis-30157	271	6	accuracy	accuracy	NOUN
fcis-30157	271	7	(	(	PUNCT
fcis-30157	271	8	92.5	92.5	NUM
fcis-30157	271	9	%	%	NOUN
fcis-30157	271	10	)	)	PUNCT
fcis-30157	271	11	,	,	PUNCT
fcis-30157	271	12	precision	precision	NOUN
fcis-30157	271	13	(	(	PUNCT
fcis-30157	271	14	91.8	91.8	NUM
fcis-30157	271	15	%	%	NOUN
fcis-30157	271	16	)	)	PUNCT
fcis-30157	271	17	,	,	PUNCT
fcis-30157	271	18	recall	recall	INTJ
fcis-30157	271	19	(	(	PUNCT
fcis-30157	271	20	93.2	93.2	NUM
fcis-30157	271	21	%	%	NOUN
fcis-30157	271	22	)	)	PUNCT
fcis-30157	271	23	,	,	PUNCT
fcis-30157	271	24	and	and	CCONJ
fcis-30157	271	25	f1	f1	NOUN
fcis-30157	271	26	score	score	NOUN
fcis-30157	271	27	(	(	PUNCT
fcis-30157	271	28	92.5	92.5	NUM
fcis-30157	271	29	%	%	NOUN
fcis-30157	271	30	)	)	PUNCT
fcis-30157	271	31	were	be	AUX
fcis-30157	271	32	used	use	VERB
fcis-30157	271	33	to	to	PART
fcis-30157	271	34	evaluate	evaluate	VERB
fcis-30157	271	35	model	model	NOUN
fcis-30157	271	36	performance	performance	NOUN
fcis-30157	271	37	,	,	PUNCT
fcis-30157	271	38	demonstrating	demonstrate	VERB
fcis-30157	271	39	the	the	DET
fcis-30157	271	40	effectiveness	effectiveness	NOUN
fcis-30157	271	41	of	of	ADP
fcis-30157	271	42	adam	adam	PROPN
fcis-30157	271	43	and	and	CCONJ
fcis-30157	271	44	early	early	ADV
fcis-30157	271	45	stopping	stop	VERB
fcis-30157	271	46	in	in	ADP
fcis-30157	271	47	improving	improve	VERB
fcis-30157	271	48	stability	stability	NOUN
fcis-30157	271	49	and	and	CCONJ
fcis-30157	271	50	generalization	generalization	NOUN
fcis-30157	271	51	.	.	PUNCT
fcis-30157	272	1	discussion	discussion	NOUN
fcis-30157	272	2	:	:	PUNCT
fcis-30157	272	3	the	the	DET
fcis-30157	272	4	results	result	NOUN
fcis-30157	272	5	highlight	highlight	VERB
fcis-30157	272	6	how	how	SCONJ
fcis-30157	272	7	adam	adam	PROPN
fcis-30157	272	8	's	's	PART
fcis-30157	272	9	adaptive	adaptive	ADJ
fcis-30157	272	10	learning	learning	NOUN
fcis-30157	272	11	rate	rate	NOUN
fcis-30157	272	12	and	and	CCONJ
fcis-30157	272	13	early	early	ADJ
fcis-30157	272	14	stopping	stopping	NOUN
fcis-30157	272	15	contribute	contribute	NOUN
fcis-30157	272	16	to	to	ADP
fcis-30157	272	17	model	model	NOUN
fcis-30157	272	18	performance	performance	NOUN
fcis-30157	272	19	.	.	PUNCT
fcis-30157	273	1	however	however	ADV
fcis-30157	273	2	,	,	PUNCT
fcis-30157	273	3	the	the	DET
fcis-30157	273	4	optimal	optimal	ADJ
fcis-30157	273	5	hyperparameter	hyperparameter	NOUN
fcis-30157	273	6	configuration	configuration	NOUN
fcis-30157	273	7	may	may	AUX
fcis-30157	273	8	vary	vary	VERB
fcis-30157	273	9	across	across	ADP
fcis-30157	273	10	datasets	dataset	NOUN
fcis-30157	273	11	and	and	CCONJ
fcis-30157	273	12	architectures	architecture	NOUN
fcis-30157	273	13	,	,	PUNCT
fcis-30157	273	14	requiring	require	VERB
fcis-30157	273	15	further	further	ADJ
fcis-30157	273	16	research	research	NOUN
fcis-30157	273	17	.	.	PUNCT
fcis-30157	274	1	conclusion	conclusion	NOUN
fcis-30157	274	2	:	:	PUNCT
fcis-30157	274	3	the	the	DET
fcis-30157	274	4	training	training	NOUN
fcis-30157	274	5	and	and	CCONJ
fcis-30157	274	6	optimization	optimization	NOUN
fcis-30157	274	7	of	of	ADP
fcis-30157	274	8	skin	skin	NOUN
fcis-30157	274	9	disease	disease	NOUN
fcis-30157	274	10	recognition	recognition	NOUN
fcis-30157	274	11	models	model	NOUN
fcis-30157	274	12	benefit	benefit	VERB
fcis-30157	274	13	from	from	ADP
fcis-30157	274	14	advanced	advanced	ADJ
fcis-30157	274	15	algorithms	algorithm	NOUN
fcis-30157	274	16	like	like	ADP
fcis-30157	274	17	adam	adam	NOUN
fcis-30157	274	18	and	and	CCONJ
fcis-30157	274	19	techniques	technique	NOUN
fcis-30157	274	20	such	such	ADJ
fcis-30157	274	21	as	as	ADP
fcis-30157	274	22	early	early	ADJ
fcis-30157	274	23	stopping	stopping	NOUN
fcis-30157	274	24	.	.	PUNCT
fcis-30157	275	1	future	future	ADJ
fcis-30157	275	2	research	research	NOUN
fcis-30157	275	3	should	should	AUX
fcis-30157	275	4	focus	focus	VERB
fcis-30157	275	5	on	on	ADP
fcis-30157	275	6	optimizing	optimize	VERB
fcis-30157	275	7	configurations	configuration	NOUN
fcis-30157	275	8	for	for	ADP
fcis-30157	275	9	different	different	ADJ
fcis-30157	275	10	tasks	task	NOUN
fcis-30157	275	11	to	to	PART
fcis-30157	275	12	develop	develop	VERB
fcis-30157	275	13	more	more	ADV
fcis-30157	275	14	robust	robust	ADJ
fcis-30157	275	15	models	model	NOUN
fcis-30157	275	16	for	for	ADP
fcis-30157	275	17	accurate	accurate	ADJ
fcis-30157	275	18	skin	skin	NOUN
fcis-30157	275	19	disease	disease	NOUN
fcis-30157	275	20	diagnosis	diagnosis	NOUN
fcis-30157	275	21	.	.	PUNCT
fcis-30157	276	1	2.2.3	2.2.3	X
fcis-30157	276	2	.	.	X
fcis-30157	277	1	validation	validation	NOUN
fcis-30157	277	2	and	and	CCONJ
fcis-30157	277	3	testing	testing	NOUN
fcis-30157	277	4	of	of	ADP
fcis-30157	277	5	the	the	DET
fcis-30157	277	6	recognition	recognition	NOUN
fcis-30157	277	7	model	model	NOUN
fcis-30157	277	8	to	to	PART
fcis-30157	277	9	validate	validate	VERB
fcis-30157	277	10	and	and	CCONJ
fcis-30157	277	11	test	test	VERB
fcis-30157	277	12	the	the	DET
fcis-30157	277	13	automatic	automatic	ADJ
fcis-30157	277	14	skin	skin	NOUN
fcis-30157	277	15	disease	disease	NOUN
fcis-30157	277	16	identification	identification	NOUN
fcis-30157	277	17	model	model	NOUN
fcis-30157	277	18	,	,	PUNCT
fcis-30157	277	19	we	we	PRON
fcis-30157	277	20	used	use	VERB
fcis-30157	277	21	cross	cross	ADJ
fcis-30157	277	22	-	-	ADJ
fcis-30157	277	23	validation	validation	ADJ
fcis-30157	277	24	and	and	CCONJ
fcis-30157	277	25	independent	independent	ADJ
fcis-30157	277	26	testing	testing	NOUN
fcis-30157	277	27	.	.	PUNCT
fcis-30157	278	1	cross	cross	ADJ
fcis-30157	278	2	-	-	NOUN
fcis-30157	278	3	validation	validation	ADJ
fcis-30157	278	4	,	,	PUNCT
fcis-30157	278	5	specifically	specifically	ADV
fcis-30157	278	6	k	k	NOUN
fcis-30157	278	7	-	-	ADJ
fcis-30157	278	8	fold	fold	ADJ
fcis-30157	278	9	,	,	PUNCT
fcis-30157	278	10	was	be	AUX
fcis-30157	278	11	employed	employ	VERB
fcis-30157	278	12	to	to	PART
fcis-30157	278	13	assess	assess	VERB
fcis-30157	278	14	model	model	NOUN
fcis-30157	278	15	robustness	robustness	NOUN
fcis-30157	278	16	by	by	ADP
fcis-30157	278	17	partitioning	partition	VERB
fcis-30157	278	18	the	the	DET
fcis-30157	278	19	data	datum	NOUN
fcis-30157	278	20	into	into	ADP
fcis-30157	278	21	k	k	PROPN
fcis-30157	278	22	folds	fold	NOUN
fcis-30157	278	23	.	.	PUNCT
fcis-30157	279	1	the	the	DET
fcis-30157	279	2	model	model	NOUN
fcis-30157	279	3	is	be	AUX
fcis-30157	279	4	trained	train	VERB
fcis-30157	279	5	on	on	ADP
fcis-30157	279	6	k-1	k-1	PROPN
fcis-30157	279	7	folds	fold	NOUN
fcis-30157	279	8	and	and	CCONJ
fcis-30157	279	9	validated	validate	VERB
fcis-30157	279	10	on	on	ADP
fcis-30157	279	11	the	the	DET
fcis-30157	279	12	remaining	remain	VERB
fcis-30157	279	13	fold	fold	NOUN
fcis-30157	279	14	,	,	PUNCT
fcis-30157	279	15	with	with	SCONJ
fcis-30157	279	16	the	the	DET
fcis-30157	279	17	process	process	NOUN
fcis-30157	279	18	repeated	repeat	VERB
fcis-30157	279	19	k	k	PROPN
fcis-30157	279	20	times	times	PROPN
fcis-30157	279	21	.	.	PUNCT
fcis-30157	280	1	for	for	ADP
fcis-30157	280	2	instance	instance	NOUN
fcis-30157	280	3	,	,	PUNCT
fcis-30157	280	4	a	a	DET
fcis-30157	280	5	10	10	NUM
fcis-30157	280	6	-	-	ADJ
fcis-30157	280	7	fold	fold	ADJ
fcis-30157	280	8	cross	cross	NOUN
fcis-30157	280	9	-	-	NOUN
fcis-30157	280	10	validation	validation	NOUN
fcis-30157	280	11	on	on	ADP
fcis-30157	280	12	dermoscopy	dermoscopy	NOUN
fcis-30157	280	13	images	image	NOUN
fcis-30157	280	14	achieved	achieve	VERB
fcis-30157	280	15	an	an	DET
fcis-30157	280	16	average	average	ADJ
fcis-30157	280	17	accuracy	accuracy	NOUN
fcis-30157	280	18	of	of	ADP
fcis-30157	280	19	87.5	87.5	NUM
fcis-30157	280	20	%	%	NOUN
fcis-30157	280	21	,	,	PUNCT
fcis-30157	280	22	indicating	indicate	VERB
fcis-30157	280	23	the	the	DET
fcis-30157	280	24	model	model	NOUN
fcis-30157	280	25	’s	’s	PART
fcis-30157	280	26	stability	stability	NOUN
fcis-30157	280	27	.	.	PUNCT
fcis-30157	281	1	in	in	ADP
fcis-30157	281	2	addition	addition	NOUN
fcis-30157	281	3	to	to	ADP
fcis-30157	281	4	cross	cross	ADJ
fcis-30157	281	5	-	-	NOUN
fcis-30157	281	6	validation	validation	ADJ
fcis-30157	281	7	,	,	PUNCT
fcis-30157	281	8	an	an	DET
fcis-30157	281	9	independent	independent	ADJ
fcis-30157	281	10	test	test	NOUN
fcis-30157	281	11	set	set	NOUN
fcis-30157	281	12	was	be	AUX
fcis-30157	281	13	used	use	VERB
fcis-30157	281	14	to	to	PART
fcis-30157	281	15	evaluate	evaluate	VERB
fcis-30157	281	16	the	the	DET
fcis-30157	281	17	model	model	NOUN
fcis-30157	281	18	's	's	PART
fcis-30157	281	19	performance	performance	NOUN
fcis-30157	281	20	on	on	ADP
fcis-30157	281	21	unseen	unseen	ADJ
fcis-30157	281	22	data	datum	NOUN
fcis-30157	281	23	.	.	PUNCT
fcis-30157	282	1	for	for	ADP
fcis-30157	282	2	example	example	NOUN
fcis-30157	282	3	,	,	PUNCT
fcis-30157	282	4	a	a	DET
fcis-30157	282	5	deep	deep	ADJ
fcis-30157	282	6	learning	learning	NOUN
fcis-30157	282	7	model	model	NOUN
fcis-30157	282	8	achieved	achieve	VERB
fcis-30157	282	9	92.3	92.3	NUM
fcis-30157	282	10	%	%	NOUN
fcis-30157	282	11	accuracy	accuracy	NOUN
fcis-30157	282	12	,	,	PUNCT
fcis-30157	282	13	89.7	89.7	NUM
fcis-30157	282	14	%	%	NOUN
fcis-30157	282	15	recall	recall	NOUN
fcis-30157	282	16	,	,	PUNCT
fcis-30157	282	17	and	and	CCONJ
fcis-30157	282	18	90.9	90.9	NUM
fcis-30157	282	19	%	%	NOUN
fcis-30157	282	20	f1	f1	NOUN
fcis-30157	282	21	score	score	NOUN
fcis-30157	282	22	on	on	ADP
fcis-30157	282	23	an	an	DET
fcis-30157	282	24	independent	independent	ADJ
fcis-30157	282	25	test	test	NOUN
fcis-30157	282	26	set	set	VERB
fcis-30157	282	27	.	.	PUNCT
fcis-30157	283	1	accuracy	accuracy	NOUN
fcis-30157	283	2	provides	provide	VERB
fcis-30157	283	3	an	an	DET
fcis-30157	283	4	overview	overview	NOUN
fcis-30157	283	5	of	of	ADP
fcis-30157	283	6	correctness	correctness	NOUN
fcis-30157	283	7	,	,	PUNCT
fcis-30157	283	8	recall	recall	VERB
fcis-30157	283	9	measures	measure	NOUN
fcis-30157	283	10	the	the	DET
fcis-30157	283	11	ability	ability	NOUN
fcis-30157	283	12	to	to	PART
fcis-30157	283	13	identify	identify	VERB
fcis-30157	283	14	positive	positive	ADJ
fcis-30157	283	15	cases	case	NOUN
fcis-30157	283	16	,	,	PUNCT
fcis-30157	283	17	and	and	CCONJ
fcis-30157	283	18	the	the	DET
fcis-30157	283	19	f1	f1	PROPN
fcis-30157	283	20	score	score	NOUN
fcis-30157	283	21	balances	balance	NOUN
fcis-30157	283	22	precision	precision	NOUN
fcis-30157	283	23	and	and	CCONJ
fcis-30157	283	24	recall	recall	NOUN
fcis-30157	283	25	.	.	PUNCT
fcis-30157	284	1	the	the	DET
fcis-30157	284	2	confusion	confusion	NOUN
fcis-30157	284	3	matrix	matrix	NOUN
fcis-30157	284	4	revealed	reveal	VERB
fcis-30157	284	5	areas	area	NOUN
fcis-30157	284	6	for	for	ADP
fcis-30157	284	7	improvement	improvement	NOUN
fcis-30157	284	8	,	,	PUNCT
fcis-30157	284	9	such	such	ADJ
fcis-30157	284	10	as	as	ADP
fcis-30157	284	11	a	a	DET
fcis-30157	284	12	high	high	ADJ
fcis-30157	284	13	false	false	ADJ
fcis-30157	284	14	-	-	PUNCT
fcis-30157	284	15	negative	negative	ADJ
fcis-30157	284	16	rate	rate	NOUN
fcis-30157	284	17	for	for	ADP
fcis-30157	284	18	melanoma	melanoma	NOUN
fcis-30157	284	19	cases	case	NOUN
fcis-30157	284	20	.	.	PUNCT
fcis-30157	285	1	a	a	DET
fcis-30157	285	2	qualitative	qualitative	ADJ
fcis-30157	285	3	analysis	analysis	NOUN
fcis-30157	285	4	of	of	ADP
fcis-30157	285	5	ambiguous	ambiguous	ADJ
fcis-30157	285	6	cases	case	NOUN
fcis-30157	285	7	showed	show	VERB
fcis-30157	285	8	the	the	DET
fcis-30157	285	9	model	model	NOUN
fcis-30157	285	10	's	's	PART
fcis-30157	285	11	strong	strong	ADJ
fcis-30157	285	12	performance	performance	NOUN
fcis-30157	285	13	,	,	PUNCT
fcis-30157	285	14	with	with	ADP
fcis-30157	285	15	85.6	85.6	NUM
fcis-30157	285	16	%	%	NOUN
fcis-30157	285	17	accuracy	accuracy	NOUN
fcis-30157	285	18	on	on	ADP
fcis-30157	285	19	hyperpigmented	hyperpigmented	ADJ
fcis-30157	285	20	lesions	lesion	NOUN
fcis-30157	285	21	.	.	PUNCT
fcis-30157	286	1	computational	computational	ADJ
fcis-30157	286	2	efficiency	efficiency	NOUN
fcis-30157	286	3	was	be	AUX
fcis-30157	286	4	also	also	ADV
fcis-30157	286	5	tested	test	VERB
fcis-30157	286	6	,	,	PUNCT
fcis-30157	286	7	with	with	ADP
fcis-30157	286	8	the	the	DET
fcis-30157	286	9	model	model	NOUN
fcis-30157	286	10	processing	process	VERB
fcis-30157	286	11	an	an	DET
fcis-30157	286	12	image	image	NOUN
fcis-30157	286	13	in	in	ADP
fcis-30157	286	14	0.3	0.3	NUM
fcis-30157	286	15	seconds	second	NOUN
fcis-30157	286	16	,	,	PUNCT
fcis-30157	286	17	suitable	suitable	ADJ
fcis-30157	286	18	for	for	ADP
fcis-30157	286	19	real	real	ADJ
fcis-30157	286	20	-	-	PUNCT
fcis-30157	286	21	time	time	NOUN
fcis-30157	286	22	clinical	clinical	ADJ
fcis-30157	286	23	applications	application	NOUN
fcis-30157	286	24	.	.	PUNCT
fcis-30157	287	1	hyperparameter	hyperparameter	NOUN
fcis-30157	287	2	optimization	optimization	NOUN
fcis-30157	287	3	,	,	PUNCT
fcis-30157	287	4	including	include	VERB
fcis-30157	287	5	adjustments	adjustment	NOUN
fcis-30157	287	6	to	to	ADP
fcis-30157	287	7	the	the	DET
fcis-30157	287	8	learning	learning	NOUN
fcis-30157	287	9	rate	rate	NOUN
fcis-30157	287	10	,	,	PUNCT
fcis-30157	287	11	batch	batch	NOUN
fcis-30157	287	12	size	size	NOUN
fcis-30157	287	13	,	,	PUNCT
fcis-30157	287	14	and	and	CCONJ
fcis-30157	287	15	layers	layer	NOUN
fcis-30157	287	16	,	,	PUNCT
fcis-30157	287	17	improved	improve	VERB
fcis-30157	287	18	the	the	DET
fcis-30157	287	19	model	model	NOUN
fcis-30157	287	20	’s	’s	PART
fcis-30157	287	21	accuracy	accuracy	NOUN
fcis-30157	287	22	.	.	PUNCT
fcis-30157	288	1	for	for	ADP
fcis-30157	288	2	example	example	NOUN
fcis-30157	288	3	,	,	PUNCT
fcis-30157	288	4	an	an	DET
fcis-30157	288	5	optimal	optimal	ADJ
fcis-30157	288	6	learning	learning	NOUN
fcis-30157	288	7	rate	rate	NOUN
fcis-30157	288	8	of	of	ADP
fcis-30157	288	9	0.001	0.001	NUM
fcis-30157	288	10	increased	increase	VERB
fcis-30157	288	11	accuracy	accuracy	NOUN
fcis-30157	288	12	from	from	ADP
fcis-30157	288	13	85.3	85.3	NUM
fcis-30157	288	14	%	%	NOUN
fcis-30157	288	15	to	to	ADP
fcis-30157	288	16	90.1	90.1	NUM
fcis-30157	288	17	%	%	NOUN
fcis-30157	288	18	.	.	PUNCT
fcis-30157	289	1	2.3	2.3	NUM
fcis-30157	289	2	.	.	PUNCT
fcis-30157	289	3	auxiliary	auxiliary	ADJ
fcis-30157	289	4	diagnosis	diagnosis	NOUN
fcis-30157	289	5	framework	framework	NOUN
fcis-30157	289	6	2.3.1	2.3.1	NUM
fcis-30157	289	7	.	.	PUNCT
fcis-30157	290	1	integration	integration	NOUN
fcis-30157	290	2	of	of	ADP
fcis-30157	290	3	recognition	recognition	NOUN
fcis-30157	290	4	results	result	NOUN
fcis-30157	290	5	integrating	integrate	VERB
fcis-30157	290	6	the	the	DET
fcis-30157	290	7	results	result	NOUN
fcis-30157	290	8	of	of	ADP
fcis-30157	290	9	automatic	automatic	ADJ
fcis-30157	290	10	identification	identification	NOUN
fcis-30157	290	11	models	model	NOUN
fcis-30157	290	12	is	be	AUX
fcis-30157	290	13	crucial	crucial	ADJ
fcis-30157	290	14	for	for	ADP
fcis-30157	290	15	improving	improve	VERB
fcis-30157	290	16	diagnostic	diagnostic	ADJ
fcis-30157	290	17	accuracy	accuracy	NOUN
fcis-30157	290	18	and	and	CCONJ
fcis-30157	290	19	reliability	reliability	NOUN
fcis-30157	290	20	.	.	PUNCT
fcis-30157	291	1	various	various	ADJ
fcis-30157	291	2	methods	method	NOUN
fcis-30157	291	3	,	,	PUNCT
fcis-30157	291	4	such	such	ADJ
fcis-30157	291	5	as	as	ADP
fcis-30157	291	6	simple	simple	ADJ
fcis-30157	291	7	voting	voting	NOUN
fcis-30157	291	8	and	and	CCONJ
fcis-30157	291	9	weighted	weighted	ADJ
fcis-30157	291	10	averaging	averaging	NOUN
fcis-30157	291	11	,	,	PUNCT
fcis-30157	291	12	have	have	VERB
fcis-30157	291	13	different	different	ADJ
fcis-30157	291	14	advantages	advantage	NOUN
fcis-30157	291	15	depending	depend	VERB
fcis-30157	291	16	on	on	ADP
fcis-30157	291	17	the	the	DET
fcis-30157	291	18	scenario	scenario	NOUN
fcis-30157	291	19	.	.	PUNCT
fcis-30157	292	1	simple	simple	ADJ
fcis-30157	292	2	voting	voting	NOUN
fcis-30157	292	3	works	work	NOUN
fcis-30157	292	4	well	well	ADV
fcis-30157	292	5	when	when	SCONJ
fcis-30157	292	6	model	model	NOUN
fcis-30157	292	7	outputs	output	NOUN
fcis-30157	292	8	are	be	AUX
fcis-30157	292	9	consistent	consistent	ADJ
fcis-30157	292	10	,	,	PUNCT
fcis-30157	292	11	such	such	ADJ
fcis-30157	292	12	as	as	ADP
fcis-30157	292	13	in	in	ADP
fcis-30157	292	14	skin	skin	NOUN
fcis-30157	292	15	disease	disease	NOUN
fcis-30157	292	16	classification	classification	NOUN
fcis-30157	292	17	,	,	PUNCT
fcis-30157	292	18	where	where	SCONJ
fcis-30157	292	19	results	result	NOUN
fcis-30157	292	20	from	from	ADP
fcis-30157	292	21	multiple	multiple	ADJ
fcis-30157	292	22	models	model	NOUN
fcis-30157	292	23	align	align	VERB
fcis-30157	292	24	.	.	PUNCT
fcis-30157	293	1	however	however	ADV
fcis-30157	293	2	,	,	PUNCT
fcis-30157	293	3	it	it	PRON
fcis-30157	293	4	becomes	become	VERB
fcis-30157	293	5	less	less	ADV
fcis-30157	293	6	reliable	reliable	ADJ
fcis-30157	293	7	when	when	SCONJ
fcis-30157	293	8	model	model	NOUN
fcis-30157	293	9	outputs	output	NOUN
fcis-30157	293	10	vary	vary	VERB
fcis-30157	293	11	significantly	significantly	ADV
fcis-30157	293	12	.	.	PUNCT
fcis-30157	294	1	the	the	DET
fcis-30157	294	2	weighted	weight	VERB
fcis-30157	294	3	average	average	ADJ
fcis-30157	294	4	method	method	NOUN
fcis-30157	294	5	,	,	PUNCT
fcis-30157	294	6	on	on	ADP
fcis-30157	294	7	the	the	DET
fcis-30157	294	8	other	other	ADJ
fcis-30157	294	9	hand	hand	NOUN
fcis-30157	294	10	,	,	PUNCT
fcis-30157	294	11	assigns	assign	VERB
fcis-30157	294	12	weights	weight	NOUN
fcis-30157	294	13	based	base	VERB
fcis-30157	294	14	on	on	ADP
fcis-30157	294	15	model	model	NOUN
fcis-30157	294	16	performance	performance	NOUN
fcis-30157	294	17	(	(	PUNCT
fcis-30157	294	18	e.g.	e.g.	ADJ
fcis-30157	294	19	,	,	PUNCT
fcis-30157	294	20	accuracy	accuracy	NOUN
fcis-30157	294	21	,	,	PUNCT
fcis-30157	294	22	recall	recall	NOUN
fcis-30157	294	23	)	)	PUNCT
fcis-30157	294	24	,	,	PUNCT
fcis-30157	294	25	improving	improve	VERB
fcis-30157	294	26	overall	overall	ADJ
fcis-30157	294	27	accuracy	accuracy	NOUN
fcis-30157	294	28	and	and	CCONJ
fcis-30157	294	29	robustness	robustness	NOUN
fcis-30157	294	30	,	,	PUNCT
fcis-30157	294	31	especially	especially	ADV
fcis-30157	294	32	in	in	ADP
fcis-30157	294	33	complex	complex	ADJ
fcis-30157	294	34	cases	case	NOUN
fcis-30157	294	35	.	.	PUNCT
fcis-30157	295	1	for	for	ADP
fcis-30157	295	2	instance	instance	NOUN
fcis-30157	295	3	,	,	PUNCT
fcis-30157	295	4	cnn	cnn	PROPN
fcis-30157	295	5	-	-	PUNCT
fcis-30157	295	6	based	base	VERB
fcis-30157	295	7	models	model	NOUN
fcis-30157	295	8	,	,	PUNCT
fcis-30157	295	9	which	which	PRON
fcis-30157	295	10	perform	perform	VERB
fcis-30157	295	11	well	well	ADV
fcis-30157	295	12	in	in	ADP
fcis-30157	295	13	specific	specific	ADJ
fcis-30157	295	14	lesion	lesion	NOUN
fcis-30157	295	15	types	type	NOUN
fcis-30157	295	16	,	,	PUNCT
fcis-30157	295	17	are	be	AUX
fcis-30157	295	18	assigned	assign	VERB
fcis-30157	295	19	higher	high	ADJ
fcis-30157	295	20	weights	weight	NOUN
fcis-30157	295	21	in	in	ADP
fcis-30157	295	22	skin	skin	NOUN
fcis-30157	295	23	cancer	cancer	NOUN
fcis-30157	295	24	classification	classification	NOUN
fcis-30157	295	25	.	.	PUNCT
fcis-30157	296	1	in	in	ADP
fcis-30157	296	2	addition	addition	NOUN
fcis-30157	296	3	to	to	ADP
fcis-30157	296	4	these	these	DET
fcis-30157	296	5	methods	method	NOUN
fcis-30157	296	6	,	,	PUNCT
fcis-30157	296	7	multimodal	multimodal	NOUN
fcis-30157	296	8	fusion	fusion	NOUN
fcis-30157	296	9	,	,	PUNCT
fcis-30157	296	10	which	which	PRON
fcis-30157	296	11	combines	combine	VERB
fcis-30157	296	12	image	image	NOUN
fcis-30157	296	13	and	and	CCONJ
fcis-30157	296	14	clinical	clinical	ADJ
fcis-30157	296	15	data	datum	NOUN
fcis-30157	296	16	(	(	PUNCT
fcis-30157	296	17	such	such	ADJ
fcis-30157	296	18	as	as	ADP
fcis-30157	296	19	age	age	NOUN
fcis-30157	296	20	and	and	CCONJ
fcis-30157	296	21	gender	gender	NOUN
fcis-30157	296	22	)	)	PUNCT
fcis-30157	296	23	,	,	PUNCT
fcis-30157	296	24	has	have	AUX
fcis-30157	296	25	shown	show	VERB
fcis-30157	296	26	to	to	PART
fcis-30157	296	27	improve	improve	VERB
fcis-30157	296	28	diagnostic	diagnostic	ADJ
fcis-30157	296	29	performance	performance	NOUN
fcis-30157	296	30	,	,	PUNCT
fcis-30157	296	31	as	as	SCONJ
fcis-30157	296	32	seen	see	VERB
fcis-30157	296	33	in	in	ADP
fcis-30157	296	34	the	the	DET
fcis-30157	296	35	mdfnet	mdfnet	NOUN
fcis-30157	296	36	model	model	NOUN
fcis-30157	296	37	for	for	ADP
fcis-30157	296	38	skin	skin	NOUN
fcis-30157	296	39	cancer	cancer	NOUN
fcis-30157	296	40	classification	classification	NOUN
fcis-30157	296	41	.	.	PUNCT
fcis-30157	297	1	this	this	DET
fcis-30157	297	2	approach	approach	NOUN
fcis-30157	297	3	compensates	compensate	VERB
fcis-30157	297	4	for	for	ADP
fcis-30157	297	5	the	the	DET
fcis-30157	297	6	limitations	limitation	NOUN
fcis-30157	297	7	of	of	ADP
fcis-30157	297	8	individual	individual	ADJ
fcis-30157	297	9	models	model	NOUN
fcis-30157	297	10	and	and	CCONJ
fcis-30157	297	11	enhances	enhance	VERB
fcis-30157	297	12	diagnosis	diagnosis	NOUN
fcis-30157	297	13	for	for	ADP
fcis-30157	297	14	complex	complex	ADJ
fcis-30157	297	15	diseases	disease	NOUN
fcis-30157	297	16	.	.	PUNCT
fcis-30157	298	1	uncertainty	uncertainty	NOUN
fcis-30157	298	2	in	in	ADP
fcis-30157	298	3	identification	identification	NOUN
fcis-30157	298	4	results	result	NOUN
fcis-30157	298	5	,	,	PUNCT
fcis-30157	298	6	due	due	ADP
fcis-30157	298	7	to	to	ADP
fcis-30157	298	8	factors	factor	NOUN
fcis-30157	298	9	like	like	ADP
fcis-30157	298	10	model	model	NOUN
fcis-30157	298	11	confidence	confidence	NOUN
fcis-30157	298	12	or	or	CCONJ
fcis-30157	298	13	data	data	NOUN
fcis-30157	298	14	noise	noise	NOUN
fcis-30157	298	15	,	,	PUNCT
fcis-30157	298	16	can	can	AUX
fcis-30157	298	17	also	also	ADV
fcis-30157	298	18	be	be	AUX
fcis-30157	298	19	addressed	address	VERB
fcis-30157	298	20	.	.	PUNCT
fcis-30157	299	1	researchers	researcher	NOUN
fcis-30157	299	2	have	have	AUX
fcis-30157	299	3	proposed	propose	VERB
fcis-30157	299	4	methods	method	NOUN
fcis-30157	299	5	based	base	VERB
fcis-30157	299	6	on	on	ADP
fcis-30157	299	7	probability	probability	NOUN
fcis-30157	299	8	distribution	distribution	NOUN
fcis-30157	299	9	to	to	PART
fcis-30157	299	10	identify	identify	VERB
fcis-30157	299	11	high	high	ADJ
fcis-30157	299	12	-	-	PUNCT
fcis-30157	299	13	uncertainty	uncertainty	NOUN
fcis-30157	299	14	areas	area	NOUN
fcis-30157	299	15	and	and	CCONJ
fcis-30157	299	16	analyze	analyze	VERB
fcis-30157	299	17	them	they	PRON
fcis-30157	299	18	using	use	VERB
fcis-30157	299	19	clinical	clinical	ADJ
fcis-30157	299	20	knowledge	knowledge	NOUN
fcis-30157	299	21	,	,	PUNCT
fcis-30157	299	22	improving	improve	VERB
fcis-30157	299	23	diagnosis	diagnosis	NOUN
fcis-30157	299	24	and	and	CCONJ
fcis-30157	299	25	offering	offer	VERB
fcis-30157	299	26	detailed	detailed	ADJ
fcis-30157	299	27	decision	decision	NOUN
fcis-30157	299	28	support	support	NOUN
fcis-30157	299	29	.	.	PUNCT
fcis-30157	300	1	dynamic	dynamic	ADJ
fcis-30157	300	2	weight	weight	NOUN
fcis-30157	300	3	adjustment	adjustment	NOUN
fcis-30157	300	4	is	be	AUX
fcis-30157	300	5	another	another	DET
fcis-30157	300	6	method	method	NOUN
fcis-30157	300	7	for	for	ADP
fcis-30157	300	8	managing	manage	VERB
fcis-30157	300	9	uncertainty	uncertainty	NOUN
fcis-30157	300	10	.	.	PUNCT
fcis-30157	301	1	for	for	ADP
fcis-30157	301	2	example	example	NOUN
fcis-30157	301	3	,	,	PUNCT
fcis-30157	301	4	in	in	ADP
fcis-30157	301	5	skin	skin	NOUN
fcis-30157	301	6	lesion	lesion	NOUN
fcis-30157	301	7	classification	classification	NOUN
fcis-30157	301	8	,	,	PUNCT
fcis-30157	301	9	the	the	DET
fcis-30157	301	10	system	system	NOUN
fcis-30157	301	11	can	can	AUX
fcis-30157	301	12	adjust	adjust	VERB
fcis-30157	301	13	weights	weight	NOUN
fcis-30157	301	14	based	base	VERB
fcis-30157	301	15	on	on	ADP
fcis-30157	301	16	real	real	ADJ
fcis-30157	301	17	-	-	PUNCT
fcis-30157	301	18	time	time	NOUN
fcis-30157	301	19	model	model	NOUN
fcis-30157	301	20	performance	performance	NOUN
fcis-30157	301	21	,	,	PUNCT
fcis-30157	301	22	reducing	reduce	VERB
fcis-30157	301	23	uncertainty	uncertainty	NOUN
fcis-30157	301	24	's	's	PART
fcis-30157	301	25	impact	impact	NOUN
fcis-30157	301	26	on	on	ADP
fcis-30157	301	27	outcomes	outcome	NOUN
fcis-30157	301	28	.	.	PUNCT
fcis-30157	302	1	this	this	DET
fcis-30157	302	2	flexibility	flexibility	NOUN
fcis-30157	302	3	allows	allow	VERB
fcis-30157	302	4	better	well	ADJ
fcis-30157	302	5	adaptation	adaptation	NOUN
fcis-30157	302	6	to	to	ADP
fcis-30157	302	7	different	different	ADJ
fcis-30157	302	8	cases	case	NOUN
fcis-30157	302	9	.	.	PUNCT
fcis-30157	303	1	the	the	DET
fcis-30157	303	2	integration	integration	NOUN
fcis-30157	303	3	of	of	ADP
fcis-30157	303	4	model	model	NOUN
fcis-30157	303	5	results	result	NOUN
fcis-30157	303	6	is	be	AUX
fcis-30157	303	7	often	often	ADV
fcis-30157	303	8	combined	combine	VERB
fcis-30157	303	9	with	with	ADP
fcis-30157	303	10	clinical	clinical	ADJ
fcis-30157	303	11	knowledge	knowledge	NOUN
fcis-30157	303	12	.	.	PUNCT
fcis-30157	304	1	in	in	ADP
fcis-30157	304	2	practice	practice	NOUN
fcis-30157	304	3	,	,	PUNCT
fcis-30157	304	4	the	the	DET
fcis-30157	304	5	collaboration	collaboration	NOUN
fcis-30157	304	6	between	between	ADP
fcis-30157	304	7	model	model	NOUN
fcis-30157	304	8	outputs	output	NOUN
fcis-30157	304	9	and	and	CCONJ
fcis-30157	304	10	a	a	DET
fcis-30157	304	11	doctor	doctor	NOUN
fcis-30157	304	12	's	's	PART
fcis-30157	304	13	experience	experience	NOUN
fcis-30157	304	14	can	can	AUX
fcis-30157	304	15	significantly	significantly	ADV
fcis-30157	304	16	increase	increase	VERB
fcis-30157	304	17	diagnostic	diagnostic	ADJ
fcis-30157	304	18	accuracy	accuracy	NOUN
fcis-30157	304	19	,	,	PUNCT
fcis-30157	304	20	as	as	SCONJ
fcis-30157	304	21	shown	show	VERB
fcis-30157	304	22	in	in	ADP
fcis-30157	304	23	a	a	DET
fcis-30157	304	24	study	study	NOUN
fcis-30157	304	25	on	on	ADP
fcis-30157	304	26	contact	contact	NOUN
fcis-30157	304	27	dermatitis	dermatitis	NOUN
fcis-30157	304	28	.	.	PUNCT
fcis-30157	305	1	real	real	ADJ
fcis-30157	305	2	-	-	PUNCT
fcis-30157	305	3	world	world	NOUN
fcis-30157	305	4	applications	application	NOUN
fcis-30157	305	5	have	have	AUX
fcis-30157	305	6	demonstrated	demonstrate	VERB
fcis-30157	305	7	the	the	DET
fcis-30157	305	8	effectiveness	effectiveness	NOUN
fcis-30157	305	9	of	of	ADP
fcis-30157	305	10	result	result	NOUN
fcis-30157	305	11	integration	integration	NOUN
fcis-30157	305	12	.	.	PUNCT
fcis-30157	306	1	in	in	ADP
fcis-30157	306	2	the	the	DET
fcis-30157	306	3	2019	2019	NUM
fcis-30157	306	4	isic	isic	PROPN
fcis-30157	306	5	challenge	challenge	NOUN
fcis-30157	306	6	for	for	ADP
fcis-30157	306	7	skin	skin	NOUN
fcis-30157	306	8	lesion	lesion	NOUN
fcis-30157	306	9	classification	classification	NOUN
fcis-30157	306	10	,	,	PUNCT
fcis-30157	306	11	using	use	VERB
fcis-30157	306	12	multi	multi	ADJ
fcis-30157	306	13	-	-	ADJ
fcis-30157	306	14	model	model	ADJ
fcis-30157	306	15	integration	integration	NOUN
fcis-30157	306	16	improved	improve	VERB
fcis-30157	306	17	diagnostic	diagnostic	ADJ
fcis-30157	306	18	accuracy	accuracy	NOUN
fcis-30157	306	19	to	to	ADP
fcis-30157	306	20	92.3	92.3	NUM
fcis-30157	306	21	%	%	NOUN
fcis-30157	306	22	,	,	PUNCT
fcis-30157	306	23	far	far	ADV
fcis-30157	306	24	surpassing	surpass	VERB
fcis-30157	306	25	single	single	ADJ
fcis-30157	306	26	model	model	NOUN
fcis-30157	306	27	performance	performance	NOUN
fcis-30157	306	28	.	.	PUNCT
fcis-30157	307	1	however	however	ADV
fcis-30157	307	2	,	,	PUNCT
fcis-30157	307	3	challenges	challenge	NOUN
fcis-30157	307	4	remain	remain	VERB
fcis-30157	307	5	,	,	PUNCT
fcis-30157	307	6	such	such	ADJ
fcis-30157	307	7	as	as	ADP
fcis-30157	307	8	how	how	SCONJ
fcis-30157	307	9	to	to	PART
fcis-30157	307	10	assign	assign	VERB
fcis-30157	307	11	appropriate	appropriate	ADJ
fcis-30157	307	12	weights	weight	NOUN
fcis-30157	307	13	to	to	ADP
fcis-30157	307	14	different	different	ADJ
fcis-30157	307	15	lesion	lesion	NOUN
fcis-30157	307	16	types	type	NOUN
fcis-30157	307	17	.	.	PUNCT
fcis-30157	308	1	dynamic	dynamic	ADJ
fcis-30157	308	2	weight	weight	NOUN
fcis-30157	308	3	allocation	allocation	NOUN
fcis-30157	308	4	based	base	VERB
fcis-30157	308	5	on	on	ADP
fcis-30157	308	6	lesion	lesion	NOUN
fcis-30157	308	7	characteristics	characteristic	NOUN
fcis-30157	308	8	can	can	AUX
fcis-30157	308	9	help	help	VERB
fcis-30157	308	10	tailor	tailor	VERB
fcis-30157	308	11	the	the	DET
fcis-30157	308	12	system	system	NOUN
fcis-30157	308	13	's	's	PART
fcis-30157	308	14	diagnostic	diagnostic	ADJ
fcis-30157	308	15	approach	approach	NOUN
fcis-30157	308	16	.	.	PUNCT
fcis-30157	309	1	the	the	DET
fcis-30157	309	2	interpretability	interpretability	NOUN
fcis-30157	309	3	of	of	ADP
fcis-30157	309	4	integrated	integrate	VERB
fcis-30157	309	5	models	model	NOUN
fcis-30157	309	6	is	be	AUX
fcis-30157	309	7	also	also	ADV
fcis-30157	309	8	crucial	crucial	ADJ
fcis-30157	309	9	for	for	ADP
fcis-30157	309	10	clinical	clinical	ADJ
fcis-30157	309	11	use	use	NOUN
fcis-30157	309	12	.	.	PUNCT
fcis-30157	310	1	doctors	doctor	NOUN
fcis-30157	310	2	require	require	VERB
fcis-30157	310	3	insights	insight	NOUN
fcis-30157	310	4	into	into	ADP
fcis-30157	310	5	how	how	SCONJ
fcis-30157	310	6	results	result	NOUN
fcis-30157	310	7	are	be	AUX
fcis-30157	310	8	generated	generate	VERB
fcis-30157	310	9	.	.	PUNCT
fcis-30157	311	1	implementing	implement	VERB
fcis-30157	311	2	explainable	explainable	ADJ
fcis-30157	311	3	ai	ai	NOUN
fcis-30157	311	4	(	(	PUNCT
fcis-30157	311	5	xai	xai	PROPN
fcis-30157	311	6	)	)	PUNCT
fcis-30157	311	7	can	can	AUX
fcis-30157	311	8	help	help	VERB
fcis-30157	311	9	by	by	ADP
fcis-30157	311	10	providing	provide	VERB
fcis-30157	311	11	interpretable	interpretable	ADJ
fcis-30157	311	12	outputs	output	NOUN
fcis-30157	311	13	,	,	PUNCT
fcis-30157	311	14	enhancing	enhance	VERB
fcis-30157	311	15	trust	trust	NOUN
fcis-30157	311	16	and	and	CCONJ
fcis-30157	311	17	ensuring	ensure	VERB
fcis-30157	311	18	more	more	ADV
fcis-30157	311	19	widespread	widespread	ADJ
fcis-30157	311	20	adoption	adoption	NOUN
fcis-30157	311	21	in	in	ADP
fcis-30157	311	22	clinical	clinical	ADJ
fcis-30157	311	23	practice	practice	NOUN
fcis-30157	311	24	.	.	PUNCT
fcis-30157	312	1	future	future	ADJ
fcis-30157	312	2	advancements	advancement	NOUN
fcis-30157	312	3	may	may	AUX
fcis-30157	312	4	include	include	VERB
fcis-30157	312	5	dynamic	dynamic	ADJ
fcis-30157	312	6	integration	integration	NOUN
fcis-30157	312	7	methods	method	NOUN
fcis-30157	312	8	that	that	PRON
fcis-30157	312	9	utilize	utilize	VERB
fcis-30157	312	10	time	time	NOUN
fcis-30157	312	11	-	-	PUNCT
fcis-30157	312	12	series	series	NOUN
fcis-30157	312	13	118	118	NUM
fcis-30157	312	14	data	datum	NOUN
fcis-30157	312	15	to	to	PART
fcis-30157	312	16	track	track	VERB
fcis-30157	312	17	changes	change	NOUN
fcis-30157	312	18	in	in	ADP
fcis-30157	312	19	chronic	chronic	ADJ
fcis-30157	312	20	skin	skin	NOUN
fcis-30157	312	21	diseases	disease	NOUN
fcis-30157	312	22	,	,	PUNCT
fcis-30157	312	23	offering	offer	VERB
fcis-30157	312	24	a	a	DET
fcis-30157	312	25	more	more	ADV
fcis-30157	312	26	detailed	detailed	ADJ
fcis-30157	312	27	diagnosis	diagnosis	NOUN
fcis-30157	312	28	.	.	PUNCT
fcis-30157	313	1	2.3.2	2.3.2	NUM
fcis-30157	313	2	.	.	X
fcis-30157	313	3	incorporation	incorporation	NOUN
fcis-30157	313	4	of	of	ADP
fcis-30157	313	5	clinical	clinical	ADJ
fcis-30157	313	6	knowledge	knowledge	NOUN
fcis-30157	313	7	the	the	DET
fcis-30157	313	8	integration	integration	NOUN
fcis-30157	313	9	of	of	ADP
fcis-30157	313	10	clinical	clinical	ADJ
fcis-30157	313	11	knowledge	knowledge	NOUN
fcis-30157	313	12	with	with	ADP
fcis-30157	313	13	data	data	NOUN
fcis-30157	313	14	-	-	PUNCT
fcis-30157	313	15	driven	drive	VERB
fcis-30157	313	16	models	model	NOUN
fcis-30157	313	17	is	be	AUX
fcis-30157	313	18	crucial	crucial	ADJ
fcis-30157	313	19	for	for	ADP
fcis-30157	313	20	improving	improve	VERB
fcis-30157	313	21	skin	skin	NOUN
fcis-30157	313	22	disease	disease	NOUN
fcis-30157	313	23	diagnosis	diagnosis	NOUN
fcis-30157	313	24	.	.	PUNCT
fcis-30157	314	1	the	the	DET
fcis-30157	314	2	mdfnet	mdfnet	NOUN
fcis-30157	314	3	team	team	NOUN
fcis-30157	314	4	used	use	VERB
fcis-30157	314	5	a	a	DET
fcis-30157	314	6	multimodal	multimodal	ADJ
fcis-30157	314	7	fusion	fusion	NOUN
fcis-30157	314	8	strategy	strategy	NOUN
fcis-30157	314	9	combining	combine	VERB
fcis-30157	314	10	12	12	NUM
fcis-30157	314	11	clinical	clinical	ADJ
fcis-30157	314	12	parameters	parameter	NOUN
fcis-30157	314	13	(	(	PUNCT
fcis-30157	314	14	e.g.	e.g.	ADV
fcis-30157	314	15	,	,	PUNCT
fcis-30157	314	16	age	age	NOUN
fcis-30157	314	17	,	,	PUNCT
fcis-30157	314	18	lesion	lesion	NOUN
fcis-30157	314	19	location	location	NOUN
fcis-30157	314	20	)	)	PUNCT
fcis-30157	314	21	with	with	ADP
fcis-30157	314	22	dermatoscope	dermatoscope	NOUN
fcis-30157	314	23	images	image	NOUN
fcis-30157	314	24	,	,	PUNCT
fcis-30157	314	25	boosting	boost	VERB
fcis-30157	314	26	classification	classification	NOUN
fcis-30157	314	27	accuracy	accuracy	NOUN
fcis-30157	314	28	from	from	ADP
fcis-30157	314	29	82.3	82.3	NUM
fcis-30157	314	30	%	%	NOUN
fcis-30157	314	31	to	to	ADP
fcis-30157	314	32	89.7	89.7	NUM
fcis-30157	314	33	%	%	NOUN
fcis-30157	314	34	.	.	PUNCT
fcis-30157	315	1	this	this	DET
fcis-30157	315	2	approach	approach	NOUN
fcis-30157	315	3	adjusts	adjust	VERB
fcis-30157	315	4	the	the	DET
fcis-30157	315	5	weight	weight	NOUN
fcis-30157	315	6	of	of	ADP
fcis-30157	315	7	each	each	DET
fcis-30157	315	8	modality	modality	NOUN
fcis-30157	315	9	using	use	VERB
fcis-30157	315	10	attention	attention	NOUN
fcis-30157	315	11	-	-	PUNCT
fcis-30157	315	12	gating	gate	VERB
fcis-30157	315	13	networks	network	NOUN
fcis-30157	315	14	,	,	PUNCT
fcis-30157	315	15	assigning	assign	VERB
fcis-30157	315	16	higher	high	ADJ
fcis-30157	315	17	weights	weight	NOUN
fcis-30157	315	18	to	to	ADP
fcis-30157	315	19	certain	certain	ADJ
fcis-30157	315	20	features	feature	NOUN
fcis-30157	315	21	,	,	PUNCT
fcis-30157	315	22	like	like	ADP
fcis-30157	315	23	age	age	NOUN
fcis-30157	315	24	for	for	ADP
fcis-30157	315	25	seborrheic	seborrheic	ADJ
fcis-30157	315	26	keratosis	keratosis	NOUN
fcis-30157	315	27	.	.	PUNCT
fcis-30157	316	1	west	west	PROPN
fcis-30157	316	2	china	china	PROPN
fcis-30157	316	3	hospital	hospital	PROPN
fcis-30157	316	4	’s	’s	PART
fcis-30157	316	5	dermatology	dermatology	NOUN
fcis-30157	316	6	diagnosis	diagnosis	NOUN
fcis-30157	316	7	system	system	NOUN
fcis-30157	316	8	integrates	integrate	VERB
fcis-30157	316	9	ontological	ontological	ADJ
fcis-30157	316	10	knowledge	knowledge	NOUN
fcis-30157	316	11	,	,	PUNCT
fcis-30157	316	12	using	use	VERB
fcis-30157	316	13	a	a	DET
fcis-30157	316	14	differential	differential	ADJ
fcis-30157	316	15	diagnosis	diagnosis	NOUN
fcis-30157	316	16	tree	tree	NOUN
fcis-30157	316	17	and	and	CCONJ
fcis-30157	316	18	a	a	DET
fcis-30157	316	19	three	three	NUM
fcis-30157	316	20	-	-	PUNCT
fcis-30157	316	21	level	level	NOUN
fcis-30157	316	22	review	review	NOUN
fcis-30157	316	23	mechanism	mechanism	NOUN
fcis-30157	316	24	when	when	SCONJ
fcis-30157	316	25	image	image	NOUN
fcis-30157	316	26	results	result	VERB
fcis-30157	316	27	conflict	conflict	NOUN
fcis-30157	316	28	with	with	ADP
fcis-30157	316	29	clinical	clinical	ADJ
fcis-30157	316	30	indications	indication	NOUN
fcis-30157	316	31	[	[	X
fcis-30157	316	32	35	35	NUM
fcis-30157	316	33	]	]	PUNCT
fcis-30157	316	34	.	.	PUNCT
fcis-30157	317	1	knowledge	knowledge	NOUN
fcis-30157	317	2	distillation	distillation	NOUN
fcis-30157	317	3	technology	technology	NOUN
fcis-30157	317	4	,	,	PUNCT
fcis-30157	317	5	used	use	VERB
fcis-30157	317	6	in	in	ADP
fcis-30157	317	7	the	the	DET
fcis-30157	317	8	isic	isic	PROPN
fcis-30157	317	9	2022	2022	NUM
fcis-30157	317	10	challenge	challenge	NOUN
fcis-30157	317	11	,	,	PUNCT
fcis-30157	317	12	translates	translate	VERB
fcis-30157	317	13	chief	chief	ADJ
fcis-30157	317	14	physicians	physician	NOUN
fcis-30157	317	15	'	'	PART
fcis-30157	317	16	diagnostic	diagnostic	ADJ
fcis-30157	317	17	logic	logic	NOUN
fcis-30157	317	18	into	into	ADP
fcis-30157	317	19	interpretable	interpretable	ADJ
fcis-30157	317	20	decision	decision	NOUN
fcis-30157	317	21	-	-	PUNCT
fcis-30157	317	22	making	make	VERB
fcis-30157	317	23	rules	rule	NOUN
fcis-30157	317	24	,	,	PUNCT
fcis-30157	317	25	reducing	reduce	VERB
fcis-30157	317	26	melanoma	melanoma	NOUN
fcis-30157	317	27	false	false	ADJ
fcis-30157	317	28	-	-	PUNCT
fcis-30157	317	29	negative	negative	ADJ
fcis-30157	317	30	rates	rate	NOUN
fcis-30157	317	31	to	to	ADP
fcis-30157	317	32	below	below	ADP
fcis-30157	317	33	4.2	4.2	NUM
fcis-30157	317	34	%	%	NOUN
fcis-30157	317	35	.	.	PUNCT
fcis-30157	318	1	for	for	ADP
fcis-30157	318	2	eczema	eczema	NOUN
fcis-30157	318	3	and	and	CCONJ
fcis-30157	318	4	contact	contact	NOUN
fcis-30157	318	5	dermatitis	dermatitis	PROPN
fcis-30157	318	6	,	,	PUNCT
fcis-30157	318	7	the	the	DET
fcis-30157	318	8	european	european	PROPN
fcis-30157	318	9	contact	contact	PROPN
fcis-30157	318	10	dermatitis	dermatitis	PROPN
fcis-30157	318	11	research	research	PROPN
fcis-30157	318	12	group	group	NOUN
fcis-30157	318	13	integrated	integrate	VERB
fcis-30157	318	14	patch	patch	NOUN
fcis-30157	318	15	test	test	NOUN
fcis-30157	318	16	results	result	NOUN
fcis-30157	318	17	,	,	PUNCT
fcis-30157	318	18	boosting	boost	VERB
fcis-30157	318	19	diagnostic	diagnostic	ADJ
fcis-30157	318	20	confidence	confidence	NOUN
fcis-30157	318	21	by	by	ADP
fcis-30157	318	22	27	27	NUM
fcis-30157	318	23	%	%	NOUN
fcis-30157	318	24	when	when	SCONJ
fcis-30157	318	25	nickel	nickel	NOUN
fcis-30157	318	26	tests	test	NOUN
fcis-30157	318	27	were	be	AUX
fcis-30157	318	28	positive	positive	ADJ
fcis-30157	318	29	[	[	X
fcis-30157	318	30	13	13	NUM
fcis-30157	318	31	]	]	PUNCT
fcis-30157	318	32	.	.	PUNCT
fcis-30157	319	1	additionally	additionally	ADV
fcis-30157	319	2	,	,	PUNCT
fcis-30157	319	3	mayo	mayo	NOUN
fcis-30157	319	4	clinic	clinic	PROPN
fcis-30157	319	5	’s	’s	PART
fcis-30157	319	6	clinical	clinical	ADJ
fcis-30157	319	7	narrative	narrative	NOUN
fcis-30157	319	8	analysis	analysis	NOUN
fcis-30157	319	9	links	link	NOUN
fcis-30157	319	10	electronic	electronic	ADJ
fcis-30157	319	11	medical	medical	ADJ
fcis-30157	319	12	record	record	NOUN
fcis-30157	319	13	descriptions	description	NOUN
fcis-30157	319	14	with	with	ADP
fcis-30157	319	15	psoriasis	psoriasis	NOUN
fcis-30157	319	16	activity	activity	NOUN
fcis-30157	319	17	scores	score	NOUN
fcis-30157	319	18	,	,	PUNCT
fcis-30157	319	19	showing	show	VERB
fcis-30157	319	20	a	a	DET
fcis-30157	319	21	significant	significant	ADJ
fcis-30157	319	22	correlation	correlation	NOUN
fcis-30157	319	23	of	of	ADP
fcis-30157	319	24	0.81	0.81	NUM
fcis-30157	319	25	[	[	SYM
fcis-30157	319	26	30	30	NUM
fcis-30157	319	27	]	]	PUNCT
fcis-30157	319	28	.	.	PUNCT
fcis-30157	320	1	however	however	ADV
fcis-30157	320	2	,	,	PUNCT
fcis-30157	320	3	technical	technical	ADJ
fcis-30157	320	4	challenges	challenge	NOUN
fcis-30157	320	5	remain	remain	VERB
fcis-30157	320	6	,	,	PUNCT
fcis-30157	320	7	such	such	ADJ
fcis-30157	320	8	as	as	ADP
fcis-30157	320	9	clinical	clinical	ADJ
fcis-30157	320	10	terminology	terminology	NOUN
fcis-30157	320	11	standardization	standardization	NOUN
fcis-30157	320	12	,	,	PUNCT
fcis-30157	320	13	data	datum	NOUN
fcis-30157	320	14	alignment	alignment	NOUN
fcis-30157	320	15	,	,	PUNCT
fcis-30157	320	16	and	and	CCONJ
fcis-30157	320	17	dynamic	dynamic	ADJ
fcis-30157	320	18	knowledge	knowledge	NOUN
fcis-30157	320	19	updates	update	NOUN
fcis-30157	320	20	.	.	PUNCT
fcis-30157	321	1	huashan	huashan	PROPN
fcis-30157	321	2	hospital	hospital	PROPN
fcis-30157	321	3	's	's	PART
fcis-30157	321	4	dermatology	dermatology	NOUN
fcis-30157	321	5	knowledge	knowledge	NOUN
fcis-30157	321	6	map	map	NOUN
fcis-30157	321	7	resolves	resolve	VERB
fcis-30157	321	8	47.6	47.6	NUM
fcis-30157	321	9	%	%	NOUN
fcis-30157	321	10	of	of	ADP
fcis-30157	321	11	cross	cross	ADJ
fcis-30157	321	12	-	-	ADJ
fcis-30157	321	13	institutional	institutional	ADJ
fcis-30157	321	14	data	datum	NOUN
fcis-30157	321	15	heterogeneity	heterogeneity	NOUN
fcis-30157	321	16	by	by	ADP
fcis-30157	321	17	using	use	VERB
fcis-30157	321	18	the	the	DET
fcis-30157	321	19	snomed	snome	VERB
fcis-30157	321	20	ct	ct	PROPN
fcis-30157	321	21	standard	standard	NOUN
fcis-30157	321	22	[	[	X
fcis-30157	321	23	28	28	NUM
fcis-30157	321	24	]	]	PUNCT
fcis-30157	321	25	.	.	PUNCT
fcis-30157	322	1	the	the	DET
fcis-30157	322	2	dual	dual	ADJ
fcis-30157	322	3	attention	attention	NOUN
fcis-30157	322	4	network	network	NOUN
fcis-30157	322	5	associates	associate	NOUN
fcis-30157	322	6	lesion	lesion	NOUN
fcis-30157	322	7	pixels	pixel	NOUN
fcis-30157	322	8	with	with	ADP
fcis-30157	322	9	medical	medical	ADJ
fcis-30157	322	10	record	record	NOUN
fcis-30157	322	11	descriptions	description	NOUN
fcis-30157	322	12	,	,	PUNCT
fcis-30157	322	13	improving	improve	VERB
fcis-30157	322	14	cross	cross	ADJ
fcis-30157	322	15	-	-	ADJ
fcis-30157	322	16	modal	modal	ADJ
fcis-30157	322	17	feature	feature	NOUN
fcis-30157	322	18	matching	matching	NOUN
fcis-30157	322	19	accuracy	accuracy	NOUN
fcis-30157	322	20	to	to	ADP
fcis-30157	322	21	91.3	91.3	NUM
fcis-30157	322	22	%	%	NOUN
fcis-30157	322	23	in	in	ADP
fcis-30157	322	24	lupus	lupus	ADJ
fcis-30157	322	25	erythematosus	erythematosus	NOUN
fcis-30157	322	26	diagnosis	diagnosis	NOUN
fcis-30157	322	27	.	.	PUNCT
fcis-30157	323	1	to	to	PART
fcis-30157	323	2	address	address	VERB
fcis-30157	323	3	knowledge	knowledge	NOUN
fcis-30157	323	4	lag	lag	PROPN
fcis-30157	323	5	,	,	PUNCT
fcis-30157	323	6	the	the	DET
fcis-30157	323	7	fs3dciot	fs3dciot	NOUN
fcis-30157	323	8	framework	framework	NOUN
fcis-30157	323	9	uses	use	VERB
fcis-30157	323	10	incremental	incremental	ADJ
fcis-30157	323	11	learning	learning	NOUN
fcis-30157	323	12	,	,	PUNCT
fcis-30157	323	13	allowing	allow	VERB
fcis-30157	323	14	updates	update	NOUN
fcis-30157	323	15	to	to	ADP
fcis-30157	323	16	the	the	DET
fcis-30157	323	17	diagnostic	diagnostic	ADJ
fcis-30157	323	18	model	model	NOUN
fcis-30157	323	19	with	with	ADP
fcis-30157	323	20	minimal	minimal	ADJ
fcis-30157	323	21	samples	sample	NOUN
fcis-30157	323	22	.	.	PUNCT
fcis-30157	324	1	practical	practical	ADJ
fcis-30157	324	2	evaluations	evaluation	NOUN
fcis-30157	324	3	show	show	VERB
fcis-30157	324	4	that	that	SCONJ
fcis-30157	324	5	integrating	integrate	VERB
fcis-30157	324	6	clinical	clinical	ADJ
fcis-30157	324	7	knowledge	knowledge	NOUN
fcis-30157	324	8	can	can	AUX
fcis-30157	324	9	improve	improve	VERB
fcis-30157	324	10	diagnostic	diagnostic	ADJ
fcis-30157	324	11	accuracy	accuracy	NOUN
fcis-30157	324	12	.	.	PUNCT
fcis-30157	325	1	in	in	ADP
fcis-30157	325	2	2023	2023	NUM
fcis-30157	325	3	,	,	PUNCT
fcis-30157	325	4	multi	multi	ADJ
fcis-30157	325	5	-	-	ADJ
fcis-30157	325	6	center	center	ADJ
fcis-30157	325	7	trials	trial	NOUN
fcis-30157	325	8	showed	show	VERB
fcis-30157	325	9	that	that	SCONJ
fcis-30157	325	10	systems	system	NOUN
fcis-30157	325	11	with	with	ADP
fcis-30157	325	12	clinical	clinical	ADJ
fcis-30157	325	13	knowledge	knowledge	NOUN
fcis-30157	325	14	had	have	VERB
fcis-30157	325	15	a	a	DET
fcis-30157	325	16	19.8	19.8	NUM
fcis-30157	325	17	%	%	NOUN
fcis-30157	325	18	lower	low	ADJ
fcis-30157	325	19	misdiagnosis	misdiagnosis	NOUN
fcis-30157	325	20	rate	rate	NOUN
fcis-30157	325	21	compared	compare	VERB
fcis-30157	325	22	to	to	ADP
fcis-30157	325	23	pure	pure	ADJ
fcis-30157	325	24	image	image	NOUN
fcis-30157	325	25	recognition	recognition	NOUN
fcis-30157	325	26	systems	system	NOUN
fcis-30157	325	27	,	,	PUNCT
fcis-30157	325	28	particularly	particularly	ADV
fcis-30157	325	29	in	in	ADP
fcis-30157	325	30	diseases	disease	NOUN
fcis-30157	325	31	with	with	ADP
fcis-30157	325	32	similar	similar	ADJ
fcis-30157	325	33	clinical	clinical	ADJ
fcis-30157	325	34	manifestations	manifestation	NOUN
fcis-30157	325	35	.	.	PUNCT
fcis-30157	326	1	however	however	ADV
fcis-30157	326	2	,	,	PUNCT
fcis-30157	326	3	data	datum	NOUN
fcis-30157	326	4	deviation	deviation	NOUN
fcis-30157	326	5	remains	remain	VERB
fcis-30157	326	6	an	an	DET
fcis-30157	326	7	issue	issue	NOUN
fcis-30157	326	8	,	,	PUNCT
fcis-30157	326	9	as	as	SCONJ
fcis-30157	326	10	seen	see	VERB
fcis-30157	326	11	in	in	ADP
fcis-30157	326	12	mit	mit	PROPN
fcis-30157	326	13	medical	medical	ADJ
fcis-30157	326	14	ai	ai	PROPN
fcis-30157	326	15	laboratory	laboratory	NOUN
