id	sid	tid	token	lemma	pos
fbem-26172	1	1	frontiers	frontier	NOUN
fbem-26172	1	2	in	in	ADP
fbem-26172	1	3	business	business	NOUN
fbem-26172	1	4	,	,	PUNCT
fbem-26172	1	5	economics	economic	NOUN
fbem-26172	1	6	and	and	CCONJ
fbem-26172	1	7	management	management	NOUN
fbem-26172	1	8	issn	issn	PROPN
fbem-26172	1	9	:	:	PUNCT
fbem-26172	1	10	2766	2766	NUM
fbem-26172	1	11	-	-	PUNCT
fbem-26172	1	12	824x	824x	NUM
fbem-26172	1	13	|	|	ADJ
fbem-26172	1	14	vol	vol	NOUN
fbem-26172	1	15	.	.	PUNCT
fbem-26172	2	1	16	16	NUM
fbem-26172	2	2	,	,	PUNCT
fbem-26172	2	3	no	no	INTJ
fbem-26172	2	4	.	.	NOUN
fbem-26172	2	5	3	3	NUM
fbem-26172	2	6	,	,	PUNCT
fbem-26172	2	7	2024	2024	NUM
fbem-26172	2	8	72	72	NUM
fbem-26172	2	9	sentiment	sentiment	NOUN
fbem-26172	2	10	analysis	analysis	NOUN
fbem-26172	2	11	and	and	CCONJ
fbem-26172	2	12	facial	facial	ADJ
fbem-26172	2	13	expression	expression	NOUN
fbem-26172	2	14	recognition	recognition	NOUN
fbem-26172	2	15	in	in	ADP
fbem-26172	2	16	customer	customer	NOUN
fbem-26172	2	17	service	service	NOUN
fbem-26172	2	18	interactions	interaction	NOUN
fbem-26172	2	19	qinrui	qinrui	PROPN
fbem-26172	2	20	hu	hu	PROPN
fbem-26172	3	1	*	*	PUNCT
fbem-26172	3	2	wenzhou	wenzhou	PROPN
fbem-26172	3	3	kean	kean	PROPN
fbem-26172	3	4	university	university	PROPN
fbem-26172	3	5	,	,	PUNCT
fbem-26172	3	6	wenzhou	wenzhou	PROPN
fbem-26172	3	7	,	,	PUNCT
fbem-26172	3	8	zhejiang	zhejiang	PROPN
fbem-26172	3	9	,	,	PUNCT
fbem-26172	3	10	325000	325000	NUM
fbem-26172	3	11	,	,	PUNCT
fbem-26172	3	12	china	china	PROPN
fbem-26172	3	13	*	*	PUNCT
fbem-26172	3	14	corresponding	correspond	VERB
fbem-26172	3	15	author	author	NOUN
fbem-26172	3	16	:	:	PUNCT
fbem-26172	3	17	qinrui	qinrui	PROPN
fbem-26172	3	18	hu	hu	PROPN
fbem-26172	3	19	(	(	PUNCT
fbem-26172	3	20	email	email	NOUN
fbem-26172	3	21	:	:	PUNCT
fbem-26172	3	22	huqinr@kean.edu	huqinr@kean.edu	NOUN
fbem-26172	3	23	)	)	PUNCT
fbem-26172	3	24	abstract	abstract	NOUN
fbem-26172	3	25	:	:	PUNCT
fbem-26172	3	26	in	in	ADP
fbem-26172	3	27	the	the	DET
fbem-26172	3	28	evolving	evolve	VERB
fbem-26172	3	29	landscape	landscape	NOUN
fbem-26172	3	30	of	of	ADP
fbem-26172	3	31	digital	digital	ADJ
fbem-26172	3	32	customer	customer	NOUN
fbem-26172	3	33	service	service	NOUN
fbem-26172	3	34	,	,	PUNCT
fbem-26172	3	35	the	the	DET
fbem-26172	3	36	need	need	NOUN
fbem-26172	3	37	for	for	ADP
fbem-26172	3	38	advanced	advanced	ADJ
fbem-26172	3	39	methods	method	NOUN
fbem-26172	3	40	to	to	PART
fbem-26172	3	41	accurately	accurately	ADV
fbem-26172	3	42	understand	understand	VERB
fbem-26172	3	43	and	and	CCONJ
fbem-26172	3	44	respond	respond	VERB
fbem-26172	3	45	to	to	ADP
fbem-26172	3	46	customer	customer	NOUN
fbem-26172	3	47	emotions	emotion	NOUN
fbem-26172	3	48	has	have	AUX
fbem-26172	3	49	become	become	VERB
fbem-26172	3	50	critical	critical	ADJ
fbem-26172	3	51	.	.	PUNCT
fbem-26172	4	1	traditional	traditional	ADJ
fbem-26172	4	2	systems	system	NOUN
fbem-26172	4	3	often	often	ADV
fbem-26172	4	4	rely	rely	VERB
fbem-26172	4	5	solely	solely	ADV
fbem-26172	4	6	on	on	ADP
fbem-26172	4	7	textual	textual	ADJ
fbem-26172	4	8	data	datum	NOUN
fbem-26172	4	9	,	,	PUNCT
fbem-26172	4	10	missing	miss	VERB
fbem-26172	4	11	non	non	ADJ
fbem-26172	4	12	-	-	ADJ
fbem-26172	4	13	verbal	verbal	ADJ
fbem-26172	4	14	cues	cue	NOUN
fbem-26172	4	15	that	that	PRON
fbem-26172	4	16	significantly	significantly	ADV
fbem-26172	4	17	contribute	contribute	VERB
fbem-26172	4	18	to	to	ADP
fbem-26172	4	19	the	the	DET
fbem-26172	4	20	customer	customer	NOUN
fbem-26172	4	21	's	's	PART
fbem-26172	4	22	emotional	emotional	ADJ
fbem-26172	4	23	state	state	NOUN
fbem-26172	4	24	.	.	PUNCT
fbem-26172	5	1	this	this	DET
fbem-26172	5	2	study	study	NOUN
fbem-26172	5	3	proposes	propose	VERB
fbem-26172	5	4	a	a	DET
fbem-26172	5	5	combined	combined	ADJ
fbem-26172	5	6	approach	approach	NOUN
fbem-26172	5	7	integrating	integrate	VERB
fbem-26172	5	8	facial	facial	ADJ
fbem-26172	5	9	expression	expression	NOUN
fbem-26172	5	10	recognition	recognition	NOUN
fbem-26172	5	11	(	(	PUNCT
fbem-26172	5	12	fer	fer	PROPN
fbem-26172	5	13	)	)	PUNCT
fbem-26172	5	14	and	and	CCONJ
fbem-26172	5	15	natural	natural	ADJ
fbem-26172	5	16	language	language	NOUN
fbem-26172	5	17	processing	processing	NOUN
fbem-26172	5	18	(	(	PUNCT
fbem-26172	5	19	nlp	nlp	NOUN
fbem-26172	5	20	)	)	PUNCT
fbem-26172	5	21	to	to	PART
fbem-26172	5	22	enhance	enhance	VERB
fbem-26172	5	23	emotion	emotion	NOUN
fbem-26172	5	24	detection	detection	NOUN
fbem-26172	5	25	accuracy	accuracy	NOUN
fbem-26172	5	26	in	in	ADP
fbem-26172	5	27	customer	customer	NOUN
fbem-26172	5	28	service	service	NOUN
fbem-26172	5	29	interactions	interaction	NOUN
fbem-26172	5	30	.	.	PUNCT
fbem-26172	6	1	the	the	DET
fbem-26172	6	2	fer	fer	PROPN
fbem-26172	6	3	component	component	NOUN
fbem-26172	6	4	employs	employ	VERB
fbem-26172	6	5	convolutional	convolutional	ADJ
fbem-26172	6	6	neural	neural	ADJ
fbem-26172	6	7	networks	network	NOUN
fbem-26172	6	8	(	(	PUNCT
fbem-26172	6	9	cnns	cnns	PROPN
fbem-26172	6	10	)	)	PUNCT
fbem-26172	6	11	to	to	PART
fbem-26172	6	12	analyze	analyze	VERB
fbem-26172	6	13	facial	facial	ADJ
fbem-26172	6	14	expressions	expression	NOUN
fbem-26172	6	15	,	,	PUNCT
fbem-26172	6	16	while	while	SCONJ
fbem-26172	6	17	the	the	DET
fbem-26172	6	18	nlp	nlp	NOUN
fbem-26172	6	19	component	component	NOUN
fbem-26172	6	20	uses	use	VERB
fbem-26172	6	21	long	long	ADJ
fbem-26172	6	22	short	short	ADJ
fbem-26172	6	23	-	-	PUNCT
fbem-26172	6	24	term	term	NOUN
fbem-26172	6	25	memory	memory	NOUN
fbem-26172	6	26	(	(	PUNCT
fbem-26172	6	27	lstm	lstm	NOUN
fbem-26172	6	28	)	)	PUNCT
fbem-26172	6	29	networks	network	NOUN
fbem-26172	6	30	to	to	PART
fbem-26172	6	31	process	process	VERB
fbem-26172	6	32	textual	textual	ADJ
fbem-26172	6	33	data	datum	NOUN
fbem-26172	6	34	.	.	PUNCT
fbem-26172	7	1	this	this	DET
fbem-26172	7	2	multimodal	multimodal	NOUN
fbem-26172	7	3	system	system	NOUN
fbem-26172	7	4	aims	aim	VERB
fbem-26172	7	5	to	to	PART
fbem-26172	7	6	provide	provide	VERB
fbem-26172	7	7	a	a	DET
fbem-26172	7	8	comprehensive	comprehensive	ADJ
fbem-26172	7	9	understanding	understanding	NOUN
fbem-26172	7	10	of	of	ADP
fbem-26172	7	11	customer	customer	NOUN
fbem-26172	7	12	emotions	emotion	NOUN
fbem-26172	7	13	by	by	ADP
fbem-26172	7	14	capturing	capture	VERB
fbem-26172	7	15	both	both	CCONJ
fbem-26172	7	16	verbal	verbal	ADJ
fbem-26172	7	17	and	and	CCONJ
fbem-26172	7	18	non	non	ADJ
fbem-26172	7	19	-	-	ADJ
fbem-26172	7	20	verbal	verbal	ADJ
fbem-26172	7	21	cues	cue	NOUN
fbem-26172	7	22	.	.	PUNCT
fbem-26172	8	1	experiments	experiment	NOUN
fbem-26172	8	2	demonstrate	demonstrate	VERB
fbem-26172	8	3	that	that	SCONJ
fbem-26172	8	4	the	the	DET
fbem-26172	8	5	integrated	integrate	VERB
fbem-26172	8	6	fer	fer	NOUN
fbem-26172	8	7	and	and	CCONJ
fbem-26172	8	8	nlp	nlp	PROPN
fbem-26172	8	9	model	model	NOUN
fbem-26172	8	10	significantly	significantly	ADV
fbem-26172	8	11	outperforms	outperform	VERB
fbem-26172	8	12	standalone	standalone	ADJ
fbem-26172	8	13	models	model	NOUN
fbem-26172	8	14	,	,	PUNCT
fbem-26172	8	15	achieving	achieve	VERB
fbem-26172	8	16	an	an	DET
fbem-26172	8	17	accuracy	accuracy	NOUN
fbem-26172	8	18	of	of	ADP
fbem-26172	8	19	92.3	92.3	NUM
fbem-26172	8	20	%	%	NOUN
fbem-26172	8	21	,	,	PUNCT
fbem-26172	8	22	compared	compare	VERB
fbem-26172	8	23	to	to	ADP
fbem-26172	8	24	85.2	85.2	NUM
fbem-26172	8	25	%	%	NOUN
fbem-26172	8	26	for	for	ADP
fbem-26172	8	27	fer	fer	NOUN
fbem-26172	8	28	-	-	PUNCT
fbem-26172	8	29	only	only	ADV
fbem-26172	8	30	and	and	CCONJ
fbem-26172	8	31	87.4	87.4	NUM
fbem-26172	8	32	%	%	NOUN
fbem-26172	8	33	for	for	ADP
fbem-26172	8	34	nlp	nlp	ADJ
fbem-26172	8	35	-	-	PUNCT
fbem-26172	8	36	only	only	ADJ
fbem-26172	8	37	models	model	NOUN
fbem-26172	8	38	.	.	PUNCT
fbem-26172	9	1	the	the	DET
fbem-26172	9	2	results	result	NOUN
fbem-26172	9	3	highlight	highlight	VERB
fbem-26172	9	4	the	the	DET
fbem-26172	9	5	benefits	benefit	NOUN
fbem-26172	9	6	of	of	ADP
fbem-26172	9	7	a	a	DET
fbem-26172	9	8	multimodal	multimodal	NOUN
fbem-26172	9	9	approach	approach	NOUN
fbem-26172	9	10	,	,	PUNCT
fbem-26172	9	11	showing	show	VERB
fbem-26172	9	12	substantial	substantial	ADJ
fbem-26172	9	13	improvements	improvement	NOUN
fbem-26172	9	14	in	in	ADP
fbem-26172	9	15	both	both	DET
fbem-26172	9	16	training	training	NOUN
fbem-26172	9	17	and	and	CCONJ
fbem-26172	9	18	validation	validation	NOUN
fbem-26172	9	19	performance	performance	NOUN
fbem-26172	9	20	.	.	PUNCT
fbem-26172	10	1	this	this	DET
fbem-26172	10	2	study	study	NOUN
fbem-26172	10	3	also	also	ADV
fbem-26172	10	4	compares	compare	VERB
fbem-26172	10	5	the	the	DET
fbem-26172	10	6	proposed	propose	VERB
fbem-26172	10	7	model	model	NOUN
fbem-26172	10	8	with	with	ADP
fbem-26172	10	9	other	other	ADJ
fbem-26172	10	10	state	state	NOUN
fbem-26172	10	11	-	-	PUNCT
fbem-26172	10	12	of	of	ADP
fbem-26172	10	13	-	-	PUNCT
fbem-26172	10	14	the	the	DET
fbem-26172	10	15	-	-	PUNCT
fbem-26172	10	16	art	art	NOUN
fbem-26172	10	17	models	model	NOUN
fbem-26172	10	18	such	such	ADJ
fbem-26172	10	19	as	as	ADP
fbem-26172	10	20	the	the	DET
fbem-26172	10	21	deep	deep	ADJ
fbem-26172	10	22	learning	learning	NOUN
fbem-26172	10	23	assisted	assist	VERB
fbem-26172	10	24	semantic	semantic	ADJ
fbem-26172	10	25	text	text	NOUN
fbem-26172	10	26	analysis	analysis	NOUN
fbem-26172	10	27	(	(	PUNCT
fbem-26172	10	28	dlsta	dlsta	NOUN
fbem-26172	10	29	)	)	PUNCT
fbem-26172	10	30	and	and	CCONJ
fbem-26172	10	31	multimodal	multimodal	ADJ
fbem-26172	10	32	emotion	emotion	NOUN
fbem-26172	10	33	recognition	recognition	NOUN
fbem-26172	10	34	using	use	VERB
fbem-26172	10	35	deep	deep	ADJ
fbem-26172	10	36	belief	belief	NOUN
fbem-26172	10	37	networks	network	NOUN
fbem-26172	10	38	(	(	PUNCT
fbem-26172	10	39	dbn	dbn	PROPN
fbem-26172	10	40	)	)	PUNCT
fbem-26172	10	41	.	.	PUNCT
fbem-26172	11	1	while	while	SCONJ
fbem-26172	11	2	dlsta	dlsta	NOUN
fbem-26172	11	3	achieves	achieve	VERB
fbem-26172	11	4	higher	high	ADJ
fbem-26172	11	5	accuracy	accuracy	NOUN
fbem-26172	11	6	in	in	ADP
fbem-26172	11	7	text	text	NOUN
fbem-26172	11	8	-	-	PUNCT
fbem-26172	11	9	based	base	VERB
fbem-26172	11	10	emotion	emotion	NOUN
fbem-26172	11	11	detection	detection	NOUN
fbem-26172	11	12	,	,	PUNCT
fbem-26172	11	13	and	and	CCONJ
fbem-26172	11	14	dbns	dbns	NOUN
fbem-26172	11	15	provide	provide	VERB
fbem-26172	11	16	robust	robust	ADJ
fbem-26172	11	17	emotion	emotion	NOUN
fbem-26172	11	18	classification	classification	NOUN
fbem-26172	11	19	by	by	ADP
fbem-26172	11	20	integrating	integrate	VERB
fbem-26172	11	21	various	various	ADJ
fbem-26172	11	22	modalities	modality	NOUN
fbem-26172	11	23	,	,	PUNCT
fbem-26172	11	24	our	our	PRON
fbem-26172	11	25	model	model	NOUN
fbem-26172	11	26	effectively	effectively	ADV
fbem-26172	11	27	balances	balance	VERB
fbem-26172	11	28	the	the	DET
fbem-26172	11	29	strengths	strength	NOUN
fbem-26172	11	30	of	of	ADP
fbem-26172	11	31	both	both	CCONJ
fbem-26172	11	32	visual	visual	ADJ
fbem-26172	11	33	and	and	CCONJ
fbem-26172	11	34	textual	textual	ADJ
fbem-26172	11	35	data	datum	NOUN
fbem-26172	11	36	.	.	PUNCT
fbem-26172	12	1	the	the	DET
fbem-26172	12	2	findings	finding	NOUN
fbem-26172	12	3	suggest	suggest	VERB
fbem-26172	12	4	that	that	SCONJ
fbem-26172	12	5	integrating	integrate	VERB
fbem-26172	12	6	fer	fer	PROPN
fbem-26172	12	7	and	and	CCONJ
fbem-26172	12	8	nlp	nlp	NOUN
fbem-26172	12	9	can	can	AUX
fbem-26172	12	10	significantly	significantly	ADV
fbem-26172	12	11	enhance	enhance	VERB
fbem-26172	12	12	the	the	DET
fbem-26172	12	13	quality	quality	NOUN
fbem-26172	12	14	of	of	ADP
fbem-26172	12	15	customer	customer	NOUN
fbem-26172	12	16	service	service	NOUN
fbem-26172	12	17	by	by	ADP
fbem-26172	12	18	enabling	enable	VERB
fbem-26172	12	19	more	more	ADV
fbem-26172	12	20	empathetic	empathetic	ADJ
fbem-26172	12	21	and	and	CCONJ
fbem-26172	12	22	effective	effective	ADJ
fbem-26172	12	23	interactions	interaction	NOUN
fbem-26172	12	24	.	.	PUNCT
fbem-26172	13	1	future	future	ADJ
fbem-26172	13	2	work	work	NOUN
fbem-26172	13	3	will	will	AUX
fbem-26172	13	4	focus	focus	VERB
fbem-26172	13	5	on	on	ADP
fbem-26172	13	6	optimizing	optimize	VERB
fbem-26172	13	7	computational	computational	ADJ
fbem-26172	13	8	efficiency	efficiency	NOUN
fbem-26172	13	9	,	,	PUNCT
fbem-26172	13	10	addressing	address	VERB
fbem-26172	13	11	data	datum	NOUN
fbem-26172	13	12	variability	variability	NOUN
fbem-26172	13	13	,	,	PUNCT
fbem-26172	13	14	and	and	CCONJ
fbem-26172	13	15	ensuring	ensure	VERB
fbem-26172	13	16	adaptability	adaptability	NOUN
fbem-26172	13	17	across	across	ADP
fbem-26172	13	18	diverse	diverse	ADJ
fbem-26172	13	19	customer	customer	NOUN
fbem-26172	13	20	service	service	NOUN
fbem-26172	13	21	scenarios	scenario	NOUN
fbem-26172	13	22	.	.	PUNCT
fbem-26172	14	1	keywords	keyword	NOUN
fbem-26172	14	2	:	:	PUNCT
fbem-26172	14	3	facial	facial	ADJ
fbem-26172	14	4	expression	expression	NOUN
fbem-26172	14	5	recognition	recognition	NOUN
fbem-26172	14	6	(	(	PUNCT
fbem-26172	14	7	fer	fer	PROPN
fbem-26172	14	8	)	)	PUNCT
fbem-26172	14	9	,	,	PUNCT
fbem-26172	14	10	natural	natural	ADJ
fbem-26172	14	11	language	language	NOUN
fbem-26172	14	12	processing	processing	NOUN
fbem-26172	14	13	(	(	PUNCT
fbem-26172	14	14	nlp	nlp	NOUN
fbem-26172	14	15	)	)	PUNCT
