id	sid	tid	token	lemma	pos
cana-5754	1	1	communications	communication	NOUN
cana-5754	1	2	on	on	ADP
cana-5754	1	3	applied	apply	VERB
cana-5754	1	4	nonlinear	nonlinear	ADJ
cana-5754	1	5	analysis	analysis	NOUN
cana-5754	1	6	issn	issn	NOUN
cana-5754	1	7	:	:	PUNCT
cana-5754	1	8	1074	1074	NUM
cana-5754	1	9	-	-	PUNCT
cana-5754	1	10	133x	133x	NUM
cana-5754	1	11	vol	vol	NOUN
cana-5754	1	12	32	32	NUM
cana-5754	1	13	no	no	NOUN
cana-5754	1	14	.	.	PUNCT
cana-5754	2	1	3s	3s	NUM
cana-5754	2	2	(	(	PUNCT
cana-5754	2	3	2025	2025	NUM
cana-5754	2	4	)	)	PUNCT
cana-5754	2	5	780	780	NUM
cana-5754	2	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-5754	2	7	computer	computer	NOUN
cana-5754	2	8	vision	vision	NOUN
cana-5754	2	9	-	-	PUNCT
cana-5754	2	10	based	base	VERB
cana-5754	2	11	deep	deep	ADJ
cana-5754	2	12	learning	learning	NOUN
cana-5754	2	13	techniques	technique	NOUN
cana-5754	2	14	for	for	ADP
cana-5754	2	15	diabetic	diabetic	ADJ
cana-5754	2	16	retinopathy	retinopathy	ADJ
cana-5754	2	17	assessment	assessment	NOUN
cana-5754	2	18	from	from	ADP
cana-5754	2	19	fundus	fundus	NOUN
cana-5754	2	20	images	image	NOUN
cana-5754	2	21	munazzah	munazzah	NOUN
cana-5754	2	22	maryam1	maryam1	PROPN
cana-5754	2	23	dr	dr	PROPN
cana-5754	2	24	.	.	PROPN
cana-5754	2	25	ruhiat	ruhiat	PROPN
cana-5754	2	26	sultana2	sultana2	PROPN
cana-5754	3	1	dr	dr	PROPN
cana-5754	3	2	.	.	PROPN
cana-5754	3	3	t.	t.	PROPN
cana-5754	3	4	k.	k.	PROPN
cana-5754	3	5	shaik	shaik	PROPN
cana-5754	3	6	shavali3	shavali3	PROPN
cana-5754	3	7	1	1	NUM
cana-5754	3	8	research	research	NOUN
cana-5754	3	9	scholar	scholar	NOUN
cana-5754	3	10	,	,	PUNCT
cana-5754	3	11	dept	dept	NOUN
cana-5754	3	12	.	.	PROPN
cana-5754	3	13	of	of	ADP
cana-5754	3	14	computer	computer	NOUN
cana-5754	3	15	science	science	NOUN
cana-5754	3	16	and	and	CCONJ
cana-5754	3	17	engineering	engineering	NOUN
cana-5754	3	18	,	,	PUNCT
cana-5754	3	19	lords	lords	PROPN
cana-5754	3	20	institute	institute	PROPN
cana-5754	3	21	of	of	ADP
cana-5754	3	22	engineering	engineering	NOUN
cana-5754	3	23	and	and	CCONJ
cana-5754	3	24	technology	technology	NOUN
cana-5754	3	25	,	,	PUNCT
cana-5754	3	26	hyderabad	hyderabad	PROPN
cana-5754	3	27	,	,	PUNCT
cana-5754	3	28	telangana	telangana	PROPN
cana-5754	3	29	2associate	2associate	NUM
cana-5754	3	30	professor	professor	NOUN
cana-5754	3	31	,	,	PUNCT
cana-5754	3	32	dept	dept	NOUN
cana-5754	3	33	.	.	PROPN
cana-5754	3	34	of	of	ADP
cana-5754	3	35	computer	computer	NOUN
cana-5754	3	36	science	science	NOUN
cana-5754	3	37	and	and	CCONJ
cana-5754	3	38	engineering	engineering	NOUN
cana-5754	3	39	,	,	PUNCT
cana-5754	3	40	lords	lords	PROPN
cana-5754	3	41	institute	institute	PROPN
cana-5754	3	42	of	of	ADP
cana-5754	3	43	engineering	engineering	NOUN
cana-5754	3	44	and	and	CCONJ
cana-5754	3	45	technology	technology	NOUN
cana-5754	3	46	,	,	PUNCT
cana-5754	3	47	hyderabad	hyderabad	PROPN
cana-5754	3	48	,	,	PUNCT
cana-5754	3	49	telangana	telangana	PROPN
cana-5754	3	50	3professor	3professor	NUM
cana-5754	3	51	,	,	PUNCT
cana-5754	3	52	dept	dept	NOUN
cana-5754	3	53	.	.	PROPN
cana-5754	3	54	of	of	ADP
cana-5754	3	55	computer	computer	NOUN
cana-5754	3	56	science	science	NOUN
cana-5754	3	57	and	and	CCONJ
cana-5754	3	58	engineering	engineering	NOUN
cana-5754	3	59	,	,	PUNCT
cana-5754	3	60	lords	lords	PROPN
cana-5754	3	61	institute	institute	PROPN
cana-5754	3	62	of	of	ADP
cana-5754	3	63	engineering	engineering	NOUN
cana-5754	3	64	and	and	CCONJ
cana-5754	3	65	technology	technology	NOUN
cana-5754	3	66	,	,	PUNCT
cana-5754	3	67	hyderabad	hyderabad	PROPN
cana-5754	3	68	,	,	PUNCT
cana-5754	3	69	telangana	telangana	PROPN
cana-5754	3	70	article	article	PROPN
cana-5754	3	71	history	history	NOUN
cana-5754	3	72	:	:	PUNCT
cana-5754	3	73	received	receive	VERB
cana-5754	3	74	:	:	PUNCT
cana-5754	3	75	12	12	NUM
cana-5754	3	76	-	-	SYM
cana-5754	3	77	01	01	NUM
cana-5754	3	78	-	-	PUNCT
cana-5754	3	79	2025	2025	NUM
cana-5754	3	80	revised	revise	VERB
cana-5754	3	81	:	:	PUNCT
cana-5754	3	82	15	15	NUM
cana-5754	3	83	-	-	NUM
cana-5754	3	84	02	02	NUM
cana-5754	3	85	-	-	PUNCT
cana-5754	3	86	2025	2025	NUM
cana-5754	3	87	accepted	accept	VERB
cana-5754	3	88	:	:	PUNCT
cana-5754	3	89	01	01	NUM
cana-5754	3	90	-	-	SYM
cana-5754	3	91	03	03	NUM
cana-5754	3	92	-	-	PUNCT
cana-5754	3	93	2025	2025	NUM
cana-5754	3	94	abstract	abstract	ADJ
cana-5754	3	95	long	long	ADJ
cana-5754	3	96	-	-	PUNCT
cana-5754	3	97	term	term	NOUN
cana-5754	3	98	diabetes	diabetes	NOUN
cana-5754	3	99	can	can	AUX
cana-5754	3	100	damage	damage	VERB
cana-5754	3	101	the	the	DET
cana-5754	3	102	blood	blood	NOUN
cana-5754	3	103	vessels	vessel	NOUN
cana-5754	3	104	in	in	ADP
cana-5754	3	105	the	the	DET
cana-5754	3	106	retina	retina	NOUN
cana-5754	3	107	,	,	PUNCT
cana-5754	3	108	resulting	result	VERB
cana-5754	3	109	in	in	ADP
cana-5754	3	110	diabetic	diabetic	ADJ
cana-5754	3	111	retinopathy	retinopathy	NOUN
cana-5754	3	112	(	(	PUNCT
cana-5754	3	113	dr	dr	PROPN
cana-5754	3	114	)	)	PUNCT
cana-5754	3	115	,	,	PUNCT
cana-5754	3	116	a	a	DET
cana-5754	3	117	form	form	NOUN
cana-5754	3	118	of	of	ADP
cana-5754	3	119	vision	vision	NOUN
cana-5754	3	120	impairment	impairment	NOUN
cana-5754	3	121	.	.	PUNCT
cana-5754	4	1	it	it	PRON
cana-5754	4	2	affects	affect	VERB
cana-5754	4	3	around	around	ADV
cana-5754	4	4	191	191	NUM
cana-5754	4	5	million	million	NUM
cana-5754	4	6	people	people	NOUN
cana-5754	4	7	worldwide	worldwide	ADV
cana-5754	4	8	and	and	CCONJ
cana-5754	4	9	is	be	AUX
cana-5754	4	10	the	the	DET
cana-5754	4	11	most	most	ADV
cana-5754	4	12	prevalent	prevalent	ADJ
cana-5754	4	13	cause	cause	NOUN
cana-5754	4	14	of	of	ADP
cana-5754	4	15	blindness	blindness	NOUN
cana-5754	4	16	.	.	PUNCT
cana-5754	5	1	although	although	SCONJ
cana-5754	5	2	earlier	early	ADJ
cana-5754	5	3	studies	study	NOUN
cana-5754	5	4	have	have	AUX
cana-5754	5	5	focused	focus	VERB
cana-5754	5	6	on	on	ADP
cana-5754	5	7	dr	dr	PROPN
cana-5754	5	8	categorization	categorization	NOUN
cana-5754	5	9	by	by	ADP
cana-5754	5	10	retinalfundus	retinalfundus	PROPN
cana-5754	5	11	pictures	picture	NOUN
cana-5754	5	12	,	,	PUNCT
cana-5754	5	13	current	current	ADJ
cana-5754	5	14	techniques	technique	NOUN
cana-5754	5	15	typically	typically	ADV
cana-5754	5	16	concentrate	concentrate	VERB
cana-5754	5	17	on	on	ADP
cana-5754	5	18	identifying	identify	VERB
cana-5754	5	19	certain	certain	ADJ
cana-5754	5	20	tumors	tumor	NOUN
cana-5754	5	21	without	without	ADP
cana-5754	5	22	providing	provide	VERB
cana-5754	5	23	a	a	DET
cana-5754	5	24	thorough	thorough	ADJ
cana-5754	5	25	basis	basis	NOUN
cana-5754	5	26	for	for	ADP
cana-5754	5	27	concurrently	concurrently	ADV
cana-5754	5	28	detecting	detect	VERB
cana-5754	5	29	each	each	DET
cana-5754	5	30	lesion	lesion	NOUN
cana-5754	5	31	.	.	PUNCT
cana-5754	6	1	prior	prior	ADJ
cana-5754	6	2	research	research	NOUN
cana-5754	6	3	focused	focus	VERB
cana-5754	6	4	on	on	ADP
cana-5754	6	5	earlystage	earlystage	NOUN
cana-5754	6	6	features	feature	NOUN
cana-5754	6	7	such	such	ADJ
cana-5754	6	8	as	as	ADP
cana-5754	6	9	blood	blood	NOUN
cana-5754	6	10	vessels	vessel	NOUN
cana-5754	6	11	,	,	PUNCT
cana-5754	6	12	hemorrhages	hemorrhage	NOUN
cana-5754	6	13	,	,	PUNCT
cana-5754	6	14	aneurysms	aneurysm	NOUN
cana-5754	6	15	,	,	PUNCT
cana-5754	6	16	and	and	CCONJ
cana-5754	6	17	exudates	exudate	VERB
cana-5754	6	18	highlight	highlight	VERB
cana-5754	6	19	lesions	lesion	NOUN
cana-5754	6	20	at	at	ADP
cana-5754	6	21	the	the	DET
cana-5754	6	22	severe	severe	ADJ
cana-5754	6	23	stage.such	stage.such	NOUN
cana-5754	6	24	as	as	ADP
cana-5754	6	25	exceedingly	exceedingly	ADV
cana-5754	6	26	severe	severe	ADJ
cana-5754	6	27	intraretinal	intraretinal	ADJ
cana-5754	6	28	microvascular	microvascular	ADJ
cana-5754	6	29	abnormalities	abnormality	NOUN
cana-5754	6	30	(	(	PUNCT
cana-5754	6	31	irma	irma	PROPN
cana-5754	6	32	)	)	PUNCT
cana-5754	6	33	,	,	PUNCT
cana-5754	6	34	venous	venous	ADJ
cana-5754	6	35	beading	beading	NOUN
cana-5754	6	36	,	,	PUNCT
cana-5754	6	37	and	and	CCONJ
cana-5754	6	38	cotton	cotton	NOUN
cana-5754	6	39	wool	wool	NOUN
cana-5754	6	40	patches	patch	NOUN
cana-5754	6	41	,	,	PUNCT
cana-5754	6	42	retinal	retinal	ADJ
cana-5754	6	43	pigment	pigment	NOUN
cana-5754	6	44	,	,	PUNCT
cana-5754	6	45	capillary	capillary	ADJ
cana-5754	6	46	degeneration	degeneration	NOUN
cana-5754	6	47	,	,	PUNCT
cana-5754	6	48	diffuse	diffuse	ADJ
cana-5754	6	49	intraretinal	intraretinal	ADJ
cana-5754	6	50	hemorrhages	hemorrhage	NOUN
cana-5754	6	51	,	,	PUNCT
cana-5754	6	52	and	and	CCONJ
cana-5754	6	53	extremely	extremely	ADV
cana-5754	6	54	active	active	ADJ
cana-5754	6	55	microglia	microglia	ADJ
cana-5754	6	56	impairment	impairment	NOUN
cana-5754	6	57	of	of	ADP
cana-5754	6	58	rpe	rpe	PROPN
cana-5754	6	59	.	.	PUNCT
cana-5754	7	1	this	this	DET
cana-5754	7	2	work	work	NOUN
cana-5754	7	3	suggests	suggest	VERB
cana-5754	7	4	that	that	SCONJ
cana-5754	7	5	deep	deep	ADJ
cana-5754	7	6	learning	learning	NOUN
cana-5754	7	7	can	can	AUX
cana-5754	7	8	be	be	AUX
cana-5754	7	9	used	use	VERB
cana-5754	7	10	to	to	PART
cana-5754	7	11	classify	classify	VERB
cana-5754	7	12	dr	dr	PROPN
cana-5754	7	13	fundus	fundus	NOUN
cana-5754	7	14	images.different	images.different	ADJ
cana-5754	7	15	degrees	degree	NOUN
cana-5754	7	16	of	of	ADP
cana-5754	7	17	severity	severity	NOUN
cana-5754	7	18	,	,	PUNCT
cana-5754	7	19	using	use	VERB
cana-5754	7	20	adaptive	adaptive	ADJ
cana-5754	7	21	particle	particle	NOUN
cana-5754	7	22	swarm	swarm	NOUN
cana-5754	7	23	optimizer	optimizer	NOUN
cana-5754	7	24	-	-	PUNCT
cana-5754	7	25	based	base	VERB
cana-5754	7	26	googlenet	googlenet	NOUN
cana-5754	7	27	and	and	CCONJ
cana-5754	7	28	resnet	resnet	NOUN
cana-5754	7	29	models	model	NOUN
cana-5754	7	30	(	(	PUNCT
cana-5754	7	31	apso	apso	NOUN
cana-5754	7	32	)	)	PUNCT
cana-5754	7	33	to	to	PART
cana-5754	7	34	improve	improve	VERB
cana-5754	7	35	feature	feature	NOUN
cana-5754	7	36	extraction	extraction	NOUN
cana-5754	7	37	.	.	PUNCT
cana-5754	8	1	after	after	ADP
cana-5754	8	2	then	then	ADV
cana-5754	8	3	,	,	PUNCT
cana-5754	8	4	the	the	DET
cana-5754	8	5	hybrid	hybrid	ADJ
cana-5754	8	6	model	model	NOUN
cana-5754	8	7	's	's	PART
cana-5754	8	8	features	feature	NOUN
cana-5754	8	9	are	be	AUX
cana-5754	8	10	put	put	VERB
cana-5754	8	11	to	to	ADP
cana-5754	8	12	use.employing	use.employe	VERB
cana-5754	8	13	a	a	DET
cana-5754	8	14	range	range	NOUN
cana-5754	8	15	of	of	ADP
cana-5754	8	16	machine	machine	NOUN
cana-5754	8	17	learning	learning	NOUN
cana-5754	8	18	methods	method	NOUN
cana-5754	8	19	,	,	PUNCT
cana-5754	8	20	such	such	ADJ
cana-5754	8	21	as	as	ADP
cana-5754	8	22	decision	decision	NOUN
cana-5754	8	23	trees	tree	NOUN
cana-5754	8	24	,	,	PUNCT
cana-5754	8	25	random	random	ADJ
cana-5754	8	26	forests	forest	NOUN
cana-5754	8	27	,	,	PUNCT
cana-5754	8	28	support	support	VERB
cana-5754	8	29	vector	vector	NOUN
cana-5754	8	30	machines	machine	NOUN
cana-5754	8	31	,	,	PUNCT
cana-5754	8	32	and	and	CCONJ
cana-5754	8	33	models	model	NOUN
cana-5754	8	34	of	of	ADP
cana-5754	8	35	linear	linear	ADJ
cana-5754	8	36	regression	regression	NOUN
cana-5754	8	37	.	.	PUNCT
cana-5754	9	1	according	accord	VERB
cana-5754	9	2	to	to	ADP
cana-5754	9	3	experimental	experimental	ADJ
cana-5754	9	4	results	result	NOUN
cana-5754	9	5	,	,	PUNCT
cana-5754	9	6	the	the	DET
cana-5754	9	7	suggested	suggest	VERB
cana-5754	9	8	hybrid	hybrid	ADJ
cana-5754	9	9	framework	framework	NOUN
cana-5754	9	10	performed	perform	VERB
cana-5754	9	11	better	well	ADV
cana-5754	9	12	than	than	ADP
cana-5754	9	13	sophisticated	sophisticated	ADJ
cana-5754	9	14	methods	method	NOUN
cana-5754	9	15	that	that	PRON
cana-5754	9	16	achieved	achieve	VERB
cana-5754	9	17	an	an	DET
cana-5754	9	18	astounding	astounding	ADJ
cana-5754	9	19	94	94	NUM
cana-5754	9	20	%	%	NOUN
cana-5754	9	21	accuracy	accuracy	NOUN
cana-5754	9	22	rate	rate	NOUN
cana-5754	9	23	on	on	ADP
cana-5754	9	24	the	the	DET
cana-5754	9	25	benchmark	benchmark	NOUN
cana-5754	9	26	dataset	dataset	NOUN
cana-5754	9	27	.	.	PUNCT
cana-5754	10	1	this	this	DET
cana-5754	10	2	approach	approach	NOUN
cana-5754	10	3	illustrates	illustrate	VERB
cana-5754	10	4	possible	possible	ADJ
cana-5754	10	5	improvements	improvement	NOUN
cana-5754	10	6	in	in	ADP
cana-5754	10	7	f1	f1	PROPN
cana-5754	10	8	score	score	NOUN
cana-5754	10	9	,	,	PUNCT
cana-5754	10	10	recall	recall	NOUN
cana-5754	10	11	,	,	PUNCT
cana-5754	10	12	accuracy	accuracy	NOUN
cana-5754	10	13	,	,	PUNCT
cana-5754	10	14	and	and	CCONJ
cana-5754	10	15	precision	precision	NOUN
cana-5754	10	16	for	for	ADP
cana-5754	10	17	varying	vary	VERB
cana-5754	10	18	degrees	degree	NOUN
cana-5754	10	19	of	of	ADP
cana-5754	10	20	dr	dr	PROPN
cana-5754	10	21	severity	severity	NOUN
cana-5754	10	22	.	.	PUNCT
cana-5754	11	1	keywords	keyword	NOUN
cana-5754	11	2	:	:	PUNCT
cana-5754	11	3	diabetic	diabetic	ADJ
cana-5754	11	4	retinopathy	retinopathy	ADJ
cana-5754	11	5	gradingfundus	gradingfundus	NOUN
cana-5754	11	6	imagesdeep	imagesdeep	NOUN
cana-5754	11	7	learningattention	learningattention	NOUN
cana-5754	11	8	mechanismexplainable	mechanismexplainable	ADJ
cana-5754	11	9	artificial	artificial	ADJ
cana-5754	11	10	intelligence	intelligence	NOUN
cana-5754	11	11	1	1	NUM
cana-5754	11	12	.	.	PUNCT
cana-5754	11	13	introduction	introduction	NOUN
cana-5754	11	14	one	one	NUM
cana-5754	11	15	microvascular	microvascular	ADJ
cana-5754	11	16	effect	effect	NOUN
cana-5754	11	17	of	of	ADP
cana-5754	11	18	diabetes	diabetes	NOUN
cana-5754	11	19	mellitus	mellitus	NOUN
cana-5754	11	20	that	that	PRON
cana-5754	11	21	might	might	AUX
cana-5754	11	22	impair	impair	VERB
cana-5754	11	23	eyesight	eyesight	NOUN
cana-5754	11	24	is	be	AUX
cana-5754	11	25	diabetic	diabetic	ADJ
cana-5754	11	26	retinopathy	retinopathy	ADJ
cana-5754	11	27	.	.	PUNCT
cana-5754	12	1	it	it	PRON
cana-5754	12	2	affects	affect	VERB
cana-5754	12	3	about	about	ADV
cana-5754	12	4	one	one	NUM
cana-5754	12	5	-	-	PUNCT
cana-5754	12	6	third	third	NOUN
cana-5754	12	7	of	of	ADP
cana-5754	12	8	the	the	DET
cana-5754	12	9	463	463	NUM
cana-5754	12	10	million	million	NUM
cana-5754	12	11	individuals	individual	NOUN
cana-5754	12	12	with	with	ADP
