id	sid	tid	token	lemma	pos
ajst-21816	1	1	academic	academic	ADJ
ajst-21816	1	2	journal	journal	NOUN
ajst-21816	1	3	of	of	ADP
ajst-21816	1	4	science	science	NOUN
ajst-21816	1	5	and	and	CCONJ
ajst-21816	1	6	technology	technology	NOUN
ajst-21816	1	7	issn	issn	NOUN
ajst-21816	1	8	:	:	PUNCT
ajst-21816	1	9	2771	2771	NUM
ajst-21816	1	10	-	-	SYM
ajst-21816	1	11	3032	3032	NUM
ajst-21816	1	12	|	|	NOUN
ajst-21816	1	13	vol	vol	NOUN
ajst-21816	1	14	.	.	PROPN
ajst-21816	2	1	11	11	NUM
ajst-21816	2	2	,	,	PUNCT
ajst-21816	2	3	no	no	INTJ
ajst-21816	2	4	.	.	NOUN
ajst-21816	2	5	1	1	NUM
ajst-21816	2	6	,	,	PUNCT
ajst-21816	2	7	2024	2024	NUM
ajst-21816	2	8	286	286	NUM
ajst-21816	2	9	research	research	NOUN
ajst-21816	2	10	on	on	ADP
ajst-21816	2	11	gait	gait	ADJ
ajst-21816	2	12	recognition	recognition	NOUN
ajst-21816	2	13	method	method	NOUN
ajst-21816	2	14	based	base	VERB
ajst-21816	2	15	on	on	ADP
ajst-21816	2	16	lfe‐	lfe‐	PROPN
ajst-21816	2	17	mgpa	mgpa	NOUN
ajst-21816	2	18	network	network	NOUN
ajst-21816	2	19	meiqi	meiqi	PROPN
ajst-21816	2	20	zhao	zhao	PROPN
ajst-21816	2	21	,	,	PUNCT
ajst-21816	2	22	hua	hua	PROPN
ajst-21816	2	23	huo	huo	PROPN
ajst-21816	2	24	school	school	NOUN
ajst-21816	2	25	of	of	ADP
ajst-21816	2	26	information	information	NOUN
ajst-21816	2	27	engineering	engineering	PROPN
ajst-21816	2	28	,	,	PUNCT
ajst-21816	2	29	henan	henan	PROPN
ajst-21816	2	30	university	university	PROPN
ajst-21816	2	31	of	of	ADP
ajst-21816	2	32	science	science	NOUN
ajst-21816	2	33	and	and	CCONJ
ajst-21816	2	34	technology	technology	NOUN
ajst-21816	2	35	,	,	PUNCT
ajst-21816	2	36	china	china	PROPN
ajst-21816	2	37	abstract	abstract	PROPN
ajst-21816	2	38	:	:	PUNCT
ajst-21816	2	39	as	as	ADP
ajst-21816	2	40	a	a	DET
ajst-21816	2	41	non	non	ADJ
ajst-21816	2	42	-	-	ADJ
ajst-21816	2	43	contact	contact	ADJ
ajst-21816	2	44	long	long	ADJ
ajst-21816	2	45	-	-	PUNCT
ajst-21816	2	46	distance	distance	NOUN
ajst-21816	2	47	biometric	biometric	ADJ
ajst-21816	2	48	recognition	recognition	NOUN
ajst-21816	2	49	technology	technology	NOUN
ajst-21816	2	50	,	,	PUNCT
ajst-21816	2	51	gait	gait	ADJ
ajst-21816	2	52	recognition	recognition	NOUN
ajst-21816	2	53	is	be	AUX
ajst-21816	2	54	tasked	task	VERB
ajst-21816	2	55	to	to	PART
ajst-21816	2	56	realize	realize	VERB
ajst-21816	2	57	the	the	DET
ajst-21816	2	58	identification	identification	NOUN
ajst-21816	2	59	of	of	ADP
ajst-21816	2	60	long	long	ADJ
ajst-21816	2	61	-	-	PUNCT
ajst-21816	2	62	distance	distance	NOUN
ajst-21816	2	63	pedestrians	pedestrian	NOUN
ajst-21816	2	64	according	accord	VERB
ajst-21816	2	65	to	to	ADP
ajst-21816	2	66	their	their	PRON
ajst-21816	2	67	walking	walk	VERB
ajst-21816	2	68	patterns	pattern	NOUN
ajst-21816	2	69	.	.	PUNCT
ajst-21816	3	1	however	however	ADV
ajst-21816	3	2	,	,	PUNCT
ajst-21816	3	3	gait	gait	ADJ
ajst-21816	3	4	recognition	recognition	NOUN
ajst-21816	3	5	is	be	AUX
ajst-21816	3	6	greatly	greatly	ADV
ajst-21816	3	7	affected	affect	VERB
ajst-21816	3	8	by	by	ADP
ajst-21816	3	9	external	external	ADJ
ajst-21816	3	10	factors	factor	NOUN
ajst-21816	3	11	such	such	ADJ
ajst-21816	3	12	as	as	ADP
ajst-21816	3	13	what	what	PRON
ajst-21816	3	14	the	the	DET
ajst-21816	3	15	subject	subject	NOUN
ajst-21816	3	16	is	be	AUX
ajst-21816	3	17	wearing	wear	VERB
ajst-21816	3	18	and	and	CCONJ
ajst-21816	3	19	carrying	carry	VERB
ajst-21816	3	20	.	.	PUNCT
ajst-21816	4	1	to	to	PART
ajst-21816	4	2	solve	solve	VERB
ajst-21816	4	3	this	this	DET
ajst-21816	4	4	problem	problem	NOUN
ajst-21816	4	5	,	,	PUNCT
ajst-21816	4	6	this	this	DET
ajst-21816	4	7	paper	paper	NOUN
ajst-21816	4	8	focuses	focus	VERB
ajst-21816	4	9	on	on	ADP
ajst-21816	4	10	gait	gait	ADJ
ajst-21816	4	11	feature	feature	NOUN
ajst-21816	4	12	extraction	extraction	NOUN
ajst-21816	4	13	and	and	CCONJ
ajst-21816	4	14	proposes	propose	VERB
ajst-21816	4	15	lfe	lfe	PROPN
ajst-21816	4	16	-	-	PUNCT
ajst-21816	4	17	mgpa	mgpa	PROPN
ajst-21816	4	18	network	network	NOUN
ajst-21816	4	19	based	base	VERB
ajst-21816	4	20	on	on	ADP
ajst-21816	4	21	convolutional	convolutional	ADJ
ajst-21816	4	22	neural	neural	ADJ
ajst-21816	4	23	network	network	NOUN
ajst-21816	4	24	(	(	PUNCT
ajst-21816	4	25	cnn	cnn	PROPN
ajst-21816	4	26	)	)	PUNCT
ajst-21816	4	27	.	.	PUNCT
ajst-21816	5	1	a	a	DET
ajst-21816	5	2	local	local	ADJ
ajst-21816	5	3	feature	feature	NOUN
ajst-21816	5	4	extractor	extractor	NOUN
ajst-21816	5	5	(	(	PUNCT
ajst-21816	5	6	lfe	lfe	PROPN
ajst-21816	5	7	)	)	PUNCT
ajst-21816	5	8	was	be	AUX
ajst-21816	5	9	constructed	construct	VERB
ajst-21816	5	10	to	to	PART
ajst-21816	5	11	extract	extract	VERB
ajst-21816	5	12	fine	fine	ADV
ajst-21816	5	13	-	-	PUNCT
ajst-21816	5	14	grained	grain	VERB
ajst-21816	5	15	features	feature	NOUN
ajst-21816	5	16	of	of	ADP
ajst-21816	5	17	gait	gait	NOUN
ajst-21816	5	18	through	through	ADP
ajst-21816	5	19	sconv	sconv	NOUN
ajst-21816	5	20	,	,	PUNCT
ajst-21816	5	21	and	and	CCONJ
ajst-21816	5	22	mgpa	mgpa	NOUN
ajst-21816	5	23	module	module	NOUN
ajst-21816	5	24	was	be	AUX
ajst-21816	5	25	integrated	integrate	VERB
ajst-21816	5	26	to	to	PART
ajst-21816	5	27	extract	extract	VERB
ajst-21816	5	28	collection	collection	NOUN
ajst-21816	5	29	-	-	PUNCT
ajst-21816	5	30	level	level	NOUN
ajst-21816	5	31	information	information	NOUN
ajst-21816	5	32	of	of	ADP
ajst-21816	5	33	different	different	ADJ
ajst-21816	5	34	layers	layer	NOUN
ajst-21816	5	35	.	.	PUNCT
ajst-21816	6	1	in	in	ADP
ajst-21816	6	2	order	order	NOUN
ajst-21816	6	3	to	to	PART
ajst-21816	6	4	improve	improve	VERB
ajst-21816	6	5	the	the	DET
ajst-21816	6	6	recognition	recognition	NOUN
ajst-21816	6	7	accuracy	accuracy	NOUN
ajst-21816	6	8	of	of	ADP
ajst-21816	6	9	the	the	DET
ajst-21816	6	10	network	network	NOUN
ajst-21816	6	11	model	model	NOUN
ajst-21816	6	12	,	,	PUNCT
ajst-21816	6	13	a	a	DET
ajst-21816	6	14	joint	joint	ADJ
ajst-21816	6	15	loss	loss	NOUN
ajst-21816	6	16	function	function	NOUN
ajst-21816	6	17	is	be	AUX
ajst-21816	6	18	constructed	construct	VERB
ajst-21816	6	19	.	.	PUNCT
ajst-21816	7	1	the	the	DET
ajst-21816	7	2	experiment	experiment	NOUN
ajst-21816	7	3	was	be	AUX
ajst-21816	7	4	carried	carry	VERB
ajst-21816	7	5	out	out	ADP
ajst-21816	7	6	on	on	ADP
ajst-21816	7	7	the	the	DET
ajst-21816	7	8	open	open	ADJ
ajst-21816	7	9	gait	gait	PROPN
ajst-21816	7	10	dataset	dataset	PROPN
ajst-21816	7	11	casia	casia	PROPN
ajst-21816	7	12	-	-	PUNCT
ajst-21816	7	13	b	b	NOUN
ajst-21816	7	14	,	,	PUNCT
ajst-21816	7	15	and	and	CCONJ
ajst-21816	7	16	the	the	DET
ajst-21816	7	17	recognition	recognition	NOUN
ajst-21816	7	18	accuracy	accuracy	NOUN
ajst-21816	7	19	reached	reach	VERB
ajst-21816	7	20	97.53	97.53	NUM
ajst-21816	7	21	%	%	NOUN
ajst-21816	7	22	,	,	PUNCT
ajst-21816	7	23	94.5	94.5	NUM
ajst-21816	7	24	%	%	NOUN
ajst-21816	7	25	and	and	CCONJ
ajst-21816	7	26	81.0	81.0	NUM
ajst-21816	7	27	%	%	NOUN
ajst-21816	7	28	respectively	respectively	ADV
ajst-21816	7	29	under	under	ADP
ajst-21816	7	30	the	the	DET
ajst-21816	7	31	three	three	NUM
ajst-21816	7	32	walking	walk	VERB
ajst-21816	7	33	states	state	NOUN
ajst-21816	7	34	of	of	ADP
ajst-21816	7	35	normal	normal	ADJ
ajst-21816	7	36	walking	walking	NOUN
ajst-21816	7	37	,	,	PUNCT
ajst-21816	7	38	backpack	backpack	NOUN
ajst-21816	7	39	and	and	CCONJ
ajst-21816	7	40	overcoat	overcoat	NOUN
ajst-21816	7	41	,	,	PUNCT
ajst-21816	7	42	which	which	PRON
ajst-21816	7	43	is	be	AUX
ajst-21816	7	44	higher	high	ADJ
ajst-21816	7	45	than	than	ADP
ajst-21816	7	46	the	the	DET
ajst-21816	7	47	current	current	ADJ
ajst-21816	7	48	mainstream	mainstream	NOUN