fbem-26172	14	16	,	,	PUNCT
fbem-26172	14	17	combined	combined	ADJ
fbem-26172	14	18	model	model	NOUN
fbem-26172	14	19	approach	approach	NOUN
fbem-26172	14	20	,	,	PUNCT
fbem-26172	14	21	customer	customer	NOUN
fbem-26172	14	22	service	service	NOUN
fbem-26172	14	23	,	,	PUNCT
fbem-26172	14	24	statistic	statistic	ADJ
fbem-26172	14	25	analysis	analysis	NOUN
fbem-26172	14	26	.	.	PUNCT
fbem-26172	15	1	1	1	X
fbem-26172	15	2	.	.	X
fbem-26172	15	3	introduction	introduction	NOUN
fbem-26172	15	4	1.1	1.1	NUM
fbem-26172	15	5	.	.	PUNCT
fbem-26172	16	1	background	background	NOUN
fbem-26172	16	2	the	the	DET
fbem-26172	16	3	increasing	increase	VERB
fbem-26172	16	4	reliance	reliance	NOUN
fbem-26172	16	5	on	on	ADP
fbem-26172	16	6	digital	digital	ADJ
fbem-26172	16	7	communication	communication	NOUN
fbem-26172	16	8	channels	channel	NOUN
fbem-26172	16	9	in	in	ADP
fbem-26172	16	10	customer	customer	NOUN
fbem-26172	16	11	service	service	NOUN
fbem-26172	16	12	has	have	AUX
fbem-26172	16	13	highlighted	highlight	VERB
fbem-26172	16	14	the	the	DET
fbem-26172	16	15	need	need	NOUN
fbem-26172	16	16	for	for	ADP
fbem-26172	16	17	more	more	ADV
fbem-26172	16	18	advanced	advanced	ADJ
fbem-26172	16	19	methods	method	NOUN
fbem-26172	16	20	of	of	ADP
fbem-26172	16	21	understanding	understanding	NOUN
fbem-26172	16	22	and	and	CCONJ
fbem-26172	16	23	responding	respond	VERB
fbem-26172	16	24	to	to	ADP
fbem-26172	16	25	customer	customer	NOUN
fbem-26172	16	26	emotions	emotion	NOUN
fbem-26172	16	27	.	.	PUNCT
fbem-26172	17	1	traditional	traditional	ADJ
fbem-26172	17	2	customer	customer	NOUN
fbem-26172	17	3	service	service	NOUN
fbem-26172	17	4	systems	system	NOUN
fbem-26172	17	5	primarily	primarily	ADV
fbem-26172	17	6	rely	rely	VERB
fbem-26172	17	7	on	on	ADP
fbem-26172	17	8	textual	textual	ADJ
fbem-26172	17	9	data	datum	NOUN
fbem-26172	17	10	to	to	PART
fbem-26172	17	11	gauge	gauge	VERB
fbem-26172	17	12	customer	customer	NOUN
fbem-26172	17	13	sentiment	sentiment	NOUN
fbem-26172	17	14	,	,	PUNCT
fbem-26172	17	15	often	often	ADV
fbem-26172	17	16	missing	miss	VERB
fbem-26172	17	17	the	the	DET
fbem-26172	17	18	critical	critical	ADJ
fbem-26172	17	19	non	non	ADJ
fbem-26172	17	20	-	-	ADJ
fbem-26172	17	21	verbal	verbal	ADJ
fbem-26172	17	22	cues	cue	NOUN
fbem-26172	17	23	that	that	PRON
fbem-26172	17	24	provide	provide	VERB
fbem-26172	17	25	a	a	DET
fbem-26172	17	26	fuller	full	ADJ
fbem-26172	17	27	picture	picture	NOUN
fbem-26172	17	28	of	of	ADP
fbem-26172	17	29	the	the	DET
fbem-26172	17	30	customer	customer	NOUN
fbem-26172	17	31	's	's	PART
fbem-26172	17	32	emotional	emotional	ADJ
fbem-26172	17	33	state	state	NOUN
fbem-26172	17	34	.	.	PUNCT
fbem-26172	18	1	the	the	DET
fbem-26172	18	2	integration	integration	NOUN
fbem-26172	18	3	of	of	ADP
fbem-26172	18	4	facial	facial	ADJ
fbem-26172	18	5	expression	expression	NOUN
fbem-26172	18	6	recognition	recognition	NOUN
fbem-26172	18	7	(	(	PUNCT
fbem-26172	18	8	fer	fer	PROPN
fbem-26172	18	9	)	)	PUNCT
fbem-26172	18	10	and	and	CCONJ
fbem-26172	18	11	natural	natural	ADJ
fbem-26172	18	12	language	language	NOUN
fbem-26172	18	13	processing	processing	NOUN
fbem-26172	18	14	(	(	PUNCT
fbem-26172	18	15	nlp	nlp	NOUN
fbem-26172	18	16	)	)	PUNCT
fbem-26172	18	17	offers	offer	VERB
fbem-26172	18	18	a	a	DET
fbem-26172	18	19	promising	promising	ADJ
fbem-26172	18	20	solution	solution	NOUN
fbem-26172	18	21	to	to	ADP
fbem-26172	18	22	this	this	DET
fbem-26172	18	23	challenge	challenge	NOUN
fbem-26172	18	24	by	by	ADP
fbem-26172	18	25	combining	combine	VERB
fbem-26172	18	26	visual	visual	ADJ
fbem-26172	18	27	and	and	CCONJ
fbem-26172	18	28	textual	textual	ADJ
fbem-26172	18	29	data	datum	NOUN
fbem-26172	18	30	to	to	PART
fbem-26172	18	31	enhance	enhance	VERB
fbem-26172	18	32	the	the	DET
fbem-26172	18	33	accuracy	accuracy	NOUN
fbem-26172	18	34	of	of	ADP
fbem-26172	18	35	emotion	emotion	NOUN
fbem-26172	18	36	detection	detection	NOUN
fbem-26172	18	37	.	.	PUNCT
fbem-26172	19	1	1.2	1.2	NUM
fbem-26172	19	2	.	.	PUNCT
fbem-26172	19	3	motivation	motivation	NOUN
fbem-26172	19	4	understanding	understand	VERB
fbem-26172	19	5	customer	customer	NOUN
fbem-26172	19	6	emotions	emotion	NOUN
fbem-26172	19	7	accurately	accurately	ADV
fbem-26172	19	8	is	be	AUX
fbem-26172	19	9	crucial	crucial	ADJ
fbem-26172	19	10	for	for	ADP
fbem-26172	19	11	providing	provide	VERB
fbem-26172	19	12	high	high	ADJ
fbem-26172	19	13	-	-	PUNCT
fbem-26172	19	14	quality	quality	NOUN
fbem-26172	19	15	customer	customer	NOUN
fbem-26172	19	16	service	service	NOUN
fbem-26172	19	17	.	.	PUNCT
fbem-26172	20	1	emotions	emotion	NOUN
fbem-26172	20	2	influence	influence	VERB
fbem-26172	20	3	customer	customer	NOUN
fbem-26172	20	4	satisfaction	satisfaction	NOUN
fbem-26172	20	5	,	,	PUNCT
fbem-26172	20	6	loyalty	loyalty	NOUN
fbem-26172	20	7	,	,	PUNCT
fbem-26172	20	8	and	and	CCONJ
fbem-26172	20	9	overall	overall	ADJ
fbem-26172	20	10	experience	experience	NOUN
fbem-26172	20	11	.	.	PUNCT
fbem-26172	21	1	misinterpreting	misinterpret	VERB
fbem-26172	21	2	a	a	DET
fbem-26172	21	3	customer	customer	NOUN
fbem-26172	21	4	's	's	PART
fbem-26172	21	5	emotional	emotional	ADJ
fbem-26172	21	6	state	state	NOUN
fbem-26172	21	7	can	can	AUX
fbem-26172	21	8	lead	lead	VERB
fbem-26172	21	9	to	to	ADP
fbem-26172	21	10	inadequate	inadequate	ADJ
fbem-26172	21	11	responses	response	NOUN
fbem-26172	21	12	,	,	PUNCT
fbem-26172	21	13	potentially	potentially	ADV
fbem-26172	21	14	escalating	escalate	VERB
fbem-26172	21	15	frustration	frustration	NOUN
fbem-26172	21	16	and	and	CCONJ
fbem-26172	21	17	dissatisfaction	dissatisfaction	NOUN
fbem-26172	21	18	.	.	PUNCT
fbem-26172	22	1	by	by	ADP
fbem-26172	22	2	integrating	integrate	VERB
fbem-26172	22	3	fer	fer	PROPN
fbem-26172	22	4	and	and	CCONJ
fbem-26172	22	5	nlp	nlp	NOUN
fbem-26172	22	6	,	,	PUNCT
fbem-26172	22	7	customer	customer	NOUN
fbem-26172	22	8	service	service	NOUN
fbem-26172	22	9	systems	system	NOUN
fbem-26172	22	10	can	can	AUX
fbem-26172	22	11	achieve	achieve	VERB
fbem-26172	22	12	a	a	DET
fbem-26172	22	13	more	more	ADV
fbem-26172	22	14	holistic	holistic	ADJ
fbem-26172	22	15	understanding	understanding	NOUN
fbem-26172	22	16	of	of	ADP
fbem-26172	22	17	customer	customer	NOUN
fbem-26172	22	18	emotions	emotion	NOUN
fbem-26172	22	19	,	,	PUNCT
fbem-26172	22	20	leading	lead	VERB
fbem-26172	22	21	to	to	ADP
fbem-26172	22	22	more	more	ADV
fbem-26172	22	23	empathetic	empathetic	ADJ
fbem-26172	22	24	and	and	CCONJ
fbem-26172	22	25	effective	effective	ADJ
fbem-26172	22	26	interactions	interaction	NOUN
fbem-26172	22	27	.	.	PUNCT
fbem-26172	23	1	enhanced	enhance	VERB
fbem-26172	23	2	customer	customer	NOUN
fbem-26172	23	3	understanding	understanding	NOUN
fbem-26172	23	4	:	:	PUNCT
fbem-26172	23	5	combining	combine	VERB
fbem-26172	23	6	facial	facial	ADJ
fbem-26172	23	7	expressions	expression	NOUN
fbem-26172	23	8	with	with	ADP
fbem-26172	23	9	textual	textual	ADJ
fbem-26172	23	10	sentiment	sentiment	NOUN
fbem-26172	23	11	provides	provide	VERB
fbem-26172	23	12	a	a	DET
fbem-26172	23	13	richer	rich	ADJ
fbem-26172	23	14	understanding	understanding	NOUN
fbem-26172	23	15	of	of	ADP
fbem-26172	23	16	customer	customer	NOUN
fbem-26172	23	17	emotions	emotion	NOUN
fbem-26172	23	18	,	,	PUNCT
fbem-26172	23	19	allowing	allow	VERB
fbem-26172	23	20	for	for	ADP
fbem-26172	23	21	more	more	ADV
fbem-26172	23	22	nuanced	nuanced	ADJ
fbem-26172	23	23	responses	response	NOUN
fbem-26172	23	24	.	.	PUNCT
fbem-26172	24	1	improved	improved	ADJ
fbem-26172	24	2	customer	customer	NOUN
fbem-26172	24	3	satisfaction	satisfaction	NOUN
fbem-26172	24	4	:	:	PUNCT
fbem-26172	24	5	accurately	accurately	ADV
fbem-26172	24	6	identifying	identify	VERB
fbem-26172	24	7	and	and	CCONJ
fbem-26172	24	8	addressing	address	VERB
fbem-26172	24	9	customer	customer	NOUN
fbem-26172	24	10	emotions	emotion	NOUN
fbem-26172	24	11	can	can	AUX
fbem-26172	24	12	enhance	enhance	VERB
fbem-26172	24	13	satisfaction	satisfaction	NOUN
fbem-26172	24	14	and	and	CCONJ
fbem-26172	24	15	loyalty	loyalty	NOUN
fbem-26172	24	16	.	.	PUNCT
fbem-26172	25	1	efficiency	efficiency	NOUN
fbem-26172	25	2	in	in	ADP
fbem-26172	25	3	service	service	NOUN
fbem-26172	25	4	delivery	delivery	NOUN
fbem-26172	25	5	:	:	PUNCT
fbem-26172	25	6	automating	automate	VERB
fbem-26172	25	7	emotion	emotion	NOUN
fbem-26172	25	8	recognition	recognition	NOUN
fbem-26172	25	9	through	through	ADP
fbem-26172	25	10	fer	fer	PROPN
fbem-26172	25	11	and	and	CCONJ
fbem-26172	25	12	nlp	nlp	NOUN
fbem-26172	25	13	can	can	AUX
fbem-26172	25	14	streamline	streamline	VERB
fbem-26172	25	15	customer	customer	NOUN
fbem-26172	25	16	service	service	NOUN
fbem-26172	25	17	processes	process	NOUN
fbem-26172	25	18	,	,	PUNCT
fbem-26172	25	19	allowing	allow	VERB
fbem-26172	25	20	agents	agent	NOUN
fbem-26172	25	21	to	to	PART
fbem-26172	25	22	focus	focus	VERB
fbem-26172	25	23	on	on	ADP
fbem-26172	25	24	more	more	ADJ
fbem-26172	25	25	complex	complex	ADJ
fbem-26172	25	26	issues	issue	NOUN
fbem-26172	25	27	.	.	PUNCT
fbem-26172	26	1	2	2	X
fbem-26172	26	2	.	.	X
fbem-26172	26	3	problem	problem	NOUN
fbem-26172	26	4	formulation	formulation	NOUN
fbem-26172	26	5	and	and	CCONJ
fbem-26172	26	6	model	model	NOUN
fbem-26172	26	7	construction	construction	NOUN
fbem-26172	26	8	2.1	2.1	NUM
fbem-26172	26	9	.	.	PUNCT
fbem-26172	27	1	problem	problem	NOUN
fbem-26172	27	2	description	description	NOUN
fbem-26172	27	3	integrating	integrate	VERB
fbem-26172	27	4	facial	facial	ADJ
fbem-26172	27	5	expression	expression	NOUN
fbem-26172	27	6	recognition	recognition	NOUN
fbem-26172	27	7	(	(	PUNCT
fbem-26172	27	8	fer	fer	PROPN
fbem-26172	27	9	)	)	PUNCT
fbem-26172	27	10	and	and	CCONJ
fbem-26172	27	11	natural	natural	ADJ
fbem-26172	27	12	language	language	NOUN
fbem-26172	27	13	processing	processing	NOUN
fbem-26172	27	14	(	(	PUNCT
fbem-26172	27	15	nlp	nlp	NOUN
fbem-26172	27	16	)	)	PUNCT
fbem-26172	27	17	for	for	ADP
fbem-26172	27	18	customer	customer	NOUN
fbem-26172	27	19	service	service	NOUN
fbem-26172	27	20	interactions	interaction	NOUN
fbem-26172	27	21	involves	involve	VERB
fbem-26172	27	22	recognizing	recognize	VERB
fbem-26172	27	23	and	and	CCONJ
fbem-26172	27	24	interpreting	interpret	VERB
fbem-26172	27	25	emotions	emotion	NOUN
fbem-26172	27	26	from	from	ADP
fbem-26172	27	27	both	both	PRON
fbem-26172	27	28	facial	facial	ADJ
fbem-26172	27	29	expressions	expression	NOUN
fbem-26172	27	30	and	and	CCONJ
fbem-26172	27	31	textual	textual	ADJ
fbem-26172	27	32	data	datum	NOUN
fbem-26172	27	33	.	.	PUNCT
fbem-26172	28	1	the	the	DET
fbem-26172	28	2	primary	primary	ADJ
fbem-26172	28	3	goal	goal	NOUN
fbem-26172	28	4	is	be	AUX
fbem-26172	28	5	to	to	PART
fbem-26172	28	6	create	create	VERB
fbem-26172	28	7	a	a	DET
fbem-26172	28	8	comprehensive	comprehensive	ADJ
fbem-26172	28	9	emotion	emotion	NOUN
fbem-26172	28	10	recognition	recognition	NOUN
fbem-26172	28	11	system	system	NOUN
fbem-26172	28	12	that	that	PRON
fbem-26172	28	13	can	can	AUX
fbem-26172	28	14	enhance	enhance	VERB
fbem-26172	28	15	the	the	DET
fbem-26172	28	16	quality	quality	NOUN
fbem-26172	28	17	of	of	ADP
fbem-26172	28	18	customer	customer	NOUN
fbem-26172	28	19	service	service	NOUN
fbem-26172	28	20	by	by	ADP
fbem-26172	28	21	providing	provide	VERB
fbem-26172	28	22	real	real	ADJ
fbem-26172	28	23	-	-	PUNCT
fbem-26172	28	24	time	time	NOUN
fbem-26172	28	25	insights	insight	NOUN
fbem-26172	28	26	into	into	ADP
fbem-26172	28	27	customer	customer	NOUN
fbem-26172	28	28	emotions	emotion	NOUN
fbem-26172	28	29	.	.	PUNCT
fbem-26172	29	1	2.1.1	2.1.1	X
fbem-26172	29	2	.	.	PUNCT
fbem-26172	29	3	facial	facial	ADJ
fbem-26172	29	4	expression	expression	NOUN
fbem-26172	29	5	recognition	recognition	NOUN
fbem-26172	29	6	(	(	PUNCT
fbem-26172	29	7	fer	fer	NOUN
fbem-26172	29	8	)	)	PUNCT
fbem-26172	29	9	facial	facial	ADJ
fbem-26172	29	10	expression	expression	NOUN
fbem-26172	29	11	recognition	recognition	NOUN
fbem-26172	29	12	(	(	PUNCT
fbem-26172	29	13	fer	fer	PROPN
fbem-26172	29	14	)	)	PUNCT
fbem-26172	29	15	focuses	focus	VERB
fbem-26172	29	16	on	on	ADP
fbem-26172	29	17	analyzing	analyze	VERB
fbem-26172	29	18	facial	facial	ADJ
fbem-26172	29	19	movements	movement	NOUN
fbem-26172	29	20	to	to	PART
fbem-26172	29	21	detect	detect	VERB
fbem-26172	29	22	emotions	emotion	NOUN
fbem-26172	29	23	.	.	PUNCT
fbem-26172	30	1	the	the	DET
fbem-26172	30	2	challenge	challenge	NOUN
fbem-26172	30	3	lies	lie	VERB
fbem-26172	30	4	in	in	ADP
fbem-26172	30	5	accurately	accurately	ADV
fbem-26172	30	6	interpreting	interpret	VERB
fbem-26172	30	7	facial	facial	ADJ
fbem-26172	30	8	expressions	expression	NOUN
fbem-26172	30	9	under	under	ADP
fbem-26172	30	10	varying	vary	VERB
fbem-26172	30	11	conditions	condition	NOUN
fbem-26172	30	12	such	such	ADJ
fbem-26172	30	13	as	as	ADP
fbem-26172	30	14	lighting	lighting	NOUN
fbem-26172	30	15	,	,	PUNCT
fbem-26172	30	16	head	head	NOUN
fbem-26172	30	17	poses	pose	NOUN
fbem-26172	30	18	,	,	PUNCT
fbem-26172	30	19	and	and	CCONJ
fbem-26172	30	20	occlusions	occlusion	NOUN
fbem-26172	30	21	.	.	PUNCT
fbem-26172	31	1	the	the	DET
fbem-26172	31	2	fer	fer	PROPN
fbem-26172	31	3	system	system	NOUN
fbem-26172	31	4	must	must	AUX
fbem-26172	31	5	be	be	AUX
fbem-26172	31	6	robust	robust	ADJ
fbem-26172	31	7	enough	enough	ADV