fcis-30157	326	16	's	's	PART
fcis-30157	326	17	finding	finding	NOUN
fcis-30157	326	18	that	that	SCONJ
fcis-30157	326	19	recognition	recognition	NOUN
fcis-30157	326	20	accuracy	accuracy	NOUN
fcis-30157	326	21	drops	drop	NOUN
fcis-30157	326	22	by	by	ADP
fcis-30157	326	23	14.7	14.7	NUM
fcis-30157	326	24	%	%	NOUN
fcis-30157	326	25	for	for	ADP
fcis-30157	326	26	darker	dark	ADJ
fcis-30157	326	27	skin	skin	NOUN
fcis-30157	326	28	types	type	NOUN
fcis-30157	326	29	(	(	PUNCT
fcis-30157	326	30	fitzpatrick	fitzpatrick	PROPN
fcis-30157	326	31	vvi	vvi	PROPN
fcis-30157	326	32	)	)	PUNCT
fcis-30157	327	1	[	[	X
fcis-30157	327	2	29].to	29].to	NUM
fcis-30157	327	3	balance	balance	NOUN
fcis-30157	327	4	performance	performance	NOUN
fcis-30157	327	5	and	and	CCONJ
fcis-30157	327	6	interpretability	interpretability	NOUN
fcis-30157	327	7	,	,	PUNCT
fcis-30157	327	8	hybrid	hybrid	ADJ
fcis-30157	327	9	inference	inference	NOUN
fcis-30157	327	10	architectures	architecture	NOUN
fcis-30157	327	11	like	like	ADP
fcis-30157	327	12	dermexpert	dermexpert	NOUN
fcis-30157	327	13	combine	combine	VERB
fcis-30157	327	14	deep	deep	ADJ
fcis-30157	327	15	neural	neural	ADJ
fcis-30157	327	16	network	network	NOUN
fcis-30157	327	17	outputs	output	NOUN
fcis-30157	327	18	with	with	ADP
fcis-30157	327	19	decision	decision	NOUN
fcis-30157	327	20	tree	tree	NOUN
fcis-30157	327	21	logic	logic	NOUN
fcis-30157	327	22	,	,	PUNCT
fcis-30157	327	23	providing	provide	VERB
fcis-30157	327	24	a	a	DET
fcis-30157	327	25	clinical	clinical	ADJ
fcis-30157	327	26	decision	decision	NOUN
fcis-30157	327	27	chain	chain	NOUN
fcis-30157	327	28	.	.	PUNCT
fcis-30157	328	1	current	current	ADJ
fcis-30157	328	2	research	research	NOUN
fcis-30157	328	3	focuses	focus	VERB
fcis-30157	328	4	on	on	ADP
fcis-30157	328	5	improving	improve	VERB
fcis-30157	328	6	dynamic	dynamic	ADJ
fcis-30157	328	7	adaptability	adaptability	NOUN
fcis-30157	328	8	in	in	ADP
fcis-30157	328	9	knowledge	knowledge	NOUN
fcis-30157	328	10	representation	representation	NOUN
fcis-30157	328	11	.	.	PUNCT
fcis-30157	329	1	for	for	ADP
fcis-30157	329	2	example	example	NOUN
fcis-30157	329	3	,	,	PUNCT
fcis-30157	329	4	stanford	stanford	PROPN
fcis-30157	329	5	’s	’s	PART
fcis-30157	329	6	clinical	clinical	ADJ
fcis-30157	329	7	decision	decision	NOUN
fcis-30157	329	8	memory	memory	NOUN
fcis-30157	329	9	network	network	NOUN
fcis-30157	329	10	(	(	PUNCT
fcis-30157	329	11	cdmn	cdmn	NOUN
fcis-30157	329	12	)	)	PUNCT
fcis-30157	329	13	uses	use	VERB
fcis-30157	329	14	reinforcement	reinforcement	NOUN
fcis-30157	329	15	learning	learning	NOUN
fcis-30157	329	16	to	to	PART
fcis-30157	329	17	simulate	simulate	VERB
fcis-30157	329	18	expert	expert	ADJ
fcis-30157	329	19	diagnostic	diagnostic	ADJ
fcis-30157	329	20	thinking	thinking	NOUN
fcis-30157	329	21	,	,	PUNCT
fcis-30157	329	22	improving	improve	VERB
fcis-30157	329	23	decision	decision	NOUN
fcis-30157	329	24	consistency	consistency	NOUN
fcis-30157	329	25	by	by	ADP
fcis-30157	329	26	22.4	22.4	NUM
fcis-30157	329	27	%	%	NOUN
fcis-30157	329	28	in	in	ADP
fcis-30157	329	29	complex	complex	ADJ
fcis-30157	329	30	cases	case	NOUN
fcis-30157	329	31	[	[	X
fcis-30157	329	32	15	15	NUM
fcis-30157	329	33	]	]	PUNCT
fcis-30157	329	34	.	.	PUNCT
fcis-30157	330	1	the	the	DET
fcis-30157	330	2	md	md	PROPN
fcis-30157	330	3	anderson	anderson	PROPN
fcis-30157	330	4	cancer	cancer	PROPN
fcis-30157	330	5	center	center	PROPN
fcis-30157	330	6	's	's	PART
fcis-30157	330	7	knowledge	knowledge	NOUN
fcis-30157	330	8	evolution	evolution	NOUN
fcis-30157	330	9	framework	framework	NOUN
fcis-30157	330	10	automatically	automatically	ADV
fcis-30157	330	11	fine	fine	ADJ
fcis-30157	330	12	-	-	PUNCT
fcis-30157	330	13	tunes	tune	NOUN
fcis-30157	330	14	models	model	NOUN
fcis-30157	330	15	based	base	VERB
fcis-30157	330	16	on	on	ADP
fcis-30157	330	17	updates	update	NOUN
fcis-30157	330	18	from	from	ADP
fcis-30157	330	19	authoritative	authoritative	ADJ
fcis-30157	330	20	databases	database	NOUN
fcis-30157	330	21	,	,	PUNCT
fcis-30157	330	22	integrating	integrate	VERB
fcis-30157	330	23	new	new	ADJ
fcis-30157	330	24	guidelines	guideline	NOUN
fcis-30157	330	25	within	within	ADP
fcis-30157	330	26	72	72	NUM
fcis-30157	330	27	hours	hour	NOUN
fcis-30157	330	28	.	.	PUNCT
fcis-30157	331	1	these	these	DET
fcis-30157	331	2	breakthroughs	breakthrough	NOUN
fcis-30157	331	3	are	be	AUX
fcis-30157	331	4	transforming	transform	VERB
fcis-30157	331	5	skin	skin	NOUN
fcis-30157	331	6	disease	disease	NOUN
fcis-30157	331	7	diagnostic	diagnostic	ADJ
fcis-30157	331	8	systems	system	NOUN
fcis-30157	331	9	from	from	ADP
fcis-30157	331	10	static	static	ADJ
fcis-30157	331	11	knowledge	knowledge	NOUN
fcis-30157	331	12	bases	basis	NOUN
fcis-30157	331	13	to	to	ADP
fcis-30157	331	14	continuous	continuous	ADJ
fcis-30157	331	15	learning	learning	NOUN
fcis-30157	331	16	agents	agent	NOUN
fcis-30157	331	17	,	,	PUNCT
fcis-30157	331	18	significantly	significantly	ADV
fcis-30157	331	19	enhancing	enhance	VERB
fcis-30157	331	20	clinical	clinical	ADJ
fcis-30157	331	21	applications	application	NOUN
fcis-30157	331	22	.	.	PUNCT
fcis-30157	332	1	2.3.3	2.3.3	NUM
fcis-30157	332	2	.	.	PUNCT
fcis-30157	332	3	decision	decision	NOUN
fcis-30157	332	4	making	make	VERB
fcis-30157	332	5	algorithms	algorithm	NOUN
fcis-30157	332	6	for	for	ADP
fcis-30157	332	7	diagnosis	diagnosis	NOUN
fcis-30157	332	8	decision	decision	NOUN
fcis-30157	332	9	-	-	PUNCT
fcis-30157	332	10	making	make	VERB
fcis-30157	332	11	algorithms	algorithm	NOUN
fcis-30157	332	12	are	be	AUX
fcis-30157	332	13	crucial	crucial	ADJ
fcis-30157	332	14	for	for	ADP
fcis-30157	332	15	integrating	integrate	VERB
fcis-30157	332	16	automatic	automatic	ADJ
fcis-30157	332	17	recognition	recognition	NOUN
fcis-30157	332	18	results	result	NOUN
fcis-30157	332	19	and	and	CCONJ
fcis-30157	332	20	clinical	clinical	ADJ
fcis-30157	332	21	knowledge	knowledge	NOUN
fcis-30157	332	22	in	in	ADP
fcis-30157	332	23	skin	skin	NOUN
fcis-30157	332	24	disease	disease	NOUN
fcis-30157	332	25	diagnosis	diagnosis	NOUN
fcis-30157	332	26	.	.	PUNCT
fcis-30157	333	1	this	this	DET
fcis-30157	333	2	section	section	NOUN
fcis-30157	333	3	examines	examine	VERB
fcis-30157	333	4	rule	rule	NOUN
fcis-30157	333	5	-	-	PUNCT
fcis-30157	333	6	based	base	VERB
fcis-30157	333	7	,	,	PUNCT
fcis-30157	333	8	probability	probability	NOUN
fcis-30157	333	9	-	-	PUNCT
fcis-30157	333	10	based	base	VERB
fcis-30157	333	11	,	,	PUNCT
fcis-30157	333	12	and	and	CCONJ
fcis-30157	333	13	hybrid	hybrid	ADJ
fcis-30157	333	14	algorithms	algorithm	NOUN
fcis-30157	333	15	,	,	PUNCT
fcis-30157	333	16	evaluating	evaluate	VERB
fcis-30157	333	17	their	their	PRON
fcis-30157	333	18	effectiveness	effectiveness	NOUN
fcis-30157	333	19	.	.	PUNCT
fcis-30157	334	1	rule	rule	NOUN
fcis-30157	334	2	-	-	PUNCT
fcis-30157	334	3	based	base	VERB
fcis-30157	334	4	algorithms	algorithm	NOUN
fcis-30157	334	5	rely	rely	VERB
fcis-30157	334	6	on	on	ADP
fcis-30157	334	7	predefined	predefine	VERB
fcis-30157	334	8	clinical	clinical	ADJ
fcis-30157	334	9	rules	rule	NOUN
fcis-30157	334	10	.	.	PUNCT
fcis-30157	335	1	for	for	ADP
fcis-30157	335	2	example	example	NOUN
fcis-30157	335	3	,	,	PUNCT
fcis-30157	335	4	a	a	DET
fcis-30157	335	5	system	system	NOUN
fcis-30157	335	6	for	for	ADP
fcis-30157	335	7	diagnosing	diagnose	VERB
fcis-30157	335	8	actinic	actinic	ADJ
fcis-30157	335	9	keratosis	keratosis	NOUN
fcis-30157	335	10	combines	combine	VERB
fcis-30157	335	11	image	image	NOUN
fcis-30157	335	12	recognition	recognition	NOUN
fcis-30157	335	13	with	with	ADP
fcis-30157	335	14	clinical	clinical	ADJ
fcis-30157	335	15	rules	rule	NOUN
fcis-30157	335	16	,	,	PUNCT
fcis-30157	335	17	achieving	achieve	VERB
fcis-30157	335	18	85	85	NUM
fcis-30157	335	19	%	%	NOUN
fcis-30157	335	20	sensitivity	sensitivity	NOUN
fcis-30157	335	21	and	and	CCONJ
fcis-30157	335	22	90	90	NUM
fcis-30157	335	23	%	%	NOUN
fcis-30157	335	24	specificity	specificity	NOUN
fcis-30157	335	25	.	.	PUNCT
fcis-30157	336	1	these	these	DET
fcis-30157	336	2	algorithms	algorithm	NOUN
fcis-30157	336	3	offer	offer	VERB
fcis-30157	336	4	interpretability	interpretability	NOUN
fcis-30157	336	5	but	but	CCONJ
fcis-30157	336	6	may	may	AUX
fcis-30157	336	7	be	be	AUX
fcis-30157	336	8	limited	limit	VERB
fcis-30157	336	9	in	in	ADP
fcis-30157	336	10	handling	handle	VERB
fcis-30157	336	11	complex	complex	ADJ
fcis-30157	336	12	cases	case	NOUN
fcis-30157	336	13	.	.	PUNCT
fcis-30157	337	1	probability	probability	NOUN
fcis-30157	337	2	-	-	PUNCT
fcis-30157	337	3	based	base	VERB
fcis-30157	337	4	algorithms	algorithm	NOUN
fcis-30157	337	5	use	use	VERB
fcis-30157	337	6	statistical	statistical	ADJ
fcis-30157	337	7	models	model	NOUN
fcis-30157	337	8	like	like	ADP
fcis-30157	337	9	bayesian	bayesian	NOUN
fcis-30157	337	10	networks	network	NOUN
fcis-30157	337	11	or	or	CCONJ
fcis-30157	337	12	decision	decision	NOUN
fcis-30157	337	13	trees	tree	NOUN
fcis-30157	337	14	to	to	PART
fcis-30157	337	15	estimate	estimate	VERB
fcis-30157	337	16	the	the	DET
fcis-30157	337	17	likelihood	likelihood	NOUN
fcis-30157	337	18	of	of	ADP
fcis-30157	337	19	diagnoses	diagnosis	NOUN
fcis-30157	337	20	.	.	PUNCT
fcis-30157	338	1	in	in	ADP
fcis-30157	338	2	the	the	DET
fcis-30157	338	3	isic	isic	PROPN
fcis-30157	338	4	grand	grand	ADJ
fcis-30157	338	5	challenge	challenge	NOUN
fcis-30157	338	6	,	,	PUNCT
fcis-30157	338	7	a	a	DET
fcis-30157	338	8	probability	probability	NOUN
fcis-30157	338	9	-	-	PUNCT
fcis-30157	338	10	based	base	VERB
fcis-30157	338	11	model	model	NOUN
fcis-30157	338	12	achieved	achieve	VERB
fcis-30157	338	13	88	88	NUM
fcis-30157	338	14	%	%	NOUN
fcis-30157	338	15	accuracy	accuracy	NOUN
fcis-30157	338	16	in	in	ADP
fcis-30157	338	17	classifying	classify	VERB
fcis-30157	338	18	skin	skin	NOUN
fcis-30157	338	19	cancer	cancer	NOUN
fcis-30157	338	20	,	,	PUNCT
fcis-30157	338	21	outperforming	outperform	VERB
fcis-30157	338	22	rule	rule	NOUN
fcis-30157	338	23	-	-	PUNCT
fcis-30157	338	24	based	base	VERB
fcis-30157	338	25	systems	system	NOUN
fcis-30157	338	26	in	in	ADP
fcis-30157	338	27	complex	complex	ADJ
fcis-30157	338	28	scenarios	scenario	NOUN
fcis-30157	338	29	.	.	PUNCT
fcis-30157	339	1	these	these	DET
fcis-30157	339	2	algorithms	algorithm	NOUN
fcis-30157	339	3	excel	excel	VERB
fcis-30157	339	4	in	in	ADP
fcis-30157	339	5	distinguishing	distinguish	VERB
fcis-30157	339	6	between	between	ADP
fcis-30157	339	7	diseases	disease	NOUN
fcis-30157	339	8	with	with	ADP
fcis-30157	339	9	similar	similar	ADJ
fcis-30157	339	10	features	feature	NOUN
fcis-30157	339	11	.	.	PUNCT
fcis-30157	340	1	hybrid	hybrid	ADJ
fcis-30157	340	2	approaches	approach	NOUN
fcis-30157	340	3	,	,	PUNCT
fcis-30157	340	4	combining	combine	VERB
fcis-30157	340	5	both	both	CCONJ
fcis-30157	340	6	rule	rule	NOUN
fcis-30157	340	7	-	-	PUNCT
fcis-30157	340	8	based	base	VERB
fcis-30157	340	9	and	and	CCONJ
fcis-30157	340	10	probabilitybased	probabilitybase	VERB
fcis-30157	340	11	methods	method	NOUN
fcis-30157	340	12	,	,	PUNCT
fcis-30157	340	13	generally	generally	ADV
fcis-30157	340	14	perform	perform	VERB
fcis-30157	340	15	best	good	ADJ
fcis-30157	340	16	.	.	PUNCT
fcis-30157	341	1	for	for	ADP
fcis-30157	341	2	instance	instance	NOUN
fcis-30157	341	3	,	,	PUNCT
fcis-30157	341	4	a	a	DET
fcis-30157	341	5	dual	dual	ADJ
fcis-30157	341	6	attention	attention	NOUN
fcis-30157	341	7	-	-	PUNCT
fcis-30157	341	8	based	base	VERB
fcis-30157	341	9	network	network	NOUN
fcis-30157	341	10	achieved	achieve	VERB
fcis-30157	341	11	92	92	NUM
fcis-30157	341	12	%	%	NOUN
fcis-30157	341	13	accuracy	accuracy	NOUN
fcis-30157	341	14	,	,	PUNCT
fcis-30157	341	15	integrating	integrate	VERB
fcis-30157	341	16	deep	deep	ADJ
fcis-30157	341	17	learning	learning	NOUN
fcis-30157	341	18	with	with	ADP
fcis-30157	341	19	clinical	clinical	ADJ
fcis-30157	341	20	rules	rule	NOUN
fcis-30157	341	21	.	.	PUNCT
fcis-30157	342	1	this	this	DET
fcis-30157	342	2	hybrid	hybrid	ADJ
fcis-30157	342	3	approach	approach	NOUN
fcis-30157	342	4	balances	balance	VERB
fcis-30157	342	5	interpretability	interpretability	NOUN
fcis-30157	342	6	with	with	ADP
fcis-30157	342	7	statistical	statistical	ADJ
fcis-30157	342	8	power	power	NOUN
fcis-30157	342	9	,	,	PUNCT
fcis-30157	342	10	leading	lead	VERB
fcis-30157	342	11	to	to	ADP
fcis-30157	342	12	more	more	ADV
fcis-30157	342	13	accurate	accurate	ADJ
fcis-30157	342	14	diagnoses	diagnosis	NOUN
fcis-30157	342	15	.	.	PUNCT
fcis-30157	343	1	performance	performance	NOUN
fcis-30157	343	2	evaluation	evaluation	NOUN
fcis-30157	343	3	of	of	ADP
fcis-30157	343	4	algorithms	algorithm	NOUN
fcis-30157	343	5	using	use	VERB
fcis-30157	343	6	a	a	DET
fcis-30157	343	7	dataset	dataset	NOUN
fcis-30157	343	8	of	of	ADP
fcis-30157	343	9	10,000	10,000	NUM
fcis-30157	343	10	skin	skin	NOUN
fcis-30157	343	11	lesions	lesion	NOUN
fcis-30157	343	12	shows	show	VERB
fcis-30157	343	13	that	that	SCONJ
fcis-30157	343	14	the	the	DET
fcis-30157	343	15	hybrid	hybrid	ADJ
fcis-30157	343	16	approach	approach	NOUN
fcis-30157	343	17	outperforms	outperform	NOUN
fcis-30157	343	18	both	both	CCONJ
fcis-30157	343	19	rulebased	rulebase	VERB
fcis-30157	343	20	and	and	CCONJ
fcis-30157	343	21	probability	probability	NOUN
fcis-30157	343	22	-	-	PUNCT
fcis-30157	343	23	based	base	VERB
fcis-30157	343	24	algorithms	algorithm	NOUN
fcis-30157	343	25	in	in	ADP
fcis-30157	343	26	accuracy	accuracy	NOUN
fcis-30157	343	27	,	,	PUNCT
fcis-30157	343	28	sensitivity	sensitivity	NOUN
fcis-30157	343	29	,	,	PUNCT
fcis-30157	343	30	and	and	CCONJ
fcis-30157	343	31	specificity	specificity	NOUN
fcis-30157	343	32	(	(	PUNCT
fcis-30157	343	33	table	table	NOUN
fcis-30157	343	34	2	2	NUM
fcis-30157	343	35	)	)	PUNCT
fcis-30157	343	36	.	.	PUNCT
fcis-30157	344	1	table	table	NOUN
fcis-30157	344	2	2	2	NUM
fcis-30157	344	3	.	.	PUNCT
fcis-30157	344	4	performance	performance	NOUN
fcis-30157	344	5	of	of	ADP
fcis-30157	344	6	decision	decision	NOUN
fcis-30157	344	7	-	-	PUNCT
fcis-30157	344	8	making	make	VERB
fcis-30157	344	9	algorithms	algorithm	NOUN
fcis-30157	344	10	in	in	ADP
fcis-30157	344	11	skin	skin	NOUN
fcis-30157	344	12	disease	disease	NOUN
fcis-30157	344	13	diagnosis	diagnosis	NOUN
fcis-30157	344	14	algorithm	algorithm	NOUN
fcis-30157	344	15	type	type	NOUN
fcis-30157	344	16	accuracy	accuracy	NOUN
fcis-30157	344	17	(	(	PUNCT
fcis-30157	344	18	%	%	INTJ
fcis-30157	344	19	)	)	PUNCT
fcis-30157	344	20	sensitivity	sensitivity	NOUN
fcis-30157	344	21	(	(	PUNCT
fcis-30157	344	22	%	%	INTJ
fcis-30157	344	23	)	)	PUNCT
fcis-30157	344	24	specificity	specificity	NOUN
fcis-30157	344	25	(	(	PUNCT
fcis-30157	344	26	%	%	INTJ
fcis-30157	344	27	)	)	PUNCT
fcis-30157	344	28	rule	rule	NOUN
fcis-30157	344	29	-	-	PUNCT
fcis-30157	344	30	based	base	VERB
fcis-30157	344	31	85	85	NUM
fcis-30157	344	32	80	80	NUM
fcis-30157	344	33	90	90	NUM
fcis-30157	344	34	probability	probability	NOUN
fcis-30157	344	35	-	-	PUNCT
fcis-30157	344	36	based	base	VERB
fcis-30157	344	37	88	88	NUM
fcis-30157	344	38	85	85	NUM
fcis-30157	344	39	91	91	NUM
fcis-30157	344	40	hybrid	hybrid	NOUN
fcis-30157	344	41	92	92	NUM
fcis-30157	344	42	90	90	NUM
fcis-30157	344	43	93	93	NUM
fcis-30157	344	44	optimization	optimization	NOUN
fcis-30157	344	45	techniques	technique	NOUN
fcis-30157	344	46	further	far	ADV
fcis-30157	344	47	enhance	enhance	VERB
fcis-30157	344	48	performance	performance	NOUN
fcis-30157	344	49	.	.	PUNCT
fcis-30157	345	1	ensemble	ensemble	ADJ
fcis-30157	345	2	methods	method	NOUN
fcis-30157	345	3	,	,	PUNCT
fcis-30157	345	4	like	like	ADP
fcis-30157	345	5	multi	multi	ADJ
fcis-30157	345	6	-	-	ADJ
fcis-30157	345	7	task	task	ADJ
fcis-30157	345	8	models	model	NOUN
fcis-30157	345	9	,	,	PUNCT
fcis-30157	345	10	combine	combine	VERB
fcis-30157	345	11	multiple	multiple	ADJ
fcis-30157	345	12	outputs	output	NOUN
fcis-30157	345	13	to	to	PART
fcis-30157	345	14	improve	improve	VERB
fcis-30157	345	15	generalization	generalization	NOUN
fcis-30157	345	16	,	,	PUNCT
fcis-30157	345	17	with	with	ADP
fcis-30157	345	18	one	one	NUM
fcis-30157	345	19	study	study	NOUN
fcis-30157	345	20	achieving	achieve	VERB
fcis-30157	345	21	94	94	NUM
fcis-30157	345	22	%	%	NOUN
fcis-30157	345	23	diagnostic	diagnostic	ADJ
fcis-30157	345	24	accuracy	accuracy	NOUN
fcis-30157	345	25	[	[	X
fcis-30157	345	26	35	35	NUM
fcis-30157	345	27	]	]	PUNCT
fcis-30157	345	28	.	.	PUNCT
fcis-30157	346	1	additionally	additionally	ADV
fcis-30157	346	2	,	,	PUNCT
fcis-30157	346	3	explainable	explainable	ADJ
fcis-30157	346	4	ai	ai	NOUN
fcis-30157	346	5	(	(	PUNCT
fcis-30157	346	6	xai	xai	PROPN
fcis-30157	346	7	)	)	PUNCT
fcis-30157	346	8	improves	improve	VERB
fcis-30157	346	9	transparency	transparency	NOUN
fcis-30157	346	10	,	,	PUNCT
fcis-30157	346	11	helping	help	VERB
fcis-30157	346	12	clinicians	clinician	NOUN
fcis-30157	346	13	understand	understand	VERB
fcis-30157	346	14	the	the	DET
fcis-30157	346	15	system	system	NOUN
fcis-30157	346	16	’s	’s	PART
fcis-30157	346	17	decisions	decision	NOUN
fcis-30157	346	18	.	.	PUNCT
fcis-30157	347	1	mathematical	mathematical	ADJ
fcis-30157	347	2	formulation	formulation	NOUN
fcis-30157	347	3	of	of	ADP
fcis-30157	347	4	decision	decision	NOUN
fcis-30157	347	5	-	-	PUNCT
fcis-30157	347	6	making	making	NOUN
fcis-30157	347	7	involves	involve	NOUN
fcis-30157	347	8	maximizing	maximize	VERB
fcis-30157	347	9	posterior	posterior	ADJ
fcis-30157	347	10	probability	probability	NOUN
fcis-30157	347	11	using	use	VERB
fcis-30157	347	12	bayes	baye	NOUN
fcis-30157	347	13	'	'	PART
fcis-30157	347	14	theorem	theorem	NOUN
fcis-30157	347	15	:	:	PUNCT
fcis-30157	347	16	this	this	DET
fcis-30157	347	17	formulation	formulation	NOUN
fcis-30157	347	18	aids	aid	NOUN
fcis-30157	347	19	in	in	ADP
fcis-30157	347	20	optimizing	optimize	VERB
fcis-30157	347	21	diagnosis	diagnosis	NOUN
fcis-30157	347	22	by	by	ADP
fcis-30157	347	23	integrating	integrate	VERB
fcis-30157	347	24	features	feature	NOUN
fcis-30157	347	25	extracted	extract	VERB
fcis-30157	347	26	from	from	ADP
fcis-30157	347	27	skin	skin	NOUN
fcis-30157	347	28	lesion	lesion	NOUN
fcis-30157	347	29	images	image	NOUN
fcis-30157	347	30	and	and	CCONJ
fcis-30157	347	31	patient	patient	ADJ
fcis-30157	347	32	data	datum	NOUN
fcis-30157	347	33	.	.	PUNCT
fcis-30157	348	1	case	case	NOUN
fcis-30157	348	2	study	study	NOUN
fcis-30157	348	3	:	:	PUNCT
fcis-30157	348	4	a	a	DET
fcis-30157	348	5	hybrid	hybrid	ADJ
fcis-30157	348	6	algorithm	algorithm	NOUN
fcis-30157	348	7	was	be	AUX
fcis-30157	348	8	used	use	VERB
fcis-30157	348	9	for	for	ADP
fcis-30157	348	10	the	the	DET
fcis-30157	348	11	differential	differential	ADJ
fcis-30157	348	12	diagnosis	diagnosis	NOUN
fcis-30157	348	13	of	of	ADP
fcis-30157	348	14	contact	contact	NOUN
fcis-30157	348	15	dermatitis	dermatitis	NOUN
fcis-30157	348	16	,	,	PUNCT
fcis-30157	348	17	with	with	ADP
fcis-30157	348	18	high	high	ADJ
fcis-30157	348	19	accuracy	accuracy	NOUN
fcis-30157	348	20	,	,	PUNCT
fcis-30157	348	21	sensitivity	sensitivity	NOUN
fcis-30157	348	22	,	,	PUNCT
fcis-30157	348	23	and	and	CCONJ
fcis-30157	348	24	specificity	specificity	NOUN
fcis-30157	348	25	for	for	ADP
fcis-30157	348	26	irritant	irritant	ADJ
fcis-30157	348	27	contact	contact	NOUN
fcis-30157	348	28	dermatitis	dermatitis	NOUN
fcis-30157	348	29	(	(	PUNCT
fcis-30157	348	30	90	90	NUM
fcis-30157	348	31	%	%	NOUN
fcis-30157	348	32	,	,	PUNCT
fcis-30157	348	33	88	88	NUM
fcis-30157	348	34	%	%	NOUN
fcis-30157	348	35	,	,	PUNCT
fcis-30157	348	36	92	92	NUM
fcis-30157	348	37	%	%	NOUN
fcis-30157	348	38	)	)	PUNCT
fcis-30157	348	39	,	,	PUNCT
fcis-30157	348	40	allergic	allergic	ADJ
fcis-30157	348	41	contact	contact	NOUN
fcis-30157	348	42	dermatitis	dermatitis	NOUN
fcis-30157	348	43	(	(	PUNCT
fcis-30157	348	44	87	87	NUM
fcis-30157	348	45	%	%	NOUN
fcis-30157	348	46	,	,	PUNCT
fcis-30157	348	47	85	85	NUM
fcis-30157	348	48	%	%	NOUN
fcis-30157	348	49	,	,	PUNCT
fcis-30157	348	50	90	90	NUM
fcis-30157	348	51	%	%	NOUN
fcis-30157	348	52	)	)	PUNCT
fcis-30157	348	53	,	,	PUNCT
fcis-30157	348	54	and	and	CCONJ
fcis-30157	348	55	atopic	atopic	NOUN
fcis-30157	348	56	dermatitis	dermatitis	NOUN
fcis-30157	348	57	(	(	PUNCT
fcis-30157	348	58	85	85	NUM
fcis-30157	348	59	%	%	NOUN
fcis-30157	348	60	,	,	PUNCT
fcis-30157	348	61	82	82	NUM
fcis-30157	348	62	%	%	NOUN
fcis-30157	348	63	,	,	PUNCT
fcis-30157	348	64	88	88	NUM
fcis-30157	348	65	%	%	NOUN
fcis-30157	348	66	)	)	PUNCT
fcis-30157	348	67	(	(	PUNCT
fcis-30157	348	68	table	table	NOUN
fcis-30157	348	69	3	3	NUM
fcis-30157	348	70	)	)	PUNCT
fcis-30157	348	71	.	.	PUNCT
fcis-30157	349	1	in	in	ADP
fcis-30157	349	2	conclusion	conclusion	NOUN
fcis-30157	349	3	,	,	PUNCT
fcis-30157	349	4	decision	decision	NOUN
fcis-30157	349	5	-	-	PUNCT
fcis-30157	349	6	making	make	VERB
fcis-30157	349	7	algorithms	algorithm	NOUN
fcis-30157	349	8	are	be	AUX
fcis-30157	349	9	essential	essential	ADJ
fcis-30157	349	10	for	for	ADP
fcis-30157	349	11	accurate	accurate	ADJ
fcis-30157	349	12	skin	skin	NOUN
fcis-30157	349	13	disease	disease	NOUN
fcis-30157	349	14	diagnosis	diagnosis	NOUN
fcis-30157	349	15	.	.	PUNCT
fcis-30157	350	1	hybrid	hybrid	ADJ
fcis-30157	350	2	approaches	approach	NOUN
fcis-30157	350	3	offer	offer	VERB
fcis-30157	350	4	the	the	DET
fcis-30157	350	5	best	good	ADJ
fcis-30157	350	6	performance	performance	NOUN
fcis-30157	350	7	,	,	PUNCT
fcis-30157	350	8	combining	combine	VERB
fcis-30157	350	9	rule	rule	NOUN
fcis-30157	350	10	-	-	PUNCT
fcis-30157	350	11	based	base	VERB
fcis-30157	350	12	interpretability	interpretability	NOUN
fcis-30157	350	13	with	with	ADP
fcis-30157	350	14	probability	probability	NOUN
fcis-30157	350	15	-	-	PUNCT
fcis-30157	350	16	based	base	VERB
fcis-30157	350	17	precision	precision	NOUN
fcis-30157	350	18	.	.	PUNCT
fcis-30157	351	1	optimization	optimization	NOUN
fcis-30157	351	2	techniques	technique	NOUN
fcis-30157	351	3	like	like	ADP
fcis-30157	351	4	ensemble	ensemble	ADJ
fcis-30157	351	5	methods	method	NOUN
fcis-30157	351	6	and	and	CCONJ
fcis-30157	351	7	explainable	explainable	ADJ
fcis-30157	351	8	ai	ai	VERB
fcis-30157	351	9	improve	improve	VERB
fcis-30157	351	10	their	their	PRON
fcis-30157	351	11	effectiveness	effectiveness	NOUN
fcis-30157	351	12	.	.	PUNCT
fcis-30157	352	1	future	future	ADJ
fcis-30157	352	2	research	research	NOUN
fcis-30157	352	3	should	should	AUX
fcis-30157	352	4	focus	focus	VERB
fcis-30157	352	5	on	on	ADP
fcis-30157	352	6	enhancing	enhance	VERB
fcis-30157	352	7	these	these	DET
fcis-30157	352	8	models	model	NOUN
fcis-30157	352	9	with	with	ADP
fcis-30157	352	10	additional	additional	ADJ
fcis-30157	352	11	data	datum	NOUN
fcis-30157	352	12	sources	source	NOUN
fcis-30157	352	13	,	,	PUNCT
fcis-30157	352	14	such	such	ADJ
fcis-30157	352	15	as	as	ADP
fcis-30157	352	16	genetic	genetic	ADJ
fcis-30157	352	17	and	and	CCONJ
fcis-30157	352	18	environmental	environmental	ADJ
fcis-30157	352	19	factors	factor	NOUN
fcis-30157	352	20	,	,	PUNCT
fcis-30157	352	21	to	to	PART
fcis-30157	352	22	further	far	ADV
fcis-30157	352	23	improve	improve	VERB
fcis-30157	352	24	diagnostic	diagnostic	ADJ
fcis-30157	352	25	accuracy	accuracy	NOUN
fcis-30157	352	26	.	.	PUNCT
fcis-30157	353	1	119	119	NUM
fcis-30157	353	2	table	table	NOUN
fcis-30157	353	3	3	3	NUM
fcis-30157	353	4	.	.	PUNCT
fcis-30157	353	5	differential	differential	ADJ
fcis-30157	353	6	diagnosis	diagnosis	NOUN
fcis-30157	353	7	of	of	ADP
fcis-30157	353	8	contact	contact	NOUN
fcis-30157	353	9	dermatitis	dermatitis	NOUN
fcis-30157	353	10	using	use	VERB
fcis-30157	353	11	hybrid	hybrid	ADJ
fcis-30157	353	12	decision	decision	NOUN
fcis-30157	353	13	-	-	PUNCT
fcis-30157	353	14	making	make	VERB
fcis-30157	353	15	algorithm	algorithm	NOUN
fcis-30157	353	16	diagnosis	diagnosis	NOUN
fcis-30157	353	17	type	type	NOUN
fcis-30157	353	18	accuracy	accuracy	NOUN
fcis-30157	353	19	(	(	PUNCT
fcis-30157	353	20	%	%	INTJ
fcis-30157	353	21	)	)	PUNCT
fcis-30157	353	22	sensitivity	sensitivity	NOUN
fcis-30157	353	23	(	(	PUNCT
fcis-30157	353	24	%	%	INTJ
fcis-30157	353	25	)	)	PUNCT
fcis-30157	353	26	specificity	specificity	NOUN
fcis-30157	353	27	(	(	PUNCT
fcis-30157	353	28	%	%	INTJ
fcis-30157	353	29	)	)	PUNCT
fcis-30157	353	30	irritant	irritant	VERB
fcis-30157	353	31	contact	contact	NOUN
fcis-30157	353	32	dermatitis	dermatitis	NOUN
fcis-30157	353	33	90	90	NUM
fcis-30157	353	34	88	88	NUM
fcis-30157	353	35	92	92	NUM
fcis-30157	353	36	allergic	allergic	ADJ
fcis-30157	353	37	contact	contact	NOUN
fcis-30157	353	38	dermatitis	dermatitis	NOUN
fcis-30157	353	39	87	87	NUM
fcis-30157	353	40	85	85	NUM
fcis-30157	353	41	90	90	NUM
fcis-30157	353	42	atopic	atopic	NOUN
fcis-30157	353	43	dermatitis	dermatitis	NOUN
fcis-30157	353	44	85	85	NUM
fcis-30157	353	45	82	82	NUM
fcis-30157	353	46	88	88	NUM
fcis-30157	353	47	3	3	NUM
fcis-30157	353	48	.	.	PUNCT
fcis-30157	353	49	experiments	experiment	NOUN
fcis-30157	353	50	and	and	CCONJ
fcis-30157	353	51	results	result	NOUN
fcis-30157	353	52	3.1	3.1	NUM
fcis-30157	353	53	.	.	PUNCT
fcis-30157	354	1	experimental	experimental	ADJ
fcis-30157	354	2	setup	setup	NOUN
fcis-30157	354	3	3.1.1	3.1.1	NUM
fcis-30157	354	4	.	.	PUNCT
fcis-30157	355	1	datasets	dataset	NOUN
fcis-30157	355	2	used	use	VERB
fcis-30157	355	3	for	for	ADP
fcis-30157	355	4	experiments	experiment	NOUN
fcis-30157	355	5	this	this	DET
fcis-30157	355	6	study	study	NOUN
fcis-30157	355	7	utilizes	utilize	VERB
fcis-30157	355	8	a	a	DET
fcis-30157	355	9	comprehensive	comprehensive	ADJ
fcis-30157	355	10	dataset	dataset	NOUN
fcis-30157	355	11	sourced	source	VERB
fcis-30157	355	12	from	from	ADP
fcis-30157	355	13	hospital	hospital	NOUN
fcis-30157	355	14	case	case	NOUN
fcis-30157	355	15	data	datum	NOUN
fcis-30157	355	16	and	and	CCONJ
fcis-30157	355	17	publicly	publicly	ADV
fcis-30157	355	18	available	available	ADJ
fcis-30157	355	19	medical	medical	ADJ
fcis-30157	355	20	image	image	NOUN
fcis-30157	355	21	databases	database	NOUN
fcis-30157	355	22	,	,	PUNCT
fcis-30157	355	23	ensuring	ensure	VERB
fcis-30157	355	24	diverse	diverse	ADJ
fcis-30157	355	25	and	and	CCONJ
fcis-30157	355	26	representative	representative	ADJ
fcis-30157	355	27	skin	skin	NOUN
fcis-30157	355	28	disease	disease	NOUN
fcis-30157	355	29	images	image	NOUN
fcis-30157	355	30	.	.	PUNCT
fcis-30157	356	1	hospital	hospital	NOUN
fcis-30157	356	2	data	datum	NOUN
fcis-30157	356	3	includes	include	VERB
fcis-30157	356	4	over	over	ADP
fcis-30157	356	5	10,000	10,000	NUM
fcis-30157	356	6	images	image	NOUN
fcis-30157	356	7	from	from	ADP
fcis-30157	356	8	three	three	NUM
fcis-30157	356	9	major	major	ADJ
fcis-30157	356	10	tertiary	tertiary	ADJ
fcis-30157	356	11	hospitals	hospital	NOUN
fcis-30157	356	12	,	,	PUNCT
fcis-30157	356	13	featuring	feature	VERB
fcis-30157	356	14	conditions	condition	NOUN
fcis-30157	356	15	like	like	ADP
fcis-30157	356	16	melanoma	melanoma	NOUN
fcis-30157	356	17	,	,	PUNCT
fcis-30157	356	18	psoriasis	psoriasis	NOUN
fcis-30157	356	19	,	,	PUNCT
fcis-30157	356	20	and	and	CCONJ
fcis-30157	356	21	eczema	eczema	NOUN
fcis-30157	356	22	,	,	PUNCT
fcis-30157	356	23	all	all	PRON
fcis-30157	356	24	annotated	annotate	VERB
fcis-30157	356	25	by	by	ADP
fcis-30157	356	26	board	board	NOUN
fcis-30157	356	27	-	-	PUNCT
fcis-30157	356	28	certified	certify	VERB
fcis-30157	356	29	dermatologists	dermatologist	NOUN
fcis-30157	356	30	.	.	PUNCT
fcis-30157	357	1	public	public	ADJ
fcis-30157	357	2	repositories	repository	NOUN
fcis-30157	357	3	such	such	ADJ
fcis-30157	357	4	as	as	ADP
fcis-30157	357	5	the	the	DET
fcis-30157	357	6	isic	isic	PROPN
fcis-30157	357	7	archive	archive	NOUN
fcis-30157	357	8	(	(	PUNCT
fcis-30157	357	9	23,000	23,000	NUM
fcis-30157	357	10	dermoscopy	dermoscopy	NOUN
fcis-30157	357	11	images	image	NOUN
fcis-30157	357	12	)	)	PUNCT
fcis-30157	357	13	and	and	CCONJ
fcis-30157	357	14	dermnet	dermnet	NOUN
fcis-30157	357	15	(	(	PUNCT
fcis-30157	357	16	diverse	diverse	ADJ
fcis-30157	357	17	clinical	clinical	ADJ
fcis-30157	357	18	images	image	NOUN
fcis-30157	357	19	)	)	PUNCT
fcis-30157	357	20	complement	complement	VERB
fcis-30157	357	21	this	this	PRON
fcis-30157	357	22	,	,	PUNCT
fcis-30157	357	23	providing	provide	VERB
fcis-30157	357	24	metadata	metadata	NOUN
fcis-30157	357	25	crucial	crucial	ADJ
fcis-30157	357	26	for	for	ADP
fcis-30157	357	27	ai	ai	PROPN
fcis-30157	357	28	model	model	NOUN
fcis-30157	357	29	training	training	NOUN
fcis-30157	357	30	.	.	PUNCT
fcis-30157	358	1	the	the	DET
fcis-30157	358	2	combined	combine	VERB
fcis-30157	358	3	dataset	dataset	NOUN
fcis-30157	358	4	totals	total	NOUN
fcis-30157	358	5	over	over	ADP
fcis-30157	358	6	35,000	35,000	NUM
fcis-30157	358	7	images	image	NOUN
fcis-30157	358	8	covering	cover	VERB
fcis-30157	358	9	more	more	ADJ
fcis-30157	358	10	than	than	ADP
fcis-30157	358	11	20	20	NUM
fcis-30157	358	12	skin	skin	NOUN
fcis-30157	358	13	diseases	disease	NOUN
fcis-30157	358	14	,	,	PUNCT
fcis-30157	358	15	with	with	ADP
fcis-30157	358	16	variation	variation	NOUN
fcis-30157	358	17	in	in	ADP
fcis-30157	358	18	resolution	resolution	NOUN
fcis-30157	358	19	,	,	PUNCT
fcis-30157	358	20	lighting	lighting	NOUN
fcis-30157	358	21	,	,	PUNCT
fcis-30157	358	22	and	and	CCONJ
fcis-30157	358	23	lesion	lesion	NOUN
fcis-30157	358	24	visibility	visibility	NOUN
fcis-30157	358	25	,	,	PUNCT
fcis-30157	358	26	reflecting	reflect	VERB
fcis-30157	358	27	real	real	ADJ
fcis-30157	358	28	-	-	PUNCT
fcis-30157	358	29	world	world	NOUN
fcis-30157	358	30	clinical	clinical	ADJ
fcis-30157	358	31	challenges	challenge	NOUN
fcis-30157	358	32	.	.	PUNCT
fcis-30157	359	1	rigorous	rigorous	ADJ
fcis-30157	359	2	preprocessing	preprocessing	NOUN
fcis-30157	359	3	steps	step	NOUN
fcis-30157	359	4	,	,	PUNCT
fcis-30157	359	5	including	include	VERB
fcis-30157	359	6	noise	noise	NOUN
fcis-30157	359	7	reduction	reduction	NOUN
fcis-30157	359	8	and	and	CCONJ
fcis-30157	359	9	contrast	contrast	NOUN
fcis-30157	359	10	enhancement	enhancement	NOUN
fcis-30157	359	11	,	,	PUNCT
fcis-30157	359	12	were	be	AUX
fcis-30157	359	13	followed	follow	VERB
fcis-30157	359	14	by	by	ADP
fcis-30157	359	15	manual	manual	ADJ
fcis-30157	359	16	review	review	NOUN
fcis-30157	359	17	to	to	PART
fcis-30157	359	18	correct	correct	VERB
fcis-30157	359	19	errors	error	NOUN
fcis-30157	359	20	in	in	ADP
fcis-30157	359	21	annotations	annotation	NOUN
fcis-30157	359	22	(	(	PUNCT
fcis-30157	359	23	e.g.	e.g.	ADV
fcis-30157	359	24	,	,	PUNCT
fcis-30157	359	25	5	5	NUM
fcis-30157	359	26	%	%	NOUN
fcis-30157	359	27	of	of	ADP
fcis-30157	359	28	500	500	NUM
fcis-30157	359	29	sample	sample	NOUN
fcis-30157	359	30	images	image	NOUN
fcis-30157	359	31	were	be	AUX
fcis-30157	359	32	corrected	correct	VERB
fcis-30157	359	33	)	)	PUNCT
fcis-30157	359	34	.	.	PUNCT
fcis-30157	360	1	the	the	DET
fcis-30157	360	2	dataset	dataset	NOUN
fcis-30157	360	3	balances	balance	NOUN
fcis-30157	360	4	common	common	ADJ
fcis-30157	360	5	and	and	CCONJ
fcis-30157	360	6	rare	rare	ADJ
fcis-30157	360	7	diseases	disease	NOUN
fcis-30157	360	8	:	:	PUNCT
fcis-30157	360	9	melanoma	melanoma	NOUN
fcis-30157	360	10	(	(	PUNCT
fcis-30157	360	11	25	25	NUM
fcis-30157	360	12	%	%	NOUN
fcis-30157	360	13	)	)	PUNCT
fcis-30157	360	14	,	,	PUNCT
fcis-30157	360	15	psoriasis	psoriasis	NOUN
fcis-30157	360	16	(	(	PUNCT
fcis-30157	360	17	20	20	NUM
fcis-30157	360	18	%	%	NOUN
fcis-30157	360	19	)	)	PUNCT
fcis-30157	360	20	,	,	PUNCT
fcis-30157	360	21	eczema	eczema	NOUN
fcis-30157	360	22	(	(	PUNCT
fcis-30157	360	23	15	15	NUM
fcis-30157	360	24	%	%	NOUN
fcis-30157	360	25	)	)	PUNCT
fcis-30157	360	26	,	,	PUNCT
fcis-30157	360	27	and	and	CCONJ
fcis-30157	360	28	others	other	NOUN
fcis-30157	360	29	like	like	ADP
fcis-30157	360	30	lichen	lichen	NOUN
fcis-30157	360	31	planus	planus	NOUN
fcis-30157	360	32	and	and	CCONJ
fcis-30157	360	33	seborrheic	seborrheic	VERB
fcis-30157	360	34	keratosis	keratosis	NOUN
fcis-30157	360	35	.	.	PUNCT
fcis-30157	361	1	it	it	PRON
fcis-30157	361	2	includes	include	VERB
fcis-30157	361	3	images	image	NOUN
fcis-30157	361	4	from	from	ADP
fcis-30157	361	5	various	various	ADJ
fcis-30157	361	6	age	age	NOUN
fcis-30157	361	7	groups	group	NOUN
fcis-30157	361	8	,	,	PUNCT
fcis-30157	361	9	genders	gender	NOUN
fcis-30157	361	10	,	,	PUNCT
fcis-30157	361	11	and	and	CCONJ
fcis-30157	361	12	ethnicities	ethnicity	NOUN
fcis-30157	361	13	,	,	PUNCT
fcis-30157	361	14	with	with	ADP
fcis-30157	361	15	about	about	ADP
fcis-30157	361	16	30	30	NUM
fcis-30157	361	17	%	%	NOUN
fcis-30157	361	18	from	from	ADP
fcis-30157	361	19	non	non	ADJ
fcis-30157	361	20	-	-	ADJ
fcis-30157	361	21	caucasian	caucasian	ADJ
fcis-30157	361	22	backgrounds	background	NOUN
fcis-30157	361	23	.	.	PUNCT
fcis-30157	362	1	analyzed	analyze	VERB
fcis-30157	362	2	image	image	NOUN
fcis-30157	362	3	features	feature	VERB
fcis-30157	362	4	such	such	ADJ
fcis-30157	362	5	as	as	ADP
fcis-30157	362	6	lesion	lesion	NOUN
fcis-30157	362	7	size	size	NOUN
fcis-30157	362	8	,	,	PUNCT
fcis-30157	362	9	shape	shape	NOUN
fcis-30157	362	10	,	,	PUNCT
fcis-30157	362	11	and	and	CCONJ
fcis-30157	362	12	texture	texture	NOUN
fcis-30157	362	13	indicate	indicate	VERB
fcis-30157	362	14	that	that	SCONJ
fcis-30157	362	15	40	40	NUM
fcis-30157	362	16	%	%	NOUN
fcis-30157	362	17	have	have	VERB
fcis-30157	362	18	irregular	irregular	ADJ
fcis-30157	362	19	borders	border	NOUN
fcis-30157	362	20	,	,	PUNCT
fcis-30157	362	21	35	35	NUM
fcis-30157	362	22	%	%	NOUN
fcis-30157	362	23	show	show	NOUN
fcis-30157	362	24	heterogeneous	heterogeneous	ADJ
fcis-30157	362	25	coloration	coloration	NOUN
fcis-30157	362	26	,	,	PUNCT
fcis-30157	362	27	and	and	CCONJ
fcis-30157	362	28	25	25	NUM
fcis-30157	362	29	%	%	NOUN
fcis-30157	362	30	combine	combine	VERB
fcis-30157	362	31	both	both	DET
fcis-30157	362	32	features	feature	NOUN
fcis-30157	362	33	.	.	PUNCT
fcis-30157	363	1	the	the	DET
fcis-30157	363	2	dataset	dataset	NOUN
fcis-30157	363	3	's	's	PART
fcis-30157	363	4	quality	quality	NOUN
fcis-30157	363	5	and	and	CCONJ
fcis-30157	363	6	diversity	diversity	NOUN
fcis-30157	363	7	ensure	ensure	VERB
fcis-30157	363	8	the	the	DET
fcis-30157	363	9	reliability	reliability	NOUN
fcis-30157	363	10	and	and	CCONJ
fcis-30157	363	11	generalization	generalization	NOUN
fcis-30157	363	12	of	of	ADP
fcis-30157	363	13	ai	ai	PROPN
fcis-30157	363	14	models	model	NOUN
fcis-30157	363	15	,	,	PUNCT
fcis-30157	363	16	minimizing	minimize	VERB
fcis-30157	363	17	bias	bias	NOUN
fcis-30157	363	18	and	and	CCONJ
fcis-30157	363	19	overfitting	overfitting	NOUN
fcis-30157	363	20	.	.	PUNCT
fcis-30157	364	1	models	model	NOUN
fcis-30157	364	2	trained	train	VERB
fcis-30157	364	3	on	on	ADP
fcis-30157	364	4	such	such	ADJ
fcis-30157	364	5	diverse	diverse	ADJ
fcis-30157	364	6	data	datum	NOUN
fcis-30157	364	7	achieved	achieve	VERB
fcis-30157	364	8	85	85	NUM
fcis-30157	364	9	%	%	NOUN
fcis-30157	364	10	accuracy	accuracy	NOUN
fcis-30157	364	11	,	,	PUNCT
fcis-30157	364	12	compared	compare	VERB
fcis-30157	364	13	to	to	ADP
fcis-30157	364	14	72	72	NUM
fcis-30157	364	15	%	%	NOUN
fcis-30157	364	16	for	for	ADP
fcis-30157	364	17	those	those	PRON
fcis-30157	364	18	trained	train	VERB
fcis-30157	364	19	on	on	ADP
fcis-30157	364	20	homogeneous	homogeneous	ADJ
fcis-30157	364	21	datasets	dataset	NOUN
fcis-30157	364	22	.	.	PUNCT
fcis-30157	365	1	the	the	DET
fcis-30157	365	2	dataset	dataset	NOUN
fcis-30157	365	3	provides	provide	VERB
fcis-30157	365	4	a	a	DET
fcis-30157	365	5	solid	solid	ADJ
fcis-30157	365	6	foundation	foundation	NOUN
fcis-30157	365	7	for	for	ADP
fcis-30157	365	8	training	training	NOUN
fcis-30157	365	9	and	and	CCONJ
fcis-30157	365	10	evaluating	evaluate	VERB
fcis-30157	365	11	skin	skin	NOUN
fcis-30157	365	12	disease	disease	NOUN
fcis-30157	365	13	recognition	recognition	NOUN
fcis-30157	365	14	models	model	NOUN
fcis-30157	365	15	,	,	PUNCT
fcis-30157	365	16	enabling	enable	VERB
fcis-30157	365	17	real	real	ADJ
fcis-30157	365	18	-	-	PUNCT
fcis-30157	365	19	world	world	NOUN
fcis-30157	365	20	applicability	applicability	NOUN
fcis-30157	365	21	and	and	CCONJ
fcis-30157	365	22	the	the	DET
fcis-30157	365	23	exploration	exploration	NOUN
fcis-30157	365	24	of	of	ADP
fcis-30157	365	25	new	new	ADJ
fcis-30157	365	26	diagnostic	diagnostic	ADJ
fcis-30157	365	27	approaches	approach	NOUN
fcis-30157	365	28	.	.	PUNCT
fcis-30157	366	1	3.1.2	3.1.2	X
fcis-30157	366	2	.	.	PUNCT
fcis-30157	366	3	evaluation	evaluation	NOUN
fcis-30157	366	4	metrics	metric	NOUN
fcis-30157	366	5	when	when	SCONJ
fcis-30157	366	6	evaluating	evaluate	VERB
fcis-30157	366	7	automatic	automatic	ADJ
fcis-30157	366	8	identification	identification	NOUN
fcis-30157	366	9	and	and	CCONJ
fcis-30157	366	10	diagnostic	diagnostic	ADJ
fcis-30157	366	11	systems	system	NOUN
fcis-30157	366	12	,	,	PUNCT
fcis-30157	366	13	selecting	select	VERB
fcis-30157	366	14	appropriate	appropriate	ADJ
fcis-30157	366	15	metrics	metric	NOUN
fcis-30157	366	16	is	be	AUX
fcis-30157	366	17	key	key	ADJ
fcis-30157	366	18	.	.	PUNCT
fcis-30157	367	1	common	common	ADJ
fcis-30157	367	2	indicators	indicator	NOUN
fcis-30157	367	3	include	include	VERB
fcis-30157	367	4	accuracy	accuracy	NOUN
fcis-30157	367	5	,	,	PUNCT
fcis-30157	367	6	recall	recall	NOUN
fcis-30157	367	7	,	,	PUNCT
fcis-30157	367	8	and	and	CCONJ
fcis-30157	367	9	f1	f1	NOUN
fcis-30157	367	10	value	value	NOUN
fcis-30157	367	11	,	,	PUNCT
fcis-30157	367	12	each	each	PRON
fcis-30157	367	13	serving	serve	VERB
fcis-30157	367	14	a	a	DET
fcis-30157	367	15	distinct	distinct	ADJ
fcis-30157	367	16	purpose	purpose	NOUN
fcis-30157	367	17	.	.	PUNCT
fcis-30157	368	1	accuracy	accuracy	NOUN
fcis-30157	368	2	measures	measure	VERB
fcis-30157	368	3	the	the	DET
fcis-30157	368	4	proportion	proportion	NOUN
fcis-30157	368	5	of	of	ADP
fcis-30157	368	6	correct	correct	ADJ
fcis-30157	368	7	classifications	classification	NOUN
fcis-30157	368	8	,	,	PUNCT
fcis-30157	368	9	but	but	CCONJ
fcis-30157	368	10	it	it	PRON
fcis-30157	368	11	can	can	AUX
fcis-30157	368	12	be	be	AUX
fcis-30157	368	13	misleading	mislead	VERB
fcis-30157	368	14	in	in	ADP
fcis-30157	368	15	imbalanced	imbalanced	ADJ
fcis-30157	368	16	datasets	dataset	NOUN
fcis-30157	368	17	,	,	PUNCT
fcis-30157	368	18	where	where	SCONJ
fcis-30157	368	19	a	a	DET
fcis-30157	368	20	model	model	NOUN
fcis-30157	368	21	may	may	AUX
fcis-30157	368	22	achieve	achieve	VERB
fcis-30157	368	23	high	high	ADJ
fcis-30157	368	24	accuracy	accuracy	NOUN
fcis-30157	368	25	by	by	ADP
fcis-30157	368	26	predicting	predict	VERB
fcis-30157	368	27	only	only	ADV
fcis-30157	368	28	the	the	DET
fcis-30157	368	29	majority	majority	NOUN
fcis-30157	368	30	class	class	NOUN
fcis-30157	368	31	.	.	PUNCT
fcis-30157	369	1	recall	recall	NOUN
fcis-30157	369	2	,	,	PUNCT
fcis-30157	369	3	or	or	CCONJ
fcis-30157	369	4	sensitivity	sensitivity	NOUN
fcis-30157	369	5	,	,	PUNCT
fcis-30157	369	6	quantifies	quantify	VERB
fcis-30157	369	7	the	the	DET
fcis-30157	369	8	system	system	NOUN
fcis-30157	369	9	's	's	PART
fcis-30157	369	10	ability	ability	NOUN
fcis-30157	369	11	to	to	PART
fcis-30157	369	12	identify	identify	VERB
fcis-30157	369	13	positive	positive	ADJ
fcis-30157	369	14	instances	instance	NOUN
fcis-30157	369	15	,	,	PUNCT
fcis-30157	369	16	critical	critical	ADJ
fcis-30157	369	17	in	in	ADP
fcis-30157	369	18	medical	medical	ADJ
fcis-30157	369	19	contexts	context	NOUN
fcis-30157	369	20	were	be	AUX
fcis-30157	369	21	missing	miss	VERB
fcis-30157	369	22	a	a	DET
fcis-30157	369	23	condition	condition	NOUN
fcis-30157	369	24	can	can	AUX
fcis-30157	369	25	be	be	AUX
fcis-30157	369	26	dangerous	dangerous	ADJ
fcis-30157	369	27	.	.	PUNCT
fcis-30157	370	1	however	however	ADV
fcis-30157	370	2	,	,	PUNCT
fcis-30157	370	3	high	high	ADJ
fcis-30157	370	4	recall	recall	NOUN
fcis-30157	370	5	can	can	AUX
fcis-30157	370	6	lead	lead	VERB
fcis-30157	370	7	to	to	ADP
fcis-30157	370	8	false	false	ADJ
fcis-30157	370	9	positives	positive	NOUN
fcis-30157	370	10	,	,	PUNCT
fcis-30157	370	11	burdening	burden	VERB
fcis-30157	370	12	clinicians	clinician	NOUN
fcis-30157	370	13	with	with	ADP
fcis-30157	370	14	unnecessary	unnecessary	ADJ
fcis-30157	370	15	follow	follow	NOUN
fcis-30157	370	16	-	-	PUNCT
fcis-30157	370	17	ups	up	NOUN
fcis-30157	370	18	.	.	PUNCT
fcis-30157	371	1	the	the	DET
fcis-30157	371	2	f1	f1	PROPN
fcis-30157	371	3	value	value	NOUN
fcis-30157	371	4	balances	balance	NOUN
fcis-30157	371	5	precision	precision	NOUN
fcis-30157	371	6	and	and	CCONJ
fcis-30157	371	7	recall	recall	NOUN
fcis-30157	371	8	,	,	PUNCT
fcis-30157	371	9	useful	useful	ADJ
fcis-30157	371	10	in	in	ADP
fcis-30157	371	11	imbalanced	imbalanced	ADJ
fcis-30157	371	12	datasets	dataset	NOUN
fcis-30157	371	13	,	,	PUNCT
fcis-30157	371	14	as	as	SCONJ
fcis-30157	371	15	it	it	PRON
fcis-30157	371	16	accounts	account	VERB
fcis-30157	371	17	for	for	ADP
fcis-30157	371	18	both	both	DET
fcis-30157	371	19	false	false	ADJ
fcis-30157	371	20	positives	positive	NOUN
fcis-30157	371	21	and	and	CCONJ
fcis-30157	371	22	false	false	ADJ
fcis-30157	371	23	negatives	negative	NOUN
fcis-30157	371	24	.	.	PUNCT
fcis-30157	372	1	other	other	ADJ
fcis-30157	372	2	metrics	metric	NOUN
fcis-30157	372	3	,	,	PUNCT
fcis-30157	372	4	such	such	ADJ
fcis-30157	372	5	as	as	ADP
fcis-30157	372	6	specificity	specificity	NOUN
fcis-30157	372	7	,	,	PUNCT
fcis-30157	372	8	auc	auc	NOUN
fcis-30157	372	9	-	-	PUNCT
fcis-30157	372	10	roc	roc	NOUN
fcis-30157	372	11	,	,	PUNCT
fcis-30157	372	12	and	and	CCONJ
fcis-30157	372	13	auc	auc	NOUN
fcis-30157	372	14	-	-	PUNCT
fcis-30157	372	15	pr	pr	NOUN
fcis-30157	372	16	,	,	PUNCT
fcis-30157	372	17	are	be	AUX
fcis-30157	372	18	also	also	ADV
fcis-30157	372	19	important	important	ADJ
fcis-30157	372	20	.	.	PUNCT
fcis-30157	373	1	specificity	specificity	NOUN
fcis-30157	373	2	measures	measure	VERB
fcis-30157	373	3	the	the	DET
fcis-30157	373	4	ability	ability	NOUN
fcis-30157	373	5	to	to	PART
fcis-30157	373	6	correctly	correctly	ADV
fcis-30157	373	7	identify	identify	VERB
fcis-30157	373	8	negative	negative	ADJ
fcis-30157	373	9	instances	instance	NOUN
fcis-30157	373	10	,	,	PUNCT
fcis-30157	373	11	while	while	SCONJ
fcis-30157	373	12	aucroc	aucroc	NOUN
fcis-30157	373	13	and	and	CCONJ
fcis-30157	373	14	auc	auc	NOUN
fcis-30157	373	15	-	-	PUNCT
fcis-30157	373	16	pr	pr	NOUN
fcis-30157	373	17	evaluate	evaluate	VERB
fcis-30157	373	18	the	the	DET
fcis-30157	373	19	system	system	NOUN
fcis-30157	373	20	across	across	ADP
fcis-30157	373	21	different	different	ADJ
fcis-30157	373	22	thresholds	threshold	NOUN
fcis-30157	373	23	.	.	PUNCT
fcis-30157	374	1	for	for	ADP
fcis-30157	374	2	example	example	NOUN
fcis-30157	374	3	,	,	PUNCT
fcis-30157	374	4	a	a	DET
fcis-30157	374	5	study	study	NOUN
fcis-30157	374	6	on	on	ADP
fcis-30157	374	7	skin	skin	NOUN
fcis-30157	374	8	cancer	cancer	NOUN
fcis-30157	374	9	classification	classification	NOUN
fcis-30157	374	10	using	use	VERB
fcis-30157	374	11	dermoscopy	dermoscopy	NOUN
fcis-30157	374	12	images	image	NOUN
fcis-30157	374	13	reported	report	VERB
fcis-30157	374	14	an	an	DET
fcis-30157	374	15	auc	auc	NOUN
fcis-30157	374	16	-	-	PUNCT
fcis-30157	374	17	roc	roc	NOUN
fcis-30157	374	18	of	of	ADP
fcis-30157	374	19	0.92	0.92	NUM
fcis-30157	374	20	,	,	PUNCT
fcis-30157	374	21	indicating	indicate	VERB
fcis-30157	374	22	strong	strong	ADJ
fcis-30157	374	23	performance	performance	NOUN
fcis-30157	374	24	in	in	ADP
fcis-30157	374	25	distinguishing	distinguish	VERB
fcis-30157	374	26	malignant	malignant	NOUN
fcis-30157	374	27	from	from	ADP
fcis-30157	374	28	benign	benign	ADJ
fcis-30157	374	29	lesions	lesion	NOUN
fcis-30157	374	30	.	.	PUNCT
fcis-30157	375	1	the	the	DET
fcis-30157	375	2	choice	choice	NOUN
fcis-30157	375	3	of	of	ADP
fcis-30157	375	4	metrics	metric	NOUN
fcis-30157	375	5	depends	depend	VERB
fcis-30157	375	6	on	on	ADP
fcis-30157	375	7	the	the	DET
fcis-30157	375	8	system	system	NOUN
fcis-30157	375	9	's	's	PART
fcis-30157	375	10	purpose	purpose	NOUN
fcis-30157	375	11	.	.	PUNCT
fcis-30157	376	1	for	for	ADP
fcis-30157	376	2	melanoma	melanoma	PROPN
fcis-30157	376	3	detection	detection	NOUN
fcis-30157	376	4	,	,	PUNCT
fcis-30157	376	5	recall	recall	NOUN
fcis-30157	376	6	is	be	AUX
fcis-30157	376	7	prioritized	prioritize	VERB
fcis-30157	376	8	,	,	PUNCT
fcis-30157	376	9	while	while	SCONJ
fcis-30157	376	10	in	in	ADP
fcis-30157	376	11	diagnosing	diagnose	VERB
fcis-30157	376	12	common	common	ADJ
fcis-30157	376	13	skin	skin	NOUN
fcis-30157	376	14	conditions	condition	NOUN
fcis-30157	376	15	,	,	PUNCT