ajst-21816	7	49	algorithms	algorithm	NOUN
ajst-21816	7	50	.	.	PUNCT
ajst-21816	8	1	in	in	ADP
ajst-21816	8	2	particular	particular	ADJ
ajst-21816	8	3	,	,	PUNCT
ajst-21816	8	4	the	the	DET
ajst-21816	8	5	recognition	recognition	NOUN
ajst-21816	8	6	accuracy	accuracy	NOUN
ajst-21816	8	7	is	be	AUX
ajst-21816	8	8	significantly	significantly	ADV
ajst-21816	8	9	improved	improve	VERB
ajst-21816	8	10	in	in	ADP
ajst-21816	8	11	the	the	DET
ajst-21816	8	12	walking	walk	VERB
ajst-21816	8	13	state	state	NOUN
ajst-21816	8	14	of	of	ADP
ajst-21816	8	15	backpack	backpack	NOUN
ajst-21816	8	16	and	and	CCONJ
ajst-21816	8	17	overcoat	overcoat	NOUN
ajst-21816	8	18	,	,	PUNCT
ajst-21816	8	19	which	which	PRON
ajst-21816	8	20	proves	prove	VERB
ajst-21816	8	21	that	that	SCONJ
ajst-21816	8	22	the	the	DET
ajst-21816	8	23	method	method	NOUN
ajst-21816	8	24	has	have	VERB
ajst-21816	8	25	strong	strong	ADJ
ajst-21816	8	26	robustness	robustness	NOUN
ajst-21816	8	27	.	.	PUNCT
ajst-21816	9	1	keywords	keyword	NOUN
ajst-21816	9	2	:	:	PUNCT
ajst-21816	9	3	gait	gait	ADJ
ajst-21816	9	4	recognition	recognition	NOUN
ajst-21816	9	5	;	;	PUNCT
ajst-21816	9	6	convolutional	convolutional	ADJ
ajst-21816	9	7	neural	neural	ADJ
ajst-21816	9	8	network	network	NOUN
ajst-21816	9	9	;	;	PUNCT
ajst-21816	9	10	feature	feature	NOUN
ajst-21816	9	11	extraction	extraction	NOUN
ajst-21816	9	12	;	;	PUNCT
ajst-21816	9	13	gait	gait	VERB
ajst-21816	9	14	contour	contour	NOUN
ajst-21816	9	15	diagram	diagram	NOUN
ajst-21816	9	16	.	.	PUNCT
ajst-21816	10	1	1	1	X
ajst-21816	10	2	.	.	X
ajst-21816	10	3	introduction	introduction	NOUN
ajst-21816	10	4	in	in	ADP
ajst-21816	10	5	recent	recent	ADJ
ajst-21816	10	6	years	year	NOUN
ajst-21816	10	7	,	,	PUNCT
ajst-21816	10	8	gait	gait	PROPN
ajst-21816	10	9	recognition	recognition	NOUN
ajst-21816	10	10	research	research	NOUN
ajst-21816	10	11	has	have	AUX
ajst-21816	10	12	gradually	gradually	ADV
ajst-21816	10	13	become	become	VERB
ajst-21816	10	14	a	a	DET
ajst-21816	10	15	popular	popular	ADJ
ajst-21816	10	16	research	research	NOUN
ajst-21816	10	17	direction	direction	NOUN
ajst-21816	10	18	in	in	ADP
ajst-21816	10	19	the	the	DET
ajst-21816	10	20	field	field	NOUN
ajst-21816	10	21	of	of	ADP
ajst-21816	10	22	computer	computer	NOUN
ajst-21816	10	23	vision	vision	NOUN
ajst-21816	10	24	and	and	CCONJ
ajst-21816	10	25	biometric	biometric	ADJ
ajst-21816	10	26	recognition	recognition	NOUN
ajst-21816	10	27	.	.	PUNCT
ajst-21816	11	1	compared	compare	VERB
ajst-21816	11	2	with	with	ADP
ajst-21816	11	3	general	general	ADJ
ajst-21816	11	4	biometric	biometric	ADJ
ajst-21816	11	5	recognition	recognition	NOUN
ajst-21816	11	6	technology	technology	NOUN
ajst-21816	11	7	(	(	PUNCT
ajst-21816	11	8	such	such	ADJ
ajst-21816	11	9	as	as	ADP
ajst-21816	11	10	fingerprints[1	fingerprints[1	PROPN
ajst-21816	11	11	]	]	X
ajst-21816	11	12	,	,	PUNCT
ajst-21816	11	13	faces[2,3	faces[2,3	NOUN
ajst-21816	11	14	]	]	X
ajst-21816	11	15	,	,	PUNCT
ajst-21816	11	16	irises[4	irises[4	PROPN
ajst-21816	11	17	]	]	PUNCT
ajst-21816	11	18	,	,	PUNCT
ajst-21816	11	19	etc	etc	X
ajst-21816	11	20	)	)	PUNCT
ajst-21816	11	21	,	,	PUNCT
ajst-21816	11	22	gait	gait	PROPN
ajst-21816	11	23	recognition	recognition	NOUN
ajst-21816	11	24	has	have	VERB
ajst-21816	11	25	the	the	DET
ajst-21816	11	26	advantages	advantage	NOUN
ajst-21816	11	27	of	of	ADP
ajst-21816	11	28	no	no	DET
ajst-21816	11	29	need	need	NOUN
ajst-21816	11	30	for	for	ADP
ajst-21816	11	31	subject	subject	ADJ
ajst-21816	11	32	cooperation	cooperation	NOUN
ajst-21816	11	33	,	,	PUNCT
ajst-21816	11	34	easy	easy	ADJ
ajst-21816	11	35	acquisition	acquisition	NOUN
ajst-21816	11	36	,	,	PUNCT
ajst-21816	11	37	non	non	ADJ
ajst-21816	11	38	-	-	NOUN
ajst-21816	11	39	contact	contact	ADJ
ajst-21816	11	40	,	,	PUNCT
ajst-21816	11	41	difficult	difficult	ADJ
ajst-21816	11	42	to	to	PART
ajst-21816	11	43	camouflage	camouflage	VERB
ajst-21816	11	44	,	,	PUNCT
ajst-21816	11	45	and	and	CCONJ
ajst-21816	11	46	low	low	ADJ
ajst-21816	11	47	image	image	NOUN
ajst-21816	11	48	resolution	resolution	NOUN
ajst-21816	11	49	requirements	requirement	NOUN
ajst-21816	11	50	.	.	PUNCT
ajst-21816	12	1	the	the	DET
ajst-21816	12	2	above	above	ADJ
ajst-21816	12	3	characteristics	characteristic	NOUN
ajst-21816	12	4	mean	mean	VERB
ajst-21816	12	5	that	that	SCONJ
ajst-21816	12	6	gait	gait	ADJ
ajst-21816	12	7	recognition	recognition	NOUN
ajst-21816	12	8	technology	technology	NOUN
ajst-21816	12	9	has	have	VERB
ajst-21816	12	10	a	a	DET
ajst-21816	12	11	wide	wide	ADJ
ajst-21816	12	12	range	range	NOUN
ajst-21816	12	13	of	of	ADP
ajst-21816	12	14	application	application	NOUN
ajst-21816	12	15	prospects	prospect	NOUN
ajst-21816	12	16	and	and	CCONJ
ajst-21816	12	17	economic	economic	ADJ
ajst-21816	12	18	value	value	NOUN
ajst-21816	12	19	in	in	ADP
ajst-21816	12	20	social	social	ADJ
ajst-21816	12	21	security	security	NOUN
ajst-21816	12	22	and	and	CCONJ
ajst-21816	12	23	medical	medical	ADJ
ajst-21816	12	24	diagnosis	diagnosis	NOUN
ajst-21816	12	25	.	.	PUNCT
ajst-21816	13	1	however	however	ADV
ajst-21816	13	2	,	,	PUNCT
ajst-21816	13	3	as	as	ADP
ajst-21816	13	4	the	the	DET
ajst-21816	13	5	basic	basic	ADJ
ajst-21816	13	6	information	information	NOUN
ajst-21816	13	7	of	of	ADP
ajst-21816	13	8	gait	gait	ADJ
ajst-21816	13	9	recognition	recognition	NOUN
ajst-21816	13	10	,	,	PUNCT
ajst-21816	13	11	the	the	DET
ajst-21816	13	12	change	change	NOUN
ajst-21816	13	13	of	of	ADP
ajst-21816	13	14	the	the	DET
ajst-21816	13	15	pedestrian	pedestrian	NOUN
ajst-21816	13	16	's	's	PART
ajst-21816	13	17	posture	posture	NOUN
ajst-21816	13	18	during	during	ADP
ajst-21816	13	19	walking	walking	NOUN
ajst-21816	13	20	is	be	AUX
ajst-21816	13	21	easily	easily	ADV
ajst-21816	13	22	affected	affect	VERB
ajst-21816	13	23	by	by	ADP
ajst-21816	13	24	external	external	ADJ
ajst-21816	13	25	factors	factor	NOUN
ajst-21816	13	26	,	,	PUNCT
ajst-21816	13	27	such	such	ADJ
ajst-21816	13	28	as	as	ADP
ajst-21816	13	29	the	the	DET
ajst-21816	13	30	walking	walk	VERB
ajst-21816	13	31	speed	speed	NOUN
ajst-21816	13	32	of	of	ADP
ajst-21816	13	33	the	the	DET
ajst-21816	13	34	pedestrian	pedestrian	NOUN
ajst-21816	13	35	,	,	PUNCT
ajst-21816	13	36	clothing	clothing	NOUN
ajst-21816	13	37	,	,	PUNCT
ajst-21816	13	38	carrying	carry	VERB
ajst-21816	13	39	items	item	NOUN
ajst-21816	13	40	,	,	PUNCT
ajst-21816	13	41	and	and	CCONJ
ajst-21816	13	42	the	the	DET
ajst-21816	13	43	camera	camera	NOUN
ajst-21816	13	44	's	's	PART
ajst-21816	13	45	angle	angle	NOUN
ajst-21816	13	46	of	of	ADP
ajst-21816	13	47	view	view	NOUN
ajst-21816	13	48	and	and	CCONJ
ajst-21816	13	49	frame	frame	NOUN
ajst-21816	13	50	rate	rate	NOUN
ajst-21816	13	51	.	.	PUNCT
ajst-21816	14	1	these	these	DET
ajst-21816	14	2	factors	factor	NOUN
ajst-21816	14	3	make	make	VERB
ajst-21816	14	4	gait	gait	ADJ
ajst-21816	14	5	recognition	recognition	NOUN
ajst-21816	14	6	very	very	ADV
ajst-21816	14	7	challenging	challenging	ADJ
ajst-21816	14	8	,	,	PUNCT
ajst-21816	14	9	especially	especially	ADV
ajst-21816	14	10	crossperspective	crossperspective	ADJ
ajst-21816	14	11	gait	gait	PROPN
ajst-21816	14	12	recognition	recognition	PROPN
ajst-21816	14	13	,	,	PUNCT
ajst-21816	14	14	that	that	ADV
ajst-21816	14	15	is	is	ADV
ajst-21816	14	16	,	,	PUNCT
ajst-21816	14	17	recognizing	recognize	VERB
ajst-21816	14	18	gait	gait	ADJ
ajst-21816	14	19	information	information	NOUN
ajst-21816	14	20	obtained	obtain	VERB
ajst-21816	14	21	from	from	ADP
ajst-21816	14	22	different	different	ADJ