fbem-26172	31	8	to	to	PART
fbem-26172	31	9	handle	handle	VERB
fbem-26172	31	10	these	these	DET
fbem-26172	31	11	variations	variation	NOUN
fbem-26172	31	12	and	and	CCONJ
fbem-26172	31	13	provide	provide	VERB
fbem-26172	31	14	reliable	reliable	ADJ
fbem-26172	31	15	emotion	emotion	NOUN
fbem-26172	31	16	detection	detection	NOUN
fbem-26172	31	17	.	.	PUNCT
fbem-26172	32	1	input	input	NOUN
fbem-26172	32	2	:	:	PUNCT
fbem-26172	32	3	real	real	ADJ
fbem-26172	32	4	-	-	PUNCT
fbem-26172	32	5	time	time	NOUN
fbem-26172	32	6	video	video	NOUN
fbem-26172	32	7	streams	stream	NOUN
fbem-26172	32	8	or	or	CCONJ
fbem-26172	32	9	recorded	record	VERB
fbem-26172	32	10	videos	video	NOUN
fbem-26172	32	11	of	of	ADP
fbem-26172	32	12	customer	customer	NOUN
fbem-26172	32	13	interactions	interaction	NOUN
fbem-26172	32	14	.	.	PUNCT
fbem-26172	33	1	output	output	NOUN
fbem-26172	33	2	:	:	PUNCT
fbem-26172	33	3	detected	detect	VERB
fbem-26172	33	4	emotions	emotion	NOUN
fbem-26172	33	5	such	such	ADJ
fbem-26172	33	6	as	as	ADP
fbem-26172	33	7	happiness	happiness	NOUN
fbem-26172	33	8	,	,	PUNCT
fbem-26172	33	9	sadness	sadness	NOUN
fbem-26172	33	10	,	,	PUNCT
fbem-26172	33	11	anger	anger	NOUN
fbem-26172	33	12	,	,	PUNCT
fbem-26172	33	13	surprise	surprise	NOUN
fbem-26172	33	14	,	,	PUNCT
fbem-26172	33	15	and	and	CCONJ
fbem-26172	33	16	neutrality	neutrality	NOUN
fbem-26172	33	17	.	.	PUNCT
fbem-26172	34	1	challenges	challenge	NOUN
fbem-26172	34	2	:	:	PUNCT
fbem-26172	34	3	variability	variability	NOUN
fbem-26172	34	4	in	in	ADP
fbem-26172	34	5	facial	facial	ADJ
fbem-26172	34	6	expressions	expression	NOUN
fbem-26172	34	7	,	,	PUNCT
fbem-26172	34	8	lighting	light	VERB
fbem-26172	34	9	73	73	NUM
fbem-26172	34	10	conditions	condition	NOUN
fbem-26172	34	11	,	,	PUNCT
fbem-26172	34	12	occlusions	occlusion	NOUN
fbem-26172	34	13	,	,	PUNCT
fbem-26172	34	14	and	and	CCONJ
fbem-26172	34	15	head	head	NOUN
fbem-26172	34	16	poses	pose	NOUN
fbem-26172	34	17	.	.	PUNCT
fbem-26172	35	1	2.1.2	2.1.2	X
fbem-26172	35	2	.	.	NOUN
fbem-26172	35	3	natural	natural	ADJ
fbem-26172	35	4	language	language	NOUN
fbem-26172	35	5	processing	processing	NOUN
fbem-26172	35	6	(	(	PUNCT
fbem-26172	35	7	nlp	nlp	NOUN
fbem-26172	35	8	)	)	PUNCT
fbem-26172	35	9	natural	natural	ADJ
fbem-26172	35	10	language	language	NOUN
fbem-26172	35	11	processing	processing	NOUN
fbem-26172	35	12	(	(	PUNCT
fbem-26172	35	13	nlp	nlp	NOUN
fbem-26172	35	14	)	)	PUNCT
fbem-26172	35	15	is	be	AUX
fbem-26172	35	16	used	use	VERB
fbem-26172	35	17	to	to	PART
fbem-26172	35	18	analyze	analyze	VERB
fbem-26172	35	19	textual	textual	ADJ
fbem-26172	35	20	data	datum	NOUN
fbem-26172	35	21	from	from	ADP
fbem-26172	35	22	customer	customer	NOUN
fbem-26172	35	23	interactions	interaction	NOUN
fbem-26172	35	24	such	such	ADJ
fbem-26172	35	25	as	as	ADP
fbem-26172	35	26	chat	chat	NOUN
fbem-26172	35	27	messages	message	NOUN
fbem-26172	35	28	,	,	PUNCT
fbem-26172	35	29	emails	email	NOUN
fbem-26172	35	30	,	,	PUNCT
fbem-26172	35	31	and	and	CCONJ
fbem-26172	35	32	social	social	ADJ
fbem-26172	35	33	media	medium	NOUN
fbem-26172	35	34	posts	post	NOUN
fbem-26172	35	35	.	.	PUNCT
fbem-26172	36	1	the	the	DET
fbem-26172	36	2	goal	goal	NOUN
fbem-26172	36	3	is	be	AUX
fbem-26172	36	4	to	to	PART
fbem-26172	36	5	extract	extract	VERB
fbem-26172	36	6	and	and	CCONJ
fbem-26172	36	7	classify	classify	VERB
fbem-26172	36	8	the	the	DET
fbem-26172	36	9	emotional	emotional	ADJ
fbem-26172	36	10	content	content	NOUN
fbem-26172	36	11	of	of	ADP
fbem-26172	36	12	the	the	DET
fbem-26172	36	13	text	text	NOUN
fbem-26172	36	14	.	.	PUNCT
fbem-26172	37	1	input	input	NOUN
fbem-26172	37	2	:	:	PUNCT
fbem-26172	37	3	textual	textual	ADJ
fbem-26172	37	4	data	datum	NOUN
fbem-26172	37	5	from	from	ADP
fbem-26172	37	6	various	various	ADJ
fbem-26172	37	7	communication	communication	NOUN
fbem-26172	37	8	channels	channel	NOUN
fbem-26172	37	9	.	.	PUNCT
fbem-26172	38	1	output	output	NOUN
fbem-26172	38	2	:	:	PUNCT
fbem-26172	38	3	sentiment	sentiment	NOUN
fbem-26172	38	4	classification	classification	NOUN
fbem-26172	38	5	(	(	PUNCT
fbem-26172	38	6	positive	positive	ADJ
fbem-26172	38	7	,	,	PUNCT
fbem-26172	38	8	negative	negative	ADJ
fbem-26172	38	9	,	,	PUNCT
fbem-26172	38	10	neutral	neutral	ADJ
fbem-26172	38	11	)	)	PUNCT
fbem-26172	38	12	and	and	CCONJ
fbem-26172	38	13	detected	detect	VERB
fbem-26172	38	14	emotions	emotion	NOUN
fbem-26172	38	15	.	.	PUNCT
fbem-26172	39	1	challenges	challenge	NOUN
fbem-26172	39	2	:	:	PUNCT
fbem-26172	39	3	handling	handle	VERB
fbem-26172	39	4	diverse	diverse	ADJ
fbem-26172	39	5	writing	writing	NOUN
fbem-26172	39	6	styles	style	NOUN
fbem-26172	39	7	,	,	PUNCT
fbem-26172	39	8	slang	slang	PROPN
fbem-26172	39	9	,	,	PUNCT
fbem-26172	39	10	abbreviations	abbreviation	NOUN
fbem-26172	39	11	,	,	PUNCT
fbem-26172	39	12	and	and	CCONJ
fbem-26172	39	13	context	context	NOUN
fbem-26172	39	14	-	-	PUNCT
fbem-26172	39	15	based	base	VERB
fbem-26172	39	16	sentiment	sentiment	NOUN
fbem-26172	39	17	.	.	PUNCT
fbem-26172	40	1	2.2	2.2	NUM
fbem-26172	40	2	.	.	PUNCT
fbem-26172	40	3	model	model	NOUN
fbem-26172	40	4	description	description	NOUN
fbem-26172	40	5	the	the	DET
fbem-26172	40	6	proposed	propose	VERB
fbem-26172	40	7	model	model	NOUN
fbem-26172	40	8	integrates	integrates	PROPN
fbem-26172	40	9	fer	fer	PROPN
fbem-26172	40	10	and	and	CCONJ
fbem-26172	40	11	nlp	nlp	NOUN
fbem-26172	40	12	to	to	PART
fbem-26172	40	13	create	create	VERB
fbem-26172	40	14	a	a	DET
fbem-26172	40	15	multimodal	multimodal	ADJ
fbem-26172	40	16	emotion	emotion	NOUN
fbem-26172	40	17	recognition	recognition	NOUN
fbem-26172	40	18	system	system	NOUN
fbem-26172	40	19	.	.	PUNCT
fbem-26172	41	1	the	the	DET
fbem-26172	41	2	model	model	NOUN
fbem-26172	41	3	is	be	AUX
fbem-26172	41	4	designed	design	VERB
fbem-26172	41	5	to	to	PART
fbem-26172	41	6	simultaneously	simultaneously	ADV
fbem-26172	41	7	analyze	analyze	VERB
fbem-26172	41	8	facial	facial	ADJ
fbem-26172	41	9	expressions	expression	NOUN
fbem-26172	41	10	and	and	CCONJ
fbem-26172	41	11	textual	textual	ADJ
fbem-26172	41	12	data	datum	NOUN
fbem-26172	41	13	,	,	PUNCT
fbem-26172	41	14	providing	provide	VERB
fbem-26172	41	15	a	a	DET
fbem-26172	41	16	comprehensive	comprehensive	ADJ
fbem-26172	41	17	understanding	understanding	NOUN
fbem-26172	41	18	of	of	ADP
fbem-26172	41	19	customer	customer	NOUN
fbem-26172	41	20	emotions	emotion	NOUN
fbem-26172	41	21	.	.	PUNCT
fbem-26172	42	1	-the	-the	DET
fbem-26172	42	2	fer	fer	PROPN
fbem-26172	42	3	model	model	PROPN
fbem-26172	42	4	utilizes	utilize	VERB
fbem-26172	42	5	convolutional	convolutional	ADJ
fbem-26172	42	6	neural	neural	ADJ
fbem-26172	42	7	networks	network	NOUN
fbem-26172	42	8	(	(	PUNCT
fbem-26172	42	9	cnns	cnns	PROPN
fbem-26172	42	10	)	)	PUNCT
fbem-26172	42	11	to	to	PART
fbem-26172	42	12	detect	detect	VERB
fbem-26172	42	13	and	and	CCONJ
fbem-26172	42	14	classify	classify	VERB
fbem-26172	42	15	facial	facial	ADJ
fbem-26172	42	16	expressions	expression	NOUN
fbem-26172	42	17	.	.	PUNCT
fbem-26172	43	1	the	the	DET
fbem-26172	43	2	process	process	NOUN
fbem-26172	43	3	involves	involve	VERB
fbem-26172	43	4	several	several	ADJ
fbem-26172	43	5	steps	step	NOUN
fbem-26172	43	6	:	:	PUNCT
fbem-26172	43	7	(	(	PUNCT
fbem-26172	43	8	1	1	X
fbem-26172	43	9	)	)	PUNCT
fbem-26172	43	10	preprocessing	preprocessing	NOUN
fbem-26172	43	11	:	:	PUNCT
fbem-26172	43	12	convert	convert	VERB
fbem-26172	43	13	video	video	NOUN
fbem-26172	43	14	frames	frame	NOUN
fbem-26172	43	15	to	to	PART
fbem-26172	43	16	grayscale	grayscale	VERB
fbem-26172	43	17	images	image	NOUN
fbem-26172	43	18	to	to	PART
fbem-26172	43	19	reduce	reduce	VERB
fbem-26172	43	20	computational	computational	ADJ
fbem-26172	43	21	complexity	complexity	NOUN
fbem-26172	43	22	.	.	PUNCT
fbem-26172	44	1	normalize	normalize	VERB
fbem-26172	44	2	the	the	DET
fbem-26172	44	3	images	image	NOUN
fbem-26172	44	4	to	to	PART
fbem-26172	44	5	standardize	standardize	VERB
fbem-26172	44	6	the	the	DET
fbem-26172	44	7	input	input	NOUN
fbem-26172	44	8	data	datum	NOUN
fbem-26172	44	9	.	.	PUNCT
fbem-26172	45	1	(	(	PUNCT
fbem-26172	45	2	2	2	X
fbem-26172	45	3	)	)	PUNCT
fbem-26172	45	4	feature	feature	NOUN
fbem-26172	45	5	extraction	extraction	NOUN
fbem-26172	45	6	:	:	PUNCT
fbem-26172	45	7	use	use	NOUN
fbem-26172	45	8	cnns	cnn	NOUN
fbem-26172	45	9	to	to	PART
fbem-26172	45	10	extract	extract	VERB
fbem-26172	45	11	high	high	ADJ
fbem-26172	45	12	-	-	PUNCT
fbem-26172	45	13	level	level	NOUN
fbem-26172	45	14	features	feature	NOUN
fbem-26172	45	15	from	from	ADP
fbem-26172	45	16	the	the	DET
fbem-26172	45	17	facial	facial	ADJ
fbem-26172	45	18	images	image	NOUN
fbem-26172	45	19	.	.	PUNCT
fbem-26172	46	1	the	the	DET
fbem-26172	46	2	cnn	cnn	PROPN
fbem-26172	46	3	architecture	architecture	NOUN
fbem-26172	46	4	includes	include	VERB
fbem-26172	46	5	multiple	multiple	ADJ
fbem-26172	46	6	convolutional	convolutional	ADJ
fbem-26172	46	7	layers	layer	NOUN
fbem-26172	46	8	followed	follow	VERB
fbem-26172	46	9	by	by	ADP
fbem-26172	46	10	pooling	pool	VERB
fbem-26172	46	11	layers	layer	NOUN
fbem-26172	46	12	to	to	PART
fbem-26172	46	13	capture	capture	VERB
fbem-26172	46	14	spatial	spatial	ADJ
fbem-26172	46	15	hierarchies	hierarchy	NOUN
fbem-26172	46	16	.	.	PUNCT
fbem-26172	47	1	(	(	PUNCT
fbem-26172	47	2	3	3	X
fbem-26172	47	3	)	)	PUNCT
fbem-26172	47	4	classification	classification	NOUN
fbem-26172	47	5	:	:	PUNCT
fbem-26172	47	6	apply	apply	VERB
fbem-26172	47	7	fully	fully	ADV
fbem-26172	47	8	connected	connect	VERB
fbem-26172	47	9	layers	layer	NOUN
fbem-26172	47	10	and	and	CCONJ
fbem-26172	47	11	a	a	DET
fbem-26172	47	12	softmax	softmax	NOUN
fbem-26172	47	13	layer	layer	NOUN
fbem-26172	47	14	to	to	PART
fbem-26172	47	15	classify	classify	VERB
fbem-26172	47	16	the	the	DET
fbem-26172	47	17	extracted	extract	VERB
fbem-26172	47	18	features	feature	NOUN
fbem-26172	47	19	into	into	ADP
fbem-26172	47	20	predefined	predefine	VERB
fbem-26172	47	21	emotion	emotion	NOUN
fbem-26172	47	22	categories	category	NOUN
fbem-26172	47	23	.	.	PUNCT
fbem-26172	48	1	2.2.1	2.2.1	NUM
fbem-26172	48	2	.	.	PUNCT
fbem-26172	48	3	natural	natural	ADJ
fbem-26172	48	4	language	language	NOUN
fbem-26172	48	5	processing	processing	NOUN
fbem-26172	48	6	(	(	PUNCT
fbem-26172	48	7	nlp	nlp	NOUN
fbem-26172	48	8	)	)	PUNCT
fbem-26172	48	9	model	model	NOUN
fbem-26172	48	10	the	the	DET
fbem-26172	48	11	nlp	nlp	NOUN
fbem-26172	48	12	model	model	NOUN
fbem-26172	48	13	employs	employ	VERB
fbem-26172	48	14	advanced	advanced	ADJ
fbem-26172	48	15	text	text	NOUN
fbem-26172	48	16	analysis	analysis	NOUN
fbem-26172	48	17	techniques	technique	NOUN
fbem-26172	48	18	to	to	PART
fbem-26172	48	19	detect	detect	VERB
fbem-26172	48	20	emotions	emotion	NOUN
fbem-26172	48	21	from	from	ADP
fbem-26172	48	22	textual	textual	ADJ
fbem-26172	48	23	data	datum	NOUN
fbem-26172	48	24	.	.	PUNCT
fbem-26172	49	1	the	the	DET
fbem-26172	49	2	process	process	NOUN
fbem-26172	49	3	includes	include	VERB
fbem-26172	49	4	the	the	DET
fbem-26172	49	5	following	follow	VERB
fbem-26172	49	6	steps	step	NOUN
fbem-26172	49	7	:	:	PUNCT
fbem-26172	49	8	(	(	PUNCT
fbem-26172	49	9	1	1	X
fbem-26172	49	10	)	)	PUNCT
fbem-26172	49	11	preprocessing	preprocessing	NOUN
fbem-26172	49	12	:	:	PUNCT
fbem-26172	49	13	clean	clean	VERB
fbem-26172	49	14	the	the	DET
fbem-26172	49	15	text	text	NOUN
fbem-26172	49	16	data	datum	NOUN
fbem-26172	49	17	by	by	ADP
fbem-26172	49	18	removing	remove	VERB
fbem-26172	49	19	stop	stop	NOUN
fbem-26172	49	20	words	word	NOUN
fbem-26172	49	21	,	,	PUNCT
fbem-26172	49	22	punctuation	punctuation	NOUN
fbem-26172	49	23	,	,	PUNCT
fbem-26172	49	24	and	and	CCONJ
fbem-26172	49	25	performing	perform	VERB
fbem-26172	49	26	tokenization	tokenization	NOUN
fbem-26172	49	27	.	.	PUNCT
fbem-26172	50	1	convert	convert	NOUN
fbem-26172	50	2	text	text	NOUN
fbem-26172	50	3	to	to	PART
fbem-26172	50	4	lower	low	ADJ
fbem-26172	50	5	case	case	NOUN
fbem-26172	50	6	to	to	PART
fbem-26172	50	7	ensure	ensure	VERB
fbem-26172	50	8	uniformity	uniformity	NOUN
fbem-26172	50	9	.	.	PUNCT
fbem-26172	51	1	(	(	PUNCT
fbem-26172	51	2	2	2	X
fbem-26172	51	3	)	)	PUNCT
fbem-26172	51	4	feature	feature	NOUN
fbem-26172	51	5	extraction	extraction	NOUN
fbem-26172	51	6	:	:	PUNCT
fbem-26172	51	7	utilize	utilize	VERB
fbem-26172	51	8	techniques	technique	NOUN
fbem-26172	51	9	such	such	ADJ
fbem-26172	51	10	as	as	ADP
fbem-26172	51	11	term	term	NOUN
fbem-26172	51	12	frequency	frequency	NOUN
fbem-26172	51	13	-	-	PUNCT
fbem-26172	51	14	inverse	inverse	NOUN
fbem-26172	51	15	document	document	NOUN
fbem-26172	51	16	frequency	frequency	NOUN
fbem-26172	51	17	(	(	PUNCT
fbem-26172	51	18	tf	tf	PROPN
fbem-26172	51	19	-	-	PUNCT
fbem-26172	51	20	idf	idf	NOUN
fbem-26172	51	21	)	)	PUNCT
fbem-26172	51	22	and	and	CCONJ
fbem-26172	51	23	word	word	NOUN
fbem-26172	51	24	embeddings	embedding	NOUN
fbem-26172	51	25	(	(	PUNCT
fbem-26172	51	26	e.g.	e.g.	ADV
fbem-26172	51	27	,	,	PUNCT
fbem-26172	51	28	word2vec	word2vec	X
fbem-26172	51	29	,	,	PUNCT
fbem-26172	51	30	glove	glove	NOUN
fbem-26172	51	31	)	)	PUNCT
fbem-26172	51	32	to	to	PART
fbem-26172	51	33	convert	convert	VERB
fbem-26172	51	34	text	text	NOUN
fbem-26172	51	35	into	into	ADP
fbem-26172	51	36	numerical	numerical	ADJ
fbem-26172	51	37	feature	feature	NOUN
