id	sid	tid	token	lemma	pos
aiti-14589	1	1	microsoft	microsoft	PROPN
aiti-14589	1	2	word	word	NOUN
aiti-14589	1	3	6	6	NUM
aiti-14589	1	4	-	-	PUNCT
aiti-14589	1	5	v10n4(2025)-aiti#14589(395	v10n4(2025)-aiti#14589(395	NOUN
aiti-14589	1	6	-	-	PUNCT
aiti-14589	1	7	406).docx	406).docx	NOUN
aiti-14589	1	8	advances	advance	NOUN
aiti-14589	1	9	in	in	ADP
aiti-14589	1	10	technology	technology	NOUN
aiti-14589	1	11	innovation	innovation	NOUN
aiti-14589	1	12	,	,	PUNCT
aiti-14589	1	13	vol	vol	NOUN
aiti-14589	1	14	.	.	PROPN
aiti-14589	1	15	10	10	NUM
aiti-14589	1	16	,	,	PUNCT
aiti-14589	1	17	no	no	INTJ
aiti-14589	1	18	.	.	NOUN
aiti-14589	1	19	4	4	NUM
aiti-14589	1	20	,	,	PUNCT
aiti-14589	1	21	2025	2025	NUM
aiti-14589	1	22	,	,	PUNCT
aiti-14589	1	23	pp	pp	ADJ
aiti-14589	1	24	.	.	PUNCT
aiti-14589	2	1	395	395	NUM
aiti-14589	2	2	-	-	SYM
aiti-14589	2	3	406	406	NUM
aiti-14589	2	4	comparative	comparative	ADJ
aiti-14589	2	5	analysis	analysis	NOUN
aiti-14589	2	6	of	of	ADP
aiti-14589	2	7	facial	facial	ADJ
aiti-14589	2	8	expression	expression	NOUN
aiti-14589	2	9	recognition	recognition	NOUN
aiti-14589	2	10	using	use	VERB
aiti-14589	2	11	imagebased	imagebase	VERB
aiti-14589	2	12	and	and	CCONJ
aiti-14589	2	13	landmark	landmark	NOUN
aiti-14589	2	14	-	-	PUNCT
aiti-14589	2	15	based	base	VERB
aiti-14589	2	16	methods	method	NOUN
aiti-14589	2	17	thanawat	thanawat	VERB
aiti-14589	2	18	srikaewsiew	srikaewsiew	NOUN
aiti-14589	2	19	,	,	PUNCT
aiti-14589	2	20	sarunya	sarunya	PROPN
aiti-14589	2	21	kanjanawattana	kanjanawattana	PROPN
aiti-14589	2	22	*	*	PROPN
aiti-14589	2	23	school	school	NOUN
aiti-14589	2	24	of	of	ADP
aiti-14589	2	25	computer	computer	NOUN
aiti-14589	2	26	engineering	engineering	NOUN
aiti-14589	2	27	,	,	PUNCT
aiti-14589	2	28	institute	institute	NOUN
aiti-14589	2	29	of	of	ADP
aiti-14589	2	30	engineering	engineering	PROPN
aiti-14589	2	31	,	,	PUNCT
aiti-14589	2	32	suranaree	suranaree	PROPN
aiti-14589	2	33	university	university	PROPN
aiti-14589	2	34	of	of	ADP
aiti-14589	2	35	technology	technology	NOUN
aiti-14589	2	36	,	,	PUNCT
aiti-14589	2	37	nakhon	nakhon	PROPN
aiti-14589	2	38	ratchasima	ratchasima	PROPN
aiti-14589	2	39	,	,	PUNCT
aiti-14589	2	40	thailand	thailand	PROPN
aiti-14589	2	41	received	receive	VERB
aiti-14589	2	42	05	05	NUM
aiti-14589	2	43	december	december	PROPN
aiti-14589	2	44	2024	2024	NUM
aiti-14589	2	45	;	;	PUNCT
aiti-14589	2	46	received	receive	VERB
aiti-14589	2	47	in	in	ADP
aiti-14589	2	48	revised	revise	VERB
aiti-14589	2	49	form	form	NOUN
aiti-14589	2	50	22	22	NUM
aiti-14589	2	51	march	march	NOUN
aiti-14589	2	52	2025	2025	NUM
aiti-14589	2	53	;	;	PUNCT
aiti-14589	2	54	accepted	accept	VERB
aiti-14589	2	55	31	31	NUM
aiti-14589	2	56	march	march	NOUN
aiti-14589	2	57	2025	2025	NUM
aiti-14589	2	58	doi	doi	NOUN
aiti-14589	2	59	:	:	PUNCT
aiti-14589	2	60	https://doi.org/10.46604/aiti.2025.14589	https://doi.org/10.46604/aiti.2025.14589	PROPN
aiti-14589	2	61	abstract	abstract	NOUN
aiti-14589	2	62	this	this	DET
aiti-14589	2	63	study	study	NOUN
aiti-14589	2	64	compares	compare	VERB
aiti-14589	2	65	the	the	DET
aiti-14589	2	66	effectiveness	effectiveness	NOUN
aiti-14589	2	67	of	of	ADP
aiti-14589	2	68	image	image	NOUN
aiti-14589	2	69	-	-	PUNCT
aiti-14589	2	70	based	base	VERB
aiti-14589	2	71	and	and	CCONJ
aiti-14589	2	72	landmark	landmark	NOUN
aiti-14589	2	73	-	-	PUNCT
aiti-14589	2	74	based	base	VERB
aiti-14589	2	75	methods	method	NOUN
aiti-14589	2	76	for	for	ADP
aiti-14589	2	77	facial	facial	ADJ
aiti-14589	2	78	expression	expression	NOUN
aiti-14589	2	79	recognition	recognition	NOUN
aiti-14589	2	80	(	(	PUNCT
aiti-14589	2	81	fer	fer	PROPN
aiti-14589	2	82	)	)	PUNCT
aiti-14589	2	83	in	in	ADP
aiti-14589	2	84	classifying	classify	VERB
aiti-14589	2	85	hurt	hurt	NOUN
aiti-14589	2	86	and	and	CCONJ
aiti-14589	2	87	normal	normal	ADJ
aiti-14589	2	88	facial	facial	ADJ
aiti-14589	2	89	expressions	expression	NOUN
aiti-14589	2	90	,	,	PUNCT
aiti-14589	2	91	utilizing	utilize	VERB
aiti-14589	2	92	datasets	dataset	NOUN
aiti-14589	2	93	from	from	ADP
aiti-14589	2	94	the	the	DET
aiti-14589	2	95	delaware	delaware	PROPN
aiti-14589	2	96	pain	pain	PROPN
aiti-14589	2	97	database	database	NOUN
aiti-14589	2	98	and	and	CCONJ
aiti-14589	2	99	utkface	utkface	NOUN
aiti-14589	2	100	.	.	PUNCT
aiti-14589	3	1	five	five	NUM
aiti-14589	3	2	machine	machine	NOUN
aiti-14589	3	3	learning	learning	NOUN
aiti-14589	3	4	models	model	NOUN
aiti-14589	3	5	are	be	AUX
aiti-14589	3	6	assessed	assess	VERB
aiti-14589	3	7	,	,	PUNCT
aiti-14589	3	8	including	include	VERB
aiti-14589	3	9	convolutional	convolutional	ADJ
aiti-14589	3	10	neural	neural	ADJ
aiti-14589	3	11	networks	network	NOUN
aiti-14589	3	12	(	(	PUNCT
aiti-14589	3	13	cnn	cnn	PROPN
aiti-14589	3	14	)	)	PUNCT
aiti-14589	3	15	,	,	PUNCT
aiti-14589	3	16	support	support	VERB
aiti-14589	3	17	vector	vector	NOUN
aiti-14589	3	18	machines	machine	NOUN
aiti-14589	3	19	(	(	PUNCT
aiti-14589	3	20	svm	svm	PROPN
aiti-14589	3	21	)	)	PUNCT
aiti-14589	3	22	,	,	PUNCT
aiti-14589	3	23	random	random	ADJ
aiti-14589	3	24	forest	forest	NOUN
aiti-14589	3	25	classifier	classifier	NOUN
aiti-14589	3	26	(	(	PUNCT
aiti-14589	3	27	rfc	rfc	PROPN
aiti-14589	3	28	)	)	PUNCT
aiti-14589	3	29	,	,	PUNCT
aiti-14589	3	30	logistic	logistic	ADJ
aiti-14589	3	31	regression	regression	NOUN
aiti-14589	3	32	classifier	classifier	NOUN
aiti-14589	3	33	(	(	PUNCT
aiti-14589	3	34	lrc	lrc	PROPN
aiti-14589	3	35	)	)	PUNCT
aiti-14589	3	36	,	,	PUNCT
aiti-14589	3	37	and	and	CCONJ
aiti-14589	3	38	gradient	gradient	NOUN
aiti-14589	3	39	boosting	boost	VERB
aiti-14589	3	40	classifier	classifier	NOUN
aiti-14589	3	41	(	(	PUNCT
aiti-14589	3	42	gbc	gbc	PROPN
aiti-14589	3	43	)	)	PUNCT
aiti-14589	3	44	.	.	PUNCT
aiti-14589	4	1	the	the	DET
aiti-14589	4	2	findings	finding	NOUN
aiti-14589	4	3	indicate	indicate	VERB
aiti-14589	4	4	that	that	SCONJ
aiti-14589	4	5	cnn	cnn	PROPN
aiti-14589	4	6	achieves	achieve	VERB
aiti-14589	4	7	the	the	DET
aiti-14589	4	8	highest	high	ADJ
aiti-14589	4	9	accuracy	accuracy	NOUN
aiti-14589	4	10	at	at	ADP
aiti-14589	4	11	95	95	NUM
aiti-14589	4	12	%	%	NOUN
aiti-14589	4	13	when	when	SCONJ
aiti-14589	4	14	using	use	VERB
aiti-14589	4	15	landmark	landmark	NOUN
aiti-14589	4	16	-	-	PUNCT
aiti-14589	4	17	based	base	VERB
aiti-14589	4	18	features	feature	NOUN
aiti-14589	4	19	,	,	PUNCT
aiti-14589	4	20	while	while	SCONJ
aiti-14589	4	21	svm	svm	PROPN
aiti-14589	4	22	and	and	CCONJ
aiti-14589	4	23	gbc	gbc	PROPN
aiti-14589	4	24	also	also	ADV
aiti-14589	4	25	perform	perform	VERB
aiti-14589	4	26	admirably	admirably	ADV
aiti-14589	4	27	with	with	ADP
aiti-14589	4	28	these	these	DET
aiti-14589	4	29	features	feature	NOUN
aiti-14589	4	30	.	.	PUNCT
aiti-14589	5	1	conversely	conversely	ADV
aiti-14589	5	2	,	,	PUNCT
aiti-14589	5	3	lrc	lrc	PROPN
aiti-14589	5	4	exhibits	exhibit	VERB
aiti-14589	5	5	inconsistent	inconsistent	ADJ
aiti-14589	5	6	results	result	NOUN
aiti-14589	5	7	,	,	PUNCT
aiti-14589	5	8	especially	especially	ADV
aiti-14589	5	9	when	when	SCONJ
aiti-14589	5	10	relying	rely	VERB
aiti-14589	5	11	on	on	ADP
aiti-14589	5	12	image	image	NOUN
aiti-14589	5	13	-	-	PUNCT
aiti-14589	5	14	based	base	VERB
aiti-14589	5	15	features	feature	NOUN
aiti-14589	5	16	.	.	PUNCT
aiti-14589	6	1	these	these	DET
aiti-14589	6	2	findings	finding	NOUN
aiti-14589	6	3	offer	offer	VERB
aiti-14589	6	4	valuable	valuable	ADJ
aiti-14589	6	5	insights	insight	NOUN
aiti-14589	6	6	into	into	ADP
aiti-14589	6	7	the	the	DET
aiti-14589	6	8	strengths	strength	NOUN
aiti-14589	6	9	and	and	CCONJ
aiti-14589	6	10	weaknesses	weakness	NOUN
aiti-14589	6	11	of	of	ADP
aiti-14589	6	12	each	each	DET
aiti-14589	6	13	approach	approach	NOUN
aiti-14589	6	14	,	,	PUNCT
aiti-14589	6	15	guiding	guide	VERB
aiti-14589	6	16	the	the	DET
aiti-14589	6	17	selection	selection	NOUN
aiti-14589	6	18	of	of	ADP
aiti-14589	6	19	effective	effective	ADJ
aiti-14589	6	20	fer	fer	PROPN
aiti-14589	6	21	techniques	technique	NOUN
aiti-14589	6	22	.	.	PUNCT
aiti-14589	7	1	keywords	keyword	NOUN
aiti-14589	7	2	:	:	PUNCT
aiti-14589	7	3	face	face	VERB
aiti-14589	7	4	expression	expression	NOUN
aiti-14589	7	5	recognition	recognition	NOUN
aiti-14589	7	6	,	,	PUNCT
aiti-14589	7	7	machine	machine	NOUN
aiti-14589	7	8	model	model	PROPN
aiti-14589	7	9	comparison	comparison	NOUN
aiti-14589	7	10	,	,	PUNCT
aiti-14589	7	11	image	image	NOUN
aiti-14589	7	12	-	-	PUNCT
aiti-14589	7	13	based	base	VERB
aiti-14589	7	14	classification	classification	NOUN
aiti-14589	7	15	,	,	PUNCT
aiti-14589	7	16	landmark	landmark	NOUN
aiti-14589	7	17	-	-	PUNCT
aiti-14589	7	18	based	base	VERB
aiti-14589	7	19	classification	classification	NOUN
aiti-14589	7	20	,	,	PUNCT
aiti-14589	7	21	expression	expression	NOUN
aiti-14589	7	22	dataset	dataset	VERB
aiti-14589	7	23	1	1	NUM
aiti-14589	7	24	.	.	PUNCT
aiti-14589	8	1	introduction	introduction	NOUN
aiti-14589	8	2	facial	facial	ADJ
aiti-14589	8	3	expression	expression	NOUN
aiti-14589	8	4	recognition	recognition	NOUN
aiti-14589	8	5	(	(	PUNCT
aiti-14589	8	6	fer	fer	PROPN
aiti-14589	8	7	)	)	PUNCT
aiti-14589	8	8	is	be	AUX
aiti-14589	8	9	a	a	DET
aiti-14589	8	10	field	field	NOUN
aiti-14589	8	11	that	that	PRON
aiti-14589	8	12	combines	combine	VERB
aiti-14589	8	13	computer	computer	NOUN
aiti-14589	8	14	vision	vision	NOUN
aiti-14589	8	15	,	,	PUNCT
aiti-14589	8	16	artificial	artificial	ADJ
aiti-14589	8	17	intelligence	intelligence	NOUN
aiti-14589	8	18	,	,	PUNCT
aiti-14589	8	19	and	and	CCONJ
aiti-14589	8	20	psychology	psychology	NOUN
aiti-14589	8	21	to	to	PART
aiti-14589	8	22	interpret	interpret	VERB
aiti-14589	8	23	human	human	ADJ
aiti-14589	8	24	emotions	emotion	NOUN
aiti-14589	8	25	from	from	ADP
aiti-14589	8	26	facial	facial	ADJ
aiti-14589	8	27	cues	cue	NOUN
aiti-14589	8	28	.	.	PUNCT
aiti-14589	9	1	although	although	SCONJ
aiti-14589	9	2	humans	human	NOUN
aiti-14589	9	3	naturally	naturally	ADV
aiti-14589	9	4	recognize	recognize	VERB
aiti-14589	9	5	expressions	expression	NOUN
aiti-14589	9	6	,	,	PUNCT
aiti-14589	9	7	training	training	NOUN
aiti-14589	9	8	machines	machine	NOUN
aiti-14589	9	9	to	to	PART
aiti-14589	9	10	do	do	VERB
aiti-14589	9	11	the	the	DET
aiti-14589	9	12	same	same	ADJ
aiti-14589	9	13	continues	continue	VERB
aiti-14589	9	14	to	to	PART
aiti-14589	9	15	be	be	AUX
aiti-14589	9	16	a	a	DET
aiti-14589	9	17	challenge	challenge	NOUN
aiti-14589	9	18	.	.	PUNCT
aiti-14589	10	1	with	with	ADP
aiti-14589	10	2	the	the	DET
aiti-14589	10	3	rise	rise	NOUN
aiti-14589	10	4	of	of	ADP
aiti-14589	10	5	human	human	ADJ
aiti-14589	10	6	-	-	PUNCT
aiti-14589	10	7	computer	computer	NOUN
aiti-14589	10	8	interaction	interaction	NOUN
aiti-14589	10	9	,	,	PUNCT
aiti-14589	10	10	emotionally	emotionally	ADV
aiti-14589	10	11	intelligent	intelligent	ADJ
aiti-14589	10	12	technologies	technology	NOUN
aiti-14589	10	13	are	be	AUX
aiti-14589	10	14	becoming	become	VERB
aiti-14589	10	15	increasingly	increasingly	ADV
aiti-14589	10	16	important	important	ADJ
aiti-14589	10	17	.	.	PUNCT
aiti-14589	11	1	fer	fer	PROPN
aiti-14589	12	1	[	[	X
aiti-14589	12	2	1	1	NUM
aiti-14589	12	3	]	]	PUNCT
aiti-14589	12	4	is	be	AUX
aiti-14589	12	5	an	an	DET
aiti-14589	12	6	interdisciplinary	interdisciplinary	ADJ
aiti-14589	12	7	field	field	NOUN
aiti-14589	12	8	that	that	PRON
aiti-14589	12	9	bridges	bridge	VERB
aiti-14589	12	10	computer	computer	NOUN
aiti-14589	12	11	vision	vision	NOUN
aiti-14589	12	12	[	[	X
aiti-14589	12	13	2	2	NUM
aiti-14589	12	14	]	]	PUNCT
aiti-14589	12	15	,	,	PUNCT
aiti-14589	12	16	artificial	artificial	ADJ
aiti-14589	12	17	intelligence	intelligence	NOUN
aiti-14589	13	1	[	[	X
aiti-14589	13	2	3	3	NUM
aiti-14589	13	3	]	]	PUNCT
aiti-14589	13	4	,	,	PUNCT
aiti-14589	13	5	and	and	CCONJ
aiti-14589	13	6	psychology	psychology	NOUN
aiti-14589	13	7	[	[	X
aiti-14589	13	8	4	4	NUM
aiti-14589	13	9	]	]	PUNCT
aiti-14589	13	10	,	,	PUNCT
aiti-14589	13	11	aiming	aim	VERB
aiti-14589	13	12	to	to	PART
aiti-14589	13	13	decode	decode	VERB
aiti-14589	13	14	the	the	DET
aiti-14589	13	15	intricate	intricate	ADJ
aiti-14589	13	16	language	language	NOUN
aiti-14589	13	17	of	of	ADP
aiti-14589	13	18	human	human	ADJ
aiti-14589	13	19	emotions	emotion	NOUN
aiti-14589	13	20	conveyed	convey	VERB
aiti-14589	13	21	through	through	ADP
aiti-14589	13	22	facial	facial	ADJ
aiti-14589	13	23	expressions	expression	NOUN
aiti-14589	13	24	.	.	PUNCT
aiti-14589	14	1	facial	facial	ADJ
aiti-14589	14	2	expressions	expression	NOUN
aiti-14589	14	3	,	,	PUNCT
aiti-14589	14	4	a	a	DET
aiti-14589	14	5	fundamental	fundamental	ADJ
aiti-14589	14	6	mode	mode	NOUN
aiti-14589	14	7	of	of	ADP
aiti-14589	14	8	nonverbal	nonverbal	ADJ
aiti-14589	14	9	communication	communication	NOUN
aiti-14589	14	10	[	[	X
aiti-14589	14	11	5	5	NUM
aiti-14589	14	12	]	]	PUNCT
aiti-14589	14	13	,	,	PUNCT
aiti-14589	14	14	help	help	VERB
aiti-14589	14	15	humans	human	NOUN
aiti-14589	14	16	express	express	VERB
aiti-14589	14	17	emotions	emotion	NOUN
aiti-14589	14	18	,	,	PUNCT
aiti-14589	14	19	intentions	intention	NOUN
aiti-14589	14	20	,	,	PUNCT
aiti-14589	14	21	and	and	CCONJ
aiti-14589	14	22	perceptions	perception	NOUN
aiti-14589	14	23	.	.	PUNCT
aiti-14589	15	1	while	while	SCONJ
aiti-14589	15	2	humans	human	NOUN
aiti-14589	15	3	can	can	AUX
aiti-14589	15	4	naturally	naturally	ADV
aiti-14589	15	5	interpret	interpret	VERB
aiti-14589	15	6	these	these	DET
aiti-14589	15	7	expressions	expression	NOUN
aiti-14589	15	8	,	,	PUNCT
aiti-14589	15	9	enabling	enable	VERB
aiti-14589	15	10	machines	machine	NOUN
aiti-14589	15	11	to	to	PART
aiti-14589	15	12	interpret	interpret	VERB
aiti-14589	15	13	them	they	PRON
aiti-14589	15	14	remains	remain	VERB
aiti-14589	15	15	a	a	DET
aiti-14589	15	16	significant	significant	ADJ
aiti-14589	15	17	challenge	challenge	NOUN
aiti-14589	15	18	[	[	X
aiti-14589	15	19	6	6	NUM
aiti-14589	15	20	]	]	PUNCT
aiti-14589	15	21	.	.	PUNCT
aiti-14589	16	1	with	with	ADP
aiti-14589	16	2	the	the	DET
aiti-14589	16	3	growing	grow	VERB
aiti-14589	16	4	importance	importance	NOUN
aiti-14589	16	5	of	of	ADP
aiti-14589	16	6	human	human	ADJ
aiti-14589	16	7	-	-	PUNCT
aiti-14589	16	8	computer	computer	NOUN
aiti-14589	16	9	interaction	interaction	NOUN
aiti-14589	16	10	,	,	PUNCT
aiti-14589	16	11	the	the	DET
aiti-14589	16	12	need	need	NOUN
aiti-14589	16	13	for	for	ADP
aiti-14589	16	14	emotionally	emotionally	ADV
aiti-14589	16	15	intelligent	intelligent	ADJ
aiti-14589	16	16	technologies	technology	NOUN
aiti-14589	16	17	capable	capable	ADJ
aiti-14589	16	18	of	of	ADP
aiti-14589	16	19	recognizing	recognize	VERB
aiti-14589	16	20	and	and	CCONJ
aiti-14589	16	21	responding	respond	VERB
aiti-14589	16	22	to	to	ADP
aiti-14589	16	23	human	human	ADJ
aiti-14589	16	24	emotions	emotion	NOUN
aiti-14589	16	25	is	be	AUX
aiti-14589	16	26	increasingly	increasingly	ADV
aiti-14589	16	27	apparent	apparent	ADJ
aiti-14589	16	28	.	.	PUNCT
aiti-14589	17	1	image	image	NOUN
aiti-14589	17	2	-	-	PUNCT
aiti-14589	17	3	based	base	VERB
aiti-14589	17	4	methods	method	NOUN
aiti-14589	17	5	leverage	leverage	VERB
aiti-14589	17	6	the	the	DET
aiti-14589	17	7	transformative	transformative	ADJ
aiti-14589	17	8	power	power	NOUN
aiti-14589	17	9	of	of	ADP
aiti-14589	17	10	deep	deep	ADJ
aiti-14589	17	11	learning	learning	NOUN
aiti-14589	17	12	[	[	X
aiti-14589	17	13	7	7	NUM
aiti-14589	17	14	]	]	PUNCT
aiti-14589	17	15	,	,	PUNCT
aiti-14589	17	16	with	with	ADP
aiti-14589	17	17	convolutional	convolutional	ADJ
aiti-14589	17	18	neural	neural	ADJ
aiti-14589	17	19	networks	network	NOUN
aiti-14589	17	20	(	(	PUNCT
aiti-14589	17	21	cnn	cnn	PROPN
aiti-14589	17	22	)	)	PUNCT
aiti-14589	17	23	at	at	ADP
aiti-14589	17	24	the	the	DET
aiti-14589	17	25	forefront	forefront	NOUN
aiti-14589	17	26	.	.	PUNCT
aiti-14589	18	1	cnn	cnn	PROPN
aiti-14589	18	2	models	model	NOUN
aiti-14589	18	3	,	,	PUNCT
aiti-14589	18	4	trained	train	VERB
aiti-14589	18	5	on	on	ADP
aiti-14589	18	6	large	large	ADJ
aiti-14589	18	7	-	-	PUNCT
aiti-14589	18	8	scale	scale	NOUN
aiti-14589	18	9	datasets	dataset	NOUN
aiti-14589	18	10	like	like	ADP
aiti-14589	18	11	the	the	DET
aiti-14589	18	12	delaware	delaware	PROPN
aiti-14589	18	13	pain	pain	PROPN
aiti-14589	18	14	database	database	NOUN
aiti-14589	18	15	[	[	X
aiti-14589	18	16	8	8	NUM
aiti-14589	18	17	]	]	PUNCT
aiti-14589	18	18	and	and	CCONJ
aiti-14589	18	19	utkface	utkface	NOUN
aiti-14589	18	20	,	,	PUNCT
aiti-14589	18	21	excel	excel	VERB
aiti-14589	18	22	at	at	ADP
aiti-14589	18	23	capturing	capture	VERB
aiti-14589	18	24	global	global	ADJ
aiti-14589	18	25	and	and	CCONJ
aiti-14589	18	26	local	local	ADJ
aiti-14589	18	27	features	feature	NOUN
aiti-14589	18	28	from	from	ADP
aiti-14589	18	29	facial	facial	ADJ
aiti-14589	18	30	images	image	NOUN
aiti-14589	18	31	,	,	PUNCT
aiti-14589	18	32	enabling	enable	VERB
aiti-14589	18	33	robust	robust	ADJ
aiti-14589	18	34	emotion	emotion	NOUN
aiti-14589	18	35	recognition	recognition	NOUN
aiti-14589	18	36	.	.	PUNCT
aiti-14589	19	1	in	in	ADP
aiti-14589	19	2	contrast	contrast	NOUN
aiti-14589	19	3	,	,	PUNCT
aiti-14589	19	4	landmark	landmark	NOUN
aiti-14589	19	5	-	-	PUNCT
aiti-14589	19	6	based	base	VERB
aiti-14589	19	7	learning	learning	NOUN
aiti-14589	19	8	[	[	X
aiti-14589	19	9	9	9	NUM
aiti-14589	19	10	]	]	PUNCT
aiti-14589	19	11	focuses	focus	VERB
aiti-14589	19	12	on	on	ADP
aiti-14589	19	13	extracting	extract	VERB
aiti-14589	19	14	key	key	ADJ
aiti-14589	19	15	facial	facial	ADJ
aiti-14589	19	16	points	point	NOUN
aiti-14589	19	17	,	,	PUNCT
aiti-14589	19	18	such	such	ADJ
aiti-14589	19	19	as	as	ADP
aiti-14589	19	20	the	the	DET
aiti-14589	19	21	eyes	eye	NOUN
aiti-14589	19	22	,	,	PUNCT
aiti-14589	19	23	nose	nose	NOUN
aiti-14589	19	24	,	,	PUNCT
aiti-14589	19	25	and	and	CCONJ
aiti-14589	19	26	mouth	mouth	NOUN
aiti-14589	19	27	,	,	PUNCT
aiti-14589	19	28	as	as	ADP
aiti-14589	19	29	representative	representative	ADJ
aiti-14589	19	30	features	feature	NOUN
aiti-14589	19	31	of	of	ADP
aiti-14589	19	32	facial	facial	ADJ
aiti-14589	19	33	expressions	expression	NOUN
aiti-14589	19	34	.	.	PUNCT
aiti-14589	20	1	these	these	DET
aiti-14589	20	2	methods	method	NOUN
aiti-14589	20	3	utilize	utilize	VERB
aiti-14589	20	4	geometric	geometric	ADJ
aiti-14589	20	5	relationships	relationship	NOUN
aiti-14589	20	6	between	between	ADP
aiti-14589	20	7	landmarks	landmark	NOUN
aiti-14589	20	8	,	,	PUNCT
aiti-14589	20	9	offering	offer	VERB
aiti-14589	20	10	robust	robust	ADJ
aiti-14589	20	11	models	model	NOUN
aiti-14589	20	12	for	for	ADP
aiti-14589	20	13	emotion	emotion	NOUN
aiti-14589	20	14	classification	classification	NOUN
aiti-14589	20	15	under	under	ADP
aiti-14589	20	16	varying	vary	VERB
aiti-14589	20	17	poses	pose	NOUN
aiti-14589	20	18	and	and	CCONJ
aiti-14589	20	19	lighting	lighting	NOUN
aiti-14589	20	20	conditions	condition	NOUN
aiti-14589	20	21	.	.	PUNCT
aiti-14589	21	1	*	*	PUNCT
aiti-14589	21	2	corresponding	correspond	VERB
aiti-14589	21	3	author	author	NOUN
aiti-14589	21	4	.	.	PUNCT
aiti-14589	22	1	e	e	X
aiti-14589	22	2	-	-	NOUN
aiti-14589	22	3	mail	mail	NOUN
aiti-14589	22	4	address	address	NOUN
aiti-14589	22	5	:	:	PUNCT
aiti-14589	22	6	sarunya.k@sut.ac.th	sarunya.k@sut.ac.th	NOUN
aiti-14589	22	7	advances	advance	VERB
aiti-14589	22	8	in	in	ADP
aiti-14589	22	9	technology	technology	NOUN
aiti-14589	22	10	innovation	innovation	NOUN
aiti-14589	22	11	,	,	PUNCT
aiti-14589	22	12	vol	vol	NOUN
aiti-14589	22	13	.	.	PROPN
aiti-14589	23	1	10	10	NUM
aiti-14589	23	2	,	,	PUNCT
aiti-14589	23	3	no	no	INTJ
aiti-14589	23	4	.	.	NOUN
aiti-14589	23	5	4	4	NUM
aiti-14589	23	6	,	,	PUNCT
aiti-14589	23	7	2025	2025	NUM
aiti-14589	23	8	,	,	PUNCT
aiti-14589	23	9	pp	pp	ADJ
aiti-14589	23	10	.	.	PUNCT
aiti-14589	24	1	395	395	NUM
aiti-14589	24	2	-	-	SYM
aiti-14589	24	3	406	406	NUM
aiti-14589	24	4	396	396	NUM
aiti-14589	24	5	while	while	SCONJ
aiti-14589	24	6	both	both	PRON
aiti-14589	24	7	image	image	NOUN
aiti-14589	24	8	-	-	PUNCT
aiti-14589	24	9	based	base	VERB
aiti-14589	24	10	[	[	X
aiti-14589	24	11	10	10	NUM
aiti-14589	24	12	]	]	PUNCT
aiti-14589	24	13	and	and	CCONJ
aiti-14589	24	14	landmark	landmark	NOUN
aiti-14589	24	15	-	-	PUNCT
aiti-14589	24	16	based	base	VERB
aiti-14589	24	17	methods	method	NOUN
aiti-14589	24	18	[	[	X
aiti-14589	24	19	11	11	NUM
aiti-14589	24	20	]	]	PUNCT
aiti-14589	24	21	for	for	ADP
aiti-14589	24	22	fer	fer	PROPN
aiti-14589	24	23	have	have	AUX
aiti-14589	24	24	been	be	AUX
aiti-14589	24	25	extensively	extensively	ADV
aiti-14589	24	26	explored	explore	VERB
aiti-14589	24	27	,	,	PUNCT
aiti-14589	24	28	there	there	PRON
aiti-14589	24	29	has	have	AUX
aiti-14589	24	30	been	be	AUX
aiti-14589	24	31	limited	limit	VERB
aiti-14589	24	32	research	research	NOUN
aiti-14589	24	33	directly	directly	ADV
aiti-14589	24	34	comparing	compare	VERB
aiti-14589	24	35	their	their	PRON
aiti-14589	24	36	effectiveness	effectiveness	NOUN
aiti-14589	24	37	across	across	ADP
aiti-14589	24	38	diverse	diverse	ADJ
aiti-14589	24	39	facial	facial	ADJ
aiti-14589	24	40	expression	expression	NOUN
aiti-14589	24	41	datasets	dataset	NOUN
aiti-14589	24	42	.	.	PUNCT
aiti-14589	25	1	most	most	ADJ
aiti-14589	25	2	studies	study	NOUN
aiti-14589	25	3	tend	tend	VERB
aiti-14589	25	4	to	to	PART
aiti-14589	25	5	focus	focus	VERB
aiti-14589	25	6	on	on	ADP
aiti-14589	25	7	one	one	NUM
aiti-14589	25	8	approach	approach	NOUN
aiti-14589	25	9	,	,	PUNCT
aiti-14589	25	10	without	without	ADP
aiti-14589	25	11	offering	offer	VERB
aiti-14589	25	12	a	a	DET
aiti-14589	25	13	clear	clear	ADJ
aiti-14589	25	14	comparison	comparison	NOUN
aiti-14589	25	15	that	that	PRON
aiti-14589	25	16	highlights	highlight	VERB
aiti-14589	25	17	the	the	DET
aiti-14589	25	18	relative	relative	ADJ
aiti-14589	25	19	advantages	advantage	NOUN
aiti-14589	25	20	and	and	CCONJ
aiti-14589	25	21	drawbacks	drawback	NOUN
aiti-14589	25	22	of	of	ADP
aiti-14589	25	23	each	each	PRON
aiti-14589	25	24	.	.	PUNCT
aiti-14589	26	1	this	this	DET
aiti-14589	26	2	study	study	NOUN
aiti-14589	26	3	addresses	address	VERB
aiti-14589	26	4	that	that	DET
aiti-14589	26	5	gap	gap	NOUN
aiti-14589	26	6	by	by	ADP
aiti-14589	26	7	conducting	conduct	VERB
aiti-14589	26	8	a	a	DET
aiti-14589	26	9	direct	direct	ADJ
aiti-14589	26	10	comparison	comparison	NOUN
aiti-14589	26	11	,	,	PUNCT
aiti-14589	26	12	providing	provide	VERB
aiti-14589	26	13	valuable	valuable	ADJ
aiti-14589	26	14	insights	insight	NOUN
aiti-14589	26	15	into	into	ADP
aiti-14589	26	16	which	which	DET
aiti-14589	26	17	method	method	NOUN
aiti-14589	26	18	performs	perform	VERB
aiti-14589	26	19	better	well	ADV
aiti-14589	26	20	under	under	ADP
aiti-14589	26	21	various	various	ADJ
aiti-14589	26	22	conditions	condition	NOUN
aiti-14589	26	23	.	.	PUNCT
aiti-14589	27	1	the	the	DET
aiti-14589	27	2	scarcity	scarcity	NOUN
aiti-14589	27	3	of	of	ADP
aiti-14589	27	4	such	such	ADJ
aiti-14589	27	5	comparisons	comparison	NOUN
aiti-14589	27	6	can	can	AUX
aiti-14589	27	7	be	be	AUX
aiti-14589	27	8	attributed	attribute	VERB
aiti-14589	27	9	to	to	ADP
aiti-14589	27	10	the	the	DET
aiti-14589	27	11	complexity	complexity	NOUN
aiti-14589	27	12	and	and	CCONJ
aiti-14589	27	13	variation	variation	NOUN
aiti-14589	27	14	in	in	ADP
aiti-14589	27	15	how	how	SCONJ
aiti-14589	27	16	facial	facial	ADJ
aiti-14589	27	17	expressions	expression	NOUN
aiti-14589	27	18	are	be	AUX
aiti-14589	27	19	captured	capture	VERB
aiti-14589	27	20	and	and	CCONJ
aiti-14589	27	21	analyzed	analyze	VERB
aiti-14589	27	22	.	.	PUNCT
aiti-14589	28	1	image	image	NOUN
aiti-14589	28	2	-	-	PUNCT
aiti-14589	28	3	based	base	VERB
aiti-14589	28	4	methods	method	NOUN
aiti-14589	28	5	typically	typically	ADV
aiti-14589	28	6	process	process	VERB
aiti-14589	28	7	raw	raw	ADJ
aiti-14589	28	8	data	datum	NOUN
aiti-14589	28	9	and	and	CCONJ
aiti-14589	28	10	rely	rely	VERB
aiti-14589	28	11	on	on	ADP
aiti-14589	28	12	deep	deep	ADJ
aiti-14589	28	13	learning	learning	NOUN
aiti-14589	28	14	for	for	ADP
aiti-14589	28	15	feature	feature	NOUN
aiti-14589	28	16	extraction	extraction	NOUN
aiti-14589	28	17	.	.	PUNCT
aiti-14589	29	1	in	in	ADP
aiti-14589	29	2	contrast	contrast	NOUN
aiti-14589	29	3	,	,	PUNCT
aiti-14589	29	4	landmark	landmark	NOUN
aiti-14589	29	5	-	-	PUNCT
aiti-14589	29	6	based	base	VERB
aiti-14589	29	7	methods	method	NOUN
aiti-14589	29	8	extract	extract	VERB
aiti-14589	29	9	key	key	ADJ
aiti-14589	29	10	facial	facial	ADJ
aiti-14589	29	11	points	point	NOUN
aiti-14589	29	12	,	,	PUNCT
aiti-14589	29	13	offering	offer	VERB
aiti-14589	29	14	a	a	DET
aiti-14589	29	15	more	more	ADV
aiti-14589	29	16	computationally	computationally	ADV
aiti-14589	29	17	efficient	efficient	ADJ
aiti-14589	29	18	approach	approach	NOUN
aiti-14589	29	19	.	.	PUNCT
aiti-14589	30	1	by	by	ADP
aiti-14589	30	2	combining	combine	VERB
aiti-14589	30	3	both	both	DET
aiti-14589	30	4	methodologies	methodology	NOUN
aiti-14589	30	5	in	in	ADP
aiti-14589	30	6	one	one	NUM
aiti-14589	30	7	study	study	NOUN
aiti-14589	30	8	,	,	PUNCT
aiti-14589	30	9	this	this	DET
aiti-14589	30	10	research	research	NOUN
aiti-14589	30	11	aims	aim	VERB
aiti-14589	30	12	to	to	PART
aiti-14589	30	13	uncover	uncover	VERB
aiti-14589	30	14	their	their	PRON
aiti-14589	30	15	complementary	complementary	ADJ
aiti-14589	30	16	strengths	strength	NOUN
aiti-14589	30	17	and	and	CCONJ
aiti-14589	30	18	limitations	limitation	NOUN
aiti-14589	30	19	.	.	PUNCT
aiti-14589	31	1	despite	despite	SCONJ
aiti-14589	31	2	progress	progress	NOUN
aiti-14589	31	3	in	in	ADP
aiti-14589	31	4	both	both	DET
aiti-14589	31	5	areas	area	NOUN
aiti-14589	31	6	,	,	PUNCT
aiti-14589	31	7	significant	significant	ADJ
aiti-14589	31	8	gaps	gap	NOUN
aiti-14589	31	9	remain	remain	VERB
aiti-14589	31	10	in	in	ADP
aiti-14589	31	11	the	the	DET
aiti-14589	31	12	literature	literature	NOUN
aiti-14589	31	13	.	.	PUNCT
aiti-14589	32	1	most	most	ADJ
aiti-14589	32	2	research	research	NOUN
aiti-14589	32	3	focuses	focus	VERB
aiti-14589	32	4	exclusively	exclusively	ADV
aiti-14589	32	5	on	on	ADP
aiti-14589	32	6	either	either	CCONJ
aiti-14589	32	7	image	image	NOUN
aiti-14589	32	8	-	-	PUNCT
aiti-14589	32	9	based	base	VERB
aiti-14589	32	10	or	or	CCONJ
aiti-14589	32	11	landmark	landmark	NOUN
aiti-14589	32	12	-	-	PUNCT
aiti-14589	32	13	based	base	VERB
aiti-14589	32	14	techniques	technique	NOUN
aiti-14589	32	15	,	,	PUNCT
aiti-14589	32	16	without	without	ADP
aiti-14589	32	17	comparing	compare	VERB
aiti-14589	32	18	their	their	PRON
aiti-14589	32	19	effectiveness	effectiveness	NOUN
aiti-14589	32	20	across	across	ADP
aiti-14589	32	21	multiple	multiple	ADJ
aiti-14589	32	22	machine	machine	NOUN
aiti-14589	32	23	learning	learning	NOUN
aiti-14589	32	24	(	(	PUNCT
aiti-14589	32	25	ml	ml	NOUN
aiti-14589	32	26	)	)	PUNCT
aiti-14589	32	27	models	model	NOUN
aiti-14589	32	28	.	.	PUNCT
aiti-14589	33	1	furthermore	furthermore	ADV
aiti-14589	33	2	,	,	PUNCT
aiti-14589	33	3	few	few	ADJ
aiti-14589	33	4	studies	study	NOUN
aiti-14589	33	5	explore	explore	VERB
aiti-14589	33	6	the	the	DET
aiti-14589	33	7	practical	practical	ADJ
aiti-14589	33	8	application	application	NOUN
aiti-14589	33	9	of	of	ADP
aiti-14589	33	10	these	these	DET
aiti-14589	33	11	methods	method	NOUN
aiti-14589	33	12	in	in	ADP
aiti-14589	33	13	real	real	ADJ
aiti-14589	33	14	-	-	PUNCT
aiti-14589	33	15	world	world	NOUN
aiti-14589	33	16	contexts	contexts	NOUN
aiti-14589	33	17	,	,	PUNCT
aiti-14589	33	18	such	such	ADJ
aiti-14589	33	19	as	as	ADP
aiti-14589	33	20	rehabilitation	rehabilitation	NOUN
aiti-14589	33	21	and	and	CCONJ
aiti-14589	33	22	diagnostic	diagnostic	ADJ
aiti-14589	33	23	systems	system	NOUN
aiti-14589	33	24	.	.	PUNCT
aiti-14589	34	1	these	these	DET
aiti-14589	34	2	gaps	gap	NOUN
aiti-14589	34	3	highlight	highlight	VERB
aiti-14589	34	4	the	the	DET
aiti-14589	34	5	need	need	NOUN
aiti-14589	34	6	for	for	SCONJ
aiti-14589	34	7	a	a	DET
aiti-14589	34	8	thorough	thorough	ADJ
aiti-14589	34	9	evaluation	evaluation	NOUN
aiti-14589	34	10	of	of	ADP
aiti-14589	34	11	these	these	DET
aiti-14589	34	12	approaches	approach	NOUN
aiti-14589	34	13	to	to	PART
aiti-14589	34	14	better	well	ADV
aiti-14589	34	15	understand	understand	VERB
aiti-14589	34	16	their	their	PRON
aiti-14589	34	17	strengths	strength	NOUN
aiti-14589	34	18	and	and	CCONJ
aiti-14589	34	19	limitations	limitation	NOUN
aiti-14589	34	20	in	in	ADP
aiti-14589	34	21	practical	practical	ADJ
aiti-14589	34	22	applications	application	NOUN
aiti-14589	34	23	.	.	PUNCT
aiti-14589	35	1	fer	fer	PROPN
aiti-14589	35	2	is	be	AUX
aiti-14589	35	3	a	a	DET
aiti-14589	35	4	crucial	crucial	ADJ
aiti-14589	35	5	task	task	NOUN
aiti-14589	35	6	in	in	ADP
aiti-14589	35	7	affective	affective	NOUN
aiti-14589	35	8	computing	computing	NOUN
aiti-14589	35	9	and	and	CCONJ
aiti-14589	35	10	human	human	ADJ
aiti-14589	35	11	-	-	PUNCT
aiti-14589	35	12	computer	computer	NOUN
aiti-14589	35	13	interaction	interaction	NOUN
aiti-14589	35	14	,	,	PUNCT
aiti-14589	35	15	with	with	ADP
aiti-14589	35	16	applications	application	NOUN
aiti-14589	35	17	in	in	ADP
aiti-14589	35	18	healthcare	healthcare	NOUN
aiti-14589	35	19	and	and	CCONJ
aiti-14589	35	20	emotional	emotional	ADJ
aiti-14589	35	21	ai	ai	NOUN
aiti-14589	35	22	.	.	PUNCT
aiti-14589	36	1	while	while	SCONJ
aiti-14589	36	2	existing	exist	VERB
aiti-14589	36	3	fer	fer	PROPN
aiti-14589	36	4	techniques	technique	NOUN
aiti-14589	36	5	leverage	leverage	NOUN
aiti-14589	36	6	both	both	DET
aiti-14589	36	7	image	image	NOUN
aiti-14589	36	8	-	-	PUNCT
aiti-14589	36	9	based	base	VERB
aiti-14589	36	10	and	and	CCONJ
aiti-14589	36	11	landmark	landmark	NOUN
aiti-14589	36	12	-	-	PUNCT
aiti-14589	36	13	based	base	VERB
aiti-14589	36	14	approaches	approach	NOUN
aiti-14589	36	15	,	,	PUNCT
aiti-14589	36	16	there	there	PRON
aiti-14589	36	17	remains	remain	VERB
aiti-14589	36	18	a	a	DET
aiti-14589	36	19	gap	gap	NOUN
aiti-14589	36	20	in	in	ADP
aiti-14589	36	21	systematically	systematically	ADV
aiti-14589	36	22	comparing	compare	VERB
aiti-14589	36	23	these	these	DET
aiti-14589	36	24	methods	method	NOUN
aiti-14589	36	25	under	under	ADP
aiti-14589	36	26	standardized	standardized	ADJ
aiti-14589	36	27	conditions	condition	NOUN
aiti-14589	36	28	using	use	VERB
aiti-14589	36	29	diverse	diverse	ADJ
aiti-14589	36	30	ml	ml	NOUN
aiti-14589	36	31	models	model	NOUN
aiti-14589	36	32	.	.	PUNCT
aiti-14589	37	1	this	this	DET
aiti-14589	37	2	study	study	NOUN
aiti-14589	37	3	aims	aim	VERB
aiti-14589	37	4	to	to	PART
aiti-14589	37	5	bridge	bridge	VERB
aiti-14589	37	6	this	this	DET
aiti-14589	37	7	gap	gap	NOUN
aiti-14589	37	8	by	by	ADP
aiti-14589	37	9	conducting	conduct	VERB
aiti-14589	37	10	a	a	DET
aiti-14589	37	11	comprehensive	comprehensive	ADJ
aiti-14589	37	12	comparison	comparison	NOUN
aiti-14589	37	13	of	of	ADP
aiti-14589	37	14	image	image	NOUN
aiti-14589	37	15	-	-	PUNCT
aiti-14589	37	16	based	base	VERB
aiti-14589	37	17	and	and	CCONJ
aiti-14589	37	18	landmark	landmark	NOUN
aiti-14589	37	19	-	-	PUNCT
aiti-14589	37	20	based	base	VERB
aiti-14589	37	21	fer	fer	PROPN
aiti-14589	37	22	techniques	technique	NOUN
aiti-14589	37	23	using	use	VERB
aiti-14589	37	24	five	five	NUM
aiti-14589	37	25	distinct	distinct	ADJ
aiti-14589	37	26	ml	ml	NOUN
aiti-14589	37	27	models	model	NOUN
aiti-14589	37	28	:	:	PUNCT
aiti-14589	38	1	cnn	cnn	PROPN
aiti-14589	39	1	[	[	X
aiti-14589	39	2	12	12	NUM
aiti-14589	39	3	]	]	PUNCT
aiti-14589	39	4	,	,	PUNCT
aiti-14589	39	5	logistic	logistic	ADJ
aiti-14589	39	6	regression	regression	NOUN
aiti-14589	39	7	classifier	classifier	NOUN
aiti-14589	39	8	(	(	PUNCT
aiti-14589	39	9	lrc	lrc	PROPN
aiti-14589	39	10	)	)	PUNCT
aiti-14589	40	1	[	[	X
aiti-14589	40	2	13	13	NUM
aiti-14589	40	3	]	]	PUNCT
aiti-14589	40	4	,	,	PUNCT
aiti-14589	40	5	support	support	NOUN
aiti-14589	40	6	vector	vector	NOUN
aiti-14589	40	7	machines	machine	NOUN
aiti-14589	40	8	(	(	PUNCT
aiti-14589	40	9	svm	svm	PROPN
aiti-14589	40	10	)	)	PUNCT
aiti-14589	41	1	[	[	X
aiti-14589	41	2	14	14	NUM
aiti-14589	41	3	]	]	X
aiti-14589	41	4	,	,	PUNCT
aiti-14589	41	5	random	random	ADJ
aiti-14589	41	6	forest	forest	NOUN
aiti-14589	41	7	classifier	classifier	NOUN
aiti-14589	41	8	(	(	PUNCT
aiti-14589	41	9	rfc	rfc	PROPN
aiti-14589	41	10	)	)	PUNCT
aiti-14589	42	1	[	[	X
aiti-14589	42	2	15	15	NUM
aiti-14589	42	3	]	]	PUNCT
aiti-14589	42	4	,	,	PUNCT
aiti-14589	42	5	and	and	CCONJ
aiti-14589	42	6	gradient	gradient	NOUN
aiti-14589	42	7	boosting	boost	VERB
aiti-14589	42	8	classifier	classifier	NOUN
aiti-14589	42	9	(	(	PUNCT
aiti-14589	42	10	gbc	gbc	PROPN
aiti-14589	42	11	)	)	PUNCT
aiti-14589	43	1	[	[	X
aiti-14589	43	2	16	16	NUM
aiti-14589	43	3	]	]	PUNCT
aiti-14589	43	4	.	.	PUNCT
aiti-14589	44	1	a	a	DET
aiti-14589	44	2	key	key	ADJ
aiti-14589	44	3	aspect	aspect	NOUN
aiti-14589	44	4	of	of	ADP
aiti-14589	44	5	this	this	DET
aiti-14589	44	6	research	research	NOUN
aiti-14589	44	7	is	be	AUX
aiti-14589	44	8	the	the	DET
aiti-14589	44	9	evaluation	evaluation	NOUN
aiti-14589	44	10	of	of	ADP
aiti-14589	44	11	image	image	NOUN
