id	sid	tid	token	lemma	pos
fcis-15262	1	1	frontiers	frontier	NOUN
fcis-15262	1	2	in	in	ADP
fcis-15262	1	3	computing	computing	NOUN
fcis-15262	1	4	and	and	CCONJ
fcis-15262	1	5	intelligent	intelligent	ADJ
fcis-15262	1	6	systems	system	NOUN
fcis-15262	1	7	issn	issn	VERB
fcis-15262	1	8	:	:	PUNCT
fcis-15262	1	9	2832	2832	NUM
fcis-15262	1	10	-	-	SYM
fcis-15262	1	11	6024	6024	NUM
fcis-15262	1	12	|	|	NOUN
fcis-15262	1	13	vol	vol	NOUN
fcis-15262	1	14	.	.	PROPN
fcis-15262	2	1	6	6	NUM
fcis-15262	2	2	,	,	PUNCT
fcis-15262	2	3	no	no	INTJ
fcis-15262	2	4	.	.	NOUN
fcis-15262	2	5	2	2	NUM
fcis-15262	2	6	,	,	PUNCT
fcis-15262	2	7	2023	2023	NUM
fcis-15262	2	8	80	80	NUM
fcis-15262	2	9	research	research	NOUN
fcis-15262	2	10	on	on	ADP
fcis-15262	2	11	voiceprint	voiceprint	NOUN
fcis-15262	2	12	recognition	recognition	NOUN
fcis-15262	2	13	based	base	VERB
fcis-15262	2	14	on	on	ADP
fcis-15262	2	15	densenet	densenet	PROPN
fcis-15262	2	16	yongkang	yongkang	PROPN
fcis-15262	2	17	tang	tang	PROPN
fcis-15262	2	18	,	,	PUNCT
fcis-15262	2	19	xiangguo	xiangguo	PROPN
fcis-15262	2	20	sun	sun	PROPN
fcis-15262	2	21	school	school	NOUN
fcis-15262	2	22	of	of	ADP
fcis-15262	2	23	mechanical	mechanical	ADJ
fcis-15262	2	24	engineering	engineering	NOUN
fcis-15262	2	25	,	,	PUNCT
fcis-15262	2	26	sichuan	sichuan	PROPN
fcis-15262	2	27	university	university	PROPN
fcis-15262	2	28	of	of	ADP
fcis-15262	2	29	science	science	PROPN
fcis-15262	2	30	&	&	CCONJ
fcis-15262	2	31	engineering	engineering	PROPN
fcis-15262	2	32	,	,	PUNCT
fcis-15262	2	33	yibin	yibin	PROPN
fcis-15262	2	34	sichuan	sichuan	PROPN
fcis-15262	2	35	,	,	PUNCT
fcis-15262	2	36	644002	644002	NUM
fcis-15262	2	37	,	,	PUNCT
fcis-15262	2	38	china	china	PROPN
fcis-15262	2	39	abstract	abstract	NOUN
fcis-15262	2	40	:	:	PUNCT
fcis-15262	2	41	to	to	ADP
fcis-15262	2	42	date	date	NOUN
fcis-15262	2	43	,	,	PUNCT
fcis-15262	2	44	most	most	ADJ
fcis-15262	2	45	methods	method	NOUN
fcis-15262	2	46	for	for	ADP
fcis-15262	2	47	voiceprint	voiceprint	NOUN
fcis-15262	2	48	recognition	recognition	NOUN
fcis-15262	2	49	are	be	AUX
fcis-15262	2	50	implemented	implement	VERB
fcis-15262	2	51	based	base	VERB
fcis-15262	2	52	on	on	ADP
fcis-15262	2	53	convolutional	convolutional	ADJ
fcis-15262	2	54	neural	neural	ADJ
fcis-15262	2	55	networks	network	NOUN
fcis-15262	2	56	.	.	PUNCT
fcis-15262	3	1	in	in	ADP
fcis-15262	3	2	order	order	NOUN
fcis-15262	3	3	to	to	PART
fcis-15262	3	4	combine	combine	VERB
fcis-15262	3	5	the	the	DET
fcis-15262	3	6	spatial	spatial	ADJ
fcis-15262	3	7	features	feature	NOUN
fcis-15262	3	8	and	and	CCONJ
fcis-15262	3	9	temporal	temporal	ADJ
fcis-15262	3	10	features	feature	NOUN
fcis-15262	3	11	of	of	ADP
fcis-15262	3	12	sound	sound	NOUN
fcis-15262	3	13	and	and	CCONJ
fcis-15262	3	14	improve	improve	VERB
fcis-15262	3	15	computational	computational	ADJ
fcis-15262	3	16	efficiency	efficiency	NOUN
fcis-15262	3	17	,	,	PUNCT
fcis-15262	3	18	a	a	DET
fcis-15262	3	19	combined	combined	ADJ
fcis-15262	3	20	model	model	NOUN
fcis-15262	3	21	of	of	ADP
fcis-15262	3	22	densenet169	densenet169	PROPN
fcis-15262	3	23	and	and	CCONJ
fcis-15262	3	24	bidirectional	bidirectional	ADJ
fcis-15262	3	25	gated	gate	VERB
fcis-15262	3	26	recurrent	recurrent	ADJ
fcis-15262	3	27	unit	unit	NOUN
fcis-15262	3	28	(	(	PUNCT
fcis-15262	3	29	bigru	bigru	NOUN
fcis-15262	3	30	)	)	PUNCT
fcis-15262	3	31	,	,	PUNCT
fcis-15262	3	32	named	name	VERB
fcis-15262	3	33	densenet169	densenet169	PROPN
fcis-15262	3	34	-	-	PUNCT
fcis-15262	3	35	bigru	bigru	NOUN
fcis-15262	3	36	,	,	PUNCT
fcis-15262	3	37	is	be	AUX
fcis-15262	3	38	proposed	propose	VERB
fcis-15262	3	39	.	.	PUNCT
fcis-15262	4	1	firstly	firstly	ADV
fcis-15262	4	2	,	,	PUNCT
fcis-15262	4	3	deep	deep	ADJ
fcis-15262	4	4	separable	separable	ADJ
fcis-15262	4	5	convolution	convolution	NOUN
fcis-15262	4	6	is	be	AUX
fcis-15262	4	7	employed	employ	VERB
fcis-15262	4	8	to	to	PART
fcis-15262	4	9	reduce	reduce	VERB
fcis-15262	4	10	the	the	DET
fcis-15262	4	11	number	number	NOUN
fcis-15262	4	12	of	of	ADP
fcis-15262	4	13	network	network	NOUN
fcis-15262	4	14	parameters	parameter	NOUN
fcis-15262	4	15	and	and	CCONJ
fcis-15262	4	16	enhance	enhance	VERB
fcis-15262	4	17	computational	computational	ADJ
fcis-15262	4	18	efficiency	efficiency	NOUN
fcis-15262	4	19	.	.	PUNCT
fcis-15262	5	1	the	the	DET
fcis-15262	5	2	speech	speech	NOUN
fcis-15262	5	3	fbank	fbank	NOUN
fcis-15262	5	4	feature	feature	NOUN
fcis-15262	5	5	map	map	NOUN
fcis-15262	5	6	is	be	AUX
fcis-15262	5	7	extracted	extract	VERB
fcis-15262	5	8	after	after	ADP
fcis-15262	5	9	preprocessing	preprocessing	NOUN
fcis-15262	5	10	,	,	PUNCT
fcis-15262	5	11	and	and	CCONJ
fcis-15262	5	12	experimental	experimental	ADJ
fcis-15262	5	13	results	result	NOUN
fcis-15262	5	14	show	show	VERB
fcis-15262	5	15	that	that	SCONJ
fcis-15262	5	16	the	the	DET
fcis-15262	5	17	densenet169	densenet169	PROPN
fcis-15262	5	18	-	-	PUNCT
fcis-15262	5	19	bigru	bigru	NOUN
fcis-15262	5	20	model	model	NOUN
fcis-15262	5	21	outperforms	outperform	NOUN
fcis-15262	5	22	both	both	PRON
fcis-15262	5	23	densenet169	densenet169	PROPN
fcis-15262	5	24	and	and	CCONJ
fcis-15262	5	25	densenet169	densenet169	PROPN
fcis-15262	5	26	-	-	PUNCT
fcis-15262	5	27	gru	gru	NOUN
fcis-15262	5	28	models	model	NOUN
fcis-15262	5	29	,	,	PUNCT
fcis-15262	5	30	with	with	ADP
fcis-15262	5	31	an	an	DET
fcis-15262	5	32	improvement	improvement	NOUN
fcis-15262	5	33	in	in	ADP
fcis-15262	5	34	accuracy	accuracy	NOUN
fcis-15262	5	35	rate	rate	NOUN
fcis-15262	5	36	of	of	ADP
fcis-15262	5	37	2.7	2.7	NUM
fcis-15262	5	38	%	%	NOUN
fcis-15262	5	39	compared	compare	VERB
fcis-15262	5	40	to	to	ADP
fcis-15262	5	41	densenet169	densenet169	PROPN
fcis-15262	5	42	,	,	PUNCT
fcis-15262	5	43	and	and	CCONJ
fcis-15262	5	44	an	an	DET
fcis-15262	5	45	improvement	improvement	NOUN
fcis-15262	5	46	of	of	ADP
fcis-15262	5	47	1.1	1.1	NUM
fcis-15262	5	48	%	%	NOUN
fcis-15262	5	49	compared	compare	VERB
fcis-15262	5	50	to	to	ADP
fcis-15262	5	51	densenet	densenet	NOUN
fcis-15262	5	52	-	-	PUNCT
fcis-15262	5	53	gru	gru	NOUN
fcis-15262	5	54	.	.	PUNCT
fcis-15262	6	1	this	this	PRON
fcis-15262	6	2	validates	validate	VERB
fcis-15262	6	3	the	the	DET
fcis-15262	6	4	effectiveness	effectiveness	NOUN
fcis-15262	6	5	of	of	ADP
fcis-15262	6	6	the	the	DET
fcis-15262	6	7	proposed	propose	VERB
fcis-15262	6	8	method	method	NOUN
fcis-15262	6	9	in	in	ADP
fcis-15262	6	10	both	both	PRON
fcis-15262	6	11	improving	improve	VERB
fcis-15262	6	12	computational	computational	ADJ
fcis-15262	6	13	efficiency	efficiency	NOUN
fcis-15262	6	14	and	and	CCONJ
fcis-15262	6	15	achieving	achieve	VERB
fcis-15262	6	16	good	good	ADJ
fcis-15262	6	17	performance	performance	NOUN
fcis-15262	6	18	.	.	PUNCT
fcis-15262	7	1	keywords	keyword	NOUN
fcis-15262	7	2	:	:	PUNCT
fcis-15262	7	3	voiceprint	voiceprint	NOUN
fcis-15262	7	4	recognition	recognition	NOUN
fcis-15262	7	5	;	;	PUNCT
fcis-15262	7	6	depth	depth	NOUN
fcis-15262	7	7	separable	separable	ADJ
fcis-15262	7	8	convolutional	convolutional	ADJ
fcis-15262	7	9	;	;	PUNCT
fcis-15262	7	10	fbank	fbank	NOUN
fcis-15262	7	11	;	;	PUNCT
fcis-15262	7	12	densenet	densenet	NOUN
fcis-15262	7	13	;	;	PUNCT
fcis-15262	7	14	bigru	bigru	NOUN
fcis-15262	7	15	.	.	NOUN
fcis-15262	8	1	1	1	X
fcis-15262	8	2	.	.	X
fcis-15262	8	3	introduction	introduction	NOUN
fcis-15262	8	4	voiceprint	voiceprint	NOUN
fcis-15262	8	5	recognition	recognition	NOUN
fcis-15262	8	6	,	,	PUNCT
fcis-15262	8	7	also	also	ADV
fcis-15262	8	8	known	know	VERB
fcis-15262	8	9	as	as	ADP
fcis-15262	8	10	speaker	speaker	NOUN
fcis-15262	8	11	identification	identification	NOUN
fcis-15262	8	12	,	,	PUNCT
fcis-15262	8	13	is	be	AUX
fcis-15262	8	14	able	able	ADJ
fcis-15262	8	15	to	to	PART
fcis-15262	8	16	extract	extract	VERB
fcis-15262	8	17	sound	sound	NOUN
fcis-15262	8	18	features	feature	NOUN
fcis-15262	8	19	from	from	ADP
fcis-15262	8	20	a	a	DET
fcis-15262	8	21	piece	piece	NOUN
fcis-15262	8	22	of	of	ADP
fcis-15262	8	23	speech	speech	NOUN
fcis-15262	8	24	and	and	CCONJ
fcis-15262	8	25	identify	identify	VERB
fcis-15262	8	26	the	the	DET
fcis-15262	8	27	speaker	speaker	NOUN
fcis-15262	8	28	's	's	PART
fcis-15262	8	29	identity[1	identity[1	PROPN
fcis-15262	8	30	]	]	PUNCT
fcis-15262	8	31	.	.	PUNCT
fcis-15262	9	1	speech	speech	NOUN
fcis-15262	9	2	is	be	AUX
fcis-15262	9	3	a	a	DET
fcis-15262	9	4	unique	unique	ADJ
fcis-15262	9	5	physiological	physiological	ADJ
fcis-15262	9	6	behavior	behavior	NOUN
fcis-15262	9	7	of	of	ADP
fcis-15262	9	8	humans	human	NOUN
fcis-15262	9	9	,	,	PUNCT
fcis-15262	9	10	and	and	CCONJ
fcis-15262	9	11	for	for	ADP
fcis-15262	9	12	each	each	DET
fcis-15262	9	13	individual	individual	NOUN
fcis-15262	9	14	,	,	PUNCT
fcis-15262	9	15	their	their	PRON
fcis-15262	9	16	speech	speech	NOUN
fcis-15262	9	17	has	have	VERB
fcis-15262	9	18	its	its	PRON
fcis-15262	9	19	own	own	ADJ
fcis-15262	9	20	biological	biological	ADJ
fcis-15262	9	21	characteristics	characteristic	NOUN
fcis-15262	9	22	,	,	PUNCT
fcis-15262	9	23	such	such	ADJ
fcis-15262	9	24	as	as	ADP
fcis-15262	9	25	physiological	physiological	ADJ
fcis-15262	9	26	structural	structural	ADJ
fcis-15262	9	27	differences	difference	NOUN
fcis-15262	9	28	caused	cause	VERB
fcis-15262	9	29	by	by	ADP
fcis-15262	9	30	genetic	genetic	ADJ
fcis-15262	9	31	variations	variation	NOUN
fcis-15262	9	32	,	,	PUNCT
fcis-15262	9	33	voice	voice	NOUN
fcis-15262	9	34	differences	difference	NOUN
fcis-15262	9	35	formed	form	VERB
fcis-15262	9	36	after	after	ADP
fcis-15262	9	37	birth	birth	NOUN
fcis-15262	9	38	,	,	PUNCT
fcis-15262	9	39	and	and	CCONJ
fcis-15262	9	40	gender	gender	NOUN
fcis-15262	9	41	differences[2	differences[2	PROPN
fcis-15262	9	42	]	]	PUNCT
fcis-15262	9	43	,	,	PUNCT
fcis-15262	9	44	among	among	ADP
fcis-15262	9	45	others	other	NOUN
fcis-15262	9	46	.	.	PUNCT
fcis-15262	10	1	similar	similar	ADJ
fcis-15262	10	2	to	to	ADP
fcis-15262	10	3	speech	speech	NOUN
fcis-15262	10	4	recognition	recognition	NOUN
fcis-15262	10	5	,	,	PUNCT
fcis-15262	10	6	both	both	PRON
fcis-15262	10	7	require	require	VERB
fcis-15262	10	8	processing	processing	NOUN
fcis-15262	10	9	of	of	ADP
fcis-15262	10	10	speech	speech	NOUN
fcis-15262	10	11	signals	signal	NOUN
fcis-15262	10	12	to	to	PART
fcis-15262	10	13	extract	extract	VERB
fcis-15262	10	14	their	their	PRON
fcis-15262	10	15	features	feature	NOUN
fcis-15262	10	16	.	.	PUNCT
fcis-15262	11	1	the	the	DET
fcis-15262	11	2	feature	feature	NOUN
fcis-15262	11	3	map	map	NOUN
fcis-15262	11	4	is	be	AUX
fcis-15262	11	5	used	use	VERB
fcis-15262	11	6	as	as	ADP
fcis-15262	11	7	input	input	NOUN
fcis-15262	11	8	to	to	PART
fcis-15262	11	9	train	train	VERB
fcis-15262	11	10	deep	deep	ADJ
fcis-15262	11	11	learning	learning	NOUN
fcis-15262	11	12	models	model	NOUN
fcis-15262	11	13	,	,	PUNCT
fcis-15262	11	14	enabling	enable	VERB
fcis-15262	11	15	the	the	DET
fcis-15262	11	16	models	model	NOUN
fcis-15262	11	17	to	to	PART
fcis-15262	11	18	learn	learn	VERB
fcis-15262	11	19	information	information	NOUN
fcis-15262	11	20	from	from	ADP
fcis-15262	11	21	the	the	DET
fcis-15262	11	22	feature	feature	NOUN
fcis-15262	11	23	map	map	NOUN
fcis-15262	11	24	to	to	PART
fcis-15262	11	25	be	be	AUX
fcis-15262	11	26	used	use	VERB
fcis-15262	11	27	for	for	ADP
fcis-15262	11	28	prediction	prediction	NOUN
fcis-15262	11	29	and	and	CCONJ
fcis-15262	11	30	classification[3,4	classification[3,4	NOUN
fcis-15262	11	31	]	]	PUNCT
fcis-15262	11	32	.	.	PUNCT
fcis-15262	12	1	these	these	DET
fcis-15262	12	2	features	feature	NOUN
fcis-15262	12	3	include	include	VERB
fcis-15262	12	4	the	the	DET
fcis-15262	12	5	vibrational	vibrational	ADJ
fcis-15262	12	6	cycle	cycle	NOUN
fcis-15262	12	7	of	of	ADP
fcis-15262	12	8	the	the	DET
fcis-15262	12	9	vocal	vocal	ADJ
fcis-15262	12	10	cords	cord	NOUN
fcis-15262	12	11	and	and	CCONJ
fcis-15262	12	12	the	the	DET
fcis-15262	12	13	frequency	frequency	NOUN
fcis-15262	12	14	range	range	NOUN
fcis-15262	12	15	of	of	ADP
fcis-15262	12	16	the	the	DET
fcis-15262	12	17	sound	sound	NOUN
fcis-15262	12	18	,	,	PUNCT
fcis-15262	12	19	among	among	ADP
fcis-15262	12	20	others	other	NOUN
fcis-15262	12	21	.	.	PUNCT
fcis-15262	13	1	due	due	ADP
fcis-15262	13	2	to	to	ADP
fcis-15262	13	3	the	the	DET
fcis-15262	13	4	high	high	ADJ
fcis-15262	13	5	security	security	NOUN
fcis-15262	13	6	and	and	CCONJ
fcis-15262	13	7	the	the	DET
fcis-15262	13	8	ability	ability	NOUN
fcis-15262	13	9	to	to	PART
fcis-15262	13	10	collect	collect	VERB
fcis-15262	13	11	speech	speech	NOUN
fcis-15262	13	12	features	feature	NOUN
fcis-15262	13	13	without	without	ADP
fcis-15262	13	14	user	user	NOUN
fcis-15262	13	15	contact	contact	NOUN
fcis-15262	13	16	,	,	PUNCT
fcis-15262	13	17	the	the	DET
fcis-15262	13	18	market	market	NOUN
fcis-15262	13	19	share	share	NOUN
fcis-15262	13	20	of	of	ADP
fcis-15262	13	21	voiceprint	voiceprint	NOUN
fcis-15262	13	22	recognition	recognition	NOUN
fcis-15262	13	23	has	have	AUX
fcis-15262	13	24	increased	increase	VERB
fcis-15262	13	25	and	and	CCONJ
fcis-15262	13	26	is	be	AUX
fcis-15262	13	27	being	be	AUX
fcis-15262	13	28	applied	apply	VERB
fcis-15262	13	29	to	to	ADP
fcis-15262	13	30	a	a	DET
fcis-15262	13	31	wider	wide	ADJ
fcis-15262	13	32	range	range	NOUN
fcis-15262	13	33	of	of	ADP
fcis-15262	13	34	fields	field	NOUN
fcis-15262	13	35	.	.	PUNCT
fcis-15262	14	1	speaker	speaker	NOUN
fcis-15262	14	2	recognition	recognition	NOUN
fcis-15262	14	3	can	can	AUX
fcis-15262	14	4	be	be	AUX
fcis-15262	14	5	divided	divide	VERB
fcis-15262	14	6	into	into	ADP
fcis-15262	14	7	two	two	NUM
fcis-15262	14	8	tasks	task	NOUN
fcis-15262	14	9	:	:	PUNCT
fcis-15262	14	10	speaker	speaker	NOUN
fcis-15262	14	11	verification	verification	NOUN
fcis-15262	14	12	and	and	CCONJ
fcis-15262	14	13	speaker	speaker	NOUN
fcis-15262	14	14	identification	identification	NOUN
fcis-15262	14	15	.	.	PUNCT
fcis-15262	15	1	speaker	speaker	NOUN
fcis-15262	15	2	verification	verification	NOUN
fcis-15262	15	3	is	be	AUX
fcis-15262	15	4	a	a	DET
fcis-15262	15	5	key	key	ADJ
fcis-15262	15	6	technology	technology	NOUN
fcis-15262	15	7	for	for	ADP
fcis-15262	15	8	intelligent	intelligent	ADJ
fcis-15262	15	9	interaction	interaction	NOUN
fcis-15262	15	10	and	and	CCONJ
fcis-15262	15	11	can	can	AUX
fcis-15262	15	12	be	be	AUX
fcis-15262	15	13	widely	widely	ADV
fcis-15262	15	14	applied	apply	VERB
fcis-15262	15	15	in	in	ADP
fcis-15262	15	16	fields	field	NOUN
fcis-15262	15	17	such	such	ADJ
fcis-15262	15	18	as	as	ADP
fcis-15262	15	19	financial	financial	ADJ
fcis-15262	15	20	payments	payment	NOUN
fcis-15262	15	21	,	,	PUNCT
fcis-15262	15	22	criminal	criminal	ADJ
fcis-15262	15	23	investigations	investigation	NOUN
fcis-15262	15	24	,	,	PUNCT
fcis-15262	15	25	and	and	CCONJ
fcis-15262	15	26	national	national	ADJ
fcis-15262	15	27	defense[5	defense[5	NUM
fcis-15262	15	28	]	]	PUNCT
fcis-15262	15	29	.	.	PUNCT
fcis-15262	16	1	speaker	speaker	NOUN
fcis-15262	16	2	identification	identification	NOUN
fcis-15262	16	3	is	be	AUX
fcis-15262	16	4	the	the	DET
fcis-15262	16	5	process	process	NOUN
fcis-15262	16	6	of	of	ADP
fcis-15262	16	7	determining	determine	VERB
fcis-15262	16	8	a	a	DET
fcis-15262	16	9	specific	specific	ADJ
fcis-15262	16	10	reference	reference	NOUN
fcis-15262	16	11	speaker	speaker	NOUN
fcis-15262	16	12	based	base	VERB
fcis-15262	16	13	on	on	ADP
fcis-15262	16	14	the	the	DET
fcis-15262	16	15	speech	speech	NOUN
fcis-15262	16	16	content	content	NOUN
fcis-15262	16	17	.	.	PUNCT
fcis-15262	17	1	voiceprint	voiceprint	NOUN
fcis-15262	17	2	recognition	recognition	NOUN
fcis-15262	17	3	with	with	ADP
fcis-15262	17	4	predefined	predefine	VERB
fcis-15262	17	5	speech	speech	NOUN
fcis-15262	17	6	content	content	NOUN
fcis-15262	17	7	is	be	AUX
fcis-15262	17	8	called	call	VERB
fcis-15262	17	9	text	text	NOUN
fcis-15262	17	10	-	-	PUNCT
fcis-15262	17	11	dependent	dependent	ADJ
fcis-15262	17	12	speaker	speaker	NOUN
fcis-15262	17	13	recognition	recognition	NOUN
fcis-15262	17	14	,	,	PUNCT
fcis-15262	17	15	while	while	SCONJ
fcis-15262	17	16	voiceprint	voiceprint	NOUN
fcis-15262	17	17	recognition	recognition	NOUN
fcis-15262	17	18	that	that	PRON
fcis-15262	17	19	does	do	AUX
fcis-15262	17	20	not	not	PART
fcis-15262	17	21	require	require	VERB
fcis-15262	17	22	predetermined	predetermined	ADJ
fcis-15262	17	23	speech	speech	NOUN
fcis-15262	17	24	content	content	NOUN
fcis-15262	17	25	is	be	AUX
fcis-15262	17	26	called	call	VERB
fcis-15262	17	27	text	text	NOUN
fcis-15262	17	28	-	-	PUNCT
fcis-15262	17	29	independent	independent	ADJ
fcis-15262	17	30	speaker	speaker	NOUN
fcis-15262	17	31	recognition	recognition	NOUN
fcis-15262	17	32	.	.	PUNCT
fcis-15262	18	1	the	the	DET
fcis-15262	18	2	focus	focus	NOUN
fcis-15262	18	3	of	of	ADP
fcis-15262	18	4	this	this	DET
fcis-15262	18	5	study	study	NOUN
fcis-15262	18	6	is	be	AUX
fcis-15262	18	7	on	on	ADP
fcis-15262	18	8	textindependent	textindependent	NOUN
fcis-15262	18	9	speaker	speaker	NOUN
fcis-15262	18	10	identification	identification	NOUN
fcis-15262	18	11	.	.	PUNCT
fcis-15262	19	1	in	in	ADP
fcis-15262	19	2	the	the	DET
fcis-15262	19	3	past	past	NOUN
fcis-15262	19	4	,	,	PUNCT
fcis-15262	19	5	the	the	DET
fcis-15262	19	6	main	main	ADJ
fcis-15262	19	7	technique	technique	NOUN
fcis-15262	19	8	used	use	VERB
fcis-15262	19	9	for	for	ADP
fcis-15262	19	10	voiceprint	voiceprint	NOUN
fcis-15262	19	11	recognition	recognition	NOUN
fcis-15262	19	12	was	be	AUX
fcis-15262	19	13	the	the	DET
fcis-15262	19	14	gaussian	gaussian	ADJ
fcis-15262	19	15	mixture	mixture	NOUN
fcis-15262	19	16	model	model	NOUN
fcis-15262	19	17	-	-	PUNCT
fcis-15262	19	18	universal	universal	ADJ
fcis-15262	19	19	background	background	NOUN
fcis-15262	19	20	model	model	NOUN
fcis-15262	19	21	(	(	PUNCT
fcis-15262	19	22	gmm	gmm	NOUN
fcis-15262	19	23	-	-	PUNCT
fcis-15262	19	24	ubm)[6,7	ubm)[6,7	NOUN
fcis-15262	19	25	]	]	PUNCT
fcis-15262	19	26	.	.	PUNCT
fcis-15262	20	1	in	in	ADP
fcis-15262	20	2	recent	recent	ADJ
fcis-15262	20	3	years	year	NOUN
fcis-15262	20	4	,	,	PUNCT
fcis-15262	20	5	with	with	ADP
fcis-15262	20	6	the	the	DET
fcis-15262	20	7	emergence	emergence	NOUN
fcis-15262	20	8	of	of	ADP
fcis-15262	20	9	spectrograms	spectrogram	NOUN
fcis-15262	20	10	,	,	PUNCT
fcis-15262	20	11	researchers	researcher	NOUN
fcis-15262	20	12	have	have	AUX
fcis-15262	20	13	proposed	propose	VERB
fcis-15262	20	14	combining	combine	VERB
fcis-15262	20	15	spectrograms	spectrogram	NOUN
fcis-15262	20	16	with	with	ADP
fcis-15262	20	17	convolutional	convolutional	ADJ
fcis-15262	20	18	neural	neural	ADJ
fcis-15262	20	19	networks	network	NOUN
fcis-15262	20	20	(	(	PUNCT
fcis-15262	20	21	cnn	cnn	PROPN
fcis-15262	20	22	)	)	PUNCT
fcis-15262	20	23	for	for	ADP
fcis-15262	20	24	speaker	speaker	NOUN
fcis-15262	20	25	identification[8–10	identification[8–10	NOUN
fcis-15262	20	26	]	]	PUNCT
fcis-15262	20	27	.	.	PUNCT
fcis-15262	21	1	spectrograms[11	spectrograms[11	PROPN
fcis-15262	21	2	]	]	PUNCT
fcis-15262	21	3	contain	contain	VERB
fcis-15262	21	4	rich	rich	ADJ
fcis-15262	21	5	information	information	NOUN
fcis-15262	21	6	about	about	ADP
fcis-15262	21	7	the	the	DET
fcis-15262	21	8	voice	voice	NOUN
fcis-15262	21	9	characteristics	characteristic	NOUN
fcis-15262	21	10	of	of	ADP
fcis-15262	21	11	speakers	speaker	NOUN
fcis-15262	21	12	.	.	PUNCT
fcis-15262	22	1	by	by	ADP
fcis-15262	22	2	extracting	extract	VERB
fcis-15262	22	3	features	feature	NOUN