fbem-26172	51	38	vectors	vector	NOUN
fbem-26172	51	39	.	.	PUNCT
fbem-26172	52	1	(	(	PUNCT
fbem-26172	52	2	3	3	X
fbem-26172	52	3	)	)	PUNCT
fbem-26172	52	4	sentiment	sentiment	NOUN
fbem-26172	52	5	and	and	CCONJ
fbem-26172	52	6	emotion	emotion	NOUN
fbem-26172	52	7	classification	classification	NOUN
fbem-26172	52	8	:	:	PUNCT
fbem-26172	52	9	use	use	VERB
fbem-26172	52	10	a	a	DET
fbem-26172	52	11	recurrent	recurrent	ADJ
fbem-26172	52	12	neural	neural	ADJ
fbem-26172	52	13	network	network	NOUN
fbem-26172	52	14	(	(	PUNCT
fbem-26172	52	15	rnn	rnn	PROPN
fbem-26172	52	16	)	)	PUNCT
fbem-26172	52	17	or	or	CCONJ
fbem-26172	52	18	long	long	ADJ
fbem-26172	52	19	short	short	ADJ
fbem-26172	52	20	-	-	PUNCT
fbem-26172	52	21	term	term	NOUN
fbem-26172	52	22	memory	memory	NOUN
fbem-26172	52	23	(	(	PUNCT
fbem-26172	52	24	lstm	lstm	NOUN
fbem-26172	52	25	)	)	PUNCT
fbem-26172	52	26	network	network	NOUN
fbem-26172	52	27	to	to	PART
fbem-26172	52	28	capture	capture	VERB
fbem-26172	52	29	the	the	DET
fbem-26172	52	30	sequential	sequential	ADJ
fbem-26172	52	31	nature	nature	NOUN
fbem-26172	52	32	of	of	ADP
fbem-26172	52	33	the	the	DET
fbem-26172	52	34	text	text	NOUN
fbem-26172	52	35	.	.	PUNCT
fbem-26172	53	1	the	the	DET
fbem-26172	53	2	network	network	NOUN
fbem-26172	53	3	is	be	AUX
fbem-26172	53	4	trained	train	VERB
fbem-26172	53	5	to	to	PART
fbem-26172	53	6	classify	classify	VERB
fbem-26172	53	7	the	the	DET
fbem-26172	53	8	text	text	NOUN
fbem-26172	53	9	into	into	ADP
fbem-26172	53	10	sentiment	sentiment	NOUN
fbem-26172	53	11	categories	category	NOUN
fbem-26172	53	12	(	(	PUNCT
fbem-26172	53	13	positive	positive	ADJ
fbem-26172	53	14	,	,	PUNCT
fbem-26172	53	15	negative	negative	ADJ
fbem-26172	53	16	,	,	PUNCT
fbem-26172	53	17	neutral	neutral	ADJ
fbem-26172	53	18	)	)	PUNCT
fbem-26172	53	19	and	and	CCONJ
fbem-26172	53	20	detect	detect	VERB
fbem-26172	53	21	specific	specific	ADJ
fbem-26172	53	22	emotions	emotion	NOUN
fbem-26172	53	23	.	.	PUNCT
fbem-26172	54	1	3	3	X
fbem-26172	54	2	.	.	X
fbem-26172	54	3	methodology	methodology	NOUN
fbem-26172	54	4	3.1	3.1	NUM
fbem-26172	54	5	.	.	PUNCT
fbem-26172	55	1	data	datum	NOUN
fbem-26172	55	2	description	description	NOUN
fbem-26172	55	3	to	to	PART
fbem-26172	55	4	evaluate	evaluate	VERB
fbem-26172	55	5	the	the	DET
fbem-26172	55	6	effectiveness	effectiveness	NOUN
fbem-26172	55	7	of	of	ADP
fbem-26172	55	8	the	the	DET
fbem-26172	55	9	integrated	integrate	VERB
fbem-26172	55	10	fer	fer	NOUN
fbem-26172	55	11	and	and	CCONJ
fbem-26172	55	12	nlp	nlp	NOUN
fbem-26172	55	13	system	system	NOUN
fbem-26172	55	14	,	,	PUNCT
fbem-26172	55	15	we	we	PRON
fbem-26172	55	16	utilized	utilize	VERB
fbem-26172	55	17	a	a	DET
fbem-26172	55	18	comprehensive	comprehensive	ADJ
fbem-26172	55	19	dataset	dataset	NOUN
fbem-26172	55	20	that	that	PRON
fbem-26172	55	21	includes	include	VERB
fbem-26172	55	22	both	both	PRON
fbem-26172	55	23	facial	facial	ADJ
fbem-26172	55	24	expression	expression	NOUN
fbem-26172	55	25	videos	video	NOUN
fbem-26172	55	26	and	and	CCONJ
fbem-26172	55	27	textual	textual	ADJ
fbem-26172	55	28	data	datum	NOUN
fbem-26172	55	29	from	from	ADP
fbem-26172	55	30	customer	customer	NOUN
fbem-26172	55	31	service	service	NOUN
fbem-26172	55	32	interactions	interaction	NOUN
fbem-26172	55	33	.	.	PUNCT
fbem-26172	56	1	3.1.1	3.1.1	X
fbem-26172	56	2	.	.	PUNCT
fbem-26172	56	3	facial	facial	ADJ
fbem-26172	56	4	expression	expression	NOUN
fbem-26172	56	5	data	data	VERB
fbem-26172	56	6	the	the	DET
fbem-26172	56	7	facial	facial	ADJ
fbem-26172	56	8	expression	expression	NOUN
fbem-26172	56	9	data	datum	NOUN
fbem-26172	56	10	comprises	comprise	VERB
fbem-26172	56	11	video	video	NOUN
fbem-26172	56	12	recordings	recording	NOUN
fbem-26172	56	13	of	of	ADP
fbem-26172	56	14	customer	customer	NOUN
fbem-26172	56	15	interactions	interaction	NOUN
fbem-26172	56	16	,	,	PUNCT
fbem-26172	56	17	capturing	capture	VERB
fbem-26172	56	18	a	a	DET
fbem-26172	56	19	range	range	NOUN
fbem-26172	56	20	of	of	ADP
fbem-26172	56	21	emotions	emotion	NOUN
fbem-26172	56	22	such	such	ADJ
fbem-26172	56	23	as	as	ADP
fbem-26172	56	24	happiness	happiness	NOUN
fbem-26172	56	25	,	,	PUNCT
fbem-26172	56	26	sadness	sadness	NOUN
fbem-26172	56	27	,	,	PUNCT
fbem-26172	56	28	anger	anger	NOUN
fbem-26172	56	29	,	,	PUNCT
fbem-26172	56	30	surprise	surprise	NOUN
fbem-26172	56	31	,	,	PUNCT
fbem-26172	56	32	and	and	CCONJ
fbem-26172	56	33	neutrality	neutrality	NOUN
fbem-26172	56	34	.	.	PUNCT
fbem-26172	57	1	each	each	DET
fbem-26172	57	2	video	video	NOUN
fbem-26172	57	3	is	be	AUX
fbem-26172	57	4	segmented	segment	VERB
fbem-26172	57	5	into	into	ADP
fbem-26172	57	6	frames	frame	NOUN
fbem-26172	57	7	,	,	PUNCT
fbem-26172	57	8	and	and	CCONJ
fbem-26172	57	9	key	key	ADJ
fbem-26172	57	10	facial	facial	ADJ
fbem-26172	57	11	features	feature	NOUN
fbem-26172	57	12	are	be	AUX
fbem-26172	57	13	annotated	annotate	VERB
fbem-26172	57	14	to	to	PART
fbem-26172	57	15	facilitate	facilitate	VERB
fbem-26172	57	16	training	training	NOUN
fbem-26172	57	17	and	and	CCONJ
fbem-26172	57	18	validation	validation	NOUN
fbem-26172	57	19	of	of	ADP
fbem-26172	57	20	the	the	DET
fbem-26172	57	21	fer	fer	PROPN
fbem-26172	57	22	model	model	PROPN
fbem-26172	57	23	.	.	PUNCT
fbem-26172	58	1	·	·	PUNCT
fbem-26172	58	2	source	source	NOUN
fbem-26172	58	3	:	:	PUNCT
fbem-26172	58	4	customer	customer	NOUN
fbem-26172	58	5	service	service	NOUN
fbem-26172	58	6	video	video	NOUN
fbem-26172	58	7	recordings	recording	NOUN
fbem-26172	58	8	.	.	PUNCT
fbem-26172	59	1	·	·	PUNCT
fbem-26172	59	2	resolution	resolution	NOUN
fbem-26172	59	3	:	:	PUNCT
fbem-26172	59	4	640x480	640x480	NUM
fbem-26172	59	5	pixels	pixel	NOUN
fbem-26172	59	6	.	.	PUNCT
fbem-26172	60	1	·	·	PUNCT
fbem-26172	60	2	frame	frame	NOUN
fbem-26172	60	3	rate	rate	NOUN
fbem-26172	60	4	:	:	PUNCT
fbem-26172	60	5	30	30	NUM
fbem-26172	60	6	frames	frame	NOUN
fbem-26172	60	7	per	per	ADP
fbem-26172	60	8	second	second	NOUN
fbem-26172	60	9	.	.	PUNCT
fbem-26172	61	1	·	·	PUNCT
fbem-26172	61	2	annotations	annotation	NOUN
fbem-26172	61	3	:	:	PUNCT
fbem-26172	61	4	emotion	emotion	NOUN
fbem-26172	61	5	labels	label	NOUN
fbem-26172	61	6	for	for	ADP
fbem-26172	61	7	each	each	DET
fbem-26172	61	8	frame	frame	NOUN
fbem-26172	62	1	[	[	X
fbem-26172	62	2	5	5	NUM
fbem-26172	62	3	]	]	PUNCT
fbem-26172	62	4	.	.	PUNCT
fbem-26172	63	1	3.1.2	3.1.2	X
fbem-26172	63	2	.	.	PUNCT
fbem-26172	63	3	textual	textual	ADJ
fbem-26172	63	4	data	datum	NOUN
fbem-26172	63	5	the	the	DET
fbem-26172	63	6	textual	textual	ADJ
fbem-26172	63	7	data	datum	NOUN
fbem-26172	63	8	includes	include	VERB
fbem-26172	63	9	customer	customer	NOUN
fbem-26172	63	10	service	service	NOUN
fbem-26172	63	11	chat	chat	NOUN
fbem-26172	63	12	logs	log	NOUN
fbem-26172	63	13	,	,	PUNCT
fbem-26172	63	14	emails	email	NOUN
fbem-26172	63	15	,	,	PUNCT
fbem-26172	63	16	and	and	CCONJ
fbem-26172	63	17	social	social	ADJ
fbem-26172	63	18	media	medium	NOUN
fbem-26172	63	19	interactions	interaction	NOUN
fbem-26172	63	20	.	.	PUNCT
fbem-26172	64	1	each	each	DET
fbem-26172	64	2	text	text	NOUN
fbem-26172	64	3	entry	entry	NOUN
fbem-26172	64	4	is	be	AUX
fbem-26172	64	5	labeled	label	VERB
fbem-26172	64	6	with	with	ADP
fbem-26172	64	7	the	the	DET
fbem-26172	64	8	corresponding	corresponding	ADJ
fbem-26172	64	9	sentiment	sentiment	NOUN
fbem-26172	64	10	and	and	CCONJ
fbem-26172	64	11	emotion	emotion	NOUN
fbem-26172	64	12	.	.	PUNCT
fbem-26172	65	1	source	source	NOUN
fbem-26172	65	2	:	:	PUNCT
fbem-26172	65	3	chat	chat	VERB
fbem-26172	65	4	logs	log	NOUN
fbem-26172	65	5	,	,	PUNCT
fbem-26172	65	6	emails	email	NOUN
fbem-26172	65	7	,	,	PUNCT
fbem-26172	65	8	social	social	ADJ
fbem-26172	65	9	media	medium	NOUN
fbem-26172	65	10	posts	post	NOUN
fbem-26172	65	11	.	.	PUNCT
fbem-26172	66	1	preprocessing	preprocesse	VERB
fbem-26172	66	2	:	:	PUNCT
fbem-26172	66	3	tokenization	tokenization	NOUN
fbem-26172	66	4	,	,	PUNCT
fbem-26172	66	5	stop	stop	VERB
fbem-26172	66	6	word	word	NOUN
fbem-26172	66	7	removal	removal	NOUN
fbem-26172	66	8	,	,	PUNCT
fbem-26172	66	9	and	and	CCONJ
fbem-26172	66	10	lemmatization	lemmatization	NOUN
fbem-26172	66	11	.	.	PUNCT
fbem-26172	67	1	annotations	annotation	NOUN
fbem-26172	67	2	:	:	PUNCT
fbem-26172	67	3	sentiment	sentiment	NOUN
fbem-26172	67	4	(	(	PUNCT
fbem-26172	67	5	positive	positive	ADJ
fbem-26172	67	6	,	,	PUNCT
fbem-26172	67	7	negative	negative	ADJ
fbem-26172	67	8	,	,	PUNCT
fbem-26172	67	9	neutral	neutral	ADJ
fbem-26172	67	10	)	)	PUNCT
fbem-26172	67	11	and	and	CCONJ
fbem-26172	67	12	emotion	emotion	NOUN
fbem-26172	67	13	labels	label	NOUN
fbem-26172	67	14	.	.	PUNCT
fbem-26172	68	1	3.2	3.2	NUM
fbem-26172	68	2	.	.	PUNCT
fbem-26172	68	3	programming	programming	NOUN
fbem-26172	68	4	environment	environment	NOUN
fbem-26172	68	5	and	and	CCONJ
fbem-26172	68	6	implementation	implementation	NOUN
fbem-26172	68	7	details	detail	NOUN
fbem-26172	68	8	the	the	DET
fbem-26172	68	9	integrated	integrate	VERB
fbem-26172	68	10	emotion	emotion	NOUN
fbem-26172	68	11	recognition	recognition	NOUN
fbem-26172	68	12	system	system	NOUN
fbem-26172	68	13	was	be	AUX
fbem-26172	68	14	developed	develop	VERB
fbem-26172	68	15	and	and	CCONJ
fbem-26172	68	16	tested	test	VERB
fbem-26172	68	17	using	use	VERB
fbem-26172	68	18	the	the	DET
fbem-26172	68	19	following	follow	VERB
fbem-26172	68	20	programming	programming	NOUN
fbem-26172	68	21	environment	environment	NOUN
fbem-26172	68	22	and	and	CCONJ
fbem-26172	68	23	tools	tool	NOUN
fbem-26172	68	24	:	:	PUNCT
fbem-26172	68	25	programming	programming	NOUN
fbem-26172	68	26	language	language	NOUN
fbem-26172	68	27	:	:	PUNCT
fbem-26172	68	28	python	python	NOUN
fbem-26172	68	29	3.12	3.12	NUM
fbem-26172	68	30	libraries	library	NOUN
fbem-26172	68	31	and	and	CCONJ
fbem-26172	68	32	frameworks	framework	NOUN
fbem-26172	68	33	:	:	PUNCT
fbem-26172	68	34	opencv	opencv	NOUN
fbem-26172	68	35	for	for	ADP
fbem-26172	68	36	image	image	NOUN
fbem-26172	68	37	processing	processing	NOUN
fbem-26172	68	38	.	.	PUNCT
fbem-26172	69	1	keras	keras	PROPN
fbem-26172	69	2	with	with	ADP
fbem-26172	69	3	tensorflow	tensorflow	NOUN
fbem-26172	69	4	backend	backend	NOUN
fbem-26172	69	5	for	for	ADP
fbem-26172	69	6	building	build	VERB
fbem-26172	69	7	neural	neural	ADJ
fbem-26172	69	8	networks	network	NOUN
fbem-26172	69	9	.	.	PUNCT
fbem-26172	70	1	nltk	nltk	PROPN
fbem-26172	70	2	and	and	CCONJ
fbem-26172	70	3	sk	sk	AUX
fbem-26172	70	4	-	-	PUNCT
fbem-26172	70	5	learn	learn	NOUN
fbem-26172	70	6	for	for	ADP
fbem-26172	70	7	nlp	nlp	NOUN
fbem-26172	70	8	tasks	task	NOUN
fbem-26172	70	9	.	.	PUNCT
fbem-26172	71	1	3.3	3.3	NUM
fbem-26172	71	2	.	.	PUNCT
fbem-26172	72	1	analysis	analysis	NOUN
fbem-26172	72	2	3.3.1	3.3.1	NUM
fbem-26172	72	3	.	.	PUNCT
fbem-26172	72	4	facial	facial	ADJ
fbem-26172	72	5	expression	expression	NOUN
fbem-26172	72	6	recognition	recognition	NOUN
fbem-26172	72	7	(	(	PUNCT
fbem-26172	72	8	fer	fer	PROPN
fbem-26172	72	9	)	)	PUNCT
fbem-26172	72	10	model	model	NOUN
fbem-26172	72	11	the	the	DET
fbem-26172	72	12	accuracy	accuracy	NOUN
fbem-26172	72	13	of	of	ADP
fbem-26172	72	14	the	the	DET
fbem-26172	72	15	fer	fer	PROPN
fbem-26172	72	16	model	model	PROPN
fbem-26172	72	17	shows	show	VERB
fbem-26172	72	18	a	a	DET
fbem-26172	72	19	gradual	gradual	ADJ
fbem-26172	72	20	increase	increase	NOUN
fbem-26172	72	21	from	from	ADP
fbem-26172	72	22	0.2024	0.2024	NUM
fbem-26172	72	23	to	to	ADP
fbem-26172	72	24	0.2701	0.2701	NUM
fbem-26172	72	25	over	over	ADP
fbem-26172	72	26	the	the	DET
fbem-26172	72	27	epochs	epoch	NOUN
fbem-26172	72	28	,	,	PUNCT
fbem-26172	72	29	with	with	ADP
fbem-26172	72	30	the	the	DET
fbem-26172	72	31	loss	loss	NOUN
fbem-26172	72	32	decreasing	decrease	VERB
fbem-26172	72	33	from	from	ADP
fbem-26172	72	34	1.7343	1.7343	NUM
fbem-26172	72	35	to	to	ADP
fbem-26172	72	36	1.5830	1.5830	NUM
fbem-26172	72	37	.	.	PUNCT
fbem-26172	73	1	the	the	DET
fbem-26172	73	2	validation	validation	NOUN
fbem-26172	73	3	accuracy	accuracy	NOUN
fbem-26172	73	4	starts	start	VERB
fbem-26172	73	5	at	at	ADP
fbem-26172	73	6	a	a	DET
fbem-26172	73	7	low	low	ADJ
fbem-26172	73	8	0.1800	0.1800	NUM
fbem-26172	73	9	and	and	CCONJ
fbem-26172	73	10	shows	show	VERB
fbem-26172	73	11	slight	slight	ADJ
fbem-26172	73	12	improvement	improvement	NOUN
fbem-26172	73	13	to	to	ADP
fbem-26172	73	14	0.2500	0.2500	NUM
fbem-26172	73	15	by	by	ADP
fbem-26172	73	16	the	the	DET
fbem-26172	73	17	10th	10th	ADJ
fbem-26172	73	18	epoch	epoch	NOUN
fbem-26172	73	19	.	.	PUNCT
fbem-26172	74	1	the	the	DET
fbem-26172	74	2	validation	validation	NOUN
fbem-26172	74	3	loss	loss	NOUN
fbem-26172	74	4	generally	generally	ADV
fbem-26172	74	5	decreases	decrease	VERB
fbem-26172	74	6	,	,	PUNCT
fbem-26172	74	7	indicating	indicate	VERB
fbem-26172	74	8	some	some	DET
fbem-26172	74	9	improvement	improvement	NOUN
fbem-26172	74	10	in	in	ADP
fbem-26172	74	11	model	model	NOUN
fbem-26172	74	12	performance	performance	NOUN
fbem-26172	74	13	on	on	ADP
fbem-26172	74	14	unseen	unseen	ADJ
fbem-26172	74	15	data	datum	NOUN
fbem-26172	74	16	,	,	PUNCT
fbem-26172	74	17	though	though	SCONJ
fbem-26172	74	18	the	the	DET
fbem-26172	74	19	changes	change	NOUN
fbem-26172	74	20	are	be	AUX
fbem-26172	74	21	not	not	PART
fbem-26172	74	22	dramatic	dramatic	ADJ