fcis-30157	376	16	precision	precision	NOUN
fcis-30157	376	17	may	may	AUX
fcis-30157	376	18	be	be	AUX
fcis-30157	376	19	more	more	ADV
fcis-30157	376	20	important	important	ADJ
fcis-30157	376	21	to	to	PART
fcis-30157	376	22	avoid	avoid	VERB
fcis-30157	376	23	unnecessary	unnecessary	ADJ
fcis-30157	376	24	follow	follow	NOUN
fcis-30157	376	25	-	-	PUNCT
fcis-30157	376	26	ups	up	NOUN
fcis-30157	376	27	.	.	PUNCT
fcis-30157	377	1	recent	recent	ADJ
fcis-30157	377	2	studies	study	NOUN
fcis-30157	377	3	,	,	PUNCT
fcis-30157	377	4	such	such	ADJ
fcis-30157	377	5	as	as	ADP
fcis-30157	377	6	those	those	PRON
fcis-30157	377	7	on	on	ADP
fcis-30157	377	8	multimodal	multimodal	ADJ
fcis-30157	377	9	fusion	fusion	NOUN
fcis-30157	377	10	methods	method	NOUN
fcis-30157	377	11	,	,	PUNCT
fcis-30157	377	12	have	have	AUX
fcis-30157	377	13	shown	show	VERB
fcis-30157	377	14	that	that	SCONJ
fcis-30157	377	15	combining	combine	VERB
fcis-30157	377	16	skin	skin	NOUN
fcis-30157	377	17	images	image	NOUN
fcis-30157	377	18	with	with	ADP
fcis-30157	377	19	clinical	clinical	ADJ
fcis-30157	377	20	data	datum	NOUN
fcis-30157	377	21	can	can	AUX
fcis-30157	377	22	improve	improve	VERB
fcis-30157	377	23	both	both	DET
fcis-30157	377	24	accuracy	accuracy	NOUN
fcis-30157	377	25	and	and	CCONJ
fcis-30157	377	26	recall	recall	NOUN
fcis-30157	377	27	.	.	PUNCT
fcis-30157	378	1	however	however	ADV
fcis-30157	378	2	,	,	PUNCT
fcis-30157	378	3	these	these	DET
fcis-30157	378	4	metrics	metric	NOUN
fcis-30157	378	5	have	have	VERB
fcis-30157	378	6	limitations	limitation	NOUN
fcis-30157	378	7	.	.	PUNCT
fcis-30157	379	1	accuracy	accuracy	NOUN
fcis-30157	379	2	does	do	AUX
fcis-30157	379	3	not	not	PART
fcis-30157	379	4	consider	consider	VERB
fcis-30157	379	5	class	class	NOUN
fcis-30157	379	6	distribution	distribution	NOUN
fcis-30157	379	7	,	,	PUNCT
fcis-30157	379	8	and	and	CCONJ
fcis-30157	379	9	recall	recall	NOUN
fcis-30157	379	10	and	and	CCONJ
fcis-30157	379	11	precision	precision	NOUN
fcis-30157	379	12	may	may	AUX
fcis-30157	379	13	conflict	conflict	VERB
fcis-30157	379	14	when	when	SCONJ
fcis-30157	379	15	false	false	ADJ
fcis-30157	379	16	positives	positive	NOUN
fcis-30157	379	17	and	and	CCONJ
fcis-30157	379	18	negatives	negative	NOUN
fcis-30157	379	19	have	have	VERB
fcis-30157	379	20	different	different	ADJ
fcis-30157	379	21	consequences	consequence	NOUN
fcis-30157	379	22	.	.	PUNCT
fcis-30157	380	1	additionally	additionally	ADV
fcis-30157	380	2	,	,	PUNCT
fcis-30157	380	3	these	these	DET
fcis-30157	380	4	metrics	metric	NOUN
fcis-30157	380	5	may	may	AUX
fcis-30157	380	6	not	not	PART
fcis-30157	380	7	assess	assess	VERB
fcis-30157	380	8	a	a	DET
fcis-30157	380	9	system	system	NOUN
fcis-30157	380	10	's	's	PART
fcis-30157	380	11	interpretability	interpretability	NOUN
fcis-30157	380	12	or	or	CCONJ
fcis-30157	380	13	actionable	actionable	ADJ
fcis-30157	380	14	insights	insight	NOUN
fcis-30157	380	15	.	.	PUNCT
fcis-30157	381	1	to	to	PART
fcis-30157	381	2	address	address	VERB
fcis-30157	381	3	this	this	PRON
fcis-30157	381	4	,	,	PUNCT
fcis-30157	381	5	some	some	DET
fcis-30157	381	6	studies	study	NOUN
fcis-30157	381	7	have	have	AUX
fcis-30157	381	8	combined	combine	VERB
fcis-30157	381	9	quantitative	quantitative	ADJ
fcis-30157	381	10	metrics	metric	NOUN
fcis-30157	381	11	with	with	ADP
fcis-30157	381	12	qualitative	qualitative	ADJ
fcis-30157	381	13	evaluations	evaluation	NOUN
fcis-30157	381	14	,	,	PUNCT
fcis-30157	381	15	like	like	ADP
fcis-30157	381	16	user	user	NOUN
fcis-30157	381	17	studies	study	NOUN
fcis-30157	381	18	on	on	ADP
fcis-30157	381	19	model	model	NOUN
fcis-30157	381	20	interpretability	interpretability	NOUN
fcis-30157	381	21	or	or	CCONJ
fcis-30157	381	22	combining	combine	VERB
fcis-30157	381	23	performance	performance	NOUN
fcis-30157	381	24	scores	score	NOUN
fcis-30157	381	25	with	with	ADP
fcis-30157	381	26	user	user	NOUN
fcis-30157	381	27	satisfaction	satisfaction	NOUN
fcis-30157	381	28	.	.	PUNCT
fcis-30157	382	1	3.1.3	3.1.3	X
fcis-30157	382	2	.	.	PUNCT
fcis-30157	382	3	experimental	experimental	ADJ
fcis-30157	382	4	environment	environment	NOUN
fcis-30157	382	5	this	this	DET
fcis-30157	382	6	study	study	NOUN
fcis-30157	382	7	's	's	PART
fcis-30157	382	8	experimental	experimental	ADJ
fcis-30157	382	9	environment	environment	NOUN
fcis-30157	382	10	was	be	AUX
fcis-30157	382	11	carefully	carefully	ADV
fcis-30157	382	12	designed	design	VERB
fcis-30157	382	13	for	for	ADP
fcis-30157	382	14	reproducibility	reproducibility	NOUN
fcis-30157	382	15	and	and	CCONJ
fcis-30157	382	16	consistency	consistency	NOUN
fcis-30157	382	17	.	.	PUNCT
fcis-30157	383	1	the	the	DET
fcis-30157	383	2	hardware	hardware	NOUN
fcis-30157	383	3	setup	setup	NOUN
fcis-30157	383	4	featured	feature	VERB
fcis-30157	383	5	a	a	DET
fcis-30157	383	6	high	high	ADJ
fcis-30157	383	7	-	-	PUNCT
fcis-30157	383	8	performance	performance	NOUN
fcis-30157	383	9	computing	computing	NOUN
fcis-30157	383	10	system	system	NOUN
fcis-30157	383	11	with	with	ADP
fcis-30157	383	12	an	an	DET
fcis-30157	383	13	nvidia	nvidia	PROPN
fcis-30157	383	14	geforce	geforce	NOUN
fcis-30157	383	15	rtx	rtx	PROPN
fcis-30157	383	16	3090	3090	NUM
fcis-30157	383	17	gpu	gpu	PROPN
fcis-30157	383	18	,	,	PUNCT
fcis-30157	383	19	64	64	NUM
fcis-30157	383	20	gb	gb	NOUN
fcis-30157	383	21	ram	ram	NOUN
fcis-30157	383	22	,	,	PUNCT
fcis-30157	383	23	and	and	CCONJ
fcis-30157	383	24	an	an	DET
fcis-30157	383	25	amd	amd	ADJ
fcis-30157	383	26	ryzen	ryzen	ADJ
fcis-30157	383	27	threadripper	threadripper	NOUN
fcis-30157	383	28	3970x	3970x	NUM
fcis-30157	383	29	processor	processor	NOUN
fcis-30157	383	30	,	,	PUNCT
fcis-30157	383	31	selected	select	VERB
fcis-30157	383	32	to	to	PART
fcis-30157	383	33	handle	handle	VERB
fcis-30157	383	34	the	the	DET
fcis-30157	383	35	demands	demand	NOUN
fcis-30157	383	36	of	of	ADP
fcis-30157	383	37	deep	deep	ADJ
fcis-30157	383	38	learning	learning	NOUN
fcis-30157	383	39	on	on	ADP
fcis-30157	383	40	large	large	ADJ
fcis-30157	383	41	skin	skin	NOUN
fcis-30157	383	42	disease	disease	NOUN
fcis-30157	383	43	image	image	NOUN
fcis-30157	383	44	datasets	dataset	NOUN
fcis-30157	383	45	.	.	PUNCT
fcis-30157	384	1	the	the	DET
fcis-30157	384	2	gpu	gpu	PROPN
fcis-30157	384	3	,	,	PUNCT
fcis-30157	384	4	with	with	ADP
fcis-30157	384	5	24	24	NUM
fcis-30157	384	6	gb	gb	NOUN
fcis-30157	384	7	vram	vram	NOUN
fcis-30157	384	8	,	,	PUNCT
fcis-30157	384	9	accelerated	accelerate	VERB
fcis-30157	384	10	the	the	DET
fcis-30157	384	11	training	training	NOUN
fcis-30157	384	12	of	of	ADP
fcis-30157	384	13	convolutional	convolutional	ADJ
fcis-30157	384	14	neural	neural	ADJ
fcis-30157	384	15	networks	network	NOUN
fcis-30157	384	16	(	(	PUNCT
fcis-30157	384	17	cnns	cnns	PROPN
fcis-30157	384	18	)	)	PUNCT
fcis-30157	384	19	,	,	PUNCT
fcis-30157	384	20	while	while	SCONJ
fcis-30157	384	21	the	the	DET
fcis-30157	384	22	high	high	ADJ
fcis-30157	384	23	-	-	PUNCT
fcis-30157	384	24	speed	speed	NOUN
fcis-30157	384	25	cpu	cpu	NOUN
fcis-30157	384	26	and	and	CCONJ
fcis-30157	384	27	ample	ample	ADJ
fcis-30157	384	28	ram	ram	NOUN
fcis-30157	384	29	ensured	ensure	VERB
fcis-30157	384	30	efficient	efficient	ADJ
fcis-30157	384	31	data	datum	NOUN
fcis-30157	384	32	preprocessing	preprocessing	NOUN
fcis-30157	384	33	and	and	CCONJ
fcis-30157	384	34	model	model	NOUN
fcis-30157	384	35	training	training	NOUN
fcis-30157	384	36	.	.	PUNCT
fcis-30157	385	1	the	the	DET
fcis-30157	385	2	software	software	NOUN
fcis-30157	385	3	environment	environment	NOUN
fcis-30157	385	4	used	use	VERB
fcis-30157	385	5	ubuntu	ubuntu	NOUN
fcis-30157	385	6	20.04	20.04	NUM
fcis-30157	385	7	lts	lts	NOUN
fcis-30157	385	8	,	,	PUNCT
fcis-30157	385	9	tensorflow	tensorflow	NOUN
fcis-30157	385	10	2.8.0	2.8.0	NUM
fcis-30157	385	11	,	,	PUNCT
fcis-30157	385	12	pytorch	pytorch	NOUN
fcis-30157	385	13	1.10.0	1.10.0	NUM
fcis-30157	385	14	,	,	PUNCT
fcis-30157	385	15	and	and	CCONJ
fcis-30157	385	16	python	python	NOUN
fcis-30157	385	17	3.8	3.8	NUM
fcis-30157	385	18	.	.	PUNCT
fcis-30157	386	1	libraries	library	NOUN
fcis-30157	386	2	like	like	ADP
fcis-30157	386	3	numpy	numpy	NOUN
fcis-30157	386	4	,	,	PUNCT
fcis-30157	386	5	opencv	opencv	NOUN
fcis-30157	386	6	,	,	PUNCT
fcis-30157	386	7	and	and	CCONJ
fcis-30157	386	8	scikitlearn	scikitlearn	ADJ
fcis-30157	386	9	supported	support	VERB
fcis-30157	386	10	data	datum	NOUN
fcis-30157	386	11	manipulation	manipulation	NOUN
fcis-30157	386	12	,	,	PUNCT
fcis-30157	386	13	image	image	NOUN
fcis-30157	386	14	processing	processing	NOUN
fcis-30157	386	15	,	,	PUNCT
fcis-30157	386	16	and	and	CCONJ
fcis-30157	386	17	evaluation	evaluation	NOUN
fcis-30157	386	18	.	.	PUNCT
fcis-30157	387	1	version	version	NOUN
fcis-30157	387	2	control	control	NOUN
fcis-30157	387	3	was	be	AUX
fcis-30157	387	4	maintained	maintain	VERB
fcis-30157	387	5	with	with	ADP
fcis-30157	387	6	git	git	NOUN
fcis-30157	387	7	,	,	PUNCT
fcis-30157	387	8	and	and	CCONJ
fcis-30157	387	9	docker	docker	PROPN
fcis-30157	387	10	ensured	ensure	VERB
fcis-30157	387	11	consistency	consistency	NOUN
fcis-30157	387	12	across	across	ADP
fcis-30157	387	13	different	different	ADJ
fcis-30157	387	14	machines	machine	NOUN
fcis-30157	387	15	.	.	PUNCT
fcis-30157	388	1	datasets	dataset	NOUN
fcis-30157	388	2	underwent	undergo	VERB
fcis-30157	388	3	standardized	standardized	ADJ
fcis-30157	388	4	preprocessing	preprocessing	NOUN
fcis-30157	388	5	,	,	PUNCT
fcis-30157	388	6	including	include	VERB
fcis-30157	388	7	resizing	resize	VERB
fcis-30157	388	8	to	to	ADP
fcis-30157	388	9	224x224	224x224	NUM
fcis-30157	388	10	pixels	pixel	NOUN
fcis-30157	388	11	,	,	PUNCT
fcis-30157	388	12	pixel	pixel	PROPN
fcis-30157	388	13	normalization	normalization	NOUN
fcis-30157	388	14	,	,	PUNCT
fcis-30157	388	15	and	and	CCONJ
fcis-30157	388	16	data	datum	NOUN
fcis-30157	388	17	augmentation	augmentation	NOUN
fcis-30157	388	18	techniques	technique	NOUN
fcis-30157	388	19	,	,	PUNCT
fcis-30157	388	20	validated	validate	VERB
fcis-30157	388	21	to	to	PART
fcis-30157	388	22	ensure	ensure	VERB
fcis-30157	388	23	identical	identical	ADJ
fcis-30157	388	24	outputs	output	NOUN
fcis-30157	388	25	across	across	ADP
fcis-30157	388	26	different	different	ADJ
fcis-30157	388	27	systems	system	NOUN
fcis-30157	388	28	.	.	PUNCT
fcis-30157	389	1	environmental	environmental	ADJ
fcis-30157	389	2	factors	factor	NOUN
fcis-30157	389	3	like	like	ADP
fcis-30157	389	4	power	power	NOUN
fcis-30157	389	5	stability	stability	NOUN
fcis-30157	389	6	,	,	PUNCT
fcis-30157	389	7	temperature	temperature	NOUN
fcis-30157	389	8	,	,	PUNCT
fcis-30157	389	9	and	and	CCONJ
fcis-30157	389	10	humidity	humidity	NOUN
fcis-30157	389	11	were	be	AUX
fcis-30157	389	12	controlled	control	VERB
fcis-30157	389	13	to	to	PART
fcis-30157	389	14	avoid	avoid	VERB
fcis-30157	389	15	hardware	hardware	NOUN
fcis-30157	389	16	performance	performance	NOUN
fcis-30157	389	17	degradation	degradation	NOUN
fcis-30157	389	18	.	.	PUNCT
fcis-30157	390	1	random	random	ADJ
fcis-30157	390	2	seeds	seed	NOUN
fcis-30157	390	3	were	be	AUX
fcis-30157	390	4	set	set	VERB
fcis-30157	390	5	for	for	ADP
fcis-30157	390	6	all	all	DET
fcis-30157	390	7	stochastic	stochastic	ADJ
fcis-30157	390	8	operations	operation	NOUN
fcis-30157	390	9	to	to	PART
fcis-30157	390	10	ensure	ensure	VERB
fcis-30157	390	11	reproducible	reproducible	ADJ
fcis-30157	390	12	results	result	NOUN
fcis-30157	390	13	across	across	ADP
fcis-30157	390	14	multiple	multiple	ADJ
fcis-30157	390	15	runs	run	NOUN
fcis-30157	390	16	.	.	PUNCT
fcis-30157	391	1	the	the	DET
fcis-30157	391	2	use	use	NOUN
fcis-30157	391	3	of	of	ADP
fcis-30157	391	4	version	version	NOUN
fcis-30157	391	5	control	control	NOUN
fcis-30157	391	6	for	for	ADP
fcis-30157	391	7	code	code	NOUN
fcis-30157	391	8	and	and	CCONJ
fcis-30157	391	9	datasets	dataset	NOUN
fcis-30157	391	10	also	also	ADV
fcis-30157	391	11	enabled	enable	VERB
fcis-30157	391	12	experiment	experiment	NOUN
fcis-30157	391	13	replication	replication	NOUN
fcis-30157	391	14	and	and	CCONJ
fcis-30157	391	15	comparison	comparison	NOUN
fcis-30157	391	16	,	,	PUNCT
fcis-30157	391	17	with	with	ADP
fcis-30157	391	18	unique	unique	ADJ
fcis-30157	391	19	identifiers	identifier	NOUN
fcis-30157	391	20	assigned	assign	VERB
fcis-30157	391	21	to	to	ADP
fcis-30157	391	22	each	each	DET
fcis-30157	391	23	experiment	experiment	NOUN
fcis-30157	391	24	.	.	PUNCT
fcis-30157	392	1	ethical	ethical	ADJ
fcis-30157	392	2	considerations	consideration	NOUN
fcis-30157	392	3	were	be	AUX
fcis-30157	392	4	addressed	address	VERB
fcis-30157	392	5	by	by	ADP
fcis-30157	392	6	anonymizing	anonymize	VERB
fcis-30157	392	7	patient	patient	NOUN
fcis-30157	392	8	images	image	NOUN
fcis-30157	392	9	and	and	CCONJ
fcis-30157	392	10	adhering	adhere	VERB
fcis-30157	392	11	to	to	ADP
fcis-30157	392	12	institutional	institutional	ADJ
fcis-30157	392	13	review	review	NOUN
fcis-30157	392	14	board	board	NOUN
fcis-30157	392	15	(	(	PUNCT
fcis-30157	392	16	irb	irb	PROPN
fcis-30157	392	17	)	)	PUNCT
fcis-30157	392	18	guidelines	guideline	NOUN
fcis-30157	392	19	.	.	PUNCT
fcis-30157	393	1	the	the	DET
fcis-30157	393	2	datasets	dataset	NOUN
fcis-30157	393	3	were	be	AUX
fcis-30157	393	4	sourced	source	VERB
fcis-30157	393	5	from	from	ADP
fcis-30157	393	6	publicly	publicly	ADV
fcis-30157	393	7	available	available	ADJ
fcis-30157	393	8	repositories	repository	NOUN
fcis-30157	393	9	like	like	ADP
fcis-30157	393	10	the	the	DET
fcis-30157	393	11	isic	isic	PROPN
fcis-30157	393	12	archive	archive	NOUN
fcis-30157	393	13	,	,	PUNCT
fcis-30157	393	14	ensuring	ensure	VERB
fcis-30157	393	15	compliance	compliance	NOUN
fcis-30157	393	16	with	with	ADP
fcis-30157	393	17	ethical	ethical	ADJ
fcis-30157	393	18	standards.to	standards.to	PROPN
fcis-30157	393	19	support	support	NOUN
fcis-30157	393	20	scalability	scalability	NOUN
fcis-30157	393	21	,	,	PUNCT
fcis-30157	393	22	cloud	cloud	ADJ
fcis-30157	393	23	resources	resource	NOUN
fcis-30157	393	24	like	like	ADP
fcis-30157	393	25	aws	aw	NOUN
fcis-30157	393	26	ec2	ec2	PROPN
fcis-30157	393	27	instances	instance	NOUN
fcis-30157	393	28	were	be	AUX
fcis-30157	393	29	used	use	VERB
fcis-30157	393	30	for	for	ADP
fcis-30157	393	31	parallelized	parallelized	ADJ
fcis-30157	393	32	experiments	experiment	NOUN
fcis-30157	393	33	,	,	PUNCT
fcis-30157	393	34	reducing	reduce	VERB
fcis-30157	393	35	training	training	NOUN
fcis-30157	393	36	time	time	NOUN
fcis-30157	393	37	.	.	PUNCT
fcis-30157	394	1	for	for	ADP
fcis-30157	394	2	example	example	NOUN
fcis-30157	394	3	,	,	PUNCT
fcis-30157	394	4	training	train	VERB
fcis-30157	394	5	a	a	DET
fcis-30157	394	6	model	model	NOUN
fcis-30157	394	7	to	to	PART
fcis-30157	394	8	classify	classify	VERB
fcis-30157	394	9	melanoma	melanoma	NOUN
fcis-30157	394	10	from	from	ADP
fcis-30157	394	11	10,000	10,000	NUM
fcis-30157	394	12	dermoscopy	dermoscopy	NOUN
fcis-30157	394	13	images	image	NOUN
fcis-30157	394	14	took	take	VERB
fcis-30157	394	15	12	12	NUM
fcis-30157	394	16	hours	hour	NOUN
fcis-30157	394	17	,	,	PUNCT
fcis-30157	394	18	achieving	achieve	VERB
fcis-30157	394	19	92.5	92.5	NUM
fcis-30157	394	20	%	%	NOUN
fcis-30157	394	21	accuracy	accuracy	NOUN
fcis-30157	394	22	,	,	PUNCT
fcis-30157	394	23	which	which	PRON
fcis-30157	394	24	was	be	AUX
fcis-30157	394	25	consistent	consistent	ADJ
fcis-30157	394	26	with	with	ADP
fcis-30157	394	27	results	result	NOUN
fcis-30157	394	28	from	from	ADP
fcis-30157	394	29	similar	similar	ADJ
fcis-30157	394	30	studies	study	NOUN
fcis-30157	394	31	.	.	PUNCT
fcis-30157	395	1	tools	tool	NOUN
fcis-30157	395	2	like	like	ADP
fcis-30157	395	3	tensorboard	tensorboard	NOUN
fcis-30157	395	4	monitored	monitor	VERB
fcis-30157	395	5	training	training	NOUN
fcis-30157	395	6	progress	progress	NOUN
fcis-30157	395	7	,	,	PUNCT
fcis-30157	395	8	and	and	CCONJ
fcis-30157	395	9	shap	shap	PROPN
fcis-30157	395	10	values	value	NOUN
fcis-30157	395	11	were	be	AUX
fcis-30157	395	12	used	use	VERB
fcis-30157	395	13	for	for	ADP
fcis-30157	395	14	model	model	NOUN
fcis-30157	395	15	interpretability	interpretability	NOUN
fcis-30157	395	16	,	,	PUNCT
fcis-30157	395	17	helping	help	VERB
fcis-30157	395	18	clinicians	clinician	NOUN
fcis-30157	395	19	understand	understand	VERB
fcis-30157	395	20	classification	classification	NOUN
fcis-30157	395	21	decisions	decision	NOUN
fcis-30157	395	22	.	.	PUNCT
fcis-30157	396	1	overall	overall	ADV
fcis-30157	396	2	,	,	PUNCT
fcis-30157	396	3	the	the	DET
fcis-30157	396	4	experimental	experimental	ADJ
fcis-30157	396	5	environment	environment	NOUN
fcis-30157	396	6	was	be	AUX
fcis-30157	396	7	optimized	optimize	VERB
fcis-30157	396	8	for	for	ADP
fcis-30157	396	9	performance	performance	NOUN
fcis-30157	396	10	,	,	PUNCT
fcis-30157	396	11	reproducibility	reproducibility	NOUN
fcis-30157	396	12	,	,	PUNCT
fcis-30157	396	13	and	and	CCONJ
fcis-30157	396	14	ethical	ethical	ADJ
fcis-30157	396	15	compliance	compliance	NOUN
fcis-30157	396	16	,	,	PUNCT
fcis-30157	396	17	contributing	contribute	VERB
fcis-30157	396	18	to	to	ADP
fcis-30157	396	19	the	the	DET
fcis-30157	396	20	successful	successful	ADJ
fcis-30157	396	21	development	development	NOUN
fcis-30157	396	22	of	of	ADP
fcis-30157	396	23	skin	skin	NOUN
fcis-30157	396	24	disease	disease	NOUN
fcis-30157	396	25	diagnosis	diagnosis	NOUN
fcis-30157	396	26	models	model	NOUN
fcis-30157	396	27	.	.	PUNCT
fcis-30157	397	1	120	120	NUM
fcis-30157	397	2	3.2	3.2	NUM
fcis-30157	397	3	.	.	PUNCT
fcis-30157	398	1	automatic	automatic	ADJ
fcis-30157	398	2	recognition	recognition	NOUN
fcis-30157	398	3	results	result	VERB
fcis-30157	398	4	3.2.1	3.2.1	NUM
fcis-30157	398	5	.	.	PUNCT
fcis-30157	399	1	accuracy	accuracy	NOUN
fcis-30157	399	2	of	of	ADP
fcis-30157	399	3	different	different	ADJ
fcis-30157	399	4	recognition	recognition	NOUN
fcis-30157	399	5	models	model	NOUN
fcis-30157	399	6	in	in	ADP
fcis-30157	399	7	this	this	DET
fcis-30157	399	8	section	section	NOUN
fcis-30157	399	9	,	,	PUNCT
fcis-30157	399	10	we	we	PRON
fcis-30157	399	11	present	present	VERB
fcis-30157	399	12	the	the	DET
fcis-30157	399	13	accuracy	accuracy	NOUN
fcis-30157	399	14	results	result	NOUN
fcis-30157	399	15	of	of	ADP
fcis-30157	399	16	various	various	ADJ
fcis-30157	399	17	recognition	recognition	NOUN
fcis-30157	399	18	models	model	NOUN
fcis-30157	399	19	applied	apply	VERB
fcis-30157	399	20	to	to	ADP
fcis-30157	399	21	our	our	PRON
fcis-30157	399	22	dataset	dataset	NOUN
fcis-30157	399	23	,	,	PUNCT
fcis-30157	399	24	focusing	focus	VERB
fcis-30157	399	25	on	on	ADP
fcis-30157	399	26	cnn	cnn	PROPN
fcis-30157	399	27	architectures	architecture	NOUN
fcis-30157	399	28	like	like	ADP
fcis-30157	399	29	resnet	resnet	NOUN
fcis-30157	399	30	,	,	PUNCT
fcis-30157	399	31	vgg	vgg	NOUN
fcis-30157	399	32	,	,	PUNCT
fcis-30157	399	33	and	and	CCONJ
fcis-30157	399	34	inception	inception	NOUN
fcis-30157	399	35	for	for	ADP
fcis-30157	399	36	skin	skin	NOUN
fcis-30157	399	37	disease	disease	NOUN
fcis-30157	399	38	classification	classification	NOUN
fcis-30157	399	39	.	.	PUNCT
fcis-30157	400	1	the	the	DET
fcis-30157	400	2	results	result	NOUN
fcis-30157	400	3	,	,	PUNCT
fcis-30157	400	4	shown	show	VERB
fcis-30157	400	5	in	in	ADP
fcis-30157	400	6	table	table	NOUN
fcis-30157	400	7	1	1	NUM
fcis-30157	400	8	,	,	PUNCT
fcis-30157	400	9	indicate	indicate	VERB
fcis-30157	400	10	significant	significant	ADJ
fcis-30157	400	11	differences	difference	NOUN
fcis-30157	400	12	in	in	ADP
fcis-30157	400	13	performance	performance	NOUN
fcis-30157	400	14	:	:	PUNCT
fcis-30157	400	15	resnet	resnet	NOUN
fcis-30157	400	16	achieved	achieve	VERB
fcis-30157	400	17	the	the	DET
fcis-30157	400	18	highest	high	ADJ
fcis-30157	400	19	accuracy	accuracy	NOUN
fcis-30157	400	20	of	of	ADP
fcis-30157	400	21	89.5	89.5	NUM
fcis-30157	400	22	%	%	NOUN
fcis-30157	400	23	,	,	PUNCT
fcis-30157	400	24	followed	follow	VERB
fcis-30157	400	25	by	by	ADP
fcis-30157	400	26	inception	inception	NOUN
fcis-30157	400	27	at	at	ADP
fcis-30157	400	28	87.8	87.8	NUM
fcis-30157	400	29	%	%	NOUN
fcis-30157	400	30	,	,	PUNCT
fcis-30157	400	31	and	and	CCONJ
fcis-30157	400	32	vgg	vgg	VERB
fcis-30157	400	33	at	at	ADP
fcis-30157	400	34	85.2	85.2	NUM
fcis-30157	400	35	%	%	NOUN
fcis-30157	400	36	.	.	PUNCT
fcis-30157	401	1	resnet	resnet	PROPN
fcis-30157	401	2	's	's	PART
fcis-30157	401	3	superior	superior	ADJ
fcis-30157	401	4	accuracy	accuracy	NOUN
fcis-30157	401	5	is	be	AUX
fcis-30157	401	6	attributed	attribute	VERB
fcis-30157	401	7	to	to	ADP
fcis-30157	401	8	its	its	PRON
fcis-30157	401	9	residual	residual	ADJ
fcis-30157	401	10	learning	learning	NOUN
fcis-30157	401	11	framework	framework	NOUN
fcis-30157	401	12	,	,	PUNCT
fcis-30157	401	13	which	which	PRON
fcis-30157	401	14	enables	enable	VERB
fcis-30157	401	15	deeper	deep	ADJ
fcis-30157	401	16	networks	network	NOUN
fcis-30157	401	17	and	and	CCONJ
fcis-30157	401	18	captures	capture	VERB
fcis-30157	401	19	more	more	ADJ
fcis-30157	401	20	complex	complex	ADJ
fcis-30157	401	21	features	feature	NOUN
fcis-30157	401	22	in	in	ADP
fcis-30157	401	23	skin	skin	NOUN
fcis-30157	401	24	disease	disease	NOUN
fcis-30157	401	25	images	image	NOUN
fcis-30157	401	26	.	.	PUNCT
fcis-30157	402	1	in	in	ADP
fcis-30157	402	2	contrast	contrast	NOUN
fcis-30157	402	3	,	,	PUNCT
fcis-30157	402	4	vgg	vgg	PROPN
fcis-30157	402	5	,	,	PUNCT
fcis-30157	402	6	with	with	ADP
fcis-30157	402	7	its	its	PRON
fcis-30157	402	8	simpler	simple	ADJ
fcis-30157	402	9	architecture	architecture	NOUN
fcis-30157	402	10	,	,	PUNCT
fcis-30157	402	11	struggles	struggle	VERB
fcis-30157	402	12	with	with	ADP
fcis-30157	402	13	high	high	ADJ
fcis-30157	402	14	-	-	PUNCT
fcis-30157	402	15	resolution	resolution	NOUN
fcis-30157	402	16	images	image	NOUN
fcis-30157	402	17	,	,	PUNCT
fcis-30157	402	18	leading	lead	VERB
fcis-30157	402	19	to	to	ADP
fcis-30157	402	20	lower	low	ADJ
fcis-30157	402	21	accuracy	accuracy	NOUN
fcis-30157	402	22	.	.	PUNCT
fcis-30157	403	1	inception	inception	NOUN
fcis-30157	403	2	performs	perform	VERB
fcis-30157	403	3	better	well	ADV
fcis-30157	403	4	than	than	ADP
fcis-30157	403	5	vgg	vgg	ADJ
fcis-30157	403	6	due	due	ADP
fcis-30157	403	7	to	to	ADP
fcis-30157	403	8	its	its	PRON
fcis-30157	403	9	multi	multi	ADJ
fcis-30157	403	10	-	-	ADJ
fcis-30157	403	11	scale	scale	ADJ
fcis-30157	403	12	feature	feature	NOUN
fcis-30157	403	13	extraction	extraction	NOUN
fcis-30157	403	14	but	but	CCONJ
fcis-30157	403	15	still	still	ADV
fcis-30157	403	16	lags	lag	VERB
fcis-30157	403	17	behind	behind	ADP
fcis-30157	403	18	resnet	resnet	NOUN
fcis-30157	403	19	.	.	PUNCT
fcis-30157	404	1	we	we	PRON
fcis-30157	404	2	further	far	ADV
fcis-30157	404	3	improved	improve	VERB
fcis-30157	404	4	model	model	NOUN
fcis-30157	404	5	performance	performance	NOUN
fcis-30157	404	6	with	with	ADP
fcis-30157	404	7	data	datum	NOUN
fcis-30157	404	8	augmentation	augmentation	NOUN
fcis-30157	404	9	(	(	PUNCT
fcis-30157	404	10	random	random	ADJ
fcis-30157	404	11	rotations	rotation	NOUN
fcis-30157	404	12	,	,	PUNCT
fcis-30157	404	13	flips	flip	NOUN
fcis-30157	404	14	,	,	PUNCT
fcis-30157	404	15	and	and	CCONJ
fcis-30157	404	16	color	color	NOUN
fcis-30157	404	17	jittering	jittere	VERB
fcis-30157	404	18	)	)	PUNCT
fcis-30157	404	19	,	,	PUNCT
fcis-30157	404	20	boosting	boost	VERB
fcis-30157	404	21	resnet	resnet	NOUN
fcis-30157	404	22	's	's	PART
fcis-30157	404	23	accuracy	accuracy	NOUN
fcis-30157	404	24	to	to	ADP
fcis-30157	404	25	91.2	91.2	NUM
fcis-30157	404	26	%	%	NOUN
fcis-30157	404	27	,	,	PUNCT
fcis-30157	404	28	inception	inception	NOUN
fcis-30157	404	29	's	's	PART
fcis-30157	404	30	to	to	ADP
fcis-30157	404	31	89.3	89.3	NUM
fcis-30157	404	32	%	%	NOUN
fcis-30157	404	33	,	,	PUNCT
fcis-30157	404	34	and	and	CCONJ
fcis-30157	404	35	vgg	vgg	NOUN
fcis-30157	404	36	's	's	PART
fcis-30157	404	37	to	to	ADP
fcis-30157	404	38	87.5	87.5	NUM
fcis-30157	404	39	%	%	NOUN
fcis-30157	404	40	.	.	PUNCT
fcis-30157	405	1	this	this	PRON
fcis-30157	405	2	highlights	highlight	VERB
fcis-30157	405	3	the	the	DET
fcis-30157	405	4	importance	importance	NOUN
fcis-30157	405	5	of	of	ADP
fcis-30157	405	6	data	datum	NOUN
fcis-30157	405	7	diversity	diversity	NOUN
fcis-30157	405	8	in	in	ADP
fcis-30157	405	9	improving	improve	VERB
fcis-30157	405	10	model	model	NOUN
fcis-30157	405	11	performance	performance	NOUN
fcis-30157	405	12	.	.	PUNCT
fcis-30157	406	1	incorporating	incorporate	VERB
fcis-30157	406	2	auxiliary	auxiliary	ADJ
fcis-30157	406	3	tasks	task	NOUN
fcis-30157	406	4	like	like	ADP
fcis-30157	406	5	segmentation	segmentation	NOUN
fcis-30157	406	6	and	and	CCONJ
fcis-30157	406	7	boundary	boundary	ADJ
fcis-30157	406	8	detection	detection	NOUN
fcis-30157	406	9	showed	show	VERB
fcis-30157	406	10	mixed	mixed	ADJ
fcis-30157	406	11	results	result	NOUN
fcis-30157	406	12	.	.	PUNCT
fcis-30157	407	1	while	while	SCONJ
fcis-30157	407	2	resnet	resnet	NOUN
fcis-30157	407	3	showed	show	VERB
fcis-30157	407	4	a	a	DET
fcis-30157	407	5	slight	slight	ADJ
fcis-30157	407	6	accuracy	accuracy	NOUN
fcis-30157	407	7	increase	increase	NOUN
fcis-30157	407	8	(	(	PUNCT
fcis-30157	407	9	90.1	90.1	NUM
fcis-30157	407	10	%	%	NOUN
fcis-30157	407	11	)	)	PUNCT
fcis-30157	407	12	,	,	PUNCT
fcis-30157	407	13	vgg	vgg	NOUN
fcis-30157	407	14	and	and	CCONJ
fcis-30157	407	15	inception	inception	NOUN
fcis-30157	407	16	saw	see	VERB
fcis-30157	407	17	negligible	negligible	ADJ
fcis-30157	407	18	changes	change	NOUN
fcis-30157	407	19	,	,	PUNCT
fcis-30157	407	20	suggesting	suggest	VERB
fcis-30157	407	21	that	that	SCONJ
fcis-30157	407	22	these	these	DET
fcis-30157	407	23	models	model	NOUN
fcis-30157	407	24	may	may	AUX
fcis-30157	407	25	not	not	PART
fcis-30157	407	26	fully	fully	ADV
fcis-30157	407	27	benefit	benefit	VERB
fcis-30157	407	28	from	from	ADP
fcis-30157	407	29	additional	additional	ADJ
fcis-30157	407	30	contextual	contextual	ADJ
fcis-30157	407	31	information	information	NOUN
fcis-30157	407	32	.	.	PUNCT
fcis-30157	408	1	we	we	PRON
fcis-30157	408	2	also	also	ADV
fcis-30157	408	3	explored	explore	VERB
fcis-30157	408	4	hyperparameter	hyperparameter	NOUN
fcis-30157	408	5	tuning	tuning	NOUN
fcis-30157	408	6	,	,	PUNCT
fcis-30157	408	7	which	which	PRON
fcis-30157	408	8	improved	improve	VERB
fcis-30157	408	9	resnet	resnet	NOUN
fcis-30157	408	10	's	's	PART
fcis-30157	408	11	accuracy	accuracy	NOUN
fcis-30157	408	12	to	to	ADP
fcis-30157	408	13	92.3	92.3	NUM
fcis-30157	408	14	%	%	NOUN
fcis-30157	408	15	with	with	ADP
fcis-30157	408	16	optimal	optimal	ADJ
fcis-30157	408	17	settings	setting	NOUN
fcis-30157	408	18	(	(	PUNCT
fcis-30157	408	19	learning	learn	VERB
fcis-30157	408	20	rate	rate	NOUN
fcis-30157	408	21	:	:	PUNCT
fcis-30157	408	22	0.001	0.001	NUM
fcis-30157	408	23	,	,	PUNCT
fcis-30157	408	24	batch	batch	NOUN
fcis-30157	408	25	size	size	NOUN
fcis-30157	408	26	:	:	PUNCT
fcis-30157	408	27	32	32	NUM
fcis-30157	408	28	)	)	PUNCT
fcis-30157	408	29	,	,	PUNCT
fcis-30157	408	30	emphasizing	emphasize	VERB
fcis-30157	408	31	the	the	DET
fcis-30157	408	32	importance	importance	NOUN
fcis-30157	408	33	of	of	ADP
fcis-30157	408	34	careful	careful	ADJ
fcis-30157	408	35	tuning	tuning	NOUN
fcis-30157	408	36	for	for	ADP
fcis-30157	408	37	optimal	optimal	ADJ
fcis-30157	408	38	performance	performance	NOUN
fcis-30157	408	39	.	.	PUNCT
fcis-30157	409	1	analyzing	analyze	VERB
fcis-30157	409	2	error	error	NOUN
fcis-30157	409	3	distribution	distribution	NOUN
fcis-30157	409	4	,	,	PUNCT
fcis-30157	409	5	we	we	PRON
fcis-30157	409	6	found	find	VERB
fcis-30157	409	7	diseases	disease	NOUN
fcis-30157	409	8	like	like	ADP
fcis-30157	409	9	melanoma	melanoma	NOUN
fcis-30157	409	10	and	and	CCONJ
fcis-30157	409	11	actinic	actinic	ADJ
fcis-30157	409	12	keratosis	keratosis	NOUN
fcis-30157	409	13	harder	hard	ADV
fcis-30157	409	14	to	to	PART
fcis-30157	409	15	classify	classify	VERB
fcis-30157	409	16	due	due	ADJ
fcis-30157	409	17	to	to	ADP
fcis-30157	409	18	subtle	subtle	ADJ
fcis-30157	409	19	visual	visual	ADJ
fcis-30157	409	20	differences	difference	NOUN
fcis-30157	409	21	.	.	PUNCT
fcis-30157	410	1	to	to	PART
fcis-30157	410	2	address	address	VERB
fcis-30157	410	3	this	this	PRON
fcis-30157	410	4	,	,	PUNCT
fcis-30157	410	5	we	we	PRON
fcis-30157	410	6	used	use	VERB
fcis-30157	410	7	a	a	DET
fcis-30157	410	8	multi	multi	ADJ
fcis-30157	410	9	-	-	ADJ
fcis-30157	410	10	modal	modal	ADJ
fcis-30157	410	11	fusion	fusion	NOUN
fcis-30157	410	12	approach	approach	NOUN
fcis-30157	410	13	,	,	PUNCT
fcis-30157	410	14	combining	combine	VERB
fcis-30157	410	15	skin	skin	NOUN
fcis-30157	410	16	images	image	NOUN
fcis-30157	410	17	with	with	ADP
fcis-30157	410	18	clinical	clinical	ADJ
fcis-30157	410	19	data	datum	NOUN
fcis-30157	410	20	.	.	PUNCT
fcis-30157	411	1	this	this	DET
fcis-30157	411	2	method	method	NOUN
fcis-30157	411	3	significantly	significantly	ADV
fcis-30157	411	4	improved	improve	VERB
fcis-30157	411	5	accuracy	accuracy	NOUN
fcis-30157	411	6	,	,	PUNCT
fcis-30157	411	7	with	with	ADP
fcis-30157	411	8	melanoma	melanoma	NOUN
fcis-30157	411	9	rising	rise	VERB
fcis-30157	411	10	from	from	ADP
fcis-30157	411	11	85	85	NUM
fcis-30157	411	12	%	%	NOUN
fcis-30157	411	13	to	to	ADP
fcis-30157	411	14	92	92	NUM
fcis-30157	411	15	%	%	NOUN
fcis-30157	411	16	,	,	PUNCT
fcis-30157	411	17	and	and	CCONJ
fcis-30157	411	18	actinic	actinic	ADJ
fcis-30157	411	19	keratosis	keratosis	NOUN
fcis-30157	411	20	from	from	ADP
fcis-30157	411	21	82	82	NUM
fcis-30157	411	22	%	%	NOUN
fcis-30157	411	23	to	to	PART
fcis-30157	411	24	89	89	NUM
fcis-30157	411	25	%	%	NOUN
fcis-30157	411	26	.	.	PUNCT
fcis-30157	412	1	3.2.2	3.2.2	NUM
fcis-30157	412	2	.	.	PUNCT
fcis-30157	412	3	comparison	comparison	NOUN
fcis-30157	412	4	with	with	ADP
fcis-30157	412	5	existing	exist	VERB
fcis-30157	412	6	methods	method	NOUN
fcis-30157	412	7	in	in	ADP
fcis-30157	412	8	this	this	DET
fcis-30157	412	9	study	study	NOUN
fcis-30157	412	10	,	,	PUNCT
fcis-30157	412	11	we	we	PRON
fcis-30157	412	12	compared	compare	VERB
fcis-30157	412	13	our	our	PRON
fcis-30157	412	14	proposed	propose	VERB
fcis-30157	412	15	method	method	NOUN
fcis-30157	412	16	with	with	ADP
fcis-30157	412	17	existing	exist	VERB
fcis-30157	412	18	approaches	approach	NOUN
fcis-30157	412	19	to	to	PART
fcis-30157	412	20	evaluate	evaluate	VERB
fcis-30157	412	21	its	its	PRON
fcis-30157	412	22	performance	performance	NOUN
fcis-30157	412	23	in	in	ADP
fcis-30157	412	24	skin	skin	NOUN
fcis-30157	412	25	disease	disease	NOUN
fcis-30157	412	26	recognition	recognition	NOUN
fcis-30157	412	27	and	and	CCONJ
fcis-30157	412	28	auxiliary	auxiliary	ADJ
fcis-30157	412	29	diagnosis	diagnosis	NOUN
fcis-30157	412	30	across	across	ADP
fcis-30157	412	31	dimensions	dimension	NOUN
fcis-30157	412	32	of	of	ADP
fcis-30157	412	33	accuracy	accuracy	NOUN
fcis-30157	412	34	,	,	PUNCT
fcis-30157	412	35	computational	computational	ADJ
fcis-30157	412	36	efficiency	efficiency	NOUN
fcis-30157	412	37	,	,	PUNCT
fcis-30157	412	38	and	and	CCONJ
fcis-30157	412	39	generalization	generalization	NOUN
fcis-30157	412	40	ability	ability	NOUN
fcis-30157	412	41	.	.	PUNCT
fcis-30157	413	1	accuracy	accuracy	NOUN
fcis-30157	413	2	comparison	comparison	NOUN
fcis-30157	413	3	our	our	PRON
fcis-30157	413	4	method	method	NOUN
fcis-30157	413	5	achieved	achieve	VERB
fcis-30157	413	6	an	an	DET
fcis-30157	413	7	overall	overall	ADJ
fcis-30157	413	8	accuracy	accuracy	NOUN
fcis-30157	413	9	of	of	ADP
fcis-30157	413	10	92.3	92.3	NUM
fcis-30157	413	11	%	%	NOUN
fcis-30157	413	12	on	on	ADP
fcis-30157	413	13	the	the	DET
fcis-30157	413	14	isic	isic	PROPN
fcis-30157	413	15	2019	2019	NUM
fcis-30157	413	16	dataset	dataset	NOUN
fcis-30157	413	17	,	,	PUNCT
fcis-30157	413	18	outperforming	outperform	VERB
fcis-30157	413	19	previous	previous	ADJ
fcis-30157	413	20	models	model	NOUN
fcis-30157	413	21	:	:	PUNCT
fcis-30157	413	22	mdfnet	mdfnet	NOUN
fcis-30157	413	23	:	:	PUNCT
fcis-30157	413	24	88.7	88.7	NUM
fcis-30157	413	25	%	%	NOUN
fcis-30157	413	26	model	model	NOUN
fcis-30157	413	27	in	in	ADP
fcis-30157	413	28	[	[	X
fcis-30157	413	29	12	12	NUM
fcis-30157	413	30	]	]	SYM
fcis-30157	413	31	:	:	PUNCT
fcis-30157	413	32	89.5	89.5	NUM
fcis-30157	413	33	%	%	NOUN
fcis-30157	413	34	hybrid	hybrid	NOUN
fcis-30157	413	35	model	model	NOUN
fcis-30157	413	36	in	in	ADP
fcis-30157	413	37	[	[	X
fcis-30157	413	38	13	13	NUM
fcis-30157	413	39	]	]	SYM
fcis-30157	413	40	:	:	PUNCT
fcis-30157	413	41	90.1	90.1	NUM
fcis-30157	413	42	%	%	NOUN
fcis-30157	413	43	the	the	DET
fcis-30157	413	44	improvement	improvement	NOUN
fcis-30157	413	45	is	be	AUX
fcis-30157	413	46	attributed	attribute	VERB
fcis-30157	413	47	to	to	ADP
fcis-30157	413	48	a	a	DET
fcis-30157	413	49	dual	dual	ADJ
fcis-30157	413	50	-	-	PUNCT
fcis-30157	413	51	attention	attention	NOUN
fcis-30157	413	52	mechanism	mechanism	NOUN
fcis-30157	413	53	focusing	focus	VERB
fcis-30157	413	54	on	on	ADP
fcis-30157	413	55	both	both	CCONJ
fcis-30157	413	56	local	local	ADJ
fcis-30157	413	57	and	and	CCONJ
fcis-30157	413	58	global	global	ADJ
fcis-30157	413	59	features	feature	NOUN
fcis-30157	413	60	,	,	PUNCT
fcis-30157	413	61	enhanced	enhance	VERB
fcis-30157	413	62	by	by	ADP
fcis-30157	413	63	auxiliary	auxiliary	ADJ
fcis-30157	413	64	learning	learning	NOUN
fcis-30157	413	65	tasks	task	NOUN
fcis-30157	413	66	.	.	PUNCT
fcis-30157	414	1	computational	computational	ADJ
fcis-30157	414	2	efficiency	efficiency	NOUN
fcis-30157	414	3	our	our	PRON
fcis-30157	414	4	method	method	NOUN
fcis-30157	414	5	balanced	balanced	ADJ
fcis-30157	414	6	accuracy	accuracy	NOUN
fcis-30157	414	7	with	with	ADP
fcis-30157	414	8	resource	resource	NOUN
fcis-30157	414	9	consumption	consumption	NOUN
fcis-30157	414	10	,	,	PUNCT
fcis-30157	414	11	achieving	achieve	VERB
fcis-30157	414	12	an	an	DET
fcis-30157	414	13	average	average	ADJ
fcis-30157	414	14	inference	inference	NOUN
fcis-30157	414	15	time	time	NOUN
fcis-30157	414	16	of	of	ADP
fcis-30157	414	17	0.12	0.12	NUM
fcis-30157	414	18	seconds	second	NOUN
fcis-30157	414	19	per	per	ADP
fcis-30157	414	20	image	image	NOUN
fcis-30157	414	21	on	on	ADP
fcis-30157	414	22	a	a	DET
fcis-30157	414	23	standard	standard	ADJ
fcis-30157	414	24	gpu	gpu	NOUN
fcis-30157	414	25	,	,	PUNCT
fcis-30157	414	26	compared	compare	VERB
fcis-30157	414	27	to	to	ADP
fcis-30157	414	28	0.15	0.15	NUM
fcis-30157	414	29	seconds	second	NOUN
fcis-30157	414	30	reported	report	VERB
fcis-30157	414	31	by	by	ADP
fcis-30157	414	32	a	a	DET
fcis-30157	414	33	mobilenet	mobilenet	NOUN
fcis-30157	414	34	-	-	PUNCT
fcis-30157	414	35	based	base	VERB
fcis-30157	414	36	model	model	NOUN
fcis-30157	414	37	.	.	PUNCT
fcis-30157	415	1	this	this	DET
fcis-30157	415	2	efficiency	efficiency	NOUN
fcis-30157	415	3	is	be	AUX
fcis-30157	415	4	notable	notable	ADJ
fcis-30157	415	5	even	even	ADV
fcis-30157	415	6	with	with	ADP
fcis-30157	415	7	higher	high	ADJ
fcis-30157	415	8	accuracy	accuracy	NOUN
fcis-30157	415	9	,	,	PUNCT
fcis-30157	415	10	unlike	unlike	ADP
fcis-30157	415	11	some	some	DET
fcis-30157	415	12	existing	exist	VERB
fcis-30157	415	13	methods	method	NOUN
fcis-30157	415	14	that	that	PRON
fcis-30157	415	15	require	require	VERB
fcis-30157	415	16	significantly	significantly	ADV
fcis-30157	415	17	more	more	ADV
fcis-30157	415	18	computational	computational	ADJ
fcis-30157	415	19	power	power	NOUN
fcis-30157	415	20	.	.	PUNCT
fcis-30157	416	1	generalization	generalization	NOUN
fcis-30157	416	2	ability	ability	NOUN
fcis-30157	416	3	our	our	PRON
fcis-30157	416	4	method	method	NOUN
fcis-30157	416	5	achieved	achieve	VERB
fcis-30157	416	6	a	a	DET
fcis-30157	416	7	cross	cross	ADJ
fcis-30157	416	8	-	-	ADJ
fcis-30157	416	9	validation	validation	ADJ
fcis-30157	416	10	accuracy	accuracy	NOUN
fcis-30157	416	11	of	of	ADP
fcis-30157	416	12	91.2	91.2	NUM
fcis-30157	416	13	%	%	NOUN
fcis-30157	416	14	on	on	ADP
fcis-30157	416	15	a	a	DET
fcis-30157	416	16	diverse	diverse	ADJ
fcis-30157	416	17	dataset	dataset	NOUN
fcis-30157	416	18	,	,	PUNCT
fcis-30157	416	19	indicating	indicate	VERB
fcis-30157	416	20	strong	strong	ADJ
fcis-30157	416	21	generalization	generalization	NOUN
fcis-30157	416	22	ability	ability	NOUN
fcis-30157	416	23	.	.	PUNCT
fcis-30157	417	1	this	this	DET
fcis-30157	417	2	performance	performance	NOUN
fcis-30157	417	3	is	be	AUX
fcis-30157	417	4	comparable	comparable	ADJ
fcis-30157	417	5	to	to	ADP
fcis-30157	417	6	the	the	DET
fcis-30157	417	7	deep	deep	ADJ
fcis-30157	417	8	learning	learning	NOUN
fcis-30157	417	9	-	-	PUNCT
fcis-30157	417	10	aided	aid	VERB
fcis-30157	417	11	system	system	NOUN
fcis-30157	417	12	described	describe	VERB
fcis-30157	417	13	in	in	ADP
fcis-30157	417	14	[	[	X
fcis-30157	417	15	15	15	NUM
fcis-30157	417	16	]	]	PUNCT
fcis-30157	417	17	.	.	PUNCT
fcis-30157	418	1	other	other	ADJ
fcis-30157	418	2	models	model	NOUN
fcis-30157	418	3	,	,	PUNCT
fcis-30157	418	4	like	like	ADP
fcis-30157	418	5	the	the	DET
fcis-30157	418	6	cnn	cnn	PROPN
fcis-30157	418	7	model	model	NOUN
fcis-30157	418	8	in	in	ADP
fcis-30157	418	9	[	[	X
fcis-30157	418	10	16	16	NUM
fcis-30157	418	11	]	]	PUNCT
fcis-30157	418	12	,	,	PUNCT
fcis-30157	418	13	showed	show	VERB
fcis-30157	418	14	limited	limited	ADJ
fcis-30157	418	15	generalization	generalization	NOUN
fcis-30157	418	16	with	with	ADP
fcis-30157	418	17	varied	varied	ADJ
fcis-30157	418	18	image	image	NOUN
fcis-30157	418	19	quality	quality	NOUN
fcis-30157	418	20	or	or	CCONJ
fcis-30157	418	21	patient	patient	ADJ
fcis-30157	418	22	demographics	demographic	NOUN
fcis-30157	418	23	.	.	PUNCT
fcis-30157	419	1	our	our	PRON
fcis-30157	419	2	method	method	NOUN
fcis-30157	419	3	shows	show	VERB
fcis-30157	419	4	significant	significant	ADJ
fcis-30157	419	5	improvements	improvement	NOUN
fcis-30157	419	6	over	over	ADP
fcis-30157	419	7	existing	exist	VERB
fcis-30157	419	8	approaches	approach	NOUN
fcis-30157	419	9	in	in	ADP
fcis-30157	419	10	terms	term	NOUN
fcis-30157	419	11	of	of	ADP
fcis-30157	419	12	accuracy	accuracy	NOUN
fcis-30157	419	13	,	,	PUNCT
fcis-30157	419	14	computational	computational	ADJ
fcis-30157	419	15	efficiency	efficiency	NOUN
fcis-30157	419	16	,	,	PUNCT
fcis-30157	419	17	and	and	CCONJ
fcis-30157	419	18	generalization	generalization	NOUN
fcis-30157	419	19	ability	ability	NOUN
fcis-30157	419	20	.	.	PUNCT
fcis-30157	420	1	it	it	PRON
fcis-30157	420	2	is	be	AUX
fcis-30157	420	3	both	both	CCONJ
fcis-30157	420	4	accurate	accurate	ADJ
fcis-30157	420	5	and	and	CCONJ
fcis-30157	420	6	practical	practical	ADJ
fcis-30157	420	7	for	for	ADP
fcis-30157	420	8	real	real	ADJ
fcis-30157	420	9	-	-	PUNCT
fcis-30157	420	10	world	world	NOUN
fcis-30157	420	11	applications	application	NOUN
fcis-30157	420	12	but	but	CCONJ
fcis-30157	420	13	has	have	VERB
fcis-30157	420	14	room	room	NOUN
fcis-30157	420	15	for	for	ADP
fcis-30157	420	16	improvement	improvement	NOUN
fcis-30157	420	17	in	in	ADP
fcis-30157	420	18	data	datum	NOUN
fcis-30157	420	19	quality	quality	NOUN
fcis-30157	420	20	and	and	CCONJ
fcis-30157	420	21	model	model	NOUN
fcis-30157	420	22	interpretability	interpretability	NOUN
fcis-30157	420	23	.	.	PUNCT
fcis-30157	421	1	show	show	VERB
fcis-30157	421	2	in	in	ADP
fcis-30157	421	3	table	table	NOUN
fcis-30157	421	4	4	4	NUM
fcis-30157	421	5	.	.	PUNCT
fcis-30157	421	6	table	table	NOUN
fcis-30157	421	7	4	4	NUM
fcis-30157	421	8	.	.	PUNCT
fcis-30157	421	9	performance	performance	NOUN
fcis-30157	421	10	comparison	comparison	NOUN
fcis-30157	421	11	of	of	ADP
fcis-30157	421	12	proposed	propose	VERB
fcis-30157	421	13	method	method	NOUN
fcis-30157	421	14	vs.	vs.	ADP
fcis-30157	421	15	existing	exist	VERB
fcis-30157	421	16	approaches	approach	NOUN
fcis-30157	421	17	comparison	comparison	NOUN
fcis-30157	421	18	dimension	dimension	NOUN
fcis-30157	421	19	our	our	PRON
fcis-30157	421	20	method	method	NOUN
fcis-30157	421	21	existing	exist	VERB
fcis-30157	421	22	approaches	approach	NOUN
fcis-30157	421	23	accuracy	accuracy	NOUN
fcis-30157	421	24	92.3	92.3	NUM
fcis-30157	421	25	%	%	NOUN
fcis-30157	421	26	on	on	ADP
fcis-30157	421	27	isic	isic	PROPN
fcis-30157	421	28	2019	2019	NUM
fcis-30157	421	29	dataset	dataset	VERB
fcis-30157	421	30	higher	high	ADJ
fcis-30157	421	31	accuracy	accuracy	NOUN
fcis-30157	421	32	due	due	ADP
fcis-30157	421	33	to	to	ADP
fcis-30157	421	34	dual	dual	ADJ
fcis-30157	421	35	-	-	PUNCT
fcis-30157	421	36	attention	attention	NOUN
fcis-30157	421	37	and	and	CCONJ
fcis-30157	421	38	auxiliary	auxiliary	ADJ
fcis-30157	421	39	tasks	task	NOUN
fcis-30157	421	40	mdfnet	mdfnet	NOUN
fcis-30157	421	41	:	:	PUNCT
fcis-30157	421	42	88.7	88.7	NUM
fcis-30157	421	43	%	%	NOUN
fcis-30157	421	44	model	model	NOUN
fcis-30157	421	45	in	in	ADP
fcis-30157	421	46	[	[	X
fcis-30157	421	47	12	12	NUM
fcis-30157	421	48	]	]	SYM
fcis-30157	421	49	:	:	PUNCT
fcis-30157	421	50	89.5	89.5	NUM
fcis-30157	421	51	%	%	NOUN
fcis-30157	421	52	hybrid	hybrid	NOUN
fcis-30157	421	53	model	model	NOUN
fcis-30157	421	54	in	in	ADP
fcis-30157	421	55	[	[	X
fcis-30157	421	56	13	13	NUM
fcis-30157	421	57	]	]	SYM
fcis-30157	421	58	:	:	PUNCT
fcis-30157	421	59	90.1	90.1	NUM
fcis-30157	421	60	%	%	NOUN
fcis-30157	421	61	computational	computational	ADJ
fcis-30157	421	62	efficiency	efficiency	NOUN
fcis-30157	421	63	0.12	0.12	NUM
fcis-30157	421	64	seconds	second	NOUN
fcis-30157	421	65	per	per	ADP
fcis-30157	421	66	image	image	NOUN
fcis-30157	421	67	on	on	ADP
fcis-30157	421	68	standard	standard	ADJ
fcis-30157	421	69	gpu	gpu	NOUN
fcis-30157	421	70	balanced	balanced	ADJ
fcis-30157	421	71	accuracy	accuracy	NOUN
fcis-30157	421	72	with	with	ADP
fcis-30157	421	73	resource	resource	NOUN
fcis-30157	421	74	consumption	consumption	NOUN
fcis-30157	421	75	mobilenet	mobilenet	NOUN
fcis-30157	421	76	-	-	PUNCT
fcis-30157	421	77	based	base	VERB
fcis-30157	421	78	model	model	NOUN
fcis-30157	421	79	in	in	ADP
fcis-30157	421	80	[	[	X
fcis-30157	421	81	14	14	NUM
fcis-30157	421	82	]	]	SYM
fcis-30157	421	83	:	:	PUNCT
fcis-30157	421	84	0.15	0.15	NUM
fcis-30157	421	85	seconds	second	NOUN
fcis-30157	421	86	multi	multi	ADJ
fcis-30157	421	87	-	-	ADJ
fcis-30157	421	88	task	task	ADJ
fcis-30157	421	89	model	model	NOUN
fcis-30157	421	90	in	in	ADP
fcis-30157	421	91	[	[	X
fcis-30157	421	92	15	15	NUM
fcis-30157	421	93	]	]	SYM
fcis-30157	421	94	:	:	PUNCT
fcis-30157	421	95	high	high	ADJ
fcis-30157	421	96	power	power	NOUN
fcis-30157	421	97	use	use	NOUN
fcis-30157	421	98	generalization	generalization	NOUN
fcis-30157	421	99	ability	ability	NOUN
fcis-30157	421	100	91.2	91.2	NUM
fcis-30157	421	101	%	%	NOUN
fcis-30157	421	102	cross	cross	ADJ
fcis-30157	421	103	-	-	ADJ
fcis-30157	421	104	validation	validation	ADJ
fcis-30157	421	105	accuracy	accuracy	NOUN
fcis-30157	421	106	on	on	ADP
fcis-30157	421	107	diverse	diverse	ADJ
fcis-30157	421	108	dataset	dataset	VERB
fcis-30157	421	109	similar	similar	ADJ
fcis-30157	421	110	accuracy	accuracy	NOUN
fcis-30157	421	111	in	in	ADP
fcis-30157	421	112	[	[	X
fcis-30157	421	113	15	15	NUM
fcis-30157	421	114	]	]	X
fcis-30157	421	115	limited	limited	ADJ
fcis-30157	421	116	generalization	generalization	NOUN
fcis-30157	421	117	in	in	ADP
fcis-30157	421	118	[	[	X
fcis-30157	421	119	16	16	NUM
fcis-30157	421	120	]	]	PUNCT
fcis-30157	421	121	3.2.3	3.2.3	NUM
fcis-30157	421	122	.	.	PUNCT
fcis-30157	422	1	analysis	analysis	NOUN
fcis-30157	422	2	of	of	ADP
fcis-30157	422	3	recognition	recognition	NOUN
fcis-30157	422	4	errors	error	NOUN
fcis-30157	422	5	in	in	ADP
fcis-30157	422	6	this	this	DET
fcis-30157	422	7	study	study	NOUN
fcis-30157	422	8	,	,	PUNCT
fcis-30157	422	9	the	the	DET
fcis-30157	422	10	causes	cause	NOUN
fcis-30157	422	11	of	of	ADP
fcis-30157	422	12	model	model	NOUN
fcis-30157	422	13	misjudgment	misjudgment	NOUN
fcis-30157	422	14	in	in	ADP
fcis-30157	422	15	skin	skin	NOUN
fcis-30157	422	16	disease	disease	NOUN
fcis-30157	422	17	diagnosis	diagnosis	NOUN
fcis-30157	422	18	were	be	AUX
fcis-30157	422	19	analyzed	analyze	VERB
fcis-30157	422	20	,	,	PUNCT
fcis-30157	422	21	with	with	ADP
fcis-30157	422	22	several	several	ADJ
fcis-30157	422	23	factors	factor	NOUN
fcis-30157	422	24	identified	identify	VERB
fcis-30157	422	25	as	as	ADP
fcis-30157	422	26	contributing	contribute	VERB
fcis-30157	422	27	to	to	ADP
fcis-30157	422	28	the	the	DET
fcis-30157	422	29	misclassifications	misclassification	NOUN
fcis-30157	422	30	.	.	PUNCT
fcis-30157	423	1	disease	disease	NOUN
fcis-30157	423	2	characteristics	characteristic	NOUN