cana-5754	12	13	diabetes	diabetes	NOUN
cana-5754	12	14	globally	globally	ADV
cana-5754	12	15	[	[	X
cana-5754	12	16	1	1	NUM
cana-5754	12	17	]	]	PUNCT
cana-5754	12	18	.	.	PUNCT
cana-5754	13	1	according	accord	VERB
cana-5754	13	2	to	to	ADP
cana-5754	13	3	the	the	DET
cana-5754	13	4	international	international	ADJ
cana-5754	13	5	diabetes	diabetes	PROPN
cana-5754	13	6	federation	federation	PROPN
cana-5754	13	7	(	(	PUNCT
cana-5754	13	8	idf	idf	PROPN
cana-5754	13	9	)	)	PUNCT
cana-5754	13	10	,	,	PUNCT
cana-5754	13	11	the	the	DET
cana-5754	13	12	number	number	NOUN
cana-5754	13	13	of	of	ADP
cana-5754	13	14	people	people	NOUN
cana-5754	13	15	with	with	ADP
cana-5754	13	16	diabetes	diabetes	NOUN
cana-5754	13	17	is	be	AUX
cana-5754	13	18	expected	expect	VERB
cana-5754	13	19	to	to	PART
cana-5754	13	20	rise	rise	VERB
cana-5754	13	21	significantly	significantly	ADV
cana-5754	13	22	from	from	ADP
cana-5754	13	23	the	the	DET
cana-5754	13	24	552	552	NUM
cana-5754	13	25	million	million	NUM
cana-5754	13	26	estimated	estimate	VERB
cana-5754	13	27	in	in	ADP
cana-5754	13	28	2035	2035	NUM
cana-5754	13	29	to	to	ADP
cana-5754	13	30	642	642	NUM
cana-5754	13	31	million	million	NUM
cana-5754	13	32	by	by	ADP
cana-5754	13	33	2040	2040	NUM
cana-5754	14	1	[	[	X
cana-5754	14	2	2	2	NUM
cana-5754	14	3	]	]	PUNCT
cana-5754	14	4	,	,	PUNCT
cana-5754	14	5	[	[	X
cana-5754	14	6	3	3	NUM
cana-5754	14	7	]	]	PUNCT
cana-5754	14	8	.	.	PUNCT
cana-5754	15	1	by	by	ADP
cana-5754	15	2	2030	2030	NUM
cana-5754	15	3	,	,	PUNCT
cana-5754	15	4	there	there	PRON
cana-5754	15	5	will	will	AUX
cana-5754	15	6	be	be	AUX
cana-5754	15	7	over	over	ADP
cana-5754	15	8	191	191	NUM
cana-5754	15	9	million	million	NUM
cana-5754	15	10	persons	person	NOUN
cana-5754	15	11	with	with	ADP
cana-5754	15	12	diabetic	diabetic	ADJ
cana-5754	15	13	retinopathy	retinopathy	NOUN
cana-5754	15	14	,	,	PUNCT
cana-5754	15	15	up	up	ADV
cana-5754	15	16	from	from	ADP
cana-5754	15	17	the	the	DET
cana-5754	15	18	present	present	ADJ
cana-5754	15	19	total	total	NOUN
cana-5754	15	20	of	of	ADP
cana-5754	15	21	over	over	ADP
cana-5754	15	22	158.2	158.2	NUM
cana-5754	15	23	million	million	NUM
cana-5754	15	24	[	[	X
cana-5754	15	25	4	4	NUM
cana-5754	15	26	]	]	PUNCT
cana-5754	15	27	,	,	PUNCT
cana-5754	16	1	[	[	X
cana-5754	16	2	5].diabetic	5].diabetic	NUM
cana-5754	16	3	retinopathy	retinopathy	ADJ
cana-5754	16	4	is	be	AUX
cana-5754	16	5	one	one	NUM
cana-5754	16	6	of	of	ADP
cana-5754	16	7	the	the	DET
cana-5754	16	8	main	main	ADJ
cana-5754	16	9	causes	cause	NOUN
cana-5754	16	10	of	of	ADP
cana-5754	16	11	blindness	blindness	NOUN
cana-5754	16	12	in	in	ADP
cana-5754	16	13	the	the	DET
cana-5754	16	14	communications	communication	NOUN
cana-5754	16	15	on	on	ADP
cana-5754	16	16	applied	apply	VERB
cana-5754	16	17	nonlinear	nonlinear	ADJ
cana-5754	16	18	analysis	analysis	NOUN
cana-5754	16	19	issn	issn	NOUN
cana-5754	16	20	:	:	PUNCT
cana-5754	16	21	1074	1074	NUM
cana-5754	16	22	-	-	PUNCT
cana-5754	16	23	133x	133x	NUM
cana-5754	16	24	vol	vol	NOUN
cana-5754	16	25	32	32	NUM
cana-5754	16	26	no	no	NOUN
cana-5754	16	27	.	.	PUNCT
cana-5754	17	1	3s	3s	NUM
cana-5754	17	2	(	(	PUNCT
cana-5754	17	3	2025	2025	NUM
cana-5754	17	4	)	)	PUNCT
cana-5754	17	5	781	781	NUM
cana-5754	17	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-5754	17	7	modern	modern	ADJ
cana-5754	17	8	world	world	NOUN
cana-5754	17	9	.	.	PUNCT
cana-5754	18	1	those	those	PRON
cana-5754	18	2	with	with	ADP
cana-5754	18	3	poorly	poorly	ADV
cana-5754	18	4	managed	manage	VERB
cana-5754	18	5	blood	blood	NOUN
cana-5754	18	6	sugar	sugar	NOUN
cana-5754	18	7	levels	level	NOUN
cana-5754	18	8	are	be	AUX
cana-5754	18	9	more	more	ADV
cana-5754	18	10	likely	likely	ADJ
cana-5754	18	11	to	to	PART
cana-5754	18	12	suffer	suffer	VERB
cana-5754	18	13	from	from	ADP
cana-5754	18	14	this	this	DET
cana-5754	18	15	condition	condition	NOUN
cana-5754	18	16	.	.	PUNCT
cana-5754	19	1	normal	normal	ADJ
cana-5754	19	2	,	,	PUNCT
cana-5754	19	3	non	non	ADJ
cana-5754	19	4	-	-	ADJ
cana-5754	19	5	proliferative	proliferative	ADJ
cana-5754	19	6	diabetic	diabetic	ADJ
cana-5754	19	7	retinopathy	retinopathy	ADJ
cana-5754	19	8	(	(	PUNCT
cana-5754	19	9	npdr	npdr	NOUN
cana-5754	19	10	)	)	PUNCT
cana-5754	19	11	,	,	PUNCT
cana-5754	19	12	and	and	CCONJ
cana-5754	19	13	proliferative	proliferative	ADJ
cana-5754	19	14	diabetic	diabetic	ADJ
cana-5754	19	15	retinopathy	retinopathy	ADJ
cana-5754	19	16	(	(	PUNCT
cana-5754	19	17	pdr	pdr	NOUN
cana-5754	19	18	)	)	PUNCT
cana-5754	19	19	are	be	AUX
cana-5754	19	20	the	the	DET
cana-5754	19	21	three	three	NUM
cana-5754	19	22	main	main	ADJ
cana-5754	19	23	categories	category	NOUN
cana-5754	19	24	into	into	ADP
cana-5754	19	25	which	which	PRON
cana-5754	19	26	diabetic	diabetic	ADJ
cana-5754	19	27	retinopathy	retinopathy	NOUN
cana-5754	19	28	is	be	AUX
cana-5754	19	29	typically	typically	ADV
cana-5754	19	30	divided	divide	VERB
cana-5754	19	31	[	[	PUNCT
cana-5754	19	32	6].the	6].the	DET
cana-5754	19	33	slowdown	slowdown	NOUN
cana-5754	19	34	of	of	ADP
cana-5754	19	35	retinal	retinal	ADJ
cana-5754	19	36	blood	blood	NOUN
cana-5754	19	37	vessel	vessel	NOUN
cana-5754	19	38	creation	creation	NOUN
cana-5754	19	39	and	and	CCONJ
cana-5754	19	40	the	the	DET
cana-5754	19	41	increasing	increase	VERB
cana-5754	19	42	deterioration	deterioration	NOUN
cana-5754	19	43	of	of	ADP
cana-5754	19	44	the	the	DET
cana-5754	19	45	vessel	vessel	NOUN
cana-5754	19	46	walls	wall	NOUN
cana-5754	19	47	are	be	AUX
cana-5754	19	48	the	the	DET
cana-5754	19	49	primary	primary	ADJ
cana-5754	19	50	characteristics	characteristic	NOUN
cana-5754	19	51	of	of	ADP
cana-5754	19	52	non	non	ADJ
cana-5754	19	53	-	-	ADJ
cana-5754	19	54	proliferative	proliferative	ADJ
cana-5754	19	55	diabetic	diabetic	ADJ
cana-5754	19	56	retinopathy	retinopathy	NOUN
cana-5754	19	57	(	(	PUNCT
cana-5754	19	58	npdr).npdr	npdr).npdr	NOUN
cana-5754	19	59	is	be	AUX
cana-5754	19	60	then	then	ADV
cana-5754	19	61	separated	separate	VERB
cana-5754	19	62	into	into	ADP
cana-5754	19	63	three	three	NUM
cana-5754	19	64	groups	group	NOUN
cana-5754	19	65	:	:	PUNCT
cana-5754	19	66	mild	mild	ADJ
cana-5754	19	67	,	,	PUNCT
cana-5754	19	68	moderate	moderate	ADJ
cana-5754	19	69	,	,	PUNCT
cana-5754	19	70	and	and	CCONJ
cana-5754	19	71	severe	severe	ADJ
cana-5754	19	72	,	,	PUNCT
cana-5754	19	73	depending	depend	VERB
cana-5754	19	74	on	on	ADP
cana-5754	19	75	how	how	SCONJ
cana-5754	19	76	serious	serious	ADJ
cana-5754	19	77	the	the	DET
cana-5754	19	78	problem	problem	NOUN
cana-5754	19	79	is	be	AUX
cana-5754	19	80	.	.	PUNCT
cana-5754	20	1	however	however	ADV
cana-5754	20	2	,	,	PUNCT
cana-5754	20	3	the	the	DET
cana-5754	20	4	formation	formation	NOUN
cana-5754	20	5	of	of	ADP
cana-5754	20	6	new	new	ADJ
cana-5754	20	7	blood	blood	NOUN
cana-5754	20	8	vessels	vessel	NOUN
cana-5754	20	9	in	in	ADP
cana-5754	20	10	the	the	DET
cana-5754	20	11	retina	retina	NOUN
cana-5754	20	12	,	,	PUNCT
cana-5754	20	13	which	which	PRON
cana-5754	20	14	prevents	prevent	VERB
cana-5754	20	15	the	the	DET
cana-5754	20	16	retina	retina	NOUN
cana-5754	20	17	's	's	PART
cana-5754	20	18	normal	normal	ADJ
cana-5754	20	19	blood	blood	NOUN
cana-5754	20	20	flow	flow	NOUN
cana-5754	20	21	,	,	PUNCT
cana-5754	20	22	is	be	AUX
cana-5754	20	23	a	a	DET
cana-5754	20	24	feature	feature	NOUN
cana-5754	20	25	of	of	ADP
cana-5754	20	26	proliferative	proliferative	ADJ
cana-5754	20	27	diabetic	diabetic	ADJ
cana-5754	20	28	retinopathy	retinopathy	ADJ
cana-5754	20	29	(	(	PUNCT
cana-5754	20	30	pdr	pdr	NOUN
cana-5754	20	31	)	)	PUNCT
cana-5754	21	1	[	[	X
cana-5754	21	2	7	7	NUM
cana-5754	21	3	]	]	PUNCT
cana-5754	21	4	,	,	PUNCT
cana-5754	21	5	[	[	X
cana-5754	21	6	8	8	NUM
cana-5754	21	7	]	]	SYM
cana-5754	21	8	.	.	PUNCT
cana-5754	22	1	2	2	X
cana-5754	22	2	.	.	NUM
cana-5754	22	3	proposed	propose	VERB
cana-5754	22	4	system	system	NOUN
cana-5754	22	5	googlenet	googlenet	NOUN
cana-5754	22	6	uses	use	VERB
cana-5754	22	7	the	the	DET
cana-5754	22	8	inception	inception	NOUN
cana-5754	22	9	module	module	NOUN
cana-5754	22	10	to	to	PART
cana-5754	22	11	efficiently	efficiently	ADV
cana-5754	22	12	reduce	reduce	VERB
cana-5754	22	13	the	the	DET
cana-5754	22	14	arithmetic	arithmetic	ADJ
cana-5754	22	15	resources	resource	NOUN
cana-5754	22	16	of	of	ADP
cana-5754	22	17	the	the	DET
cana-5754	22	18	while	while	NOUN
cana-5754	22	19	gaining	gain	VERB
cana-5754	22	20	critical	critical	ADJ
cana-5754	22	21	roomand	roomand	NOUN
cana-5754	22	22	local	local	ADJ
cana-5754	22	23	features	feature	NOUN
cana-5754	22	24	.	.	PUNCT
cana-5754	23	1	resnet-16	resnet-16	PROPN
cana-5754	23	2	includes	include	VERB
cana-5754	23	3	a	a	DET
cana-5754	23	4	skip	skip	ADJ
cana-5754	23	5	connectionto	connectionto	NOUN
cana-5754	23	6	reduce	reduce	VERB
cana-5754	23	7	the	the	DET
cana-5754	23	8	disposal	disposal	NOUN
cana-5754	23	9	of	of	ADP
cana-5754	23	10	issues	issue	NOUN
cana-5754	23	11	and	and	CCONJ
cana-5754	23	12	minimize	minimize	VERB
cana-5754	23	13	training	training	NOUN
cana-5754	23	14	errors	error	NOUN
cana-5754	23	15	.	.	PUNCT
cana-5754	24	1	combining	combine	VERB
cana-5754	24	2	the	the	DET
cana-5754	24	3	features	feature	NOUN
cana-5754	24	4	from	from	ADP
cana-5754	24	5	both	both	DET
cana-5754	24	6	models	model	NOUN
cana-5754	24	7	create	create	VERB
cana-5754	24	8	a	a	DET
cana-5754	24	9	hybrid	hybrid	ADJ
cana-5754	24	10	feature	feature	NOUN
cana-5754	24	11	vector	vector	NOUN
cana-5754	24	12	with	with	ADP
cana-5754	24	13	2000	2000	NUM
cana-5754	24	14	features	feature	NOUN
cana-5754	24	15	.	.	PUNCT
cana-5754	25	1	this	this	DET
cana-5754	25	2	hybrid	hybrid	ADJ
cana-5754	25	3	vectorserves	vectorserve	NOUN
cana-5754	25	4	as	as	ADP
cana-5754	25	5	an	an	DET
cana-5754	25	6	input	input	NOUN
cana-5754	25	7	for	for	ADP
cana-5754	25	8	different	different	ADJ
cana-5754	25	9	classifiers	classifier	NOUN
cana-5754	25	10	,	,	PUNCT
cana-5754	25	11	allowing	allow	VERB
cana-5754	25	12	for	for	ADP
cana-5754	25	13	a	a	DET
cana-5754	25	14	comparative	comparative	ADJ
cana-5754	25	15	evaluation	evaluation	NOUN
cana-5754	25	16	of	of	ADP
cana-5754	25	17	performance	performance	NOUN
cana-5754	25	18	in	in	ADP
cana-5754	25	19	classifying	classify	VERB
cana-5754	25	20	diabetic	diabetic	ADJ
cana-5754	25	21	retinopathy.search	retinopathy.search	NOUN
cana-5754	25	22	photos	photo	NOUN
cana-5754	25	23	are	be	AUX
cana-5754	25	24	divided	divide	VERB
cana-5754	25	25	into	into	ADP
cana-5754	25	26	four	four	NUM
cana-5754	25	27	groups	group	NOUN
cana-5754	25	28	for	for	ADP
cana-5754	25	29	classification	classification	NOUN
cana-5754	25	30	.	.	PUNCT
cana-5754	26	1	there	there	PRON
cana-5754	26	2	is	be	VERB
cana-5754	26	3	no	no	DET
cana-5754	26	4	evidence	evidence	NOUN
cana-5754	26	5	of	of	ADP
cana-5754	26	6	diabetic	diabetic	ADJ
cana-5754	26	7	retinopathy	retinopathy	ADJ
cana-5754	26	8	(	(	PUNCT
cana-5754	26	9	ndr	ndr	PROPN
cana-5754	26	10	)	)	PUNCT
cana-5754	26	11	.	.	PUNCT
cana-5754	27	1	this	this	PRON
cana-5754	27	2	represents	represent	VERB
cana-5754	27	3	levels	level	NOUN
cana-5754	27	4	0	0	PUNCT
cana-5754	27	5	and	and	CCONJ
cana-5754	27	6	level	level	NOUN
cana-5754	27	7	0	0	NUM
cana-5754	27	8	of	of	ADP
cana-5754	27	9	icdrscherskala	icdrscherskala	NOUN
cana-5754	27	10	at	at	ADP
cana-5754	27	11	several	several	ADJ
cana-5754	27	12	stages	stage	NOUN
cana-5754	27	13	of	of	ADP
cana-5754	27	14	diabetic	diabetic	ADJ
cana-5754	27	15	retinopathy	retinopathy	NOUN
cana-5754	27	16	.	.	PUNCT
cana-5754	28	1	multiclas	multiclas	ADJ
cana-5754	28	2	classification	classification	NOUN
cana-5754	28	3	divides	divide	VERB
cana-5754	28	4	the	the	DET
cana-5754	28	5	photo	photo	NOUN
cana-5754	28	6	into	into	ADP
cana-5754	28	7	three	three	NUM
cana-5754	28	8	classes	class	NOUN
cana-5754	28	9	of	of	ADP
cana-5754	28	10	three	three	NUM
cana-5754	28	11	thick	thick	ADJ
cana-5754	28	12	levels	level	NOUN
cana-5754	28	13	:	:	PUNCT
cana-5754	28	14	icdr	icdr	VERB
cana-5754	28	15	:	:	PUNCT
cana-5754	28	16	ndr	ndr	PROPN
cana-5754	28	17	(	(	PUNCT
cana-5754	28	18	level	level	NOUN
cana-5754	28	19	0	0	NUM
cana-5754	28	20	)	)	PUNCT
cana-5754	28	21	,	,	PUNCT
cana-5754	28	22	mdr	mdr	PROPN
cana-5754	28	23	(	(	PUNCT
cana-5754	28	24	level	level	NOUN
cana-5754	28	25	1	1	NUM
cana-5754	28	26	and	and	CCONJ
cana-5754	28	27	2	2	NUM
cana-5754	28	28	)	)	PUNCT
cana-5754	28	29	,	,	PUNCT
cana-5754	28	30	pdr	pdr	PROPN
cana-5754	28	31	(	(	PUNCT
cana-5754	28	32	level	level	NOUN
cana-5754	28	33	3	3	NUM
cana-5754	28	34	and).the	and).the	DET
cana-5754	28	35	methodology	methodology	NOUN
cana-5754	28	36	proposed	propose	VERB
cana-5754	28	37	in	in	ADP
cana-5754	28	38	figure	figure	NOUN
cana-5754	28	39	1	1	NUM
cana-5754	28	40	described	describe	VERB
cana-5754	28	41	the	the	DET
cana-5754	28	42	hybrid	hybrid	ADJ
cana-5754	28	43	property	property	NOUN
cana-5754	28	44	extraction	extraction	NOUN
cana-5754	28	45	and	and	CCONJ
cana-5754	28	46	classification	classification	NOUN
cana-5754	28	47	process	process	NOUN
cana-5754	28	48	of	of	ADP
cana-5754	28	49	the	the	DET
cana-5754	28	50	basic	basic	ADJ
cana-5754	28	51	image	image	NOUN
cana-5754	28	52	.	.	PUNCT
cana-5754	29	1	according	accord	VERB
cana-5754	29	2	to	to	ADP
cana-5754	29	3	temporary	temporary	ADJ
cana-5754	29	4	preprocessing	preprocessing	NOUN
cana-5754	29	5	level	level	NOUN
cana-5754	29	6	images	image	NOUN
cana-5754	29	7	such	such	ADJ
cana-5754	29	8	as	as	ADP
cana-5754	29	9	google	google	PROPN
cana-5754	29	10	and	and	CCONJ
cana-5754	29	11	reset-16	reset-16	PROPN
cana-5754	29	12	are	be	AUX
cana-5754	29	13	input	input	VERB
cana-5754	29	14	into	into	ADP
cana-5754	29	15	the	the	DET
cana-5754	29	16	transfer	transfer	NOUN
cana-5754	29	17	learning	learning	NOUN
cana-5754	29	18	model	model	NOUN
cana-5754	29	19	that	that	PRON
cana-5754	29	20	performs	perform	VERB
cana-5754	29	21	the	the	DET
cana-5754	29	22	functions	function	NOUN
cana-5754	29	23	independently	independently	ADV
cana-5754	29	24	.	.	PUNCT
cana-5754	30	1	the	the	DET
cana-5754	30	2	resulting	result	VERB
cana-5754	30	3	properties	property	NOUN
cana-5754	30	4	are	be	AUX
cana-5754	30	5	used	use	VERB
cana-5754	30	6	as	as	ADP