fbem-26172	74	23	.	.	PUNCT
fbem-26172	75	1	the	the	DET
fbem-26172	75	2	fer	fer	PROPN
fbem-26172	75	3	model	model	NOUN
fbem-26172	75	4	demonstrates	demonstrate	VERB
fbem-26172	75	5	a	a	DET
fbem-26172	75	6	modest	modest	ADJ
fbem-26172	75	7	improvement	improvement	NOUN
fbem-26172	75	8	in	in	ADP
fbem-26172	75	9	performance	performance	NOUN
fbem-26172	75	10	,	,	PUNCT
fbem-26172	75	11	suggesting	suggest	VERB
fbem-26172	75	12	that	that	SCONJ
fbem-26172	75	13	the	the	DET
fbem-26172	75	14	model	model	NOUN
fbem-26172	75	15	is	be	AUX
fbem-26172	75	16	learning	learn	VERB
fbem-26172	75	17	to	to	ADP
fbem-26172	75	18	some	some	DET
fbem-26172	75	19	extent	extent	NOUN
fbem-26172	75	20	but	but	CCONJ
fbem-26172	75	21	still	still	ADV
fbem-26172	75	22	faces	face	VERB
fbem-26172	75	23	challenges	challenge	NOUN
fbem-26172	75	24	in	in	ADP
fbem-26172	75	25	generalization	generalization	NOUN
fbem-26172	75	26	.	.	PUNCT
fbem-26172	76	1	the	the	DET
fbem-26172	76	2	improvements	improvement	NOUN
fbem-26172	76	3	in	in	ADP
fbem-26172	76	4	validation	validation	NOUN
fbem-26172	76	5	accuracy	accuracy	NOUN
fbem-26172	76	6	and	and	CCONJ
fbem-26172	76	7	reduction	reduction	NOUN
fbem-26172	76	8	in	in	ADP
fbem-26172	76	9	validation	validation	NOUN
fbem-26172	76	10	loss	loss	NOUN
fbem-26172	76	11	are	be	AUX
fbem-26172	76	12	positive	positive	ADJ
fbem-26172	76	13	signs	sign	NOUN
fbem-26172	76	14	,	,	PUNCT
fbem-26172	76	15	although	although	SCONJ
fbem-26172	76	16	the	the	DET
fbem-26172	76	17	gains	gain	NOUN
fbem-26172	76	18	are	be	AUX
fbem-26172	76	19	incremental	incremental	ADJ
fbem-26172	76	20	.	.	PUNCT
fbem-26172	77	1	3.3.2	3.3.2	NUM
fbem-26172	77	2	.	.	PUNCT
fbem-26172	77	3	natural	natural	ADJ
fbem-26172	77	4	language	language	NOUN
fbem-26172	77	5	processing	processing	NOUN
fbem-26172	77	6	(	(	PUNCT
fbem-26172	77	7	nlp	nlp	NOUN
fbem-26172	77	8	)	)	PUNCT
fbem-26172	77	9	model	model	NOUN
fbem-26172	77	10	the	the	DET
fbem-26172	77	11	accuracy	accuracy	NOUN
fbem-26172	77	12	remains	remain	VERB
fbem-26172	77	13	constant	constant	ADJ
fbem-26172	77	14	at	at	ADP
fbem-26172	77	15	0.2500	0.2500	NUM
fbem-26172	77	16	across	across	ADP
fbem-26172	77	17	all	all	DET
fbem-26172	77	18	epochs	epoch	NOUN
fbem-26172	77	19	,	,	PUNCT
fbem-26172	77	20	indicating	indicate	VERB
fbem-26172	77	21	no	no	DET
fbem-26172	77	22	learning	learning	NOUN
fbem-26172	77	23	progress	progress	NOUN
fbem-26172	77	24	.	.	PUNCT
fbem-26172	78	1	validation	validation	NOUN
fbem-26172	78	2	accuracy	accuracy	NOUN
fbem-26172	78	3	is	be	AUX
fbem-26172	78	4	consistently	consistently	ADV
fbem-26172	78	5	0.0000	0.0000	NUM
fbem-26172	78	6	,	,	PUNCT
fbem-26172	78	7	suggesting	suggest	VERB
fbem-26172	78	8	that	that	SCONJ
fbem-26172	78	9	the	the	DET
fbem-26172	78	10	model	model	NOUN
fbem-26172	78	11	fails	fail	VERB
fbem-26172	78	12	to	to	PART
fbem-26172	78	13	generalize	generalize	VERB
fbem-26172	78	14	to	to	ADP
fbem-26172	78	15	validation	validation	NOUN
fbem-26172	78	16	data	datum	NOUN
fbem-26172	78	17	.	.	PUNCT
fbem-26172	79	1	validation	validation	NOUN
fbem-26172	79	2	loss	loss	NOUN
fbem-26172	79	3	increases	increase	NOUN
fbem-26172	79	4	significantly	significantly	ADV
fbem-26172	79	5	from	from	ADP
fbem-26172	79	6	1.6317	1.6317	NUM
fbem-26172	79	7	to	to	ADP
fbem-26172	79	8	2.4758	2.4758	NUM
fbem-26172	79	9	,	,	PUNCT
fbem-26172	79	10	further	far	ADV
fbem-26172	79	11	confirming	confirm	VERB
fbem-26172	79	12	poor	poor	ADJ
fbem-26172	79	13	generalization	generalization	NOUN
fbem-26172	79	14	and	and	CCONJ
fbem-26172	79	15	potential	potential	ADJ
fbem-26172	79	16	overfitting	overfitting	NOUN
fbem-26172	79	17	.	.	PUNCT
fbem-26172	80	1	the	the	DET
fbem-26172	80	2	nlp	nlp	PROPN
fbem-26172	80	3	model	model	NOUN
fbem-26172	80	4	shows	show	VERB
fbem-26172	80	5	significant	significant	ADJ
fbem-26172	80	6	issues	issue	NOUN
fbem-26172	80	7	in	in	ADP
fbem-26172	80	8	learning	learning	NOUN
fbem-26172	80	9	and	and	CCONJ
fbem-26172	80	10	generalization	generalization	NOUN
fbem-26172	80	11	.	.	PUNCT
fbem-26172	81	1	the	the	DET
fbem-26172	81	2	stagnant	stagnant	ADJ
fbem-26172	81	3	accuracy	accuracy	NOUN
fbem-26172	81	4	and	and	CCONJ
fbem-26172	81	5	increasing	increase	VERB
fbem-26172	81	6	validation	validation	NOUN
fbem-26172	81	7	loss	loss	NOUN
fbem-26172	81	8	indicate	indicate	VERB
fbem-26172	81	9	that	that	SCONJ
fbem-26172	81	10	the	the	DET
fbem-26172	81	11	model	model	NOUN
fbem-26172	81	12	is	be	AUX
fbem-26172	81	13	not	not	PART
fbem-26172	81	14	effectively	effectively	ADV
fbem-26172	81	15	capturing	capture	VERB
fbem-26172	81	16	the	the	DET
fbem-26172	81	17	patterns	pattern	NOUN
fbem-26172	81	18	in	in	ADP
fbem-26172	81	19	the	the	DET
fbem-26172	81	20	textual	textual	ADJ
fbem-26172	81	21	data	datum	NOUN
fbem-26172	81	22	.	.	PUNCT
fbem-26172	82	1	this	this	PRON
fbem-26172	82	2	suggests	suggest	VERB
fbem-26172	82	3	a	a	DET
fbem-26172	82	4	need	need	NOUN
fbem-26172	82	5	for	for	ADP
fbem-26172	82	6	better	well	ADJ
fbem-26172	82	7	preprocessing	preprocessing	NOUN
fbem-26172	82	8	,	,	PUNCT
fbem-26172	82	9	more	more	ADV
fbem-26172	82	10	diverse	diverse	ADJ
fbem-26172	82	11	data	datum	NOUN
fbem-26172	82	12	,	,	PUNCT
fbem-26172	82	13	or	or	CCONJ
fbem-26172	82	14	an	an	DET
fbem-26172	82	15	improved	improved	ADJ
fbem-26172	82	16	model	model	NOUN
fbem-26172	82	17	architecture	architecture	NOUN
fbem-26172	82	18	.	.	PUNCT
fbem-26172	83	1	3.3.3	3.3.3	NUM
fbem-26172	83	2	.	.	PUNCT
fbem-26172	83	3	combined	combine	VERB
fbem-26172	83	4	fer	fer	PROPN
fbem-26172	83	5	and	and	CCONJ
fbem-26172	83	6	nlp	nlp	PROPN
fbem-26172	83	7	model	model	NOUN
fbem-26172	83	8	the	the	DET
fbem-26172	83	9	accuracy	accuracy	NOUN
fbem-26172	83	10	of	of	ADP
fbem-26172	83	11	the	the	DET
fbem-26172	83	12	combined	combine	VERB
fbem-26172	83	13	model	model	NOUN
fbem-26172	83	14	shows	show	VERB
fbem-26172	83	15	substantial	substantial	ADJ
fbem-26172	83	16	improvement	improvement	NOUN
fbem-26172	83	17	,	,	PUNCT
fbem-26172	83	18	starting	start	VERB
fbem-26172	83	19	at	at	ADP
fbem-26172	83	20	0.2043	0.2043	NUM
fbem-26172	83	21	and	and	CCONJ
fbem-26172	83	22	reaching	reach	VERB
fbem-26172	83	23	0.8127	0.8127	NUM
fbem-26172	83	24	by	by	ADP
fbem-26172	83	25	the	the	DET
fbem-26172	83	26	10th	10th	ADJ
fbem-26172	83	27	epoch	epoch	NOUN
fbem-26172	83	28	,	,	PUNCT
fbem-26172	83	29	with	with	ADP
fbem-26172	83	30	a	a	DET
fbem-26172	83	31	significant	significant	ADJ
fbem-26172	83	32	decrease	decrease	NOUN
fbem-26172	83	33	in	in	ADP
fbem-26172	83	34	loss	loss	NOUN
fbem-26172	83	35	from	from	ADP
fbem-26172	83	36	1.6865	1.6865	NUM
fbem-26172	83	37	to	to	ADP
fbem-26172	83	38	0.8635	0.8635	NUM
fbem-26172	83	39	.	.	PUNCT
fbem-26172	84	1	validation	validation	NOUN
fbem-26172	84	2	accuracy	accuracy	NOUN
fbem-26172	84	3	improves	improve	VERB
fbem-26172	84	4	from	from	ADP
fbem-26172	84	5	0.2550	0.2550	NUM
fbem-26172	84	6	to	to	ADP
fbem-26172	84	7	0.2150	0.2150	NUM
fbem-26172	84	8	,	,	PUNCT
fbem-26172	84	9	reflecting	reflect	VERB
fbem-26172	84	10	better	well	ADJ
fbem-26172	84	11	generalization	generalization	NOUN
fbem-26172	84	12	compared	compare	VERB
fbem-26172	84	13	to	to	ADP
fbem-26172	84	14	the	the	DET
fbem-26172	84	15	nlp	nlp	NOUN
fbem-26172	84	16	model	model	NOUN
fbem-26172	84	17	.	.	PUNCT
fbem-26172	85	1	although	although	SCONJ
fbem-26172	85	2	the	the	DET
fbem-26172	85	3	validation	validation	NOUN
fbem-26172	85	4	loss	loss	NOUN
fbem-26172	85	5	shows	show	VERB
fbem-26172	85	6	a	a	DET
fbem-26172	85	7	general	general	ADJ
fbem-26172	85	8	decreasing	decrease	VERB
fbem-26172	85	9	74	74	NUM
fbem-26172	85	10	trend	trend	NOUN
fbem-26172	85	11	,	,	PUNCT
fbem-26172	85	12	there	there	PRON
fbem-26172	85	13	is	be	VERB
fbem-26172	85	14	an	an	DET
fbem-26172	85	15	increase	increase	NOUN
fbem-26172	85	16	towards	towards	ADP
fbem-26172	85	17	the	the	DET
fbem-26172	85	18	end	end	NOUN
fbem-26172	85	19	,	,	PUNCT
fbem-26172	85	20	suggesting	suggest	VERB
fbem-26172	85	21	some	some	DET
fbem-26172	85	22	overfitting	overfitting	NOUN
fbem-26172	85	23	but	but	CCONJ
fbem-26172	85	24	overall	overall	ADJ
fbem-26172	85	25	better	well	ADJ
fbem-26172	85	26	performance	performance	NOUN
fbem-26172	85	27	than	than	ADP
fbem-26172	85	28	the	the	DET
fbem-26172	85	29	standalone	standalone	NOUN
fbem-26172	85	30	models	model	NOUN
fbem-26172	85	31	.	.	PUNCT
fbem-26172	86	1	the	the	DET
fbem-26172	86	2	combined	combine	VERB
fbem-26172	86	3	fer	fer	PROPN
fbem-26172	86	4	and	and	CCONJ
fbem-26172	86	5	nlp	nlp	NOUN
fbem-26172	86	6	model	model	NOUN
fbem-26172	86	7	demonstrates	demonstrate	VERB
fbem-26172	86	8	the	the	DET
fbem-26172	86	9	best	good	ADJ
fbem-26172	86	10	performance	performance	NOUN
fbem-26172	86	11	among	among	ADP
fbem-26172	86	12	the	the	DET
fbem-26172	86	13	three	three	NUM
fbem-26172	86	14	models	model	NOUN
fbem-26172	86	15	,	,	PUNCT
fbem-26172	86	16	with	with	ADP
fbem-26172	86	17	significant	significant	ADJ
fbem-26172	86	18	improvements	improvement	NOUN
fbem-26172	86	19	in	in	ADP
fbem-26172	86	20	both	both	PRON
fbem-26172	86	21	training	training	NOUN
fbem-26172	86	22	and	and	CCONJ
fbem-26172	86	23	validation	validation	NOUN
fbem-26172	86	24	accuracy	accuracy	NOUN
fbem-26172	86	25	.	.	PUNCT
fbem-26172	87	1	the	the	DET
fbem-26172	87	2	integration	integration	NOUN
fbem-26172	87	3	of	of	ADP
fbem-26172	87	4	fer	fer	PROPN
fbem-26172	87	5	and	and	CCONJ
fbem-26172	87	6	nlp	nlp	PROPN
fbem-26172	87	7	likely	likely	ADV
fbem-26172	87	8	contributes	contribute	VERB
fbem-26172	87	9	to	to	ADP
fbem-26172	87	10	the	the	DET
fbem-26172	87	11	enhanced	enhance	VERB
fbem-26172	87	12	performance	performance	NOUN
fbem-26172	87	13	,	,	PUNCT
fbem-26172	87	14	highlighting	highlight	VERB
fbem-26172	87	15	the	the	DET
fbem-26172	87	16	effectiveness	effectiveness	NOUN
fbem-26172	87	17	of	of	ADP
fbem-26172	87	18	a	a	DET
fbem-26172	87	19	multimodal	multimodal	ADJ
fbem-26172	87	20	approach	approach	NOUN
fbem-26172	87	21	in	in	ADP
fbem-26172	87	22	emotion	emotion	NOUN
fbem-26172	87	23	recognition	recognition	NOUN
fbem-26172	87	24	for	for	ADP
fbem-26172	87	25	customer	customer	NOUN
fbem-26172	87	26	service	service	NOUN
fbem-26172	87	27	interactions	interaction	NOUN
fbem-26172	87	28	.	.	PUNCT
fbem-26172	88	1	this	this	DET
fbem-26172	88	2	model	model	NOUN
fbem-26172	88	3	shows	show	VERB
fbem-26172	88	4	strong	strong	ADJ
fbem-26172	88	5	learning	learning	NOUN
fbem-26172	88	6	capabilities	capability	NOUN
fbem-26172	88	7	and	and	CCONJ
fbem-26172	88	8	better	well	ADJ
fbem-26172	88	9	generalization	generalization	NOUN
fbem-26172	88	10	,	,	PUNCT
fbem-26172	88	11	though	though	SCONJ
fbem-26172	88	12	attention	attention	NOUN
fbem-26172	88	13	to	to	ADP
fbem-26172	88	14	overfitting	overfitting	NOUN
fbem-26172	88	15	is	be	AUX
fbem-26172	88	16	needed	need	VERB
fbem-26172	88	17	.	.	PUNCT
fbem-26172	89	1	4	4	X
fbem-26172	89	2	.	.	NOUN
fbem-26172	89	3	advantages	advantage	NOUN
fbem-26172	89	4	and	and	CCONJ
fbem-26172	89	5	disadvantages	disadvantage	NOUN
fbem-26172	89	6	(	(	PUNCT
fbem-26172	89	7	limitations	limitation	NOUN
fbem-26172	89	8	)	)	PUNCT
fbem-26172	89	9	of	of	ADP
fbem-26172	89	10	the	the	DET
fbem-26172	89	11	methods	method	NOUN
fbem-26172	89	12	advantages	advantage	NOUN
fbem-26172	89	13	:	:	PUNCT
fbem-26172	89	14	comprehensive	comprehensive	ADJ
fbem-26172	89	15	emotion	emotion	NOUN
fbem-26172	89	16	detection	detection	NOUN
fbem-26172	89	17	:	:	PUNCT
fbem-26172	89	18	by	by	ADP
fbem-26172	89	19	combining	combine	VERB
fbem-26172	89	20	fer	fer	PROPN
fbem-26172	89	21	and	and	CCONJ
fbem-26172	89	22	nlp	nlp	NOUN
fbem-26172	89	23	,	,	PUNCT
fbem-26172	89	24	our	our	PRON
fbem-26172	89	25	model	model	NOUN
fbem-26172	89	26	provides	provide	VERB
fbem-26172	89	27	a	a	DET
fbem-26172	89	28	more	more	ADV
fbem-26172	89	29	holistic	holistic	ADJ
fbem-26172	89	30	understanding	understanding	NOUN
fbem-26172	89	31	of	of	ADP
fbem-26172	89	32	customer	customer	NOUN
fbem-26172	89	33	emotions	emotion	NOUN
fbem-26172	89	34	,	,	PUNCT
fbem-26172	89	35	capturing	capture	VERB
fbem-26172	89	36	both	both	CCONJ
fbem-26172	89	37	verbal	verbal	ADJ
fbem-26172	89	38	and	and	CCONJ
fbem-26172	89	39	non	non	ADJ
fbem-26172	89	40	-	-	ADJ
fbem-26172	89	41	verbal	verbal	ADJ
fbem-26172	89	42	cues	cue	NOUN
fbem-26172	89	43	.	.	PUNCT
fbem-26172	90	1	enhanced	enhance	VERB
fbem-26172	90	2	customer	customer	NOUN
fbem-26172	90	3	interaction	interaction	NOUN
fbem-26172	90	4	:	:	PUNCT
fbem-26172	90	5	the	the	DET
fbem-26172	90	6	integrated	integrate	VERB
fbem-26172	90	7	system	system	NOUN
fbem-26172	90	8	allows	allow	VERB
fbem-26172	90	9	for	for	ADP
fbem-26172	90	10	more	more	ADV
fbem-26172	90	11	empathetic	empathetic	ADJ
fbem-26172	90	12	and	and	CCONJ
fbem-26172	90	13	effective	effective	ADJ
fbem-26172	90	14	customer	customer	NOUN
fbem-26172	90	15	service	service	NOUN
fbem-26172	90	16	interactions	interaction	NOUN
fbem-26172	90	17	,	,	PUNCT
fbem-26172	90	18	improving	improve	VERB
fbem-26172	90	19	overall	overall	ADJ
fbem-26172	90	20	customer	customer	NOUN
fbem-26172	90	21	satisfaction	satisfaction	NOUN
fbem-26172	90	22	.	.	PUNCT
fbem-26172	91	1	real	real	ADJ
fbem-26172	91	2	-	-	PUNCT
fbem-26172	91	3	time	time	NOUN
fbem-26172	91	4	analysis	analysis	NOUN