fcis-15262	22	4	from	from	ADP
fcis-15262	22	5	spectrograms	spectrogram	NOUN
fcis-15262	22	6	using	use	VERB
fcis-15262	22	7	cnn	cnn	PROPN
fcis-15262	22	8	,	,	PUNCT
fcis-15262	22	9	the	the	DET
fcis-15262	22	10	accuracy	accuracy	NOUN
fcis-15262	22	11	of	of	ADP
fcis-15262	22	12	voiceprint	voiceprint	NOUN
fcis-15262	22	13	recognition	recognition	NOUN
fcis-15262	22	14	has	have	AUX
fcis-15262	22	15	been	be	AUX
fcis-15262	22	16	significantly	significantly	ADV
fcis-15262	22	17	improved	improve	VERB
fcis-15262	22	18	.	.	PUNCT
fcis-15262	23	1	in	in	ADP
fcis-15262	23	2	literature[12	literature[12	PROPN
fcis-15262	23	3	]	]	PUNCT
fcis-15262	23	4	,	,	PUNCT
fcis-15262	23	5	an	an	DET
fcis-15262	23	6	lfbank	lfbank	NOUN
fcis-15262	23	7	feature	feature	NOUN
fcis-15262	23	8	was	be	AUX
fcis-15262	23	9	designed	design	VERB
fcis-15262	23	10	specifically	specifically	ADV
fcis-15262	23	11	for	for	ADP
fcis-15262	23	12	female	female	ADJ
fcis-15262	23	13	voiceprint	voiceprint	NOUN
fcis-15262	23	14	recognition	recognition	NOUN
fcis-15262	23	15	,	,	PUNCT
fcis-15262	23	16	and	and	CCONJ
fcis-15262	23	17	the	the	DET
fcis-15262	23	18	results	result	NOUN
fcis-15262	23	19	show	show	VERB
fcis-15262	23	20	that	that	SCONJ
fcis-15262	23	21	this	this	DET
fcis-15262	23	22	feature	feature	NOUN
fcis-15262	23	23	has	have	VERB
fcis-15262	23	24	advantages	advantage	NOUN
fcis-15262	23	25	in	in	ADP
fcis-15262	23	26	recognizing	recognize	VERB
fcis-15262	23	27	female	female	ADJ
fcis-15262	23	28	voices	voice	NOUN
fcis-15262	23	29	.	.	PUNCT
fcis-15262	24	1	literature[13	literature[13	NOUN
fcis-15262	24	2	]	]	PUNCT
fcis-15262	24	3	proposed	propose	VERB
fcis-15262	24	4	a	a	DET
fcis-15262	24	5	cnn	cnn	PROPN
fcis-15262	24	6	-	-	PUNCT
fcis-15262	24	7	lstm	lstm	PROPN
fcis-15262	24	8	model	model	NOUN
fcis-15262	24	9	based	base	VERB
fcis-15262	24	10	on	on	ADP
fcis-15262	24	11	global	global	ADJ
fcis-15262	24	12	attention	attention	NOUN
fcis-15262	24	13	mechanism	mechanism	NOUN
fcis-15262	24	14	,	,	PUNCT
fcis-15262	24	15	which	which	PRON
fcis-15262	24	16	simultaneously	simultaneously	ADV
fcis-15262	24	17	considers	consider	VERB
fcis-15262	24	18	spatial	spatial	ADJ
fcis-15262	24	19	and	and	CCONJ
fcis-15262	24	20	temporal	temporal	ADJ
fcis-15262	24	21	features	feature	NOUN
fcis-15262	24	22	of	of	ADP
fcis-15262	24	23	speech	speech	NOUN
fcis-15262	24	24	,	,	PUNCT
fcis-15262	24	25	and	and	CCONJ
fcis-15262	24	26	uses	use	VERB
fcis-15262	24	27	different	different	ADJ
fcis-15262	24	28	weights	weight	NOUN
fcis-15262	24	29	in	in	ADP
fcis-15262	24	30	lstm	lstm	NOUN
fcis-15262	24	31	hidden	hide	VERB
fcis-15262	24	32	layers	layer	NOUN
fcis-15262	24	33	through	through	ADP
fcis-15262	24	34	attention	attention	NOUN
fcis-15262	24	35	mechanism	mechanism	NOUN
fcis-15262	24	36	,	,	PUNCT
fcis-15262	24	37	resulting	result	VERB
fcis-15262	24	38	in	in	ADP
fcis-15262	24	39	noticeable	noticeable	ADJ
fcis-15262	24	40	improvement	improvement	NOUN
fcis-15262	24	41	in	in	ADP
fcis-15262	24	42	accuracy	accuracy	NOUN
fcis-15262	24	43	compared	compare	VERB
fcis-15262	24	44	to	to	ADP
fcis-15262	24	45	cnn	cnn	PROPN
fcis-15262	24	46	and	and	CCONJ
fcis-15262	24	47	cnn	cnn	PROPN
fcis-15262	24	48	-	-	PUNCT
fcis-15262	24	49	lstm	lstm	ADJ
fcis-15262	24	50	models	model	NOUN
fcis-15262	24	51	.	.	PUNCT
fcis-15262	25	1	literature[14	literature[14	PROPN
fcis-15262	25	2	]	]	PUNCT
fcis-15262	25	3	introduced	introduce	VERB
fcis-15262	25	4	a	a	DET
fcis-15262	25	5	voiceprint	voiceprint	NOUN
fcis-15262	25	6	recognition	recognition	NOUN
fcis-15262	25	7	method	method	NOUN
fcis-15262	25	8	combining	combine	VERB
fcis-15262	25	9	knowledge	knowledge	NOUN
fcis-15262	25	10	distillation	distillation	NOUN
fcis-15262	25	11	and	and	CCONJ
fcis-15262	25	12	resnet	resnet	NOUN
fcis-15262	25	13	,	,	PUNCT
fcis-15262	25	14	which	which	PRON
fcis-15262	25	15	achieved	achieve	VERB
fcis-15262	25	16	significant	significant	ADJ
fcis-15262	25	17	error	error	NOUN
fcis-15262	25	18	rate	rate	NOUN
fcis-15262	25	19	reduction	reduction	NOUN
fcis-15262	25	20	.	.	PUNCT
fcis-15262	26	1	literature[15	literature[15	NOUN
fcis-15262	26	2	]	]	PUNCT
fcis-15262	26	3	presented	present	VERB
fcis-15262	26	4	an	an	DET
fcis-15262	26	5	end	end	NOUN
fcis-15262	26	6	-	-	PUNCT
fcis-15262	26	7	to	to	ADP
fcis-15262	26	8	-	-	PUNCT
fcis-15262	26	9	end	end	NOUN
fcis-15262	26	10	voiceprint	voiceprint	NOUN
fcis-15262	26	11	recognition	recognition	NOUN
fcis-15262	26	12	algorithm	algorithm	NOUN
fcis-15262	26	13	based	base	VERB
fcis-15262	26	14	on	on	ADP
fcis-15262	26	15	lstm	lstm	PROPN
fcis-15262	26	16	,	,	PUNCT
fcis-15262	26	17	which	which	PRON
fcis-15262	26	18	shortened	shorten	VERB
fcis-15262	26	19	the	the	DET
fcis-15262	26	20	model	model	NOUN
fcis-15262	26	21	training	training	NOUN
fcis-15262	26	22	time	time	NOUN
fcis-15262	26	23	and	and	CCONJ
fcis-15262	26	24	achieved	achieve	VERB
fcis-15262	26	25	good	good	ADJ
fcis-15262	26	26	results	result	NOUN
fcis-15262	26	27	.	.	PUNCT
fcis-15262	27	1	literature[16	literature[16	PROPN
fcis-15262	27	2	]	]	PUNCT
fcis-15262	27	3	proposed	propose	VERB
fcis-15262	27	4	an	an	DET
fcis-15262	27	5	end	end	NOUN
fcis-15262	27	6	-	-	PUNCT
fcis-15262	27	7	to	to	ADP
fcis-15262	27	8	-	-	PUNCT
fcis-15262	27	9	end	end	NOUN
fcis-15262	27	10	voiceprint	voiceprint	NOUN
fcis-15262	27	11	recognition	recognition	NOUN
fcis-15262	27	12	method	method	NOUN
fcis-15262	27	13	based	base	VERB
fcis-15262	27	14	on	on	ADP
fcis-15262	27	15	stacked	stack	VERB
fcis-15262	27	16	bidirectional	bidirectional	ADJ
fcis-15262	27	17	lstm	lstm	NOUN
fcis-15262	27	18	with	with	ADP
fcis-15262	27	19	non	non	ADJ
fcis-15262	27	20	-	-	ADJ
fcis-15262	27	21	linear	linear	ADJ
fcis-15262	27	22	layers	layer	NOUN
fcis-15262	27	23	,	,	PUNCT
fcis-15262	27	24	which	which	PRON
fcis-15262	27	25	extracts	extract	VERB
fcis-15262	27	26	deeper	deep	ADJ
fcis-15262	27	27	abstract	abstract	ADJ
fcis-15262	27	28	features	feature	NOUN
fcis-15262	27	29	of	of	ADP
fcis-15262	27	30	speech	speech	NOUN
fcis-15262	27	31	signals	signal	NOUN
fcis-15262	27	32	using	use	VERB
fcis-15262	27	33	stacked	stack	VERB
fcis-15262	27	34	multiple	multiple	ADJ
fcis-15262	27	35	layers	layer	NOUN
fcis-15262	27	36	of	of	ADP
fcis-15262	27	37	bidirectional	bidirectional	ADJ
fcis-15262	27	38	lstm	lstm	NOUN
fcis-15262	27	39	and	and	CCONJ
fcis-15262	27	40	non	non	ADJ
fcis-15262	27	41	-	-	ADJ
fcis-15262	27	42	linear	linear	ADJ
fcis-15262	27	43	layers	layer	NOUN
fcis-15262	27	44	,	,	PUNCT
fcis-15262	27	45	resulting	result	VERB
fcis-15262	27	46	in	in	ADP
fcis-15262	27	47	better	well	ADJ
fcis-15262	27	48	recognition	recognition	NOUN
fcis-15262	27	49	performance	performance	NOUN
fcis-15262	27	50	than	than	ADP
fcis-15262	27	51	unidirectional	unidirectional	ADJ
fcis-15262	27	52	gru	gru	NOUN
fcis-15262	27	53	and	and	CCONJ
fcis-15262	27	54	unidirectional	unidirectional	ADJ
fcis-15262	27	55	lstm	lstm	NOUN
fcis-15262	27	56	.	.	PUNCT
fcis-15262	28	1	2	2	X
fcis-15262	28	2	.	.	NOUN
fcis-15262	28	3	speech	speech	NOUN
fcis-15262	28	4	feature	feature	NOUN
fcis-15262	28	5	extraction	extraction	NOUN
fcis-15262	28	6	currently	currently	ADV
fcis-15262	28	7	,	,	PUNCT
fcis-15262	28	8	two	two	NUM
fcis-15262	28	9	commonly	commonly	ADV
fcis-15262	28	10	used	use	VERB
fcis-15262	28	11	features	feature	NOUN
fcis-15262	28	12	in	in	ADP
fcis-15262	28	13	voiceprint	voiceprint	NOUN
fcis-15262	28	14	recognition	recognition	NOUN
fcis-15262	28	15	are	be	AUX
fcis-15262	28	16	mfcc	mfcc	NOUN
fcis-15262	28	17	and	and	CCONJ
fcis-15262	28	18	fbank	fbank	NOUN
fcis-15262	28	19	.	.	PUNCT
fcis-15262	29	1	the	the	DET
fcis-15262	29	2	human	human	ADJ
fcis-15262	29	3	ear	ear	NOUN
fcis-15262	29	4	's	's	PART
fcis-15262	29	5	response	response	NOUN
fcis-15262	29	6	to	to	PART
fcis-15262	29	7	sound	sound	VERB
fcis-15262	29	8	spectra	spectra	NOUN
fcis-15262	29	9	is	be	AUX
fcis-15262	29	10	nonlinear	nonlinear	ADJ
fcis-15262	29	11	.	.	PUNCT
fcis-15262	30	1	the	the	DET
fcis-15262	30	2	fbank	fbank	NOUN
fcis-15262	30	3	algorithm	algorithm	NOUN
fcis-15262	30	4	processes	process	VERB
fcis-15262	30	5	audio	audio	NOUN
fcis-15262	30	6	in	in	ADP
fcis-15262	30	7	a	a	DET
fcis-15262	30	8	similar	similar	ADJ
fcis-15262	30	9	way	way	NOUN
fcis-15262	30	10	to	to	ADP
fcis-15262	30	11	the	the	DET
fcis-15262	30	12	human	human	ADJ
fcis-15262	30	13	ear	ear	NOUN
fcis-15262	30	14	,	,	PUNCT
fcis-15262	30	15	fitting	fit	VERB
fcis-15262	30	16	the	the	DET
fcis-15262	30	17	characteristics	characteristic	NOUN
fcis-15262	30	18	received	receive	VERB
fcis-15262	30	19	by	by	ADP
fcis-15262	30	20	the	the	DET
fcis-15262	30	21	ear	ear	NOUN
fcis-15262	30	22	.	.	PUNCT
fcis-15262	31	1	mfcc	mfcc	NOUN
fcis-15262	31	2	is	be	AUX
fcis-15262	31	3	actually	actually	ADV
fcis-15262	31	4	obtained	obtain	VERB
fcis-15262	31	5	from	from	ADP
fcis-15262	31	6	fbank	fbank	NOUN
fcis-15262	31	7	through	through	ADP
fcis-15262	31	8	discrete	discrete	ADJ
fcis-15262	31	9	cosine	cosine	NOUN
fcis-15262	31	10	transform	transform	NOUN
fcis-15262	31	11	(	(	PUNCT
fcis-15262	31	12	dct	dct	PROPN
fcis-15262	31	13	)	)	PUNCT
fcis-15262	31	14	,	,	PUNCT
fcis-15262	31	15	which	which	PRON
fcis-15262	31	16	is	be	AUX
fcis-15262	31	17	a	a	DET
fcis-15262	31	18	linear	linear	ADJ
fcis-15262	31	19	transformation	transformation	NOUN
fcis-15262	31	20	that	that	PRON
fcis-15262	31	21	leads	lead	VERB
fcis-15262	31	22	to	to	ADP
fcis-15262	31	23	the	the	DET
fcis-15262	31	24	loss	loss	NOUN
fcis-15262	31	25	of	of	ADP
fcis-15262	31	26	many	many	ADJ
fcis-15262	31	27	nonlinear	nonlinear	ADJ
fcis-15262	31	28	components	component	NOUN
fcis-15262	31	29	in	in	ADP
fcis-15262	31	30	speech	speech	NOUN
fcis-15262	31	31	signals	signal	NOUN
fcis-15262	31	32	.	.	PUNCT
fcis-15262	32	1	in	in	ADP
fcis-15262	32	2	the	the	DET
fcis-15262	32	3	past	past	NOUN
fcis-15262	32	4	,	,	PUNCT
fcis-15262	32	5	mfcc	mfcc	NOUN
fcis-15262	32	6	features	feature	VERB
fcis-15262	32	7	combined	combine	VERB
fcis-15262	32	8	with	with	ADP
fcis-15262	32	9	gmms	gmms	ADJ
fcis-15262	32	10	-	-	PUNCT
fcis-15262	32	11	hmms	hmms	ADJ
fcis-15262	32	12	methods	method	NOUN
fcis-15262	32	13	were	be	AUX
fcis-15262	32	14	frequently	frequently	ADV
fcis-15262	32	15	used	use	VERB
fcis-15262	32	16	in	in	ADP
fcis-15262	32	17	speech	speech	NOUN
fcis-15262	32	18	recognition	recognition	NOUN
fcis-15262	32	19	.	.	PUNCT
fcis-15262	33	1	with	with	ADP
fcis-15262	33	2	the	the	DET
fcis-15262	33	3	emergence	emergence	NOUN
fcis-15262	33	4	of	of	ADP
fcis-15262	33	5	deep	deep	ADJ
fcis-15262	33	6	learning	learning	NOUN
fcis-15262	33	7	algorithms	algorithm	NOUN
fcis-15262	33	8	,	,	PUNCT
fcis-15262	33	9	neural	neural	ADJ
fcis-15262	33	10	networks	network	NOUN
fcis-15262	33	11	are	be	AUX
fcis-15262	33	12	not	not	PART
fcis-15262	33	13	sensitive	sensitive	ADJ
fcis-15262	33	14	to	to	ADP
fcis-15262	33	15	highly	highly	ADV
fcis-15262	33	16	correlated	correlate	VERB
fcis-15262	33	17	information	information	NOUN
fcis-15262	33	18	.	.	PUNCT
fcis-15262	34	1	as	as	ADP
fcis-15262	34	2	a	a	DET
fcis-15262	34	3	result	result	NOUN
fcis-15262	34	4	,	,	PUNCT
fcis-15262	34	5	fbank	fbank	NOUN
fcis-15262	34	6	has	have	AUX
fcis-15262	34	7	gained	gain	VERB
fcis-15262	34	8	broader	broad	ADJ
fcis-15262	34	9	application	application	NOUN
fcis-15262	34	10	.	.	PUNCT
fcis-15262	35	1	in	in	ADP
fcis-15262	35	2	this	this	DET
fcis-15262	35	3	paper	paper	NOUN
fcis-15262	35	4	,	,	PUNCT
fcis-15262	35	5	fbank	fbank	NOUN
fcis-15262	35	6	is	be	AUX
fcis-15262	35	7	used	use	VERB
fcis-15262	35	8	as	as	ADP
fcis-15262	35	9	the	the	DET
fcis-15262	35	10	speech	speech	NOUN
fcis-15262	35	11	feature	feature	NOUN
fcis-15262	35	12	.	.	PUNCT
fcis-15262	36	1	the	the	DET
fcis-15262	36	2	process	process	NOUN
fcis-15262	36	3	of	of	ADP
fcis-15262	36	4	extracting	extract	VERB
fcis-15262	36	5	fbank	fbank	NOUN
fcis-15262	36	6	feature	feature	NOUN
fcis-15262	36	7	maps	map	NOUN
fcis-15262	36	8	is	be	AUX
fcis-15262	36	9	shown	show	VERB
fcis-15262	36	10	in	in	ADP
fcis-15262	36	11	the	the	DET
fcis-15262	36	12	following	follow	VERB
fcis-15262	36	13	figure	figure	NOUN
fcis-15262	36	14	.	.	PUNCT
fcis-15262	37	1	pre	pre	ADJ
fcis-15262	37	2	-	-	ADJ
fcis-15262	37	3	emphasis	emphasis	ADJ
fcis-15262	37	4	fast	fast	ADJ
fcis-15262	37	5	fourier	fourier	NOUN
fcis-15262	37	6	transform	transform	NOUN
fcis-15262	37	7	magnitude	magnitude	NOUN
fcis-15262	37	8	spectrum	spectrum	NOUN
fcis-15262	37	9	mel	mel	PROPN
fcis-15262	37	10	filters	filter	NOUN
fcis-15262	37	11	taking	take	VERB
fcis-15262	37	12	the	the	DET
fcis-15262	37	13	logarithm	logarithm	NOUN
fcis-15262	37	14	speech	speech	NOUN
fcis-15262	37	15	signal	signal	NOUN
fcis-15262	37	16	fbank	fbank	NOUN
fcis-15262	37	17	feature	feature	NOUN
fcis-15262	37	18	maps	map	NOUN
fcis-15262	37	19	frame	frame	VERB
fcis-15262	37	20	segmentation	segmentation	NOUN
fcis-15262	37	21	windowing	windowe	VERB
fcis-15262	37	22	figure	figure	NOUN
fcis-15262	37	23	1	1	NUM
fcis-15262	37	24	.	.	PUNCT
fcis-15262	37	25	speech	speech	NOUN
fcis-15262	37	26	feature	feature	NOUN
fcis-15262	37	27	extraction	extraction	NOUN
fcis-15262	37	28	process	process	NOUN
fcis-15262	37	29	81	81	NUM
fcis-15262	37	30	taking	take	VERB
fcis-15262	37	31	a	a	DET
fcis-15262	37	32	speech	speech	NOUN
fcis-15262	37	33	segment	segment	NOUN
fcis-15262	37	34	from	from	ADP
fcis-15262	37	35	the	the	DET
fcis-15262	37	36	dataset	dataset	NOUN
fcis-15262	37	37	used	use	VERB
fcis-15262	37	38	in	in	ADP
fcis-15262	37	39	this	this	DET
fcis-15262	37	40	paper	paper	NOUN
fcis-15262	37	41	as	as	ADP
fcis-15262	37	42	an	an	DET
fcis-15262	37	43	example	example	NOUN
fcis-15262	37	44	,	,	PUNCT
fcis-15262	37	45	the	the	DET
fcis-15262	37	46	speech	speech	NOUN
fcis-15262	37	47	content	content	NOUN
fcis-15262	37	48	is	be	AUX
fcis-15262	37	49	"	"	PUNCT
fcis-15262	37	50	green	green	ADJ
fcis-15262	37	51	is	be	AUX
fcis-15262	37	52	the	the	DET
fcis-15262	37	53	background	background	NOUN
fcis-15262	37	54	color	color	NOUN
fcis-15262	37	55	of	of	ADP
fcis-15262	37	56	the	the	DET
fcis-15262	37	57	great	great	ADJ
fcis-15262	37	58	spring	spring	NOUN
fcis-15262	37	59	scenery	scenery	NOUN
fcis-15262	37	60	"	"	PUNCT
fcis-15262	37	61	.	.	PUNCT
fcis-15262	38	1	the	the	DET
fcis-15262	38	2	duration	duration	NOUN
fcis-15262	38	3	of	of	ADP
fcis-15262	38	4	the	the	DET
fcis-15262	38	5	speech	speech	NOUN
fcis-15262	38	6	is	be	AUX
fcis-15262	38	7	3.5	3.5	NUM
fcis-15262	38	8	seconds	second	NOUN
fcis-15262	38	9	.	.	PUNCT
fcis-15262	39	1	first	first	ADV
fcis-15262	39	2	,	,	PUNCT
fcis-15262	39	3	the	the	DET
fcis-15262	39	4	speech	speech	NOUN
fcis-15262	39	5	signal	signal	NOUN
fcis-15262	39	6	is	be	AUX
fcis-15262	39	7	processed	process	VERB
fcis-15262	39	8	to	to	PART
fcis-15262	39	9	obtain	obtain	VERB
fcis-15262	39	10	the	the	DET
fcis-15262	39	11	temporal	temporal	ADJ
fcis-15262	39	12	information	information	NOUN
fcis-15262	39	13	,	,	PUNCT
fcis-15262	39	14	and	and	CCONJ
fcis-15262	39	15	the	the	DET
fcis-15262	39	16	corresponding	corresponding	ADJ
fcis-15262	39	17	temporal	temporal	ADJ
fcis-15262	39	18	graph	graph	NOUN
fcis-15262	39	19	is	be	AUX
fcis-15262	39	20	shown	show	VERB
fcis-15262	39	21	in	in	ADP
fcis-15262	39	22	figure	figure	NOUN
fcis-15262	39	23	2	2	NUM
fcis-15262	39	24	.	.	PUNCT
fcis-15262	40	1	the	the	DET
fcis-15262	40	2	horizontal	horizontal	ADJ
fcis-15262	40	3	axis	axis	NOUN
fcis-15262	40	4	represents	represent	VERB
fcis-15262	40	5	time	time	NOUN
fcis-15262	40	6	,	,	PUNCT
fcis-15262	40	7	and	and	CCONJ
fcis-15262	40	8	the	the	DET
fcis-15262	40	9	vertical	vertical	ADJ
fcis-15262	40	10	axis	axis	NOUN
fcis-15262	40	11	represents	represent	VERB
fcis-15262	40	12	amplitude	amplitude	NOUN
fcis-15262	40	13	.	.	PUNCT
fcis-15262	41	1	it	it	PRON
fcis-15262	41	2	can	can	AUX
fcis-15262	41	3	be	be	AUX
fcis-15262	41	4	observed	observe	VERB
fcis-15262	41	5	from	from	ADP
fcis-15262	41	6	the	the	DET
fcis-15262	41	7	graph	graph	NOUN
fcis-15262	41	8	that	that	PRON
fcis-15262	41	9	the	the	DET
fcis-15262	41	10	amplitude	amplitude	NOUN
fcis-15262	41	11	varies	vary	VERB
fcis-15262	41	12	over	over	ADP
fcis-15262	41	13	time	time	NOUN
fcis-15262	41	14	.	.	PUNCT
fcis-15262	42	1	the	the	DET
fcis-15262	42	2	temporal	temporal	ADJ
fcis-15262	42	3	signal	signal	NOUN
fcis-15262	42	4	is	be	AUX
fcis-15262	42	5	then	then	ADV
fcis-15262	42	6	subjected	subject	VERB
fcis-15262	42	7	to	to	ADP
fcis-15262	42	8	the	the	DET
fcis-15262	42	9	following	follow	VERB
fcis-15262	42	10	processing	processing	NOUN
fcis-15262	42	11	steps	step	NOUN
fcis-15262	42	12	as	as	ADP
fcis-15262	42	13	input	input	NOUN
fcis-15262	42	14	.	.	PUNCT
fcis-15262	43	1	figure	figure	NOUN
fcis-15262	43	2	2	2	NUM
fcis-15262	43	3	.	.	NOUN
fcis-15262	43	4	time	time	NOUN
fcis-15262	43	5	-	-	PUNCT
fcis-15262	43	6	domain	domain	NOUN
fcis-15262	43	7	graph	graph	NOUN
fcis-15262	43	8	(	(	PUNCT
fcis-15262	43	9	1	1	NUM
fcis-15262	43	10	)	)	PUNCT
fcis-15262	43	11	pre	pre	ADJ
fcis-15262	43	12	-	-	NOUN
fcis-15262	43	13	emphasis	emphasis	NOUN
fcis-15262	43	14	is	be	AUX
fcis-15262	43	15	a	a	DET
fcis-15262	43	16	process	process	NOUN
fcis-15262	43	17	where	where	SCONJ
fcis-15262	43	18	the	the	DET
fcis-15262	43	19	input	input	NOUN
fcis-15262	43	20	speech	speech	NOUN
fcis-15262	43	21	signal	signal	NOUN
fcis-15262	43	22	is	be	AUX
fcis-15262	43	23	passed	pass	VERB
fcis-15262	43	24	through	through	ADP
fcis-15262	43	25	a	a	DET
fcis-15262	43	26	high	high	ADJ
fcis-15262	43	27	-	-	PUNCT
fcis-15262	43	28	pass	pass	NOUN
fcis-15262	43	29	filter	filter	NOUN
fcis-15262	43	30	to	to	PART
fcis-15262	43	31	compensate	compensate	VERB
fcis-15262	43	32	for	for	ADP
fcis-15262	43	33	highfrequency	highfrequency	NOUN
fcis-15262	43	34	signals	signal	NOUN
fcis-15262	43	35	in	in	ADP
fcis-15262	43	36	the	the	DET
fcis-15262	43	37	speech	speech	NOUN
fcis-15262	43	38	signal	signal	NOUN
fcis-15262	43	39	and	and	CCONJ
fcis-15262	43	40	flatten	flatten	VERB
fcis-15262	43	41	the	the	DET
fcis-15262	43	42	spectrum	spectrum	NOUN
fcis-15262	43	43	.	.	PUNCT
fcis-15262	44	1	the	the	DET
fcis-15262	44	2	formula	formula	NOUN
fcis-15262	44	3	is	be	AUX
fcis-15262	44	4	as	as	SCONJ
fcis-15262	44	5	follows	follow	VERB
fcis-15262	44	6	:	:	PUNCT
fcis-15262	44	7	(	(	PUNCT
fcis-15262	44	8	)	)	PUNCT
fcis-15262	44	9	(	(	PUNCT
fcis-15262	44	10	)	)	PUNCT
fcis-15262	44	11	(	(	PUNCT
fcis-15262	44	12	1)g	1)g	PROPN
fcis-15262	44	13	t	t	PROPN
fcis-15262	45	1	f	f	PROPN
fcis-15262	45	2	t	t	PROPN
fcis-15262	45	3	f	f	PROPN
fcis-15262	45	4	t	t	PROPN
fcis-15262	45	5			PROPN
fcis-15262	45	6			NOUN
fcis-15262	45	7	(	(	PUNCT
fcis-15262	45	8	1	1	NUM
fcis-15262	45	9	)	)	PUNCT
fcis-15262	45	10	f(t	f(t	NOUN
fcis-15262	45	11	)	)	PUNCT
fcis-15262	45	12	represents	represent	VERB
fcis-15262	45	13	the	the	DET
fcis-15262	45	14	original	original	ADJ
fcis-15262	45	15	speech	speech	NOUN
fcis-15262	45	16	signal	signal	NOUN
fcis-15262	45	17	,	,	PUNCT
fcis-15262	45	18	g(t	g(t	PROPN
fcis-15262	45	19	)	)	PUNCT
fcis-15262	45	20	represents	represent	VERB
fcis-15262	45	21	the	the	DET
fcis-15262	45	22	pre	pre	ADJ
fcis-15262	45	23	-	-	ADJ
fcis-15262	45	24	emphasis	emphasis	ADJ
fcis-15262	45	25	result	result	NOUN
fcis-15262	45	26	,	,	PUNCT
fcis-15262	45	27	α	α	PRON
fcis-15262	45	28	represents	represent	VERB
fcis-15262	45	29	the	the	DET