aiti-14589	44	12	-	-	PUNCT
aiti-14589	44	13	based	base	VERB
aiti-14589	44	14	features	feature	NOUN
aiti-14589	44	15	versus	versus	ADP
aiti-14589	44	16	landmark	landmark	NOUN
aiti-14589	44	17	-	-	PUNCT
aiti-14589	44	18	based	base	VERB
aiti-14589	44	19	feature	feature	NOUN
aiti-14589	44	20	extraction	extraction	NOUN
aiti-14589	44	21	,	,	PUNCT
aiti-14589	44	22	assessing	assess	VERB
aiti-14589	44	23	their	their	PRON
aiti-14589	44	24	effectiveness	effectiveness	NOUN
aiti-14589	44	25	in	in	ADP
aiti-14589	44	26	recognizing	recognize	VERB
aiti-14589	44	27	both	both	CCONJ
aiti-14589	44	28	normal	normal	ADJ
aiti-14589	44	29	and	and	CCONJ
aiti-14589	44	30	hurt	hurt	VERB
aiti-14589	44	31	emotions	emotion	NOUN
aiti-14589	44	32	.	.	PUNCT
aiti-14589	45	1	the	the	DET
aiti-14589	45	2	specific	specific	ADJ
aiti-14589	45	3	objectives	objective	NOUN
aiti-14589	45	4	of	of	ADP
aiti-14589	45	5	this	this	DET
aiti-14589	45	6	study	study	NOUN
aiti-14589	45	7	include	include	VERB
aiti-14589	45	8	:	:	PUNCT
aiti-14589	45	9	(	(	PUNCT
aiti-14589	45	10	1	1	X
aiti-14589	45	11	)	)	PUNCT
aiti-14589	45	12	comparing	compare	VERB
aiti-14589	45	13	the	the	DET
aiti-14589	45	14	performance	performance	NOUN
aiti-14589	45	15	of	of	ADP
aiti-14589	45	16	image	image	NOUN
aiti-14589	45	17	-	-	PUNCT
aiti-14589	45	18	based	base	VERB
aiti-14589	45	19	and	and	CCONJ
aiti-14589	45	20	landmark	landmark	NOUN
aiti-14589	45	21	-	-	PUNCT
aiti-14589	45	22	based	base	VERB
aiti-14589	45	23	fer	fer	PROPN
aiti-14589	45	24	approaches	approach	NOUN
aiti-14589	45	25	using	use	VERB
aiti-14589	45	26	standardized	standardized	ADJ
aiti-14589	45	27	datasets	dataset	NOUN
aiti-14589	45	28	.	.	PUNCT
aiti-14589	46	1	(	(	PUNCT
aiti-14589	46	2	2	2	X
aiti-14589	46	3	)	)	PUNCT
aiti-14589	46	4	assessing	assess	VERB
aiti-14589	46	5	the	the	DET
aiti-14589	46	6	strengths	strength	NOUN
aiti-14589	46	7	and	and	CCONJ
aiti-14589	46	8	weaknesses	weakness	NOUN
aiti-14589	46	9	of	of	ADP
aiti-14589	46	10	cnn	cnn	PROPN
aiti-14589	46	11	,	,	PUNCT
aiti-14589	46	12	lrc	lrc	PROPN
aiti-14589	46	13	,	,	PUNCT
aiti-14589	46	14	svm	svm	PROPN
aiti-14589	46	15	,	,	PUNCT
aiti-14589	46	16	rfc	rfc	PRON
aiti-14589	46	17	,	,	PUNCT
aiti-14589	46	18	and	and	CCONJ
aiti-14589	46	19	gbc	gbc	PROPN
aiti-14589	46	20	in	in	ADP
aiti-14589	46	21	processing	process	VERB
aiti-14589	46	22	different	different	ADJ
aiti-14589	46	23	facial	facial	ADJ
aiti-14589	46	24	features	feature	NOUN
aiti-14589	46	25	.	.	PUNCT
aiti-14589	47	1	(	(	PUNCT
aiti-14589	47	2	3	3	X
aiti-14589	47	3	)	)	PUNCT
aiti-14589	47	4	measuring	measure	VERB
aiti-14589	47	5	accuracy	accuracy	NOUN
aiti-14589	47	6	,	,	PUNCT
aiti-14589	47	7	precision	precision	NOUN
aiti-14589	47	8	,	,	PUNCT
aiti-14589	47	9	recall	recall	NOUN
aiti-14589	47	10	,	,	PUNCT
aiti-14589	47	11	and	and	CCONJ
aiti-14589	47	12	f1	f1	NOUN
aiti-14589	47	13	-	-	PUNCT
aiti-14589	47	14	score	score	NOUN
aiti-14589	47	15	to	to	PART
aiti-14589	47	16	determine	determine	VERB
aiti-14589	47	17	the	the	DET
aiti-14589	47	18	robustness	robustness	NOUN
aiti-14589	47	19	and	and	CCONJ
aiti-14589	47	20	efficiency	efficiency	NOUN
aiti-14589	47	21	of	of	ADP
aiti-14589	47	22	each	each	DET
aiti-14589	47	23	approach	approach	NOUN
aiti-14589	47	24	.	.	PUNCT
aiti-14589	48	1	the	the	DET
aiti-14589	48	2	research	research	NOUN
aiti-14589	48	3	contributes	contribute	VERB
aiti-14589	48	4	to	to	ADP
aiti-14589	48	5	the	the	DET
aiti-14589	48	6	field	field	NOUN
aiti-14589	48	7	in	in	ADP
aiti-14589	48	8	several	several	ADJ
aiti-14589	48	9	ways	way	NOUN
aiti-14589	48	10	.	.	PUNCT
aiti-14589	49	1	firstly	firstly	ADV
aiti-14589	49	2	,	,	PUNCT
aiti-14589	49	3	it	it	PRON
aiti-14589	49	4	provides	provide	VERB
aiti-14589	49	5	a	a	DET
aiti-14589	49	6	direct	direct	ADJ
aiti-14589	49	7	comparison	comparison	NOUN
aiti-14589	49	8	of	of	ADP
aiti-14589	49	9	the	the	DET
aiti-14589	49	10	performance	performance	NOUN
aiti-14589	49	11	of	of	ADP
aiti-14589	49	12	imagebased	imagebase	VERB
aiti-14589	49	13	methods	method	NOUN
aiti-14589	49	14	,	,	PUNCT
aiti-14589	49	15	which	which	PRON
aiti-14589	49	16	use	use	VERB
aiti-14589	49	17	a	a	DET
aiti-14589	49	18	histogram	histogram	NOUN
aiti-14589	49	19	of	of	ADP
aiti-14589	49	20	oriented	orient	VERB
aiti-14589	49	21	gradients	gradient	NOUN
aiti-14589	49	22	(	(	PUNCT
aiti-14589	49	23	hog	hog	NOUN
aiti-14589	49	24	)	)	PUNCT
aiti-14589	49	25	features	feature	NOUN
aiti-14589	49	26	,	,	PUNCT
aiti-14589	49	27	and	and	CCONJ
aiti-14589	49	28	landmark	landmark	NOUN
aiti-14589	49	29	-	-	PUNCT
aiti-14589	49	30	based	base	VERB
aiti-14589	49	31	methods	method	NOUN
aiti-14589	49	32	,	,	PUNCT
aiti-14589	49	33	which	which	PRON
aiti-14589	49	34	rely	rely	VERB
aiti-14589	49	35	on	on	ADP
aiti-14589	49	36	facial	facial	ADJ
aiti-14589	49	37	landmarks	landmark	NOUN
aiti-14589	49	38	for	for	ADP
aiti-14589	49	39	feature	feature	NOUN
aiti-14589	49	40	extraction	extraction	NOUN
aiti-14589	49	41	.	.	PUNCT
aiti-14589	50	1	this	this	DET
aiti-14589	50	2	comparison	comparison	NOUN
aiti-14589	50	3	is	be	AUX
aiti-14589	50	4	based	base	VERB
aiti-14589	50	5	on	on	ADP
aiti-14589	50	6	a	a	DET
aiti-14589	50	7	large	large	ADJ
aiti-14589	50	8	and	and	CCONJ
aiti-14589	50	9	diverse	diverse	ADJ
aiti-14589	50	10	dataset	dataset	NOUN
aiti-14589	50	11	,	,	PUNCT
aiti-14589	50	12	including	include	VERB
aiti-14589	50	13	normal	normal	ADJ
aiti-14589	50	14	(	(	PUNCT
aiti-14589	50	15	neutral	neutral	ADJ
aiti-14589	50	16	)	)	PUNCT
aiti-14589	50	17	and	and	CCONJ
aiti-14589	50	18	hurt	hurt	VERB
aiti-14589	50	19	emotions	emotion	NOUN
aiti-14589	50	20	.	.	PUNCT
aiti-14589	51	1	secondly	secondly	ADV
aiti-14589	51	2	,	,	PUNCT
aiti-14589	51	3	the	the	DET
aiti-14589	51	4	study	study	NOUN
aiti-14589	51	5	assesses	assess	VERB
aiti-14589	51	6	the	the	DET
aiti-14589	51	7	performance	performance	NOUN
aiti-14589	51	8	of	of	ADP
aiti-14589	51	9	a	a	DET
aiti-14589	51	10	variety	variety	NOUN
aiti-14589	51	11	of	of	ADP
aiti-14589	51	12	widely	widely	ADV
aiti-14589	51	13	used	use	VERB
aiti-14589	51	14	ml	ml	NOUN
aiti-14589	51	15	algorithms	algorithm	NOUN
aiti-14589	51	16	,	,	PUNCT
aiti-14589	51	17	identifying	identify	VERB
aiti-14589	51	18	the	the	DET
aiti-14589	51	19	most	most	ADV
aiti-14589	51	20	effective	effective	ADJ
aiti-14589	51	21	models	model	NOUN
aiti-14589	51	22	for	for	ADP
aiti-14589	51	23	each	each	DET
aiti-14589	51	24	technique	technique	NOUN
aiti-14589	51	25	.	.	PUNCT
aiti-14589	52	1	this	this	DET
aiti-14589	52	2	evaluation	evaluation	NOUN
aiti-14589	52	3	serves	serve	VERB
aiti-14589	52	4	as	as	ADP
aiti-14589	52	5	a	a	DET
aiti-14589	52	6	comprehensive	comprehensive	ADJ
aiti-14589	52	7	guide	guide	NOUN
aiti-14589	52	8	for	for	ADP
aiti-14589	52	9	selecting	select	VERB
aiti-14589	52	10	the	the	DET
aiti-14589	52	11	optimal	optimal	ADJ
aiti-14589	52	12	method	method	NOUN
aiti-14589	52	13	for	for	ADP
aiti-14589	52	14	different	different	ADJ
aiti-14589	52	15	fer	fer	PROPN
aiti-14589	52	16	tasks	task	NOUN
aiti-14589	52	17	.	.	PUNCT
aiti-14589	53	1	lastly	lastly	ADV
aiti-14589	53	2	,	,	PUNCT
aiti-14589	53	3	the	the	DET
aiti-14589	53	4	study	study	NOUN
aiti-14589	53	5	employs	employ	VERB
aiti-14589	53	6	a	a	DET
aiti-14589	53	7	range	range	NOUN
aiti-14589	53	8	of	of	ADP
aiti-14589	53	9	performance	performance	NOUN
aiti-14589	53	10	metrics	metric	NOUN
aiti-14589	53	11	,	,	PUNCT
aiti-14589	53	12	including	include	VERB
aiti-14589	53	13	accuracy	accuracy	NOUN
aiti-14589	53	14	,	,	PUNCT
aiti-14589	53	15	precision	precision	NOUN
aiti-14589	53	16	,	,	PUNCT
aiti-14589	53	17	recall	recall	NOUN
aiti-14589	53	18	,	,	PUNCT
aiti-14589	53	19	and	and	CCONJ
aiti-14589	53	20	f1score	f1score	ADV
aiti-14589	53	21	,	,	PUNCT
aiti-14589	53	22	ensuring	ensure	VERB
aiti-14589	53	23	a	a	DET
aiti-14589	53	24	thorough	thorough	ADJ
aiti-14589	53	25	and	and	CCONJ
aiti-14589	53	26	balanced	balanced	ADJ
aiti-14589	53	27	assessment	assessment	NOUN
aiti-14589	53	28	of	of	ADP
aiti-14589	53	29	the	the	DET
aiti-14589	53	30	model	model	NOUN
aiti-14589	53	31	’s	’s	PART
aiti-14589	53	32	effectiveness	effectiveness	NOUN
aiti-14589	53	33	in	in	ADP
aiti-14589	53	34	fer	fer	PROPN
aiti-14589	53	35	.	.	PUNCT
aiti-14589	54	1	while	while	SCONJ
aiti-14589	54	2	both	both	DET
aiti-14589	54	3	approaches	approach	NOUN
aiti-14589	54	4	in	in	ADP
aiti-14589	54	5	this	this	DET
aiti-14589	54	6	study	study	NOUN
aiti-14589	54	7	offer	offer	VERB
aiti-14589	54	8	distinct	distinct	ADJ
aiti-14589	54	9	advantages	advantage	NOUN
aiti-14589	54	10	,	,	PUNCT
aiti-14589	54	11	they	they	PRON
aiti-14589	54	12	also	also	ADV
aiti-14589	54	13	come	come	VERB
aiti-14589	54	14	with	with	ADP
aiti-14589	54	15	limitations	limitation	NOUN
aiti-14589	54	16	that	that	PRON
aiti-14589	54	17	could	could	AUX
aiti-14589	54	18	affect	affect	VERB
aiti-14589	54	19	the	the	DET
aiti-14589	54	20	results	result	NOUN
aiti-14589	54	21	.	.	PUNCT
aiti-14589	55	1	the	the	DET
aiti-14589	55	2	image	image	NOUN
aiti-14589	55	3	-	-	PUNCT
aiti-14589	55	4	based	base	VERB
aiti-14589	55	5	approach	approach	NOUN
aiti-14589	55	6	,	,	PUNCT
aiti-14589	55	7	particularly	particularly	ADV
aiti-14589	55	8	when	when	SCONJ
aiti-14589	55	9	using	use	VERB
aiti-14589	55	10	deep	deep	ADJ
aiti-14589	55	11	learning	learning	NOUN
aiti-14589	55	12	models	model	NOUN
aiti-14589	55	13	like	like	ADP
aiti-14589	55	14	cnn	cnn	PROPN
aiti-14589	55	15	,	,	PUNCT
aiti-14589	55	16	demands	demand	VERB
aiti-14589	55	17	significant	significant	ADJ
aiti-14589	55	18	computational	computational	ADJ
aiti-14589	55	19	resources	resource	NOUN
aiti-14589	55	20	and	and	CCONJ
aiti-14589	55	21	may	may	AUX
aiti-14589	55	22	be	be	AUX
aiti-14589	55	23	sensitive	sensitive	ADJ
aiti-14589	55	24	to	to	ADP
aiti-14589	55	25	variations	variation	NOUN
aiti-14589	55	26	in	in	ADP
aiti-14589	55	27	image	image	NOUN
aiti-14589	55	28	quality	quality	NOUN
aiti-14589	55	29	and	and	CCONJ
aiti-14589	55	30	lighting	lighting	NOUN
aiti-14589	55	31	conditions	condition	NOUN
aiti-14589	55	32	.	.	PUNCT
aiti-14589	56	1	additionally	additionally	ADV
aiti-14589	56	2	,	,	PUNCT
aiti-14589	56	3	the	the	DET
aiti-14589	56	4	feature	feature	NOUN
aiti-14589	56	5	extraction	extraction	NOUN
aiti-14589	56	6	process	process	NOUN
aiti-14589	56	7	,	,	PUNCT
aiti-14589	56	8	such	such	ADJ
aiti-14589	56	9	as	as	ADP
aiti-14589	56	10	with	with	ADP
aiti-14589	56	11	hog	hog	NOUN
aiti-14589	56	12	,	,	PUNCT
aiti-14589	56	13	might	might	AUX
aiti-14589	56	14	fail	fail	VERB
aiti-14589	56	15	to	to	PART
aiti-14589	56	16	capture	capture	VERB
aiti-14589	56	17	subtle	subtle	ADJ
aiti-14589	56	18	facial	facial	ADJ
aiti-14589	56	19	expressions	expression	NOUN
aiti-14589	56	20	better	well	ADV
aiti-14589	56	21	represented	represent	VERB
aiti-14589	56	22	by	by	ADP
aiti-14589	56	23	the	the	DET
aiti-14589	56	24	geometry	geometry	NOUN
aiti-14589	56	25	of	of	ADP
aiti-14589	56	26	facial	facial	ADJ
aiti-14589	56	27	landmarks	landmark	NOUN
aiti-14589	56	28	.	.	PUNCT
aiti-14589	57	1	on	on	ADP
aiti-14589	57	2	the	the	DET
aiti-14589	57	3	other	other	ADJ
aiti-14589	57	4	hand	hand	NOUN
aiti-14589	57	5	,	,	PUNCT
aiti-14589	57	6	the	the	DET
aiti-14589	57	7	accuracy	accuracy	NOUN
aiti-14589	57	8	of	of	ADP
aiti-14589	57	9	the	the	DET
aiti-14589	57	10	landmark	landmark	NOUN
aiti-14589	57	11	-	-	PUNCT
aiti-14589	57	12	based	base	VERB
aiti-14589	57	13	approach	approach	NOUN
aiti-14589	57	14	is	be	AUX
aiti-14589	57	15	heavily	heavily	ADV
aiti-14589	57	16	reliant	reliant	ADJ
aiti-14589	57	17	on	on	ADP
aiti-14589	57	18	the	the	DET
aiti-14589	57	19	quality	quality	NOUN
aiti-14589	57	20	and	and	CCONJ
aiti-14589	57	21	precision	precision	NOUN
aiti-14589	57	22	of	of	ADP
aiti-14589	57	23	the	the	DET
aiti-14589	57	24	landmark	landmark	NOUN
aiti-14589	57	25	detection	detection	NOUN
aiti-14589	57	26	process	process	NOUN
aiti-14589	57	27	.	.	PUNCT
aiti-14589	58	1	factors	factor	NOUN
aiti-14589	58	2	like	like	ADP
aiti-14589	58	3	occlusions	occlusion	NOUN
aiti-14589	58	4	,	,	PUNCT
aiti-14589	58	5	misalignments	misalignment	NOUN
aiti-14589	58	6	,	,	PUNCT
aiti-14589	58	7	or	or	CCONJ
aiti-14589	58	8	low	low	ADJ
aiti-14589	58	9	-	-	PUNCT
aiti-14589	58	10	resolution	resolution	NOUN
aiti-14589	58	11	images	image	NOUN
aiti-14589	58	12	can	can	AUX
aiti-14589	58	13	compromise	compromise	VERB
aiti-14589	58	14	the	the	DET
aiti-14589	58	15	reliability	reliability	NOUN
aiti-14589	58	16	of	of	ADP
aiti-14589	58	17	feature	feature	NOUN
aiti-14589	58	18	extraction	extraction	NOUN
aiti-14589	58	19	.	.	PUNCT
aiti-14589	59	1	furthermore	furthermore	ADV
aiti-14589	59	2	,	,	PUNCT
aiti-14589	59	3	landmark	landmark	NOUN
aiti-14589	59	4	-	-	PUNCT
aiti-14589	59	5	based	base	VERB
aiti-14589	59	6	methods	method	NOUN
aiti-14589	59	7	may	may	AUX
aiti-14589	59	8	not	not	PART
aiti-14589	59	9	effectively	effectively	ADV
aiti-14589	59	10	capture	capture	VERB
aiti-14589	59	11	the	the	DET
aiti-14589	59	12	dynamic	dynamic	ADJ
aiti-14589	59	13	changes	change	NOUN
aiti-14589	59	14	in	in	ADP
aiti-14589	59	15	facial	facial	ADJ
aiti-14589	59	16	expressions	expression	NOUN
aiti-14589	59	17	,	,	PUNCT
aiti-14589	59	18	as	as	SCONJ
aiti-14589	59	19	image	image	NOUN
aiti-14589	59	20	-	-	PUNCT
aiti-14589	59	21	based	base	VERB
aiti-14589	59	22	methods	method	NOUN
aiti-14589	59	23	can	can	AUX
aiti-14589	59	24	.	.	PUNCT
aiti-14589	60	1	to	to	PART
aiti-14589	60	2	address	address	VERB
aiti-14589	60	3	these	these	DET
aiti-14589	60	4	limitations	limitation	NOUN
aiti-14589	60	5	,	,	PUNCT
aiti-14589	60	6	this	this	DET
aiti-14589	60	7	study	study	NOUN
aiti-14589	60	8	employs	employ	VERB
aiti-14589	60	9	a	a	DET
aiti-14589	60	10	diverse	diverse	ADJ
aiti-14589	60	11	dataset	dataset	NOUN
aiti-14589	60	12	and	and	CCONJ
aiti-14589	60	13	carefully	carefully	ADV
aiti-14589	60	14	selected	select	VERB
aiti-14589	60	15	models	model	NOUN
aiti-14589	60	16	,	,	PUNCT
aiti-14589	60	17	ensuring	ensure	VERB
aiti-14589	60	18	that	that	SCONJ
aiti-14589	60	19	the	the	DET
aiti-14589	60	20	comparison	comparison	NOUN
aiti-14589	60	21	remains	remain	VERB
aiti-14589	60	22	as	as	ADV
aiti-14589	60	23	balanced	balanced	ADJ
aiti-14589	60	24	and	and	CCONJ
aiti-14589	60	25	robust	robust	ADJ
aiti-14589	60	26	as	as	ADP
aiti-14589	60	27	possible	possible	ADJ
aiti-14589	60	28	.	.	PUNCT
aiti-14589	61	1	advances	advance	NOUN
aiti-14589	61	2	in	in	ADP
aiti-14589	61	3	technology	technology	NOUN
aiti-14589	61	4	innovation	innovation	NOUN
aiti-14589	61	5	,	,	PUNCT
aiti-14589	61	6	vol	vol	NOUN
aiti-14589	61	7	.	.	PROPN
aiti-14589	62	1	10	10	NUM
aiti-14589	62	2	,	,	PUNCT
aiti-14589	62	3	no	no	INTJ
aiti-14589	62	4	.	.	NOUN
aiti-14589	62	5	4	4	NUM
aiti-14589	62	6	,	,	PUNCT
aiti-14589	62	7	2025	2025	NUM
aiti-14589	62	8	,	,	PUNCT
aiti-14589	62	9	pp	pp	ADJ
aiti-14589	62	10	.	.	PUNCT
aiti-14589	63	1	395	395	NUM
aiti-14589	63	2	-	-	SYM
aiti-14589	63	3	406	406	NUM
aiti-14589	63	4	397	397	NUM
aiti-14589	63	5	the	the	DET
aiti-14589	63	6	research	research	NOUN
aiti-14589	63	7	concept	concept	NOUN
aiti-14589	63	8	is	be	AUX
aiti-14589	63	9	shown	show	VERB
aiti-14589	63	10	in	in	ADP
aiti-14589	63	11	fig	fig	NOUN
aiti-14589	63	12	.	.	PUNCT
aiti-14589	64	1	1	1	NUM
aiti-14589	64	2	,	,	PUNCT
aiti-14589	64	3	which	which	PRON
aiti-14589	64	4	visually	visually	ADV
aiti-14589	64	5	represents	represent	VERB
aiti-14589	64	6	the	the	DET
aiti-14589	64	7	comparative	comparative	ADJ
aiti-14589	64	8	framework	framework	NOUN
aiti-14589	64	9	for	for	ADP
aiti-14589	64	10	fer	fer	PROPN
aiti-14589	64	11	using	use	VERB
aiti-14589	64	12	imagebased	imagebase	VERB
aiti-14589	64	13	and	and	CCONJ
aiti-14589	64	14	landmark	landmark	NOUN
aiti-14589	64	15	-	-	PUNCT
aiti-14589	64	16	based	base	VERB
aiti-14589	64	17	approaches	approach	NOUN
aiti-14589	64	18	.	.	PUNCT
aiti-14589	65	1	it	it	PRON
aiti-14589	65	2	highlights	highlight	VERB
aiti-14589	65	3	key	key	ADJ
aiti-14589	65	4	features	feature	NOUN
aiti-14589	65	5	such	such	ADJ
aiti-14589	65	6	as	as	ADP
aiti-14589	65	7	“	"	PUNCT
aiti-14589	65	8	normal	normal	ADJ
aiti-14589	65	9	”	"	PUNCT
aiti-14589	65	10	and	and	CCONJ
aiti-14589	65	11	“	"	PUNCT
aiti-14589	65	12	hurt	hurt	VERB
aiti-14589	65	13	”	"	PUNCT
aiti-14589	65	14	as	as	ADP
aiti-14589	65	15	critical	critical	ADJ
aiti-14589	65	16	markers	marker	NOUN
aiti-14589	65	17	for	for	ADP
aiti-14589	65	18	evaluating	evaluate	VERB
aiti-14589	65	19	the	the	DET
aiti-14589	65	20	effectiveness	effectiveness	NOUN
aiti-14589	65	21	of	of	ADP
aiti-14589	65	22	different	different	ADJ
aiti-14589	65	23	methods	method	NOUN
aiti-14589	65	24	.	.	PUNCT
aiti-14589	66	1	fig	fig	NOUN
aiti-14589	66	2	.	.	PUNCT
aiti-14589	67	1	1	1	NUM
aiti-14589	67	2	fer	fer	PROPN
aiti-14589	67	3	comparative	comparative	ADJ
aiti-14589	67	4	framework	framework	NOUN
aiti-14589	67	5	2	2	NUM
aiti-14589	67	6	.	.	PUNCT
aiti-14589	67	7	literature	literature	PROPN
aiti-14589	67	8	review	review	PROPN
aiti-14589	67	9	fer	fer	PROPN
aiti-14589	67	10	has	have	AUX
aiti-14589	67	11	been	be	AUX
aiti-14589	67	12	an	an	DET
aiti-14589	67	13	important	important	ADJ
aiti-14589	67	14	area	area	NOUN
aiti-14589	67	15	of	of	ADP
aiti-14589	67	16	research	research	NOUN
aiti-14589	67	17	in	in	ADP
aiti-14589	67	18	computer	computer	NOUN
aiti-14589	67	19	vision	vision	NOUN
aiti-14589	67	20	,	,	PUNCT
aiti-14589	67	21	artificial	artificial	ADJ
aiti-14589	67	22	intelligence	intelligence	NOUN
aiti-14589	67	23	,	,	PUNCT
aiti-14589	67	24	and	and	CCONJ
aiti-14589	67	25	human	human	ADJ
aiti-14589	67	26	-	-	PUNCT
aiti-14589	67	27	computer	computer	NOUN
aiti-14589	67	28	interaction	interaction	NOUN
aiti-14589	67	29	.	.	PUNCT
aiti-14589	68	1	this	this	DET
aiti-14589	68	2	section	section	NOUN
aiti-14589	68	3	reviews	review	VERB
aiti-14589	68	4	significant	significant	ADJ
aiti-14589	68	5	contributions	contribution	NOUN
aiti-14589	68	6	in	in	ADP
aiti-14589	68	7	fer	fer	PROPN
aiti-14589	68	8	areas	area	NOUN
aiti-14589	68	9	,	,	PUNCT
aiti-14589	68	10	aligning	align	VERB
aiti-14589	68	11	with	with	ADP
aiti-14589	68	12	the	the	DET
aiti-14589	68	13	comparative	comparative	ADJ
aiti-14589	68	14	works	work	NOUN
aiti-14589	68	15	in	in	ADP
aiti-14589	68	16	table	table	NOUN
aiti-14589	68	17	1	1	NUM
aiti-14589	68	18	and	and	CCONJ
aiti-14589	68	19	positioning	position	VERB
aiti-14589	68	20	the	the	DET
aiti-14589	68	21	current	current	ADJ
aiti-14589	68	22	study	study	NOUN
aiti-14589	68	23	within	within	ADP
aiti-14589	68	24	this	this	DET
aiti-14589	68	25	research	research	NOUN
aiti-14589	68	26	landscape	landscape	NOUN
aiti-14589	68	27	.	.	PUNCT
aiti-14589	69	1	early	early	ADJ
aiti-14589	69	2	work	work	NOUN
aiti-14589	69	3	,	,	PUNCT
aiti-14589	69	4	buck	buck	NOUN
aiti-14589	69	5	et	et	PROPN
aiti-14589	69	6	al	al	PROPN
aiti-14589	69	7	.	.	PUNCT
aiti-14589	70	1	[	[	X
aiti-14589	70	2	5	5	NUM
aiti-14589	70	3	]	]	PUNCT
aiti-14589	70	4	,	,	PUNCT
aiti-14589	70	5	focused	focus	VERB
aiti-14589	70	6	on	on	ADP
aiti-14589	70	7	manual	manual	ADJ
aiti-14589	70	8	coding	coding	NOUN
aiti-14589	70	9	of	of	ADP
aiti-14589	70	10	facial	facial	ADJ
aiti-14589	70	11	expressions	expression	NOUN
aiti-14589	70	12	,	,	PUNCT
aiti-14589	70	13	which	which	PRON
aiti-14589	70	14	provided	provide	VERB
aiti-14589	70	15	foundational	foundational	ADJ
aiti-14589	70	16	insights	insight	NOUN
aiti-14589	70	17	but	but	CCONJ
aiti-14589	70	18	lacked	lack	VERB
aiti-14589	70	19	automation	automation	NOUN
aiti-14589	70	20	.	.	PUNCT
aiti-14589	71	1	the	the	DET
aiti-14589	71	2	introduction	introduction	NOUN
aiti-14589	71	3	of	of	ADP
aiti-14589	71	4	ml	ml	NOUN
aiti-14589	71	5	in	in	ADP
aiti-14589	71	6	fer	fer	PROPN
aiti-14589	71	7	began	begin	VERB
aiti-14589	71	8	with	with	ADP
aiti-14589	71	9	vapnik	vapnik	NOUN
aiti-14589	72	1	[	[	X
aiti-14589	72	2	17	17	NUM
aiti-14589	72	3	]	]	PUNCT
aiti-14589	72	4	,	,	PUNCT
aiti-14589	72	5	who	who	PRON
aiti-14589	72	6	developed	develop	VERB
aiti-14589	72	7	svm	svm	PROPN
aiti-14589	72	8	,	,	PUNCT
aiti-14589	72	9	a	a	DET
aiti-14589	72	10	key	key	ADJ
aiti-14589	72	11	advancement	advancement	NOUN
aiti-14589	72	12	in	in	ADP
aiti-14589	72	13	classification	classification	NOUN
aiti-14589	72	14	methods	method	NOUN
aiti-14589	72	15	.	.	PUNCT
aiti-14589	73	1	however	however	ADV
aiti-14589	73	2	,	,	PUNCT
aiti-14589	73	3	early	early	ADJ
aiti-14589	73	4	ml	ml	NOUN
aiti-14589	73	5	applications	application	NOUN
aiti-14589	73	6	in	in	ADP
aiti-14589	73	7	fer	fer	PROPN
aiti-14589	73	8	were	be	AUX
aiti-14589	73	9	limited	limit	VERB
aiti-14589	73	10	in	in	ADP
aiti-14589	73	11	scope	scope	NOUN
aiti-14589	73	12	and	and	CCONJ
aiti-14589	73	13	required	require	VERB
aiti-14589	73	14	further	further	ADJ
aiti-14589	73	15	refinement	refinement	NOUN
aiti-14589	73	16	.	.	PUNCT
aiti-14589	74	1	as	as	SCONJ
aiti-14589	74	2	artificial	artificial	ADJ
aiti-14589	74	3	intelligence	intelligence	NOUN
aiti-14589	74	4	advanced	advance	VERB
aiti-14589	74	5	,	,	PUNCT
aiti-14589	74	6	various	various	ADJ
aiti-14589	74	7	approaches	approach	NOUN
aiti-14589	74	8	to	to	ADP
aiti-14589	74	9	fer	fer	PROPN
aiti-14589	74	10	were	be	AUX
aiti-14589	74	11	explored	explore	VERB
aiti-14589	74	12	.	.	PUNCT
aiti-14589	75	1	lisetti	lisetti	NOUN
aiti-14589	75	2	and	and	CCONJ
aiti-14589	75	3	schiano	schiano	X
aiti-14589	76	1	[	[	X
aiti-14589	76	2	6	6	NUM
aiti-14589	76	3	]	]	PUNCT
aiti-14589	76	4	,	,	PUNCT
aiti-14589	76	5	along	along	ADP
aiti-14589	76	6	with	with	ADP
aiti-14589	76	7	picard	picard	PROPN
aiti-14589	76	8	et	et	PROPN
aiti-14589	76	9	al	al	PROPN
aiti-14589	76	10	.	.	PUNCT
aiti-14589	77	1	[	[	X
aiti-14589	77	2	18	18	NUM
aiti-14589	77	3	]	]	PUNCT
aiti-14589	77	4	,	,	PUNCT
aiti-14589	77	5	expanded	expand	VERB
aiti-14589	77	6	the	the	DET
aiti-14589	77	7	field	field	NOUN
aiti-14589	77	8	by	by	ADP
aiti-14589	77	9	integrating	integrate	VERB
aiti-14589	77	10	affective	affective	ADJ
aiti-14589	77	11	computing	computing	NOUN
aiti-14589	77	12	,	,	PUNCT
aiti-14589	77	13	enabling	enable	VERB
aiti-14589	77	14	machines	machine	NOUN
aiti-14589	77	15	to	to	PART
aiti-14589	77	16	automatically	automatically	ADV
aiti-14589	77	17	interpret	interpret	VERB
aiti-14589	77	18	emotional	emotional	ADJ
aiti-14589	77	19	states	state	NOUN
aiti-14589	77	20	.	.	PUNCT
aiti-14589	78	1	bartlett	bartlett	PROPN
aiti-14589	78	2	et	et	PROPN
aiti-14589	78	3	al	al	PROPN
aiti-14589	78	4	.	.	PUNCT
aiti-14589	79	1	[	[	X
aiti-14589	79	2	19	19	NUM
aiti-14589	79	3	]	]	PUNCT
aiti-14589	79	4	later	later	ADV
aiti-14589	79	5	demonstrated	demonstrate	VERB
aiti-14589	79	6	the	the	DET
aiti-14589	79	7	effectiveness	effectiveness	NOUN
aiti-14589	79	8	of	of	ADP
aiti-14589	79	9	cnn	cnn	PROPN
aiti-14589	79	10	for	for	ADP
aiti-14589	79	11	real	real	ADJ
aiti-14589	79	12	-	-	PUNCT
aiti-14589	79	13	time	time	NOUN
aiti-14589	79	14	fer	fer	NOUN
aiti-14589	79	15	,	,	PUNCT
aiti-14589	79	16	though	though	SCONJ
aiti-14589	79	17	their	their	PRON
aiti-14589	79	18	study	study	NOUN
aiti-14589	79	19	was	be	AUX
aiti-14589	79	20	constrained	constrain	VERB
aiti-14589	79	21	by	by	ADP
aiti-14589	79	22	the	the	DET
aiti-14589	79	23	availability	availability	NOUN
aiti-14589	79	24	of	of	ADP
aiti-14589	79	25	labeled	label	VERB
aiti-14589	79	26	data	datum	NOUN
aiti-14589	79	27	.	.	PUNCT
aiti-14589	80	1	other	other	ADJ
aiti-14589	80	2	foundational	foundational	ADJ
aiti-14589	80	3	works	work	NOUN
aiti-14589	80	4	,	,	PUNCT
aiti-14589	80	5	such	such	ADJ
aiti-14589	80	6	as	as	ADP
aiti-14589	80	7	those	those	PRON
aiti-14589	80	8	of	of	ADP
aiti-14589	80	9	lu	lu	PROPN
aiti-14589	80	10	and	and	CCONJ
aiti-14589	80	11	weng	weng	PROPN
aiti-14589	81	1	[	[	X
aiti-14589	81	2	10	10	NUM
aiti-14589	81	3	]	]	PUNCT
aiti-14589	81	4	,	,	PUNCT
aiti-14589	81	5	along	along	ADP
aiti-14589	81	6	with	with	ADP
aiti-14589	81	7	hilbe	hilbe	NOUN
aiti-14589	81	8	[	[	X
aiti-14589	81	9	13	13	NUM
aiti-14589	81	10	]	]	PUNCT
aiti-14589	81	11	,	,	PUNCT
aiti-14589	81	12	examined	examine	VERB
aiti-14589	81	13	classification	classification	NOUN
aiti-14589	81	14	techniques	technique	NOUN
aiti-14589	81	15	and	and	CCONJ
aiti-14589	81	16	statistical	statistical	ADJ
aiti-14589	81	17	modeling	modeling	NOUN
aiti-14589	81	18	,	,	PUNCT
aiti-14589	81	19	but	but	CCONJ
aiti-14589	81	20	did	do	AUX
aiti-14589	81	21	not	not	PART
aiti-14589	81	22	provide	provide	VERB
aiti-14589	81	23	direct	direct	ADJ
aiti-14589	81	24	comparisons	comparison	NOUN
aiti-14589	81	25	between	between	ADP
aiti-14589	81	26	traditional	traditional	ADJ
aiti-14589	81	27	ml	ml	NOUN
aiti-14589	81	28	and	and	CCONJ
aiti-14589	81	29	deep	deep	ADJ
aiti-14589	81	30	learning	learning	NOUN
aiti-14589	81	31	models	model	NOUN
aiti-14589	81	32	in	in	ADP
aiti-14589	81	33	fer	fer	PROPN
aiti-14589	81	34	.	.	PUNCT
aiti-14589	82	1	beyond	beyond	ADP
aiti-14589	82	2	general	general	PROPN
aiti-14589	82	3	fer	fer	PROPN
aiti-14589	82	4	applications	application	NOUN
aiti-14589	82	5	,	,	PUNCT
aiti-14589	82	6	researchers	researcher	NOUN
aiti-14589	82	7	have	have	AUX
aiti-14589	82	8	explored	explore	VERB
aiti-14589	82	9	its	its	PRON
aiti-14589	82	10	use	use	NOUN
aiti-14589	82	11	in	in	ADP
aiti-14589	82	12	healthcare	healthcare	NOUN
aiti-14589	82	13	and	and	CCONJ
aiti-14589	82	14	specialized	specialized	ADJ
aiti-14589	82	15	domains	domain	NOUN
aiti-14589	82	16	.	.	PUNCT
aiti-14589	83	1	lee	lee	PROPN
aiti-14589	83	2	and	and	CCONJ
aiti-14589	83	3	park	park	NOUN
aiti-14589	84	1	[	[	X
aiti-14589	84	2	20	20	NUM
aiti-14589	84	3	]	]	PUNCT
aiti-14589	84	4	developed	develop	VERB
aiti-14589	84	5	a	a	DET
aiti-14589	84	6	fast	fast	ADJ
aiti-14589	84	7	region	region	NOUN
aiti-14589	84	8	-	-	PUNCT
aiti-14589	84	9	based	base	VERB
aiti-14589	84	10	convolutional	convolutional	ADJ
aiti-14589	84	11	neural	neural	ADJ
aiti-14589	84	12	networks	network	NOUN
aiti-14589	84	13	(	(	PUNCT
aiti-14589	84	14	r	r	NOUN
aiti-14589	84	15	-	-	PUNCT
aiti-14589	84	16	cnn)-based	cnn)-base	VERB
aiti-14589	84	17	fer	fer	PROPN
aiti-14589	84	18	model	model	NOUN
aiti-14589	84	19	for	for	ADP
aiti-14589	84	20	diagnosing	diagnose	VERB
aiti-14589	84	21	depressive	depressive	ADJ
aiti-14589	84	22	disorders	disorder	NOUN
aiti-14589	84	23	,	,	PUNCT
aiti-14589	84	24	demonstrating	demonstrate	VERB
aiti-14589	84	25	its	its	PRON
aiti-14589	84	26	potential	potential	NOUN
aiti-14589	84	27	in	in	ADP
aiti-14589	84	28	mental	mental	ADJ
aiti-14589	84	29	health	health	NOUN
aiti-14589	84	30	assessments	assessment	NOUN
aiti-14589	84	31	.	.	PUNCT
aiti-14589	85	1	however	however	ADV
aiti-14589	85	2	,	,	PUNCT
aiti-14589	85	3	their	their	PRON
aiti-14589	85	4	study	study	NOUN
aiti-14589	85	5	relied	rely	VERB
aiti-14589	85	6	only	only	ADV
aiti-14589	85	7	on	on	ADP
aiti-14589	85	8	image	image	NOUN
aiti-14589	85	9	-	-	PUNCT
aiti-14589	85	10	based	base	VERB
aiti-14589	85	11	features	feature	NOUN
aiti-14589	85	12	,	,	PUNCT
aiti-14589	85	13	without	without	ADP
aiti-14589	85	14	considering	consider	VERB
aiti-14589	85	15	how	how	SCONJ
aiti-14589	85	16	landmark	landmark	NOUN
aiti-14589	85	17	-	-	PUNCT
aiti-14589	85	18	based	base	VERB
aiti-14589	85	19	features	feature	NOUN
aiti-14589	85	20	might	might	AUX
aiti-14589	85	21	enhance	enhance	VERB
aiti-14589	85	22	diagnostic	diagnostic	ADJ
aiti-14589	85	23	accuracy	accuracy	NOUN
aiti-14589	85	24	.	.	PUNCT
aiti-14589	86	1	similarly	similarly	ADV
aiti-14589	86	2	,	,	PUNCT
aiti-14589	86	3	bekele	bekele	PROPN
aiti-14589	86	4	et	et	PROPN
aiti-14589	86	5	al	al	PROPN
aiti-14589	86	6	.	.	PUNCT
aiti-14589	87	1	[	[	X
aiti-14589	87	2	21	21	NUM
aiti-14589	87	3	]	]	PUNCT
aiti-14589	87	4	introduced	introduce	VERB
aiti-14589	87	5	a	a	DET
aiti-14589	87	6	virtual	virtual	ADJ
aiti-14589	87	7	reality	reality	NOUN
aiti-14589	87	8	-	-	PUNCT
aiti-14589	87	9	based	base	VERB
aiti-14589	87	10	fer	fer	PROPN
aiti-14589	87	11	system	system	NOUN
aiti-14589	87	12	,	,	PUNCT
aiti-14589	87	13	while	while	SCONJ
aiti-14589	87	14	assari	assari	NOUN
aiti-14589	87	15	and	and	CCONJ
aiti-14589	87	16	rahmati	rahmati	NOUN
aiti-14589	87	17	[	[	X
aiti-14589	87	18	22	22	NUM
aiti-14589	87	19	]	]	PUNCT
aiti-14589	87	20	focused	focus	VERB
aiti-14589	87	21	on	on	ADP
aiti-14589	87	22	drowsiness	drowsiness	NOUN
aiti-14589	87	23	detection	detection	NOUN
aiti-14589	87	24	in	in	ADP
aiti-14589	87	25	driver	driver	NOUN
aiti-14589	87	26	monitoring	monitoring	NOUN
aiti-14589	87	27	systems	system	NOUN
aiti-14589	87	28	.	.	PUNCT
aiti-14589	88	1	while	while	SCONJ
aiti-14589	88	2	these	these	DET
aiti-14589	88	3	studies	study	NOUN
aiti-14589	88	4	highlighted	highlight	VERB
aiti-14589	88	5	important	important	ADJ
aiti-14589	88	6	applications	application	NOUN
aiti-14589	88	7	,	,	PUNCT
aiti-14589	88	8	they	they	PRON
aiti-14589	88	9	did	do	AUX
aiti-14589	88	10	not	not	PART
aiti-14589	88	11	compare	compare	VERB
aiti-14589	88	12	landmark	landmark	NOUN
aiti-14589	88	13	-	-	PUNCT
aiti-14589	88	14	based	base	VERB
aiti-14589	88	15	and	and	CCONJ
aiti-14589	88	16	image	image	NOUN
aiti-14589	88	17	-	-	PUNCT
aiti-14589	88	18	based	base	VERB
aiti-14589	88	19	methods	method	NOUN
aiti-14589	88	20	,	,	PUNCT
aiti-14589	88	21	limiting	limit	VERB
aiti-14589	88	22	their	their	PRON
aiti-14589	88	23	scope	scope	NOUN
aiti-14589	88	24	.	.	PUNCT
aiti-14589	89	1	many	many	ADJ
aiti-14589	89	2	studies	study	NOUN
aiti-14589	89	3	have	have	AUX
aiti-14589	89	4	concentrated	concentrate	VERB
aiti-14589	89	5	on	on	ADP
aiti-14589	89	6	image	image	NOUN
aiti-14589	89	7	-	-	PUNCT
aiti-14589	89	8	based	base	VERB
aiti-14589	89	9	fer	fer	PROPN
aiti-14589	89	10	techniques	technique	NOUN
aiti-14589	89	11	,	,	PUNCT
aiti-14589	89	12	refining	refine	VERB
aiti-14589	89	13	classification	classification	NOUN
aiti-14589	89	14	models	model	NOUN
aiti-14589	89	15	to	to	PART
aiti-14589	89	16	improve	improve	VERB
aiti-14589	89	17	accuracy	accuracy	NOUN
aiti-14589	89	18	.	.	PUNCT
aiti-14589	90	1	hamester	hamester	NOUN
aiti-14589	90	2	et	et	PROPN
aiti-14589	90	3	al	al	PROPN
aiti-14589	90	4	.	.	PUNCT
aiti-14589	91	1	[	[	X
aiti-14589	91	2	23	23	NUM
aiti-14589	91	3	]	]	PUNCT
aiti-14589	91	4	proposed	propose	VERB
aiti-14589	91	5	a	a	DET
aiti-14589	91	6	two	two	NUM
aiti-14589	91	7	-	-	PUNCT
aiti-14589	91	8	channel	channel	NOUN
aiti-14589	91	9	cnn	cnn	NOUN
aiti-14589	91	10	architecture	architecture	NOUN
aiti-14589	91	11	,	,	PUNCT
aiti-14589	91	12	where	where	SCONJ
aiti-14589	91	13	one	one	NUM
aiti-14589	91	14	channel	channel	NOUN
aiti-14589	91	15	processed	process	VERB
aiti-14589	91	16	raw	raw	ADJ
aiti-14589	91	17	image	image	NOUN
aiti-14589	91	18	data	datum	NOUN
aiti-14589	91	19	and	and	CCONJ
aiti-14589	91	20	the	the	DET
aiti-14589	91	21	other	other	ADJ
aiti-14589	91	22	extracted	extract	VERB
aiti-14589	91	23	features	feature	NOUN
aiti-14589	91	24	using	use	VERB
aiti-14589	91	25	an	an	DET
aiti-14589	91	26	autoencoder	autoencoder	NOUN
aiti-14589	91	27	.	.	PUNCT
aiti-14589	92	1	while	while	SCONJ
aiti-14589	92	2	this	this	DET
aiti-14589	92	3	approach	approach	NOUN
aiti-14589	92	4	improved	improve	VERB
aiti-14589	92	5	classification	classification	NOUN
aiti-14589	92	6	accuracy	accuracy	NOUN
aiti-14589	92	7	,	,	PUNCT
aiti-14589	92	8	it	it	PRON
aiti-14589	92	9	lacked	lack	VERB
aiti-14589	92	10	comparisons	comparison	NOUN
aiti-14589	92	11	with	with	ADP
aiti-14589	92	12	traditional	traditional	ADJ
aiti-14589	92	13	ml	ml	NOUN
aiti-14589	92	14	classifiers	classifier	NOUN
aiti-14589	92	15	.	.	PUNCT
aiti-14589	93	1	alsubari	alsubari	PROPN
aiti-14589	93	2	et	et	PROPN
aiti-14589	93	3	al	al	PROPN
aiti-14589	93	4	.	.	PUNCT
aiti-14589	94	1	[	[	X
aiti-14589	94	2	24	24	NUM
aiti-14589	94	3	]	]	PUNCT
aiti-14589	94	4	incorporated	incorporate	VERB
aiti-14589	94	5	wavelet	wavelet	NOUN
aiti-14589	94	6	transforms	transform	VERB
aiti-14589	94	7	and	and	CCONJ
aiti-14589	94	8	local	local	ADJ
aiti-14589	94	9	binary	binary	ADJ
aiti-14589	94	10	patterns	pattern	NOUN
aiti-14589	94	11	(	(	PUNCT
aiti-14589	94	12	lbp	lbp	PROPN
aiti-14589	94	13	)	)	PUNCT
aiti-14589	94	14	to	to	PART
aiti-14589	94	15	enhance	enhance	VERB
aiti-14589	94	16	feature	feature	NOUN
aiti-14589	94	17	extraction	extraction	NOUN
aiti-14589	94	18	and	and	CCONJ
aiti-14589	94	19	fusion	fusion	NOUN
aiti-14589	94	20	,	,	PUNCT
aiti-14589	94	21	improving	improve	VERB
aiti-14589	94	22	recognition	recognition	NOUN
aiti-14589	94	23	rates	rate	NOUN
aiti-14589	94	24	.	.	PUNCT
aiti-14589	95	1	however	however	ADV
aiti-14589	95	2	,	,	PUNCT
aiti-14589	95	3	the	the	DET
aiti-14589	95	4	study	study	NOUN
aiti-14589	95	5	did	do	AUX
aiti-14589	95	6	not	not	PART
aiti-14589	95	7	explicitly	explicitly	ADV
aiti-14589	95	8	analyze	analyze	VERB
aiti-14589	95	9	the	the	DET
aiti-14589	95	10	individual	individual	ADJ
aiti-14589	95	11	contribution	contribution	NOUN
aiti-14589	95	12	of	of	ADP
aiti-14589	95	13	landmark	landmark	NOUN
aiti-14589	95	14	-	-	PUNCT