fcis-30157	423	3	and	and	CCONJ
fcis-30157	423	4	misjudgments	misjudgment	NOUN
fcis-30157	423	5	melanoma	melanoma	PROPN
fcis-30157	423	6	vs.	vs.	ADP
fcis-30157	423	7	seborrheic	seborrheic	VERB
fcis-30157	423	8	keratosis	keratosis	NOUN
fcis-30157	423	9	:	:	PUNCT
fcis-30157	423	10	the	the	DET
fcis-30157	423	11	confusion	confusion	NOUN
fcis-30157	423	12	rate	rate	NOUN
fcis-30157	423	13	between	between	ADP
fcis-30157	423	14	these	these	DET
fcis-30157	423	15	two	two	NUM
fcis-30157	423	16	diseases	disease	NOUN
fcis-30157	423	17	was	be	AUX
fcis-30157	423	18	18.7	18.7	NUM
fcis-30157	423	19	%	%	NOUN
fcis-30157	423	20	,	,	PUNCT
fcis-30157	423	21	mainly	mainly	ADV
fcis-30157	423	22	due	due	ADP
fcis-30157	423	23	to	to	ADP
fcis-30157	423	24	similarities	similarity	NOUN
fcis-30157	423	25	in	in	ADP
fcis-30157	423	26	lesion	lesion	NOUN
fcis-30157	423	27	edge	edge	NOUN
fcis-30157	423	28	ambiguity	ambiguity	NOUN
fcis-30157	423	29	and	and	CCONJ
fcis-30157	423	30	pigment	pigment	NOUN
fcis-30157	423	31	distribution	distribution	NOUN
fcis-30157	423	32	.	.	PUNCT
fcis-30157	424	1	small	small	ADJ
fcis-30157	424	2	lesions	lesion	NOUN
fcis-30157	424	3	:	:	PUNCT
fcis-30157	424	4	when	when	SCONJ
fcis-30157	424	5	lesions	lesion	NOUN
fcis-30157	424	6	are	be	AUX
fcis-30157	424	7	<	<	X
fcis-30157	424	8	6	6	NUM
fcis-30157	424	9	mm	mm	NOUN
fcis-30157	424	10	in	in	ADP
fcis-30157	424	11	diameter	diameter	NOUN
fcis-30157	424	12	,	,	PUNCT
fcis-30157	424	13	the	the	DET
fcis-30157	424	14	accuracy	accuracy	NOUN
fcis-30157	424	15	of	of	ADP
fcis-30157	424	16	detecting	detect	VERB
fcis-30157	424	17	basal	basal	ADJ
fcis-30157	424	18	cell	cell	NOUN
fcis-30157	424	19	carcinoma	carcinoma	NOUN
fcis-30157	424	20	dropped	drop	VERB
fcis-30157	424	21	by	by	ADP
fcis-30157	424	22	12.3	12.3	NUM
fcis-30157	424	23	%	%	NOUN
fcis-30157	424	24	due	due	ADP
fcis-30157	424	25	to	to	ADP
fcis-30157	424	26	loss	loss	NOUN
fcis-30157	424	27	of	of	ADP
fcis-30157	424	28	texture	texture	ADJ
fcis-30157	424	29	details	detail	NOUN
fcis-30157	424	30	in	in	ADP
fcis-30157	424	31	small	small	ADJ
fcis-30157	424	32	-	-	PUNCT
fcis-30157	424	33	sized	sized	ADJ
fcis-30157	424	34	lesions	lesion	NOUN
fcis-30157	424	35	.	.	PUNCT
fcis-30157	425	1	discoid	discoid	PROPN
fcis-30157	425	2	lupus	lupus	PROPN
fcis-30157	425	3	erythematosus	erythematosus	PROPN
fcis-30157	425	4	:	:	PUNCT
fcis-30157	425	5	the	the	DET
fcis-30157	425	6	efficientnetb7	efficientnetb7	NOUN
fcis-30157	425	7	system	system	NOUN
fcis-30157	425	8	misdiagnosed	misdiagnose	VERB
fcis-30157	425	9	15	15	NUM
fcis-30157	425	10	cases	case	NOUN
fcis-30157	425	11	as	as	ADP
fcis-30157	425	12	psoriasis	psoriasis	NOUN
fcis-30157	425	13	,	,	PUNCT
fcis-30157	425	14	failing	fail	VERB
fcis-30157	425	15	to	to	PART
fcis-30157	425	16	capture	capture	VERB
fcis-30157	425	17	horn	horn	NOUN
fcis-30157	425	18	plug	plug	NOUN
fcis-30157	425	19	features	feature	NOUN
fcis-30157	425	20	in	in	ADP
fcis-30157	425	21	the	the	DET
fcis-30157	425	22	lesions	lesion	NOUN
fcis-30157	425	23	.	.	PUNCT
fcis-30157	426	1	model	model	NOUN
fcis-30157	426	2	attention	attention	NOUN
fcis-30157	426	3	bias	bias	NOUN
fcis-30157	426	4	and	and	CCONJ
fcis-30157	426	5	misjudgment	misjudgment	NOUN
fcis-30157	426	6	psoriasis	psoriasis	NOUN
fcis-30157	426	7	:	:	PUNCT
fcis-30157	426	8	the	the	DET
fcis-30157	426	9	model	model	NOUN
fcis-30157	426	10	focused	focus	VERB
fcis-30157	426	11	only	only	ADV
fcis-30157	426	12	23	23	NUM
fcis-30157	426	13	%	%	NOUN
fcis-30157	426	14	of	of	ADP
fcis-30157	426	15	its	its	PRON
fcis-30157	426	16	attention	attention	NOUN
fcis-30157	426	17	on	on	ADP
fcis-30157	426	18	psoriasis	psoriasis	NOUN
fcis-30157	426	19	scale	scale	NOUN
fcis-30157	426	20	features	feature	NOUN
fcis-30157	426	21	,	,	PUNCT
fcis-30157	427	1	while	while	SCONJ
fcis-30157	427	2	clinicians	clinician	NOUN
fcis-30157	427	3	assigned	assign	VERB
fcis-30157	427	4	58	58	NUM
fcis-30157	427	5	%	%	NOUN
fcis-30157	427	6	of	of	ADP
fcis-30157	427	7	their	their	PRON
fcis-30157	427	8	decision	decision	NOUN
fcis-30157	427	9	weight	weight	NOUN
fcis-30157	427	10	to	to	ADP
fcis-30157	427	11	this	this	DET
fcis-30157	427	12	feature	feature	NOUN
fcis-30157	427	13	,	,	PUNCT
fcis-30157	427	14	leading	lead	VERB
fcis-30157	427	15	to	to	ADP
fcis-30157	427	16	a	a	DET
fcis-30157	427	17	4.8x	4.8x	NUM
fcis-30157	427	18	higher	high	ADJ
fcis-30157	427	19	misdiagnosis	misdiagnosis	NOUN
fcis-30157	427	20	rate	rate	NOUN
fcis-30157	427	21	when	when	SCONJ
fcis-30157	427	22	psoriasis	psoriasis	NOUN
fcis-30157	427	23	was	be	AUX
fcis-30157	427	24	accompanied	accompany	VERB
fcis-30157	427	25	by	by	ADP
fcis-30157	427	26	eczema	eczema	NOUN
fcis-30157	427	27	-	-	PUNCT
fcis-30157	427	28	like	like	ADJ
fcis-30157	427	29	changes	change	NOUN
fcis-30157	427	30	.	.	PUNCT
fcis-30157	428	1	vascular	vascular	ADJ
fcis-30157	428	2	morphology	morphology	NOUN
fcis-30157	428	3	:	:	PUNCT
fcis-30157	428	4	the	the	DET
fcis-30157	428	5	model	model	NOUN
fcis-30157	428	6	captured	capture	VERB
fcis-30157	428	7	only	only	ADV
fcis-30157	428	8	34.2	34.2	NUM
fcis-30157	428	9	%	%	NOUN
fcis-30157	428	10	of	of	ADP
fcis-30157	428	11	branched	branched	ADJ
fcis-30157	428	12	blood	blood	NOUN
fcis-30157	428	13	vessels	vessel	NOUN
fcis-30157	428	14	when	when	SCONJ
fcis-30157	428	15	diagnosing	diagnose	VERB
fcis-30157	428	16	hemangioma	hemangioma	NOUN
fcis-30157	428	17	and	and	CCONJ
fcis-30157	428	18	suppurative	suppurative	ADJ
fcis-30157	428	19	granuloma	granuloma	NOUN
fcis-30157	428	20	,	,	PUNCT
fcis-30157	428	21	leading	lead	VERB
fcis-30157	428	22	to	to	ADP
fcis-30157	428	23	misjudgments	misjudgment	NOUN
fcis-30157	428	24	due	due	ADP
fcis-30157	428	25	to	to	ADP
fcis-30157	428	26	incorrect	incorrect	ADJ
fcis-30157	428	27	vascular	vascular	ADJ
fcis-30157	428	28	morphology	morphology	NOUN
fcis-30157	428	29	identification	identification	NOUN
fcis-30157	428	30	.	.	PUNCT
fcis-30157	429	1	data	datum	NOUN
fcis-30157	429	2	and	and	CCONJ
fcis-30157	429	3	skin	skin	NOUN
fcis-30157	429	4	color	color	NOUN
fcis-30157	429	5	bias	bias	NOUN
fcis-30157	429	6	skin	skin	NOUN
fcis-30157	429	7	color	color	NOUN
fcis-30157	429	8	differences	difference	NOUN
fcis-30157	429	9	:	:	PUNCT
fcis-30157	429	10	the	the	DET
fcis-30157	429	11	misdiagnosis	misdiagnosis	NOUN
fcis-30157	429	12	rate	rate	NOUN
fcis-30157	429	13	for	for	ADP
fcis-30157	429	14	seborrheic	seborrheic	ADJ
fcis-30157	429	15	keratosis	keratosis	NOUN
fcis-30157	429	16	in	in	ADP
fcis-30157	429	17	dark	dark	ADJ
fcis-30157	429	18	-	-	PUNCT
fcis-30157	429	19	skinned	skin	VERB
fcis-30157	429	20	patients	patient	NOUN
fcis-30157	429	21	was	be	AUX
fcis-30157	429	22	9.6	9.6	NUM
fcis-30157	429	23	%	%	NOUN
fcis-30157	429	24	higher	high	ADJ
fcis-30157	429	25	than	than	ADP
fcis-30157	429	26	in	in	ADP
fcis-30157	429	27	light	light	ADJ
fcis-30157	429	28	-	-	PUNCT
fcis-30157	429	29	skinned	skin	VERB
fcis-30157	429	30	patients	patient	NOUN
fcis-30157	429	31	due	due	ADJ
fcis-30157	429	32	to	to	ADP
fcis-30157	429	33	limited	limited	ADJ
fcis-30157	429	34	dark	dark	ADJ
fcis-30157	429	35	skin	skin	NOUN
fcis-30157	429	36	samples	sample	NOUN
fcis-30157	429	37	in	in	ADP
fcis-30157	429	38	the	the	DET
fcis-30157	429	39	training	training	NOUN
fcis-30157	429	40	data	datum	NOUN
fcis-30157	429	41	.	.	PUNCT
fcis-30157	430	1	the	the	DET
fcis-30157	430	2	mdfnet	mdfnet	NOUN
fcis-30157	430	3	system	system	NOUN
fcis-30157	430	4	increased	increase	VERB
fcis-30157	430	5	accuracy	accuracy	NOUN
fcis-30157	430	6	to	to	ADP
fcis-30157	430	7	89.4	89.4	NUM
fcis-30157	430	8	%	%	NOUN
fcis-30157	430	9	by	by	ADP
fcis-30157	430	10	integrating	integrate	VERB
fcis-30157	430	11	clinical	clinical	ADJ
fcis-30157	430	12	metadata	metadata	NOUN
fcis-30157	430	13	.	.	PUNCT
fcis-30157	431	1	other	other	ADJ
fcis-30157	431	2	misjudgments	misjudgment	NOUN
fcis-30157	431	3	and	and	CCONJ
fcis-30157	431	4	solutions	solution	NOUN
fcis-30157	431	5	bleeding	bleed	VERB
fcis-30157	431	6	dermatomas	dermatoma	NOUN
fcis-30157	431	7	:	:	PUNCT
fcis-30157	431	8	when	when	SCONJ
fcis-30157	431	9	diagnosing	diagnose	VERB
fcis-30157	431	10	dermatomas	dermatoma	NOUN
fcis-30157	431	11	with	with	ADP
fcis-30157	431	12	bleeding	bleeding	NOUN
fcis-30157	431	13	,	,	PUNCT
fcis-30157	431	14	the	the	DET
fcis-30157	431	15	model	model	NOUN
fcis-30157	431	16	focused	focus	VERB
fcis-30157	431	17	on	on	ADP
fcis-30157	431	18	the	the	DET
fcis-30157	431	19	bleeding	bleeding	NOUN
fcis-30157	431	20	area	area	NOUN
fcis-30157	431	21	rather	rather	ADV
fcis-30157	431	22	than	than	ADP
fcis-30157	431	23	the	the	DET
fcis-30157	431	24	peripheral	peripheral	ADJ
fcis-30157	431	25	pigmentation	pigmentation	NOUN
fcis-30157	431	26	band	band	NOUN
fcis-30157	431	27	,	,	PUNCT
fcis-30157	431	28	resulting	result	VERB
fcis-30157	431	29	in	in	ADP
fcis-30157	431	30	23.6	23.6	NUM
fcis-30157	431	31	%	%	NOUN
fcis-30157	431	32	misdiagnosis	misdiagnosis	NOUN
fcis-30157	431	33	as	as	ADP
fcis-30157	431	34	melanoma	melanoma	NOUN
fcis-30157	431	35	.	.	PUNCT
fcis-30157	432	1	the	the	DET
fcis-30157	432	2	dual	dual	ADJ
fcis-30157	432	3	attention	attention	NOUN
fcis-30157	432	4	121	121	NUM
fcis-30157	432	5	network	network	NOUN
fcis-30157	432	6	reduced	reduce	VERB
fcis-30157	432	7	this	this	DET
fcis-30157	432	8	rate	rate	NOUN
fcis-30157	432	9	to	to	ADP
fcis-30157	432	10	9.8	9.8	NUM
fcis-30157	432	11	%	%	NOUN
fcis-30157	432	12	.	.	PUNCT
fcis-30157	433	1	data	datum	NOUN
fcis-30157	433	2	quality	quality	NOUN
fcis-30157	433	3	:	:	PUNCT
fcis-30157	433	4	the	the	DET
fcis-30157	433	5	isic	isic	PROPN
fcis-30157	433	6	2019	2019	NUM
fcis-30157	433	7	dataset	dataset	NOUN
fcis-30157	433	8	contained	contain	VERB
fcis-30157	433	9	31.2	31.2	NUM
fcis-30157	433	10	%	%	NOUN
fcis-30157	433	11	blurry	blurry	NOUN
fcis-30157	433	12	images	image	NOUN
fcis-30157	433	13	,	,	PUNCT
fcis-30157	433	14	resulting	result	VERB
fcis-30157	433	15	in	in	ADP
fcis-30157	433	16	only	only	ADV
fcis-30157	433	17	72.3	72.3	NUM
fcis-30157	433	18	%	%	NOUN
fcis-30157	433	19	accuracy	accuracy	NOUN
fcis-30157	433	20	in	in	ADP
fcis-30157	433	21	distinguishing	distinguish	VERB
fcis-30157	433	22	eczema	eczema	NOUN
fcis-30157	433	23	from	from	ADP
fcis-30157	433	24	contact	contact	NOUN
fcis-30157	433	25	dermatitis	dermatitis	NOUN
fcis-30157	433	26	.	.	PUNCT
fcis-30157	434	1	a	a	DET
fcis-30157	434	2	standardized	standardized	ADJ
fcis-30157	434	3	shooting	shooting	NOUN
fcis-30157	434	4	protocol	protocol	NOUN
fcis-30157	434	5	in	in	ADP
fcis-30157	434	6	hospital	hospital	NOUN
fcis-30157	434	7	datasets	dataset	NOUN
fcis-30157	434	8	increased	increase	VERB
fcis-30157	434	9	accuracy	accuracy	NOUN
fcis-30157	434	10	to	to	ADP
fcis-30157	434	11	85.6	85.6	NUM
fcis-30157	434	12	%	%	NOUN
fcis-30157	434	13	.	.	PUNCT
fcis-30157	435	1	cnn	cnn	PROPN
fcis-30157	435	2	deficiencies	deficiency	NOUN
fcis-30157	435	3	:	:	PUNCT
fcis-30157	435	4	traditional	traditional	ADJ
fcis-30157	435	5	cnn	cnn	PROPN
fcis-30157	435	6	models	model	NOUN
fcis-30157	435	7	captured	capture	VERB
fcis-30157	435	8	only	only	ADV
fcis-30157	435	9	42	42	NUM
fcis-30157	435	10	%	%	NOUN
fcis-30157	435	11	of	of	ADP
fcis-30157	435	12	mast	mast	NOUN
fcis-30157	435	13	cell	cell	NOUN
fcis-30157	435	14	infiltration	infiltration	NOUN
fcis-30157	435	15	characteristics	characteristic	NOUN
fcis-30157	435	16	compared	compare	VERB
fcis-30157	435	17	to	to	ADP
fcis-30157	435	18	pathologists	pathologist	NOUN
fcis-30157	435	19	.	.	PUNCT
fcis-30157	436	1	using	use	VERB
fcis-30157	436	2	high	high	ADJ
fcis-30157	436	3	-	-	PUNCT
fcis-30157	436	4	resolution	resolution	NOUN
fcis-30157	436	5	dermatoscope	dermatoscope	NOUN
fcis-30157	436	6	images	image	NOUN
fcis-30157	436	7	with	with	ADP
fcis-30157	436	8	transformer	transformer	ADJ
fcis-30157	436	9	architecture	architecture	NOUN
fcis-30157	436	10	improved	improve	VERB
fcis-30157	436	11	capture	capture	NOUN
fcis-30157	436	12	efficiency	efficiency	NOUN
fcis-30157	436	13	to	to	ADP
fcis-30157	436	14	78.9	78.9	NUM
fcis-30157	436	15	%	%	NOUN
fcis-30157	436	16	.	.	PUNCT
fcis-30157	437	1	hair	hair	NOUN
fcis-30157	437	2	occlusion	occlusion	NOUN
fcis-30157	437	3	:	:	PUNCT
fcis-30157	437	4	the	the	DET
fcis-30157	437	5	unet++	unet++	PROPN
fcis-30157	437	6	segmentation	segmentation	NOUN
fcis-30157	437	7	network	network	NOUN
fcis-30157	437	8	reduced	reduce	VERB
fcis-30157	437	9	misjudgment	misjudgment	NOUN
fcis-30157	437	10	by	by	ADP
fcis-30157	437	11	17.4	17.4	NUM
fcis-30157	437	12	%	%	NOUN
fcis-30157	437	13	in	in	ADP
fcis-30157	437	14	images	image	NOUN
fcis-30157	437	15	with	with	ADP
fcis-30157	437	16	hair	hair	NOUN
fcis-30157	437	17	occlusion	occlusion	NOUN
fcis-30157	437	18	.	.	PUNCT
fcis-30157	438	1	parameter	parameter	NOUN
fcis-30157	438	2	and	and	CCONJ
fcis-30157	438	3	model	model	NOUN
fcis-30157	438	4	architecture	architecture	NOUN
fcis-30157	438	5	adjustments	adjustment	NOUN
fcis-30157	438	6	inappropriate	inappropriate	ADJ
fcis-30157	438	7	thresholds	threshold	NOUN
fcis-30157	438	8	:	:	PUNCT
fcis-30157	438	9	in	in	ADP
fcis-30157	438	10	diagnosing	diagnose	VERB
fcis-30157	438	11	skin	skin	NOUN
fcis-30157	438	12	t	t	PROPN
fcis-30157	438	13	-	-	PUNCT
fcis-30157	438	14	cell	cell	NOUN
fcis-30157	438	15	lymphoma	lymphoma	NOUN
fcis-30157	438	16	,	,	PUNCT
fcis-30157	438	17	the	the	DET
fcis-30157	438	18	model	model	NOUN
fcis-30157	438	19	's	's	PART
fcis-30157	438	20	response	response	NOUN
fcis-30157	438	21	threshold	threshold	NOUN
fcis-30157	438	22	was	be	AUX
fcis-30157	438	23	too	too	ADV
fcis-30157	438	24	high	high	ADJ
fcis-30157	438	25	,	,	PUNCT
fcis-30157	438	26	with	with	ADP
fcis-30157	438	27	early	early	ADJ
fcis-30157	438	28	recognition	recognition	NOUN
fcis-30157	438	29	sensitivity	sensitivity	NOUN
fcis-30157	438	30	at	at	ADP
fcis-30157	438	31	only	only	ADV
fcis-30157	438	32	63.2	63.2	NUM
fcis-30157	438	33	%	%	NOUN
fcis-30157	438	34	.	.	PUNCT
fcis-30157	439	1	after	after	ADP
fcis-30157	439	2	adjusting	adjust	VERB
fcis-30157	439	3	the	the	DET
fcis-30157	439	4	feature	feature	NOUN
fcis-30157	439	5	activation	activation	NOUN
fcis-30157	439	6	function	function	NOUN
fcis-30157	439	7	,	,	PUNCT
fcis-30157	439	8	sensitivity	sensitivity	NOUN
fcis-30157	439	9	increased	increase	VERB
fcis-30157	439	10	to	to	ADP
fcis-30157	439	11	79.8	79.8	NUM
fcis-30157	439	12	%	%	NOUN
fcis-30157	439	13	.	.	PUNCT
fcis-30157	440	1	small	small	ADJ
fcis-30157	440	2	sample	sample	NOUN
fcis-30157	440	3	diseases	disease	NOUN
fcis-30157	440	4	:	:	PUNCT
fcis-30157	440	5	traditional	traditional	ADJ
fcis-30157	440	6	models	model	NOUN
fcis-30157	440	7	had	have	VERB
fcis-30157	440	8	poor	poor	ADJ
fcis-30157	440	9	performance	performance	NOUN
fcis-30157	440	10	on	on	ADP
fcis-30157	440	11	keratosis	keratosis	NOUN
fcis-30157	440	12	with	with	ADP
fcis-30157	440	13	only	only	ADV
fcis-30157	440	14	58.3	58.3	NUM
fcis-30157	440	15	%	%	NOUN
fcis-30157	440	16	accuracy	accuracy	NOUN
fcis-30157	440	17	.	.	PUNCT
fcis-30157	441	1	the	the	DET
fcis-30157	441	2	fs3dciot	fs3dciot	NOUN
fcis-30157	441	3	framework	framework	NOUN
fcis-30157	441	4	with	with	ADP
fcis-30157	441	5	incremental	incremental	ADJ
fcis-30157	441	6	learning	learning	NOUN
fcis-30157	441	7	improved	improve	VERB
fcis-30157	441	8	accuracy	accuracy	NOUN
fcis-30157	441	9	to	to	ADP
fcis-30157	441	10	82.1	82.1	NUM
fcis-30157	441	11	%	%	NOUN
fcis-30157	441	12	.	.	PUNCT
fcis-30157	442	1	model	model	NOUN
fcis-30157	442	2	optimization	optimization	NOUN
fcis-30157	442	3	and	and	CCONJ
fcis-30157	442	4	enhancements	enhancement	NOUN
fcis-30157	442	5	atopic	atopic	NOUN
fcis-30157	442	6	dermatitis	dermatitis	NOUN
fcis-30157	442	7	:	:	PUNCT
fcis-30157	442	8	integrating	integrate	VERB
fcis-30157	442	9	clinical	clinical	ADJ
fcis-30157	442	10	data	datum	NOUN
fcis-30157	442	11	(	(	PUNCT
fcis-30157	442	12	age	age	NOUN
fcis-30157	442	13	,	,	PUNCT
fcis-30157	442	14	course	course	NOUN
fcis-30157	442	15	)	)	PUNCT
fcis-30157	442	16	increased	increase	VERB
fcis-30157	442	17	diagnostic	diagnostic	ADJ
fcis-30157	442	18	specificity	specificity	NOUN
fcis-30157	442	19	from	from	ADP
fcis-30157	442	20	76.4	76.4	NUM
fcis-30157	442	21	%	%	NOUN
fcis-30157	442	22	to	to	ADP
fcis-30157	442	23	89.3	89.3	NUM
fcis-30157	442	24	%	%	NOUN
fcis-30157	442	25	.	.	PUNCT
fcis-30157	443	1	data	datum	NOUN
fcis-30157	443	2	augmentation	augmentation	NOUN
fcis-30157	443	3	:	:	PUNCT
fcis-30157	443	4	attention	attention	NOUN
fcis-30157	443	5	-	-	PUNCT
fcis-30157	443	6	guided	guide	VERB
fcis-30157	443	7	data	data	NOUN
fcis-30157	443	8	augmentation	augmentation	NOUN
fcis-30157	443	9	strategies	strategy	NOUN
fcis-30157	443	10	reduced	reduce	VERB
fcis-30157	443	11	misjudgment	misjudgment	NOUN
fcis-30157	443	12	for	for	ADP
fcis-30157	443	13	alpha	alpha	NOUN
fcis-30157	443	14	fungi	fungus	NOUN
fcis-30157	443	15	from	from	ADP
fcis-30157	443	16	21.7	21.7	NUM
fcis-30157	443	17	%	%	NOUN
fcis-30157	443	18	to	to	ADP
fcis-30157	443	19	13.4	13.4	NUM
fcis-30157	443	20	%	%	NOUN
fcis-30157	443	21	,	,	PUNCT
fcis-30157	443	22	improving	improve	VERB
fcis-30157	443	23	recognition	recognition	NOUN
fcis-30157	443	24	accuracy	accuracy	NOUN
fcis-30157	443	25	by	by	ADP
fcis-30157	443	26	19.2	19.2	NUM
fcis-30157	443	27	%	%	NOUN
fcis-30157	443	28	.	.	PUNCT
fcis-30157	444	1	loss	loss	NOUN
fcis-30157	444	2	function	function	NOUN
fcis-30157	444	3	:	:	PUNCT
fcis-30157	444	4	using	use	VERB
fcis-30157	444	5	a	a	DET
fcis-30157	444	6	mixed	mixed	ADJ
fcis-30157	444	7	loss	loss	NOUN
fcis-30157	444	8	function	function	NOUN
fcis-30157	444	9	(	(	PUNCT
fcis-30157	444	10	focal	focal	ADJ
fcis-30157	444	11	loss	loss	NOUN
fcis-30157	444	12	+	+	CCONJ
fcis-30157	444	13	dice	dice	NOUN
fcis-30157	444	14	loss	loss	NOUN
fcis-30157	444	15	)	)	PUNCT
fcis-30157	444	16	reduced	reduce	VERB
fcis-30157	444	17	the	the	DET
fcis-30157	444	18	misdiagnosis	misdiagnosis	NOUN
fcis-30157	444	19	rate	rate	NOUN
fcis-30157	444	20	for	for	ADP
fcis-30157	444	21	ulcerative	ulcerative	ADJ
fcis-30157	444	22	skin	skin	NOUN
fcis-30157	444	23	tuberculosis	tuberculosis	NOUN
fcis-30157	444	24	by	by	ADP
fcis-30157	444	25	8.9	8.9	NUM
fcis-30157	444	26	%	%	NOUN
fcis-30157	444	27	.	.	PUNCT
fcis-30157	445	1	3.3	3.3	NUM
fcis-30157	445	2	.	.	PUNCT
fcis-30157	446	1	auxiliary	auxiliary	ADJ
fcis-30157	446	2	diagnosis	diagnosis	NOUN
fcis-30157	446	3	performance	performance	NOUN
fcis-30157	446	4	3.3.1	3.3.1	NUM
fcis-30157	446	5	.	.	PUNCT
fcis-30157	446	6	sensitivity	sensitivity	NOUN
fcis-30157	446	7	and	and	CCONJ
fcis-30157	446	8	specificity	specificity	NOUN
fcis-30157	446	9	of	of	ADP
fcis-30157	446	10	diagnosis	diagnosis	NOUN
fcis-30157	446	11	in	in	ADP
fcis-30157	446	12	auxiliary	auxiliary	ADJ
fcis-30157	446	13	diagnosis	diagnosis	NOUN
fcis-30157	446	14	,	,	PUNCT
fcis-30157	446	15	sensitivity	sensitivity	NOUN
fcis-30157	446	16	and	and	CCONJ
fcis-30157	446	17	specificity	specificity	NOUN
fcis-30157	446	18	are	be	AUX
fcis-30157	446	19	key	key	ADJ
fcis-30157	446	20	metrics	metric	NOUN
fcis-30157	446	21	that	that	PRON
fcis-30157	446	22	evaluate	evaluate	VERB
fcis-30157	446	23	a	a	DET
fcis-30157	446	24	system	system	NOUN
fcis-30157	446	25	's	's	PART
fcis-30157	446	26	ability	ability	NOUN
fcis-30157	446	27	to	to	PART
fcis-30157	446	28	correctly	correctly	ADV
fcis-30157	446	29	identify	identify	VERB
fcis-30157	446	30	diseases	disease	NOUN
fcis-30157	446	31	and	and	CCONJ
fcis-30157	446	32	exclude	exclude	VERB
fcis-30157	446	33	non	non	NOUN
fcis-30157	446	34	-	-	NOUN
fcis-30157	446	35	diseases	disease	NOUN
fcis-30157	446	36	,	,	PUNCT
fcis-30157	446	37	respectively	respectively	ADV
fcis-30157	446	38	.	.	PUNCT
fcis-30157	447	1	for	for	ADP
fcis-30157	447	2	skin	skin	NOUN
fcis-30157	447	3	disease	disease	NOUN
fcis-30157	447	4	diagnosis	diagnosis	NOUN
fcis-30157	447	5	,	,	PUNCT
fcis-30157	447	6	the	the	DET
fcis-30157	447	7	isic	isic	PROPN
fcis-30157	447	8	2023	2023	NUM
fcis-30157	447	9	challenge	challenge	NOUN
fcis-30157	447	10	showed	show	VERB
fcis-30157	447	11	that	that	SCONJ
fcis-30157	447	12	a	a	DET
fcis-30157	447	13	deep	deep	ADJ
fcis-30157	447	14	learning	learning	NOUN
fcis-30157	447	15	model	model	NOUN
fcis-30157	447	16	achieved	achieve	VERB
fcis-30157	447	17	89.5	89.5	NUM
fcis-30157	447	18	%	%	NOUN
fcis-30157	447	19	sensitivity	sensitivity	NOUN
fcis-30157	447	20	and	and	CCONJ
fcis-30157	447	21	92.3	92.3	NUM
fcis-30157	447	22	%	%	NOUN
fcis-30157	447	23	specificity	specificity	NOUN
fcis-30157	447	24	for	for	ADP
fcis-30157	447	25	skin	skin	NOUN
fcis-30157	447	26	cancer	cancer	NOUN
fcis-30157	447	27	detection	detection	NOUN
fcis-30157	447	28	,	,	PUNCT
fcis-30157	447	29	indicating	indicate	VERB
fcis-30157	447	30	strong	strong	ADJ
fcis-30157	447	31	performance	performance	NOUN
fcis-30157	447	32	in	in	ADP
fcis-30157	447	33	both	both	DET
fcis-30157	447	34	disease	disease	NOUN
fcis-30157	447	35	identification	identification	NOUN
fcis-30157	447	36	and	and	CCONJ
fcis-30157	447	37	non	non	ADJ
fcis-30157	447	38	-	-	ADJ
fcis-30157	447	39	disease	disease	ADJ
fcis-30157	447	40	exclusion	exclusion	NOUN
fcis-30157	447	41	.	.	PUNCT
fcis-30157	448	1	the	the	DET
fcis-30157	448	2	choice	choice	NOUN
fcis-30157	448	3	of	of	ADP
fcis-30157	448	4	model	model	NOUN
fcis-30157	448	5	parameters	parameter	NOUN
fcis-30157	448	6	and	and	CCONJ
fcis-30157	448	7	structure	structure	NOUN
fcis-30157	448	8	significantly	significantly	ADV
fcis-30157	448	9	impacts	impact	VERB
fcis-30157	448	10	these	these	DET
fcis-30157	448	11	metrics	metric	NOUN
fcis-30157	448	12	.	.	PUNCT
fcis-30157	449	1	for	for	ADP
fcis-30157	449	2	instance	instance	NOUN
fcis-30157	449	3	,	,	PUNCT
fcis-30157	449	4	using	use	VERB
fcis-30157	449	5	multimodal	multimodal	ADJ
fcis-30157	449	6	fusion	fusion	NOUN
fcis-30157	449	7	(	(	PUNCT
fcis-30157	449	8	combining	combine	VERB
fcis-30157	449	9	skin	skin	NOUN
fcis-30157	449	10	images	image	NOUN
fcis-30157	449	11	and	and	CCONJ
fcis-30157	449	12	clinical	clinical	ADJ
fcis-30157	449	13	data	datum	NOUN
fcis-30157	449	14	)	)	PUNCT
fcis-30157	449	15	improved	improve	VERB
fcis-30157	449	16	sensitivity	sensitivity	NOUN
fcis-30157	449	17	to	to	ADP
fcis-30157	449	18	91.2	91.2	NUM
fcis-30157	449	19	%	%	NOUN
fcis-30157	449	20	but	but	CCONJ
fcis-30157	449	21	slightly	slightly	ADV
fcis-30157	449	22	reduced	reduce	VERB
fcis-30157	449	23	specificity	specificity	NOUN
fcis-30157	449	24	to	to	ADP
fcis-30157	449	25	90.8	90.8	NUM
fcis-30157	449	26	%	%	NOUN
fcis-30157	449	27	.	.	PUNCT
fcis-30157	450	1	adjusting	adjust	VERB
fcis-30157	450	2	the	the	DET
fcis-30157	450	3	classification	classification	NOUN
fcis-30157	450	4	threshold	threshold	NOUN
fcis-30157	450	5	can	can	AUX
fcis-30157	450	6	balance	balance	VERB
fcis-30157	450	7	these	these	DET
fcis-30157	450	8	two	two	NUM
fcis-30157	450	9	indicators	indicator	NOUN
fcis-30157	450	10	:	:	PUNCT
fcis-30157	450	11	increasing	increase	VERB
fcis-30157	450	12	the	the	DET
fcis-30157	450	13	threshold	threshold	NOUN
fcis-30157	450	14	from	from	ADP
fcis-30157	450	15	0.5	0.5	NUM
fcis-30157	450	16	to	to	PART
fcis-30157	450	17	0.6	0.6	NUM
fcis-30157	450	18	raised	raise	VERB
fcis-30157	450	19	specificity	specificity	NOUN
fcis-30157	450	20	to	to	ADP
fcis-30157	450	21	94.7	94.7	NUM
fcis-30157	450	22	%	%	NOUN
fcis-30157	450	23	,	,	PUNCT
fcis-30157	450	24	but	but	CCONJ
fcis-30157	450	25	lowered	lower	VERB
fcis-30157	450	26	sensitivity	sensitivity	NOUN
fcis-30157	450	27	to	to	ADP
fcis-30157	450	28	85.3	85.3	NUM
fcis-30157	450	29	%	%	NOUN
fcis-30157	450	30	.	.	PUNCT
fcis-30157	451	1	this	this	DET
fcis-30157	451	2	highlights	highlight	VERB
fcis-30157	451	3	the	the	DET
fcis-30157	451	4	trade	trade	NOUN
fcis-30157	451	5	-	-	PUNCT
fcis-30157	451	6	off	off	NOUN
fcis-30157	451	7	between	between	ADP
fcis-30157	451	8	reducing	reduce	VERB
fcis-30157	451	9	misdiagnoses	misdiagnose	NOUN
fcis-30157	451	10	and	and	CCONJ
fcis-30157	451	11	maintaining	maintain	VERB
fcis-30157	451	12	high	high	ADJ
fcis-30157	451	13	sensitivity	sensitivity	NOUN
fcis-30157	451	14	.	.	PUNCT
fcis-30157	452	1	auxiliary	auxiliary	ADJ
fcis-30157	452	2	task	task	PROPN
fcis-30157	452	3	learning	learning	PROPN
fcis-30157	452	4	,	,	PUNCT
fcis-30157	452	5	such	such	ADJ
fcis-30157	452	6	as	as	ADP
fcis-30157	452	7	predicting	predict	VERB
fcis-30157	452	8	lesion	lesion	NOUN
fcis-30157	452	9	boundaries	boundary	NOUN
fcis-30157	452	10	along	along	ADP
fcis-30157	452	11	with	with	ADP
fcis-30157	452	12	classifications	classification	NOUN
fcis-30157	452	13	,	,	PUNCT
fcis-30157	452	14	improved	improve	VERB
fcis-30157	452	15	both	both	DET
fcis-30157	452	16	sensitivity	sensitivity	NOUN
fcis-30157	452	17	(	(	PUNCT
fcis-30157	452	18	92.1	92.1	NUM
fcis-30157	452	19	%	%	NOUN
fcis-30157	452	20	)	)	PUNCT
fcis-30157	452	21	and	and	CCONJ
fcis-30157	452	22	specificity	specificity	NOUN
fcis-30157	452	23	(	(	PUNCT
fcis-30157	452	24	93.5	93.5	NUM
fcis-30157	452	25	%	%	NOUN
fcis-30157	452	26	)	)	PUNCT
fcis-30157	452	27	.	.	PUNCT
fcis-30157	453	1	additionally	additionally	ADV
fcis-30157	453	2	,	,	PUNCT
fcis-30157	453	3	cnn	cnn	PROPN
fcis-30157	453	4	models	model	NOUN
fcis-30157	453	5	often	often	ADV
fcis-30157	453	6	show	show	VERB
fcis-30157	453	7	higher	high	ADJ
fcis-30157	453	8	sensitivity	sensitivity	NOUN
fcis-30157	453	9	but	but	CCONJ
fcis-30157	453	10	lower	low	ADJ
fcis-30157	453	11	specificity	specificity	NOUN
fcis-30157	453	12	,	,	PUNCT
fcis-30157	453	13	while	while	SCONJ
fcis-30157	453	14	models	model	NOUN
fcis-30157	453	15	using	use	VERB
fcis-30157	453	16	attention	attention	NOUN
fcis-30157	453	17	mechanisms	mechanism	NOUN
fcis-30157	453	18	can	can	AUX
fcis-30157	453	19	improve	improve	VERB
fcis-30157	453	20	specificity	specificity	NOUN
fcis-30157	453	21	while	while	SCONJ
fcis-30157	453	22	maintaining	maintain	VERB
fcis-30157	453	23	high	high	ADJ
fcis-30157	453	24	sensitivity	sensitivity	NOUN
fcis-30157	453	25	,	,	PUNCT
fcis-30157	453	26	reaching	reach	VERB
fcis-30157	453	27	95.2	95.2	NUM
fcis-30157	453	28	%	%	NOUN
fcis-30157	453	29	sensitivity	sensitivity	NOUN
fcis-30157	453	30	.	.	PUNCT
fcis-30157	454	1	this	this	PRON
fcis-30157	454	2	suggests	suggest	VERB
fcis-30157	454	3	the	the	DET
fcis-30157	454	4	importance	importance	NOUN
fcis-30157	454	5	of	of	ADP
fcis-30157	454	6	model	model	NOUN
fcis-30157	454	7	design	design	NOUN
fcis-30157	454	8	in	in	ADP
fcis-30157	454	9	balancing	balance	VERB
fcis-30157	454	10	disease	disease	NOUN
fcis-30157	454	11	identification	identification	NOUN
fcis-30157	454	12	and	and	CCONJ
fcis-30157	454	13	nondisease	nondisease	ADJ
fcis-30157	454	14	exclusion	exclusion	NOUN
fcis-30157	454	15	.	.	PUNCT
fcis-30157	455	1	in	in	ADP
fcis-30157	455	2	practical	practical	ADJ
fcis-30157	455	3	settings	setting	NOUN
fcis-30157	455	4	,	,	PUNCT
fcis-30157	455	5	a	a	DET
fcis-30157	455	6	dermatology	dermatology	NOUN
fcis-30157	455	7	hospital	hospital	NOUN
fcis-30157	455	8	saw	see	VERB
fcis-30157	455	9	a	a	DET
fcis-30157	455	10	20	20	NUM
fcis-30157	455	11	%	%	NOUN
fcis-30157	455	12	decrease	decrease	NOUN
fcis-30157	455	13	in	in	ADP
fcis-30157	455	14	misdiagnoses	misdiagnose	NOUN
fcis-30157	455	15	but	but	CCONJ
fcis-30157	455	16	a	a	DET
fcis-30157	455	17	5	5	NUM
fcis-30157	455	18	%	%	NOUN
fcis-30157	455	19	increase	increase	NOUN
fcis-30157	455	20	in	in	ADP
fcis-30157	455	21	false	false	ADJ
fcis-30157	455	22	positives	positive	NOUN
fcis-30157	455	23	after	after	ADP
fcis-30157	455	24	implementing	implement	VERB
fcis-30157	455	25	deep	deep	ADJ
fcis-30157	455	26	learning	learning	NOUN
fcis-30157	455	27	systems	system	NOUN
fcis-30157	455	28	.	.	PUNCT
fcis-30157	456	1	this	this	PRON
fcis-30157	456	2	indicates	indicate	VERB
fcis-30157	456	3	that	that	SCONJ
fcis-30157	456	4	doctors	doctor	NOUN
fcis-30157	456	5	should	should	AUX
fcis-30157	456	6	combine	combine	VERB
fcis-30157	456	7	system	system	NOUN
fcis-30157	456	8	recommendations	recommendation	NOUN
fcis-30157	456	9	with	with	ADP
fcis-30157	456	10	clinical	clinical	ADJ
fcis-30157	456	11	judgment	judgment	NOUN
fcis-30157	456	12	to	to	PART
fcis-30157	456	13	reduce	reduce	VERB
fcis-30157	456	14	missed	missed	ADJ
fcis-30157	456	15	diagnoses	diagnosis	NOUN
fcis-30157	456	16	.	.	PUNCT
fcis-30157	457	1	similarly	similarly	ADV
fcis-30157	457	2	,	,	PUNCT
fcis-30157	457	3	in	in	ADP
fcis-30157	457	4	primary	primary	ADJ
fcis-30157	457	5	care	care	NOUN
fcis-30157	457	6	settings	setting	NOUN
fcis-30157	457	7	,	,	PUNCT
fcis-30157	457	8	adjusting	adjust	VERB
fcis-30157	457	9	model	model	NOUN
fcis-30157	457	10	parameters	parameter	NOUN
fcis-30157	457	11	boosted	boost	VERB
fcis-30157	457	12	both	both	DET
fcis-30157	457	13	sensitivity	sensitivity	NOUN
fcis-30157	457	14	(	(	PUNCT
fcis-30157	457	15	90.5	90.5	NUM
fcis-30157	457	16	%	%	NOUN
fcis-30157	457	17	)	)	PUNCT
fcis-30157	457	18	and	and	CCONJ
fcis-30157	457	19	specificity	specificity	NOUN
fcis-30157	457	20	(	(	PUNCT
fcis-30157	457	21	93.8	93.8	NUM
fcis-30157	457	22	%	%	NOUN
fcis-30157	457	23	)	)	PUNCT
fcis-30157	457	24	,	,	PUNCT
fcis-30157	457	25	improving	improve	VERB
fcis-30157	457	26	system	system	NOUN
fcis-30157	457	27	practicality	practicality	NOUN
fcis-30157	457	28	.	.	PUNCT
fcis-30157	458	1	experiments	experiment	NOUN
fcis-30157	458	2	showed	show	VERB
fcis-30157	458	3	significant	significant	ADJ
fcis-30157	458	4	variation	variation	NOUN
fcis-30157	458	5	in	in	ADP
fcis-30157	458	6	performance	performance	NOUN
fcis-30157	458	7	across	across	ADP
fcis-30157	458	8	datasets	dataset	NOUN
fcis-30157	458	9	.	.	PUNCT
fcis-30157	459	1	for	for	ADP
fcis-30157	459	2	multiple	multiple	ADJ
fcis-30157	459	3	skin	skin	NOUN
fcis-30157	459	4	diseases	disease	NOUN
fcis-30157	459	5	,	,	PUNCT
fcis-30157	459	6	a	a	DET
fcis-30157	459	7	cnn	cnn	PROPN
fcis-30157	459	8	model	model	NOUN
fcis-30157	459	9	achieved	achieve	VERB
fcis-30157	459	10	88.7	88.7	NUM
fcis-30157	459	11	%	%	NOUN
fcis-30157	459	12	sensitivity	sensitivity	NOUN
fcis-30157	459	13	and	and	CCONJ
fcis-30157	459	14	91.2	91.2	NUM
fcis-30157	459	15	%	%	NOUN
fcis-30157	459	16	specificity	specificity	NOUN
fcis-30157	459	17	,	,	PUNCT
fcis-30157	459	18	while	while	SCONJ
fcis-30157	459	19	for	for	ADP
fcis-30157	459	20	skin	skin	NOUN
fcis-30157	459	21	cancer	cancer	NOUN
fcis-30157	459	22	,	,	PUNCT
fcis-30157	459	23	sensitivity	sensitivity	NOUN
fcis-30157	459	24	rose	rise	VERB
fcis-30157	459	25	to	to	ADP
fcis-30157	459	26	92.3	92.3	NUM
fcis-30157	459	27	%	%	NOUN
fcis-30157	459	28	,	,	PUNCT
fcis-30157	459	29	with	with	ADP
fcis-30157	459	30	a	a	DET
fcis-30157	459	31	slight	slight	ADJ
fcis-30157	459	32	drop	drop	NOUN
fcis-30157	459	33	in	in	ADP
fcis-30157	459	34	specificity	specificity	NOUN
fcis-30157	459	35	to	to	ADP
fcis-30157	459	36	90.5	90.5	NUM
fcis-30157	459	37	%	%	NOUN
fcis-30157	459	38	.	.	PUNCT
fcis-30157	460	1	this	this	PRON
fcis-30157	460	2	demonstrates	demonstrate	VERB
fcis-30157	460	3	the	the	DET
fcis-30157	460	4	impact	impact	NOUN
fcis-30157	460	5	of	of	ADP
fcis-30157	460	6	dataset	dataset	ADJ
fcis-30157	460	7	diversity	diversity	NOUN
fcis-30157	460	8	and	and	CCONJ
fcis-30157	460	9	sample	sample	NOUN
fcis-30157	460	10	distribution	distribution	NOUN
fcis-30157	460	11	on	on	ADP
fcis-30157	460	12	model	model	NOUN
fcis-30157	460	13	performance	performance	NOUN
fcis-30157	460	14	.	.	PUNCT
fcis-30157	461	1	table	table	NOUN
fcis-30157	461	2	5	5	NUM
fcis-30157	461	3	.	.	PUNCT
fcis-30157	462	1	summary	summary	NOUN
fcis-30157	462	2	of	of	ADP
fcis-30157	462	3	model	model	NOUN
fcis-30157	462	4	performance	performance	NOUN
fcis-30157	462	5	model	model	NOUN
fcis-30157	462	6	type	type	NOUN
fcis-30157	462	7	sensitivity	sensitivity	NOUN
fcis-30157	462	8	(	(	PUNCT
fcis-30157	462	9	%	%	INTJ
fcis-30157	462	10	)	)	PUNCT
fcis-30157	462	11	specificity	specificity	NOUN
fcis-30157	462	12	(	(	PUNCT
fcis-30157	462	13	%	%	INTJ
fcis-30157	462	14	)	)	PUNCT
fcis-30157	462	15	dataset	dataset	NOUN
fcis-30157	462	16	type	type	NOUN
fcis-30157	462	17	cnn	cnn	PROPN
fcis-30157	462	18	88.7	88.7	NUM
fcis-30157	462	19	91.2	91.2	NUM
fcis-30157	462	20	multiple	multiple	ADJ
fcis-30157	462	21	diseases	disease	NOUN
fcis-30157	462	22	cnn	cnn	NOUN
fcis-30157	462	23	+	+	CCONJ
fcis-30157	462	24	attention	attention	NOUN
fcis-30157	462	25	92.3	92.3	NUM
fcis-30157	462	26	90.5	90.5	NUM
fcis-30157	462	27	skin	skin	NOUN
fcis-30157	462	28	cancer	cancer	NOUN
fcis-30157	462	29	multimodal	multimodal	NOUN
fcis-30157	462	30	fusion	fusion	NOUN
fcis-30157	462	31	91.2	91.2	NUM
fcis-30157	462	32	90.8	90.8	NUM
fcis-30157	462	33	multiple	multiple	ADJ
fcis-30157	462	34	diseases	disease	NOUN
fcis-30157	462	35	assisted	assist	VERB
fcis-30157	462	36	task	task	NOUN
fcis-30157	462	37	learning	learn	VERB
fcis-30157	462	38	92.1	92.1	NUM
fcis-30157	462	39	93.5	93.5	NUM
fcis-30157	462	40	multiple	multiple	ADJ
fcis-30157	462	41	diseases	disease	NOUN
fcis-30157	462	42	balancing	balance	VERB
fcis-30157	462	43	sensitivity	sensitivity	NOUN
fcis-30157	462	44	and	and	CCONJ
fcis-30157	462	45	specificity	specificity	NOUN
fcis-30157	462	46	are	be	AUX
fcis-30157	462	47	crucial	crucial	ADJ
fcis-30157	462	48	for	for	ADP
fcis-30157	462	49	accurate	accurate	ADJ
fcis-30157	462	50	diagnostic	diagnostic	ADJ
fcis-30157	462	51	performance	performance	NOUN
fcis-30157	462	52	.	.	PUNCT
fcis-30157	463	1	by	by	ADP
fcis-30157	463	2	optimizing	optimize	VERB
fcis-30157	463	3	model	model	NOUN
fcis-30157	463	4	structures	structure	NOUN
fcis-30157	463	5	,	,	PUNCT
fcis-30157	463	6	adjusting	adjust	VERB
fcis-30157	463	7	parameters	parameter	NOUN
fcis-30157	463	8	,	,	PUNCT
fcis-30157	463	9	and	and	CCONJ
fcis-30157	463	10	incorporating	incorporate	VERB
fcis-30157	463	11	multimodal	multimodal	NOUN
fcis-30157	463	12	data	datum	NOUN
fcis-30157	463	13	and	and	CCONJ
fcis-30157	463	14	auxiliary	auxiliary	ADJ
fcis-30157	463	15	task	task	PROPN
fcis-30157	463	16	learning	learning	PROPN
fcis-30157	463	17	,	,	PUNCT
fcis-30157	463	18	systems	system	NOUN
fcis-30157	463	19	can	can	AUX
fcis-30157	463	20	be	be	AUX
fcis-30157	463	21	tailored	tailor	VERB
fcis-30157	463	22	to	to	PART
fcis-30157	463	23	achieve	achieve	VERB
fcis-30157	463	24	better	well	ADJ
fcis-30157	463	25	diagnostic	diagnostic	ADJ
fcis-30157	463	26	outcomes	outcome	NOUN
fcis-30157	463	27	.	.	PUNCT
fcis-30157	464	1	the	the	DET
fcis-30157	464	2	configuration	configuration	NOUN
fcis-30157	464	3	should	should	AUX
fcis-30157	464	4	be	be	AUX
fcis-30157	464	5	adapted	adapt	VERB
fcis-30157	464	6	based	base	VERB
fcis-30157	464	7	on	on	ADP
fcis-30157	464	8	specific	specific	ADJ
fcis-30157	464	9	clinical	clinical	ADJ
fcis-30157	464	10	needs	need	NOUN
fcis-30157	464	11	and	and	CCONJ
fcis-30157	464	12	scenarios	scenario	NOUN
fcis-30157	464	13	.	.	PUNCT
fcis-30157	465	1	3.3.2	3.3.2	X
fcis-30157	465	2	.	.	X
fcis-30157	465	3	impact	impact	NOUN
fcis-30157	465	4	of	of	ADP
fcis-30157	465	5	different	different	ADJ
fcis-30157	465	6	factors	factor	NOUN
fcis-30157	465	7	on	on	ADP
fcis-30157	465	8	diagnosis	diagnosis	NOUN
fcis-30157	465	9	in	in	ADP
fcis-30157	465	10	auxiliary	auxiliary	ADJ
fcis-30157	465	11	diagnostic	diagnostic	ADJ
fcis-30157	465	12	systems	system	NOUN
fcis-30157	465	13	,	,	PUNCT
fcis-30157	465	14	factors	factor	NOUN
fcis-30157	465	15	like	like	ADP
fcis-30157	465	16	image	image	NOUN
fcis-30157	465	17	quality	quality	NOUN
fcis-30157	465	18	,	,	PUNCT
fcis-30157	465	19	patient	patient	ADJ
fcis-30157	465	20	information	information	NOUN
fcis-30157	465	21	,	,	PUNCT
fcis-30157	465	22	and	and	CCONJ
fcis-30157	465	23	identification	identification	NOUN
fcis-30157	465	24	accuracy	accuracy	NOUN
fcis-30157	465	25	significantly	significantly	ADV
fcis-30157	465	26	influence	influence	VERB
fcis-30157	465	27	diagnosis	diagnosis	NOUN
fcis-30157	465	28	results	result	NOUN
fcis-30157	465	29	.	.	PUNCT
fcis-30157	466	1	image	image	NOUN
fcis-30157	466	2	quality	quality	NOUN
fcis-30157	466	3	high	high	ADJ
fcis-30157	466	4	-	-	PUNCT
fcis-30157	466	5	quality	quality	NOUN
fcis-30157	466	6	images	image	NOUN
fcis-30157	466	7	,	,	PUNCT
fcis-30157	466	8	with	with	ADP
fcis-30157	466	9	clear	clear	ADJ
fcis-30157	466	10	details	detail	NOUN
fcis-30157	466	11	,	,	PUNCT
fcis-30157	466	12	enhance	enhance	VERB
fcis-30157	466	13	the	the	DET
fcis-30157	466	14	model	model	NOUN
fcis-30157	466	15	’s	’s	PART
fcis-30157	466	16	ability	ability	NOUN
fcis-30157	466	17	to	to	PART
fcis-30157	466	18	extract	extract	VERB
fcis-30157	466	19	features	feature	NOUN
fcis-30157	466	20	and	and	CCONJ
fcis-30157	466	21	classify	classify	VERB
fcis-30157	466	22	them	they	PRON
fcis-30157	466	23	accurately	accurately	ADV
fcis-30157	466	24	.	.	PUNCT
fcis-30157	467	1	studies	study	NOUN
fcis-30157	467	2	show	show	VERB
fcis-30157	467	3	that	that	SCONJ
fcis-30157	467	4	low	low	ADJ
fcis-30157	467	5	-	-	PUNCT
fcis-30157	467	6	resolution	resolution	NOUN
fcis-30157	467	7	images	image	NOUN
fcis-30157	467	8	reduce	reduce	VERB
fcis-30157	467	9	model	model	NOUN
fcis-30157	467	10	accuracy	accuracy	NOUN
fcis-30157	467	11	by	by	ADP
fcis-30157	467	12	around	around	ADP
fcis-30157	467	13	15	15	NUM
fcis-30157	467	14	%	%	NOUN
fcis-30157	467	15	compared	compare	VERB
fcis-30157	467	16	to	to	ADP
fcis-30157	467	17	high	high	ADJ
fcis-30157	467	18	-	-	PUNCT
fcis-30157	467	19	resolution	resolution	NOUN
fcis-30157	467	20	images	image	NOUN
fcis-30157	467	21	.	.	PUNCT
fcis-30157	468	1	lighting	lighting	NOUN
fcis-30157	468	2	,	,	PUNCT
fcis-30157	468	3	shooting	shooting	NOUN
fcis-30157	468	4	angle	angle	NOUN
fcis-30157	468	5	,	,	PUNCT
fcis-30157	468	6	and	and	CCONJ
fcis-30157	468	7	clarity	clarity	NOUN
fcis-30157	468	8	also	also	ADV
fcis-30157	468	9	affect	affect	VERB
fcis-30157	468	10	results	result	NOUN
fcis-30157	468	11	,	,	PUNCT
fcis-30157	468	12	especially	especially	ADV
fcis-30157	468	13	when	when	SCONJ
fcis-30157	468	14	the	the	DET
fcis-30157	468	15	lesion	lesion	NOUN
fcis-30157	468	16	blends	blend	VERB
fcis-30157	468	17	with	with	ADP
fcis-30157	468	18	surrounding	surround	VERB
fcis-30157	468	19	skin	skin	NOUN
fcis-30157	468	20	.	.	PUNCT
fcis-30157	469	1	optimizing	optimize	VERB
fcis-30157	469	2	image	image	NOUN
fcis-30157	469	3	acquisition	acquisition	NOUN
fcis-30157	469	4	methods	method	NOUN
fcis-30157	469	5	,	,	PUNCT
fcis-30157	469	6	such	such	ADJ
fcis-30157	469	7	as	as	ADP
fcis-30157	469	8	using	use	VERB
fcis-30157	469	9	highresolution	highresolution	NOUN
fcis-30157	469	10	equipment	equipment	NOUN
fcis-30157	469	11	and	and	CCONJ
fcis-30157	469	12	consistent	consistent	ADJ
fcis-30157	469	13	lighting	lighting	NOUN
fcis-30157	469	14	,	,	PUNCT
fcis-30157	469	15	can	can	AUX
fcis-30157	469	16	improve	improve	VERB
fcis-30157	469	17	diagnostic	diagnostic	ADJ
fcis-30157	469	18	performance	performance	NOUN
fcis-30157	469	19	.	.	PUNCT
fcis-30157	470	1	patient	patient	ADJ
fcis-30157	470	2	information	information	NOUN
fcis-30157	470	3	clinical	clinical	ADJ
fcis-30157	470	4	data	datum	NOUN
fcis-30157	470	5	like	like	ADP
fcis-30157	470	6	age	age	NOUN
fcis-30157	470	7	,	,	PUNCT
fcis-30157	470	8	gender	gender	NOUN
fcis-30157	470	9	,	,	PUNCT
fcis-30157	470	10	medical	medical	ADJ
fcis-30157	470	11	history	history	NOUN
fcis-30157	470	12	,	,	PUNCT
fcis-30157	470	13	and	and	CCONJ
fcis-30157	470	14	genetic	genetic	ADJ
fcis-30157	470	15	information	information	NOUN
fcis-30157	470	16	are	be	AUX
fcis-30157	470	17	crucial	crucial	ADJ
fcis-30157	470	18	for	for	ADP
fcis-30157	470	19	accurate	accurate	ADJ
fcis-30157	470	20	diagnosis	diagnosis	NOUN
fcis-30157	470	21	.	.	PUNCT
fcis-30157	471	1	for	for	ADP
fcis-30157	471	2	example	example	NOUN
fcis-30157	471	3	,	,	PUNCT
fcis-30157	471	4	psoriasis	psoriasis	NOUN
fcis-30157	471	5	is	be	AUX
fcis-30157	471	6	more	more	ADV
fcis-30157	471	7	common	common	ADJ
fcis-30157	471	8	in	in	ADP
fcis-30157	471	9	people	people	NOUN
fcis-30157	471	10	aged	age	VERB
fcis-30157	471	11	30	30	NUM
fcis-30157	471	12	-	-	SYM
fcis-30157	471	13	50	50	NUM
fcis-30157	471	14	,	,	PUNCT
fcis-30157	471	15	and	and	CCONJ
fcis-30157	471	16	acne	acne	NOUN
fcis-30157	471	17	in	in	ADP
fcis-30157	471	18	adolescents	adolescent	NOUN
fcis-30157	471	19	.	.	PUNCT
fcis-30157	472	1	integrating	integrate	VERB
fcis-30157	472	2	this	this	DET
fcis-30157	472	3	information	information	NOUN
fcis-30157	472	4	with	with	ADP
fcis-30157	472	5	image	image	NOUN
fcis-30157	472	6	data	datum	NOUN
fcis-30157	472	7	can	can	AUX
fcis-30157	472	8	improve	improve	VERB
fcis-30157	472	9	model	model	NOUN
fcis-30157	472	10	accuracy	accuracy	NOUN
fcis-30157	472	11	.	.	PUNCT
fcis-30157	473	1	studies	study	NOUN
fcis-30157	473	2	show	show	VERB
fcis-30157	473	3	models	model	NOUN
fcis-30157	473	4	using	use	VERB
fcis-30157	473	5	both	both	CCONJ
fcis-30157	473	6	clinical	clinical	ADJ
fcis-30157	473	7	and	and	CCONJ
fcis-30157	473	8	image	image	NOUN
fcis-30157	473	9	data	datum	NOUN
fcis-30157	473	10	achieve	achieve	VERB
fcis-30157	473	11	higher	high	ADJ
fcis-30157	473	12	accuracy	accuracy	NOUN
fcis-30157	473	13	(	(	PUNCT
fcis-30157	473	14	92.3	92.3	NUM
fcis-30157	473	15	%	%	NOUN
fcis-30157	473	16	)	)	PUNCT
fcis-30157	473	17	compared	compare	VERB
fcis-30157	473	18	to	to	ADP
fcis-30157	473	19	those	those	PRON
fcis-30157	473	20	relying	rely	VERB
fcis-30157	473	21	on	on	ADP
fcis-30157	473	22	images	image	NOUN
fcis-30157	473	23	alone	alone	ADV
fcis-30157	473	24	(	(	PUNCT
fcis-30157	473	25	85.7	85.7	NUM
fcis-30157	473	26	%	%	NOUN
fcis-30157	473	27	)	)	PUNCT
fcis-30157	473	28	.	.	PUNCT
fcis-30157	474	1	this	this	DET
fcis-30157	474	2	integration	integration	NOUN
fcis-30157	474	3	helps	help	VERB
fcis-30157	474	4	the	the	DET
fcis-30157	474	5	model	model	NOUN
fcis-30157	474	6	better	well	ADV
fcis-30157	474	7	evaluate	evaluate	VERB
fcis-30157	474	8	the	the	DET
fcis-30157	474	9	patient	patient	NOUN
fcis-30157	474	10	’s	’s	PART
fcis-30157	474	11	condition	condition	NOUN
fcis-30157	474	12	.	.	PUNCT
fcis-30157	475	1	identification	identification	NOUN
fcis-30157	475	2	accuracy	accuracy	NOUN
fcis-30157	475	3	the	the	DET
fcis-30157	475	4	model	model	NOUN
fcis-30157	475	5	's	's	PART
fcis-30157	475	6	identification	identification	NOUN
fcis-30157	475	7	accuracy	accuracy	NOUN
fcis-30157	475	8	directly	directly	ADV
fcis-30157	475	9	affects	affect	VERB
fcis-30157	475	10	the	the	DET
fcis-30157	475	11	diagnostic	diagnostic	ADJ
fcis-30157	475	12	system	system	NOUN
fcis-30157	475	13	’s	’s	PART
fcis-30157	475	14	reliability	reliability	NOUN
fcis-30157	475	15	.	.	PUNCT
fcis-30157	476	1	a	a	DET
fcis-30157	476	2	1	1	NUM
fcis-30157	476	3	%	%	NOUN
fcis-30157	476	4	improvement	improvement	NOUN
fcis-30157	476	5	in	in	ADP
fcis-30157	476	6	model	model	NOUN
fcis-30157	476	7	accuracy	accuracy	NOUN
fcis-30157	476	8	reduces	reduce	VERB
fcis-30157	476	9	the	the	DET
fcis-30157	476	10	misdiagnosis	misdiagnosis	NOUN
fcis-30157	476	11	rate	rate	NOUN
fcis-30157	476	12	by	by	ADP
fcis-30157	476	13	2	2	NUM
fcis-30157	476	14	%	%	NOUN
fcis-30157	476	15	.	.	PUNCT
fcis-30157	477	1	factors	factor	NOUN
fcis-30157	477	2	such	such	ADJ
fcis-30157	477	3	as	as	ADP
fcis-30157	477	4	dataset	dataset	ADJ
fcis-30157	477	5	diversity	diversity	NOUN
fcis-30157	477	6	,	,	PUNCT
fcis-30157	477	7	model	model	NOUN
fcis-30157	477	8	complexity	complexity	NOUN
fcis-30157	477	9	,	,	PUNCT
fcis-30157	477	10	and	and	CCONJ
fcis-30157	477	11	adequate	adequate	ADJ
fcis-30157	477	12	training	training	NOUN
fcis-30157	477	13	data	datum	NOUN
fcis-30157	477	14	influence	influence	NOUN
fcis-30157	477	15	performance	performance	NOUN
fcis-30157	477	16	.	.	PUNCT
fcis-30157	478	1	for	for	ADP
fcis-30157	478	2	instance	instance	NOUN
fcis-30157	478	3	,	,	PUNCT
fcis-30157	478	4	models	model	NOUN
fcis-30157	478	5	using	use	VERB
fcis-30157	478	6	multimodal	multimodal	ADJ
fcis-30157	478	7	data	datum	NOUN
fcis-30157	478	8	(	(	PUNCT
fcis-30157	478	9	images	image	NOUN
fcis-30157	478	10	and	and	CCONJ
fcis-30157	478	11	clinical	clinical	ADJ
fcis-30157	478	12	data	datum	NOUN
fcis-30157	478	13	)	)	PUNCT
fcis-30157	478	14	perform	perform	VERB
fcis-30157	478	15	better	well	ADV
fcis-30157	478	16	,	,	PUNCT
fcis-30157	478	17	achieving	achieve	VERB
fcis-30157	478	18	92.3	92.3	NUM
fcis-30157	478	19	%	%	NOUN
fcis-30157	478	20	accuracy	accuracy	NOUN
fcis-30157	478	21	compared	compare	VERB
fcis-30157	478	22	to	to	ADP
fcis-30157	478	23	image	image	NOUN