cana-5754	30	7	the	the	DET
cana-5754	30	8	input	input	NOUN
cana-5754	30	9	for	for	ADP
cana-5754	30	10	various	various	ADJ
cana-5754	30	11	classifiers	classifier	NOUN
cana-5754	30	12	,	,	PUNCT
cana-5754	30	13	such	such	ADJ
cana-5754	30	14	as	as	ADP
cana-5754	30	15	:	:	PUNCT
cana-5754	30	16	the	the	DET
cana-5754	30	17	results	result	NOUN
cana-5754	30	18	are	be	AUX
cana-5754	30	19	compared	compare	VERB
cana-5754	30	20	with	with	ADP
cana-5754	30	21	existing	exist	VERB
cana-5754	30	22	literature	literature	NOUN
cana-5754	30	23	and	and	CCONJ
cana-5754	30	24	provide	provide	VERB
cana-5754	30	25	a	a	DET
cana-5754	30	26	progress	progress	NOUN
cana-5754	30	27	in	in	ADP
cana-5754	30	28	diabetic	diabetic	ADJ
cana-5754	30	29	retinal	retinal	ADJ
cana-5754	30	30	classification	classification	NOUN
cana-5754	30	31	technology	technology	NOUN
cana-5754	30	32	.	.	PUNCT
cana-5754	31	1	3	3	X
cana-5754	31	2	.	.	X
cana-5754	31	3	system	system	NOUN
cana-5754	31	4	architecture	architecture	NOUN
cana-5754	31	5	fig	fig	NOUN
cana-5754	31	6	4.1	4.1	NUM
cana-5754	31	7	system	system	NOUN
cana-5754	31	8	architecture	architecture	NOUN
cana-5754	31	9	communications	communication	NOUN
cana-5754	31	10	on	on	ADP
cana-5754	31	11	applied	apply	VERB
cana-5754	31	12	nonlinear	nonlinear	ADJ
cana-5754	31	13	analysis	analysis	NOUN
cana-5754	31	14	issn	issn	NOUN
cana-5754	31	15	:	:	PUNCT
cana-5754	31	16	1074	1074	NUM
cana-5754	31	17	-	-	PUNCT
cana-5754	31	18	133x	133x	NUM
cana-5754	31	19	vol	vol	NOUN
cana-5754	31	20	32	32	NUM
cana-5754	31	21	no	no	NOUN
cana-5754	31	22	.	.	PUNCT
cana-5754	32	1	3s	3s	NUM
cana-5754	32	2	(	(	PUNCT
cana-5754	32	3	2025	2025	NUM
cana-5754	32	4	)	)	PUNCT
cana-5754	32	5	782	782	NUM
cana-5754	32	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-5754	32	7	fig	fig	NOUN
cana-5754	32	8	4	4	NUM
cana-5754	32	9	.	.	NOUN
cana-5754	32	10	1	1	NUM
cana-5754	32	11	:	:	PUNCT
cana-5754	32	12	architecture	architecture	NOUN
cana-5754	32	13	of	of	ADP
cana-5754	32	14	the	the	DET
cana-5754	32	15	proposed	propose	VERB
cana-5754	32	16	methodology	methodology	NOUN
cana-5754	32	17	for	for	ADP
cana-5754	32	18	diabetic	diabetic	ADJ
cana-5754	32	19	retinopathy	retinopathy	ADJ
cana-5754	32	20	detection	detection	NOUN
cana-5754	32	21	.	.	PUNCT
cana-5754	33	1	retinal	retinal	ADJ
cana-5754	33	2	images	image	NOUN
cana-5754	33	3	are	be	AUX
cana-5754	33	4	preprocessed	preprocesse	VERB
cana-5754	33	5	—	—	PUNCT
cana-5754	33	6	resized	resize	VERB
cana-5754	33	7	,	,	PUNCT
cana-5754	33	8	denoised	denoise	VERB
cana-5754	33	9	,	,	PUNCT
cana-5754	33	10	and	and	CCONJ
cana-5754	33	11	normalized	normalize	VERB
cana-5754	33	12	—	—	PUNCT
cana-5754	33	13	then	then	ADV
cana-5754	33	14	partitioned	partition	VERB
cana-5754	33	15	into	into	ADP
cana-5754	33	16	training	training	NOUN
cana-5754	33	17	,	,	PUNCT
cana-5754	33	18	validation	validation	NOUN
cana-5754	33	19	,	,	PUNCT
cana-5754	33	20	and	and	CCONJ
cana-5754	33	21	test	test	NOUN
cana-5754	33	22	sets	set	NOUN
cana-5754	33	23	to	to	PART
cana-5754	33	24	ensure	ensure	VERB
cana-5754	33	25	robust	robust	ADJ
cana-5754	33	26	model	model	NOUN
cana-5754	33	27	development	development	NOUN
cana-5754	33	28	and	and	CCONJ
cana-5754	33	29	evaluation	evaluation	NOUN
cana-5754	33	30	4	4	NUM
cana-5754	33	31	.	.	PUNCT
cana-5754	33	32	methodology	methodology	NOUN
cana-5754	33	33	.	.	PUNCT
cana-5754	34	1	data	datum	NOUN
cana-5754	34	2	preprocessing	preprocessing	NOUN
cana-5754	34	3	and	and	CCONJ
cana-5754	34	4	enhanced	enhance	VERB
cana-5754	34	5	predicting	predict	VERB
cana-5754	34	6	diabetic	diabetic	ADJ
cana-5754	34	7	retinopathy	retinopathy	NOUN
cana-5754	34	8	typically	typically	ADV
cana-5754	34	9	requires	require	VERB
cana-5754	34	10	preprocessing	preprocesse	VERB
cana-5754	34	11	of	of	ADP
cana-5754	34	12	retinal	retinal	ADJ
cana-5754	34	13	images	image	NOUN
cana-5754	34	14	to	to	PART
cana-5754	34	15	improve	improve	VERB
cana-5754	34	16	model	model	NOUN
cana-5754	34	17	output	output	NOUN
cana-5754	34	18	and	and	CCONJ
cana-5754	34	19	accuracy	accuracy	NOUN
cana-5754	34	20	.	.	PUNCT
cana-5754	35	1	below	below	ADP
cana-5754	35	2	you	you	PRON
cana-5754	35	3	will	will	AUX
cana-5754	35	4	find	find	VERB
cana-5754	35	5	a	a	DET
cana-5754	35	6	pretreatment	pretreatment	NOUN
cana-5754	35	7	method	method	NOUN
cana-5754	35	8	:	:	PUNCT
cana-5754	36	1	1	1	X
cana-5754	36	2	.	.	PUNCT
cana-5754	36	3	temp	temp	NOUN
cana-5754	36	4	:	:	PUNCT
cana-5754	36	5	harves	harve	VERB
cana-5754	36	6	up	up	ADP
cana-5754	36	7	non	non	ADJ
cana-5754	36	8	-	-	ADJ
cana-5754	36	9	reticular	reticular	ADJ
cana-5754	36	10	parts	part	NOUN
cana-5754	36	11	of	of	ADP
cana-5754	36	12	the	the	DET
cana-5754	36	13	image	image	NOUN
cana-5754	36	14	so	so	SCONJ
cana-5754	36	15	that	that	SCONJ
cana-5754	36	16	the	the	DET
cana-5754	36	17	model	model	NOUN
cana-5754	36	18	focuses	focus	VERB
cana-5754	36	19	only	only	ADV
cana-5754	36	20	on	on	ADP
cana-5754	36	21	areas	area	NOUN
cana-5754	36	22	of	of	ADP
cana-5754	36	23	interest	interest	NOUN
cana-5754	36	24	,	,	PUNCT
cana-5754	36	25	increasing	increase	VERB
cana-5754	36	26	processing	processing	NOUN
cana-5754	36	27	efficiency	efficiency	NOUN
cana-5754	36	28	.	.	PUNCT
cana-5754	37	1	2	2	X
cana-5754	37	2	.	.	X
cana-5754	37	3	try	try	VERB
cana-5754	37	4	gausschenblur	gausschenblur	NOUN
cana-5754	37	5	's	's	PART
cana-5754	37	6	pre	pre	ADJ
cana-5754	37	7	-	-	ADJ
cana-5754	37	8	processing	processing	ADJ
cana-5754	37	9	method	method	NOUN
cana-5754	37	10	:	:	PUNCT
cana-5754	37	11	the	the	DET
cana-5754	37	12	brightness	brightness	NOUN
cana-5754	37	13	and	and	CCONJ
cana-5754	37	14	contrast	contrast	NOUN
cana-5754	37	15	of	of	ADP
cana-5754	37	16	diabetic	diabetic	ADJ
cana-5754	37	17	retinopathy	retinopathy	NOUN
cana-5754	37	18	are	be	AUX
cana-5754	37	19	very	very	ADV
cana-5754	37	20	different	different	ADJ
cana-5754	37	21	,	,	PUNCT
cana-5754	37	22	so	so	CCONJ
cana-5754	37	23	gaußsche	gaußsche	NOUN
cana-5754	37	24	blurs	blur	VERB
cana-5754	37	25	can	can	AUX
cana-5754	37	26	be	be	AUX
cana-5754	37	27	used	use	VERB
cana-5754	37	28	to	to	PART
cana-5754	37	29	improve	improve	VERB
cana-5754	37	30	image	image	NOUN
cana-5754	37	31	improvements	improvement	NOUN
cana-5754	37	32	to	to	PART
cana-5754	37	33	improve	improve	VERB
cana-5754	37	34	image	image	NOUN
cana-5754	37	35	contrast	contrast	NOUN
cana-5754	37	36	and	and	CCONJ
cana-5754	37	37	visualization	visualization	NOUN
cana-5754	37	38	.	.	PUNCT
cana-5754	38	1	3	3	X
cana-5754	38	2	.	.	X
cana-5754	38	3	tripping	trip	VERB
cana-5754	38	4	and	and	CCONJ
cana-5754	38	5	gaussian	gaussian	ADJ
cana-5754	38	6	blur	blur	NOUN
cana-5754	38	7	color	color	NOUN
cana-5754	38	8	version	version	NOUN
cana-5754	38	9	pre	pre	ADJ
cana-5754	38	10	-	-	NOUN
cana-5754	38	11	processing	processing	ADJ
cana-5754	38	12	:	:	PUNCT
cana-5754	38	13	while	while	SCONJ
cana-5754	38	14	cutting	cut	VERB
cana-5754	38	15	retina	retina	NOUN
cana-5754	38	16	images	image	NOUN
cana-5754	38	17	,	,	PUNCT
cana-5754	38	18	you	you	PRON
cana-5754	38	19	can	can	AUX
cana-5754	38	20	also	also	ADV
cana-5754	38	21	use	use	VERB
cana-5754	38	22	the	the	DET
cana-5754	38	23	gaußsche	gaußsche	PROPN
cana-5754	38	24	blur	blur	PROPN
cana-5754	38	25	pre	pre	ADJ
cana-5754	38	26	-	-	ADJ
cana-5754	38	27	processing	processing	ADJ
cana-5754	38	28	method	method	NOUN
cana-5754	38	29	to	to	PART
cana-5754	38	30	improve	improve	VERB
cana-5754	38	31	the	the	DET
cana-5754	38	32	image	image	NOUN
cana-5754	38	33	and	and	CCONJ
cana-5754	38	34	prevent	prevent	VERB
cana-5754	38	35	corrupting	corrupt	VERB
cana-5754	38	36	the	the	DET
cana-5754	38	37	color	color	NOUN
cana-5754	38	38	information	information	NOUN
cana-5754	38	39	of	of	ADP
cana-5754	38	40	the	the	DET
cana-5754	38	41	image	image	NOUN
cana-5754	38	42	.	.	PUNCT
cana-5754	39	1	4	4	X
cana-5754	39	2	.	.	X
cana-5754	39	3	auto	auto	NOUN
cana-5754	39	4	crop	crop	NOUN
cana-5754	39	5	:	:	PUNCT
cana-5754	39	6	the	the	DET
cana-5754	39	7	automatic	automatic	ADJ
cana-5754	39	8	cutting	cutting	NOUN
cana-5754	39	9	algorithm	algorithm	NOUN
cana-5754	39	10	can	can	AUX
cana-5754	39	11	automatically	automatically	ADV
cana-5754	39	12	dig	dig	VERB
cana-5754	39	13	areas	area	NOUN
cana-5754	39	14	of	of	ADP
cana-5754	39	15	interest	interest	NOUN
cana-5754	39	16	according	accord	VERB
cana-5754	39	17	to	to	ADP
cana-5754	39	18	the	the	DET
cana-5754	39	19	characteristics	characteristic	NOUN
cana-5754	39	20	of	of	ADP
cana-5754	39	21	the	the	DET
cana-5754	39	22	retinal	retinal	ADJ
cana-5754	39	23	image	image	NOUN
cana-5754	39	24	,	,	PUNCT
cana-5754	39	25	improving	improve	VERB
cana-5754	39	26	the	the	DET
cana-5754	39	27	efficiency	efficiency	NOUN
cana-5754	39	28	and	and	CCONJ
cana-5754	39	29	accuracy	accuracy	NOUN
cana-5754	39	30	of	of	ADP
cana-5754	39	31	preprocessing	preprocessing	NOUN
cana-5754	39	32	.	.	PUNCT
cana-5754	40	1	5	5	X
cana-5754	40	2	.	.	X
cana-5754	40	3	central	central	ADJ
cana-5754	40	4	access	access	NOUN
cana-5754	40	5	:	:	PUNCT
cana-5754	40	6	for	for	ADP
cana-5754	40	7	large	large	ADJ
cana-5754	40	8	-	-	PUNCT
cana-5754	40	9	scale	scale	NOUN
cana-5754	40	10	retinal	retinal	ADJ
cana-5754	40	11	images	image	NOUN
cana-5754	40	12	,	,	PUNCT
cana-5754	40	13	you	you	PRON
cana-5754	40	14	can	can	AUX
cana-5754	40	15	choose	choose	VERB
cana-5754	40	16	the	the	DET
cana-5754	40	17	center	center	ADJ
cana-5754	40	18	cultivation	cultivation	NOUN
cana-5754	40	19	method	method	NOUN
cana-5754	40	20	.	.	PUNCT
cana-5754	41	1	this	this	PRON
cana-5754	41	2	means	mean	VERB
cana-5754	41	3	that	that	SCONJ
cana-5754	41	4	the	the	DET
cana-5754	41	5	central	central	ADJ
cana-5754	41	6	portion	portion	NOUN
cana-5754	41	7	of	of	ADP
cana-5754	41	8	the	the	DET
cana-5754	41	9	image	image	NOUN
cana-5754	41	10	is	be	AUX
cana-5754	41	11	preserved	preserve	VERB
cana-5754	41	12	and	and	CCONJ
cana-5754	41	13	other	other	ADJ
cana-5754	41	14	preprocessing	preprocessing	NOUN
cana-5754	41	15	methods	method	NOUN
cana-5754	41	16	can	can	AUX
cana-5754	41	17	be	be	AUX
cana-5754	41	18	used	use	VERB
cana-5754	41	19	to	to	PART
cana-5754	41	20	improve	improve	VERB
cana-5754	41	21	image	image	NOUN
cana-5754	41	22	contrast	contrast	NOUN
cana-5754	41	23	and	and	CCONJ
cana-5754	41	24	visualization	visualization	NOUN
cana-5754	41	25	.	.	PUNCT
cana-5754	42	1	professor	professor	NOUN
cana-5754	42	2	.	.	PUNCT
cana-5754	43	1	(	(	PUNCT
cana-5754	43	2	a	a	X
cana-5754	43	3	)	)	PUNCT
cana-5754	43	4	the	the	DET
cana-5754	43	5	first	first	ADJ
cana-5754	43	6	row	row	NOUN
cana-5754	43	7	indicates	indicate	VERB
cana-5754	43	8	that	that	SCONJ
cana-5754	43	9	the	the	DET
cana-5754	43	10	original	original	ADJ
cana-5754	43	11	image	image	NOUN
cana-5754	43	12	was	be	AUX
cana-5754	43	13	not	not	PART
cana-5754	43	14	processed	process	VERB
cana-5754	43	15	by	by	ADP
cana-5754	43	16	the	the	DET
cana-5754	43	17	image	image	NOUN
cana-5754	43	18	preprocessing	preprocesse	VERB
cana-5754	43	19	algorithm	algorithm	NOUN
cana-5754	43	20	.	.	PUNCT
cana-5754	44	1	(	(	PUNCT
cana-5754	44	2	b	b	X
cana-5754	44	3	)	)	PUNCT
cana-5754	44	4	line	line	NOUN
cana-5754	44	5	2	2	NUM
cana-5754	44	6	represents	represent	VERB
cana-5754	44	7	the	the	DET
cana-5754	44	8	results	result	NOUN
cana-5754	44	9	of	of	ADP
cana-5754	44	10	the	the	DET
cana-5754	44	11	original	original	ADJ
cana-5754	44	12	image	image	NOUN
cana-5754	44	13	processed	process	VERB
cana-5754	44	14	by	by	ADP
cana-5754	44	15	the	the	DET
cana-5754	44	16	image	image	NOUN
cana-5754	44	17	preprocessing	preprocesse	VERB
cana-5754	44	18	algorithm	algorithm	NOUN
cana-5754	44	19	proposed	propose	VERB
cana-5754	44	20	in	in	ADP
cana-5754	44	21	this	this	DET
cana-5754	44	22	article	article	NOUN
cana-5754	44	23	.	.	PUNCT
cana-5754	45	1	an	an	DET
cana-5754	45	2	architecture	architecture	NOUN
cana-5754	45	3	of	of	ADP
cana-5754	45	4	a	a	DET
cana-5754	45	5	hybrid	hybrid	ADJ
cana-5754	45	6	model	model	NOUN
cana-5754	45	7	with	with	ADP
cana-5754	45	8	efficient	efficient	ADJ
cana-5754	45	9	swin	swin	PROPN
cana-5754	45	10	transformer	transformer	NOUN
cana-5754	45	11	this	this	DET
cana-5754	45	12	study	study	NOUN
cana-5754	45	13	proposes	propose	VERB
cana-5754	45	14	a	a	DET
cana-5754	45	15	hybrid	hybrid	ADJ
cana-5754	45	16	model	model	NOUN
cana-5754	45	17	combining	combine	VERB
cana-5754	45	18	efficient	efficient	ADJ
cana-5754	45	19	net	net	NOUN
cana-5754	45	20	and	and	CCONJ
cana-5754	45	21	swin	swin	PROPN
cana-5754	45	22	transformer	transformer	NOUN
cana-5754	45	23	for	for	ADP
cana-5754	45	24	prediction	prediction	NOUN
cana-5754	45	25	of	of	ADP
cana-5754	45	26	diabetic	diabetic	ADJ
cana-5754	45	27	retinopathy	retinopathy	NOUN
cana-5754	45	28	.	.	PUNCT
cana-5754	46	1	this	this	DET
cana-5754	46	2	model	model	NOUN
cana-5754	46	3	is	be	AUX
cana-5754	46	4	divided	divide	VERB
cana-5754	46	5	into	into	ADP
cana-5754	46	6	two	two	NUM
cana-5754	46	7	branches	branch	NOUN
cana-5754	46	8	,	,	PUNCT
cana-5754	46	9	one	one	NUM
cana-5754	46	10	of	of	ADP
cana-5754	46	11	which	which	PRON
cana-5754	46	12	is	be	AUX
cana-5754	46	13	used	use	VERB
cana-5754	46	14	as	as	ADP
cana-5754	46	15	a	a	DET
cana-5754	46	16	backbone	backbone	NOUN
cana-5754	46	17	network	network	NOUN
cana-5754	46	18	and	and	CCONJ
cana-5754	46	19	the	the	DET
cana-5754	46	20	other	other	ADJ
cana-5754	46	21	uses	use	VERB
cana-5754	46	22	a	a	DET
cana-5754	46	23	swin	swin	PROPN
cana-5754	46	24	transformer	transformer	NOUN
cana-5754	46	25	.	.	PUNCT
cana-5754	47	1	the	the	DET
cana-5754	47	2	outputs	output	NOUN
cana-5754	47	3	of	of	ADP
cana-5754	47	4	the	the	DET
cana-5754	47	5	two	two	NUM
cana-5754	47	6	branches	branch	NOUN
cana-5754	47	7	are	be	AUX
cana-5754	47	8	chained	chain	VERB
cana-5754	47	9	and	and	CCONJ
cana-5754	47	10	fused	fuse	VERB
cana-5754	47	11	by	by	ADP
cana-5754	47	12	fully	fully	ADV
cana-5754	47	13	connected	connected	ADJ
cana-5754	47	14	layers	layer	NOUN
cana-5754	47	15	.	.	PUNCT
cana-5754	48	1	framework	framework	NOUN
cana-5754	48	2	structure	structure	NOUN
cana-5754	48	3	for	for	ADP
cana-5754	48	4	a	a	DET
cana-5754	48	5	hybrid	hybrid	ADJ