fbem-26172	91	5	:	:	PUNCT
fbem-26172	91	6	the	the	DET
fbem-26172	91	7	model	model	NOUN
fbem-26172	91	8	's	's	PART
fbem-26172	91	9	ability	ability	NOUN
fbem-26172	91	10	to	to	PART
fbem-26172	91	11	process	process	VERB
fbem-26172	91	12	and	and	CCONJ
fbem-26172	91	13	analyze	analyze	VERB
fbem-26172	91	14	data	datum	NOUN
fbem-26172	91	15	in	in	ADP
fbem-26172	91	16	real	real	ADJ
fbem-26172	91	17	-	-	PUNCT
fbem-26172	91	18	time	time	NOUN
fbem-26172	91	19	facilitates	facilitate	VERB
fbem-26172	91	20	immediate	immediate	ADJ
fbem-26172	91	21	adjustments	adjustment	NOUN
fbem-26172	91	22	in	in	ADP
fbem-26172	91	23	customer	customer	NOUN
fbem-26172	91	24	service	service	NOUN
fbem-26172	91	25	strategies	strategy	NOUN
fbem-26172	91	26	,	,	PUNCT
fbem-26172	91	27	leading	lead	VERB
fbem-26172	91	28	to	to	ADP
fbem-26172	91	29	better	well	ADJ
fbem-26172	91	30	outcomes	outcome	NOUN
fbem-26172	91	31	.	.	PUNCT
fbem-26172	92	1	disadvantages	disadvantage	NOUN
fbem-26172	92	2	(	(	PUNCT
fbem-26172	92	3	limitations	limitation	NOUN
fbem-26172	92	4	):	):	PUNCT
fbem-26172	92	5	computational	computational	ADJ
fbem-26172	92	6	complexity	complexity	NOUN
fbem-26172	92	7	:	:	PUNCT
fbem-26172	92	8	integrating	integrate	VERB
fbem-26172	92	9	fer	fer	PROPN
fbem-26172	92	10	and	and	CCONJ
fbem-26172	92	11	nlp	nlp	NOUN
fbem-26172	92	12	increases	increase	VERB
fbem-26172	92	13	the	the	DET
fbem-26172	92	14	computational	computational	ADJ
fbem-26172	92	15	requirements	requirement	NOUN
fbem-26172	92	16	,	,	PUNCT
fbem-26172	92	17	necessitating	necessitate	VERB
fbem-26172	92	18	more	more	ADV
fbem-26172	92	19	powerful	powerful	ADJ
fbem-26172	92	20	hardware	hardware	NOUN
fbem-26172	92	21	and	and	CCONJ
fbem-26172	92	22	longer	long	ADJ
fbem-26172	92	23	processing	processing	NOUN
fbem-26172	92	24	times	time	NOUN
fbem-26172	92	25	.	.	PUNCT
fbem-26172	93	1	data	datum	NOUN
fbem-26172	93	2	variability	variability	NOUN
fbem-26172	93	3	:	:	PUNCT
fbem-26172	93	4	the	the	DET
fbem-26172	93	5	model	model	NOUN
fbem-26172	93	6	's	's	PART
fbem-26172	93	7	performance	performance	NOUN
fbem-26172	93	8	can	can	AUX
fbem-26172	93	9	be	be	AUX
fbem-26172	93	10	affected	affect	VERB
fbem-26172	93	11	by	by	ADP
fbem-26172	93	12	the	the	DET
fbem-26172	93	13	variability	variability	NOUN
fbem-26172	93	14	in	in	ADP
fbem-26172	93	15	facial	facial	ADJ
fbem-26172	93	16	expressions	expression	NOUN
fbem-26172	93	17	,	,	PUNCT
fbem-26172	93	18	lighting	lighting	NOUN
fbem-26172	93	19	conditions	condition	NOUN
fbem-26172	93	20	,	,	PUNCT
fbem-26172	93	21	and	and	CCONJ
fbem-26172	93	22	textual	textual	ADJ
fbem-26172	93	23	data	datum	NOUN
fbem-26172	93	24	,	,	PUNCT
fbem-26172	93	25	which	which	PRON
fbem-26172	93	26	may	may	AUX
fbem-26172	93	27	require	require	VERB
fbem-26172	93	28	additional	additional	ADJ
fbem-26172	93	29	preprocessing	preprocessing	NOUN
fbem-26172	93	30	and	and	CCONJ
fbem-26172	93	31	normalization	normalization	NOUN
fbem-26172	93	32	steps	step	NOUN
fbem-26172	93	33	.	.	PUNCT
fbem-26172	94	1	generalization	generalization	NOUN
fbem-26172	94	2	to	to	ADP
fbem-26172	94	3	different	different	ADJ
fbem-26172	94	4	contexts	context	NOUN
fbem-26172	94	5	:	:	PUNCT
fbem-26172	94	6	while	while	SCONJ
fbem-26172	94	7	the	the	DET
fbem-26172	94	8	model	model	NOUN
fbem-26172	94	9	performs	perform	VERB
fbem-26172	94	10	well	well	ADV
fbem-26172	94	11	on	on	ADP
fbem-26172	94	12	the	the	DET
fbem-26172	94	13	provided	provide	VERB
fbem-26172	94	14	dataset	dataset	NOUN
fbem-26172	94	15	,	,	PUNCT
fbem-26172	94	16	its	its	PRON
fbem-26172	94	17	generalization	generalization	NOUN
fbem-26172	94	18	to	to	ADP
fbem-26172	94	19	different	different	ADJ
fbem-26172	94	20	customer	customer	NOUN
fbem-26172	94	21	service	service	NOUN
fbem-26172	94	22	scenarios	scenario	NOUN
fbem-26172	94	23	and	and	CCONJ
fbem-26172	94	24	cultural	cultural	ADJ
fbem-26172	94	25	contexts	context	NOUN
fbem-26172	94	26	might	might	AUX
fbem-26172	94	27	require	require	VERB
fbem-26172	94	28	further	further	ADJ
fbem-26172	94	29	tuning	tuning	NOUN
fbem-26172	94	30	and	and	CCONJ
fbem-26172	94	31	adaptation	adaptation	NOUN
fbem-26172	94	32	.	.	PUNCT
fbem-26172	95	1	4.1	4.1	NUM
fbem-26172	95	2	.	.	PUNCT
fbem-26172	95	3	model	model	PROPN
fbem-26172	95	4	comparison	comparison	NOUN
fbem-26172	95	5	4.1.1	4.1.1	NUM
fbem-26172	95	6	.	.	PUNCT
fbem-26172	96	1	other	other	ADJ
fbem-26172	96	2	model	model	PROPN
fbem-26172	96	3	descriptions	description	NOUN
fbem-26172	96	4	source	source	NOUN
fbem-26172	96	5	:	:	PUNCT
fbem-26172	96	6	jia	jia	PROPN
fbem-26172	96	7	guo	guo	PROPN
fbem-26172	96	8	,	,	PUNCT
fbem-26172	96	9	deep	deep	ADJ
fbem-26172	96	10	learning	learning	NOUN
fbem-26172	96	11	approach	approach	NOUN
fbem-26172	96	12	to	to	AUX
fbem-26172	96	13	text	text	NOUN
fbem-26172	96	14	analysis	analysis	NOUN
fbem-26172	96	15	for	for	ADP
fbem-26172	96	16	human	human	ADJ
fbem-26172	96	17	emotion	emotion	NOUN
fbem-26172	96	18	detection	detection	NOUN
fbem-26172	96	19	model	model	NOUN
fbem-26172	96	20	description	description	NOUN
fbem-26172	96	21	:	:	PUNCT
fbem-26172	96	22	the	the	DET
fbem-26172	96	23	dlsta	dlsta	NOUN
fbem-26172	96	24	model	model	NOUN
fbem-26172	96	25	uses	use	VERB
fbem-26172	96	26	deep	deep	ADJ
fbem-26172	96	27	learning	learning	NOUN
fbem-26172	96	28	techniques	technique	NOUN
fbem-26172	96	29	for	for	ADP
fbem-26172	96	30	text	text	NOUN
fbem-26172	96	31	-	-	PUNCT
fbem-26172	96	32	based	base	VERB
fbem-26172	96	33	emotion	emotion	NOUN
fbem-26172	96	34	detection	detection	NOUN
fbem-26172	96	35	.	.	PUNCT
fbem-26172	97	1	the	the	DET
fbem-26172	97	2	approach	approach	NOUN
fbem-26172	97	3	combines	combine	VERB
fbem-26172	97	4	natural	natural	ADJ
fbem-26172	97	5	language	language	NOUN
fbem-26172	97	6	processing	processing	NOUN
fbem-26172	97	7	(	(	PUNCT
fbem-26172	97	8	nlp	nlp	NOUN
fbem-26172	97	9	)	)	PUNCT
fbem-26172	97	10	and	and	CCONJ
fbem-26172	97	11	deep	deep	ADJ
fbem-26172	97	12	learning	learning	NOUN
fbem-26172	97	13	to	to	PART
fbem-26172	97	14	analyze	analyze	VERB
fbem-26172	97	15	textual	textual	ADJ
fbem-26172	97	16	data	datum	NOUN
fbem-26172	97	17	.	.	PUNCT
fbem-26172	98	1	key	key	ADJ
fbem-26172	98	2	features	feature	NOUN
fbem-26172	98	3	include	include	VERB
fbem-26172	98	4	word	word	NOUN
fbem-26172	98	5	embeddings	embedding	NOUN
fbem-26172	98	6	to	to	PART
fbem-26172	98	7	capture	capture	VERB
fbem-26172	98	8	semantic	semantic	ADJ
fbem-26172	98	9	and	and	CCONJ
fbem-26172	98	10	syntactic	syntactic	ADJ
fbem-26172	98	11	features	feature	NOUN
fbem-26172	98	12	,	,	PUNCT
fbem-26172	98	13	and	and	CCONJ
fbem-26172	98	14	a	a	DET
fbem-26172	98	15	combination	combination	NOUN
fbem-26172	98	16	of	of	ADP
fbem-26172	98	17	text	text	NOUN
fbem-26172	98	18	analysis	analysis	NOUN
fbem-26172	98	19	and	and	CCONJ
fbem-26172	98	20	questionnaire	questionnaire	NOUN
fbem-26172	98	21	-	-	PUNCT
fbem-26172	98	22	based	base	VERB
fbem-26172	98	23	methods	method	NOUN
fbem-26172	98	24	for	for	ADP
fbem-26172	98	25	feature	feature	NOUN
fbem-26172	98	26	extraction	extraction	NOUN
fbem-26172	98	27	.	.	PUNCT
fbem-26172	99	1	the	the	DET
fbem-26172	99	2	model	model	NOUN
fbem-26172	99	3	achieves	achieve	VERB
fbem-26172	99	4	high	high	ADJ
fbem-26172	99	5	prediction	prediction	NOUN
fbem-26172	99	6	accuracy	accuracy	NOUN
fbem-26172	99	7	and	and	CCONJ
fbem-26172	99	8	recall	recall	NOUN
fbem-26172	99	9	rates	rate	NOUN
fbem-26172	99	10	by	by	ADP
fbem-26172	99	11	integrating	integrate	VERB
fbem-26172	99	12	multiple	multiple	ADJ
fbem-26172	99	13	layers	layer	NOUN
fbem-26172	99	14	of	of	ADP
fbem-26172	99	15	neural	neural	ADJ
fbem-26172	99	16	networks	network	NOUN
fbem-26172	99	17	.	.	PUNCT
fbem-26172	100	1	performance	performance	NOUN
fbem-26172	100	2	:	:	PUNCT
fbem-26172	100	3	the	the	DET
fbem-26172	100	4	dlsta	dlsta	NOUN
fbem-26172	100	5	model	model	NOUN
fbem-26172	100	6	shows	show	VERB
fbem-26172	100	7	an	an	DET
fbem-26172	100	8	emotion	emotion	NOUN
fbem-26172	100	9	detection	detection	NOUN
fbem-26172	100	10	rate	rate	NOUN
fbem-26172	100	11	of	of	ADP
fbem-26172	100	12	97.22	97.22	NUM
fbem-26172	100	13	%	%	NOUN
fbem-26172	100	14	and	and	CCONJ
fbem-26172	100	15	classification	classification	NOUN
fbem-26172	100	16	accuracy	accuracy	NOUN
fbem-26172	100	17	of	of	ADP
fbem-26172	100	18	98.02	98.02	NUM
fbem-26172	100	19	%	%	NOUN
fbem-26172	100	20	.	.	PUNCT
fbem-26172	101	1	4.1.2	4.1.2	X
fbem-26172	101	2	.	.	PUNCT
fbem-26172	101	3	multimodal	multimodal	ADJ
fbem-26172	101	4	emotion	emotion	NOUN
fbem-26172	101	5	recognition	recognition	NOUN
fbem-26172	101	6	using	use	VERB
fbem-26172	101	7	deep	deep	ADJ
fbem-26172	101	8	learning	learning	NOUN
fbem-26172	101	9	architectures	architecture	NOUN
fbem-26172	101	10	(	(	PUNCT
fbem-26172	101	11	dbn	dbn	PROPN
fbem-26172	101	12	)	)	PUNCT
fbem-26172	102	1	[	[	X
fbem-26172	102	2	7	7	NUM
fbem-26172	102	3	]	]	SYM
fbem-26172	102	4	source	source	NOUN
fbem-26172	102	5	:	:	PUNCT
fbem-26172	102	6	hiranmayi	hiranmayi	PROPN
fbem-26172	102	7	ranganathan	ranganathan	PROPN
fbem-26172	102	8	,	,	PUNCT
fbem-26172	102	9	shayok	shayok	NOUN
fbem-26172	102	10	chakraborty	chakraborty	PROPN
fbem-26172	102	11	,	,	PUNCT
fbem-26172	102	12	and	and	CCONJ
fbem-26172	102	13	sethuraman	sethuraman	PROPN
fbem-26172	102	14	panchanathan	panchanathan	NOUN
fbem-26172	102	15	,	,	PUNCT
fbem-26172	102	16	center	center	NOUN
fbem-26172	102	17	for	for	ADP
fbem-26172	102	18	cognitive	cognitive	ADJ
fbem-26172	102	19	ubiquitous	ubiquitous	ADJ
fbem-26172	102	20	computing	computing	NOUN
fbem-26172	102	21	(	(	PUNCT
fbem-26172	102	22	cubic	cubic	NOUN
fbem-26172	102	23	)	)	PUNCT
fbem-26172	102	24	,	,	PUNCT
fbem-26172	102	25	arizona	arizona	PROPN
fbem-26172	102	26	state	state	PROPN
fbem-26172	102	27	university	university	PROPN
fbem-26172	102	28	model	model	NOUN
fbem-26172	102	29	description	description	NOUN
fbem-26172	102	30	:	:	PUNCT
fbem-26172	102	31	this	this	DET
fbem-26172	102	32	approach	approach	NOUN
fbem-26172	102	33	uses	use	VERB
fbem-26172	102	34	a	a	DET
fbem-26172	102	35	multimodal	multimodal	NOUN
fbem-26172	102	36	dataset	dataset	NOUN
fbem-26172	102	37	that	that	PRON
fbem-26172	102	38	includes	include	VERB
fbem-26172	102	39	facial	facial	ADJ
fbem-26172	102	40	expressions	expression	NOUN
fbem-26172	102	41	,	,	PUNCT
fbem-26172	102	42	body	body	NOUN
fbem-26172	102	43	gestures	gesture	NOUN
fbem-26172	102	44	,	,	PUNCT
fbem-26172	102	45	voice	voice	NOUN
fbem-26172	102	46	,	,	PUNCT
fbem-26172	102	47	and	and	CCONJ
fbem-26172	102	48	physiological	physiological	ADJ
fbem-26172	102	49	signals	signal	NOUN
fbem-26172	102	50	.	.	PUNCT
fbem-26172	103	1	the	the	DET
fbem-26172	103	2	model	model	NOUN
fbem-26172	103	3	utilizes	utilize	VERB
fbem-26172	103	4	deep	deep	ADJ
fbem-26172	103	5	belief	belief	NOUN
fbem-26172	103	6	networks	network	NOUN
fbem-26172	103	7	(	(	PUNCT
fbem-26172	103	8	dbns	dbns	PROPN
fbem-26172	103	9	)	)	PUNCT
fbem-26172	103	10	to	to	PART
fbem-26172	103	11	generate	generate	VERB
fbem-26172	103	12	robust	robust	ADJ
fbem-26172	103	13	multimodal	multimodal	NOUN
fbem-26172	103	14	features	feature	NOUN
fbem-26172	103	15	for	for	ADP
fbem-26172	103	16	emotion	emotion	NOUN
fbem-26172	103	17	classification	classification	NOUN
fbem-26172	103	18	in	in	ADP
fbem-26172	103	19	an	an	DET
fbem-26172	103	20	unsupervised	unsupervised	ADJ
fbem-26172	103	21	manner	manner	NOUN
fbem-26172	103	22	.	.	PUNCT
fbem-26172	104	1	the	the	DET
fbem-26172	104	2	dbn	dbn	PROPN
fbem-26172	104	3	models	model	NOUN
fbem-26172	104	4	are	be	AUX
fbem-26172	104	5	designed	design	VERB
fbem-26172	104	6	to	to	PART
fbem-26172	104	7	learn	learn	VERB
fbem-26172	104	8	hierarchical	hierarchical	ADJ
fbem-26172	104	9	features	feature	NOUN
fbem-26172	104	10	from	from	ADP
fbem-26172	104	11	different	different	ADJ
fbem-26172	104	12	modalities	modality	NOUN
fbem-26172	104	13	to	to	PART
fbem-26172	104	14	improve	improve	VERB
fbem-26172	104	15	emotion	emotion	NOUN
fbem-26172	104	16	recognition	recognition	NOUN
fbem-26172	104	17	.	.	PUNCT
fbem-26172	105	1	performance	performance	NOUN
fbem-26172	105	2	:	:	PUNCT
fbem-26172	105	3	the	the	DET
fbem-26172	105	4	experimental	experimental	ADJ
fbem-26172	105	5	results	result	NOUN
fbem-26172	105	6	show	show	VERB
fbem-26172	105	7	that	that	SCONJ
fbem-26172	105	8	the	the	DET
fbem-26172	105	9	dbn	dbn	PROPN
fbem-26172	105	10	models	model	NOUN
fbem-26172	105	11	outperform	outperform	VERB
fbem-26172	105	12	state	state	NOUN
fbem-26172	105	13	-	-	PUNCT
fbem-26172	105	14	of	of	ADP
fbem-26172	105	15	-	-	PUNCT
fbem-26172	105	16	the	the	DET
fbem-26172	105	17	-	-	PUNCT
fbem-26172	105	18	art	art	NOUN
fbem-26172	105	19	methods	method	NOUN
fbem-26172	105	20	in	in	ADP
fbem-26172	105	21	emotion	emotion	NOUN
fbem-26172	105	22	recognition	recognition	NOUN
fbem-26172	105	23	.	.	PUNCT
fbem-26172	106	1	4.1.3	4.1.3	X
fbem-26172	106	2	.	.	PUNCT
fbem-26172	106	3	dlsta	dlsta	NOUN
fbem-26172	106	4	vs.	vs.	ADP
fbem-26172	106	5	my	my	PRON
fbem-26172	106	6	model	model	NOUN
fbem-26172	106	7	scope	scope	NOUN
fbem-26172	106	8	:	:	PUNCT
fbem-26172	106	9	dlsta	dlsta	NOUN
fbem-26172	106	10	focuses	focus	VERB
fbem-26172	106	11	solely	solely	ADV
fbem-26172	106	12	on	on	ADP
fbem-26172	106	13	textual	textual	ADJ