fcis-30157	478	24	-	-	PUNCT
fcis-30157	478	25	only	only	ADJ
fcis-30157	478	26	models	model	NOUN
fcis-30157	478	27	122	122	NUM
fcis-30157	478	28	at	at	ADP
fcis-30157	478	29	85.7	85.7	NUM
fcis-30157	478	30	%	%	NOUN
fcis-30157	478	31	.	.	PUNCT
fcis-30157	479	1	additionally	additionally	ADV
fcis-30157	479	2	,	,	PUNCT
fcis-30157	479	3	models	model	NOUN
fcis-30157	479	4	should	should	AUX
fcis-30157	479	5	be	be	AUX
fcis-30157	479	6	adaptable	adaptable	ADJ
fcis-30157	479	7	to	to	ADP
fcis-30157	479	8	various	various	ADJ
fcis-30157	479	9	skin	skin	NOUN
fcis-30157	479	10	colors	color	NOUN
fcis-30157	479	11	,	,	PUNCT
fcis-30157	479	12	as	as	SCONJ
fcis-30157	479	13	shown	show	VERB
fcis-30157	479	14	in	in	ADP
fcis-30157	479	15	studies	study	NOUN
fcis-30157	479	16	where	where	SCONJ
fcis-30157	479	17	models	model	NOUN
fcis-30157	479	18	performed	perform	VERB
fcis-30157	479	19	better	well	ADV
fcis-30157	479	20	on	on	ADP
fcis-30157	479	21	light	light	ADJ
fcis-30157	479	22	-	-	PUNCT
fcis-30157	479	23	skinned	skin	VERB
fcis-30157	479	24	patients	patient	NOUN
fcis-30157	479	25	compared	compare	VERB
fcis-30157	479	26	to	to	ADP
fcis-30157	479	27	dark	dark	ADJ
fcis-30157	479	28	-	-	PUNCT
fcis-30157	479	29	skinned	skinned	ADJ
fcis-30157	479	30	ones	one	NOUN
fcis-30157	479	31	.	.	PUNCT
fcis-30157	480	1	optimization	optimization	NOUN
fcis-30157	480	2	strategies	strategy	NOUN
fcis-30157	480	3	and	and	CCONJ
fcis-30157	480	4	challenges	challenge	NOUN
fcis-30157	480	5	based	base	VERB
fcis-30157	480	6	on	on	ADP
fcis-30157	480	7	the	the	DET
fcis-30157	480	8	analysis	analysis	NOUN
fcis-30157	480	9	,	,	PUNCT
fcis-30157	480	10	optimization	optimization	NOUN
fcis-30157	480	11	strategies	strategy	NOUN
fcis-30157	480	12	include	include	VERB
fcis-30157	480	13	:	:	PUNCT
fcis-30157	480	14	image	image	NOUN
fcis-30157	480	15	quality	quality	NOUN
fcis-30157	480	16	:	:	PUNCT
fcis-30157	480	17	use	use	VERB
fcis-30157	480	18	high	high	ADJ
fcis-30157	480	19	-	-	PUNCT
fcis-30157	480	20	resolution	resolution	NOUN
fcis-30157	480	21	equipment	equipment	NOUN
fcis-30157	480	22	and	and	CCONJ
fcis-30157	480	23	ensure	ensure	VERB
fcis-30157	480	24	consistent	consistent	ADJ
fcis-30157	480	25	lighting	lighting	NOUN
fcis-30157	480	26	.	.	PUNCT
fcis-30157	481	1	patient	patient	ADJ
fcis-30157	481	2	information	information	NOUN
fcis-30157	481	3	:	:	PUNCT
fcis-30157	481	4	integrate	integrate	VERB
fcis-30157	481	5	clinical	clinical	ADJ
fcis-30157	481	6	data	datum	NOUN
fcis-30157	481	7	and	and	CCONJ
fcis-30157	481	8	build	build	VERB
fcis-30157	481	9	multimodal	multimodal	NOUN
fcis-30157	481	10	datasets	dataset	NOUN
fcis-30157	481	11	to	to	PART
fcis-30157	481	12	enhance	enhance	VERB
fcis-30157	481	13	model	model	NOUN
fcis-30157	481	14	analysis	analysis	NOUN
fcis-30157	481	15	.	.	PUNCT
fcis-30157	482	1	identification	identification	NOUN
fcis-30157	482	2	accuracy	accuracy	NOUN
fcis-30157	482	3	:	:	PUNCT
fcis-30157	482	4	use	use	VERB
fcis-30157	482	5	diverse	diverse	ADJ
fcis-30157	482	6	training	training	NOUN
fcis-30157	482	7	data	datum	NOUN
fcis-30157	482	8	to	to	PART
fcis-30157	482	9	improve	improve	VERB
fcis-30157	482	10	model	model	NOUN
fcis-30157	482	11	generalization	generalization	NOUN
fcis-30157	482	12	and	and	CCONJ
fcis-30157	482	13	reduce	reduce	VERB
fcis-30157	482	14	bias	bias	NOUN
fcis-30157	482	15	.	.	PUNCT
fcis-30157	483	1	these	these	DET
fcis-30157	483	2	strategies	strategy	NOUN
fcis-30157	483	3	significantly	significantly	ADV
fcis-30157	483	4	improve	improve	VERB
fcis-30157	483	5	system	system	NOUN
fcis-30157	483	6	accuracy	accuracy	NOUN
fcis-30157	483	7	,	,	PUNCT
fcis-30157	483	8	as	as	SCONJ
fcis-30157	483	9	seen	see	VERB
fcis-30157	483	10	in	in	ADP
fcis-30157	483	11	studies	study	NOUN
fcis-30157	483	12	where	where	SCONJ
fcis-30157	483	13	multimodal	multimodal	NOUN
fcis-30157	483	14	models	model	NOUN
fcis-30157	483	15	achieved	achieve	VERB
fcis-30157	483	16	94.5	94.5	NUM
fcis-30157	483	17	%	%	NOUN
fcis-30157	483	18	accuracy	accuracy	NOUN
fcis-30157	483	19	.	.	PUNCT
fcis-30157	484	1	in	in	ADP
fcis-30157	484	2	practical	practical	ADJ
fcis-30157	484	3	applications	application	NOUN
fcis-30157	484	4	,	,	PUNCT
fcis-30157	484	5	optimized	optimize	VERB
fcis-30157	484	6	systems	system	NOUN
fcis-30157	484	7	have	have	AUX
fcis-30157	484	8	reduced	reduce	VERB
fcis-30157	484	9	skin	skin	NOUN
fcis-30157	484	10	cancer	cancer	NOUN
fcis-30157	484	11	misdiagnosis	misdiagnosis	NOUN
fcis-30157	484	12	by	by	ADP
fcis-30157	484	13	12	12	NUM
fcis-30157	484	14	%	%	NOUN
fcis-30157	484	15	.	.	PUNCT
fcis-30157	485	1	however	however	ADV
fcis-30157	485	2	,	,	PUNCT
fcis-30157	485	3	challenges	challenge	NOUN
fcis-30157	485	4	like	like	ADP
fcis-30157	485	5	data	datum	NOUN
fcis-30157	485	6	privacy	privacy	NOUN
fcis-30157	485	7	,	,	PUNCT
fcis-30157	485	8	incomplete	incomplete	ADJ
fcis-30157	485	9	data	datum	NOUN
fcis-30157	485	10	,	,	PUNCT
fcis-30157	485	11	and	and	CCONJ
fcis-30157	485	12	standardization	standardization	NOUN
fcis-30157	485	13	across	across	ADP
fcis-30157	485	14	institutions	institution	NOUN
fcis-30157	485	15	need	need	VERB
fcis-30157	485	16	to	to	PART
fcis-30157	485	17	be	be	AUX
fcis-30157	485	18	addressed	address	VERB
fcis-30157	485	19	for	for	ADP
fcis-30157	485	20	broader	broad	ADJ
fcis-30157	485	21	system	system	NOUN
fcis-30157	485	22	deployment	deployment	NOUN
fcis-30157	485	23	.	.	PUNCT
fcis-30157	486	1	table	table	NOUN
fcis-30157	486	2	6	6	NUM
fcis-30157	486	3	.	.	PUNCT
fcis-30157	487	1	summary	summary	NOUN
fcis-30157	487	2	of	of	ADP
fcis-30157	487	3	key	key	ADJ
fcis-30157	487	4	factors	factor	NOUN
fcis-30157	487	5	and	and	CCONJ
fcis-30157	487	6	strategies	strategy	NOUN
fcis-30157	487	7	factor	factor	NOUN
fcis-30157	487	8	impact	impact	NOUN
fcis-30157	487	9	optimization	optimization	NOUN
fcis-30157	487	10	strategy	strategy	NOUN
fcis-30157	487	11	challenges	challenge	VERB
fcis-30157	487	12	image	image	NOUN
fcis-30157	487	13	quality	quality	NOUN
fcis-30157	487	14	high	high	ADJ
fcis-30157	487	15	-	-	PUNCT
fcis-30157	487	16	quality	quality	NOUN
fcis-30157	487	17	images	image	NOUN
fcis-30157	487	18	enhance	enhance	VERB
fcis-30157	487	19	feature	feature	NOUN
fcis-30157	487	20	extraction	extraction	NOUN
fcis-30157	487	21	.	.	PUNCT
fcis-30157	488	1	low	low	ADJ
fcis-30157	488	2	-	-	PUNCT
fcis-30157	488	3	quality	quality	NOUN
fcis-30157	488	4	images	image	NOUN
fcis-30157	488	5	lead	lead	VERB
fcis-30157	488	6	to	to	ADP
fcis-30157	488	7	increased	increase	VERB
fcis-30157	488	8	error	error	NOUN
fcis-30157	488	9	rates	rate	NOUN
fcis-30157	488	10	.	.	PUNCT
fcis-30157	489	1	use	use	VERB
fcis-30157	489	2	high	high	ADJ
fcis-30157	489	3	-	-	PUNCT
fcis-30157	489	4	resolution	resolution	NOUN
fcis-30157	489	5	equipment	equipment	NOUN
fcis-30157	489	6	and	and	CCONJ
fcis-30157	489	7	consistent	consistent	ADJ
fcis-30157	489	8	lighting	lighting	NOUN
fcis-30157	489	9	.	.	PUNCT
fcis-30157	490	1	privacy	privacy	NOUN
fcis-30157	490	2	,	,	PUNCT
fcis-30157	490	3	data	datum	NOUN
fcis-30157	490	4	standards	standard	NOUN
fcis-30157	490	5	,	,	PUNCT
fcis-30157	490	6	and	and	CCONJ
fcis-30157	490	7	system	system	NOUN
fcis-30157	490	8	adaptability	adaptability	NOUN
fcis-30157	490	9	across	across	ADP
fcis-30157	490	10	institutions	institution	NOUN
fcis-30157	490	11	.	.	PUNCT
fcis-30157	491	1	patient	patient	ADJ
fcis-30157	491	2	info	info	NOUN
fcis-30157	491	3	clinical	clinical	ADJ
fcis-30157	491	4	data	datum	NOUN
fcis-30157	491	5	provides	provide	VERB
fcis-30157	491	6	context	context	NOUN
fcis-30157	491	7	to	to	PART
fcis-30157	491	8	improve	improve	VERB
fcis-30157	491	9	diagnosis	diagnosis	NOUN
fcis-30157	491	10	accuracy	accuracy	NOUN
fcis-30157	491	11	.	.	PUNCT
fcis-30157	492	1	integrate	integrate	VERB
fcis-30157	492	2	clinical	clinical	ADJ
fcis-30157	492	3	data	datum	NOUN
fcis-30157	492	4	into	into	ADP
fcis-30157	492	5	multimodal	multimodal	ADJ
fcis-30157	492	6	datasets	dataset	NOUN
fcis-30157	492	7	.	.	PUNCT
fcis-30157	493	1	handling	handle	VERB
fcis-30157	493	2	missing	miss	VERB
fcis-30157	493	3	data	datum	NOUN
fcis-30157	493	4	and	and	CCONJ
fcis-30157	493	5	ensuring	ensure	VERB
fcis-30157	493	6	privacy	privacy	NOUN
fcis-30157	493	7	.	.	PUNCT
fcis-30157	494	1	identification	identification	NOUN
fcis-30157	494	2	accuracy	accuracy	NOUN
fcis-30157	494	3	high	high	ADJ
fcis-30157	494	4	accuracy	accuracy	NOUN
fcis-30157	494	5	reduces	reduce	VERB
fcis-30157	494	6	misdiagnosis	misdiagnosis	NOUN
fcis-30157	494	7	.	.	PUNCT
fcis-30157	495	1	affected	affect	VERB
fcis-30157	495	2	by	by	ADP
fcis-30157	495	3	model	model	NOUN
fcis-30157	495	4	complexity	complexity	NOUN
fcis-30157	495	5	and	and	CCONJ
fcis-30157	495	6	training	training	NOUN
fcis-30157	495	7	data	datum	NOUN
fcis-30157	495	8	.	.	PUNCT
fcis-30157	496	1	use	use	VERB
fcis-30157	496	2	diverse	diverse	ADJ
fcis-30157	496	3	datasets	dataset	NOUN
fcis-30157	496	4	to	to	PART
fcis-30157	496	5	improve	improve	VERB
fcis-30157	496	6	generalization	generalization	NOUN
fcis-30157	496	7	ability	ability	NOUN
fcis-30157	496	8	.	.	PUNCT
fcis-30157	497	1	data	datum	NOUN
fcis-30157	497	2	privacy	privacy	NOUN
fcis-30157	497	3	and	and	CCONJ
fcis-30157	497	4	standards	standard	NOUN
fcis-30157	497	5	variation	variation	NOUN
fcis-30157	497	6	.	.	PUNCT
fcis-30157	498	1	optimizing	optimize	VERB
fcis-30157	498	2	image	image	NOUN
fcis-30157	498	3	quality	quality	NOUN
fcis-30157	498	4	,	,	PUNCT
fcis-30157	498	5	incorporating	incorporate	VERB
fcis-30157	498	6	patient	patient	ADJ
fcis-30157	498	7	information	information	NOUN
fcis-30157	498	8	,	,	PUNCT
fcis-30157	498	9	and	and	CCONJ
fcis-30157	498	10	improving	improve	VERB
fcis-30157	498	11	model	model	NOUN
fcis-30157	498	12	accuracy	accuracy	NOUN
fcis-30157	498	13	are	be	AUX
fcis-30157	498	14	essential	essential	ADJ
fcis-30157	498	15	for	for	ADP
fcis-30157	498	16	enhancing	enhance	VERB
fcis-30157	498	17	auxiliary	auxiliary	ADJ
fcis-30157	498	18	diagnostic	diagnostic	ADJ
fcis-30157	498	19	systems	system	NOUN
fcis-30157	498	20	.	.	PUNCT
fcis-30157	499	1	however	however	ADV
fcis-30157	499	2	,	,	PUNCT
fcis-30157	499	3	practical	practical	ADJ
fcis-30157	499	4	challenges	challenge	NOUN
fcis-30157	499	5	like	like	ADP
fcis-30157	499	6	privacy	privacy	NOUN
fcis-30157	499	7	and	and	CCONJ
fcis-30157	499	8	data	datum	NOUN
fcis-30157	499	9	standardization	standardization	NOUN
fcis-30157	499	10	must	must	AUX
fcis-30157	499	11	be	be	AUX
fcis-30157	499	12	addressed	address	VERB
fcis-30157	499	13	to	to	PART
fcis-30157	499	14	ensure	ensure	VERB
fcis-30157	499	15	widespread	widespread	ADJ
fcis-30157	499	16	application	application	NOUN
fcis-30157	499	17	.	.	PUNCT
fcis-30157	500	1	future	future	ADJ
fcis-30157	500	2	research	research	NOUN
fcis-30157	500	3	should	should	AUX
fcis-30157	500	4	focus	focus	VERB
fcis-30157	500	5	on	on	ADP
fcis-30157	500	6	these	these	DET
fcis-30157	500	7	issues	issue	NOUN
fcis-30157	500	8	to	to	PART
fcis-30157	500	9	advance	advance	VERB
fcis-30157	500	10	the	the	DET
fcis-30157	500	11	effectiveness	effectiveness	NOUN
fcis-30157	500	12	of	of	ADP
fcis-30157	500	13	these	these	DET
fcis-30157	500	14	technologies	technology	NOUN
fcis-30157	500	15	in	in	ADP
fcis-30157	500	16	diagnosing	diagnose	VERB
fcis-30157	500	17	skin	skin	NOUN
fcis-30157	500	18	diseases	disease	NOUN
fcis-30157	500	19	.	.	PUNCT
fcis-30157	501	1	3.3.3	3.3.3	X
fcis-30157	501	2	.	.	PUNCT
fcis-30157	501	3	clinical	clinical	ADJ
fcis-30157	501	4	application	application	NOUN
fcis-30157	501	5	assessment	assessment	NOUN
fcis-30157	501	6	the	the	DET
fcis-30157	501	7	clinical	clinical	ADJ
fcis-30157	501	8	assessment	assessment	NOUN
fcis-30157	501	9	of	of	ADP
fcis-30157	501	10	an	an	DET
fcis-30157	501	11	auxiliary	auxiliary	ADJ
fcis-30157	501	12	diagnostic	diagnostic	ADJ
fcis-30157	501	13	system	system	NOUN
fcis-30157	501	14	focuses	focus	VERB
fcis-30157	501	15	on	on	ADP
fcis-30157	501	16	its	its	PRON
fcis-30157	501	17	real	real	ADJ
fcis-30157	501	18	-	-	PUNCT
fcis-30157	501	19	world	world	NOUN
fcis-30157	501	20	utility	utility	NOUN
fcis-30157	501	21	in	in	ADP
fcis-30157	501	22	medical	medical	ADJ
fcis-30157	501	23	settings	setting	NOUN
fcis-30157	501	24	,	,	PUNCT
fcis-30157	501	25	including	include	VERB
fcis-30157	501	26	operational	operational	ADJ
fcis-30157	501	27	ease	ease	NOUN
fcis-30157	501	28	,	,	PUNCT
fcis-30157	501	29	diagnostic	diagnostic	ADJ
fcis-30157	501	30	efficiency	efficiency	NOUN
fcis-30157	501	31	,	,	PUNCT
fcis-30157	501	32	and	and	CCONJ
fcis-30157	501	33	its	its	PRON
fcis-30157	501	34	role	role	NOUN
fcis-30157	501	35	in	in	ADP
fcis-30157	501	36	supporting	support	VERB
fcis-30157	501	37	doctors	doctor	NOUN
fcis-30157	501	38	’	’	PART
fcis-30157	501	39	decision	decision	NOUN
fcis-30157	501	40	-	-	PUNCT
fcis-30157	501	41	making	making	NOUN
fcis-30157	501	42	.	.	PUNCT
fcis-30157	502	1	in	in	ADP
fcis-30157	502	2	a	a	DET
fcis-30157	502	3	study	study	NOUN
fcis-30157	502	4	at	at	ADP
fcis-30157	502	5	a	a	DET
fcis-30157	502	6	tertiary	tertiary	ADJ
fcis-30157	502	7	hospital	hospital	NOUN
fcis-30157	502	8	,	,	PUNCT
fcis-30157	502	9	the	the	DET
fcis-30157	502	10	system	system	NOUN
fcis-30157	502	11	processed	process	VERB
fcis-30157	502	12	over	over	ADP
fcis-30157	502	13	5,000	5,000	NUM
fcis-30157	502	14	skin	skin	NOUN
fcis-30157	502	15	images	image	NOUN
fcis-30157	502	16	in	in	ADP
fcis-30157	502	17	a	a	DET
fcis-30157	502	18	month	month	NOUN
fcis-30157	502	19	,	,	PUNCT
fcis-30157	502	20	reducing	reduce	VERB
fcis-30157	502	21	the	the	DET
fcis-30157	502	22	average	average	ADJ
fcis-30157	502	23	diagnostic	diagnostic	ADJ
fcis-30157	502	24	time	time	NOUN
fcis-30157	502	25	to	to	ADP
fcis-30157	502	26	2.5	2.5	NUM
fcis-30157	502	27	minutes	minute	NOUN
fcis-30157	502	28	per	per	ADP
fcis-30157	502	29	case	case	NOUN
fcis-30157	502	30	,	,	PUNCT
fcis-30157	502	31	compared	compare	VERB
fcis-30157	502	32	to	to	ADP
fcis-30157	502	33	10	10	NUM
fcis-30157	502	34	-	-	SYM
fcis-30157	502	35	15	15	NUM
fcis-30157	502	36	minutes	minute	NOUN
fcis-30157	502	37	by	by	ADP
fcis-30157	502	38	human	human	ADJ
fcis-30157	502	39	dermatologists	dermatologist	NOUN
fcis-30157	502	40	.	.	PUNCT
fcis-30157	503	1	this	this	DET
fcis-30157	503	2	time	time	NOUN
fcis-30157	503	3	-	-	PUNCT
fcis-30157	503	4	saving	save	VERB
fcis-30157	503	5	improved	improve	VERB
fcis-30157	503	6	patient	patient	ADJ
fcis-30157	503	7	throughput	throughput	NOUN
fcis-30157	503	8	and	and	CCONJ
fcis-30157	503	9	reduced	reduce	VERB
fcis-30157	503	10	clinician	clinician	ADJ
fcis-30157	503	11	workload	workload	NOUN
fcis-30157	503	12	.	.	PUNCT
fcis-30157	504	1	user	user	NOUN
fcis-30157	504	2	feedback	feedback	NOUN
fcis-30157	504	3	from	from	ADP
fcis-30157	504	4	dermatologists	dermatologist	NOUN
fcis-30157	504	5	showed	show	VERB
fcis-30157	504	6	that	that	SCONJ
fcis-30157	504	7	92	92	NUM
fcis-30157	504	8	%	%	NOUN
fcis-30157	504	9	found	find	VERB
fcis-30157	504	10	the	the	DET
fcis-30157	504	11	system	system	NOUN
fcis-30157	504	12	intuitive	intuitive	ADJ
fcis-30157	504	13	,	,	PUNCT
fcis-30157	504	14	with	with	ADP
fcis-30157	504	15	only	only	ADJ
fcis-30157	504	16	minor	minor	ADJ
fcis-30157	504	17	difficulties	difficulty	NOUN
fcis-30157	504	18	reported	report	VERB
fcis-30157	504	19	by	by	ADP
fcis-30157	504	20	8	8	NUM
fcis-30157	504	21	%	%	NOUN
fcis-30157	504	22	.	.	PUNCT
fcis-30157	505	1	these	these	DET
fcis-30157	505	2	results	result	NOUN
fcis-30157	505	3	suggest	suggest	VERB
fcis-30157	505	4	the	the	DET
fcis-30157	505	5	system	system	NOUN
fcis-30157	505	6	is	be	AUX
fcis-30157	505	7	easy	easy	ADJ
fcis-30157	505	8	to	to	PART
fcis-30157	505	9	adopt	adopt	VERB
fcis-30157	505	10	and	and	CCONJ
fcis-30157	505	11	minimizes	minimize	VERB
fcis-30157	505	12	the	the	DET
fcis-30157	505	13	learning	learning	NOUN
fcis-30157	505	14	curve	curve	NOUN
fcis-30157	505	15	.	.	PUNCT
fcis-30157	506	1	while	while	SCONJ
fcis-30157	506	2	diagnostic	diagnostic	ADJ
fcis-30157	506	3	speed	speed	NOUN
fcis-30157	506	4	is	be	AUX
fcis-30157	506	5	important	important	ADJ
fcis-30157	506	6	,	,	PUNCT
fcis-30157	506	7	accuracy	accuracy	NOUN
fcis-30157	506	8	is	be	AUX
fcis-30157	506	9	crucial	crucial	ADJ
fcis-30157	506	10	.	.	PUNCT
fcis-30157	507	1	in	in	ADP
fcis-30157	507	2	a	a	DET
fcis-30157	507	3	study	study	NOUN
fcis-30157	507	4	with	with	ADP
fcis-30157	507	5	100	100	NUM
fcis-30157	507	6	patients	patient	NOUN
fcis-30157	507	7	,	,	PUNCT
fcis-30157	507	8	the	the	DET
fcis-30157	507	9	system	system	NOUN
fcis-30157	507	10	achieved	achieve	VERB
fcis-30157	507	11	89.7	89.7	NUM
fcis-30157	507	12	%	%	NOUN
fcis-30157	507	13	diagnostic	diagnostic	ADJ
fcis-30157	507	14	accuracy	accuracy	NOUN
fcis-30157	507	15	,	,	PUNCT
fcis-30157	507	16	compared	compare	VERB
fcis-30157	507	17	to	to	ADP
fcis-30157	507	18	92.3	92.3	NUM
fcis-30157	507	19	%	%	NOUN
fcis-30157	507	20	by	by	ADP
fcis-30157	507	21	experienced	experienced	ADJ
fcis-30157	507	22	dermatologists	dermatologist	NOUN
fcis-30157	507	23	.	.	PUNCT
fcis-30157	508	1	however	however	ADV
fcis-30157	508	2	,	,	PUNCT
fcis-30157	508	3	it	it	PRON
fcis-30157	508	4	outperformed	outperform	VERB
fcis-30157	508	5	junior	junior	ADJ
fcis-30157	508	6	dermatologists	dermatologist	NOUN
fcis-30157	508	7	(	(	PUNCT
fcis-30157	508	8	82.5	82.5	NUM
fcis-30157	508	9	%	%	NOUN
fcis-30157	508	10	accuracy	accuracy	NOUN
fcis-30157	508	11	)	)	PUNCT
fcis-30157	508	12	,	,	PUNCT
fcis-30157	508	13	making	make	VERB
fcis-30157	508	14	it	it	PRON
fcis-30157	508	15	a	a	DET
fcis-30157	508	16	valuable	valuable	ADJ
fcis-30157	508	17	tool	tool	NOUN
fcis-30157	508	18	for	for	ADP
fcis-30157	508	19	providing	provide	VERB
fcis-30157	508	20	a	a	DET
fcis-30157	508	21	second	second	ADJ
fcis-30157	508	22	opinion	opinion	NOUN
fcis-30157	508	23	,	,	PUNCT
fcis-30157	508	24	particularly	particularly	ADV
fcis-30157	508	25	in	in	ADP
fcis-30157	508	26	complex	complex	ADJ
fcis-30157	508	27	cases	case	NOUN
fcis-30157	508	28	.	.	PUNCT
fcis-30157	509	1	the	the	DET
fcis-30157	509	2	system	system	NOUN
fcis-30157	509	3	also	also	ADV
fcis-30157	509	4	correctly	correctly	ADV
fcis-30157	509	5	identified	identify	VERB
fcis-30157	509	6	a	a	DET
fcis-30157	509	7	melanoma	melanoma	NOUN
fcis-30157	509	8	missed	miss	VERB
fcis-30157	509	9	by	by	ADP
fcis-30157	509	10	the	the	DET
fcis-30157	509	11	physician	physician	NOUN
fcis-30157	509	12	,	,	PUNCT
fcis-30157	509	13	highlighting	highlight	VERB
fcis-30157	509	14	its	its	PRON
fcis-30157	509	15	role	role	NOUN
fcis-30157	509	16	in	in	ADP
fcis-30157	509	17	reducing	reduce	VERB
fcis-30157	509	18	diagnostic	diagnostic	ADJ
fcis-30157	509	19	errors	error	NOUN
fcis-30157	509	20	.	.	PUNCT
fcis-30157	510	1	despite	despite	SCONJ
fcis-30157	510	2	strengths	strength	NOUN
fcis-30157	510	3	,	,	PUNCT
fcis-30157	510	4	clinicians	clinician	NOUN
fcis-30157	510	5	noted	note	VERB
fcis-30157	510	6	areas	area	NOUN
fcis-30157	510	7	for	for	ADP
fcis-30157	510	8	improvement	improvement	NOUN
fcis-30157	510	9	.	.	PUNCT
fcis-30157	511	1	the	the	DET
fcis-30157	511	2	system	system	NOUN
fcis-30157	511	3	’s	’s	PART
fcis-30157	511	4	performance	performance	NOUN
fcis-30157	511	5	dropped	drop	VERB
fcis-30157	511	6	to	to	ADP
fcis-30157	511	7	78	78	NUM
fcis-30157	511	8	%	%	NOUN
fcis-30157	511	9	accuracy	accuracy	NOUN
fcis-30157	511	10	when	when	SCONJ
fcis-30157	511	11	lesions	lesion	NOUN
fcis-30157	511	12	were	be	AUX
fcis-30157	511	13	obscured	obscure	VERB
fcis-30157	511	14	by	by	ADP
fcis-30157	511	15	hair	hair	NOUN
fcis-30157	511	16	or	or	CCONJ
fcis-30157	511	17	tattoos	tattoo	NOUN
fcis-30157	511	18	,	,	PUNCT
fcis-30157	511	19	compared	compare	VERB
fcis-30157	511	20	to	to	ADP
fcis-30157	511	21	92	92	NUM
fcis-30157	511	22	%	%	NOUN
fcis-30157	511	23	for	for	ADP
fcis-30157	511	24	clear	clear	ADJ
fcis-30157	511	25	images	image	NOUN
fcis-30157	511	26	.	.	PUNCT
fcis-30157	512	1	this	this	PRON
fcis-30157	512	2	points	point	VERB
fcis-30157	512	3	to	to	ADP
fcis-30157	512	4	the	the	DET
fcis-30157	512	5	need	need	NOUN
fcis-30157	512	6	for	for	ADP
fcis-30157	512	7	better	well	ADJ
fcis-30157	512	8	preprocessing	preprocessing	NOUN
fcis-30157	512	9	techniques	technique	NOUN
fcis-30157	512	10	.	.	PUNCT
fcis-30157	513	1	additionally	additionally	ADV
fcis-30157	513	2	,	,	PUNCT
fcis-30157	513	3	the	the	DET
fcis-30157	513	4	system	system	NOUN
fcis-30157	513	5	lacked	lack	VERB
fcis-30157	513	6	explainability	explainability	NOUN
fcis-30157	513	7	,	,	PUNCT
fcis-30157	513	8	which	which	PRON
fcis-30157	513	9	limited	limit	VERB
fcis-30157	513	10	clinician	clinician	NOUN
fcis-30157	513	11	trust	trust	NOUN
fcis-30157	513	12	.	.	PUNCT
fcis-30157	514	1	future	future	ADJ
fcis-30157	514	2	versions	version	NOUN
fcis-30157	514	3	could	could	AUX
fcis-30157	514	4	integrate	integrate	VERB
fcis-30157	514	5	explainable	explainable	ADJ
fcis-30157	514	6	ai	ai	NOUN
fcis-30157	514	7	(	(	PUNCT
fcis-30157	514	8	xai	xai	PROPN
fcis-30157	514	9	)	)	PUNCT
fcis-30157	514	10	,	,	PUNCT
fcis-30157	514	11	such	such	ADJ
fcis-30157	514	12	as	as	ADP
fcis-30157	514	13	heatmaps	heatmap	NOUN
fcis-30157	514	14	or	or	CCONJ
fcis-30157	514	15	decision	decision	NOUN
fcis-30157	514	16	trees	tree	NOUN
fcis-30157	514	17	,	,	PUNCT
fcis-30157	514	18	to	to	PART
fcis-30157	514	19	clarify	clarify	VERB
fcis-30157	514	20	diagnosis	diagnosis	NOUN
fcis-30157	514	21	reasoning	reasoning	NOUN
fcis-30157	514	22	.	.	PUNCT
fcis-30157	515	1	the	the	DET
fcis-30157	515	2	system	system	NOUN
fcis-30157	515	3	also	also	ADV
fcis-30157	515	4	showed	show	VERB
fcis-30157	515	5	disparities	disparity	NOUN
fcis-30157	515	6	in	in	ADP
fcis-30157	515	7	performance	performance	NOUN
fcis-30157	515	8	across	across	ADP
fcis-30157	515	9	skin	skin	NOUN
fcis-30157	515	10	tones	tone	NOUN
fcis-30157	515	11	,	,	PUNCT
fcis-30157	515	12	with	with	ADP
fcis-30157	515	13	87	87	NUM
fcis-30157	515	14	%	%	NOUN
fcis-30157	515	15	accuracy	accuracy	NOUN
fcis-30157	515	16	for	for	ADP
fcis-30157	515	17	light	light	ADJ
fcis-30157	515	18	-	-	PUNCT
fcis-30157	515	19	skinned	skinned	ADJ
fcis-30157	515	20	patients	patient	NOUN
fcis-30157	515	21	,	,	PUNCT
fcis-30157	515	22	compared	compare	VERB
fcis-30157	515	23	to	to	ADP
fcis-30157	515	24	79	79	NUM
fcis-30157	515	25	%	%	NOUN
fcis-30157	515	26	for	for	ADP
fcis-30157	515	27	dark	dark	ADJ
fcis-30157	515	28	-	-	PUNCT
fcis-30157	515	29	skinned	skinned	ADJ
fcis-30157	515	30	ones	one	NOUN
fcis-30157	515	31	.	.	PUNCT
fcis-30157	516	1	this	this	DET
fcis-30157	516	2	highlights	highlight	VERB
fcis-30157	516	3	the	the	DET
fcis-30157	516	4	need	need	NOUN
fcis-30157	516	5	for	for	ADP
fcis-30157	516	6	more	more	ADJ
fcis-30157	516	7	inclusive	inclusive	ADJ
fcis-30157	516	8	training	training	NOUN
fcis-30157	516	9	datasets	dataset	NOUN
fcis-30157	516	10	.	.	PUNCT
fcis-30157	517	1	in	in	ADP
fcis-30157	517	2	terms	term	NOUN
fcis-30157	517	3	of	of	ADP
fcis-30157	517	4	scalability	scalability	NOUN
fcis-30157	517	5	,	,	PUNCT
fcis-30157	517	6	the	the	DET
fcis-30157	517	7	system	system	NOUN
fcis-30157	517	8	performed	perform	VERB
fcis-30157	517	9	well	well	ADV
fcis-30157	517	10	in	in	ADP
fcis-30157	517	11	a	a	DET
fcis-30157	517	12	rural	rural	ADJ
fcis-30157	517	13	clinic	clinic	NOUN
fcis-30157	517	14	using	use	VERB
fcis-30157	517	15	smartphone	smartphone	NOUN
fcis-30157	517	16	images	image	NOUN
fcis-30157	517	17	,	,	PUNCT
fcis-30157	517	18	achieving	achieve	VERB
fcis-30157	517	19	84	84	NUM
fcis-30157	517	20	%	%	NOUN
fcis-30157	517	21	accuracy	accuracy	NOUN
fcis-30157	517	22	,	,	PUNCT
fcis-30157	517	23	showing	show	VERB
fcis-30157	517	24	its	its	PRON
fcis-30157	517	25	potential	potential	NOUN
fcis-30157	517	26	to	to	PART
fcis-30157	517	27	support	support	VERB
fcis-30157	517	28	dermatological	dermatological	ADJ
fcis-30157	517	29	care	care	NOUN
fcis-30157	517	30	in	in	ADP
fcis-30157	517	31	underserved	underserved	ADJ
fcis-30157	517	32	areas	area	NOUN
fcis-30157	517	33	.	.	PUNCT
fcis-30157	518	1	economically	economically	ADV
fcis-30157	518	2	,	,	PUNCT
fcis-30157	518	3	the	the	DET
fcis-30157	518	4	system	system	NOUN
fcis-30157	518	5	reduced	reduce	VERB
fcis-30157	518	6	diagnostic	diagnostic	ADJ
fcis-30157	518	7	costs	cost	NOUN
fcis-30157	518	8	by	by	ADP
fcis-30157	518	9	30	30	NUM
fcis-30157	518	10	%	%	NOUN
fcis-30157	518	11	,	,	PUNCT
fcis-30157	518	12	largely	largely	ADV
fcis-30157	518	13	by	by	ADP
fcis-30157	518	14	reducing	reduce	VERB
fcis-30157	518	15	biopsies	biopsy	NOUN
fcis-30157	518	16	and	and	CCONJ
fcis-30157	518	17	specialist	specialist	NOUN
fcis-30157	518	18	consultations	consultation	NOUN
fcis-30157	518	19	.	.	PUNCT
fcis-30157	519	1	this	this	PRON
fcis-30157	519	2	makes	make	VERB
fcis-30157	519	3	dermatological	dermatological	ADJ
fcis-30157	519	4	care	care	NOUN
fcis-30157	519	5	more	more	ADV
fcis-30157	519	6	affordable	affordable	ADJ
fcis-30157	519	7	for	for	ADP
fcis-30157	519	8	patients	patient	NOUN
fcis-30157	519	9	with	with	ADP
fcis-30157	519	10	limited	limited	ADJ
fcis-30157	519	11	resources	resource	NOUN
fcis-30157	519	12	.	.	PUNCT
fcis-30157	520	1	however	however	ADV
fcis-30157	520	2	,	,	PUNCT
fcis-30157	520	3	challenges	challenge	NOUN
fcis-30157	520	4	remain	remain	VERB
fcis-30157	520	5	in	in	ADP
fcis-30157	520	6	initial	initial	ADJ
fcis-30157	520	7	investment	investment	NOUN
fcis-30157	520	8	,	,	PUNCT
fcis-30157	520	9	regulatory	regulatory	ADJ
fcis-30157	520	10	approval	approval	NOUN
fcis-30157	520	11	,	,	PUNCT
fcis-30157	520	12	and	and	CCONJ
fcis-30157	520	13	data	datum	NOUN
fcis-30157	520	14	privacy	privacy	NOUN
fcis-30157	520	15	,	,	PUNCT
fcis-30157	520	16	which	which	PRON
fcis-30157	520	17	need	need	VERB
fcis-30157	520	18	addressing	address	VERB
fcis-30157	520	19	for	for	ADP
fcis-30157	520	20	broader	broad	ADJ
fcis-30157	520	21	adoption	adoption	NOUN
fcis-30157	520	22	.	.	PUNCT
fcis-30157	521	1	4	4	X
fcis-30157	521	2	.	.	X
fcis-30157	521	3	experimental	experimental	ADJ
fcis-30157	521	4	summary	summary	NOUN
fcis-30157	521	5	4.1	4.1	NUM
fcis-30157	521	6	.	.	PUNCT
fcis-30157	522	1	summary	summary	NOUN
fcis-30157	522	2	of	of	ADP
fcis-30157	522	3	findings	finding	NOUN
fcis-30157	522	4	4.1.1	4.1.1	NUM
fcis-30157	522	5	.	.	PUNCT
fcis-30157	523	1	key	key	ADJ
fcis-30157	523	2	results	result	NOUN
fcis-30157	523	3	of	of	ADP
fcis-30157	523	4	automatic	automatic	ADJ
fcis-30157	523	5	recognition	recognition	NOUN
fcis-30157	523	6	the	the	DET
fcis-30157	523	7	automatic	automatic	ADJ
fcis-30157	523	8	recognition	recognition	NOUN
fcis-30157	523	9	model	model	NOUN
fcis-30157	523	10	for	for	ADP
fcis-30157	523	11	skin	skin	NOUN
fcis-30157	523	12	diseases	disease	NOUN
fcis-30157	523	13	achieved	achieve	VERB
fcis-30157	523	14	significant	significant	ADJ
fcis-30157	523	15	performance	performance	NOUN
fcis-30157	523	16	metrics	metric	NOUN
fcis-30157	523	17	,	,	PUNCT
fcis-30157	523	18	including	include	VERB
fcis-30157	523	19	92.3	92.3	NUM
fcis-30157	523	20	%	%	NOUN
fcis-30157	523	21	accuracy	accuracy	NOUN
fcis-30157	523	22	,	,	PUNCT
fcis-30157	523	23	89.7	89.7	NUM
fcis-30157	523	24	%	%	NOUN
fcis-30157	523	25	recall	recall	NOUN
fcis-30157	523	26	,	,	PUNCT
fcis-30157	523	27	and	and	CCONJ
fcis-30157	523	28	a	a	DET
fcis-30157	523	29	90.9	90.9	NUM
fcis-30157	523	30	%	%	NOUN
fcis-30157	523	31	f1	f1	NOUN
fcis-30157	523	32	score	score	NOUN
fcis-30157	523	33	.	.	PUNCT
fcis-30157	524	1	notably	notably	ADV
fcis-30157	524	2	,	,	PUNCT
fcis-30157	524	3	it	it	PRON
fcis-30157	524	4	classified	classify	VERB
fcis-30157	524	5	melanoma	melanoma	NOUN
fcis-30157	524	6	and	and	CCONJ
fcis-30157	524	7	benign	benign	ADJ
fcis-30157	524	8	nevi	nevi	NOUN
fcis-30157	524	9	with	with	ADP
fcis-30157	524	10	94.5	94.5	NUM
fcis-30157	524	11	%	%	NOUN
fcis-30157	524	12	accuracy	accuracy	NOUN
fcis-30157	524	13	,	,	PUNCT
fcis-30157	524	14	comparable	comparable	ADJ
fcis-30157	524	15	to	to	ADP
fcis-30157	524	16	experienced	experienced	ADJ
fcis-30157	524	17	dermatologists	dermatologist	NOUN
fcis-30157	524	18	.	.	PUNCT
fcis-30157	525	1	among	among	ADP
fcis-30157	525	2	various	various	ADJ
fcis-30157	525	3	deep	deep	ADJ
fcis-30157	525	4	learning	learning	NOUN
fcis-30157	525	5	architectures	architecture	NOUN
fcis-30157	525	6	,	,	PUNCT
fcis-30157	525	7	resnet	resnet	NOUN
fcis-30157	525	8	outperformed	outperform	VERB
fcis-30157	525	9	others	other	NOUN
fcis-30157	525	10	,	,	PUNCT
fcis-30157	525	11	with	with	ADP
fcis-30157	525	12	93.1	93.1	NUM
fcis-30157	525	13	%	%	NOUN
fcis-30157	525	14	accuracy	accuracy	NOUN
fcis-30157	525	15	,	,	PUNCT
fcis-30157	525	16	while	while	SCONJ
fcis-30157	525	17	vgg	vgg	NOUN
fcis-30157	525	18	and	and	CCONJ
fcis-30157	525	19	inception	inception	NOUN
fcis-30157	525	20	achieved	achieve	VERB
fcis-30157	525	21	90.2	90.2	NUM
fcis-30157	525	22	%	%	NOUN
fcis-30157	525	23	and	and	CCONJ
fcis-30157	525	24	88.7	88.7	NUM
fcis-30157	525	25	%	%	NOUN
fcis-30157	525	26	,	,	PUNCT
fcis-30157	525	27	respectively	respectively	ADV
fcis-30157	525	28	,	,	PUNCT
fcis-30157	525	29	demonstrating	demonstrate	VERB
fcis-30157	525	30	the	the	DET
fcis-30157	525	31	advantage	advantage	NOUN
fcis-30157	525	32	of	of	ADP
fcis-30157	525	33	resnet	resnet	NOUN
fcis-30157	525	34	’s	’s	PART
fcis-30157	525	35	deeper	deep	ADJ
fcis-30157	525	36	architecture	architecture	NOUN
fcis-30157	525	37	and	and	CCONJ
fcis-30157	525	38	residual	residual	ADJ
fcis-30157	525	39	connections	connection	NOUN
fcis-30157	525	40	in	in	ADP
fcis-30157	525	41	capturing	capture	VERB
fcis-30157	525	42	complex	complex	ADJ
fcis-30157	525	43	skin	skin	NOUN
fcis-30157	525	44	disease	disease	NOUN
fcis-30157	525	45	features	feature	VERB
fcis-30157	525	46	.	.	PUNCT
fcis-30157	526	1	the	the	DET
fcis-30157	526	2	model	model	NOUN
fcis-30157	526	3	utilized	utilize	VERB
fcis-30157	526	4	key	key	NOUN
fcis-30157	526	5	features	feature	NOUN
fcis-30157	526	6	such	such	ADJ
fcis-30157	526	7	as	as	ADP
fcis-30157	526	8	color	color	NOUN
fcis-30157	526	9	,	,	PUNCT
fcis-30157	526	10	texture	texture	NOUN
fcis-30157	526	11	,	,	PUNCT
fcis-30157	526	12	and	and	CCONJ
fcis-30157	526	13	shape	shape	NOUN
fcis-30157	526	14	,	,	PUNCT
fcis-30157	526	15	successfully	successfully	ADV
fcis-30157	526	16	identifying	identify	VERB
fcis-30157	526	17	melanoma	melanoma	NOUN
fcis-30157	526	18	characteristics	characteristic	NOUN
fcis-30157	526	19	like	like	ADP
fcis-30157	526	20	irregular	irregular	ADJ
fcis-30157	526	21	borders	border	NOUN
fcis-30157	526	22	and	and	CCONJ
fcis-30157	526	23	asymmetry	asymmetry	NOUN
fcis-30157	526	24	.	.	PUNCT
fcis-30157	527	1	integrating	integrate	VERB
fcis-30157	527	2	multimodal	multimodal	NOUN
fcis-30157	527	3	data	datum	NOUN
fcis-30157	527	4	,	,	PUNCT
fcis-30157	527	5	including	include	VERB
fcis-30157	527	6	patient	patient	ADJ
fcis-30157	527	7	age	age	NOUN
fcis-30157	527	8	and	and	CCONJ
fcis-30157	527	9	lesion	lesion	NOUN
fcis-30157	527	10	location	location	NOUN
fcis-30157	527	11	,	,	PUNCT
fcis-30157	527	12	boosted	boost	VERB
fcis-30157	527	13	accuracy	accuracy	NOUN
fcis-30157	527	14	to	to	ADP
fcis-30157	527	15	95.2	95.2	NUM
fcis-30157	527	16	%	%	NOUN
fcis-30157	527	17	,	,	PUNCT
fcis-30157	527	18	enhancing	enhance	VERB
fcis-30157	527	19	the	the	DET
fcis-30157	527	20	model	model	NOUN
fcis-30157	527	21	’s	’s	PART
fcis-30157	527	22	clinical	clinical	ADJ
fcis-30157	527	23	relevance	relevance	NOUN
fcis-30157	527	24	.	.	PUNCT
fcis-30157	528	1	furthermore	furthermore	ADV
fcis-30157	528	2	,	,	PUNCT
fcis-30157	528	3	the	the	DET
fcis-30157	528	4	model	model	NOUN
fcis-30157	528	5	maintained	maintain	VERB
fcis-30157	528	6	high	high	ADJ
fcis-30157	528	7	accuracy	accuracy	NOUN
fcis-30157	528	8	of	of	ADP
fcis-30157	528	9	91.8	91.8	NUM
fcis-30157	528	10	%	%	NOUN
fcis-30157	528	11	across	across	ADP
fcis-30157	528	12	diverse	diverse	ADJ
fcis-30157	528	13	skin	skin	NOUN
fcis-30157	528	14	tones	tone	NOUN
fcis-30157	528	15	,	,	PUNCT
fcis-30157	528	16	addressing	address	VERB
fcis-30157	528	17	the	the	DET
fcis-30157	528	18	underrepresentation	underrepresentation	NOUN
fcis-30157	528	19	of	of	ADP
fcis-30157	528	20	darker	dark	ADJ
fcis-30157	528	21	skin	skin	NOUN
fcis-30157	528	22	tones	tone	NOUN
fcis-30157	528	23	in	in	ADP
fcis-30157	528	24	many	many	ADJ
fcis-30157	528	25	medical	medical	ADJ
fcis-30157	528	26	datasets	dataset	NOUN
fcis-30157	528	27	.	.	PUNCT
fcis-30157	529	1	in	in	ADP
fcis-30157	529	2	a	a	DET
fcis-30157	529	3	real	real	ADJ
fcis-30157	529	4	-	-	PUNCT
fcis-30157	529	5	world	world	NOUN
fcis-30157	529	6	pilot	pilot	NOUN
fcis-30157	529	7	study	study	NOUN
fcis-30157	529	8	at	at	ADP
fcis-30157	529	9	a	a	DET
fcis-30157	529	10	dermatology	dermatology	NOUN
fcis-30157	529	11	clinic	clinic	NOUN
fcis-30157	529	12	with	with	ADP
fcis-30157	529	13	500	500	NUM
fcis-30157	529	14	patients	patient	NOUN
fcis-30157	529	15	,	,	PUNCT
fcis-30157	529	16	the	the	DET
fcis-30157	529	17	model	model	NOUN
fcis-30157	529	18	identified	identify	VERB
fcis-30157	529	19	93	93	NUM
fcis-30157	529	20	%	%	NOUN
fcis-30157	529	21	of	of	ADP
fcis-30157	529	22	cases	case	NOUN
fcis-30157	529	23	correctly	correctly	ADV
fcis-30157	529	24	,	,	PUNCT
fcis-30157	529	25	with	with	ADP
fcis-30157	529	26	misclassifications	misclassification	NOUN
fcis-30157	529	27	primarily	primarily	ADV
fcis-30157	529	28	involving	involve	VERB
fcis-30157	529	29	overlapping	overlap	VERB
fcis-30157	529	30	conditions	condition	NOUN
fcis-30157	529	31	like	like	ADP
fcis-30157	529	32	psoriasis	psoriasis	NOUN
fcis-30157	529	33	and	and	CCONJ
fcis-30157	529	34	eczema	eczema	NOUN
fcis-30157	529	35	.	.	PUNCT
fcis-30157	530	1	to	to	PART
fcis-30157	530	2	improve	improve	VERB
fcis-30157	530	3	transparency	transparency	NOUN
fcis-30157	530	4	,	,	PUNCT
fcis-30157	530	5	explainable	explainable	ADJ
fcis-30157	530	6	ai	ai	NOUN
fcis-30157	530	7	techniques	technique	NOUN
fcis-30157	530	8	were	be	AUX
fcis-30157	530	9	incorporated	incorporate	VERB
fcis-30157	530	10	,	,	PUNCT
fcis-30157	530	11	allowing	allow	VERB
fcis-30157	530	12	clinicians	clinician	NOUN
fcis-30157	530	13	to	to	PART
fcis-30157	530	14	understand	understand	VERB
fcis-30157	530	15	misclassifications	misclassification	NOUN
fcis-30157	530	16	by	by	ADP
fcis-30157	530	17	highlighting	highlight	VERB
fcis-30157	530	18	key	key	ADJ
fcis-30157	530	19	features	feature	NOUN
fcis-30157	530	20	.	.	PUNCT
fcis-30157	531	1	the	the	DET
fcis-30157	531	2	model	model	NOUN
fcis-30157	531	3	reduced	reduce	VERB
fcis-30157	531	4	diagnosis	diagnosis	NOUN
fcis-30157	531	5	time	time	NOUN
fcis-30157	531	6	from	from	ADP
fcis-30157	531	7	30	30	NUM
fcis-30157	531	8	minutes	minute	NOUN
fcis-30157	531	9	to	to	ADP
fcis-30157	531	10	just	just	ADV
fcis-30157	531	11	5	5	NUM
fcis-30157	531	12	minutes	minute	NOUN
fcis-30157	531	13	and	and	CCONJ
fcis-30157	531	14	minimized	minimize	VERB
fcis-30157	531	15	human	human	ADJ
fcis-30157	531	16	subjectivity	subjectivity	NOUN
fcis-30157	531	17	,	,	PUNCT
fcis-30157	531	18	making	make	VERB
fcis-30157	531	19	it	it	PRON
fcis-30157	531	20	particularly	particularly	ADV
fcis-30157	531	21	useful	useful	ADJ
fcis-30157	531	22	in	in	ADP
fcis-30157	531	23	resource	resource	NOUN
fcis-30157	531	24	-	-	PUNCT
fcis-30157	531	25	limited	limit	VERB
fcis-30157	531	26	areas	area	NOUN
fcis-30157	531	27	.	.	PUNCT
fcis-30157	532	1	however	however	ADV
fcis-30157	532	2	,	,	PUNCT
fcis-30157	532	3	performance	performance	NOUN
fcis-30157	532	4	on	on	ADP
fcis-30157	532	5	rare	rare	ADJ
fcis-30157	532	6	skin	skin	NOUN
fcis-30157	532	7	diseases	disease	NOUN
fcis-30157	532	8	,	,	PUNCT
fcis-30157	532	9	like	like	ADP
fcis-30157	532	10	cutaneous	cutaneous	ADJ
fcis-30157	532	11	t	t	PROPN
fcis-30157	532	12	-	-	PUNCT
fcis-30157	532	13	cell	cell	NOUN
fcis-30157	532	14	lymphoma	lymphoma	NOUN
fcis-30157	532	15	,	,	PUNCT
fcis-30157	532	16	was	be	AUX
fcis-30157	532	17	lower	low	ADJ
fcis-30157	532	18	(	(	PUNCT
fcis-30157	532	19	85	85	NUM
fcis-30157	532	20	%	%	NOUN
fcis-30157	532	21	accuracy	accuracy	NOUN
fcis-30157	532	22	)	)	PUNCT
fcis-30157	532	23	,	,	PUNCT
fcis-30157	532	24	likely	likely	ADJ
fcis-30157	532	25	due	due	ADP
fcis-30157	532	26	to	to	ADP
fcis-30157	532	27	limited	limit	VERB
fcis-30157	532	28	training	training	NOUN
fcis-30157	532	29	data	datum	NOUN
fcis-30157	532	30	.	.	PUNCT
fcis-30157	533	1	additionally	additionally	ADV
fcis-30157	533	2	,	,	PUNCT
fcis-30157	533	3	image	image	NOUN
fcis-30157	533	4	quality	quality	NOUN
fcis-30157	533	5	issues	issue	NOUN
fcis-30157	533	6	,	,	PUNCT
fcis-30157	533	7	such	such	ADJ
fcis-30157	533	8	as	as	ADP
fcis-30157	533	9	blurriness	blurriness	NOUN
fcis-30157	533	10	or	or	CCONJ
fcis-30157	533	11	poor	poor	ADJ
fcis-30157	533	12	lighting	lighting	NOUN
fcis-30157	533	13	,	,	PUNCT
fcis-30157	533	14	reduced	reduce	VERB
fcis-30157	533	15	accuracy	accuracy	NOUN
fcis-30157	533	16	to	to	ADP
fcis-30157	533	17	87	87	NUM
fcis-30157	533	18	%	%	NOUN
fcis-30157	533	19	,	,	PUNCT
fcis-30157	533	20	highlighting	highlight	VERB
fcis-30157	533	21	the	the	DET
fcis-30157	533	22	need	need	NOUN
fcis-30157	533	23	for	for	ADP
fcis-30157	533	24	improved	improved	ADJ
fcis-30157	533	25	image	image	NOUN
fcis-30157	533	26	acquisition	acquisition	NOUN
fcis-30157	533	27	protocols	protocol	NOUN
fcis-30157	533	28	.	.	PUNCT
fcis-30157	534	1	123	123	NUM
fcis-30157	534	2	4.1.2	4.1.2	NUM
fcis-30157	534	3	.	.	PUNCT
fcis-30157	534	4	achievements	achievement	NOUN
fcis-30157	534	5	in	in	ADP
fcis-30157	534	6	auxiliary	auxiliary	ADJ
fcis-30157	534	7	diagnosis	diagnosis	NOUN
fcis-30157	534	8	the	the	DET
fcis-30157	534	9	auxiliary	auxiliary	ADJ
fcis-30157	534	10	diagnostic	diagnostic	ADJ
fcis-30157	534	11	framework	framework	NOUN
fcis-30157	534	12	developed	develop	VERB
fcis-30157	534	13	in	in	ADP
fcis-30157	534	14	this	this	DET
fcis-30157	534	15	study	study	NOUN
fcis-30157	534	16	combines	combine	VERB
fcis-30157	534	17	automatic	automatic	ADJ
fcis-30157	534	18	skin	skin	NOUN
fcis-30157	534	19	disease	disease	NOUN
fcis-30157	534	20	recognition	recognition	NOUN
fcis-30157	534	21	with	with	ADP
fcis-30157	534	22	clinical	clinical	ADJ
fcis-30157	534	23	knowledge	knowledge	NOUN
fcis-30157	534	24	to	to	PART
fcis-30157	534	25	improve	improve	VERB
fcis-30157	534	26	diagnostic	diagnostic	ADJ
fcis-30157	534	27	accuracy	accuracy	NOUN
fcis-30157	534	28	.	.	PUNCT
fcis-30157	535	1	by	by	ADP
fcis-30157	535	2	integrating	integrate	VERB
fcis-30157	535	3	image	image	NOUN
fcis-30157	535	4	-	-	PUNCT
fcis-30157	535	5	based	base	VERB
fcis-30157	535	6	recognition	recognition	NOUN
fcis-30157	535	7	with	with	ADP
fcis-30157	535	8	patient	patient	ADJ
fcis-30157	535	9	history	history	NOUN
fcis-30157	535	10	,	,	PUNCT
fcis-30157	535	11	symptoms	symptom	NOUN
fcis-30157	535	12	,	,	PUNCT
fcis-30157	535	13	and	and	CCONJ
fcis-30157	535	14	physical	physical	ADJ
fcis-30157	535	15	examination	examination	NOUN
fcis-30157	535	16	findings	finding	NOUN
fcis-30157	535	17	,	,	PUNCT
fcis-30157	535	18	the	the	DET
fcis-30157	535	19	framework	framework	NOUN
fcis-30157	535	20	handles	handle	VERB
fcis-30157	535	21	complex	complex	ADJ
fcis-30157	535	22	cases	case	NOUN
fcis-30157	535	23	where	where	SCONJ
fcis-30157	535	24	visual	visual	ADJ
fcis-30157	535	25	cues	cue	NOUN
fcis-30157	535	26	alone	alone	ADV
fcis-30157	535	27	may	may	AUX
fcis-30157	535	28	be	be	AUX
fcis-30157	535	29	insufficient	insufficient	ADJ
fcis-30157	535	30	.	.	PUNCT
fcis-30157	536	1	in	in	ADP
fcis-30157	536	2	a	a	DET
fcis-30157	536	3	study	study	NOUN
fcis-30157	536	4	of	of	ADP
fcis-30157	536	5	1,000	1,000	NUM
fcis-30157	536	6	skin	skin	NOUN
fcis-30157	536	7	disease	disease	NOUN
fcis-30157	536	8	cases	case	NOUN
fcis-30157	536	9	,	,	PUNCT
fcis-30157	536	10	the	the	DET
fcis-30157	536	11	framework	framework	NOUN
fcis-30157	536	12	achieved	achieve	VERB
fcis-30157	536	13	92	92	NUM
fcis-30157	536	14	%	%	NOUN
fcis-30157	536	15	sensitivity	sensitivity	NOUN
fcis-30157	536	16	and	and	CCONJ
fcis-30157	536	17	95	95	NUM
fcis-30157	536	18	%	%	NOUN
fcis-30157	536	19	specificity	specificity	NOUN
fcis-30157	536	20	,	,	PUNCT
fcis-30157	536	21	outperforming	outperform	VERB
fcis-30157	536	22	traditional	traditional	ADJ
fcis-30157	536	23	methods	method	NOUN
fcis-30157	536	24	that	that	PRON
fcis-30157	536	25	struggle	struggle	VERB
fcis-30157	536	26	with	with	ADP
fcis-30157	536	27	balancing	balance	VERB
fcis-30157	536	28	these	these	DET
fcis-30157	536	29	metrics	metric	NOUN
fcis-30157	536	30	.	.	PUNCT
fcis-30157	537	1	it	it	PRON
fcis-30157	537	2	was	be	AUX
fcis-30157	537	3	particularly	particularly	ADV
fcis-30157	537	4	effective	effective	ADJ
fcis-30157	537	5	in	in	ADP
fcis-30157	537	6	primary	primary	ADJ
fcis-30157	537	7	care	care	NOUN
fcis-30157	537	8	,	,	PUNCT
fcis-30157	537	9	reducing	reduce	VERB
fcis-30157	537	10	diagnosis	diagnosis	NOUN
fcis-30157	537	11	time	time	NOUN
fcis-30157	537	12	from	from	ADP
fcis-30157	537	13	20	20	NUM
fcis-30157	537	14	minutes	minute	NOUN
fcis-30157	537	15	to	to	ADP
fcis-30157	537	16	5	5	NUM
fcis-30157	537	17	minutes	minute	NOUN
fcis-30157	537	18	,	,	PUNCT
fcis-30157	537	19	and	and	CCONJ
fcis-30157	537	20	in	in	ADP
fcis-30157	537	21	specialized	specialized	ADJ
fcis-30157	537	22	dermatology	dermatology	NOUN
fcis-30157	537	23	clinics	clinic	NOUN
fcis-30157	537	24	,	,	PUNCT
fcis-30157	537	25	it	it	PRON
fcis-30157	537	26	helped	help	VERB
fcis-30157	537	27	diagnose	diagnose	VERB
fcis-30157	537	28	rare	rare	ADJ
fcis-30157	537	29	conditions	condition	NOUN
fcis-30157	537	30	like	like	ADP
fcis-30157	537	31	melanoma	melanoma	NOUN
fcis-30157	537	32	,	,	PUNCT
fcis-30157	537	33	improving	improve	VERB
fcis-30157	537	34	accuracy	accuracy	NOUN
fcis-30157	537	35	in	in	ADP
fcis-30157	537	36	challenging	challenging	ADJ
fcis-30157	537	37	cases	case	NOUN
fcis-30157	537	38	.	.	PUNCT
fcis-30157	538	1	the	the	DET
fcis-30157	538	2	framework	framework	NOUN
fcis-30157	538	3	’s	’s	PART
fcis-30157	538	4	integration	integration	NOUN
fcis-30157	538	5	of	of	ADP
fcis-30157	538	6	automatic	automatic	ADJ
fcis-30157	538	7	identification	identification	NOUN
fcis-30157	538	8	with	with	ADP
fcis-30157	538	9	clinical	clinical	ADJ
fcis-30157	538	10	knowledge	knowledge	NOUN
fcis-30157	538	11	reduces	reduce	VERB
fcis-30157	538	12	clinician	clinician	NOUN
fcis-30157	538	13	cognitive	cognitive	ADJ
fcis-30157	538	14	load	load	NOUN
fcis-30157	538	15	and	and	CCONJ
fcis-30157	538	16	enables	enable	VERB
fcis-30157	538	17	faster	fast	ADV
fcis-30157	538	18	,	,	PUNCT
fcis-30157	538	19	more	more	ADV
fcis-30157	538	20	accurate	accurate	ADJ
fcis-30157	538	21	decisions	decision	NOUN
fcis-30157	538	22	.	.	PUNCT
fcis-30157	539	1	in	in	ADP
fcis-30157	539	2	a	a	DET
fcis-30157	539	3	comparison	comparison	NOUN
fcis-30157	539	4	with	with	ADP
fcis-30157	539	5	dermatologists	dermatologist	NOUN
fcis-30157	539	6	,	,	PUNCT
fcis-30157	539	7	the	the	DET
fcis-30157	539	8	system	system	NOUN
fcis-30157	539	9	achieved	achieve	VERB
fcis-30157	539	10	93	93	NUM
fcis-30157	539	11	%	%	NOUN
fcis-30157	539	12	accuracy	accuracy	NOUN
fcis-30157	539	13	,	,	PUNCT
fcis-30157	539	14	compared	compare	VERB
fcis-30157	539	15	to	to	ADP
fcis-30157	539	16	85	85	NUM
fcis-30157	539	17	%	%	NOUN
fcis-30157	539	18	for	for	ADP
fcis-30157	539	19	human	human	ADJ
fcis-30157	539	20	experts	expert	NOUN
fcis-30157	539	21	,	,	PUNCT
fcis-30157	539	22	especially	especially	ADV
fcis-30157	539	23	in	in	ADP
fcis-30157	539	24	time	time	NOUN
fcis-30157	539	25	-	-	PUNCT
fcis-30157	539	26	sensitive	sensitive	ADJ
fcis-30157	539	27	diagnoses	diagnosis	NOUN
fcis-30157	539	28	such	such	ADJ
fcis-30157	539	29	as	as	ADP
fcis-30157	539	30	skin	skin	NOUN
fcis-30157	539	31	cancers	cancer	NOUN
fcis-30157	539	32	.	.	PUNCT
fcis-30157	540	1	it	it	PRON
fcis-30157	540	2	also	also	ADV
fcis-30157	540	3	learns	learn	VERB
fcis-30157	540	4	and	and	CCONJ
fcis-30157	540	5	adapts	adapt	VERB
fcis-30157	540	6	over	over	ADP
fcis-30157	540	7	time	time	NOUN
fcis-30157	540	8	,	,	PUNCT
fcis-30157	540	9	continuously	continuously	ADV
fcis-30157	540	10	updating	update	VERB
fcis-30157	540	11	its	its	PRON
fcis-30157	540	12	knowledge	knowledge	NOUN
fcis-30157	540	13	base	base	NOUN
fcis-30157	540	14	to	to	PART
fcis-30157	540	15	stay	stay	VERB
fcis-30157	540	16	current	current	ADJ
fcis-30157	540	17	with	with	ADP
fcis-30157	540	18	advancements	advancement	NOUN
fcis-30157	540	19	in	in	ADP