cana-5754	48	6	model	model	NOUN
cana-5754	48	7	with	with	ADP
cana-5754	48	8	two	two	NUM
cana-5754	48	9	branches	branch	NOUN
cana-5754	48	10	.	.	PUNCT
cana-5754	49	1	using	use	VERB
cana-5754	49	2	the	the	DET
cana-5754	49	3	benefits	benefit	NOUN
cana-5754	49	4	of	of	ADP
cana-5754	49	5	swin	swin	PROPN
cana-5754	49	6	transformer	transformer	PROPN
cana-5754	49	7	and	and	CCONJ
cana-5754	49	8	efficient	efficient	ADJ
cana-5754	49	9	network	network	NOUN
cana-5754	49	10	models	model	NOUN
cana-5754	49	11	,	,	PUNCT
cana-5754	49	12	accurately	accurately	ADV
cana-5754	49	13	record	record	VERB
cana-5754	49	14	global	global	ADJ
cana-5754	49	15	and	and	CCONJ
cana-5754	49	16	local	local	ADJ
cana-5754	49	17	lesion	lesion	NOUN
cana-5754	49	18	features	feature	NOUN
cana-5754	49	19	,	,	PUNCT
cana-5754	49	20	and	and	CCONJ
cana-5754	49	21	combine	combine	VERB
cana-5754	49	22	global	global	ADJ
cana-5754	49	23	and	and	CCONJ
cana-5754	49	24	local	local	ADJ
cana-5754	49	25	features	feature	NOUN
cana-5754	49	26	to	to	PART
cana-5754	49	27	connect	connect	VERB
cana-5754	49	28	classification	classification	NOUN
cana-5754	49	29	layers	layer	NOUN
cana-5754	49	30	to	to	PART
cana-5754	49	31	classify	classify	VERB
cana-5754	49	32	classification	classification	NOUN
cana-5754	49	33	retinal	retinal	ADJ
cana-5754	49	34	images	image	NOUN
cana-5754	49	35	.	.	PUNCT
cana-5754	50	1	in	in	ADP
cana-5754	50	2	particular	particular	ADJ
cana-5754	50	3	,	,	PUNCT
cana-5754	50	4	efficient	efficient	ADJ
cana-5754	50	5	net	net	ADJ
cana-5754	50	6	branches	branch	NOUN
cana-5754	50	7	are	be	AUX
cana-5754	50	8	primarily	primarily	ADV
cana-5754	50	9	used	use	VERB
cana-5754	50	10	to	to	PART
cana-5754	50	11	extract	extract	VERB
cana-5754	50	12	local	local	ADJ
cana-5754	50	13	features	feature	NOUN
cana-5754	50	14	of	of	ADP
cana-5754	50	15	images	image	NOUN
cana-5754	50	16	,	,	PUNCT
cana-5754	50	17	while	while	SCONJ
cana-5754	50	18	swin	swin	PROPN
cana-5754	50	19	trans	trans	PROPN
cana-5754	50	20	branches	branch	NOUN
cana-5754	50	21	are	be	AUX
cana-5754	50	22	used	use	VERB
cana-5754	50	23	to	to	PART
cana-5754	50	24	extract	extract	VERB
cana-5754	50	25	global	global	ADJ
cana-5754	50	26	properties	property	NOUN
cana-5754	50	27	of	of	ADP
cana-5754	50	28	images	image	NOUN
cana-5754	50	29	.	.	PUNCT
cana-5754	51	1	an	an	DET
cana-5754	51	2	efficient	efficient	ADJ
cana-5754	51	3	net	net	ADJ
cana-5754	51	4	branch	branch	NOUN
cana-5754	51	5	consists	consist	VERB
cana-5754	51	6	of	of	ADP
cana-5754	51	7	a	a	DET
cana-5754	51	8	basic	basic	ADJ
cana-5754	51	9	network	network	NOUN
cana-5754	51	10	,	,	PUNCT
cana-5754	51	11	feature	feature	NOUN
cana-5754	51	12	extraction	extraction	NOUN
cana-5754	51	13	layer	layer	NOUN
cana-5754	51	14	,	,	PUNCT
cana-5754	51	15	global	global	ADJ
cana-5754	51	16	average	average	ADJ
cana-5754	51	17	pooling	pool	VERB
cana-5754	51	18	https://www.researchgate.net/publication/359023281_a_deep_learning_approach_for_the_diabetic_retinopathy_detection?utm_source=chatgpt.com	https://www.researchgate.net/publication/359023281_a_deep_learning_approach_for_the_diabetic_retinopathy_detection?utm_source=chatgpt.com	ADJ
cana-5754	51	19	communications	communication	NOUN
cana-5754	51	20	on	on	ADP
cana-5754	51	21	applied	apply	VERB
cana-5754	51	22	nonlinear	nonlinear	ADJ
cana-5754	51	23	analysis	analysis	NOUN
cana-5754	51	24	issn	issn	NOUN
cana-5754	51	25	:	:	PUNCT
cana-5754	51	26	1074	1074	NUM
cana-5754	51	27	-	-	PUNCT
cana-5754	51	28	133x	133x	NUM
cana-5754	51	29	vol	vol	NOUN
cana-5754	51	30	32	32	NUM
cana-5754	51	31	no	no	NOUN
cana-5754	51	32	.	.	PUNCT
cana-5754	52	1	3s	3s	NUM
cana-5754	52	2	(	(	PUNCT
cana-5754	52	3	2025	2025	NUM
cana-5754	52	4	)	)	PUNCT
cana-5754	52	5	783	783	NUM
cana-5754	52	6	https://internationalpubls.com	https://internationalpubls.com	NUM
cana-5754	52	7	levels	level	NOUN
cana-5754	52	8	,	,	PUNCT
cana-5754	52	9	and	and	CCONJ
cana-5754	52	10	some	some	DET
cana-5754	52	11	head	head	NOUN
cana-5754	52	12	attention	attention	NOUN
cana-5754	52	13	.	.	PUNCT
cana-5754	53	1	the	the	DET
cana-5754	53	2	swin	swin	PROPN
cana-5754	53	3	transformer	transformer	PROPN
cana-5754	53	4	branch	branch	PROPN
cana-5754	53	5	consists	consist	VERB
cana-5754	53	6	of	of	ADP
cana-5754	53	7	swin	swin	PROPN
cana-5754	53	8	transformer	transformer	PROPN
cana-5754	53	9	blocks	block	NOUN
cana-5754	53	10	,	,	PUNCT
cana-5754	53	11	adaptive	adaptive	ADJ
cana-5754	53	12	pooling	pool	VERB
cana-5754	53	13	layers	layer	NOUN
cana-5754	53	14	,	,	PUNCT
cana-5754	53	15	attention	attention	NOUN
cana-5754	53	16	coordinates	coordinate	NOUN
cana-5754	53	17	,	,	PUNCT
cana-5754	53	18	and	and	CCONJ
cana-5754	53	19	fully	fully	ADV
cana-5754	53	20	connected	connected	ADJ
cana-5754	53	21	layers	layer	NOUN
cana-5754	53	22	.	.	PUNCT
cana-5754	54	1	a	a	DET
cana-5754	54	2	very	very	ADV
cana-5754	54	3	small	small	ADJ
cana-5754	54	4	window	window	NOUN
cana-5754	54	5	size	size	NOUN
cana-5754	54	6	is	be	AUX
cana-5754	54	7	used	use	VERB
cana-5754	54	8	in	in	ADP
cana-5754	54	9	swin	swin	PROPN
cana-5754	54	10	transformer	transformer	NOUN
cana-5754	54	11	blocks	block	NOUN
cana-5754	54	12	to	to	PART
cana-5754	54	13	improve	improve	VERB
cana-5754	54	14	the	the	DET
cana-5754	54	15	ability	ability	NOUN
cana-5754	54	16	to	to	PART
cana-5754	54	17	share	share	VERB
cana-5754	54	18	images	image	NOUN
cana-5754	54	19	and	and	CCONJ
cana-5754	54	20	recognize	recognize	VERB
cana-5754	54	21	details	detail	NOUN
cana-5754	54	22	.	.	PUNCT
cana-5754	55	1	diabetic	diabetic	ADJ
cana-5754	55	2	retinal	retinal	ADJ
cana-5754	55	3	images	image	NOUN
cana-5754	55	4	are	be	AUX
cana-5754	55	5	efficiently	efficiently	ADV
cana-5754	55	6	delivered	deliver	VERB
cana-5754	55	7	.	.	PUNCT
cana-5754	56	1	efficientnet	efficientnet	PROPN
cana-5754	56	2	consists	consist	VERB
cana-5754	56	3	of	of	ADP
cana-5754	56	4	several	several	ADJ
cana-5754	56	5	layers	layer	NOUN
cana-5754	56	6	of	of	ADP
cana-5754	56	7	multipurpose	multipurpose	NOUN
cana-5754	56	8	attention	attention	NOUN
cana-5754	56	9	that	that	PRON
cana-5754	56	10	achieves	achieve	VERB
cana-5754	56	11	varying	vary	VERB
cana-5754	56	12	depths	depth	NOUN
cana-5754	56	13	and	and	CCONJ
cana-5754	56	14	widths	width	NOUN
cana-5754	56	15	,	,	PUNCT
cana-5754	56	16	batch	batch	VERB
cana-5754	56	17	normalization	normalization	NOUN
cana-5754	56	18	levels	level	NOUN
cana-5754	56	19	,	,	PUNCT
cana-5754	56	20	activation	activation	NOUN
cana-5754	56	21	layers	layer	NOUN
cana-5754	56	22	,	,	PUNCT
cana-5754	56	23	pooling	pool	VERB
cana-5754	56	24	layers	layer	NOUN
cana-5754	56	25	,	,	PUNCT
cana-5754	56	26	and	and	CCONJ
cana-5754	56	27	localized	localized	ADJ
cana-5754	56	28	image	image	NOUN
cana-5754	56	29	characterization	characterization	NOUN
cana-5754	56	30	extraction	extraction	NOUN
cana-5754	56	31	.	.	PUNCT
cana-5754	57	1	5	5	X
cana-5754	57	2	.	.	NOUN
cana-5754	57	3	results	result	NOUN
cana-5754	57	4	and	and	CCONJ
cana-5754	57	5	discussion	discussion	NOUN
cana-5754	57	6	classification	classification	NOUN
cana-5754	57	7	report	report	NOUN
cana-5754	57	8	:	:	PUNCT
cana-5754	57	9	the	the	DET
cana-5754	57	10	classification	classification	NOUN
cana-5754	57	11	report	report	NOUN
cana-5754	57	12	evaluates	evaluate	VERB
cana-5754	57	13	a	a	DET
cana-5754	57	14	cnn	cnn	PROPN
cana-5754	57	15	-	-	PUNCT
cana-5754	57	16	mobilenet	mobilenet	ADJ
cana-5754	57	17	model	model	NOUN
cana-5754	57	18	for	for	ADP
cana-5754	57	19	diabetic	diabetic	ADJ
cana-5754	57	20	retinopathy	retinopathy	ADJ
cana-5754	57	21	detection	detection	NOUN
cana-5754	57	22	across	across	ADP
cana-5754	57	23	five	five	NUM
cana-5754	57	24	stages	stage	NOUN
cana-5754	57	25	.	.	PUNCT
cana-5754	58	1	the	the	DET
cana-5754	58	2	model	model	NOUN
cana-5754	58	3	performs	perform	VERB
cana-5754	58	4	exceptionally	exceptionally	ADV
cana-5754	58	5	for	for	ADP
cana-5754	58	6	no_dr	no_dr	PROPN
cana-5754	58	7	with	with	ADP
cana-5754	58	8	a	a	DET
cana-5754	58	9	precision	precision	NOUN
cana-5754	58	10	,	,	PUNCT
cana-5754	58	11	recall	recall	NOUN
cana-5754	58	12	,	,	PUNCT
cana-5754	58	13	and	and	CCONJ
cana-5754	58	14	f1	f1	NOUN
cana-5754	58	15	-	-	PUNCT
cana-5754	58	16	score	score	NOUN
cana-5754	58	17	of	of	ADP
cana-5754	58	18	0.95	0.95	NUM
cana-5754	58	19	,	,	PUNCT
cana-5754	58	20	reflecting	reflect	VERB
cana-5754	58	21	strong	strong	ADJ
cana-5754	58	22	identification	identification	NOUN
cana-5754	58	23	of	of	ADP
cana-5754	58	24	healthy	healthy	ADJ
cana-5754	58	25	cases	case	NOUN
cana-5754	58	26	.	.	PUNCT
cana-5754	59	1	however	however	ADV
cana-5754	59	2	,	,	PUNCT
cana-5754	59	3	performance	performance	NOUN
cana-5754	59	4	drops	drop	VERB
cana-5754	59	5	significantly	significantly	ADV
cana-5754	59	6	for	for	ADP
cana-5754	59	7	advanced	advanced	ADJ
cana-5754	59	8	stages	stage	NOUN
cana-5754	59	9	like	like	ADP
cana-5754	59	10	proliferate_dr	proliferate_dr	NOUN
cana-5754	59	11	(	(	PUNCT
cana-5754	59	12	f1	f1	NOUN
cana-5754	59	13	-	-	PUNCT
cana-5754	59	14	score	score	NOUN
cana-5754	59	15	:	:	PUNCT
cana-5754	59	16	0.33	0.33	NUM
cana-5754	59	17	)	)	PUNCT
cana-5754	59	18	and	and	CCONJ
cana-5754	59	19	severe	severe	ADJ
cana-5754	59	20	(	(	PUNCT
cana-5754	59	21	f1	f1	NOUN
cana-5754	59	22	-	-	PUNCT
cana-5754	59	23	score	score	NOUN
cana-5754	59	24	:	:	PUNCT
cana-5754	59	25	0.37	0.37	NUM
cana-5754	59	26	)	)	PUNCT
cana-5754	59	27	,	,	PUNCT
cana-5754	59	28	indicating	indicate	VERB
cana-5754	59	29	challenges	challenge	NOUN
cana-5754	59	30	in	in	ADP
cana-5754	59	31	correctly	correctly	ADV
cana-5754	59	32	classifying	classify	VERB
cana-5754	59	33	critical	critical	ADJ
cana-5754	59	34	disease	disease	NOUN
cana-5754	59	35	stages	stage	NOUN
cana-5754	59	36	.	.	PUNCT
cana-5754	60	1	the	the	DET
cana-5754	60	2	macro	macro	ADJ
cana-5754	60	3	average	average	ADJ
cana-5754	60	4	f1	f1	NOUN
cana-5754	60	5	-	-	PUNCT
cana-5754	60	6	score	score	NOUN
cana-5754	60	7	is	be	AUX
cana-5754	60	8	0.54	0.54	NUM
cana-5754	60	9	,	,	PUNCT
cana-5754	60	10	revealing	reveal	VERB
cana-5754	60	11	imbalanced	imbalanced	ADJ
cana-5754	60	12	performance	performance	NOUN
cana-5754	60	13	across	across	ADP
cana-5754	60	14	classes	class	NOUN
cana-5754	60	15	.	.	PUNCT
cana-5754	61	1	despite	despite	SCONJ
cana-5754	61	2	an	an	DET
cana-5754	61	3	overall	overall	ADJ
cana-5754	61	4	accuracy	accuracy	NOUN
cana-5754	61	5	of	of	ADP
cana-5754	61	6	71	71	NUM
cana-5754	61	7	%	%	NOUN
cana-5754	61	8	,	,	PUNCT
cana-5754	61	9	the	the	DET
cana-5754	61	10	model	model	NOUN
cana-5754	61	11	demonstrates	demonstrate	VERB
cana-5754	61	12	a	a	DET
cana-5754	61	13	tendency	tendency	NOUN
cana-5754	61	14	to	to	PART
cana-5754	61	15	favor	favor	VERB
cana-5754	61	16	the	the	DET
cana-5754	61	17	majority	majority	NOUN
cana-5754	61	18	class	class	NOUN
cana-5754	61	19	(	(	PUNCT
cana-5754	61	20	no_dr	no_dr	PROPN
cana-5754	61	21	)	)	PUNCT
cana-5754	61	22	,	,	PUNCT
cana-5754	61	23	highlighting	highlight	VERB
cana-5754	61	24	the	the	DET
cana-5754	61	25	need	need	NOUN
cana-5754	61	26	for	for	ADP
cana-5754	61	27	better	well	ADJ
cana-5754	61	28	class	class	NOUN
cana-5754	61	29	balance	balance	NOUN
cana-5754	61	30	and	and	CCONJ
cana-5754	61	31	feature	feature	NOUN
cana-5754	61	32	discrimination	discrimination	NOUN
cana-5754	61	33	,	,	PUNCT
cana-5754	61	34	especially	especially	ADV
cana-5754	61	35	for	for	ADP
cana-5754	61	36	minority	minority	NOUN
cana-5754	61	37	severe	severe	ADJ
cana-5754	61	38	categories	category	NOUN
cana-5754	61	39	.	.	PUNCT
cana-5754	62	1	communications	communication	NOUN
cana-5754	62	2	on	on	ADP
cana-5754	62	3	applied	apply	VERB
cana-5754	62	4	nonlinear	nonlinear	ADJ
cana-5754	62	5	analysis	analysis	NOUN
cana-5754	62	6	issn	issn	NOUN
cana-5754	62	7	:	:	PUNCT
cana-5754	62	8	1074	1074	NUM
cana-5754	62	9	-	-	PUNCT
cana-5754	62	10	133x	133x	NUM
cana-5754	62	11	vol	vol	NOUN
cana-5754	62	12	32	32	NUM
cana-5754	62	13	no	no	NOUN
cana-5754	62	14	.	.	PUNCT
cana-5754	63	1	3s	3s	NUM
cana-5754	63	2	(	(	PUNCT
cana-5754	63	3	2025	2025	NUM
cana-5754	63	4	)	)	PUNCT
cana-5754	63	5	784	784	NUM
cana-5754	63	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-5754	63	7	confusion	confusion	NOUN
cana-5754	63	8	matrix	matrix	NOUN
cana-5754	63	9	:	:	PUNCT
cana-5754	63	10	the	the	DET
cana-5754	63	11	confusion	confusion	NOUN
cana-5754	63	12	matrix	matrix	NOUN
cana-5754	63	13	highlights	highlight	VERB
cana-5754	63	14	the	the	DET
cana-5754	63	15	performance	performance	NOUN
cana-5754	63	16	of	of	ADP
cana-5754	63	17	a	a	DET
cana-5754	63	18	cnn	cnn	PROPN
cana-5754	63	19	-	-	PUNCT
cana-5754	63	20	mobilenet	mobilenet	ADJ
cana-5754	63	21	model	model	NOUN
cana-5754	63	22	in	in	ADP
cana-5754	63	23	detecting	detect	VERB
cana-5754	63	24	diabetic	diabetic	ADJ
cana-5754	63	25	retinopathy	retinopathy	ADJ
cana-5754	63	26	stages	stage	NOUN
cana-5754	63	27	.	.	PUNCT
cana-5754	64	1	the	the	DET
cana-5754	64	2	model	model	NOUN
cana-5754	64	3	classifies	classify	VERB
cana-5754	64	4	no_dr	no_dr	PROPN
cana-5754	64	5	cases	case	NOUN
cana-5754	64	6	very	very	ADV
cana-5754	64	7	accurately	accurately	ADV
cana-5754	64	8	with	with	ADP
cana-5754	64	9	340	340	NUM
cana-5754	64	10	out	out	ADP
cana-5754	64	11	of	of	ADP
cana-5754	64	12	361	361	NUM
cana-5754	64	13	correct	correct	ADJ
cana-5754	64	14	predictions	prediction	NOUN
cana-5754	64	15	,	,	PUNCT
cana-5754	64	16	indicating	indicate	VERB
cana-5754	64	17	high	high	ADJ
cana-5754	64	18	sensitivity	sensitivity	NOUN
cana-5754	64	19	to	to	ADP
cana-5754	64	20	healthy	healthy	ADJ
cana-5754	64	21	retinas	retina	NOUN
cana-5754	64	22	.	.	PUNCT
cana-5754	65	1	however	however	ADV
cana-5754	65	2	,	,	PUNCT
cana-5754	65	3	performance	performance	NOUN
cana-5754	65	4	on	on	ADP
cana-5754	65	5	other	other	ADJ
cana-5754	65	6	classes	class	NOUN
cana-5754	65	7	,	,	PUNCT
cana-5754	65	8	especially	especially	ADV
cana-5754	65	9	moderate	moderate	ADJ
cana-5754	65	10	,	,	PUNCT
cana-5754	65	11	severe	severe	ADJ
cana-5754	65	12	,	,	PUNCT
cana-5754	65	13	and	and	CCONJ
cana-5754	65	14	proliferate_dr	proliferate_dr	NOUN
cana-5754	65	15	,	,	PUNCT
cana-5754	65	16	shows	show	VERB
cana-5754	65	17	significant	significant	ADJ