aiti-14589	95	15	based	base	VERB
aiti-14589	95	16	features	feature	NOUN
aiti-14589	95	17	to	to	PART
aiti-14589	95	18	fer	fer	PROPN
aiti-14589	95	19	performance	performance	NOUN
aiti-14589	95	20	.	.	PUNCT
aiti-14589	96	1	advances	advance	NOUN
aiti-14589	96	2	in	in	ADP
aiti-14589	96	3	technology	technology	NOUN
aiti-14589	96	4	innovation	innovation	NOUN
aiti-14589	96	5	,	,	PUNCT
aiti-14589	96	6	vol	vol	NOUN
aiti-14589	96	7	.	.	PROPN
aiti-14589	97	1	10	10	NUM
aiti-14589	97	2	,	,	PUNCT
aiti-14589	97	3	no	no	INTJ
aiti-14589	97	4	.	.	NOUN
aiti-14589	97	5	4	4	NUM
aiti-14589	97	6	,	,	PUNCT
aiti-14589	97	7	2025	2025	NUM
aiti-14589	97	8	,	,	PUNCT
aiti-14589	97	9	pp	pp	ADJ
aiti-14589	97	10	.	.	PUNCT
aiti-14589	98	1	395	395	NUM
aiti-14589	98	2	-	-	SYM
aiti-14589	98	3	406	406	NUM
aiti-14589	98	4	398	398	NUM
aiti-14589	98	5	alongside	alongside	ADP
aiti-14589	98	6	image	image	NOUN
aiti-14589	98	7	-	-	PUNCT
aiti-14589	98	8	based	base	VERB
aiti-14589	98	9	approaches	approach	NOUN
aiti-14589	98	10	,	,	PUNCT
aiti-14589	98	11	scholars	scholar	NOUN
aiti-14589	98	12	have	have	AUX
aiti-14589	98	13	also	also	ADV
aiti-14589	98	14	explored	explore	VERB
aiti-14589	98	15	landmark	landmark	NOUN
aiti-14589	98	16	-	-	PUNCT
aiti-14589	98	17	based	base	VERB
aiti-14589	98	18	fer	fer	PROPN
aiti-14589	98	19	.	.	PUNCT
aiti-14589	99	1	söylemez	söylemez	PROPN
aiti-14589	99	2	et	et	PROPN
aiti-14589	99	3	al	al	PROPN
aiti-14589	99	4	.	.	PUNCT
aiti-14589	100	1	[	[	X
aiti-14589	100	2	25	25	NUM
aiti-14589	100	3	]	]	PUNCT
aiti-14589	100	4	applied	apply	VERB
aiti-14589	100	5	threedimensional	threedimensional	ADJ
aiti-14589	100	6	facial	facial	ADJ
aiti-14589	100	7	landmark	landmark	NOUN
aiti-14589	100	8	distances	distance	NOUN
aiti-14589	100	9	with	with	ADP
aiti-14589	100	10	svm	svm	PROPN
aiti-14589	100	11	,	,	PUNCT
aiti-14589	100	12	showing	show	VERB
aiti-14589	100	13	strong	strong	ADJ
aiti-14589	100	14	performance	performance	NOUN
aiti-14589	100	15	in	in	ADP
aiti-14589	100	16	recognizing	recognize	VERB
aiti-14589	100	17	expressions	expression	NOUN
aiti-14589	100	18	from	from	ADP
aiti-14589	100	19	3d	3d	NUM
aiti-14589	100	20	face	face	NOUN
aiti-14589	100	21	models	model	NOUN
aiti-14589	100	22	.	.	PUNCT
aiti-14589	101	1	however	however	ADV
aiti-14589	101	2	,	,	PUNCT
aiti-14589	101	3	their	their	PRON
aiti-14589	101	4	study	study	NOUN
aiti-14589	101	5	did	do	AUX
aiti-14589	101	6	not	not	PART
aiti-14589	101	7	compare	compare	VERB
aiti-14589	101	8	landmark	landmark	NOUN
aiti-14589	101	9	-	-	PUNCT
aiti-14589	101	10	based	base	VERB
aiti-14589	101	11	methods	method	NOUN
aiti-14589	101	12	to	to	ADP
aiti-14589	101	13	traditional	traditional	ADJ
aiti-14589	101	14	image	image	NOUN
aiti-14589	101	15	-	-	PUNCT
aiti-14589	101	16	based	base	VERB
aiti-14589	101	17	techniques	technique	NOUN
aiti-14589	101	18	.	.	PUNCT
aiti-14589	102	1	munasinghe	munasinghe	PROPN
aiti-14589	103	1	[	[	X
aiti-14589	103	2	26	26	NUM
aiti-14589	103	3	]	]	PUNCT
aiti-14589	103	4	used	use	VERB
aiti-14589	103	5	rfc	rfc	NOUN
aiti-14589	103	6	for	for	ADP
aiti-14589	103	7	landmark	landmark	NOUN
aiti-14589	103	8	-	-	PUNCT
aiti-14589	103	9	based	base	VERB
aiti-14589	103	10	fer	fer	PROPN
aiti-14589	103	11	,	,	PUNCT
aiti-14589	103	12	demonstrating	demonstrate	VERB
aiti-14589	103	13	efficient	efficient	ADJ
aiti-14589	103	14	training	training	NOUN
aiti-14589	103	15	times	time	NOUN
aiti-14589	103	16	but	but	CCONJ
aiti-14589	103	17	relying	rely	VERB
aiti-14589	103	18	on	on	ADP
aiti-14589	103	19	a	a	DET
aiti-14589	103	20	relatively	relatively	ADV
aiti-14589	103	21	small	small	ADJ
aiti-14589	103	22	dataset	dataset	NOUN
aiti-14589	103	23	.	.	PUNCT
aiti-14589	104	1	more	more	ADV
aiti-14589	104	2	recently	recently	ADV
aiti-14589	104	3	,	,	PUNCT
aiti-14589	104	4	sharma	sharma	PROPN
aiti-14589	104	5	et	et	PROPN
aiti-14589	104	6	al	al	PROPN
aiti-14589	104	7	.	.	PUNCT
aiti-14589	105	1	[	[	X
aiti-14589	105	2	9	9	NUM
aiti-14589	105	3	]	]	PUNCT
aiti-14589	105	4	optimized	optimize	VERB
aiti-14589	105	5	landmark	landmark	NOUN
aiti-14589	105	6	-	-	PUNCT
aiti-14589	105	7	based	base	VERB
aiti-14589	105	8	svm	svm	ADJ
aiti-14589	105	9	models	model	NOUN
aiti-14589	105	10	,	,	PUNCT
aiti-14589	105	11	improving	improve	VERB
aiti-14589	105	12	robustness	robustness	NOUN
aiti-14589	105	13	but	but	CCONJ
aiti-14589	105	14	highlighting	highlight	VERB
aiti-14589	105	15	their	their	PRON
aiti-14589	105	16	sensitivity	sensitivity	NOUN
aiti-14589	105	17	to	to	ADP
aiti-14589	105	18	errors	error	NOUN
aiti-14589	105	19	in	in	ADP
aiti-14589	105	20	facial	facial	ADJ
aiti-14589	105	21	landmark	landmark	NOUN
aiti-14589	105	22	detection	detection	NOUN
aiti-14589	105	23	.	.	PUNCT
aiti-14589	106	1	di	di	NOUN
aiti-14589	106	2	luzio	luzio	PROPN
aiti-14589	106	3	et	et	PROPN
aiti-14589	106	4	al	al	PROPN
aiti-14589	106	5	.	.	PUNCT
aiti-14589	107	1	[	[	X
aiti-14589	107	2	27	27	NUM
aiti-14589	107	3	]	]	PUNCT
aiti-14589	107	4	proposed	propose	VERB
aiti-14589	107	5	deep	deep	ADJ
aiti-14589	107	6	neural	neural	ADJ
aiti-14589	107	7	networks	network	NOUN
aiti-14589	107	8	(	(	PUNCT
aiti-14589	107	9	dnns	dnn	NOUN
aiti-14589	107	10	)	)	PUNCT
aiti-14589	107	11	for	for	ADP
aiti-14589	107	12	processing	process	VERB
aiti-14589	107	13	hierarchical	hierarchical	ADJ
aiti-14589	107	14	landmark	landmark	NOUN
aiti-14589	107	15	features	feature	NOUN
aiti-14589	107	16	,	,	PUNCT
aiti-14589	107	17	though	though	SCONJ
aiti-14589	107	18	their	their	PRON
aiti-14589	107	19	study	study	NOUN
aiti-14589	107	20	did	do	AUX
aiti-14589	107	21	not	not	PART
aiti-14589	107	22	include	include	VERB
aiti-14589	107	23	a	a	DET
aiti-14589	107	24	direct	direct	ADJ
aiti-14589	107	25	comparison	comparison	NOUN
aiti-14589	107	26	between	between	ADP
aiti-14589	107	27	landmark	landmark	NOUN
aiti-14589	107	28	-	-	PUNCT
aiti-14589	107	29	based	base	VERB
aiti-14589	107	30	and	and	CCONJ
aiti-14589	107	31	image	image	NOUN
aiti-14589	107	32	-	-	PUNCT
aiti-14589	107	33	based	base	VERB
aiti-14589	107	34	models	model	NOUN
aiti-14589	107	35	.	.	PUNCT
aiti-14589	108	1	hangaragi	hangaragi	PROPN
aiti-14589	108	2	et	et	PROPN
aiti-14589	108	3	al	al	PROPN
aiti-14589	108	4	.	.	PUNCT
aiti-14589	109	1	[	[	X
aiti-14589	109	2	28	28	NUM
aiti-14589	109	3	]	]	PUNCT
aiti-14589	109	4	addressed	address	VERB
aiti-14589	109	5	this	this	DET
aiti-14589	109	6	limitation	limitation	NOUN
aiti-14589	109	7	by	by	ADP
aiti-14589	109	8	implementing	implement	VERB
aiti-14589	109	9	a	a	DET
aiti-14589	109	10	face	face	NOUN
aiti-14589	109	11	mesh	mesh	NOUN
aiti-14589	109	12	-	-	PUNCT
aiti-14589	109	13	based	base	VERB
aiti-14589	109	14	deep	deep	ADJ
aiti-14589	109	15	learning	learning	NOUN
aiti-14589	109	16	model	model	NOUN
aiti-14589	109	17	,	,	PUNCT
aiti-14589	109	18	which	which	PRON
aiti-14589	109	19	achieved	achieve	VERB
aiti-14589	109	20	realtime	realtime	NOUN
aiti-14589	109	21	efficiency	efficiency	NOUN
aiti-14589	109	22	but	but	CCONJ
aiti-14589	109	23	lacked	lack	VERB
aiti-14589	109	24	comparative	comparative	ADJ
aiti-14589	109	25	analysis	analysis	NOUN
aiti-14589	109	26	with	with	ADP
aiti-14589	109	27	landmark	landmark	NOUN
aiti-14589	109	28	-	-	PUNCT
aiti-14589	109	29	based	base	VERB
aiti-14589	109	30	classifiers	classifier	NOUN
aiti-14589	109	31	.	.	PUNCT
aiti-14589	110	1	despite	despite	SCONJ
aiti-14589	110	2	extensive	extensive	ADJ
aiti-14589	110	3	research	research	NOUN
aiti-14589	110	4	on	on	ADP
aiti-14589	110	5	both	both	DET
aiti-14589	110	6	image	image	NOUN
aiti-14589	110	7	-	-	PUNCT
aiti-14589	110	8	based	base	VERB
aiti-14589	110	9	and	and	CCONJ
aiti-14589	110	10	landmark	landmark	NOUN
aiti-14589	110	11	-	-	PUNCT
aiti-14589	110	12	based	base	VERB
aiti-14589	110	13	fer	fer	PROPN
aiti-14589	110	14	,	,	PUNCT
aiti-14589	110	15	a	a	DET
aiti-14589	110	16	notable	notable	ADJ
aiti-14589	110	17	gap	gap	NOUN
aiti-14589	110	18	remains	remain	VERB
aiti-14589	110	19	:	:	PUNCT
aiti-14589	110	20	few	few	ADJ
aiti-14589	110	21	studies	study	NOUN
aiti-14589	110	22	provide	provide	VERB
aiti-14589	110	23	direct	direct	ADJ
aiti-14589	110	24	performance	performance	NOUN
aiti-14589	110	25	comparisons	comparison	NOUN
aiti-14589	110	26	between	between	ADP
aiti-14589	110	27	the	the	DET
aiti-14589	110	28	two	two	NUM
aiti-14589	110	29	approaches	approach	NOUN
aiti-14589	110	30	.	.	PUNCT
aiti-14589	111	1	most	most	ADJ
aiti-14589	111	2	prior	prior	ADJ
aiti-14589	111	3	work	work	NOUN
aiti-14589	111	4	has	have	AUX
aiti-14589	111	5	focused	focus	VERB
aiti-14589	111	6	on	on	ADP
aiti-14589	111	7	a	a	DET
aiti-14589	111	8	single	single	ADJ
aiti-14589	111	9	classifier	classifier	NOUN
aiti-14589	111	10	,	,	PUNCT
aiti-14589	111	11	making	make	VERB
aiti-14589	111	12	it	it	PRON
aiti-14589	111	13	difficult	difficult	ADJ
aiti-14589	111	14	to	to	PART
aiti-14589	111	15	assess	assess	VERB
aiti-14589	111	16	how	how	SCONJ
aiti-14589	111	17	different	different	ADJ
aiti-14589	111	18	models	model	NOUN
aiti-14589	111	19	generalize	generalize	VERB
aiti-14589	111	20	across	across	ADP
aiti-14589	111	21	datasets	dataset	NOUN
aiti-14589	111	22	.	.	PUNCT
aiti-14589	112	1	additionally	additionally	ADV
aiti-14589	112	2	,	,	PUNCT
aiti-14589	112	3	while	while	SCONJ
aiti-14589	112	4	studies	study	NOUN
aiti-14589	112	5	such	such	ADJ
aiti-14589	112	6	as	as	ADP
aiti-14589	112	7	lee	lee	PROPN
aiti-14589	112	8	and	and	CCONJ
aiti-14589	112	9	park	park	NOUN
aiti-14589	112	10	[	[	X
aiti-14589	112	11	20	20	NUM
aiti-14589	112	12	]	]	PUNCT
aiti-14589	112	13	demonstrated	demonstrate	VERB
aiti-14589	112	14	the	the	DET
aiti-14589	112	15	potential	potential	NOUN
aiti-14589	112	16	of	of	ADP
aiti-14589	112	17	image	image	NOUN
aiti-14589	112	18	-	-	PUNCT
aiti-14589	112	19	based	base	VERB
aiti-14589	112	20	fer	fer	NOUN
aiti-14589	112	21	for	for	ADP
aiti-14589	112	22	medical	medical	ADJ
aiti-14589	112	23	diagnosis	diagnosis	NOUN
aiti-14589	112	24	,	,	PUNCT
aiti-14589	112	25	they	they	PRON
aiti-14589	112	26	did	do	AUX
aiti-14589	112	27	not	not	PART
aiti-14589	112	28	explore	explore	VERB
aiti-14589	112	29	landmark	landmark	NOUN
aiti-14589	112	30	-	-	PUNCT
aiti-14589	112	31	based	base	VERB
aiti-14589	112	32	techniques	technique	NOUN
aiti-14589	112	33	,	,	PUNCT
aiti-14589	112	34	which	which	PRON
aiti-14589	112	35	could	could	AUX
aiti-14589	112	36	enhance	enhance	VERB
aiti-14589	112	37	interpretability	interpretability	NOUN
aiti-14589	112	38	and	and	CCONJ
aiti-14589	112	39	robustness	robustness	NOUN
aiti-14589	112	40	in	in	ADP
aiti-14589	112	41	clinical	clinical	ADJ
aiti-14589	112	42	settings	setting	NOUN
aiti-14589	112	43	.	.	PUNCT
aiti-14589	113	1	to	to	PART
aiti-14589	113	2	address	address	VERB
aiti-14589	113	3	these	these	DET
aiti-14589	113	4	gaps	gap	NOUN
aiti-14589	113	5	,	,	PUNCT
aiti-14589	113	6	this	this	DET
aiti-14589	113	7	study	study	NOUN
aiti-14589	113	8	conducts	conduct	VERB
aiti-14589	113	9	a	a	DET
aiti-14589	113	10	detailed	detailed	ADJ
aiti-14589	113	11	evaluation	evaluation	NOUN
aiti-14589	113	12	of	of	ADP
aiti-14589	113	13	both	both	DET
aiti-14589	113	14	image	image	NOUN
aiti-14589	113	15	-	-	PUNCT
aiti-14589	113	16	based	base	VERB
aiti-14589	113	17	and	and	CCONJ
aiti-14589	113	18	landmark	landmark	NOUN
aiti-14589	113	19	-	-	PUNCT
aiti-14589	113	20	based	base	VERB
aiti-14589	113	21	fer	fer	PROPN
aiti-14589	113	22	methods	method	NOUN
aiti-14589	113	23	across	across	ADP
aiti-14589	113	24	multiple	multiple	ADJ
aiti-14589	113	25	classifiers	classifier	NOUN
aiti-14589	113	26	,	,	PUNCT
aiti-14589	113	27	including	include	VERB
aiti-14589	113	28	cnn	cnn	PROPN
aiti-14589	113	29	,	,	PUNCT
aiti-14589	113	30	svm	svm	PROPN
aiti-14589	113	31	,	,	PUNCT
aiti-14589	113	32	lrc	lrc	PROPN
aiti-14589	113	33	,	,	PUNCT
aiti-14589	113	34	rfc	rfc	PROPN
aiti-14589	113	35	,	,	PUNCT
aiti-14589	113	36	and	and	CCONJ
aiti-14589	113	37	gbc	gbc	NOUN
aiti-14589	113	38	.	.	PUNCT
aiti-14589	114	1	furthermore	furthermore	ADV
aiti-14589	114	2	,	,	PUNCT
aiti-14589	114	3	by	by	ADP
aiti-14589	114	4	using	use	VERB
aiti-14589	114	5	diverse	diverse	ADJ
aiti-14589	114	6	datasets	dataset	NOUN
aiti-14589	114	7	such	such	ADJ
aiti-14589	114	8	as	as	ADP
aiti-14589	114	9	the	the	DET
aiti-14589	114	10	delaware	delaware	PROPN
aiti-14589	114	11	pain	pain	PROPN
aiti-14589	114	12	database	database	NOUN
aiti-14589	114	13	and	and	CCONJ
aiti-14589	114	14	utkface	utkface	NOUN
aiti-14589	114	15	,	,	PUNCT
aiti-14589	114	16	the	the	DET
aiti-14589	114	17	study	study	NOUN
aiti-14589	114	18	aims	aim	VERB
aiti-14589	114	19	to	to	PART
aiti-14589	114	20	improve	improve	VERB
aiti-14589	114	21	generalizability	generalizability	NOUN
aiti-14589	114	22	across	across	ADP
aiti-14589	114	23	different	different	ADJ
aiti-14589	114	24	demographics	demographic	NOUN
aiti-14589	114	25	and	and	CCONJ
aiti-14589	114	26	application	application	NOUN
aiti-14589	114	27	domains	domain	NOUN
aiti-14589	114	28	.	.	PUNCT
aiti-14589	115	1	this	this	DET
aiti-14589	115	2	research	research	NOUN
aiti-14589	115	3	extends	extend	VERB
aiti-14589	115	4	the	the	DET
aiti-14589	115	5	findings	finding	NOUN
aiti-14589	115	6	of	of	ADP
aiti-14589	115	7	lee	lee	PROPN
aiti-14589	115	8	and	and	CCONJ
aiti-14589	115	9	park	park	NOUN
aiti-14589	115	10	[	[	X
aiti-14589	115	11	20	20	NUM
aiti-14589	115	12	]	]	PUNCT
aiti-14589	115	13	,	,	PUNCT
aiti-14589	115	14	hangaragi	hangaragi	PROPN
aiti-14589	115	15	et	et	PROPN
aiti-14589	115	16	al	al	PROPN
aiti-14589	115	17	.	.	PUNCT
aiti-14589	116	1	[	[	X
aiti-14589	116	2	28	28	NUM
aiti-14589	116	3	]	]	PUNCT
aiti-14589	116	4	,	,	PUNCT
aiti-14589	116	5	and	and	CCONJ
aiti-14589	116	6	di	di	PROPN
aiti-14589	116	7	luzio	luzio	PROPN
aiti-14589	116	8	et	et	PROPN
aiti-14589	116	9	al	al	PROPN
aiti-14589	116	10	.	.	PUNCT
aiti-14589	117	1	[	[	X
aiti-14589	117	2	27	27	NUM
aiti-14589	117	3	]	]	PUNCT
aiti-14589	117	4	,	,	PUNCT
aiti-14589	117	5	providing	provide	VERB
aiti-14589	117	6	a	a	DET
aiti-14589	117	7	comprehensive	comprehensive	ADJ
aiti-14589	117	8	performance	performance	NOUN
aiti-14589	117	9	analysis	analysis	NOUN
aiti-14589	117	10	of	of	ADP
aiti-14589	117	11	image	image	NOUN
aiti-14589	117	12	-	-	PUNCT
aiti-14589	117	13	based	base	VERB
aiti-14589	117	14	and	and	CCONJ
aiti-14589	117	15	landmark	landmark	NOUN
aiti-14589	117	16	-	-	PUNCT
aiti-14589	117	17	based	base	VERB
aiti-14589	117	18	fer	fer	PROPN
aiti-14589	117	19	techniques	technique	NOUN
aiti-14589	117	20	across	across	ADP
aiti-14589	117	21	multiple	multiple	ADJ
aiti-14589	117	22	ml	ml	NOUN
aiti-14589	117	23	classifiers	classifier	NOUN
aiti-14589	117	24	.	.	PUNCT
aiti-14589	118	1	table	table	NOUN
aiti-14589	118	2	1	1	NUM
aiti-14589	118	3	chronological	chronological	ADJ
aiti-14589	118	4	comparison	comparison	NOUN
aiti-14589	118	5	of	of	ADP
aiti-14589	118	6	studies	study	NOUN
aiti-14589	118	7	study	study	NOUN
aiti-14589	118	8	methodology	methodology	NOUN
aiti-14589	118	9	models	model	NOUN
aiti-14589	118	10	used	use	VERB
aiti-14589	118	11	key	key	ADJ
aiti-14589	118	12	findings	finding	NOUN
aiti-14589	118	13	limitation	limitation	NOUN
aiti-14589	118	14	novelty	novelty	NOUN
aiti-14589	118	15	in	in	ADP
aiti-14589	118	16	study	study	NOUN
aiti-14589	118	17	buck	buck	NOUN
aiti-14589	118	18	et	et	PROPN
aiti-14589	118	19	al	al	PROPN
aiti-14589	118	20	.	.	PUNCT
aiti-14589	119	1	[	[	X
aiti-14589	119	2	5	5	NUM
aiti-14589	119	3	]	]	PUNCT
aiti-14589	119	4	,	,	PUNCT
aiti-14589	119	5	1972	1972	NUM
aiti-14589	119	6	image	image	NOUN
aiti-14589	119	7	-	-	PUNCT
aiti-14589	119	8	based	base	VERB
aiti-14589	119	9	human	human	NOUN
aiti-14589	119	10	-	-	PUNCT
aiti-14589	119	11	coded	code	VERB
aiti-14589	119	12	expressions	expression	NOUN
aiti-14589	119	13	early	early	ADJ
aiti-14589	119	14	facial	facial	ADJ
aiti-14589	119	15	expression	expression	NOUN
aiti-14589	119	16	analysis	analysis	NOUN
aiti-14589	119	17	no	no	DET
aiti-14589	119	18	automated	automate	VERB
aiti-14589	119	19	recognition	recognition	NOUN
aiti-14589	119	20	introduces	introduce	VERB
aiti-14589	119	21	ml	ml	VERB
aiti-14589	119	22	-	-	PUNCT
aiti-14589	119	23	based	base	VERB
aiti-14589	119	24	automated	automate	VERB
aiti-14589	119	25	recognition	recognition	NOUN
aiti-14589	119	26	vapnik	vapnik	X
aiti-14589	120	1	[	[	X
aiti-14589	120	2	17	17	NUM
aiti-14589	120	3	]	]	PUNCT
aiti-14589	120	4	,	,	PUNCT
aiti-14589	120	5	1999	1999	NUM
aiti-14589	120	6	image	image	NOUN
aiti-14589	120	7	-	-	PUNCT
aiti-14589	120	8	based	base	VERB
aiti-14589	120	9	svm	svm	PROPN
aiti-14589	120	10	foundational	foundational	PROPN
aiti-14589	120	11	ml	ml	PROPN
aiti-14589	120	12	theory	theory	NOUN
aiti-14589	120	13	no	no	DET
aiti-14589	120	14	fer	fer	PROPN
aiti-14589	120	15	application	application	NOUN
aiti-14589	120	16	test	test	NOUN
aiti-14589	120	17	svm	svm	NOUN
aiti-14589	120	18	on	on	ADP
aiti-14589	120	19	fer	fer	PROPN
aiti-14589	120	20	tasks	task	NOUN
aiti-14589	120	21	lisetti	lisetti	NOUN
aiti-14589	120	22	and	and	CCONJ
aiti-14589	120	23	schiano	schiano	X
aiti-14589	121	1	[	[	X
aiti-14589	121	2	6	6	NUM
aiti-14589	121	3	]	]	PUNCT
aiti-14589	121	4	,	,	PUNCT
aiti-14589	121	5	2000	2000	NUM
aiti-14589	121	6	image	image	NOUN
aiti-14589	121	7	-	-	PUNCT
aiti-14589	121	8	based	base	VERB
aiti-14589	121	9	ai	ai	NOUN
aiti-14589	121	10	-	-	PUNCT
aiti-14589	121	11	based	base	VERB
aiti-14589	121	12	interpretation	interpretation	NOUN
aiti-14589	121	13	human	human	ADJ
aiti-14589	121	14	-	-	PUNCT
aiti-14589	121	15	computer	computer	NOUN
aiti-14589	121	16	interaction	interaction	NOUN
aiti-14589	121	17	focus	focus	VERB
aiti-14589	121	18	no	no	DET
aiti-14589	121	19	ml	ml	INTJ
aiti-14589	121	20	comparison	comparison	NOUN
aiti-14589	121	21	broadens	broaden	VERB
aiti-14589	121	22	scope	scope	NOUN
aiti-14589	121	23	to	to	ADP
aiti-14589	121	24	modern	modern	ADJ
aiti-14589	121	25	ml	ml	NOUN
aiti-14589	121	26	models	model	NOUN
aiti-14589	121	27	picard	picard	PROPN
aiti-14589	121	28	et	et	PROPN
aiti-14589	121	29	al	al	PROPN
aiti-14589	121	30	.	.	PUNCT
aiti-14589	122	1	[	[	X
aiti-14589	122	2	18	18	NUM
aiti-14589	122	3	]	]	PUNCT
aiti-14589	122	4	,	,	PUNCT
aiti-14589	122	5	2001	2001	NUM
aiti-14589	122	6	image	image	NOUN
aiti-14589	122	7	-	-	PUNCT
aiti-14589	122	8	based	base	VERB
aiti-14589	122	9	physiological	physiological	ADJ
aiti-14589	122	10	state	state	NOUN
aiti-14589	122	11	analysis	analysis	NOUN
aiti-14589	122	12	early	early	ADV
aiti-14589	122	13	ml	ml	NOUN
aiti-14589	122	14	in	in	ADP
aiti-14589	122	15	affective	affective	NOUN
aiti-14589	122	16	computing	computing	NOUN
aiti-14589	122	17	no	no	DET
aiti-14589	122	18	landmark	landmark	NOUN
aiti-14589	122	19	feature	feature	NOUN
aiti-14589	122	20	use	use	NOUN
aiti-14589	122	21	evaluates	evaluate	NOUN
aiti-14589	122	22	landmark	landmark	PROPN
aiti-14589	122	23	features	feature	VERB
aiti-14589	122	24	bartlett	bartlett	PROPN
aiti-14589	122	25	et	et	PROPN
aiti-14589	122	26	al	al	PROPN
aiti-14589	122	27	.	.	PUNCT
aiti-14589	123	1	[	[	X
aiti-14589	123	2	19	19	NUM
aiti-14589	123	3	]	]	PUNCT
aiti-14589	123	4	,	,	PUNCT
aiti-14589	123	5	2003	2003	NUM
aiti-14589	123	6	image	image	NOUN
aiti-14589	123	7	-	-	PUNCT
aiti-14589	123	8	based	base	VERB
aiti-14589	123	9	cnn	cnn	PROPN
aiti-14589	123	10	real	real	ADJ
aiti-14589	123	11	-	-	PUNCT
aiti-14589	123	12	time	time	NOUN
aiti-14589	123	13	fer	fer	PROPN
aiti-14589	123	14	limited	limit	VERB
aiti-14589	123	15	dataset	dataset	NOUN
aiti-14589	123	16	uses	use	VERB
aiti-14589	123	17	multiple	multiple	ADJ
aiti-14589	123	18	datasets	dataset	NOUN
aiti-14589	123	19	for	for	ADP
aiti-14589	123	20	comparison	comparison	NOUN
aiti-14589	123	21	lu	lu	NOUN
aiti-14589	123	22	and	and	CCONJ
aiti-14589	123	23	weng	weng	PROPN
aiti-14589	124	1	[	[	X
aiti-14589	124	2	10	10	NUM
aiti-14589	124	3	]	]	PUNCT
aiti-14589	124	4	,	,	PUNCT
aiti-14589	124	5	2007	2007	NUM
aiti-14589	124	6	image	image	NOUN
aiti-14589	124	7	-	-	PUNCT
aiti-14589	124	8	based	base	VERB
aiti-14589	124	9	image	image	NOUN
aiti-14589	124	10	classification	classification	NOUN
aiti-14589	124	11	survey	survey	NOUN
aiti-14589	124	12	overview	overview	NOUN
aiti-14589	124	13	of	of	ADP
aiti-14589	124	14	classification	classification	NOUN
aiti-14589	124	15	methods	method	NOUN
aiti-14589	124	16	no	no	DET
aiti-14589	124	17	specific	specific	ADJ
aiti-14589	124	18	fer	fer	PROPN
aiti-14589	124	19	application	application	NOUN
aiti-14589	124	20	direct	direct	ADJ
aiti-14589	124	21	application	application	NOUN
aiti-14589	124	22	to	to	PART
aiti-14589	124	23	fer	fer	PROPN
aiti-14589	124	24	hilbe	hilbe	PROPN
aiti-14589	124	25	[	[	X
aiti-14589	124	26	13	13	NUM
aiti-14589	124	27	]	]	PUNCT
aiti-14589	124	28	,	,	PUNCT
aiti-14589	124	29	2009	2009	NUM
aiti-14589	124	30	image	image	NOUN
aiti-14589	124	31	-	-	PUNCT
aiti-14589	124	32	based	base	VERB
aiti-14589	124	33	lrc	lrc	PROPN
aiti-14589	124	34	foundational	foundational	PROPN
aiti-14589	124	35	statistical	statistical	ADJ
aiti-14589	124	36	modeling	modeling	NOUN
aiti-14589	124	37	no	no	DET
aiti-14589	124	38	comparison	comparison	NOUN
aiti-14589	124	39	with	with	ADP
aiti-14589	124	40	deep	deep	ADJ
aiti-14589	124	41	learning	learning	NOUN
aiti-14589	124	42	compares	compare	VERB
aiti-14589	124	43	classical	classical	ADJ
aiti-14589	124	44	ml	ml	NOUN
aiti-14589	124	45	vs.	vs.	ADP
aiti-14589	124	46	dl	dl	PROPN
aiti-14589	124	47	models	model	NOUN
aiti-14589	124	48	assari	assari	ADV
aiti-14589	124	49	and	and	CCONJ
aiti-14589	124	50	rahmati	rahmati	NOUN
aiti-14589	124	51	[	[	X
aiti-14589	124	52	22	22	NUM
aiti-14589	124	53	]	]	PUNCT
aiti-14589	124	54	,	,	PUNCT
aiti-14589	124	55	2011	2011	NUM
aiti-14589	124	56	image	image	NOUN
aiti-14589	124	57	-	-	PUNCT
aiti-14589	124	58	based	base	VERB
aiti-14589	124	59	fer	fer	NOUN
aiti-14589	124	60	-	-	PUNCT
aiti-14589	124	61	based	base	VERB
aiti-14589	124	62	drowsiness	drowsiness	NOUN
aiti-14589	124	63	detection	detection	NOUN
aiti-14589	124	64	application	application	NOUN
aiti-14589	124	65	-	-	PUNCT
aiti-14589	124	66	specific	specific	ADJ
aiti-14589	124	67	no	no	DET
aiti-14589	124	68	direct	direct	ADJ
aiti-14589	124	69	comparison	comparison	NOUN
aiti-14589	124	70	with	with	ADP
aiti-14589	124	71	landmark	landmark	NOUN
aiti-14589	124	72	-	-	PUNCT
aiti-14589	124	73	based	base	VERB
aiti-14589	124	74	methods	method	NOUN
aiti-14589	124	75	generalizes	generalize	VERB
aiti-14589	124	76	findings	finding	NOUN
aiti-14589	124	77	beyond	beyond	ADP
aiti-14589	124	78	drowsiness	drowsiness	NOUN
aiti-14589	124	79	detection	detection	NOUN
aiti-14589	124	80	hamester	hamester	NOUN
aiti-14589	124	81	et	et	PROPN
aiti-14589	124	82	al	al	PROPN
aiti-14589	124	83	.	.	PUNCT
aiti-14589	125	1	[	[	X
aiti-14589	125	2	23	23	NUM
aiti-14589	125	3	]	]	PUNCT
aiti-14589	125	4	,	,	PUNCT
aiti-14589	125	5	2015	2015	NUM
aiti-14589	125	6	image	image	NOUN
aiti-14589	125	7	-	-	PUNCT
aiti-14589	125	8	based	base	VERB
aiti-14589	125	9	cnn	cnn	PROPN
aiti-14589	125	10	2	2	NUM
aiti-14589	125	11	-	-	PUNCT
aiti-14589	125	12	channel	channel	NOUN
aiti-14589	125	13	cnn	cnn	NOUN
aiti-14589	125	14	for	for	ADP
aiti-14589	125	15	expression	expression	NOUN
aiti-14589	125	16	recognition	recognition	NOUN
aiti-14589	125	17	no	no	DET
aiti-14589	125	18	comparison	comparison	NOUN
aiti-14589	125	19	with	with	ADP
aiti-14589	125	20	classical	classical	ADJ
aiti-14589	125	21	ml	ml	NOUN
aiti-14589	125	22	models	model	NOUN
aiti-14589	125	23	broadens	broaden	VERB
aiti-14589	125	24	evaluation	evaluation	NOUN
aiti-14589	125	25	across	across	ADP
aiti-14589	125	26	ml	ml	NOUN
aiti-14589	125	27	models	model	NOUN
aiti-14589	125	28	bekele	bekele	PROPN
aiti-14589	125	29	et	et	PROPN
aiti-14589	125	30	al	al	PROPN
aiti-14589	125	31	.	.	PUNCT
aiti-14589	126	1	[	[	X
aiti-14589	126	2	21	21	NUM
aiti-14589	126	3	]	]	X
aiti-14589	126	4	,	,	PUNCT
aiti-14589	126	5	2017	2017	NUM
aiti-14589	126	6	image	image	NOUN
aiti-14589	126	7	-	-	PUNCT
aiti-14589	126	8	based	base	VERB
aiti-14589	126	9	vr	vr	NOUN
aiti-14589	126	10	-	-	NOUN
aiti-14589	126	11	saafe	saafe	NOUN
aiti-14589	126	12	for	for	ADP
aiti-14589	126	13	fer	fer	PROPN
aiti-14589	126	14	application	application	NOUN
aiti-14589	126	15	-	-	PUNCT
aiti-14589	126	16	specific	specific	ADJ
aiti-14589	126	17	no	no	DET
aiti-14589	126	18	classical	classical	ADJ
aiti-14589	126	19	ml	ml	NOUN
aiti-14589	126	20	comparison	comparison	NOUN
aiti-14589	126	21	expands	expand	VERB
aiti-14589	126	22	analysis	analysis	NOUN
aiti-14589	126	23	to	to	ADP
aiti-14589	126	24	classical	classical	ADJ
aiti-14589	126	25	ml	ml	NOUN
aiti-14589	126	26	models	model	NOUN
aiti-14589	126	27	söylemez	söylemez	PROPN
aiti-14589	126	28	et	et	PROPN
aiti-14589	126	29	al	al	PROPN
aiti-14589	126	30	.	.	PUNCT
aiti-14589	127	1	[	[	X
aiti-14589	127	2	25	25	NUM
aiti-14589	127	3	]	]	PUNCT
aiti-14589	127	4	,	,	PUNCT
aiti-14589	127	5	2017	2017	NUM
aiti-14589	127	6	landmarkbased	landmarkbase	VERB
aiti-14589	127	7	svm	svm	ADJ
aiti-14589	127	8	distance	distance	NOUN
aiti-14589	127	9	-	-	PUNCT
aiti-14589	127	10	based	base	VERB
aiti-14589	127	11	features	feature	NOUN
aiti-14589	127	12	for	for	ADP
aiti-14589	127	13	3d	3d	PROPN
aiti-14589	127	14	fer	fer	PROPN
aiti-14589	128	1	no	no	DET
aiti-14589	128	2	image	image	NOUN
aiti-14589	128	3	-	-	PUNCT
aiti-14589	128	4	based	base	VERB
aiti-14589	128	5	comparison	comparison	NOUN
aiti-14589	128	6	extends	extend	VERB
aiti-14589	128	7	analysis	analysis	NOUN
aiti-14589	128	8	to	to	ADP
aiti-14589	128	9	both	both	DET
aiti-14589	128	10	2d	2d	NOUN
aiti-14589	128	11	and	and	CCONJ
aiti-14589	128	12	image	image	NOUN
aiti-14589	128	13	-	-	PUNCT
aiti-14589	128	14	based	base	VERB
aiti-14589	128	15	methods	method	NOUN
aiti-14589	128	16	advances	advance	VERB
aiti-14589	128	17	in	in	ADP
aiti-14589	128	18	technology	technology	NOUN
aiti-14589	128	19	innovation	innovation	NOUN
aiti-14589	128	20	,	,	PUNCT
aiti-14589	128	21	vol	vol	NOUN
aiti-14589	128	22	.	.	PROPN
aiti-14589	129	1	10	10	NUM
aiti-14589	129	2	,	,	PUNCT
aiti-14589	129	3	no	no	INTJ
aiti-14589	129	4	.	.	NOUN
aiti-14589	129	5	4	4	NUM
aiti-14589	129	6	,	,	PUNCT
aiti-14589	129	7	2025	2025	NUM
aiti-14589	129	8	,	,	PUNCT
aiti-14589	129	9	pp	pp	ADJ
aiti-14589	129	10	.	.	PUNCT
aiti-14589	130	1	395	395	NUM
aiti-14589	130	2	-	-	SYM
aiti-14589	130	3	406	406	NUM
aiti-14589	130	4	399	399	NUM
aiti-14589	130	5	table	table	NOUN
aiti-14589	130	6	1	1	NUM
aiti-14589	130	7	chronological	chronological	ADJ
aiti-14589	130	8	comparison	comparison	NOUN
aiti-14589	130	9	of	of	ADP
aiti-14589	130	10	studies	study	NOUN
aiti-14589	130	11	(	(	PUNCT
aiti-14589	130	12	continued	continued	ADJ
aiti-14589	130	13	)	)	PUNCT
aiti-14589	130	14	study	study	NOUN
aiti-14589	130	15	methodology	methodology	NOUN
aiti-14589	130	16	models	model	NOUN
aiti-14589	130	17	used	use	VERB
aiti-14589	130	18	key	key	ADJ
aiti-14589	130	19	findings	finding	NOUN
aiti-14589	130	20	limitation	limitation	NOUN
aiti-14589	130	21	novelty	novelty	NOUN
aiti-14589	130	22	in	in	ADP
aiti-14589	130	23	study	study	NOUN
aiti-14589	130	24	alsubari	alsubari	PROPN
aiti-14589	130	25	et	et	PROPN
aiti-14589	130	26	al	al	PROPN
aiti-14589	130	27	.	.	PUNCT
aiti-14589	131	1	[	[	X
aiti-14589	131	2	24	24	NUM
aiti-14589	131	3	]	]	PUNCT
aiti-14589	131	4	,	,	PUNCT
aiti-14589	131	5	2017	2017	NUM
aiti-14589	131	6	image	image	NOUN
aiti-14589	131	7	-	-	PUNCT
aiti-14589	131	8	based	base	VERB
aiti-14589	131	9	wavelet	wavelet	NOUN
aiti-14589	131	10	transform	transform	NOUN
aiti-14589	131	11	+	+	CCONJ
aiti-14589	131	12	lbp	lbp	NOUN
aiti-14589	131	13	feature	feature	NOUN
aiti-14589	131	14	fusion	fusion	NOUN
aiti-14589	131	15	improves	improve	VERB
aiti-14589	131	16	recognition	recognition	NOUN
aiti-14589	131	17	no	no	DET
aiti-14589	131	18	landmark	landmark	NOUN
aiti-14589	131	19	feature	feature	NOUN
aiti-14589	131	20	analysis	analysis	NOUN
aiti-14589	131	21	incorporates	incorporate	VERB
aiti-14589	131	22	landmark	landmark	NOUN
aiti-14589	131	23	features	feature	NOUN
aiti-14589	131	24	for	for	ADP
aiti-14589	131	25	comparison	comparison	NOUN
aiti-14589	131	26	munasinghe	munasinghe	NOUN
aiti-14589	132	1	[	[	X
aiti-14589	132	2	26	26	NUM
aiti-14589	132	3	]	]	PUNCT
aiti-14589	132	4	,	,	PUNCT
aiti-14589	132	5	2018	2018	NUM
aiti-14589	132	6	landmarkbased	landmarkbase	VERB
aiti-14589	132	7	rfc	rfc	NOUN
aiti-14589	132	8	fast	fast	ADJ
aiti-14589	132	9	training	training	NOUN
aiti-14589	132	10	limited	limited	ADJ
aiti-14589	132	11	dataset	dataset	NOUN
aiti-14589	132	12	extends	extend	VERB
aiti-14589	132	13	analysis	analysis	NOUN
aiti-14589	132	14	with	with	ADP
aiti-14589	132	15	a	a	DET
aiti-14589	132	16	larger	large	ADJ
aiti-14589	132	17	dataset	dataset	NOUN
aiti-14589	132	18	michael	michael	PROPN
aiti-14589	132	19	revina	revina	PROPN
aiti-14589	132	20	and	and	CCONJ
aiti-14589	132	21	sam	sam	PROPN
aiti-14589	132	22	emmanuel	emmanuel	PROPN
aiti-14589	133	1	[	[	X
aiti-14589	133	2	1	1	NUM
aiti-14589	133	3	]	]	PUNCT
aiti-14589	133	4	,	,	PUNCT
aiti-14589	133	5	2021	2021	NUM
aiti-14589	133	6	image	image	NOUN
aiti-14589	133	7	-	-	PUNCT
aiti-14589	133	8	based	base	VERB
aiti-14589	133	9	cnn	cnn	PROPN
aiti-14589	133	10	high	high	ADJ
aiti-14589	133	11	accuracy	accuracy	NOUN
aiti-14589	133	12	computationally	computationally	ADV
aiti-14589	133	13	expensive	expensive	ADJ
aiti-14589	133	14	direct	direct	ADJ
aiti-14589	133	15	comparison	comparison	NOUN
aiti-14589	133	16	with	with	ADP
aiti-14589	133	17	landmark	landmark	NOUN
aiti-14589	133	18	-	-	PUNCT
aiti-14589	133	19	based	base	VERB
aiti-14589	133	20	methods	method	NOUN
aiti-14589	133	21	lee	lee	PROPN
aiti-14589	133	22	and	and	CCONJ
aiti-14589	133	23	park	park	NOUN
aiti-14589	133	24	[	[	X
aiti-14589	133	25	20	20	NUM
aiti-14589	133	26	]	]	PUNCT
aiti-14589	133	27	,	,	PUNCT
aiti-14589	133	28	2022	2022	NUM
aiti-14589	133	29	image	image	NOUN
aiti-14589	133	30	-	-	PUNCT
aiti-14589	133	31	based	base	VERB
aiti-14589	133	32	fast	fast	ADJ
aiti-14589	133	33	r	r	NOUN
aiti-14589	133	34	-	-	PUNCT
aiti-14589	133	35	cnn	cnn	NOUN
aiti-14589	133	36	fer	fer	PROPN
aiti-14589	133	37	-	-	PUNCT
aiti-14589	133	38	based	base	VERB
aiti-14589	133	39	depressive	depressive	ADJ
aiti-14589	133	40	disorder	disorder	NOUN
aiti-14589	133	41	diagnosis	diagnosis	NOUN
aiti-14589	133	42	no	no	DET
aiti-14589	133	43	landmark	landmark	NOUN
aiti-14589	133	44	-	-	PUNCT
aiti-14589	133	45	based	base	VERB
aiti-14589	133	46	evaluation	evaluation	NOUN
aiti-14589	133	47	applied	apply	VERB
aiti-14589	133	48	fer	fer	PROPN
aiti-14589	133	49	to	to	ADP
aiti-14589	133	50	clinical	clinical	ADJ
aiti-14589	133	51	diagnostics	diagnostic	NOUN
aiti-14589	133	52	sharma	sharma	PROPN
aiti-14589	133	53	et	et	PROPN
aiti-14589	133	54	al	al	PROPN
aiti-14589	133	55	.	.	PUNCT
aiti-14589	134	1	[	[	X
aiti-14589	134	2	9	9	NUM
aiti-14589	134	3	]	]	PUNCT
aiti-14589	134	4	,	,	PUNCT
aiti-14589	134	5	2023	2023	NUM
aiti-14589	134	6	landmarkbased	landmarkbase	VERB
aiti-14589	134	7	svm	svm	NOUN
aiti-14589	134	8	improved	improve	VERB
aiti-14589	134	9	robustness	robustness	NOUN
aiti-14589	134	10	sensitive	sensitive	ADJ
aiti-14589	134	11	to	to	ADP
aiti-14589	134	12	detection	detection	NOUN
aiti-14589	134	13	errors	error	NOUN
aiti-14589	134	14	evaluates	evaluate	VERB
aiti-14589	134	15	multiple	multiple	ADJ
aiti-14589	134	16	classifiers	classifier	NOUN
aiti-14589	134	17	di	di	VERB
aiti-14589	134	18	luzio	luzio	PROPN
aiti-14589	134	19	et	et	PROPN
aiti-14589	134	20	al	al	PROPN
aiti-14589	134	21	.	.	PUNCT
aiti-14589	135	1	[	[	X
aiti-14589	135	2	27	27	NUM
aiti-14589	135	3	]	]	PUNCT
aiti-14589	135	4	,	,	PUNCT
aiti-14589	135	5	2023	2023	NUM
aiti-14589	135	6	landmarkbased	landmarkbase	VERB
aiti-14589	135	7	dnn	dnn	PROPN
aiti-14589	135	8	hierarchical	hierarchical	ADJ
aiti-14589	135	9	features	feature	VERB
aiti-14589	135	10	no	no	DET
aiti-14589	135	11	comparison	comparison	NOUN
aiti-14589	135	12	with	with	ADP
aiti-14589	135	13	image	image	NOUN
aiti-14589	135	14	-	-	PUNCT
aiti-14589	135	15	based	base	VERB
aiti-14589	135	16	methods	method	NOUN
aiti-14589	135	17	direct	direct	ADJ
aiti-14589	135	18	performance	performance	NOUN
aiti-14589	135	19	comparison	comparison	NOUN
aiti-14589	135	20	hangaragi	hangaragi	PROPN
aiti-14589	135	21	et	et	PROPN
aiti-14589	135	22	al	al	PROPN
aiti-14589	135	23	.	.	PUNCT
aiti-14589	136	1	[	[	X
aiti-14589	136	2	28	28	NUM
aiti-14589	136	3	]	]	PUNCT
aiti-14589	136	4	,	,	PUNCT
aiti-14589	136	5	2023	2023	NUM
aiti-14589	136	6	image	image	NOUN
aiti-14589	136	7	-	-	PUNCT
aiti-14589	136	8	based	base	VERB
aiti-14589	136	9	face	face	NOUN
aiti-14589	136	10	mesh	mesh	NOUN
aiti-14589	136	11	+	+	CCONJ
aiti-14589	136	12	dnn	dnn	PROPN
aiti-14589	136	13	efficient	efficient	ADJ
aiti-14589	136	14	for	for	ADP
aiti-14589	136	15	real	real	ADJ
aiti-14589	136	16	-	-	PUNCT
aiti-14589	136	17	time	time	NOUN
aiti-14589	136	18	applications	application	NOUN