fcis-15262	45	30	pre	pre	ADJ
fcis-15262	45	31	-	-	ADJ
fcis-15262	45	32	emphasis	emphasis	ADJ
fcis-15262	45	33	coefficient	coefficient	NOUN
fcis-15262	45	34	,	,	PUNCT
fcis-15262	45	35	with	with	ADP
fcis-15262	45	36	typical	typical	ADJ
fcis-15262	45	37	values	value	NOUN
fcis-15262	45	38	of	of	ADP
fcis-15262	45	39	0.95	0.95	NUM
fcis-15262	45	40	or	or	CCONJ
fcis-15262	45	41	0.97	0.97	NUM
fcis-15262	45	42	.	.	PUNCT
fcis-15262	46	1	in	in	ADP
fcis-15262	46	2	this	this	DET
fcis-15262	46	3	paper	paper	NOUN
fcis-15262	46	4	,	,	PUNCT
fcis-15262	46	5	the	the	DET
fcis-15262	46	6	value	value	NOUN
fcis-15262	46	7	of	of	ADP
fcis-15262	46	8	α	α	PROPN
fcis-15262	46	9	is	be	AUX
fcis-15262	46	10	taken	take	VERB
fcis-15262	46	11	as	as	ADP
fcis-15262	46	12	0.97	0.97	NUM
fcis-15262	46	13	.	.	PUNCT
fcis-15262	47	1	after	after	ADP
fcis-15262	47	2	pre	pre	ADJ
fcis-15262	47	3	-	-	ADJ
fcis-15262	47	4	emphasis	emphasis	ADJ
fcis-15262	47	5	processing	processing	NOUN
fcis-15262	47	6	,	,	PUNCT
fcis-15262	47	7	the	the	DET
fcis-15262	47	8	temporal	temporal	ADJ
fcis-15262	47	9	graph	graph	NOUN
fcis-15262	47	10	is	be	AUX
fcis-15262	47	11	shown	show	VERB
fcis-15262	47	12	in	in	ADP
fcis-15262	47	13	figure	figure	NOUN
fcis-15262	47	14	3	3	NUM
fcis-15262	47	15	,	,	PUNCT
fcis-15262	47	16	which	which	PRON
fcis-15262	47	17	is	be	AUX
fcis-15262	47	18	flatter	flatter	NOUN
fcis-15262	47	19	compared	compare	VERB
fcis-15262	47	20	to	to	PART
fcis-15262	47	21	figure	figure	VERB
fcis-15262	47	22	2	2	NUM
fcis-15262	47	23	.	.	PUNCT
fcis-15262	47	24	figure	figure	VERB
fcis-15262	47	25	3	3	NUM
fcis-15262	47	26	.	.	NOUN
fcis-15262	47	27	time	time	NOUN
fcis-15262	47	28	-	-	PUNCT
fcis-15262	47	29	domain	domain	NOUN
fcis-15262	47	30	graph	graph	NOUN
fcis-15262	47	31	after	after	ADP
fcis-15262	47	32	pre	pre	ADJ
fcis-15262	47	33	-	-	ADJ
fcis-15262	47	34	emphasis	emphasis	ADJ
fcis-15262	47	35	processing	processing	NOUN
fcis-15262	47	36	(	(	PUNCT
fcis-15262	47	37	2	2	NUM
fcis-15262	47	38	)	)	PUNCT
fcis-15262	47	39	frame	frame	NOUN
fcis-15262	47	40	segmentation	segmentation	NOUN
fcis-15262	47	41	is	be	AUX
fcis-15262	47	42	performed	perform	VERB
fcis-15262	47	43	to	to	PART
fcis-15262	47	44	facilitate	facilitate	VERB
fcis-15262	47	45	subsequent	subsequent	ADJ
fcis-15262	47	46	short	short	ADJ
fcis-15262	47	47	-	-	PUNCT
fcis-15262	47	48	time	time	NOUN
fcis-15262	47	49	fourier	fourier	NOUN
fcis-15262	47	50	transform	transform	NOUN
fcis-15262	47	51	.	.	PUNCT
fcis-15262	48	1	the	the	DET
fcis-15262	48	2	speech	speech	NOUN
fcis-15262	48	3	signal	signal	NOUN
fcis-15262	48	4	is	be	AUX
fcis-15262	48	5	divided	divide	VERB
fcis-15262	48	6	into	into	ADP
fcis-15262	48	7	small	small	ADJ
fcis-15262	48	8	segments	segment	NOUN
fcis-15262	48	9	of	of	ADP
fcis-15262	48	10	equal	equal	ADJ
fcis-15262	48	11	duration	duration	NOUN
fcis-15262	48	12	,	,	PUNCT
fcis-15262	48	13	with	with	SCONJ
fcis-15262	48	14	each	each	DET
fcis-15262	48	15	segment	segment	NOUN
fcis-15262	48	16	referred	refer	VERB
fcis-15262	48	17	to	to	ADP
fcis-15262	48	18	as	as	ADP
fcis-15262	48	19	a	a	DET
fcis-15262	48	20	frame	frame	NOUN
fcis-15262	48	21	.	.	PUNCT
fcis-15262	49	1	to	to	PART
fcis-15262	49	2	ensure	ensure	VERB
fcis-15262	49	3	smooth	smooth	ADJ
fcis-15262	49	4	transitions	transition	NOUN
fcis-15262	49	5	between	between	ADP
fcis-15262	49	6	adjacent	adjacent	ADJ
fcis-15262	49	7	frames	frame	NOUN
fcis-15262	49	8	,	,	PUNCT
fcis-15262	49	9	there	there	PRON
fcis-15262	49	10	is	be	VERB
fcis-15262	49	11	an	an	DET
fcis-15262	49	12	overlap	overlap	NOUN
fcis-15262	49	13	region	region	NOUN
fcis-15262	49	14	between	between	ADP
fcis-15262	49	15	them	they	PRON
fcis-15262	49	16	.	.	PUNCT
fcis-15262	50	1	in	in	ADP
fcis-15262	50	2	this	this	DET
fcis-15262	50	3	paper	paper	NOUN
fcis-15262	50	4	,	,	PUNCT
fcis-15262	50	5	the	the	DET
fcis-15262	50	6	duration	duration	NOUN
fcis-15262	50	7	of	of	ADP
fcis-15262	50	8	each	each	DET
fcis-15262	50	9	frame	frame	NOUN
fcis-15262	50	10	is	be	AUX
fcis-15262	50	11	25ms	25ms	ADV
fcis-15262	50	12	,	,	PUNCT
fcis-15262	50	13	and	and	CCONJ
fcis-15262	50	14	the	the	DET
fcis-15262	50	15	overlap	overlap	NOUN
fcis-15262	50	16	region	region	NOUN
fcis-15262	50	17	is	be	AUX
fcis-15262	50	18	10ms	10ms	ADJ
fcis-15262	50	19	.	.	PUNCT
fcis-15262	51	1	the	the	DET
fcis-15262	51	2	waveform	waveform	NOUN
fcis-15262	51	3	of	of	ADP
fcis-15262	51	4	the	the	DET
fcis-15262	51	5	first	first	ADJ
fcis-15262	51	6	frame	frame	NOUN
fcis-15262	51	7	after	after	SCONJ
fcis-15262	51	8	frame	frame	NOUN
fcis-15262	51	9	segmentation	segmentation	NOUN
fcis-15262	51	10	is	be	AUX
fcis-15262	51	11	shown	show	VERB
fcis-15262	51	12	in	in	ADP
fcis-15262	51	13	figure	figure	NOUN
fcis-15262	51	14	4	4	NUM
fcis-15262	51	15	.	.	PUNCT
fcis-15262	51	16	figure	figure	VERB
fcis-15262	51	17	4	4	NUM
fcis-15262	51	18	.	.	PUNCT
fcis-15262	52	1	waveform	waveform	VERB
fcis-15262	52	2	graph	graph	NOUN
fcis-15262	52	3	of	of	ADP
fcis-15262	52	4	the	the	DET
fcis-15262	52	5	first	first	ADJ
fcis-15262	52	6	frame	frame	NOUN
fcis-15262	52	7	(	(	PUNCT
fcis-15262	52	8	3	3	X
fcis-15262	52	9	)	)	PUNCT
fcis-15262	52	10	windowing	windowing	NOUN
fcis-15262	52	11	is	be	AUX
fcis-15262	52	12	performed	perform	VERB
fcis-15262	52	13	in	in	ADP
fcis-15262	52	14	this	this	DET
fcis-15262	52	15	paper	paper	NOUN
fcis-15262	52	16	using	use	VERB
fcis-15262	52	17	a	a	DET
fcis-15262	52	18	hamming	hamming	NOUN
fcis-15262	52	19	window	window	NOUN
fcis-15262	52	20	.	.	PUNCT
fcis-15262	53	1	each	each	DET
fcis-15262	53	2	frame	frame	NOUN
fcis-15262	53	3	of	of	ADP
fcis-15262	53	4	the	the	DET
fcis-15262	53	5	speech	speech	NOUN
fcis-15262	53	6	signal	signal	NOUN
fcis-15262	53	7	is	be	AUX
fcis-15262	53	8	multiplied	multiply	VERB
fcis-15262	53	9	by	by	ADP
fcis-15262	53	10	the	the	DET
fcis-15262	53	11	window	window	NOUN
fcis-15262	53	12	function	function	NOUN
fcis-15262	53	13	to	to	PART
fcis-15262	53	14	enhance	enhance	VERB
fcis-15262	53	15	continuity	continuity	NOUN
fcis-15262	53	16	at	at	ADP
fcis-15262	53	17	the	the	DET
fcis-15262	53	18	edges	edge	NOUN
fcis-15262	53	19	of	of	ADP
fcis-15262	53	20	each	each	DET
fcis-15262	53	21	frame	frame	NOUN
fcis-15262	53	22	.	.	PUNCT
fcis-15262	54	1	the	the	DET
fcis-15262	54	2	formula	formula	NOUN
fcis-15262	54	3	is	be	AUX
fcis-15262	54	4	shown	show	VERB
fcis-15262	54	5	in	in	ADP
fcis-15262	54	6	equation	equation	NOUN
fcis-15262	54	7	(	(	PUNCT
fcis-15262	54	8	2	2	NUM
fcis-15262	54	9	)	)	PUNCT
fcis-15262	54	10	,	,	PUNCT
fcis-15262	54	11	where	where	SCONJ
fcis-15262	54	12	(	(	PUNCT
fcis-15262	54	13	)	)	PUNCT
fcis-15262	54	14	represents	represent	VERB
fcis-15262	54	15	the	the	DET
fcis-15262	54	16	hamming	hamming	NOUN
fcis-15262	54	17	window	window	NOUN
fcis-15262	54	18	and	and	CCONJ
fcis-15262	54	19	is	be	AUX
fcis-15262	54	20	set	set	VERB
fcis-15262	54	21	to	to	ADP
fcis-15262	54	22	0.46	0.46	NUM
fcis-15262	54	23	.	.	PUNCT
fcis-15262	55	1	figure	figure	NOUN
fcis-15262	55	2	5	5	NUM
fcis-15262	55	3	shows	show	VERB
fcis-15262	55	4	the	the	DET
fcis-15262	55	5	result	result	NOUN
fcis-15262	55	6	of	of	ADP
fcis-15262	55	7	the	the	DET
fcis-15262	55	8	first	first	ADJ
fcis-15262	55	9	frame	frame	NOUN
fcis-15262	55	10	of	of	ADP
fcis-15262	55	11	the	the	DET
fcis-15262	55	12	speech	speech	NOUN
fcis-15262	55	13	signal	signal	NOUN
fcis-15262	55	14	after	after	ADP
fcis-15262	55	15	windowing	windowe	VERB
fcis-15262	55	16	.	.	PUNCT
fcis-15262	56	1	it	it	PRON
fcis-15262	56	2	can	can	AUX
fcis-15262	56	3	be	be	AUX
fcis-15262	56	4	observed	observe	VERB
fcis-15262	56	5	that	that	SCONJ
fcis-15262	56	6	the	the	DET
fcis-15262	56	7	peaks	peak	NOUN
fcis-15262	56	8	in	in	ADP
fcis-15262	56	9	the	the	DET
fcis-15262	56	10	spectrum	spectrum	NOUN
fcis-15262	56	11	become	become	VERB
fcis-15262	56	12	narrower	narrow	ADJ
fcis-15262	56	13	,	,	PUNCT
fcis-15262	56	14	which	which	PRON
fcis-15262	56	15	helps	help	VERB
fcis-15262	56	16	reduce	reduce	VERB
fcis-15262	56	17	spectral	spectral	ADJ
fcis-15262	56	18	leakage	leakage	NOUN
fcis-15262	56	19	.	.	PUNCT
fcis-15262	57	1	however	however	ADV
fcis-15262	57	2	,	,	PUNCT
fcis-15262	57	3	the	the	DET
fcis-15262	57	4	amplitude	amplitude	NOUN
fcis-15262	57	5	of	of	ADP
fcis-15262	57	6	the	the	DET
fcis-15262	57	7	signal	signal	NOUN
fcis-15262	57	8	at	at	ADP
fcis-15262	57	9	the	the	DET
fcis-15262	57	10	edges	edge	NOUN
fcis-15262	57	11	of	of	ADP
fcis-15262	57	12	each	each	DET
fcis-15262	57	13	frame	frame	NOUN
fcis-15262	57	14	is	be	AUX
fcis-15262	57	15	attenuated	attenuate	VERB
fcis-15262	57	16	.	.	PUNCT
fcis-15262	58	1	to	to	PART
fcis-15262	58	2	compensate	compensate	VERB
fcis-15262	58	3	for	for	ADP
fcis-15262	58	4	this	this	PRON
fcis-15262	58	5	,	,	PUNCT
fcis-15262	58	6	overlapping	overlap	VERB
fcis-15262	58	7	regions	region	NOUN
fcis-15262	58	8	are	be	AUX
fcis-15262	58	9	introduced	introduce	VERB
fcis-15262	58	10	between	between	ADP
fcis-15262	58	11	adjacent	adjacent	ADJ
fcis-15262	58	12	frames	frame	NOUN
fcis-15262	58	13	during	during	ADP
fcis-15262	58	14	the	the	DET
fcis-15262	58	15	frame	frame	NOUN
fcis-15262	58	16	segmentation	segmentation	NOUN
fcis-15262	58	17	process	process	NOUN
fcis-15262	58	18	.	.	PUNCT
fcis-15262	59	1	2	2	NUM
fcis-15262	59	2	w	w	NOUN
fcis-15262	59	3	(	(	PUNCT
fcis-15262	59	4	,	,	PUNCT
fcis-15262	59	5	)	)	PUNCT
fcis-15262	59	6	(	(	PUNCT
fcis-15262	59	7	1	1	X
fcis-15262	59	8	)	)	PUNCT
fcis-15262	59	9	cos	cos	PROPN
fcis-15262	59	10	(	(	PUNCT
fcis-15262	59	11	)	)	PUNCT
fcis-15262	59	12	1	1	NUM
fcis-15262	59	13	n	n	ADP
fcis-15262	59	14	n	n	CCONJ
fcis-15262	59	15	n	n	CCONJ
fcis-15262	59	16			PROPN
fcis-15262	59	17			X
fcis-15262	59	18			X
fcis-15262	59	19			PROPN
fcis-15262	59	20			PROPN
fcis-15262	59	21			ADP
fcis-15262	59	22			PROPN
fcis-15262	59	23	(	(	PUNCT
fcis-15262	59	24	2	2	NUM
fcis-15262	59	25	)	)	PUNCT
fcis-15262	59	26	figure	figure	NOUN
fcis-15262	59	27	5	5	NUM
fcis-15262	59	28	.	.	PUNCT
fcis-15262	59	29	waveform	waveform	VERB
fcis-15262	59	30	graph	graph	NOUN
fcis-15262	59	31	of	of	ADP
fcis-15262	59	32	the	the	DET
fcis-15262	59	33	first	first	ADJ
fcis-15262	59	34	frame	frame	NOUN
fcis-15262	59	35	after	after	ADP
fcis-15262	59	36	windowing	windowe	VERB
fcis-15262	59	37	processing	processing	NOUN
fcis-15262	59	38	(	(	PUNCT
fcis-15262	59	39	4)fourier	4)fourier	NUM
fcis-15262	59	40	transform	transform	NOUN
fcis-15262	59	41	is	be	AUX
fcis-15262	59	42	used	use	VERB
fcis-15262	59	43	in	in	ADP
fcis-15262	59	44	this	this	DET
fcis-15262	59	45	case	case	NOUN
fcis-15262	59	46	to	to	PART
fcis-15262	59	47	convert	convert	VERB
fcis-15262	59	48	the	the	DET
fcis-15262	59	49	signal	signal	NOUN
fcis-15262	59	50	in	in	ADP
fcis-15262	59	51	the	the	DET
fcis-15262	59	52	time	time	NOUN
fcis-15262	59	53	domain	domain	NOUN
fcis-15262	59	54	into	into	ADP
fcis-15262	59	55	the	the	DET
fcis-15262	59	56	energy	energy	NOUN
fcis-15262	59	57	distribution	distribution	NOUN
fcis-15262	59	58	in	in	ADP
fcis-15262	59	59	the	the	DET
fcis-15262	59	60	frequency	frequency	NOUN
fcis-15262	59	61	domain	domain	NOUN
fcis-15262	59	62	.	.	PUNCT
fcis-15262	60	1	since	since	SCONJ
fcis-15262	60	2	the	the	DET
fcis-15262	60	3	speech	speech	NOUN
fcis-15262	60	4	waveform	waveform	VERB
fcis-15262	60	5	exhibits	exhibit	NOUN
fcis-15262	60	6	periodicity	periodicity	NOUN
fcis-15262	60	7	only	only	ADV
fcis-15262	60	8	in	in	ADP
fcis-15262	60	9	the	the	DET
fcis-15262	60	10	short	short	ADJ
fcis-15262	60	11	-	-	PUNCT
fcis-15262	60	12	time	time	NOUN
fcis-15262	60	13	domain	domain	NOUN
fcis-15262	60	14	,	,	PUNCT
fcis-15262	60	15	short	short	ADJ
fcis-15262	60	16	-	-	PUNCT
fcis-15262	60	17	time	time	NOUN
fcis-15262	60	18	fourier	fourier	NOUN
fcis-15262	60	19	transform	transform	NOUN
fcis-15262	60	20	(	(	PUNCT
fcis-15262	60	21	stft	stft	PROPN
fcis-15262	60	22	)	)	PUNCT
fcis-15262	60	23	is	be	AUX
fcis-15262	60	24	employed	employ	VERB
fcis-15262	60	25	.	.	PUNCT
fcis-15262	61	1	the	the	DET
fcis-15262	61	2	formula	formula	NOUN
fcis-15262	61	3	is	be	AUX
fcis-15262	61	4	as	as	SCONJ
fcis-15262	61	5	follows	follow	VERB
fcis-15262	61	6	:	:	PUNCT
fcis-15262	61	7	1	1	NUM
fcis-15262	61	8	(	(	PUNCT
fcis-15262	61	9	)	)	PUNCT
fcis-15262	61	10	(	(	PUNCT
fcis-15262	61	11	)	)	PUNCT
fcis-15262	61	12	exp	exp	NOUN
fcis-15262	61	13	{	{	PUNCT
fcis-15262	61	14	2	2	NUM
fcis-15262	61	15	}	}	PUNCT
fcis-15262	61	16	,	,	PUNCT
fcis-15262	62	1	1	1	NUM
fcis-15262	62	2	n	n	NOUN
fcis-15262	62	3	i	i	PRON
fcis-15262	62	4	i	i	PRON
fcis-15262	62	5	n	n	VERB
fcis-15262	62	6	kn	kn	PROPN
fcis-15262	62	7	s	s	PROPN
fcis-15262	62	8	k	k	PROPN
fcis-15262	62	9	s	s	X
fcis-15262	62	10	n	n	PRON
fcis-15262	62	11	j	j	PROPN
fcis-15262	62	12	k	k	PROPN
fcis-15262	62	13	k	k	PROPN
fcis-15262	63	1	n	n	PROPN
fcis-15262	63	2			PROPN
fcis-15262	64	1			PROPN
fcis-15262	64	2			PROPN
fcis-15262	64	3			PROPN
fcis-15262	64	4			NUM
fcis-15262	64	5			X
fcis-15262	64	6	(	(	PUNCT
fcis-15262	64	7	3	3	NUM
fcis-15262	64	8	)	)	PUNCT
fcis-15262	64	9	(	(	PUNCT
fcis-15262	64	10	5)the	5)the	DET
fcis-15262	64	11	magnitude	magnitude	NOUN
fcis-15262	64	12	spectrum	spectrum	NOUN
fcis-15262	64	13	is	be	AUX
fcis-15262	64	14	calculated	calculate	VERB
fcis-15262	64	15	using	use	VERB
fcis-15262	64	16	the	the	DET
fcis-15262	64	17	following	follow	VERB
fcis-15262	64	18	formula	formula	NOUN
fcis-15262	64	19	,	,	PUNCT
fcis-15262	64	20	where	where	SCONJ
fcis-15262	64	21	fft(x	fft(x	PROPN
fcis-15262	64	22	)	)	PUNCT
fcis-15262	64	23	represents	represent	VERB
fcis-15262	64	24	the	the	DET
fcis-15262	64	25	result	result	NOUN
fcis-15262	64	26	obtained	obtain	VERB
fcis-15262	64	27	from	from	ADP
fcis-15262	64	28	the	the	DET
fcis-15262	64	29	fourier	fourier	NOUN
fcis-15262	64	30	transform	transform	NOUN
fcis-15262	64	31	.	.	PUNCT
fcis-15262	64	32	2	2	NUM
fcis-15262	64	33	(	(	PUNCT
fcis-15262	64	34	)	)	PUNCT
fcis-15262	64	35	ifft	ifft	PROPN
fcis-15262	64	36	x	x	PUNCT
fcis-15262	64	37	p	p	NOUN
fcis-15262	64	38	n	n	ADV
fcis-15262	64	39			NUM
fcis-15262	64	40	(	(	PUNCT
fcis-15262	64	41	4	4	NUM
fcis-15262	64	42	)	)	PUNCT
fcis-15262	64	43	(	(	PUNCT
fcis-15262	64	44	6)mel	6)mel	NUM
fcis-15262	64	45	filter	filter	NOUN
fcis-15262	64	46	bank	bank	NOUN
fcis-15262	64	47	is	be	AUX
fcis-15262	64	48	used	use	VERB
fcis-15262	64	49	to	to	PART
fcis-15262	64	50	simulate	simulate	VERB
fcis-15262	64	51	the	the	DET
fcis-15262	64	52	characteristics	characteristic	NOUN
fcis-15262	64	53	of	of	ADP
fcis-15262	64	54	human	human	ADJ
fcis-15262	64	55	hearing	hearing	NOUN
fcis-15262	64	56	and	and	CCONJ
fcis-15262	64	57	extract	extract	VERB
fcis-15262	64	58	frequency	frequency	NOUN
fcis-15262	64	59	bands	band	NOUN
fcis-15262	64	60	.	.	PUNCT
fcis-15262	65	1	the	the	DET
fcis-15262	65	2	formula	formula	NOUN
fcis-15262	65	3	for	for	ADP
fcis-15262	65	4	converting	convert	VERB
fcis-15262	65	5	mel	mel	PROPN
fcis-15262	65	6	frequency	frequency	NOUN
fcis-15262	65	7	to	to	ADP
fcis-15262	65	8	speech	speech	NOUN
fcis-15262	65	9	frequency	frequency	NOUN
fcis-15262	65	10	is	be	AUX
fcis-15262	65	11	as	as	SCONJ
fcis-15262	65	12	follows	follow	VERB
fcis-15262	65	13	:	:	PUNCT
fcis-15262	65	14	102595*log	102595*log	NUM
fcis-15262	65	15	(	(	PUNCT
fcis-15262	65	16	1	1	NUM
fcis-15262	65	17	)	)	PUNCT
fcis-15262	65	18	700mel	700mel	NUM
fcis-15262	66	1	f	f	X
fcis-15262	66	2	f	f	PROPN
fcis-15262	66	3			PROPN
fcis-15262	66	4			PROPN
fcis-15262	66	5	(	(	PUNCT
fcis-15262	66	6	5	5	NUM
fcis-15262	66	7	)	)	PUNCT
fcis-15262	66	8	2595700(10	2595700(10	NUM
fcis-15262	66	9	1	1	NUM
fcis-15262	66	10	)	)	PUNCT
fcis-15262	66	11	melf	melf	NOUN
fcis-15262	66	12	f	f	PROPN
fcis-15262	67	1			PROPN
fcis-15262	67	2			PROPN
fcis-15262	67	3	(	(	PUNCT
fcis-15262	67	4	6	6	NUM
fcis-15262	67	5	)	)	PUNCT
fcis-15262	67	6	after	after	ADP
fcis-15262	67	7	the	the	DET
fcis-15262	67	8	above	above	ADJ
fcis-15262	67	9	steps	step	NOUN
fcis-15262	67	10	,	,	PUNCT
fcis-15262	67	11	the	the	DET
fcis-15262	67	12	fbank	fbank	NOUN
fcis-15262	67	13	feature	feature	NOUN
fcis-15262	67	14	map	map	NOUN
fcis-15262	67	15	can	can	AUX
fcis-15262	67	16	be	be	AUX
fcis-15262	67	17	obtained	obtain	VERB
fcis-15262	67	18	.	.	PUNCT
fcis-15262	68	1	the	the	DET
fcis-15262	68	2	following	follow	VERB
fcis-15262	68	3	image	image	NOUN
fcis-15262	68	4	shows	show	VERB
fcis-15262	68	5	the	the	DET
fcis-15262	68	6	fbank	fbank	NOUN
fcis-15262	68	7	feature	feature	NOUN
fcis-15262	68	8	maps	map	NOUN
fcis-15262	68	9	obtained	obtain	VERB
fcis-15262	68	10	when	when	SCONJ
fcis-15262	68	11	two	two	NUM
fcis-15262	68	12	people	people	NOUN
fcis-15262	68	13	pronounce	pronounce	VERB
fcis-15262	68	14	the	the	DET
fcis-15262	68	15	same	same	ADJ
fcis-15262	68	16	speech	speech	NOUN
fcis-15262	68	17	.	.	PUNCT
fcis-15262	69	1	the	the	DET
fcis-15262	69	2	feature	feature	NOUN
fcis-15262	69	3	map	map	NOUN
fcis-15262	69	4	contains	contain	VERB
fcis-15262	69	5	a	a	DET
fcis-15262	69	6	lot	lot	NOUN
fcis-15262	69	7	of	of	ADP
fcis-15262	69	8	important	important	ADJ
fcis-15262	69	9	information	information	NOUN
fcis-15262	69	10	.	.	PUNCT
fcis-15262	70	1	the	the	DET
fcis-15262	70	2	color	color	NOUN
fcis-15262	70	3	variation	variation	NOUN
fcis-15262	70	4	in	in	ADP
fcis-15262	70	5	the	the	DET
fcis-15262	70	6	image	image	NOUN
fcis-15262	70	7	represents	represent	VERB
fcis-15262	70	8	the	the	DET
fcis-15262	70	9	change	change	NOUN
fcis-15262	70	10	in	in	ADP
fcis-15262	70	11	speech	speech	NOUN
fcis-15262	70	12	energy	energy	NOUN
fcis-15262	70	13	,	,	PUNCT
fcis-15262	70	14	where	where	SCONJ
fcis-15262	70	15	darker	dark	ADJ
fcis-15262	70	16	colors	color	NOUN
fcis-15262	70	17	indicate	indicate	VERB
fcis-15262	70	18	stronger	strong	ADJ
fcis-15262	70	19	energy	energy	NOUN
fcis-15262	70	20	.	.	PUNCT
fcis-15262	71	1	the	the	DET
fcis-15262	71	2	vertical	vertical	ADJ
fcis-15262	71	3	axis	axis	NOUN
fcis-15262	71	4	corresponding	correspond	VERB
fcis-15262	71	5	to	to	ADP
fcis-15262	71	6	the	the	DET
fcis-15262	71	7	lowest	low	ADJ
fcis-15262	71	8	frequency	frequency	NOUN
fcis-15262	71	9	range	range	NOUN
fcis-15262	71	10	of	of	ADP
fcis-15262	71	11	the	the	DET
fcis-15262	71	12	horizontal	horizontal	ADJ
fcis-15262	71	13	stripes	stripe	NOUN
fcis-15262	71	14	represents	represent	VERB
fcis-15262	71	15	the	the	DET
fcis-15262	71	16	fundamental	fundamental	ADJ
fcis-15262	71	17	frequency	frequency	NOUN
fcis-15262	71	18	,	,	PUNCT
fcis-15262	71	19	which	which	PRON
fcis-15262	71	20	can	can	AUX
fcis-15262	71	21	be	be	AUX
fcis-15262	71	22	used	use	VERB
fcis-15262	71	23	to	to	PART
fcis-15262	71	24	determine	determine	VERB
fcis-15262	71	25	the	the	DET
fcis-15262	71	26	fundamental	fundamental	ADJ
fcis-15262	71	27	period	period	NOUN
fcis-15262	71	28	,	,	PUNCT
fcis-15262	71	29	i.e.	i.e.	X
fcis-15262	71	30	,	,	PUNCT
fcis-15262	71	31	the	the	DET
fcis-15262	71	32	vibration	vibration	ADJ
fcis-15262	71	33	period	period	NOUN
fcis-15262	71	34	of	of	ADP
fcis-15262	71	35	the	the	DET
fcis-15262	71	36	vocal	vocal	ADJ
fcis-15262	71	37	cords	cord	NOUN
fcis-15262	71	38	.	.	PUNCT
fcis-15262	72	1	additionally	additionally	ADV
fcis-15262	72	2	,	,	PUNCT
fcis-15262	72	3	the	the	DET
fcis-15262	72	4	darker	dark	ADJ
fcis-15262	72	5	stripes	stripe	NOUN
fcis-15262	72	6	at	at	ADP
fcis-15262	72	7	a	a	DET
fcis-15262	72	8	particular	particular	ADJ
fcis-15262	72	9	moment	moment	NOUN
fcis-15262	72	10	represent	represent	VERB