cana-5754	65	18	confusion	confusion	NOUN
cana-5754	65	19	.	.	PUNCT
cana-5754	66	1	for	for	ADP
cana-5754	66	2	example	example	NOUN
cana-5754	66	3	,	,	PUNCT
cana-5754	66	4	only	only	ADV
cana-5754	66	5	88	88	NUM
cana-5754	66	6	out	out	ADP
cana-5754	66	7	of	of	ADP
cana-5754	66	8	199	199	NUM
cana-5754	66	9	moderate	moderate	ADJ
cana-5754	66	10	cases	case	NOUN
cana-5754	66	11	were	be	AUX
cana-5754	66	12	correctly	correctly	ADV
cana-5754	66	13	identified	identify	VERB
cana-5754	66	14	,	,	PUNCT
cana-5754	66	15	with	with	ADP
cana-5754	66	16	many	many	ADJ
cana-5754	66	17	misclassified	misclassified	ADJ
cana-5754	66	18	as	as	ADP
cana-5754	66	19	mild	mild	ADJ
cana-5754	66	20	or	or	CCONJ
cana-5754	66	21	severe	severe	ADJ
cana-5754	66	22	.	.	PUNCT
cana-5754	67	1	severe	severe	ADJ
cana-5754	67	2	and	and	CCONJ
cana-5754	67	3	proliferate_dr	proliferate_dr	NOUN
cana-5754	67	4	stages	stage	NOUN
cana-5754	67	5	suffer	suffer	VERB
cana-5754	67	6	from	from	ADP
cana-5754	67	7	high	high	ADJ
cana-5754	67	8	misclassification	misclassification	NOUN
cana-5754	67	9	,	,	PUNCT
cana-5754	67	10	often	often	ADV
cana-5754	67	11	confused	confuse	VERB
cana-5754	67	12	with	with	ADP
cana-5754	67	13	moderate	moderate	ADJ
cana-5754	67	14	or	or	CCONJ
cana-5754	67	15	mild	mild	ADJ
cana-5754	67	16	,	,	PUNCT
cana-5754	67	17	suggesting	suggest	VERB
cana-5754	67	18	poor	poor	ADJ
cana-5754	67	19	class	class	NOUN
cana-5754	67	20	separation	separation	NOUN
cana-5754	67	21	.	.	PUNCT
cana-5754	68	1	these	these	DET
cana-5754	68	2	results	result	NOUN
cana-5754	68	3	underline	underline	VERB
cana-5754	68	4	the	the	DET
cana-5754	68	5	need	need	NOUN
cana-5754	68	6	for	for	ADP
cana-5754	68	7	data	data	NOUN
cana-5754	68	8	balancing	balancing	NOUN
cana-5754	68	9	and	and	CCONJ
cana-5754	68	10	enhanced	enhance	VERB
cana-5754	68	11	feature	feature	NOUN
cana-5754	68	12	learning	learning	NOUN
cana-5754	68	13	for	for	ADP
cana-5754	68	14	minority	minority	NOUN
cana-5754	68	15	and	and	CCONJ
cana-5754	68	16	advanced	advanced	ADJ
cana-5754	68	17	disease	disease	NOUN
cana-5754	68	18	stages	stage	NOUN
cana-5754	68	19	.	.	PUNCT
cana-5754	69	1	precision	precision	NOUN
cana-5754	69	2	-	-	PUNCT
cana-5754	69	3	recall	recall	NOUN
cana-5754	69	4	(	(	PUNCT
cana-5754	69	5	pr	pr	NOUN
cana-5754	69	6	)	)	PUNCT
cana-5754	69	7	curve	curve	NOUN
cana-5754	69	8	:	:	PUNCT
cana-5754	69	9	the	the	DET
cana-5754	69	10	precision	precision	NOUN
cana-5754	69	11	-	-	PUNCT
cana-5754	69	12	recall	recall	NOUN
cana-5754	69	13	(	(	PUNCT
cana-5754	69	14	pr	pr	NOUN
cana-5754	69	15	)	)	PUNCT
cana-5754	69	16	curve	curve	NOUN
cana-5754	69	17	visualizes	visualize	VERB
cana-5754	69	18	the	the	DET
cana-5754	69	19	classification	classification	NOUN
cana-5754	69	20	performance	performance	NOUN
cana-5754	69	21	of	of	ADP
cana-5754	69	22	the	the	DET
cana-5754	69	23	cnnmobilenet	cnnmobilenet	ADJ
cana-5754	69	24	model	model	NOUN
cana-5754	69	25	for	for	ADP
cana-5754	69	26	different	different	ADJ
cana-5754	69	27	diabetic	diabetic	ADJ
cana-5754	69	28	retinopathy	retinopathy	ADJ
cana-5754	69	29	stages	stage	NOUN
cana-5754	69	30	.	.	PUNCT
cana-5754	70	1	the	the	DET
cana-5754	70	2	no_dr	no_dr	PROPN
cana-5754	70	3	class	class	NOUN
cana-5754	70	4	(	(	PUNCT
cana-5754	70	5	blue	blue	ADJ
cana-5754	70	6	curve	curve	NOUN
cana-5754	70	7	)	)	PUNCT
cana-5754	70	8	achieves	achieve	VERB
cana-5754	70	9	consistently	consistently	ADV
cana-5754	70	10	high	high	ADJ
cana-5754	70	11	precision	precision	NOUN
cana-5754	70	12	and	and	CCONJ
cana-5754	70	13	recall	recall	NOUN
cana-5754	70	14	,	,	PUNCT
cana-5754	70	15	indicating	indicate	VERB
cana-5754	70	16	excellent	excellent	ADJ
cana-5754	70	17	classification	classification	NOUN
cana-5754	70	18	reliability	reliability	NOUN
cana-5754	70	19	for	for	ADP
cana-5754	70	20	healthy	healthy	ADJ
cana-5754	70	21	cases	case	NOUN
cana-5754	70	22	.	.	PUNCT
cana-5754	71	1	in	in	ADP
cana-5754	71	2	contrast	contrast	NOUN
cana-5754	71	3	,	,	PUNCT
cana-5754	71	4	classes	class	NOUN
cana-5754	71	5	like	like	ADP
cana-5754	71	6	proliferate_dr	proliferate_dr	NOUN
cana-5754	71	7	(	(	PUNCT
cana-5754	71	8	red	red	ADJ
cana-5754	71	9	)	)	PUNCT
cana-5754	71	10	and	and	CCONJ
cana-5754	71	11	severe	severe	ADJ
cana-5754	71	12	(	(	PUNCT
cana-5754	71	13	purple	purple	ADJ
cana-5754	71	14	)	)	PUNCT
cana-5754	71	15	exhibit	exhibit	VERB
cana-5754	71	16	low	low	ADJ
cana-5754	71	17	and	and	CCONJ
cana-5754	71	18	unstable	unstable	ADJ
cana-5754	71	19	precision	precision	NOUN
cana-5754	71	20	across	across	ADP
cana-5754	71	21	all	all	DET
cana-5754	71	22	recall	recall	NOUN
cana-5754	71	23	levels	level	NOUN
cana-5754	71	24	,	,	PUNCT
cana-5754	71	25	reflecting	reflect	VERB
cana-5754	71	26	weak	weak	ADJ
cana-5754	71	27	model	model	NOUN
cana-5754	71	28	performance	performance	NOUN
cana-5754	71	29	for	for	ADP
cana-5754	71	30	advanced	advanced	ADJ
cana-5754	71	31	communications	communication	NOUN
cana-5754	71	32	on	on	ADP
cana-5754	71	33	applied	apply	VERB
cana-5754	71	34	nonlinear	nonlinear	ADJ
cana-5754	71	35	analysis	analysis	NOUN
cana-5754	71	36	issn	issn	NOUN
cana-5754	71	37	:	:	PUNCT
cana-5754	71	38	1074	1074	NUM
cana-5754	71	39	-	-	PUNCT
cana-5754	71	40	133x	133x	NUM
cana-5754	71	41	vol	vol	NOUN
cana-5754	71	42	32	32	NUM
cana-5754	71	43	no	no	NOUN
cana-5754	71	44	.	.	PUNCT
cana-5754	72	1	3s	3s	NUM
cana-5754	72	2	(	(	PUNCT
cana-5754	72	3	2025	2025	NUM
cana-5754	72	4	)	)	PUNCT
cana-5754	72	5	785	785	NUM
cana-5754	72	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-5754	72	7	dr	dr	NOUN
cana-5754	72	8	stages	stage	NOUN
cana-5754	72	9	.	.	PUNCT
cana-5754	73	1	the	the	DET
cana-5754	73	2	moderate	moderate	ADJ
cana-5754	73	3	and	and	CCONJ
cana-5754	73	4	mild	mild	ADJ
cana-5754	73	5	stages	stage	NOUN
cana-5754	73	6	show	show	VERB
cana-5754	73	7	moderate	moderate	ADJ
cana-5754	73	8	pr	pr	NOUN
cana-5754	73	9	balance	balance	NOUN
cana-5754	73	10	but	but	CCONJ
cana-5754	73	11	still	still	ADV
cana-5754	73	12	suffer	suffer	VERB
cana-5754	73	13	from	from	ADP
cana-5754	73	14	gradual	gradual	ADJ
cana-5754	73	15	precision	precision	NOUN
cana-5754	73	16	decline	decline	NOUN
cana-5754	73	17	as	as	ADP
cana-5754	73	18	recall	recall	NOUN
cana-5754	73	19	increases	increase	NOUN
cana-5754	73	20	.	.	PUNCT
cana-5754	74	1	this	this	DET
cana-5754	74	2	imbalance	imbalance	NOUN
cana-5754	74	3	highlights	highlight	VERB
cana-5754	74	4	the	the	DET
cana-5754	74	5	model	model	NOUN
cana-5754	74	6	's	's	PART
cana-5754	74	7	difficulty	difficulty	NOUN
cana-5754	74	8	in	in	ADP
cana-5754	74	9	discriminating	discriminate	VERB
cana-5754	74	10	between	between	ADP
cana-5754	74	11	pathological	pathological	ADJ
cana-5754	74	12	stages	stage	NOUN
cana-5754	74	13	and	and	CCONJ
cana-5754	74	14	emphasizes	emphasize	VERB
cana-5754	74	15	the	the	DET
cana-5754	74	16	need	need	NOUN
cana-5754	74	17	for	for	ADP
cana-5754	74	18	further	further	ADJ
cana-5754	74	19	training	training	NOUN
cana-5754	74	20	refinement	refinement	NOUN
cana-5754	74	21	and	and	CCONJ
cana-5754	74	22	class	class	NOUN
cana-5754	74	23	balancing	balancing	NOUN
cana-5754	74	24	.	.	PUNCT
cana-5754	75	1	roc	roc	PROPN
cana-5754	75	2	curve	curve	PROPN
cana-5754	75	3	:	:	PUNCT
cana-5754	75	4	the	the	DET
cana-5754	75	5	roc	roc	PROPN
cana-5754	75	6	curve	curve	NOUN
cana-5754	75	7	illustrates	illustrate	VERB
cana-5754	75	8	the	the	DET
cana-5754	75	9	discriminative	discriminative	NOUN
cana-5754	75	10	ability	ability	NOUN
cana-5754	75	11	of	of	ADP
cana-5754	75	12	the	the	DET
cana-5754	75	13	cnn	cnn	PROPN
cana-5754	75	14	-	-	PUNCT
cana-5754	75	15	mobilenet	mobilenet	PROPN
cana-5754	75	16	model	model	NOUN
cana-5754	75	17	across	across	ADP
cana-5754	75	18	all	all	DET
cana-5754	75	19	diabetic	diabetic	ADJ
cana-5754	75	20	retinopathy	retinopathy	ADJ
cana-5754	75	21	classes	class	NOUN
cana-5754	75	22	.	.	PUNCT
cana-5754	76	1	the	the	DET
cana-5754	76	2	no_dr	no_dr	PROPN
cana-5754	76	3	class	class	NOUN
cana-5754	76	4	demonstrates	demonstrate	VERB
cana-5754	76	5	superior	superior	ADJ
cana-5754	76	6	performance	performance	NOUN
cana-5754	76	7	with	with	ADP
cana-5754	76	8	an	an	DET
cana-5754	76	9	auc	auc	NOUN
cana-5754	76	10	of	of	ADP
cana-5754	76	11	0.98	0.98	NUM
cana-5754	76	12	,	,	PUNCT
cana-5754	76	13	indicating	indicate	VERB
cana-5754	76	14	excellent	excellent	ADJ
cana-5754	76	15	separation	separation	NOUN
cana-5754	76	16	between	between	ADP
cana-5754	76	17	healthy	healthy	ADJ
cana-5754	76	18	and	and	CCONJ
cana-5754	76	19	diseased	diseased	ADJ
cana-5754	76	20	cases	case	NOUN
cana-5754	76	21	.	.	PUNCT
cana-5754	77	1	the	the	DET
cana-5754	77	2	remaining	remain	VERB
cana-5754	77	3	stages	stage	NOUN
cana-5754	77	4	—	—	PUNCT
cana-5754	77	5	mild	mild	ADJ
cana-5754	77	6	(	(	PUNCT
cana-5754	77	7	0.88	0.88	NUM
cana-5754	77	8	)	)	PUNCT
cana-5754	77	9	,	,	PUNCT
cana-5754	77	10	moderate	moderate	ADJ
cana-5754	77	11	(	(	PUNCT
cana-5754	77	12	0.87	0.87	NUM
cana-5754	77	13	)	)	PUNCT
cana-5754	77	14	,	,	PUNCT
cana-5754	77	15	severe	severe	ADJ
cana-5754	77	16	(	(	PUNCT
cana-5754	77	17	0.88	0.88	NUM
cana-5754	77	18	)	)	PUNCT
cana-5754	77	19	,	,	PUNCT
cana-5754	77	20	and	and	CCONJ
cana-5754	77	21	proliferate_dr	proliferate_dr	NOUN
cana-5754	77	22	(	(	PUNCT
cana-5754	77	23	0.87	0.87	NUM
cana-5754	77	24	)	)	PUNCT
cana-5754	77	25	—	—	PUNCT
cana-5754	77	26	show	show	VERB
cana-5754	77	27	strong	strong	ADJ
cana-5754	77	28	but	but	CCONJ
cana-5754	77	29	slightly	slightly	ADV
cana-5754	77	30	lower	low	ADJ
cana-5754	77	31	aucs	aucs	NOUN
cana-5754	77	32	,	,	PUNCT
cana-5754	77	33	suggesting	suggest	VERB
cana-5754	77	34	competent	competent	ADJ
cana-5754	77	35	yet	yet	CCONJ
cana-5754	77	36	imperfect	imperfect	ADJ
cana-5754	77	37	class	class	NOUN
cana-5754	77	38	separation	separation	NOUN
cana-5754	77	39	.	.	PUNCT
cana-5754	78	1	the	the	DET
cana-5754	78	2	proximity	proximity	NOUN
cana-5754	78	3	of	of	ADP
cana-5754	78	4	these	these	DET
cana-5754	78	5	auc	auc	ADJ
cana-5754	78	6	values	value	NOUN
cana-5754	78	7	indicates	indicate	VERB
cana-5754	78	8	the	the	DET
cana-5754	78	9	model	model	NOUN
cana-5754	78	10	is	be	AUX
cana-5754	78	11	fairly	fairly	ADV
cana-5754	78	12	consistent	consistent	ADJ
cana-5754	78	13	across	across	ADP
cana-5754	78	14	stages	stage	NOUN
cana-5754	78	15	,	,	PUNCT
cana-5754	78	16	though	though	SCONJ
cana-5754	78	17	it	it	PRON
cana-5754	78	18	performs	perform	VERB
cana-5754	78	19	best	good	ADJ
cana-5754	78	20	in	in	ADP
cana-5754	78	21	detecting	detect	VERB
cana-5754	78	22	the	the	DET
cana-5754	78	23	absence	absence	NOUN
cana-5754	78	24	of	of	ADP
cana-5754	78	25	disease	disease	NOUN
cana-5754	78	26	.	.	PUNCT
cana-5754	79	1	this	this	DET
cana-5754	79	2	performance	performance	NOUN
cana-5754	79	3	highlights	highlight	VERB
cana-5754	79	4	the	the	DET
cana-5754	79	5	model	model	NOUN
cana-5754	79	6	’s	’s	PART
cana-5754	79	7	robustness	robustness	NOUN
cana-5754	79	8	but	but	CCONJ
cana-5754	79	9	also	also	ADV
cana-5754	79	10	suggests	suggest	VERB
cana-5754	79	11	a	a	DET
cana-5754	79	12	need	need	NOUN
cana-5754	79	13	for	for	ADP
cana-5754	79	14	improvement	improvement	NOUN
cana-5754	79	15	in	in	ADP
cana-5754	79	16	distinguishing	distinguish	VERB
cana-5754	79	17	between	between	ADP
cana-5754	79	18	more	more	ADV
cana-5754	79	19	advanced	advanced	ADJ
cana-5754	79	20	dr	dr	NOUN
cana-5754	79	21	stages	stage	NOUN
cana-5754	79	22	.	.	PUNCT
cana-5754	80	1	training	training	NOUN
cana-5754	80	2	vs	vs	ADP
cana-5754	80	3	validation	validation	NOUN
cana-5754	80	4	accuracy	accuracy	NOUN
cana-5754	80	5	analysis	analysis	NOUN
cana-5754	80	6	communications	communication	NOUN
cana-5754	80	7	on	on	ADP
cana-5754	80	8	applied	apply	VERB
cana-5754	80	9	nonlinear	nonlinear	ADJ
cana-5754	80	10	analysis	analysis	NOUN
cana-5754	80	11	issn	issn	NOUN
cana-5754	80	12	:	:	PUNCT
cana-5754	80	13	1074	1074	NUM
cana-5754	80	14	-	-	PUNCT
cana-5754	80	15	133x	133x	NUM
cana-5754	80	16	vol	vol	NOUN
cana-5754	80	17	32	32	NUM
cana-5754	80	18	no	no	NOUN
cana-5754	80	19	.	.	PUNCT
cana-5754	81	1	3s	3s	NUM
cana-5754	81	2	(	(	PUNCT
cana-5754	81	3	2025	2025	NUM
cana-5754	81	4	)	)	PUNCT
cana-5754	81	5	786	786	NUM
cana-5754	81	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-5754	81	7	the	the	DET
cana-5754	81	8	plot	plot	NOUN
cana-5754	81	9	demonstrates	demonstrate	VERB
cana-5754	81	10	the	the	DET
cana-5754	81	11	progression	progression	NOUN
cana-5754	81	12	of	of	ADP
cana-5754	81	13	training	training	NOUN
cana-5754	81	14	and	and	CCONJ
cana-5754	81	15	validation	validation	NOUN
cana-5754	81	16	accuracy	accuracy	NOUN
cana-5754	81	17	across	across	ADP
cana-5754	81	18	14	14	NUM
cana-5754	81	19	epochs	epoch	NOUN
cana-5754	81	20	for	for	ADP
cana-5754	81	21	the	the	DET
cana-5754	81	22	cnn	cnn	PROPN
cana-5754	81	23	-	-	PUNCT
cana-5754	81	24	mobilenet	mobilenet	PROPN
cana-5754	81	25	model	model	NOUN
cana-5754	81	26	.	.	PUNCT
cana-5754	82	1	initially	initially	ADV
cana-5754	82	2	,	,	PUNCT
cana-5754	82	3	both	both	DET
cana-5754	82	4	accuracies	accuracy	NOUN
cana-5754	82	5	rise	rise	VERB
cana-5754	82	6	rapidly	rapidly	ADV
cana-5754	82	7	,	,	PUNCT
cana-5754	82	8	indicating	indicate	VERB
cana-5754	82	9	effective	effective	ADJ
cana-5754	82	10	learning	learning	NOUN
cana-5754	82	11	.	.	PUNCT
cana-5754	83	1	the	the	DET
cana-5754	83	2	validation	validation	NOUN
cana-5754	83	3	accuracy	accuracy	NOUN
cana-5754	83	4	consistently	consistently	ADV
cana-5754	83	5	outperforms	outperform	VERB
cana-5754	83	6	the	the	DET
cana-5754	83	7	training	training	NOUN
cana-5754	83	8	accuracy	accuracy	NOUN
cana-5754	83	9	until	until	ADP
cana-5754	83	10	around	around	ADP
cana-5754	83	11	epoch	epoch	NOUN
cana-5754	83	12	11	11	NUM
cana-5754	83	13	,	,	PUNCT
cana-5754	83	14	suggesting	suggest	VERB
cana-5754	83	15	that	that	SCONJ
cana-5754	83	16	the	the	DET
cana-5754	83	17	model	model	NOUN
cana-5754	83	18	generalizes	generalize	VERB
cana-5754	83	19	well	well	ADV
cana-5754	83	20	and	and	CCONJ
cana-5754	83	21	does	do	AUX
cana-5754	83	22	not	not	PART
cana-5754	83	23	suffer	suffer	VERB
cana-5754	83	24	from	from	ADP
cana-5754	83	25	early	early	ADJ
cana-5754	83	26	overfitting	overfitting	NOUN