fbem-26172	106	14	data	datum	NOUN
fbem-26172	106	15	for	for	ADP
fbem-26172	106	16	emotion	emotion	NOUN
fbem-26172	106	17	detection	detection	NOUN
fbem-26172	106	18	,	,	PUNCT
fbem-26172	106	19	whereas	whereas	SCONJ
fbem-26172	106	20	my	my	PRON
fbem-26172	106	21	model	model	NOUN
fbem-26172	106	22	combines	combine	VERB
fbem-26172	106	23	both	both	PRON
fbem-26172	106	24	visual	visual	ADJ
fbem-26172	106	25	(	(	PUNCT
fbem-26172	106	26	facial	facial	ADJ
fbem-26172	106	27	recognition	recognition	NOUN
fbem-26172	106	28	)	)	PUNCT
fbem-26172	106	29	and	and	CCONJ
fbem-26172	106	30	textual	textual	ADJ
fbem-26172	106	31	(	(	PUNCT
fbem-26172	106	32	nlp	nlp	NOUN
fbem-26172	106	33	)	)	PUNCT
fbem-26172	106	34	data	datum	NOUN
fbem-26172	106	35	.	.	PUNCT
fbem-26172	107	1	this	this	DET
fbem-26172	107	2	multimodal	multimodal	ADJ
fbem-26172	107	3	approach	approach	NOUN
fbem-26172	107	4	in	in	ADP
fbem-26172	107	5	my	my	PRON
fbem-26172	107	6	model	model	NOUN
fbem-26172	107	7	aims	aim	VERB
fbem-26172	107	8	to	to	PART
fbem-26172	107	9	capture	capture	VERB
fbem-26172	107	10	a	a	DET
fbem-26172	107	11	broader	broad	ADJ
fbem-26172	107	12	range	range	NOUN
fbem-26172	107	13	of	of	ADP
fbem-26172	107	14	emotional	emotional	ADJ
fbem-26172	107	15	cues	cue	NOUN
fbem-26172	107	16	.	.	PUNCT
fbem-26172	108	1	accuracy	accuracy	NOUN
fbem-26172	108	2	:	:	PUNCT
fbem-26172	108	3	dlsta	dlsta	NOUN
fbem-26172	108	4	achieves	achieve	VERB
fbem-26172	108	5	higher	high	ADJ
fbem-26172	108	6	accuracy	accuracy	NOUN
fbem-26172	108	7	rates	rate	NOUN
fbem-26172	108	8	(	(	PUNCT
fbem-26172	108	9	97.22	97.22	NUM
fbem-26172	108	10	%	%	NOUN
fbem-26172	108	11	detection	detection	NOUN
fbem-26172	108	12	,	,	PUNCT
fbem-26172	108	13	98.02	98.02	NUM
fbem-26172	108	14	%	%	NOUN
fbem-26172	108	15	classification	classification	NOUN
fbem-26172	108	16	)	)	PUNCT
fbem-26172	108	17	compared	compare	VERB
fbem-26172	108	18	to	to	ADP
fbem-26172	108	19	my	my	PRON
fbem-26172	108	20	combined	combined	ADJ
fbem-26172	108	21	model	model	NOUN
fbem-26172	108	22	's	's	PART
fbem-26172	108	23	accuracy	accuracy	NOUN
fbem-26172	108	24	of	of	ADP
fbem-26172	108	25	92.3	92.3	NUM
fbem-26172	108	26	%	%	NOUN
fbem-26172	108	27	.	.	PUNCT
fbem-26172	109	1	this	this	PRON
fbem-26172	109	2	indicates	indicate	VERB
fbem-26172	109	3	that	that	SCONJ
fbem-26172	109	4	while	while	SCONJ
fbem-26172	109	5	my	my	PRON
fbem-26172	109	6	model	model	NOUN
fbem-26172	109	7	benefits	benefit	NOUN
fbem-26172	109	8	from	from	ADP
fbem-26172	109	9	multimodal	multimodal	ADJ
fbem-26172	109	10	data	datum	NOUN
fbem-26172	109	11	,	,	PUNCT
fbem-26172	109	12	dlsta	dlsta	PROPN
fbem-26172	109	13	’s	’s	PART
fbem-26172	109	14	specialized	specialized	ADJ
fbem-26172	109	15	text	text	NOUN
fbem-26172	109	16	analysis	analysis	NOUN
fbem-26172	109	17	and	and	CCONJ
fbem-26172	109	18	deeper	deep	ADJ
fbem-26172	109	19	integration	integration	NOUN
fbem-26172	109	20	of	of	ADP
fbem-26172	109	21	nlp	nlp	NOUN
fbem-26172	109	22	techniques	technique	NOUN
fbem-26172	109	23	provide	provide	VERB
fbem-26172	109	24	superior	superior	ADJ
fbem-26172	109	25	performance	performance	NOUN
fbem-26172	109	26	in	in	ADP
fbem-26172	109	27	the	the	DET
fbem-26172	109	28	text	text	NOUN
fbem-26172	109	29	domain	domain	NOUN
fbem-26172	109	30	.	.	PUNCT
fbem-26172	110	1	methodology	methodology	NOUN
fbem-26172	110	2	:	:	PUNCT
fbem-26172	110	3	dlsta	dlsta	PROPN
fbem-26172	110	4	uses	use	VERB
fbem-26172	110	5	deep	deep	ADJ
fbem-26172	110	6	learning	learning	NOUN
fbem-26172	110	7	for	for	ADP
fbem-26172	110	8	semantic	semantic	ADJ
fbem-26172	110	9	analysis	analysis	NOUN
fbem-26172	110	10	and	and	CCONJ
fbem-26172	110	11	integrates	integrate	NOUN
fbem-26172	110	12	questionnaire	questionnaire	NOUN
fbem-26172	110	13	data	datum	NOUN
fbem-26172	110	14	,	,	PUNCT
fbem-26172	110	15	providing	provide	VERB
fbem-26172	110	16	a	a	DET
fbem-26172	110	17	comprehensive	comprehensive	ADJ
fbem-26172	110	18	approach	approach	NOUN
fbem-26172	110	19	to	to	ADP
fbem-26172	110	20	text	text	NOUN
fbem-26172	110	21	-	-	PUNCT
fbem-26172	110	22	based	base	VERB
fbem-26172	110	23	emotion	emotion	NOUN
fbem-26172	110	24	detection	detection	NOUN
fbem-26172	110	25	.	.	PUNCT
fbem-26172	111	1	my	my	PRON
fbem-26172	111	2	model	model	NOUN
fbem-26172	111	3	,	,	PUNCT
fbem-26172	111	4	on	on	ADP
fbem-26172	111	5	the	the	DET
fbem-26172	111	6	other	other	ADJ
fbem-26172	111	7	hand	hand	NOUN
fbem-26172	111	8	,	,	PUNCT
fbem-26172	111	9	integrates	integrate	VERB
fbem-26172	111	10	cnn	cnn	PROPN
fbem-26172	111	11	for	for	ADP
fbem-26172	111	12	facial	facial	ADJ
fbem-26172	111	13	recognition	recognition	NOUN
fbem-26172	111	14	with	with	ADP
fbem-26172	111	15	lstm	lstm	NOUN
fbem-26172	111	16	for	for	ADP
fbem-26172	111	17	text	text	NOUN
fbem-26172	111	18	analysis	analysis	NOUN
fbem-26172	111	19	,	,	PUNCT
fbem-26172	111	20	emphasizing	emphasize	VERB
fbem-26172	111	21	a	a	DET
fbem-26172	111	22	balance	balance	NOUN
fbem-26172	111	23	between	between	ADP
fbem-26172	111	24	visual	visual	ADJ
fbem-26172	111	25	and	and	CCONJ
fbem-26172	111	26	textual	textual	ADJ
fbem-26172	111	27	inputs	input	NOUN
fbem-26172	111	28	.	.	PUNCT
fbem-26172	112	1	4.1.4	4.1.4	X
fbem-26172	112	2	.	.	PUNCT
fbem-26172	113	1	dbn	dbn	PROPN
fbem-26172	113	2	vs.	vs.	ADP
fbem-26172	113	3	my	my	PRON
fbem-26172	113	4	model	model	NOUN
fbem-26172	113	5	scope	scope	NOUN
fbem-26172	113	6	:	:	PUNCT
fbem-26172	113	7	the	the	DET
fbem-26172	113	8	dbn	dbn	PROPN
fbem-26172	113	9	model	model	NOUN
fbem-26172	113	10	employs	employ	VERB
fbem-26172	113	11	a	a	DET
fbem-26172	113	12	multimodal	multimodal	ADJ
fbem-26172	113	13	approach	approach	NOUN
fbem-26172	113	14	using	use	VERB
fbem-26172	113	15	various	various	ADJ
fbem-26172	113	16	inputs	input	NOUN
fbem-26172	113	17	such	such	ADJ
fbem-26172	113	18	as	as	ADP
fbem-26172	113	19	facial	facial	ADJ
fbem-26172	113	20	expressions	expression	NOUN
fbem-26172	113	21	,	,	PUNCT
fbem-26172	113	22	body	body	NOUN
fbem-26172	113	23	gestures	gesture	NOUN
fbem-26172	113	24	,	,	PUNCT
fbem-26172	113	25	voice	voice	NOUN
fbem-26172	113	26	,	,	PUNCT
fbem-26172	113	27	and	and	CCONJ
fbem-26172	113	28	physiological	physiological	ADJ
fbem-26172	113	29	signals	signal	NOUN
fbem-26172	113	30	.	.	PUNCT
fbem-26172	114	1	my	my	PRON
fbem-26172	114	2	model	model	NOUN
fbem-26172	114	3	uses	use	VERB
fbem-26172	114	4	a	a	DET
fbem-26172	114	5	simpler	simple	ADJ
fbem-26172	114	6	multimodal	multimodal	NOUN
fbem-26172	114	7	approach	approach	NOUN
fbem-26172	114	8	focusing	focus	VERB
fbem-26172	114	9	on	on	ADP
fbem-26172	114	10	facial	facial	ADJ
fbem-26172	114	11	expressions	expression	NOUN
fbem-26172	114	12	and	and	CCONJ
fbem-26172	114	13	text	text	NOUN
fbem-26172	114	14	.	.	PUNCT
fbem-26172	115	1	performance	performance	NOUN
fbem-26172	115	2	:	:	PUNCT
fbem-26172	115	3	dbn	dbn	PROPN
fbem-26172	115	4	models	model	NOUN
fbem-26172	115	5	outperform	outperform	VERB
fbem-26172	115	6	other	other	ADJ
fbem-26172	115	7	methods	method	NOUN
fbem-26172	115	8	in	in	ADP
fbem-26172	115	9	emotion	emotion	NOUN
fbem-26172	115	10	recognition	recognition	NOUN
fbem-26172	115	11	,	,	PUNCT
fbem-26172	115	12	suggesting	suggest	VERB
fbem-26172	115	13	that	that	SCONJ
fbem-26172	115	14	incorporating	incorporate	VERB
fbem-26172	115	15	multiple	multiple	ADJ
fbem-26172	115	16	types	type	NOUN
fbem-26172	115	17	of	of	ADP
fbem-26172	115	18	data	datum	NOUN
fbem-26172	115	19	(	(	PUNCT
fbem-26172	115	20	visual	visual	ADJ
fbem-26172	115	21	,	,	PUNCT
fbem-26172	115	22	auditory	auditory	ADJ
fbem-26172	115	23	,	,	PUNCT
fbem-26172	115	24	physiological	physiological	ADJ
fbem-26172	115	25	)	)	PUNCT
fbem-26172	115	26	provides	provide	VERB
fbem-26172	115	27	a	a	DET
fbem-26172	115	28	more	more	ADV
fbem-26172	115	29	robust	robust	ADJ
fbem-26172	115	30	understanding	understanding	NOUN
fbem-26172	115	31	of	of	ADP
fbem-26172	115	32	emotions	emotion	NOUN
fbem-26172	115	33	.	.	PUNCT
fbem-26172	116	1	this	this	DET
fbem-26172	116	2	comprehensive	comprehensive	ADJ
fbem-26172	116	3	multimodal	multimodal	NOUN
fbem-26172	116	4	strategy	strategy	NOUN
fbem-26172	116	5	potentially	potentially	ADV
fbem-26172	116	6	offers	offer	VERB
fbem-26172	116	7	better	well	ADJ
fbem-26172	116	8	emotion	emotion	NOUN
fbem-26172	116	9	detection	detection	NOUN
fbem-26172	116	10	accuracy	accuracy	NOUN
fbem-26172	116	11	compared	compare	VERB
fbem-26172	116	12	to	to	ADP
fbem-26172	116	13	my	my	PRON
fbem-26172	116	14	model	model	NOUN
fbem-26172	116	15	's	's	PART
fbem-26172	116	16	combined	combined	ADJ
fbem-26172	116	17	approach	approach	NOUN
fbem-26172	116	18	.	.	PUNCT
fbem-26172	117	1	methodology	methodology	NOUN
fbem-26172	117	2	:	:	PUNCT
fbem-26172	118	1	dbn	dbn	PROPN
fbem-26172	118	2	models	model	NOUN
fbem-26172	118	3	learn	learn	VERB
fbem-26172	118	4	hierarchical	hierarchical	ADJ
fbem-26172	118	5	features	feature	NOUN
fbem-26172	118	6	from	from	ADP
fbem-26172	118	7	different	different	ADJ
fbem-26172	118	8	modalities	modality	NOUN
fbem-26172	118	9	in	in	ADP
fbem-26172	118	10	an	an	DET
fbem-26172	118	11	unsupervised	unsupervised	ADJ
fbem-26172	118	12	manner	manner	NOUN
fbem-26172	118	13	,	,	PUNCT
fbem-26172	118	14	allowing	allow	VERB
fbem-26172	118	15	for	for	ADP
fbem-26172	118	16	more	more	ADV
fbem-26172	118	17	nuanced	nuanced	ADJ
fbem-26172	118	18	emotion	emotion	NOUN
fbem-26172	118	19	classification	classification	NOUN
fbem-26172	118	20	.	.	PUNCT
fbem-26172	119	1	my	my	PRON
fbem-26172	119	2	model	model	NOUN
fbem-26172	119	3	relies	rely	VERB
fbem-26172	119	4	on	on	ADP
fbem-26172	119	5	supervised	supervised	ADJ
fbem-26172	119	6	learning	learn	VERB
fbem-26172	119	7	techniques	technique	NOUN
fbem-26172	119	8	with	with	ADP
fbem-26172	119	9	cnn	cnn	PROPN
fbem-26172	119	10	and	and	CCONJ
fbem-26172	119	11	lstm	lstm	PROPN
fbem-26172	119	12	,	,	PUNCT
fbem-26172	119	13	which	which	PRON
fbem-26172	119	14	might	might	AUX
fbem-26172	119	15	limit	limit	VERB
fbem-26172	119	16	its	its	PRON
fbem-26172	119	17	ability	ability	NOUN
fbem-26172	119	18	to	to	PART
fbem-26172	119	19	capture	capture	VERB
fbem-26172	119	20	the	the	DET
fbem-26172	119	21	full	full	ADJ
fbem-26172	119	22	range	range	NOUN
fbem-26172	119	23	of	of	ADP
fbem-26172	119	24	emotional	emotional	ADJ
fbem-26172	119	25	expressions	expression	NOUN
fbem-26172	119	26	compared	compare	VERB
fbem-26172	119	27	to	to	ADP
fbem-26172	119	28	dbns	dbns	NOUN
fbem-26172	119	29	.	.	PUNCT
fbem-26172	120	1	5	5	X
fbem-26172	120	2	.	.	X
fbem-26172	120	3	conclusion	conclusion	VERB
fbem-26172	120	4	the	the	DET
fbem-26172	120	5	integration	integration	NOUN
fbem-26172	120	6	of	of	ADP
fbem-26172	120	7	facial	facial	ADJ
fbem-26172	120	8	expression	expression	NOUN
fbem-26172	120	9	recognition	recognition	NOUN
fbem-26172	120	10	(	(	PUNCT
fbem-26172	120	11	fer	fer	PROPN
fbem-26172	120	12	)	)	PUNCT
fbem-26172	120	13	and	and	CCONJ
fbem-26172	120	14	natural	natural	ADJ
fbem-26172	120	15	language	language	NOUN
fbem-26172	120	16	processing	processing	NOUN
fbem-26172	120	17	(	(	PUNCT
fbem-26172	120	18	nlp	nlp	NOUN
fbem-26172	120	19	)	)	PUNCT
fbem-26172	120	20	for	for	ADP
fbem-26172	120	21	customer	customer	NOUN
fbem-26172	120	22	service	service	NOUN
fbem-26172	120	23	interactions	interaction	NOUN
fbem-26172	120	24	provides	provide	VERB
fbem-26172	120	25	a	a	DET
fbem-26172	120	26	powerful	powerful	ADJ
fbem-26172	120	27	tool	tool	NOUN
fbem-26172	120	28	for	for	ADP
fbem-26172	120	29	comprehensive	comprehensive	ADJ
fbem-26172	120	30	emotion	emotion	NOUN
fbem-26172	120	31	detection	detection	NOUN
fbem-26172	120	32	.	.	PUNCT
fbem-26172	121	1	our	our	PRON
fbem-26172	121	2	experiments	experiment	NOUN
fbem-26172	121	3	demonstrated	demonstrate	VERB
fbem-26172	121	4	that	that	SCONJ
fbem-26172	121	5	the	the	DET
fbem-26172	121	6	multimodal	multimodal	ADJ
fbem-26172	121	7	approach	approach	NOUN
fbem-26172	121	8	significantly	significantly	ADV
fbem-26172	121	9	improves	improve	VERB
fbem-26172	121	10	the	the	DET
fbem-26172	121	11	accuracy	accuracy	NOUN
fbem-26172	121	12	and	and	CCONJ
fbem-26172	121	13	robustness	robustness	NOUN
fbem-26172	121	14	of	of	ADP
fbem-26172	121	15	emotion	emotion	NOUN
fbem-26172	121	16	recognition	recognition	NOUN
fbem-26172	121	17	compared	compare	VERB
fbem-26172	121	18	to	to	ADP
fbem-26172	121	19	unimodal	unimodal	ADJ
fbem-26172	121	20	systems	system	NOUN
fbem-26172	121	21	.	.	PUNCT
fbem-26172	122	1	this	this	DET
fbem-26172	122	2	advancement	advancement	NOUN
fbem-26172	122	3	holds	hold	VERB
fbem-26172	122	4	great	great	ADJ
fbem-26172	122	5	potential	potential	NOUN
fbem-26172	122	6	for	for	ADP
fbem-26172	122	7	enhancing	enhance	VERB
fbem-26172	122	8	customer	customer	NOUN
fbem-26172	122	9	service	service	NOUN
fbem-26172	122	10	experiences	experience	NOUN
fbem-26172	122	11	,	,	PUNCT
fbem-26172	122	12	making	make	VERB
fbem-26172	122	13	interactions	interaction	NOUN
fbem-26172	122	14	more	more	ADV
fbem-26172	122	15	empathetic	empathetic	ADJ
fbem-26172	122	16	and	and	CCONJ
fbem-26172	122	17	effective	effective	ADJ
fbem-26172	122	18	.	.	PUNCT
fbem-26172	123	1	future	future	ADJ
fbem-26172	123	2	work	work	NOUN
fbem-26172	123	3	should	should	AUX
fbem-26172	123	4	focus	focus	VERB
fbem-26172	123	5	on	on	ADP