fcis-30157	540	20	dermatology	dermatology	NOUN
fcis-30157	540	21	.	.	PUNCT
fcis-30157	541	1	additionally	additionally	ADV
fcis-30157	541	2	,	,	PUNCT
fcis-30157	541	3	the	the	DET
fcis-30157	541	4	framework	framework	NOUN
fcis-30157	541	5	compensates	compensate	VERB
fcis-30157	541	6	for	for	ADP
fcis-30157	541	7	limitations	limitation	NOUN
fcis-30157	541	8	of	of	ADP
fcis-30157	541	9	purely	purely	ADV
fcis-30157	541	10	image	image	NOUN
fcis-30157	541	11	-	-	PUNCT
fcis-30157	541	12	based	base	VERB
fcis-30157	541	13	systems	system	NOUN
fcis-30157	541	14	,	,	PUNCT
fcis-30157	541	15	like	like	ADP
fcis-30157	541	16	identifying	identify	VERB
fcis-30157	541	17	subtle	subtle	ADJ
fcis-30157	541	18	lesions	lesion	NOUN
fcis-30157	541	19	or	or	CCONJ
fcis-30157	541	20	visually	visually	ADV
fcis-30157	541	21	ambiguous	ambiguous	ADJ
fcis-30157	541	22	conditions	condition	NOUN
fcis-30157	541	23	,	,	PUNCT
fcis-30157	541	24	by	by	ADP
fcis-30157	541	25	using	use	VERB
fcis-30157	541	26	clinical	clinical	ADJ
fcis-30157	541	27	data	datum	NOUN
fcis-30157	541	28	to	to	PART
fcis-30157	541	29	provide	provide	VERB
fcis-30157	541	30	more	more	ADV
fcis-30157	541	31	holistic	holistic	ADJ
fcis-30157	541	32	insights	insight	NOUN
fcis-30157	541	33	.	.	PUNCT
fcis-30157	542	1	its	its	PRON
fcis-30157	542	2	educational	educational	ADJ
fcis-30157	542	3	value	value	NOUN
fcis-30157	542	4	is	be	AUX
fcis-30157	542	5	significant	significant	ADJ
fcis-30157	542	6	,	,	PUNCT
fcis-30157	542	7	as	as	SCONJ
fcis-30157	542	8	it	it	PRON
fcis-30157	542	9	helps	help	VERB
fcis-30157	542	10	medical	medical	ADJ
fcis-30157	542	11	students	student	NOUN
fcis-30157	542	12	and	and	CCONJ
fcis-30157	542	13	junior	junior	ADJ
fcis-30157	542	14	clinicians	clinician	NOUN
fcis-30157	542	15	understand	understand	VERB
fcis-30157	542	16	diagnostic	diagnostic	ADJ
fcis-30157	542	17	reasoning	reasoning	NOUN
fcis-30157	542	18	.	.	PUNCT
fcis-30157	543	1	in	in	ADP
fcis-30157	543	2	rural	rural	ADJ
fcis-30157	543	3	india	india	PROPN
fcis-30157	543	4	,	,	PUNCT
fcis-30157	543	5	the	the	DET
fcis-30157	543	6	framework	framework	NOUN
fcis-30157	543	7	led	lead	VERB
fcis-30157	543	8	to	to	ADP
fcis-30157	543	9	a	a	DET
fcis-30157	543	10	30	30	NUM
fcis-30157	543	11	%	%	NOUN
fcis-30157	543	12	reduction	reduction	NOUN
fcis-30157	543	13	in	in	ADP
fcis-30157	543	14	referral	referral	ADJ
fcis-30157	543	15	rates	rate	NOUN
fcis-30157	543	16	to	to	ADP
fcis-30157	543	17	urban	urban	ADJ
fcis-30157	543	18	hospitals	hospital	NOUN
fcis-30157	543	19	,	,	PUNCT
fcis-30157	543	20	highlighting	highlight	VERB
fcis-30157	543	21	its	its	PRON
fcis-30157	543	22	ability	ability	NOUN
fcis-30157	543	23	to	to	PART
fcis-30157	543	24	improve	improve	VERB
fcis-30157	543	25	patient	patient	ADJ
fcis-30157	543	26	outcomes	outcome	NOUN
fcis-30157	543	27	and	and	CCONJ
fcis-30157	543	28	reduce	reduce	VERB
fcis-30157	543	29	healthcare	healthcare	NOUN
fcis-30157	543	30	system	system	NOUN
fcis-30157	543	31	burdens	burden	NOUN
fcis-30157	543	32	.	.	PUNCT
fcis-30157	544	1	however	however	ADV
fcis-30157	544	2	,	,	PUNCT
fcis-30157	544	3	challenges	challenge	NOUN
fcis-30157	544	4	remain	remain	VERB
fcis-30157	544	5	,	,	PUNCT
fcis-30157	544	6	such	such	ADJ
fcis-30157	544	7	as	as	ADP
fcis-30157	544	8	the	the	DET
fcis-30157	544	9	need	need	NOUN
fcis-30157	544	10	for	for	ADP
fcis-30157	544	11	high	high	ADJ
fcis-30157	544	12	-	-	PUNCT
fcis-30157	544	13	quality	quality	NOUN
fcis-30157	544	14	training	training	NOUN
fcis-30157	544	15	data	datum	NOUN
fcis-30157	544	16	and	and	CCONJ
fcis-30157	544	17	integration	integration	NOUN
fcis-30157	544	18	into	into	ADP
fcis-30157	544	19	existing	exist	VERB
fcis-30157	544	20	clinical	clinical	ADJ
fcis-30157	544	21	workflows	workflow	NOUN
fcis-30157	544	22	.	.	PUNCT
fcis-30157	545	1	to	to	PART
fcis-30157	545	2	address	address	VERB
fcis-30157	545	3	this	this	PRON
fcis-30157	545	4	,	,	PUNCT
fcis-30157	545	5	the	the	DET
fcis-30157	545	6	system	system	NOUN
fcis-30157	545	7	was	be	AUX
fcis-30157	545	8	designed	design	VERB
fcis-30157	545	9	to	to	PART
fcis-30157	545	10	complement	complement	VERB
fcis-30157	545	11	rather	rather	ADV
fcis-30157	545	12	than	than	ADP
fcis-30157	545	13	replace	replace	VERB
fcis-30157	545	14	human	human	ADJ
fcis-30157	545	15	judgment	judgment	NOUN
fcis-30157	545	16	,	,	PUNCT
fcis-30157	545	17	increasing	increase	VERB
fcis-30157	545	18	its	its	PRON
fcis-30157	545	19	adoption	adoption	NOUN
fcis-30157	545	20	in	in	ADP
fcis-30157	545	21	clinical	clinical	ADJ
fcis-30157	545	22	practice	practice	NOUN
fcis-30157	545	23	.	.	PUNCT
fcis-30157	546	1	4.1.3	4.1.3	X
fcis-30157	546	2	.	.	PUNCT
fcis-30157	547	1	overall	overall	ADJ
fcis-30157	547	2	contributions	contribution	NOUN
fcis-30157	547	3	of	of	ADP
fcis-30157	547	4	the	the	DET
fcis-30157	547	5	research	research	NOUN
fcis-30157	547	6	this	this	DET
fcis-30157	547	7	study	study	NOUN
fcis-30157	547	8	achieved	achieve	VERB
fcis-30157	547	9	significant	significant	ADJ
fcis-30157	547	10	breakthroughs	breakthrough	NOUN
fcis-30157	547	11	in	in	ADP
fcis-30157	547	12	the	the	DET
fcis-30157	547	13	accuracy	accuracy	NOUN
fcis-30157	547	14	and	and	CCONJ
fcis-30157	547	15	clinical	clinical	ADJ
fcis-30157	547	16	applicability	applicability	NOUN
fcis-30157	547	17	of	of	ADP
fcis-30157	547	18	skin	skin	NOUN
fcis-30157	547	19	disease	disease	NOUN
fcis-30157	547	20	diagnosis	diagnosis	NOUN
fcis-30157	547	21	through	through	ADP
fcis-30157	547	22	an	an	DET
fcis-30157	547	23	adaptive	adaptive	ADJ
fcis-30157	547	24	neural	neural	ADJ
fcis-30157	547	25	network	network	NOUN
fcis-30157	547	26	architecture	architecture	NOUN
fcis-30157	547	27	that	that	PRON
fcis-30157	547	28	integrates	integrate	VERB
fcis-30157	547	29	multimodal	multimodal	NOUN
fcis-30157	547	30	data	datum	NOUN
fcis-30157	547	31	.	.	PUNCT
fcis-30157	548	1	the	the	DET
fcis-30157	548	2	proposed	propose	VERB
fcis-30157	548	3	cross	cross	ADJ
fcis-30157	548	4	-	-	ADJ
fcis-30157	548	5	modal	modal	ADJ
fcis-30157	548	6	feature	feature	NOUN
fcis-30157	548	7	alignment	alignment	NOUN
fcis-30157	548	8	algorithm	algorithm	NOUN
fcis-30157	548	9	improved	improve	VERB
fcis-30157	548	10	the	the	DET
fcis-30157	548	11	distinction	distinction	NOUN
fcis-30157	548	12	accuracy	accuracy	NOUN
fcis-30157	548	13	between	between	ADP
fcis-30157	548	14	eczema	eczema	NOUN
fcis-30157	548	15	and	and	CCONJ
fcis-30157	548	16	psoriasis	psoriasis	NOUN
fcis-30157	548	17	from	from	ADP
fcis-30157	548	18	72.3	72.3	NUM
fcis-30157	548	19	%	%	NOUN
fcis-30157	548	20	to	to	ADP
fcis-30157	548	21	89.7	89.7	NUM
fcis-30157	548	22	%	%	NOUN
fcis-30157	548	23	.	.	PUNCT
fcis-30157	549	1	to	to	PART
fcis-30157	549	2	address	address	VERB
fcis-30157	549	3	blurred	blurred	ADJ
fcis-30157	549	4	lesions	lesion	NOUN
fcis-30157	549	5	in	in	ADP
fcis-30157	549	6	melanoma	melanoma	ADJ
fcis-30157	549	7	diagnosis	diagnosis	NOUN
fcis-30157	549	8	,	,	PUNCT
fcis-30157	549	9	a	a	DET
fcis-30157	549	10	dynamic	dynamic	ADJ
fcis-30157	549	11	attention	attention	NOUN
fcis-30157	549	12	mechanism	mechanism	NOUN
fcis-30157	549	13	achieved	achieve	VERB
fcis-30157	549	14	an	an	DET
fcis-30157	549	15	auc	auc	NOUN
fcis-30157	549	16	of	of	ADP
fcis-30157	549	17	0.912	0.912	NUM
fcis-30157	549	18	in	in	ADP
fcis-30157	549	19	the	the	DET
fcis-30157	549	20	isic	isic	PROPN
fcis-30157	549	21	2020	2020	NUM
fcis-30157	549	22	challenge	challenge	NOUN
fcis-30157	549	23	,	,	PUNCT
fcis-30157	549	24	a	a	DET
fcis-30157	549	25	14.6	14.6	NUM
fcis-30157	549	26	%	%	NOUN
fcis-30157	549	27	improvement	improvement	NOUN
fcis-30157	549	28	.	.	PUNCT
fcis-30157	550	1	aligning	align	VERB
fcis-30157	550	2	patient	patient	ADJ
fcis-30157	550	3	history	history	NOUN
fcis-30157	550	4	data	datum	NOUN
fcis-30157	550	5	with	with	ADP
fcis-30157	550	6	dermatoscope	dermatoscope	NOUN
fcis-30157	550	7	images	image	NOUN
fcis-30157	550	8	corrected	correct	VERB
fcis-30157	550	9	17	17	NUM
fcis-30157	550	10	cases	case	NOUN
fcis-30157	550	11	of	of	ADP
fcis-30157	550	12	seborrheic	seborrheic	ADJ
fcis-30157	550	13	keratosis	keratosis	NOUN
fcis-30157	550	14	misdiagnosed	misdiagnose	VERB
fcis-30157	550	15	as	as	ADP
fcis-30157	550	16	basal	basal	ADJ
fcis-30157	550	17	cell	cell	NOUN
fcis-30157	550	18	carcinoma	carcinoma	NOUN
fcis-30157	550	19	in	in	ADP
fcis-30157	550	20	302	302	NUM
fcis-30157	550	21	clinical	clinical	ADJ
fcis-30157	550	22	verifications	verification	NOUN
fcis-30157	550	23	.	.	PUNCT
fcis-30157	551	1	in	in	ADP
fcis-30157	551	2	an	an	DET
fcis-30157	551	3	8	8	NUM
fcis-30157	551	4	-	-	PUNCT
fcis-30157	551	5	month	month	NOUN
fcis-30157	551	6	trial	trial	NOUN
fcis-30157	551	7	at	at	ADP
fcis-30157	551	8	shanghai	shanghai	PROPN
fcis-30157	551	9	huashan	huashan	PROPN
fcis-30157	551	10	hospital	hospital	PROPN
fcis-30157	551	11	,	,	PUNCT
fcis-30157	551	12	the	the	DET
fcis-30157	551	13	auxiliary	auxiliary	ADJ
fcis-30157	551	14	decision	decision	NOUN
fcis-30157	551	15	-	-	PUNCT
fcis-30157	551	16	making	make	VERB
fcis-30157	551	17	system	system	NOUN
fcis-30157	551	18	reduced	reduce	VERB
fcis-30157	551	19	the	the	DET
fcis-30157	551	20	average	average	ADJ
fcis-30157	551	21	diagnosis	diagnosis	NOUN
fcis-30157	551	22	time	time	NOUN
fcis-30157	551	23	from	from	ADP
fcis-30157	551	24	12.6	12.6	NUM
fcis-30157	551	25	minutes	minute	NOUN
fcis-30157	551	26	to	to	ADP
fcis-30157	551	27	7.8	7.8	NUM
fcis-30157	551	28	minutes	minute	NOUN
fcis-30157	551	29	and	and	CCONJ
fcis-30157	551	30	increased	increase	VERB
fcis-30157	551	31	diagnostic	diagnostic	ADJ
fcis-30157	551	32	compliance	compliance	NOUN
fcis-30157	551	33	in	in	ADP
fcis-30157	551	34	psoriasis	psoriasis	NOUN
fcis-30157	551	35	classification	classification	NOUN
fcis-30157	551	36	from	from	ADP
fcis-30157	551	37	64.2	64.2	NUM
fcis-30157	551	38	%	%	NOUN
fcis-30157	551	39	to	to	ADP
fcis-30157	551	40	83.5	83.5	NUM
fcis-30157	551	41	%	%	NOUN
fcis-30157	551	42	.	.	PUNCT
fcis-30157	552	1	a	a	DET
fcis-30157	552	2	transfer	transfer	NOUN
fcis-30157	552	3	learning	learn	VERB
fcis-30157	552	4	framework	framework	NOUN
fcis-30157	552	5	maintained	maintain	VERB
fcis-30157	552	6	87.4	87.4	NUM
fcis-30157	552	7	%	%	NOUN
fcis-30157	552	8	accuracy	accuracy	NOUN
fcis-30157	552	9	in	in	ADP
fcis-30157	552	10	cross	cross	ADJ
fcis-30157	552	11	-	-	ADJ
fcis-30157	552	12	device	device	ADJ
fcis-30157	552	13	testing	testing	NOUN
fcis-30157	552	14	,	,	PUNCT
fcis-30157	552	15	and	and	CCONJ
fcis-30157	552	16	a	a	DET
fcis-30157	552	17	balanced	balanced	ADJ
fcis-30157	552	18	training	training	NOUN
fcis-30157	552	19	dataset	dataset	NOUN
fcis-30157	552	20	improved	improve	VERB
fcis-30157	552	21	diagnostic	diagnostic	ADJ
fcis-30157	552	22	specificity	specificity	NOUN
fcis-30157	552	23	in	in	ADP
fcis-30157	552	24	dark	dark	ADJ
fcis-30157	552	25	-	-	PUNCT
fcis-30157	552	26	skinned	skin	VERB
fcis-30157	552	27	patients	patient	NOUN
fcis-30157	552	28	to	to	ADP
fcis-30157	552	29	91.2	91.2	NUM
fcis-30157	552	30	%	%	NOUN
fcis-30157	552	31	,	,	PUNCT
fcis-30157	552	32	a	a	DET
fcis-30157	552	33	19.3	19.3	NUM
fcis-30157	552	34	%	%	NOUN
fcis-30157	552	35	increase	increase	NOUN
fcis-30157	552	36	.	.	PUNCT
fcis-30157	553	1	in	in	ADP
fcis-30157	553	2	a	a	DET
fcis-30157	553	3	randomized	randomized	ADJ
fcis-30157	553	4	controlled	control	VERB
fcis-30157	553	5	trial	trial	NOUN
fcis-30157	553	6	in	in	ADP
fcis-30157	553	7	12	12	NUM
fcis-30157	553	8	community	community	NOUN
fcis-30157	553	9	hospitals	hospital	NOUN
fcis-30157	553	10	in	in	ADP
fcis-30157	553	11	zhejiang	zhejiang	PROPN
fcis-30157	553	12	province	province	PROPN
fcis-30157	553	13	,	,	PUNCT
fcis-30157	553	14	mobile	mobile	ADJ
fcis-30157	553	15	diagnostic	diagnostic	ADJ
fcis-30157	553	16	equipment	equipment	NOUN
fcis-30157	553	17	improved	improve	VERB
fcis-30157	553	18	first	first	ADJ
fcis-30157	553	19	diagnosis	diagnosis	NOUN
fcis-30157	553	20	accuracy	accuracy	NOUN
fcis-30157	553	21	of	of	ADP
fcis-30157	553	22	skin	skin	NOUN
fcis-30157	553	23	infectious	infectious	ADJ
fcis-30157	553	24	diseases	disease	NOUN
fcis-30157	553	25	from	from	ADP
fcis-30157	553	26	58.7	58.7	NUM
fcis-30157	553	27	%	%	NOUN
fcis-30157	553	28	to	to	ADP
fcis-30157	553	29	79.4	79.4	NUM
fcis-30157	553	30	%	%	NOUN
fcis-30157	553	31	and	and	CCONJ
fcis-30157	553	32	reduced	reduce	VERB
fcis-30157	553	33	the	the	DET
fcis-30157	553	34	scabies	scabie	NOUN
fcis-30157	553	35	and	and	CCONJ
fcis-30157	553	36	eczema	eczema	VERB
fcis-30157	553	37	misdiagnosis	misdiagnosis	NOUN
fcis-30157	553	38	rate	rate	NOUN
fcis-30157	553	39	by	by	ADP
fcis-30157	553	40	62	62	NUM
fcis-30157	553	41	%	%	NOUN
fcis-30157	553	42	.	.	PUNCT
fcis-30157	554	1	the	the	DET
fcis-30157	554	2	distributed	distribute	VERB
fcis-30157	554	3	learning	learning	NOUN
fcis-30157	554	4	framework	framework	NOUN
fcis-30157	554	5	enabled	enable	VERB
fcis-30157	554	6	grassroots	grassroots	ADJ
fcis-30157	554	7	units	unit	NOUN
fcis-30157	554	8	to	to	PART
fcis-30157	554	9	achieve	achieve	VERB
fcis-30157	554	10	85	85	NUM
fcis-30157	554	11	%	%	NOUN
fcis-30157	554	12	diagnostic	diagnostic	ADJ
fcis-30157	554	13	accuracy	accuracy	NOUN
fcis-30157	554	14	of	of	ADP
fcis-30157	554	15	grade	grade	NOUN
fcis-30157	554	16	a	a	DET
fcis-30157	554	17	hospitals	hospital	NOUN
fcis-30157	554	18	with	with	ADP
fcis-30157	554	19	only	only	ADV
fcis-30157	554	20	500	500	NUM
fcis-30157	554	21	local	local	ADJ
fcis-30157	554	22	training	training	NOUN
fcis-30157	554	23	data	datum	NOUN
fcis-30157	554	24	.	.	PUNCT
fcis-30157	555	1	the	the	DET
fcis-30157	555	2	model	model	NOUN
fcis-30157	555	3	's	's	PART
fcis-30157	555	4	interpretability	interpretability	NOUN
fcis-30157	555	5	,	,	PUNCT
fcis-30157	555	6	based	base	VERB
fcis-30157	555	7	on	on	ADP
fcis-30157	555	8	gradientweighted	gradientweighted	ADJ
fcis-30157	555	9	activation	activation	NOUN
fcis-30157	555	10	maps	map	NOUN
fcis-30157	555	11	,	,	PUNCT
fcis-30157	555	12	achieved	achieve	VERB
fcis-30157	555	13	an	an	DET
fcis-30157	555	14	82.4	82.4	NUM
fcis-30157	555	15	%	%	NOUN
fcis-30157	555	16	overlap	overlap	NOUN
fcis-30157	555	17	with	with	ADP
fcis-30157	555	18	dermatologist	dermatologist	NOUN
fcis-30157	555	19	labels	label	NOUN
fcis-30157	555	20	,	,	PUNCT
fcis-30157	555	21	increasing	increase	VERB
fcis-30157	555	22	physician	physician	NOUN
fcis-30157	555	23	acceptance	acceptance	NOUN
fcis-30157	555	24	to	to	ADP
fcis-30157	555	25	79.8	79.8	NUM
fcis-30157	555	26	%	%	NOUN
fcis-30157	555	27	.	.	PUNCT
fcis-30157	556	1	a	a	DET
fcis-30157	556	2	third	third	ADJ
fcis-30157	556	3	-	-	PUNCT
fcis-30157	556	4	level	level	NOUN
fcis-30157	556	5	risk	risk	NOUN
fcis-30157	556	6	warning	warning	NOUN
fcis-30157	556	7	mechanism	mechanism	NOUN
fcis-30157	556	8	intercepted	intercept	VERB
fcis-30157	556	9	all	all	DET
fcis-30157	556	10	6	6	NUM
fcis-30157	556	11	misdiagnosed	misdiagnose	VERB
fcis-30157	556	12	systemic	systemic	ADJ
fcis-30157	556	13	lupus	lupus	PROPN
fcis-30157	556	14	erythematosus	erythematosus	PROPN
fcis-30157	556	15	cases	case	NOUN
fcis-30157	556	16	in	in	ADP
fcis-30157	556	17	clinical	clinical	ADJ
fcis-30157	556	18	trials	trial	NOUN
fcis-30157	556	19	.	.	PUNCT
fcis-30157	557	1	technological	technological	ADJ
fcis-30157	557	2	innovations	innovation	NOUN
fcis-30157	557	3	also	also	ADV
fcis-30157	557	4	supported	support	VERB
fcis-30157	557	5	multi	multi	ADJ
fcis-30157	557	6	-	-	ADJ
fcis-30157	557	7	task	task	ADJ
fcis-30157	557	8	learning	learning	NOUN
fcis-30157	557	9	,	,	PUNCT
fcis-30157	557	10	achieving	achieve	VERB
fcis-30157	557	11	both	both	DET
fcis-30157	557	12	disease	disease	NOUN
fcis-30157	557	13	classification	classification	NOUN
fcis-30157	557	14	and	and	CCONJ
fcis-30157	557	15	treatment	treatment	NOUN
fcis-30157	557	16	recommendations	recommendation	NOUN
fcis-30157	557	17	.	.	PUNCT
fcis-30157	558	1	for	for	ADP
fcis-30157	558	2	psoriasis	psoriasis	NOUN
fcis-30157	558	3	,	,	PUNCT
fcis-30157	558	4	the	the	DET
fcis-30157	558	5	system	system	NOUN
fcis-30157	558	6	's	's	PART
fcis-30157	558	7	treatment	treatment	NOUN
fcis-30157	558	8	timing	timing	NOUN
fcis-30157	558	9	recommendations	recommendation	NOUN
fcis-30157	558	10	matched	match	VERB
fcis-30157	558	11	expert	expert	ADJ
fcis-30157	558	12	consensus	consensus	NOUN
fcis-30157	558	13	.	.	PUNCT
fcis-30157	559	1	the	the	DET
fcis-30157	559	2	recurrence	recurrence	NOUN
fcis-30157	559	3	warning	warning	NOUN
fcis-30157	559	4	accuracy	accuracy	NOUN
fcis-30157	559	5	for	for	ADP
fcis-30157	559	6	atopic	atopic	NOUN
fcis-30157	559	7	dermatitis	dermatitis	NOUN
fcis-30157	559	8	increased	increase	VERB
fcis-30157	559	9	to	to	ADP
fcis-30157	559	10	76.8	76.8	NUM
fcis-30157	559	11	%	%	NOUN
fcis-30157	559	12	,	,	PUNCT
fcis-30157	559	13	a	a	DET
fcis-30157	559	14	29.2	29.2	NUM
fcis-30157	559	15	%	%	NOUN
fcis-30157	559	16	improvement	improvement	NOUN
fcis-30157	559	17	over	over	ADP
fcis-30157	559	18	traditional	traditional	ADJ
fcis-30157	559	19	models	model	NOUN
fcis-30157	559	20	.	.	PUNCT
fcis-30157	560	1	despite	despite	SCONJ
fcis-30157	560	2	these	these	DET
fcis-30157	560	3	advancements	advancement	NOUN
fcis-30157	560	4	,	,	PUNCT
fcis-30157	560	5	the	the	DET
fcis-30157	560	6	study	study	NOUN
fcis-30157	560	7	cautions	caution	VERB
fcis-30157	560	8	that	that	SCONJ
fcis-30157	560	9	over	over	NOUN
fcis-30157	560	10	-	-	PUNCT
fcis-30157	560	11	reliance	reliance	NOUN
fcis-30157	560	12	on	on	ADP
fcis-30157	560	13	technology	technology	NOUN
fcis-30157	560	14	may	may	AUX
fcis-30157	560	15	weaken	weaken	VERB
fcis-30157	560	16	clinical	clinical	ADJ
fcis-30157	560	17	thinking	thinking	NOUN
fcis-30157	560	18	,	,	PUNCT
fcis-30157	560	19	particularly	particularly	ADV
fcis-30157	560	20	in	in	ADP
fcis-30157	560	21	diagnosing	diagnose	VERB
fcis-30157	560	22	rare	rare	ADJ
fcis-30157	560	23	diseases	disease	NOUN
fcis-30157	560	24	,	,	PUNCT
fcis-30157	560	25	emphasizing	emphasize	VERB
fcis-30157	560	26	the	the	DET
fcis-30157	560	27	need	need	NOUN
fcis-30157	560	28	for	for	ADP
fcis-30157	560	29	clearly	clearly	ADV
fcis-30157	560	30	defined	define	VERB
fcis-30157	560	31	human	human	ADJ
fcis-30157	560	32	-	-	PUNCT
fcis-30157	560	33	computer	computer	NOUN
fcis-30157	560	34	collaboration	collaboration	NOUN
fcis-30157	560	35	boundaries	boundary	NOUN
fcis-30157	560	36	.	.	PUNCT
fcis-30157	561	1	4.2	4.2	NUM
fcis-30157	561	2	.	.	PUNCT
fcis-30157	561	3	implications	implication	NOUN
fcis-30157	561	4	and	and	CCONJ
fcis-30157	561	5	future	future	ADJ
fcis-30157	561	6	work	work	NOUN
fcis-30157	561	7	4.2.1	4.2.1	NUM
fcis-30157	561	8	.	.	PUNCT
fcis-30157	562	1	clinical	clinical	ADJ
fcis-30157	562	2	implications	implication	NOUN
fcis-30157	562	3	of	of	ADP
fcis-30157	562	4	the	the	DET
fcis-30157	562	5	study	study	NOUN
fcis-30157	562	6	this	this	DET
fcis-30157	562	7	study	study	NOUN
fcis-30157	562	8	has	have	VERB
fcis-30157	562	9	significant	significant	ADJ
fcis-30157	562	10	clinical	clinical	ADJ
fcis-30157	562	11	implications	implication	NOUN
fcis-30157	562	12	,	,	PUNCT
fcis-30157	562	13	particularly	particularly	ADV
fcis-30157	562	14	in	in	ADP
fcis-30157	562	15	integrating	integrate	VERB
fcis-30157	562	16	automatic	automatic	ADJ
fcis-30157	562	17	identification	identification	NOUN
fcis-30157	562	18	and	and	CCONJ
fcis-30157	562	19	auxiliary	auxiliary	ADJ
fcis-30157	562	20	diagnostic	diagnostic	ADJ
fcis-30157	562	21	techniques	technique	NOUN
fcis-30157	562	22	into	into	ADP
fcis-30157	562	23	real	real	ADJ
fcis-30157	562	24	-	-	PUNCT
fcis-30157	562	25	world	world	NOUN
fcis-30157	562	26	clinical	clinical	ADJ
fcis-30157	562	27	practice	practice	NOUN
fcis-30157	562	28	.	.	PUNCT
fcis-30157	563	1	a	a	DET
fcis-30157	563	2	study	study	NOUN
fcis-30157	563	3	demonstrated	demonstrate	VERB
fcis-30157	563	4	that	that	SCONJ
fcis-30157	563	5	a	a	DET
fcis-30157	563	6	multimodal	multimodal	ADJ
fcis-30157	563	7	fusion	fusion	NOUN
fcis-30157	563	8	method	method	NOUN
fcis-30157	563	9	combining	combine	VERB
fcis-30157	563	10	skin	skin	NOUN
fcis-30157	563	11	images	image	NOUN
fcis-30157	563	12	and	and	CCONJ
fcis-30157	563	13	clinical	clinical	ADJ
fcis-30157	563	14	data	datum	NOUN
fcis-30157	563	15	achieved	achieve	VERB
fcis-30157	563	16	87	87	NUM
fcis-30157	563	17	%	%	NOUN
fcis-30157	563	18	accuracy	accuracy	NOUN
fcis-30157	563	19	in	in	ADP
fcis-30157	563	20	skin	skin	NOUN
fcis-30157	563	21	cancer	cancer	NOUN
fcis-30157	563	22	classification	classification	NOUN
fcis-30157	563	23	,	,	PUNCT
fcis-30157	563	24	matching	match	VERB
fcis-30157	563	25	experienced	experienced	ADJ
fcis-30157	563	26	dermatologists	dermatologist	NOUN
fcis-30157	563	27	.	.	PUNCT
fcis-30157	564	1	this	this	PRON
fcis-30157	564	2	suggests	suggest	VERB
fcis-30157	564	3	that	that	SCONJ
fcis-30157	564	4	such	such	ADJ
fcis-30157	564	5	technologies	technology	NOUN
fcis-30157	564	6	can	can	AUX
fcis-30157	564	7	be	be	AUX
fcis-30157	564	8	seamlessly	seamlessly	ADV
fcis-30157	564	9	integrated	integrate	VERB
fcis-30157	564	10	into	into	ADP
fcis-30157	564	11	current	current	ADJ
fcis-30157	564	12	medical	medical	ADJ
fcis-30157	564	13	practices	practice	NOUN
fcis-30157	564	14	.	.	PUNCT
fcis-30157	565	1	integration	integration	NOUN
fcis-30157	565	2	involves	involve	VERB
fcis-30157	565	3	several	several	ADJ
fcis-30157	565	4	key	key	ADJ
fcis-30157	565	5	steps	step	NOUN
fcis-30157	565	6	.	.	PUNCT
fcis-30157	566	1	the	the	DET
fcis-30157	566	2	system	system	NOUN
fcis-30157	566	3	must	must	AUX
fcis-30157	566	4	be	be	AUX
fcis-30157	566	5	user	user	NOUN
fcis-30157	566	6	-	-	PUNCT
fcis-30157	566	7	friendly	friendly	ADJ
fcis-30157	566	8	,	,	PUNCT
fcis-30157	566	9	allowing	allow	VERB
fcis-30157	566	10	physicians	physician	NOUN
fcis-30157	566	11	to	to	PART
fcis-30157	566	12	input	input	VERB
fcis-30157	566	13	data	datum	NOUN
fcis-30157	566	14	and	and	CCONJ
fcis-30157	566	15	receive	receive	VERB
fcis-30157	566	16	diagnostic	diagnostic	ADJ
fcis-30157	566	17	suggestions	suggestion	NOUN
fcis-30157	566	18	within	within	ADP
fcis-30157	566	19	minutes	minute	NOUN
fcis-30157	566	20	.	.	PUNCT
fcis-30157	567	1	for	for	ADP
fcis-30157	567	2	example	example	NOUN
fcis-30157	567	3	,	,	PUNCT
fcis-30157	567	4	a	a	DET
fcis-30157	567	5	deep	deep	ADJ
fcis-30157	567	6	learning	learning	NOUN
fcis-30157	567	7	model	model	NOUN
fcis-30157	567	8	for	for	ADP
fcis-30157	567	9	skin	skin	NOUN
fcis-30157	567	10	lesion	lesion	NOUN
fcis-30157	567	11	classification	classification	NOUN
fcis-30157	567	12	can	can	AUX
fcis-30157	567	13	process	process	VERB
fcis-30157	567	14	dermoscopy	dermoscopy	VERB
fcis-30157	567	15	images	image	NOUN
fcis-30157	567	16	in	in	ADP
fcis-30157	567	17	under	under	ADP
fcis-30157	567	18	30	30	NUM
fcis-30157	567	19	seconds	second	NOUN
fcis-30157	567	20	.	.	PUNCT
fcis-30157	568	1	additionally	additionally	ADV
fcis-30157	568	2	,	,	PUNCT
fcis-30157	568	3	the	the	DET
fcis-30157	568	4	system	system	NOUN
fcis-30157	568	5	must	must	AUX
fcis-30157	568	6	provide	provide	VERB
fcis-30157	568	7	explainable	explainable	ADJ
fcis-30157	568	8	results	result	NOUN
fcis-30157	568	9	,	,	PUNCT
fcis-30157	568	10	helping	help	VERB
fcis-30157	568	11	physicians	physician	NOUN
fcis-30157	568	12	understand	understand	VERB
fcis-30157	568	13	the	the	DET
fcis-30157	568	14	rationale	rationale	NOUN
fcis-30157	568	15	behind	behind	ADP
fcis-30157	568	16	diagnoses	diagnosis	NOUN
fcis-30157	568	17	,	,	PUNCT
fcis-30157	568	18	especially	especially	ADV
fcis-30157	568	19	when	when	SCONJ
fcis-30157	568	20	the	the	DET
fcis-30157	568	21	system	system	NOUN
fcis-30157	568	22	's	's	PART
fcis-30157	568	23	suggestions	suggestion	NOUN
fcis-30157	568	24	differ	differ	VERB
fcis-30157	568	25	from	from	ADP
fcis-30157	568	26	the	the	DET
fcis-30157	568	27	physician	physician	NOUN
fcis-30157	568	28	’s	’s	PART
fcis-30157	568	29	initial	initial	ADJ
fcis-30157	568	30	assessment	assessment	NOUN
fcis-30157	568	31	.	.	PUNCT
fcis-30157	569	1	these	these	DET
fcis-30157	569	2	technologies	technology	NOUN
fcis-30157	569	3	significantly	significantly	ADV
fcis-30157	569	4	aid	aid	VERB
fcis-30157	569	5	physician	physician	NOUN
fcis-30157	569	6	decision	decision	NOUN
fcis-30157	569	7	-	-	PUNCT
fcis-30157	569	8	making	making	NOUN
fcis-30157	569	9	by	by	ADP
fcis-30157	569	10	automating	automate	VERB
fcis-30157	569	11	the	the	DET
fcis-30157	569	12	initial	initial	ADJ
fcis-30157	569	13	screening	screening	NOUN
fcis-30157	569	14	process	process	NOUN
fcis-30157	569	15	,	,	PUNCT
fcis-30157	569	16	reducing	reduce	VERB
fcis-30157	569	17	cognitive	cognitive	ADJ
fcis-30157	569	18	load	load	NOUN
fcis-30157	569	19	,	,	PUNCT
fcis-30157	569	20	and	and	CCONJ
fcis-30157	569	21	enabling	enable	VERB
fcis-30157	569	22	focus	focus	NOUN
fcis-30157	569	23	on	on	ADP
fcis-30157	569	24	complex	complex	ADJ
fcis-30157	569	25	cases	case	NOUN
fcis-30157	569	26	.	.	PUNCT
fcis-30157	570	1	for	for	ADP
fcis-30157	570	2	instance	instance	NOUN
fcis-30157	570	3	,	,	PUNCT
fcis-30157	570	4	a	a	DET
fcis-30157	570	5	machine	machine	NOUN
fcis-30157	570	6	learning	learn	VERB
fcis-30157	570	7	approach	approach	NOUN
fcis-30157	570	8	achieved	achieve	VERB
fcis-30157	570	9	92	92	NUM
fcis-30157	570	10	%	%	NOUN
fcis-30157	570	11	precision	precision	NOUN
fcis-30157	570	12	in	in	ADP
fcis-30157	570	13	detecting	detect	VERB
fcis-30157	570	14	early	early	ADJ
fcis-30157	570	15	signs	sign	NOUN
fcis-30157	570	16	of	of	ADP
fcis-30157	570	17	skin	skin	NOUN
fcis-30157	570	18	diseases	disease	NOUN
fcis-30157	570	19	,	,	PUNCT
fcis-30157	570	20	crucial	crucial	ADJ
fcis-30157	570	21	for	for	ADP
fcis-30157	570	22	timely	timely	ADJ
fcis-30157	570	23	treatment	treatment	NOUN
fcis-30157	570	24	of	of	ADP
fcis-30157	570	25	melanoma	melanoma	NOUN
fcis-30157	570	26	.	.	PUNCT
fcis-30157	571	1	such	such	ADJ
fcis-30157	571	2	systems	system	NOUN
fcis-30157	571	3	can	can	AUX
fcis-30157	571	4	also	also	ADV
fcis-30157	571	5	improve	improve	VERB
fcis-30157	571	6	diagnostic	diagnostic	ADJ
fcis-30157	571	7	accuracy	accuracy	NOUN
fcis-30157	571	8	across	across	ADP
fcis-30157	571	9	different	different	ADJ
fcis-30157	571	10	skin	skin	NOUN
fcis-30157	571	11	tones	tone	NOUN
fcis-30157	571	12	,	,	PUNCT
fcis-30157	571	13	where	where	SCONJ
fcis-30157	571	14	traditional	traditional	ADJ
fcis-30157	571	15	methods	method	NOUN
fcis-30157	571	16	often	often	ADV
fcis-30157	571	17	struggle	struggle	VERB
fcis-30157	571	18	,	,	PUNCT
fcis-30157	571	19	as	as	SCONJ
fcis-30157	571	20	demonstrated	demonstrate	VERB
fcis-30157	571	21	by	by	ADP
fcis-30157	571	22	a	a	DET
fcis-30157	571	23	study	study	NOUN
fcis-30157	571	24	using	use	VERB
fcis-30157	571	25	a	a	DET
fcis-30157	571	26	deep	deep	ADJ
fcis-30157	571	27	learning	learning	NOUN
fcis-30157	571	28	-	-	PUNCT
fcis-30157	571	29	aided	aid	VERB
fcis-30157	571	30	decision	decision	NOUN
fcis-30157	571	31	support	support	NOUN
fcis-30157	571	32	system	system	NOUN
fcis-30157	571	33	.	.	PUNCT
fcis-30157	572	1	in	in	ADP
fcis-30157	572	2	resource	resource	NOUN
fcis-30157	572	3	-	-	PUNCT
fcis-30157	572	4	constrained	constrain	VERB
fcis-30157	572	5	settings	setting	NOUN
fcis-30157	572	6	,	,	PUNCT
fcis-30157	572	7	where	where	SCONJ
fcis-30157	572	8	access	access	NOUN
fcis-30157	572	9	to	to	ADP
fcis-30157	572	10	dermatologists	dermatologist	NOUN
fcis-30157	572	11	is	be	AUX
fcis-30157	572	12	limited	limited	ADJ
fcis-30157	572	13	,	,	PUNCT
fcis-30157	572	14	these	these	DET
fcis-30157	572	15	automated	automate	VERB
fcis-30157	572	16	systems	system	NOUN
fcis-30157	572	17	can	can	AUX
fcis-30157	572	18	serve	serve	VERB
fcis-30157	572	19	as	as	ADP
fcis-30157	572	20	a	a	DET
fcis-30157	572	21	first	first	ADJ
fcis-30157	572	22	line	line	NOUN
fcis-30157	572	23	of	of	ADP
fcis-30157	572	24	defense	defense	NOUN
fcis-30157	572	25	.	.	PUNCT
fcis-30157	573	1	a	a	DET
fcis-30157	573	2	mobile	mobile	NOUN
fcis-30157	573	3	-	-	PUNCT
fcis-30157	573	4	based	base	VERB
fcis-30157	573	5	diagnostic	diagnostic	ADJ
fcis-30157	573	6	tool	tool	NOUN
fcis-30157	573	7	in	in	ADP
fcis-30157	573	8	rural	rural	ADJ
fcis-30157	573	9	india	india	PROPN
fcis-30157	573	10	achieved	achieve	VERB
fcis-30157	573	11	85	85	NUM
fcis-30157	573	12	%	%	NOUN
fcis-30157	573	13	accuracy	accuracy	NOUN
fcis-30157	573	14	,	,	PUNCT
fcis-30157	573	15	improving	improve	VERB
fcis-30157	573	16	diagnostic	diagnostic	ADJ
fcis-30157	573	17	outcomes	outcome	NOUN
fcis-30157	573	18	while	while	SCONJ
fcis-30157	573	19	reducing	reduce	VERB
fcis-30157	573	20	referral	referral	ADJ
fcis-30157	573	21	costs	cost	NOUN
fcis-30157	573	22	.	.	PUNCT
fcis-30157	574	1	in	in	ADP
fcis-30157	574	2	advanced	advanced	ADJ
fcis-30157	574	3	healthcare	healthcare	NOUN
fcis-30157	574	4	settings	setting	NOUN
fcis-30157	574	5	,	,	PUNCT
fcis-30157	574	6	these	these	DET
fcis-30157	574	7	systems	system	NOUN
fcis-30157	574	8	complement	complement	NOUN
fcis-30157	574	9	dermatologists	dermatologist	NOUN
fcis-30157	574	10	by	by	ADP
fcis-30157	574	11	providing	provide	VERB
fcis-30157	574	12	second	second	ADJ
fcis-30157	574	13	opinions	opinion	NOUN
fcis-30157	574	14	or	or	CCONJ
fcis-30157	574	15	assisting	assist	VERB
fcis-30157	574	16	in	in	ADP
fcis-30157	574	17	complex	complex	ADJ
fcis-30157	574	18	case	case	NOUN
fcis-30157	574	19	diagnoses	diagnose	VERB
fcis-30157	574	20	.	.	PUNCT
fcis-30157	575	1	a	a	DET
fcis-30157	575	2	novel	novel	ADJ
fcis-30157	575	3	multi	multi	ADJ
fcis-30157	575	4	-	-	ADJ
fcis-30157	575	5	task	task	ADJ
fcis-30157	575	6	model	model	NOUN
fcis-30157	575	7	showed	show	VERB
fcis-30157	575	8	improved	improve	VERB
fcis-30157	575	9	diagnostic	diagnostic	ADJ
fcis-30157	575	10	accuracy	accuracy	NOUN
fcis-30157	575	11	compared	compare	VERB
fcis-30157	575	12	to	to	ADP
fcis-30157	575	13	traditional	traditional	ADJ
fcis-30157	575	14	methods	method	NOUN
fcis-30157	575	15	.	.	PUNCT
fcis-30157	576	1	however	however	ADV
fcis-30157	576	2	,	,	PUNCT
fcis-30157	576	3	the	the	DET
fcis-30157	576	4	integration	integration	NOUN
fcis-30157	576	5	of	of	ADP
fcis-30157	576	6	ai	ai	NOUN
fcis-30157	576	7	in	in	ADP
fcis-30157	576	8	diagnosis	diagnosis	NOUN
fcis-30157	576	9	raises	raise	VERB
fcis-30157	576	10	ethical	ethical	ADJ
fcis-30157	576	11	and	and	CCONJ
fcis-30157	576	12	practical	practical	ADJ
fcis-30157	576	13	concerns	concern	NOUN
fcis-30157	576	14	.	.	PUNCT
fcis-30157	577	1	transparency	transparency	NOUN
fcis-30157	577	2	and	and	CCONJ
fcis-30157	577	3	accountability	accountability	NOUN
fcis-30157	577	4	are	be	AUX
fcis-30157	577	5	crucial	crucial	ADJ
fcis-30157	577	6	,	,	PUNCT
fcis-30157	577	7	as	as	SCONJ
fcis-30157	577	8	ai	ai	PROPN
fcis-30157	577	9	models	model	NOUN
fcis-30157	577	10	must	must	AUX
fcis-30157	577	11	be	be	AUX
fcis-30157	577	12	continuously	continuously	ADV
fcis-30157	577	13	monitored	monitor	VERB
fcis-30157	577	14	and	and	CCONJ
fcis-30157	577	15	validated	validate	VERB
fcis-30157	577	16	to	to	PART
fcis-30157	577	17	ensure	ensure	VERB
fcis-30157	577	18	their	their	PRON
fcis-30157	577	19	reliability	reliability	NOUN
fcis-30157	577	20	in	in	ADP
fcis-30157	577	21	the	the	DET
fcis-30157	577	22	dynamic	dynamic	ADJ
fcis-30157	577	23	field	field	NOUN
fcis-30157	577	24	of	of	ADP
fcis-30157	577	25	skin	skin	NOUN
fcis-30157	577	26	diseases	disease	NOUN
fcis-30157	577	27	.	.	PUNCT
fcis-30157	578	1	these	these	DET
fcis-30157	578	2	technologies	technology	NOUN
fcis-30157	578	3	can	can	AUX
fcis-30157	578	4	greatly	greatly	ADV
fcis-30157	578	5	enhance	enhance	VERB
fcis-30157	578	6	diagnostic	diagnostic	ADJ
fcis-30157	578	7	accuracy	accuracy	NOUN
fcis-30157	578	8	,	,	PUNCT
fcis-30157	578	9	reduce	reduce	VERB
fcis-30157	578	10	physician	physician	NOUN
fcis-30157	578	11	burden	burden	NOUN
fcis-30157	578	12	,	,	PUNCT
fcis-30157	578	13	and	and	CCONJ
fcis-30157	578	14	improve	improve	VERB
fcis-30157	578	15	patient	patient	ADJ
fcis-30157	578	16	outcomes	outcome	NOUN
fcis-30157	578	17	,	,	PUNCT
fcis-30157	578	18	but	but	CCONJ
fcis-30157	578	19	successful	successful	ADJ
fcis-30157	578	20	implementation	implementation	NOUN
fcis-30157	578	21	requires	require	VERB
fcis-30157	578	22	collaboration	collaboration	NOUN
fcis-30157	578	23	between	between	ADP
fcis-30157	578	24	technologists	technologist	NOUN
fcis-30157	578	25	,	,	PUNCT
fcis-30157	578	26	clinicians	clinician	NOUN
fcis-30157	578	27	,	,	PUNCT
fcis-30157	578	28	and	and	CCONJ
fcis-30157	578	29	policymakers	policymaker	NOUN
fcis-30157	578	30	to	to	PART
fcis-30157	578	31	address	address	VERB
fcis-30157	578	32	ethical	ethical	ADJ
fcis-30157	578	33	and	and	CCONJ
fcis-30157	578	34	practical	practical	ADJ
fcis-30157	578	35	challenges	challenge	NOUN
fcis-30157	578	36	.	.	PUNCT
fcis-30157	579	1	4.2.2	4.2.2	X
fcis-30157	579	2	.	.	PUNCT
fcis-30157	579	3	potential	potential	ADJ
fcis-30157	579	4	improvements	improvement	NOUN
fcis-30157	579	5	and	and	CCONJ
fcis-30157	579	6	extensions	extension	NOUN
fcis-30157	579	7	the	the	DET
fcis-30157	579	8	current	current	ADJ
fcis-30157	579	9	research	research	NOUN
fcis-30157	579	10	in	in	ADP
fcis-30157	579	11	automatic	automatic	ADJ
fcis-30157	579	12	skin	skin	NOUN
fcis-30157	579	13	disease	disease	NOUN
fcis-30157	579	14	identification	identification	NOUN
fcis-30157	579	15	and	and	CCONJ
fcis-30157	579	16	auxiliary	auxiliary	ADJ
fcis-30157	579	17	diagnosis	diagnosis	NOUN
fcis-30157	579	18	has	have	AUX
fcis-30157	579	19	made	make	VERB
fcis-30157	579	20	notable	notable	ADJ
fcis-30157	579	21	advances	advance	NOUN
fcis-30157	579	22	,	,	PUNCT
fcis-30157	579	23	but	but	CCONJ
fcis-30157	579	24	optimization	optimization	NOUN
fcis-30157	579	25	is	be	AUX
fcis-30157	579	26	still	still	ADV
fcis-30157	579	27	needed	need	VERB
fcis-30157	579	28	in	in	ADP
fcis-30157	579	29	several	several	ADJ
fcis-30157	579	30	areas	area	NOUN
fcis-30157	579	31	.	.	PUNCT
fcis-30157	580	1	for	for	ADP
fcis-30157	580	2	model	model	NOUN
fcis-30157	580	3	improvement	improvement	NOUN
fcis-30157	580	4	,	,	PUNCT
fcis-30157	580	5	optimizing	optimize	VERB
fcis-30157	580	6	both	both	PRON
fcis-30157	580	7	data	datum	NOUN
fcis-30157	580	8	124	124	NUM
fcis-30157	580	9	quality	quality	NOUN
fcis-30157	580	10	and	and	CCONJ
fcis-30157	580	11	algorithm	algorithm	NOUN
fcis-30157	580	12	architecture	architecture	NOUN
fcis-30157	580	13	is	be	AUX
fcis-30157	580	14	essential	essential	ADJ
fcis-30157	580	15	.	.	PUNCT
fcis-30157	581	1	cross	cross	ADJ
fcis-30157	581	2	-	-	ADJ
fcis-30157	581	3	center	center	ADJ
fcis-30157	581	4	verification	verification	NOUN
fcis-30157	581	5	from	from	ADP
fcis-30157	581	6	the	the	DET
fcis-30157	581	7	isic	isic	PROPN
fcis-30157	581	8	2019	2019	NUM
fcis-30157	581	9	challenge	challenge	NOUN
fcis-30157	581	10	showed	show	VERB
fcis-30157	581	11	that	that	SCONJ
fcis-30157	581	12	training	train	VERB
fcis-30157	581	13	a	a	DET
fcis-30157	581	14	model	model	NOUN
fcis-30157	581	15	on	on	ADP
fcis-30157	581	16	high	high	ADJ
fcis-30157	581	17	-	-	PUNCT
fcis-30157	581	18	definition	definition	NOUN
fcis-30157	581	19	dermatoscope	dermatoscope	NOUN
fcis-30157	581	20	images	image	NOUN
fcis-30157	581	21	from	from	ADP
fcis-30157	581	22	a	a	DET
fcis-30157	581	23	single	single	ADJ
fcis-30157	581	24	device	device	NOUN
fcis-30157	581	25	resulted	result	VERB
fcis-30157	581	26	in	in	ADP
fcis-30157	581	27	a	a	DET
fcis-30157	581	28	14.3	14.3	NUM
fcis-30157	581	29	%	%	NOUN
fcis-30157	581	30	accuracy	accuracy	NOUN
fcis-30157	581	31	drop	drop	NOUN
fcis-30157	581	32	when	when	SCONJ
fcis-30157	581	33	tested	test	VERB
fcis-30157	581	34	with	with	ADP
fcis-30157	581	35	low	low	ADJ
fcis-30157	581	36	-	-	PUNCT
fcis-30157	581	37	resolution	resolution	NOUN
fcis-30157	581	38	images	image	NOUN
fcis-30157	581	39	from	from	ADP
fcis-30157	581	40	smartphones	smartphone	NOUN
fcis-30157	581	41	.	.	PUNCT
fcis-30157	582	1	this	this	PRON
fcis-30157	582	2	emphasizes	emphasize	VERB
fcis-30157	582	3	the	the	DET
fcis-30157	582	4	need	need	NOUN
fcis-30157	582	5	for	for	ADP
fcis-30157	582	6	multi	multi	ADJ
fcis-30157	582	7	-	-	ADJ
fcis-30157	582	8	source	source	NOUN
fcis-30157	582	9	datasets	dataset	NOUN
fcis-30157	582	10	.	.	PUNCT
fcis-30157	583	1	the	the	DET
fcis-30157	583	2	mdfnet	mdfnet	NOUN
fcis-30157	583	3	team	team	NOUN
fcis-30157	583	4	improved	improve	VERB
fcis-30157	583	5	eczema	eczema	NOUN
fcis-30157	583	6	and	and	CCONJ
fcis-30157	583	7	contact	contact	NOUN
fcis-30157	583	8	dermatitis	dermatitis	NOUN
fcis-30157	583	9	differential	differential	ADJ
fcis-30157	583	10	diagnosis	diagnosis	NOUN
fcis-30157	583	11	accuracy	accuracy	NOUN
fcis-30157	583	12	to	to	ADP
fcis-30157	583	13	91.7	91.7	NUM
fcis-30157	583	14	%	%	NOUN
fcis-30157	583	15	by	by	ADP
fcis-30157	583	16	integrating	integrate	VERB
fcis-30157	583	17	dermatoscope	dermatoscope	NOUN
fcis-30157	583	18	images	image	NOUN
fcis-30157	583	19	with	with	ADP
fcis-30157	583	20	structured	structured	ADJ
fcis-30157	583	21	clinical	clinical	ADJ
fcis-30157	583	22	data	datum	NOUN
fcis-30157	583	23	,	,	PUNCT
fcis-30157	583	24	offering	offer	VERB
fcis-30157	583	25	new	new	ADJ
fcis-30157	583	26	directions	direction	NOUN
fcis-30157	583	27	for	for	ADP
fcis-30157	583	28	clinical	clinical	ADJ
fcis-30157	583	29	applicability	applicability	NOUN
fcis-30157	583	30	.	.	PUNCT
fcis-30157	584	1	in	in	ADP
fcis-30157	584	2	feature	feature	NOUN
fcis-30157	584	3	extraction	extraction	NOUN
fcis-30157	584	4	,	,	PUNCT
fcis-30157	584	5	a	a	DET
fcis-30157	584	6	dualattention	dualattention	NOUN
fcis-30157	584	7	network	network	NOUN
fcis-30157	584	8	increased	increase	VERB
fcis-30157	584	9	basal	basal	NOUN
fcis-30157	584	10	cell	cell	NOUN
fcis-30157	584	11	carcinoma	carcinoma	NOUN
fcis-30157	584	12	recognition	recognition	NOUN
fcis-30157	584	13	sensitivity	sensitivity	NOUN
fcis-30157	584	14	to	to	ADP
fcis-30157	584	15	89.5	89.5	NUM
fcis-30157	584	16	%	%	NOUN
fcis-30157	584	17	,	,	PUNCT
fcis-30157	584	18	though	though	SCONJ
fcis-30157	584	19	false	false	ADJ
fcis-30157	584	20	positives	positive	NOUN
fcis-30157	584	21	for	for	ADP
fcis-30157	584	22	pigmentation	pigmentation	NOUN
fcis-30157	584	23	diseases	disease	NOUN
fcis-30157	584	24	remained	remain	VERB
fcis-30157	584	25	at	at	ADP
fcis-30157	584	26	16.2	16.2	NUM
fcis-30157	584	27	%	%	NOUN
fcis-30157	584	28	,	,	PUNCT
fcis-30157	584	29	primarily	primarily	ADV
fcis-30157	584	30	due	due	ADP
fcis-30157	584	31	to	to	ADP
fcis-30157	584	32	skin	skin	NOUN
fcis-30157	584	33	tone	tone	NOUN
fcis-30157	584	34	differences	difference	NOUN
fcis-30157	584	35	.	.	PUNCT
fcis-30157	585	1	a	a	DET
fcis-30157	585	2	study	study	NOUN
fcis-30157	585	3	at	at	ADP
fcis-30157	585	4	harvard	harvard	PROPN
fcis-30157	585	5	medical	medical	PROPN
fcis-30157	585	6	school	school	PROPN
fcis-30157	585	7	found	find	VERB
fcis-30157	585	8	that	that	SCONJ
fcis-30157	585	9	when	when	SCONJ
fcis-30157	585	10	less	less	ADJ
fcis-30157	585	11	than	than	ADP
fcis-30157	585	12	15	15	NUM
fcis-30157	585	13	%	%	NOUN
fcis-30157	585	14	of	of	ADP
fcis-30157	585	15	training	training	NOUN
fcis-30157	585	16	data	datum	NOUN
fcis-30157	585	17	included	include	VERB
fcis-30157	585	18	dark	dark	ADJ
fcis-30157	585	19	-	-	PUNCT
fcis-30157	585	20	skinned	skin	VERB
fcis-30157	585	21	samples	sample	NOUN
fcis-30157	585	22	,	,	PUNCT
fcis-30157	585	23	melanoma	melanoma	NOUN
fcis-30157	585	24	diagnostic	diagnostic	ADJ
fcis-30157	585	25	specificity	specificity	NOUN
fcis-30157	585	26	dropped	drop	VERB
fcis-30157	585	27	by	by	ADP
fcis-30157	585	28	23	23	NUM
fcis-30157	585	29	%	%	NOUN
fcis-30157	585	30	.	.	PUNCT
fcis-30157	586	1	hence	hence	ADV
fcis-30157	586	2	,	,	PUNCT
fcis-30157	586	3	creating	create	VERB
fcis-30157	586	4	a	a	DET
fcis-30157	586	5	demographically	demographically	ADV
fcis-30157	586	6	representative	representative	ADJ
fcis-30157	586	7	dataset	dataset	NOUN
fcis-30157	586	8	is	be	AUX
fcis-30157	586	9	crucial	crucial	ADJ
fcis-30157	586	10	.	.	PUNCT
fcis-30157	587	1	the	the	DET
fcis-30157	587	2	fs3dciot	fs3dciot	NOUN
fcis-30157	587	3	framework	framework	NOUN
fcis-30157	587	4	’s	’s	PART
fcis-30157	587	5	incremental	incremental	ADJ
fcis-30157	587	6	learning	learning	NOUN
fcis-30157	587	7	strategy	strategy	NOUN
fcis-30157	587	8	could	could	AUX
fcis-30157	587	9	dynamically	dynamically	ADV
fcis-30157	587	10	optimize	optimize	VERB
fcis-30157	587	11	models	model	NOUN
fcis-30157	587	12	across	across	ADP
fcis-30157	587	13	regions	region	NOUN
fcis-30157	587	14	and	and	CCONJ
fcis-30157	587	15	races	race	NOUN
fcis-30157	587	16	.	.	PUNCT
fcis-30157	588	1	to	to	PART
fcis-30157	588	2	enhance	enhance	VERB
fcis-30157	588	3	diagnostic	diagnostic	ADJ
fcis-30157	588	4	scalability	scalability	NOUN
fcis-30157	588	5	,	,	PUNCT
fcis-30157	588	6	combining	combine	VERB
fcis-30157	588	7	multi	multi	ADJ
fcis-30157	588	8	-	-	ADJ
fcis-30157	588	9	dimensional	dimensional	ADJ
fcis-30157	588	10	information	information	NOUN
fcis-30157	588	11	has	have	AUX
fcis-30157	588	12	proven	prove	VERB
fcis-30157	588	13	effective	effective	ADJ
fcis-30157	588	14	.	.	PUNCT
fcis-30157	589	1	at	at	ADP
fcis-30157	589	2	cleveland	cleveland	PROPN
fcis-30157	589	3	medical	medical	ADJ
fcis-30157	589	4	center	center	NOUN
fcis-30157	589	5	,	,	PUNCT
fcis-30157	589	6	adding	add	VERB
fcis-30157	589	7	patient	patient	ADJ
fcis-30157	589	8	complaint	complaint	NOUN
fcis-30157	589	9	data	datum	NOUN
fcis-30157	589	10	(	(	PUNCT
fcis-30157	589	11	e.g.	e.g.	ADV
fcis-30157	589	12	,	,	PUNCT
fcis-30157	589	13	itching	itching	NOUN
fcis-30157	589	14	degree	degree	NOUN
fcis-30157	589	15	,	,	PUNCT
fcis-30157	589	16	disease	disease	NOUN
fcis-30157	589	17	cycle	cycle	NOUN
fcis-30157	589	18	)	)	PUNCT
fcis-30157	589	19	improved	improve	VERB
fcis-30157	589	20	differential	differential	ADJ
fcis-30157	589	21	diagnosis	diagnosis	NOUN
fcis-30157	589	22	of	of	ADP
fcis-30157	589	23	seborrheic	seborrheic	VERB
fcis-30157	589	24	dermatitis	dermatitis	NOUN
fcis-30157	589	25	and	and	CCONJ
fcis-30157	589	26	psoriasis	psoriasis	NOUN
fcis-30157	589	27	from	from	ADP
fcis-30157	589	28	78.4	78.4	NUM
fcis-30157	589	29	%	%	NOUN
fcis-30157	589	30	to	to	ADP
fcis-30157	589	31	86.9	86.9	NUM
fcis-30157	589	32	%	%	NOUN
fcis-30157	589	33	.	.	PUNCT
fcis-30157	590	1	however	however	ADV
fcis-30157	590	2	,	,	PUNCT
fcis-30157	590	3	synchronizing	synchronize	VERB
fcis-30157	590	4	heterogeneous	heterogeneous	ADJ
fcis-30157	590	5	data	datum	NOUN
fcis-30157	590	6	remains	remain	VERB
fcis-30157	590	7	challenging	challenge	VERB
fcis-30157	590	8	.	.	PUNCT
fcis-30157	591	1	recent	recent	ADJ
fcis-30157	591	2	attempts	attempt	NOUN
fcis-30157	591	3	with	with	ADP
fcis-30157	591	4	timing	time	VERB
fcis-30157	591	5	convolutional	convolutional	ADJ
fcis-30157	591	6	networks	network	NOUN
fcis-30157	591	7	have	have	AUX
fcis-30157	591	8	performed	perform	VERB
fcis-30157	591	9	well	well	ADV
fcis-30157	591	10	in	in	ADP
fcis-30157	591	11	chronic	chronic	ADJ
fcis-30157	591	12	eczema	eczema	NOUN
fcis-30157	591	13	prediction	prediction	NOUN
fcis-30157	591	14	,	,	PUNCT
fcis-30157	591	15	offering	offer	VERB
fcis-30157	591	16	a	a	DET
fcis-30157	591	17	path	path	NOUN
fcis-30157	591	18	for	for	ADP
fcis-30157	591	19	dynamic	dynamic	ADJ
fcis-30157	591	20	diagnostics	diagnostic	NOUN
fcis-30157	591	21	.	.	PUNCT
fcis-30157	592	1	transfer	transfer	NOUN
fcis-30157	592	2	learning	learn	VERB
fcis-30157	592	3	across	across	ADP
fcis-30157	592	4	diseases	disease	NOUN
fcis-30157	592	5	shows	show	VERB
fcis-30157	592	6	promise	promise	NOUN
fcis-30157	592	7	.	.	PUNCT
fcis-30157	593	1	stanford	stanford	PROPN
fcis-30157	593	2	’s	’s	PART
fcis-30157	593	3	team	team	NOUN
fcis-30157	593	4	applied	apply	VERB
fcis-30157	593	5	a	a	DET
fcis-30157	593	6	skin	skin	NOUN
fcis-30157	593	7	disease	disease	NOUN
fcis-30157	593	8	model	model	NOUN
fcis-30157	593	9	to	to	PART
fcis-30157	593	10	assist	assist	VERB
fcis-30157	593	11	with	with	ADP
fcis-30157	593	12	covid-19	covid-19	PROPN
fcis-30157	593	13	ct	ct	PROPN
fcis-30157	593	14	imaging	imaging	NOUN
fcis-30157	593	15	,	,	PUNCT
fcis-30157	593	16	achieving	achieve	VERB
fcis-30157	593	17	93.2	93.2	NUM
fcis-30157	593	18	%	%	NOUN
fcis-30157	593	19	accuracy	accuracy	NOUN
fcis-30157	593	20	.	.	PUNCT
fcis-30157	594	1	transfer	transfer	NOUN
fcis-30157	594	2	learning	learning	NOUN
fcis-30157	594	3	can	can	AUX
fcis-30157	594	4	also	also	ADV
fcis-30157	594	5	address	address	VERB
fcis-30157	594	6	rare	rare	ADJ