cana-5754	83	27	.	.	PUNCT
cana-5754	84	1	post	post	VERB
cana-5754	84	2	epoch	epoch	PROPN
cana-5754	84	3	11	11	NUM
cana-5754	84	4	,	,	PUNCT
cana-5754	84	5	the	the	DET
cana-5754	84	6	curves	curve	NOUN
cana-5754	84	7	converge	converge	VERB
cana-5754	84	8	,	,	PUNCT
cana-5754	84	9	and	and	CCONJ
cana-5754	84	10	training	training	NOUN
cana-5754	84	11	accuracy	accuracy	NOUN
cana-5754	84	12	slightly	slightly	ADV
cana-5754	84	13	dips	dip	VERB
cana-5754	84	14	at	at	ADP
cana-5754	84	15	the	the	DET
cana-5754	84	16	final	final	ADJ
cana-5754	84	17	epoch	epoch	NOUN
cana-5754	84	18	,	,	PUNCT
cana-5754	84	19	while	while	SCONJ
cana-5754	84	20	validation	validation	NOUN
cana-5754	84	21	accuracy	accuracy	NOUN
cana-5754	84	22	stabilizes	stabilize	VERB
cana-5754	84	23	around	around	ADV
cana-5754	84	24	0.69	0.69	NUM
cana-5754	84	25	,	,	PUNCT
cana-5754	84	26	showing	show	VERB
cana-5754	84	27	relatively	relatively	ADV
cana-5754	84	28	consistent	consistent	ADJ
cana-5754	84	29	model	model	NOUN
cana-5754	84	30	behavior	behavior	NOUN
cana-5754	84	31	.	.	PUNCT
cana-5754	85	1	overall	overall	ADV
cana-5754	85	2	,	,	PUNCT
cana-5754	85	3	the	the	DET
cana-5754	85	4	model	model	NOUN
cana-5754	85	5	exhibits	exhibit	VERB
cana-5754	85	6	a	a	DET
cana-5754	85	7	steady	steady	ADJ
cana-5754	85	8	learning	learning	NOUN
cana-5754	85	9	trajectory	trajectory	NOUN
cana-5754	85	10	,	,	PUNCT
cana-5754	85	11	with	with	ADP
cana-5754	85	12	minor	minor	ADJ
cana-5754	85	13	fluctuations	fluctuation	NOUN
cana-5754	85	14	and	and	CCONJ
cana-5754	85	15	no	no	DET
cana-5754	85	16	significant	significant	ADJ
cana-5754	85	17	overfitting	overfitting	NOUN
cana-5754	85	18	,	,	PUNCT
cana-5754	85	19	reflecting	reflect	VERB
cana-5754	85	20	good	good	ADJ
cana-5754	85	21	training	training	NOUN
cana-5754	85	22	stability	stability	NOUN
cana-5754	85	23	and	and	CCONJ
cana-5754	85	24	generalization	generalization	NOUN
cana-5754	85	25	capability	capability	NOUN
cana-5754	85	26	.	.	PUNCT
cana-5754	86	1	training	training	NOUN
cana-5754	86	2	vs	vs	ADP
cana-5754	86	3	validation	validation	NOUN
cana-5754	86	4	loss	loss	NOUN
cana-5754	86	5	analysis	analysis	NOUN
cana-5754	86	6	:	:	PUNCT
cana-5754	86	7	this	this	DET
cana-5754	86	8	plot	plot	NOUN
cana-5754	86	9	illustrates	illustrate	VERB
cana-5754	86	10	the	the	DET
cana-5754	86	11	comparison	comparison	NOUN
cana-5754	86	12	between	between	ADP
cana-5754	86	13	training	training	NOUN
cana-5754	86	14	and	and	CCONJ
cana-5754	86	15	validation	validation	NOUN
cana-5754	86	16	loss	loss	NOUN
cana-5754	86	17	over	over	ADP
cana-5754	86	18	14	14	NUM
cana-5754	86	19	epochs	epoch	NOUN
cana-5754	86	20	.	.	PUNCT
cana-5754	87	1	the	the	DET
cana-5754	87	2	training	training	NOUN
cana-5754	87	3	loss	loss	NOUN
cana-5754	87	4	shows	show	VERB
cana-5754	87	5	a	a	DET
cana-5754	87	6	steady	steady	ADJ
cana-5754	87	7	downward	downward	ADJ
cana-5754	87	8	trend	trend	NOUN
cana-5754	87	9	,	,	PUNCT
cana-5754	87	10	indicating	indicate	VERB
cana-5754	87	11	that	that	SCONJ
cana-5754	87	12	the	the	DET
cana-5754	87	13	model	model	NOUN
cana-5754	87	14	is	be	AUX
cana-5754	87	15	learning	learn	VERB
cana-5754	87	16	and	and	CCONJ
cana-5754	87	17	optimizing	optimize	VERB
cana-5754	87	18	effectively	effectively	ADV
cana-5754	87	19	on	on	ADP
cana-5754	87	20	the	the	DET
cana-5754	87	21	training	training	NOUN
cana-5754	87	22	data	datum	NOUN
cana-5754	87	23	.	.	PUNCT
cana-5754	88	1	the	the	DET
cana-5754	88	2	validation	validation	NOUN
cana-5754	88	3	loss	loss	NOUN
cana-5754	88	4	also	also	ADV
cana-5754	88	5	decreases	decrease	VERB
cana-5754	88	6	and	and	CCONJ
cana-5754	88	7	stabilizes	stabilize	VERB
cana-5754	88	8	around	around	ADP
cana-5754	88	9	0.78	0.78	NUM
cana-5754	88	10	,	,	PUNCT
cana-5754	88	11	reflecting	reflect	VERB
cana-5754	88	12	improved	improved	ADJ
cana-5754	88	13	generalization	generalization	NOUN
cana-5754	88	14	and	and	CCONJ
cana-5754	88	15	reduced	reduce	VERB
cana-5754	88	16	error	error	NOUN
cana-5754	88	17	on	on	ADP
cana-5754	88	18	unseen	unseen	ADJ
cana-5754	88	19	data	datum	NOUN
cana-5754	88	20	.	.	PUNCT
cana-5754	89	1	interestingly	interestingly	ADV
cana-5754	89	2	,	,	PUNCT
cana-5754	89	3	the	the	DET
cana-5754	89	4	validation	validation	NOUN
cana-5754	89	5	loss	loss	NOUN
cana-5754	89	6	remains	remain	VERB
cana-5754	89	7	consistently	consistently	ADV
cana-5754	89	8	lower	low	ADJ
cana-5754	89	9	than	than	ADP
cana-5754	89	10	the	the	DET
cana-5754	89	11	training	training	NOUN
cana-5754	89	12	loss	loss	NOUN
cana-5754	89	13	,	,	PUNCT
cana-5754	89	14	which	which	PRON
cana-5754	89	15	may	may	AUX
cana-5754	89	16	suggest	suggest	VERB
cana-5754	89	17	regularization	regularization	NOUN
cana-5754	89	18	effectiveness	effectiveness	NOUN
cana-5754	89	19	or	or	CCONJ
cana-5754	89	20	early	early	ADJ
cana-5754	89	21	stopping	stopping	NOUN
cana-5754	89	22	techniques	technique	NOUN
cana-5754	89	23	.	.	PUNCT
cana-5754	90	1	the	the	DET
cana-5754	90	2	lack	lack	NOUN
cana-5754	90	3	of	of	ADP
cana-5754	90	4	divergence	divergence	NOUN
cana-5754	90	5	between	between	ADP
cana-5754	90	6	the	the	DET
cana-5754	90	7	two	two	NUM
cana-5754	90	8	curves	curve	NOUN
cana-5754	90	9	indicates	indicate	VERB
cana-5754	90	10	minimal	minimal	ADJ
cana-5754	90	11	overfitting	overfitting	NOUN
cana-5754	90	12	,	,	PUNCT
cana-5754	90	13	and	and	CCONJ
cana-5754	90	14	the	the	DET
cana-5754	90	15	model	model	NOUN
cana-5754	90	16	appears	appear	VERB
cana-5754	90	17	well	well	ADV
cana-5754	90	18	-	-	PUNCT
cana-5754	90	19	balanced	balanced	ADJ
cana-5754	90	20	in	in	ADP
cana-5754	90	21	learning	learn	VERB
cana-5754	90	22	from	from	ADP
cana-5754	90	23	both	both	DET
cana-5754	90	24	training	training	NOUN
cana-5754	90	25	and	and	CCONJ
cana-5754	90	26	validation	validation	NOUN
cana-5754	90	27	sets	set	NOUN
cana-5754	90	28	.	.	PUNCT
cana-5754	91	1	6	6	X
cana-5754	91	2	.	.	X
cana-5754	91	3	conclusions	conclusion	NOUN
cana-5754	91	4	and	and	CCONJ
cana-5754	91	5	future	future	ADJ
cana-5754	91	6	work	work	NOUN
cana-5754	91	7	this	this	DET
cana-5754	91	8	study	study	NOUN
cana-5754	91	9	illustrates	illustrate	VERB
cana-5754	91	10	a	a	DET
cana-5754	91	11	hybrid	hybrid	ADJ
cana-5754	91	12	approach	approach	NOUN
cana-5754	91	13	to	to	PART
cana-5754	91	14	recognize	recognize	VERB
cana-5754	91	15	early	early	ADJ
cana-5754	91	16	diabetes	diabetes	NOUN
cana-5754	91	17	retinopathy	retinopathy	ADJ
cana-5754	91	18	by	by	ADP
cana-5754	91	19	extracting	extract	VERB
cana-5754	91	20	features	feature	NOUN
cana-5754	91	21	from	from	ADP
cana-5754	91	22	fundus	fundus	NOUN
cana-5754	91	23	images	image	NOUN
cana-5754	91	24	by	by	ADP
cana-5754	91	25	sending	send	VERB
cana-5754	91	26	reset	reset	NOUN
cana-5754	91	27	16	16	NUM
cana-5754	91	28	and	and	CCONJ
cana-5754	91	29	google	google	PROPN
cana-5754	91	30	models	model	NOUN
cana-5754	91	31	based	base	VERB
cana-5754	91	32	on	on	ADP
cana-5754	91	33	adaptive	adaptive	ADJ
cana-5754	91	34	particle	particle	NOUN
cana-5754	91	35	swarm	swarm	NOUN
cana-5754	91	36	optimization	optimization	NOUN
cana-5754	91	37	.	.	PUNCT
cana-5754	92	1	these	these	DET
cana-5754	92	2	functions	function	NOUN
cana-5754	92	3	are	be	AUX
cana-5754	92	4	entered	enter	VERB
cana-5754	92	5	into	into	ADP
cana-5754	92	6	ipax	ipax	NOUN
cana-5754	92	7	records	record	NOUN
cana-5754	92	8	of	of	ADP
cana-5754	92	9	various	various	ADJ
cana-5754	92	10	classifiers	classifier	NOUN
cana-5754	92	11	for	for	ADP
cana-5754	92	12	classification	classification	NOUN
cana-5754	92	13	of	of	ADP
cana-5754	92	14	diabetic	diabetic	ADJ
cana-5754	92	15	patients	patient	NOUN
cana-5754	92	16	from	from	ADP
cana-5754	92	17	retinopathy	retinopathy	ADJ
cana-5754	92	18	with	with	ADP
cana-5754	92	19	several	several	ADJ
cana-5754	92	20	classes	class	NOUN
cana-5754	92	21	the	the	DET
cana-5754	92	22	combination	combination	NOUN
cana-5754	92	23	of	of	ADP
cana-5754	92	24	characteristics	characteristic	NOUN
cana-5754	92	25	of	of	ADP
cana-5754	92	26	the	the	PRON
cana-5754	92	27	modelsignificantly	modelsignificantly	ADV
cana-5754	92	28	increases	increase	VERB
cana-5754	92	29	the	the	DET
cana-5754	92	30	power	power	NOUN
cana-5754	92	31	metrics	metric	NOUN
cana-5754	92	32	for	for	ADP
cana-5754	92	33	classifying	classify	VERB
cana-5754	92	34	diabetes	diabetes	NOUN
cana-5754	92	35	disorders	disorder	NOUN
cana-5754	92	36	and	and	CCONJ
cana-5754	92	37	improves	improve	VERB
cana-5754	92	38	overall	overall	ADJ
cana-5754	92	39	system	system	NOUN
cana-5754	92	40	effectiveness.the	effectiveness.the	PRON
cana-5754	92	41	proposed	propose	VERB
cana-5754	92	42	technique	technique	NOUN
cana-5754	92	43	is	be	AUX
cana-5754	92	44	promised	promise	VERB
cana-5754	92	45	with	with	ADP
cana-5754	92	46	the	the	DET
cana-5754	92	47	help	help	NOUN
cana-5754	92	48	of	of	ADP
cana-5754	92	49	an	an	DET
cana-5754	92	50	ophthalmologist	ophthalmologist	NOUN
cana-5754	92	51	as	as	ADP
cana-5754	92	52	knowledge	knowledge	NOUN
cana-5754	92	53	of	of	ADP
cana-5754	92	54	early	early	ADJ
cana-5754	92	55	diabetic	diabetic	ADJ
cana-5754	92	56	retinopathy	retinopathy	NOUN
cana-5754	92	57	.	.	PUNCT
cana-5754	93	1	results	result	NOUN
cana-5754	93	2	highlights	highlight	VERB
cana-5754	93	3	the	the	DET
cana-5754	93	4	effectiveness	effectiveness	NOUN
cana-5754	93	5	of	of	ADP
cana-5754	93	6	the	the	DET
cana-5754	93	7	combination	combination	NOUN
cana-5754	93	8	of	of	ADP
cana-5754	93	9	cnn	cnn	PROPN
cana-5754	93	10	feature	feature	NOUN
cana-5754	93	11	extraction	extraction	NOUN
cana-5754	93	12	and	and	CCONJ
cana-5754	93	13	classifiers	classifier	NOUN
cana-5754	93	14	for	for	ADP
cana-5754	93	15	machine	machine	NOUN
cana-5754	93	16	learning	learning	NOUN
cana-5754	93	17	,	,	PUNCT
cana-5754	93	18	providing	provide	VERB
cana-5754	93	19	quick	quick	ADJ
cana-5754	93	20	and	and	CCONJ
cana-5754	93	21	incredible	incredible	ADJ
cana-5754	93	22	results	result	NOUN
cana-5754	93	23	.	.	PUNCT
cana-5754	94	1	surprisingly	surprisingly	ADV
cana-5754	94	2	,	,	PUNCT
cana-5754	94	3	the	the	DET
cana-5754	94	4	hybrid	hybrid	ADJ
cana-5754	94	5	model	model	NOUN
cana-5754	94	6	with	with	ADP
cana-5754	94	7	svm	svm	PROPN
cana-5754	94	8	current	current	ADJ
cana-5754	94	9	and	and	CCONJ
cana-5754	94	10	multiclass	multiclass	ADJ
cana-5754	94	11	diabetic	diabetic	ADJ
cana-5754	94	12	retinopathy	retinopathy	NOUN
cana-5754	94	13	exceeds	exceed	VERB
cana-5754	94	14	the	the	DET
cana-5754	94	15	detection	detection	NOUN
cana-5754	94	16	technique	technique	NOUN
cana-5754	94	17	of	of	ADP
cana-5754	94	18	at	at	ADP
cana-5754	94	19	9	9	NUM
cana-5754	94	20	%	%	NOUN
cana-5754	94	21	with	with	ADP
cana-5754	94	22	an	an	DET
cana-5754	94	23	communications	communication	NOUN
cana-5754	94	24	on	on	ADP
cana-5754	94	25	applied	apply	VERB
cana-5754	94	26	nonlinear	nonlinear	ADJ
cana-5754	94	27	analysis	analysis	NOUN
cana-5754	94	28	issn	issn	NOUN
cana-5754	94	29	:	:	PUNCT
cana-5754	94	30	1074	1074	NUM
cana-5754	94	31	-	-	PUNCT
cana-5754	94	32	133x	133x	NUM
cana-5754	94	33	vol	vol	NOUN
cana-5754	94	34	32	32	NUM
cana-5754	94	35	no	no	NOUN
cana-5754	94	36	.	.	PUNCT
cana-5754	95	1	3s	3s	NUM
cana-5754	95	2	(	(	PUNCT
cana-5754	95	3	2025	2025	NUM
cana-5754	95	4	)	)	PUNCT
cana-5754	95	5	787	787	NUM
cana-5754	95	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-5754	95	7	astonishing	astonishing	ADJ
cana-5754	95	8	average	average	ADJ
cana-5754	95	9	accuracy	accuracy	NOUN
cana-5754	95	10	.	.	PUNCT
cana-5754	96	1	the	the	DET
cana-5754	96	2	results	result	NOUN
cana-5754	96	3	show	show	VERB
cana-5754	96	4	the	the	DET
cana-5754	96	5	efficiency	efficiency	NOUN
cana-5754	96	6	of	of	ADP
cana-5754	96	7	the	the	DET
cana-5754	96	8	combination	combination	NOUN
cana-5754	96	9	of	of	ADP
cana-5754	96	10	cnn	cnn	PROPN
cana-5754	96	11	feature	feature	NOUN
cana-5754	96	12	extraction	extraction	NOUN
cana-5754	96	13	for	for	ADP
cana-5754	96	14	machine	machine	NOUN
cana-5754	96	15	learning	learning	NOUN
cana-5754	96	16	,	,	PUNCT
cana-5754	96	17	classification	classification	NOUN
cana-5754	96	18	using	use	VERB
cana-5754	96	19	this	this	PRON
cana-5754	96	20	gives	give	VERB
cana-5754	96	21	you	you	PRON
cana-5754	96	22	quick	quick	ADJ
cana-5754	96	23	and	and	CCONJ
cana-5754	96	24	highly	highly	ADV
cana-5754	96	25	accurate	accurate	ADJ
cana-5754	96	26	results	result	NOUN
cana-5754	96	27	.	.	PUNCT
cana-5754	97	1	surprisingly	surprisingly	ADV
cana-5754	97	2	,	,	PUNCT
cana-5754	97	3	the	the	DET
cana-5754	97	4	average	average	ADJ
cana-5754	97	5	accuracy	accuracy	NOUN
cana-5754	97	6	of	of	ADP
cana-5754	97	7	the	the	DET
cana-5754	97	8	svm	svm	PROPN
cana-5754	97	9	hybrid	hybrid	NOUN
cana-5754	97	10	model	model	NOUN
cana-5754	97	11	is	be	AUX
cana-5754	97	12	i	i	PRON
cana-5754	97	13	look	look	VERB
cana-5754	97	14	forward	forward	ADV
cana-5754	97	15	to	to	ADP
cana-5754	97	16	testing	test	VERB
cana-5754	97	17	algorithms	algorithm	NOUN
cana-5754	97	18	for	for	ADP
cana-5754	97	19	diagnosis	diagnosis	NOUN
cana-5754	97	20	at	at	ADP
cana-5754	97	21	and	and	CCONJ
cana-5754	97	22	to	to	PART
cana-5754	97	23	successfully	successfully	ADV
cana-5754	97	24	remove	remove	VERB
cana-5754	97	25	noise	noise	NOUN
cana-5754	97	26	and	and	CCONJ
cana-5754	97	27	artifacts	artifact	NOUN
cana-5754	97	28	,	,	PUNCT
cana-5754	97	29	we	we	PRON
cana-5754	97	30	need	need	VERB
cana-5754	97	31	to	to	PART
cana-5754	97	32	focus	focus	VERB
cana-5754	97	33	on	on	ADP
cana-5754	97	34	improving	improve	VERB
cana-5754	97	35	strategies	strategy	NOUN
cana-5754	97	36	,	,	PUNCT
cana-5754	97	37	such	such	ADJ
cana-5754	97	38	as	as	ADP
cana-5754	97	39	data	data	NOUN
cana-5754	97	40	expansion	expansion	NOUN
cana-5754	97	41	and	and	CCONJ
cana-5754	97	42	other	other	ADJ
cana-5754	97	43	processing	processing	NOUN
cana-5754	97	44	approaches	approach	NOUN
cana-5754	97	45	.	.	PUNCT
cana-5754	98	1	this	this	DET
cana-5754	98	2	field	field	NOUN
cana-5754	98	3	is	be	AUX
cana-5754	98	4	also	also	ADV
cana-5754	98	5	important	important	ADJ
cana-5754	98	6	to	to	PART
cana-5754	98	7	improve	improve	VERB
cana-5754	98	8	and	and	CCONJ
cana-5754	98	9	achieve	achieve	VERB
cana-5754	98	10	the	the	PRON