aiti-14589	136	19	lacks	lack	VERB
aiti-14589	136	20	comparison	comparison	NOUN
aiti-14589	136	21	with	with	ADP
aiti-14589	136	22	landmark	landmark	NOUN
aiti-14589	136	23	-	-	PUNCT
aiti-14589	136	24	based	base	VERB
aiti-14589	136	25	models	model	NOUN
aiti-14589	136	26	extends	extend	VERB
aiti-14589	136	27	analysis	analysis	NOUN
aiti-14589	136	28	with	with	ADP
aiti-14589	136	29	landmark	landmark	NOUN
aiti-14589	136	30	-	-	PUNCT
aiti-14589	136	31	based	base	VERB
aiti-14589	136	32	methods	method	NOUN
aiti-14589	136	33	this	this	DET
aiti-14589	136	34	study	study	NOUN
aiti-14589	136	35	image	image	NOUN
aiti-14589	136	36	-	-	PUNCT
aiti-14589	136	37	based	base	VERB
aiti-14589	136	38	and	and	CCONJ
aiti-14589	136	39	landmarkbased	landmarkbase	VERB
aiti-14589	136	40	cnn	cnn	PROPN
aiti-14589	136	41	,	,	PUNCT
aiti-14589	136	42	svm	svm	PROPN
aiti-14589	136	43	,	,	PUNCT
aiti-14589	136	44	lrc	lrc	PROPN
aiti-14589	136	45	,	,	PUNCT
aiti-14589	136	46	rfc	rfc	PROPN
aiti-14589	136	47	,	,	PUNCT
aiti-14589	136	48	and	and	CCONJ
aiti-14589	136	49	gbc	gbc	PROPN
aiti-14589	136	50	landmark	landmark	NOUN
aiti-14589	136	51	-	-	PUNCT
aiti-14589	136	52	based	base	VERB
aiti-14589	136	53	features	feature	NOUN
aiti-14589	136	54	enhance	enhance	VERB
aiti-14589	136	55	performance	performance	NOUN
aiti-14589	136	56	computational	computational	ADJ
aiti-14589	136	57	tradeoffs	tradeoff	NOUN
aiti-14589	136	58	first	first	ADV
aiti-14589	136	59	detailed	detailed	ADJ
aiti-14589	136	60	evaluation	evaluation	NOUN
aiti-14589	136	61	of	of	ADP
aiti-14589	136	62	both	both	DET
aiti-14589	136	63	methods	method	NOUN
aiti-14589	136	64	across	across	ADP
aiti-14589	136	65	multiple	multiple	ADJ
aiti-14589	136	66	classifiers	classifier	NOUN
aiti-14589	136	67	3	3	NUM
aiti-14589	136	68	.	.	PUNCT
aiti-14589	136	69	machine	machine	NOUN
aiti-14589	136	70	learning	learn	VERB
aiti-14589	136	71	techniques	technique	NOUN
aiti-14589	136	72	this	this	DET
aiti-14589	136	73	section	section	NOUN
aiti-14589	136	74	provides	provide	VERB
aiti-14589	136	75	an	an	DET
aiti-14589	136	76	overview	overview	NOUN
aiti-14589	136	77	of	of	ADP
aiti-14589	136	78	the	the	DET
aiti-14589	136	79	ml	ml	PROPN
aiti-14589	136	80	algorithms	algorithm	NOUN
aiti-14589	136	81	used	use	VERB
aiti-14589	136	82	in	in	ADP
aiti-14589	136	83	this	this	DET
aiti-14589	136	84	study	study	NOUN
aiti-14589	136	85	,	,	PUNCT
aiti-14589	136	86	the	the	DET
aiti-14589	136	87	preprocessing	preprocessing	NOUN
aiti-14589	136	88	steps	step	NOUN
aiti-14589	136	89	,	,	PUNCT
aiti-14589	136	90	and	and	CCONJ
aiti-14589	136	91	training	training	NOUN
aiti-14589	136	92	procedures	procedure	NOUN
aiti-14589	136	93	for	for	ADP
aiti-14589	136	94	each	each	DET
aiti-14589	136	95	approach	approach	NOUN
aiti-14589	136	96	.	.	PUNCT
aiti-14589	137	1	the	the	DET
aiti-14589	137	2	comparative	comparative	ADJ
aiti-14589	137	3	setup	setup	NOUN
aiti-14589	137	4	of	of	ADP
aiti-14589	137	5	these	these	DET
aiti-14589	137	6	methods	method	NOUN
aiti-14589	137	7	is	be	AUX
aiti-14589	137	8	summarized	summarize	VERB
aiti-14589	137	9	in	in	ADP
aiti-14589	137	10	table	table	NOUN
aiti-14589	137	11	2	2	NUM
aiti-14589	137	12	.	.	PUNCT
aiti-14589	138	1	each	each	DET
aiti-14589	138	2	algorithm	algorithm	NOUN
aiti-14589	138	3	,	,	PUNCT
aiti-14589	138	4	including	include	VERB
aiti-14589	138	5	cnn	cnn	PROPN
aiti-14589	138	6	,	,	PUNCT
aiti-14589	138	7	rfc	rfc	PROPN
aiti-14589	138	8	,	,	PUNCT
aiti-14589	138	9	lrc	lrc	PROPN
aiti-14589	138	10	,	,	PUNCT
aiti-14589	138	11	svm	svm	PROPN
aiti-14589	138	12	,	,	PUNCT
aiti-14589	138	13	and	and	CCONJ
aiti-14589	138	14	gbc	gbc	PROPN
aiti-14589	138	15	,	,	PUNCT
aiti-14589	138	16	was	be	AUX
aiti-14589	138	17	selected	select	VERB
aiti-14589	138	18	for	for	ADP
aiti-14589	138	19	its	its	PRON
aiti-14589	138	20	ability	ability	NOUN
aiti-14589	138	21	to	to	PART
aiti-14589	138	22	handle	handle	VERB
aiti-14589	138	23	both	both	CCONJ
aiti-14589	138	24	image	image	NOUN
aiti-14589	138	25	and	and	CCONJ
aiti-14589	138	26	landmark	landmark	NOUN
aiti-14589	138	27	data	datum	NOUN
aiti-14589	138	28	,	,	PUNCT
aiti-14589	138	29	with	with	ADP
aiti-14589	138	30	distinct	distinct	ADJ
aiti-14589	138	31	preprocessing	preprocessing	NOUN
aiti-14589	138	32	techniques	technique	NOUN
aiti-14589	138	33	applied	apply	VERB
aiti-14589	138	34	to	to	PART
aiti-14589	138	35	optimize	optimize	VERB
aiti-14589	138	36	model	model	NOUN
aiti-14589	138	37	performance	performance	NOUN
aiti-14589	138	38	.	.	PUNCT
aiti-14589	139	1	(	(	PUNCT
aiti-14589	139	2	1	1	X
aiti-14589	139	3	)	)	PUNCT
aiti-14589	139	4	convolutional	convolutional	ADJ
aiti-14589	139	5	neural	neural	ADJ
aiti-14589	139	6	networks	network	NOUN
aiti-14589	139	7	(	(	PUNCT
aiti-14589	139	8	cnn	cnn	PROPN
aiti-14589	139	9	)	)	PUNCT
aiti-14589	139	10	cnn	cnn	PROPN
aiti-14589	139	11	are	be	AUX
aiti-14589	139	12	widely	widely	ADV
aiti-14589	139	13	used	use	VERB
aiti-14589	139	14	in	in	ADP
aiti-14589	139	15	image	image	NOUN
aiti-14589	139	16	processing	processing	NOUN
aiti-14589	139	17	due	due	ADP
aiti-14589	139	18	to	to	ADP
aiti-14589	139	19	their	their	PRON
aiti-14589	139	20	ability	ability	NOUN
aiti-14589	139	21	to	to	PART
aiti-14589	139	22	learn	learn	VERB
aiti-14589	139	23	spatial	spatial	ADJ
aiti-14589	139	24	hierarchies	hierarchy	NOUN
aiti-14589	139	25	from	from	ADP
aiti-14589	139	26	pixel	pixel	PROPN
aiti-14589	139	27	data	data	PROPN
aiti-14589	139	28	.	.	PUNCT
aiti-14589	140	1	in	in	ADP
aiti-14589	140	2	this	this	DET
aiti-14589	140	3	study	study	NOUN
aiti-14589	140	4	,	,	PUNCT
aiti-14589	140	5	cnns	cnn	NOUN
aiti-14589	140	6	were	be	AUX
aiti-14589	140	7	applied	apply	VERB
aiti-14589	140	8	in	in	ADP
aiti-14589	140	9	both	both	PRON
aiti-14589	140	10	image	image	NOUN
aiti-14589	140	11	-	-	PUNCT
aiti-14589	140	12	based	base	VERB
aiti-14589	140	13	and	and	CCONJ
aiti-14589	140	14	landmark	landmark	NOUN
aiti-14589	140	15	-	-	PUNCT
aiti-14589	140	16	based	base	VERB
aiti-14589	140	17	approaches	approach	NOUN
aiti-14589	140	18	.	.	PUNCT
aiti-14589	141	1	for	for	ADP
aiti-14589	141	2	the	the	DET
aiti-14589	141	3	image	image	NOUN
aiti-14589	141	4	-	-	PUNCT
aiti-14589	141	5	based	base	VERB
aiti-14589	141	6	cnn	cnn	PROPN
aiti-14589	141	7	,	,	PUNCT
aiti-14589	141	8	raw	raw	ADJ
aiti-14589	141	9	images	image	NOUN
aiti-14589	141	10	were	be	AUX
aiti-14589	141	11	resized	resize	VERB
aiti-14589	141	12	to	to	ADP
aiti-14589	141	13	128×128	128×128	NUM
aiti-14589	141	14	pixels	pixel	NOUN
aiti-14589	141	15	and	and	CCONJ
aiti-14589	141	16	normalized	normalize	VERB
aiti-14589	141	17	.	.	PUNCT
aiti-14589	142	1	the	the	DET
aiti-14589	142	2	architecture	architecture	NOUN
aiti-14589	142	3	consisted	consist	VERB
aiti-14589	142	4	of	of	ADP
aiti-14589	142	5	three	three	NUM
aiti-14589	142	6	convolutional	convolutional	ADJ
aiti-14589	142	7	layers	layer	NOUN
aiti-14589	142	8	with	with	ADP
aiti-14589	142	9	rectified	rectified	ADJ
aiti-14589	142	10	linear	linear	NOUN
aiti-14589	142	11	unit	unit	NOUN
aiti-14589	142	12	(	(	PUNCT
aiti-14589	142	13	relu	relu	NOUN
aiti-14589	142	14	)	)	PUNCT
aiti-14589	142	15	activation	activation	NOUN
aiti-14589	142	16	,	,	PUNCT
aiti-14589	142	17	followed	follow	VERB
aiti-14589	142	18	by	by	ADP
aiti-14589	142	19	max	max	PROPN
aiti-14589	142	20	-	-	PUNCT
aiti-14589	142	21	pooling	pool	VERB
aiti-14589	142	22	operations	operation	NOUN
aiti-14589	142	23	to	to	PART
aiti-14589	142	24	reduce	reduce	VERB
aiti-14589	142	25	spatial	spatial	ADJ
aiti-14589	142	26	dimensions	dimension	NOUN
aiti-14589	142	27	.	.	PUNCT
aiti-14589	143	1	the	the	DET
aiti-14589	143	2	output	output	NOUN
aiti-14589	143	3	was	be	AUX
aiti-14589	143	4	flattened	flatten	VERB
aiti-14589	143	5	and	and	CCONJ
aiti-14589	143	6	passed	pass	VERB
aiti-14589	143	7	through	through	ADP
aiti-14589	143	8	fully	fully	ADV
aiti-14589	143	9	connected	connected	ADJ
aiti-14589	143	10	layers	layer	NOUN
aiti-14589	143	11	,	,	PUNCT
aiti-14589	143	12	ending	end	VERB
aiti-14589	143	13	with	with	ADP
aiti-14589	143	14	a	a	DET
aiti-14589	143	15	sigmoid	sigmoid	NOUN
aiti-14589	143	16	-	-	PUNCT
aiti-14589	143	17	activated	activate	VERB
aiti-14589	143	18	neuron	neuron	NOUN
aiti-14589	143	19	for	for	ADP
aiti-14589	143	20	binary	binary	ADJ
aiti-14589	143	21	classification	classification	NOUN
aiti-14589	143	22	.	.	PUNCT
aiti-14589	144	1	the	the	DET
aiti-14589	144	2	adam	adam	PROPN
aiti-14589	144	3	optimizer	optimizer	NOUN
aiti-14589	144	4	and	and	CCONJ
aiti-14589	144	5	binary	binary	PROPN
aiti-14589	144	6	crossentropy	crossentropy	NOUN
aiti-14589	144	7	loss	loss	NOUN
aiti-14589	144	8	were	be	AUX
aiti-14589	144	9	used	use	VERB
aiti-14589	144	10	for	for	ADP
aiti-14589	144	11	training	training	NOUN
aiti-14589	144	12	.	.	PUNCT
aiti-14589	145	1	for	for	ADP
aiti-14589	145	2	the	the	DET
aiti-14589	145	3	landmark	landmark	NOUN
aiti-14589	145	4	-	-	PUNCT
aiti-14589	145	5	based	base	VERB
aiti-14589	145	6	cnn	cnn	PROPN
aiti-14589	145	7	,	,	PUNCT
aiti-14589	145	8	images	image	NOUN
aiti-14589	145	9	were	be	AUX
aiti-14589	145	10	processed	process	VERB
aiti-14589	145	11	using	use	VERB
aiti-14589	145	12	mediapipe	mediapipe	NOUN
aiti-14589	145	13	to	to	PART
aiti-14589	145	14	extract	extract	VERB
aiti-14589	145	15	468	468	NUM
aiti-14589	145	16	facial	facial	ADJ
aiti-14589	145	17	landmarks	landmark	NOUN
aiti-14589	145	18	,	,	PUNCT
aiti-14589	145	19	which	which	PRON
aiti-14589	145	20	were	be	AUX
aiti-14589	145	21	reshaped	reshape	VERB
aiti-14589	145	22	into	into	ADP
aiti-14589	145	23	a	a	DET
aiti-14589	145	24	1d	1d	NUM
aiti-14589	145	25	input	input	NOUN
aiti-14589	145	26	format	format	NOUN
aiti-14589	145	27	.	.	PUNCT
aiti-14589	146	1	the	the	DET
aiti-14589	146	2	cnn	cnn	PROPN
aiti-14589	146	3	architecture	architecture	NOUN
aiti-14589	146	4	included	include	VERB
aiti-14589	146	5	a	a	DET
aiti-14589	146	6	1d	1d	NUM
aiti-14589	146	7	convolutional	convolutional	ADJ
aiti-14589	146	8	layer	layer	NOUN
aiti-14589	146	9	,	,	PUNCT
aiti-14589	146	10	followed	follow	VERB
aiti-14589	146	11	by	by	ADP
aiti-14589	146	12	max	max	PROPN
aiti-14589	146	13	-	-	PUNCT
aiti-14589	146	14	pooling	pooling	ADJ
aiti-14589	146	15	,	,	PUNCT
aiti-14589	146	16	flattening	flattening	NOUN
aiti-14589	146	17	,	,	PUNCT
aiti-14589	146	18	and	and	CCONJ
aiti-14589	146	19	dense	dense	ADJ
aiti-14589	146	20	layers	layer	NOUN
aiti-14589	146	21	,	,	PUNCT
aiti-14589	146	22	similar	similar	ADJ
aiti-14589	146	23	to	to	ADP
aiti-14589	146	24	the	the	DET
aiti-14589	146	25	image	image	NOUN
aiti-14589	146	26	-	-	PUNCT
aiti-14589	146	27	based	base	VERB
aiti-14589	146	28	approach	approach	NOUN
aiti-14589	146	29	.	.	PUNCT
aiti-14589	147	1	feature	feature	NOUN
aiti-14589	147	2	standardization	standardization	NOUN
aiti-14589	147	3	was	be	AUX
aiti-14589	147	4	applied	apply	VERB
aiti-14589	147	5	using	use	VERB
aiti-14589	147	6	the	the	DET
aiti-14589	147	7	standardscaler	standardscaler	NOUN
aiti-14589	147	8	.	.	PUNCT
aiti-14589	148	1	(	(	PUNCT
aiti-14589	148	2	2	2	X
aiti-14589	148	3	)	)	PUNCT
aiti-14589	148	4	random	random	ADJ
aiti-14589	148	5	forest	forest	NOUN
aiti-14589	148	6	classifier	classifier	NOUN
aiti-14589	148	7	(	(	PUNCT
aiti-14589	148	8	rfc	rfc	PROPN
aiti-14589	148	9	)	)	PUNCT
aiti-14589	148	10	rfc	rfc	PROPN
aiti-14589	148	11	is	be	AUX
aiti-14589	148	12	an	an	DET
aiti-14589	148	13	ensemble	ensemble	ADJ
aiti-14589	148	14	learning	learning	NOUN
aiti-14589	148	15	method	method	NOUN
aiti-14589	148	16	that	that	PRON
aiti-14589	148	17	constructs	construct	VERB
aiti-14589	148	18	multiple	multiple	ADJ
aiti-14589	148	19	decision	decision	NOUN
aiti-14589	148	20	trees	tree	NOUN
aiti-14589	148	21	and	and	CCONJ
aiti-14589	148	22	aggregates	aggregate	VERB
aiti-14589	148	23	their	their	PRON
aiti-14589	148	24	predictions	prediction	NOUN
aiti-14589	148	25	to	to	PART
aiti-14589	148	26	enhance	enhance	VERB
aiti-14589	148	27	accuracy	accuracy	NOUN
aiti-14589	148	28	and	and	CCONJ
aiti-14589	148	29	reduce	reduce	VERB
aiti-14589	148	30	overfitting	overfitte	VERB
aiti-14589	148	31	.	.	PUNCT
aiti-14589	149	1	for	for	ADP
aiti-14589	149	2	the	the	DET
aiti-14589	149	3	image	image	NOUN
aiti-14589	149	4	-	-	PUNCT
aiti-14589	149	5	based	base	VERB
aiti-14589	149	6	approach	approach	NOUN
aiti-14589	149	7	,	,	PUNCT
aiti-14589	149	8	hog	hog	PROPN
aiti-14589	149	9	was	be	AUX
aiti-14589	149	10	used	use	VERB
aiti-14589	149	11	to	to	PART
aiti-14589	149	12	extract	extract	VERB
aiti-14589	149	13	texture	texture	NOUN
aiti-14589	149	14	features	feature	NOUN
aiti-14589	149	15	before	before	ADP
aiti-14589	149	16	classification	classification	NOUN
aiti-14589	149	17	.	.	PUNCT
aiti-14589	150	1	in	in	ADP
aiti-14589	150	2	the	the	DET
aiti-14589	150	3	landmark	landmark	NOUN
aiti-14589	150	4	-	-	PUNCT
aiti-14589	150	5	based	base	VERB
aiti-14589	150	6	approach	approach	NOUN
aiti-14589	150	7	,	,	PUNCT
aiti-14589	150	8	the	the	DET
aiti-14589	150	9	extracted	extract	VERB
aiti-14589	150	10	facial	facial	ADJ
aiti-14589	150	11	landmarks	landmark	NOUN
aiti-14589	150	12	were	be	AUX
aiti-14589	150	13	used	use	VERB
aiti-14589	150	14	directly	directly	ADV
aiti-14589	150	15	.	.	PUNCT
aiti-14589	151	1	standard	standard	ADJ
aiti-14589	151	2	scaling	scaling	NOUN
aiti-14589	151	3	was	be	AUX
aiti-14589	151	4	applied	apply	VERB
aiti-14589	151	5	to	to	ADP
aiti-14589	151	6	the	the	DET
aiti-14589	151	7	extracted	extract	VERB
aiti-14589	151	8	features	feature	NOUN
aiti-14589	151	9	before	before	ADP
aiti-14589	151	10	training	training	NOUN
aiti-14589	151	11	.	.	PUNCT
aiti-14589	152	1	the	the	DET
aiti-14589	152	2	model	model	NOUN
aiti-14589	152	3	was	be	AUX
aiti-14589	152	4	trained	train	VERB
aiti-14589	152	5	using	use	VERB
aiti-14589	152	6	default	default	NOUN
aiti-14589	152	7	rfc	rfc	NOUN
aiti-14589	152	8	parameters	parameter	NOUN
aiti-14589	152	9	,	,	PUNCT
aiti-14589	152	10	with	with	ADP
aiti-14589	152	11	the	the	DET
aiti-14589	152	12	number	number	NOUN
aiti-14589	152	13	of	of	ADP
aiti-14589	152	14	trees	tree	NOUN
aiti-14589	152	15	set	set	VERB
aiti-14589	152	16	to	to	ADP
aiti-14589	152	17	100	100	NUM
aiti-14589	152	18	.	.	PUNCT
aiti-14589	153	1	(	(	PUNCT
aiti-14589	153	2	3	3	X
aiti-14589	153	3	)	)	PUNCT
aiti-14589	153	4	logistic	logistic	ADJ
aiti-14589	153	5	regression	regression	NOUN
aiti-14589	153	6	classification	classification	NOUN
aiti-14589	153	7	(	(	PUNCT
aiti-14589	153	8	lrc	lrc	PROPN
aiti-14589	153	9	)	)	PUNCT
aiti-14589	153	10	lrc	lrc	PROPN
aiti-14589	153	11	is	be	AUX
aiti-14589	153	12	a	a	DET
aiti-14589	153	13	simple	simple	ADJ
aiti-14589	153	14	yet	yet	CCONJ
aiti-14589	153	15	effective	effective	ADJ
aiti-14589	153	16	classifier	classifier	NOUN
aiti-14589	153	17	for	for	ADP
aiti-14589	153	18	binary	binary	ADJ
aiti-14589	153	19	classification	classification	NOUN
aiti-14589	153	20	problems	problem	NOUN
aiti-14589	153	21	.	.	PUNCT
aiti-14589	154	1	it	it	PRON
aiti-14589	154	2	models	model	VERB
aiti-14589	154	3	the	the	DET
aiti-14589	154	4	probability	probability	NOUN
aiti-14589	154	5	of	of	ADP
aiti-14589	154	6	an	an	DET
aiti-14589	154	7	instance	instance	NOUN
aiti-14589	154	8	belonging	belong	VERB
aiti-14589	154	9	to	to	ADP
aiti-14589	154	10	a	a	DET
aiti-14589	154	11	specific	specific	ADJ
aiti-14589	154	12	class	class	NOUN
aiti-14589	154	13	using	use	VERB
aiti-14589	154	14	the	the	DET
aiti-14589	154	15	logistic	logistic	ADJ
aiti-14589	154	16	sigmoid	sigmoid	NOUN
aiti-14589	154	17	function	function	NOUN
aiti-14589	154	18	.	.	PUNCT
aiti-14589	155	1	for	for	ADP
aiti-14589	155	2	the	the	DET
aiti-14589	155	3	image	image	NOUN
aiti-14589	155	4	-	-	PUNCT
aiti-14589	155	5	based	base	VERB
aiti-14589	155	6	approach	approach	NOUN
aiti-14589	155	7	,	,	PUNCT
aiti-14589	155	8	hog	hog	PROPN
aiti-14589	155	9	was	be	AUX
aiti-14589	155	10	used	use	VERB
aiti-14589	155	11	for	for	ADP
aiti-14589	155	12	feature	feature	NOUN
aiti-14589	155	13	extraction	extraction	NOUN
aiti-14589	155	14	,	,	PUNCT
aiti-14589	155	15	while	while	SCONJ
aiti-14589	155	16	for	for	ADP
aiti-14589	155	17	the	the	DET
aiti-14589	155	18	landmark	landmark	NOUN
aiti-14589	155	19	-	-	PUNCT
aiti-14589	155	20	based	base	VERB
aiti-14589	155	21	approach	approach	NOUN
aiti-14589	155	22	,	,	PUNCT
aiti-14589	155	23	facial	facial	ADJ
aiti-14589	155	24	landmarks	landmark	NOUN
aiti-14589	155	25	were	be	AUX
aiti-14589	155	26	directly	directly	ADV
aiti-14589	155	27	utilized	utilize	VERB
aiti-14589	155	28	.	.	PUNCT
aiti-14589	156	1	standardization	standardization	NOUN
aiti-14589	156	2	was	be	AUX
aiti-14589	156	3	applied	apply	VERB
aiti-14589	156	4	using	use	VERB
aiti-14589	156	5	the	the	DET
aiti-14589	156	6	standardscaler	standardscaler	NOUN
aiti-14589	156	7	.	.	PUNCT
aiti-14589	157	1	the	the	DET
aiti-14589	157	2	model	model	NOUN
aiti-14589	157	3	was	be	AUX
aiti-14589	157	4	trained	train	VERB
aiti-14589	157	5	with	with	ADP
aiti-14589	157	6	default	default	NOUN
aiti-14589	157	7	parameters	parameter	NOUN
aiti-14589	157	8	,	,	PUNCT
aiti-14589	157	9	including	include	VERB
aiti-14589	157	10	l2	l2	NOUN
aiti-14589	157	11	regularization	regularization	NOUN
aiti-14589	157	12	to	to	PART
aiti-14589	157	13	prevent	prevent	VERB
aiti-14589	157	14	overfitting	overfitting	NOUN
aiti-14589	157	15	.	.	PUNCT
aiti-14589	158	1	advances	advance	NOUN
aiti-14589	158	2	in	in	ADP
aiti-14589	158	3	technology	technology	NOUN
aiti-14589	158	4	innovation	innovation	NOUN
aiti-14589	158	5	,	,	PUNCT
aiti-14589	158	6	vol	vol	NOUN
aiti-14589	158	7	.	.	PROPN
aiti-14589	159	1	10	10	NUM
aiti-14589	159	2	,	,	PUNCT
aiti-14589	159	3	no	no	INTJ
aiti-14589	159	4	.	.	NOUN
aiti-14589	159	5	4	4	NUM
aiti-14589	159	6	,	,	PUNCT
aiti-14589	159	7	2025	2025	NUM
aiti-14589	159	8	,	,	PUNCT
aiti-14589	159	9	pp	pp	ADJ
aiti-14589	159	10	.	.	PUNCT
aiti-14589	160	1	395	395	NUM
aiti-14589	160	2	-	-	SYM
aiti-14589	160	3	406	406	NUM
aiti-14589	160	4	400	400	NUM
aiti-14589	160	5	(	(	PUNCT
aiti-14589	160	6	4	4	X
aiti-14589	160	7	)	)	PUNCT
aiti-14589	160	8	support	support	NOUN
aiti-14589	160	9	vector	vector	NOUN
aiti-14589	160	10	machine	machine	NOUN
aiti-14589	160	11	(	(	PUNCT
aiti-14589	160	12	svm	svm	ADJ
aiti-14589	160	13	)	)	PUNCT
aiti-14589	160	14	svm	svm	NOUN
aiti-14589	160	15	aim	aim	VERB
aiti-14589	160	16	to	to	PART
aiti-14589	160	17	find	find	VERB
aiti-14589	160	18	the	the	DET
aiti-14589	160	19	optimal	optimal	ADJ
aiti-14589	160	20	hyperplane	hyperplane	NOUN
aiti-14589	160	21	that	that	PRON
aiti-14589	160	22	maximizes	maximize	VERB
aiti-14589	160	23	the	the	DET
aiti-14589	160	24	margin	margin	NOUN
aiti-14589	160	25	between	between	ADP
aiti-14589	160	26	different	different	ADJ
aiti-14589	160	27	classes	class	NOUN
aiti-14589	160	28	.	.	PUNCT
aiti-14589	161	1	for	for	ADP
aiti-14589	161	2	the	the	DET
aiti-14589	161	3	image	image	NOUN
aiti-14589	161	4	-	-	PUNCT
aiti-14589	161	5	based	base	VERB
aiti-14589	161	6	approach	approach	NOUN
aiti-14589	161	7	,	,	PUNCT
aiti-14589	161	8	hog	hog	PROPN
aiti-14589	161	9	was	be	AUX
aiti-14589	161	10	used	use	VERB
aiti-14589	161	11	for	for	ADP
aiti-14589	161	12	feature	feature	NOUN
aiti-14589	161	13	extraction	extraction	NOUN
aiti-14589	161	14	,	,	PUNCT
aiti-14589	161	15	while	while	SCONJ
aiti-14589	161	16	for	for	ADP
aiti-14589	161	17	the	the	DET
aiti-14589	161	18	landmark	landmark	NOUN
aiti-14589	161	19	-	-	PUNCT
aiti-14589	161	20	based	base	VERB
aiti-14589	161	21	approach	approach	NOUN
aiti-14589	161	22	,	,	PUNCT
aiti-14589	161	23	landmark	landmark	PROPN
aiti-14589	161	24	coordinates	coordinate	NOUN
aiti-14589	161	25	were	be	AUX
aiti-14589	161	26	used	use	VERB
aiti-14589	161	27	.	.	PUNCT
aiti-14589	162	1	feature	feature	NOUN
aiti-14589	162	2	normalization	normalization	NOUN
aiti-14589	162	3	was	be	AUX
aiti-14589	162	4	applied	apply	VERB
aiti-14589	162	5	using	use	VERB
aiti-14589	162	6	the	the	DET
aiti-14589	162	7	standardscaler	standardscaler	NOUN
aiti-14589	162	8	.	.	PUNCT
aiti-14589	163	1	the	the	DET
aiti-14589	163	2	model	model	NOUN
aiti-14589	163	3	was	be	AUX
aiti-14589	163	4	trained	train	VERB
aiti-14589	163	5	using	use	VERB
aiti-14589	163	6	default	default	NOUN
aiti-14589	163	7	parameters	parameter	NOUN
aiti-14589	163	8	with	with	ADP
aiti-14589	163	9	a	a	DET
aiti-14589	163	10	radial	radial	ADJ
aiti-14589	163	11	basis	basis	NOUN
aiti-14589	163	12	function	function	NOUN
aiti-14589	163	13	(	(	PUNCT
aiti-14589	163	14	rbf	rbf	PROPN
aiti-14589	163	15	)	)	PUNCT
aiti-14589	163	16	kernel	kernel	PROPN
aiti-14589	163	17	to	to	PART
aiti-14589	163	18	handle	handle	VERB
aiti-14589	163	19	non	non	ADJ
aiti-14589	163	20	-	-	ADJ
aiti-14589	163	21	linear	linear	ADJ
aiti-14589	163	22	decision	decision	NOUN
aiti-14589	163	23	boundaries	boundary	NOUN
aiti-14589	163	24	effectively	effectively	ADV
aiti-14589	163	25	.	.	PUNCT
aiti-14589	164	1	(	(	PUNCT
aiti-14589	164	2	5	5	X
aiti-14589	164	3	)	)	PUNCT
aiti-14589	164	4	gradient	gradient	NOUN
aiti-14589	164	5	boosting	boost	VERB
aiti-14589	164	6	classifier	classifier	NOUN
aiti-14589	164	7	(	(	PUNCT
aiti-14589	164	8	gbc	gbc	PROPN
aiti-14589	164	9	)	)	PUNCT
aiti-14589	164	10	gbc	gbc	PROPN
aiti-14589	164	11	is	be	AUX
aiti-14589	164	12	an	an	DET
aiti-14589	164	13	ensemble	ensemble	ADJ
aiti-14589	164	14	method	method	NOUN
aiti-14589	164	15	that	that	PRON
aiti-14589	164	16	sequentially	sequentially	ADV
aiti-14589	164	17	builds	build	VERB
aiti-14589	164	18	decision	decision	NOUN
aiti-14589	164	19	trees	tree	NOUN
aiti-14589	164	20	to	to	PART
aiti-14589	164	21	correct	correct	VERB
aiti-14589	164	22	the	the	DET
aiti-14589	164	23	errors	error	NOUN
aiti-14589	164	24	of	of	ADP
aiti-14589	164	25	previous	previous	ADJ
aiti-14589	164	26	iterations	iteration	NOUN
aiti-14589	164	27	.	.	PUNCT
aiti-14589	165	1	for	for	ADP
aiti-14589	165	2	the	the	DET
aiti-14589	165	3	image	image	NOUN
aiti-14589	165	4	-	-	PUNCT
aiti-14589	165	5	based	base	VERB
aiti-14589	165	6	approach	approach	NOUN
aiti-14589	165	7	,	,	PUNCT
aiti-14589	165	8	hog	hog	PROPN
aiti-14589	165	9	was	be	AUX
aiti-14589	165	10	used	use	VERB
aiti-14589	165	11	for	for	ADP
aiti-14589	165	12	feature	feature	NOUN
aiti-14589	165	13	extraction	extraction	NOUN
aiti-14589	165	14	,	,	PUNCT
aiti-14589	165	15	whereas	whereas	SCONJ
aiti-14589	165	16	the	the	DET
aiti-14589	165	17	landmark	landmark	NOUN
aiti-14589	165	18	-	-	PUNCT
aiti-14589	165	19	based	base	VERB
aiti-14589	165	20	approach	approach	NOUN
aiti-14589	165	21	used	use	VERB
aiti-14589	165	22	facial	facial	ADJ
aiti-14589	165	23	landmark	landmark	NOUN
aiti-14589	165	24	coordinates	coordinate	NOUN
aiti-14589	165	25	.	.	PUNCT
aiti-14589	166	1	standardization	standardization	NOUN
aiti-14589	166	2	was	be	AUX
aiti-14589	166	3	applied	apply	VERB
aiti-14589	166	4	to	to	ADP
aiti-14589	166	5	the	the	DET
aiti-14589	166	6	extracted	extract	VERB
aiti-14589	166	7	features	feature	NOUN
aiti-14589	166	8	.	.	PUNCT
aiti-14589	167	1	the	the	DET
aiti-14589	167	2	model	model	NOUN
aiti-14589	167	3	was	be	AUX
aiti-14589	167	4	trained	train	VERB
aiti-14589	167	5	using	use	VERB
aiti-14589	167	6	default	default	NOUN
aiti-14589	167	7	parameters	parameter	NOUN
aiti-14589	167	8	,	,	PUNCT
aiti-14589	167	9	including	include	VERB
aiti-14589	167	10	a	a	DET
aiti-14589	167	11	learning	learn	VERB
aiti-14589	167	12	rate	rate	NOUN
aiti-14589	167	13	of	of	ADP
aiti-14589	167	14	0.1	0.1	NUM
aiti-14589	167	15	and	and	CCONJ
aiti-14589	167	16	100	100	NUM
aiti-14589	167	17	estimators	estimator	NOUN
aiti-14589	167	18	.	.	PUNCT
aiti-14589	168	1	table	table	NOUN
aiti-14589	168	2	2	2	NUM
aiti-14589	168	3	summary	summary	NOUN
aiti-14589	168	4	of	of	ADP
aiti-14589	168	5	machine	machine	NOUN
aiti-14589	168	6	learning	learn	VERB
aiti-14589	168	7	methods	method	NOUN
aiti-14589	168	8	and	and	CCONJ
aiti-14589	168	9	preprocessing	preprocesse	VERB
aiti-14589	168	10	steps	step	NOUN
aiti-14589	168	11	model	model	NOUN
aiti-14589	168	12	feature	feature	NOUN
aiti-14589	168	13	extraction	extraction	NOUN
aiti-14589	168	14	preprocessing	preprocesse	VERB
aiti-14589	168	15	hyperparameters	hyperparameter	NOUN
aiti-14589	168	16	cnn	cnn	PROPN
aiti-14589	168	17	(	(	PUNCT
aiti-14589	168	18	image	image	NOUN
aiti-14589	168	19	-	-	PUNCT
aiti-14589	168	20	based	base	VERB
aiti-14589	168	21	)	)	PUNCT
aiti-14589	168	22	raw	raw	ADJ
aiti-14589	168	23	image	image	NOUN
aiti-14589	168	24	data	datum	NOUN
aiti-14589	168	25	resizing	resizing	NOUN
aiti-14589	168	26	,	,	PUNCT
aiti-14589	168	27	normalization	normalization	NOUN
aiti-14589	168	28	custom	custom	NOUN
aiti-14589	168	29	architecture	architecture	NOUN
aiti-14589	168	30	(	(	PUNCT
aiti-14589	168	31	conv2d	conv2d	VERB
aiti-14589	168	32	layers	layer	NOUN
aiti-14589	168	33	,	,	PUNCT
aiti-14589	168	34	adam	adam	PROPN
aiti-14589	168	35	optimizer	optimizer	PROPN
aiti-14589	168	36	)	)	PUNCT
aiti-14589	168	37	cnn	cnn	PROPN
aiti-14589	168	38	(	(	PUNCT
aiti-14589	168	39	landmark	landmark	NOUN
aiti-14589	168	40	-	-	PUNCT
aiti-14589	168	41	based	base	VERB
aiti-14589	168	42	)	)	PUNCT
aiti-14589	168	43	468	468	NUM
aiti-14589	168	44	facial	facial	ADJ
aiti-14589	168	45	landmarks	landmark	NOUN
aiti-14589	168	46	standardization	standardization	NOUN
aiti-14589	168	47	custom	custom	NOUN
aiti-14589	168	48	architecture	architecture	NOUN
aiti-14589	168	49	(	(	PUNCT
aiti-14589	168	50	conv1d	conv1d	ADJ
aiti-14589	168	51	layers	layer	NOUN
aiti-14589	168	52	,	,	PUNCT
aiti-14589	168	53	adam	adam	PROPN
aiti-14589	168	54	optimizer	optimizer	NOUN
aiti-14589	168	55	)	)	PUNCT
aiti-14589	168	56	rfc	rfc	PROPN
aiti-14589	168	57	hog	hog	NOUN
aiti-14589	168	58	(	(	PUNCT
aiti-14589	168	59	image	image	NOUN
aiti-14589	168	60	)	)	PUNCT
aiti-14589	168	61	;	;	PUNCT
aiti-14589	168	62	landmarks	landmark	NOUN
aiti-14589	168	63	(	(	PUNCT
aiti-14589	168	64	landmark	landmark	NOUN
aiti-14589	168	65	-	-	PUNCT
aiti-14589	168	66	based	base	VERB
aiti-14589	168	67	)	)	PUNCT
aiti-14589	168	68	standardization	standardization	NOUN
aiti-14589	168	69	default	default	NOUN
aiti-14589	168	70	(	(	PUNCT
aiti-14589	168	71	100	100	NUM
aiti-14589	168	72	trees	tree	NOUN
aiti-14589	168	73	)	)	PUNCT
aiti-14589	168	74	lrc	lrc	PROPN
aiti-14589	168	75	hog	hog	PROPN
aiti-14589	168	76	(	(	PUNCT
aiti-14589	168	77	image	image	NOUN
aiti-14589	168	78	)	)	PUNCT
aiti-14589	168	79	;	;	PUNCT
aiti-14589	168	80	landmarks	landmark	NOUN
aiti-14589	168	81	(	(	PUNCT
aiti-14589	168	82	landmark	landmark	NOUN
aiti-14589	168	83	-	-	PUNCT
aiti-14589	168	84	based	base	VERB
aiti-14589	168	85	)	)	PUNCT
aiti-14589	168	86	standardization	standardization	NOUN
aiti-14589	168	87	default	default	NOUN
aiti-14589	168	88	(	(	PUNCT
aiti-14589	168	89	l2	l2	NOUN
aiti-14589	168	90	regularization	regularization	NOUN
aiti-14589	168	91	)	)	PUNCT
aiti-14589	168	92	svm	svm	PROPN
aiti-14589	168	93	hog	hog	NOUN
aiti-14589	168	94	(	(	PUNCT
aiti-14589	168	95	image	image	NOUN
aiti-14589	168	96	)	)	PUNCT
aiti-14589	168	97	;	;	PUNCT
aiti-14589	168	98	landmarks	landmark	NOUN
aiti-14589	168	99	(	(	PUNCT
aiti-14589	168	100	landmark	landmark	NOUN
aiti-14589	168	101	-	-	PUNCT
aiti-14589	168	102	based	base	VERB
aiti-14589	168	103	)	)	PUNCT
aiti-14589	168	104	standardization	standardization	NOUN
aiti-14589	168	105	default	default	NOUN
aiti-14589	168	106	(	(	PUNCT
aiti-14589	168	107	rbf	rbf	PROPN
aiti-14589	168	108	kernel	kernel	PROPN
aiti-14589	168	109	)	)	PUNCT
aiti-14589	168	110	gbc	gbc	PROPN
aiti-14589	168	111	hog	hog	NOUN
aiti-14589	168	112	(	(	PUNCT
aiti-14589	168	113	image	image	NOUN
aiti-14589	168	114	)	)	PUNCT
aiti-14589	168	115	;	;	PUNCT
aiti-14589	168	116	landmarks	landmark	NOUN
aiti-14589	168	117	(	(	PUNCT
aiti-14589	168	118	landmark	landmark	NOUN
aiti-14589	168	119	-	-	PUNCT
aiti-14589	168	120	based	base	VERB
aiti-14589	168	121	)	)	PUNCT
aiti-14589	168	122	standardization	standardization	NOUN
aiti-14589	168	123	default	default	NOUN
aiti-14589	168	124	(	(	PUNCT
aiti-14589	168	125	learning	learn	VERB
aiti-14589	168	126	rate	rate	NOUN
aiti-14589	168	127	=	=	NOUN
aiti-14589	168	128	0.1	0.1	NUM
aiti-14589	168	129	)	)	PUNCT
aiti-14589	168	130	4	4	NUM
aiti-14589	168	131	.	.	PUNCT
aiti-14589	168	132	experiment	experiment	NOUN
aiti-14589	168	133	fig	fig	NOUN
aiti-14589	168	134	.	.	PUNCT
aiti-14589	169	1	2	2	NUM
aiti-14589	169	2	experiment	experiment	NOUN
aiti-14589	169	3	pipeline	pipeline	NOUN
aiti-14589	169	4	this	this	DET
aiti-14589	169	5	section	section	NOUN
aiti-14589	169	6	outlines	outline	VERB
aiti-14589	169	7	the	the	DET
aiti-14589	169	8	experimental	experimental	ADJ
aiti-14589	169	9	framework	framework	NOUN
aiti-14589	169	10	used	use	VERB
aiti-14589	169	11	to	to	PART
aiti-14589	169	12	evaluate	evaluate	VERB
aiti-14589	169	13	the	the	DET
aiti-14589	169	14	image	image	NOUN
aiti-14589	169	15	-	-	PUNCT
aiti-14589	169	16	based	base	VERB
aiti-14589	169	17	and	and	CCONJ
aiti-14589	169	18	landmark	landmark	NOUN
aiti-14589	169	19	-	-	PUNCT
aiti-14589	169	20	based	base	VERB
aiti-14589	169	21	fer	fer	PROPN
aiti-14589	169	22	methods	method	NOUN
aiti-14589	169	23	.	.	PUNCT
aiti-14589	170	1	the	the	DET
aiti-14589	170	2	experimental	experimental	ADJ
aiti-14589	170	3	pipeline	pipeline	NOUN
aiti-14589	170	4	,	,	PUNCT
aiti-14589	170	5	depicted	depict	VERB
aiti-14589	170	6	in	in	ADP
aiti-14589	170	7	fig	fig	NOUN
aiti-14589	170	8	.	.	PUNCT
aiti-14589	171	1	2	2	NUM
aiti-14589	171	2	,	,	PUNCT
aiti-14589	171	3	demonstrates	demonstrate	VERB
aiti-14589	171	4	the	the	DET
aiti-14589	171	5	step	step	NOUN
aiti-14589	171	6	-	-	PUNCT
aiti-14589	171	7	by	by	ADP
aiti-14589	171	8	-	-	PUNCT
aiti-14589	171	9	step	step	NOUN
aiti-14589	171	10	process	process	NOUN
aiti-14589	171	11	,	,	PUNCT
aiti-14589	171	12	from	from	ADP
aiti-14589	171	13	preprocessing	preprocesse	VERB
aiti-14589	171	14	and	and	CCONJ
aiti-14589	171	15	feature	feature	NOUN
aiti-14589	171	16	extraction	extraction	NOUN
aiti-14589	171	17	to	to	ADP
aiti-14589	171	18	classification	classification	NOUN
aiti-14589	171	19	and	and	CCONJ
aiti-14589	171	20	performance	performance	NOUN
aiti-14589	171	21	evaluation	evaluation	NOUN
aiti-14589	171	22	.	.	PUNCT
aiti-14589	172	1	this	this	DET
aiti-14589	172	2	experiment	experiment	NOUN
aiti-14589	172	3	aims	aim	VERB
aiti-14589	172	4	to	to	PART
aiti-14589	172	5	compare	compare	VERB
aiti-14589	172	6	the	the	DET
aiti-14589	172	7	effectiveness	effectiveness	NOUN
aiti-14589	172	8	of	of	ADP
aiti-14589	172	9	five	five	NUM
aiti-14589	172	10	classification	classification	NOUN
aiti-14589	172	11	models	model	NOUN
aiti-14589	172	12	—	—	PUNCT
aiti-14589	172	13	advances	advance	NOUN
aiti-14589	172	14	in	in	ADP
aiti-14589	172	15	technology	technology	NOUN
aiti-14589	172	16	innovation	innovation	NOUN
aiti-14589	172	17	,	,	PUNCT
aiti-14589	172	18	vol	vol	NOUN
aiti-14589	172	19	.	.	PROPN
aiti-14589	173	1	10	10	NUM
aiti-14589	173	2	,	,	PUNCT
aiti-14589	173	3	no	no	INTJ
aiti-14589	173	4	.	.	NOUN
aiti-14589	173	5	4	4	NUM
aiti-14589	173	6	,	,	PUNCT
aiti-14589	173	7	2025	2025	NUM
aiti-14589	173	8	,	,	PUNCT
aiti-14589	173	9	pp	pp	ADJ
aiti-14589	173	10	.	.	PUNCT
aiti-14589	174	1	395	395	NUM
aiti-14589	174	2	-	-	SYM
aiti-14589	174	3	406	406	NUM
aiti-14589	174	4	401	401	NUM
aiti-14589	174	5	cnn	cnn	PROPN
aiti-14589	174	6	,	,	PUNCT
aiti-14589	174	7	lrc	lrc	PROPN
aiti-14589	174	8	,	,	PUNCT
aiti-14589	174	9	svm	svm	PROPN
aiti-14589	174	10	,	,	PUNCT
aiti-14589	174	11	rfc	rfc	PROPN
aiti-14589	174	12	,	,	PUNCT
aiti-14589	174	13	and	and	CCONJ
aiti-14589	174	14	gbc	gbc	NOUN
aiti-14589	174	15	—	—	PUNCT
aiti-14589	174	16	on	on	ADP
aiti-14589	174	17	two	two	NUM
aiti-14589	174	18	distinct	distinct	ADJ
aiti-14589	174	19	input	input	NOUN
aiti-14589	174	20	types	type	NOUN
aiti-14589	174	21	:	:	PUNCT
aiti-14589	174	22	image	image	NOUN
aiti-14589	174	23	-	-	PUNCT
aiti-14589	174	24	based	base	VERB
aiti-14589	174	25	features	feature	NOUN
aiti-14589	174	26	using	use	VERB
aiti-14589	174	27	hog	hog	NOUN
aiti-14589	174	28	and	and	CCONJ
aiti-14589	174	29	landmark	landmark	NOUN
aiti-14589	174	30	-	-	PUNCT
aiti-14589	174	31	based	base	VERB
aiti-14589	174	32	features	feature	NOUN
aiti-14589	174	33	extracted	extract	VERB
aiti-14589	174	34	from	from	ADP
aiti-14589	174	35	facial	facial	ADJ
aiti-14589	174	36	landmarks	landmark	NOUN
aiti-14589	174	37	.	.	PUNCT
aiti-14589	175	1	both	both	DET
aiti-14589	175	2	approaches	approach	NOUN
aiti-14589	175	3	are	be	AUX
aiti-14589	175	4	assessed	assess	VERB
aiti-14589	175	5	through	through	ADP
aiti-14589	175	6	a	a	DET
aiti-14589	175	7	comprehensive	comprehensive	ADJ
aiti-14589	175	8	set	set	NOUN
aiti-14589	175	9	of	of	ADP
aiti-14589	175	10	performance	performance	NOUN
aiti-14589	175	11	metrics	metric	NOUN
aiti-14589	175	12	,	,	PUNCT
aiti-14589	175	13	including	include	VERB
aiti-14589	175	14	precision	precision	NOUN
aiti-14589	175	15	,	,	PUNCT
aiti-14589	175	16	recall	recall	NOUN
aiti-14589	175	17	,	,	PUNCT