fcis-15262	72	11	the	the	DET
fcis-15262	72	12	formants	formant	NOUN
fcis-15262	72	13	,	,	PUNCT
fcis-15262	72	14	which	which	PRON
fcis-15262	72	15	refer	refer	VERB
fcis-15262	72	16	to	to	ADP
fcis-15262	72	17	the	the	DET
fcis-15262	72	18	resonant	resonant	ADJ
fcis-15262	72	19	frequencies	frequency	NOUN
fcis-15262	72	20	of	of	ADP
fcis-15262	72	21	the	the	DET
fcis-15262	72	22	vocal	vocal	ADJ
fcis-15262	72	23	tract	tract	NOUN
fcis-15262	72	24	and	and	CCONJ
fcis-15262	72	25	reflect	reflect	VERB
fcis-15262	72	26	the	the	DET
fcis-15262	72	27	physical	physical	ADJ
fcis-15262	72	28	characteristics	characteristic	NOUN
fcis-15262	72	29	of	of	ADP
fcis-15262	72	30	the	the	DET
fcis-15262	72	31	vocal	vocal	ADJ
fcis-15262	72	32	tract	tract	NOUN
fcis-15262	72	33	.	.	PUNCT
fcis-15262	73	1	82	82	NUM
fcis-15262	73	2	figure	figure	NOUN
fcis-15262	73	3	6	6	NUM
fcis-15262	73	4	.	.	PUNCT
fcis-15262	73	5	fbank	fbank	NOUN
fcis-15262	73	6	feature	feature	NOUN
fcis-15262	73	7	map	map	NOUN
fcis-15262	73	8	of	of	ADP
fcis-15262	73	9	speaker	speaker	NOUN
fcis-15262	73	10	1	1	NUM
fcis-15262	73	11	figure	figure	NOUN
fcis-15262	73	12	7	7	NUM
fcis-15262	73	13	.	.	PUNCT
fcis-15262	73	14	fbank	fbank	NOUN
fcis-15262	73	15	feature	feature	NOUN
fcis-15262	73	16	map	map	NOUN
fcis-15262	73	17	of	of	ADP
fcis-15262	73	18	speaker	speaker	NOUN
fcis-15262	73	19	2	2	NUM
fcis-15262	73	20	3	3	NUM
fcis-15262	73	21	.	.	PUNCT
fcis-15262	73	22	voiceprint	voiceprint	NOUN
fcis-15262	73	23	recognition	recognition	NOUN
fcis-15262	73	24	based	base	VERB
fcis-15262	73	25	on	on	ADP
fcis-15262	73	26	densenet169	densenet169	PROPN
fcis-15262	73	27	-	-	PUNCT
fcis-15262	73	28	bigru	bigru	NOUN
fcis-15262	73	29	3.1	3.1	NUM
fcis-15262	73	30	.	.	PUNCT
fcis-15262	73	31	densenet	densenet	NOUN
fcis-15262	73	32	network	network	NOUN
fcis-15262	73	33	densenet	densenet	NOUN
fcis-15262	73	34	,	,	PUNCT
fcis-15262	73	35	which	which	PRON
fcis-15262	73	36	belongs	belong	VERB
fcis-15262	73	37	to	to	ADP
fcis-15262	73	38	a	a	DET
fcis-15262	73	39	type	type	NOUN
fcis-15262	73	40	of	of	ADP
fcis-15262	73	41	convolutional	convolutional	ADJ
fcis-15262	73	42	neural	neural	ADJ
fcis-15262	73	43	networks	network	NOUN
fcis-15262	73	44	,	,	PUNCT
fcis-15262	73	45	can	can	AUX
fcis-15262	73	46	be	be	AUX
fcis-15262	73	47	considered	consider	VERB
fcis-15262	73	48	as	as	ADP
fcis-15262	73	49	an	an	DET
fcis-15262	73	50	extension	extension	NOUN
fcis-15262	73	51	of	of	ADP
fcis-15262	73	52	residual	residual	ADJ
fcis-15262	73	53	networks	network	NOUN
fcis-15262	73	54	.	.	PUNCT
fcis-15262	74	1	similar	similar	ADJ
fcis-15262	74	2	to	to	ADP
fcis-15262	74	3	residual	residual	ADJ
fcis-15262	74	4	networks	network	NOUN
fcis-15262	74	5	,	,	PUNCT
fcis-15262	74	6	the	the	DET
fcis-15262	74	7	fundamental	fundamental	ADJ
fcis-15262	74	8	building	building	NOUN
fcis-15262	74	9	block	block	NOUN
fcis-15262	74	10	of	of	ADP
fcis-15262	74	11	densenet	densenet	NOUN
fcis-15262	74	12	is	be	AUX
fcis-15262	74	13	a	a	DET
fcis-15262	74	14	residual	residual	ADJ
fcis-15262	74	15	block	block	NOUN
fcis-15262	74	16	.	.	PUNCT
fcis-15262	75	1	the	the	DET
fcis-15262	75	2	basic	basic	ADJ
fcis-15262	75	3	structure	structure	NOUN
fcis-15262	75	4	is	be	AUX
fcis-15262	75	5	shown	show	VERB
fcis-15262	75	6	in	in	ADP
fcis-15262	75	7	figure	figure	NOUN
fcis-15262	75	8	8	8	NUM
fcis-15262	75	9	,	,	PUNCT
fcis-15262	75	10	where	where	SCONJ
fcis-15262	75	11	it	it	PRON
fcis-15262	75	12	can	can	AUX
fcis-15262	75	13	be	be	AUX
fcis-15262	75	14	observed	observe	VERB
fcis-15262	75	15	that	that	SCONJ
fcis-15262	75	16	the	the	DET
fcis-15262	75	17	residual	residual	ADJ
fcis-15262	75	18	block	block	NOUN
fcis-15262	75	19	allows	allow	VERB
fcis-15262	75	20	for	for	ADP
fcis-15262	75	21	faster	fast	ADJ
fcis-15262	75	22	forward	forward	ADJ
fcis-15262	75	23	propagation	propagation	NOUN
fcis-15262	75	24	of	of	ADP
fcis-15262	75	25	the	the	DET
fcis-15262	75	26	input	input	NOUN
fcis-15262	75	27	.	.	PUNCT
fcis-15262	76	1	figure	figure	NOUN
fcis-15262	76	2	8	8	NUM
fcis-15262	76	3	.	.	PUNCT
fcis-15262	76	4	residual	residual	ADJ
fcis-15262	76	5	block	block	NOUN
fcis-15262	76	6	the	the	DET
fcis-15262	76	7	difference	difference	NOUN
fcis-15262	76	8	between	between	ADP
fcis-15262	76	9	densenet	densenet	NOUN
fcis-15262	76	10	and	and	CCONJ
fcis-15262	76	11	resnet	resnet	NOUN
fcis-15262	76	12	lies	lie	NOUN
fcis-15262	76	13	in	in	ADP
fcis-15262	76	14	the	the	DET
fcis-15262	76	15	way	way	NOUN
fcis-15262	76	16	their	their	PRON
fcis-15262	76	17	outputs	output	NOUN
fcis-15262	76	18	are	be	AUX
fcis-15262	76	19	combined	combine	VERB
fcis-15262	76	20	.	.	PUNCT
fcis-15262	77	1	while	while	SCONJ
fcis-15262	77	2	resnet	resnet	NOUN
fcis-15262	77	3	uses	use	VERB
fcis-15262	77	4	summation	summation	NOUN
fcis-15262	77	5	,	,	PUNCT
fcis-15262	77	6	densenet	densenet	NOUN
fcis-15262	77	7	uses	use	VERB
fcis-15262	77	8	concatenation	concatenation	NOUN
fcis-15262	77	9	.	.	PUNCT
fcis-15262	78	1	all	all	DET
fcis-15262	78	2	layers	layer	NOUN
fcis-15262	78	3	are	be	AUX
fcis-15262	78	4	connected	connect	VERB
fcis-15262	78	5	together	together	ADV
fcis-15262	78	6	,	,	PUNCT
fcis-15262	78	7	meaning	mean	VERB
fcis-15262	78	8	that	that	SCONJ
fcis-15262	78	9	if	if	SCONJ
fcis-15262	78	10	a	a	DET
fcis-15262	78	11	densenet	densenet	NOUN
fcis-15262	78	12	has	have	VERB
fcis-15262	78	13	k	k	PROPN
fcis-15262	78	14	layers	layer	NOUN
fcis-15262	78	15	,	,	PUNCT
fcis-15262	78	16	there	there	PRON
fcis-15262	78	17	will	will	AUX
fcis-15262	78	18	be	be	AUX
fcis-15262	78	19	k(k+1)/2	k(k+1)/2	PROPN
fcis-15262	78	20	connections	connection	NOUN
fcis-15262	78	21	in	in	ADP
fcis-15262	78	22	the	the	DET
fcis-15262	78	23	network	network	NOUN
fcis-15262	78	24	.	.	PUNCT
fcis-15262	79	1	it	it	PRON
fcis-15262	79	2	is	be	AUX
fcis-15262	79	3	because	because	SCONJ
fcis-15262	79	4	all	all	DET
fcis-15262	79	5	layers	layer	NOUN
fcis-15262	79	6	are	be	AUX
fcis-15262	79	7	connected	connect	VERB
fcis-15262	79	8	that	that	SCONJ
fcis-15262	79	9	features	feature	NOUN
fcis-15262	79	10	are	be	AUX
fcis-15262	79	11	fully	fully	ADV
fcis-15262	79	12	utilized	utilize	VERB
fcis-15262	79	13	,	,	PUNCT
fcis-15262	79	14	the	the	DET
fcis-15262	79	15	number	number	NOUN
fcis-15262	79	16	of	of	ADP
fcis-15262	79	17	parameters	parameter	NOUN
fcis-15262	79	18	is	be	AUX
fcis-15262	79	19	reduced	reduce	VERB
fcis-15262	79	20	,	,	PUNCT
fcis-15262	79	21	and	and	CCONJ
fcis-15262	79	22	the	the	DET
fcis-15262	79	23	issue	issue	NOUN
fcis-15262	79	24	of	of	ADP
fcis-15262	79	25	gradient	gradient	ADJ
fcis-15262	79	26	vanishing	vanishing	NOUN
fcis-15262	79	27	is	be	AUX
fcis-15262	79	28	alleviated	alleviate	VERB
fcis-15262	79	29	.	.	PUNCT
fcis-15262	80	1	the	the	DET
fcis-15262	80	2	densenet	densenet	NOUN
fcis-15262	80	3	connectivity	connectivity	NOUN
fcis-15262	80	4	pattern	pattern	NOUN
fcis-15262	80	5	is	be	AUX
fcis-15262	80	6	shown	show	VERB
fcis-15262	80	7	in	in	ADP
fcis-15262	80	8	figure	figure	NOUN
fcis-15262	80	9	9	9	NUM
fcis-15262	80	10	.	.	PUNCT
fcis-15262	81	1	figure	figure	NOUN
fcis-15262	81	2	9	9	NUM
fcis-15262	81	3	.	.	PUNCT
fcis-15262	81	4	dense	dense	ADJ
fcis-15262	81	5	connection	connection	NOUN
fcis-15262	81	6	in	in	ADP
fcis-15262	81	7	this	this	DET
fcis-15262	81	8	study	study	NOUN
fcis-15262	81	9	,	,	PUNCT
fcis-15262	81	10	the	the	DET
fcis-15262	81	11	densenet169	densenet169	PROPN
fcis-15262	81	12	network	network	NOUN
fcis-15262	81	13	is	be	AUX
fcis-15262	81	14	adopted	adopt	VERB
fcis-15262	81	15	,	,	PUNCT
fcis-15262	81	16	and	and	CCONJ
fcis-15262	81	17	its	its	PRON
fcis-15262	81	18	simplified	simplified	ADJ
fcis-15262	81	19	structure	structure	NOUN
fcis-15262	81	20	diagram	diagram	NOUN
fcis-15262	81	21	is	be	AUX
fcis-15262	81	22	shown	show	VERB
fcis-15262	81	23	in	in	ADP
fcis-15262	81	24	figure	figure	NOUN
fcis-15262	81	25	10	10	NUM
fcis-15262	81	26	,	,	PUNCT
fcis-15262	81	27	where	where	SCONJ
fcis-15262	81	28	169	169	NUM
fcis-15262	81	29	represents	represent	VERB
fcis-15262	81	30	the	the	DET
fcis-15262	81	31	sum	sum	NOUN
fcis-15262	81	32	of	of	ADP
fcis-15262	81	33	convolutional	convolutional	ADJ
fcis-15262	81	34	layers	layer	NOUN
fcis-15262	81	35	and	and	CCONJ
fcis-15262	81	36	fully	fully	ADV
fcis-15262	81	37	connected	connected	ADJ
fcis-15262	81	38	layers	layer	NOUN
fcis-15262	81	39	.	.	PUNCT
fcis-15262	82	1	figure	figure	VERB
fcis-15262	82	2	10	10	NUM
fcis-15262	82	3	.	.	PUNCT
fcis-15262	83	1	schematic	schematic	ADJ
fcis-15262	83	2	diagram	diagram	NOUN
fcis-15262	83	3	of	of	ADP
fcis-15262	83	4	densenet169	densenet169	PROPN
fcis-15262	83	5	structure	structure	NOUN
fcis-15262	83	6	3.2	3.2	NUM
fcis-15262	83	7	.	.	PUNCT
fcis-15262	84	1	optimization	optimization	NOUN
fcis-15262	84	2	of	of	ADP
fcis-15262	84	3	densenet	densenet	NOUN
fcis-15262	84	4	network	network	NOUN
fcis-15262	84	5	structure	structure	NOUN
fcis-15262	84	6	as	as	ADP
fcis-15262	84	7	the	the	DET
fcis-15262	84	8	number	number	NOUN
fcis-15262	84	9	of	of	ADP
fcis-15262	84	10	network	network	NOUN
fcis-15262	84	11	layers	layer	NOUN
fcis-15262	84	12	increases	increase	VERB
fcis-15262	84	13	,	,	PUNCT
fcis-15262	84	14	the	the	DET
fcis-15262	84	15	number	number	NOUN
fcis-15262	84	16	of	of	ADP
fcis-15262	84	17	parameters	parameter	NOUN
fcis-15262	84	18	also	also	ADV
fcis-15262	84	19	increases	increase	VERB
fcis-15262	84	20	.	.	PUNCT
fcis-15262	85	1	to	to	PART
fcis-15262	85	2	address	address	VERB
fcis-15262	85	3	this	this	DET
fcis-15262	85	4	issue	issue	NOUN
fcis-15262	85	5	,	,	PUNCT
fcis-15262	85	6	this	this	DET
fcis-15262	85	7	study	study	NOUN
fcis-15262	85	8	optimizes	optimize	VERB
fcis-15262	85	9	the	the	DET
fcis-15262	85	10	densenet169	densenet169	PROPN
fcis-15262	85	11	network	network	NOUN
fcis-15262	85	12	using	use	VERB
fcis-15262	85	13	the	the	DET
fcis-15262	85	14	depth	depth	NOUN
fcis-15262	85	15	separable	separable	ADJ
fcis-15262	85	16	convolution	convolution	NOUN
fcis-15262	85	17	(	(	PUNCT
fcis-15262	85	18	dsc	dsc	NOUN
fcis-15262	85	19	)	)	PUNCT
fcis-15262	85	20	method	method	NOUN
fcis-15262	85	21	.	.	PUNCT
fcis-15262	86	1	dsc	dsc	NOUN
fcis-15262	86	2	is	be	AUX
fcis-15262	86	3	an	an	DET
fcis-15262	86	4	effective	effective	ADJ
fcis-15262	86	5	approach	approach	NOUN
fcis-15262	86	6	to	to	PART
fcis-15262	86	7	reduce	reduce	VERB
fcis-15262	86	8	the	the	DET
fcis-15262	86	9	number	number	NOUN
fcis-15262	86	10	of	of	ADP
fcis-15262	86	11	network	network	NOUN
fcis-15262	86	12	parameters[17	parameters[17	PROPN
fcis-15262	86	13	]	]	PUNCT
fcis-15262	86	14	.	.	PUNCT
fcis-15262	87	1	the	the	DET
fcis-15262	87	2	method	method	NOUN
fcis-15262	87	3	of	of	ADP
fcis-15262	87	4	depth	depth	NOUN
fcis-15262	87	5	separable	separable	ADJ
fcis-15262	87	6	convolution	convolution	NOUN
fcis-15262	87	7	involves	involve	VERB
fcis-15262	87	8	performing	perform	VERB
fcis-15262	87	9	a	a	DET
fcis-15262	87	10	depthwise	depthwise	NOUN
fcis-15262	87	11	convolution	convolution	NOUN
fcis-15262	87	12	on	on	ADP
fcis-15262	87	13	the	the	DET
fcis-15262	87	14	input	input	NOUN
fcis-15262	87	15	,	,	PUNCT
fcis-15262	87	16	where	where	SCONJ
fcis-15262	87	17	each	each	DET
fcis-15262	87	18	channel	channel	NOUN
fcis-15262	87	19	of	of	ADP
fcis-15262	87	20	the	the	DET
fcis-15262	87	21	input	input	NOUN
fcis-15262	87	22	is	be	AUX
fcis-15262	87	23	convolved	convolve	VERB
fcis-15262	87	24	with	with	ADP
fcis-15262	87	25	a	a	DET
fcis-15262	87	26	separate	separate	ADJ
fcis-15262	87	27	convolutional	convolutional	ADJ
fcis-15262	87	28	kernel	kernel	NOUN
fcis-15262	87	29	.	.	PUNCT
fcis-15262	88	1	however	however	ADV
fcis-15262	88	2	,	,	PUNCT
fcis-15262	88	3	due	due	ADP
fcis-15262	88	4	to	to	ADP
fcis-15262	88	5	the	the	DET
fcis-15262	88	6	individual	individual	ADJ
fcis-15262	88	7	convolutions	convolution	NOUN
fcis-15262	88	8	across	across	ADP
fcis-15262	88	9	channels	channel	NOUN
fcis-15262	88	10	,	,	PUNCT
fcis-15262	88	11	the	the	DET
fcis-15262	88	12	feature	feature	NOUN
fcis-15262	88	13	information	information	NOUN
fcis-15262	88	14	of	of	ADP
fcis-15262	88	15	different	different	ADJ
fcis-15262	88	16	channels	channel	NOUN
fcis-15262	88	17	at	at	ADP
fcis-15262	88	18	the	the	DET
fcis-15262	88	19	same	same	ADJ
fcis-15262	88	20	location	location	NOUN
fcis-15262	88	21	is	be	AUX
fcis-15262	88	22	not	not	PART
fcis-15262	88	23	fully	fully	ADV
fcis-15262	88	24	utilized	utilize	VERB
fcis-15262	88	25	.	.	PUNCT
fcis-15262	89	1	therefore	therefore	ADV
fcis-15262	89	2	,	,	PUNCT
fcis-15262	89	3	after	after	ADP
fcis-15262	89	4	the	the	DET
fcis-15262	89	5	depthwise	depthwise	NOUN
fcis-15262	89	6	convolution	convolution	NOUN
fcis-15262	89	7	,	,	PUNCT
fcis-15262	89	8	a	a	DET
fcis-15262	89	9	pointwise	pointwise	ADJ
fcis-15262	89	10	convolution	convolution	NOUN
fcis-15262	89	11	is	be	AUX
fcis-15262	89	12	applied	apply	VERB
fcis-15262	89	13	,	,	PUNCT
fcis-15262	89	14	where	where	SCONJ
fcis-15262	89	15	the	the	DET
fcis-15262	89	16	kernel	kernel	NOUN
fcis-15262	89	17	size	size	NOUN
fcis-15262	89	18	is	be	AUX
fcis-15262	89	19	1	1	NUM
fcis-15262	89	20	*	*	SYM
fcis-15262	89	21	1	1	NUM
fcis-15262	89	22	,	,	PUNCT
fcis-15262	89	23	aiming	aim	VERB
fcis-15262	89	24	to	to	PART
fcis-15262	89	25	reduce	reduce	VERB
fcis-15262	89	26	the	the	DET
fcis-15262	89	27	number	number	NOUN
fcis-15262	89	28	of	of	ADP
fcis-15262	89	29	channels	channel	NOUN
fcis-15262	89	30	and	and	CCONJ
fcis-15262	89	31	combine	combine	VERB
fcis-15262	89	32	the	the	DET
fcis-15262	89	33	results	result	NOUN
fcis-15262	89	34	of	of	ADP
fcis-15262	89	35	the	the	DET
fcis-15262	89	36	depthwise	depthwise	NOUN
fcis-15262	89	37	convolution	convolution	NOUN
fcis-15262	89	38	.	.	PUNCT
fcis-15262	90	1	the	the	DET
fcis-15262	90	2	improved	improved	ADJ
fcis-15262	90	3	densenet169	densenet169	PROPN
fcis-15262	90	4	network	network	NOUN
fcis-15262	90	5	,	,	PUNCT
fcis-15262	90	6	compared	compare	VERB
fcis-15262	90	7	to	to	ADP
fcis-15262	90	8	the	the	DET
fcis-15262	90	9	original	original	ADJ
fcis-15262	90	10	densenet169	densenet169	PROPN
fcis-15262	90	11	structure	structure	NOUN
fcis-15262	90	12	before	before	ADP
fcis-15262	90	13	optimization	optimization	NOUN
fcis-15262	90	14	,	,	PUNCT
fcis-15262	90	15	is	be	AUX
fcis-15262	90	16	shown	show	VERB
fcis-15262	90	17	in	in	ADP
fcis-15262	90	18	figure	figure	NOUN
fcis-15262	90	19	11	11	NUM
fcis-15262	90	20	.	.	PUNCT
fcis-15262	91	1	the	the	DET
fcis-15262	91	2	improvement	improvement	NOUN
fcis-15262	91	3	involves	involve	VERB
fcis-15262	91	4	replacing	replace	VERB
fcis-15262	91	5	each	each	DET
fcis-15262	91	6	3	3	NUM
fcis-15262	91	7	*	*	SYM
fcis-15262	91	8	3	3	NUM
fcis-15262	91	9	convolution	convolution	NOUN
fcis-15262	91	10	in	in	ADP
fcis-15262	91	11	each	each	DET
fcis-15262	91	12	denseblock	denseblock	NOUN
fcis-15262	91	13	with	with	ADP
fcis-15262	91	14	a	a	DET
fcis-15262	91	15	3	3	NUM
fcis-15262	91	16	*	*	SYM
fcis-15262	91	17	3	3	NUM
fcis-15262	91	18	depthwise	depthwise	NOUN
fcis-15262	91	19	convolution	convolution	NOUN
fcis-15262	91	20	and	and	CCONJ
fcis-15262	91	21	a	a	DET
fcis-15262	91	22	1	1	NUM
fcis-15262	91	23	*	*	SYM
fcis-15262	91	24	1	1	NUM
fcis-15262	91	25	pointwise	pointwise	NOUN
fcis-15262	91	26	convolution	convolution	NOUN
fcis-15262	91	27	.	.	PUNCT
fcis-15262	92	1	layers	layer	NOUN
fcis-15262	92	2	convolution	convolution	VERB
fcis-15262	92	3	pooling	pool	VERB
fcis-15262	92	4	denseblock	denseblock	NOUN
fcis-15262	92	5	（	（	PUNCT
fcis-15262	92	6	1	1	NUM
fcis-15262	92	7	）	）	NOUN
fcis-15262	92	8	tansition	tansition	NOUN
fcis-15262	92	9	layer	layer	NOUN
fcis-15262	92	10	（	（	PUNCT
fcis-15262	92	11	1	1	NUM
fcis-15262	92	12	）	）	PUNCT
fcis-15262	92	13	denseblock	denseblock	X
fcis-15262	92	14	（	（	PUNCT
fcis-15262	92	15	2	2	NUM
fcis-15262	92	16	）	）	NOUN
fcis-15262	92	17	tansition	tansition	NOUN
fcis-15262	92	18	layer	layer	NOUN
fcis-15262	92	19	（	（	NUM
fcis-15262	92	20	2	2	NUM
fcis-15262	92	21	）	）	NOUN
fcis-15262	92	22	denseblock	denseblock	X
fcis-15262	92	23	（	（	PUNCT
fcis-15262	92	24	3	3	NUM
fcis-15262	92	25	）	）	NOUN
fcis-15262	92	26	tansition	tansition	NOUN
fcis-15262	92	27	layer	layer	NOUN
fcis-15262	92	28	（	（	PUNCT
fcis-15262	92	29	3	3	NUM
fcis-15262	92	30	）	）	NOUN
fcis-15262	92	31	denseblock	denseblock	X
fcis-15262	92	32	（	（	PUNCT
fcis-15262	92	33	4	4	NUM
fcis-15262	92	34	）	）	NOUN
fcis-15262	92	35	classfication	classfication	NOUN
fcis-15262	92	36	layer	layer	NOUN
fcis-15262	92	37	output	output	NOUN
fcis-15262	92	38	size	size	NOUN
fcis-15262	92	39	112	112	NUM
fcis-15262	92	40	*	*	SYM
fcis-15262	92	41	112	112	NUM
fcis-15262	92	42	56	56	NUM
fcis-15262	92	43	*	*	NUM
fcis-15262	92	44	56	56	NUM
fcis-15262	92	45	56	56	NUM
fcis-15262	92	46	*	*	SYM
fcis-15262	92	47	56	56	NUM
fcis-15262	92	48	56	56	NUM
fcis-15262	92	49	*	*	SYM
fcis-15262	92	50	56	56	NUM
fcis-15262	92	51	28	28	NUM
fcis-15262	92	52	*	*	SYM
fcis-15262	92	53	28	28	NUM
fcis-15262	92	54	28	28	NUM
fcis-15262	92	55	*	*	SYM
fcis-15262	92	56	28	28	NUM
fcis-15262	92	57	28	28	NUM
fcis-15262	92	58	*	*	SYM
fcis-15262	92	59	28	28	NUM
fcis-15262	92	60	14	14	NUM
fcis-15262	92	61	*	*	SYM
fcis-15262	92	62	14	14	NUM
fcis-15262	92	63	14	14	NUM
fcis-15262	92	64	*	*	SYM
fcis-15262	92	65	14	14	NUM
fcis-15262	92	66	14	14	NUM
fcis-15262	92	67	*	*	SYM
fcis-15262	92	68	14	14	NUM
fcis-15262	92	69	7	7	NUM
fcis-15262	92	70	*	*	NUM
fcis-15262	92	71	7	7	NUM
fcis-15262	92	72	7	7	NUM
fcis-15262	92	73	*	*	SYM
fcis-15262	92	74	7	7	NUM
fcis-15262	92	75	1	1	NUM
fcis-15262	92	76	*	*	SYM
fcis-15262	92	77	1	1	NUM
fcis-15262	92	78	densenet169	densenet169	PROPN
fcis-15262	92	79	7	7	NUM
fcis-15262	92	80	*	*	SYM
fcis-15262	92	81	7conv，stride2	7conv，stride2	NUM
fcis-15262	92	82	3	3	NUM
fcis-15262	92	83	*	*	SYM
fcis-15262	92	84	3maxpool，stride2	3maxpool，stride2	NUM
fcis-15262	92	85	densenet169+dsc	densenet169+dsc	NOUN
fcis-15262	92	86	6	6	NUM
fcis-15262	92	87	conv33	conv33	ADV
fcis-15262	92	88	conv11	conv11	ADJ
fcis-15262	92	89			PROPN
fcis-15262	92	90			PROPN
fcis-15262	92	91			PRON
fcis-15262	92	92			NOUN
fcis-15262	92	93			VERB
fcis-15262	92	94			PROPN
fcis-15262	92	95			NOUN
fcis-15262	92	96			PROPN
fcis-15262	92	97	12	12	NUM
fcis-15262	92	98	conv33	conv33	ADV
fcis-15262	92	99	conv11	conv11	PROPN
fcis-15262	92	100			PROPN
fcis-15262	92	101			PROPN
fcis-15262	92	102			PRON
fcis-15262	93	1			NOUN
fcis-15262	93	2			VERB
fcis-15262	93	3			PROPN
fcis-15262	93	4			NOUN
fcis-15262	93	5			PROPN
fcis-15262	93	6	32	32	NUM
fcis-15262	93	7	conv33	conv33	ADV
fcis-15262	93	8	conv11	conv11	PROPN
fcis-15262	93	9			PROPN
fcis-15262	93	10			PROPN
fcis-15262	93	11			PRON
fcis-15262	93	12			NOUN
fcis-15262	93	13			VERB
fcis-15262	93	14			PROPN
fcis-15262	93	15			NOUN
fcis-15262	93	16			PROPN
fcis-15262	93	17	32	32	NUM
fcis-15262	93	18	conv33	conv33	ADV
fcis-15262	93	19	conv11	conv11	PROPN
fcis-15262	93	20			PROPN
fcis-15262	93	21			PROPN
fcis-15262	93	22			PRON
fcis-15262	93	23			NOUN
fcis-15262	93	24			VERB
fcis-15262	93	25			PROPN
fcis-15262	93	26			NOUN
fcis-15262	93	27			PROPN
fcis-15262	93	28	6	6	NUM
fcis-15262	93	29	11	11	NUM
fcis-15262	93	30	dwcon33	dwcon33	NOUN
fcis-15262	93	31	conv11	conv11	ADJ
fcis-15262	93	32			PROPN
fcis-15262	93	33			PROPN
fcis-15262	93	34			PROPN
fcis-15262	94	1			PROPN
fcis-15262	94	2			PROPN
fcis-15262	94	3			X
fcis-15262	94	4			NOUN