fcis-30157	594	7	skin	skin	NOUN
fcis-30157	594	8	diseases	disease	NOUN
fcis-30157	594	9	;	;	PUNCT
fcis-30157	594	10	fine	fine	ADV
fcis-30157	594	11	-	-	PUNCT
fcis-30157	594	12	tuning	tune	VERB
fcis-30157	594	13	a	a	DET
fcis-30157	594	14	vitiligo	vitiligo	NOUN
fcis-30157	594	15	model	model	NOUN
fcis-30157	594	16	with	with	ADP
fcis-30157	594	17	300	300	NUM
fcis-30157	594	18	new	new	ADJ
fcis-30157	594	19	samples	sample	NOUN
fcis-30157	594	20	increased	increase	VERB
fcis-30157	594	21	recognition	recognition	NOUN
fcis-30157	594	22	accuracy	accuracy	NOUN
fcis-30157	594	23	from	from	ADP
fcis-30157	594	24	68	68	NUM
fcis-30157	594	25	%	%	NOUN
fcis-30157	594	26	to	to	PART
fcis-30157	594	27	82	82	NUM
fcis-30157	594	28	%	%	NOUN
fcis-30157	594	29	.	.	PUNCT
fcis-30157	595	1	however	however	ADV
fcis-30157	595	2	,	,	PUNCT
fcis-30157	595	3	care	care	NOUN
fcis-30157	595	4	is	be	AUX
fcis-30157	595	5	needed	need	VERB
fcis-30157	595	6	regarding	regard	VERB
fcis-30157	595	7	disease	disease	NOUN
fcis-30157	595	8	-	-	PUNCT
fcis-30157	595	9	specific	specific	ADJ
fcis-30157	595	10	characteristics	characteristic	NOUN
fcis-30157	595	11	to	to	PART
fcis-30157	595	12	avoid	avoid	VERB
fcis-30157	595	13	spatial	spatial	ADJ
fcis-30157	595	14	offsets	offset	NOUN
fcis-30157	595	15	during	during	ADP
fcis-30157	595	16	migration	migration	NOUN
fcis-30157	595	17	.	.	PUNCT
fcis-30157	596	1	model	model	NOUN
fcis-30157	596	2	compression	compression	NOUN
fcis-30157	596	3	and	and	CCONJ
fcis-30157	596	4	hardware	hardware	NOUN
fcis-30157	596	5	adaptation	adaptation	NOUN
fcis-30157	596	6	are	be	AUX
fcis-30157	596	7	critical	critical	ADJ
fcis-30157	596	8	for	for	ADP
fcis-30157	596	9	clinical	clinical	ADJ
fcis-30157	596	10	deployment	deployment	NOUN
fcis-30157	596	11	.	.	PUNCT
fcis-30157	597	1	the	the	DET
fcis-30157	597	2	enhanced	enhanced	ADJ
fcis-30157	597	3	mobilenet	mobilenet	NOUN
fcis-30157	597	4	architecture	architecture	NOUN
fcis-30157	597	5	reduced	reduce	VERB
fcis-30157	597	6	model	model	NOUN
fcis-30157	597	7	size	size	NOUN
fcis-30157	597	8	to	to	ADP
fcis-30157	597	9	4.2	4.2	NUM
fcis-30157	597	10	mb	mb	NOUN
fcis-30157	597	11	while	while	SCONJ
fcis-30157	597	12	maintaining	maintain	VERB
fcis-30157	597	13	91.3	91.3	NUM
fcis-30157	597	14	%	%	NOUN
fcis-30157	597	15	accuracy	accuracy	NOUN
fcis-30157	597	16	,	,	PUNCT
fcis-30157	597	17	facilitating	facilitate	VERB
fcis-30157	597	18	mobile	mobile	ADJ
fcis-30157	597	19	deployment	deployment	NOUN
fcis-30157	597	20	.	.	PUNCT
fcis-30157	598	1	privacy	privacy	NOUN
fcis-30157	598	2	protection	protection	NOUN
fcis-30157	598	3	remains	remain	VERB
fcis-30157	598	4	an	an	DET
fcis-30157	598	5	issue	issue	NOUN
fcis-30157	598	6	,	,	PUNCT
fcis-30157	598	7	though	though	SCONJ
fcis-30157	598	8	federated	federated	ADJ
fcis-30157	598	9	learning	learning	NOUN
fcis-30157	598	10	has	have	AUX
fcis-30157	598	11	proven	prove	VERB
fcis-30157	598	12	effective	effective	ADJ
fcis-30157	598	13	,	,	PUNCT
fcis-30157	598	14	allowing	allow	VERB
fcis-30157	598	15	multiple	multiple	ADJ
fcis-30157	598	16	institutions	institution	NOUN
fcis-30157	598	17	to	to	PART
fcis-30157	598	18	share	share	VERB
fcis-30157	598	19	knowledge	knowledge	NOUN
fcis-30157	598	20	without	without	ADP
fcis-30157	598	21	violating	violate	VERB
fcis-30157	598	22	data	data	NOUN
fcis-30157	598	23	privacy	privacy	NOUN
fcis-30157	598	24	,	,	PUNCT
fcis-30157	598	25	as	as	SCONJ
fcis-30157	598	26	shown	show	VERB
fcis-30157	598	27	by	by	ADP
fcis-30157	598	28	a	a	DET
fcis-30157	598	29	collaboration	collaboration	NOUN
fcis-30157	598	30	between	between	ADP
fcis-30157	598	31	23	23	NUM
fcis-30157	598	32	german	german	ADJ
fcis-30157	598	33	clinics	clinic	NOUN
fcis-30157	598	34	.	.	PUNCT
fcis-30157	599	1	ethical	ethical	ADJ
fcis-30157	599	2	considerations	consideration	NOUN
fcis-30157	599	3	are	be	AUX
fcis-30157	599	4	also	also	ADV
fcis-30157	599	5	vital	vital	ADJ
fcis-30157	599	6	.	.	PUNCT
fcis-30157	600	1	in	in	ADP
fcis-30157	600	2	2022	2022	NUM
fcis-30157	600	3	,	,	PUNCT
fcis-30157	600	4	63	63	NUM
fcis-30157	600	5	%	%	NOUN
fcis-30157	600	6	of	of	ADP
fcis-30157	600	7	ai	ai	PROPN
fcis-30157	600	8	systems	system	NOUN
fcis-30157	600	9	in	in	ADP
fcis-30157	600	10	skin	skin	NOUN
fcis-30157	600	11	disease	disease	NOUN
fcis-30157	600	12	diagnosis	diagnosis	NOUN
fcis-30157	600	13	lacked	lack	VERB
fcis-30157	600	14	sufficient	sufficient	ADJ
fcis-30157	600	15	decision	decision	NOUN
fcis-30157	600	16	-	-	PUNCT
fcis-30157	600	17	making	make	VERB
fcis-30157	600	18	transparency	transparency	NOUN
fcis-30157	600	19	.	.	PUNCT
fcis-30157	601	1	improvements	improvement	NOUN
fcis-30157	601	2	using	use	VERB
fcis-30157	601	3	class	class	NOUN
fcis-30157	601	4	activation	activation	NOUN
fcis-30157	601	5	mapping	mapping	NOUN
fcis-30157	601	6	(	(	PUNCT
fcis-30157	601	7	grad	grad	NOUN
fcis-30157	601	8	-	-	PUNCT
fcis-30157	601	9	cam	cam	NOUN
fcis-30157	601	10	)	)	PUNCT
fcis-30157	601	11	increased	increase	VERB
fcis-30157	601	12	physician	physician	NOUN
fcis-30157	601	13	acceptance	acceptance	NOUN
fcis-30157	601	14	by	by	ADP
fcis-30157	601	15	41	41	NUM
fcis-30157	601	16	%	%	NOUN
fcis-30157	601	17	.	.	PUNCT
fcis-30157	602	1	a	a	DET
fcis-30157	602	2	robust	robust	ADJ
fcis-30157	602	3	error	error	NOUN
fcis-30157	602	4	traceability	traceability	NOUN
fcis-30157	602	5	mechanism	mechanism	NOUN
fcis-30157	602	6	,	,	PUNCT
fcis-30157	602	7	such	such	ADJ
fcis-30157	602	8	as	as	ADP
fcis-30157	602	9	the	the	DET
fcis-30157	602	10	mayo	mayo	NOUN
fcis-30157	602	11	clinic	clinic	PROPN
fcis-30157	602	12	’s	’s	PART
fcis-30157	602	13	diagnostic	diagnostic	ADJ
fcis-30157	602	14	deviation	deviation	NOUN
fcis-30157	602	15	analysis	analysis	NOUN
fcis-30157	602	16	module	module	NOUN
fcis-30157	602	17	,	,	PUNCT
fcis-30157	602	18	is	be	AUX
fcis-30157	602	19	necessary	necessary	ADJ
fcis-30157	602	20	to	to	PART
fcis-30157	602	21	ensure	ensure	VERB
fcis-30157	602	22	trust	trust	NOUN
fcis-30157	602	23	and	and	CCONJ
fcis-30157	602	24	facilitate	facilitate	NOUN
fcis-30157	602	25	system	system	NOUN
fcis-30157	602	26	iterations	iteration	NOUN
fcis-30157	602	27	.	.	PUNCT
fcis-30157	603	1	from	from	ADP
fcis-30157	603	2	an	an	DET
fcis-30157	603	3	economic	economic	ADJ
fcis-30157	603	4	perspective	perspective	NOUN
fcis-30157	603	5	,	,	PUNCT
fcis-30157	603	6	cost	cost	NOUN
fcis-30157	603	7	-	-	PUNCT
fcis-30157	603	8	effectiveness	effectiveness	NOUN
fcis-30157	603	9	is	be	AUX
fcis-30157	603	10	key	key	ADJ
fcis-30157	603	11	.	.	PUNCT
fcis-30157	604	1	a	a	DET
fcis-30157	604	2	model	model	NOUN
fcis-30157	604	3	from	from	ADP
fcis-30157	604	4	cambridge	cambridge	PROPN
fcis-30157	604	5	university	university	PROPN
fcis-30157	604	6	showed	show	VERB
fcis-30157	604	7	that	that	SCONJ
fcis-30157	604	8	ai	ai	VERB
fcis-30157	604	9	systems	system	NOUN
fcis-30157	604	10	with	with	ADP
fcis-30157	604	11	85	85	NUM
fcis-30157	604	12	%	%	NOUN
fcis-30157	604	13	diagnostic	diagnostic	ADJ
fcis-30157	604	14	accuracy	accuracy	NOUN
fcis-30157	604	15	can	can	AUX
fcis-30157	604	16	reduce	reduce	VERB
fcis-30157	604	17	diagnosis	diagnosis	NOUN
fcis-30157	604	18	costs	cost	NOUN
fcis-30157	604	19	by	by	ADP
fcis-30157	604	20	£	£	SYM
fcis-30157	604	21	34	34	NUM
fcis-30157	604	22	,	,	PUNCT
fcis-30157	604	23	provided	provide	VERB
fcis-30157	604	24	daily	daily	ADJ
fcis-30157	604	25	treatments	treatment	NOUN
fcis-30157	604	26	exceed	exceed	VERB
fcis-30157	604	27	50	50	NUM
fcis-30157	604	28	cases	case	NOUN
fcis-30157	604	29	.	.	PUNCT
fcis-30157	605	1	this	this	PRON
fcis-30157	605	2	suggests	suggest	VERB
fcis-30157	605	3	tailored	tailor	VERB
fcis-30157	605	4	solutions	solution	NOUN
fcis-30157	605	5	for	for	ADP
fcis-30157	605	6	different	different	ADJ
fcis-30157	605	7	healthcare	healthcare	NOUN
fcis-30157	605	8	levels	level	NOUN
fcis-30157	605	9	:	:	PUNCT
fcis-30157	605	10	high	high	ADJ
fcis-30157	605	11	-	-	PUNCT
fcis-30157	605	12	precision	precision	NOUN
fcis-30157	605	13	systems	system	NOUN
fcis-30157	605	14	for	for	ADP
fcis-30157	605	15	level	level	NOUN
fcis-30157	605	16	3	3	NUM
fcis-30157	605	17	hospitals	hospital	NOUN
fcis-30157	605	18	and	and	CCONJ
fcis-30157	605	19	lightweight	lightweight	ADJ
fcis-30157	605	20	tools	tool	NOUN
fcis-30157	605	21	for	for	ADP
fcis-30157	605	22	community	community	NOUN
fcis-30157	605	23	clinics	clinic	NOUN
fcis-30157	605	24	.	.	PUNCT
fcis-30157	606	1	interdisciplinary	interdisciplinary	ADJ
fcis-30157	606	2	collaboration	collaboration	NOUN
fcis-30157	606	3	is	be	AUX
fcis-30157	606	4	essential	essential	ADJ
fcis-30157	606	5	for	for	ADP
fcis-30157	606	6	overcoming	overcome	VERB
fcis-30157	606	7	technical	technical	ADJ
fcis-30157	606	8	barriers	barrier	NOUN
fcis-30157	606	9	.	.	PUNCT
fcis-30157	607	1	the	the	DET
fcis-30157	607	2	mhorunet	mhorunet	ADJ
fcis-30157	607	3	segmentation	segmentation	NOUN
fcis-30157	607	4	network	network	NOUN
fcis-30157	607	5	,	,	PUNCT
fcis-30157	607	6	inspired	inspire	VERB
fcis-30157	607	7	by	by	ADP
fcis-30157	607	8	skin	skin	NOUN
fcis-30157	607	9	histopathology	histopathology	NOUN
fcis-30157	607	10	,	,	PUNCT
fcis-30157	607	11	achieved	achieve	VERB
fcis-30157	607	12	82.1	82.1	NUM
fcis-30157	607	13	%	%	NOUN
fcis-30157	607	14	in	in	ADP
fcis-30157	607	15	lesion	lesion	NOUN
fcis-30157	607	16	boundary	boundary	ADJ
fcis-30157	607	17	positioning	positioning	NOUN
fcis-30157	607	18	.	.	PUNCT
fcis-30157	608	1	expanding	expand	VERB
fcis-30157	608	2	on	on	ADP
fcis-30157	608	3	bionic	bionic	ADJ
fcis-30157	608	4	design	design	NOUN
fcis-30157	608	5	ideas	idea	NOUN
fcis-30157	608	6	and	and	CCONJ
fcis-30157	608	7	leveraging	leverage	VERB
fcis-30157	608	8	immune	immune	ADJ
fcis-30157	608	9	cell	cell	NOUN
fcis-30157	608	10	recognition	recognition	NOUN
fcis-30157	608	11	mechanisms	mechanism	NOUN
fcis-30157	608	12	can	can	AUX
fcis-30157	608	13	optimize	optimize	VERB
fcis-30157	608	14	model	model	NOUN
fcis-30157	608	15	performance	performance	NOUN
fcis-30157	608	16	.	.	PUNCT
fcis-30157	609	1	integration	integration	NOUN
fcis-30157	609	2	with	with	ADP
fcis-30157	609	3	medical	medical	ADJ
fcis-30157	609	4	iot	iot	NOUN
fcis-30157	609	5	technology	technology	NOUN
fcis-30157	609	6	,	,	PUNCT
fcis-30157	609	7	like	like	ADP
fcis-30157	609	8	wearable	wearable	ADJ
fcis-30157	609	9	sensors	sensor	NOUN
fcis-30157	609	10	for	for	ADP
fcis-30157	609	11	continuous	continuous	ADJ
fcis-30157	609	12	psoriasis	psoriasis	NOUN
fcis-30157	609	13	monitoring	monitoring	NOUN
fcis-30157	609	14	,	,	PUNCT
fcis-30157	609	15	offers	offer	VERB
fcis-30157	609	16	new	new	ADJ
fcis-30157	609	17	diagnostic	diagnostic	ADJ
fcis-30157	609	18	possibilities	possibility	NOUN
fcis-30157	609	19	.	.	PUNCT
fcis-30157	610	1	a	a	DET
fcis-30157	610	2	significant	significant	ADJ
fcis-30157	610	3	gap	gap	NOUN
fcis-30157	610	4	exists	exist	VERB
fcis-30157	610	5	in	in	ADP
fcis-30157	610	6	standardization	standardization	NOUN
fcis-30157	610	7	for	for	ADP
fcis-30157	610	8	technology	technology	NOUN
fcis-30157	610	9	transformation	transformation	NOUN
fcis-30157	610	10	.	.	PUNCT
fcis-30157	611	1	the	the	DET
fcis-30157	611	2	isic	isic	PROPN
fcis-30157	611	3	framework	framework	NOUN
fcis-30157	611	4	currently	currently	ADV
fcis-30157	611	5	covers	cover	VERB
fcis-30157	611	6	only	only	ADV
fcis-30157	611	7	7	7	NUM
fcis-30157	611	8	common	common	ADJ
fcis-30157	611	9	skin	skin	NOUN
fcis-30157	611	10	tumors	tumor	NOUN
fcis-30157	611	11	,	,	PUNCT
fcis-30157	611	12	leaving	leave	VERB
fcis-30157	611	13	over	over	ADP
fcis-30157	611	14	3,000	3,000	NUM
fcis-30157	611	15	rare	rare	ADJ
fcis-30157	611	16	skin	skin	NOUN
fcis-30157	611	17	diseases	disease	NOUN
fcis-30157	611	18	underrepresented	underrepresente	VERB
fcis-30157	611	19	.	.	PUNCT
fcis-30157	612	1	building	build	VERB
fcis-30157	612	2	a	a	DET
fcis-30157	612	3	more	more	ADV
fcis-30157	612	4	fine	fine	ADV
fcis-30157	612	5	-	-	PUNCT
fcis-30157	612	6	grained	grain	VERB
fcis-30157	612	7	ontology	ontology	NOUN
fcis-30157	612	8	library	library	NOUN
fcis-30157	612	9	,	,	PUNCT
fcis-30157	612	10	as	as	SCONJ
fcis-30157	612	11	seen	see	VERB
fcis-30157	612	12	in	in	ADP
fcis-30157	612	13	yonsei	yonsei	PROPN
fcis-30157	612	14	university	university	PROPN
fcis-30157	612	15	’s	’s	PART
fcis-30157	612	16	skin	skin	NOUN
fcis-30157	612	17	disease	disease	NOUN
fcis-30157	612	18	ontology	ontology	NOUN
fcis-30157	612	19	mapping	mapping	NOUN
fcis-30157	612	20	,	,	PUNCT
fcis-30157	612	21	could	could	AUX
fcis-30157	612	22	improve	improve	VERB
fcis-30157	612	23	zero	zero	NUM
fcis-30157	612	24	-	-	PUNCT
fcis-30157	612	25	sample	sample	NOUN
fcis-30157	612	26	recognition	recognition	NOUN
fcis-30157	612	27	accuracy	accuracy	NOUN
fcis-30157	612	28	to	to	ADP
fcis-30157	612	29	67.3	67.3	NUM
fcis-30157	612	30	%	%	NOUN
fcis-30157	612	31	,	,	PUNCT
fcis-30157	612	32	expanding	expand	VERB
fcis-30157	612	33	system	system	NOUN
fcis-30157	612	34	capabilities	capability	NOUN
fcis-30157	612	35	.	.	PUNCT
fcis-30157	613	1	future	future	ADJ
fcis-30157	613	2	research	research	NOUN
fcis-30157	613	3	should	should	AUX
fcis-30157	613	4	focus	focus	VERB
fcis-30157	613	5	on	on	ADP
fcis-30157	613	6	creating	create	VERB
fcis-30157	613	7	a	a	DET
fcis-30157	613	8	closed	closed	ADJ
fcis-30157	613	9	-	-	PUNCT
fcis-30157	613	10	loop	loop	NOUN
fcis-30157	613	11	intelligent	intelligent	ADJ
fcis-30157	613	12	system	system	NOUN
fcis-30157	613	13	.	.	PUNCT
fcis-30157	614	1	a	a	DET
fcis-30157	614	2	mayo	mayo	NOUN
fcis-30157	614	3	clinic	clinic	PROPN
fcis-30157	614	4	pilot	pilot	NOUN
fcis-30157	614	5	project	project	NOUN
fcis-30157	614	6	integrating	integrate	VERB
fcis-30157	614	7	ai	ai	VERB
fcis-30157	614	8	diagnosis	diagnosis	NOUN
fcis-30157	614	9	,	,	PUNCT
fcis-30157	614	10	treatment	treatment	NOUN
fcis-30157	614	11	recommendations	recommendation	NOUN
fcis-30157	614	12	,	,	PUNCT
fcis-30157	614	13	and	and	CCONJ
fcis-30157	614	14	efficacy	efficacy	NOUN
fcis-30157	614	15	predictions	prediction	NOUN
fcis-30157	614	16	shortened	shorten	VERB
fcis-30157	614	17	chronic	chronic	ADJ
fcis-30157	614	18	urticaria	urticaria	NOUN
fcis-30157	614	19	diagnosis	diagnosis	NOUN
fcis-30157	614	20	time	time	NOUN
fcis-30157	614	21	from	from	ADP
fcis-30157	614	22	6.2	6.2	NUM
fcis-30157	614	23	weeks	week	NOUN
fcis-30157	614	24	to	to	ADP
fcis-30157	614	25	2.1	2.1	NUM
fcis-30157	614	26	weeks	week	NOUN
fcis-30157	614	27	,	,	PUNCT
fcis-30157	614	28	improving	improve	VERB
fcis-30157	614	29	treatment	treatment	NOUN
fcis-30157	614	30	efficiency	efficiency	NOUN
fcis-30157	614	31	by	by	ADP
fcis-30157	614	32	28	28	NUM
fcis-30157	614	33	%	%	NOUN
fcis-30157	614	34	.	.	PUNCT
fcis-30157	615	1	multi	multi	ADJ
fcis-30157	615	2	-	-	ADJ
fcis-30157	615	3	task	task	ADJ
fcis-30157	615	4	learning	learn	VERB
fcis-30157	615	5	frameworks	framework	NOUN
fcis-30157	615	6	that	that	PRON
fcis-30157	615	7	combine	combine	VERB
fcis-30157	615	8	disease	disease	NOUN
fcis-30157	615	9	classification	classification	NOUN
fcis-30157	615	10	and	and	CCONJ
fcis-30157	615	11	severity	severity	NOUN
fcis-30157	615	12	assessment	assessment	NOUN
fcis-30157	615	13	,	,	PUNCT
fcis-30157	615	14	like	like	ADP
fcis-30157	615	15	the	the	DET
fcis-30157	615	16	model	model	NOUN
fcis-30157	615	17	from	from	ADP
fcis-30157	615	18	drexel	drexel	PROPN
fcis-30157	615	19	university	university	PROPN
fcis-30157	615	20	,	,	PUNCT
fcis-30157	615	21	offer	offer	VERB
fcis-30157	615	22	a	a	DET
fcis-30157	615	23	promising	promising	ADJ
fcis-30157	615	24	path	path	NOUN
fcis-30157	615	25	forward	forward	ADV
fcis-30157	615	26	,	,	PUNCT
fcis-30157	615	27	achieving	achieve	VERB
fcis-30157	615	28	an	an	DET
fcis-30157	615	29	87.9	87.9	NUM
fcis-30157	615	30	%	%	NOUN
fcis-30157	615	31	weighted	weight	VERB
fcis-30157	615	32	accuracy	accuracy	NOUN
fcis-30157	615	33	rate	rate	NOUN
fcis-30157	615	34	across	across	ADP
fcis-30157	615	35	40	40	NUM
fcis-30157	615	36	inflammatory	inflammatory	ADJ
fcis-30157	615	37	skin	skin	NOUN
fcis-30157	615	38	diseases	disease	NOUN
fcis-30157	615	39	,	,	PUNCT
fcis-30157	615	40	a	a	DET
fcis-30157	615	41	11.6	11.6	NUM
fcis-30157	615	42	%	%	NOUN
fcis-30157	615	43	improvement	improvement	NOUN
fcis-30157	615	44	over	over	ADP
fcis-30157	615	45	single	single	ADJ
fcis-30157	615	46	-	-	PUNCT
fcis-30157	615	47	task	task	NOUN
fcis-30157	615	48	models	model	NOUN
fcis-30157	615	49	.	.	PUNCT
fcis-30157	616	1	4.2.3	4.2.3	X
fcis-30157	616	2	.	.	PUNCT
fcis-30157	617	1	directions	direction	NOUN
fcis-30157	617	2	for	for	ADP
fcis-30157	617	3	future	future	ADJ
fcis-30157	617	4	research	research	NOUN
fcis-30157	617	5	future	future	ADJ
fcis-30157	617	6	research	research	NOUN
fcis-30157	617	7	in	in	ADP
fcis-30157	617	8	ai	ai	ADJ
fcis-30157	617	9	-	-	PUNCT
fcis-30157	617	10	driven	drive	VERB
fcis-30157	617	11	skin	skin	NOUN
fcis-30157	617	12	disease	disease	NOUN
fcis-30157	617	13	diagnosis	diagnosis	NOUN
fcis-30157	617	14	should	should	AUX
fcis-30157	617	15	focus	focus	VERB
fcis-30157	617	16	on	on	ADP
fcis-30157	617	17	several	several	ADJ
fcis-30157	617	18	key	key	ADJ
fcis-30157	617	19	areas	area	NOUN
fcis-30157	617	20	to	to	PART
fcis-30157	617	21	enhance	enhance	VERB
fcis-30157	617	22	model	model	NOUN
fcis-30157	617	23	capabilities	capability	NOUN
fcis-30157	617	24	.	.	PUNCT
fcis-30157	618	1	one	one	NUM
fcis-30157	618	2	promising	promise	VERB
fcis-30157	618	3	direction	direction	NOUN
fcis-30157	618	4	is	be	AUX
fcis-30157	618	5	the	the	DET
fcis-30157	618	6	exploration	exploration	NOUN
fcis-30157	618	7	of	of	ADP
fcis-30157	618	8	advanced	advanced	ADJ
fcis-30157	618	9	deep	deep	ADJ
fcis-30157	618	10	learning	learning	NOUN
fcis-30157	618	11	architectures	architecture	NOUN
fcis-30157	618	12	,	,	PUNCT
fcis-30157	618	13	such	such	ADJ
fcis-30157	618	14	as	as	ADP
fcis-30157	618	15	multimodal	multimodal	ADJ
fcis-30157	618	16	fusion	fusion	NOUN
fcis-30157	618	17	methods	method	NOUN
fcis-30157	618	18	like	like	ADP
fcis-30157	618	19	mdfnet	mdfnet	NOUN
fcis-30157	618	20	,	,	PUNCT
fcis-30157	618	21	which	which	PRON
fcis-30157	618	22	combine	combine	VERB
fcis-30157	618	23	skin	skin	NOUN
fcis-30157	618	24	images	image	NOUN
fcis-30157	618	25	with	with	ADP
fcis-30157	618	26	clinical	clinical	ADJ
fcis-30157	618	27	data	datum	NOUN
fcis-30157	618	28	to	to	PART
fcis-30157	618	29	improve	improve	VERB
fcis-30157	618	30	skin	skin	NOUN
fcis-30157	618	31	cancer	cancer	NOUN
fcis-30157	618	32	classification	classification	NOUN
fcis-30157	618	33	.	.	PUNCT
fcis-30157	619	1	this	this	DET
fcis-30157	619	2	approach	approach	NOUN
fcis-30157	619	3	integrates	integrate	VERB
fcis-30157	619	4	both	both	DET
fcis-30157	619	5	visual	visual	ADJ
fcis-30157	619	6	and	and	CCONJ
fcis-30157	619	7	patient	patient	ADJ
fcis-30157	619	8	history	history	NOUN
fcis-30157	619	9	data	datum	NOUN
fcis-30157	619	10	for	for	ADP
fcis-30157	619	11	more	more	ADJ
fcis-30157	619	12	comprehensive	comprehensive	ADJ
fcis-30157	619	13	diagnostics	diagnostic	NOUN
fcis-30157	619	14	.	.	PUNCT
fcis-30157	620	1	another	another	DET
fcis-30157	620	2	critical	critical	ADJ
fcis-30157	620	3	area	area	NOUN
fcis-30157	620	4	is	be	AUX
fcis-30157	620	5	leveraging	leverage	VERB
fcis-30157	620	6	big	big	ADJ
fcis-30157	620	7	data	datum	NOUN
fcis-30157	620	8	and	and	CCONJ
fcis-30157	620	9	ai	ai	VERB
fcis-30157	620	10	for	for	ADP
fcis-30157	620	11	personalized	personalized	ADJ
fcis-30157	620	12	diagnosis	diagnosis	NOUN
fcis-30157	620	13	.	.	PUNCT
fcis-30157	621	1	large	large	ADJ
fcis-30157	621	2	,	,	PUNCT
fcis-30157	621	3	diverse	diverse	ADJ
fcis-30157	621	4	datasets	dataset	NOUN
fcis-30157	621	5	combined	combine	VERB
fcis-30157	621	6	with	with	ADP
fcis-30157	621	7	advanced	advanced	ADJ
fcis-30157	621	8	algorithms	algorithm	NOUN
fcis-30157	621	9	can	can	AUX
fcis-30157	621	10	tailor	tailor	VERB
fcis-30157	621	11	models	model	NOUN
fcis-30157	621	12	to	to	ADP
fcis-30157	621	13	individual	individual	ADJ
fcis-30157	621	14	patients	patient	NOUN
fcis-30157	621	15	,	,	PUNCT
fcis-30157	621	16	as	as	SCONJ
fcis-30157	621	17	demonstrated	demonstrate	VERB
fcis-30157	621	18	by	by	ADP
fcis-30157	621	19	isic	isic	PROPN
fcis-30157	621	20	2019	2019	NUM
fcis-30157	621	21	,	,	PUNCT
fcis-30157	621	22	where	where	SCONJ
fcis-30157	621	23	models	model	NOUN
fcis-30157	621	24	trained	train	VERB
fcis-30157	621	25	on	on	ADP
fcis-30157	621	26	diverse	diverse	ADJ
fcis-30157	621	27	datasets	dataset	NOUN
fcis-30157	621	28	achieved	achieve	VERB
fcis-30157	621	29	over	over	ADP
fcis-30157	621	30	90	90	NUM
fcis-30157	621	31	%	%	NOUN
fcis-30157	621	32	accuracy	accuracy	NOUN
fcis-30157	621	33	in	in	ADP
fcis-30157	621	34	skin	skin	NOUN
fcis-30157	621	35	cancer	cancer	NOUN
fcis-30157	621	36	diagnosis	diagnosis	NOUN
fcis-30157	621	37	.	.	PUNCT
fcis-30157	622	1	however	however	ADV
fcis-30157	622	2	,	,	PUNCT
fcis-30157	622	3	challenges	challenge	NOUN
fcis-30157	622	4	remain	remain	VERB
fcis-30157	622	5	in	in	ADP
fcis-30157	622	6	ensuring	ensure	VERB
fcis-30157	622	7	these	these	DET
fcis-30157	622	8	models	model	NOUN
fcis-30157	622	9	are	be	AUX
fcis-30157	622	10	robust	robust	ADJ
fcis-30157	622	11	enough	enough	ADV
fcis-30157	622	12	to	to	PART
fcis-30157	622	13	account	account	VERB
fcis-30157	622	14	for	for	ADP
fcis-30157	622	15	variations	variation	NOUN
fcis-30157	622	16	in	in	ADP
fcis-30157	622	17	skin	skin	NOUN
fcis-30157	622	18	tones	tone	NOUN
fcis-30157	622	19	,	,	PUNCT
fcis-30157	622	20	ages	age	NOUN
fcis-30157	622	21	,	,	PUNCT
fcis-30157	622	22	and	and	CCONJ
fcis-30157	622	23	ethnicities	ethnicity	NOUN
fcis-30157	622	24	.	.	PUNCT
fcis-30157	623	1	multi	multi	ADJ
fcis-30157	623	2	-	-	ADJ
fcis-30157	623	3	center	center	ADJ
fcis-30157	623	4	clinical	clinical	ADJ
fcis-30157	623	5	research	research	NOUN
fcis-30157	623	6	is	be	AUX
fcis-30157	623	7	essential	essential	ADJ
fcis-30157	623	8	to	to	PART
fcis-30157	623	9	validate	validate	VERB
fcis-30157	623	10	ai	ai	NOUN
fcis-30157	623	11	models	model	NOUN
fcis-30157	623	12	across	across	ADP
fcis-30157	623	13	varied	varied	ADJ
fcis-30157	623	14	clinical	clinical	ADJ
fcis-30157	623	15	scenarios	scenario	NOUN
fcis-30157	623	16	.	.	PUNCT
fcis-30157	624	1	studies	study	NOUN
fcis-30157	624	2	across	across	ADP
fcis-30157	624	3	multiple	multiple	ADJ
fcis-30157	624	4	dermatology	dermatology	NOUN
fcis-30157	624	5	clinics	clinic	NOUN
fcis-30157	624	6	have	have	AUX
fcis-30157	624	7	shown	show	VERB
fcis-30157	624	8	that	that	SCONJ
fcis-30157	624	9	deep	deep	ADJ
fcis-30157	624	10	learning	learning	NOUN
fcis-30157	624	11	models	model	NOUN
fcis-30157	624	12	can	can	AUX
fcis-30157	624	13	classify	classify	VERB
fcis-30157	624	14	skin	skin	NOUN
fcis-30157	624	15	lesions	lesion	NOUN
fcis-30157	624	16	accurately	accurately	ADV
fcis-30157	624	17	across	across	ADP
fcis-30157	624	18	different	different	ADJ
fcis-30157	624	19	skin	skin	NOUN
fcis-30157	624	20	tones	tone	NOUN
fcis-30157	624	21	,	,	PUNCT
fcis-30157	624	22	highlighting	highlight	VERB
fcis-30157	624	23	their	their	PRON
fcis-30157	624	24	broader	broad	ADJ
fcis-30157	624	25	application	application	NOUN
fcis-30157	624	26	potential	potential	NOUN
fcis-30157	624	27	.	.	PUNCT
fcis-30157	625	1	explainable	explainable	ADJ
fcis-30157	625	2	ai	ai	PROPN
fcis-30157	625	3	(	(	PUNCT
fcis-30157	625	4	xai	xai	PROPN
fcis-30157	625	5	)	)	PUNCT
fcis-30157	625	6	is	be	AUX
fcis-30157	625	7	another	another	DET
fcis-30157	625	8	crucial	crucial	ADJ
fcis-30157	625	9	area	area	NOUN
fcis-30157	625	10	.	.	PUNCT
fcis-30157	626	1	as	as	SCONJ
fcis-30157	626	2	ai	ai	PROPN
fcis-30157	626	3	models	model	NOUN
fcis-30157	626	4	become	become	VERB
fcis-30157	626	5	more	more	ADV
fcis-30157	626	6	complex	complex	ADJ
fcis-30157	626	7	,	,	PUNCT
fcis-30157	626	8	transparency	transparency	NOUN
fcis-30157	626	9	in	in	ADP
fcis-30157	626	10	decision	decision	NOUN
fcis-30157	626	11	-	-	PUNCT
fcis-30157	626	12	making	making	NOUN
fcis-30157	626	13	is	be	AUX
fcis-30157	626	14	necessary	necessary	ADJ
fcis-30157	626	15	.	.	PUNCT
fcis-30157	627	1	techniques	technique	NOUN
fcis-30157	627	2	like	like	ADP
fcis-30157	627	3	dual	dual	ADJ
fcis-30157	627	4	attention	attention	NOUN
fcis-30157	627	5	networks	network	NOUN
fcis-30157	627	6	can	can	AUX
fcis-30157	627	7	provide	provide	VERB
fcis-30157	627	8	interpretable	interpretable	ADJ
fcis-30157	627	9	results	result	NOUN
fcis-30157	627	10	,	,	PUNCT
fcis-30157	627	11	allowing	allow	VERB
fcis-30157	627	12	dermatologists	dermatologist	NOUN
fcis-30157	627	13	to	to	PART
fcis-30157	627	14	better	well	ADV
fcis-30157	627	15	understand	understand	VERB
fcis-30157	627	16	ai	ai	VERB
fcis-30157	627	17	’s	’s	PART
fcis-30157	627	18	recommendations	recommendation	NOUN
fcis-30157	627	19	and	and	CCONJ
fcis-30157	627	20	increase	increase	VERB
fcis-30157	627	21	trust	trust	NOUN
fcis-30157	627	22	in	in	ADP
fcis-30157	627	23	the	the	DET
fcis-30157	627	24	system	system	NOUN
fcis-30157	627	25	.	.	PUNCT
fcis-30157	628	1	the	the	DET
fcis-30157	628	2	integration	integration	NOUN
fcis-30157	628	3	of	of	ADP
fcis-30157	628	4	auxiliary	auxiliary	ADJ
fcis-30157	628	5	tasks	task	NOUN
fcis-30157	628	6	such	such	ADJ
fcis-30157	628	7	as	as	ADP
fcis-30157	628	8	lesion	lesion	NOUN
fcis-30157	628	9	segmentation	segmentation	NOUN
fcis-30157	628	10	,	,	PUNCT
fcis-30157	628	11	feature	feature	NOUN
fcis-30157	628	12	extraction	extraction	NOUN
fcis-30157	628	13	,	,	PUNCT
fcis-30157	628	14	and	and	CCONJ
fcis-30157	628	15	patient	patient	ADJ
fcis-30157	628	16	history	history	NOUN
fcis-30157	628	17	analysis	analysis	NOUN
fcis-30157	628	18	can	can	AUX
fcis-30157	628	19	further	far	ADV
fcis-30157	628	20	enhance	enhance	VERB
fcis-30157	628	21	ai	ai	VERB
fcis-30157	628	22	diagnostic	diagnostic	ADJ
fcis-30157	628	23	support	support	NOUN
fcis-30157	628	24	.	.	PUNCT
fcis-30157	629	1	a	a	DET
fcis-30157	629	2	multi	multi	ADJ
fcis-30157	629	3	-	-	ADJ
fcis-30157	629	4	task	task	ADJ
fcis-30157	629	5	model	model	NOUN
fcis-30157	629	6	developed	develop	VERB
fcis-30157	629	7	for	for	ADP
fcis-30157	629	8	differentiating	differentiate	VERB
fcis-30157	629	9	skin	skin	NOUN
fcis-30157	629	10	diseases	disease	NOUN
fcis-30157	629	11	not	not	PART
fcis-30157	629	12	only	only	ADV
fcis-30157	629	13	classifies	classify	VERB
fcis-30157	629	14	but	but	CCONJ
fcis-30157	629	15	also	also	ADV
fcis-30157	629	16	provides	provide	VERB
fcis-30157	629	17	critical	critical	ADJ
fcis-30157	629	18	information	information	NOUN
fcis-30157	629	19	,	,	PUNCT
fcis-30157	629	20	such	such	ADJ
fcis-30157	629	21	as	as	ADP
fcis-30157	629	22	lesion	lesion	NOUN
fcis-30157	629	23	boundaries	boundary	NOUN
fcis-30157	629	24	and	and	CCONJ
fcis-30157	629	25	risk	risk	NOUN
fcis-30157	629	26	factors	factor	NOUN
fcis-30157	629	27	,	,	PUNCT
fcis-30157	629	28	aiding	aid	VERB
fcis-30157	629	29	clinical	clinical	ADJ
fcis-30157	629	30	decision	decision	NOUN
fcis-30157	629	31	-	-	PUNCT
fcis-30157	629	32	making	making	NOUN
fcis-30157	629	33	.	.	PUNCT
fcis-30157	630	1	personalized	personalized	ADJ
fcis-30157	630	2	diagnosis	diagnosis	NOUN
fcis-30157	630	3	,	,	PUNCT
fcis-30157	630	4	considering	consider	VERB
fcis-30157	630	5	factors	factor	NOUN
fcis-30157	630	6	like	like	ADP
fcis-30157	630	7	age	age	NOUN
fcis-30157	630	8	,	,	PUNCT
fcis-30157	630	9	gender	gender	NOUN
fcis-30157	630	10	,	,	PUNCT
fcis-30157	630	11	and	and	CCONJ
fcis-30157	630	12	genetics	genetic	NOUN
fcis-30157	630	13	,	,	PUNCT
fcis-30157	630	14	is	be	AUX
fcis-30157	630	15	a	a	DET
fcis-30157	630	16	promising	promising	ADJ
fcis-30157	630	17	area	area	NOUN
fcis-30157	630	18	for	for	ADP
fcis-30157	630	19	ai	ai	PROPN
fcis-30157	630	20	.	.	PUNCT
fcis-30157	630	21	models	model	NOUN
fcis-30157	630	22	trained	train	VERB
fcis-30157	630	23	to	to	PART
fcis-30157	630	24	incorporate	incorporate	VERB
fcis-30157	630	25	these	these	DET
fcis-30157	630	26	characteristics	characteristic	NOUN
fcis-30157	630	27	can	can	AUX
fcis-30157	630	28	improve	improve	VERB
fcis-30157	630	29	diagnostic	diagnostic	ADJ
fcis-30157	630	30	accuracy	accuracy	NOUN
fcis-30157	630	31	.	.	PUNCT
fcis-30157	631	1	for	for	ADP
fcis-30157	631	2	example	example	NOUN
fcis-30157	631	3	,	,	PUNCT
fcis-30157	631	4	an	an	DET
fcis-30157	631	5	enhanced	enhanced	ADJ
fcis-30157	631	6	mobilenet	mobilenet	NOUN
fcis-30157	631	7	model	model	NOUN
fcis-30157	631	8	improved	improve	VERB
fcis-30157	631	9	actinic	actinic	ADJ
fcis-30157	631	10	keratosis	keratosis	NOUN
fcis-30157	631	11	diagnosis	diagnosis	NOUN
fcis-30157	631	12	by	by	ADP
fcis-30157	631	13	factoring	factor	VERB
fcis-30157	631	14	in	in	ADP
fcis-30157	631	15	patient	patient	NOUN
fcis-30157	631	16	-	-	PUNCT
fcis-30157	631	17	specific	specific	ADJ
fcis-30157	631	18	features	feature	NOUN
fcis-30157	631	19	.	.	PUNCT
fcis-30157	632	1	data	datum	NOUN
fcis-30157	632	2	augmentation	augmentation	NOUN
fcis-30157	632	3	and	and	CCONJ
fcis-30157	632	4	preprocessing	preprocessing	NOUN
fcis-30157	632	5	are	be	AUX
fcis-30157	632	6	essential	essential	ADJ
fcis-30157	632	7	to	to	PART
fcis-30157	632	8	overcome	overcome	VERB
fcis-30157	632	9	issues	issue	NOUN
fcis-30157	632	10	like	like	ADP
fcis-30157	632	11	poor	poor	ADJ
fcis-30157	632	12	resolution	resolution	NOUN
fcis-30157	632	13	and	and	CCONJ
fcis-30157	632	14	variability	variability	NOUN
fcis-30157	632	15	in	in	ADP
fcis-30157	632	16	skin	skin	NOUN
fcis-30157	632	17	tones	tone	NOUN
fcis-30157	632	18	.	.	PUNCT
fcis-30157	633	1	techniques	technique	NOUN
fcis-30157	633	2	such	such	ADJ
fcis-30157	633	3	as	as	ADP
fcis-30157	633	4	rotation	rotation	NOUN
fcis-30157	633	5	,	,	PUNCT
fcis-30157	633	6	scaling	scaling	NOUN
fcis-30157	633	7	,	,	PUNCT
fcis-30157	633	8	and	and	CCONJ
fcis-30157	633	9	color	color	NOUN
fcis-30157	633	10	normalization	normalization	NOUN
fcis-30157	633	11	can	can	AUX
fcis-30157	633	12	improve	improve	VERB
fcis-30157	633	13	model	model	NOUN
fcis-30157	633	14	performance	performance	NOUN
fcis-30157	633	15	and	and	CCONJ
fcis-30157	633	16	generalization	generalization	NOUN
fcis-30157	633	17	to	to	ADP
fcis-30157	633	18	new	new	ADJ
fcis-30157	633	19	images	image	NOUN
fcis-30157	633	20	.	.	PUNCT
fcis-30157	634	1	collaborative	collaborative	ADJ
fcis-30157	634	2	research	research	NOUN
fcis-30157	634	3	between	between	ADP
fcis-30157	634	4	ai	ai	VERB
fcis-30157	634	5	developers	developer	NOUN
fcis-30157	634	6	and	and	CCONJ
fcis-30157	634	7	clinicians	clinician	NOUN
fcis-30157	634	8	is	be	AUX
fcis-30157	634	9	vital	vital	ADJ
fcis-30157	634	10	.	.	PUNCT
fcis-30157	635	1	close	close	ADJ
fcis-30157	635	2	collaboration	collaboration	NOUN
fcis-30157	635	3	ensures	ensure	NOUN
fcis-30157	635	4	that	that	PRON
fcis-30157	635	5	ai	ai	VERB
fcis-30157	635	6	solutions	solution	NOUN
fcis-30157	635	7	address	address	VERB
fcis-30157	635	8	clinical	clinical	ADJ
fcis-30157	635	9	challenges	challenge	NOUN
fcis-30157	635	10	effectively	effectively	ADV
fcis-30157	635	11	.	.	PUNCT
fcis-30157	636	1	a	a	DET
fcis-30157	636	2	study	study	NOUN
fcis-30157	636	3	on	on	ADP
fcis-30157	636	4	contact	contact	NOUN
fcis-30157	636	5	dermatitis	dermatitis	NOUN
fcis-30157	636	6	showed	show	VERB
fcis-30157	636	7	125	125	NUM
fcis-30157	636	8	how	how	SCONJ
fcis-30157	636	9	integrating	integrating	NOUN
fcis-30157	636	10	ai	ai	VERB
fcis-30157	636	11	with	with	ADP
fcis-30157	636	12	clinical	clinical	ADJ
fcis-30157	636	13	expertise	expertise	NOUN
fcis-30157	636	14	can	can	AUX
fcis-30157	636	15	enhance	enhance	VERB
fcis-30157	636	16	diagnostic	diagnostic	ADJ
fcis-30157	636	17	accuracy	accuracy	NOUN
fcis-30157	636	18	and	and	CCONJ
fcis-30157	636	19	support	support	NOUN
fcis-30157	636	20	dermatologists	dermatologist	NOUN
fcis-30157	636	21	in	in	ADP
fcis-30157	636	22	daily	daily	ADJ
fcis-30157	636	23	practice	practice	NOUN
fcis-30157	636	24	.	.	PUNCT
fcis-30157	637	1	finally	finally	ADV
fcis-30157	637	2	,	,	PUNCT
fcis-30157	637	3	the	the	DET
fcis-30157	637	4	ethical	ethical	ADJ
fcis-30157	637	5	implications	implication	NOUN
fcis-30157	637	6	of	of	ADP
fcis-30157	637	7	ai	ai	NOUN
fcis-30157	637	8	,	,	PUNCT
fcis-30157	637	9	such	such	ADJ
fcis-30157	637	10	as	as	ADP
fcis-30157	637	11	data	data	NOUN
fcis-30157	637	12	privacy	privacy	NOUN
fcis-30157	637	13	,	,	PUNCT
fcis-30157	637	14	bias	bias	NOUN
fcis-30157	637	15	,	,	PUNCT
fcis-30157	637	16	and	and	CCONJ
fcis-30157	637	17	accountability	accountability	NOUN
fcis-30157	637	18	,	,	PUNCT
fcis-30157	637	19	must	must	AUX
fcis-30157	637	20	be	be	AUX
fcis-30157	637	21	addressed	address	VERB
fcis-30157	637	22	.	.	PUNCT
fcis-30157	638	1	bias	bias	NOUN
fcis-30157	638	2	in	in	ADP
fcis-30157	638	3	ai	ai	PROPN
fcis-30157	638	4	models	model	NOUN
fcis-30157	638	5	,	,	PUNCT
fcis-30157	638	6	particularly	particularly	ADV
fcis-30157	638	7	when	when	SCONJ
fcis-30157	638	8	trained	train	VERB
fcis-30157	638	9	on	on	ADP
fcis-30157	638	10	non	non	ADJ
fcis-30157	638	11	-	-	ADJ
fcis-30157	638	12	representative	representative	ADJ
fcis-30157	638	13	datasets	dataset	NOUN
fcis-30157	638	14	,	,	PUNCT
fcis-30157	638	15	is	be	AUX
fcis-30157	638	16	a	a	DET
fcis-30157	638	17	significant	significant	ADJ
fcis-30157	638	18	concern	concern	NOUN
fcis-30157	638	19	.	.	PUNCT
fcis-30157	639	1	ensuring	ensure	VERB
fcis-30157	639	2	fairness	fairness	NOUN
fcis-30157	639	3	,	,	PUNCT
fcis-30157	639	4	transparency	transparency	NOUN
fcis-30157	639	5	,	,	PUNCT
fcis-30157	639	6	and	and	CCONJ
fcis-30157	639	7	trust	trust	NOUN
fcis-30157	639	8	in	in	ADP
fcis-30157	639	9	ai	ai	PROPN
fcis-30157	639	10	systems	system	NOUN
fcis-30157	639	11	is	be	AUX
fcis-30157	639	12	crucial	crucial	ADJ
fcis-30157	639	13	for	for	ADP
fcis-30157	639	14	their	their	PRON
fcis-30157	639	15	successful	successful	ADJ
fcis-30157	639	16	integration	integration	NOUN
fcis-30157	639	17	into	into	ADP
fcis-30157	639	18	clinical	clinical	ADJ
fcis-30157	639	19	practice	practice	NOUN
fcis-30157	639	20	.	.	PUNCT
fcis-30157	640	1	in	in	ADP
fcis-30157	640	2	summary	summary	NOUN
fcis-30157	640	3	,	,	PUNCT
fcis-30157	640	4	future	future	ADJ
fcis-30157	640	5	research	research	NOUN
fcis-30157	640	6	should	should	AUX
fcis-30157	640	7	focus	focus	VERB
fcis-30157	640	8	on	on	ADP
fcis-30157	640	9	advancing	advance	VERB
fcis-30157	640	10	deep	deep	ADJ
fcis-30157	640	11	learning	learning	NOUN
fcis-30157	640	12	models	model	NOUN
fcis-30157	640	13	,	,	PUNCT
fcis-30157	640	14	utilizing	utilize	VERB
fcis-30157	640	15	big	big	ADJ
fcis-30157	640	16	data	datum	NOUN
fcis-30157	640	17	for	for	ADP
fcis-30157	640	18	personalized	personalized	ADJ
fcis-30157	640	19	diagnosis	diagnosis	NOUN
fcis-30157	640	20	,	,	PUNCT
fcis-30157	640	21	conducting	conduct	VERB
fcis-30157	640	22	multi	multi	ADJ
fcis-30157	640	23	-	-	ADJ
fcis-30157	640	24	center	center	ADJ
fcis-30157	640	25	studies	study	NOUN
fcis-30157	640	26	,	,	PUNCT
fcis-30157	640	27	and	and	CCONJ
fcis-30157	640	28	integrating	integrate	VERB
fcis-30157	640	29	explainable	explainable	ADJ
fcis-30157	640	30	ai	ai	NOUN
fcis-30157	640	31	.	.	PUNCT
fcis-30157	641	1	developing	develop	VERB
fcis-30157	641	2	multi	multi	ADJ
fcis-30157	641	3	-	-	ADJ
fcis-30157	641	4	task	task	NOUN
fcis-30157	641	5	models	model	NOUN
fcis-30157	641	6	,	,	PUNCT
fcis-30157	641	7	improving	improve	VERB
fcis-30157	641	8	data	datum	NOUN
fcis-30157	641	9	preprocessing	preprocessing	NOUN
fcis-30157	641	10	,	,	PUNCT
fcis-30157	641	11	fostering	foster	VERB
fcis-30157	641	12	collaboration	collaboration	NOUN
fcis-30157	641	13	,	,	PUNCT
fcis-30157	641	14	and	and	CCONJ
fcis-30157	641	15	addressing	address	VERB
fcis-30157	641	16	ethical	ethical	ADJ
fcis-30157	641	17	concerns	concern	NOUN
fcis-30157	641	18	will	will	AUX
fcis-30157	641	19	be	be	AUX
fcis-30157	641	20	essential	essential	ADJ
fcis-30157	641	21	for	for	ADP
fcis-30157	641	22	the	the	DET
fcis-30157	641	23	adoption	adoption	NOUN
fcis-30157	641	24	and	and	CCONJ
fcis-30157	641	25	effectiveness	effectiveness	NOUN
fcis-30157	641	26	of	of	ADP
fcis-30157	641	27	ai	ai	VERB
fcis-30157	641	28	in	in	ADP
fcis-30157	641	29	clinical	clinical	ADJ
fcis-30157	641	30	skin	skin	NOUN
fcis-30157	641	31	disease	disease	NOUN
fcis-30157	641	32	diagnosis	diagnosis	NOUN
fcis-30157	641	33	.	.	PUNCT
fcis-30157	642	1	acknowledgments	acknowledgment	NOUN
fcis-30157	642	2	this	this	DET
fcis-30157	642	3	study	study	NOUN
fcis-30157	642	4	was	be	AUX
fcis-30157	642	5	supported	support	VERB
fcis-30157	642	6	by	by	ADP
fcis-30157	642	7	the	the	DET
fcis-30157	642	8	"	"	PUNCT
fcis-30157	642	9	research	research	NOUN
fcis-30157	642	10	on	on	ADP
fcis-30157	642	11	automatic	automatic	ADJ
fcis-30157	642	12	recognition	recognition	NOUN
fcis-30157	642	13	and	and	CCONJ
fcis-30157	642	14	auxiliary	auxiliary	ADJ
fcis-30157	642	15	diagnosis	diagnosis	NOUN
fcis-30157	642	16	of	of	ADP
fcis-30157	642	17	artificial	artificial	ADJ
fcis-30157	642	18	intelligence	intelligence	NOUN
fcis-30157	642	19	in	in	ADP
fcis-30157	642	20	skin	skin	NOUN
fcis-30157	642	21	diseases	disease	NOUN
fcis-30157	642	22	"	"	PUNCT
fcis-30157	642	23	project	project	NOUN
fcis-30157	642	24	of	of	ADP
fcis-30157	642	25	guangxi	guangxi	PROPN
fcis-30157	642	26	college	college	PROPN
fcis-30157	642	27	students	student	NOUN
fcis-30157	642	28	'	'	PART
fcis-30157	642	29	innovation	innovation	NOUN
fcis-30157	642	30	and	and	CCONJ
fcis-30157	642	31	entrepreneurship	entrepreneurship	NOUN
fcis-30157	642	32	training	training	NOUN
fcis-30157	642	33	program	program	NOUN
fcis-30157	642	34	in	in	ADP
fcis-30157	642	35	2024	2024	NUM
fcis-30157	642	36	,	,	PUNCT
fcis-30157	642	37	project	project	NOUN
fcis-30157	642	38	number	number	NOUN
fcis-30157	642	39	:	:	PUNCT
fcis-30157	642	40	202410599047	202410599047	NUM
fcis-30157	642	41	.	.	PUNCT
fcis-30157	643	1	references	reference	NOUN
fcis-30157	643	2	[	[	X
fcis-30157	643	3	1	1	NUM
fcis-30157	643	4	]	]	SYM
fcis-30157	643	5	li	li	PROPN
fcis-30157	643	6	,	,	PUNCT
fcis-30157	643	7	xuhui	xuhui	PROPN
fcis-30157	643	8	,	,	PUNCT
fcis-30157	643	9	et	et	PROPN
fcis-30157	643	10	al	al	PROPN
fcis-30157	643	11	.	.	PUNCT
fcis-30157	644	1	image	image	NOUN
fcis-30157	644	2	analysis	analysis	NOUN
fcis-30157	644	3	and	and	CCONJ
fcis-30157	644	4	diagnosis	diagnosis	NOUN
fcis-30157	644	5	of	of	ADP
fcis-30157	644	6	skin	skin	NOUN
fcis-30157	644	7	diseasesa	diseasesa	PROPN
fcis-30157	644	8	review	review	NOUN
fcis-30157	644	9	.	.	PUNCT
fcis-30157	645	1	current	current	ADJ
fcis-30157	645	2	medical	medical	ADJ
fcis-30157	645	3	imaging	imaging	NOUN
fcis-30157	645	4	reviews	review	NOUN
fcis-30157	645	5	19.3	19.3	NUM
fcis-30157	645	6	(	(	PUNCT
fcis-30157	645	7	2023	2023	NUM
fcis-30157	645	8	):	):	PUNCT
fcis-30157	645	9	199242	199242	NUM
fcis-30157	645	10	.	.	PUNCT
fcis-30157	646	1	[	[	X
fcis-30157	646	2	2	2	NUM
fcis-30157	646	3	]	]	X
fcis-30157	646	4	gautam	gautam	PROPN
fcis-30157	646	5	,	,	PUNCT
fcis-30157	646	6	vinay	vinay	PROPN
fcis-30157	646	7	,	,	PUNCT
fcis-30157	646	8	et	et	PROPN
fcis-30157	646	9	al	al	PROPN
fcis-30157	646	10	.	.	PROPN
fcis-30157	647	1	early	early	ADJ
fcis-30157	647	2	skin	skin	NOUN
fcis-30157	647	3	disease	disease	NOUN
fcis-30157	647	4	identification	identification	NOUN
fcis-30157	647	5	using	use	VERB
fcis-30157	647	6	eep	eep	NOUN
fcis-30157	647	7	neural	neural	ADJ
fcis-30157	647	8	network	network	NOUN
fcis-30157	647	9	.	.	PUNCT
fcis-30157	648	1	comput	comput	NOUN
fcis-30157	648	2	.	.	PUNCT
fcis-30157	649	1	syst	syst	PROPN
fcis-30157	649	2	.	.	PUNCT
fcis-30157	650	1	sci	sci	PROPN
fcis-30157	650	2	.	.	PUNCT
fcis-30157	651	1	eng	eng	PROPN
fcis-30157	651	2	.	.	PROPN
fcis-30157	652	1	44.3	44.3	NUM
fcis-30157	652	2	(	(	PUNCT
fcis-30157	652	3	2023	2023	NUM
fcis-30157	652	4	):	):	PUNCT
fcis-30157	652	5	2259	2259	NUM
fcis-30157	652	6	-	-	SYM
fcis-30157	652	7	2275	2275	NUM
fcis-30157	652	8	.	.	PUNCT
fcis-30157	653	1	[	[	X
fcis-30157	653	2	3	3	NUM
fcis-30157	653	3	]	]	X
fcis-30157	653	4	anand	anand	PROPN
fcis-30157	653	5	,	,	PUNCT
fcis-30157	653	6	vatsala	vatsala	NOUN
fcis-30157	653	7	,	,	PUNCT
fcis-30157	653	8	et	et	PROPN
fcis-30157	653	9	al	al	PROPN
fcis-30157	653	10	.	.	PUNCT
fcis-30157	654	1	an	an	DET
fcis-30157	654	2	automated	automate	VERB
fcis-30157	654	3	deep	deep	ADJ
fcis-30157	654	4	learning	learning	NOUN
fcis-30157	654	5	models	model	NOUN
fcis-30157	654	6	for	for	ADP
fcis-30157	654	7	classification	classification	NOUN
fcis-30157	654	8	of	of	ADP
fcis-30157	654	9	skin	skin	NOUN
fcis-30157	654	10	disease	disease	NOUN
fcis-30157	654	11	using	use	VERB
fcis-30157	654	12	dermoscopy	dermoscopy	NOUN
fcis-30157	654	13	images	image	NOUN
fcis-30157	654	14	:	:	PUNCT
fcis-30157	654	15	a	a	DET
fcis-30157	654	16	comprehensive	comprehensive	ADJ
fcis-30157	654	17	study	study	NOUN
fcis-30157	654	18	.	.	PUNCT
fcis-30157	655	1	multimedia	multimedia	NOUN
fcis-30157	655	2	tools	tool	NOUN
fcis-30157	655	3	and	and	CCONJ
fcis-30157	655	4	applications	application	NOUN
fcis-30157	655	5	81.26	81.26	NUM
fcis-30157	655	6	(	(	PUNCT
fcis-30157	655	7	2022	2022	NUM
fcis-30157	655	8	):	):	PUNCT
fcis-30157	655	9	37379	37379	NUM
fcis-30157	655	10	-	-	SYM
fcis-30157	655	11	37401	37401	NUM
fcis-30157	655	12	.	.	PUNCT
fcis-30157	656	1	[	[	X
fcis-30157	656	2	4	4	NUM
fcis-30157	656	3	]	]	PUNCT
fcis-30157	656	4	combalia	combalia	PROPN
fcis-30157	656	5	,	,	PUNCT
fcis-30157	656	6	marc	marc	PROPN
fcis-30157	656	7	,	,	PUNCT
fcis-30157	656	8	et	et	PROPN
fcis-30157	656	9	al	al	PROPN
fcis-30157	656	10	.	.	PUNCT
fcis-30157	656	11	validation	validation	NOUN
fcis-30157	656	12	of	of	ADP
fcis-30157	656	13	artificial	artificial	ADJ
fcis-30157	656	14	intelligence	intelligence	NOUN
fcis-30157	656	15	prediction	prediction	NOUN
fcis-30157	656	16	models	model	NOUN
fcis-30157	656	17	for	for	ADP
fcis-30157	656	18	skin	skin	NOUN
fcis-30157	656	19	cancer	cancer	NOUN
fcis-30157	656	20	diagnosis	diagnosis	NOUN
fcis-30157	656	21	using	use	VERB
fcis-30157	656	22	dermoscopy	dermoscopy	NOUN
fcis-30157	656	23	images	image	NOUN
fcis-30157	656	24	:	:	PUNCT
fcis-30157	656	25	the	the	DET
fcis-30157	656	26	2019	2019	NUM
fcis-30157	656	27	international	international	ADJ
fcis-30157	656	28	skin	skin	NOUN
fcis-30157	656	29	imaging	imaging	NOUN
fcis-30157	656	30	collaboration	collaboration	NOUN
fcis-30157	656	31	grand	grand	ADJ
fcis-30157	656	32	challenge	challenge	NOUN
fcis-30157	656	33	.	.	PUNCT
fcis-30157	657	1	the	the	DET
fcis-30157	657	2	lancet	lancet	PROPN
fcis-30157	657	3	digital	digital	PROPN
fcis-30157	657	4	health	health	PROPN
fcis-30157	657	5	4.5	4.5	NUM
fcis-30157	657	6	(	(	PUNCT
fcis-30157	657	7	2022	2022	NUM
fcis-30157	657	8	):	):	PUNCT
fcis-30157	657	9	e330e339	e330e339	X
fcis-30157	657	10	.	.	PUNCT
fcis-30157	658	1	[	[	X
fcis-30157	658	2	5	5	NUM
fcis-30157	658	3	]	]	X
fcis-30157	658	4	nigar	nigar	PROPN
fcis-30157	658	5	,	,	PUNCT
fcis-30157	658	6	natasha	natasha	PROPN
fcis-30157	658	7	,	,	PUNCT
fcis-30157	658	8	et	et	PROPN
fcis-30157	658	9	al	al	PROPN
fcis-30157	658	10	.	.	PUNCT
fcis-30157	659	1	a	a	DET
fcis-30157	659	2	deep	deep	ADJ
fcis-30157	659	3	learning	learning	NOUN
fcis-30157	659	4	approach	approach	NOUN
fcis-30157	659	5	based	base	VERB
fcis-30157	659	6	on	on	ADP
fcis-30157	659	7	explainable	explainable	ADJ
fcis-30157	659	8	artificial	artificial	ADJ
fcis-30157	659	9	intelligence	intelligence	NOUN
fcis-30157	659	10	for	for	ADP
fcis-30157	659	11	skin	skin	NOUN
fcis-30157	659	12	lesion	lesion	NOUN
fcis-30157	659	13	classification	classification	NOUN
fcis-30157	659	14	.	.	PUNCT
fcis-30157	660	1	ieee	ieee	NOUN
fcis-30157	660	2	access	access	NOUN
fcis-30157	660	3	10	10	NUM
fcis-30157	660	4	(	(	PUNCT
fcis-30157	660	5	2022	2022	NUM
fcis-30157	660	6	):	):	PUNCT
fcis-30157	660	7	113715	113715	NUM
fcis-30157	660	8	-	-	SYM
fcis-30157	660	9	113725	113725	NUM
fcis-30157	660	10	.	.	PUNCT
fcis-30157	661	1	[	[	X
fcis-30157	661	2	6	6	NUM
fcis-30157	661	3	]	]	PUNCT
fcis-30157	661	4	aldera	aldera	NOUN
fcis-30157	661	5	,	,	PUNCT
fcis-30157	661	6	shaden	shaden	ADJ
fcis-30157	661	7	abdulaziz	abdulaziz	PROPN
fcis-30157	661	8	,	,	PUNCT
fcis-30157	661	9	and	and	CCONJ
fcis-30157	661	10	mohamed	mohamed	PROPN
fcis-30157	661	11	tahar	tahar	PROPN
fcis-30157	661	12	ben	ben	PROPN
fcis-30157	661	13	othman	othman	PROPN
fcis-30157	661	14	.	.	PUNCT
fcis-30157	662	1	a	a	DET
fcis-30157	662	2	model	model	NOUN
fcis-30157	662	3	for	for	ADP