ajst-21816	14	23	angles	angle	NOUN
ajst-21816	14	24	.	.	PUNCT
ajst-21816	15	1	in	in	ADP
ajst-21816	15	2	the	the	DET
ajst-21816	15	3	early	early	ADJ
ajst-21816	15	4	stage	stage	NOUN
ajst-21816	15	5	of	of	ADP
ajst-21816	15	6	gait	gait	ADJ
ajst-21816	15	7	recognition	recognition	NOUN
ajst-21816	15	8	research	research	NOUN
ajst-21816	15	9	,	,	PUNCT
ajst-21816	15	10	static	static	ADJ
ajst-21816	15	11	or	or	CCONJ
ajst-21816	15	12	dynamic	dynamic	ADJ
ajst-21816	15	13	features	feature	NOUN
ajst-21816	15	14	related	relate	VERB
ajst-21816	15	15	to	to	ADP
ajst-21816	15	16	gait	gait	NOUN
ajst-21816	15	17	were	be	AUX
ajst-21816	15	18	manually	manually	ADV
ajst-21816	15	19	extracted	extract	VERB
ajst-21816	15	20	from	from	ADP
ajst-21816	15	21	contour	contour	NOUN
ajst-21816	15	22	sequence	sequence	NOUN
ajst-21816	15	23	,	,	PUNCT
ajst-21816	15	24	and	and	CCONJ
ajst-21816	15	25	features	feature	NOUN
ajst-21816	15	26	were	be	AUX
ajst-21816	15	27	dimensionality	dimensionality	NOUN
ajst-21816	15	28	reduced	reduce	VERB
ajst-21816	15	29	or	or	CCONJ
ajst-21816	15	30	matched	match	VERB
ajst-21816	15	31	by	by	ADP
ajst-21816	15	32	machine	machine	NOUN
ajst-21816	15	33	learning	learning	NOUN
ajst-21816	15	34	.	.	PUNCT
ajst-21816	16	1	alternatively	alternatively	ADV
ajst-21816	16	2	,	,	PUNCT
ajst-21816	16	3	a	a	DET
ajst-21816	16	4	template	template	NOUN
ajst-21816	16	5	is	be	AUX
ajst-21816	16	6	constructed	construct	VERB
ajst-21816	16	7	to	to	PART
ajst-21816	16	8	maintain	maintain	VERB
ajst-21816	16	9	the	the	DET
ajst-21816	16	10	temporal	temporal	ADJ
ajst-21816	16	11	and	and	CCONJ
ajst-21816	16	12	spatial	spatial	ADJ
ajst-21816	16	13	information	information	NOUN
ajst-21816	16	14	in	in	ADP
ajst-21816	16	15	the	the	DET
ajst-21816	16	16	gait	gait	ADJ
ajst-21816	16	17	sequence	sequence	NOUN
ajst-21816	16	18	directly	directly	ADV
ajst-21816	16	19	through	through	ADP
ajst-21816	16	20	the	the	DET
ajst-21816	16	21	gait	gait	ADJ
ajst-21816	16	22	contour	contour	NOUN
ajst-21816	16	23	diagram	diagram	NOUN
ajst-21816	16	24	,	,	PUNCT
ajst-21816	16	25	and	and	CCONJ
ajst-21816	16	26	the	the	DET
ajst-21816	16	27	discriminative	discriminative	NOUN
ajst-21816	16	28	representation	representation	NOUN
ajst-21816	16	29	is	be	AUX
ajst-21816	16	30	learned	learn	VERB
ajst-21816	16	31	by	by	ADP
ajst-21816	16	32	machine	machine	NOUN
ajst-21816	16	33	learning	learning	NOUN
ajst-21816	16	34	.	.	PUNCT
ajst-21816	17	1	han[5]et	han[5]et	PROPN
ajst-21816	17	2	al	al	PROPN
ajst-21816	17	3	.	.	PROPN
ajst-21816	17	4	proposed	propose	VERB
ajst-21816	17	5	a	a	DET
ajst-21816	17	6	spatial	spatial	ADJ
ajst-21816	17	7	-	-	PUNCT
ajst-21816	17	8	temporal	temporal	ADJ
ajst-21816	17	9	gait	gait	NOUN
ajst-21816	17	10	representation	representation	NOUN
ajst-21816	17	11	called	call	VERB
ajst-21816	17	12	gait	gait	PROPN
ajst-21816	17	13	energy	energy	NOUN
ajst-21816	17	14	diagram	diagram	NOUN
ajst-21816	17	15	(	(	PUNCT
ajst-21816	17	16	gei	gei	PROPN
ajst-21816	17	17	)	)	PUNCT
ajst-21816	17	18	,	,	PUNCT
ajst-21816	17	19	which	which	PRON
ajst-21816	17	20	has	have	VERB
ajst-21816	17	21	low	low	ADJ
ajst-21816	17	22	computational	computational	ADJ
ajst-21816	17	23	complexity	complexity	NOUN
ajst-21816	17	24	and	and	CCONJ
ajst-21816	17	25	can	can	AUX
ajst-21816	17	26	effectively	effectively	ADV
ajst-21816	17	27	maintain	maintain	VERB
ajst-21816	17	28	the	the	DET
ajst-21816	17	29	spatial	spatial	ADJ
ajst-21816	17	30	information	information	NOUN
ajst-21816	17	31	in	in	ADP
ajst-21816	17	32	the	the	DET
ajst-21816	17	33	gait	gait	ADJ
ajst-21816	17	34	sequence	sequence	NOUN
ajst-21816	17	35	,	,	PUNCT
ajst-21816	17	36	but	but	CCONJ
ajst-21816	17	37	loses	lose	VERB
ajst-21816	17	38	the	the	DET
ajst-21816	17	39	temporal	temporal	ADJ
ajst-21816	17	40	information	information	NOUN
ajst-21816	17	41	.	.	PUNCT
ajst-21816	18	1	wu[6	wu[6	X
ajst-21816	18	2	]	]	PUNCT
ajst-21816	18	3	et	et	PROPN
ajst-21816	18	4	al	al	PROPN
ajst-21816	18	5	.	.	PROPN
ajst-21816	18	6	proposed	propose	VERB
ajst-21816	18	7	a	a	DET
ajst-21816	18	8	gait	gait	ADJ
ajst-21816	18	9	recognition	recognition	NOUN
ajst-21816	18	10	method	method	NOUN
ajst-21816	18	11	based	base	VERB
ajst-21816	18	12	on	on	ADP
ajst-21816	18	13	deep	deep	ADJ
ajst-21816	18	14	convolutional	convolutional	ADJ
ajst-21816	18	15	neural	neural	ADJ
ajst-21816	18	16	networks	network	NOUN
ajst-21816	18	17	.	.	PUNCT
ajst-21816	19	1	in	in	ADP
ajst-21816	19	2	this	this	DET
ajst-21816	19	3	method	method	NOUN
ajst-21816	19	4	,	,	PUNCT
ajst-21816	19	5	three	three	NUM
ajst-21816	19	6	different	different	ADJ
ajst-21816	19	7	network	network	NOUN
ajst-21816	19	8	structures	structure	NOUN
ajst-21816	19	9	with	with	ADP
ajst-21816	19	10	gei	gei	PROPN
ajst-21816	19	11	as	as	ADP
ajst-21816	19	12	input	input	NOUN
ajst-21816	19	13	were	be	AUX
ajst-21816	19	14	designed	design	VERB
ajst-21816	19	15	,	,	PUNCT
ajst-21816	19	16	and	and	CCONJ
ajst-21816	19	17	finally	finally	ADV
ajst-21816	19	18	these	these	DET
ajst-21816	19	19	networks	network	NOUN
ajst-21816	19	20	were	be	AUX
ajst-21816	19	21	fused	fuse	VERB
ajst-21816	19	22	to	to	PART
ajst-21816	19	23	obtain	obtain	VERB
ajst-21816	19	24	recognition	recognition	NOUN
ajst-21816	19	25	results	result	NOUN
ajst-21816	19	26	.	.	PUNCT
ajst-21816	20	1	compared	compare	VERB
ajst-21816	20	2	with	with	ADP
ajst-21816	20	3	the	the	DET
ajst-21816	20	4	existing	exist	VERB
ajst-21816	20	5	methods	method	NOUN
ajst-21816	20	6	,	,	PUNCT
ajst-21816	20	7	this	this	DET
ajst-21816	20	8	method	method	NOUN
ajst-21816	20	9	has	have	VERB
ajst-21816	20	10	obvious	obvious	ADJ
ajst-21816	20	11	advantages	advantage	NOUN
ajst-21816	20	12	when	when	SCONJ
ajst-21816	20	13	the	the	DET
ajst-21816	20	14	angle	angle	NOUN
ajst-21816	20	15	of	of	ADP
ajst-21816	20	16	view	view	NOUN
ajst-21816	20	17	changes	change	VERB
ajst-21816	20	18	greatly	greatly	ADV
ajst-21816	20	19	.	.	PUNCT
ajst-21816	21	1	in	in	ADP
ajst-21816	21	2	order	order	NOUN
ajst-21816	21	3	to	to	PART
ajst-21816	21	4	solve	solve	VERB
ajst-21816	21	5	this	this	DET
ajst-21816	21	6	problem	problem	NOUN
ajst-21816	21	7	,	,	PUNCT
ajst-21816	21	8	wang[7]et	wang[7]et	PROPN
ajst-21816	21	9	al	al	PROPN
ajst-21816	21	10	proposed	propose	VERB
ajst-21816	21	11	a	a	DET
ajst-21816	21	12	timing	timing	NOUN
ajst-21816	21	13	template	template	NOUN
ajst-21816	21	14	called	call	VERB
ajst-21816	21	15	chrono	chrono	NOUN
ajst-21816	21	16	-	-	PUNCT
ajst-21816	21	17	gait	gait	NOUN
ajst-21816	21	18	image(cgi	image(cgi	PROPN
ajst-21816	21	19	)	)	PUNCT
ajst-21816	21	20	,	,	PUNCT
ajst-21816	21	21	which	which	PRON
ajst-21816	21	22	can	can	AUX
ajst-21816	21	23	effectively	effectively	ADV
ajst-21816	21	24	maintain	maintain	VERB
ajst-21816	21	25	the	the	DET
ajst-21816	21	26	timing	timing	NOUN
ajst-21816	21	27	information	information	NOUN
ajst-21816	21	28	in	in	ADP
ajst-21816	21	29	gait	gait	NOUN
ajst-21816	21	30	.	.	PUNCT
ajst-21816	22	1	at	at	ADP
ajst-21816	22	2	present	present	ADJ
ajst-21816	22	3	,	,	PUNCT
ajst-21816	22	4	most	most	ADJ
ajst-21816	22	5	contour	contour	NOUN
ajst-21816	22	6	-	-	PUNCT
ajst-21816	22	7	based	base	VERB
ajst-21816	22	8	gait	gait	NOUN
ajst-21816	22	9	recognition	recognition	NOUN
ajst-21816	22	10	methods	method	NOUN
ajst-21816	22	11	are	be	AUX
ajst-21816	22	12	based	base	VERB
ajst-21816	22	13	on	on	ADP
ajst-21816	22	14	deep	deep	ADJ
ajst-21816	22	15	learning[8	learning[8	NOUN
ajst-21816	22	16	]	]	PUNCT
ajst-21816	22	17	,	,	PUNCT