fcis-15262	94	5			NOUN
fcis-15262	94	6			NOUN
fcis-15262	94	7			NOUN
fcis-15262	94	8			PROPN
fcis-15262	94	9			NOUN
fcis-15262	94	10			PROPN
fcis-15262	94	11			PROPN
fcis-15262	94	12	pwconv	pwconv	PROPN
fcis-15262	94	13	v	v	ADP
fcis-15262	94	14	12	12	NUM
fcis-15262	94	15	11	11	NUM
fcis-15262	94	16	dwcon33	dwcon33	NOUN
fcis-15262	94	17	conv11	conv11	ADJ
fcis-15262	94	18			PROPN
fcis-15262	94	19			PROPN
fcis-15262	94	20			PROPN
fcis-15262	94	21			PROPN
fcis-15262	94	22			PROPN
fcis-15262	94	23			X
fcis-15262	94	24			NOUN
fcis-15262	94	25			NOUN
fcis-15262	94	26			NOUN
fcis-15262	94	27			NOUN
fcis-15262	94	28			PROPN
fcis-15262	94	29			NOUN
fcis-15262	94	30			PROPN
fcis-15262	94	31			PROPN
fcis-15262	94	32	pwconv	pwconv	PROPN
fcis-15262	94	33	v	v	ADP
fcis-15262	94	34	32	32	NUM
fcis-15262	94	35	11	11	NUM
fcis-15262	94	36	dwcon33	dwcon33	NOUN
fcis-15262	94	37	conv11	conv11	ADJ
fcis-15262	94	38			PROPN
fcis-15262	94	39			PROPN
fcis-15262	94	40			PROPN
fcis-15262	94	41			PROPN
fcis-15262	94	42			PROPN
fcis-15262	94	43			X
fcis-15262	94	44			NOUN
fcis-15262	94	45			NOUN
fcis-15262	94	46			NOUN
fcis-15262	94	47			NOUN
fcis-15262	94	48			PROPN
fcis-15262	94	49			NOUN
fcis-15262	94	50			PROPN
fcis-15262	94	51			PROPN
fcis-15262	94	52	pwconv	pwconv	PROPN
fcis-15262	94	53	v	v	ADP
fcis-15262	94	54	32	32	NUM
fcis-15262	94	55	11	11	NUM
fcis-15262	94	56	dwcon33	dwcon33	NOUN
fcis-15262	94	57	conv11	conv11	ADJ
fcis-15262	94	58			PROPN
fcis-15262	94	59			PROPN
fcis-15262	94	60			PROPN
fcis-15262	94	61			PROPN
fcis-15262	94	62			PROPN
fcis-15262	94	63			X
fcis-15262	94	64			NOUN
fcis-15262	94	65			NOUN
fcis-15262	94	66			NOUN
fcis-15262	94	67			NOUN
fcis-15262	94	68			PROPN
fcis-15262	94	69			NOUN
fcis-15262	94	70			PROPN
fcis-15262	94	71			PROPN
fcis-15262	94	72	pwconv	pwconv	PROPN
fcis-15262	94	73	v	v	ADP
fcis-15262	94	74	1	1	NUM
fcis-15262	94	75	*	*	SYM
fcis-15262	94	76	1conv	1conv	NUM
fcis-15262	94	77	2	2	NUM
fcis-15262	94	78	*	*	NUM
fcis-15262	94	79	2average，stride2	2average，stride2	NUM
fcis-15262	94	80	1	1	NUM
fcis-15262	94	81	*	*	SYM
fcis-15262	94	82	1conv	1conv	NUM
fcis-15262	94	83	2	2	NUM
fcis-15262	94	84	*	*	NUM
fcis-15262	94	85	2average，stride2	2average，stride2	NUM
fcis-15262	94	86	1	1	NUM
fcis-15262	94	87	*	*	SYM
fcis-15262	94	88	1conv	1conv	NUM
fcis-15262	94	89	2	2	NUM
fcis-15262	94	90	*	*	NUM
fcis-15262	94	91	2average，stride2	2average，stride2	NUM
fcis-15262	94	92	7	7	NUM
fcis-15262	94	93	*	*	SYM
fcis-15262	94	94	7	7	NUM
fcis-15262	94	95	globle	globle	ADJ
fcis-15262	94	96	average	average	ADJ
fcis-15262	94	97	pool	pool	NOUN
fcis-15262	94	98	1000d	1000d	NOUN
fcis-15262	94	99	fully	fully	ADV
fcis-15262	94	100	connected，softmax	connected，softmax	ADV
fcis-15262	94	101	figure	figure	NOUN
fcis-15262	94	102	11	11	NUM
fcis-15262	94	103	.	.	PUNCT
fcis-15262	95	1	network	network	NOUN
fcis-15262	95	2	structure	structure	NOUN
fcis-15262	95	3	comparison	comparison	NOUN
fcis-15262	95	4	graph	graph	NOUN
fcis-15262	95	5	assuming	assume	VERB
fcis-15262	95	6	the	the	DET
fcis-15262	95	7	input	input	NOUN
fcis-15262	95	8	size	size	NOUN
fcis-15262	95	9	is	be	AUX
fcis-15262	95	10	h	h	NOUN
fcis-15262	95	11	*	*	ADJ
fcis-15262	95	12	w*m	w*m	PROPN
fcis-15262	95	13	and	and	CCONJ
fcis-15262	95	14	the	the	DET
fcis-15262	95	15	number	number	NOUN
fcis-15262	95	16	of	of	ADP
fcis-15262	95	17	output	output	NOUN
fcis-15262	95	18	channels	channel	NOUN
fcis-15262	95	19	is	be	AUX
fcis-15262	95	20	n	n	CCONJ
fcis-15262	95	21	,	,	PUNCT
fcis-15262	95	22	with	with	ADP
fcis-15262	95	23	a	a	DET
fcis-15262	95	24	convolution	convolution	NOUN
fcis-15262	95	25	kernel	kernel	NOUN
fcis-15262	95	26	size	size	NOUN
fcis-15262	95	27	of	of	ADP
fcis-15262	95	28	k*k	k*k	PROPN
fcis-15262	95	29	and	and	CCONJ
fcis-15262	95	30	padding	padding	NOUN
fcis-15262	95	31	size	size	NOUN
fcis-15262	95	32	of	of	ADP
fcis-15262	95	33	p.	p.	NOUN
fcis-15262	95	34	for	for	ADP
fcis-15262	95	35	regular	regular	ADJ
fcis-15262	95	36	convolution	convolution	NOUN
fcis-15262	95	37	,	,	PUNCT
fcis-15262	95	38	the	the	DET
fcis-15262	95	39	formula	formula	NOUN
fcis-15262	95	40	for	for	ADP
fcis-15262	95	41	calculating	calculate	VERB
fcis-15262	95	42	the	the	DET
fcis-15262	95	43	number	number	NOUN
fcis-15262	95	44	of	of	ADP
fcis-15262	95	45	parameters	parameter	NOUN
fcis-15262	95	46	yc	yc	VERB
fcis-15262	95	47	is	be	AUX
fcis-15262	95	48	given	give	VERB
fcis-15262	95	49	by	by	ADP
fcis-15262	95	50	83	83	NUM
fcis-15262	95	51	equation	equation	NOUN
fcis-15262	95	52	7	7	NUM
fcis-15262	95	53	,	,	PUNCT
fcis-15262	95	54	and	and	CCONJ
fcis-15262	95	55	the	the	DET
fcis-15262	95	56	formula	formula	NOUN
fcis-15262	95	57	for	for	ADP
fcis-15262	95	58	calculating	calculate	VERB
fcis-15262	95	59	the	the	DET
fcis-15262	95	60	computational	computational	ADJ
fcis-15262	95	61	cost	cost	NOUN
fcis-15262	95	62	fc	fc	PROPN
fcis-15262	95	63	is	be	AUX
fcis-15262	95	64	given	give	VERB
fcis-15262	95	65	by	by	ADP
fcis-15262	95	66	equation	equation	NOUN
fcis-15262	95	67	8	8	NUM
fcis-15262	95	68	.	.	PUNCT
fcis-15262	96	1	m	m	VERB
fcis-15262	97	1	k	k	NOUN
fcis-15262	97	2	k	k	PROPN
fcis-15262	97	3	n	n	PROPN
fcis-15262	97	4			PROPN
fcis-15262	97	5			PROPN
fcis-15262	97	6	(	(	PUNCT
fcis-15262	97	7	7	7	NUM
fcis-15262	97	8	)	)	PUNCT
fcis-15262	97	9	(	(	PUNCT
fcis-15262	97	10	2	2	NUM
fcis-15262	97	11	1	1	NUM
fcis-15262	97	12	)	)	PUNCT
fcis-15262	97	13	(	(	PUNCT
fcis-15262	97	14	2	2	NUM
fcis-15262	97	15	1)h	1)h	NOUN
fcis-15262	97	16	k	k	NOUN
fcis-15262	98	1	p	p	X
fcis-15262	98	2	w	w	PROPN
fcis-15262	99	1	k	k	PROPN
fcis-15262	99	2	p	p	X
fcis-15262	99	3	m	m	VERB
fcis-15262	99	4	k	k	PROPN
fcis-15262	99	5	k	k	PROPN
fcis-15262	99	6	n	n	PROPN
fcis-15262	99	7			PROPN
fcis-15262	99	8			VERB
fcis-15262	99	9			PROPN
fcis-15262	99	10			PROPN
fcis-15262	99	11			PUNCT
fcis-15262	99	12			VERB
fcis-15262	99	13			PROPN
fcis-15262	99	14			PROPN
fcis-15262	99	15			PROPN
fcis-15262	99	16			PROPN
fcis-15262	99	17	(	(	PUNCT
fcis-15262	99	18	8)	8)	NUM
fcis-15262	99	19	the	the	DET
fcis-15262	99	20	parameter	parameter	NOUN
fcis-15262	99	21	quantity	quantity	NOUN
fcis-15262	99	22	ydsc	ydsc	ADV
fcis-15262	99	23	of	of	ADP
fcis-15262	99	24	depth	depth	NOUN
fcis-15262	99	25	separable	separable	ADJ
fcis-15262	99	26	convolution	convolution	NOUN
fcis-15262	99	27	consists	consist	VERB
fcis-15262	99	28	of	of	ADP
fcis-15262	99	29	the	the	DET
fcis-15262	99	30	parameter	parameter	NOUN
fcis-15262	99	31	quantities	quantity	NOUN
fcis-15262	99	32	of	of	ADP
fcis-15262	99	33	depthwise	depthwise	NOUN
fcis-15262	99	34	convolution	convolution	NOUN
fcis-15262	99	35	and	and	CCONJ
fcis-15262	99	36	pointwise	pointwise	NOUN
fcis-15262	99	37	convolution	convolution	NOUN
fcis-15262	99	38	,	,	PUNCT
fcis-15262	99	39	and	and	CCONJ
fcis-15262	99	40	the	the	DET
fcis-15262	99	41	formula	formula	NOUN
fcis-15262	99	42	is	be	AUX
fcis-15262	99	43	as	as	SCONJ
fcis-15262	99	44	follows	follow	VERB
fcis-15262	99	45	(	(	PUNCT
fcis-15262	99	46	equation	equation	NOUN
fcis-15262	99	47	9	9	NUM
fcis-15262	99	48	):	):	PUNCT
fcis-15262	99	49	1	1	NUM
fcis-15262	99	50	1	1	NUM
fcis-15262	99	51	m	m	NOUN
fcis-15262	100	1	k	k	NOUN
fcis-15262	101	1	k	k	PROPN
fcis-15262	102	1	n	n	INTJ
fcis-15262	102	2	m	m	PROPN
fcis-15262	102	3			PROPN
fcis-15262	102	4			PROPN
fcis-15262	102	5			PROPN
fcis-15262	102	6			PROPN
fcis-15262	102	7			PROPN
fcis-15262	102	8	(	(	PUNCT
fcis-15262	102	9	9	9	NUM
fcis-15262	102	10	)	)	PUNCT
fcis-15262	102	11	similarly	similarly	ADV
fcis-15262	102	12	,	,	PUNCT
fcis-15262	102	13	the	the	DET
fcis-15262	102	14	computational	computational	ADJ
fcis-15262	102	15	cost	cost	NOUN
fcis-15262	102	16	fdsc	fdsc	NOUN
fcis-15262	102	17	is	be	AUX
fcis-15262	102	18	calculated	calculate	VERB
fcis-15262	102	19	as	as	ADP
fcis-15262	102	20	equation	equation	NOUN
fcis-15262	102	21	10	10	NUM
fcis-15262	102	22	:	:	PUNCT
fcis-15262	102	23	(	(	PUNCT
fcis-15262	102	24	2	2	NUM
fcis-15262	102	25	1	1	NUM
fcis-15262	102	26	)	)	PUNCT
fcis-15262	102	27	(	(	PUNCT
fcis-15262	102	28	2	2	NUM
fcis-15262	102	29	1	1	NUM
fcis-15262	102	30	)	)	PUNCT
fcis-15262	102	31	1	1	NUM
fcis-15262	102	32	1h	1h	NUM
fcis-15262	102	33	k	k	X
fcis-15262	103	1	p	p	X
fcis-15262	103	2	w	w	PROPN
fcis-15262	104	1	k	k	PROPN
fcis-15262	104	2	p	p	X
fcis-15262	105	1	k	k	PROPN
fcis-15262	105	2	k	k	PROPN
fcis-15262	105	3	m	m	PROPN
fcis-15262	105	4	h	h	PROPN
fcis-15262	105	5	w	w	PROPN
fcis-15262	105	6	m	m	PROPN
fcis-15262	105	7	n	n	PROPN
fcis-15262	105	8			ADV
fcis-15262	105	9			VERB
fcis-15262	105	10			PROPN
fcis-15262	105	11			PROPN
fcis-15262	105	12			PUNCT
fcis-15262	105	13			VERB
fcis-15262	105	14			PROPN
fcis-15262	105	15			PROPN
fcis-15262	105	16			PROPN
fcis-15262	105	17			PROPN
fcis-15262	105	18			PROPN
fcis-15262	105	19			PROPN
fcis-15262	105	20			PROPN
fcis-15262	105	21			PROPN
fcis-15262	105	22			PROPN
fcis-15262	105	23	(	(	PUNCT
fcis-15262	105	24	10	10	NUM
fcis-15262	105	25	)	)	PUNCT
fcis-15262	105	26	in	in	ADP
fcis-15262	105	27	this	this	DET
fcis-15262	105	28	model	model	NOUN
fcis-15262	105	29	,	,	PUNCT
fcis-15262	105	30	k	k	PROPN
fcis-15262	105	31	is	be	AUX
fcis-15262	105	32	3	3	NUM
fcis-15262	105	33	and	and	CCONJ
fcis-15262	105	34	p	p	NOUN
fcis-15262	105	35	is	be	AUX
fcis-15262	105	36	1	1	NUM
fcis-15262	105	37	.	.	PUNCT
fcis-15262	105	38	substituting	substitute	VERB
fcis-15262	105	39	the	the	DET
fcis-15262	105	40	values	value	NOUN
fcis-15262	105	41	into	into	ADP
fcis-15262	105	42	the	the	DET
fcis-15262	105	43	formula	formula	NOUN
fcis-15262	105	44	,	,	PUNCT
fcis-15262	105	45	we	we	PRON
fcis-15262	105	46	can	can	AUX
fcis-15262	105	47	obtain	obtain	VERB
fcis-15262	105	48	the	the	DET
fcis-15262	105	49	ratio	ratio	NOUN
fcis-15262	105	50	of	of	ADP
fcis-15262	105	51	computational	computational	ADJ
fcis-15262	105	52	cost	cost	NOUN
fcis-15262	105	53	:	:	PUNCT
fcis-15262	105	54	1	1	NUM
fcis-15262	105	55	1	1	NUM
fcis-15262	105	56	9	9	NUM
fcis-15262	105	57	dsc	dsc	NOUN
fcis-15262	105	58	dsc	dsc	NOUN
fcis-15262	106	1	c	c	PROPN
fcis-15262	106	2	c	c	PROPN
fcis-15262	107	1	f	f	PROPN
fcis-15262	107	2	y	y	PROPN
fcis-15262	107	3	f	f	PROPN
fcis-15262	107	4	y	y	PROPN
fcis-15262	107	5	n	n	CCONJ
fcis-15262	107	6			PROPN
fcis-15262	107	7			PROPN
fcis-15262	107	8			PUNCT
fcis-15262	107	9	(	(	PUNCT
fcis-15262	107	10	11	11	NUM
fcis-15262	107	11	)	)	PUNCT
fcis-15262	107	12	from	from	ADP
fcis-15262	107	13	the	the	DET
fcis-15262	107	14	above	above	ADJ
fcis-15262	107	15	equation	equation	NOUN
fcis-15262	107	16	,	,	PUNCT
fcis-15262	107	17	it	it	PRON
fcis-15262	107	18	can	can	AUX
fcis-15262	107	19	be	be	AUX
fcis-15262	107	20	seen	see	VERB
fcis-15262	107	21	that	that	SCONJ
fcis-15262	107	22	using	use	VERB
fcis-15262	107	23	depth	depth	NOUN
fcis-15262	107	24	separable	separable	ADJ
fcis-15262	107	25	convolution	convolution	NOUN
fcis-15262	107	26	can	can	AUX
fcis-15262	107	27	significantly	significantly	ADV
fcis-15262	107	28	reduce	reduce	VERB
fcis-15262	107	29	the	the	DET
fcis-15262	107	30	computational	computational	ADJ
fcis-15262	107	31	cost	cost	NOUN
fcis-15262	107	32	.	.	PUNCT
fcis-15262	108	1	the	the	DET
fcis-15262	108	2	parameter	parameter	PROPN
fcis-15262	108	3	comparison	comparison	NOUN
fcis-15262	108	4	obtained	obtain	VERB
fcis-15262	108	5	from	from	ADP
fcis-15262	108	6	experimental	experimental	ADJ
fcis-15262	108	7	results	result	NOUN
fcis-15262	108	8	is	be	AUX
fcis-15262	108	9	shown	show	VERB
fcis-15262	108	10	in	in	ADP
fcis-15262	108	11	table	table	NOUN
fcis-15262	108	12	1	1	NUM
fcis-15262	108	13	.	.	PUNCT
fcis-15262	108	14	table	table	NOUN
fcis-15262	108	15	1	1	NUM
fcis-15262	108	16	.	.	PUNCT
fcis-15262	109	1	parameter	parameter	PROPN
fcis-15262	109	2	comparison	comparison	NOUN
fcis-15262	109	3	net	net	ADJ
fcis-15262	109	4	batch	batch	NOUN
fcis-15262	109	5	size	size	NOUN
fcis-15262	109	6	input	input	NOUN
fcis-15262	109	7	size	size	NOUN
fcis-15262	109	8	(	(	PUNCT
fcis-15262	109	9	mb	mb	NOUN
fcis-15262	109	10	)	)	PUNCT
fcis-15262	109	11	total	total	ADJ
fcis-15262	109	12	params	param	NOUN
fcis-15262	109	13	params	param	NOUN
fcis-15262	109	14	size	size	NOUN
fcis-15262	109	15	(	(	PUNCT
fcis-15262	109	16	mb	mb	NOUN
fcis-15262	109	17	)	)	PUNCT
fcis-15262	109	18	densenet169	densenet169	PROPN
fcis-15262	109	19	64	64	NUM
fcis-15262	109	20	39.35	39.35	NUM
fcis-15262	109	21	12617680	12617680	NUM
fcis-15262	109	22	50.47	50.47	NUM
fcis-15262	109	23	densenet169+dsc	densenet169+dsc	NOUN
fcis-15262	109	24	64	64	NUM
fcis-15262	109	25	29.35	29.35	NUM
fcis-15262	109	26	10368560	10368560	NUM
fcis-15262	109	27	41.23	41.23	NUM
fcis-15262	109	28	3.3	3.3	NUM
fcis-15262	109	29	.	.	PUNCT
fcis-15262	110	1	gru	gru	PROPN
fcis-15262	110	2	network	network	NOUN
fcis-15262	110	3	the	the	DET
fcis-15262	110	4	gated	gate	VERB
fcis-15262	110	5	recurrent	recurrent	ADJ
fcis-15262	110	6	unit	unit	NOUN
fcis-15262	110	7	(	(	PUNCT
fcis-15262	110	8	gru	gru	PROPN
fcis-15262	110	9	)	)	PUNCT
fcis-15262	110	10	belongs	belong	VERB
fcis-15262	110	11	to	to	ADP
fcis-15262	110	12	the	the	DET
fcis-15262	110	13	family	family	NOUN
fcis-15262	110	14	of	of	ADP
fcis-15262	110	15	recurrent	recurrent	ADJ
fcis-15262	110	16	neural	neural	ADJ
fcis-15262	110	17	networks	network	NOUN
fcis-15262	110	18	(	(	PUNCT
fcis-15262	110	19	rnn	rnn	PROPN
fcis-15262	110	20	)	)	PUNCT
fcis-15262	111	1	[	[	X
fcis-15262	111	2	18	18	NUM
fcis-15262	111	3	]	]	PUNCT
fcis-15262	111	4	.	.	PUNCT
fcis-15262	112	1	it	it	PRON
fcis-15262	112	2	is	be	AUX
fcis-15262	112	3	an	an	DET
fcis-15262	112	4	improved	improved	ADJ
fcis-15262	112	5	version	version	NOUN
fcis-15262	112	6	based	base	VERB
fcis-15262	112	7	on	on	ADP
fcis-15262	112	8	the	the	DET
fcis-15262	112	9	long	long	ADJ
fcis-15262	112	10	short	short	ADJ
fcis-15262	112	11	-	-	PUNCT
fcis-15262	112	12	term	term	NOUN
fcis-15262	112	13	memory	memory	NOUN
fcis-15262	112	14	(	(	PUNCT
fcis-15262	112	15	lstm	lstm	NOUN
fcis-15262	112	16	)	)	PUNCT
fcis-15262	112	17	network	network	NOUN
fcis-15262	112	18	.	.	PUNCT
fcis-15262	113	1	both	both	DET
fcis-15262	113	2	gru	gru	PROPN
fcis-15262	113	3	and	and	CCONJ
fcis-15262	113	4	lstm	lstm	NOUN
fcis-15262	113	5	introduce	introduce	VERB
fcis-15262	113	6	a	a	DET
fcis-15262	113	7	"	"	PUNCT
fcis-15262	113	8	gate	gate	NOUN
fcis-15262	113	9	"	"	PUNCT
fcis-15262	113	10	mechanism	mechanism	NOUN
fcis-15262	113	11	on	on	ADP
fcis-15262	113	12	top	top	NOUN
fcis-15262	113	13	of	of	ADP
fcis-15262	113	14	the	the	DET
fcis-15262	113	15	basic	basic	ADJ
fcis-15262	113	16	rnn[18	rnn[18	PROPN
fcis-15262	113	17	]	]	PUNCT
fcis-15262	113	18	,	,	PUNCT
fcis-15262	113	19	which	which	PRON
fcis-15262	113	20	allows	allow	VERB
fcis-15262	113	21	for	for	ADP
fcis-15262	113	22	determining	determine	VERB
fcis-15262	113	23	which	which	DET
fcis-15262	113	24	data	datum	NOUN
fcis-15262	113	25	should	should	AUX
fcis-15262	113	26	be	be	AUX
fcis-15262	113	27	retained	retain	VERB
fcis-15262	113	28	or	or	CCONJ
fcis-15262	113	29	discarded	discard	VERB
fcis-15262	113	30	.	.	PUNCT
fcis-15262	114	1	the	the	DET
fcis-15262	114	2	simplified	simplified	ADJ
fcis-15262	114	3	structure	structure	NOUN
fcis-15262	114	4	of	of	ADP
fcis-15262	114	5	gru	gru	NOUN
fcis-15262	114	6	is	be	AUX
fcis-15262	114	7	shown	show	VERB
fcis-15262	114	8	in	in	ADP
fcis-15262	114	9	figure	figure	NOUN
fcis-15262	114	10	12	12	NUM
fcis-15262	114	11	.	.	PUNCT
fcis-15262	115	1	gru	gru	NOUN
fcis-15262	115	2	simplifies	simplifie	NOUN
fcis-15262	115	3	the	the	DET
fcis-15262	115	4	three	three	NUM
fcis-15262	115	5	gate	gate	NOUN
fcis-15262	115	6	functions	function	NOUN
fcis-15262	115	7	in	in	ADP
fcis-15262	115	8	lstm	lstm	NOUN
fcis-15262	115	9	to	to	ADP
fcis-15262	115	10	two	two	NUM
fcis-15262	115	11	gate	gate	NOUN
fcis-15262	115	12	functions	function	NOUN
fcis-15262	115	13	.	.	PUNCT
fcis-15262	116	1	the	the	DET
fcis-15262	116	2	function	function	NOUN
fcis-15262	116	3	zt	zt	PROPN
fcis-15262	116	4	represents	represent	VERB
fcis-15262	116	5	the	the	DET
fcis-15262	116	6	update	update	NOUN
fcis-15262	116	7	gate	gate	NOUN
fcis-15262	116	8	,	,	PUNCT
fcis-15262	116	9	where	where	SCONJ
fcis-15262	116	10	a	a	DET
fcis-15262	116	11	higher	high	ADJ
fcis-15262	116	12	value	value	NOUN
fcis-15262	116	13	indicates	indicate	VERB
fcis-15262	116	14	more	more	ADJ
fcis-15262	116	15	information	information	NOUN
fcis-15262	116	16	from	from	ADP
fcis-15262	116	17	the	the	DET
fcis-15262	116	18	previous	previous	ADJ
fcis-15262	116	19	state	state	NOUN
fcis-15262	116	20	being	be	AUX
fcis-15262	116	21	incorporated	incorporate	VERB
fcis-15262	116	22	into	into	ADP
fcis-15262	116	23	the	the	DET
fcis-15262	116	24	current	current	ADJ
fcis-15262	116	25	state	state	NOUN
fcis-15262	116	26	.	.	PUNCT
fcis-15262	117	1	the	the	DET
fcis-15262	117	2	function	function	NOUN
fcis-15262	117	3	rt	rt	PROPN
fcis-15262	117	4	represents	represent	VERB
fcis-15262	117	5	the	the	DET
fcis-15262	117	6	reset	reset	NOUN
fcis-15262	117	7	gate	gate	NOUN
fcis-15262	117	8	,	,	PUNCT
fcis-15262	117	9	which	which	PRON
fcis-15262	117	10	determines	determine	VERB
fcis-15262	117	11	how	how	SCONJ
fcis-15262	117	12	much	much	ADJ
fcis-15262	117	13	information	information	NOUN
fcis-15262	117	14	from	from	ADP
fcis-15262	117	15	the	the	DET
fcis-15262	117	16	previous	previous	ADJ
fcis-15262	117	17	state	state	NOUN
fcis-15262	117	18	needs	need	VERB
fcis-15262	117	19	to	to	PART
fcis-15262	117	20	be	be	AUX
fcis-15262	117	21	forgotten	forget	VERB
fcis-15262	117	22	.	.	PUNCT
fcis-15262	118	1	1tanh	1tanh	NUM
fcis-15262	118	2	ht-1	ht-1	NUM
fcis-15262	118	3	xt	xt	NUM
fcis-15262	119	1	ht	ht	INTJ
fcis-15262	119	2	rt	rt	PROPN
fcis-15262	119	3	zt	zt	PROPN
fcis-15262	119	4			PROPN
fcis-15262	119	5			PROPN
fcis-15262	119	6	figure	figure	NOUN
fcis-15262	119	7	12	12	NUM
fcis-15262	119	8	.	.	PUNCT
fcis-15262	120	1	schematic	schematic	ADJ
fcis-15262	120	2	diagram	diagram	NOUN
fcis-15262	120	3	of	of	ADP
fcis-15262	120	4	the	the	DET
fcis-15262	120	5	gru	gru	NOUN
fcis-15262	120	6	structure	structure	NOUN
fcis-15262	120	7	compared	compare	VERB
fcis-15262	120	8	to	to	ADP
fcis-15262	120	9	lstm	lstm	PROPN
fcis-15262	120	10	,	,	PUNCT
fcis-15262	120	11	gru	gru	PROPN
fcis-15262	120	12	has	have	VERB
fcis-15262	120	13	fewer	few	ADJ
fcis-15262	120	14	parameters	parameter	NOUN
fcis-15262	120	15	,	,	PUNCT
fcis-15262	120	16	higher	high	ADJ
fcis-15262	120	17	computational	computational	ADJ
fcis-15262	120	18	efficiency	efficiency	NOUN
fcis-15262	120	19	,	,	PUNCT
fcis-15262	120	20	and	and	CCONJ
fcis-15262	120	21	faster	fast	ADJ
fcis-15262	120	22	training	training	NOUN
fcis-15262	120	23	time	time	NOUN
fcis-15262	120	24	.	.	PUNCT
fcis-15262	121	1	however	however	ADV
fcis-15262	121	2	,	,	PUNCT
fcis-15262	121	3	since	since	SCONJ
fcis-15262	121	4	the	the	DET
fcis-15262	121	5	data	datum	NOUN
fcis-15262	121	6	in	in	ADP
fcis-15262	121	7	gru	gru	NOUN
fcis-15262	121	8	is	be	AUX
fcis-15262	121	9	unidirectional	unidirectional	ADJ
fcis-15262	121	10	,	,	PUNCT
fcis-15262	121	11	predictions	prediction	NOUN
fcis-15262	121	12	can	can	AUX
fcis-15262	121	13	only	only	ADV
fcis-15262	121	14	be	be	AUX
fcis-15262	121	15	based	base	VERB
fcis-15262	121	16	on	on	ADP
fcis-15262	121	17	previous	previous	ADJ
fcis-15262	121	18	data	datum	NOUN
fcis-15262	121	19	.	.	PUNCT
fcis-15262	122	1	bidirectional	bidirectional	ADJ