fcis-30157	662	4	classification	classification	NOUN
fcis-30157	662	5	and	and	CCONJ
fcis-30157	662	6	diagnosis	diagnosis	NOUN
fcis-30157	662	7	of	of	ADP
fcis-30157	662	8	skin	skin	NOUN
fcis-30157	662	9	disease	disease	NOUN
fcis-30157	662	10	using	use	VERB
fcis-30157	662	11	machine	machine	NOUN
fcis-30157	662	12	learning	learning	NOUN
fcis-30157	662	13	and	and	CCONJ
fcis-30157	662	14	image	image	NOUN
fcis-30157	662	15	processing	processing	NOUN
fcis-30157	662	16	techniques	technique	NOUN
fcis-30157	662	17	.	.	PUNCT
fcis-30157	663	1	international	international	ADJ
fcis-30157	663	2	journal	journal	NOUN
fcis-30157	663	3	of	of	ADP
fcis-30157	663	4	advanced	advanced	ADJ
fcis-30157	663	5	computer	computer	NOUN
fcis-30157	663	6	science	science	NOUN
fcis-30157	663	7	and	and	CCONJ
fcis-30157	663	8	applications	application	NOUN
fcis-30157	663	9	13.5	13.5	NUM
fcis-30157	663	10	(	(	PUNCT
fcis-30157	663	11	2022	2022	NUM
fcis-30157	663	12	)	)	PUNCT
fcis-30157	663	13	.	.	PUNCT
fcis-30157	664	1	[	[	X
fcis-30157	664	2	7	7	NUM
fcis-30157	664	3	]	]	X
fcis-30157	664	4	wei	wei	PROPN
fcis-30157	664	5	,	,	PUNCT
fcis-30157	664	6	zenghui	zenghui	PROPN
fcis-30157	664	7	,	,	PUNCT
fcis-30157	664	8	qiang	qiang	PROPN
fcis-30157	664	9	li	li	PROPN
fcis-30157	664	10	,	,	PUNCT
fcis-30157	664	11	and	and	CCONJ
fcis-30157	664	12	hong	hong	PROPN
fcis-30157	664	13	song	song	NOUN
fcis-30157	664	14	.	.	PUNCT
fcis-30157	665	1	dual	dual	ADJ
fcis-30157	665	2	attention	attention	NOUN
fcis-30157	665	3	-	-	PUNCT
fcis-30157	665	4	based	base	VERB
fcis-30157	665	5	network	network	NOUN
fcis-30157	665	6	for	for	ADP
fcis-30157	665	7	skin	skin	NOUN
fcis-30157	665	8	lesion	lesion	NOUN
fcis-30157	665	9	classification	classification	NOUN
fcis-30157	665	10	with	with	ADP
fcis-30157	665	11	auxiliary	auxiliary	ADJ
fcis-30157	665	12	learning	learning	NOUN
fcis-30157	665	13	.	.	PUNCT
fcis-30157	666	1	biomedical	biomedical	ADJ
fcis-30157	666	2	signal	signal	NOUN
fcis-30157	666	3	processing	processing	NOUN
fcis-30157	666	4	and	and	CCONJ
fcis-30157	666	5	control	control	NOUN
fcis-30157	666	6	74	74	NUM
fcis-30157	666	7	(	(	PUNCT
fcis-30157	666	8	2022	2022	NUM
fcis-30157	666	9	):	):	PUNCT
fcis-30157	666	10	103549	103549	NUM
fcis-30157	666	11	.	.	PUNCT
fcis-30157	667	1	[	[	X
fcis-30157	667	2	8	8	NUM
fcis-30157	667	3	]	]	X
fcis-30157	667	4	chen	chen	PROPN
fcis-30157	667	5	,	,	PUNCT
fcis-30157	667	6	qian	qian	PROPN
fcis-30157	667	7	,	,	PUNCT
fcis-30157	667	8	et	et	PROPN
fcis-30157	667	9	al	al	PROPN
fcis-30157	667	10	.	.	PUNCT
fcis-30157	667	11	mdfnet	mdfnet	NOUN
fcis-30157	667	12	:	:	PUNCT
fcis-30157	667	13	application	application	NOUN
fcis-30157	667	14	of	of	ADP
fcis-30157	667	15	multimodal	multimodal	NOUN
fcis-30157	667	16	fusion	fusion	NOUN
fcis-30157	667	17	method	method	NOUN
fcis-30157	667	18	based	base	VERB
fcis-30157	667	19	on	on	ADP
fcis-30157	667	20	skin	skin	NOUN
fcis-30157	667	21	image	image	NOUN
fcis-30157	667	22	and	and	CCONJ
fcis-30157	667	23	clinical	clinical	ADJ
fcis-30157	667	24	data	datum	NOUN
fcis-30157	667	25	to	to	ADP
fcis-30157	667	26	skin	skin	NOUN
fcis-30157	667	27	cancer	cancer	NOUN
fcis-30157	667	28	classification	classification	NOUN
fcis-30157	667	29	.	.	PUNCT
fcis-30157	668	1	journal	journal	PROPN
fcis-30157	668	2	of	of	ADP
fcis-30157	668	3	cancer	cancer	NOUN
fcis-30157	668	4	research	research	NOUN
fcis-30157	668	5	and	and	CCONJ
fcis-30157	668	6	clinical	clinical	ADJ
fcis-30157	668	7	oncology	oncology	NOUN
fcis-30157	668	8	149.7	149.7	NUM
fcis-30157	668	9	(	(	PUNCT
fcis-30157	668	10	2023	2023	NUM
fcis-30157	668	11	):	):	PUNCT
fcis-30157	668	12	3287	3287	NUM
fcis-30157	668	13	-	-	SYM
fcis-30157	668	14	3299	3299	NUM
fcis-30157	668	15	.	.	PUNCT
fcis-30157	669	1	[	[	X
fcis-30157	669	2	9	9	NUM
fcis-30157	669	3	]	]	X
fcis-30157	669	4	xiao	xiao	PROPN
fcis-30157	669	5	,	,	PUNCT
fcis-30157	669	6	junsheng	junsheng	PROPN
fcis-30157	669	7	,	,	PUNCT
fcis-30157	669	8	jirui	jirui	PROPN
fcis-30157	669	9	li	li	PROPN
fcis-30157	669	10	,	,	PUNCT
fcis-30157	669	11	and	and	CCONJ
fcis-30157	669	12	honghao	honghao	PROPN
fcis-30157	669	13	gao	gao	PROPN
fcis-30157	669	14	.	.	PUNCT
fcis-30157	670	1	fs3dciot	fs3dciot	PROPN
fcis-30157	670	2	:	:	PUNCT
fcis-30157	670	3	a	a	DET
fcis-30157	670	4	few	few	ADJ
fcis-30157	670	5	-	-	PUNCT
fcis-30157	670	6	shot	shot	NOUN
fcis-30157	670	7	incremental	incremental	ADJ
fcis-30157	670	8	learning	learning	NOUN
fcis-30157	670	9	network	network	NOUN
fcis-30157	670	10	for	for	ADP
fcis-30157	670	11	skin	skin	NOUN
fcis-30157	670	12	disease	disease	NOUN
fcis-30157	670	13	differential	differential	ADJ
fcis-30157	670	14	diagnosis	diagnosis	NOUN
fcis-30157	670	15	in	in	ADP
fcis-30157	670	16	the	the	DET
fcis-30157	670	17	consumer	consumer	NOUN
fcis-30157	670	18	iot	iot	NOUN
fcis-30157	670	19	.	.	PUNCT
fcis-30157	671	1	ieee	ieee	NOUN
fcis-30157	671	2	transactions	transaction	NOUN
fcis-30157	671	3	on	on	ADP
fcis-30157	671	4	consumer	consumer	NOUN
fcis-30157	671	5	electronics	electronic	NOUN
fcis-30157	671	6	69.4	69.4	NUM
fcis-30157	671	7	(	(	PUNCT
fcis-30157	671	8	2023	2023	NUM
fcis-30157	671	9	):	):	PUNCT
fcis-30157	671	10	786	786	NUM
fcis-30157	671	11	-	-	SYM
fcis-30157	671	12	797	797	NUM
fcis-30157	671	13	.	.	PUNCT
fcis-30157	672	1	[	[	X
fcis-30157	672	2	10	10	NUM
fcis-30157	672	3	]	]	X
fcis-30157	672	4	groh	groh	NOUN
fcis-30157	672	5	,	,	PUNCT
fcis-30157	672	6	matthew	matthew	PROPN
fcis-30157	672	7	,	,	PUNCT
fcis-30157	672	8	et	et	PROPN
fcis-30157	672	9	al	al	PROPN
fcis-30157	672	10	.	.	PUNCT
fcis-30157	673	1	deep	deep	ADJ
fcis-30157	673	2	learning	learning	NOUN
fcis-30157	673	3	-	-	PUNCT
fcis-30157	673	4	aided	aid	VERB
fcis-30157	673	5	decision	decision	NOUN
fcis-30157	673	6	support	support	NOUN
fcis-30157	673	7	for	for	ADP
fcis-30157	673	8	diagnosis	diagnosis	NOUN
fcis-30157	673	9	of	of	ADP
fcis-30157	673	10	skin	skin	NOUN
fcis-30157	673	11	disease	disease	NOUN
fcis-30157	673	12	across	across	ADP
fcis-30157	673	13	skin	skin	NOUN
fcis-30157	673	14	tones	tone	NOUN
fcis-30157	673	15	.	.	PUNCT
fcis-30157	674	1	nature	nature	NOUN
fcis-30157	674	2	medicine	medicine	NOUN
fcis-30157	674	3	30.2	30.2	NUM
fcis-30157	674	4	(	(	PUNCT
fcis-30157	674	5	2024	2024	NUM
fcis-30157	674	6	):	):	PUNCT
fcis-30157	674	7	573	573	NUM
fcis-30157	674	8	-	-	SYM
fcis-30157	674	9	583	583	NUM
fcis-30157	674	10	.	.	PUNCT
fcis-30157	675	1	[	[	X
fcis-30157	675	2	11	11	NUM
fcis-30157	675	3	]	]	SYM
fcis-30157	675	4	jones	jones	PROPN
fcis-30157	675	5	,	,	PUNCT
fcis-30157	675	6	o.	o.	PROPN
fcis-30157	675	7	t.	t.	PROPN
fcis-30157	675	8	,	,	PUNCT
fcis-30157	675	9	et	et	PROPN
fcis-30157	675	10	al	al	PROPN
fcis-30157	675	11	.	.	PUNCT
fcis-30157	675	12	artificial	artificial	ADJ
fcis-30157	675	13	intelligence	intelligence	NOUN
fcis-30157	675	14	and	and	CCONJ
fcis-30157	675	15	machine	machine	NOUN
fcis-30157	675	16	learning	learn	VERB
fcis-30157	675	17	algorithms	algorithm	NOUN
fcis-30157	675	18	for	for	ADP
fcis-30157	675	19	early	early	ADJ
fcis-30157	675	20	detection	detection	NOUN
fcis-30157	675	21	of	of	ADP
fcis-30157	675	22	skin	skin	NOUN
fcis-30157	675	23	cancer	cancer	NOUN
fcis-30157	675	24	in	in	ADP
fcis-30157	675	25	community	community	NOUN
fcis-30157	675	26	and	and	CCONJ
fcis-30157	675	27	primary	primary	ADJ
fcis-30157	675	28	care	care	NOUN
fcis-30157	675	29	settings	setting	NOUN
fcis-30157	675	30	:	:	PUNCT
fcis-30157	675	31	a	a	DET
fcis-30157	675	32	systematic	systematic	ADJ
fcis-30157	675	33	review	review	NOUN
fcis-30157	675	34	.	.	PUNCT
fcis-30157	676	1	the	the	DET
fcis-30157	676	2	lancet	lancet	PROPN
fcis-30157	676	3	digital	digital	PROPN
fcis-30157	676	4	health	health	PROPN
fcis-30157	676	5	4.6	4.6	NUM
fcis-30157	676	6	(	(	PUNCT
fcis-30157	676	7	2022	2022	NUM
fcis-30157	676	8	):	):	PUNCT
fcis-30157	676	9	e466	e466	PROPN
fcis-30157	676	10	-	-	PUNCT
fcis-30157	676	11	e476	e476	PROPN
fcis-30157	676	12	.	.	PUNCT
fcis-30157	677	1	[	[	X
fcis-30157	677	2	12	12	NUM
fcis-30157	677	3	]	]	X
fcis-30157	677	4	wu	wu	PROPN
fcis-30157	677	5	,	,	PUNCT
fcis-30157	677	6	renkai	renkai	NOUN
fcis-30157	677	7	,	,	PUNCT
fcis-30157	677	8	et	et	PROPN
fcis-30157	677	9	al	al	PROPN
fcis-30157	677	10	.	.	PROPN
fcis-30157	677	11	mhorunet	mhorunet	PROPN
fcis-30157	677	12	:	:	PUNCT
fcis-30157	677	13	high	high	ADJ
fcis-30157	677	14	-	-	PUNCT
fcis-30157	677	15	order	order	NOUN
fcis-30157	677	16	spatial	spatial	ADJ
fcis-30157	677	17	interaction	interaction	NOUN
fcis-30157	677	18	unet	unet	NOUN
fcis-30157	677	19	for	for	ADP
fcis-30157	677	20	skin	skin	NOUN
fcis-30157	677	21	lesion	lesion	NOUN
fcis-30157	677	22	segmentation	segmentation	NOUN
fcis-30157	677	23	.	.	PUNCT
fcis-30157	678	1	biomedical	biomedical	ADJ
fcis-30157	678	2	signal	signal	NOUN
fcis-30157	678	3	processing	processing	NOUN
fcis-30157	678	4	and	and	CCONJ
fcis-30157	678	5	control	control	NOUN
fcis-30157	678	6	88	88	NUM
fcis-30157	678	7	(	(	PUNCT
fcis-30157	678	8	2024	2024	NUM
fcis-30157	678	9	):	):	PUNCT
fcis-30157	678	10	105517	105517	NUM
fcis-30157	678	11	.	.	PUNCT
fcis-30157	679	1	[	[	X
fcis-30157	679	2	13	13	NUM
fcis-30157	679	3	]	]	SYM
fcis-30157	679	4	almuayqil	almuayqil	PROPN
fcis-30157	679	5	,	,	PUNCT
fcis-30157	679	6	saleh	saleh	NOUN
fcis-30157	679	7	naif	naif	NOUN
fcis-30157	679	8	,	,	PUNCT
fcis-30157	679	9	sameh	sameh	PROPN
fcis-30157	679	10	abd	abd	PROPN
fcis-30157	679	11	el	el	PROPN
fcis-30157	679	12	-	-	PUNCT
fcis-30157	679	13	ghany	ghany	NOUN
fcis-30157	679	14	,	,	PUNCT
fcis-30157	679	15	and	and	CCONJ
fcis-30157	679	16	mohammed	mohammed	PROPN
fcis-30157	679	17	elmogy	elmogy	NOUN
fcis-30157	679	18	.	.	PUNCT
fcis-30157	680	1	computer	computer	NOUN
fcis-30157	680	2	-	-	PUNCT
fcis-30157	680	3	aided	aid	VERB
fcis-30157	680	4	diagnosis	diagnosis	NOUN
fcis-30157	680	5	for	for	ADP
fcis-30157	680	6	early	early	ADJ
fcis-30157	680	7	signs	sign	NOUN
fcis-30157	680	8	of	of	ADP
fcis-30157	680	9	skin	skin	NOUN
fcis-30157	680	10	diseases	disease	NOUN
fcis-30157	680	11	using	use	VERB
fcis-30157	680	12	multi	multi	ADJ
fcis-30157	680	13	types	type	NOUN
fcis-30157	680	14	feature	feature	VERB
fcis-30157	680	15	fusion	fusion	NOUN
fcis-30157	680	16	based	base	VERB
fcis-30157	680	17	on	on	ADP
fcis-30157	680	18	a	a	DET
fcis-30157	680	19	hybrid	hybrid	ADJ
fcis-30157	680	20	deep	deep	ADJ
fcis-30157	680	21	learning	learning	NOUN
fcis-30157	680	22	model	model	NOUN
fcis-30157	680	23	.	.	PUNCT
fcis-30157	681	1	electronics	electronic	NOUN
fcis-30157	681	2	11.23	11.23	NUM
fcis-30157	681	3	(	(	PUNCT
fcis-30157	681	4	2022	2022	NUM
fcis-30157	681	5	):	):	PUNCT
fcis-30157	681	6	4009	4009	NUM
fcis-30157	681	7	.	.	PUNCT
fcis-30157	682	1	[	[	X
fcis-30157	682	2	14	14	NUM
fcis-30157	682	3	]	]	X
fcis-30157	682	4	khurshid	khurshid	PROPN
fcis-30157	682	5	,	,	PUNCT
fcis-30157	682	6	mahapara	mahapara	PROPN
fcis-30157	682	7	,	,	PUNCT
fcis-30157	682	8	mayank	mayank	PROPN
fcis-30157	682	9	vatsa	vatsa	PROPN
fcis-30157	682	10	,	,	PUNCT
fcis-30157	682	11	and	and	CCONJ
fcis-30157	682	12	richa	richa	PROPN
fcis-30157	682	13	singh	singh	PROPN
fcis-30157	682	14	.	.	PUNCT
fcis-30157	683	1	optimizing	optimize	VERB
fcis-30157	683	2	skin	skin	NOUN
fcis-30157	683	3	lesion	lesion	NOUN
fcis-30157	683	4	classification	classification	NOUN
fcis-30157	683	5	via	via	ADP
fcis-30157	683	6	multimodal	multimodal	NOUN
fcis-30157	683	7	data	datum	NOUN
fcis-30157	683	8	and	and	CCONJ
fcis-30157	683	9	auxiliary	auxiliary	ADJ
fcis-30157	683	10	task	task	NOUN
fcis-30157	683	11	integration	integration	NOUN
fcis-30157	683	12	.	.	PUNCT
fcis-30157	684	1	2024	2024	NUM
fcis-30157	684	2	ieee	ieee	PROPN
fcis-30157	684	3	international	international	ADJ
fcis-30157	684	4	symposium	symposium	NOUN
fcis-30157	684	5	on	on	ADP
fcis-30157	684	6	biomedical	biomedical	ADJ
fcis-30157	684	7	imaging	imaging	NOUN
fcis-30157	684	8	(	(	PUNCT
fcis-30157	684	9	isbi	isbi	NOUN
fcis-30157	684	10	)	)	PUNCT
fcis-30157	684	11	.	.	PUNCT
fcis-30157	685	1	ieee	ieee	NOUN
fcis-30157	685	2	,	,	PUNCT
fcis-30157	685	3	2024	2024	NUM
fcis-30157	685	4	.	.	PUNCT
fcis-30157	686	1	[	[	X
fcis-30157	686	2	15	15	NUM
fcis-30157	686	3	]	]	X
fcis-30157	686	4	sun	sun	PROPN
fcis-30157	686	5	,	,	PUNCT
fcis-30157	686	6	jie	jie	PROPN
fcis-30157	686	7	,	,	PUNCT
fcis-30157	686	8	et	et	PROPN
fcis-30157	686	9	al	al	PROPN
fcis-30157	686	10	.	.	PROPN
fcis-30157	686	11	machine	machine	NOUN
fcis-30157	686	12	learning	learn	VERB
fcis-30157	686	13	methods	method	NOUN
fcis-30157	686	14	in	in	ADP
fcis-30157	686	15	skin	skin	NOUN
fcis-30157	686	16	disease	disease	NOUN
fcis-30157	686	17	recognition	recognition	NOUN
fcis-30157	686	18	:	:	PUNCT
fcis-30157	686	19	a	a	DET
fcis-30157	686	20	systematic	systematic	ADJ
fcis-30157	686	21	review	review	NOUN
fcis-30157	686	22	.	.	PUNCT
fcis-30157	687	1	processes	process	VERB
fcis-30157	687	2	11.4	11.4	NUM
fcis-30157	687	3	(	(	PUNCT
fcis-30157	687	4	2023	2023	NUM
fcis-30157	687	5	):	):	PUNCT
fcis-30157	687	6	1003	1003	NUM
fcis-30157	687	7	.	.	PUNCT
fcis-30157	688	1	[	[	X
fcis-30157	688	2	16	16	NUM
fcis-30157	688	3	]	]	X
fcis-30157	688	4	agarwal	agarwal	PROPN
fcis-30157	688	5	,	,	PUNCT
fcis-30157	688	6	raghav	raghav	PROPN
fcis-30157	688	7	,	,	PUNCT
fcis-30157	688	8	and	and	CCONJ
fcis-30157	688	9	deepthi	deepthi	PROPN
fcis-30157	688	10	godavarthi	godavarthi	PROPN
fcis-30157	688	11	.	.	PUNCT
fcis-30157	689	1	skin	skin	NOUN
fcis-30157	689	2	disease	disease	NOUN
fcis-30157	689	3	classification	classification	NOUN
fcis-30157	689	4	using	use	VERB
fcis-30157	689	5	cnn	cnn	PROPN
fcis-30157	689	6	algorithms	algorithm	NOUN
fcis-30157	689	7	.	.	PUNCT
fcis-30157	690	1	eai	eai	PROPN
fcis-30157	690	2	endorsed	endorse	VERB
fcis-30157	690	3	transactions	transaction	NOUN
fcis-30157	690	4	on	on	ADP
fcis-30157	690	5	pervasive	pervasive	ADJ
fcis-30157	690	6	health	health	NOUN
fcis-30157	690	7	and	and	CCONJ
fcis-30157	690	8	technology	technology	NOUN
fcis-30157	690	9	9.1	9.1	NUM
fcis-30157	690	10	(	(	PUNCT
fcis-30157	690	11	2023	2023	NUM
fcis-30157	690	12	)	)	PUNCT
fcis-30157	690	13	.	.	PUNCT
fcis-30157	691	1	[	[	X
fcis-30157	691	2	17	17	NUM
fcis-30157	691	3	]	]	X
fcis-30157	691	4	singh	singh	ADJ
fcis-30157	691	5	,	,	PUNCT
fcis-30157	691	6	jagandeep	jagandeep	ADJ
fcis-30157	691	7	,	,	PUNCT
fcis-30157	691	8	jasminder	jasminder	PROPN
fcis-30157	691	9	kaur	kaur	PROPN
fcis-30157	691	10	sandhu	sandhu	PROPN
fcis-30157	691	11	,	,	PUNCT
fcis-30157	691	12	and	and	CCONJ
fcis-30157	691	13	yogesh	yogesh	PROPN
fcis-30157	691	14	kumar	kumar	PROPN
fcis-30157	691	15	.	.	PUNCT
fcis-30157	692	1	an	an	DET
fcis-30157	692	2	analysis	analysis	NOUN
fcis-30157	692	3	of	of	ADP
fcis-30157	692	4	detection	detection	NOUN
fcis-30157	692	5	and	and	CCONJ
fcis-30157	692	6	diagnosis	diagnosis	NOUN
fcis-30157	692	7	of	of	ADP
fcis-30157	692	8	different	different	ADJ
fcis-30157	692	9	classes	class	NOUN
fcis-30157	692	10	of	of	ADP
fcis-30157	692	11	skin	skin	NOUN
fcis-30157	692	12	diseases	disease	NOUN
fcis-30157	692	13	using	use	VERB
fcis-30157	692	14	artificial	artificial	ADJ
fcis-30157	692	15	intelligence	intelligence	NOUN
fcis-30157	692	16	-	-	PUNCT
fcis-30157	692	17	based	base	VERB
fcis-30157	692	18	learning	learning	NOUN
fcis-30157	692	19	approaches	approach	NOUN
fcis-30157	692	20	with	with	ADP
fcis-30157	692	21	hyper	hyper	ADJ
fcis-30157	692	22	parameters	parameter	NOUN
fcis-30157	692	23	.	.	PUNCT
fcis-30157	693	1	archives	archive	NOUN
fcis-30157	693	2	of	of	ADP
fcis-30157	693	3	computational	computational	ADJ
fcis-30157	693	4	methods	method	NOUN
fcis-30157	693	5	in	in	ADP
fcis-30157	693	6	engineering	engineering	NOUN
fcis-30157	693	7	31.2	31.2	NUM
fcis-30157	693	8	(	(	PUNCT
fcis-30157	693	9	2024	2024	NUM
fcis-30157	693	10	):	):	PUNCT
fcis-30157	693	11	1051	1051	NUM
fcis-30157	693	12	-	-	SYM
fcis-30157	693	13	1078	1078	NUM
fcis-30157	693	14	.	.	PUNCT
fcis-30157	694	1	[	[	X
fcis-30157	694	2	18	18	NUM
fcis-30157	694	3	]	]	PUNCT
fcis-30157	694	4	fraiwan	fraiwan	NOUN
fcis-30157	694	5	,	,	PUNCT
fcis-30157	694	6	mohammad	mohammad	PROPN
fcis-30157	694	7	,	,	PUNCT
fcis-30157	694	8	and	and	CCONJ
fcis-30157	694	9	esraa	esraa	VERB
fcis-30157	694	10	faouri	faouri	NOUN
fcis-30157	694	11	.	.	PUNCT
fcis-30157	695	1	on	on	ADP
fcis-30157	695	2	the	the	DET
fcis-30157	695	3	automatic	automatic	ADJ
fcis-30157	695	4	detection	detection	NOUN
fcis-30157	695	5	and	and	CCONJ
fcis-30157	695	6	classification	classification	NOUN
fcis-30157	695	7	of	of	ADP
fcis-30157	695	8	skin	skin	NOUN
fcis-30157	695	9	cancer	cancer	NOUN
fcis-30157	695	10	using	use	VERB
fcis-30157	695	11	deep	deep	ADJ
fcis-30157	695	12	transfer	transfer	NOUN
fcis-30157	695	13	learning	learning	NOUN
fcis-30157	695	14	.	.	PUNCT
fcis-30157	696	1	sensors	sensor	NOUN
fcis-30157	696	2	22.13	22.13	NUM
fcis-30157	696	3	(	(	PUNCT
fcis-30157	696	4	2022	2022	NUM
fcis-30157	696	5	):	):	PUNCT
fcis-30157	696	6	4963	4963	NUM
fcis-30157	696	7	.	.	PUNCT
fcis-30157	697	1	[	[	X
fcis-30157	697	2	19	19	NUM
fcis-30157	697	3	]	]	X
fcis-30157	697	4	escalé	escalé	NOUN
fcis-30157	697	5	-	-	PUNCT
fcis-30157	697	6	besa	besa	PROPN
fcis-30157	697	7	,	,	PUNCT
fcis-30157	697	8	anna	anna	PROPN
fcis-30157	697	9	,	,	PUNCT
fcis-30157	697	10	et	et	PROPN
fcis-30157	697	11	al	al	PROPN
fcis-30157	697	12	.	.	PUNCT
fcis-30157	697	13	exploring	explore	VERB
fcis-30157	697	14	the	the	DET
fcis-30157	697	15	potential	potential	NOUN
fcis-30157	697	16	of	of	ADP
fcis-30157	697	17	artificial	artificial	ADJ
fcis-30157	697	18	intelligence	intelligence	NOUN
fcis-30157	697	19	in	in	ADP
fcis-30157	697	20	improving	improve	VERB
fcis-30157	697	21	skin	skin	NOUN
fcis-30157	697	22	lesion	lesion	NOUN
fcis-30157	697	23	diagnosis	diagnosis	NOUN
fcis-30157	697	24	in	in	ADP
fcis-30157	697	25	primary	primary	ADJ
fcis-30157	697	26	care	care	NOUN
fcis-30157	697	27	.	.	PUNCT
fcis-30157	698	1	scientific	scientific	ADJ
fcis-30157	698	2	reports	report	NOUN
fcis-30157	698	3	13.1	13.1	NUM
fcis-30157	698	4	(	(	PUNCT
fcis-30157	698	5	2023	2023	NUM
fcis-30157	698	6	):	):	PUNCT
fcis-30157	698	7	4293	4293	NUM
fcis-30157	698	8	.	.	PUNCT
fcis-30157	699	1	[	[	X
fcis-30157	699	2	20	20	NUM
fcis-30157	699	3	]	]	X
fcis-30157	699	4	pesqué	pesqué	PROPN
fcis-30157	699	5	,	,	PUNCT
fcis-30157	699	6	david	david	PROPN
fcis-30157	699	7	,	,	PUNCT
fcis-30157	699	8	et	et	PROPN
fcis-30157	699	9	al	al	PROPN
fcis-30157	699	10	.	.	PROPN
fcis-30157	699	11	differential	differential	ADJ
fcis-30157	699	12	diagnosis	diagnosis	NOUN
fcis-30157	699	13	of	of	ADP
fcis-30157	699	14	contact	contact	NOUN
fcis-30157	699	15	dermatitis	dermatitis	PROPN
fcis-30157	699	16	:	:	PUNCT
fcis-30157	699	17	a	a	DET
fcis-30157	699	18	practical‐approach	practical‐approach	NOUN
fcis-30157	699	19	review	review	NOUN
fcis-30157	699	20	by	by	ADP
fcis-30157	699	21	the	the	DET
fcis-30157	699	22	eadv	eadv	ADJ
fcis-30157	699	23	task	task	NOUN
fcis-30157	699	24	force	force	NOUN
fcis-30157	699	25	on	on	ADP
fcis-30157	699	26	contact	contact	NOUN
fcis-30157	699	27	dermatitis	dermatitis	PROPN
fcis-30157	699	28	.	.	PUNCT
fcis-30157	700	1	journal	journal	PROPN
fcis-30157	700	2	of	of	ADP
fcis-30157	700	3	the	the	DET
fcis-30157	700	4	european	european	PROPN
fcis-30157	700	5	academy	academy	PROPN
fcis-30157	700	6	of	of	ADP
fcis-30157	700	7	dermatology	dermatology	NOUN
fcis-30157	700	8	and	and	CCONJ
fcis-30157	700	9	venereology	venereology	NOUN
fcis-30157	700	10	38.9	38.9	NUM
fcis-30157	700	11	(	(	PUNCT
fcis-30157	700	12	2024	2024	NUM
fcis-30157	700	13	):	):	PUNCT
fcis-30157	700	14	1704	1704	NUM
fcis-30157	700	15	-	-	SYM
fcis-30157	700	16	1722	1722	NUM
fcis-30157	700	17	.	.	PUNCT
fcis-30157	701	1	[	[	X
fcis-30157	701	2	21	21	NUM
fcis-30157	701	3	]	]	X
fcis-30157	701	4	kshirsagar	kshirsagar	NOUN
fcis-30157	701	5	,	,	PUNCT
fcis-30157	701	6	pravin	pravin	PROPN
fcis-30157	701	7	r.	r.	PROPN
fcis-30157	701	8	,	,	PUNCT
fcis-30157	701	9	et	et	PROPN
fcis-30157	701	10	al	al	PROPN
fcis-30157	701	11	.	.	PUNCT
fcis-30157	702	1	deep	deep	ADJ
fcis-30157	702	2	learning	learning	NOUN
fcis-30157	702	3	approaches	approach	VERB
fcis-30157	702	4	for	for	ADP
fcis-30157	702	5	prognosis	prognosis	NOUN
fcis-30157	702	6	of	of	ADP
fcis-30157	702	7	automated	automate	VERB
fcis-30157	702	8	skin	skin	NOUN
fcis-30157	702	9	disease	disease	NOUN
fcis-30157	702	10	.	.	PUNCT
fcis-30157	703	1	life	life	NOUN
fcis-30157	703	2	12.3	12.3	NUM
fcis-30157	703	3	(	(	PUNCT
fcis-30157	703	4	2022	2022	NUM
fcis-30157	703	5	):	):	PUNCT
fcis-30157	703	6	426	426	NUM
fcis-30157	703	7	.	.	PUNCT
fcis-30157	704	1	[	[	X
fcis-30157	704	2	22	22	NUM
fcis-30157	704	3	]	]	SYM
fcis-30157	704	4	anand	anand	PROPN
fcis-30157	704	5	,	,	PUNCT
fcis-30157	704	6	vatsala	vatsala	NOUN
fcis-30157	704	7	,	,	PUNCT
fcis-30157	704	8	sheifali	sheifali	PROPN
fcis-30157	704	9	gupta	gupta	PROPN
fcis-30157	704	10	,	,	PUNCT
fcis-30157	704	11	and	and	CCONJ
fcis-30157	704	12	deepika	deepika	PROPN
fcis-30157	704	13	koundal	koundal	PROPN
fcis-30157	704	14	.	.	PUNCT
fcis-30157	705	1	skin	skin	NOUN
fcis-30157	705	2	disease	disease	NOUN
fcis-30157	705	3	diagnosis	diagnosis	NOUN
fcis-30157	705	4	:	:	PUNCT
fcis-30157	705	5	challenges	challenge	NOUN
fcis-30157	705	6	and	and	CCONJ
fcis-30157	705	7	opportunities	opportunity	NOUN
fcis-30157	705	8	.	.	PUNCT
fcis-30157	706	1	proceedings	proceeding	NOUN
fcis-30157	706	2	of	of	ADP
fcis-30157	706	3	second	second	ADJ
fcis-30157	706	4	doctoral	doctoral	ADJ
fcis-30157	706	5	symposium	symposium	NOUN
fcis-30157	706	6	on	on	ADP
fcis-30157	706	7	computational	computational	ADJ
fcis-30157	706	8	intelligence	intelligence	NOUN
fcis-30157	706	9	:	:	PUNCT
fcis-30157	706	10	dosci	dosci	PROPN
fcis-30157	706	11	2021	2021	NUM
fcis-30157	706	12	.	.	PUNCT
fcis-30157	707	1	springer	springer	PROPN
fcis-30157	707	2	singapore	singapore	PROPN
fcis-30157	707	3	,	,	PUNCT
fcis-30157	707	4	2022	2022	NUM
fcis-30157	707	5	.	.	PUNCT
fcis-30157	708	1	[	[	X
fcis-30157	708	2	23	23	NUM
fcis-30157	708	3	]	]	PUNCT
fcis-30157	708	4	ahammed	ahamme	VERB
fcis-30157	708	5	,	,	PUNCT
fcis-30157	708	6	mostafiz	mostafiz	PROPN
fcis-30157	708	7	,	,	PUNCT
fcis-30157	708	8	md	md	PROPN
fcis-30157	708	9	al	al	PROPN
fcis-30157	708	10	mamun	mamun	PROPN
fcis-30157	708	11	,	,	PUNCT
fcis-30157	708	12	and	and	CCONJ
fcis-30157	708	13	mohammad	mohammad	PROPN
fcis-30157	708	14	shorif	shorif	PROPN
fcis-30157	708	15	uddin	uddin	PROPN
fcis-30157	708	16	.	.	PUNCT
fcis-30157	709	1	a	a	DET
fcis-30157	709	2	machine	machine	NOUN
fcis-30157	709	3	learning	learn	VERB
fcis-30157	709	4	approach	approach	NOUN
fcis-30157	709	5	for	for	ADP
fcis-30157	709	6	skin	skin	NOUN
fcis-30157	709	7	disease	disease	NOUN
fcis-30157	709	8	detection	detection	NOUN
fcis-30157	709	9	and	and	CCONJ
fcis-30157	709	10	classification	classification	NOUN
fcis-30157	709	11	using	use	VERB
fcis-30157	709	12	image	image	NOUN
fcis-30157	709	13	segmentation	segmentation	NOUN
fcis-30157	709	14	.	.	PUNCT
fcis-30157	710	1	healthcare	healthcare	NOUN
fcis-30157	710	2	analytics	analytic	NOUN
fcis-30157	710	3	2	2	NUM
fcis-30157	710	4	(	(	PUNCT
fcis-30157	710	5	2022	2022	NUM
fcis-30157	710	6	):	):	PUNCT
fcis-30157	710	7	100122	100122	NUM
fcis-30157	710	8	.	.	PUNCT
fcis-30157	711	1	[	[	X
fcis-30157	711	2	24	24	NUM
fcis-30157	711	3	]	]	PUNCT
fcis-30157	711	4	sangers	sanger	NOUN
fcis-30157	711	5	,	,	PUNCT
fcis-30157	711	6	tobias	tobias	PROPN
fcis-30157	711	7	e.	e.	PROPN
fcis-30157	711	8	,	,	PUNCT
fcis-30157	711	9	et	et	PROPN
fcis-30157	711	10	al	al	PROPN
fcis-30157	711	11	.	.	PUNCT
fcis-30157	712	1	position	position	NOUN
fcis-30157	712	2	statement	statement	NOUN
fcis-30157	712	3	of	of	ADP
fcis-30157	712	4	the	the	DET
fcis-30157	712	5	eadv	eadv	ADJ
fcis-30157	712	6	artificial	artificial	ADJ
fcis-30157	712	7	intelligence	intelligence	NOUN
fcis-30157	712	8	(	(	PUNCT
fcis-30157	712	9	ai	ai	NOUN
fcis-30157	712	10	)	)	PUNCT
fcis-30157	712	11	task	task	NOUN
fcis-30157	712	12	force	force	NOUN
fcis-30157	712	13	on	on	ADP
fcis-30157	712	14	ai‐assisted	ai‐assiste	VERB
fcis-30157	712	15	smartphone	smartphone	NOUN
fcis-30157	712	16	apps	app	NOUN
fcis-30157	712	17	and	and	CCONJ
fcis-30157	712	18	web‐based	web‐base	VERB
fcis-30157	712	19	services	service	NOUN
fcis-30157	712	20	for	for	ADP
fcis-30157	712	21	skin	skin	NOUN
fcis-30157	712	22	disease	disease	NOUN
fcis-30157	712	23	.	.	PUNCT
fcis-30157	713	1	journal	journal	NOUN
fcis-30157	713	2	of	of	ADP
fcis-30157	713	3	the	the	DET
fcis-30157	713	4	european	european	PROPN
fcis-30157	713	5	academy	academy	PROPN
fcis-30157	713	6	of	of	ADP
fcis-30157	713	7	dermatology	dermatology	NOUN
fcis-30157	713	8	and	and	CCONJ
fcis-30157	713	9	venereology	venereology	NOUN
fcis-30157	713	10	38.1	38.1	NUM
fcis-30157	713	11	(	(	PUNCT
fcis-30157	713	12	2024	2024	NUM
fcis-30157	713	13	):	):	PUNCT
fcis-30157	713	14	22	22	NUM
fcis-30157	713	15	-	-	SYM
fcis-30157	713	16	30	30	NUM
fcis-30157	713	17	.	.	PUNCT
fcis-30157	714	1	[	[	X
fcis-30157	714	2	25	25	NUM
fcis-30157	714	3	]	]	SYM
fcis-30157	714	4	li	li	PROPN
fcis-30157	714	5	,	,	PUNCT
fcis-30157	714	6	zhouxiao	zhouxiao	PROPN
fcis-30157	714	7	,	,	PUNCT
fcis-30157	714	8	et	et	PROPN
fcis-30157	714	9	al	al	PROPN
fcis-30157	714	10	.	.	PUNCT
fcis-30157	714	11	artificial	artificial	ADJ
fcis-30157	714	12	intelligence	intelligence	NOUN
fcis-30157	714	13	in	in	ADP
fcis-30157	714	14	dermatology	dermatology	NOUN
fcis-30157	714	15	image	image	NOUN
fcis-30157	714	16	analysis	analysis	NOUN
fcis-30157	714	17	:	:	PUNCT
fcis-30157	714	18	current	current	ADJ
fcis-30157	714	19	developments	development	NOUN
fcis-30157	714	20	and	and	CCONJ
fcis-30157	714	21	future	future	ADJ
fcis-30157	714	22	trends	trend	NOUN
fcis-30157	714	23	.	.	PUNCT
fcis-30157	715	1	journal	journal	NOUN
fcis-30157	715	2	of	of	ADP
fcis-30157	715	3	clinical	clinical	ADJ
fcis-30157	715	4	medicine	medicine	NOUN
fcis-30157	715	5	11.22	11.22	NUM
fcis-30157	715	6	(	(	PUNCT
fcis-30157	715	7	2022	2022	NUM
fcis-30157	715	8	):	):	PUNCT
fcis-30157	715	9	6826	6826	NUM
fcis-30157	715	10	.	.	PUNCT
fcis-30157	716	1	[	[	X
fcis-30157	716	2	26	26	NUM
fcis-30157	716	3	]	]	X
fcis-30157	716	4	inthiyaz	inthiyaz	NOUN
fcis-30157	716	5	,	,	PUNCT
fcis-30157	716	6	syed	syed	ADJ
fcis-30157	716	7	,	,	PUNCT
fcis-30157	716	8	et	et	PROPN
fcis-30157	716	9	al	al	PROPN
fcis-30157	716	10	.	.	PROPN
fcis-30157	716	11	skin	skin	NOUN
fcis-30157	716	12	disease	disease	NOUN
fcis-30157	716	13	detection	detection	NOUN
fcis-30157	716	14	using	use	VERB
fcis-30157	716	15	deep	deep	ADJ
fcis-30157	716	16	learning	learning	NOUN
fcis-30157	716	17	.	.	PUNCT
fcis-30157	717	1	advances	advance	NOUN
fcis-30157	717	2	in	in	ADP
fcis-30157	717	3	engineering	engineering	NOUN
fcis-30157	717	4	software	software	NOUN
fcis-30157	717	5	175	175	NUM
fcis-30157	717	6	(	(	PUNCT
fcis-30157	717	7	2023	2023	NUM
fcis-30157	717	8	):	):	PUNCT
fcis-30157	717	9	103361	103361	NUM
fcis-30157	717	10	.	.	PUNCT
fcis-30157	718	1	[	[	X
fcis-30157	718	2	27	27	NUM
fcis-30157	718	3	]	]	SYM
fcis-30157	718	4	li	li	PROPN
fcis-30157	718	5	,	,	PUNCT
fcis-30157	718	6	shiyang	shiyang	PROPN
fcis-30157	718	7	,	,	PUNCT
fcis-30157	718	8	et	et	PROPN
fcis-30157	718	9	al	al	PROPN
fcis-30157	718	10	.	.	PUNCT
fcis-30157	719	1	an	an	DET
fcis-30157	719	2	actinic	actinic	ADJ
fcis-30157	719	3	keratosis	keratosis	NOUN
fcis-30157	719	4	auxiliary	auxiliary	ADJ
fcis-30157	719	5	diagnosis	diagnosis	NOUN
fcis-30157	719	6	method	method	NOUN
fcis-30157	719	7	based	base	VERB
fcis-30157	719	8	on	on	ADP
fcis-30157	719	9	an	an	DET
fcis-30157	719	10	enhanced	enhanced	ADJ
fcis-30157	719	11	mobilenet	mobilenet	NOUN
fcis-30157	719	12	model	model	NOUN
fcis-30157	719	13	.	.	PUNCT
fcis-30157	720	1	bioengineering	bioengineere	VERB
fcis-30157	720	2	10.6	10.6	NUM
fcis-30157	720	3	(	(	PUNCT
fcis-30157	720	4	2023	2023	NUM
fcis-30157	720	5	):	):	PUNCT
fcis-30157	720	6	732	732	NUM
fcis-30157	720	7	.	.	PUNCT
fcis-30157	721	1	[	[	X
fcis-30157	721	2	28	28	NUM
fcis-30157	721	3	]	]	X
fcis-30157	721	4	lu	lu	PROPN
fcis-30157	721	5	,	,	PUNCT
fcis-30157	721	6	jianyun	jianyun	PROPN
fcis-30157	721	7	,	,	PUNCT
fcis-30157	721	8	et	et	PROPN
fcis-30157	721	9	al	al	PROPN
fcis-30157	721	10	.	.	PUNCT
fcis-30157	722	1	image	image	NOUN
fcis-30157	722	2	classification	classification	NOUN
fcis-30157	722	3	and	and	CCONJ
fcis-30157	722	4	auxiliary	auxiliary	ADJ
fcis-30157	722	5	diagnosis	diagnosis	NOUN
fcis-30157	722	6	system	system	NOUN
fcis-30157	722	7	for	for	ADP
fcis-30157	722	8	hyperpigmented	hyperpigmente	VERB
fcis-30157	722	9	skin	skin	NOUN
fcis-30157	722	10	diseases	disease	NOUN
fcis-30157	722	11	based	base	VERB
fcis-30157	722	12	on	on	ADP
fcis-30157	722	13	deep	deep	ADJ
fcis-30157	722	14	learning	learning	NOUN
fcis-30157	722	15	.	.	PUNCT
fcis-30157	723	1	heliyon	heliyon	NOUN
fcis-30157	723	2	9.9	9.9	NUM
fcis-30157	723	3	(	(	PUNCT
fcis-30157	723	4	2023	2023	NUM
fcis-30157	723	5	)	)	PUNCT
fcis-30157	723	6	.	.	PUNCT
fcis-30157	724	1	[	[	X
fcis-30157	724	2	29	29	NUM
fcis-30157	724	3	]	]	SYM
fcis-30157	724	4	ye	ye	NOUN
fcis-30157	724	5	,	,	PUNCT
fcis-30157	724	6	shengzhen	shengzhen	ADV
fcis-30157	724	7	,	,	PUNCT
fcis-30157	724	8	and	and	CCONJ
fcis-30157	724	9	mingling	mingle	VERB
fcis-30157	724	10	chen	chen	PROPN
fcis-30157	724	11	.	.	PUNCT
fcis-30157	725	1	the	the	DET
fcis-30157	725	2	emerging	emerge	VERB
fcis-30157	725	3	role	role	NOUN
fcis-30157	725	4	of	of	ADP
fcis-30157	725	5	artificial	artificial	ADJ
fcis-30157	725	6	intelligence	intelligence	NOUN
fcis-30157	725	7	in	in	ADP
fcis-30157	725	8	diagnosis	diagnosis	NOUN
fcis-30157	725	9	and	and	CCONJ
fcis-30157	725	10	clinical	clinical	ADJ
fcis-30157	725	11	analysis	analysis	NOUN
fcis-30157	725	12	of	of	ADP
fcis-30157	725	13	dermatology	dermatology	NOUN
fcis-30157	725	14	.	.	PUNCT
fcis-30157	726	1	dermatologica	dermatologica	PROPN
fcis-30157	726	2	sinica	sinica	PROPN
fcis-30157	726	3	41.3	41.3	NUM
fcis-30157	726	4	(	(	PUNCT
fcis-30157	726	5	2023	2023	NUM
fcis-30157	726	6	):	):	PUNCT
fcis-30157	726	7	145	145	NUM
fcis-30157	726	8	-	-	SYM
fcis-30157	726	9	152	152	NUM
fcis-30157	726	10	.	.	PUNCT
fcis-30157	727	1	126	126	NUM
fcis-30157	728	1	[	[	SYM
fcis-30157	728	2	30	30	NUM
fcis-30157	728	3	]	]	X
fcis-30157	728	4	muhaba	muhaba	PROPN
fcis-30157	728	5	,	,	PUNCT
fcis-30157	728	6	kedir	kedir	PROPN
fcis-30157	728	7	ali	ali	PROPN
fcis-30157	728	8	,	,	PUNCT
fcis-30157	728	9	et	et	PROPN
fcis-30157	728	10	al	al	PROPN
fcis-30157	728	11	.	.	PROPN
fcis-30157	729	1	automatic	automatic	ADJ
fcis-30157	729	2	skin	skin	NOUN
fcis-30157	729	3	disease	disease	NOUN
fcis-30157	729	4	diagnosis	diagnosis	NOUN
fcis-30157	729	5	using	use	VERB
fcis-30157	729	6	deep	deep	ADJ
fcis-30157	729	7	learning	learning	NOUN
fcis-30157	729	8	from	from	ADP
fcis-30157	729	9	clinical	clinical	ADJ
fcis-30157	729	10	image	image	NOUN
fcis-30157	729	11	and	and	CCONJ
fcis-30157	729	12	patient	patient	ADJ
fcis-30157	729	13	information	information	NOUN
fcis-30157	729	14	.	.	PUNCT
fcis-30157	730	1	skin	skin	NOUN
fcis-30157	730	2	health	health	NOUN
fcis-30157	730	3	and	and	CCONJ
fcis-30157	730	4	disease	disease	NOUN
fcis-30157	730	5	2.1	2.1	NUM
fcis-30157	730	6	(	(	PUNCT
fcis-30157	730	7	2022	2022	NUM
fcis-30157	730	8	):	):	PUNCT
fcis-30157	730	9	ski2	ski2	PROPN
fcis-30157	730	10	-	-	PUNCT
fcis-30157	730	11	81	81	NUM
fcis-30157	730	12	.	.	PUNCT
fcis-30157	731	1	[	[	X
fcis-30157	731	2	31	31	NUM
fcis-30157	731	3	]	]	PUNCT
fcis-30157	731	4	parashar	parashar	NOUN
fcis-30157	731	5	,	,	PUNCT
fcis-30157	731	6	tarun	tarun	PROPN
fcis-30157	731	7	,	,	PUNCT
fcis-30157	731	8	et	et	PROPN
fcis-30157	731	9	al	al	PROPN
fcis-30157	731	10	.	.	PROPN
fcis-30157	731	11	skin	skin	NOUN
fcis-30157	731	12	disease	disease	NOUN
fcis-30157	731	13	detection	detection	NOUN
fcis-30157	731	14	using	use	VERB
fcis-30157	731	15	deep	deep	ADJ
fcis-30157	731	16	learning	learning	NOUN
fcis-30157	731	17	.	.	PUNCT
fcis-30157	732	1	2022	2022	NUM
fcis-30157	732	2	11th	11th	ADJ
fcis-30157	732	3	international	international	ADJ
fcis-30157	732	4	conference	conference	NOUN
fcis-30157	732	5	on	on	ADP
fcis-30157	732	6	system	system	NOUN
fcis-30157	732	7	modeling	modeling	NOUN
fcis-30157	732	8	&	&	CCONJ
fcis-30157	732	9	advancement	advancement	NOUN
fcis-30157	732	10	in	in	ADP
fcis-30157	732	11	research	research	NOUN
fcis-30157	732	12	trends	trend	NOUN
fcis-30157	732	13	(	(	PUNCT
fcis-30157	732	14	smart	smart	ADJ
fcis-30157	732	15	)	)	PUNCT
fcis-30157	732	16	.	.	PUNCT
fcis-30157	733	1	ieee	ieee	PROPN
fcis-30157	733	2	,	,	PUNCT
fcis-30157	733	3	2022	2022	NUM
fcis-30157	733	4	.	.	PUNCT
fcis-30157	734	1	[	[	X
fcis-30157	734	2	32	32	NUM
fcis-30157	734	3	]	]	SYM
fcis-30157	734	4	ghaffar	ghaffar	PROPN
fcis-30157	734	5	nia	nia	PROPN
fcis-30157	734	6	,	,	PUNCT
fcis-30157	734	7	nafiseh	nafiseh	PROPN
fcis-30157	734	8	,	,	PUNCT
fcis-30157	734	9	erkan	erkan	PROPN
fcis-30157	734	10	kaplanoglu	kaplanoglu	PROPN
fcis-30157	734	11	,	,	PUNCT
fcis-30157	734	12	and	and	CCONJ
fcis-30157	734	13	ahad	ahad	PROPN
fcis-30157	734	14	nasab	nasab	PROPN
fcis-30157	734	15	.	.	PUNCT
fcis-30157	735	1	evaluation	evaluation	NOUN
fcis-30157	735	2	of	of	ADP
fcis-30157	735	3	artificial	artificial	ADJ
fcis-30157	735	4	intelligence	intelligence	NOUN
fcis-30157	735	5	techniques	technique	NOUN
fcis-30157	735	6	in	in	ADP
fcis-30157	735	7	disease	disease	NOUN
fcis-30157	735	8	diagnosis	diagnosis	NOUN
fcis-30157	735	9	and	and	CCONJ
fcis-30157	735	10	prediction	prediction	NOUN
fcis-30157	735	11	.	.	PUNCT
fcis-30157	736	1	discover	discover	VERB
fcis-30157	736	2	artificial	artificial	ADJ
fcis-30157	736	3	intelligence	intelligence	NOUN
fcis-30157	736	4	3.1	3.1	NUM
fcis-30157	736	5	(	(	PUNCT
fcis-30157	736	6	2023	2023	NUM
fcis-30157	736	7	):	):	PUNCT
fcis-30157	736	8	5	5	NUM
fcis-30157	736	9	.	.	PUNCT
fcis-30157	737	1	[	[	X
fcis-30157	737	2	33	33	NUM
fcis-30157	737	3	]	]	SYM
fcis-30157	737	4	li	li	PROPN
fcis-30157	737	5	,	,	PUNCT
fcis-30157	737	6	huanyu	huanyu	PROPN
fcis-30157	737	7	,	,	PUNCT
fcis-30157	737	8	et	et	PROPN
fcis-30157	737	9	al	al	PROPN
fcis-30157	737	10	.	.	PUNCT
fcis-30157	738	1	deep	deep	ADJ
fcis-30157	738	2	skin	skin	NOUN
fcis-30157	738	3	diseases	disease	VERB
fcis-30157	738	4	diagnostic	diagnostic	ADJ
fcis-30157	738	5	system	system	NOUN
fcis-30157	738	6	with	with	ADP
fcis-30157	738	7	dual	dual	ADJ
fcis-30157	738	8	-	-	PUNCT
fcis-30157	738	9	channel	channel	NOUN
fcis-30157	738	10	image	image	NOUN
fcis-30157	738	11	and	and	CCONJ
fcis-30157	738	12	extracted	extract	VERB
fcis-30157	738	13	text	text	NOUN
fcis-30157	738	14	.	.	PUNCT
fcis-30157	739	1	frontiers	frontier	NOUN
fcis-30157	739	2	in	in	ADP
fcis-30157	739	3	artificial	artificial	ADJ
fcis-30157	739	4	intelligence	intelligence	NOUN
fcis-30157	739	5	6	6	NUM
fcis-30157	739	6	(	(	PUNCT
fcis-30157	739	7	2023	2023	NUM
fcis-30157	739	8	):	):	PUNCT
fcis-30157	739	9	1213620	1213620	NUM
fcis-30157	739	10	.	.	PUNCT
fcis-30157	740	1	[	[	X
fcis-30157	740	2	34	34	NUM
fcis-30157	740	3	]	]	SYM
fcis-30157	740	4	li	li	PROPN
fcis-30157	740	5	,	,	PUNCT
fcis-30157	740	6	xiaoqiong	xiaoqiong	PROPN
fcis-30157	740	7	,	,	PUNCT
fcis-30157	740	8	et	et	PROPN
fcis-30157	740	9	al	al	PROPN
fcis-30157	740	10	.	.	PROPN
fcis-30157	740	11	electret	electret	PROPN
fcis-30157	740	12	-	-	PUNCT
fcis-30157	740	13	based	base	VERB
fcis-30157	740	14	flexible	flexible	ADJ
fcis-30157	740	15	pressure	pressure	NOUN
fcis-30157	740	16	sensor	sensor	NOUN
fcis-30157	740	17	for	for	ADP
fcis-30157	740	18	respiratory	respiratory	ADJ
fcis-30157	740	19	diseases	disease	NOUN
fcis-30157	740	20	auxiliary	auxiliary	ADJ
fcis-30157	740	21	diagnosis	diagnosis	NOUN
fcis-30157	740	22	system	system	NOUN
fcis-30157	740	23	using	use	VERB
fcis-30157	740	24	machine	machine	NOUN
fcis-30157	740	25	learning	learning	NOUN
fcis-30157	740	26	technique	technique	NOUN
fcis-30157	740	27	.	.	PUNCT
fcis-30157	741	1	nano	nano	NOUN
fcis-30157	741	2	energy	energy	NOUN
fcis-30157	741	3	114	114	NUM
fcis-30157	741	4	(	(	PUNCT
fcis-30157	741	5	2023	2023	NUM
fcis-30157	741	6	):	):	PUNCT
fcis-30157	741	7	108652	108652	NUM
fcis-30157	741	8	.	.	PUNCT
fcis-30157	742	1	[	[	X
fcis-30157	742	2	35	35	NUM
fcis-30157	742	3	]	]	X
fcis-30157	742	4	zhou	zhou	X
fcis-30157	742	5	,	,	PUNCT
fcis-30157	742	6	yan	yan	PROPN
fcis-30157	742	7	-	-	PUNCT
fcis-30157	742	8	jie	jie	PROPN
fcis-30157	742	9	,	,	PUNCT
fcis-30157	742	10	et	et	PROPN
fcis-30157	742	11	al	al	PROPN
fcis-30157	742	12	.	.	PUNCT
fcis-30157	743	1	a	a	DET
fcis-30157	743	2	novel	novel	ADJ
fcis-30157	743	3	multi	multi	ADJ
fcis-30157	743	4	-	-	ADJ
fcis-30157	743	5	task	task	ADJ
fcis-30157	743	6	model	model	NOUN
fcis-30157	743	7	imitating	imitating	NOUN
fcis-30157	743	8	dermatologists	dermatologist	NOUN
fcis-30157	743	9	for	for	ADP
fcis-30157	743	10	accurate	accurate	ADJ
fcis-30157	743	11	differential	differential	ADJ
fcis-30157	743	12	diagnosis	diagnosis	NOUN
fcis-30157	743	13	of	of	ADP
fcis-30157	743	14	skin	skin	NOUN
fcis-30157	743	15	diseases	disease	NOUN
fcis-30157	743	16	in	in	ADP
fcis-30157	743	17	clinical	clinical	ADJ
fcis-30157	743	18	images	image	NOUN
fcis-30157	743	19	.	.	PUNCT
fcis-30157	744	1	international	international	ADJ
fcis-30157	744	2	conference	conference	NOUN
fcis-30157	744	3	on	on	ADP
fcis-30157	744	4	medical	medical	ADJ
fcis-30157	744	5	image	image	NOUN
fcis-30157	744	6	computing	computing	NOUN
fcis-30157	744	7	and	and	CCONJ
fcis-30157	744	8	computer	computer	NOUN
fcis-30157	744	9	-	-	PUNCT
fcis-30157	744	10	assisted	assist	VERB
fcis-30157	744	11	intervention	intervention	NOUN
fcis-30157	744	12	.	.	PUNCT
fcis-30157	745	1	cham	cham	PROPN
fcis-30157	745	2	:	:	PUNCT
fcis-30157	745	3	springer	springer	NOUN
fcis-30157	745	4	nature	nature	PROPN
fcis-30157	745	5	switzerland	switzerland	PROPN
fcis-30157	745	6	,	,	PUNCT
fcis-30157	745	7	2023	2023	NUM
fcis-30157	745	8	.	.	PUNCT
fcis-30157	746	1	[	[	X
fcis-30157	746	2	36	36	NUM
fcis-30157	746	3	]	]	PUNCT
fcis-30157	746	4	vayadande	vayadande	NOUN
fcis-30157	746	5	,	,	PUNCT
fcis-30157	746	6	kuldeep	kuldeep	PROPN
fcis-30157	746	7	.	.	PUNCT
fcis-30157	747	1	innovative	innovative	ADJ
fcis-30157	747	2	approaches	approach	NOUN
fcis-30157	747	3	for	for	ADP
fcis-30157	747	4	skin	skin	NOUN
fcis-30157	747	5	disease	disease	NOUN
fcis-30157	747	6	identification	identification	NOUN
fcis-30157	747	7	in	in	ADP
fcis-30157	747	8	machine	machine	NOUN
fcis-30157	747	9	learning	learning	NOUN
fcis-30157	747	10	:	:	PUNCT
fcis-30157	747	11	a	a	DET
fcis-30157	747	12	comprehensive	comprehensive	ADJ
fcis-30157	747	13	study	study	NOUN
fcis-30157	747	14	.	.	PUNCT
fcis-30157	748	1	oral	oral	ADJ
fcis-30157	748	2	oncology	oncology	NOUN
fcis-30157	748	3	reports	report	NOUN
fcis-30157	748	4	(	(	PUNCT
fcis-30157	748	5	2024	2024	NUM
fcis-30157	748	6	):	):	PUNCT
fcis-30157	748	7	100365	100365	NUM
fcis-30157	748	8	.	.	PUNCT
fcis-30157	749	1	[	[	X
fcis-30157	749	2	37	37	NUM
fcis-30157	749	3	]	]	X
fcis-30157	749	4	aruna	aruna	PROPN
fcis-30157	749	5	r	r	PROPN
fcis-30157	749	6	,	,	PUNCT
fcis-30157	749	7	dr	dr	PROPN
fcis-30157	749	8	,	,	PUNCT
fcis-30157	749	9	et	et	PROPN
fcis-30157	749	10	al	al	PROPN
fcis-30157	749	11	.	.	PUNCT
fcis-30157	750	1	[	[	X
fcis-30157	750	2	retracted	retract	VERB
fcis-30157	750	3	]	]	PUNCT
fcis-30157	750	4	an	an	DET
fcis-30157	750	5	enhancement	enhancement	NOUN
fcis-30157	750	6	on	on	ADP
fcis-30157	750	7	convolutional	convolutional	ADJ
fcis-30157	750	8	artificial	artificial	ADJ
fcis-30157	750	9	intelligent	intelligent	ADJ
fcis-30157	750	10	based	base	VERB
fcis-30157	750	11	diagnosis	diagnosis	NOUN
fcis-30157	750	12	for	for	ADP
fcis-30157	750	13	skin	skin	NOUN
fcis-30157	750	14	disease	disease	NOUN
fcis-30157	750	15	using	use	VERB
fcis-30157	750	16	nanotechnology	nanotechnology	NOUN
fcis-30157	750	17	sensors	sensor	NOUN
fcis-30157	750	18	.	.	PUNCT
fcis-30157	751	1	computational	computational	ADJ
fcis-30157	751	2	intelligence	intelligence	NOUN
fcis-30157	751	3	and	and	CCONJ
fcis-30157	751	4	neuroscience	neuroscience	NOUN
fcis-30157	751	5	2022.1	2022.1	NUM
fcis-30157	751	6	(	(	PUNCT
fcis-30157	751	7	2022	2022	NUM
fcis-30157	751	8	):	):	PUNCT
fcis-30157	751	9	9539503	9539503	NUM
fcis-30157	751	10	.	.	PUNCT
fcis-30157	752	1	[	[	X
fcis-30157	752	2	38	38	NUM
fcis-30157	752	3	]	]	X
fcis-30157	752	4	zheng	zheng	PROPN
fcis-30157	752	5	,	,	PUNCT
fcis-30157	752	6	guiyue	guiyue	PROPN
fcis-30157	752	7	,	,	PUNCT
fcis-30157	752	8	et	et	PROPN
fcis-30157	752	9	al	al	PROPN
fcis-30157	752	10	.	.	PROPN
fcis-30157	752	11	research	research	NOUN
fcis-30157	752	12	on	on	ADP
fcis-30157	752	13	skin	skin	NOUN
fcis-30157	752	14	disease	disease	NOUN
fcis-30157	752	15	health	health	NOUN
fcis-30157	752	16	detection	detection	NOUN
fcis-30157	752	17	of	of	ADP
fcis-30157	752	18	college	college	NOUN
fcis-30157	752	19	students	student	NOUN
fcis-30157	752	20	based	base	VERB
fcis-30157	752	21	on	on	ADP
fcis-30157	752	22	deep	deep	ADJ
fcis-30157	752	23	learning	learning	NOUN
fcis-30157	752	24	.	.	PUNCT
fcis-30157	753	1	journal	journal	PROPN
fcis-30157	753	2	of	of	ADP
fcis-30157	753	3	physics	physics	PROPN
fcis-30157	753	4	:	:	PUNCT
fcis-30157	753	5	conference	conference	NOUN
fcis-30157	753	6	series	series	NOUN
fcis-30157	753	7	.	.	PUNCT
fcis-30157	754	1	vol	vol	NOUN
fcis-30157	754	2	.	.	PROPN
fcis-30157	754	3	2289	2289	NUM
fcis-30157	754	4	.	.	PUNCT
fcis-30157	755	1	no	no	INTJ
fcis-30157	755	2	.	.	NOUN
fcis-30157	756	1	1	1	X
fcis-30157	756	2	.	.	X
fcis-30157	756	3	iop	iop	NOUN
fcis-30157	756	4	publishing	publishing	NOUN
fcis-30157	756	5	,	,	PUNCT
fcis-30157	756	6	2022	2022	NUM
fcis-30157	756	7	.	.	PUNCT
fcis-30157	757	1	[	[	X
fcis-30157	757	2	39	39	NUM
fcis-30157	757	3	]	]	PUNCT
fcis-30157	757	4	maduranga	maduranga	NOUN
fcis-30157	757	5	,	,	PUNCT
fcis-30157	757	6	m.	m.	PROPN
fcis-30157	757	7	w.	w.	PROPN
fcis-30157	757	8	p.	p.	PROPN
fcis-30157	757	9	,	,	PUNCT
fcis-30157	757	10	and	and	CCONJ
fcis-30157	757	11	dilshan	dilshan	PROPN
fcis-30157	757	12	nandasena	nandasena	PROPN
fcis-30157	757	13	.	.	PUNCT
fcis-30157	758	1	mobile	mobile	NOUN
fcis-30157	758	2	-	-	PUNCT
fcis-30157	758	3	based	base	VERB
fcis-30157	758	4	skin	skin	NOUN
fcis-30157	758	5	disease	disease	NOUN
fcis-30157	758	6	diagnosis	diagnosis	NOUN
fcis-30157	758	7	system	system	NOUN
fcis-30157	758	8	using	use	VERB
fcis-30157	758	9	convolutional	convolutional	ADJ
fcis-30157	758	10	neural	neural	ADJ
fcis-30157	758	11	networks	network	NOUN
fcis-30157	758	12	(	(	PUNCT
fcis-30157	758	13	cnn	cnn	PROPN
fcis-30157	758	14	)	)	PUNCT
fcis-30157	758	15	.	.	PUNCT
fcis-30157	759	1	international	international	ADJ
fcis-30157	759	2	journal	journal	PROPN
fcis-30157	759	3	of	of	ADP
fcis-30157	759	4	image	image	NOUN
fcis-30157	759	5	,	,	PUNCT
fcis-30157	759	6	graphics	graphic	NOUN
fcis-30157	759	7	and	and	CCONJ
fcis-30157	759	8	signal	signal	VERB
fcis-30157	759	9	processing	process	VERB
fcis-30157	759	10	12.3	12.3	NUM
fcis-30157	759	11	(	(	PUNCT
fcis-30157	759	12	2022	2022	NUM
fcis-30157	759	13	):	):	PUNCT
fcis-30157	759	14	47	47	NUM
fcis-30157	759	15	.	.	PUNCT
fcis-30157	760	1	[	[	X
fcis-30157	760	2	40	40	NUM
fcis-30157	760	3	]	]	X
fcis-30157	760	4	ke	ke	PROPN
fcis-30157	760	5	,	,	PUNCT
fcis-30157	760	6	c.	c.	PROPN
fcis-30157	760	7	h.	h.	PROPN
fcis-30157	760	8	a.	a.	PROPN
fcis-30157	760	9	i.	i.	PROPN
fcis-30157	760	10	,	,	PUNCT
fcis-30157	760	11	et	et	PROPN
fcis-30157	760	12	al	al	PROPN
fcis-30157	760	13	.	.	PROPN
fcis-30157	760	14	updated	update	VERB
fcis-30157	760	15	role	role	NOUN
fcis-30157	760	16	of	of	ADP
fcis-30157	760	17	high	high	ADJ
fcis-30157	760	18	-	-	PUNCT
fcis-30157	760	19	frequency	frequency	NOUN
fcis-30157	760	20	ultrasound	ultrasound	NOUN
fcis-30157	760	21	in	in	ADP
fcis-30157	760	22	assessing	assess	VERB
fcis-30157	760	23	dermatological	dermatological	ADJ
fcis-30157	760	24	manifestations	manifestation	NOUN
fcis-30157	760	25	in	in	ADP
fcis-30157	760	26	autoimmune	autoimmune	ADJ
fcis-30157	760	27	skin	skin	NOUN
fcis-30157	760	28	diseases	disease	NOUN
fcis-30157	760	29	.	.	PUNCT
fcis-30157	761	1	acta	acta	PROPN
fcis-30157	761	2	dermato	dermato	PROPN
fcis-30157	761	3	-	-	PUNCT
fcis-30157	761	4	venereologica	venereologica	PROPN
fcis-30157	761	5	102	102	NUM
fcis-30157	761	6	(	(	PUNCT
fcis-30157	761	7	2022	2022	NUM
fcis-30157	761	8	):	):	PUNCT
fcis-30157	761	9	1969	1969	NUM
fcis-30157	761	10	.	.	PUNCT