aiti-14589	175	18	f1	f1	NOUN
aiti-14589	175	19	-	-	PUNCT
aiti-14589	175	20	score	score	NOUN
aiti-14589	175	21	,	,	PUNCT
aiti-14589	175	22	and	and	CCONJ
aiti-14589	175	23	training	training	NOUN
aiti-14589	175	24	time	time	NOUN
aiti-14589	175	25	,	,	PUNCT
aiti-14589	175	26	to	to	PART
aiti-14589	175	27	determine	determine	VERB
aiti-14589	175	28	the	the	DET
aiti-14589	175	29	most	most	ADV
aiti-14589	175	30	suitable	suitable	ADJ
aiti-14589	175	31	method	method	NOUN
aiti-14589	175	32	for	for	ADP
aiti-14589	175	33	reliable	reliable	ADJ
aiti-14589	175	34	fer	fer	PROPN
aiti-14589	175	35	.	.	PUNCT
aiti-14589	176	1	4.1	4.1	NUM
aiti-14589	176	2	.	.	PUNCT
aiti-14589	176	3	dataset	dataset	VERB
aiti-14589	176	4	this	this	DET
aiti-14589	176	5	study	study	NOUN
aiti-14589	176	6	utilized	utilize	VERB
aiti-14589	176	7	publicly	publicly	ADV
aiti-14589	176	8	available	available	ADJ
aiti-14589	176	9	human	human	ADJ
aiti-14589	176	10	image	image	NOUN
aiti-14589	176	11	datasets	dataset	NOUN
aiti-14589	176	12	,	,	PUNCT
aiti-14589	176	13	namely	namely	ADV
aiti-14589	176	14	the	the	DET
aiti-14589	176	15	delaware	delaware	PROPN
aiti-14589	176	16	pain	pain	PROPN
aiti-14589	176	17	database	database	NOUN
aiti-14589	176	18	and	and	CCONJ
aiti-14589	176	19	utkface	utkface	NOUN
aiti-14589	176	20	,	,	PUNCT
aiti-14589	176	21	which	which	PRON
aiti-14589	176	22	contain	contain	VERB
aiti-14589	176	23	images	image	NOUN
aiti-14589	176	24	depicting	depict	VERB
aiti-14589	176	25	various	various	ADJ
aiti-14589	176	26	facial	facial	ADJ
aiti-14589	176	27	expressions	expression	NOUN
aiti-14589	176	28	.	.	PUNCT
aiti-14589	177	1	the	the	DET
aiti-14589	177	2	delaware	delaware	PROPN
aiti-14589	177	3	pain	pain	PROPN
aiti-14589	177	4	database	database	NOUN
aiti-14589	177	5	consists	consist	VERB
aiti-14589	177	6	of	of	ADP
aiti-14589	177	7	high	high	ADJ
aiti-14589	177	8	-	-	PUNCT
aiti-14589	177	9	resolution	resolution	NOUN
aiti-14589	177	10	images	image	NOUN
aiti-14589	177	11	(	(	PUNCT
aiti-14589	177	12	1200×1200	1200×1200	NUM
aiti-14589	177	13	pixels	pixel	NOUN
aiti-14589	177	14	)	)	PUNCT
aiti-14589	177	15	,	,	PUNCT
aiti-14589	177	16	while	while	SCONJ
aiti-14589	177	17	utkface	utkface	NOUN
aiti-14589	177	18	provides	provide	VERB
aiti-14589	177	19	lower	low	ADJ
aiti-14589	177	20	-	-	PUNCT
aiti-14589	177	21	resolution	resolution	NOUN
aiti-14589	177	22	images	image	NOUN
aiti-14589	177	23	(	(	PUNCT
aiti-14589	177	24	200×200	200×200	NUM
aiti-14589	177	25	pixels	pixel	NOUN
aiti-14589	177	26	)	)	PUNCT
aiti-14589	177	27	.	.	PUNCT
aiti-14589	178	1	to	to	PART
aiti-14589	178	2	ensure	ensure	VERB
aiti-14589	178	3	consistency	consistency	NOUN
aiti-14589	178	4	in	in	ADP
aiti-14589	178	5	model	model	NOUN
aiti-14589	178	6	training	training	NOUN
aiti-14589	178	7	,	,	PUNCT
aiti-14589	178	8	all	all	DET
aiti-14589	178	9	images	image	NOUN
aiti-14589	178	10	were	be	AUX
aiti-14589	178	11	resized	resize	VERB
aiti-14589	178	12	to	to	ADP
aiti-14589	178	13	128×128	128×128	NUM
aiti-14589	178	14	pixels	pixel	NOUN
aiti-14589	178	15	.	.	PUNCT
aiti-14589	179	1	for	for	ADP
aiti-14589	179	2	this	this	DET
aiti-14589	179	3	study	study	NOUN
aiti-14589	179	4	,	,	PUNCT
aiti-14589	179	5	only	only	ADV
aiti-14589	179	6	images	image	NOUN
aiti-14589	179	7	displaying	display	VERB
aiti-14589	179	8	normal	normal	ADJ
aiti-14589	179	9	and	and	CCONJ
aiti-14589	179	10	hurt	hurt	ADJ
aiti-14589	179	11	expressions	expression	NOUN
aiti-14589	179	12	were	be	AUX
aiti-14589	179	13	selected	select	VERB
aiti-14589	179	14	,	,	PUNCT
aiti-14589	179	15	resulting	result	VERB
aiti-14589	179	16	in	in	ADP
aiti-14589	179	17	1,200	1,200	NUM
aiti-14589	179	18	images	image	NOUN
aiti-14589	179	19	.	.	PUNCT
aiti-14589	180	1	the	the	DET
aiti-14589	180	2	dataset	dataset	NOUN
aiti-14589	180	3	was	be	AUX
aiti-14589	180	4	divided	divide	VERB
aiti-14589	180	5	into	into	ADP
aiti-14589	180	6	800	800	NUM
aiti-14589	180	7	images	image	NOUN
aiti-14589	180	8	for	for	ADP
aiti-14589	180	9	training	training	NOUN
aiti-14589	180	10	,	,	PUNCT
aiti-14589	180	11	200	200	NUM
aiti-14589	180	12	for	for	ADP
aiti-14589	180	13	validation	validation	NOUN
aiti-14589	180	14	,	,	PUNCT
aiti-14589	180	15	and	and	CCONJ
aiti-14589	180	16	200	200	NUM
aiti-14589	180	17	for	for	ADP
aiti-14589	180	18	testing	testing	NOUN
aiti-14589	180	19	.	.	PUNCT
aiti-14589	181	1	example	example	NOUN
aiti-14589	181	2	images	image	NOUN
aiti-14589	181	3	from	from	ADP
aiti-14589	181	4	each	each	DET
aiti-14589	181	5	category	category	NOUN
aiti-14589	181	6	are	be	AUX
aiti-14589	181	7	shown	show	VERB
aiti-14589	181	8	in	in	ADP
aiti-14589	181	9	fig	fig	NOUN
aiti-14589	181	10	.	.	PUNCT
aiti-14589	182	1	3	3	NUM
aiti-14589	182	2	and	and	CCONJ
aiti-14589	182	3	fig	fig	NOUN
aiti-14589	182	4	.	.	PUNCT
aiti-14589	183	1	4	4	X
aiti-14589	183	2	.	.	X
aiti-14589	183	3	fig	fig	NOUN
aiti-14589	183	4	.	.	PUNCT
aiti-14589	184	1	3	3	NUM
aiti-14589	184	2	normal	normal	ADJ
aiti-14589	184	3	face	face	NOUN
aiti-14589	184	4	example	example	NOUN
aiti-14589	184	5	(	(	PUNCT
aiti-14589	184	6	the	the	DET
aiti-14589	184	7	delaware	delaware	PROPN
aiti-14589	184	8	pain	pain	PROPN
aiti-14589	184	9	database	database	NOUN
aiti-14589	184	10	and	and	CCONJ
aiti-14589	184	11	utkface	utkface	NOUN
aiti-14589	184	12	)	)	PUNCT
aiti-14589	184	13	fig	fig	NOUN
aiti-14589	184	14	.	.	PUNCT
aiti-14589	185	1	4	4	NUM
aiti-14589	185	2	hurt	hurt	ADJ
aiti-14589	185	3	face	face	NOUN
aiti-14589	185	4	example	example	NOUN
aiti-14589	185	5	(	(	PUNCT
aiti-14589	185	6	the	the	DET
aiti-14589	185	7	delaware	delaware	PROPN
aiti-14589	185	8	pain	pain	PROPN
aiti-14589	185	9	database	database	NOUN
aiti-14589	185	10	and	and	CCONJ
aiti-14589	185	11	utkface	utkface	NOUN
aiti-14589	185	12	)	)	PUNCT
aiti-14589	185	13	4.2	4.2	NUM
aiti-14589	185	14	.	.	PUNCT
aiti-14589	185	15	image	image	NOUN
aiti-14589	185	16	-	-	PUNCT
aiti-14589	185	17	based	base	VERB
aiti-14589	185	18	method	method	NOUN
aiti-14589	185	19	the	the	DET
aiti-14589	185	20	image	image	NOUN
aiti-14589	185	21	-	-	PUNCT
aiti-14589	185	22	based	base	VERB
aiti-14589	185	23	method	method	NOUN
aiti-14589	185	24	involved	involve	VERB
aiti-14589	185	25	five	five	NUM
aiti-14589	185	26	classification	classification	NOUN
aiti-14589	185	27	models	model	NOUN
aiti-14589	185	28	,	,	PUNCT
aiti-14589	185	29	like	like	ADP
aiti-14589	185	30	cnn	cnn	PROPN
aiti-14589	185	31	,	,	PUNCT
aiti-14589	185	32	lrc	lrc	PROPN
aiti-14589	185	33	,	,	PUNCT
aiti-14589	185	34	svm	svm	PROPN
aiti-14589	185	35	,	,	PUNCT
aiti-14589	185	36	rfc	rfc	PROPN
aiti-14589	185	37	,	,	PUNCT
aiti-14589	185	38	and	and	CCONJ
aiti-14589	185	39	gbc	gbc	PROPN
aiti-14589	185	40	.	.	PUNCT
aiti-14589	186	1	since	since	SCONJ
aiti-14589	186	2	lrc	lrc	PROPN
aiti-14589	186	3	,	,	PUNCT
aiti-14589	186	4	svm	svm	PROPN
aiti-14589	186	5	,	,	PUNCT
aiti-14589	186	6	rfc	rfc	PRON
aiti-14589	186	7	,	,	PUNCT
aiti-14589	186	8	and	and	CCONJ
aiti-14589	186	9	gbc	gbc	NOUN
aiti-14589	186	10	do	do	AUX
aiti-14589	186	11	not	not	PART
aiti-14589	186	12	inherently	inherently	ADV
aiti-14589	186	13	process	process	VERB
aiti-14589	186	14	raw	raw	ADJ
aiti-14589	186	15	image	image	NOUN
aiti-14589	186	16	data	datum	NOUN
aiti-14589	186	17	,	,	PUNCT
aiti-14589	186	18	a	a	DET
aiti-14589	186	19	feature	feature	NOUN
aiti-14589	186	20	extraction	extraction	NOUN
aiti-14589	186	21	technique	technique	NOUN
aiti-14589	186	22	was	be	AUX
aiti-14589	186	23	required	require	VERB
aiti-14589	186	24	to	to	PART
aiti-14589	186	25	convert	convert	VERB
aiti-14589	186	26	images	image	NOUN
aiti-14589	186	27	into	into	ADP
aiti-14589	186	28	numerical	numerical	ADJ
aiti-14589	186	29	representations	representation	NOUN
aiti-14589	186	30	.	.	PUNCT
aiti-14589	187	1	for	for	ADP
aiti-14589	187	2	preprocessing	preprocessing	NOUN
aiti-14589	187	3	,	,	PUNCT
aiti-14589	187	4	all	all	DET
aiti-14589	187	5	images	image	NOUN
aiti-14589	187	6	were	be	AUX
aiti-14589	187	7	converted	convert	VERB
aiti-14589	187	8	to	to	ADP
aiti-14589	187	9	grayscale	grayscale	NOUN
aiti-14589	187	10	and	and	CCONJ
aiti-14589	187	11	normalized	normalize	VERB
aiti-14589	187	12	by	by	ADP
aiti-14589	187	13	scaling	scale	VERB
aiti-14589	187	14	pixel	pixel	ADJ
aiti-14589	187	15	values	value	NOUN
aiti-14589	187	16	between	between	ADP
aiti-14589	187	17	0	0	NUM
aiti-14589	187	18	and	and	CCONJ
aiti-14589	187	19	1	1	NUM
aiti-14589	187	20	.	.	NOUN
aiti-14589	187	21	feature	feature	NOUN
aiti-14589	187	22	extraction	extraction	NOUN
aiti-14589	187	23	was	be	AUX
aiti-14589	187	24	performed	perform	VERB
aiti-14589	187	25	using	use	VERB
aiti-14589	187	26	the	the	DET
aiti-14589	187	27	hog	hog	NOUN
aiti-14589	187	28	technique	technique	NOUN
aiti-14589	187	29	,	,	PUNCT
aiti-14589	187	30	which	which	PRON
aiti-14589	187	31	captures	capture	VERB
aiti-14589	187	32	edge	edge	NOUN
aiti-14589	187	33	and	and	CCONJ
aiti-14589	187	34	texture	texture	ADJ
aiti-14589	187	35	information	information	NOUN
aiti-14589	187	36	by	by	ADP
aiti-14589	187	37	analyzing	analyze	VERB
aiti-14589	187	38	the	the	DET
aiti-14589	187	39	distribution	distribution	NOUN
aiti-14589	187	40	of	of	ADP
aiti-14589	187	41	gradient	gradient	ADJ
aiti-14589	187	42	orientations	orientation	NOUN
aiti-14589	187	43	.	.	PUNCT
aiti-14589	188	1	in	in	ADP
aiti-14589	188	2	this	this	DET
aiti-14589	188	3	study	study	NOUN
aiti-14589	188	4	,	,	PUNCT
aiti-14589	188	5	hog	hog	NOUN
aiti-14589	188	6	parameters	parameter	NOUN
aiti-14589	188	7	were	be	AUX
aiti-14589	188	8	set	set	VERB
aiti-14589	188	9	to	to	ADP
aiti-14589	188	10	a	a	DET
aiti-14589	188	11	cell	cell	NOUN
aiti-14589	188	12	size	size	NOUN
aiti-14589	188	13	of	of	ADP
aiti-14589	188	14	(	(	PUNCT
aiti-14589	188	15	8×8	8×8	NUM
aiti-14589	188	16	)	)	PUNCT
aiti-14589	188	17	,	,	PUNCT
aiti-14589	188	18	a	a	DET
aiti-14589	188	19	block	block	NOUN
aiti-14589	188	20	size	size	NOUN
aiti-14589	188	21	of	of	ADP
aiti-14589	188	22	(	(	PUNCT
aiti-14589	188	23	2×2	2×2	NUM
aiti-14589	188	24	cells	cell	NOUN
aiti-14589	188	25	)	)	PUNCT
aiti-14589	188	26	,	,	PUNCT
aiti-14589	188	27	and	and	CCONJ
aiti-14589	188	28	nine	nine	NUM
aiti-14589	188	29	orientation	orientation	NOUN
aiti-14589	188	30	bins	bin	NOUN
aiti-14589	188	31	.	.	PUNCT
aiti-14589	189	1	the	the	DET
aiti-14589	189	2	extracted	extract	VERB
aiti-14589	189	3	feature	feature	NOUN
aiti-14589	189	4	vectors	vector	NOUN
aiti-14589	189	5	were	be	AUX
aiti-14589	189	6	then	then	ADV
aiti-14589	189	7	standardized	standardize	VERB
aiti-14589	189	8	to	to	PART
aiti-14589	189	9	improve	improve	VERB
aiti-14589	189	10	model	model	NOUN
aiti-14589	189	11	performance	performance	NOUN
aiti-14589	189	12	and	and	CCONJ
aiti-14589	189	13	ensure	ensure	VERB
aiti-14589	189	14	compatibility	compatibility	NOUN
aiti-14589	189	15	with	with	ADP
aiti-14589	189	16	classification	classification	NOUN
aiti-14589	189	17	algorithms	algorithm	NOUN
aiti-14589	189	18	.	.	PUNCT
aiti-14589	190	1	once	once	SCONJ
aiti-14589	190	2	features	feature	NOUN
aiti-14589	190	3	were	be	AUX
aiti-14589	190	4	extracted	extract	VERB
aiti-14589	190	5	,	,	PUNCT
aiti-14589	190	6	the	the	DET
aiti-14589	190	7	lrc	lrc	PROPN
aiti-14589	190	8	,	,	PUNCT
aiti-14589	190	9	svm	svm	PROPN
aiti-14589	190	10	,	,	PUNCT
aiti-14589	190	11	rfc	rfc	PRON
aiti-14589	190	12	,	,	PUNCT
aiti-14589	190	13	and	and	CCONJ
aiti-14589	190	14	gbc	gbc	NOUN
aiti-14589	190	15	models	model	NOUN
aiti-14589	190	16	were	be	AUX
aiti-14589	190	17	implemented	implement	VERB
aiti-14589	190	18	using	use	VERB
aiti-14589	190	19	scikit	scikit	NOUN
aiti-14589	190	20	-	-	PUNCT
aiti-14589	190	21	learn	learn	VERB
aiti-14589	190	22	.	.	PUNCT
aiti-14589	191	1	these	these	DET
aiti-14589	191	2	models	model	NOUN
aiti-14589	191	3	were	be	AUX
aiti-14589	191	4	trained	train	VERB
aiti-14589	191	5	with	with	ADP
aiti-14589	191	6	their	their	PRON
aiti-14589	191	7	default	default	NOUN
aiti-14589	191	8	hyperparameters	hyperparameter	NOUN
aiti-14589	191	9	,	,	PUNCT
aiti-14589	191	10	as	as	SCONJ
aiti-14589	191	11	the	the	DET
aiti-14589	191	12	focus	focus	NOUN
aiti-14589	191	13	of	of	ADP
aiti-14589	191	14	this	this	DET
aiti-14589	191	15	study	study	NOUN
aiti-14589	191	16	was	be	AUX
aiti-14589	191	17	not	not	PART
aiti-14589	191	18	on	on	ADP
aiti-14589	191	19	fine	fine	ADV
aiti-14589	191	20	-	-	PUNCT
aiti-14589	191	21	tuning	tuning	NOUN
aiti-14589	191	22	but	but	CCONJ
aiti-14589	191	23	rather	rather	ADV
aiti-14589	191	24	on	on	ADP
aiti-14589	191	25	evaluating	evaluate	VERB
aiti-14589	191	26	their	their	PRON
aiti-14589	191	27	baseline	baseline	ADJ
aiti-14589	191	28	performance	performance	NOUN
aiti-14589	191	29	for	for	ADP
aiti-14589	191	30	this	this	DET
aiti-14589	191	31	classification	classification	NOUN
aiti-14589	191	32	task	task	NOUN
aiti-14589	191	33	.	.	PUNCT
aiti-14589	192	1	for	for	ADP
aiti-14589	192	2	cnn	cnn	PROPN
aiti-14589	192	3	-	-	PUNCT
aiti-14589	192	4	based	base	VERB
aiti-14589	192	5	classification	classification	NOUN
aiti-14589	192	6	,	,	PUNCT
aiti-14589	192	7	a	a	DET
aiti-14589	192	8	deep	deep	ADJ
aiti-14589	192	9	learning	learning	NOUN
aiti-14589	192	10	model	model	NOUN
aiti-14589	192	11	was	be	AUX
aiti-14589	192	12	developed	develop	VERB
aiti-14589	192	13	using	use	VERB
aiti-14589	192	14	the	the	DET
aiti-14589	192	15	keras	keras	PROPN
aiti-14589	192	16	framework	framework	NOUN
aiti-14589	192	17	.	.	PUNCT
aiti-14589	193	1	the	the	DET
aiti-14589	193	2	cnn	cnn	PROPN
aiti-14589	193	3	architecture	architecture	NOUN
aiti-14589	193	4	consisted	consist	VERB
aiti-14589	193	5	of	of	ADP
aiti-14589	193	6	three	three	NUM
aiti-14589	193	7	convolutional	convolutional	ADJ
aiti-14589	193	8	layers	layer	NOUN
aiti-14589	193	9	,	,	PUNCT
aiti-14589	193	10	designed	design	VERB
aiti-14589	193	11	to	to	PART
aiti-14589	193	12	progressively	progressively	ADV
aiti-14589	193	13	capture	capture	VERB
aiti-14589	193	14	hierarchical	hierarchical	ADJ
aiti-14589	193	15	features	feature	NOUN
aiti-14589	193	16	within	within	ADP
aiti-14589	193	17	the	the	DET
aiti-14589	193	18	image	image	NOUN
aiti-14589	193	19	.	.	PUNCT
aiti-14589	194	1	the	the	DET
aiti-14589	194	2	first	first	ADJ
aiti-14589	194	3	layer	layer	NOUN
aiti-14589	194	4	employed	employ	VERB
aiti-14589	194	5	16	16	NUM
aiti-14589	194	6	filters	filter	NOUN
aiti-14589	194	7	of	of	ADP
aiti-14589	194	8	size	size	NOUN
aiti-14589	194	9	(	(	PUNCT
aiti-14589	194	10	3×3	3×3	NUM
aiti-14589	194	11	)	)	PUNCT
aiti-14589	194	12	with	with	ADP
aiti-14589	194	13	a	a	DET
aiti-14589	194	14	stride	stride	NOUN
aiti-14589	194	15	of	of	ADP
aiti-14589	194	16	1	1	NUM
aiti-14589	194	17	,	,	PUNCT
aiti-14589	194	18	followed	follow	VERB
aiti-14589	194	19	by	by	ADP
aiti-14589	194	20	a	a	DET
aiti-14589	194	21	relu	relu	NOUN
aiti-14589	194	22	activation	activation	NOUN
aiti-14589	194	23	function	function	NOUN
aiti-14589	194	24	.	.	PUNCT
aiti-14589	195	1	the	the	DET
aiti-14589	195	2	subsequent	subsequent	ADJ
aiti-14589	195	3	layers	layer	NOUN
aiti-14589	195	4	increased	increase	VERB
aiti-14589	195	5	advances	advance	NOUN
aiti-14589	195	6	in	in	ADP
aiti-14589	195	7	technology	technology	NOUN
aiti-14589	195	8	innovation	innovation	NOUN
aiti-14589	195	9	,	,	PUNCT
aiti-14589	195	10	vol	vol	NOUN
aiti-14589	195	11	.	.	PROPN
aiti-14589	196	1	10	10	NUM
aiti-14589	196	2	,	,	PUNCT
aiti-14589	196	3	no	no	INTJ
aiti-14589	196	4	.	.	NOUN
aiti-14589	196	5	4	4	NUM
aiti-14589	196	6	,	,	PUNCT
aiti-14589	196	7	2025	2025	NUM
aiti-14589	196	8	,	,	PUNCT
aiti-14589	196	9	pp	pp	ADJ
aiti-14589	196	10	.	.	PUNCT
aiti-14589	197	1	395	395	NUM
aiti-14589	197	2	-	-	SYM
aiti-14589	197	3	406	406	NUM
aiti-14589	197	4	402	402	NUM
aiti-14589	197	5	the	the	DET
aiti-14589	197	6	number	number	NOUN
aiti-14589	197	7	of	of	ADP
aiti-14589	197	8	filters	filter	NOUN
aiti-14589	197	9	to	to	ADP
aiti-14589	197	10	32	32	NUM
aiti-14589	197	11	and	and	CCONJ
aiti-14589	197	12	16	16	NUM
aiti-14589	197	13	,	,	PUNCT
aiti-14589	197	14	respectively	respectively	ADV
aiti-14589	197	15	.	.	PUNCT
aiti-14589	198	1	max	max	PROPN
aiti-14589	198	2	-	-	PUNCT
aiti-14589	198	3	pooling	pool	VERB
aiti-14589	198	4	layers	layer	NOUN
aiti-14589	198	5	were	be	AUX
aiti-14589	198	6	incorporated	incorporate	VERB
aiti-14589	198	7	to	to	PART
aiti-14589	198	8	reduce	reduce	VERB
aiti-14589	198	9	spatial	spatial	ADJ
aiti-14589	198	10	dimensions	dimension	NOUN
aiti-14589	198	11	while	while	SCONJ
aiti-14589	198	12	preserving	preserve	VERB
aiti-14589	198	13	essential	essential	ADJ
aiti-14589	198	14	information	information	NOUN
aiti-14589	198	15	.	.	PUNCT
aiti-14589	199	1	the	the	DET
aiti-14589	199	2	extracted	extract	VERB
aiti-14589	199	3	features	feature	NOUN
aiti-14589	199	4	were	be	AUX
aiti-14589	199	5	flattened	flatten	VERB
aiti-14589	199	6	and	and	CCONJ
aiti-14589	199	7	passed	pass	VERB
aiti-14589	199	8	through	through	ADP
aiti-14589	199	9	a	a	DET
aiti-14589	199	10	fully	fully	ADV
aiti-14589	199	11	connected	connect	VERB
aiti-14589	199	12	layer	layer	NOUN
aiti-14589	199	13	with	with	ADP
aiti-14589	199	14	256	256	NUM
aiti-14589	199	15	neurons	neuron	NOUN
aiti-14589	199	16	,	,	PUNCT
aiti-14589	199	17	utilizing	utilize	VERB
aiti-14589	199	18	relu	relu	NOUN
aiti-14589	199	19	activation	activation	NOUN
aiti-14589	199	20	to	to	PART
aiti-14589	199	21	capture	capture	VERB
aiti-14589	199	22	high	high	ADJ
aiti-14589	199	23	-	-	PUNCT
aiti-14589	199	24	level	level	NOUN
aiti-14589	199	25	patterns	pattern	NOUN
aiti-14589	199	26	.	.	PUNCT
aiti-14589	200	1	finally	finally	ADV
aiti-14589	200	2	,	,	PUNCT
aiti-14589	200	3	the	the	DET
aiti-14589	200	4	output	output	NOUN
aiti-14589	200	5	layer	layer	NOUN
aiti-14589	200	6	consisted	consist	VERB
aiti-14589	200	7	of	of	ADP
aiti-14589	200	8	a	a	DET
aiti-14589	200	9	single	single	ADJ
aiti-14589	200	10	neuron	neuron	NOUN
aiti-14589	200	11	with	with	ADP
aiti-14589	200	12	a	a	DET
aiti-14589	200	13	sigmoid	sigmoid	NOUN
aiti-14589	200	14	activation	activation	NOUN
aiti-14589	200	15	function	function	NOUN
aiti-14589	200	16	,	,	PUNCT
aiti-14589	200	17	making	make	VERB
aiti-14589	200	18	it	it	PRON
aiti-14589	200	19	suitable	suitable	ADJ
aiti-14589	200	20	for	for	ADP
aiti-14589	200	21	binary	binary	ADJ
aiti-14589	200	22	classification	classification	NOUN
aiti-14589	200	23	.	.	PUNCT
aiti-14589	201	1	the	the	DET
aiti-14589	201	2	cnn	cnn	PROPN
aiti-14589	201	3	model	model	NOUN
aiti-14589	201	4	was	be	AUX
aiti-14589	201	5	trained	train	VERB
aiti-14589	201	6	using	use	VERB
aiti-14589	201	7	the	the	DET
aiti-14589	201	8	adam	adam	PROPN
aiti-14589	201	9	optimizer	optimizer	NOUN
aiti-14589	201	10	with	with	ADP
aiti-14589	201	11	a	a	DET
aiti-14589	201	12	learning	learn	VERB
aiti-14589	201	13	rate	rate	NOUN
aiti-14589	201	14	of	of	ADP
aiti-14589	201	15	0.001	0.001	NUM
aiti-14589	201	16	,	,	PUNCT
aiti-14589	201	17	and	and	CCONJ
aiti-14589	201	18	binary	binary	ADJ
aiti-14589	201	19	cross	cross	NOUN
aiti-14589	201	20	-	-	ADJ
aiti-14589	201	21	entropy	entropy	ADJ
aiti-14589	201	22	loss	loss	NOUN
aiti-14589	201	23	was	be	AUX
aiti-14589	201	24	used	use	VERB
aiti-14589	201	25	to	to	PART
aiti-14589	201	26	measure	measure	VERB
aiti-14589	201	27	classification	classification	NOUN
aiti-14589	201	28	error	error	NOUN
aiti-14589	201	29	.	.	PUNCT
aiti-14589	202	1	the	the	DET
aiti-14589	202	2	training	training	NOUN
aiti-14589	202	3	process	process	NOUN
aiti-14589	202	4	spanned	span	VERB
aiti-14589	202	5	50	50	NUM
aiti-14589	202	6	epochs	epoch	NOUN
aiti-14589	202	7	with	with	ADP
aiti-14589	202	8	a	a	DET
aiti-14589	202	9	batch	batch	NOUN
aiti-14589	202	10	size	size	NOUN
aiti-14589	202	11	of	of	ADP
aiti-14589	202	12	32	32	NUM
aiti-14589	202	13	,	,	PUNCT
aiti-14589	202	14	incorporating	incorporate	VERB
aiti-14589	202	15	early	early	ADV
aiti-14589	202	16	stopping	stop	VERB
aiti-14589	202	17	to	to	PART
aiti-14589	202	18	prevent	prevent	VERB
aiti-14589	202	19	overfitting	overfitting	NOUN
aiti-14589	202	20	.	.	PUNCT
aiti-14589	203	1	the	the	DET
aiti-14589	203	2	cnn	cnn	PROPN
aiti-14589	203	3	architecture	architecture	NOUN
aiti-14589	203	4	used	use	VERB
aiti-14589	203	5	in	in	ADP
aiti-14589	203	6	this	this	DET
aiti-14589	203	7	study	study	NOUN
aiti-14589	203	8	is	be	AUX
aiti-14589	203	9	illustrated	illustrate	VERB
aiti-14589	203	10	in	in	ADP
aiti-14589	203	11	fig	fig	NOUN
aiti-14589	203	12	.	.	PUNCT
aiti-14589	204	1	5	5	X
aiti-14589	204	2	.	.	PUNCT
aiti-14589	204	3	the	the	DET
aiti-14589	204	4	performance	performance	NOUN
aiti-14589	204	5	of	of	ADP
aiti-14589	204	6	all	all	DET
aiti-14589	204	7	models	model	NOUN
aiti-14589	204	8	was	be	AUX
aiti-14589	204	9	evaluated	evaluate	VERB
aiti-14589	204	10	using	use	VERB
aiti-14589	204	11	precision	precision	NOUN
aiti-14589	204	12	,	,	PUNCT
aiti-14589	204	13	recall	recall	NOUN
aiti-14589	204	14	,	,	PUNCT
aiti-14589	204	15	and	and	CCONJ
aiti-14589	204	16	f1	f1	NOUN
aiti-14589	204	17	-	-	PUNCT
aiti-14589	204	18	score	score	NOUN
aiti-14589	204	19	,	,	PUNCT
aiti-14589	204	20	providing	provide	VERB
aiti-14589	204	21	a	a	DET
aiti-14589	204	22	comprehensive	comprehensive	ADJ
aiti-14589	204	23	assessment	assessment	NOUN
aiti-14589	204	24	of	of	ADP
aiti-14589	204	25	classification	classification	NOUN
aiti-14589	204	26	effectiveness	effectiveness	NOUN
aiti-14589	204	27	.	.	PUNCT
aiti-14589	205	1	additionally	additionally	ADV
aiti-14589	205	2	,	,	PUNCT
aiti-14589	205	3	confusion	confusion	NOUN
aiti-14589	205	4	matrices	matrix	NOUN
aiti-14589	205	5	were	be	AUX
aiti-14589	205	6	generated	generate	VERB
aiti-14589	205	7	to	to	PART
aiti-14589	205	8	analyze	analyze	VERB
aiti-14589	205	9	misclassifications	misclassification	NOUN
aiti-14589	205	10	and	and	CCONJ
aiti-14589	205	11	model	model	NOUN
aiti-14589	205	12	behavior	behavior	NOUN
aiti-14589	205	13	.	.	PUNCT
aiti-14589	206	1	to	to	PART
aiti-14589	206	2	assess	assess	VERB
aiti-14589	206	3	computational	computational	ADJ
aiti-14589	206	4	efficiency	efficiency	NOUN
aiti-14589	206	5	,	,	PUNCT
aiti-14589	206	6	the	the	DET
aiti-14589	206	7	training	training	NOUN
aiti-14589	206	8	time	time	NOUN
aiti-14589	206	9	for	for	ADP
aiti-14589	206	10	each	each	DET
aiti-14589	206	11	model	model	NOUN
aiti-14589	206	12	was	be	AUX
aiti-14589	206	13	recorded	record	VERB
aiti-14589	206	14	.	.	PUNCT
aiti-14589	207	1	fig	fig	NOUN
aiti-14589	207	2	.	.	PUNCT
aiti-14589	208	1	5	5	NUM
aiti-14589	208	2	image	image	NOUN
aiti-14589	208	3	-	-	PUNCT
aiti-14589	208	4	based	base	VERB
aiti-14589	208	5	cnn	cnn	PROPN
aiti-14589	208	6	model	model	NOUN
aiti-14589	208	7	architecture	architecture	NOUN
aiti-14589	208	8	4.3	4.3	NUM
aiti-14589	208	9	.	.	PUNCT
aiti-14589	208	10	landmark	landmark	NOUN
aiti-14589	208	11	-	-	PUNCT
aiti-14589	208	12	based	base	VERB
aiti-14589	208	13	method	method	NOUN
aiti-14589	208	14	in	in	ADP
aiti-14589	208	15	the	the	DET
aiti-14589	208	16	landmark	landmark	NOUN
aiti-14589	208	17	-	-	PUNCT
aiti-14589	208	18	based	base	VERB
aiti-14589	208	19	method	method	NOUN
aiti-14589	208	20	,	,	PUNCT
aiti-14589	208	21	facial	facial	ADJ
aiti-14589	208	22	images	image	NOUN
aiti-14589	208	23	were	be	AUX
aiti-14589	208	24	transformed	transform	VERB
aiti-14589	208	25	into	into	ADP
aiti-14589	208	26	structured	structured	ADJ
aiti-14589	208	27	numerical	numerical	ADJ
aiti-14589	208	28	representations	representation	NOUN
aiti-14589	208	29	by	by	ADP
aiti-14589	208	30	extracting	extract	VERB
aiti-14589	208	31	key	key	ADJ
aiti-14589	208	32	facial	facial	ADJ
aiti-14589	208	33	points	point	NOUN
aiti-14589	208	34	.	.	PUNCT
aiti-14589	209	1	the	the	DET
aiti-14589	209	2	mediapipe	mediapipe	NOUN
aiti-14589	209	3	library	library	NOUN
aiti-14589	209	4	was	be	AUX
aiti-14589	209	5	employed	employ	VERB
aiti-14589	209	6	to	to	PART
aiti-14589	209	7	extract	extract	VERB
aiti-14589	209	8	468	468	NUM
aiti-14589	209	9	facial	facial	ADJ
aiti-14589	209	10	landmarks	landmark	NOUN
aiti-14589	209	11	,	,	PUNCT
aiti-14589	209	12	capturing	capture	VERB
aiti-14589	209	13	the	the	DET
aiti-14589	209	14	relative	relative	ADJ
aiti-14589	209	15	positions	position	NOUN
aiti-14589	209	16	of	of	ADP
aiti-14589	209	17	key	key	ADJ
aiti-14589	209	18	facial	facial	ADJ
aiti-14589	209	19	features	feature	NOUN
aiti-14589	209	20	along	along	ADP
aiti-14589	209	21	the	the	DET
aiti-14589	209	22	x	x	NOUN
aiti-14589	209	23	and	and	CCONJ
aiti-14589	209	24	y	y	PROPN
aiti-14589	209	25	axes	axis	NOUN
aiti-14589	209	26	.	.	PUNCT
aiti-14589	210	1	this	this	DET
aiti-14589	210	2	approach	approach	NOUN
aiti-14589	210	3	enabled	enable	VERB
aiti-14589	210	4	the	the	DET
aiti-14589	210	5	conversion	conversion	NOUN
aiti-14589	210	6	of	of	ADP
aiti-14589	210	7	unstructured	unstructured	ADJ
aiti-14589	210	8	image	image	NOUN
aiti-14589	210	9	data	datum	NOUN
aiti-14589	210	10	into	into	ADP
aiti-14589	210	11	a	a	DET
aiti-14589	210	12	standardized	standardized	ADJ
aiti-14589	210	13	format	format	NOUN
aiti-14589	210	14	for	for	ADP
aiti-14589	210	15	ml	ml	NOUN
aiti-14589	210	16	classification	classification	NOUN
aiti-14589	210	17	.	.	PUNCT
aiti-14589	211	1	following	follow	VERB
aiti-14589	211	2	preprocessing	preprocesse	VERB
aiti-14589	211	3	,	,	PUNCT
aiti-14589	211	4	the	the	DET
aiti-14589	211	5	extracted	extract	VERB
aiti-14589	211	6	landmarks	landmark	NOUN
aiti-14589	211	7	were	be	AUX
aiti-14589	211	8	stored	store	VERB
aiti-14589	211	9	as	as	ADP
aiti-14589	211	10	structured	structure	VERB
aiti-14589	211	11	data	datum	NOUN
aiti-14589	211	12	and	and	CCONJ
aiti-14589	211	13	normalized	normalize	VERB
aiti-14589	211	14	using	use	VERB
aiti-14589	211	15	feature	feature	NOUN
aiti-14589	211	16	standardization	standardization	NOUN
aiti-14589	211	17	,	,	PUNCT
aiti-14589	211	18	ensuring	ensure	VERB
aiti-14589	211	19	values	value	NOUN
aiti-14589	211	20	were	be	AUX
aiti-14589	211	21	scaled	scale	VERB
aiti-14589	211	22	to	to	ADP
aiti-14589	211	23	a	a	DET
aiti-14589	211	24	comparable	comparable	ADJ
aiti-14589	211	25	range	range	NOUN
aiti-14589	211	26	.	.	PUNCT
aiti-14589	212	1	the	the	DET
aiti-14589	212	2	labels	label	NOUN
aiti-14589	212	3	were	be	AUX
aiti-14589	212	4	encoded	encode	VERB
aiti-14589	212	5	as	as	ADP
aiti-14589	212	6	binary	binary	ADJ
aiti-14589	212	7	values	value	NOUN
aiti-14589	212	8	,	,	PUNCT
aiti-14589	212	9	with	with	ADP
aiti-14589	212	10	normal	normal	ADJ
aiti-14589	212	11	expressions	expression	NOUN
aiti-14589	212	12	labeled	label	VERB
aiti-14589	212	13	as	as	ADP
aiti-14589	212	14	0	0	NUM
aiti-14589	212	15	and	and	CCONJ
aiti-14589	212	16	hurt	hurt	VERB
aiti-14589	212	17	expressions	expression	NOUN
aiti-14589	212	18	labeled	label	VERB
aiti-14589	212	19	as	as	ADP
aiti-14589	212	20	1	1	NUM
aiti-14589	212	21	.	.	PUNCT
aiti-14589	213	1	for	for	ADP
aiti-14589	213	2	cnn	cnn	PROPN
aiti-14589	213	3	-	-	PUNCT
aiti-14589	213	4	based	base	VERB
aiti-14589	213	5	classification	classification	NOUN
aiti-14589	213	6	,	,	PUNCT
aiti-14589	213	7	the	the	DET
aiti-14589	213	8	landmark	landmark	NOUN
aiti-14589	213	9	data	datum	NOUN
aiti-14589	213	10	was	be	AUX
aiti-14589	213	11	reshaped	reshape	VERB
aiti-14589	213	12	into	into	ADP
aiti-14589	213	13	a	a	DET
aiti-14589	213	14	format	format	NOUN
aiti-14589	213	15	compatible	compatible	ADJ
aiti-14589	213	16	with	with	ADP
aiti-14589	213	17	1d	1d	NUM
aiti-14589	213	18	convolutional	convolutional	ADJ
aiti-14589	213	19	layers	layer	NOUN
aiti-14589	213	20	,	,	PUNCT
aiti-14589	213	21	allowing	allow	VERB
aiti-14589	213	22	the	the	DET
aiti-14589	213	23	model	model	NOUN
aiti-14589	213	24	to	to	PART
aiti-14589	213	25	process	process	VERB
aiti-14589	213	26	sequential	sequential	ADJ
aiti-14589	213	27	patterns	pattern	NOUN
aiti-14589	213	28	in	in	ADP
aiti-14589	213	29	the	the	DET
aiti-14589	213	30	landmark	landmark	NOUN
aiti-14589	213	31	coordinates	coordinate	NOUN
aiti-14589	213	32	.	.	PUNCT
aiti-14589	214	1	as	as	ADP
aiti-14589	214	2	in	in	ADP
aiti-14589	214	3	the	the	DET
aiti-14589	214	4	image	image	NOUN
aiti-14589	214	5	-	-	PUNCT
aiti-14589	214	6	based	base	VERB
aiti-14589	214	7	method	method	NOUN
aiti-14589	214	8	,	,	PUNCT
aiti-14589	214	9	the	the	DET
aiti-14589	214	10	lrc	lrc	PROPN
aiti-14589	214	11	,	,	PUNCT
aiti-14589	214	12	svm	svm	PROPN
aiti-14589	214	13	,	,	PUNCT
aiti-14589	214	14	rfc	rfc	PRON
aiti-14589	214	15	,	,	PUNCT
aiti-14589	214	16	and	and	CCONJ
aiti-14589	214	17	gbc	gbc	NOUN
aiti-14589	214	18	models	model	NOUN
aiti-14589	214	19	were	be	AUX
aiti-14589	214	20	trained	train	VERB
aiti-14589	214	21	using	use	VERB
aiti-14589	214	22	default	default	NOUN
aiti-14589	214	23	parameters	parameter	NOUN
aiti-14589	214	24	,	,	PUNCT
aiti-14589	214	25	ensuring	ensure	VERB
aiti-14589	214	26	that	that	SCONJ
aiti-14589	214	27	the	the	DET
aiti-14589	214	28	evaluation	evaluation	NOUN
aiti-14589	214	29	remained	remain	VERB
aiti-14589	214	30	focused	focused	ADJ
aiti-14589	214	31	on	on	ADP
aiti-14589	214	32	baseline	baseline	ADJ
aiti-14589	214	33	performance	performance	NOUN
aiti-14589	214	34	rather	rather	ADV
aiti-14589	214	35	than	than	ADP
aiti-14589	214	36	fine	fine	ADV
aiti-14589	214	37	-	-	PUNCT
aiti-14589	214	38	tuning	tuning	NOUN
aiti-14589	214	39	.	.	PUNCT
aiti-14589	215	1	for	for	ADP
aiti-14589	215	2	the	the	DET
aiti-14589	215	3	cnn	cnn	PROPN
aiti-14589	215	4	model	model	NOUN
aiti-14589	215	5	,	,	PUNCT
aiti-14589	215	6	the	the	DET
aiti-14589	215	7	architecture	architecture	NOUN
aiti-14589	215	8	was	be	AUX
aiti-14589	215	9	adapted	adapt	VERB
aiti-14589	215	10	to	to	PART
aiti-14589	215	11	handle	handle	VERB
aiti-14589	215	12	1d	1d	NUM
aiti-14589	215	13	landmark	landmark	NOUN
aiti-14589	215	14	data	datum	NOUN
aiti-14589	215	15	,	,	PUNCT
aiti-14589	215	16	replacing	replace	VERB
aiti-14589	215	17	2d	2d	NUM
aiti-14589	215	18	convolutional	convolutional	ADJ
aiti-14589	215	19	layers	layer	NOUN
aiti-14589	215	20	with	with	ADP
aiti-14589	215	21	1d	1d	NUM
aiti-14589	215	22	convolutional	convolutional	ADJ
aiti-14589	215	23	layers	layer	NOUN
aiti-14589	215	24	.	.	PUNCT
aiti-14589	216	1	the	the	DET
aiti-14589	216	2	model	model	NOUN
aiti-14589	216	3	consisted	consist	VERB
aiti-14589	216	4	of	of	ADP
aiti-14589	216	5	an	an	DET
aiti-14589	216	6	initial	initial	ADJ
aiti-14589	216	7	1d	1d	NUM
aiti-14589	216	8	convolutional	convolutional	ADJ
aiti-14589	216	9	layer	layer	NOUN
aiti-14589	216	10	with	with	ADP
aiti-14589	216	11	32	32	NUM
aiti-14589	216	12	filters	filter	NOUN
aiti-14589	216	13	and	and	CCONJ
aiti-14589	216	14	a	a	DET
aiti-14589	216	15	kernel	kernel	NOUN
aiti-14589	216	16	size	size	NOUN
aiti-14589	216	17	of	of	ADP
aiti-14589	216	18	3	3	NUM
aiti-14589	216	19	,	,	PUNCT
aiti-14589	216	20	followed	follow	VERB
aiti-14589	216	21	by	by	ADP
aiti-14589	216	22	relu	relu	NOUN
aiti-14589	216	23	activation	activation	NOUN
aiti-14589	216	24	and	and	CCONJ
aiti-14589	216	25	max	max	PROPN
aiti-14589	216	26	-	-	PUNCT
aiti-14589	216	27	pooling	pooling	NOUN
aiti-14589	216	28	to	to	PART
aiti-14589	216	29	downsample	downsample	VERB
aiti-14589	216	30	the	the	DET
aiti-14589	216	31	feature	feature	NOUN
aiti-14589	216	32	representation	representation	NOUN
aiti-14589	216	33	.	.	PUNCT
aiti-14589	217	1	a	a	DET
aiti-14589	217	2	flattening	flattening	NOUN
aiti-14589	217	3	layer	layer	NOUN
aiti-14589	217	4	was	be	AUX
aiti-14589	217	5	then	then	ADV
aiti-14589	217	6	used	use	VERB
aiti-14589	217	7	to	to	PART
aiti-14589	217	8	convert	convert	VERB
aiti-14589	217	9	the	the	DET
aiti-14589	217	10	extracted	extract	VERB
aiti-14589	217	11	spatial	spatial	ADJ
aiti-14589	217	12	features	feature	NOUN
aiti-14589	217	13	into	into	ADP
aiti-14589	217	14	a	a	DET
aiti-14589	217	15	one	one	NUM
aiti-14589	217	16	-	-	PUNCT
aiti-14589	217	17	dimensional	dimensional	ADJ
aiti-14589	217	18	vector	vector	NOUN
aiti-14589	217	19	.	.	PUNCT
aiti-14589	218	1	the	the	DET
aiti-14589	218	2	fully	fully	ADV
aiti-14589	218	3	connected	connect	VERB
aiti-14589	218	4	layer	layer	NOUN
aiti-14589	218	5	comprised	comprise	VERB
aiti-14589	218	6	128	128	NUM
aiti-14589	218	7	neurons	neuron	NOUN
aiti-14589	218	8	with	with	ADP
aiti-14589	218	9	relu	relu	NOUN
aiti-14589	218	10	activation	activation	NOUN
aiti-14589	218	11	,	,	PUNCT
aiti-14589	218	12	followed	follow	VERB
aiti-14589	218	13	by	by	ADP
aiti-14589	218	14	an	an	DET
aiti-14589	218	15	output	output	NOUN
aiti-14589	218	16	layer	layer	NOUN
aiti-14589	218	17	with	with	ADP
aiti-14589	218	18	sigmoid	sigmoid	NOUN
aiti-14589	218	19	activation	activation	NOUN
aiti-14589	218	20	for	for	ADP
aiti-14589	218	21	binary	binary	ADJ
aiti-14589	218	22	classification	classification	NOUN
aiti-14589	218	23	.	.	PUNCT
aiti-14589	219	1	the	the	DET
aiti-14589	219	2	model	model	NOUN
aiti-14589	219	3	was	be	AUX
aiti-14589	219	4	compiled	compile	VERB
aiti-14589	219	5	using	use	VERB
aiti-14589	219	6	the	the	DET
aiti-14589	219	7	adam	adam	PROPN
aiti-14589	219	8	optimizer	optimizer	NOUN