fcis-15262	122	2	gated	gate	VERB
fcis-15262	122	3	recurrent	recurrent	ADJ
fcis-15262	122	4	unit	unit	NOUN
fcis-15262	122	5	(	(	PUNCT
fcis-15262	122	6	bigru	bigru	NOUN
fcis-15262	122	7	)	)	PUNCT
fcis-15262	122	8	consists	consist	VERB
fcis-15262	122	9	of	of	ADP
fcis-15262	122	10	two	two	NUM
fcis-15262	122	11	grus	grus	NOUN
fcis-15262	122	12	with	with	ADP
fcis-15262	122	13	opposite	opposite	ADJ
fcis-15262	122	14	directions	direction	NOUN
fcis-15262	122	15	,	,	PUNCT
fcis-15262	122	16	allowing	allow	VERB
fcis-15262	122	17	the	the	DET
fcis-15262	122	18	utilization	utilization	NOUN
fcis-15262	122	19	of	of	ADP
fcis-15262	122	20	information	information	NOUN
fcis-15262	122	21	from	from	ADP
fcis-15262	122	22	both	both	CCONJ
fcis-15262	122	23	previous	previous	ADJ
fcis-15262	122	24	and	and	CCONJ
fcis-15262	122	25	future	future	ADJ
fcis-15262	122	26	contexts	contexts	NOUN
fcis-15262	122	27	.	.	PUNCT
fcis-15262	123	1	this	this	PRON
fcis-15262	123	2	leads	lead	VERB
fcis-15262	123	3	to	to	ADP
fcis-15262	123	4	more	more	ADV
fcis-15262	123	5	accurate	accurate	ADJ
fcis-15262	123	6	predictions	prediction	NOUN
fcis-15262	123	7	.	.	PUNCT
fcis-15262	124	1	the	the	DET
fcis-15262	124	2	structure	structure	NOUN
fcis-15262	124	3	of	of	ADP
fcis-15262	124	4	bigru	bigru	NOUN
fcis-15262	124	5	is	be	AUX
fcis-15262	124	6	shown	show	VERB
fcis-15262	124	7	in	in	ADP
fcis-15262	124	8	figure	figure	NOUN
fcis-15262	124	9	13	13	NUM
fcis-15262	124	10	,	,	PUNCT
fcis-15262	124	11	where	where	SCONJ
fcis-15262	124	12	the	the	DET
fcis-15262	124	13	input	input	NOUN
fcis-15262	124	14	information	information	NOUN
fcis-15262	124	15	is	be	AUX
fcis-15262	124	16	simultaneously	simultaneously	ADV
fcis-15262	124	17	fed	feed	VERB
fcis-15262	124	18	into	into	ADP
fcis-15262	124	19	two	two	NUM
fcis-15262	124	20	gru	gru	NOUN
fcis-15262	124	21	networks	network	NOUN
fcis-15262	124	22	with	with	ADP
fcis-15262	124	23	opposite	opposite	ADJ
fcis-15262	124	24	directions	direction	NOUN
fcis-15262	124	25	,	,	PUNCT
fcis-15262	124	26	and	and	CCONJ
fcis-15262	124	27	the	the	DET
fcis-15262	124	28	output	output	NOUN
fcis-15262	124	29	is	be	AUX
fcis-15262	124	30	determined	determine	VERB
fcis-15262	124	31	by	by	ADP
fcis-15262	124	32	both	both	DET
fcis-15262	124	33	gru	gru	PROPN
fcis-15262	124	34	networks	network	NOUN
fcis-15262	124	35	.	.	PUNCT
fcis-15262	125	1	h1	h1	PROPN
fcis-15262	125	2	h2	h2	PROPN
fcis-15262	125	3	ht	ht	PROPN
fcis-15262	125	4	gru	gru	PROPN
fcis-15262	125	5	gru	gru	PROPN
fcis-15262	125	6	gru	gru	PROPN
fcis-15262	125	7	gru	gru	PROPN
fcis-15262	125	8	gru	gru	PROPN
fcis-15262	125	9	gru	gru	PROPN
fcis-15262	126	1	x1	x1	PROPN
fcis-15262	126	2	x2	x2	PROPN
fcis-15262	126	3	xt	xt	PROPN
fcis-15262	126	4	...	...	PUNCT
fcis-15262	126	5	...	...	PUNCT
fcis-15262	127	1	output	output	NOUN
fcis-15262	127	2	layer	layer	NOUN
fcis-15262	127	3	backward	backward	ADV
fcis-15262	127	4	hidden	hide	VERB
fcis-15262	127	5	layer	layer	NOUN
fcis-15262	127	6	forward	forward	ADV
fcis-15262	127	7	hidden	hide	VERB
fcis-15262	127	8	layer	layer	NOUN
fcis-15262	127	9	input	input	NOUN
fcis-15262	127	10	layer	layer	NOUN
fcis-15262	127	11	figure	figure	NOUN
fcis-15262	127	12	13	13	NUM
fcis-15262	127	13	.	.	PUNCT
fcis-15262	128	1	schematic	schematic	ADJ
fcis-15262	128	2	diagram	diagram	NOUN
fcis-15262	128	3	of	of	ADP
fcis-15262	128	4	the	the	DET
fcis-15262	128	5	bigru	bigru	NOUN
fcis-15262	128	6	structure	structure	NOUN
fcis-15262	128	7	3.4	3.4	NUM
fcis-15262	128	8	.	.	PUNCT
fcis-15262	129	1	densenet169	densenet169	PROPN
fcis-15262	129	2	-	-	PUNCT
fcis-15262	129	3	bigru	bigru	NOUN
fcis-15262	129	4	network	network	NOUN
fcis-15262	129	5	the	the	DET
fcis-15262	129	6	working	work	VERB
fcis-15262	129	7	principle	principle	NOUN
fcis-15262	129	8	of	of	ADP
fcis-15262	129	9	densenet169	densenet169	PROPN
fcis-15262	129	10	-	-	PUNCT
fcis-15262	129	11	bigru	bigru	NOUN
fcis-15262	129	12	network	network	NOUN
fcis-15262	129	13	is	be	AUX
fcis-15262	129	14	shown	show	VERB
fcis-15262	129	15	in	in	ADP
fcis-15262	129	16	figure	figure	NOUN
fcis-15262	129	17	14	14	NUM
fcis-15262	129	18	.	.	PUNCT
fcis-15262	130	1	the	the	DET
fcis-15262	130	2	two	two	NUM
fcis-15262	130	3	networks	network	NOUN
fcis-15262	130	4	are	be	AUX
fcis-15262	130	5	concatenated	concatenate	VERB
fcis-15262	130	6	,	,	PUNCT
fcis-15262	130	7	where	where	SCONJ
fcis-15262	130	8	the	the	DET
fcis-15262	130	9	feature	feature	NOUN
fcis-15262	130	10	information	information	NOUN
fcis-15262	130	11	is	be	AUX
fcis-15262	130	12	input	input	VERB
fcis-15262	130	13	into	into	ADP
fcis-15262	130	14	densenet169	densenet169	PROPN
fcis-15262	130	15	for	for	ADP
fcis-15262	130	16	feature	feature	NOUN
fcis-15262	130	17	extraction	extraction	NOUN
fcis-15262	130	18	and	and	CCONJ
fcis-15262	130	19	dimensionality	dimensionality	NOUN
fcis-15262	130	20	reduction	reduction	NOUN
fcis-15262	130	21	through	through	ADP
fcis-15262	130	22	convolutional	convolutional	ADJ
fcis-15262	130	23	calculations	calculation	NOUN
fcis-15262	130	24	.	.	PUNCT
fcis-15262	131	1	the	the	DET
fcis-15262	131	2	output	output	NOUN
fcis-15262	131	3	information	information	NOUN
fcis-15262	131	4	serves	serve	VERB
fcis-15262	131	5	as	as	ADP
fcis-15262	131	6	the	the	DET
fcis-15262	131	7	input	input	NOUN
fcis-15262	131	8	for	for	ADP
fcis-15262	131	9	bigru	bigru	NOUN
fcis-15262	131	10	network	network	NOUN
fcis-15262	131	11	,	,	PUNCT
fcis-15262	131	12	which	which	PRON
fcis-15262	131	13	learns	learn	VERB
fcis-15262	131	14	the	the	DET
fcis-15262	131	15	temporal	temporal	ADJ
fcis-15262	131	16	dependencies	dependency	NOUN
fcis-15262	131	17	among	among	ADP
fcis-15262	131	18	the	the	DET
fcis-15262	131	19	data	datum	NOUN
fcis-15262	131	20	.	.	PUNCT
fcis-15262	132	1	the	the	DET
fcis-15262	132	2	conca	conca	PROPN
fcis-15262	132	3	layer	layer	NOUN
fcis-15262	132	4	integrates	integrate	VERB
fcis-15262	132	5	the	the	DET
fcis-15262	132	6	output	output	NOUN
fcis-15262	132	7	information	information	NOUN
fcis-15262	132	8	from	from	ADP
fcis-15262	132	9	the	the	DET
fcis-15262	132	10	two	two	NUM
fcis-15262	132	11	gru	gru	NOUN
fcis-15262	132	12	networks	network	NOUN
fcis-15262	132	13	with	with	ADP
fcis-15262	132	14	opposite	opposite	ADJ
fcis-15262	132	15	directions	direction	NOUN
fcis-15262	132	16	,	,	PUNCT
fcis-15262	132	17	and	and	CCONJ
fcis-15262	132	18	the	the	DET
fcis-15262	132	19	results	result	NOUN
fcis-15262	132	20	are	be	AUX
fcis-15262	132	21	produced	produce	VERB
fcis-15262	132	22	through	through	ADP
fcis-15262	132	23	a	a	DET
fcis-15262	132	24	classifier	classifier	NOUN
fcis-15262	132	25	.	.	PUNCT
fcis-15262	133	1	gru	gru	PROPN
fcis-15262	133	2	gru	gru	PROPN
fcis-15262	133	3	gru	gru	PROPN
fcis-15262	133	4	gru	gru	PROPN
fcis-15262	133	5	gru	gru	PROPN
fcis-15262	133	6	gru	gru	PROPN
fcis-15262	133	7	...	...	PUNCT
fcis-15262	133	8	...	...	PUNCT
fcis-15262	134	1	input	input	NOUN
fcis-15262	134	2	densenet169	densenet169	PROPN
fcis-15262	134	3	concatenation	concatenation	PROPN
fcis-15262	134	4	output	output	NOUN
fcis-15262	134	5	bigru	bigru	NOUN
fcis-15262	134	6	figure	figure	NOUN
fcis-15262	134	7	14	14	NUM
fcis-15262	134	8	.	.	PUNCT
fcis-15262	135	1	schematic	schematic	ADJ
fcis-15262	135	2	diagram	diagram	NOUN
fcis-15262	135	3	of	of	ADP
fcis-15262	135	4	the	the	DET
fcis-15262	135	5	densenet169	densenet169	PROPN
fcis-15262	135	6	-	-	PUNCT
fcis-15262	135	7	bigru	bigru	NOUN
fcis-15262	135	8	structure	structure	NOUN
fcis-15262	135	9	4	4	NUM
fcis-15262	135	10	.	.	PUNCT
fcis-15262	135	11	voiceprint	voiceprint	NOUN
fcis-15262	135	12	recognition	recognition	PROPN
fcis-15262	135	13	experiment	experiment	NOUN
fcis-15262	135	14	4.1	4.1	NUM
fcis-15262	135	15	.	.	PUNCT
fcis-15262	136	1	data	datum	NOUN
fcis-15262	136	2	preparation	preparation	NOUN
fcis-15262	136	3	the	the	DET
fcis-15262	136	4	speech	speech	NOUN
fcis-15262	136	5	dataset	dataset	NOUN
fcis-15262	136	6	used	use	VERB
fcis-15262	136	7	in	in	ADP
fcis-15262	136	8	the	the	DET
fcis-15262	136	9	experiments	experiment	NOUN
fcis-15262	136	10	includes	include	VERB
fcis-15262	136	11	84	84	NUM
fcis-15262	136	12	thchs-30	thchs-30	PROPN
fcis-15262	136	13	from	from	ADP
fcis-15262	136	14	open	open	ADJ
fcis-15262	136	15	speech	speech	NOUN
fcis-15262	136	16	and	and	CCONJ
fcis-15262	136	17	language	language	NOUN
fcis-15262	136	18	resources	resource	NOUN
fcis-15262	136	19	and	and	CCONJ
fcis-15262	136	20	free	free	PROPN
fcis-15262	136	21	st	st	PROPN
fcis-15262	136	22	chinese	chinese	PROPN
fcis-15262	136	23	mandarin	mandarin	PROPN
fcis-15262	136	24	corpus	corpus	NOUN
fcis-15262	136	25	.	.	PUNCT
fcis-15262	137	1	thchs-30	thchs-30	NUM
fcis-15262	137	2	,	,	PUNCT
fcis-15262	137	3	released	release	VERB
fcis-15262	137	4	by	by	ADP
fcis-15262	137	5	the	the	DET
fcis-15262	137	6	center	center	NOUN
fcis-15262	137	7	for	for	ADP
fcis-15262	137	8	speech	speech	NOUN
fcis-15262	137	9	and	and	CCONJ
fcis-15262	137	10	language	language	NOUN
fcis-15262	137	11	technology	technology	NOUN
fcis-15262	137	12	(	(	PUNCT
fcis-15262	137	13	cslt	cslt	PROPN
fcis-15262	137	14	)	)	PUNCT
fcis-15262	137	15	at	at	ADP
fcis-15262	137	16	tsinghua	tsinghua	PROPN
fcis-15262	137	17	university	university	PROPN
fcis-15262	137	18	,	,	PUNCT
fcis-15262	137	19	is	be	AUX
fcis-15262	137	20	an	an	DET
fcis-15262	137	21	open	open	ADJ
fcis-15262	137	22	chinese	chinese	ADJ
fcis-15262	137	23	speech	speech	NOUN
fcis-15262	137	24	database	database	NOUN
fcis-15262	137	25	.	.	PUNCT
fcis-15262	138	1	the	the	DET
fcis-15262	138	2	participants	participant	NOUN
fcis-15262	138	3	in	in	ADP
fcis-15262	138	4	the	the	DET
fcis-15262	138	5	thchs30	thchs30	ADJ
fcis-15262	138	6	chinese	chinese	ADJ
fcis-15262	138	7	dataset	dataset	PROPN
fcis-15262	138	8	are	be	AUX
fcis-15262	138	9	university	university	NOUN
fcis-15262	138	10	students	student	NOUN
fcis-15262	138	11	who	who	PRON
fcis-15262	138	12	are	be	AUX
fcis-15262	138	13	fluent	fluent	ADJ
fcis-15262	138	14	in	in	ADP
fcis-15262	138	15	mandarin	mandarin	NOUN
fcis-15262	138	16	.	.	PUNCT
fcis-15262	139	1	the	the	DET
fcis-15262	139	2	recorded	record	VERB
fcis-15262	139	3	texts	text	NOUN
fcis-15262	139	4	are	be	AUX
fcis-15262	139	5	derived	derive	VERB
fcis-15262	139	6	from	from	ADP
fcis-15262	139	7	a	a	DET
fcis-15262	139	8	large	large	ADJ
fcis-15262	139	9	corpus	corpus	NOUN
fcis-15262	139	10	of	of	ADP
fcis-15262	139	11	news	news	NOUN
fcis-15262	139	12	,	,	PUNCT
fcis-15262	139	13	with	with	ADP
fcis-15262	139	14	a	a	DET
fcis-15262	139	15	total	total	ADJ
fcis-15262	139	16	duration	duration	NOUN
fcis-15262	139	17	of	of	ADP
fcis-15262	139	18	30	30	NUM
fcis-15262	139	19	hours[19	hours[19	NOUN
fcis-15262	139	20	]	]	PUNCT
fcis-15262	139	21	.	.	PUNCT
fcis-15262	140	1	the	the	DET
fcis-15262	140	2	latter	latter	ADJ
fcis-15262	140	3	dataset	dataset	NOUN
fcis-15262	140	4	,	,	PUNCT
fcis-15262	140	5	free	free	PROPN
fcis-15262	140	6	st	st	PROPN
fcis-15262	140	7	chinese	chinese	PROPN
fcis-15262	140	8	mandarin	mandarin	PROPN
fcis-15262	140	9	corpus	corpus	PROPN
fcis-15262	140	10	,	,	PUNCT
fcis-15262	140	11	is	be	AUX
fcis-15262	140	12	provided	provide	VERB
fcis-15262	140	13	free	free	ADJ
fcis-15262	140	14	-	-	PUNCT
fcis-15262	140	15	of	of	ADP
fcis-15262	140	16	-	-	PUNCT
fcis-15262	140	17	charge	charge	NOUN
fcis-15262	140	18	by	by	ADP
fcis-15262	140	19	surfingtech	surfingtech	NOUN
fcis-15262	140	20	(	(	PUNCT
fcis-15262	140	21	www.surfing.ai	www.surfing.ai	PROPN
fcis-15262	140	22	)	)	PUNCT
fcis-15262	140	23	.	.	PUNCT
fcis-15262	141	1	it	it	PRON
fcis-15262	141	2	consists	consist	VERB
fcis-15262	141	3	of	of	ADP
fcis-15262	141	4	speech	speech	NOUN
fcis-15262	141	5	data	datum	NOUN
fcis-15262	141	6	from	from	ADP
fcis-15262	141	7	855	855	NUM
fcis-15262	141	8	speakers	speaker	NOUN
fcis-15262	141	9	,	,	PUNCT
fcis-15262	141	10	with	with	ADP
fcis-15262	141	11	120	120	NUM
fcis-15262	141	12	speech	speech	NOUN
fcis-15262	141	13	segments	segment	NOUN
fcis-15262	141	14	per	per	ADP
fcis-15262	141	15	person	person	NOUN
fcis-15262	141	16	,	,	PUNCT
fcis-15262	141	17	totaling	total	VERB
fcis-15262	141	18	102,600	102,600	NUM
fcis-15262	141	19	speech	speech	NOUN
fcis-15262	141	20	segments	segment	NOUN
fcis-15262	141	21	.	.	PUNCT
fcis-15262	142	1	the	the	DET
fcis-15262	142	2	audio	audio	NOUN
fcis-15262	142	3	is	be	AUX
fcis-15262	142	4	sampled	sample	VERB
fcis-15262	142	5	at	at	ADP
fcis-15262	142	6	a	a	DET
fcis-15262	142	7	frequency	frequency	NOUN
fcis-15262	142	8	of	of	ADP
fcis-15262	142	9	16khz	16khz	NOUN
fcis-15262	142	10	.	.	PUNCT
fcis-15262	143	1	in	in	ADP
fcis-15262	143	2	this	this	DET
fcis-15262	143	3	study	study	NOUN
fcis-15262	143	4	,	,	PUNCT
fcis-15262	143	5	80	80	NUM
fcis-15262	143	6	participants	participant	NOUN
fcis-15262	143	7	(	(	PUNCT
fcis-15262	143	8	equal	equal	ADJ
fcis-15262	143	9	number	number	NOUN
fcis-15262	143	10	of	of	ADP
fcis-15262	143	11	male	male	ADJ
fcis-15262	143	12	and	and	CCONJ
fcis-15262	143	13	female	female	NOUN
fcis-15262	143	14	)	)	PUNCT
fcis-15262	143	15	were	be	AUX
fcis-15262	143	16	involved	involve	VERB
fcis-15262	143	17	.	.	PUNCT
fcis-15262	144	1	each	each	DET
fcis-15262	144	2	participant	participant	NOUN
fcis-15262	144	3	had	have	VERB
fcis-15262	144	4	110	110	NUM
fcis-15262	144	5	speech	speech	NOUN
fcis-15262	144	6	segments	segment	NOUN
fcis-15262	144	7	,	,	PUNCT
fcis-15262	144	8	with	with	ADP
fcis-15262	144	9	the	the	DET
fcis-15262	144	10	same	same	ADJ
fcis-15262	144	11	content	content	NOUN
fcis-15262	144	12	.	.	PUNCT
fcis-15262	145	1	the	the	DET
fcis-15262	145	2	distribution	distribution	NOUN
fcis-15262	145	3	of	of	ADP
fcis-15262	145	4	speech	speech	NOUN
fcis-15262	145	5	segments	segment	NOUN
fcis-15262	145	6	for	for	ADP
fcis-15262	145	7	each	each	DET
fcis-15262	145	8	individual	individual	NOUN
fcis-15262	145	9	is	be	AUX
fcis-15262	145	10	shown	show	VERB
fcis-15262	145	11	in	in	ADP
fcis-15262	145	12	table	table	NOUN
fcis-15262	145	13	2	2	NUM
fcis-15262	145	14	.	.	PUNCT
fcis-15262	146	1	the	the	DET
fcis-15262	146	2	training	training	NOUN
fcis-15262	146	3	,	,	PUNCT
fcis-15262	146	4	validation	validation	NOUN
fcis-15262	146	5	,	,	PUNCT
fcis-15262	146	6	and	and	CCONJ
fcis-15262	146	7	testing	test	VERB
fcis-15262	146	8	set	set	NOUN
fcis-15262	146	9	proportions	proportion	NOUN
fcis-15262	146	10	are	be	AUX
fcis-15262	146	11	set	set	VERB
fcis-15262	146	12	to	to	ADP
fcis-15262	146	13	9:1:1	9:1:1	NUM
fcis-15262	146	14	.	.	PUNCT
fcis-15262	147	1	table	table	NOUN
fcis-15262	147	2	2	2	NUM
fcis-15262	147	3	.	.	PUNCT
fcis-15262	148	1	distribution	distribution	NOUN
fcis-15262	148	2	of	of	ADP
fcis-15262	148	3	speech	speech	NOUN
fcis-15262	148	4	data	datum	NOUN
fcis-15262	148	5	speaker	speaker	NOUN
fcis-15262	148	6	i	i	PROPN
fcis-15262	148	7	d	d	PROPN
fcis-15262	148	8	quantity	quantity	NOUN
fcis-15262	148	9	average	average	ADJ
fcis-15262	148	10	duration(seconds	duration(second	NOUN
fcis-15262	148	11	)	)	PUNCT
fcis-15262	148	12	sampling	sample	VERB
fcis-15262	148	13	frequency	frequency	NOUN
fcis-15262	148	14	(	(	PUNCT
fcis-15262	148	15	khz	khz	NOUN
fcis-15262	148	16	)	)	PUNCT
fcis-15262	148	17	1	1	NUM
fcis-15262	148	18	110	110	NUM
fcis-15262	148	19	3.5	3.5	NUM
fcis-15262	148	20	16	16	NUM
fcis-15262	148	21	2	2	NUM
fcis-15262	148	22	110	110	NUM
fcis-15262	148	23	3.5	3.5	NUM
fcis-15262	148	24	16	16	NUM
fcis-15262	148	25	...	...	PUNCT
fcis-15262	149	1	80	80	NUM
fcis-15262	149	2	...	...	SYM
fcis-15262	150	1	110	110	NUM
fcis-15262	150	2	...	...	SYM
fcis-15262	150	3	3.5	3.5	NUM
fcis-15262	150	4	...	...	PUNCT
fcis-15262	150	5	16	16	NUM
fcis-15262	150	6	4.2	4.2	NUM
fcis-15262	150	7	.	.	PUNCT
fcis-15262	151	1	evaluation	evaluation	NOUN
fcis-15262	151	2	criteria	criterion	NOUN
fcis-15262	151	3	this	this	DET
fcis-15262	151	4	article	article	NOUN
fcis-15262	151	5	selects	select	VERB
fcis-15262	151	6	accuracy	accuracy	NOUN
fcis-15262	151	7	as	as	ADP
fcis-15262	151	8	the	the	DET
fcis-15262	151	9	evaluation	evaluation	NOUN
fcis-15262	151	10	criteria	criterion	NOUN
fcis-15262	151	11	,	,	PUNCT
fcis-15262	151	12	and	and	CCONJ
fcis-15262	151	13	the	the	DET
fcis-15262	151	14	calculation	calculation	NOUN
fcis-15262	151	15	formula	formula	NOUN
fcis-15262	151	16	is	be	AUX
fcis-15262	151	17	as	as	SCONJ
fcis-15262	151	18	follows	follow	VERB
fcis-15262	151	19	:	:	PUNCT
fcis-15262	151	20	tp	tp	NOUN
fcis-15262	151	21	represents	represent	VERB
fcis-15262	151	22	the	the	DET
fcis-15262	151	23	samples	sample	NOUN
fcis-15262	151	24	that	that	PRON
fcis-15262	151	25	are	be	AUX
fcis-15262	151	26	positive	positive	ADJ
fcis-15262	151	27	and	and	CCONJ
fcis-15262	151	28	are	be	AUX
fcis-15262	151	29	predicted	predict	VERB
fcis-15262	151	30	as	as	ADP
fcis-15262	151	31	positive	positive	ADJ
fcis-15262	151	32	by	by	ADP
fcis-15262	151	33	the	the	DET
fcis-15262	151	34	model	model	NOUN
fcis-15262	151	35	.	.	PUNCT
fcis-15262	152	1	fn	fn	PROPN
fcis-15262	152	2	represents	represent	VERB
fcis-15262	152	3	the	the	DET
fcis-15262	152	4	data	data	NOUN
fcis-15262	152	5	samples	sample	NOUN
fcis-15262	152	6	that	that	PRON
fcis-15262	152	7	are	be	AUX
fcis-15262	152	8	positive	positive	ADJ
fcis-15262	152	9	but	but	CCONJ
fcis-15262	152	10	predicted	predict	VERB
fcis-15262	152	11	as	as	ADP
fcis-15262	152	12	negative	negative	ADJ
fcis-15262	152	13	by	by	ADP
fcis-15262	152	14	the	the	DET
fcis-15262	152	15	model	model	NOUN
fcis-15262	152	16	.	.	PUNCT
fcis-15262	153	1	fp	fp	PROPN
fcis-15262	153	2	represents	represent	VERB
fcis-15262	153	3	the	the	DET
fcis-15262	153	4	data	data	NOUN
fcis-15262	153	5	samples	sample	NOUN
fcis-15262	153	6	that	that	PRON
fcis-15262	153	7	are	be	AUX
fcis-15262	153	8	negative	negative	ADJ
fcis-15262	153	9	but	but	CCONJ
fcis-15262	153	10	identified	identify	VERB
fcis-15262	153	11	as	as	ADP
fcis-15262	153	12	positive	positive	ADJ
fcis-15262	153	13	by	by	ADP
fcis-15262	153	14	the	the	DET
fcis-15262	153	15	model	model	NOUN
fcis-15262	153	16	.	.	PUNCT
fcis-15262	154	1	tn	tn	PROPN
fcis-15262	154	2	represents	represent	VERB
fcis-15262	154	3	the	the	DET
fcis-15262	154	4	data	data	NOUN
fcis-15262	154	5	samples	sample	NOUN
fcis-15262	154	6	that	that	PRON
fcis-15262	154	7	are	be	AUX
fcis-15262	154	8	negative	negative	ADJ
fcis-15262	154	9	and	and	CCONJ
fcis-15262	154	10	predicted	predict	VERB
fcis-15262	154	11	as	as	ADP
fcis-15262	154	12	negative	negative	ADJ
fcis-15262	154	13	by	by	ADP
fcis-15262	154	14	the	the	DET
fcis-15262	154	15	model	model	NOUN
fcis-15262	154	16	.	.	PUNCT
fcis-15262	155	1	acc	acc	PROPN
fcis-15262	155	2	tp	tp	PROPN
fcis-15262	155	3	tn	tn	PROPN
fcis-15262	155	4	tp	tp	ADP
fcis-15262	155	5	tn	tn	PROPN
fcis-15262	156	1	fp	fp	PROPN
fcis-15262	156	2	fn	fn	PROPN
fcis-15262	156	3			PROPN
fcis-15262	156	4			PROPN
fcis-15262	156	5			PROPN
fcis-15262	156	6			PUNCT
fcis-15262	156	7			X
fcis-15262	156	8	(	(	PUNCT
fcis-15262	156	9	12	12	NUM
fcis-15262	156	10	)	)	PUNCT
fcis-15262	156	11	4.3	4.3	NUM
fcis-15262	156	12	.	.	PUNCT
fcis-15262	157	1	experimental	experimental	ADJ
fcis-15262	157	2	results	result	NOUN
fcis-15262	157	3	and	and	CCONJ
fcis-15262	157	4	analysis	analysis	NOUN
fcis-15262	157	5	in	in	ADP
fcis-15262	157	6	order	order	NOUN
fcis-15262	157	7	to	to	PART
fcis-15262	157	8	verify	verify	VERB
fcis-15262	157	9	the	the	DET