ajst-21816	22	18	among	among	ADP
ajst-21816	22	19	which	which	PRON
ajst-21816	22	20	convolutional	convolutional	ADJ
ajst-21816	22	21	neural	neural	ADJ
ajst-21816	22	22	network	network	NOUN
ajst-21816	22	23	has	have	AUX
ajst-21816	22	24	received	receive	VERB
ajst-21816	22	25	great	great	ADJ
ajst-21816	22	26	attention	attention	NOUN
ajst-21816	22	27	.	.	PUNCT
ajst-21816	23	1	shiraga[9	shiraga[9	PUNCT
ajst-21816	23	2	]	]	X
ajst-21816	23	3	et	et	PROPN
ajst-21816	23	4	al	al	PROPN
ajst-21816	23	5	.	.	PROPN
ajst-21816	23	6	proposed	propose	VERB
ajst-21816	23	7	geinet	geinet	NOUN
ajst-21816	23	8	,	,	PUNCT
ajst-21816	23	9	which	which	PRON
ajst-21816	23	10	takes	take	VERB
ajst-21816	23	11	gei	gei	PROPN
ajst-21816	23	12	as	as	ADP
ajst-21816	23	13	the	the	DET
ajst-21816	23	14	input	input	NOUN
ajst-21816	23	15	of	of	ADP
ajst-21816	23	16	cnn	cnn	PROPN
ajst-21816	23	17	network	network	NOUN
ajst-21816	23	18	,	,	PUNCT
ajst-21816	23	19	and	and	CCONJ
ajst-21816	23	20	completes	complete	VERB
ajst-21816	23	21	identification	identification	NOUN
ajst-21816	23	22	through	through	ADP
ajst-21816	23	23	two	two	NUM
ajst-21816	23	24	continuous	continuous	ADJ
ajst-21816	23	25	convolution	convolution	NOUN
ajst-21816	23	26	layers	layer	NOUN
ajst-21816	23	27	,	,	PUNCT
ajst-21816	23	28	pooling	pool	VERB
ajst-21816	23	29	layers	layer	NOUN
ajst-21816	23	30	,	,	PUNCT
ajst-21816	23	31	normalization	normalization	NOUN
ajst-21816	23	32	layers	layer	NOUN
ajst-21816	23	33	,	,	PUNCT
ajst-21816	23	34	and	and	CCONJ
ajst-21816	23	35	finally	finally	ADV
ajst-21816	23	36	two	two	NUM
ajst-21816	23	37	fully	fully	ADV
ajst-21816	23	38	connected	connected	ADJ
ajst-21816	23	39	layers	layer	NOUN
ajst-21816	23	40	.	.	PUNCT
ajst-21816	24	1	fan[10	fan[10	NOUN
ajst-21816	24	2	]	]	PUNCT
ajst-21816	24	3	et	et	PROPN
ajst-21816	24	4	al	al	PROPN
ajst-21816	24	5	.	.	PROPN
ajst-21816	24	6	proposed	propose	VERB
ajst-21816	24	7	a	a	DET
ajst-21816	24	8	component	component	NOUN
ajst-21816	24	9	-	-	PUNCT
ajst-21816	24	10	based	base	VERB
ajst-21816	24	11	gait	gait	NOUN
ajst-21816	24	12	recognition	recognition	NOUN
ajst-21816	24	13	method	method	NOUN
ajst-21816	24	14	,	,	PUNCT
ajst-21816	24	15	gaitpart	gaitpart	NOUN
ajst-21816	24	16	,	,	PUNCT
ajst-21816	24	17	which	which	PRON
ajst-21816	24	18	improved	improve	VERB
ajst-21816	24	19	the	the	DET
ajst-21816	24	20	gait	gait	ADJ
ajst-21816	24	21	recognition	recognition	NOUN
ajst-21816	24	22	rate	rate	NOUN
ajst-21816	24	23	from	from	ADP
ajst-21816	24	24	the	the	DET
ajst-21816	24	25	perspective	perspective	NOUN
ajst-21816	24	26	of	of	ADP
ajst-21816	24	27	space	space	NOUN
ajst-21816	24	28	and	and	CCONJ
ajst-21816	24	29	time	time	NOUN
ajst-21816	24	30	,	,	PUNCT
ajst-21816	24	31	extracted	extract	VERB
ajst-21816	24	32	fine	fine	ADV
ajst-21816	24	33	-	-	PUNCT
ajst-21816	24	34	grained	grain	VERB
ajst-21816	24	35	spatial	spatial	ADJ
ajst-21816	24	36	information	information	NOUN
ajst-21816	24	37	from	from	ADP
ajst-21816	24	38	a	a	DET
ajst-21816	24	39	single	single	ADJ
ajst-21816	24	40	image	image	NOUN
ajst-21816	24	41	by	by	ADP
ajst-21816	24	42	using	use	VERB
ajst-21816	24	43	focus	focus	NOUN
ajst-21816	24	44	conver	conver	NOUN
ajst-21816	24	45	(	(	PUNCT
ajst-21816	24	46	fconv	fconv	ADJ
ajst-21816	24	47	)	)	PUNCT
ajst-21816	24	48	,	,	PUNCT
ajst-21816	24	49	and	and	CCONJ
ajst-21816	24	50	extracted	extract	VERB
ajst-21816	24	51	and	and	CCONJ
ajst-21816	24	52	aggregated	aggregate	VERB
ajst-21816	24	53	short	short	ADJ
ajst-21816	24	54	-	-	PUNCT
ajst-21816	24	55	time	time	NOUN
ajst-21816	24	56	gait	gait	ADJ
ajst-21816	24	57	information	information	NOUN
ajst-21816	24	58	from	from	ADP
ajst-21816	24	59	each	each	DET
ajst-21816	24	60	component	component	NOUN
ajst-21816	24	61	by	by	ADP
ajst-21816	24	62	using	use	VERB
ajst-21816	24	63	a	a	DET
ajst-21816	24	64	micro	micro	ADJ
ajst-21816	24	65	-	-	NOUN
ajst-21816	24	66	motion	motion	NOUN
ajst-21816	24	67	capture	capture	NOUN
ajst-21816	24	68	module	module	NOUN
ajst-21816	24	69	.	.	PUNCT
ajst-21816	25	1	ghorbani[11	ghorbani[11	PROPN
ajst-21816	25	2	]	]	PUNCT
ajst-21816	25	3	et	et	PROPN
ajst-21816	25	4	al	al	PROPN
ajst-21816	25	5	.	.	PROPN
ajst-21816	25	6	proposed	propose	VERB
ajst-21816	25	7	a	a	DET
ajst-21816	25	8	novel	novel	ADJ
ajst-21816	25	9	deep	deep	ADJ
ajst-21816	25	10	convolutional	convolutional	ADJ
ajst-21816	25	11	network	network	NOUN
ajst-21816	25	12	that	that	PRON
ajst-21816	25	13	uses	use	VERB
ajst-21816	25	14	a	a	DET
ajst-21816	25	15	capsule	capsule	NOUN
ajst-21816	25	16	network	network	NOUN
ajst-21816	25	17	(	(	PUNCT
ajst-21816	25	18	capsnet	capsnet	NOUN
ajst-21816	25	19	)	)	PUNCT
ajst-21816	25	20	to	to	PART
ajst-21816	25	21	learn	learn	VERB
ajst-21816	25	22	deeper	deep	ADJ
ajst-21816	25	23	partial	partial	ADJ
ajst-21816	25	24	and	and	CCONJ
ajst-21816	25	25	global	global	ADJ
ajst-21816	25	26	relationships	relationship	NOUN
ajst-21816	25	27	and	and	CCONJ
ajst-21816	25	28	assign	assign	VERB
ajst-21816	25	29	weights	weight	NOUN
ajst-21816	25	30	to	to	ADP
ajst-21816	25	31	related	relate	VERB
ajst-21816	25	32	features	feature	NOUN
ajst-21816	25	33	to	to	PART
ajst-21816	25	34	obtain	obtain	VERB
ajst-21816	25	35	a	a	DET
ajst-21816	25	36	gait	gait	ADJ
ajst-21816	25	37	representation	representation	NOUN
ajst-21816	25	38	that	that	PRON
ajst-21816	25	39	is	be	AUX
ajst-21816	25	40	more	more	ADV
ajst-21816	25	41	robust	robust	ADJ
ajst-21816	25	42	to	to	ADP
ajst-21816	25	43	changes	change	NOUN
ajst-21816	25	44	in	in	ADP
ajst-21816	25	45	appearance	appearance	NOUN
ajst-21816	25	46	.	.	PUNCT
ajst-21816	26	1	chao[12	chao[12	X
ajst-21816	26	2	]	]	X
ajst-21816	26	3	et	et	PROPN
ajst-21816	26	4	al	al	PROPN
ajst-21816	26	5	.	.	PUNCT
ajst-21816	27	1	propose	propose	VERB
ajst-21816	27	2	an	an	DET
ajst-21816	27	3	end	end	NOUN
ajst-21816	27	4	-	-	PUNCT
ajst-21816	27	5	to	to	ADP
ajst-21816	27	6	-	-	PUNCT
ajst-21816	27	7	end	end	NOUN
ajst-21816	27	8	gait	gait	NOUN
ajst-21816	27	9	recognition	recognition	NOUN
ajst-21816	27	10	method	method	NOUN
ajst-21816	27	11	,	,	PUNCT
ajst-21816	27	12	gaitset	gaitset	NOUN
ajst-21816	27	13	,	,	PUNCT
ajst-21816	27	14	which	which	PRON
ajst-21816	27	15	takes	take	VERB
ajst-21816	27	16	gait	gait	NOUN
ajst-21816	27	17	sequence	sequence	NOUN
ajst-21816	27	18	as	as	ADP
ajst-21816	27	19	input	input	NOUN
ajst-21816	27	20	,	,	PUNCT
ajst-21816	27	21	uses	use	VERB
ajst-21816	27	22	convolutional	convolutional	ADJ
ajst-21816	27	23	network	network	NOUN
ajst-21816	27	24	to	to	PART
ajst-21816	27	25	directly	directly	ADV
ajst-21816	27	26	extract	extract	VERB
ajst-21816	27	27	spatial	spatial	ADJ
ajst-21816	27	28	features	feature	NOUN
ajst-21816	27	29	from	from	ADP
ajst-21816	27	30	the	the	DET
ajst-21816	27	31	original	original	ADJ
ajst-21816	27	32	gait	gait	NOUN
ajst-21816	27	33	contour	contour	NOUN
ajst-21816	27	34	,	,	PUNCT
ajst-21816	27	35	and	and	CCONJ
ajst-21816	27	36	then	then	ADV
ajst-21816	27	37	adopts	adopt	VERB
ajst-21816	27	38	set	set	VERB
ajst-21816	27	39	pooling	pooling	NOUN
ajst-21816	27	40	(	(	PUNCT
ajst-21816	27	41	sp	sp	NOUN
ajst-21816	27	42	)	)	PUNCT
ajst-21816	27	43	operation	operation	NOUN
ajst-21816	27	44	to	to	PART
ajst-21816	27	45	conduct	conduct	VERB
ajst-21816	27	46	a	a	DET
ajst-21816	27	47	deep	deep	ADJ
ajst-21816	27	48	set	set	NOUN
ajst-21816	27	49	of	of	ADP
ajst-21816	27	50	spatial	spatial	ADJ
ajst-21816	27	51	features	feature	NOUN
ajst-21816	27	52	.	.	PUNCT
ajst-21816	28	1	the	the	DET
ajst-21816	28	2	multilayer	multilayer	ADJ