aiti-14589	219	9	with	with	ADP
aiti-14589	219	10	a	a	DET
aiti-14589	219	11	learning	learn	VERB
aiti-14589	219	12	rate	rate	NOUN
aiti-14589	219	13	of	of	ADP
aiti-14589	219	14	0.001	0.001	NUM
aiti-14589	219	15	,	,	PUNCT
aiti-14589	219	16	and	and	CCONJ
aiti-14589	219	17	training	training	NOUN
aiti-14589	219	18	was	be	AUX
aiti-14589	219	19	performed	perform	VERB
aiti-14589	219	20	with	with	ADP
aiti-14589	219	21	binary	binary	PROPN
aiti-14589	219	22	cross	cross	PROPN
aiti-14589	219	23	-	-	ADJ
aiti-14589	219	24	entropy	entropy	ADJ
aiti-14589	219	25	loss	loss	NOUN
aiti-14589	219	26	.	.	PUNCT
aiti-14589	220	1	advances	advance	NOUN
aiti-14589	220	2	in	in	ADP
aiti-14589	220	3	technology	technology	NOUN
aiti-14589	220	4	innovation	innovation	NOUN
aiti-14589	220	5	,	,	PUNCT
aiti-14589	220	6	vol	vol	NOUN
aiti-14589	220	7	.	.	PROPN
aiti-14589	221	1	10	10	NUM
aiti-14589	221	2	,	,	PUNCT
aiti-14589	221	3	no	no	INTJ
aiti-14589	221	4	.	.	NOUN
aiti-14589	221	5	4	4	NUM
aiti-14589	221	6	,	,	PUNCT
aiti-14589	221	7	2025	2025	NUM
aiti-14589	221	8	,	,	PUNCT
aiti-14589	221	9	pp	pp	ADJ
aiti-14589	221	10	.	.	PUNCT
aiti-14589	222	1	395	395	NUM
aiti-14589	222	2	-	-	SYM
aiti-14589	222	3	406	406	NUM
aiti-14589	222	4	403	403	NUM
aiti-14589	222	5	the	the	DET
aiti-14589	222	6	cnn	cnn	PROPN
aiti-14589	222	7	architecture	architecture	NOUN
aiti-14589	222	8	for	for	ADP
aiti-14589	222	9	the	the	DET
aiti-14589	222	10	landmark	landmark	NOUN
aiti-14589	222	11	-	-	PUNCT
aiti-14589	222	12	based	base	VERB
aiti-14589	222	13	method	method	NOUN
aiti-14589	222	14	is	be	AUX
aiti-14589	222	15	illustrated	illustrate	VERB
aiti-14589	222	16	in	in	ADP
aiti-14589	222	17	fig	fig	NOUN
aiti-14589	222	18	.	.	PUNCT
aiti-14589	223	1	6	6	NUM
aiti-14589	223	2	.	.	X
aiti-14589	223	3	as	as	ADP
aiti-14589	223	4	with	with	ADP
aiti-14589	223	5	the	the	DET
aiti-14589	223	6	image	image	NOUN
aiti-14589	223	7	-	-	PUNCT
aiti-14589	223	8	based	base	VERB
aiti-14589	223	9	approach	approach	NOUN
aiti-14589	223	10	,	,	PUNCT
aiti-14589	223	11	model	model	NOUN
aiti-14589	223	12	performance	performance	NOUN
aiti-14589	223	13	was	be	AUX
aiti-14589	223	14	assessed	assess	VERB
aiti-14589	223	15	using	use	VERB
aiti-14589	223	16	precision	precision	NOUN
aiti-14589	223	17	,	,	PUNCT
aiti-14589	223	18	recall	recall	NOUN
aiti-14589	223	19	,	,	PUNCT
aiti-14589	223	20	and	and	CCONJ
aiti-14589	223	21	f1	f1	NOUN
aiti-14589	223	22	-	-	PUNCT
aiti-14589	223	23	score	score	NOUN
aiti-14589	223	24	.	.	PUNCT
aiti-14589	224	1	confusion	confusion	NOUN
aiti-14589	224	2	matrices	matrix	NOUN
aiti-14589	224	3	were	be	AUX
aiti-14589	224	4	utilized	utilize	VERB
aiti-14589	224	5	to	to	PART
aiti-14589	224	6	further	far	ADV
aiti-14589	224	7	evaluate	evaluate	VERB
aiti-14589	224	8	classification	classification	NOUN
aiti-14589	224	9	accuracy	accuracy	NOUN
aiti-14589	224	10	,	,	PUNCT
aiti-14589	224	11	while	while	SCONJ
aiti-14589	224	12	training	training	NOUN
aiti-14589	224	13	time	time	NOUN
aiti-14589	224	14	was	be	AUX
aiti-14589	224	15	recorded	record	VERB
aiti-14589	224	16	to	to	PART
aiti-14589	224	17	compare	compare	VERB
aiti-14589	224	18	computational	computational	ADJ
aiti-14589	224	19	efficiency	efficiency	NOUN
aiti-14589	224	20	.	.	PUNCT
aiti-14589	225	1	fig	fig	NOUN
aiti-14589	225	2	.	.	PUNCT
aiti-14589	226	1	6	6	NUM
aiti-14589	226	2	landmark	landmark	NOUN
aiti-14589	226	3	-	-	PUNCT
aiti-14589	226	4	based	base	VERB
aiti-14589	226	5	cnn	cnn	PROPN
aiti-14589	226	6	model	model	NOUN
aiti-14589	226	7	architecture	architecture	NOUN
aiti-14589	226	8	5	5	NUM
aiti-14589	226	9	.	.	PUNCT
aiti-14589	226	10	experimental	experimental	ADJ
aiti-14589	226	11	results	result	NOUN
aiti-14589	226	12	this	this	DET
aiti-14589	226	13	study	study	NOUN
aiti-14589	226	14	compares	compare	VERB
aiti-14589	226	15	image	image	NOUN
aiti-14589	226	16	-	-	PUNCT
aiti-14589	226	17	based	base	VERB
aiti-14589	226	18	and	and	CCONJ
aiti-14589	226	19	landmark	landmark	NOUN
aiti-14589	226	20	-	-	PUNCT
aiti-14589	226	21	based	base	VERB
aiti-14589	226	22	fer	fer	PROPN
aiti-14589	226	23	techniques	technique	NOUN
aiti-14589	226	24	using	use	VERB
aiti-14589	226	25	various	various	ADJ
aiti-14589	226	26	ml	ml	NOUN
aiti-14589	226	27	models	model	NOUN
aiti-14589	226	28	,	,	PUNCT
aiti-14589	226	29	highlighting	highlight	VERB
aiti-14589	226	30	the	the	DET
aiti-14589	226	31	distinct	distinct	ADJ
aiti-14589	226	32	advantages	advantage	NOUN
aiti-14589	226	33	and	and	CCONJ
aiti-14589	226	34	challenges	challenge	NOUN
aiti-14589	226	35	of	of	ADP
aiti-14589	226	36	each	each	DET
aiti-14589	226	37	approach	approach	NOUN
aiti-14589	226	38	,	,	PUNCT
aiti-14589	226	39	with	with	ADP
aiti-14589	226	40	the	the	DET
aiti-14589	226	41	input	input	NOUN
aiti-14589	226	42	data	datum	NOUN
aiti-14589	226	43	from	from	ADP
aiti-14589	226	44	both	both	DET
aiti-14589	226	45	methods	method	NOUN
aiti-14589	226	46	illustrated	illustrate	VERB
aiti-14589	226	47	in	in	ADP
aiti-14589	226	48	fig	fig	NOUN
aiti-14589	226	49	.	.	PUNCT
aiti-14589	227	1	7	7	NUM
aiti-14589	227	2	,	,	PUNCT
aiti-14589	227	3	and	and	CCONJ
aiti-14589	227	4	the	the	DET
aiti-14589	227	5	results	result	NOUN
aiti-14589	227	6	,	,	PUNCT
aiti-14589	227	7	including	include	VERB
aiti-14589	227	8	performance	performance	NOUN
aiti-14589	227	9	metrics	metric	NOUN
aiti-14589	227	10	and	and	CCONJ
aiti-14589	227	11	model	model	NOUN
aiti-14589	227	12	comparisons	comparison	NOUN
aiti-14589	227	13	,	,	PUNCT
aiti-14589	227	14	summarized	summarize	VERB
aiti-14589	227	15	in	in	ADP
aiti-14589	227	16	table	table	NOUN
aiti-14589	227	17	3	3	NUM
aiti-14589	227	18	.	.	PUNCT
aiti-14589	227	19	fig	fig	NOUN
aiti-14589	227	20	.	.	PUNCT
aiti-14589	228	1	7	7	NUM
aiti-14589	228	2	research	research	NOUN
aiti-14589	228	3	input	input	NOUN
aiti-14589	228	4	data	data	NOUN
aiti-14589	228	5	illusion	illusion	NOUN
aiti-14589	228	6	table	table	NOUN
aiti-14589	228	7	3	3	NUM
aiti-14589	228	8	machine	machine	NOUN
aiti-14589	228	9	learning	learning	NOUN
aiti-14589	228	10	model	model	NOUN
aiti-14589	228	11	performance	performance	NOUN
aiti-14589	228	12	model	model	NOUN
aiti-14589	228	13	techniques	technique	NOUN
aiti-14589	228	14	training	training	NOUN
aiti-14589	228	15	time	time	NOUN
aiti-14589	228	16	(	(	PUNCT
aiti-14589	228	17	seconds	second	NOUN
aiti-14589	228	18	)	)	PUNCT
aiti-14589	228	19	accuracy	accuracy	NOUN
aiti-14589	228	20	precision	precision	NOUN
aiti-14589	228	21	recall	recall	VERB
aiti-14589	228	22	f1	f1	NOUN
aiti-14589	228	23	-	-	PUNCT
aiti-14589	228	24	score	score	NOUN
aiti-14589	228	25	svm	svm	PROPN
aiti-14589	228	26	landmark	landmark	PROPN
aiti-14589	228	27	0.24	0.24	NUM
aiti-14589	228	28	0.8550	0.8550	NUM
aiti-14589	228	29	0.8876	0.8876	NUM
aiti-14589	228	30	0.8550	0.8550	NUM
aiti-14589	228	31	0.8519	0.8519	NUM
aiti-14589	228	32	image	image	NOUN
aiti-14589	228	33	-	-	PUNCT
aiti-14589	228	34	hog	hog	NOUN
aiti-14589	228	35	-	-	PUNCT
aiti-14589	228	36	scaling	scale	VERB
aiti-14589	228	37	0.19	0.19	NUM
aiti-14589	228	38	0.8200	0.8200	NUM
aiti-14589	228	39	0.8363	0.8363	NUM
aiti-14589	228	40	0.8200	0.8200	NUM
aiti-14589	228	41	0.8178	0.8178	NUM
aiti-14589	228	42	gbc	gbc	PROPN
aiti-14589	228	43	landmark	landmark	NOUN
aiti-14589	228	44	31.72	31.72	NUM
aiti-14589	228	45	0.8850	0.8850	NUM
aiti-14589	228	46	0.8965	0.8965	NUM
aiti-14589	228	47	0.8850	0.8850	NUM
aiti-14589	228	48	0.8842	0.8842	NUM
aiti-14589	228	49	image	image	NOUN
aiti-14589	228	50	-	-	PUNCT
aiti-14589	228	51	hog	hog	NOUN
aiti-14589	228	52	-	-	PUNCT
aiti-14589	228	53	scaling	scale	VERB
aiti-14589	228	54	20.93	20.93	NUM
aiti-14589	228	55	0.8700	0.8700	NUM
aiti-14589	228	56	0.8724	0.8724	NUM
aiti-14589	228	57	0.8700	0.8700	NUM
aiti-14589	228	58	0.8698	0.8698	NUM
aiti-14589	228	59	rfc	rfc	NOUN
aiti-14589	228	60	landmark	landmark	NOUN
aiti-14589	228	61	1.34	1.34	NUM
aiti-14589	228	62	0.8950	0.8950	NUM
aiti-14589	228	63	0.9018	0.9018	NUM
aiti-14589	228	64	0.8950	0.8950	NUM
aiti-14589	228	65	0.8946	0.8946	NUM
aiti-14589	228	66	image	image	NOUN
aiti-14589	228	67	-	-	PUNCT
aiti-14589	228	68	hog	hog	NOUN
aiti-14589	228	69	-	-	PUNCT
aiti-14589	228	70	scaling	scale	VERB
aiti-14589	228	71	1.36	1.36	NUM
aiti-14589	228	72	0.8600	0.8600	NUM
aiti-14589	228	73	0.8601	0.8601	NUM
aiti-14589	228	74	0.8600	0.8600	NUM
aiti-14589	228	75	0.8600	0.8600	NUM
aiti-14589	228	76	lrc	lrc	PROPN
aiti-14589	228	77	landmark	landmark	PROPN
aiti-14589	228	78	0.11	0.11	NUM
aiti-14589	228	79	0.8850	0.8850	NUM
aiti-14589	228	80	0.8939	0.8939	NUM
aiti-14589	228	81	0.8850	0.8850	NUM
aiti-14589	228	82	0.8843	0.8843	NUM
aiti-14589	228	83	image	image	NOUN
aiti-14589	228	84	-	-	PUNCT
aiti-14589	228	85	hog	hog	NOUN
aiti-14589	228	86	-	-	PUNCT
aiti-14589	228	87	scaling	scale	VERB
aiti-14589	228	88	0.09	0.09	NUM
aiti-14589	228	89	0.7900	0.7900	NUM
aiti-14589	228	90	0.8047	0.8047	NUM
aiti-14589	228	91	0.7900	0.7900	NUM
aiti-14589	228	92	0.7874	0.7874	NUM
aiti-14589	228	93	cnn	cnn	PROPN
aiti-14589	228	94	landmark	landmark	NOUN
aiti-14589	228	95	2.01	2.01	NUM
aiti-14589	228	96	0.9500	0.9500	NUM
aiti-14589	229	1	0.9541	0.9541	NUM
aiti-14589	229	2	0.9541	0.9541	NUM
aiti-14589	229	3	0.9541	0.9541	NUM
aiti-14589	229	4	image	image	NOUN
aiti-14589	229	5	-	-	PUNCT
aiti-14589	229	6	hog	hog	NOUN
aiti-14589	229	7	-	-	PUNCT
aiti-14589	229	8	scaling	scale	VERB
aiti-14589	229	9	88.89s	88.89s	NUM
aiti-14589	229	10	0.8700	0.8700	NUM
aiti-14589	229	11	0.8558	0.8558	NUM
aiti-14589	229	12	0.8900	0.8900	NUM
aiti-14589	229	13	0.8725	0.8725	NUM
aiti-14589	229	14	advances	advance	NOUN
aiti-14589	229	15	in	in	ADP
aiti-14589	229	16	technology	technology	NOUN
aiti-14589	229	17	innovation	innovation	NOUN
aiti-14589	229	18	,	,	PUNCT
aiti-14589	229	19	vol	vol	NOUN
aiti-14589	229	20	.	.	PROPN
aiti-14589	229	21	10	10	NUM
aiti-14589	229	22	,	,	PUNCT
aiti-14589	229	23	no	no	INTJ
aiti-14589	229	24	.	.	NOUN
aiti-14589	229	25	4	4	NUM
aiti-14589	229	26	,	,	PUNCT
aiti-14589	229	27	2025	2025	NUM
aiti-14589	229	28	,	,	PUNCT
aiti-14589	229	29	pp	pp	ADJ
aiti-14589	229	30	.	.	PUNCT
aiti-14589	230	1	395	395	NUM
aiti-14589	230	2	-	-	SYM
aiti-14589	230	3	406	406	NUM
aiti-14589	230	4	404	404	NUM
aiti-14589	230	5	the	the	DET
aiti-14589	230	6	classification	classification	NOUN
aiti-14589	230	7	results	result	NOUN
aiti-14589	230	8	of	of	ADP
aiti-14589	230	9	the	the	DET
aiti-14589	230	10	five	five	NUM
aiti-14589	230	11	models	model	NOUN
aiti-14589	230	12	on	on	ADP
aiti-14589	230	13	landmark	landmark	NOUN
aiti-14589	230	14	-	-	PUNCT
aiti-14589	230	15	based	base	VERB
aiti-14589	230	16	and	and	CCONJ
aiti-14589	230	17	image	image	NOUN
aiti-14589	230	18	-	-	PUNCT
aiti-14589	230	19	hog	hog	NOUN
aiti-14589	230	20	-	-	PUNCT
aiti-14589	230	21	scaling	scale	VERB
aiti-14589	230	22	features	feature	NOUN
aiti-14589	230	23	in	in	ADP
aiti-14589	230	24	table	table	NOUN
aiti-14589	230	25	3	3	NUM
aiti-14589	230	26	are	be	AUX
aiti-14589	230	27	summarized	summarize	VERB
aiti-14589	230	28	below	below	ADP
aiti-14589	230	29	:	:	PUNCT
aiti-14589	230	30	(	(	PUNCT
aiti-14589	230	31	1	1	X
aiti-14589	230	32	)	)	PUNCT
aiti-14589	230	33	the	the	DET
aiti-14589	230	34	svm	svm	PROPN
aiti-14589	230	35	model	model	NOUN
aiti-14589	230	36	performed	perform	VERB
aiti-14589	230	37	strongly	strongly	ADV
aiti-14589	230	38	with	with	ADP
aiti-14589	230	39	the	the	DET
aiti-14589	230	40	landmark	landmark	NOUN
aiti-14589	230	41	-	-	PUNCT
aiti-14589	230	42	based	base	VERB
aiti-14589	230	43	technique	technique	NOUN
aiti-14589	230	44	(	(	PUNCT
aiti-14589	230	45	accuracy	accuracy	NOUN
aiti-14589	230	46	0.8550	0.8550	NUM
aiti-14589	230	47	,	,	PUNCT
aiti-14589	230	48	precision	precision	NOUN
aiti-14589	230	49	0.8876	0.8876	NUM
aiti-14589	230	50	,	,	PUNCT
aiti-14589	230	51	recall	recall	VERB
aiti-14589	230	52	0.8550	0.8550	NUM
aiti-14589	230	53	,	,	PUNCT
aiti-14589	230	54	and	and	CCONJ
aiti-14589	230	55	f1	f1	NOUN
aiti-14589	230	56	-	-	PUNCT
aiti-14589	230	57	score	score	NOUN
aiti-14589	230	58	0.8519	0.8519	NUM
aiti-14589	230	59	)	)	PUNCT
aiti-14589	230	60	but	but	CCONJ
aiti-14589	230	61	slightly	slightly	ADV
aiti-14589	230	62	decreased	decrease	VERB
aiti-14589	230	63	with	with	ADP
aiti-14589	230	64	the	the	DET
aiti-14589	230	65	image	image	NOUN
aiti-14589	230	66	-	-	PUNCT
aiti-14589	230	67	hog	hog	NOUN
aiti-14589	230	68	-	-	PUNCT
aiti-14589	230	69	scaling	scale	VERB
aiti-14589	230	70	method	method	NOUN
aiti-14589	230	71	(	(	PUNCT
aiti-14589	230	72	accuracy	accuracy	NOUN
aiti-14589	230	73	0.8200	0.8200	NUM
aiti-14589	230	74	,	,	PUNCT
aiti-14589	230	75	precision	precision	NOUN
aiti-14589	230	76	0.8363	0.8363	NUM
aiti-14589	230	77	,	,	PUNCT
aiti-14589	230	78	recall	recall	VERB
aiti-14589	230	79	0.8200	0.8200	NUM
aiti-14589	230	80	,	,	PUNCT
aiti-14589	230	81	and	and	CCONJ
aiti-14589	230	82	f1	f1	NOUN
aiti-14589	230	83	-	-	PUNCT
aiti-14589	230	84	score	score	NOUN
aiti-14589	230	85	0.8178	0.8178	NUM
aiti-14589	230	86	)	)	PUNCT
aiti-14589	230	87	.	.	PUNCT
aiti-14589	231	1	(	(	PUNCT
aiti-14589	231	2	2	2	X
aiti-14589	231	3	)	)	PUNCT
aiti-14589	231	4	the	the	DET
aiti-14589	231	5	gbc	gbc	PROPN
aiti-14589	231	6	showed	show	VERB
aiti-14589	231	7	high	high	ADJ
aiti-14589	231	8	performance	performance	NOUN
aiti-14589	231	9	with	with	ADP
aiti-14589	231	10	the	the	DET
aiti-14589	231	11	landmark	landmark	NOUN
aiti-14589	231	12	-	-	PUNCT
aiti-14589	231	13	based	base	VERB
aiti-14589	231	14	technique	technique	NOUN
aiti-14589	231	15	(	(	PUNCT
aiti-14589	231	16	accuracy	accuracy	NOUN
aiti-14589	231	17	0.8850	0.8850	NUM
aiti-14589	231	18	,	,	PUNCT
aiti-14589	231	19	precision	precision	NOUN
aiti-14589	231	20	0.8965	0.8965	NUM
aiti-14589	231	21	,	,	PUNCT
aiti-14589	231	22	recall	recall	VERB
aiti-14589	231	23	0.8850	0.8850	NUM
aiti-14589	231	24	,	,	PUNCT
aiti-14589	231	25	and	and	CCONJ
aiti-14589	231	26	f1	f1	NOUN
aiti-14589	231	27	-	-	PUNCT
aiti-14589	231	28	score	score	NOUN
aiti-14589	231	29	0.8842	0.8842	NUM
aiti-14589	231	30	)	)	PUNCT
aiti-14589	231	31	and	and	CCONJ
aiti-14589	231	32	also	also	ADV
aiti-14589	231	33	performed	perform	VERB
aiti-14589	231	34	well	well	ADV
aiti-14589	231	35	with	with	ADP
aiti-14589	231	36	the	the	DET
aiti-14589	231	37	image	image	NOUN
aiti-14589	231	38	-	-	PUNCT
aiti-14589	231	39	hog	hog	NOUN
aiti-14589	231	40	-	-	PUNCT
aiti-14589	231	41	scaling	scale	VERB
aiti-14589	231	42	method	method	NOUN
aiti-14589	231	43	(	(	PUNCT
aiti-14589	231	44	accuracy	accuracy	NOUN
aiti-14589	231	45	0.8700	0.8700	NUM
aiti-14589	231	46	,	,	PUNCT
aiti-14589	231	47	precision	precision	NOUN
aiti-14589	231	48	0.8724	0.8724	NUM
aiti-14589	231	49	,	,	PUNCT
aiti-14589	231	50	recall	recall	VERB
aiti-14589	231	51	0.8700	0.8700	NUM
aiti-14589	231	52	,	,	PUNCT
aiti-14589	231	53	and	and	CCONJ
aiti-14589	231	54	f1	f1	NOUN
aiti-14589	231	55	-	-	PUNCT
aiti-14589	231	56	score	score	NOUN
aiti-14589	231	57	0.8698	0.8698	NUM
aiti-14589	231	58	)	)	PUNCT
aiti-14589	231	59	.	.	PUNCT
aiti-14589	232	1	(	(	PUNCT
aiti-14589	232	2	3	3	X
aiti-14589	232	3	)	)	PUNCT
aiti-14589	232	4	the	the	DET
aiti-14589	232	5	rfc	rfc	NOUN
aiti-14589	232	6	achieved	achieve	VERB
aiti-14589	232	7	strong	strong	ADJ
aiti-14589	232	8	results	result	NOUN
aiti-14589	232	9	with	with	ADP
aiti-14589	232	10	the	the	DET
aiti-14589	232	11	landmark	landmark	NOUN
aiti-14589	232	12	-	-	PUNCT
aiti-14589	232	13	based	base	VERB
aiti-14589	232	14	technique	technique	NOUN
aiti-14589	232	15	(	(	PUNCT
aiti-14589	232	16	accuracy	accuracy	NOUN
aiti-14589	232	17	0.8950	0.8950	NUM
aiti-14589	232	18	and	and	CCONJ
aiti-14589	232	19	precision	precision	NOUN
aiti-14589	232	20	0.9018	0.9018	NUM
aiti-14589	232	21	)	)	PUNCT
aiti-14589	232	22	but	but	CCONJ
aiti-14589	232	23	lower	low	ADJ
aiti-14589	232	24	performance	performance	NOUN
aiti-14589	232	25	with	with	ADP
aiti-14589	232	26	the	the	DET
aiti-14589	232	27	image	image	NOUN
aiti-14589	232	28	-	-	PUNCT
aiti-14589	232	29	hog	hog	NOUN
aiti-14589	232	30	-	-	PUNCT
aiti-14589	232	31	scaling	scale	VERB
aiti-14589	232	32	method	method	NOUN
aiti-14589	232	33	(	(	PUNCT
aiti-14589	232	34	accuracy	accuracy	NOUN
aiti-14589	232	35	0.8600	0.8600	NUM
aiti-14589	232	36	and	and	CCONJ
aiti-14589	232	37	precision	precision	NOUN
aiti-14589	232	38	0.8601	0.8601	NUM
aiti-14589	232	39	)	)	PUNCT
aiti-14589	232	40	.	.	PUNCT
aiti-14589	233	1	(	(	PUNCT
aiti-14589	233	2	4	4	X
aiti-14589	233	3	)	)	PUNCT
aiti-14589	233	4	the	the	DET
aiti-14589	233	5	lrc	lrc	NOUN
aiti-14589	233	6	was	be	AUX
aiti-14589	233	7	efficient	efficient	ADJ
aiti-14589	233	8	in	in	ADP
aiti-14589	233	9	training	training	NOUN
aiti-14589	233	10	time	time	NOUN
aiti-14589	233	11	with	with	ADP
aiti-14589	233	12	the	the	DET
aiti-14589	233	13	landmark	landmark	NOUN
aiti-14589	233	14	-	-	PUNCT
aiti-14589	233	15	based	base	VERB
aiti-14589	233	16	technique	technique	NOUN
aiti-14589	233	17	(	(	PUNCT
aiti-14589	233	18	accuracy	accuracy	NOUN
aiti-14589	233	19	0.8850	0.8850	NUM
aiti-14589	233	20	,	,	PUNCT
aiti-14589	233	21	precision	precision	NOUN
aiti-14589	233	22	0.8939	0.8939	NUM
aiti-14589	233	23	,	,	PUNCT
aiti-14589	233	24	recall	recall	NOUN
aiti-14589	233	25	0.8850	0.8850	NUM
aiti-14589	233	26	,	,	PUNCT
aiti-14589	233	27	and	and	CCONJ
aiti-14589	233	28	f1	f1	NOUN
aiti-14589	233	29	-	-	PUNCT
aiti-14589	233	30	score	score	NOUN
aiti-14589	233	31	0.8843	0.8843	NUM
aiti-14589	233	32	)	)	PUNCT
aiti-14589	233	33	but	but	CCONJ
aiti-14589	233	34	underperformed	underperform	VERB
aiti-14589	233	35	with	with	ADP
aiti-14589	233	36	the	the	DET
aiti-14589	233	37	image	image	NOUN
aiti-14589	233	38	-	-	PUNCT
aiti-14589	233	39	hog	hog	NOUN
aiti-14589	233	40	-	-	PUNCT
aiti-14589	233	41	scaling	scale	VERB
aiti-14589	233	42	method	method	NOUN
aiti-14589	233	43	(	(	PUNCT
aiti-14589	233	44	accuracy	accuracy	NOUN
aiti-14589	233	45	0.7900	0.7900	NUM
aiti-14589	233	46	,	,	PUNCT
aiti-14589	233	47	precision	precision	NOUN
aiti-14589	233	48	0.8047	0.8047	NUM
aiti-14589	233	49	,	,	PUNCT
aiti-14589	233	50	recall	recall	NOUN
aiti-14589	233	51	0.7900	0.7900	NUM
aiti-14589	233	52	,	,	PUNCT
aiti-14589	233	53	and	and	CCONJ
aiti-14589	233	54	f1	f1	NOUN
aiti-14589	233	55	-	-	PUNCT
aiti-14589	233	56	score	score	NOUN
aiti-14589	233	57	0.7874	0.7874	NUM
aiti-14589	233	58	)	)	PUNCT
aiti-14589	233	59	.	.	PUNCT
aiti-14589	234	1	(	(	PUNCT
aiti-14589	234	2	5	5	X
aiti-14589	234	3	)	)	PUNCT
aiti-14589	234	4	the	the	DET
aiti-14589	234	5	cnn	cnn	PROPN
aiti-14589	234	6	outperformed	outperform	VERB
aiti-14589	234	7	other	other	ADJ
aiti-14589	234	8	models	model	NOUN
aiti-14589	234	9	with	with	ADP
aiti-14589	234	10	the	the	DET
aiti-14589	234	11	landmark	landmark	NOUN
aiti-14589	234	12	-	-	PUNCT
aiti-14589	234	13	based	base	VERB
aiti-14589	234	14	technique	technique	NOUN
aiti-14589	234	15	(	(	PUNCT
aiti-14589	234	16	accuracy	accuracy	NOUN
aiti-14589	234	17	0.9500	0.9500	NUM
aiti-14589	234	18	,	,	PUNCT
aiti-14589	234	19	precision	precision	NOUN
aiti-14589	234	20	0.9541	0.9541	NUM
aiti-14589	234	21	,	,	PUNCT
aiti-14589	234	22	recall	recall	VERB
aiti-14589	234	23	0.9541	0.9541	NUM
aiti-14589	234	24	,	,	PUNCT
aiti-14589	234	25	and	and	CCONJ
aiti-14589	234	26	f1	f1	NOUN
aiti-14589	234	27	-	-	PUNCT
aiti-14589	234	28	score	score	NOUN
aiti-14589	234	29	0.9541	0.9541	NUM
aiti-14589	234	30	)	)	PUNCT
aiti-14589	234	31	,	,	PUNCT
aiti-14589	234	32	while	while	SCONJ
aiti-14589	234	33	the	the	DET
aiti-14589	234	34	image	image	NOUN
aiti-14589	234	35	-	-	PUNCT
aiti-14589	234	36	hog	hog	NOUN
aiti-14589	234	37	-	-	PUNCT
aiti-14589	234	38	scaling	scale	VERB
aiti-14589	234	39	method	method	NOUN
aiti-14589	234	40	also	also	ADV
aiti-14589	234	41	yielded	yield	VERB
aiti-14589	234	42	strong	strong	ADJ
aiti-14589	234	43	results	result	NOUN
aiti-14589	234	44	(	(	PUNCT
aiti-14589	234	45	accuracy	accuracy	NOUN
aiti-14589	234	46	0.8700	0.8700	NUM
aiti-14589	234	47	,	,	PUNCT
aiti-14589	234	48	precision	precision	NOUN
aiti-14589	234	49	0.8558	0.8558	NUM
aiti-14589	234	50	,	,	PUNCT
aiti-14589	234	51	recall	recall	NOUN
aiti-14589	234	52	0.8900	0.8900	NUM
aiti-14589	234	53	,	,	PUNCT
aiti-14589	234	54	and	and	CCONJ
aiti-14589	234	55	f1	f1	NOUN
aiti-14589	234	56	-	-	PUNCT
aiti-14589	234	57	score	score	NOUN
aiti-14589	234	58	0.8725	0.8725	NUM
aiti-14589	234	59	)	)	PUNCT
aiti-14589	234	60	.	.	PUNCT
aiti-14589	235	1	these	these	DET
aiti-14589	235	2	results	result	NOUN
aiti-14589	235	3	meet	meet	VERB
aiti-14589	235	4	the	the	DET
aiti-14589	235	5	objective	objective	NOUN
aiti-14589	235	6	of	of	ADP
aiti-14589	235	7	comparing	compare	VERB
aiti-14589	235	8	the	the	DET
aiti-14589	235	9	performance	performance	NOUN
aiti-14589	235	10	of	of	ADP
aiti-14589	235	11	image	image	NOUN
aiti-14589	235	12	-	-	PUNCT
aiti-14589	235	13	based	base	VERB
aiti-14589	235	14	and	and	CCONJ
aiti-14589	235	15	landmark	landmark	NOUN
aiti-14589	235	16	-	-	PUNCT
aiti-14589	235	17	based	base	VERB
aiti-14589	235	18	fer	fer	PROPN
aiti-14589	235	19	techniques	technique	NOUN
aiti-14589	235	20	,	,	PUNCT
aiti-14589	235	21	providing	provide	VERB
aiti-14589	235	22	insight	insight	NOUN
aiti-14589	235	23	into	into	ADP
aiti-14589	235	24	the	the	DET
aiti-14589	235	25	effectiveness	effectiveness	NOUN
aiti-14589	235	26	of	of	ADP
aiti-14589	235	27	different	different	ADJ
aiti-14589	235	28	ml	ml	NOUN
aiti-14589	235	29	models	model	NOUN
aiti-14589	235	30	for	for	ADP
aiti-14589	235	31	emotion	emotion	NOUN
aiti-14589	235	32	recognition	recognition	NOUN
aiti-14589	235	33	.	.	PUNCT
aiti-14589	236	1	the	the	DET
aiti-14589	236	2	findings	finding	NOUN
aiti-14589	236	3	suggest	suggest	VERB
aiti-14589	236	4	that	that	SCONJ
aiti-14589	236	5	while	while	SCONJ
aiti-14589	236	6	landmark	landmark	NOUN
aiti-14589	236	7	-	-	PUNCT
aiti-14589	236	8	based	base	VERB
aiti-14589	236	9	methods	method	NOUN
aiti-14589	236	10	,	,	PUNCT
aiti-14589	236	11	such	such	ADJ
aiti-14589	236	12	as	as	ADP
aiti-14589	236	13	those	those	PRON
aiti-14589	236	14	using	use	VERB
aiti-14589	236	15	facial	facial	ADJ
aiti-14589	236	16	feature	feature	NOUN
aiti-14589	236	17	points	point	NOUN
aiti-14589	236	18	,	,	PUNCT
aiti-14589	236	19	offer	offer	VERB
aiti-14589	236	20	higher	high	ADJ
aiti-14589	236	21	accuracy	accuracy	NOUN
aiti-14589	236	22	and	and	CCONJ
aiti-14589	236	23	precision	precision	NOUN
aiti-14589	236	24	,	,	PUNCT
aiti-14589	236	25	image	image	NOUN
aiti-14589	236	26	-	-	PUNCT
aiti-14589	236	27	based	base	VERB
aiti-14589	236	28	methods	method	NOUN
aiti-14589	236	29	like	like	ADP
aiti-14589	236	30	hog	hog	NOUN
aiti-14589	236	31	-	-	PUNCT
aiti-14589	236	32	scaling	scaling	NOUN
aiti-14589	236	33	remain	remain	VERB
aiti-14589	236	34	viable	viable	ADJ
aiti-14589	236	35	alternatives	alternative	NOUN
aiti-14589	236	36	,	,	PUNCT
aiti-14589	236	37	although	although	SCONJ
aiti-14589	236	38	slightly	slightly	ADV
aiti-14589	236	39	lower	low	ADJ
aiti-14589	236	40	in	in	ADP
aiti-14589	236	41	performance	performance	NOUN
aiti-14589	236	42	.	.	PUNCT
aiti-14589	237	1	this	this	DET
aiti-14589	237	2	comparison	comparison	NOUN
aiti-14589	237	3	informs	inform	VERB
aiti-14589	237	4	future	future	ADJ
aiti-14589	237	5	work	work	NOUN
aiti-14589	237	6	on	on	ADP
aiti-14589	237	7	optimizing	optimize	VERB
aiti-14589	237	8	emotion	emotion	NOUN
aiti-14589	237	9	recognition	recognition	NOUN
aiti-14589	237	10	systems	system	NOUN
aiti-14589	237	11	for	for	ADP
aiti-14589	237	12	real	real	ADJ
aiti-14589	237	13	-	-	PUNCT
aiti-14589	237	14	world	world	NOUN
aiti-14589	237	15	applications	application	NOUN
aiti-14589	237	16	across	across	ADP
aiti-14589	237	17	various	various	ADJ
aiti-14589	237	18	datasets	dataset	NOUN
aiti-14589	237	19	and	and	CCONJ
aiti-14589	237	20	conditions	condition	NOUN
aiti-14589	237	21	.	.	PUNCT
aiti-14589	238	1	6	6	X
aiti-14589	238	2	.	.	X
aiti-14589	238	3	discussion	discussion	NOUN
aiti-14589	238	4	this	this	DET
aiti-14589	238	5	study	study	NOUN
aiti-14589	238	6	emphasizes	emphasize	VERB
aiti-14589	238	7	the	the	DET
aiti-14589	238	8	critical	critical	ADJ
aiti-14589	238	9	role	role	NOUN
aiti-14589	238	10	of	of	ADP
aiti-14589	238	11	selecting	select	VERB
aiti-14589	238	12	appropriate	appropriate	ADJ
aiti-14589	238	13	ml	ml	NOUN
aiti-14589	238	14	models	model	NOUN
aiti-14589	238	15	and	and	CCONJ
aiti-14589	238	16	feature	feature	NOUN
aiti-14589	238	17	extraction	extraction	NOUN
aiti-14589	238	18	methods	method	NOUN
aiti-14589	238	19	to	to	PART
aiti-14589	238	20	achieve	achieve	VERB
aiti-14589	238	21	optimal	optimal	ADJ
aiti-14589	238	22	classification	classification	NOUN
aiti-14589	238	23	outcomes	outcome	NOUN
aiti-14589	238	24	.	.	PUNCT
aiti-14589	239	1	the	the	DET
aiti-14589	239	2	performance	performance	NOUN
aiti-14589	239	3	of	of	ADP
aiti-14589	239	4	each	each	DET
aiti-14589	239	5	model	model	NOUN
aiti-14589	239	6	is	be	AUX
aiti-14589	239	7	summarized	summarize	VERB
aiti-14589	239	8	as	as	SCONJ
aiti-14589	239	9	follows	follow	VERB
aiti-14589	239	10	:	:	PUNCT
aiti-14589	239	11	(	(	PUNCT
aiti-14589	239	12	1	1	X
aiti-14589	239	13	)	)	PUNCT
aiti-14589	239	14	the	the	DET
aiti-14589	239	15	svm	svm	PROPN
aiti-14589	239	16	model	model	NOUN
aiti-14589	239	17	exhibited	exhibit	VERB
aiti-14589	239	18	sensitivity	sensitivity	NOUN
aiti-14589	239	19	to	to	ADP
aiti-14589	239	20	the	the	DET
aiti-14589	239	21	feature	feature	NOUN
aiti-14589	239	22	extraction	extraction	NOUN
aiti-14589	239	23	techniques	technique	NOUN
aiti-14589	239	24	employed	employ	VERB
aiti-14589	239	25	,	,	PUNCT
aiti-14589	239	26	with	with	ADP
aiti-14589	239	27	the	the	DET
aiti-14589	239	28	landmark	landmark	NOUN
aiti-14589	239	29	-	-	PUNCT
aiti-14589	239	30	based	base	VERB
aiti-14589	239	31	approach	approach	NOUN
aiti-14589	239	32	significantly	significantly	ADV
aiti-14589	239	33	outperforming	outperform	VERB
aiti-14589	239	34	the	the	DET
aiti-14589	239	35	image	image	NOUN
aiti-14589	239	36	-	-	PUNCT
aiti-14589	239	37	based	base	VERB
aiti-14589	239	38	method	method	NOUN
aiti-14589	239	39	.	.	PUNCT
aiti-14589	240	1	this	this	DET
aiti-14589	240	2	finding	finding	NOUN
aiti-14589	240	3	underscores	underscore	VERB
aiti-14589	240	4	the	the	DET
aiti-14589	240	5	importance	importance	NOUN
aiti-14589	240	6	of	of	ADP
aiti-14589	240	7	landmark	landmark	NOUN
aiti-14589	240	8	features	feature	NOUN
aiti-14589	240	9	in	in	ADP
aiti-14589	240	10	enhancing	enhance	VERB
aiti-14589	240	11	the	the	DET
aiti-14589	240	12	svm	svm	PROPN
aiti-14589	240	13	’s	’s	PART
aiti-14589	240	14	ability	ability	NOUN
aiti-14589	240	15	to	to	PART
aiti-14589	240	16	identify	identify	VERB
aiti-14589	240	17	complex	complex	ADJ
aiti-14589	240	18	patterns	pattern	NOUN
aiti-14589	240	19	in	in	ADP
aiti-14589	240	20	facial	facial	ADJ
aiti-14589	240	21	expression	expression	NOUN
aiti-14589	240	22	data	datum	NOUN
aiti-14589	240	23	,	,	PUNCT
aiti-14589	240	24	corroborating	corroborate	VERB
aiti-14589	240	25	the	the	DET
aiti-14589	240	26	conclusions	conclusion	NOUN
aiti-14589	240	27	of	of	ADP
aiti-14589	240	28	both	both	DET
aiti-14589	240	29	sharma	sharma	PROPN
aiti-14589	240	30	et	et	PROPN
aiti-14589	240	31	al	al	PROPN
aiti-14589	240	32	.	.	PUNCT
aiti-14589	241	1	[	[	X
aiti-14589	241	2	9	9	NUM
aiti-14589	241	3	]	]	PUNCT
aiti-14589	241	4	and	and	CCONJ
aiti-14589	241	5	michael	michael	PROPN
aiti-14589	241	6	revina	revina	PROPN
aiti-14589	241	7	and	and	CCONJ
aiti-14589	241	8	sam	sam	PROPN
aiti-14589	241	9	emmanuel	emmanuel	PROPN
aiti-14589	242	1	[	[	X
aiti-14589	242	2	1	1	NUM
aiti-14589	242	3	]	]	PUNCT
aiti-14589	242	4	.	.	PUNCT
aiti-14589	243	1	(	(	PUNCT
aiti-14589	243	2	2	2	X
aiti-14589	243	3	)	)	PUNCT
aiti-14589	243	4	the	the	DET
aiti-14589	243	5	gbc	gbc	PROPN
aiti-14589	243	6	demonstrated	demonstrate	VERB
aiti-14589	243	7	robust	robust	ADJ
aiti-14589	243	8	performance	performance	NOUN
aiti-14589	243	9	across	across	ADP
aiti-14589	243	10	both	both	DET
aiti-14589	243	11	feature	feature	NOUN
aiti-14589	243	12	extraction	extraction	NOUN
aiti-14589	243	13	techniques	technique	NOUN
aiti-14589	243	14	,	,	PUNCT
aiti-14589	243	15	indicating	indicate	VERB
aiti-14589	243	16	its	its	PRON
aiti-14589	243	17	adaptability	adaptability	NOUN
aiti-14589	243	18	to	to	PART
aiti-14589	243	19	diverse	diverse	VERB
aiti-14589	243	20	data	datum	NOUN
aiti-14589	243	21	representations	representation	NOUN
aiti-14589	243	22	.	.	PUNCT
aiti-14589	244	1	although	although	SCONJ
aiti-14589	244	2	the	the	DET
aiti-14589	244	3	training	training	NOUN
aiti-14589	244	4	duration	duration	NOUN
aiti-14589	244	5	was	be	AUX
aiti-14589	244	6	longer	long	ADJ
aiti-14589	244	7	,	,	PUNCT
aiti-14589	244	8	particularly	particularly	ADV
aiti-14589	244	9	when	when	SCONJ
aiti-14589	244	10	using	use	VERB
aiti-14589	244	11	the	the	DET
aiti-14589	244	12	landmark	landmark	NOUN
aiti-14589	244	13	-	-	PUNCT
aiti-14589	244	14	based	base	VERB
aiti-14589	244	15	method	method	NOUN
aiti-14589	244	16	,	,	PUNCT
aiti-14589	244	17	the	the	DET
aiti-14589	244	18	gbc	gbc	PROPN
aiti-14589	244	19	’s	’s	PART
aiti-14589	244	20	superior	superior	ADJ
aiti-14589	244	21	accuracy	accuracy	NOUN
aiti-14589	244	22	,	,	PUNCT
aiti-14589	244	23	precision	precision	NOUN
aiti-14589	244	24	,	,	PUNCT
aiti-14589	244	25	recall	recall	NOUN
aiti-14589	244	26	,	,	PUNCT
aiti-14589	244	27	and	and	CCONJ
aiti-14589	244	28	f1	f1	NOUN
aiti-14589	244	29	-	-	PUNCT
aiti-14589	244	30	scores	score	NOUN
aiti-14589	244	31	justified	justify	VERB
aiti-14589	244	32	the	the	DET
aiti-14589	244	33	extended	extended	ADJ
aiti-14589	244	34	training	training	NOUN
aiti-14589	244	35	period	period	NOUN
aiti-14589	244	36	.	.	PUNCT
aiti-14589	245	1	this	this	DET
aiti-14589	245	2	finding	finding	NOUN
aiti-14589	245	3	aligns	align	VERB
aiti-14589	245	4	with	with	ADP
aiti-14589	245	5	di	di	NOUN
aiti-14589	245	6	luzio	luzio	PROPN
aiti-14589	245	7	et	et	PROPN
aiti-14589	245	8	al	al	PROPN
aiti-14589	245	9	.	.	PUNCT
aiti-14589	246	1	[	[	X
aiti-14589	246	2	27	27	NUM
aiti-14589	246	3	]	]	PUNCT
aiti-14589	246	4	,	,	PUNCT
aiti-14589	246	5	who	who	PRON
aiti-14589	246	6	highlighted	highlight	VERB
aiti-14589	246	7	gbc	gbc	PROPN
aiti-14589	246	8	’s	’s	PART
aiti-14589	246	9	efficacy	efficacy	NOUN
aiti-14589	246	10	in	in	ADP
aiti-14589	246	11	managing	manage	VERB
aiti-14589	246	12	complex	complex	ADJ
aiti-14589	246	13	datasets	dataset	NOUN
aiti-14589	246	14	.	.	PUNCT
aiti-14589	247	1	(	(	PUNCT
aiti-14589	247	2	3	3	X
aiti-14589	247	3	)	)	PUNCT
aiti-14589	247	4	the	the	DET
aiti-14589	247	5	rfc	rfc	NOUN
aiti-14589	247	6	displayed	display	VERB
aiti-14589	247	7	consistent	consistent	ADJ
aiti-14589	247	8	performance	performance	NOUN
aiti-14589	247	9	across	across	ADP
aiti-14589	247	10	both	both	DET
aiti-14589	247	11	feature	feature	NOUN
aiti-14589	247	12	extraction	extraction	NOUN
aiti-14589	247	13	approaches	approach	NOUN
aiti-14589	247	14	,	,	PUNCT
aiti-14589	247	15	achieving	achieve	VERB
aiti-14589	247	16	high	high	ADJ
aiti-14589	247	17	precision	precision	NOUN
aiti-14589	247	18	,	,	PUNCT
aiti-14589	247	19	recall	recall	NOUN
aiti-14589	247	20	,	,	PUNCT
aiti-14589	247	21	and	and	CCONJ
aiti-14589	247	22	f1	f1	NOUN
aiti-14589	247	23	-	-	PUNCT
aiti-14589	247	24	scores	score	NOUN
aiti-14589	247	25	while	while	SCONJ
aiti-14589	247	26	requiring	require	VERB
aiti-14589	247	27	relatively	relatively	ADV
aiti-14589	247	28	short	short	ADJ
aiti-14589	247	29	training	training	NOUN
aiti-14589	247	30	times	time	NOUN
aiti-14589	247	31	.	.	PUNCT
aiti-14589	248	1	this	this	DET
aiti-14589	248	2	combination	combination	NOUN
aiti-14589	248	3	of	of	ADP
aiti-14589	248	4	efficiency	efficiency	NOUN
aiti-14589	248	5	and	and	CCONJ
aiti-14589	248	6	effectiveness	effectiveness	NOUN
aiti-14589	248	7	positions	position	NOUN
aiti-14589	248	8	rfc	rfc	PRON
aiti-14589	248	9	as	as	ADP
aiti-14589	248	10	a	a	DET
aiti-14589	248	11	reliable	reliable	ADJ
aiti-14589	248	12	model	model	NOUN
aiti-14589	248	13	for	for	ADP
aiti-14589	248	14	classification	classification	NOUN
aiti-14589	248	15	tasks	task	NOUN
aiti-14589	248	16	,	,	PUNCT
aiti-14589	248	17	echoing	echo	VERB
aiti-14589	248	18	the	the	DET
aiti-14589	248	19	results	result	NOUN
aiti-14589	248	20	observed	observe	VERB
aiti-14589	248	21	by	by	ADP
aiti-14589	248	22	munasinghe	munasinghe	NOUN
aiti-14589	248	23	[	[	X
aiti-14589	248	24	26	26	NUM
aiti-14589	248	25	]	]	PUNCT
aiti-14589	248	26	and	and	CCONJ
aiti-14589	248	27	canedo	canedo	PROPN
aiti-14589	248	28	and	and	CCONJ
aiti-14589	248	29	neves	neve	NOUN
aiti-14589	248	30	[	[	X
aiti-14589	248	31	2	2	NUM
aiti-14589	248	32	]	]	PUNCT
aiti-14589	248	33	,	,	PUNCT
aiti-14589	248	34	who	who	PRON
aiti-14589	248	35	demonstrated	demonstrate	VERB
aiti-14589	248	36	rfc	rfc	PROPN
aiti-14589	248	37	’s	’s	PART
aiti-14589	248	38	reliability	reliability	NOUN
aiti-14589	248	39	in	in	ADP
aiti-14589	248	40	image	image	NOUN