fcis-15262	157	10	performance	performance	NOUN
fcis-15262	157	11	of	of	ADP
fcis-15262	157	12	the	the	DET
fcis-15262	157	13	model	model	NOUN
fcis-15262	157	14	,	,	PUNCT
fcis-15262	157	15	the	the	DET
fcis-15262	157	16	network	network	NOUN
fcis-15262	157	17	model	model	NOUN
fcis-15262	157	18	constructed	construct	VERB
fcis-15262	157	19	in	in	ADP
fcis-15262	157	20	this	this	DET
fcis-15262	157	21	paper	paper	NOUN
fcis-15262	157	22	was	be	AUX
fcis-15262	157	23	compared	compare	VERB
fcis-15262	157	24	with	with	ADP
fcis-15262	157	25	densenet169	densenet169	PROPN
fcis-15262	157	26	and	and	CCONJ
fcis-15262	157	27	densenet169	densenet169	PROPN
fcis-15262	157	28	-	-	PUNCT
fcis-15262	157	29	gru	gru	PROPN
fcis-15262	157	30	in	in	ADP
fcis-15262	157	31	terms	term	NOUN
fcis-15262	157	32	of	of	ADP
fcis-15262	157	33	validation	validation	NOUN
fcis-15262	157	34	set	set	VERB
fcis-15262	157	35	accuracy	accuracy	NOUN
fcis-15262	157	36	.	.	PUNCT
fcis-15262	158	1	the	the	DET
fcis-15262	158	2	results	result	NOUN
fcis-15262	158	3	are	be	AUX
fcis-15262	158	4	shown	show	VERB
fcis-15262	158	5	in	in	ADP
fcis-15262	158	6	figure	figure	NOUN
fcis-15262	158	7	15	15	NUM
fcis-15262	158	8	.	.	PUNCT
fcis-15262	159	1	the	the	DET
fcis-15262	159	2	validation	validation	NOUN
fcis-15262	159	3	accuracy	accuracy	NOUN
fcis-15262	159	4	of	of	ADP
fcis-15262	159	5	dnesenet169	dnesenet169	PROPN
fcis-15262	159	6	-	-	PUNCT
fcis-15262	159	7	bigru	bigru	NOUN
fcis-15262	159	8	with	with	ADP
fcis-15262	159	9	a	a	DET
fcis-15262	159	10	dual	dual	ADJ
fcis-15262	159	11	-	-	PUNCT
fcis-15262	159	12	layer	layer	NOUN
fcis-15262	159	13	gru	gru	PROPN
fcis-15262	159	14	reached	reach	VERB
fcis-15262	159	15	98.9	98.9	NUM
fcis-15262	159	16	%	%	NOUN
fcis-15262	159	17	,	,	PUNCT
fcis-15262	159	18	and	and	CCONJ
fcis-15262	159	19	the	the	DET
fcis-15262	159	20	test	test	NOUN
fcis-15262	159	21	accuracy	accuracy	NOUN
fcis-15262	159	22	was	be	AUX
fcis-15262	159	23	97.3	97.3	NUM
fcis-15262	159	24	%	%	NOUN
fcis-15262	159	25	.	.	PUNCT
fcis-15262	160	1	the	the	DET
fcis-15262	160	2	validation	validation	NOUN
fcis-15262	160	3	accuracy	accuracy	NOUN
fcis-15262	160	4	of	of	ADP
fcis-15262	160	5	dnesenet169	dnesenet169	PROPN
fcis-15262	160	6	-	-	PUNCT
fcis-15262	160	7	gru	gru	PROPN
fcis-15262	160	8	with	with	ADP
fcis-15262	160	9	a	a	DET
fcis-15262	160	10	single	single	ADJ
fcis-15262	160	11	-	-	PUNCT
fcis-15262	160	12	layer	layer	NOUN
fcis-15262	160	13	gru	gru	NOUN
fcis-15262	160	14	was	be	AUX
fcis-15262	160	15	97.3	97.3	NUM
fcis-15262	160	16	%	%	NOUN
fcis-15262	160	17	,	,	PUNCT
fcis-15262	160	18	and	and	CCONJ
fcis-15262	160	19	the	the	DET
fcis-15262	160	20	test	test	NOUN
fcis-15262	160	21	accuracy	accuracy	NOUN
fcis-15262	160	22	was	be	AUX
fcis-15262	160	23	96.3	96.3	NUM
fcis-15262	160	24	%	%	NOUN
fcis-15262	160	25	.	.	PUNCT
fcis-15262	161	1	the	the	DET
fcis-15262	161	2	validation	validation	NOUN
fcis-15262	161	3	accuracy	accuracy	NOUN
fcis-15262	161	4	of	of	ADP
fcis-15262	161	5	densenet169	densenet169	PROPN
fcis-15262	161	6	was	be	AUX
fcis-15262	161	7	95.9	95.9	NUM
fcis-15262	161	8	%	%	NOUN
fcis-15262	161	9	.	.	PUNCT
fcis-15262	162	1	the	the	DET
fcis-15262	162	2	test	test	NOUN
fcis-15262	162	3	accuracy	accuracy	NOUN
fcis-15262	162	4	was	be	AUX
fcis-15262	162	5	94.6	94.6	NUM
fcis-15262	162	6	%	%	NOUN
fcis-15262	162	7	.	.	PUNCT
fcis-15262	163	1	this	this	PRON
fcis-15262	163	2	also	also	ADV
fcis-15262	163	3	indicates	indicate	VERB
fcis-15262	163	4	that	that	SCONJ
fcis-15262	163	5	the	the	DET
fcis-15262	163	6	recognition	recognition	NOUN
fcis-15262	163	7	of	of	ADP
fcis-15262	163	8	voice	voice	NOUN
fcis-15262	163	9	features	feature	VERB
fcis-15262	163	10	not	not	PART
fcis-15262	163	11	only	only	ADV
fcis-15262	163	12	involves	involve	VERB
fcis-15262	163	13	the	the	DET
fcis-15262	163	14	temporal	temporal	ADJ
fcis-15262	163	15	aspect	aspect	NOUN
fcis-15262	163	16	but	but	CCONJ
fcis-15262	163	17	also	also	ADV
fcis-15262	163	18	the	the	DET
fcis-15262	163	19	importance	importance	NOUN
fcis-15262	163	20	of	of	ADP
fcis-15262	163	21	data	datum	NOUN
fcis-15262	163	22	before	before	ADV
fcis-15262	163	23	and	and	CCONJ
fcis-15262	163	24	after	after	ADP
fcis-15262	163	25	a	a	DET
fcis-15262	163	26	certain	certain	ADJ
fcis-15262	163	27	moment	moment	NOUN
fcis-15262	163	28	.	.	PUNCT
fcis-15262	164	1	moreover	moreover	ADV
fcis-15262	164	2	,	,	PUNCT
fcis-15262	164	3	using	use	VERB
fcis-15262	164	4	this	this	DET
fcis-15262	164	5	information	information	NOUN
fcis-15262	164	6	can	can	AUX
fcis-15262	164	7	further	far	ADV
fcis-15262	164	8	improve	improve	VERB
fcis-15262	164	9	the	the	DET
fcis-15262	164	10	accuracy	accuracy	NOUN
fcis-15262	164	11	.	.	PUNCT
fcis-15262	165	1	figure	figure	NOUN
fcis-15262	165	2	16	16	NUM
fcis-15262	165	3	shows	show	VERB
fcis-15262	165	4	the	the	DET
fcis-15262	165	5	confusion	confusion	NOUN
fcis-15262	165	6	matrix	matrix	NOUN
fcis-15262	165	7	of	of	ADP
fcis-15262	165	8	the	the	DET
fcis-15262	165	9	test	test	NOUN
fcis-15262	165	10	results	result	NOUN
fcis-15262	165	11	of	of	ADP
fcis-15262	165	12	dnesenet169	dnesenet169	PROPN
fcis-15262	165	13	-	-	PUNCT
fcis-15262	165	14	bigru	bigru	NOUN
fcis-15262	165	15	,	,	PUNCT
fcis-15262	165	16	and	and	CCONJ
fcis-15262	165	17	the	the	DET
fcis-15262	165	18	accuracy	accuracy	NOUN
fcis-15262	165	19	is	be	AUX
fcis-15262	165	20	the	the	DET
fcis-15262	165	21	ratio	ratio	NOUN
fcis-15262	165	22	of	of	ADP
fcis-15262	165	23	the	the	DET
fcis-15262	165	24	sum	sum	NOUN
fcis-15262	165	25	of	of	ADP
fcis-15262	165	26	the	the	DET
fcis-15262	165	27	diagonal	diagonal	ADJ
fcis-15262	165	28	numbers	number	NOUN
fcis-15262	165	29	to	to	ADP
fcis-15262	165	30	the	the	DET
fcis-15262	165	31	total	total	ADJ
fcis-15262	165	32	number	number	NOUN
fcis-15262	165	33	of	of	ADP
fcis-15262	165	34	the	the	DET
fcis-15262	165	35	test	test	NOUN
fcis-15262	165	36	samples	sample	NOUN
fcis-15262	165	37	.	.	PUNCT
fcis-15262	166	1	next	next	ADV
fcis-15262	166	2	,	,	PUNCT
fcis-15262	166	3	the	the	DET
fcis-15262	166	4	accuracy	accuracy	NOUN
fcis-15262	166	5	of	of	ADP
fcis-15262	166	6	the	the	DET
fcis-15262	166	7	test	test	NOUN
fcis-15262	166	8	set	set	NOUN
fcis-15262	166	9	is	be	AUX
fcis-15262	166	10	compared	compare	VERB
fcis-15262	166	11	with	with	ADP
fcis-15262	166	12	the	the	DET
fcis-15262	166	13	proposed	propose	VERB
fcis-15262	166	14	methods	method	NOUN
fcis-15262	166	15	in	in	ADP
fcis-15262	166	16	voiceprint	voiceprint	NOUN
fcis-15262	166	17	recognition	recognition	NOUN
fcis-15262	166	18	research	research	NOUN
fcis-15262	166	19	,	,	PUNCT
fcis-15262	166	20	and	and	CCONJ
fcis-15262	166	21	the	the	DET
fcis-15262	166	22	results	result	NOUN
fcis-15262	166	23	are	be	AUX
fcis-15262	166	24	shown	show	VERB
fcis-15262	166	25	in	in	ADP
fcis-15262	166	26	table	table	NOUN
fcis-15262	166	27	3	3	NUM
fcis-15262	166	28	.	.	PUNCT
fcis-15262	166	29	figure	figure	NOUN
fcis-15262	166	30	15	15	NUM
fcis-15262	166	31	.	.	PUNCT
fcis-15262	167	1	validation	validation	NOUN
fcis-15262	167	2	accuracy	accuracy	NOUN
fcis-15262	167	3	curves	curve	NOUN
fcis-15262	167	4	of	of	ADP
fcis-15262	167	5	three	three	NUM
fcis-15262	167	6	network	network	NOUN
fcis-15262	167	7	models	model	NOUN
fcis-15262	167	8	table	table	VERB
fcis-15262	167	9	3	3	NUM
fcis-15262	167	10	.	.	PUNCT
fcis-15262	168	1	distribution	distribution	NOUN
fcis-15262	168	2	of	of	ADP
fcis-15262	168	3	speech	speech	NOUN
fcis-15262	168	4	data	datum	NOUN
fcis-15262	168	5	net	net	ADJ
fcis-15262	168	6	accuracy	accuracy	NOUN
fcis-15262	168	7	(	(	PUNCT
fcis-15262	168	8	%	%	INTJ
fcis-15262	168	9	)	)	PUNCT
fcis-15262	168	10	dnesenet169bigru	dnesenet169bigru	NOUN
fcis-15262	168	11	97.3	97.3	NUM
fcis-15262	168	12	%	%	NOUN
fcis-15262	168	13	densenet169	densenet169	PROPN
fcis-15262	168	14	-	-	PUNCT
fcis-15262	168	15	gru	gru	PROPN
fcis-15262	168	16	96.2	96.2	NUM
fcis-15262	168	17	%	%	NOUN
fcis-15262	168	18	densenet121bigru	densenet121bigru	PUNCT
fcis-15262	168	19	cnn	cnn	PROPN
fcis-15262	168	20	-	-	PUNCT
fcis-15262	168	21	lstm	lstm	PROPN
fcis-15262	168	22	resnet34	resnet34	NOUN
fcis-15262	168	23	-	-	PUNCT
fcis-15262	168	24	se	se	PROPN
fcis-15262	168	25	96.3	96.3	NUM
fcis-15262	168	26	%	%	NOUN
fcis-15262	168	27	95.42	95.42	NUM
fcis-15262	168	28	%	%	NOUN
fcis-15262	168	29	90.3	90.3	NUM
fcis-15262	168	30	figure	figure	NOUN
fcis-15262	168	31	16	16	NUM
fcis-15262	168	32	.	.	PUNCT
fcis-15262	169	1	confusion	confusion	NOUN
fcis-15262	169	2	matrix	matrix	NOUN
fcis-15262	169	3	of	of	ADP
fcis-15262	169	4	densenet169	densenet169	PROPN
fcis-15262	169	5	-	-	PUNCT
fcis-15262	169	6	bigru	bigru	NOUN
fcis-15262	169	7	5	5	NUM
fcis-15262	169	8	.	.	PUNCT
fcis-15262	169	9	conclusion	conclusion	NOUN
fcis-15262	169	10	this	this	DET
fcis-15262	169	11	paper	paper	NOUN
fcis-15262	169	12	proposes	propose	VERB
fcis-15262	169	13	a	a	DET
fcis-15262	169	14	joint	joint	ADJ
fcis-15262	169	15	model	model	NOUN
fcis-15262	169	16	combining	combine	VERB
fcis-15262	169	17	densenet169	densenet169	PROPN
fcis-15262	169	18	and	and	CCONJ
fcis-15262	169	19	a	a	DET
fcis-15262	169	20	two	two	NUM
fcis-15262	169	21	-	-	PUNCT
fcis-15262	169	22	layer	layer	NOUN
fcis-15262	169	23	gru	gru	NOUN
fcis-15262	169	24	network	network	NOUN
fcis-15262	169	25	,	,	PUNCT
fcis-15262	169	26	and	and	CCONJ
fcis-15262	169	27	adopts	adopt	VERB
fcis-15262	169	28	depth	depth	NOUN
fcis-15262	169	29	-	-	PUNCT
fcis-15262	169	30	wise	wise	ADJ
fcis-15262	169	31	separable	separable	ADJ
fcis-15262	169	32	convolution	convolution	NOUN
fcis-15262	169	33	with	with	ADP
fcis-15262	169	34	fewer	few	ADJ
fcis-15262	169	35	model	model	NOUN
fcis-15262	169	36	parameters	parameter	NOUN
fcis-15262	169	37	to	to	PART
fcis-15262	169	38	improve	improve	VERB
fcis-15262	169	39	computational	computational	ADJ
fcis-15262	169	40	efficiency	efficiency	NOUN
fcis-15262	169	41	.	.	PUNCT
fcis-15262	170	1	the	the	DET
fcis-15262	170	2	experimental	experimental	ADJ
fcis-15262	170	3	results	result	NOUN
fcis-15262	170	4	show	show	VERB
fcis-15262	170	5	that	that	SCONJ
fcis-15262	170	6	the	the	DET
fcis-15262	170	7	model	model	NOUN
fcis-15262	170	8	performs	perform	VERB
fcis-15262	170	9	well	well	ADV
fcis-15262	170	10	with	with	ADP
fcis-15262	170	11	a	a	DET
fcis-15262	170	12	test	test	NOUN
fcis-15262	170	13	accuracy	accuracy	NOUN
fcis-15262	170	14	of	of	ADP
fcis-15262	170	15	97.3	97.3	NUM
fcis-15262	170	16	%	%	NOUN
fcis-15262	170	17	.	.	PUNCT
fcis-15262	171	1	in	in	ADP
fcis-15262	171	2	order	order	NOUN
fcis-15262	171	3	to	to	PART
fcis-15262	171	4	verify	verify	VERB
fcis-15262	171	5	the	the	DET
fcis-15262	171	6	performance	performance	NOUN
fcis-15262	171	7	,	,	PUNCT
fcis-15262	171	8	experiments	experiment	NOUN
fcis-15262	171	9	were	be	AUX
fcis-15262	171	10	conducted	conduct	VERB
fcis-15262	171	11	comparing	compare	VERB
fcis-15262	171	12	the	the	DET
fcis-15262	171	13	model	model	NOUN
fcis-15262	171	14	with	with	ADP
fcis-15262	171	15	densenet169	densenet169	PROPN
fcis-15262	171	16	and	and	CCONJ
fcis-15262	171	17	densenet169	densenet169	PROPN
fcis-15262	171	18	-	-	PUNCT
fcis-15262	171	19	bigru	bigru	NOUN
fcis-15262	171	20	,	,	PUNCT
fcis-15262	171	21	as	as	ADV
fcis-15262	171	22	well	well	ADV
fcis-15262	171	23	as	as	ADP
fcis-15262	171	24	models	model	NOUN
fcis-15262	171	25	proposed	propose	VERB
fcis-15262	171	26	by	by	ADP
fcis-15262	171	27	other	other	ADJ
fcis-15262	171	28	authors	author	NOUN
fcis-15262	171	29	.	.	PUNCT
fcis-15262	172	1	the	the	DET
fcis-15262	172	2	advantages	advantage	NOUN
fcis-15262	172	3	of	of	ADP
fcis-15262	172	4	the	the	DET
fcis-15262	172	5	model	model	NOUN
fcis-15262	172	6	in	in	ADP
fcis-15262	172	7	this	this	DET
fcis-15262	172	8	paper	paper	NOUN
fcis-15262	172	9	can	can	AUX
fcis-15262	172	10	also	also	ADV
fcis-15262	172	11	be	be	AUX
fcis-15262	172	12	seen	see	VERB
fcis-15262	172	13	from	from	ADP
fcis-15262	172	14	the	the	DET
fcis-15262	172	15	table	table	NOUN
fcis-15262	172	16	.	.	PUNCT
fcis-15262	173	1	references	reference	NOUN
fcis-15262	173	2	[	[	X
fcis-15262	173	3	1	1	NUM
fcis-15262	173	4	]	]	PUNCT
fcis-15262	173	5	chen	chen	PROPN
fcis-15262	173	6	guo	guo	PROPN
fcis-15262	173	7	-	-	PUNCT
fcis-15262	173	8	guo	guo	PROPN
fcis-15262	173	9	,	,	PUNCT
fcis-15262	173	10	chen	chen	PROPN
fcis-15262	173	11	jia	jia	PROPN
fcis-15262	173	12	-	-	PUNCT
fcis-15262	173	13	yu	yu	PROPN
fcis-15262	173	14	,	,	PUNCT
fcis-15262	173	15	na	na	ADP
fcis-15262	173	16	xing	xing	PROPN
fcis-15262	173	17	-	-	PUNCT
fcis-15262	173	18	yu’et	yu’et	PROPN
fcis-15262	173	19	al	al	PROPN
fcis-15262	173	20	.	.	PUNCT
fcis-15262	174	1	kaldi	kaldi	PROPN
fcis-15262	174	2	speech	speech	NOUN
fcis-15262	174	3	recognition	recognition	NOUN
fcis-15262	174	4	combat[m	combat[m	NOUN
fcis-15262	174	5	]	]	PUNCT
fcis-15262	174	6	.	.	PUNCT
fcis-15262	175	1	1st	1st	ADJ
fcis-15262	175	2	ed	ed	NOUN
fcis-15262	175	3	.	.	PUNCT
fcis-15262	176	1	beijing	beijing	PROPN
fcis-15262	176	2	:	:	PUNCT
fcis-15262	176	3	publishing	publish	VERB
fcis-15262	176	4	house	house	NOUN
fcis-15262	176	5	of	of	ADP
fcis-15262	176	6	electronics	electronic	NOUN
fcis-15262	176	7	industry	industry	NOUN
fcis-15262	176	8	2020	2020	NUM
fcis-15262	176	9	.	.	PUNCT
fcis-15262	177	1	(	(	PUNCT
fcis-15262	177	2	in	in	ADP
fcis-15262	177	3	chinese	chinese	PROPN
fcis-15262	177	4	)	)	PUNCT
fcis-15262	178	1	[	[	X
fcis-15262	178	2	2	2	X
fcis-15262	178	3	]	]	PUNCT
fcis-15262	178	4	hansen	hansen	PROPN
fcis-15262	178	5	j	j	PROPN
fcis-15262	178	6	h	h	PROPN
fcis-15262	178	7	l	l	PROPN
fcis-15262	178	8	,	,	PUNCT
fcis-15262	178	9	hasan	hasan	PROPN
fcis-15262	178	10	t.	t.	PROPN
fcis-15262	178	11	speaker	speaker	PROPN
fcis-15262	178	12	recognition	recognition	PROPN
fcis-15262	178	13	by	by	ADP
fcis-15262	178	14	machines	machine	NOUN
fcis-15262	178	15	and	and	CCONJ
fcis-15262	178	16	humans	human	NOUN
fcis-15262	178	17	:	:	PUNCT
fcis-15262	178	18	a	a	DET
fcis-15262	178	19	tutorial	tutorial	NOUN
fcis-15262	178	20	review[j	review[j	PROPN
fcis-15262	178	21	/	/	SYM
fcis-15262	178	22	ol	ol	PROPN
fcis-15262	178	23	]	]	PUNCT
fcis-15262	178	24	.	.	PUNCT
fcis-15262	179	1	ieee	ieee	PROPN
fcis-15262	179	2	signal	signal	PROPN
fcis-15262	179	3	processing	processing	NOUN
fcis-15262	179	4	magazine	magazine	NOUN
fcis-15262	179	5	,	,	PUNCT
fcis-15262	179	6	2015	2015	NUM
fcis-15262	179	7	,	,	PUNCT
fcis-15262	179	8	32(6	32(6	NUM
fcis-15262	179	9	):	):	PUNCT
fcis-15262	179	10	74	74	NUM
fcis-15262	179	11	-	-	SYM
fcis-15262	179	12	99	99	NUM
fcis-15262	179	13	.	.	PUNCT
fcis-15262	180	1	doi:10.1109/	doi:10.1109/	PROPN
fcis-15262	180	2	msp	msp	PROPN
fcis-15262	180	3	.	.	PROPN
fcis-15262	180	4	2015	2015	NUM
fcis-15262	180	5	.	.	PUNCT
fcis-15262	181	1	246	246	NUM
fcis-15262	181	2	2851	2851	NUM
fcis-15262	181	3	.	.	PUNCT
fcis-15262	182	1	[	[	X
fcis-15262	182	2	3	3	X
fcis-15262	182	3	]	]	X
fcis-15262	182	4	liu	liu	PROPN
fcis-15262	182	5	d	d	PROPN
fcis-15262	182	6	,	,	PUNCT
fcis-15262	182	7	xu	xu	PROPN
fcis-15262	182	8	j	j	PROPN
fcis-15262	182	9	,	,	PUNCT
fcis-15262	182	10	zhang	zhang	PROPN
fcis-15262	182	11	p	p	PROPN
fcis-15262	182	12	,	,	PUNCT
fcis-15262	182	13	et	et	PROPN
fcis-15262	182	14	al	al	PROPN
fcis-15262	182	15	.	.	PUNCT
fcis-15262	183	1	a	a	DET
fcis-15262	183	2	unified	unified	ADJ
fcis-15262	183	3	system	system	NOUN
fcis-15262	183	4	for	for	ADP
fcis-15262	183	5	multilingual	multilingual	ADJ
fcis-15262	183	6	speech	speech	NOUN
fcis-15262	183	7	recognition	recognition	NOUN
fcis-15262	183	8	and	and	CCONJ
fcis-15262	183	9	language	language	NOUN
fcis-15262	183	10	identification	identification	NOUN
fcis-15262	183	11	85	85	NUM
fcis-15262	183	12	[	[	X
fcis-15262	183	13	j	j	X
fcis-15262	183	14	/	/	SYM
fcis-15262	183	15	ol	ol	PROPN
fcis-15262	183	16	]	]	PUNCT
fcis-15262	183	17	.	.	PUNCT
fcis-15262	184	1	speech	speech	NOUN
fcis-15262	184	2	communication	communication	NOUN
fcis-15262	184	3	,	,	PUNCT
fcis-15262	184	4	2021	2021	NUM
fcis-15262	184	5	,	,	PUNCT
fcis-15262	184	6	127	127	NUM
fcis-15262	184	7	:	:	PUNCT
fcis-15262	184	8	17	17	NUM
fcis-15262	184	9	-	-	SYM
fcis-15262	184	10	28	28	NUM
fcis-15262	184	11	.	.	PUNCT
fcis-15262	185	1	doi	doi	NOUN
fcis-15262	185	2	:	:	PUNCT
fcis-15262	185	3	10	10	NUM
fcis-15262	185	4	.	.	X
fcis-15262	186	1	1016/	1016/	NUM
fcis-15262	186	2	j.specom.2020.12.008	j.specom.2020.12.008	NOUN
fcis-15262	186	3	.	.	PUNCT
fcis-15262	187	1	[	[	X
fcis-15262	187	2	4	4	NUM
fcis-15262	187	3	]	]	X
fcis-15262	187	4	ablimit	ablimit	NOUN
fcis-15262	187	5	m	m	PROPN
fcis-15262	187	6	,	,	PUNCT
fcis-15262	187	7	xueli	xueli	PROPN
fcis-15262	187	8	m	m	PROPN
fcis-15262	187	9	,	,	PUNCT
fcis-15262	187	10	hamdulla	hamdulla	PROPN
fcis-15262	187	11	a.	a.	NOUN
fcis-15262	187	12	language	language	PROPN
fcis-15262	187	13	identification	identification	NOUN
fcis-15262	187	14	research	research	NOUN
fcis-15262	187	15	based	base	VERB
fcis-15262	187	16	on	on	ADP
fcis-15262	187	17	dual	dual	ADJ
fcis-15262	187	18	attention	attention	NOUN
fcis-15262	187	19	mechanism[c	mechanism[c	NOUN
fcis-15262	187	20	/	/	SYM
fcis-15262	187	21	ol]//2021	ol]//2021	NOUN
fcis-15262	187	22	ieee	ieee	PROPN
fcis-15262	187	23	2nd	2nd	PROPN
fcis-15262	187	24	international	international	ADJ
fcis-15262	187	25	conference	conference	NOUN
fcis-15262	187	26	on	on	ADP
fcis-15262	187	27	pattern	pattern	NOUN
fcis-15262	187	28	recognition	recognition	NOUN
fcis-15262	187	29	and	and	CCONJ
fcis-15262	187	30	machine	machine	NOUN
fcis-15262	187	31	learning	learning	NOUN
fcis-15262	187	32	(	(	PUNCT
fcis-15262	187	33	prml	prml	PROPN
fcis-15262	187	34	)	)	PUNCT
fcis-15262	187	35	.	.	PUNCT
fcis-15262	188	1	2021	2021	NUM
fcis-15262	188	2	:	:	PUNCT
fcis-15262	188	3	241	241	NUM
fcis-15262	188	4	-	-	SYM
fcis-15262	188	5	246[2023	246[2023	NUM
fcis-15262	188	6	-	-	PUNCT
fcis-15262	188	7	12	12	NUM
fcis-15262	188	8	-	-	SYM
fcis-15262	188	9	10	10	NUM
fcis-15262	188	10	]	]	PUNCT
fcis-15262	188	11	.	.	PUNCT
fcis-15262	189	1	https://	https://	PROPN
fcis-15262	189	2	ieeexplore	ieeexplore	PROPN
fcis-15262	189	3	.	.	PUNCT
fcis-15262	190	1	ieee.org/	ieee.org/	VERB
fcis-15262	190	2	document/	document/	NUM
fcis-15262	190	3	9520699	9520699	NUM
fcis-15262	190	4	.	.	PUNCT
fcis-15262	191	1	doi:10.1109	doi:10.1109	VERB
fcis-15262	191	2	/	/	SYM
fcis-15262	191	3	prml52754.2021.9520699	prml52754.2021.9520699	NOUN
fcis-15262	191	4	.	.	PUNCT
fcis-15262	192	1	[	[	X
fcis-15262	192	2	5	5	NUM
fcis-15262	192	3	]	]	X
fcis-15262	192	4	zhang	zhang	PROPN
fcis-15262	192	5	c	c	PROPN
fcis-15262	192	6	,	,	PUNCT
fcis-15262	192	7	koishida	koishida	PROPN
fcis-15262	192	8	k	k	PROPN
fcis-15262	192	9	,	,	PUNCT
fcis-15262	192	10	hansen	hansen	PROPN
fcis-15262	192	11	j.	j.	PROPN
fcis-15262	192	12	text	text	PROPN
fcis-15262	192	13	-	-	PUNCT
fcis-15262	192	14	independent	independent	ADJ
fcis-15262	192	15	speaker	speaker	NOUN
fcis-15262	192	16	verification	verification	NOUN
fcis-15262	192	17	based	base	VERB
fcis-15262	192	18	on	on	ADP
fcis-15262	192	19	triplet	triplet	NOUN
fcis-15262	192	20	convolutional	convolutional	ADJ
fcis-15262	192	21	neural	neural	ADJ
fcis-15262	192	22	network	network	NOUN
fcis-15262	192	23	embeddings[j	embeddings[j	PROPN
fcis-15262	192	24	/	/	SYM
fcis-15262	192	25	ol	ol	PROPN
fcis-15262	192	26	]	]	PUNCT
fcis-15262	192	27	.	.	PUNCT
fcis-15262	193	1	ieee	ieee	PROPN
fcis-15262	193	2	/	/	SYM