ajst-21816	28	3	global	global	ADJ
ajst-21816	28	4	pipeline	pipeline	NOUN
ajst-21816	28	5	with	with	ADP
ajst-21816	28	6	attention	attention	NOUN
ajst-21816	28	7	(	(	PUNCT
ajst-21816	28	8	mgpa	mgpa	NOUN
ajst-21816	28	9	)	)	PUNCT
ajst-21816	28	10	module	module	NOUN
ajst-21816	28	11	is	be	AUX
ajst-21816	28	12	used	use	VERB
ajst-21816	28	13	to	to	PART
ajst-21816	28	14	extract	extract	VERB
ajst-21816	28	15	the	the	DET
ajst-21816	28	16	collection	collection	NOUN
ajst-21816	28	17	-	-	PUNCT
ajst-21816	28	18	level	level	NOUN
ajst-21816	28	19	information	information	NOUN
ajst-21816	28	20	of	of	ADP
ajst-21816	28	21	each	each	DET
ajst-21816	28	22	convolution	convolution	NOUN
ajst-21816	28	23	layer	layer	NOUN
ajst-21816	28	24	.	.	PUNCT
ajst-21816	29	1	the	the	DET
ajst-21816	29	2	contour	contour	NOUN
ajst-21816	29	3	-	-	PUNCT
ajst-21816	29	4	based	base	VERB
ajst-21816	29	5	gait	gait	NOUN
ajst-21816	29	6	recognition	recognition	NOUN
ajst-21816	29	7	method	method	NOUN
ajst-21816	29	8	is	be	AUX
ajst-21816	29	9	seriously	seriously	ADV
ajst-21816	29	10	disturbed	disturb	VERB
ajst-21816	29	11	in	in	ADP
ajst-21816	29	12	the	the	DET
ajst-21816	29	13	actual	actual	ADJ
ajst-21816	29	14	complex	complex	ADJ
ajst-21816	29	15	scene	scene	NOUN
ajst-21816	29	16	(	(	PUNCT
ajst-21816	29	17	such	such	ADJ
ajst-21816	29	18	as	as	ADP
ajst-21816	29	19	cross	cross	ADJ
ajst-21816	29	20	-	-	ADJ
ajst-21816	29	21	viewing	view	VERB
ajst-21816	29	22	angle	angle	NOUN
ajst-21816	29	23	)	)	PUNCT
ajst-21816	29	24	,	,	PUNCT
ajst-21816	29	25	and	and	CCONJ
ajst-21816	29	26	can	can	AUX
ajst-21816	29	27	not	not	PART
ajst-21816	29	28	achieve	achieve	VERB
ajst-21816	29	29	ideal	ideal	ADJ
ajst-21816	29	30	results	result	NOUN
ajst-21816	29	31	.	.	PUNCT
ajst-21816	30	1	some	some	DET
ajst-21816	30	2	researchers[1315	researchers[1315	NOUN
ajst-21816	30	3	]	]	PUNCT
ajst-21816	30	4	use	use	NOUN
ajst-21816	30	5	matrix	matrix	NOUN
ajst-21816	30	6	transformation	transformation	NOUN
ajst-21816	30	7	to	to	PART
ajst-21816	30	8	normalize	normalize	VERB
ajst-21816	30	9	gait	gait	NOUN
ajst-21816	30	10	images	image	NOUN
ajst-21816	30	11	from	from	ADP
ajst-21816	30	12	different	different	ADJ
ajst-21816	30	13	perspectives	perspective	NOUN
ajst-21816	30	14	to	to	ADP
ajst-21816	30	15	a	a	DET
ajst-21816	30	16	specific	specific	ADJ
ajst-21816	30	17	perspective	perspective	NOUN
ajst-21816	30	18	,	,	PUNCT
ajst-21816	30	19	and	and	CCONJ
ajst-21816	30	20	then	then	ADV
ajst-21816	30	21	carry	carry	VERB
ajst-21816	30	22	out	out	ADP
ajst-21816	30	23	identity	identity	NOUN
ajst-21816	30	24	authentication	authentication	NOUN
ajst-21816	30	25	under	under	ADP
ajst-21816	30	26	this	this	DET
ajst-21816	30	27	perspective	perspective	NOUN
ajst-21816	30	28	.	.	PUNCT
ajst-21816	31	1	however	however	ADV
ajst-21816	31	2	,	,	PUNCT
ajst-21816	31	3	information	information	NOUN
ajst-21816	31	4	loss	loss	NOUN
ajst-21816	31	5	in	in	ADP
ajst-21816	31	6	the	the	DET
ajst-21816	31	7	conversion	conversion	NOUN
ajst-21816	31	8	process	process	NOUN
ajst-21816	31	9	will	will	AUX
ajst-21816	31	10	lead	lead	VERB
ajst-21816	31	11	to	to	ADP
ajst-21816	31	12	a	a	DET
ajst-21816	31	13	decline	decline	NOUN
ajst-21816	31	14	in	in	ADP
ajst-21816	31	15	recognition	recognition	NOUN
ajst-21816	31	16	accuracy	accuracy	NOUN
ajst-21816	31	17	,	,	PUNCT
ajst-21816	31	18	because	because	SCONJ
ajst-21816	31	19	this	this	DET
ajst-21816	31	20	method	method	NOUN
ajst-21816	31	21	is	be	AUX
ajst-21816	31	22	usually	usually	ADV
ajst-21816	31	23	affected	affect	VERB
ajst-21816	31	24	by	by	ADP
ajst-21816	31	25	changes	change	NOUN
ajst-21816	31	26	in	in	ADP
ajst-21816	31	27	perspective	perspective	NOUN
ajst-21816	31	28	,	,	PUNCT
ajst-21816	31	29	changes	change	NOUN
ajst-21816	31	30	in	in	ADP
ajst-21816	31	31	objects	object	NOUN
ajst-21816	31	32	and	and	CCONJ
ajst-21816	31	33	other	other	ADJ
ajst-21816	31	34	factors	factor	NOUN
ajst-21816	31	35	.	.	PUNCT
ajst-21816	32	1	as	as	ADP
ajst-21816	32	2	a	a	DET
ajst-21816	32	3	result	result	NOUN
ajst-21816	32	4	,	,	PUNCT
ajst-21816	32	5	some	some	DET
ajst-21816	32	6	effective	effective	ADJ
ajst-21816	32	7	information	information	NOUN
ajst-21816	32	8	is	be	AUX
ajst-21816	32	9	lost	lose	VERB
ajst-21816	32	10	during	during	ADP
ajst-21816	32	11	the	the	DET
ajst-21816	32	12	walking	walking	NOUN
ajst-21816	32	13	process	process	NOUN
ajst-21816	32	14	,	,	PUNCT
ajst-21816	32	15	which	which	PRON
ajst-21816	32	16	will	will	AUX
ajst-21816	32	17	reduce	reduce	VERB
ajst-21816	32	18	the	the	DET
ajst-21816	32	19	287	287	NUM
ajst-21816	32	20	recognition	recognition	NOUN
ajst-21816	32	21	rate	rate	NOUN
ajst-21816	32	22	.	.	PUNCT
ajst-21816	33	1	to	to	PART
ajst-21816	33	2	solve	solve	VERB
ajst-21816	33	3	the	the	DET
ajst-21816	33	4	above	above	ADJ
ajst-21816	33	5	problems	problem	NOUN
ajst-21816	33	6	,	,	PUNCT
ajst-21816	33	7	this	this	DET
ajst-21816	33	8	paper	paper	NOUN
ajst-21816	33	9	proposes	propose	VERB
ajst-21816	33	10	a	a	DET
ajst-21816	33	11	convolutional	convolutional	ADJ
ajst-21816	33	12	neural	neural	ADJ
ajst-21816	33	13	network	network	NOUN
ajst-21816	33	14	-	-	PUNCT
ajst-21816	33	15	based	base	VERB
ajst-21816	33	16	lfe	lfe	PROPN
ajst-21816	33	17	-	-	PUNCT
ajst-21816	33	18	mgpa	mgpa	PROPN
ajst-21816	33	19	network	network	NOUN
ajst-21816	33	20	model	model	NOUN
ajst-21816	33	21	,	,	PUNCT
ajst-21816	33	22	in	in	ADP
ajst-21816	33	23	which	which	PRON
ajst-21816	33	24	specific	specific	ADJ
ajst-21816	33	25	proportional	proportional	ADJ
ajst-21816	33	26	segmentation	segmentation	NOUN
ajst-21816	33	27	convolution	convolution	NOUN
ajst-21816	33	28	(	(	PUNCT
ajst-21816	33	29	sconv	sconv	NOUN
ajst-21816	33	30	)	)	PUNCT
ajst-21816	33	31	is	be	AUX
ajst-21816	33	32	proposed	propose	VERB
ajst-21816	33	33	.	.	PUNCT
ajst-21816	34	1	the	the	DET
ajst-21816	34	2	input	input	NOUN
ajst-21816	34	3	pedestrian	pedestrian	NOUN
ajst-21816	34	4	contour	contour	NOUN
ajst-21816	34	5	map	map	NOUN
ajst-21816	34	6	is	be	AUX
ajst-21816	34	7	segmented	segment	VERB
ajst-21816	34	8	according	accord	VERB
ajst-21816	34	9	to	to	ADP
ajst-21816	34	10	a	a	DET
ajst-21816	34	11	specific	specific	ADJ
ajst-21816	34	12	ratio	ratio	NOUN
ajst-21816	34	13	(	(	PUNCT
ajst-21816	34	14	1:2:1:2	1:2:1:2	NUM
ajst-21816	34	15	)	)	PUNCT
ajst-21816	34	16	,	,	PUNCT
ajst-21816	34	17	and	and	CCONJ
ajst-21816	34	18	on	on	ADP
ajst-21816	34	19	this	this	DET
ajst-21816	34	20	basis	basis	NOUN
ajst-21816	34	21	,	,	PUNCT
ajst-21816	34	22	mgpa	mgpa	VERB
ajst-21816	34	23	module	module	NOUN
ajst-21816	34	24	in	in	ADP
ajst-21816	34	25	gaitset	gaitset	NOUN
ajst-21816	34	26	is	be	AUX
ajst-21816	34	27	integrated	integrate	VERB
ajst-21816	34	28	to	to	PART
ajst-21816	34	29	construct	construct	VERB
ajst-21816	34	30	a	a	DET
ajst-21816	34	31	joint	joint	ADJ
ajst-21816	34	32	loss	loss	NOUN
ajst-21816	34	33	function	function	NOUN
ajst-21816	34	34	,	,	PUNCT
ajst-21816	34	35	and	and	CCONJ
ajst-21816	34	36	experiments	experiment	NOUN
ajst-21816	34	37	are	be	AUX
ajst-21816	34	38	conducted	conduct	VERB
ajst-21816	34	39	on	on	ADP
ajst-21816	34	40	the	the	DET
ajst-21816	34	41	open	open	ADJ
ajst-21816	34	42	gait	gait	PROPN
ajst-21816	34	43	dataset	dataset	PROPN
ajst-21816	34	44	casia	casia	PROPN
ajst-21816	34	45	-	-	PUNCT
ajst-21816	34	46	b[16	b[16	PROPN
ajst-21816	34	47	]	]	PUNCT
ajst-21816	34	48	.	.	PUNCT
ajst-21816	35	1	the	the	DET
ajst-21816	35	2	results	result	NOUN
ajst-21816	35	3	show	show	VERB
ajst-21816	35	4	that	that	SCONJ