cana-5754	98	11	with	with	ADP
cana-5754	98	12	continuous	continuous	ADJ
cana-5754	98	13	efforts	effort	NOUN
cana-5754	98	14	to	to	PART
cana-5754	98	15	improve	improve	VERB
cana-5754	98	16	and	and	CCONJ
cana-5754	98	17	achieve	achieve	VERB
cana-5754	98	18	the	the	DET
cana-5754	98	19	accuracy	accuracy	NOUN
cana-5754	98	20	of	of	ADP
cana-5754	98	21	the	the	PRON
cana-5754	98	22	and	and	CCONJ
cana-5754	98	23	the	the	DET
cana-5754	98	24	continued	continue	VERB
cana-5754	98	25	efforts	effort	NOUN
cana-5754	98	26	to	to	PART
cana-5754	98	27	achieve	achieve	VERB
cana-5754	98	28	the	the	DET
cana-5754	98	29	effectiveness	effectiveness	NOUN
cana-5754	98	30	of	of	ADP
cana-5754	98	31	the	the	DET
cana-5754	98	32	dr	dr	PROPN
cana-5754	98	33	recognition	recognition	PROPN
cana-5754	98	34	system	system	NOUN
cana-5754	98	35	.	.	PUNCT
cana-5754	99	1	the	the	DET
cana-5754	99	2	possibility	possibility	NOUN
cana-5754	99	3	that	that	PRON
cana-5754	99	4	may	may	AUX
cana-5754	99	5	change	change	VERB
cana-5754	99	6	the	the	DET
cana-5754	99	7	early	early	ADJ
cana-5754	99	8	detection	detection	NOUN
cana-5754	99	9	and	and	CCONJ
cana-5754	99	10	treatment	treatment	NOUN
cana-5754	99	11	of	of	ADP
cana-5754	99	12	diabetes	diabetes	NOUN
cana-5754	99	13	retinopathy	retinopathy	NOUN
cana-5754	99	14	is	be	AUX
cana-5754	99	15	working	work	VERB
cana-5754	99	16	together	together	ADV
cana-5754	99	17	,	,	PUNCT
cana-5754	99	18	especially	especially	ADV
cana-5754	99	19	when	when	SCONJ
cana-5754	99	20	technology	technology	NOUN
cana-5754	99	21	progresses	progress	VERB
cana-5754	99	22	.	.	PUNCT
cana-5754	100	1	future	future	ADJ
cana-5754	100	2	studies	study	NOUN
cana-5754	100	3	should	should	AUX
cana-5754	100	4	treat	treat	VERB
cana-5754	100	5	these	these	DET
cana-5754	100	6	possible	possible	ADJ
cana-5754	100	7	distortions	distortion	NOUN
cana-5754	100	8	in	in	ADP
cana-5754	100	9	of	of	ADP
cana-5754	100	10	the	the	DET
cana-5754	100	11	data	data	NOUN
cana-5754	100	12	record	record	NOUN
cana-5754	100	13	to	to	PART
cana-5754	100	14	improve	improve	VERB
cana-5754	100	15	the	the	DET
cana-5754	100	16	robustness	robustness	NOUN
cana-5754	100	17	of	of	ADP
cana-5754	100	18	the	the	DET
cana-5754	100	19	proposed	propose	VERB
cana-5754	100	20	model	model	NOUN
cana-5754	100	21	and	and	CCONJ
cana-5754	100	22	generalizability	generalizability	NOUN
cana-5754	100	23	of	of	ADP
cana-5754	100	24	numerous	numerous	ADJ
cana-5754	100	25	patient	patient	ADJ
cana-5754	100	26	groups	group	NOUN
cana-5754	100	27	.	.	PUNCT
cana-5754	101	1	can	can	AUX
cana-5754	101	2	be	be	AUX
cana-5754	101	3	used	use	VERB
cana-5754	101	4	using	use	VERB
cana-5754	101	5	methods	method	NOUN
cana-5754	101	6	such	such	ADJ
cana-5754	101	7	as	as	ADP
cana-5754	101	8	sample	sample	NOUN
cana-5754	101	9	layer	layer	NOUN
cana-5754	101	10	,	,	PUNCT
cana-5754	101	11	data	datum	NOUN
cana-5754	101	12	extension	extension	NOUN
cana-5754	101	13	,	,	PUNCT
cana-5754	101	14	careful	careful	ADJ
cana-5754	101	15	assessment	assessment	NOUN
cana-5754	101	16	of	of	ADP
cana-5754	101	17	demographic	demographic	ADJ
cana-5754	101	18	parameters	parameter	NOUN
cana-5754	101	19	,	,	PUNCT
cana-5754	101	20	facilities	facility	NOUN
cana-5754	101	21	,	,	PUNCT
cana-5754	101	22	and	and	CCONJ
cana-5754	101	23	fair	fair	ADJ
cana-5754	101	24	and	and	CCONJ
cana-5754	101	25	clinically	clinically	ADV
cana-5754	101	26	useful	useful	ADJ
cana-5754	101	27	models	model	NOUN
cana-5754	101	28	for	for	ADP
cana-5754	101	29	the	the	DET
cana-5754	101	30	detection	detection	NOUN
cana-5754	101	31	of	of	ADP
cana-5754	101	32	diabetic	diabetic	ADJ
cana-5754	101	33	retinopathy	retinopathy	NOUN
cana-5754	101	34	.	.	PUNCT
cana-5754	102	1	is	be	AUX
cana-5754	102	2	also	also	ADV
cana-5754	102	3	a	a	DET
cana-5754	102	4	disclosure	disclosure	NOUN
cana-5754	102	5	of	of	ADP
cana-5754	102	6	dataset	dataset	ADJ
cana-5754	102	7	distortions	distortion	NOUN
cana-5754	102	8	and	and	CCONJ
cana-5754	102	9	distortion	distortion	NOUN
cana-5754	102	10	openness	openness	NOUN
cana-5754	102	11	for	for	ADP
cana-5754	102	12	a	a	DET
cana-5754	102	13	more	more	ADV
cana-5754	102	14	comprehensive	comprehensive	ADJ
cana-5754	102	15	understandingand	understandingand	NOUN
cana-5754	102	16	analysis	analysis	NOUN
cana-5754	102	17	of	of	ADP
cana-5754	102	18	research	research	NOUN
cana-5754	102	19	findings	finding	NOUN
cana-5754	102	20	.	.	PUNCT
cana-5754	103	1	references	reference	NOUN
cana-5754	103	2	[	[	X
cana-5754	103	3	1	1	NUM
cana-5754	103	4	]	]	PUNCT
cana-5754	103	5	w.	w.	PROPN
cana-5754	103	6	l.	l.	PROPN
cana-5754	103	7	alyoubi	alyoubi	PROPN
cana-5754	103	8	,	,	PUNCT
cana-5754	103	9	w.	w.	PROPN
cana-5754	103	10	m.	m.	PROPN
cana-5754	103	11	shalash	shalash	PROPN
cana-5754	103	12	,	,	PUNCT
cana-5754	103	13	and	and	CCONJ
cana-5754	103	14	m.	m.	PROPN
cana-5754	103	15	f.	f.	PROPN
cana-5754	103	16	abulkhair	abulkhair	PROPN
cana-5754	103	17	,	,	PUNCT
cana-5754	103	18	“	"	PUNCT
cana-5754	103	19	diabetic	diabetic	ADJ
cana-5754	103	20	retinopathydetection	retinopathydetection	NOUN
cana-5754	103	21	through	through	ADP
cana-5754	103	22	deep	deep	ADJ
cana-5754	103	23	learning	learning	NOUN
cana-5754	103	24	techniques	technique	NOUN
cana-5754	103	25	:	:	PUNCT
cana-5754	103	26	a	a	DET
cana-5754	103	27	review	review	NOUN
cana-5754	103	28	,	,	PUNCT
cana-5754	103	29	”	"	PUNCT
cana-5754	103	30	informatics	informatic	NOUN
cana-5754	103	31	in	in	ADP
cana-5754	103	32	medicine	medicine	NOUN
cana-5754	103	33	unlocked	unlock	VERB
cana-5754	103	34	,	,	PUNCT
cana-5754	103	35	vol	vol	NOUN
cana-5754	103	36	.	.	PROPN
cana-5754	103	37	20	20	NUM
cana-5754	103	38	,	,	PUNCT
cana-5754	103	39	p.	p.	NOUN
cana-5754	103	40	100377	100377	NUM
cana-5754	103	41	,	,	PUNCT
cana-5754	103	42	2020	2020	NUM
cana-5754	103	43	.	.	PUNCT
cana-5754	104	1	[	[	X
cana-5754	104	2	2	2	NUM
cana-5754	104	3	]	]	PUNCT
cana-5754	104	4	r.	r.	NOUN
cana-5754	104	5	rajalakshmi	rajalakshmi	PROPN
cana-5754	104	6	,	,	PUNCT
cana-5754	104	7	r.	r.	PROPN
cana-5754	104	8	subashini	subashini	PROPN
cana-5754	104	9	,	,	PUNCT
cana-5754	104	10	r.	r.	PROPN
cana-5754	104	11	m.	m.	PROPN
cana-5754	104	12	anjana	anjana	PROPN
cana-5754	104	13	,	,	PUNCT
cana-5754	104	14	and	and	CCONJ
cana-5754	104	15	v.	v.	ADP
cana-5754	104	16	mohan	mohan	PROPN
cana-5754	104	17	,	,	PUNCT
cana-5754	104	18	“	"	PUNCT
cana-5754	104	19	automateddiabetic	automateddiabetic	ADJ
cana-5754	104	20	retinopathy	retinopathy	ADJ
cana-5754	104	21	detection	detection	NOUN
cana-5754	104	22	in	in	ADP
cana-5754	104	23	smartphone	smartphone	NOUN
cana-5754	104	24	-	-	PUNCT
cana-5754	104	25	based	base	VERB
cana-5754	104	26	fundus	fundus	NOUN
cana-5754	104	27	photography	photography	NOUN
cana-5754	104	28	using	use	VERB
cana-5754	104	29	artificial	artificial	ADJ
cana-5754	104	30	intelligence	intelligence	NOUN
cana-5754	104	31	,	,	PUNCT
cana-5754	104	32	”	"	PUNCT
cana-5754	104	33	eye	eye	NOUN
cana-5754	104	34	,	,	PUNCT
cana-5754	104	35	vol	vol	NOUN
cana-5754	104	36	.	.	PROPN
cana-5754	104	37	32	32	NUM
cana-5754	104	38	,	,	PUNCT
cana-5754	104	39	no	no	INTJ
cana-5754	104	40	.	.	NOUN
cana-5754	104	41	6	6	NUM
cana-5754	104	42	,	,	PUNCT
cana-5754	104	43	pp	pp	ADJ
cana-5754	104	44	.	.	PUNCT
cana-5754	105	1	1138–1144	1138–1144	NUM
cana-5754	105	2	,	,	PUNCT
cana-5754	105	3	2018	2018	NUM
cana-5754	105	4	.	.	PUNCT
cana-5754	106	1	[	[	X
cana-5754	106	2	3	3	X
cana-5754	106	3	]	]	PUNCT
cana-5754	106	4	world	world	PROPN
cana-5754	106	5	health	health	NOUN
cana-5754	106	6	organization	organization	NOUN
cana-5754	106	7	,	,	PUNCT
cana-5754	106	8	“	"	PUNCT
cana-5754	106	9	diabetes	diabetes	NOUN
cana-5754	106	10	.	.	PUNCT
cana-5754	106	11	”	"	PUNCT
cana-5754	107	1	https://www.who.int/healthtopics/diabetes#tab=tab\_1	https://www.who.int/healthtopics/diabetes#tab=tab\_1	PROPN
cana-5754	107	2	,	,	PUNCT
cana-5754	107	3	2022	2022	NUM
cana-5754	107	4	.	.	PUNCT
cana-5754	108	1	[	[	X
cana-5754	108	2	4	4	X
cana-5754	108	3	]	]	X
cana-5754	108	4	s.	s.	PROPN
cana-5754	108	5	bhandari	bhandari	PROPN
cana-5754	108	6	,	,	PUNCT
cana-5754	108	7	s.	s.	PROPN
cana-5754	108	8	pathak	pathak	PROPN
cana-5754	108	9	,	,	PUNCT
cana-5754	108	10	and	and	CCONJ
cana-5754	108	11	s.	s.	PROPN
cana-5754	108	12	a.	a.	PROPN
cana-5754	108	13	jain	jain	PROPN
cana-5754	108	14	,	,	PUNCT
cana-5754	108	15	“	"	PUNCT
cana-5754	108	16	a	a	DET
cana-5754	108	17	literature	literature	NOUN
cana-5754	108	18	review	review	NOUN
cana-5754	108	19	of	of	ADP
cana-5754	108	20	early	early	ADJ
cana-5754	108	21	-	-	PUNCT
cana-5754	108	22	stagediabetic	stagediabetic	ADJ
cana-5754	108	23	retinopathy	retinopathy	ADJ
cana-5754	108	24	detection	detection	NOUN
cana-5754	108	25	using	use	VERB
cana-5754	108	26	deep	deep	ADJ
cana-5754	108	27	learning	learning	NOUN
cana-5754	108	28	and	and	CCONJ
cana-5754	108	29	evolutionary	evolutionary	ADJ
cana-5754	108	30	computing	computing	NOUN
cana-5754	108	31	techniques	technique	NOUN
cana-5754	108	32	,	,	PUNCT
cana-5754	108	33	”	"	PUNCT
cana-5754	108	34	archives	archive	NOUN
cana-5754	108	35	of	of	ADP
cana-5754	108	36	computational	computational	ADJ
cana-5754	108	37	methods	method	NOUN
cana-5754	108	38	in	in	ADP
cana-5754	108	39	engineering	engineering	NOUN
cana-5754	108	40	,	,	PUNCT
cana-5754	108	41	vol	vol	NOUN
cana-5754	108	42	.	.	PROPN
cana-5754	108	43	30	30	NUM
cana-5754	108	44	,	,	PUNCT
cana-5754	108	45	no	no	INTJ
cana-5754	108	46	.	.	NOUN
cana-5754	108	47	2	2	NUM
cana-5754	108	48	,	,	PUNCT
cana-5754	108	49	pp	pp	ADJ
cana-5754	108	50	.	.	PUNCT
cana-5754	109	1	799–810	799–810	NUM
cana-5754	109	2	,	,	PUNCT
cana-5754	109	3	2023	2023	NUM
cana-5754	109	4	.	.	PUNCT
cana-5754	110	1	[	[	X
cana-5754	110	2	5	5	NUM
cana-5754	110	3	]	]	PUNCT
cana-5754	110	4	a.	a.	NOUN
cana-5754	110	5	k.	k.	PROPN
cana-5754	110	6	gangwar	gangwar	PROPN
cana-5754	110	7	and	and	CCONJ
cana-5754	110	8	v.	v.	ADP
cana-5754	110	9	ravi	ravi	PROPN
cana-5754	110	10	,	,	PUNCT
cana-5754	110	11	“	"	PUNCT
cana-5754	110	12	diabetic	diabetic	ADJ
cana-5754	110	13	retinopathy	retinopathy	ADJ
cana-5754	110	14	detection	detection	NOUN
cana-5754	110	15	using	use	VERB
cana-5754	110	16	transfer	transfer	NOUN
cana-5754	110	17	learning	learning	NOUN
cana-5754	110	18	and	and	CCONJ
cana-5754	110	19	deep	deep	ADJ
cana-5754	110	20	learning	learning	NOUN
cana-5754	110	21	,	,	PUNCT
cana-5754	110	22	”	"	PUNCT
cana-5754	110	23	in	in	ADP
cana-5754	110	24	evolution	evolution	NOUN
cana-5754	110	25	in	in	ADP
cana-5754	110	26	computational	computational	ADJ
cana-5754	110	27	intelligence	intelligence	NOUN
cana-5754	110	28	:	:	PUNCT
cana-5754	110	29	frontiers	frontier	NOUN
cana-5754	110	30	in	in	ADP
cana-5754	110	31	intelligent	intelligent	ADJ
cana-5754	110	32	computing	computing	NOUN
cana-5754	110	33	:	:	PUNCT
cana-5754	110	34	theory	theory	NOUN
cana-5754	110	35	and	and	CCONJ
cana-5754	110	36	applications	application	NOUN
cana-5754	110	37	(	(	PUNCT
cana-5754	110	38	ficta	ficta	NOUN
cana-5754	110	39	2020	2020	NUM
cana-5754	110	40	)	)	PUNCT
cana-5754	110	41	,	,	PUNCT
cana-5754	110	42	volume	volume	NOUN
cana-5754	110	43	1	1	NUM
cana-5754	110	44	,	,	PUNCT
cana-5754	110	45	pp	pp	ADJ
cana-5754	110	46	.	.	PUNCT
cana-5754	111	1	679–689	679–689	NUM
cana-5754	111	2	,	,	PUNCT
cana-5754	111	3	springer	springer	NOUN
cana-5754	111	4	,	,	PUNCT
cana-5754	111	5	2021	2021	NUM
cana-5754	111	6	.	.	PUNCT
cana-5754	112	1	[	[	X
cana-5754	112	2	6	6	NUM
cana-5754	112	3	]	]	PUNCT
cana-5754	112	4	z.	z.	PROPN
cana-5754	112	5	khan	khan	PROPN
cana-5754	112	6	,	,	PUNCT
cana-5754	112	7	f.	f.	PROPN
cana-5754	112	8	g.	g.	PROPN
cana-5754	112	9	khan	khan	PROPN
cana-5754	112	10	,	,	PUNCT
cana-5754	112	11	a.	a.	PROPN
cana-5754	112	12	khan	khan	PROPN
cana-5754	112	13	,	,	PUNCT
cana-5754	112	14	z.	z.	PROPN
cana-5754	112	15	u.	u.	PROPN
cana-5754	112	16	rehman	rehman	PROPN
cana-5754	112	17	,	,	PUNCT
cana-5754	112	18	s.	s.	PROPN
cana-5754	112	19	shah	shah	PROPN
cana-5754	112	20	,	,	PUNCT
cana-5754	112	21	s.	s.	PROPN
cana-5754	112	22	qummar	qummar	PROPN
cana-5754	112	23	,	,	PUNCT
cana-5754	112	24	f.	f.	PROPN
cana-5754	112	25	ali	ali	PROPN
cana-5754	112	26	,	,	PUNCT
cana-5754	112	27	and	and	CCONJ
cana-5754	112	28	s.	s.	PROPN
cana-5754	112	29	pack	pack	PROPN
cana-5754	112	30	,	,	PUNCT
cana-5754	112	31	“	"	PUNCT
cana-5754	112	32	diabetic	diabetic	ADJ
cana-5754	112	33	retinopathy	retinopathy	ADJ
cana-5754	112	34	detection	detection	NOUN
cana-5754	112	35	using	use	VERB
cana-5754	112	36	vgg	vgg	PROPN
cana-5754	112	37	-	-	PUNCT
cana-5754	112	38	nin	nin	PROPN
cana-5754	112	39	a	a	DET
cana-5754	112	40	deep	deep	ADJ
cana-5754	112	41	learning	learning	NOUN
cana-5754	112	42	architecture	architecture	NOUN
cana-5754	112	43	,	,	PUNCT
cana-5754	112	44	”	"	PUNCT
cana-5754	112	45	ieee	ieee	NOUN
cana-5754	112	46	access	access	NOUN
cana-5754	112	47	,	,	PUNCT
cana-5754	112	48	vol	vol	NOUN
cana-5754	112	49	.	.	NOUN
cana-5754	112	50	9	9	NUM
cana-5754	112	51	,	,	PUNCT
cana-5754	112	52	pp	pp	ADJ
cana-5754	112	53	.	.	PUNCT
cana-5754	112	54	61408–61416	61408–61416	NUM
cana-5754	112	55	,	,	PUNCT
cana-5754	112	56	2021	2021	NUM
cana-5754	112	57	.	.	PUNCT
cana-5754	113	1	[	[	X
cana-5754	113	2	7	7	X
cana-5754	113	3	]	]	X
cana-5754	113	4	e.	e.	PROPN
cana-5754	113	5	özbay	özbay	PROPN
cana-5754	113	6	,	,	PUNCT
cana-5754	113	7	“	"	PUNCT
cana-5754	113	8	an	an	DET
cana-5754	113	9	active	active	ADJ
cana-5754	113	10	deep	deep	ADJ
cana-5754	113	11	learning	learning	NOUN
cana-5754	113	12	method	method	NOUN
cana-5754	113	13	for	for	ADP
cana-5754	113	14	diabetic	diabetic	ADJ
cana-5754	113	15	retinopathy	retinopathy	ADJ
cana-5754	113	16	detection	detection	NOUN
cana-5754	113	17	in	in	ADP
cana-5754	113	18	segmented	segment	VERB
cana-5754	113	19	fundus	fundus	NOUN
cana-5754	113	20	images	image	NOUN
cana-5754	113	21	using	use	VERB
cana-5754	113	22	artificial	artificial	ADJ
cana-5754	113	23	bee	bee	NOUN
cana-5754	113	24	colony	colony	NOUN
cana-5754	113	25	algorithm,”artificial	algorithm,”artificial	ADJ
cana-5754	113	26	intelligence	intelligence	NOUN
cana-5754	113	27	review	review	NOUN
cana-5754	113	28	,	,	PUNCT
cana-5754	113	29	vol	vol	NOUN