aiti-14589	248	41	classification	classification	NOUN
aiti-14589	248	42	.	.	PUNCT
aiti-14589	249	1	(	(	PUNCT
aiti-14589	249	2	4	4	X
aiti-14589	249	3	)	)	PUNCT
aiti-14589	249	4	the	the	DET
aiti-14589	249	5	lrc	lrc	PROPN
aiti-14589	249	6	exhibited	exhibit	VERB
aiti-14589	249	7	remarkable	remarkable	ADJ
aiti-14589	249	8	efficiency	efficiency	NOUN
aiti-14589	249	9	in	in	ADP
aiti-14589	249	10	training	training	NOUN
aiti-14589	249	11	time	time	NOUN
aiti-14589	249	12	,	,	PUNCT
aiti-14589	249	13	especially	especially	ADV
aiti-14589	249	14	with	with	ADP
aiti-14589	249	15	the	the	DET
aiti-14589	249	16	landmark	landmark	NOUN
aiti-14589	249	17	-	-	PUNCT
aiti-14589	249	18	based	base	VERB
aiti-14589	249	19	technique	technique	NOUN
aiti-14589	249	20	,	,	PUNCT
aiti-14589	249	21	where	where	SCONJ
aiti-14589	249	22	training	training	NOUN
aiti-14589	249	23	required	require	VERB
aiti-14589	249	24	only	only	ADV
aiti-14589	249	25	0.11	0.11	NUM
aiti-14589	249	26	seconds	second	NOUN
aiti-14589	249	27	.	.	PUNCT
aiti-14589	250	1	however	however	ADV
aiti-14589	250	2	,	,	PUNCT
aiti-14589	250	3	the	the	DET
aiti-14589	250	4	notable	notable	ADJ
aiti-14589	250	5	decline	decline	NOUN
aiti-14589	250	6	in	in	ADP
aiti-14589	250	7	performance	performance	NOUN
aiti-14589	250	8	with	with	ADP
aiti-14589	250	9	the	the	DET
aiti-14589	250	10	image	image	NOUN
aiti-14589	250	11	-	-	PUNCT
aiti-14589	250	12	based	base	VERB
aiti-14589	250	13	method	method	NOUN
aiti-14589	250	14	suggests	suggest	VERB
aiti-14589	250	15	that	that	SCONJ
aiti-14589	250	16	lrc	lrc	PROPN
aiti-14589	250	17	is	be	AUX
aiti-14589	250	18	more	more	ADV
aiti-14589	250	19	sensitive	sensitive	ADJ
aiti-14589	250	20	to	to	ADP
aiti-14589	250	21	the	the	DET
aiti-14589	250	22	quality	quality	NOUN
aiti-14589	250	23	of	of	ADP
aiti-14589	250	24	feature	feature	NOUN
aiti-14589	250	25	selection	selection	NOUN
aiti-14589	250	26	.	.	PUNCT
aiti-14589	251	1	this	this	PRON
aiti-14589	251	2	aligns	align	VERB
aiti-14589	251	3	with	with	ADP
aiti-14589	251	4	findings	finding	NOUN
aiti-14589	251	5	by	by	ADP
aiti-14589	251	6	kanjanawattana	kanjanawattana	PROPN
aiti-14589	251	7	et	et	PROPN
aiti-14589	251	8	al	al	PROPN
aiti-14589	251	9	.	.	PUNCT
aiti-14589	252	1	[	[	X
aiti-14589	252	2	3	3	NUM
aiti-14589	252	3	]	]	PUNCT
aiti-14589	252	4	,	,	PUNCT
aiti-14589	252	5	who	who	PRON
aiti-14589	252	6	identified	identify	VERB
aiti-14589	252	7	limitations	limitation	NOUN
aiti-14589	252	8	in	in	ADP
aiti-14589	252	9	the	the	DET
aiti-14589	252	10	use	use	NOUN
aiti-14589	252	11	of	of	ADP
aiti-14589	252	12	lrc	lrc	NOUN
aiti-14589	252	13	for	for	ADP
aiti-14589	252	14	more	more	ADJ
aiti-14589	252	15	complex	complex	ADJ
aiti-14589	252	16	classification	classification	NOUN
aiti-14589	252	17	tasks	task	NOUN
aiti-14589	252	18	.	.	PUNCT
aiti-14589	253	1	advances	advance	NOUN
aiti-14589	253	2	in	in	ADP
aiti-14589	253	3	technology	technology	NOUN
aiti-14589	253	4	innovation	innovation	NOUN
aiti-14589	253	5	,	,	PUNCT
aiti-14589	253	6	vol	vol	NOUN
aiti-14589	253	7	.	.	PROPN
aiti-14589	254	1	10	10	NUM
aiti-14589	254	2	,	,	PUNCT
aiti-14589	254	3	no	no	INTJ
aiti-14589	254	4	.	.	NOUN
aiti-14589	254	5	4	4	NUM
aiti-14589	254	6	,	,	PUNCT
aiti-14589	254	7	2025	2025	NUM
aiti-14589	254	8	,	,	PUNCT
aiti-14589	254	9	pp	pp	ADJ
aiti-14589	254	10	.	.	PUNCT
aiti-14589	255	1	395	395	NUM
aiti-14589	255	2	-	-	SYM
aiti-14589	255	3	406	406	NUM
aiti-14589	255	4	405	405	NUM
aiti-14589	255	5	(	(	PUNCT
aiti-14589	255	6	5	5	NUM
aiti-14589	255	7	)	)	PUNCT
aiti-14589	255	8	the	the	DET
aiti-14589	255	9	cnn	cnn	PROPN
aiti-14589	255	10	emerged	emerge	VERB
aiti-14589	255	11	as	as	ADP
aiti-14589	255	12	the	the	DET
aiti-14589	255	13	top	top	ADV
aiti-14589	255	14	-	-	PUNCT
aiti-14589	255	15	performing	perform	VERB
aiti-14589	255	16	model	model	NOUN
aiti-14589	255	17	,	,	PUNCT
aiti-14589	255	18	particularly	particularly	ADV
aiti-14589	255	19	with	with	ADP
aiti-14589	255	20	the	the	DET
aiti-14589	255	21	landmark	landmark	NOUN
aiti-14589	255	22	technique	technique	NOUN
aiti-14589	255	23	,	,	PUNCT
aiti-14589	255	24	achieving	achieve	VERB
aiti-14589	255	25	superior	superior	ADJ
aiti-14589	255	26	accuracy	accuracy	NOUN
aiti-14589	255	27	,	,	PUNCT
aiti-14589	255	28	precision	precision	NOUN
aiti-14589	255	29	,	,	PUNCT
aiti-14589	255	30	recall	recall	NOUN
aiti-14589	255	31	,	,	PUNCT
aiti-14589	255	32	and	and	CCONJ
aiti-14589	255	33	f1	f1	NOUN
aiti-14589	255	34	-	-	PUNCT
aiti-14589	255	35	scores	score	NOUN
aiti-14589	255	36	.	.	PUNCT
aiti-14589	256	1	despite	despite	SCONJ
aiti-14589	256	2	the	the	DET
aiti-14589	256	3	longer	long	ADJ
aiti-14589	256	4	training	training	NOUN
aiti-14589	256	5	time	time	NOUN
aiti-14589	256	6	,	,	PUNCT
aiti-14589	256	7	particularly	particularly	ADV
aiti-14589	256	8	with	with	ADP
aiti-14589	256	9	the	the	DET
aiti-14589	256	10	image	image	NOUN
aiti-14589	256	11	-	-	PUNCT
aiti-14589	256	12	based	base	VERB
aiti-14589	256	13	technique	technique	NOUN
aiti-14589	256	14	(	(	PUNCT
aiti-14589	256	15	88.89	88.89	NUM
aiti-14589	256	16	seconds	second	NOUN
aiti-14589	256	17	)	)	PUNCT
aiti-14589	256	18	,	,	PUNCT
aiti-14589	256	19	the	the	DET
aiti-14589	256	20	substantial	substantial	ADJ
aiti-14589	256	21	improvements	improvement	NOUN
aiti-14589	256	22	in	in	ADP
aiti-14589	256	23	key	key	ADJ
aiti-14589	256	24	performance	performance	NOUN
aiti-14589	256	25	metrics	metric	NOUN
aiti-14589	256	26	justify	justify	VERB
aiti-14589	256	27	the	the	DET
aiti-14589	256	28	increased	increase	VERB
aiti-14589	256	29	computational	computational	ADJ
aiti-14589	256	30	cost	cost	NOUN
aiti-14589	256	31	.	.	PUNCT
aiti-14589	257	1	the	the	DET
aiti-14589	257	2	cnn	cnn	PROPN
aiti-14589	257	3	’s	’s	PART
aiti-14589	257	4	ability	ability	NOUN
aiti-14589	257	5	to	to	PART
aiti-14589	257	6	automatically	automatically	ADV
aiti-14589	257	7	learn	learn	VERB
aiti-14589	257	8	hierarchical	hierarchical	ADJ
aiti-14589	257	9	features	feature	NOUN
aiti-14589	257	10	from	from	ADP
aiti-14589	257	11	raw	raw	ADJ
aiti-14589	257	12	image	image	NOUN
aiti-14589	257	13	data	datum	NOUN
aiti-14589	257	14	positions	position	NOUN
aiti-14589	257	15	it	it	PRON
aiti-14589	257	16	as	as	ADP
aiti-14589	257	17	an	an	DET
aiti-14589	257	18	effective	effective	ADJ
aiti-14589	257	19	choice	choice	NOUN
aiti-14589	257	20	for	for	ADP
aiti-14589	257	21	fer	fer	NOUN
aiti-14589	257	22	,	,	PUNCT
aiti-14589	257	23	as	as	SCONJ
aiti-14589	257	24	demonstrated	demonstrate	VERB
aiti-14589	257	25	by	by	ADP
aiti-14589	257	26	hilbe	hilbe	NOUN
aiti-14589	257	27	[	[	X
aiti-14589	257	28	13	13	NUM
aiti-14589	257	29	]	]	PUNCT
aiti-14589	257	30	and	and	CCONJ
aiti-14589	257	31	bodini	bodini	VERB
aiti-14589	257	32	[	[	X
aiti-14589	257	33	11	11	NUM
aiti-14589	257	34	]	]	PUNCT
aiti-14589	257	35	.	.	PUNCT
aiti-14589	258	1	in	in	ADP
aiti-14589	258	2	summary	summary	NOUN
aiti-14589	258	3	,	,	PUNCT
aiti-14589	258	4	these	these	DET
aiti-14589	258	5	findings	finding	NOUN
aiti-14589	258	6	emphasize	emphasize	VERB
aiti-14589	258	7	the	the	DET
aiti-14589	258	8	importance	importance	NOUN
aiti-14589	258	9	of	of	ADP
aiti-14589	258	10	careful	careful	ADJ
aiti-14589	258	11	model	model	NOUN
aiti-14589	258	12	and	and	CCONJ
aiti-14589	258	13	feature	feature	NOUN
aiti-14589	258	14	selection	selection	NOUN
aiti-14589	258	15	.	.	PUNCT
aiti-14589	259	1	the	the	DET
aiti-14589	259	2	researcher	researcher	NOUN
aiti-14589	259	3	highlights	highlight	VERB
aiti-14589	259	4	the	the	DET
aiti-14589	259	5	trade	trade	NOUN
aiti-14589	259	6	-	-	PUNCT
aiti-14589	259	7	offs	off	NOUN
aiti-14589	259	8	between	between	ADP
aiti-14589	259	9	computational	computational	ADJ
aiti-14589	259	10	time	time	NOUN
aiti-14589	259	11	and	and	CCONJ
aiti-14589	259	12	performance	performance	NOUN
aiti-14589	259	13	.	.	PUNCT
aiti-14589	260	1	for	for	ADP
aiti-14589	260	2	higher	high	ADJ
aiti-14589	260	3	accuracy	accuracy	NOUN
aiti-14589	260	4	,	,	PUNCT
aiti-14589	260	5	computationally	computationally	ADV
aiti-14589	260	6	intensive	intensive	ADJ
aiti-14589	260	7	models	model	NOUN
aiti-14589	260	8	like	like	ADP
aiti-14589	260	9	cnn	cnn	PROPN
aiti-14589	260	10	,	,	PUNCT
aiti-14589	260	11	particularly	particularly	ADV
aiti-14589	260	12	when	when	SCONJ
aiti-14589	260	13	paired	pair	VERB
aiti-14589	260	14	with	with	ADP
aiti-14589	260	15	landmark	landmark	NOUN
aiti-14589	260	16	-	-	PUNCT
aiti-14589	260	17	based	base	VERB
aiti-14589	260	18	features	feature	NOUN
aiti-14589	260	19	,	,	PUNCT
aiti-14589	260	20	offer	offer	VERB
aiti-14589	260	21	substantial	substantial	ADJ
aiti-14589	260	22	benefits	benefit	NOUN
aiti-14589	260	23	.	.	PUNCT
aiti-14589	261	1	7	7	X
aiti-14589	261	2	.	.	X
aiti-14589	261	3	conclusions	conclusion	NOUN
aiti-14589	261	4	this	this	DET
aiti-14589	261	5	study	study	NOUN
aiti-14589	261	6	compared	compare	VERB
aiti-14589	261	7	five	five	NUM
aiti-14589	261	8	machine	machine	NOUN
aiti-14589	261	9	learning	learning	NOUN
aiti-14589	261	10	models	model	NOUN
aiti-14589	261	11	(	(	PUNCT
aiti-14589	261	12	cnn	cnn	PROPN
aiti-14589	261	13	,	,	PUNCT
aiti-14589	261	14	svm	svm	PROPN
aiti-14589	261	15	,	,	PUNCT
aiti-14589	261	16	rfc	rfc	PROPN
aiti-14589	261	17	,	,	PUNCT
aiti-14589	261	18	lrc	lrc	PROPN
aiti-14589	261	19	,	,	PUNCT
aiti-14589	261	20	and	and	CCONJ
aiti-14589	261	21	gbc	gbc	PROPN
aiti-14589	261	22	)	)	PUNCT
aiti-14589	261	23	for	for	ADP
aiti-14589	261	24	recognizing	recognize	VERB
aiti-14589	261	25	facial	facial	ADJ
aiti-14589	261	26	expressions	expression	NOUN
aiti-14589	261	27	.	.	PUNCT
aiti-14589	262	1	both	both	PRON
aiti-14589	262	2	image	image	NOUN
aiti-14589	262	3	-	-	PUNCT
aiti-14589	262	4	based	base	VERB
aiti-14589	262	5	hog	hog	NOUN
aiti-14589	262	6	features	feature	NOUN
aiti-14589	262	7	and	and	CCONJ
aiti-14589	262	8	landmark	landmark	NOUN
aiti-14589	262	9	-	-	PUNCT
aiti-14589	262	10	based	base	VERB
aiti-14589	262	11	methods	method	NOUN
aiti-14589	262	12	were	be	AUX
aiti-14589	262	13	evaluated	evaluate	VERB
aiti-14589	262	14	to	to	PART
aiti-14589	262	15	classify	classify	VERB
aiti-14589	262	16	normal	normal	ADJ
aiti-14589	262	17	and	and	CCONJ
aiti-14589	262	18	hurt	hurt	VERB
aiti-14589	262	19	expressions	expression	NOUN
aiti-14589	262	20	using	use	VERB
aiti-14589	262	21	data	datum	NOUN
aiti-14589	262	22	from	from	ADP
aiti-14589	262	23	the	the	DET
aiti-14589	262	24	delaware	delaware	PROPN
aiti-14589	262	25	pain	pain	PROPN
aiti-14589	262	26	database	database	NOUN
aiti-14589	262	27	and	and	CCONJ
aiti-14589	262	28	utkface	utkface	NOUN
aiti-14589	262	29	.	.	PUNCT
aiti-14589	263	1	the	the	DET
aiti-14589	263	2	main	main	ADJ
aiti-14589	263	3	findings	finding	NOUN
aiti-14589	263	4	from	from	ADP
aiti-14589	263	5	this	this	DET
aiti-14589	263	6	research	research	NOUN
aiti-14589	263	7	are	be	AUX
aiti-14589	263	8	:	:	PUNCT
aiti-14589	263	9	(	(	PUNCT
aiti-14589	263	10	1	1	X
aiti-14589	263	11	)	)	PUNCT
aiti-14589	263	12	landmark	landmark	NOUN
aiti-14589	263	13	-	-	PUNCT
aiti-14589	263	14	based	base	VERB
aiti-14589	263	15	methods	method	NOUN
aiti-14589	263	16	consistently	consistently	ADV
aiti-14589	263	17	performed	perform	VERB
aiti-14589	263	18	better	well	ADV
aiti-14589	263	19	than	than	ADP
aiti-14589	263	20	image	image	NOUN
aiti-14589	263	21	-	-	PUNCT
aiti-14589	263	22	based	base	VERB
aiti-14589	263	23	hog	hog	NOUN
aiti-14589	263	24	methods	method	NOUN
aiti-14589	263	25	across	across	ADP
aiti-14589	263	26	all	all	DET
aiti-14589	263	27	models	model	NOUN
aiti-14589	263	28	,	,	PUNCT
aiti-14589	263	29	demonstrating	demonstrate	VERB
aiti-14589	263	30	greater	great	ADJ
aiti-14589	263	31	effectiveness	effectiveness	NOUN
aiti-14589	263	32	in	in	ADP
aiti-14589	263	33	distinguishing	distinguish	VERB
aiti-14589	263	34	facial	facial	ADJ
aiti-14589	263	35	expressions	expression	NOUN
aiti-14589	263	36	.	.	PUNCT
aiti-14589	264	1	(	(	PUNCT
aiti-14589	264	2	2	2	X
aiti-14589	264	3	)	)	PUNCT
aiti-14589	264	4	cnn	cnn	PROPN
aiti-14589	264	5	achieved	achieve	VERB
aiti-14589	264	6	the	the	DET
aiti-14589	264	7	best	good	ADJ
aiti-14589	264	8	results	result	NOUN
aiti-14589	264	9	with	with	ADP
aiti-14589	264	10	landmark	landmark	NOUN
aiti-14589	264	11	-	-	PUNCT
aiti-14589	264	12	based	base	VERB
aiti-14589	264	13	features	feature	NOUN
aiti-14589	264	14	,	,	PUNCT
aiti-14589	264	15	reaching	reach	VERB
aiti-14589	264	16	95	95	NUM
aiti-14589	264	17	%	%	NOUN
aiti-14589	264	18	accuracy	accuracy	NOUN
aiti-14589	264	19	,	,	PUNCT
aiti-14589	264	20	95.41	95.41	NUM
aiti-14589	264	21	%	%	NOUN
aiti-14589	264	22	precision	precision	NOUN
aiti-14589	264	23	,	,	PUNCT
aiti-14589	264	24	recall	recall	NOUN
aiti-14589	264	25	,	,	PUNCT
aiti-14589	264	26	and	and	CCONJ
aiti-14589	264	27	f1score	f1score	ADV
aiti-14589	264	28	,	,	PUNCT
aiti-14589	264	29	making	make	VERB
aiti-14589	264	30	it	it	PRON
aiti-14589	264	31	the	the	DET
aiti-14589	264	32	most	most	ADV
aiti-14589	264	33	effective	effective	ADJ
aiti-14589	264	34	method	method	NOUN
aiti-14589	264	35	for	for	ADP
aiti-14589	264	36	applications	application	NOUN
aiti-14589	264	37	requiring	require	VERB
aiti-14589	264	38	high	high	ADJ
aiti-14589	264	39	accuracy	accuracy	NOUN
aiti-14589	264	40	.	.	PUNCT
aiti-14589	265	1	(	(	PUNCT
aiti-14589	265	2	3	3	X
aiti-14589	265	3	)	)	PUNCT
aiti-14589	265	4	traditional	traditional	ADJ
aiti-14589	265	5	machine	machine	NOUN
aiti-14589	265	6	learning	learning	NOUN
aiti-14589	265	7	models	model	NOUN
aiti-14589	265	8	(	(	PUNCT
aiti-14589	265	9	svm	svm	PROPN
aiti-14589	265	10	,	,	PUNCT
aiti-14589	265	11	rfc	rfc	PROPN
aiti-14589	265	12	,	,	PUNCT
aiti-14589	265	13	gbc	gbc	PROPN
aiti-14589	265	14	)	)	PUNCT
aiti-14589	265	15	showed	show	VERB
aiti-14589	265	16	strong	strong	ADJ
aiti-14589	265	17	performance	performance	NOUN
aiti-14589	265	18	with	with	ADP
aiti-14589	265	19	landmark	landmark	NOUN
aiti-14589	265	20	features	feature	NOUN
aiti-14589	265	21	,	,	PUNCT
aiti-14589	265	22	achieving	achieve	VERB
aiti-14589	265	23	accuracies	accuracy	NOUN
aiti-14589	265	24	between	between	ADP
aiti-14589	265	25	85.5	85.5	NUM
aiti-14589	265	26	%	%	NOUN
aiti-14589	265	27	and	and	CCONJ
aiti-14589	265	28	89.5	89.5	NUM
aiti-14589	265	29	%	%	NOUN
aiti-14589	265	30	,	,	PUNCT
aiti-14589	265	31	making	make	VERB
aiti-14589	265	32	them	they	PRON
aiti-14589	265	33	good	good	ADJ
aiti-14589	265	34	alternatives	alternative	NOUN
aiti-14589	265	35	when	when	SCONJ
aiti-14589	265	36	computing	computing	NOUN
aiti-14589	265	37	resources	resource	NOUN
aiti-14589	265	38	are	be	AUX
aiti-14589	265	39	limited	limited	ADJ
aiti-14589	265	40	.	.	PUNCT
aiti-14589	266	1	(	(	PUNCT
aiti-14589	266	2	4	4	X
aiti-14589	266	3	)	)	PUNCT
aiti-14589	266	4	large	large	ADJ
aiti-14589	266	5	performance	performance	NOUN
aiti-14589	266	6	differences	difference	NOUN
aiti-14589	266	7	existed	exist	VERB
aiti-14589	266	8	between	between	ADP
aiti-14589	266	9	feature	feature	NOUN
aiti-14589	266	10	extraction	extraction	NOUN
aiti-14589	266	11	methods	method	NOUN
aiti-14589	266	12	,	,	PUNCT
aiti-14589	266	13	with	with	ADP
aiti-14589	266	14	lrc	lrc	PROPN
aiti-14589	266	15	showing	show	VERB
aiti-14589	266	16	the	the	DET
aiti-14589	266	17	biggest	big	ADJ
aiti-14589	266	18	gap	gap	NOUN
aiti-14589	266	19	(	(	PUNCT
aiti-14589	266	20	88.5	88.5	NUM
aiti-14589	266	21	%	%	NOUN
aiti-14589	266	22	vs.	vs.	ADP
aiti-14589	266	23	79	79	NUM
aiti-14589	266	24	%	%	NOUN
aiti-14589	266	25	accuracy	accuracy	NOUN
aiti-14589	266	26	for	for	ADP
aiti-14589	266	27	landmark	landmark	NOUN
aiti-14589	266	28	vs.	vs.	ADP
aiti-14589	266	29	image	image	NOUN
aiti-14589	266	30	-	-	PUNCT
aiti-14589	266	31	based	base	VERB
aiti-14589	266	32	features	feature	NOUN
aiti-14589	266	33	)	)	PUNCT
aiti-14589	266	34	.	.	PUNCT
aiti-14589	267	1	(	(	PUNCT
aiti-14589	267	2	5	5	X
aiti-14589	267	3	)	)	PUNCT
aiti-14589	267	4	analysis	analysis	NOUN
aiti-14589	267	5	of	of	ADP
aiti-14589	267	6	computing	compute	VERB
aiti-14589	267	7	efficiency	efficiency	NOUN
aiti-14589	267	8	showed	show	VERB
aiti-14589	267	9	trade	trade	NOUN
aiti-14589	267	10	-	-	PUNCT
aiti-14589	267	11	offs	off	NOUN
aiti-14589	267	12	between	between	ADP
aiti-14589	267	13	accuracy	accuracy	NOUN
aiti-14589	267	14	and	and	CCONJ
aiti-14589	267	15	training	training	NOUN
aiti-14589	267	16	time	time	NOUN
aiti-14589	267	17	,	,	PUNCT
aiti-14589	267	18	with	with	ADP
aiti-14589	267	19	lrc	lrc	PROPN
aiti-14589	267	20	having	have	VERB
aiti-14589	267	21	the	the	DET
aiti-14589	267	22	fastest	fast	ADJ
aiti-14589	267	23	training	training	NOUN
aiti-14589	267	24	(	(	PUNCT
aiti-14589	267	25	0.11	0.11	NUM
aiti-14589	267	26	seconds	second	NOUN
aiti-14589	267	27	)	)	PUNCT
aiti-14589	267	28	while	while	SCONJ
aiti-14589	267	29	cnn	cnn	PROPN
aiti-14589	267	30	needed	need	VERB
aiti-14589	267	31	longer	long	ADV
aiti-14589	267	32	training	training	NOUN
aiti-14589	267	33	time	time	NOUN
aiti-14589	267	34	but	but	CCONJ
aiti-14589	267	35	provided	provide	VERB
aiti-14589	267	36	better	well	ADJ
aiti-14589	267	37	classification	classification	NOUN
aiti-14589	267	38	results	result	NOUN
aiti-14589	267	39	.	.	PUNCT
aiti-14589	268	1	(	(	PUNCT
aiti-14589	268	2	6	6	NUM
aiti-14589	268	3	)	)	PUNCT
aiti-14589	268	4	facial	facial	ADJ
aiti-14589	268	5	landmark	landmark	NOUN
aiti-14589	268	6	points	point	NOUN
aiti-14589	268	7	offer	offer	VERB
aiti-14589	268	8	more	more	ADV
aiti-14589	268	9	reliable	reliable	ADJ
aiti-14589	268	10	geometric	geometric	ADJ
aiti-14589	268	11	information	information	NOUN
aiti-14589	268	12	for	for	ADP
aiti-14589	268	13	expression	expression	NOUN
aiti-14589	268	14	recognition	recognition	NOUN
aiti-14589	268	15	compared	compare	VERB
aiti-14589	268	16	to	to	ADP
aiti-14589	268	17	texture	texture	NOUN
aiti-14589	268	18	-	-	PUNCT
aiti-14589	268	19	based	base	VERB
aiti-14589	268	20	hog	hog	NOUN
aiti-14589	268	21	features	feature	NOUN
aiti-14589	268	22	.	.	PUNCT
aiti-14589	269	1	future	future	ADJ
aiti-14589	269	2	research	research	NOUN
aiti-14589	269	3	should	should	AUX
aiti-14589	269	4	examine	examine	VERB
aiti-14589	269	5	hybrid	hybrid	ADJ
aiti-14589	269	6	approaches	approach	NOUN
aiti-14589	269	7	that	that	PRON
aiti-14589	269	8	integrate	integrate	VERB
aiti-14589	269	9	geometric	geometric	ADJ
aiti-14589	269	10	and	and	CCONJ
aiti-14589	269	11	appearance	appearance	NOUN
aiti-14589	269	12	-	-	PUNCT
aiti-14589	269	13	based	base	VERB
aiti-14589	269	14	features	feature	NOUN
aiti-14589	269	15	,	,	PUNCT
aiti-14589	269	16	explore	explore	VERB
aiti-14589	269	17	deep	deep	ADJ
aiti-14589	269	18	learning	learning	NOUN
aiti-14589	269	19	architectures	architecture	NOUN
aiti-14589	269	20	tailored	tailor	VERB
aiti-14589	269	21	to	to	ADP
aiti-14589	269	22	landmark	landmark	NOUN
aiti-14589	269	23	-	-	PUNCT
aiti-14589	269	24	based	base	VERB
aiti-14589	269	25	classification	classification	NOUN
aiti-14589	269	26	,	,	PUNCT
aiti-14589	269	27	and	and	CCONJ
aiti-14589	269	28	evaluate	evaluate	VERB
aiti-14589	269	29	performance	performance	NOUN
aiti-14589	269	30	on	on	ADP
aiti-14589	269	31	larger	large	ADJ
aiti-14589	269	32	,	,	PUNCT
aiti-14589	269	33	more	more	ADV
aiti-14589	269	34	diverse	diverse	ADJ
aiti-14589	269	35	datasets	dataset	NOUN
aiti-14589	269	36	,	,	PUNCT
aiti-14589	269	37	including	include	VERB
aiti-14589	269	38	cultural	cultural	ADJ
aiti-14589	269	39	variations	variation	NOUN
aiti-14589	269	40	in	in	ADP
aiti-14589	269	41	pain	pain	NOUN
aiti-14589	269	42	expression	expression	NOUN
aiti-14589	269	43	.	.	PUNCT
aiti-14589	270	1	additionally	additionally	ADV
aiti-14589	270	2	,	,	PUNCT
aiti-14589	270	3	real	real	ADJ
aiti-14589	270	4	-	-	PUNCT
aiti-14589	270	5	time	time	NOUN
aiti-14589	270	6	implementation	implementation	NOUN
aiti-14589	270	7	studies	study	NOUN
aiti-14589	270	8	and	and	CCONJ
aiti-14589	270	9	cross	cross	ADJ
aiti-14589	270	10	-	-	ADJ
aiti-14589	270	11	dataset	dataset	ADJ
aiti-14589	270	12	generalization	generalization	NOUN
aiti-14589	270	13	capabilities	capability	NOUN
aiti-14589	270	14	warrant	warrant	VERB
aiti-14589	270	15	further	further	ADJ
aiti-14589	270	16	investigation	investigation	NOUN
aiti-14589	270	17	to	to	PART
aiti-14589	270	18	enhance	enhance	VERB
aiti-14589	270	19	practical	practical	ADJ
aiti-14589	270	20	applicability	applicability	NOUN
aiti-14589	270	21	.	.	PUNCT
aiti-14589	271	1	statement	statement	NOUN
aiti-14589	271	2	of	of	ADP
aiti-14589	271	3	ethical	ethical	ADJ
aiti-14589	271	4	approval	approval	NOUN
aiti-14589	271	5	for	for	ADP
aiti-14589	271	6	this	this	DET
aiti-14589	271	7	type	type	NOUN
aiti-14589	271	8	of	of	ADP
aiti-14589	271	9	study	study	NOUN
aiti-14589	271	10	,	,	PUNCT
aiti-14589	271	11	statement	statement	NOUN
aiti-14589	271	12	of	of	ADP
aiti-14589	271	13	human	human	ADJ
aiti-14589	271	14	rights	right	NOUN
aiti-14589	271	15	is	be	AUX
aiti-14589	271	16	not	not	PART
aiti-14589	271	17	required	require	VERB
aiti-14589	271	18	.	.	PUNCT
aiti-14589	272	1	statement	statement	NOUN
aiti-14589	272	2	of	of	ADP
aiti-14589	272	3	informed	informed	ADJ
aiti-14589	272	4	consent	consent	NOUN
aiti-14589	272	5	for	for	ADP
aiti-14589	272	6	this	this	DET
aiti-14589	272	7	type	type	NOUN
aiti-14589	272	8	of	of	ADP
aiti-14589	272	9	study	study	NOUN
aiti-14589	272	10	,	,	PUNCT
aiti-14589	272	11	informed	inform	VERB
aiti-14589	272	12	consent	consent	NOUN
aiti-14589	272	13	is	be	AUX
aiti-14589	272	14	not	not	PART
aiti-14589	272	15	required	require	VERB
aiti-14589	272	16	.	.	PUNCT
aiti-14589	273	1	references	reference	NOUN
aiti-14589	273	2	[	[	X
aiti-14589	273	3	1	1	NUM
aiti-14589	273	4	]	]	PUNCT
aiti-14589	273	5	i.	i.	PROPN
aiti-14589	273	6	michael	michael	PROPN
aiti-14589	273	7	revina	revina	PROPN
aiti-14589	273	8	and	and	CCONJ
aiti-14589	273	9	w.	w.	PROPN
aiti-14589	273	10	r.	r.	PROPN
aiti-14589	273	11	sam	sam	PROPN
aiti-14589	273	12	emmanuel	emmanuel	PROPN
aiti-14589	273	13	,	,	PUNCT
aiti-14589	273	14	“	"	PUNCT
aiti-14589	273	15	a	a	DET
aiti-14589	273	16	survey	survey	NOUN
aiti-14589	273	17	on	on	ADP
aiti-14589	273	18	human	human	ADJ
aiti-14589	273	19	face	face	NOUN
aiti-14589	273	20	expression	expression	NOUN
aiti-14589	273	21	recognition	recognition	NOUN
aiti-14589	273	22	techniques	technique	NOUN
aiti-14589	273	23	,	,	PUNCT
aiti-14589	273	24	”	"	PUNCT
aiti-14589	273	25	journal	journal	NOUN
aiti-14589	273	26	of	of	ADP
aiti-14589	273	27	king	king	PROPN
aiti-14589	273	28	saud	saud	PROPN
aiti-14589	273	29	university	university	PROPN
aiti-14589	273	30	computer	computer	NOUN
aiti-14589	273	31	and	and	CCONJ
aiti-14589	273	32	information	information	NOUN
aiti-14589	273	33	sciences	science	NOUN
aiti-14589	273	34	,	,	PUNCT
aiti-14589	273	35	vol	vol	NOUN
aiti-14589	273	36	.	.	PROPN
aiti-14589	273	37	33	33	NUM
aiti-14589	273	38	,	,	PUNCT
aiti-14589	273	39	no	no	INTJ
aiti-14589	273	40	.	.	NOUN
aiti-14589	273	41	6	6	NUM
aiti-14589	273	42	,	,	PUNCT
aiti-14589	273	43	pp	pp	ADJ
aiti-14589	273	44	.	.	PUNCT
aiti-14589	274	1	619	619	NUM
aiti-14589	274	2	-	-	SYM
aiti-14589	274	3	628	628	NUM
aiti-14589	274	4	,	,	PUNCT
aiti-14589	274	5	2021	2021	NUM
aiti-14589	274	6	.	.	PUNCT
aiti-14589	275	1	advances	advance	NOUN
aiti-14589	275	2	in	in	ADP
aiti-14589	275	3	technology	technology	NOUN
aiti-14589	275	4	innovation	innovation	NOUN
aiti-14589	275	5	,	,	PUNCT
aiti-14589	275	6	vol	vol	NOUN
aiti-14589	275	7	.	.	PROPN
aiti-14589	276	1	10	10	NUM
aiti-14589	276	2	,	,	PUNCT
aiti-14589	276	3	no	no	INTJ
aiti-14589	276	4	.	.	NOUN
aiti-14589	276	5	4	4	NUM
aiti-14589	276	6	,	,	PUNCT
aiti-14589	276	7	2025	2025	NUM
aiti-14589	276	8	,	,	PUNCT
aiti-14589	276	9	pp	pp	ADJ
aiti-14589	276	10	.	.	PUNCT
aiti-14589	277	1	395	395	NUM
aiti-14589	277	2	-	-	SYM
aiti-14589	277	3	406	406	NUM
aiti-14589	277	4	406	406	NUM
aiti-14589	277	5	[	[	X
aiti-14589	277	6	2	2	NUM
aiti-14589	277	7	]	]	PUNCT
aiti-14589	277	8	d.	d.	PROPN
aiti-14589	277	9	canedo	canedo	PROPN
aiti-14589	277	10	and	and	CCONJ
aiti-14589	277	11	a.	a.	PROPN
aiti-14589	277	12	j.	j.	PROPN
aiti-14589	277	13	r.	r.	PROPN
aiti-14589	277	14	neves	neves	PROPN
aiti-14589	277	15	,	,	PUNCT
aiti-14589	277	16	“	"	PUNCT
aiti-14589	277	17	facial	facial	ADJ
aiti-14589	277	18	expression	expression	NOUN
aiti-14589	277	19	recognition	recognition	NOUN
aiti-14589	277	20	using	use	VERB
aiti-14589	277	21	computer	computer	NOUN
aiti-14589	277	22	vision	vision	NOUN
aiti-14589	277	23	:	:	PUNCT
aiti-14589	277	24	a	a	DET
aiti-14589	277	25	systematic	systematic	ADJ
aiti-14589	277	26	review	review	NOUN
aiti-14589	277	27	,	,	PUNCT
aiti-14589	277	28	”	"	PUNCT
aiti-14589	277	29	applied	apply	VERB
aiti-14589	277	30	sciences	science	NOUN
aiti-14589	277	31	,	,	PUNCT
aiti-14589	277	32	vol	vol	NOUN
aiti-14589	277	33	.	.	NOUN
aiti-14589	277	34	9	9	NUM
aiti-14589	277	35	,	,	PUNCT
aiti-14589	277	36	no	no	INTJ
aiti-14589	277	37	.	.	NOUN
aiti-14589	277	38	21	21	NUM
aiti-14589	277	39	,	,	PUNCT
aiti-14589	277	40	article	article	NOUN
aiti-14589	277	41	no	no	NOUN
aiti-14589	277	42	.	.	PROPN
aiti-14589	277	43	4678	4678	NUM
aiti-14589	277	44	,	,	PUNCT
aiti-14589	277	45	2019	2019	NUM
aiti-14589	277	46	.	.	PUNCT
aiti-14589	278	1	[	[	X
aiti-14589	278	2	3	3	X
aiti-14589	278	3	]	]	X
aiti-14589	278	4	s.	s.	PROPN
aiti-14589	278	5	kanjanawattana	kanjanawattana	PROPN
aiti-14589	278	6	,	,	PUNCT
aiti-14589	278	7	p.	p.	PROPN
aiti-14589	278	8	kittichaiwatthana	kittichaiwatthana	PROPN
aiti-14589	278	9	,	,	PUNCT
aiti-14589	278	10	k.	k.	PROPN
aiti-14589	278	11	srivisut	srivisut	PROPN
aiti-14589	278	12	,	,	PUNCT
aiti-14589	278	13	and	and	CCONJ
aiti-14589	278	14	p.	p.	PROPN
aiti-14589	278	15	praneetpholkrang	praneetpholkrang	PROPN
aiti-14589	278	16	,	,	PUNCT
aiti-14589	278	17	“	"	PUNCT
aiti-14589	278	18	deep	deep	ADJ
aiti-14589	278	19	learning	learning	NOUN
aiti-14589	278	20	-	-	PUNCT
aiti-14589	278	21	based	base	VERB
aiti-14589	278	22	emotion	emotion	NOUN
aiti-14589	278	23	recognition	recognition	NOUN
aiti-14589	278	24	through	through	ADP
aiti-14589	278	25	facial	facial	ADJ
aiti-14589	278	26	expressions	expression	NOUN
aiti-14589	278	27	,	,	PUNCT
aiti-14589	278	28	”	"	PUNCT
aiti-14589	278	29	journal	journal	NOUN
aiti-14589	278	30	of	of	ADP
aiti-14589	278	31	image	image	NOUN
aiti-14589	278	32	and	and	CCONJ
aiti-14589	278	33	graphics	graphic	NOUN
aiti-14589	278	34	,	,	PUNCT
aiti-14589	278	35	vol	vol	NOUN
aiti-14589	278	36	.	.	PROPN
aiti-14589	278	37	11	11	NUM
aiti-14589	278	38	,	,	PUNCT
aiti-14589	278	39	no	no	INTJ
aiti-14589	278	40	.	.	NOUN
aiti-14589	278	41	2	2	NUM
aiti-14589	278	42	,	,	PUNCT
aiti-14589	278	43	pp	pp	ADJ
aiti-14589	278	44	.	.	PUNCT
aiti-14589	279	1	140	140	NUM
aiti-14589	279	2	-	-	SYM
aiti-14589	279	3	145	145	NUM
aiti-14589	279	4	,	,	PUNCT
aiti-14589	279	5	2023	2023	NUM
aiti-14589	279	6	.	.	PUNCT
aiti-14589	280	1	[	[	X
aiti-14589	280	2	4	4	X
aiti-14589	280	3	]	]	PUNCT
aiti-14589	280	4	s.	s.	PROPN
aiti-14589	280	5	schindler	schindler	PROPN
aiti-14589	280	6	,	,	PUNCT
aiti-14589	280	7	c.	c.	PROPN
aiti-14589	280	8	tirloni	tirloni	PROPN
aiti-14589	280	9	,	,	PUNCT
aiti-14589	280	10	m.	m.	NOUN
aiti-14589	280	11	bruchmann	bruchmann	PROPN
aiti-14589	280	12	,	,	PUNCT
aiti-14589	280	13	and	and	CCONJ
aiti-14589	280	14	t.	t.	PROPN
aiti-14589	280	15	straube	straube	NOUN
aiti-14589	280	16	,	,	PUNCT
aiti-14589	280	17	“	"	PUNCT
aiti-14589	280	18	face	face	NOUN
aiti-14589	280	19	and	and	CCONJ
aiti-14589	280	20	emotional	emotional	ADJ
aiti-14589	280	21	expression	expression	NOUN
aiti-14589	280	22	processing	processing	NOUN
aiti-14589	280	23	under	under	ADP
aiti-14589	280	24	continuous	continuous	ADJ
aiti-14589	280	25	perceptual	perceptual	ADJ
aiti-14589	280	26	load	load	NOUN
aiti-14589	280	27	tasks	task	NOUN
aiti-14589	280	28	:	:	PUNCT
aiti-14589	280	29	an	an	DET
aiti-14589	280	30	erp	erp	PROPN
aiti-14589	280	31	study	study	NOUN
aiti-14589	280	32	,	,	PUNCT
aiti-14589	280	33	”	"	PUNCT
aiti-14589	280	34	biological	biological	ADJ
aiti-14589	280	35	psychology	psychology	NOUN
aiti-14589	280	36	,	,	PUNCT
aiti-14589	280	37	vol	vol	NOUN
aiti-14589	280	38	.	.	PROPN
aiti-14589	280	39	161	161	NUM
aiti-14589	280	40	,	,	PUNCT
aiti-14589	280	41	article	article	NOUN
aiti-14589	280	42	no	no	NOUN
aiti-14589	280	43	.	.	PROPN
aiti-14589	280	44	108056	108056	NUM
aiti-14589	280	45	,	,	PUNCT
aiti-14589	280	46	2021	2021	NUM
aiti-14589	280	47	.	.	PUNCT
aiti-14589	281	1	[	[	X
aiti-14589	281	2	5	5	NUM
aiti-14589	281	3	]	]	PUNCT
aiti-14589	281	4	r.	r.	PROPN
aiti-14589	281	5	w.	w.	PROPN
aiti-14589	281	6	buck	buck	PROPN
aiti-14589	281	7	,	,	PUNCT
aiti-14589	281	8	v.	v.	PROPN
aiti-14589	281	9	j.	j.	PROPN
aiti-14589	281	10	savin	savin	PROPN
aiti-14589	281	11	,	,	PUNCT
aiti-14589	281	12	r.	r.	PROPN
aiti-14589	281	13	e.	e.	PROPN
aiti-14589	281	14	miller	miller	PROPN
aiti-14589	281	15	,	,	PUNCT
aiti-14589	281	16	and	and	CCONJ
aiti-14589	281	17	w.	w.	PROPN
aiti-14589	281	18	f.	f.	PROPN
aiti-14589	281	19	caul	caul	PROPN
aiti-14589	281	20	,	,	PUNCT
aiti-14589	281	21	“	"	PUNCT
aiti-14589	281	22	communication	communication	NOUN
aiti-14589	281	23	of	of	ADP
aiti-14589	281	24	affect	affect	VERB
aiti-14589	281	25	through	through	ADP
aiti-14589	281	26	facial	facial	ADJ
aiti-14589	281	27	expressions	expression	NOUN
aiti-14589	281	28	in	in	ADP
aiti-14589	281	29	humans	human	NOUN
aiti-14589	281	30	,	,	PUNCT
aiti-14589	281	31	”	"	PUNCT
aiti-14589	281	32	journal	journal	NOUN
aiti-14589	281	33	of	of	ADP
aiti-14589	281	34	personality	personality	NOUN
aiti-14589	281	35	and	and	CCONJ
aiti-14589	281	36	social	social	ADJ
aiti-14589	281	37	psychology	psychology	NOUN
aiti-14589	281	38	,	,	PUNCT
aiti-14589	281	39	vol	vol	NOUN
aiti-14589	281	40	.	.	PROPN
aiti-14589	282	1	23	23	NUM
aiti-14589	282	2	,	,	PUNCT
aiti-14589	283	1	no	no	INTJ
aiti-14589	283	2	.	.	NOUN
aiti-14589	283	3	3	3	NUM
aiti-14589	283	4	,	,	PUNCT
aiti-14589	283	5	pp	pp	ADJ
aiti-14589	283	6	.	.	PUNCT
aiti-14589	284	1	362	362	NUM
aiti-14589	284	2	-	-	SYM
aiti-14589	284	3	371	371	NUM
aiti-14589	284	4	,	,	PUNCT
aiti-14589	284	5	1972	1972	NUM
aiti-14589	284	6	.	.	PUNCT
aiti-14589	285	1	[	[	X
aiti-14589	285	2	6	6	NUM
aiti-14589	285	3	]	]	X
aiti-14589	285	4	c.	c.	PROPN
aiti-14589	285	5	l.	l.	PROPN
aiti-14589	285	6	lisetti	lisetti	PROPN
aiti-14589	285	7	and	and	CCONJ
aiti-14589	285	8	d.	d.	PROPN
aiti-14589	285	9	j.	j.	PROPN
aiti-14589	285	10	schiano	schiano	PROPN
aiti-14589	285	11	,	,	PUNCT
aiti-14589	285	12	“	"	PUNCT
aiti-14589	285	13	automatic	automatic	ADJ
aiti-14589	285	14	facial	facial	ADJ
aiti-14589	285	15	expression	expression	NOUN
aiti-14589	285	16	interpretation	interpretation	NOUN
aiti-14589	285	17	:	:	PUNCT
aiti-14589	285	18	where	where	SCONJ
aiti-14589	285	19	human	human	NOUN
aiti-14589	285	20	-	-	PUNCT
aiti-14589	285	21	computer	computer	NOUN
aiti-14589	285	22	interaction	interaction	NOUN
aiti-14589	285	23	,	,	PUNCT
aiti-14589	285	24	artificial	artificial	ADJ
aiti-14589	285	25	intelligence	intelligence	NOUN
aiti-14589	285	26	and	and	CCONJ
aiti-14589	285	27	cognitive	cognitive	ADJ
aiti-14589	285	28	science	science	NOUN
aiti-14589	285	29	intersect	intersect	ADJ
aiti-14589	285	30	,	,	PUNCT
aiti-14589	285	31	”	"	PUNCT
aiti-14589	285	32	pragmatics	pragmatic	NOUN
aiti-14589	285	33	&	&	CCONJ
aiti-14589	285	34	cognition	cognition	NOUN
aiti-14589	285	35	,	,	PUNCT
aiti-14589	285	36	vol	vol	NOUN
aiti-14589	285	37	.	.	PROPN
aiti-14589	285	38	8	8	NUM
aiti-14589	285	39	,	,	PUNCT
aiti-14589	285	40	no	no	INTJ
aiti-14589	285	41	.	.	NOUN
aiti-14589	285	42	1	1	NUM
aiti-14589	285	43	,	,	PUNCT
aiti-14589	285	44	pp	pp	ADJ
aiti-14589	285	45	.	.	PUNCT
aiti-14589	286	1	185	185	NUM
aiti-14589	286	2	-	-	SYM
aiti-14589	286	3	235	235	NUM
aiti-14589	286	4	,	,	PUNCT
aiti-14589	286	5	2000	2000	NUM
aiti-14589	286	6	.	.	PUNCT
aiti-14589	287	1	[	[	X
aiti-14589	287	2	7	7	X
aiti-14589	287	3	]	]	X
aiti-14589	287	4	c.	c.	PROPN
aiti-14589	287	5	affonso	affonso	PROPN
aiti-14589	287	6	,	,	PUNCT
aiti-14589	287	7	a.	a.	PROPN
aiti-14589	287	8	l.	l.	PROPN
aiti-14589	287	9	d.	d.	PROPN
aiti-14589	287	10	rossi	rossi	PROPN
aiti-14589	287	11	,	,	PUNCT
aiti-14589	287	12	f.	f.	PROPN
aiti-14589	287	13	h.	h.	PROPN
aiti-14589	287	14	a.	a.	PROPN
aiti-14589	287	15	vieira	vieira	PROPN
aiti-14589	287	16	,	,	PUNCT
aiti-14589	287	17	and	and	CCONJ
aiti-14589	287	18	a.	a.	PROPN
aiti-14589	287	19	c.	c.	PROPN
aiti-14589	288	1	p.	p.	PROPN
aiti-14589	288	2	de	de	PROPN
aiti-14589	288	3	leon	leon	PROPN
aiti-14589	288	4	ferreira	ferreira	PROPN
aiti-14589	288	5	de	de	PROPN
aiti-14589	288	6	carvalho	carvalho	PROPN
aiti-14589	288	7	,	,	PUNCT
aiti-14589	288	8	“	"	PUNCT
aiti-14589	288	9	deep	deep	ADJ
aiti-14589	288	10	learning	learning	NOUN
aiti-14589	288	11	for	for	ADP
aiti-14589	288	12	biological	biological	ADJ
aiti-14589	288	13	image	image	NOUN
aiti-14589	288	14	classification	classification	NOUN
aiti-14589	288	15	,	,	PUNCT
aiti-14589	288	16	”	"	PUNCT
aiti-14589	288	17	expert	expert	NOUN