ajst-21816	35	5	,	,	PUNCT
ajst-21816	35	6	the	the	DET
ajst-21816	35	7	recognition	recognition	NOUN
ajst-21816	35	8	accuracy	accuracy	NOUN
ajst-21816	35	9	of	of	ADP
ajst-21816	35	10	the	the	DET
ajst-21816	35	11	proposed	propose	VERB
ajst-21816	35	12	method	method	NOUN
ajst-21816	35	13	is	be	AUX
ajst-21816	35	14	superior	superior	ADJ
ajst-21816	35	15	to	to	ADP
ajst-21816	35	16	the	the	DET
ajst-21816	35	17	existing	exist	VERB
ajst-21816	35	18	mainstream	mainstream	NOUN
ajst-21816	35	19	algorithms	algorithm	NOUN
ajst-21816	35	20	in	in	ADP
ajst-21816	35	21	the	the	DET
ajst-21816	35	22	cross	cross	NOUN
ajst-21816	35	23	-	-	ADJ
ajst-21816	35	24	view	view	ADJ
ajst-21816	35	25	and	and	CCONJ
ajst-21816	35	26	multi	multi	ADJ
ajst-21816	35	27	-	-	ADJ
ajst-21816	35	28	walk	walk	ADJ
ajst-21816	35	29	state	state	NOUN
ajst-21816	35	30	.	.	PUNCT
ajst-21816	36	1	2	2	X
ajst-21816	36	2	.	.	X
ajst-21816	36	3	textual	textual	ADJ
ajst-21816	36	4	method	method	NOUN
ajst-21816	36	5	2.1	2.1	NUM
ajst-21816	36	6	.	.	PUNCT
ajst-21816	37	1	lfe	lfe	PROPN
ajst-21816	37	2	-	-	PUNCT
ajst-21816	37	3	mgpa	mgpa	PROPN
ajst-21816	37	4	network	network	NOUN
ajst-21816	37	5	model	model	NOUN
ajst-21816	37	6	architecture	architecture	NOUN
ajst-21816	37	7	figure	figure	NOUN
ajst-21816	37	8	1	1	NUM
ajst-21816	37	9	.	.	PUNCT
ajst-21816	38	1	lfe	lfe	PROPN
ajst-21816	38	2	-	-	PUNCT
ajst-21816	38	3	mgpa	mgpa	PROPN
ajst-21816	38	4	network	network	NOUN
ajst-21816	38	5	overall	overall	ADJ
ajst-21816	38	6	frame	frame	NOUN
ajst-21816	38	7	diagram	diagram	NOUN
ajst-21816	38	8	convolutional	convolutional	ADJ
ajst-21816	38	9	neural	neural	ADJ
ajst-21816	38	10	networks	network	NOUN
ajst-21816	38	11	can	can	AUX
ajst-21816	38	12	enhance	enhance	VERB
ajst-21816	38	13	the	the	DET
ajst-21816	38	14	fitting	fitting	ADJ
ajst-21816	38	15	and	and	CCONJ
ajst-21816	38	16	generalization	generalization	NOUN
ajst-21816	38	17	ability	ability	NOUN
ajst-21816	38	18	by	by	ADP
ajst-21816	38	19	learning	learn	VERB
ajst-21816	38	20	complex	complex	ADJ
ajst-21816	38	21	samples	sample	NOUN
ajst-21816	38	22	,	,	PUNCT
ajst-21816	38	23	so	so	SCONJ
ajst-21816	38	24	this	this	DET
ajst-21816	38	25	paper	paper	NOUN
ajst-21816	38	26	proposes	propose	VERB
ajst-21816	38	27	the	the	DET
ajst-21816	38	28	lfe	lfe	PROPN
ajst-21816	38	29	-	-	PUNCT
ajst-21816	38	30	mgpa	mgpa	PROPN
ajst-21816	38	31	model	model	NOUN
ajst-21816	38	32	based	base	VERB
ajst-21816	38	33	on	on	ADP
ajst-21816	38	34	neural	neural	ADJ
ajst-21816	38	35	networks	network	NOUN
ajst-21816	38	36	,	,	PUNCT
ajst-21816	38	37	and	and	CCONJ
ajst-21816	38	38	its	its	PRON
ajst-21816	38	39	overall	overall	ADJ
ajst-21816	38	40	frame	frame	NOUN
ajst-21816	38	41	diagram	diagram	NOUN
ajst-21816	38	42	is	be	AUX
ajst-21816	38	43	shown	show	VERB
ajst-21816	38	44	in	in	ADP
ajst-21816	38	45	figure	figure	NOUN
ajst-21816	38	46	1	1	NUM
ajst-21816	38	47	.	.	PUNCT
ajst-21816	39	1	the	the	DET
ajst-21816	39	2	network	network	NOUN
ajst-21816	39	3	consists	consist	VERB
ajst-21816	39	4	of	of	ADP
ajst-21816	39	5	three	three	NUM
ajst-21816	39	6	parts	part	NOUN
ajst-21816	39	7	:	:	PUNCT
ajst-21816	39	8	the	the	DET
ajst-21816	39	9	first	first	ADJ
ajst-21816	39	10	part	part	NOUN
ajst-21816	39	11	is	be	AUX
ajst-21816	39	12	the	the	DET
ajst-21816	39	13	local	local	ADJ
ajst-21816	39	14	feature	feature	NOUN
ajst-21816	39	15	extraction	extraction	NOUN
ajst-21816	39	16	which	which	PRON
ajst-21816	39	17	contains	contain	VERB
ajst-21816	39	18	multiple	multiple	ADJ
ajst-21816	39	19	specific	specific	ADJ
ajst-21816	39	20	proportional	proportional	ADJ
ajst-21816	39	21	segmentation	segmentation	NOUN
ajst-21816	39	22	convolution	convolution	NOUN
ajst-21816	39	23	(	(	PUNCT
ajst-21816	39	24	sconv	sconv	NOUN
ajst-21816	39	25	)	)	PUNCT
ajst-21816	39	26	;	;	PUNCT
ajst-21816	39	27	the	the	DET
ajst-21816	39	28	second	second	ADJ
ajst-21816	39	29	part	part	NOUN
ajst-21816	39	30	is	be	AUX
ajst-21816	39	31	mgpa	mgpa	NOUN
ajst-21816	39	32	(	(	PUNCT
ajst-21816	39	33	multilayer	multilayer	ADJ
ajst-21816	39	34	global	global	ADJ
ajst-21816	39	35	pipeline	pipeline	NOUN
ajst-21816	39	36	)	)	PUNCT
ajst-21816	39	37	,	,	PUNCT
ajst-21816	39	38	which	which	PRON
ajst-21816	39	39	is	be	AUX
ajst-21816	39	40	used	use	VERB
ajst-21816	39	41	to	to	PART
ajst-21816	39	42	collect	collect	VERB
ajst-21816	39	43	the	the	DET
ajst-21816	39	44	collective	collective	ADJ
ajst-21816	39	45	-	-	PUNCT
ajst-21816	39	46	level	level	NOUN
ajst-21816	39	47	characteristics	characteristic	NOUN
ajst-21816	39	48	of	of	ADP
ajst-21816	39	49	different	different	ADJ
ajst-21816	39	50	layers	layer	NOUN
ajst-21816	39	51	.	.	PUNCT
ajst-21816	40	1	the	the	DET
ajst-21816	40	2	third	third	ADJ
ajst-21816	40	3	part	part	NOUN
ajst-21816	40	4	is	be	AUX
ajst-21816	40	5	the	the	DET
ajst-21816	40	6	loss	loss	NOUN
ajst-21816	40	7	function	function	NOUN
ajst-21816	40	8	.	.	PUNCT
ajst-21816	41	1	firstly	firstly	ADV
ajst-21816	41	2	,	,	PUNCT
ajst-21816	41	3	we	we	PRON
ajst-21816	41	4	take	take	VERB
ajst-21816	41	5	the	the	DET
ajst-21816	41	6	unordered	unordered	ADJ
ajst-21816	41	7	gait	gait	NOUN
ajst-21816	41	8	sequence	sequence	NOUN
ajst-21816	41	9	as	as	ADP
ajst-21816	41	10	input	input	NOUN
ajst-21816	41	11	,	,	PUNCT
ajst-21816	41	12	and	and	CCONJ
ajst-21816	41	13	use	use	NOUN
ajst-21816	41	14	sp	sp	ADP
ajst-21816	41	15	(	(	PUNCT
ajst-21816	41	16	set	set	NOUN
ajst-21816	41	17	pooling	pooling	NOUN
ajst-21816	41	18	)	)	PUNCT
ajst-21816	41	19	to	to	PART
ajst-21816	41	20	aggregate	aggregate	VERB
ajst-21816	41	21	the	the	DET
ajst-21816	41	22	frame	frame	NOUN
ajst-21816	41	23	level	level	NOUN
ajst-21816	41	24	features	feature	NOUN
ajst-21816	41	25	of	of	ADP
ajst-21816	41	26	different	different	ADJ
ajst-21816	41	27	lfe	lfe	PROPN
ajst-21816	41	28	layers	layer	NOUN
ajst-21816	41	29	into	into	ADP
ajst-21816	41	30	the	the	DET
ajst-21816	41	31	set	set	VERB
ajst-21816	41	32	level	level	NOUN
ajst-21816	41	33	features	feature	NOUN
ajst-21816	41	34	.	.	PUNCT
ajst-21816	42	1	secondly	secondly	ADV
ajst-21816	42	2	,	,	PUNCT
ajst-21816	42	3	lfe	lfe	PROPN
ajst-21816	42	4	and	and	CCONJ
ajst-21816	42	5	mgpa	mgpa	NOUN
ajst-21816	42	6	are	be	AUX
ajst-21816	42	7	subsampled	subsample	VERB
ajst-21816	42	8	respectively	respectively	ADV
ajst-21816	42	9	,	,	PUNCT
ajst-21816	42	10	and	and	CCONJ
ajst-21816	42	11	feature	feature	NOUN
ajst-21816	42	12	fusion	fusion	NOUN
ajst-21816	42	13	is	be	AUX
ajst-21816	42	14	carried	carry	VERB
ajst-21816	42	15	out	out	ADP
ajst-21816	42	16	after	after	ADP
ajst-21816	42	17	subsampling	subsample	VERB
ajst-21816	42	18	.	.	PUNCT
ajst-21816	43	1	then	then	ADV
ajst-21816	43	2	,	,	PUNCT
ajst-21816	43	3	the	the	DET
ajst-21816	43	4	feature	feature	NOUN
ajst-21816	43	5	mapping	mapping	NOUN
ajst-21816	43	6	is	be	AUX
ajst-21816	43	7	carried	carry	VERB
ajst-21816	43	8	out	out	ADP
ajst-21816	43	9	by	by	ADP
ajst-21816	43	10	horizontal	horizontal	ADJ
ajst-21816	43	11	pyramid	pyramid	PROPN
ajst-21816	43	12	matching	matching	NOUN
ajst-21816	43	13	.	.	PUNCT
ajst-21816	44	1	finally	finally	ADV
ajst-21816	44	2	,	,	PUNCT
ajst-21816	44	3	we	we	PRON
ajst-21816	44	4	use	use	VERB
ajst-21816	44	5	the	the	DET
ajst-21816	44	6	joint	joint	ADJ
ajst-21816	44	7	loss	loss	NOUN
ajst-21816	44	8	function	function	NOUN
ajst-21816	44	9	to	to	PART
ajst-21816	44	10	train	train	VERB
ajst-21816	44	11	the	the	DET
ajst-21816	44	12	model	model	NOUN
ajst-21816	44	13	.	.	PUNCT
ajst-21816	45	1	sp	sp	ADP
ajst-21816	45	2	and	and	CCONJ