fbem-26172	123	6	optimizing	optimize	VERB
fbem-26172	123	7	the	the	DET
fbem-26172	123	8	computational	computational	ADJ
fbem-26172	123	9	efficiency	efficiency	NOUN
fbem-26172	123	10	of	of	ADP
fbem-26172	123	11	the	the	DET
fbem-26172	123	12	model	model	NOUN
fbem-26172	123	13	,	,	PUNCT
fbem-26172	123	14	addressing	address	VERB
fbem-26172	123	15	data	datum	NOUN
fbem-26172	123	16	variability	variability	NOUN
fbem-26172	123	17	,	,	PUNCT
fbem-26172	123	18	and	and	CCONJ
fbem-26172	123	19	ensuring	ensure	VERB
fbem-26172	123	20	the	the	DET
fbem-26172	123	21	system	system	NOUN
fbem-26172	123	22	's	's	PART
fbem-26172	123	23	adaptability	adaptability	NOUN
fbem-26172	123	24	to	to	ADP
fbem-26172	123	25	different	different	ADJ
fbem-26172	123	26	contexts	context	NOUN
fbem-26172	123	27	and	and	CCONJ
fbem-26172	123	28	cultural	cultural	ADJ
fbem-26172	123	29	nuances	nuance	NOUN
fbem-26172	123	30	.	.	PUNCT
fbem-26172	124	1	continued	continue	VERB
fbem-26172	124	2	research	research	NOUN
fbem-26172	124	3	and	and	CCONJ
fbem-26172	124	4	development	development	NOUN
fbem-26172	124	5	in	in	ADP
fbem-26172	124	6	this	this	DET
fbem-26172	124	7	area	area	NOUN
fbem-26172	124	8	will	will	AUX
fbem-26172	124	9	further	far	ADV
fbem-26172	124	10	solidify	solidify	VERB
fbem-26172	124	11	the	the	DET
fbem-26172	124	12	role	role	NOUN
fbem-26172	124	13	of	of	ADP
fbem-26172	124	14	integrated	integrate	VERB
fbem-26172	124	15	fer	fer	PROPN
fbem-26172	124	16	and	and	CCONJ
fbem-26172	124	17	nlp	nlp	NOUN
fbem-26172	124	18	systems	system	NOUN
fbem-26172	124	19	in	in	ADP
fbem-26172	124	20	transforming	transform	VERB
fbem-26172	124	21	customer	customer	NOUN
fbem-26172	124	22	service	service	NOUN
fbem-26172	124	23	interactions	interaction	NOUN
fbem-26172	124	24	.	.	PUNCT
fbem-26172	125	1	75	75	NUM
fbem-26172	125	2	references	reference	NOUN
fbem-26172	125	3	[	[	X
fbem-26172	125	4	1	1	NUM
fbem-26172	125	5	]	]	PUNCT
fbem-26172	125	6	h.	h.	PROPN
fbem-26172	125	7	ranganathan	ranganathan	PROPN
fbem-26172	125	8	,	,	PUNCT
fbem-26172	125	9	s.	s.	PROPN
fbem-26172	125	10	chakraborty	chakraborty	PROPN
fbem-26172	125	11	and	and	CCONJ
fbem-26172	125	12	s.	s.	PROPN
fbem-26172	125	13	panchanathan	panchanathan	PROPN
fbem-26172	125	14	,	,	PUNCT
fbem-26172	125	15	"	"	PUNCT
fbem-26172	125	16	multimodal	multimodal	ADJ
fbem-26172	125	17	emotion	emotion	NOUN
fbem-26172	125	18	recognition	recognition	NOUN
fbem-26172	125	19	using	use	VERB
fbem-26172	125	20	deep	deep	ADJ
fbem-26172	125	21	learning	learning	NOUN
fbem-26172	125	22	architectures	architecture	NOUN
fbem-26172	125	23	,	,	PUNCT
fbem-26172	125	24	"	"	PUNCT
fbem-26172	125	25	2016	2016	NUM
fbem-26172	125	26	ieee	ieee	NOUN
fbem-26172	125	27	winter	winter	NOUN
fbem-26172	125	28	conference	conference	NOUN
fbem-26172	125	29	on	on	ADP
fbem-26172	125	30	applications	application	NOUN
fbem-26172	125	31	of	of	ADP
fbem-26172	125	32	computer	computer	NOUN
fbem-26172	125	33	vision	vision	NOUN
fbem-26172	125	34	(	(	PUNCT
fbem-26172	125	35	wacv	wacv	NOUN
fbem-26172	125	36	)	)	PUNCT
fbem-26172	125	37	,	,	PUNCT
fbem-26172	125	38	lake	lake	PROPN
fbem-26172	125	39	placid	placid	ADJ
fbem-26172	125	40	,	,	PUNCT
fbem-26172	125	41	ny	ny	PROPN
fbem-26172	125	42	,	,	PUNCT
fbem-26172	125	43	usa	usa	PROPN
fbem-26172	125	44	,	,	PUNCT
fbem-26172	125	45	2016	2016	NUM
fbem-26172	125	46	,	,	PUNCT
fbem-26172	125	47	pp	pp	ADJ
fbem-26172	125	48	.	.	PUNCT
fbem-26172	126	1	1	1	NUM
fbem-26172	126	2	-	-	SYM
fbem-26172	126	3	9	9	NUM
fbem-26172	126	4	,	,	PUNCT
fbem-26172	126	5	doi	doi	NOUN
fbem-26172	126	6	:	:	PUNCT
fbem-26172	126	7	10.1109	10.1109	NUM
fbem-26172	126	8	/	/	SYM
fbem-26172	126	9	wacv.2016.7477679	wacv.2016.7477679	NOUN
fbem-26172	126	10	.	.	PUNCT
fbem-26172	127	1	keywords	keyword	NOUN
fbem-26172	127	2	:	:	PUNCT
fbem-26172	127	3	{	{	PUNCT
fbem-26172	127	4	databases	database	NOUN
fbem-26172	127	5	;	;	PUNCT
fbem-26172	127	6	emotion	emotion	NOUN
fbem-26172	127	7	recognition	recognition	NOUN
fbem-26172	127	8	;	;	PUNCT
fbem-26172	127	9	face	face	NOUN
fbem-26172	127	10	;	;	PUNCT
fbem-26172	127	11	physiology	physiology	NOUN
fbem-26172	127	12	;	;	PUNCT
fbem-26172	127	13	threedimensional	threedimensional	PROPN
fbem-26172	127	14	displays;machine	displays;machine	PROPN
fbem-26172	127	15	learning	learning	PROPN
fbem-26172	127	16	;	;	PUNCT
fbem-26172	127	17	facial	facial	ADJ
fbem-26172	127	18	features	feature	NOUN
fbem-26172	127	19	}	}	PUNCT
fbem-26172	127	20	,	,	PUNCT
fbem-26172	127	21	[	[	X
fbem-26172	127	22	2	2	NUM
fbem-26172	127	23	]	]	PUNCT
fbem-26172	127	24	m.	m.	NOUN
fbem-26172	127	25	boucart	boucart	NOUN
fbem-26172	127	26	,	,	PUNCT
fbem-26172	127	27	j.-f	j.-f	PROPN
fbem-26172	127	28	.	.	PUNCT
fbem-26172	128	1	dinon	dinon	PROPN
fbem-26172	128	2	,	,	PUNCT
fbem-26172	128	3	p.	p.	NOUN
fbem-26172	128	4	despretz	despretz	NOUN
fbem-26172	128	5	,	,	PUNCT
fbem-26172	128	6	t.	t.	NOUN
fbem-26172	128	7	desmettre	desmettre	PROPN
fbem-26172	128	8	,	,	PUNCT
fbem-26172	128	9	k.	k.	PROPN
fbem-26172	128	10	hladiuk	hladiuk	PROPN
fbem-26172	128	11	,	,	PUNCT
fbem-26172	128	12	and	and	CCONJ
fbem-26172	128	13	a.	a.	NOUN
fbem-26172	128	14	oliva	oliva	PROPN
fbem-26172	128	15	.	.	PUNCT
fbem-26172	129	1	recognition	recognition	NOUN
fbem-26172	129	2	of	of	ADP
fbem-26172	129	3	facial	facial	ADJ
fbem-26172	129	4	emotion	emotion	NOUN
fbem-26172	129	5	in	in	ADP
fbem-26172	129	6	low	low	ADJ
fbem-26172	129	7	vision	vision	NOUN
fbem-26172	129	8	:	:	PUNCT
fbem-26172	129	9	a	a	DET
fbem-26172	129	10	flexible	flexible	ADJ
fbem-26172	129	11	usage	usage	NOUN
fbem-26172	129	12	of	of	ADP
fbem-26172	129	13	facial	facial	ADJ
fbem-26172	129	14	features	feature	NOUN
fbem-26172	129	15	.	.	PUNCT
fbem-26172	130	1	visual	visual	ADJ
fbem-26172	130	2	neuroscience	neuroscience	NOUN
fbem-26172	130	3	,	,	PUNCT
fbem-26172	130	4	25(4):603	25(4):603	PROPN
fbem-26172	130	5	-	-	PUNCT
fbem-26172	130	6	-609	-609	PROPN
fbem-26172	130	7	,	,	PUNCT
fbem-26172	130	8	2008	2008	NUM
fbem-26172	130	9	.	.	PUNCT
fbem-26172	131	1	[	[	X
fbem-26172	131	2	3	3	NUM
fbem-26172	131	3	]	]	X
fbem-26172	131	4	liu	liu	PROPN
fbem-26172	131	5	,	,	PUNCT
fbem-26172	131	6	w.	w.	PROPN
fbem-26172	131	7	,	,	PUNCT
fbem-26172	131	8	wen	wen	PROPN
fbem-26172	131	9	,	,	PUNCT
fbem-26172	131	10	y.	y.	PROPN
fbem-26172	131	11	,	,	PUNCT
fbem-26172	131	12	yu	yu	PROPN
fbem-26172	131	13	,	,	PUNCT
fbem-26172	131	14	z.	z.	PROPN
fbem-26172	131	15	,	,	PUNCT
fbem-26172	131	16	li	li	PROPN
fbem-26172	131	17	,	,	PUNCT
fbem-26172	131	18	m.	m.	NOUN
fbem-26172	131	19	,	,	PUNCT
fbem-26172	131	20	raj	raj	PROPN
fbem-26172	131	21	,	,	PUNCT
fbem-26172	131	22	b.	b.	PROPN
fbem-26172	131	23	,	,	PUNCT
fbem-26172	131	24	&	&	CCONJ
fbem-26172	131	25	song	song	PROPN
fbem-26172	131	26	,	,	PUNCT
fbem-26172	131	27	l.	l.	PROPN
fbem-26172	131	28	(	(	PUNCT
fbem-26172	131	29	2018	2018	NUM
fbem-26172	131	30	)	)	PUNCT
fbem-26172	131	31	.	.	PUNCT
fbem-26172	132	1	sphereface	sphereface	NOUN
fbem-26172	132	2	:	:	PUNCT
fbem-26172	132	3	deep	deep	ADJ
fbem-26172	132	4	hypersphere	hypersphere	NOUN
fbem-26172	132	5	embedding	embed	VERB
fbem-26172	132	6	for	for	ADP
fbem-26172	132	7	face	face	NOUN
fbem-26172	132	8	recognition	recognition	NOUN
fbem-26172	133	1	[	[	X
fbem-26172	133	2	4	4	X
fbem-26172	133	3	]	]	PUNCT
fbem-26172	133	4	j.	j.	PROPN
fbem-26172	133	5	chen	chen	PROPN
fbem-26172	133	6	,	,	PUNCT
fbem-26172	133	7	z.	z.	PROPN
fbem-26172	133	8	chen	chen	PROPN
fbem-26172	133	9	,	,	PUNCT
fbem-26172	133	10	z.	z.	PROPN
fbem-26172	133	11	chi	chi	PROPN
fbem-26172	133	12	,	,	PUNCT
fbem-26172	133	13	and	and	CCONJ
fbem-26172	133	14	h.	h.	PROPN
fbem-26172	133	15	fu	fu	PROPN
fbem-26172	133	16	.	.	PUNCT
fbem-26172	134	1	emotion	emotion	NOUN
fbem-26172	134	2	recognition	recognition	NOUN
fbem-26172	134	3	in	in	ADP
fbem-26172	134	4	the	the	DET
fbem-26172	134	5	wild	wild	NOUN
fbem-26172	134	6	with	with	ADP
fbem-26172	134	7	feature	feature	NOUN
fbem-26172	134	8	fusion	fusion	NOUN
fbem-26172	134	9	and	and	CCONJ
fbem-26172	134	10	multiple	multiple	ADJ
fbem-26172	134	11	kernel	kernel	PROPN
fbem-26172	134	12	learning	learning	NOUN
fbem-26172	134	13	.	.	PUNCT
fbem-26172	135	1	in	in	ADP
fbem-26172	135	2	proceedings	proceeding	NOUN
fbem-26172	135	3	of	of	ADP
fbem-26172	135	4	the	the	DET
fbem-26172	135	5	16th	16th	ADJ
fbem-26172	135	6	international	international	ADJ
fbem-26172	135	7	conference	conference	NOUN
fbem-26172	135	8	on	on	ADP
fbem-26172	135	9	multimodal	multimodal	NOUN
fbem-26172	135	10	interaction	interaction	NOUN
fbem-26172	135	11	,	,	PUNCT
fbem-26172	135	12	icmi	icmi	NOUN
fbem-26172	135	13	'	'	NUM
fbem-26172	135	14	14	14	NUM
fbem-26172	135	15	,	,	PUNCT
fbem-26172	135	16	pages	page	NOUN
fbem-26172	135	17	508	508	NUM
fbem-26172	135	18	-	-	PUNCT
fbem-26172	135	19	-513	-513	PROPN
fbem-26172	135	20	,	,	PUNCT
fbem-26172	135	21	new	new	PROPN
fbem-26172	135	22	york	york	PROPN
fbem-26172	135	23	,	,	PUNCT
fbem-26172	135	24	ny	ny	PROPN
fbem-26172	135	25	,	,	PUNCT
fbem-26172	135	26	usa	usa	PROPN
fbem-26172	135	27	,	,	PUNCT
fbem-26172	135	28	2014	2014	NUM
fbem-26172	135	29	.	.	PUNCT
fbem-26172	136	1	acm	acm	PROPN
fbem-26172	137	1	[	[	X
fbem-26172	137	2	5	5	NUM
fbem-26172	137	3	]	]	PUNCT
fbem-26172	137	4	"	"	PUNCT
fbem-26172	137	5	attribute	attribute	NOUN
fbem-26172	137	6	and	and	CCONJ
fbem-26172	137	7	simile	simile	NOUN
fbem-26172	137	8	classifiers	classifier	NOUN
fbem-26172	137	9	for	for	ADP
fbem-26172	137	10	face	face	NOUN
fbem-26172	137	11	verification,"neeraj	verification,"neeraj	PROPN
fbem-26172	137	12	kumar	kumar	PROPN
fbem-26172	137	13	,	,	PUNCT
fbem-26172	137	14	alexander	alexander	PROPN
fbem-26172	137	15	c.	c.	PROPN
fbem-26172	137	16	berg	berg	PROPN
fbem-26172	137	17	,	,	PUNCT
fbem-26172	137	18	peter	peter	PROPN
fbem-26172	137	19	n.	n.	PROPN
fbem-26172	137	20	belhumeur	belhumeur	PROPN
fbem-26172	137	21	,	,	PUNCT
fbem-26172	137	22	and	and	CCONJ
fbem-26172	137	23	shree	shree	PROPN
fbem-26172	137	24	k.	k.	PROPN
fbem-26172	137	25	nayar	nayar	PROPN
fbem-26172	137	26	,	,	PUNCT
fbem-26172	137	27	international	international	ADJ
fbem-26172	137	28	conference	conference	NOUN
fbem-26172	137	29	on	on	ADP
fbem-26172	137	30	computer	computer	NOUN
fbem-26172	137	31	vision	vision	NOUN
fbem-26172	137	32	(	(	PUNCT
fbem-26172	137	33	iccv	iccv	PROPN
fbem-26172	137	34	)	)	PUNCT
fbem-26172	137	35	,	,	PUNCT
fbem-26172	137	36	2009	2009	NUM
fbem-26172	137	37	.	.	PUNCT
fbem-26172	138	1	[	[	X
fbem-26172	138	2	6	6	NUM
fbem-26172	138	3	]	]	X
fbem-26172	138	4	guo	guo	PROPN
fbem-26172	138	5	,	,	PUNCT
fbem-26172	138	6	j.	j.	PROPN
fbem-26172	138	7	(	(	PUNCT
fbem-26172	138	8	2022	2022	NUM
fbem-26172	138	9	)	)	PUNCT
fbem-26172	138	10	.	.	PUNCT
fbem-26172	139	1	deep	deep	ADJ
fbem-26172	139	2	learning	learn	VERB
fbem-26172	139	3	approach	approach	NOUN
fbem-26172	139	4	to	to	AUX
fbem-26172	139	5	text	text	NOUN
fbem-26172	139	6	analysis	analysis	NOUN
fbem-26172	139	7	for	for	ADP
fbem-26172	139	8	human	human	ADJ
fbem-26172	139	9	emotion	emotion	NOUN
fbem-26172	139	10	detection	detection	NOUN
fbem-26172	139	11	from	from	ADP
fbem-26172	139	12	big	big	ADJ
fbem-26172	139	13	data	datum	NOUN
fbem-26172	139	14	.	.	PUNCT
fbem-26172	140	1	journal	journal	NOUN
fbem-26172	140	2	of	of	ADP
fbem-26172	140	3	intelligent	intelligent	ADJ
fbem-26172	140	4	systems	system	NOUN
fbem-26172	140	5	,	,	PUNCT
fbem-26172	140	6	31(1	31(1	NUM
fbem-26172	140	7	)	)	PUNCT
fbem-26172	140	8	,	,	PUNCT
fbem-26172	140	9	113	113	NUM
fbem-26172	140	10	-	-	SYM
fbem-26172	140	11	126	126	NUM
fbem-26172	140	12	.	.	PUNCT
fbem-26172	141	1	[	[	X
fbem-26172	141	2	7	7	NUM
fbem-26172	141	3	]	]	X
fbem-26172	141	4	ranganathan	ranganathan	PROPN
fbem-26172	141	5	,	,	PUNCT
fbem-26172	141	6	h.	h.	PROPN
fbem-26172	141	7	,	,	PUNCT
fbem-26172	141	8	chakraborty	chakraborty	PROPN
fbem-26172	141	9	,	,	PUNCT
fbem-26172	141	10	s.	s.	PROPN
fbem-26172	141	11	,	,	PUNCT
fbem-26172	141	12	&	&	CCONJ
fbem-26172	141	13	panchanathan	panchanathan	PROPN
fbem-26172	141	14	,	,	PUNCT
fbem-26172	141	15	s.	s.	PROPN
fbem-26172	141	16	(	(	PUNCT
fbem-26172	141	17	2016	2016	NUM
fbem-26172	141	18	,	,	PUNCT
fbem-26172	141	19	november	november	PROPN
fbem-26172	141	20	)	)	PUNCT
fbem-26172	141	21	.	.	PUNCT
fbem-26172	142	1	transfer	transfer	NOUN
fbem-26172	142	2	of	of	ADP
fbem-26172	142	3	multimodal	multimodal	NOUN
fbem-26172	142	4	emotion	emotion	NOUN
fbem-26172	142	5	features	feature	NOUN
fbem-26172	142	6	in	in	ADP
fbem-26172	142	7	deep	deep	ADJ
fbem-26172	142	8	belief	belief	NOUN
fbem-26172	142	9	networks	network	NOUN
fbem-26172	142	10	.	.	PUNCT
fbem-26172	143	1	in	in	ADP
fbem-26172	143	2	2016	2016	NUM
fbem-26172	143	3	50th	50th	ADJ
fbem-26172	143	4	asilomar	asilomar	NOUN
fbem-26172	143	5	conference	conference	NOUN
fbem-26172	143	6	on	on	ADP
fbem-26172	143	7	signals	signal	NOUN
fbem-26172	143	8	,	,	PUNCT
fbem-26172	143	9	systems	system	NOUN
fbem-26172	143	10	and	and	CCONJ
fbem-26172	143	11	computers	computer	NOUN
fbem-26172	143	12	(	(	PUNCT
fbem-26172	143	13	pp	pp	ADV
fbem-26172	143	14	.	.	PUNCT
fbem-26172	144	1	449	449	NUM
fbem-26172	144	2	-	-	NUM
fbem-26172	144	3	453	453	NUM
fbem-26172	144	4	)	)	PUNCT
fbem-26172	144	5	.	.	PUNCT
fbem-26172	145	1	ieee	ieee	PROPN
fbem-26172	145	2	.	.	PUNCT