fcis-15262	193	3	acm	acm	PROPN
fcis-15262	193	4	transactions	transaction	NOUN
fcis-15262	193	5	on	on	ADP
fcis-15262	193	6	audio	audio	NOUN
fcis-15262	193	7	,	,	PUNCT
fcis-15262	193	8	speech	speech	NOUN
fcis-15262	193	9	,	,	PUNCT
fcis-15262	193	10	and	and	CCONJ
fcis-15262	193	11	language	language	NOUN
fcis-15262	193	12	processing	processing	NOUN
fcis-15262	193	13	,	,	PUNCT
fcis-15262	193	14	2018	2018	NUM
fcis-15262	193	15	,	,	PUNCT
fcis-15262	193	16	26	26	NUM
fcis-15262	193	17	:	:	SYM
fcis-15262	193	18	1	1	NUM
fcis-15262	193	19	-	-	SYM
fcis-15262	193	20	1	1	NUM
fcis-15262	193	21	.	.	PUNCT
fcis-15262	194	1	doi:10.1109	doi:10.1109	VERB
fcis-15262	194	2	/	/	SYM
fcis-15262	194	3	taslp.2018.2831456	taslp.2018.2831456	NOUN
fcis-15262	194	4	.	.	PUNCT
fcis-15262	195	1	[	[	X
fcis-15262	195	2	6	6	NUM
fcis-15262	195	3	]	]	X
fcis-15262	195	4	sarkar	sarkar	PROPN
fcis-15262	195	5	a	a	PROPN
fcis-15262	195	6	,	,	PUNCT
fcis-15262	195	7	tan	tan	PROPN
fcis-15262	195	8	z	z	PROPN
fcis-15262	195	9	h.	h.	PROPN
fcis-15262	195	10	text	text	NOUN
fcis-15262	195	11	dependent	dependent	ADJ
fcis-15262	195	12	speaker	speaker	NOUN
fcis-15262	195	13	verification	verification	NOUN
fcis-15262	195	14	using	use	VERB
fcis-15262	195	15	un	un	ADJ
fcis-15262	195	16	-	-	ADJ
fcis-15262	195	17	supervised	supervised	ADJ
fcis-15262	195	18	hmm	hmm	NOUN
fcis-15262	195	19	-	-	PUNCT
fcis-15262	195	20	ubm	ubm	NOUN
fcis-15262	195	21	and	and	CCONJ
fcis-15262	195	22	temporal	temporal	ADJ
fcis-15262	195	23	gmmubm[c	gmmubm[c	PROPN
fcis-15262	195	24	/	/	SYM
fcis-15262	195	25	ol	ol	ADJ
fcis-15262	195	26	]	]	X
fcis-15262	195	27	.	.	PUNCT
fcis-15262	196	1	2016	2016	NUM
fcis-15262	196	2	.	.	PUNCT
fcis-15262	197	1	doi:10.21437	doi:10.21437	PROPN
fcis-15262	197	2	/	/	SYM
fcis-15262	197	3	interspeech.2016	interspeech.2016	PROPN
fcis-15262	197	4	-	-	PUNCT
fcis-15262	197	5	362	362	NUM
fcis-15262	197	6	.	.	PUNCT
fcis-15262	198	1	[	[	X
fcis-15262	198	2	7	7	X
fcis-15262	198	3	]	]	X
fcis-15262	198	4	hao	hao	PROPN
fcis-15262	198	5	m	m	PROPN
fcis-15262	198	6	,	,	PUNCT
fcis-15262	198	7	liu	liu	PROPN
fcis-15262	198	8	h	h	PROPN
fcis-15262	198	9	,	,	PUNCT
fcis-15262	198	10	li	li	PROPN
fcis-15262	198	11	y	y	PROPN
fcis-15262	198	12	,	,	PUNCT
fcis-15262	198	13	et	et	PROPN
fcis-15262	198	14	al	al	PROPN
fcis-15262	198	15	.	.	PROPN
fcis-15262	198	16	speech	speech	NOUN
fcis-15262	198	17	tracking	tracking	NOUN
fcis-15262	198	18	based	base	VERB
fcis-15262	198	19	on	on	ADP
fcis-15262	198	20	cluster	cluster	NOUN
fcis-15262	198	21	analysis	analysis	NOUN
fcis-15262	198	22	and	and	CCONJ
fcis-15262	198	23	speaker	speaker	NOUN
fcis-15262	198	24	recognition[j	recognition[j	PROPN
fcis-15262	198	25	/	/	SYM
fcis-15262	198	26	ol	ol	PROPN
fcis-15262	198	27	]	]	PUNCT
fcis-15262	198	28	.	.	PUNCT
fcis-15262	199	1	computer	computer	NOUN
fcis-15262	199	2	and	and	CCONJ
fcis-15262	199	3	modernization	modernization	NOUN
fcis-15262	199	4	,	,	PUNCT
fcis-15262	199	5	2020	2020	NUM
fcis-15262	199	6	,	,	PUNCT
fcis-15262	199	7	4	4	NUM
fcis-15262	199	8	:	:	SYM
fcis-15262	199	9	11	11	NUM
fcis-15262	199	10	-	-	SYM
fcis-15262	199	11	17	17	NUM
fcis-15262	199	12	.	.	PUNCT
fcis-15262	199	13	doi:10.3969	doi:10.3969	NOUN
fcis-15262	199	14	/	/	SYM
fcis-15262	199	15	j.issn.1006	j.issn.1006	NOUN
fcis-15262	199	16	-	-	PUNCT
fcis-15262	199	17	2475	2475	NUM
fcis-15262	199	18	.	.	PUNCT
fcis-15262	200	1	2020.04.002	2020.04.002	X
fcis-15262	200	2	.	.	PUNCT
fcis-15262	201	1	[	[	X
fcis-15262	201	2	8	8	NUM
fcis-15262	201	3	]	]	X
fcis-15262	201	4	liu	liu	PROPN
fcis-15262	201	5	z	z	PROPN
fcis-15262	201	6	,	,	PUNCT
fcis-15262	201	7	wu	wu	PROPN
fcis-15262	201	8	z	z	PROPN
fcis-15262	201	9	,	,	PUNCT
fcis-15262	201	10	li	li	PROPN
fcis-15262	201	11	t	t	PROPN
fcis-15262	201	12	,	,	PUNCT
fcis-15262	201	13	et	et	PROPN
fcis-15262	201	14	al	al	PROPN
fcis-15262	201	15	.	.	PROPN
fcis-15262	201	16	gmm	gmm	PROPN
fcis-15262	201	17	and	and	CCONJ
fcis-15262	201	18	cnn	cnn	PROPN
fcis-15262	201	19	hybrid	hybrid	NOUN
fcis-15262	201	20	method	method	NOUN
fcis-15262	201	21	for	for	ADP
fcis-15262	201	22	short	short	ADJ
fcis-15262	201	23	utterance	utterance	NOUN
fcis-15262	201	24	speaker	speaker	NOUN
fcis-15262	201	25	recognition[j	recognition[j	PROPN
fcis-15262	201	26	/	/	SYM
fcis-15262	201	27	ol	ol	PROPN
fcis-15262	201	28	]	]	PUNCT
fcis-15262	201	29	.	.	PUNCT
fcis-15262	202	1	ieee	ieee	NOUN
fcis-15262	202	2	transactions	transaction	NOUN
fcis-15262	202	3	on	on	ADP
fcis-15262	202	4	industrial	industrial	ADJ
fcis-15262	202	5	informatics	informatic	NOUN
fcis-15262	202	6	,	,	PUNCT
fcis-15262	202	7	2018	2018	NUM
fcis-15262	202	8	,	,	PUNCT
fcis-15262	202	9	pp	pp	CCONJ
fcis-15262	202	10	:	:	PUNCT
fcis-15262	202	11	1	1	NUM
fcis-15262	202	12	-	-	SYM
fcis-15262	202	13	1	1	NUM
fcis-15262	202	14	.	.	PUNCT
fcis-15262	203	1	doi	doi	NOUN
fcis-15262	203	2	:	:	PUNCT
fcis-15262	203	3	10	10	NUM
fcis-15262	203	4	.	.	X
fcis-15262	204	1	1109/	1109/	NUM
fcis-15262	204	2	tii.2018.2799928	tii.2018.2799928	NUM
fcis-15262	204	3	.	.	PUNCT
fcis-15262	205	1	[	[	X
fcis-15262	205	2	9	9	NUM
fcis-15262	205	3	]	]	X
fcis-15262	205	4	luo	luo	PROPN
fcis-15262	205	5	k	k	PROPN
fcis-15262	205	6	,	,	PUNCT
fcis-15262	205	7	fu	fu	PROPN
fcis-15262	205	8	l.	l.	PROPN
fcis-15262	205	9	research	research	PROPN
fcis-15262	205	10	and	and	CCONJ
fcis-15262	205	11	application	application	NOUN
fcis-15262	205	12	of	of	ADP
fcis-15262	205	13	voiceprint	voiceprint	NOUN
fcis-15262	205	14	recognition	recognition	NOUN
fcis-15262	205	15	based	base	VERB
fcis-15262	205	16	on	on	ADP
fcis-15262	205	17	a	a	DET
fcis-15262	205	18	deep	deep	ADJ
fcis-15262	205	19	recurrent	recurrent	ADJ
fcis-15262	205	20	neural	neural	ADJ
fcis-15262	205	21	network[m	network[m	PROPN
fcis-15262	205	22	/	/	SYM
fcis-15262	205	23	ol	ol	PROPN
fcis-15262	205	24	]	]	X
fcis-15262	205	25	.	.	PUNCT
fcis-15262	206	1	2019	2019	NUM
fcis-15262	206	2	:	:	PUNCT
fcis-15262	206	3	309	309	NUM
fcis-15262	206	4	-	-	SYM
fcis-15262	206	5	316	316	NUM
fcis-15262	206	6	.	.	PUNCT
fcis-15262	206	7	doi:10.1201/9780429468605	doi:10.1201/9780429468605	NOUN
fcis-15262	206	8	-	-	PUNCT
fcis-15262	206	9	43	43	NUM
fcis-15262	206	10	.	.	PUNCT
fcis-15262	207	1	[	[	X
fcis-15262	207	2	10	10	NUM
fcis-15262	207	3	]	]	PUNCT
fcis-15262	207	4	he	he	PRON
fcis-15262	207	5	y	y	PROPN
fcis-15262	207	6	,	,	PUNCT
fcis-15262	207	7	yingyan	yingyan	PROPN
fcis-15262	208	1	d	d	NOUN
fcis-15262	208	2	,	,	PUNCT
fcis-15262	208	3	peng	peng	PROPN
fcis-15262	208	4	w	w	PROPN
fcis-15262	208	5	,	,	PUNCT
fcis-15262	208	6	et	et	PROPN
fcis-15262	208	7	al	al	PROPN
fcis-15262	208	8	.	.	PUNCT
fcis-15262	208	9	voiceprint	voiceprint	PROPN
fcis-15262	208	10	recognition	recognition	NOUN
fcis-15262	208	11	based	base	VERB
fcis-15262	208	12	on	on	ADP
fcis-15262	208	13	cnn	cnn	PROPN
fcis-15262	208	14	-	-	PUNCT
fcis-15262	208	15	lstm	lstm	PROPN
fcis-15262	208	16	network	network	NOUN
fcis-15262	208	17	[	[	X
fcis-15262	208	18	j	j	PROPN
fcis-15262	208	19	/	/	SYM
fcis-15262	208	20	ol	ol	PROPN
fcis-15262	208	21	]	]	PUNCT
fcis-15262	208	22	.	.	PUNCT
fcis-15262	209	1	computer	computer	NOUN
fcis-15262	209	2	applications	application	NOUN
fcis-15262	209	3	and	and	CCONJ
fcis-15262	209	4	software	software	NOUN
fcis-15262	209	5	,	,	PUNCT
fcis-15262	209	6	2019	2019	NUM
fcis-15262	210	1	[	[	X
fcis-15262	210	2	2023	2023	NUM
fcis-15262	210	3	-	-	SYM
fcis-15262	210	4	1215	1215	NUM
fcis-15262	210	5	]	]	PUNCT
fcis-15262	210	6	.	.	PUNCT
fcis-15262	211	1	http://en.cnki	http://en.cnki	NOUN
fcis-15262	211	2	.	.	PUNCT
fcis-15262	212	1	com.cn/article	com.cn/article	X
fcis-15262	212	2	_	_	PUNCT
fcis-15262	212	3	en	en	PROPN
fcis-15262	212	4	/	/	SYM
fcis-15262	212	5	cjfd	cjfd	ADJ
fcis-15262	212	6	total	total	ADJ
fcis-15262	212	7	-	-	PUNCT
fcis-15262	212	8	jyrj	jyrj	NOUN
fcis-15262	212	9	2019	2019	NUM
fcis-15262	212	10	04027.htm	04027.htm	NOUN
fcis-15262	212	11	.	.	PUNCT
fcis-15262	213	1	[	[	X
fcis-15262	213	2	11	11	NUM
fcis-15262	213	3	]	]	X
fcis-15262	213	4	ma	ma	PROPN
fcis-15262	213	5	y	y	PROPN
fcis-15262	213	6	,	,	PUNCT
fcis-15262	213	7	yuan	yuan	PROPN
fcis-15262	213	8	m	m	PROPN
fcis-15262	213	9	,	,	PUNCT
fcis-15262	213	10	qi	qi	PROPN
fcis-15262	213	11	c	c	PROPN
fcis-15262	213	12	,	,	PUNCT
fcis-15262	213	13	et	et	PROPN
fcis-15262	213	14	al	al	PROPN
fcis-15262	213	15	.	.	PROPN
fcis-15262	213	16	reacher	reacher	PROPN
fcis-15262	213	17	of	of	ADP
fcis-15262	213	18	feature	feature	NOUN
fcis-15262	213	19	extratcion	extratcion	NOUN
fcis-15262	213	20	from	from	ADP
fcis-15262	213	21	spectrogram	spectrogram	NOUN
fcis-15262	213	22	based	base	VERB
fcis-15262	213	23	on	on	ADP
fcis-15262	213	24	pulse	pulse	NOUN
fcis-15262	213	25	coupled	couple	VERB
fcis-15262	213	26	neural	neural	ADJ
fcis-15262	213	27	network	network	NOUN
fcis-15262	213	28	in	in	ADP
fcis-15262	213	29	speaker	speaker	NOUN
fcis-15262	213	30	recognition[j	recognition[j	PROPN
fcis-15262	213	31	/	/	SYM
fcis-15262	213	32	ol	ol	PROPN
fcis-15262	213	33	]	]	PUNCT
fcis-15262	213	34	.	.	PUNCT
fcis-15262	214	1	2005(20)[2023	2005(20)[2023	NUM
fcis-15262	214	2	-	-	SYM
fcis-15262	214	3	12	12	NUM
fcis-15262	214	4	-	-	SYM
fcis-15262	214	5	11	11	NUM
fcis-15262	214	6	]	]	PUNCT
fcis-15262	214	7	.	.	PUNCT
fcis-15262	215	1	https://	https://	PROPN
fcis-15262	215	2	ir	ir	PROPN
fcis-15262	215	3	.	.	PROPN
fcis-15262	215	4	lzu	lzu	PROPN
fcis-15262	215	5	.	.	PUNCT
fcis-15262	216	1	edu.cn//handle/262010/127329	edu.cn//handle/262010/127329	PROPN
fcis-15262	216	2	.	.	PUNCT
fcis-15262	217	1	doi:10/127329	doi:10/127329	PROPN
fcis-15262	217	2	.	.	PUNCT
fcis-15262	218	1	[	[	X
fcis-15262	218	2	12	12	NUM
fcis-15262	218	3	]	]	PUNCT
fcis-15262	218	4	cui	cui	NOUN
fcis-15262	218	5	l	l	PROPN
fcis-15262	218	6	,	,	PUNCT
fcis-15262	218	7	wang	wang	PROPN
fcis-15262	218	8	z.	z.	PROPN
fcis-15262	218	9	study	study	PROPN
fcis-15262	218	10	on	on	ADP
fcis-15262	218	11	voiceprint	voiceprint	NOUN
fcis-15262	218	12	recognition	recognition	NOUN
fcis-15262	218	13	based	base	VERB
fcis-15262	218	14	on	on	ADP
fcis-15262	218	15	mixed	mixed	ADJ
fcis-15262	218	16	features	feature	NOUN
fcis-15262	218	17	of	of	ADP
fcis-15262	218	18	lfbank	lfbank	NOUN
fcis-15262	218	19	and	and	CCONJ
fcis-15262	218	20	fbank[j	fbank[j	NOUN
fcis-15262	218	21	]	]	PUNCT
fcis-15262	218	22	.	.	PUNCT
fcis-15262	219	1	computer	computer	NOUN
fcis-15262	219	2	science	science	NOUN
fcis-15262	219	3	,	,	PUNCT
fcis-15262	219	4	2022	2022	NUM
fcis-15262	219	5	,	,	PUNCT
fcis-15262	219	6	49(s2	49(s2	NUM
fcis-15262	219	7	):	):	PUNCT
fcis-15262	219	8	621	621	NUM
fcis-15262	219	9	-	-	SYM
fcis-15262	219	10	625	625	NUM
fcis-15262	219	11	.	.	PUNCT
fcis-15262	220	1	[	[	X
fcis-15262	220	2	13	13	NUM
fcis-15262	220	3	]	]	X
fcis-15262	220	4	chu	chu	PROPN
fcis-15262	220	5	x	x	PROPN
fcis-15262	220	6	,	,	PUNCT
fcis-15262	220	7	yang	yang	PROPN
fcis-15262	220	8	h	h	PROPN
fcis-15262	220	9	,	,	PUNCT
fcis-15262	220	10	yan	yan	PROPN
fcis-15262	220	11	d	d	NOUN
fcis-15262	220	12	,	,	PUNCT
fcis-15262	220	13	et	et	PROPN
fcis-15262	220	14	al	al	PROPN
fcis-15262	220	15	.	.	PROPN
fcis-15262	220	16	research	research	NOUN
fcis-15262	220	17	on	on	ADP
fcis-15262	220	18	cnn	cnn	PROPN
fcis-15262	220	19	-	-	PUNCT
fcis-15262	220	20	lstm	lstm	PROPN
fcis-15262	220	21	speaker	speaker	NOUN
fcis-15262	220	22	recognition	recognition	NOUN
fcis-15262	220	23	based	base	VERB
fcis-15262	220	24	on	on	ADP
fcis-15262	220	25	global	global	ADJ
fcis-15262	220	26	attention	attention	NOUN
fcis-15262	220	27	mechanism	mechanism	NOUN
fcis-15262	221	1	[	[	X
fcis-15262	221	2	j	j	X
fcis-15262	221	3	/	/	SYM
fcis-15262	221	4	ol	ol	PROPN
fcis-15262	221	5	]	]	PUNCT
fcis-15262	221	6	.	.	PUNCT
fcis-15262	222	1	audio	audio	ADJ
fcis-15262	222	2	engineering	engineering	NOUN
fcis-15262	222	3	,	,	PUNCT
fcis-15262	222	4	2022	2022	NUM
fcis-15262	222	5	,	,	PUNCT
fcis-15262	222	6	46(2	46(2	NUM
fcis-15262	222	7	):	):	PUNCT
fcis-15262	222	8	38	38	NUM
fcis-15262	222	9	-	-	SYM
fcis-15262	222	10	45	45	NUM
fcis-15262	222	11	.	.	PUNCT
fcis-15262	223	1	doi:10	doi:10	PROPN
fcis-15262	223	2	.	.	PUNCT
fcis-15262	224	1	16311/	16311/	NUM
fcis-15262	224	2	j.audioe.2022.02.009	j.audioe.2022.02.009	NOUN
fcis-15262	224	3	.	.	PUNCT
fcis-15262	225	1	[	[	X
fcis-15262	225	2	14	14	NUM
fcis-15262	225	3	]	]	X
fcis-15262	225	4	rong	rong	PROPN
fcis-15262	225	5	y	y	PROPN
fcis-15262	225	6	,	,	PUNCT
fcis-15262	225	7	fang	fang	PROPN
fcis-15262	225	8	y	y	PROPN
fcis-15262	225	9	,	,	PUNCT
fcis-15262	225	10	tian	tian	PROPN
fcis-15262	225	11	p	p	X
fcis-15262	225	12	,	,	PUNCT
fcis-15262	225	13	et	et	PROPN
fcis-15262	225	14	al	al	PROPN
fcis-15262	225	15	.	.	PUNCT
fcis-15262	225	16	voiceprint	voiceprint	PROPN
fcis-15262	225	17	recognition	recognition	NOUN
fcis-15262	225	18	based	base	VERB
fcis-15262	225	19	on	on	ADP
fcis-15262	225	20	knowledge	knowledge	NOUN
fcis-15262	225	21	distillation	distillation	NOUN
fcis-15262	225	22	and	and	CCONJ
fcis-15262	225	23	resnet[j	resnet[j	PRON
fcis-15262	225	24	]	]	SYM
fcis-15262	225	25	.	.	PUNCT
fcis-15262	226	1	journal	journal	PROPN
fcis-15262	226	2	of	of	ADP
fcis-15262	226	3	chongqing	chongqing	PROPN
fcis-15262	226	4	university	university	NOUN
fcis-15262	226	5	,	,	PUNCT
fcis-15262	226	6	2023	2023	NUM
fcis-15262	226	7	,	,	PUNCT
fcis-15262	226	8	46(1	46(1	NUM
fcis-15262	226	9	):	):	PUNCT
fcis-15262	226	10	113	113	NUM
fcis-15262	226	11	-	-	SYM
fcis-15262	226	12	124	124	NUM
fcis-15262	226	13	.	.	PUNCT
fcis-15262	227	1	[	[	X
fcis-15262	227	2	15	15	NUM
fcis-15262	227	3	]	]	X
fcis-15262	227	4	wang	wang	PROPN
fcis-15262	227	5	f	f	PROPN
fcis-15262	227	6	,	,	PUNCT
fcis-15262	227	7	xu	xu	PROPN
fcis-15262	227	8	y.	y.	PROPN
fcis-15262	227	9	lstm	lstm	PROPN
fcis-15262	227	10	-	-	PUNCT
fcis-15262	227	11	based	base	VERB
fcis-15262	227	12	end	end	NOUN
fcis-15262	227	13	-	-	PUNCT
fcis-15262	227	14	to	to	ADP
fcis-15262	227	15	-	-	PUNCT
fcis-15262	227	16	end	end	NOUN
fcis-15262	227	17	voiceprint	voiceprint	NOUN
fcis-15262	227	18	recognition	recognition	NOUN
fcis-15262	227	19	algorithm	algorithm	PROPN
fcis-15262	227	20	implementation[j	implementation[j	PROPN
fcis-15262	227	21	/	/	SYM
fcis-15262	227	22	ol	ol	PROPN
fcis-15262	227	23	]	]	PUNCT
fcis-15262	227	24	.	.	PUNCT
fcis-15262	228	1	software	software	NOUN
fcis-15262	228	2	engineering	engineering	NOUN
fcis-15262	228	3	and	and	CCONJ
fcis-15262	228	4	applications	application	NOUN
fcis-15262	228	5	,	,	PUNCT
fcis-15262	228	6	2021	2021	NUM
fcis-15262	228	7	,	,	PUNCT
fcis-15262	228	8	10	10	NUM
fcis-15262	228	9	:	:	SYM
fcis-15262	228	10	467	467	NUM
fcis-15262	228	11	.	.	PUNCT
fcis-15262	228	12	doi:10	doi:10	PROPN
fcis-15262	228	13	.	.	PUNCT
fcis-15262	229	1	12677/	12677/	NUM
fcis-15262	229	2	sea	sea	NOUN
fcis-15262	229	3	.	.	PUNCT
fcis-15262	230	1	2021.104052	2021.104052	NOUN
fcis-15262	230	2	.	.	PUNCT
fcis-15262	231	1	[	[	X
fcis-15262	231	2	16	16	NUM
fcis-15262	231	3	]	]	PUNCT
fcis-15262	231	4	wang	wang	PROPN
fcis-15262	231	5	zhi	zhi	PROPN
fcis-15262	231	6	-	-	PROPN
fcis-15262	231	7	yue	yue	PROPN
fcis-15262	231	8	c	c	PROPN
fcis-15262	231	9	l.	l.	PROPN
fcis-15262	231	10	end	end	VERB
fcis-15262	231	11	to	to	PART
fcis-15262	231	12	end	end	VERB
fcis-15262	231	13	voiceprint	voiceprint	NOUN
fcis-15262	231	14	recgonition	recgonition	NOUN
fcis-15262	231	15	based	base	VERB
fcis-15262	231	16	on	on	ADP
fcis-15262	231	17	nonlinear	nonlinear	ADJ
fcis-15262	231	18	stacked	stack	VERB
fcis-15262	231	19	bidirectional	bidirectional	ADJ
fcis-15262	231	20	network[j	network[j	PROPN
fcis-15262	231	21	]	]	PUNCT
fcis-15262	231	22	.	.	PUNCT
fcis-15262	232	1	computer	computer	NOUN
fcis-15262	232	2	and	and	CCONJ
fcis-15262	232	3	modernization	modernization	NOUN
fcis-15262	232	4	,	,	PUNCT
fcis-15262	232	5	2022	2022	NUM
fcis-15262	232	6	,	,	PUNCT
fcis-15262	232	7	0(03	0(03	NUM
fcis-15262	232	8	):	):	PUNCT
fcis-15262	232	9	13	13	NUM
fcis-15262	232	10	.	.	PUNCT
fcis-15262	233	1	[	[	X
fcis-15262	233	2	17	17	NUM
fcis-15262	233	3	]	]	X
fcis-15262	233	4	zhang	zhang	PROPN
fcis-15262	233	5	y	y	PROPN
fcis-15262	233	6	,	,	PUNCT
fcis-15262	233	7	zhang	zhang	PROPN
fcis-15262	233	8	z.	z.	PROPN
fcis-15262	233	9	application	application	NOUN
fcis-15262	233	10	of	of	ADP
fcis-15262	233	11	densenet	densenet	NOUN
fcis-15262	233	12	in	in	ADP
fcis-15262	233	13	voiceprint	voiceprint	NOUN
fcis-15262	233	14	recognition[j	recognition[j	PROPN
fcis-15262	233	15	]	]	PUNCT
fcis-15262	233	16	.	.	PUNCT
fcis-15262	234	1	computer	computer	NOUN
fcis-15262	234	2	engineering	engineering	PROPN
fcis-15262	234	3	&	&	CCONJ
fcis-15262	234	4	science	science	NOUN
fcis-15262	234	5	,	,	PUNCT
fcis-15262	234	6	2022	2022	NUM
fcis-15262	234	7	,	,	PUNCT
fcis-15262	234	8	44(1	44(1	PROPN
fcis-15262	234	9	):	):	PUNCT
fcis-15262	234	10	132	132	NUM
fcis-15262	234	11	-	-	SYM
fcis-15262	234	12	137	137	NUM
fcis-15262	234	13	.	.	PUNCT
fcis-15262	235	1	[	[	X
fcis-15262	235	2	18	18	NUM
fcis-15262	235	3	]	]	X
fcis-15262	235	4	jiang	jiang	PROPN
fcis-15262	235	5	g	g	PROPN
fcis-15262	235	6	,	,	PUNCT
fcis-15262	235	7	fu	fu	PROPN
fcis-15262	235	8	y.	y.	PROPN
fcis-15262	235	9	simulation	simulation	PROPN
fcis-15262	235	10	of	of	ADP
fcis-15262	235	11	user	user	NOUN
fcis-15262	235	12	login	login	NOUN
fcis-15262	235	13	speech	speech	NOUN
fcis-15262	235	14	recognition	recognition	NOUN
fcis-15262	235	15	model	model	NOUN
fcis-15262	235	16	based	base	VERB
fcis-15262	235	17	on	on	ADP
fcis-15262	235	18	multi	multi	ADJ
fcis-15262	235	19	-	-	ADJ
fcis-15262	235	20	task	task	ADJ
fcis-15262	235	21	training[j	training[j	NOUN
fcis-15262	235	22	]	]	PUNCT
fcis-15262	235	23	.	.	PUNCT
fcis-15262	236	1	computer	computer	NOUN
fcis-15262	236	2	simulation	simulation	NOUN
fcis-15262	236	3	,	,	PUNCT
fcis-15262	236	4	2022	2022	NUM
fcis-15262	236	5	,	,	PUNCT
fcis-15262	236	6	39(9	39(9	NUM
fcis-15262	236	7	):	):	PUNCT
fcis-15262	236	8	190	190	NUM
fcis-15262	236	9	-	-	SYM
fcis-15262	236	10	194	194	NUM
fcis-15262	236	11	.	.	PUNCT
fcis-15262	237	1	[	[	X
fcis-15262	237	2	19	19	NUM
fcis-15262	237	3	]	]	X
fcis-15262	237	4	liu	liu	PROPN
fcis-15262	237	5	x	x	PROPN
fcis-15262	237	6	,	,	PUNCT
fcis-15262	237	7	song	song	PROPN
fcis-15262	237	8	w	w	PROPN
fcis-15262	237	9	,	,	PUNCT
fcis-15262	237	10	chen	chen	PROPN
fcis-15262	237	11	x	x	PROPN
fcis-15262	237	12	,	,	PUNCT
fcis-15262	237	13	et	et	PROPN
fcis-15262	237	14	al	al	PROPN
fcis-15262	237	15	.	.	PUNCT
fcis-15262	237	16	blstm	blstm	PROPN
fcis-15262	237	17	-	-	PUNCT
fcis-15262	237	18	ctc	ctc	PROPN
fcis-15262	237	19	speech	speech	NOUN
fcis-15262	237	20	recognition	recognition	NOUN
fcis-15262	237	21	based	base	VERB
fcis-15262	237	22	on	on	ADP
fcis-15262	237	23	multi	multi	ADJ
fcis-15262	237	24	-	-	ADJ
fcis-15262	237	25	core	core	ADJ
fcis-15262	237	26	convolutional	convolutional	ADJ
fcis-15262	237	27	fusion	fusion	NOUN
fcis-15262	237	28	network[j	network[j	NOUN
fcis-15262	237	29	]	]	PUNCT
fcis-15262	237	30	.	.	PUNCT
fcis-15262	238	1	computer	computer	NOUN
fcis-15262	238	2	applications	application	NOUN
fcis-15262	238	3	and	and	CCONJ
fcis-15262	238	4	software	software	NOUN
fcis-15262	238	5	,	,	PUNCT
fcis-15262	238	6	2021	2021	NUM
fcis-15262	238	7	,	,	PUNCT
fcis-15262	238	8	38(11	38(11	NUM
fcis-15262	238	9	):	):	PUNCT
fcis-15262	238	10	167	167	NUM
fcis-15262	238	11	-	-	SYM
fcis-15262	238	12	173	173	NUM
fcis-15262	238	13	.	.	PUNCT