ajst-21816	45	3	lfe	lfe	PROPN
ajst-21816	45	4	are	be	AUX
ajst-21816	45	5	described	describe	VERB
ajst-21816	45	6	in	in	ADP
ajst-21816	45	7	detail	detail	NOUN
ajst-21816	45	8	in	in	ADP
ajst-21816	45	9	sections	section	NOUN
ajst-21816	45	10	2.2	2.2	NUM
ajst-21816	45	11	and	and	CCONJ
ajst-21816	45	12	2.3	2.3	NUM
ajst-21816	45	13	,	,	PUNCT
ajst-21816	45	14	respectively	respectively	ADV
ajst-21816	45	15	.	.	PUNCT
ajst-21816	46	1	2.2	2.2	NUM
ajst-21816	46	2	.	.	PUNCT
ajst-21816	47	1	mgp	mgp	NOUN
ajst-21816	47	2	with	with	ADP
ajst-21816	47	3	attention（mgpa	attention（mgpa	NOUN
ajst-21816	47	4	）	）	PUNCT
ajst-21816	47	5	generally	generally	ADV
ajst-21816	47	6	speaking	speak	VERB
ajst-21816	47	7	,	,	PUNCT
ajst-21816	47	8	in	in	ADP
ajst-21816	47	9	convolutional	convolutional	ADJ
ajst-21816	47	10	networks	network	NOUN
ajst-21816	47	11	,	,	PUNCT
ajst-21816	47	12	the	the	DET
ajst-21816	47	13	receptive	receptive	ADJ
ajst-21816	47	14	fields	field	NOUN
ajst-21816	47	15	of	of	ADP
ajst-21816	47	16	different	different	ADJ
ajst-21816	47	17	convolutional	convolutional	ADJ
ajst-21816	47	18	layers	layer	NOUN
ajst-21816	47	19	are	be	AUX
ajst-21816	47	20	different	different	ADJ
ajst-21816	47	21	.	.	PUNCT
ajst-21816	48	1	as	as	SCONJ
ajst-21816	48	2	the	the	DET
ajst-21816	48	3	layer	layer	NOUN
ajst-21816	48	4	deepens	deepen	VERB
ajst-21816	48	5	,	,	PUNCT
ajst-21816	48	6	the	the	DET
ajst-21816	48	7	receptive	receptive	ADJ
ajst-21816	48	8	field	field	NOUN
ajst-21816	48	9	becomes	become	VERB
ajst-21816	48	10	larger	large	ADJ
ajst-21816	48	11	.	.	PUNCT
ajst-21816	49	1	this	this	PRON
ajst-21816	49	2	also	also	ADV
ajst-21816	49	3	means	mean	VERB
ajst-21816	49	4	that	that	SCONJ
ajst-21816	49	5	after	after	SCONJ
ajst-21816	49	6	the	the	DET
ajst-21816	49	7	input	input	NOUN
ajst-21816	49	8	image	image	NOUN
ajst-21816	49	9	passes	pass	VERB
ajst-21816	49	10	through	through	ADP
ajst-21816	49	11	the	the	DET
ajst-21816	49	12	shallow	shallow	ADJ
ajst-21816	49	13	convolution	convolution	NOUN
ajst-21816	49	14	layer	layer	NOUN
ajst-21816	49	15	,	,	PUNCT
ajst-21816	49	16	each	each	DET
ajst-21816	49	17	pixel	pixel	NOUN
ajst-21816	49	18	of	of	ADP
ajst-21816	49	19	the	the	DET
ajst-21816	49	20	image	image	NOUN
ajst-21816	49	21	only	only	ADV
ajst-21816	49	22	extracts	extract	VERB
ajst-21816	49	23	the	the	DET
ajst-21816	49	24	local	local	ADJ
ajst-21816	49	25	information	information	NOUN
ajst-21816	49	26	of	of	ADP
ajst-21816	49	27	the	the	DET
ajst-21816	49	28	original	original	ADJ
ajst-21816	49	29	image	image	NOUN
ajst-21816	49	30	.	.	PUNCT
ajst-21816	50	1	the	the	DET
ajst-21816	50	2	extracted	extract	VERB
ajst-21816	50	3	feature	feature	NOUN
ajst-21816	50	4	details	detail	NOUN
ajst-21816	50	5	are	be	AUX
ajst-21816	50	6	rich	rich	ADJ
ajst-21816	50	7	,	,	PUNCT
ajst-21816	50	8	but	but	CCONJ
ajst-21816	50	9	the	the	DET
ajst-21816	50	10	context	context	NOUN
ajst-21816	50	11	information	information	NOUN
ajst-21816	50	12	of	of	ADP
ajst-21816	50	13	the	the	DET
ajst-21816	50	14	image	image	NOUN
ajst-21816	50	15	is	be	AUX
ajst-21816	50	16	little	little	ADJ
ajst-21816	50	17	.	.	PUNCT
ajst-21816	51	1	on	on	ADP
ajst-21816	51	2	the	the	DET
ajst-21816	51	3	other	other	ADJ
ajst-21816	51	4	hand	hand	NOUN
ajst-21816	51	5	,	,	PUNCT
ajst-21816	51	6	the	the	DET
ajst-21816	51	7	input	input	NOUN
ajst-21816	51	8	image	image	NOUN
ajst-21816	51	9	can	can	AUX
ajst-21816	51	10	be	be	AUX
ajst-21816	51	11	seen	see	VERB
ajst-21816	51	12	in	in	ADP
ajst-21816	51	13	a	a	DET
ajst-21816	51	14	larger	large	ADJ
ajst-21816	51	15	range	range	NOUN
ajst-21816	51	16	after	after	ADP
ajst-21816	51	17	the	the	DET
ajst-21816	51	18	action	action	NOUN
ajst-21816	51	19	of	of	ADP
ajst-21816	51	20	the	the	DET
ajst-21816	51	21	deep	deep	ADJ
ajst-21816	51	22	convolutional	convolutional	ADJ
ajst-21816	51	23	layer	layer	NOUN
ajst-21816	51	24	,	,	PUNCT
ajst-21816	51	25	but	but	CCONJ
ajst-21816	51	26	the	the	DET
ajst-21816	51	27	details	detail	NOUN
ajst-21816	51	28	are	be	AUX
ajst-21816	51	29	lost	lose	VERB
ajst-21816	51	30	.	.	PUNCT
ajst-21816	52	1	in	in	ADP
ajst-21816	52	2	view	view	NOUN
ajst-21816	52	3	of	of	ADP
ajst-21816	52	4	the	the	DET
ajst-21816	52	5	above	above	ADJ
ajst-21816	52	6	situation	situation	NOUN
ajst-21816	52	7	,	,	PUNCT
ajst-21816	52	8	mgpa	mgpa	NOUN
ajst-21816	52	9	module	module	NOUN
ajst-21816	52	10	is	be	AUX
ajst-21816	52	11	used	use	VERB
ajst-21816	52	12	to	to	PART
ajst-21816	52	13	extract	extract	VERB
ajst-21816	52	14	the	the	DET
ajst-21816	52	15	collective	collective	ADJ
ajst-21816	52	16	-	-	PUNCT
ajst-21816	52	17	level	level	NOUN
ajst-21816	52	18	features	feature	NOUN
ajst-21816	52	19	of	of	ADP
ajst-21816	52	20	different	different	ADJ
ajst-21816	52	21	layers	layer	NOUN
ajst-21816	52	22	.	.	PUNCT
ajst-21816	53	1	the	the	DET
ajst-21816	53	2	structure	structure	NOUN
ajst-21816	53	3	diagram	diagram	NOUN
ajst-21816	53	4	of	of	ADP
ajst-21816	53	5	this	this	DET
ajst-21816	53	6	module	module	NOUN
ajst-21816	53	7	is	be	AUX
ajst-21816	53	8	shown	show	VERB
ajst-21816	53	9	in	in	ADP
ajst-21816	53	10	figure	figure	NOUN
ajst-21816	53	11	2	2	NUM
ajst-21816	53	12	.	.	PUNCT
ajst-21816	53	13	figure	figure	NOUN
ajst-21816	53	14	2	2	NUM
ajst-21816	53	15	.	.	PUNCT
ajst-21816	54	1	mgpa	mgpa	PROPN
ajst-21816	54	2	architecture	architecture	NOUN
ajst-21816	54	3	diagram	diagram	NOUN
ajst-21816	54	4	288	288	NUM
ajst-21816	54	5	the	the	DET
ajst-21816	54	6	structure	structure	NOUN
ajst-21816	54	7	of	of	ADP
ajst-21816	54	8	mgpa	mgpa	NOUN
ajst-21816	54	9	modules	module	NOUN
ajst-21816	54	10	is	be	AUX
ajst-21816	54	11	similar	similar	ADJ
ajst-21816	54	12	to	to	ADP
ajst-21816	54	13	that	that	PRON
ajst-21816	54	14	of	of	ADP
ajst-21816	54	15	lfe	lfe	PROPN
ajst-21816	54	16	,	,	PUNCT
ajst-21816	54	17	where	where	SCONJ
ajst-21816	54	18	sconv	sconv	NOUN
ajst-21816	54	19	block2	block2	PROPN
ajst-21816	54	20	and	and	CCONJ
ajst-21816	54	21	sconv	sconv	PROPN
ajst-21816	54	22	block3	block3	PROPN
ajst-21816	54	23	have	have	VERB
ajst-21816	54	24	the	the	DET
ajst-21816	54	25	same	same	ADJ
ajst-21816	54	26	structure	structure	NOUN
ajst-21816	54	27	as	as	ADP
ajst-21816	54	28	sconv	sconv	NOUN
ajst-21816	54	29	block2	block2	NOUN
ajst-21816	54	30	and	and	CCONJ
ajst-21816	54	31	sconv	sconv	PROPN
ajst-21816	54	32	block3	block3	PROPN
ajst-21816	54	33	in	in	ADP
ajst-21816	54	34	lfe	lfe	PROPN
ajst-21816	54	35	,	,	PUNCT
ajst-21816	54	36	but	but	CCONJ
ajst-21816	54	37	the	the	DET
ajst-21816	54	38	parameters	parameter	NOUN
ajst-21816	54	39	between	between	ADP
ajst-21816	54	40	the	the	DET
ajst-21816	54	41	two	two	NUM
ajst-21816	54	42	modules	module	NOUN
ajst-21816	54	43	are	be	AUX
ajst-21816	54	44	not	not	PART
ajst-21816	54	45	shared	share	VERB
ajst-21816	54	46	.	.	PUNCT
ajst-21816	55	1	firstly	firstly	ADV
ajst-21816	55	2	,	,	PUNCT
ajst-21816	55	3	the	the	DET
ajst-21816	55	4	frame	frame	NOUN
ajst-21816	55	5	level	level	NOUN
ajst-21816	55	6	features	feature	NOUN
ajst-21816	55	7	extracted	extract	VERB
ajst-21816	55	8	by	by	ADP
ajst-21816	55	9	sconv	sconv	NOUN
ajst-21816	55	10	block1	block1	NOUN
ajst-21816	55	11	in	in	ADP
ajst-21816	55	12	lfe	lfe	PROPN
ajst-21816	55	13	are	be	AUX
ajst-21816	55	14	pooled	pool	VERB
ajst-21816	55	15	(	(	PUNCT
ajst-21816	55	16	sp	sp	NOUN
ajst-21816	55	17	)	)	PUNCT
ajst-21816	55	18	,	,	PUNCT
ajst-21816	55	19	and	and	CCONJ
ajst-21816	55	20	the	the	DET
ajst-21816	55	21	result	result	NOUN
ajst-21816	55	22	is	be	AUX
ajst-21816	55	23	input	input	NOUN