cana-5754	113	30	.	.	PROPN
cana-5754	113	31	56	56	NUM
cana-5754	113	32	,	,	PUNCT
cana-5754	113	33	no	no	INTJ
cana-5754	113	34	.	.	NOUN
cana-5754	113	35	4	4	NUM
cana-5754	113	36	,	,	PUNCT
cana-5754	113	37	pp	pp	ADJ
cana-5754	113	38	.	.	PUNCT
cana-5754	114	1	3291–3318	3291–3318	NUM
cana-5754	114	2	,	,	PUNCT
cana-5754	114	3	2023	2023	NUM
cana-5754	114	4	.	.	PUNCT
cana-5754	115	1	[	[	X
cana-5754	115	2	8	8	NUM
cana-5754	115	3	]	]	PUNCT
cana-5754	115	4	a.	a.	NOUN
cana-5754	115	5	panwar	panwar	PROPN
cana-5754	115	6	,	,	PUNCT
cana-5754	115	7	g.	g.	PROPN
cana-5754	115	8	semwal	semwal	PROPN
cana-5754	115	9	,	,	PUNCT
cana-5754	115	10	s.	s.	PROPN
cana-5754	115	11	goel	goel	PROPN
cana-5754	115	12	,	,	PUNCT
cana-5754	115	13	and	and	CCONJ
cana-5754	115	14	s.	s.	PROPN
cana-5754	115	15	gupta	gupta	PROPN
cana-5754	115	16	,	,	PUNCT
cana-5754	115	17	“	"	PUNCT
cana-5754	115	18	stratification	stratification	NOUN
cana-5754	115	19	of	of	ADP
cana-5754	115	20	the	the	DET
cana-5754	115	21	lesions	lesion	NOUN
cana-5754	115	22	in	in	ADP
cana-5754	115	23	color	color	NOUN
cana-5754	115	24	fundus	fundus	NOUN
cana-5754	115	25	images	image	NOUN
cana-5754	115	26	of	of	ADP
cana-5754	115	27	diabetic	diabetic	ADJ
cana-5754	115	28	retinopathy	retinopathy	ADJ
cana-5754	115	29	patients	patient	NOUN
cana-5754	115	30	using	use	VERB
cana-5754	115	31	deep	deep	ADJ
cana-5754	115	32	learning	learning	NOUN
cana-5754	115	33	models	model	NOUN
cana-5754	115	34	and	and	CCONJ
cana-5754	115	35	machine	machine	NOUN
cana-5754	115	36	learning	learn	VERB
cana-5754	115	37	communications	communication	NOUN
cana-5754	115	38	on	on	ADP
cana-5754	115	39	applied	apply	VERB
cana-5754	115	40	nonlinear	nonlinear	ADJ
cana-5754	115	41	analysis	analysis	NOUN
cana-5754	115	42	issn	issn	NOUN
cana-5754	115	43	:	:	PUNCT
cana-5754	115	44	1074	1074	NUM
cana-5754	115	45	-	-	PUNCT
cana-5754	115	46	133x	133x	NUM
cana-5754	115	47	vol	vol	NOUN
cana-5754	115	48	32	32	NUM
cana-5754	115	49	no	no	NOUN
cana-5754	115	50	.	.	PUNCT
cana-5754	116	1	3s	3s	NUM
cana-5754	116	2	(	(	PUNCT
cana-5754	116	3	2025	2025	NUM
cana-5754	116	4	)	)	PUNCT
cana-5754	116	5	788	788	NUM
cana-5754	116	6	https://internationalpubls.com	https://internationalpubls.com	NUM
cana-5754	116	7	classifiers	classifier	NOUN
cana-5754	116	8	,	,	PUNCT
cana-5754	116	9	”	"	PUNCT
cana-5754	116	10	in	in	ADP
cana-5754	116	11	edge	edge	NOUN
cana-5754	116	12	analytics	analytic	NOUN
cana-5754	116	13	:	:	PUNCT
cana-5754	116	14	select	select	ADJ
cana-5754	116	15	proceedings	proceeding	NOUN
cana-5754	116	16	of	of	ADP
cana-5754	116	17	26th	26th	ADJ
cana-5754	116	18	international	international	ADJ
cana-5754	116	19	conference	conference	NOUN
cana-5754	116	20	—	—	PUNCT
cana-5754	116	21	adcom	adcom	PROPN
cana-5754	116	22	2020	2020	NUM
cana-5754	116	23	,	,	PUNCT
cana-5754	116	24	pp	pp	ADP
cana-5754	116	25	.	.	PUNCT
cana-5754	117	1	653–666	653–666	NUM
cana-5754	117	2	,	,	PUNCT
cana-5754	117	3	springer	springer	NOUN
cana-5754	117	4	,	,	PUNCT
cana-5754	117	5	2022	2022	NUM
cana-5754	117	6	.	.	PUNCT
cana-5754	118	1	[	[	X
cana-5754	118	2	9	9	NUM
cana-5754	118	3	]	]	X
cana-5754	118	4	g.	g.	NOUN
cana-5754	118	5	parthasharathi	parthasharathi	PROPN
cana-5754	118	6	,	,	PUNCT
cana-5754	118	7	r.	r.	PROPN
cana-5754	118	8	premnivas	premnivas	PROPN
cana-5754	118	9	,	,	PUNCT
cana-5754	118	10	k.	k.	PROPN
cana-5754	118	11	jasmine	jasmine	PROPN
cana-5754	118	12	,	,	PUNCT
cana-5754	118	13	et	et	PROPN
cana-5754	118	14	al	al	PROPN
cana-5754	118	15	.	.	PROPN
cana-5754	118	16	,	,	PUNCT
cana-5754	118	17	“	"	PUNCT
cana-5754	118	18	diabetic	diabetic	ADJ
cana-5754	118	19	retinopathy	retinopathy	ADJ
cana-5754	118	20	detection	detection	NOUN
cana-5754	118	21	using	use	VERB
cana-5754	118	22	machine	machine	NOUN
cana-5754	118	23	learning	learning	NOUN
cana-5754	118	24	,	,	PUNCT
cana-5754	118	25	”	"	PUNCT
cana-5754	118	26	journal	journal	NOUN
cana-5754	118	27	of	of	ADP
cana-5754	118	28	innovative	innovative	ADJ
cana-5754	118	29	image	image	NOUN
cana-5754	118	30	processing	processing	NOUN
cana-5754	118	31	,	,	PUNCT
cana-5754	118	32	vol	vol	NOUN
cana-5754	118	33	.	.	PROPN
cana-5754	118	34	4	4	NUM
cana-5754	118	35	,	,	PUNCT
cana-5754	118	36	no	no	INTJ
cana-5754	118	37	.	.	NOUN
cana-5754	118	38	1	1	NUM
cana-5754	118	39	,	,	PUNCT
cana-5754	118	40	pp	pp	ADJ
cana-5754	118	41	.	.	PUNCT
cana-5754	118	42	26–33	26–33	NUM
cana-5754	118	43	,	,	PUNCT
cana-5754	118	44	2022	2022	NUM
cana-5754	118	45	.	.	PUNCT
cana-5754	119	1	[	[	X
cana-5754	119	2	10	10	NUM
cana-5754	119	3	]	]	X
cana-5754	119	4	n.	n.	NOUN
cana-5754	119	5	tsiknakis	tsiknakis	PROPN
cana-5754	119	6	,	,	PUNCT
cana-5754	119	7	d.	d.	PROPN
cana-5754	119	8	theodoropoulos	theodoropoulos	PROPN
cana-5754	119	9	,	,	PUNCT
cana-5754	119	10	g.	g.	PROPN
cana-5754	119	11	manikis	manikis	PROPN
cana-5754	119	12	,	,	PUNCT
cana-5754	119	13	e.	e.	PROPN
cana-5754	119	14	ktistakis	ktistakis	PROPN
cana-5754	119	15	,	,	PUNCT
cana-5754	119	16	o.	o.	PROPN
cana-5754	119	17	boutsora	boutsora	PROPN
cana-5754	119	18	,	,	PUNCT
cana-5754	119	19	a.	a.	PROPN
cana-5754	119	20	berto	berto	PROPN
cana-5754	119	21	,	,	PUNCT
cana-5754	119	22	f.	f.	PROPN
cana-5754	119	23	scarpa	scarpa	PROPN
cana-5754	119	24	,	,	PUNCT
cana-5754	119	25	a.	a.	PROPN
cana-5754	119	26	scarpa	scarpa	PROPN
cana-5754	119	27	,	,	PUNCT
cana-5754	119	28	d.	d.	PROPN
cana-5754	119	29	i.	i.	PROPN
cana-5754	119	30	fotiadis	fotiadis	PROPN
cana-5754	119	31	,	,	PUNCT
cana-5754	119	32	and	and	CCONJ
cana-5754	119	33	k.	k.	PROPN
cana-5754	119	34	marias	marias	PROPN
cana-5754	119	35	,	,	PUNCT
cana-5754	119	36	“	"	PUNCT
cana-5754	119	37	deep	deep	ADJ
cana-5754	119	38	learning	learning	NOUN
cana-5754	119	39	for	for	ADP
cana-5754	119	40	diabetic	diabetic	ADJ
cana-5754	119	41	retinopathy	retinopathy	ADJ
cana-5754	119	42	detection	detection	NOUN
cana-5754	119	43	and	and	CCONJ
cana-5754	119	44	classification	classification	NOUN
cana-5754	119	45	based	base	VERB
cana-5754	119	46	on	on	ADP
cana-5754	119	47	fundus	fundus	NOUN
cana-5754	119	48	images	image	NOUN
cana-5754	119	49	:	:	PUNCT
cana-5754	119	50	a	a	DET
cana-5754	119	51	review	review	NOUN
cana-5754	119	52	,	,	PUNCT
cana-5754	119	53	”	"	PUNCT
cana-5754	119	54	computers	computer	NOUN
cana-5754	119	55	in	in	ADP
cana-5754	119	56	biology	biology	NOUN
cana-5754	119	57	and	and	CCONJ
cana-5754	119	58	medicine	medicine	NOUN
cana-5754	119	59	,	,	PUNCT
cana-5754	119	60	vol	vol	NOUN
cana-5754	119	61	.	.	PUNCT
cana-5754	120	1	135,p	135,p	ADJ
cana-5754	120	2	.	.	PUNCT
cana-5754	121	1	104599	104599	NUM
cana-5754	121	2	,	,	PUNCT
cana-5754	121	3	2021	2021	NUM
cana-5754	121	4	.	.	PUNCT
cana-5754	122	1	[	[	X
cana-5754	122	2	11	11	NUM
cana-5754	122	3	]	]	PUNCT
cana-5754	122	4	r.	r.	PROPN
cana-5754	122	5	pires	pires	PROPN
cana-5754	122	6	,	,	PUNCT
cana-5754	122	7	s.	s.	PROPN
cana-5754	122	8	avila	avila	PROPN
cana-5754	122	9	,	,	PUNCT
cana-5754	122	10	j.	j.	PROPN
cana-5754	122	11	wainer	wainer	PROPN
cana-5754	122	12	,	,	PUNCT
cana-5754	122	13	e.	e.	PROPN
cana-5754	122	14	valle	valle	PROPN
cana-5754	122	15	,	,	PUNCT
cana-5754	122	16	m.	m.	PROPN
cana-5754	122	17	d.	d.	PROPN
cana-5754	122	18	abramoff	abramoff	PROPN
cana-5754	122	19	,	,	PUNCT
cana-5754	122	20	and	and	CCONJ
cana-5754	122	21	a.	a.	PROPN
cana-5754	122	22	rocha	rocha	PROPN
cana-5754	122	23	,	,	PUNCT
cana-5754	122	24	“	"	PUNCT
cana-5754	122	25	a	a	DET
cana-5754	122	26	data	data	NOUN
cana-5754	122	27	-	-	PUNCT
cana-5754	122	28	driven	drive	VERB
cana-5754	122	29	approach	approach	NOUN
cana-5754	122	30	to	to	ADP
cana-5754	122	31	referable	referable	ADJ
cana-5754	122	32	diabetic	diabetic	ADJ
cana-5754	122	33	retinopathy	retinopathy	ADJ
cana-5754	122	34	detection	detection	NOUN
cana-5754	122	35	,	,	PUNCT
cana-5754	122	36	”	"	PUNCT
cana-5754	122	37	artificial	artificial	ADJ
cana-5754	122	38	intelligence	intelligence	NOUN
cana-5754	122	39	in	in	ADP
cana-5754	122	40	medicine	medicine	NOUN
cana-5754	122	41	,	,	PUNCT
cana-5754	122	42	vol	vol	NOUN
cana-5754	122	43	.	.	PROPN
cana-5754	123	1	96	96	NUM
cana-5754	123	2	,	,	PUNCT
cana-5754	123	3	pp	pp	ADJ
cana-5754	123	4	.	.	PUNCT
cana-5754	124	1	93–106	93–106	NUM
cana-5754	124	2	,	,	PUNCT
cana-5754	124	3	2019	2019	NUM
cana-5754	124	4	.	.	PUNCT
cana-5754	125	1	[	[	X
cana-5754	125	2	12	12	NUM
cana-5754	125	3	]	]	X
cana-5754	125	4	y.	y.	PROPN
cana-5754	125	5	yang	yang	PROPN
cana-5754	125	6	,	,	PUNCT
cana-5754	125	7	t.	t.	PROPN
cana-5754	125	8	li	li	PROPN
cana-5754	125	9	,	,	PUNCT
cana-5754	125	10	w.	w.	PROPN
cana-5754	125	11	li	li	PROPN
cana-5754	125	12	,	,	PUNCT
cana-5754	125	13	h.	h.	PROPN
cana-5754	125	14	wu	wu	PROPN
cana-5754	125	15	,	,	PUNCT
cana-5754	125	16	w.	w.	PROPN
cana-5754	125	17	fan	fan	PROPN
cana-5754	125	18	,	,	PUNCT
cana-5754	125	19	and	and	CCONJ
cana-5754	125	20	w.	w.	PROPN
cana-5754	125	21	zhang	zhang	PROPN
cana-5754	125	22	,	,	PUNCT
cana-5754	125	23	“	"	PUNCT
cana-5754	125	24	lesion	lesion	NOUN
cana-5754	125	25	detection	detection	NOUN
cana-5754	125	26	and	and	CCONJ
cana-5754	125	27	grading	grading	NOUN
cana-5754	125	28	of	of	ADP
cana-5754	125	29	diabetic	diabetic	ADJ
cana-5754	125	30	retinopathy	retinopathy	NOUN
cana-5754	125	31	via	via	ADP
cana-5754	125	32	two	two	NUM
cana-5754	125	33	-	-	PUNCT
cana-5754	125	34	stages	stage	NOUN
cana-5754	125	35	deep	deep	ADJ
cana-5754	125	36	convolutional	convolutional	ADJ
cana-5754	125	37	neural	neural	ADJ
cana-5754	125	38	networks	network	NOUN
cana-5754	125	39	,	,	PUNCT
cana-5754	125	40	”	"	PUNCT
cana-5754	125	41	in	in	ADP
cana-5754	125	42	medical	medical	ADJ
cana-5754	125	43	image	image	NOUN
cana-5754	125	44	computing	computing	NOUN
cana-5754	125	45	and	and	CCONJ
cana-5754	125	46	computer	computer	NOUN
cana-5754	125	47	assisted	assist	VERB
cana-5754	125	48	interventionmiccai	interventionmiccai	PROPN
cana-5754	125	49	2017	2017	NUM
cana-5754	125	50	:	:	PUNCT
cana-5754	126	1	20th	20th	ADJ
cana-5754	126	2	international	international	ADJ
cana-5754	126	3	conference	conference	NOUN
cana-5754	126	4	,	,	PUNCT
cana-5754	126	5	quebec	quebec	PROPN
cana-5754	126	6	city	city	PROPN
cana-5754	126	7	,	,	PUNCT
cana-5754	126	8	qc	qc	PROPN
cana-5754	126	9	,	,	PUNCT
cana-5754	126	10	canada	canada	PROPN
cana-5754	126	11	,	,	PUNCT
cana-5754	126	12	september	september	PROPN
cana-5754	126	13	11	11	NUM
cana-5754	126	14	-	-	SYM
cana-5754	126	15	13	13	NUM
cana-5754	126	16	,	,	PUNCT
cana-5754	126	17	2017	2017	NUM
cana-5754	126	18	,	,	PUNCT
cana-5754	126	19	proceedings	proceeding	NOUN
cana-5754	126	20	,	,	PUNCT
cana-5754	126	21	part	part	NOUN
cana-5754	126	22	iii	iii	PROPN
cana-5754	126	23	20	20	NUM
cana-5754	126	24	,	,	PUNCT
cana-5754	126	25	pp	pp	ADJ
cana-5754	126	26	.	.	PUNCT
cana-5754	127	1	533	533	NUM
cana-5754	127	2	–	–	PUNCT
cana-5754	127	3	540	540	NUM
cana-5754	127	4	,	,	PUNCT
cana-5754	127	5	springer	springer	NOUN
cana-5754	127	6	,	,	PUNCT
cana-5754	127	7	2017	2017	NUM
cana-5754	127	8	.	.	PUNCT
cana-5754	128	1	[	[	X
cana-5754	128	2	13	13	NUM
cana-5754	128	3	]	]	PUNCT
cana-5754	128	4	p.	p.	NOUN
cana-5754	128	5	saranya	saranya	PROPN
cana-5754	128	6	and	and	CCONJ
cana-5754	128	7	s.	s.	PROPN
cana-5754	128	8	prabakaran	prabakaran	PROPN
cana-5754	128	9	,	,	PUNCT
cana-5754	128	10	“	"	PUNCT
cana-5754	128	11	automatic	automatic	ADJ
cana-5754	128	12	detection	detection	NOUN
cana-5754	128	13	of	of	ADP
cana-5754	128	14	non	non	ADJ
cana-5754	128	15	-	-	ADJ
cana-5754	128	16	proliferative	proliferative	ADJ
cana-5754	128	17	diabetic	diabetic	ADJ
cana-5754	128	18	retinopathy	retinopathy	NOUN
cana-5754	128	19	in	in	ADP
cana-5754	128	20	retinal	retinal	ADJ
cana-5754	128	21	fundus	fundus	NOUN
cana-5754	128	22	images	image	NOUN
cana-5754	128	23	using	use	VERB
cana-5754	128	24	convolution	convolution	NOUN
cana-5754	128	25	neural	neural	ADJ
cana-5754	128	26	network	network	NOUN
cana-5754	128	27	,	,	PUNCT
cana-5754	128	28	”	"	PUNCT
cana-5754	128	29	journal	journal	NOUN
cana-5754	128	30	of	of	ADP
cana-5754	128	31	ambient	ambient	ADJ
cana-5754	128	32	intelligence	intelligence	NOUN
cana-5754	128	33	and	and	CCONJ
cana-5754	128	34	humanized	humanize	VERB
cana-5754	128	35	computing	computing	NOUN
cana-5754	128	36	,	,	PUNCT
cana-5754	128	37	pp	pp	PROPN
cana-5754	128	38	.	.	PUNCT
cana-5754	129	1	1–10	1–10	NOUN
cana-5754	129	2	,	,	PUNCT
cana-5754	129	3	2020	2020	NUM
cana-5754	129	4	.	.	PUNCT
cana-5754	130	1	[	[	X
cana-5754	130	2	14	14	NUM
cana-5754	130	3	]	]	PUNCT
cana-5754	130	4	t.	t.	PROPN
cana-5754	130	5	m.	m.	PROPN
cana-5754	130	6	usman	usman	PROPN
cana-5754	130	7	,	,	PUNCT
cana-5754	130	8	y.	y.	PROPN
cana-5754	130	9	k.	k.	PROPN
cana-5754	130	10	saheed	saheed	PROPN
cana-5754	130	11	,	,	PUNCT
cana-5754	130	12	d.	d.	PROPN
cana-5754	130	13	ignace	ignace	PROPN
cana-5754	130	14	,	,	PUNCT
cana-5754	130	15	and	and	CCONJ
cana-5754	130	16	a.	a.	PROPN
cana-5754	130	17	nsang	nsang	PROPN
cana-5754	130	18	,	,	PUNCT
cana-5754	130	19	“	"	PUNCT
cana-5754	130	20	diabetic	diabetic	ADJ
cana-5754	130	21	retinopathy	retinopathy	ADJ
cana-5754	130	22	detection	detection	NOUN
cana-5754	130	23	using	use	VERB
cana-5754	130	24	principal	principal	ADJ
cana-5754	130	25	component	component	NOUN
cana-5754	130	26	analysis	analysis	NOUN
cana-5754	130	27	multi	multi	ADJ
cana-5754	130	28	-	-	ADJ
cana-5754	130	29	label	label	ADJ
cana-5754	130	30	feature	feature	NOUN
cana-5754	130	31	extraction	extraction	NOUN
cana-5754	130	32	and	and	CCONJ
cana-5754	130	33	classification	classification	NOUN
cana-5754	130	34	,	,	PUNCT
cana-5754	130	35	”	"	PUNCT
cana-5754	130	36	international	international	ADJ
cana-5754	130	37	journal	journal	NOUN
cana-5754	130	38	of	of	ADP
cana-5754	130	39	cognitive	cognitive	ADJ
cana-5754	130	40	computing	computing	NOUN
cana-5754	130	41	in	in	ADP
cana-5754	130	42	engineering	engineering	NOUN
cana-5754	130	43	,	,	PUNCT
cana-5754	130	44	vol	vol	NOUN
cana-5754	130	45	.	.	PROPN
cana-5754	130	46	4	4	NUM
cana-5754	130	47	,	,	PUNCT
cana-5754	130	48	pp	pp	ADJ
cana-5754	130	49	.	.	PUNCT
cana-5754	131	1	78–88	78–88	NUM
cana-5754	131	2	,	,	PUNCT
cana-5754	131	3	2023	2023	NUM
cana-5754	131	4	.	.	PUNCT
cana-5754	131	5	.	.	PUNCT