aiti-14589	288	18	systems	system	NOUN
aiti-14589	288	19	with	with	ADP
aiti-14589	288	20	applications	application	NOUN
aiti-14589	288	21	,	,	PUNCT
aiti-14589	288	22	vol	vol	NOUN
aiti-14589	288	23	.	.	PROPN
aiti-14589	288	24	85	85	NUM
aiti-14589	288	25	,	,	PUNCT
aiti-14589	288	26	pp	pp	ADJ
aiti-14589	288	27	.	.	PUNCT
aiti-14589	289	1	114	114	NUM
aiti-14589	289	2	-	-	SYM
aiti-14589	289	3	122	122	NUM
aiti-14589	289	4	,	,	PUNCT
aiti-14589	289	5	2017	2017	NUM
aiti-14589	289	6	.	.	PUNCT
aiti-14589	290	1	[	[	X
aiti-14589	290	2	8	8	NUM
aiti-14589	290	3	]	]	X
aiti-14589	290	4	p.	p.	PROPN
aiti-14589	290	5	mende	mende	PROPN
aiti-14589	290	6	-	-	PUNCT
aiti-14589	290	7	siedlecki	siedlecki	PROPN
aiti-14589	290	8	,	,	PUNCT
aiti-14589	290	9	j.	j.	PROPN
aiti-14589	290	10	qu	qu	PROPN
aiti-14589	290	11	-	-	PROPN
aiti-14589	290	12	lee	lee	PROPN
aiti-14589	290	13	,	,	PUNCT
aiti-14589	290	14	j.	j.	PROPN
aiti-14589	290	15	lin	lin	PROPN
aiti-14589	290	16	,	,	PUNCT
aiti-14589	290	17	a.	a.	NOUN
aiti-14589	290	18	drain	drain	NOUN
aiti-14589	290	19	,	,	PUNCT
aiti-14589	290	20	and	and	CCONJ
aiti-14589	290	21	a.	a.	NOUN
aiti-14589	290	22	goharzad	goharzad	NOUN
aiti-14589	290	23	,	,	PUNCT
aiti-14589	290	24	“	"	PUNCT
aiti-14589	290	25	the	the	DET
aiti-14589	290	26	delaware	delaware	PROPN
aiti-14589	290	27	pain	pain	PROPN
aiti-14589	290	28	database	database	PROPN
aiti-14589	290	29	:	:	PUNCT
aiti-14589	290	30	a	a	DET
aiti-14589	290	31	set	set	NOUN
aiti-14589	290	32	of	of	ADP
aiti-14589	290	33	painful	painful	ADJ
aiti-14589	290	34	expressions	expression	NOUN
aiti-14589	290	35	and	and	CCONJ
aiti-14589	290	36	corresponding	correspond	VERB
aiti-14589	290	37	norming	norming	NOUN
aiti-14589	290	38	data	datum	NOUN
aiti-14589	290	39	,	,	PUNCT
aiti-14589	290	40	”	"	PUNCT
aiti-14589	290	41	pain	pain	NOUN
aiti-14589	290	42	reports	report	NOUN
aiti-14589	290	43	,	,	PUNCT
aiti-14589	290	44	vol	vol	NOUN
aiti-14589	290	45	.	.	PROPN
aiti-14589	290	46	5	5	NUM
aiti-14589	290	47	,	,	PUNCT
aiti-14589	290	48	no	no	INTJ
aiti-14589	290	49	.	.	NOUN
aiti-14589	290	50	6	6	NUM
aiti-14589	290	51	,	,	PUNCT
aiti-14589	290	52	article	article	NOUN
aiti-14589	290	53	no	no	NOUN
aiti-14589	290	54	.	.	PUNCT
aiti-14589	291	1	e853	e853	ADJ
aiti-14589	291	2	,	,	PUNCT
aiti-14589	291	3	2020	2020	NUM
aiti-14589	291	4	.	.	PUNCT
aiti-14589	292	1	[	[	X
aiti-14589	292	2	9	9	NUM
aiti-14589	292	3	]	]	X
aiti-14589	292	4	u.	u.	PROPN
aiti-14589	292	5	sharma	sharma	PROPN
aiti-14589	292	6	,	,	PUNCT
aiti-14589	292	7	k.	k.	PROPN
aiti-14589	292	8	n.	n.	PROPN
aiti-14589	292	9	faisal	faisal	PROPN
aiti-14589	292	10	,	,	PUNCT
aiti-14589	292	11	r.	r.	PROPN
aiti-14589	292	12	r.	r.	PROPN
aiti-14589	292	13	sharma	sharma	PROPN
aiti-14589	292	14	,	,	PUNCT
aiti-14589	292	15	and	and	CCONJ
aiti-14589	292	16	k.	k.	PROPN
aiti-14589	292	17	v.	v.	PROPN
aiti-14589	292	18	arya	arya	PROPN
aiti-14589	292	19	,	,	PUNCT
aiti-14589	292	20	“	"	PUNCT
aiti-14589	292	21	facial	facial	ADJ
aiti-14589	292	22	landmark	landmark	NOUN
aiti-14589	292	23	-	-	PUNCT
aiti-14589	292	24	based	base	VERB
aiti-14589	292	25	human	human	ADJ
aiti-14589	292	26	emotion	emotion	NOUN
aiti-14589	292	27	recognition	recognition	NOUN
aiti-14589	292	28	technique	technique	NOUN
aiti-14589	292	29	for	for	ADP
aiti-14589	292	30	oriented	orient	VERB
aiti-14589	292	31	viewpoints	viewpoint	NOUN
aiti-14589	292	32	in	in	ADP
aiti-14589	292	33	the	the	DET
aiti-14589	292	34	presence	presence	NOUN
aiti-14589	292	35	of	of	ADP
aiti-14589	292	36	facial	facial	ADJ
aiti-14589	292	37	attributes	attribute	NOUN
aiti-14589	292	38	,	,	PUNCT
aiti-14589	292	39	”	"	PUNCT
aiti-14589	292	40	sn	sn	PROPN
aiti-14589	292	41	computer	computer	NOUN
aiti-14589	292	42	science	science	NOUN
aiti-14589	292	43	,	,	PUNCT
aiti-14589	292	44	vol	vol	NOUN
aiti-14589	292	45	.	.	PROPN
aiti-14589	292	46	4	4	NUM
aiti-14589	292	47	,	,	PUNCT
aiti-14589	292	48	no	no	INTJ
aiti-14589	292	49	.	.	NOUN
aiti-14589	292	50	3	3	NUM
aiti-14589	292	51	,	,	PUNCT
aiti-14589	292	52	article	article	NOUN
aiti-14589	292	53	no	no	NOUN
aiti-14589	292	54	.	.	PROPN
aiti-14589	292	55	273	273	NUM
aiti-14589	292	56	,	,	PUNCT
aiti-14589	292	57	2023	2023	NUM
aiti-14589	292	58	.	.	PUNCT
aiti-14589	293	1	[	[	X
aiti-14589	293	2	10	10	NUM
aiti-14589	293	3	]	]	X
aiti-14589	293	4	d.	d.	PROPN
aiti-14589	293	5	lu	lu	PROPN
aiti-14589	293	6	and	and	CCONJ
aiti-14589	293	7	q.	q.	PROPN
aiti-14589	293	8	weng	weng	PROPN
aiti-14589	293	9	,	,	PUNCT
aiti-14589	293	10	“	"	PUNCT
aiti-14589	293	11	a	a	DET
aiti-14589	293	12	survey	survey	NOUN
aiti-14589	293	13	of	of	ADP
aiti-14589	293	14	image	image	NOUN
aiti-14589	293	15	classification	classification	NOUN
aiti-14589	293	16	methods	method	NOUN
aiti-14589	293	17	and	and	CCONJ
aiti-14589	293	18	techniques	technique	NOUN
aiti-14589	293	19	for	for	ADP
aiti-14589	293	20	improving	improve	VERB
aiti-14589	293	21	classification	classification	NOUN
aiti-14589	293	22	performance	performance	NOUN
aiti-14589	293	23	,	,	PUNCT
aiti-14589	293	24	”	"	PUNCT
aiti-14589	293	25	international	international	ADJ
aiti-14589	293	26	journal	journal	NOUN
aiti-14589	293	27	of	of	ADP
aiti-14589	293	28	remote	remote	ADJ
aiti-14589	293	29	sensing	sensing	NOUN
aiti-14589	293	30	,	,	PUNCT
aiti-14589	293	31	vol	vol	NOUN
aiti-14589	293	32	.	.	PROPN
aiti-14589	293	33	28	28	NUM
aiti-14589	293	34	,	,	PUNCT
aiti-14589	293	35	no	no	INTJ
aiti-14589	293	36	.	.	NOUN
aiti-14589	293	37	5	5	NUM
aiti-14589	293	38	,	,	PUNCT
aiti-14589	293	39	pp	pp	ADJ
aiti-14589	293	40	.	.	PUNCT
aiti-14589	294	1	823	823	NUM
aiti-14589	294	2	-	-	SYM
aiti-14589	294	3	870	870	NUM
aiti-14589	294	4	,	,	PUNCT
aiti-14589	294	5	2007	2007	NUM
aiti-14589	294	6	.	.	PUNCT
aiti-14589	295	1	[	[	X
aiti-14589	295	2	11	11	NUM
aiti-14589	295	3	]	]	PUNCT
aiti-14589	295	4	m.	m.	NOUN
aiti-14589	295	5	bodini	bodini	PROPN
aiti-14589	295	6	,	,	PUNCT
aiti-14589	295	7	“	"	PUNCT
aiti-14589	295	8	a	a	DET
aiti-14589	295	9	review	review	NOUN
aiti-14589	295	10	of	of	ADP
aiti-14589	295	11	facial	facial	ADJ
aiti-14589	295	12	landmark	landmark	NOUN
aiti-14589	295	13	extraction	extraction	NOUN
aiti-14589	295	14	in	in	ADP
aiti-14589	295	15	2d	2d	NOUN
aiti-14589	295	16	images	image	NOUN
aiti-14589	295	17	and	and	CCONJ
aiti-14589	295	18	videos	video	NOUN
aiti-14589	295	19	using	use	VERB
aiti-14589	295	20	deep	deep	ADJ
aiti-14589	295	21	learning	learning	NOUN
aiti-14589	295	22	,	,	PUNCT
aiti-14589	295	23	”	"	PUNCT
aiti-14589	295	24	big	big	ADJ
aiti-14589	295	25	data	datum	NOUN
aiti-14589	295	26	and	and	CCONJ
aiti-14589	295	27	cognitive	cognitive	ADJ
aiti-14589	295	28	computing	computing	NOUN
aiti-14589	295	29	,	,	PUNCT
aiti-14589	295	30	vol	vol	NOUN
aiti-14589	295	31	.	.	PROPN
aiti-14589	296	1	3	3	NUM
aiti-14589	296	2	,	,	PUNCT
aiti-14589	296	3	no	no	INTJ
aiti-14589	296	4	.	.	NOUN
aiti-14589	296	5	1	1	NUM
aiti-14589	296	6	,	,	PUNCT
aiti-14589	296	7	article	article	NOUN
aiti-14589	296	8	no	no	NOUN
aiti-14589	296	9	.	.	PROPN
aiti-14589	296	10	14	14	NUM
aiti-14589	296	11	,	,	PUNCT
aiti-14589	296	12	2019	2019	NUM
aiti-14589	296	13	.	.	PUNCT
aiti-14589	297	1	[	[	X
aiti-14589	297	2	12	12	NUM
aiti-14589	297	3	]	]	X
aiti-14589	297	4	r.	r.	PROPN
aiti-14589	297	5	chauhan	chauhan	PROPN
aiti-14589	297	6	,	,	PUNCT
aiti-14589	297	7	k.	k.	PROPN
aiti-14589	297	8	k.	k.	PROPN
aiti-14589	297	9	ghanshala	ghanshala	PROPN
aiti-14589	297	10	,	,	PUNCT
aiti-14589	297	11	and	and	CCONJ
aiti-14589	297	12	r.	r.	PROPN
aiti-14589	297	13	c.	c.	PROPN
aiti-14589	297	14	joshi	joshi	PROPN
aiti-14589	297	15	,	,	PUNCT
aiti-14589	297	16	“	"	PUNCT
aiti-14589	297	17	convolutional	convolutional	ADJ
aiti-14589	297	18	neural	neural	ADJ
aiti-14589	297	19	network	network	NOUN
aiti-14589	297	20	(	(	PUNCT
aiti-14589	297	21	cnn	cnn	PROPN
aiti-14589	297	22	)	)	PUNCT
aiti-14589	297	23	for	for	ADP
aiti-14589	297	24	image	image	NOUN
aiti-14589	297	25	detection	detection	NOUN
aiti-14589	297	26	and	and	CCONJ
aiti-14589	297	27	recognition	recognition	NOUN
aiti-14589	297	28	,	,	PUNCT
aiti-14589	297	29	”	"	PUNCT
aiti-14589	297	30	first	first	ADJ
aiti-14589	297	31	international	international	ADJ
aiti-14589	297	32	conference	conference	NOUN
aiti-14589	297	33	on	on	ADP
aiti-14589	297	34	secure	secure	ADJ
aiti-14589	297	35	cyber	cyber	NOUN
aiti-14589	297	36	computing	computing	NOUN
aiti-14589	297	37	and	and	CCONJ
aiti-14589	297	38	communication	communication	NOUN
aiti-14589	297	39	,	,	PUNCT
aiti-14589	297	40	pp	pp	ADJ
aiti-14589	297	41	.	.	PUNCT
aiti-14589	298	1	278	278	NUM
aiti-14589	298	2	-	-	SYM
aiti-14589	298	3	282	282	NUM
aiti-14589	298	4	,	,	PUNCT
aiti-14589	298	5	2018	2018	NUM
aiti-14589	298	6	.	.	PUNCT
aiti-14589	299	1	[	[	X
aiti-14589	299	2	13	13	NUM
aiti-14589	299	3	]	]	PUNCT
aiti-14589	299	4	j.	j.	PROPN
aiti-14589	299	5	hilbe	hilbe	PROPN
aiti-14589	299	6	,	,	PUNCT
aiti-14589	299	7	logistic	logistic	ADJ
aiti-14589	299	8	regression	regression	NOUN
aiti-14589	299	9	models	model	NOUN
aiti-14589	299	10	,	,	PUNCT
aiti-14589	299	11	boca	boca	PROPN
aiti-14589	299	12	raton	raton	PROPN
aiti-14589	299	13	:	:	PUNCT
aiti-14589	299	14	chapman	chapman	PROPN
aiti-14589	299	15	&	&	CCONJ
aiti-14589	299	16	hall	hall	PROPN
aiti-14589	299	17	/	/	SYM
aiti-14589	299	18	crc	crc	PROPN
aiti-14589	299	19	,	,	PUNCT
aiti-14589	299	20	2009	2009	NUM
aiti-14589	299	21	.	.	PUNCT
aiti-14589	300	1	[	[	X
aiti-14589	300	2	14	14	NUM
aiti-14589	300	3	]	]	PUNCT
aiti-14589	300	4	m.	m.	NOUN
aiti-14589	300	5	a.	a.	PROPN
aiti-14589	300	6	chandra	chandra	PROPN
aiti-14589	300	7	and	and	CCONJ
aiti-14589	300	8	s.	s.	PROPN
aiti-14589	300	9	s.	s.	PROPN
aiti-14589	300	10	bedi	bedi	PROPN
aiti-14589	300	11	,	,	PUNCT
aiti-14589	300	12	“	"	PUNCT
aiti-14589	300	13	survey	survey	NOUN
aiti-14589	300	14	on	on	ADP
aiti-14589	300	15	svm	svm	PROPN
aiti-14589	300	16	and	and	CCONJ
aiti-14589	300	17	their	their	PRON
aiti-14589	300	18	application	application	NOUN
aiti-14589	300	19	in	in	ADP
aiti-14589	300	20	image	image	NOUN
aiti-14589	300	21	classification	classification	NOUN
aiti-14589	300	22	,	,	PUNCT
aiti-14589	300	23	”	"	PUNCT
aiti-14589	300	24	international	international	ADJ
aiti-14589	300	25	journal	journal	NOUN
aiti-14589	300	26	of	of	ADP
aiti-14589	300	27	information	information	NOUN
aiti-14589	300	28	technology	technology	NOUN
aiti-14589	300	29	,	,	PUNCT
aiti-14589	300	30	vol	vol	NOUN
aiti-14589	300	31	.	.	PROPN
aiti-14589	300	32	13	13	NUM
aiti-14589	300	33	,	,	PUNCT
aiti-14589	300	34	no	no	INTJ
aiti-14589	300	35	.	.	NOUN
aiti-14589	300	36	5	5	NUM
aiti-14589	300	37	,	,	PUNCT
aiti-14589	300	38	pp	pp	ADJ
aiti-14589	300	39	.	.	PUNCT
aiti-14589	301	1	1	1	NUM
aiti-14589	301	2	-	-	SYM
aiti-14589	301	3	11	11	NUM
aiti-14589	301	4	,	,	PUNCT
aiti-14589	301	5	2021	2021	NUM
aiti-14589	301	6	.	.	PUNCT
aiti-14589	302	1	[	[	X
aiti-14589	302	2	15	15	NUM
aiti-14589	302	3	]	]	X
aiti-14589	302	4	n.	n.	NOUN
aiti-14589	302	5	m.	m.	NOUN
aiti-14589	302	6	abdulkareem	abdulkareem	PROPN
aiti-14589	302	7	and	and	CCONJ
aiti-14589	302	8	a.	a.	NOUN
aiti-14589	302	9	m.	m.	PROPN
aiti-14589	302	10	abdulazeez	abdulazeez	PROPN
aiti-14589	302	11	,	,	PUNCT
aiti-14589	302	12	“	"	PUNCT
aiti-14589	302	13	machine	machine	NOUN
aiti-14589	302	14	learning	learn	VERB
aiti-14589	302	15	classification	classification	NOUN
aiti-14589	302	16	based	base	VERB
aiti-14589	302	17	on	on	ADP
aiti-14589	302	18	radom	radom	NOUN
aiti-14589	302	19	forest	forest	NOUN
aiti-14589	302	20	algorithm	algorithm	NOUN
aiti-14589	302	21	:	:	PUNCT
aiti-14589	302	22	a	a	DET
aiti-14589	302	23	review	review	NOUN
aiti-14589	302	24	,	,	PUNCT
aiti-14589	302	25	”	"	PUNCT
aiti-14589	302	26	international	international	ADJ
aiti-14589	302	27	journal	journal	NOUN
aiti-14589	302	28	of	of	ADP
aiti-14589	302	29	science	science	NOUN
aiti-14589	302	30	and	and	CCONJ
aiti-14589	302	31	business	business	NOUN
aiti-14589	302	32	,	,	PUNCT
aiti-14589	302	33	vol	vol	NOUN
aiti-14589	302	34	.	.	PROPN
aiti-14589	303	1	5	5	NUM
aiti-14589	303	2	,	,	PUNCT
aiti-14589	303	3	no	no	INTJ
aiti-14589	303	4	.	.	NOUN
aiti-14589	303	5	2	2	NUM
aiti-14589	303	6	,	,	PUNCT
aiti-14589	303	7	pp	pp	ADJ
aiti-14589	303	8	.	.	PUNCT
aiti-14589	304	1	128	128	NUM
aiti-14589	304	2	-	-	SYM
aiti-14589	304	3	142	142	NUM
aiti-14589	304	4	,	,	PUNCT
aiti-14589	304	5	2021	2021	NUM
aiti-14589	304	6	.	.	PUNCT
aiti-14589	305	1	[	[	X
aiti-14589	305	2	16	16	NUM
aiti-14589	305	3	]	]	PUNCT
aiti-14589	305	4	z.	z.	PROPN
aiti-14589	305	5	he	he	PROPN
aiti-14589	305	6	,	,	PUNCT
aiti-14589	305	7	d.	d.	PROPN
aiti-14589	305	8	lin	lin	PROPN
aiti-14589	305	9	,	,	PUNCT
aiti-14589	305	10	t.	t.	PROPN
aiti-14589	305	11	lau	lau	PROPN
aiti-14589	305	12	,	,	PUNCT
aiti-14589	305	13	and	and	CCONJ
aiti-14589	305	14	m.	m.	PROPN
aiti-14589	305	15	wu	wu	PROPN
aiti-14589	305	16	,	,	PUNCT
aiti-14589	305	17	“	"	PUNCT
aiti-14589	305	18	gradient	gradient	ADJ
aiti-14589	305	19	boosting	boost	VERB
aiti-14589	305	20	machine	machine	NOUN
aiti-14589	305	21	:	:	PUNCT
aiti-14589	305	22	a	a	DET
aiti-14589	305	23	survey	survey	NOUN
aiti-14589	305	24	,	,	PUNCT
aiti-14589	305	25	”	"	PUNCT
aiti-14589	305	26	https://doi.org/10.48550/arxiv.1908.06951	https://doi.org/10.48550/arxiv.1908.06951	PROPN
aiti-14589	305	27	,	,	PUNCT
aiti-14589	305	28	2019	2019	NUM
aiti-14589	305	29	.	.	PUNCT
aiti-14589	306	1	[	[	X
aiti-14589	306	2	17	17	NUM
aiti-14589	306	3	]	]	X
aiti-14589	306	4	v.	v.	ADP
aiti-14589	306	5	n.	n.	PROPN
aiti-14589	306	6	vapnik	vapnik	PROPN
aiti-14589	306	7	,	,	PUNCT
aiti-14589	306	8	the	the	DET
aiti-14589	306	9	nature	nature	NOUN
aiti-14589	306	10	of	of	ADP
aiti-14589	306	11	statistical	statistical	ADJ
aiti-14589	306	12	learning	learning	NOUN
aiti-14589	306	13	theory	theory	NOUN
aiti-14589	306	14	,	,	PUNCT
aiti-14589	306	15	2nd	2nd	ADJ
aiti-14589	306	16	ed	ed	NOUN
aiti-14589	306	17	.	.	PROPN
aiti-14589	306	18	,	,	PUNCT
aiti-14589	306	19	new	new	PROPN
aiti-14589	306	20	york	york	PROPN
aiti-14589	306	21	:	:	PUNCT
aiti-14589	306	22	springer	springer	NOUN
aiti-14589	306	23	,	,	PUNCT
aiti-14589	306	24	1999	1999	NUM
aiti-14589	306	25	.	.	PUNCT
aiti-14589	307	1	[	[	X
aiti-14589	307	2	18	18	NUM
aiti-14589	307	3	]	]	PUNCT
aiti-14589	307	4	r.	r.	PROPN
aiti-14589	307	5	w.	w.	PROPN
aiti-14589	307	6	picard	picard	PROPN
aiti-14589	307	7	,	,	PUNCT
aiti-14589	307	8	e.	e.	PROPN
aiti-14589	307	9	vyzas	vyzas	PROPN
aiti-14589	307	10	,	,	PUNCT
aiti-14589	307	11	and	and	CCONJ
aiti-14589	307	12	j.	j.	PROPN
aiti-14589	307	13	healey	healey	PROPN
aiti-14589	307	14	,	,	PUNCT
aiti-14589	307	15	“	"	PUNCT
aiti-14589	307	16	toward	toward	ADP
aiti-14589	307	17	machine	machine	NOUN
aiti-14589	307	18	emotional	emotional	ADJ
aiti-14589	307	19	intelligence	intelligence	NOUN
aiti-14589	307	20	:	:	PUNCT
aiti-14589	307	21	analysis	analysis	NOUN
aiti-14589	307	22	of	of	ADP
aiti-14589	307	23	affective	affective	ADJ
aiti-14589	307	24	physiological	physiological	ADJ
aiti-14589	307	25	state	state	NOUN
aiti-14589	307	26	,	,	PUNCT
aiti-14589	307	27	”	"	PUNCT
aiti-14589	307	28	ieee	ieee	NOUN
aiti-14589	307	29	transactions	transaction	NOUN
aiti-14589	307	30	on	on	ADP
aiti-14589	307	31	pattern	pattern	NOUN
aiti-14589	307	32	analysis	analysis	NOUN
aiti-14589	307	33	and	and	CCONJ
aiti-14589	307	34	machine	machine	NOUN
aiti-14589	307	35	intelligence	intelligence	NOUN
aiti-14589	307	36	,	,	PUNCT
aiti-14589	307	37	vol	vol	NOUN
aiti-14589	307	38	.	.	PROPN
aiti-14589	307	39	23	23	NUM
aiti-14589	307	40	,	,	PUNCT
aiti-14589	307	41	no	no	INTJ
aiti-14589	307	42	.	.	NOUN
aiti-14589	307	43	10	10	NUM
aiti-14589	307	44	,	,	PUNCT
aiti-14589	307	45	pp	pp	ADJ
aiti-14589	307	46	.	.	PUNCT
aiti-14589	307	47	1175	1175	NUM
aiti-14589	307	48	-	-	SYM
aiti-14589	307	49	1191	1191	NUM
aiti-14589	307	50	,	,	PUNCT
aiti-14589	307	51	2001	2001	NUM
aiti-14589	307	52	.	.	PUNCT
aiti-14589	308	1	[	[	X
aiti-14589	308	2	19	19	NUM
aiti-14589	308	3	]	]	PUNCT
aiti-14589	308	4	m.	m.	NOUN
aiti-14589	308	5	s.	s.	PROPN
aiti-14589	308	6	bartlett	bartlett	PROPN
aiti-14589	308	7	,	,	PUNCT
aiti-14589	308	8	g.	g.	PROPN
aiti-14589	308	9	littlewort	littlewort	PROPN
aiti-14589	308	10	,	,	PUNCT
aiti-14589	308	11	i.	i.	PROPN
aiti-14589	308	12	fasel	fasel	PROPN
aiti-14589	308	13	,	,	PUNCT
aiti-14589	308	14	and	and	CCONJ
aiti-14589	308	15	j.	j.	PROPN
aiti-14589	308	16	r.	r.	PROPN
aiti-14589	308	17	movellan	movellan	PROPN
aiti-14589	308	18	,	,	PUNCT
aiti-14589	308	19	“	"	PUNCT
aiti-14589	308	20	real	real	ADJ
aiti-14589	308	21	time	time	NOUN
aiti-14589	308	22	face	face	NOUN
aiti-14589	308	23	detection	detection	NOUN
aiti-14589	308	24	and	and	CCONJ
aiti-14589	308	25	facial	facial	ADJ
aiti-14589	308	26	expression	expression	NOUN
aiti-14589	308	27	recognition	recognition	NOUN
aiti-14589	308	28	:	:	PUNCT
aiti-14589	308	29	development	development	NOUN
aiti-14589	308	30	and	and	CCONJ
aiti-14589	308	31	applications	application	NOUN
aiti-14589	308	32	to	to	ADP
aiti-14589	308	33	human	human	ADJ
aiti-14589	308	34	computer	computer	NOUN
aiti-14589	308	35	interaction	interaction	NOUN
aiti-14589	308	36	,	,	PUNCT
aiti-14589	308	37	”	"	PUNCT
aiti-14589	308	38	conference	conference	NOUN
aiti-14589	308	39	on	on	ADP
aiti-14589	308	40	computer	computer	NOUN
aiti-14589	308	41	vision	vision	NOUN
aiti-14589	308	42	and	and	CCONJ
aiti-14589	308	43	pattern	pattern	NOUN
aiti-14589	308	44	recognition	recognition	NOUN
aiti-14589	308	45	workshop	workshop	NOUN
aiti-14589	308	46	,	,	PUNCT
aiti-14589	308	47	pp	pp	ADJ
aiti-14589	308	48	.	.	PUNCT
aiti-14589	309	1	53	53	NUM
aiti-14589	309	2	-	-	SYM
aiti-14589	309	3	53	53	NUM
aiti-14589	309	4	,	,	PUNCT
aiti-14589	309	5	2003	2003	NUM
aiti-14589	309	6	.	.	PUNCT
aiti-14589	310	1	[	[	X
aiti-14589	310	2	20	20	NUM
aiti-14589	310	3	]	]	X
aiti-14589	310	4	y.	y.	PROPN
aiti-14589	310	5	s.	s.	PROPN
aiti-14589	310	6	lee	lee	PROPN
aiti-14589	310	7	and	and	CCONJ
aiti-14589	310	8	w.	w.	PROPN
aiti-14589	310	9	h.	h.	PROPN
aiti-14589	310	10	park	park	PROPN
aiti-14589	310	11	,	,	PUNCT
aiti-14589	310	12	“	"	PUNCT
aiti-14589	310	13	diagnosis	diagnosis	NOUN
aiti-14589	310	14	of	of	ADP
aiti-14589	310	15	depressive	depressive	ADJ
aiti-14589	310	16	disorder	disorder	NOUN
aiti-14589	310	17	model	model	NOUN
aiti-14589	310	18	on	on	ADP
aiti-14589	310	19	facial	facial	ADJ
aiti-14589	310	20	expression	expression	NOUN
aiti-14589	310	21	based	base	VERB
aiti-14589	310	22	on	on	ADP
aiti-14589	310	23	fast	fast	ADJ
aiti-14589	310	24	r	r	NOUN
aiti-14589	310	25	-	-	PUNCT
aiti-14589	310	26	cnn	cnn	PROPN
aiti-14589	310	27	,	,	PUNCT
aiti-14589	310	28	”	"	PUNCT
aiti-14589	310	29	diagnostics	diagnostic	NOUN
aiti-14589	310	30	,	,	PUNCT
aiti-14589	310	31	vol	vol	NOUN
aiti-14589	310	32	.	.	PROPN
aiti-14589	310	33	12	12	NUM
aiti-14589	310	34	,	,	PUNCT
aiti-14589	310	35	no	no	INTJ
aiti-14589	310	36	.	.	NOUN
aiti-14589	310	37	2	2	NUM
aiti-14589	310	38	,	,	PUNCT
aiti-14589	310	39	article	article	NOUN
aiti-14589	310	40	no	no	NOUN
aiti-14589	310	41	.	.	PROPN
aiti-14589	310	42	317	317	NUM
aiti-14589	310	43	,	,	PUNCT
aiti-14589	310	44	2022	2022	NUM
aiti-14589	310	45	.	.	PUNCT
aiti-14589	311	1	[	[	X
aiti-14589	311	2	21	21	NUM
aiti-14589	311	3	]	]	X
aiti-14589	311	4	e.	e.	PROPN
aiti-14589	311	5	bekele	bekele	PROPN
aiti-14589	311	6	,	,	PUNCT
aiti-14589	311	7	d.	d.	PROPN
aiti-14589	311	8	bian	bian	PROPN
aiti-14589	311	9	,	,	PUNCT
aiti-14589	311	10	j.	j.	PROPN
aiti-14589	311	11	peterman	peterman	PROPN
aiti-14589	311	12	,	,	PUNCT
aiti-14589	311	13	s.	s.	PROPN
aiti-14589	311	14	park	park	PROPN
aiti-14589	311	15	,	,	PUNCT
aiti-14589	311	16	and	and	CCONJ
aiti-14589	311	17	n.	n.	PROPN
aiti-14589	311	18	sarkar	sarkar	PROPN
aiti-14589	311	19	,	,	PUNCT
aiti-14589	311	20	“	"	PUNCT
aiti-14589	311	21	design	design	NOUN
aiti-14589	311	22	of	of	ADP
aiti-14589	311	23	a	a	DET
aiti-14589	311	24	virtual	virtual	ADJ
aiti-14589	311	25	reality	reality	NOUN
aiti-14589	311	26	system	system	NOUN
aiti-14589	311	27	for	for	AUX
aiti-14589	311	28	affect	affect	VERB
aiti-14589	311	29	analysis	analysis	NOUN
aiti-14589	311	30	in	in	ADP
aiti-14589	311	31	facial	facial	ADJ
aiti-14589	311	32	expressions	expression	NOUN
aiti-14589	311	33	(	(	PUNCT
aiti-14589	311	34	vr	vr	NOUN
aiti-14589	311	35	-	-	PUNCT
aiti-14589	311	36	saafe	saafe	NOUN
aiti-14589	311	37	)	)	PUNCT
aiti-14589	311	38	;	;	PUNCT
aiti-14589	311	39	application	application	NOUN
aiti-14589	311	40	to	to	ADP
aiti-14589	311	41	schizophrenia	schizophrenia	NOUN
aiti-14589	311	42	,	,	PUNCT
aiti-14589	311	43	”	"	PUNCT
aiti-14589	311	44	ieee	ieee	NOUN
aiti-14589	311	45	transactions	transaction	NOUN
aiti-14589	311	46	on	on	ADP
aiti-14589	311	47	neural	neural	ADJ
aiti-14589	311	48	systems	system	NOUN
aiti-14589	311	49	and	and	CCONJ
aiti-14589	311	50	rehabilitation	rehabilitation	NOUN
aiti-14589	311	51	engineering	engineering	NOUN
aiti-14589	311	52	,	,	PUNCT
aiti-14589	311	53	vol	vol	NOUN
aiti-14589	311	54	.	.	PROPN
aiti-14589	311	55	25	25	NUM
aiti-14589	311	56	,	,	PUNCT
aiti-14589	311	57	no	no	INTJ
aiti-14589	311	58	.	.	NOUN
aiti-14589	311	59	6	6	NUM
aiti-14589	311	60	,	,	PUNCT
aiti-14589	311	61	pp	pp	ADJ
aiti-14589	311	62	.	.	PUNCT
aiti-14589	312	1	739	739	NUM
aiti-14589	312	2	-	-	SYM
aiti-14589	312	3	749	749	NUM
aiti-14589	312	4	,	,	PUNCT
aiti-14589	312	5	2017	2017	NUM
aiti-14589	312	6	.	.	PUNCT
aiti-14589	313	1	[	[	X
aiti-14589	313	2	22	22	NUM
aiti-14589	313	3	]	]	PUNCT
aiti-14589	313	4	m.	m.	NOUN
aiti-14589	313	5	a.	a.	NOUN
aiti-14589	313	6	assari	assari	PROPN
aiti-14589	313	7	and	and	CCONJ
aiti-14589	313	8	m.	m.	NOUN
aiti-14589	313	9	rahmati	rahmati	PROPN
aiti-14589	313	10	,	,	PUNCT
aiti-14589	313	11	“	"	PUNCT
aiti-14589	313	12	driver	driver	NOUN
aiti-14589	313	13	drowsiness	drowsiness	NOUN
aiti-14589	313	14	detection	detection	NOUN
aiti-14589	313	15	using	use	VERB
aiti-14589	313	16	face	face	NOUN
aiti-14589	313	17	expression	expression	NOUN
aiti-14589	313	18	recognition	recognition	NOUN
aiti-14589	313	19	,	,	PUNCT
aiti-14589	313	20	”	"	PUNCT
aiti-14589	313	21	ieee	ieee	PROPN
aiti-14589	313	22	international	international	ADJ
aiti-14589	313	23	conference	conference	NOUN
aiti-14589	313	24	on	on	ADP
aiti-14589	313	25	signal	signal	NOUN
aiti-14589	313	26	and	and	CCONJ
aiti-14589	313	27	image	image	NOUN
aiti-14589	313	28	processing	processing	NOUN
aiti-14589	313	29	applications	application	NOUN
aiti-14589	313	30	,	,	PUNCT
aiti-14589	313	31	pp	pp	ADV
aiti-14589	313	32	.	.	PUNCT
aiti-14589	313	33	337	337	NUM
aiti-14589	313	34	-	-	SYM
aiti-14589	313	35	341	341	NUM
aiti-14589	313	36	,	,	PUNCT
aiti-14589	313	37	2011	2011	NUM
aiti-14589	313	38	.	.	PUNCT
aiti-14589	314	1	[	[	X
aiti-14589	314	2	23	23	NUM
aiti-14589	314	3	]	]	PUNCT
aiti-14589	314	4	d.	d.	PROPN
aiti-14589	314	5	hamester	hamester	PROPN
aiti-14589	314	6	,	,	PUNCT
aiti-14589	314	7	p.	p.	PROPN
aiti-14589	314	8	barros	barros	PROPN
aiti-14589	314	9	,	,	PUNCT
aiti-14589	314	10	and	and	CCONJ
aiti-14589	314	11	s.	s.	PROPN
aiti-14589	314	12	wermter	wermter	PROPN
aiti-14589	314	13	,	,	PUNCT
aiti-14589	314	14	“	"	PUNCT
aiti-14589	314	15	face	face	VERB
aiti-14589	314	16	expression	expression	NOUN
aiti-14589	314	17	recognition	recognition	NOUN
aiti-14589	314	18	with	with	ADP
aiti-14589	314	19	a	a	DET
aiti-14589	314	20	2	2	NUM
aiti-14589	314	21	-	-	PUNCT
aiti-14589	314	22	channel	channel	NOUN
aiti-14589	314	23	convolutional	convolutional	ADJ
aiti-14589	314	24	neural	neural	ADJ
aiti-14589	314	25	network	network	NOUN
aiti-14589	314	26	,	,	PUNCT
aiti-14589	314	27	”	"	PUNCT
aiti-14589	314	28	international	international	ADJ
aiti-14589	314	29	joint	joint	ADJ
aiti-14589	314	30	conference	conference	NOUN
aiti-14589	314	31	on	on	ADP
aiti-14589	314	32	neural	neural	ADJ
aiti-14589	314	33	networks	network	NOUN
aiti-14589	314	34	,	,	PUNCT
aiti-14589	314	35	pp	pp	ADJ
aiti-14589	314	36	.	.	PUNCT
aiti-14589	315	1	1	1	NUM
aiti-14589	315	2	-	-	SYM
aiti-14589	315	3	8	8	NUM
aiti-14589	315	4	,	,	PUNCT
aiti-14589	315	5	2015	2015	NUM
aiti-14589	315	6	.	.	PUNCT
aiti-14589	316	1	[	[	X
aiti-14589	316	2	24	24	NUM
aiti-14589	316	3	]	]	PUNCT
aiti-14589	316	4	a.	a.	NOUN
aiti-14589	316	5	alsubari	alsubari	PROPN
aiti-14589	316	6	,	,	PUNCT
aiti-14589	316	7	d.	d.	PROPN
aiti-14589	316	8	n.	n.	PROPN
aiti-14589	316	9	satange	satange	PROPN
aiti-14589	316	10	,	,	PUNCT
aiti-14589	316	11	and	and	CCONJ
aiti-14589	316	12	r.	r.	PROPN
aiti-14589	316	13	j.	j.	PROPN
aiti-14589	316	14	ramteke	ramteke	PROPN
aiti-14589	316	15	,	,	PUNCT
aiti-14589	316	16	“	"	PUNCT
aiti-14589	316	17	facial	facial	ADJ
aiti-14589	316	18	expression	expression	NOUN
aiti-14589	316	19	recognition	recognition	NOUN
aiti-14589	316	20	using	use	VERB
aiti-14589	316	21	wavelet	wavelet	NOUN
aiti-14589	316	22	transform	transform	NOUN
aiti-14589	316	23	and	and	CCONJ
aiti-14589	316	24	local	local	ADJ
aiti-14589	316	25	binary	binary	NOUN
aiti-14589	316	26	pattern	pattern	NOUN
aiti-14589	316	27	,	,	PUNCT
aiti-14589	316	28	”	"	PUNCT
aiti-14589	316	29	2nd	2nd	ADJ
aiti-14589	316	30	international	international	ADJ
aiti-14589	316	31	conference	conference	NOUN
aiti-14589	316	32	for	for	ADP
aiti-14589	316	33	convergence	convergence	NOUN
aiti-14589	316	34	in	in	ADP
aiti-14589	316	35	technology	technology	NOUN
aiti-14589	316	36	,	,	PUNCT
aiti-14589	316	37	pp	pp	ADP
aiti-14589	316	38	.	.	PUNCT
aiti-14589	317	1	338	338	NUM
aiti-14589	317	2	-	-	SYM
aiti-14589	317	3	342	342	NUM
aiti-14589	317	4	,	,	PUNCT
aiti-14589	317	5	2017	2017	NUM
aiti-14589	317	6	.	.	PUNCT
aiti-14589	318	1	[	[	X
aiti-14589	318	2	25	25	NUM
aiti-14589	318	3	]	]	PUNCT
aiti-14589	318	4	ö.	ö.	PROPN
aiti-14589	318	5	f.	f.	PROPN
aiti-14589	318	6	söylemez	söylemez	PROPN
aiti-14589	318	7	,	,	PUNCT
aiti-14589	318	8	b.	b.	PROPN
aiti-14589	318	9	ergen	ergen	PROPN
aiti-14589	318	10	,	,	PUNCT
aiti-14589	318	11	and	and	CCONJ
aiti-14589	318	12	n.	n.	PROPN
aiti-14589	318	13	h.	h.	PROPN
aiti-14589	318	14	söylemez	söylemez	PROPN
aiti-14589	318	15	,	,	PUNCT
aiti-14589	318	16	“	"	PUNCT
aiti-14589	318	17	a	a	DET
aiti-14589	318	18	3d	3d	NUM
aiti-14589	318	19	facial	facial	ADJ
aiti-14589	318	20	expression	expression	NOUN
aiti-14589	318	21	recognition	recognition	NOUN
aiti-14589	318	22	system	system	NOUN
aiti-14589	318	23	based	base	VERB
aiti-14589	318	24	on	on	ADP
aiti-14589	318	25	svm	svm	ADJ
aiti-14589	318	26	classifier	classifier	NOUN
aiti-14589	318	27	using	use	VERB
aiti-14589	318	28	distance	distance	NOUN
aiti-14589	318	29	based	base	VERB
aiti-14589	318	30	features	feature	NOUN
aiti-14589	318	31	,	,	PUNCT
aiti-14589	318	32	”	"	PUNCT
aiti-14589	318	33	25th	25th	ADJ
aiti-14589	318	34	signal	signal	NOUN
aiti-14589	318	35	processing	processing	NOUN
aiti-14589	318	36	and	and	CCONJ
aiti-14589	318	37	communications	communication	NOUN
aiti-14589	318	38	applications	application	NOUN
aiti-14589	318	39	conference	conference	NOUN
aiti-14589	318	40	,	,	PUNCT
aiti-14589	318	41	pp	pp	ADJ
aiti-14589	318	42	.	.	PUNCT
aiti-14589	319	1	1	1	NUM
aiti-14589	319	2	-	-	SYM
aiti-14589	319	3	3	3	NUM
aiti-14589	319	4	,	,	PUNCT
aiti-14589	319	5	2017	2017	NUM
aiti-14589	319	6	.	.	PUNCT
aiti-14589	320	1	[	[	X
aiti-14589	320	2	26	26	NUM
aiti-14589	320	3	]	]	PUNCT
aiti-14589	320	4	m.	m.	NOUN
aiti-14589	320	5	i.	i.	PROPN
aiti-14589	320	6	n.	n.	PROPN
aiti-14589	320	7	p.	p.	PROPN
aiti-14589	320	8	munasinghe	munasinghe	NOUN
aiti-14589	320	9	,	,	PUNCT
aiti-14589	320	10	“	"	PUNCT
aiti-14589	320	11	facial	facial	ADJ
aiti-14589	320	12	expression	expression	NOUN
aiti-14589	320	13	recognition	recognition	NOUN
aiti-14589	320	14	using	use	VERB
aiti-14589	320	15	facial	facial	ADJ
aiti-14589	320	16	landmarks	landmark	NOUN
aiti-14589	320	17	and	and	CCONJ
aiti-14589	320	18	random	random	ADJ
aiti-14589	320	19	forest	forest	NOUN
aiti-14589	320	20	classifier	classifier	NOUN
aiti-14589	320	21	,	,	PUNCT
aiti-14589	320	22	”	"	PUNCT
aiti-14589	320	23	ieee	ieee	PROPN
aiti-14589	320	24	/	/	SYM
aiti-14589	320	25	acis	acis	PROPN
aiti-14589	320	26	17th	17th	PROPN
aiti-14589	320	27	international	international	ADJ
aiti-14589	320	28	conference	conference	NOUN
aiti-14589	320	29	on	on	ADP
aiti-14589	320	30	computer	computer	NOUN
aiti-14589	320	31	and	and	CCONJ
aiti-14589	320	32	information	information	NOUN
aiti-14589	320	33	science	science	NOUN
aiti-14589	320	34	,	,	PUNCT
aiti-14589	320	35	pp	pp	ADP
aiti-14589	320	36	.	.	PUNCT
aiti-14589	321	1	423	423	NUM
aiti-14589	321	2	-	-	SYM
aiti-14589	321	3	427	427	NUM
aiti-14589	321	4	,	,	PUNCT
aiti-14589	321	5	2018	2018	NUM
aiti-14589	321	6	.	.	PUNCT
aiti-14589	322	1	[	[	X
aiti-14589	322	2	27	27	NUM
aiti-14589	322	3	]	]	X
aiti-14589	322	4	f.	f.	PROPN
aiti-14589	322	5	di	di	PROPN
aiti-14589	322	6	luzio	luzio	PROPN
aiti-14589	322	7	,	,	PUNCT
aiti-14589	322	8	a.	a.	NOUN
aiti-14589	322	9	rosato	rosato	NOUN
aiti-14589	322	10	,	,	PUNCT
aiti-14589	322	11	and	and	CCONJ
aiti-14589	322	12	m.	m.	NOUN
aiti-14589	322	13	panella	panella	NOUN
aiti-14589	322	14	,	,	PUNCT
aiti-14589	322	15	“	"	PUNCT
aiti-14589	322	16	a	a	DET
aiti-14589	322	17	randomized	randomized	ADJ
aiti-14589	322	18	deep	deep	ADJ
aiti-14589	322	19	neural	neural	ADJ
aiti-14589	322	20	network	network	NOUN
aiti-14589	322	21	for	for	ADP
aiti-14589	322	22	emotion	emotion	NOUN
aiti-14589	322	23	recognition	recognition	NOUN
aiti-14589	322	24	with	with	ADP
aiti-14589	322	25	landmarks	landmark	NOUN
aiti-14589	322	26	detection	detection	NOUN
aiti-14589	322	27	,	,	PUNCT
aiti-14589	322	28	”	"	PUNCT
aiti-14589	322	29	biomedical	biomedical	ADJ
aiti-14589	322	30	signal	signal	NOUN
aiti-14589	322	31	processing	processing	NOUN
aiti-14589	322	32	and	and	CCONJ
aiti-14589	322	33	control	control	NOUN
aiti-14589	322	34	,	,	PUNCT
aiti-14589	322	35	vol	vol	NOUN
aiti-14589	322	36	.	.	PROPN
aiti-14589	322	37	81	81	NUM
aiti-14589	322	38	,	,	PUNCT
aiti-14589	322	39	article	article	NOUN
aiti-14589	322	40	no	no	NOUN
aiti-14589	322	41	.	.	PROPN
aiti-14589	322	42	104418	104418	NUM
aiti-14589	322	43	,	,	PUNCT
aiti-14589	322	44	2023	2023	NUM
aiti-14589	322	45	.	.	PUNCT
aiti-14589	323	1	[	[	X
aiti-14589	323	2	28	28	NUM
aiti-14589	323	3	]	]	X
aiti-14589	323	4	s.	s.	PROPN
aiti-14589	323	5	hangaragi	hangaragi	PROPN
aiti-14589	323	6	,	,	PUNCT
aiti-14589	323	7	t.	t.	PROPN
aiti-14589	323	8	singh	singh	PROPN
aiti-14589	323	9	,	,	PUNCT
aiti-14589	323	10	and	and	CCONJ
aiti-14589	323	11	n.	n.	NOUN
aiti-14589	323	12	n	n	CCONJ
aiti-14589	323	13	,	,	PUNCT
aiti-14589	323	14	“	"	PUNCT
aiti-14589	323	15	face	face	NOUN
aiti-14589	323	16	detection	detection	NOUN
aiti-14589	323	17	and	and	CCONJ
aiti-14589	323	18	recognition	recognition	NOUN
aiti-14589	323	19	using	use	VERB
aiti-14589	323	20	face	face	NOUN
aiti-14589	323	21	mesh	mesh	NOUN
aiti-14589	323	22	and	and	CCONJ
aiti-14589	323	23	deep	deep	ADJ
aiti-14589	323	24	neural	neural	ADJ
aiti-14589	323	25	network	network	NOUN
aiti-14589	323	26	,	,	PUNCT
aiti-14589	323	27	”	"	PUNCT
aiti-14589	323	28	procedia	procedia	NOUN
aiti-14589	323	29	computer	computer	NOUN
aiti-14589	323	30	science	science	NOUN
aiti-14589	323	31	,	,	PUNCT
aiti-14589	323	32	vol	vol	NOUN
aiti-14589	323	33	.	.	PROPN
aiti-14589	323	34	218	218	NUM
aiti-14589	323	35	,	,	PUNCT
aiti-14589	323	36	pp	pp	ADJ
aiti-14589	323	37	.	.	PUNCT
aiti-14589	324	1	741	741	NUM
aiti-14589	324	2	-	-	SYM
aiti-14589	324	3	749	749	NUM
aiti-14589	324	4	,	,	PUNCT
aiti-14589	324	5	2023	2023	NUM
aiti-14589	324	6	.	.	PUNCT
aiti-14589	325	1	copyright	copyright	NOUN
aiti-14589	325	2	©	©	PROPN
aiti-14589	325	3	by	by	ADP
aiti-14589	325	4	the	the	DET
aiti-14589	325	5	authors	author	NOUN
aiti-14589	325	6	.	.	PUNCT
aiti-14589	326	1	licensee	licensee	PROPN
aiti-14589	326	2	taeti	taeti	PROPN
aiti-14589	326	3	,	,	PUNCT
aiti-14589	326	4	taiwan	taiwan	PROPN
aiti-14589	326	5	.	.	PUNCT
aiti-14589	327	1	this	this	DET
aiti-14589	327	2	article	article	NOUN
aiti-14589	327	3	is	be	AUX
aiti-14589	327	4	an	an	DET
aiti-14589	327	5	open	open	ADJ
aiti-14589	327	6	-	-	PUNCT
aiti-14589	327	7	access	access	NOUN
aiti-14589	327	8	article	article	NOUN
aiti-14589	327	9	distributed	distribute	VERB
aiti-14589	327	10	under	under	ADP
aiti-14589	327	11	the	the	DET
aiti-14589	327	12	terms	term	NOUN
aiti-14589	327	13	and	and	CCONJ
aiti-14589	327	14	conditions	condition	NOUN
aiti-14589	327	15	of	of	ADP
aiti-14589	327	16	the	the	DET
aiti-14589	327	17	creative	creative	ADJ
aiti-14589	327	18	commons	common	NOUN
aiti-14589	327	19	attribution	attribution	NOUN
aiti-14589	327	20	(	(	PUNCT
aiti-14589	327	21	cc	cc	NOUN
aiti-14589	327	22	by	by	ADP
aiti-14589	327	23	-	-	PUNCT
aiti-14589	327	24	nc	nc	NOUN
aiti-14589	327	25	)	)	PUNCT
aiti-14589	327	26	license	license	NOUN
aiti-14589	327	27	(	(	PUNCT
aiti-14589	327	28	https://creativecommons.org/licenses/by-nc/4.0/	https://creativecommons.org/licenses/by-nc/4.0/	NOUN
aiti-14589	327	29	)	)	PUNCT
aiti-14589	327	30	.	.	PUNCT