ajst-21816	55	24	as	as	ADP
ajst-21816	55	25	the	the	DET
ajst-21816	55	26	first	first	ADJ
ajst-21816	55	27	layer	layer	NOUN
ajst-21816	55	28	.	.	PUNCT
ajst-21816	56	1	secondly	secondly	ADV
ajst-21816	56	2	,	,	PUNCT
ajst-21816	56	3	mgpa	mgpa	PROPN
ajst-21816	56	4	aggregates	aggregate	VERB
ajst-21816	56	5	frame	frame	NOUN
ajst-21816	56	6	-	-	PUNCT
ajst-21816	56	7	level	level	NOUN
ajst-21816	56	8	features	feature	NOUN
ajst-21816	56	9	extracted	extract	VERB
ajst-21816	56	10	from	from	ADP
ajst-21816	56	11	sconv	sconv	NOUN
ajst-21816	56	12	block2	block2	PROPN
ajst-21816	56	13	in	in	ADP
ajst-21816	56	14	lfe	lfe	PROPN
ajst-21816	56	15	into	into	ADP
ajst-21816	56	16	set	set	NOUN
ajst-21816	56	17	-	-	PUNCT
ajst-21816	56	18	level	level	NOUN
ajst-21816	56	19	features	feature	VERB
ajst-21816	56	20	through	through	ADP
ajst-21816	56	21	sp	sp	ADP
ajst-21816	56	22	operation	operation	NOUN
ajst-21816	56	23	.	.	PUNCT
ajst-21816	57	1	the	the	DET
ajst-21816	57	2	features	feature	NOUN
ajst-21816	57	3	of	of	ADP
ajst-21816	57	4	the	the	DET
ajst-21816	57	5	set	set	VERB
ajst-21816	57	6	level	level	NOUN
ajst-21816	57	7	are	be	AUX
ajst-21816	57	8	fused	fuse	VERB
ajst-21816	57	9	with	with	ADP
ajst-21816	57	10	those	those	PRON
ajst-21816	57	11	of	of	ADP
ajst-21816	57	12	the	the	DET
ajst-21816	57	13	upper	upper	ADJ
ajst-21816	57	14	layer	layer	NOUN
ajst-21816	57	15	.	.	PUNCT
ajst-21816	58	1	finally	finally	ADV
ajst-21816	58	2	,	,	PUNCT
ajst-21816	58	3	the	the	DET
ajst-21816	58	4	convolution	convolution	NOUN
ajst-21816	58	5	feature	feature	NOUN
ajst-21816	58	6	between	between	ADP
ajst-21816	58	7	channels	channel	NOUN
ajst-21816	58	8	is	be	AUX
ajst-21816	58	9	extracted	extract	VERB
ajst-21816	58	10	by	by	ADP
ajst-21816	58	11	se	se	X
ajst-21816	58	12	block	block	NOUN
ajst-21816	58	13	to	to	PART
ajst-21816	58	14	correct	correct	VERB
ajst-21816	58	15	the	the	DET
ajst-21816	58	16	original	original	ADJ
ajst-21816	58	17	output	output	NOUN
ajst-21816	58	18	.	.	PUNCT
ajst-21816	59	1	the	the	DET
ajst-21816	59	2	third	third	ADJ
ajst-21816	59	3	layer	layer	NOUN
ajst-21816	59	4	is	be	AUX
ajst-21816	59	5	the	the	DET
ajst-21816	59	6	same	same	ADJ
ajst-21816	59	7	as	as	ADP
ajst-21816	59	8	the	the	DET
ajst-21816	59	9	second	second	ADJ
ajst-21816	59	10	layer	layer	NOUN
ajst-21816	59	11	structure	structure	NOUN
ajst-21816	59	12	with	with	ADP
ajst-21816	59	13	sp	sp	NOUN
ajst-21816	59	14	,	,	PUNCT
ajst-21816	59	15	feature	feature	NOUN
ajst-21816	59	16	fusion	fusion	NOUN
ajst-21816	59	17	and	and	CCONJ
ajst-21816	59	18	se	se	X
ajst-21816	59	19	block	block	VERB
ajst-21816	59	20	respectively	respectively	ADV
ajst-21816	59	21	.	.	PUNCT
ajst-21816	60	1	figure	figure	VERB
ajst-21816	60	2	3	3	NUM
ajst-21816	60	3	.	.	PUNCT
ajst-21816	60	4	se	se	X
ajst-21816	60	5	block	block	VERB
ajst-21816	60	6	overall	overall	ADJ
ajst-21816	60	7	structure	structure	NOUN
ajst-21816	60	8	the	the	DET
ajst-21816	60	9	set	set	NOUN
ajst-21816	60	10	level	level	NOUN
ajst-21816	60	11	features	feature	NOUN
ajst-21816	60	12	aggregated	aggregate	VERB
ajst-21816	60	13	by	by	ADP
ajst-21816	60	14	sp	sp	ADP
ajst-21816	60	15	operation	operation	NOUN
ajst-21816	60	16	are	be	AUX
ajst-21816	60	17	fused	fuse	VERB
ajst-21816	60	18	with	with	ADP
ajst-21816	60	19	the	the	DET
ajst-21816	60	20	features	feature	NOUN
ajst-21816	60	21	of	of	ADP
ajst-21816	60	22	the	the	DET
ajst-21816	60	23	previous	previous	ADJ
ajst-21816	60	24	layer	layer	NOUN
ajst-21816	60	25	,	,	PUNCT
ajst-21816	60	26	and	and	CCONJ
ajst-21816	60	27	the	the	DET
ajst-21816	60	28	fused	fuse	VERB
ajst-21816	60	29	features	feature	NOUN
ajst-21816	60	30	are	be	AUX
ajst-21816	60	31	represented	represent	VERB
ajst-21816	60	32	by	by	ADP
ajst-21816	60	33	x	x	PUNCT
ajst-21816	60	34	as	as	ADP
ajst-21816	60	35	input	input	NOUN
ajst-21816	60	36	.	.	PUNCT
ajst-21816	61	1	suppose	suppose	VERB
ajst-21816	61	2	1	1	NUM
ajst-21816	61	3	2	2	NUM
ajst-21816	61	4	[	[	PUNCT
ajst-21816	61	5	,	,	PUNCT
ajst-21816	61	6	,	,	PUNCT
ajst-21816	61	7	...	...	PUNCT
ajst-21816	61	8	,	,	PUNCT
ajst-21816	61	9	]	]	PUNCT
ajst-21816	61	10	cv	cv	PROPN
ajst-21816	61	11	v	v	NUM
ajst-21816	61	12	v	v	NOUN
ajst-21816	61	13	v	v	ADV
ajst-21816	61	14	represents	represent	VERB
ajst-21816	61	15	the	the	DET
ajst-21816	61	16	learned	learn	VERB
ajst-21816	61	17	set	set	NOUN
ajst-21816	61	18	of	of	ADP
ajst-21816	61	19	filters	filter	NOUN
ajst-21816	61	20	,	,	PUNCT
ajst-21816	61	21	where	where	SCONJ
ajst-21816	61	22	cv	cv	PROPN
ajst-21816	61	23	represents	represent	VERB
ajst-21816	61	24	the	the	DET
ajst-21816	61	25	parameter	parameter	NOUN
ajst-21816	61	26	of	of	ADP
ajst-21816	61	27	the	the	DET
ajst-21816	61	28	c	c	PROPN
ajst-21816	61	29	filter	filter	NOUN
ajst-21816	61	30	.	.	PUNCT
ajst-21816	62	1	trf	trf	PROPN
ajst-21816	62	2	is	be	AUX
ajst-21816	62	3	a	a	DET
ajst-21816	62	4	function	function	NOUN
ajst-21816	62	5	transformation	transformation	NOUN
ajst-21816	62	6	,	,	PUNCT
ajst-21816	62	7	x	x	PUNCT
ajst-21816	62	8	through	through	ADP
ajst-21816	62	9	trf	trf	PROPN
ajst-21816	62	10	output	output	NOUN
ajst-21816	62	11	1	1	NUM
ajst-21816	62	12	2	2	NUM
ajst-21816	62	13	[	[	PUNCT
ajst-21816	62	14	,	,	PUNCT
ajst-21816	62	15	,	,	PUNCT
ajst-21816	62	16	...	...	PUNCT
ajst-21816	62	17	,	,	PUNCT
ajst-21816	62	18	]	]	PUNCT
ajst-21816	62	19	cu	cu	NOUN
ajst-21816	62	20	u	u	X
ajst-21816	62	21	u	u	NOUN
ajst-21816	62	22	u	u	X
ajst-21816	62	23	.	.	PUNCT
ajst-21816	63	1	the	the	DET
ajst-21816	63	2	dimension	dimension	NOUN
ajst-21816	63	3	of	of	ADP
ajst-21816	63	4	x	x	PUNCT
ajst-21816	63	5	is	be	AUX
ajst-21816	63	6	'	'	PUNCT
ajst-21816	63	7	'	'	PUNCT
ajst-21816	63	8	'	'	PUNCT
ajst-21816	63	9	h	h	NOUN
ajst-21816	63	10	w	w	PROPN
ajst-21816	63	11	c	c	PROPN
ajst-21816	63	12			NOUN
ajst-21816	63	13	,	,	PUNCT
ajst-21816	63	14	and	and	CCONJ
ajst-21816	63	15	the	the	DET
ajst-21816	63	16	dimension	dimension	NOUN
ajst-21816	63	17	of	of	ADP
ajst-21816	63	18	u	u	NOUN
ajst-21816	63	19	ish	ish	VERB
ajst-21816	63	20	w	w	PROPN
ajst-21816	63	21	c	c	PROPN
ajst-21816	63	22			PROPN
ajst-21816	63	23	.	.	PUNCT
ajst-21816	64	1	the	the	DET
ajst-21816	64	2	formula	formula	NOUN
ajst-21816	64	3	for	for	ADP
ajst-21816	64	4	trf	trf	NOUN
ajst-21816	64	5	is	be	AUX
ajst-21816	64	6	as	as	SCONJ
ajst-21816	64	7	follows	follow	VERB
ajst-21816	64	8	:	:	PUNCT
ajst-21816	64	9	(	(	PUNCT
ajst-21816	64	10	1	1	NUM
ajst-21816	64	11	)	)	PUNCT
ajst-21816	64	12	.	.	PUNCT
ajst-21816	65	1	where	where	SCONJ
ajst-21816	65	2	*	*	PUNCT
ajst-21816	65	3	represents	represent	VERB
ajst-21816	65	4	convolution	convolution	NOUN
ajst-21816	65	5	,	,	PUNCT
ajst-21816	65	6	1	1	NUM
ajst-21816	65	7	2	2	NUM
ajst-21816	65	8	'	'	NOUN
ajst-21816	65	9	[	[	PUNCT
ajst-21816	65	10	,	,	PUNCT
ajst-21816	65	11	,	,	PUNCT
ajst-21816	65	12	...	...	PUNCT
ajst-21816	65	13	,	,	PUNCT
ajst-21816	65	14	]	]	X
ajst-21816	65	15	c	c	X
ajst-21816	65	16	c	c	NOUN
ajst-21816	65	17	c	c	PROPN
ajst-21816	65	18	c	c	PROPN
ajst-21816	65	19	cv	cv	PROPN
ajst-21816	65	20	v	v	X
ajst-21816	65	21	v	v	ADP
ajst-21816	65	22	v	v	ADV
ajst-21816	65	23	,	,	PUNCT
ajst-21816	65	24	1	1	NUM
ajst-21816	65	25	2	2	NUM
ajst-21816	65	26	'	'	NOUN
ajst-21816	65	27	[	[	PUNCT
ajst-21816	65	28	,	,	PUNCT
ajst-21816	65	29	,	,	PUNCT
ajst-21816	65	30	...	...	PUNCT
ajst-21816	65	31	,	,	PUNCT
ajst-21816	65	32	]	]	X
ajst-21816	65	33	cx	cx	X
ajst-21816	65	34	x	x	PUNCT
ajst-21816	65	35	x	x	X
ajst-21816	65	36	x	x	X
ajst-21816	65	37	,	,	PUNCT
ajst-21816	65	38	and	and	CCONJ
ajst-21816	65	39	the	the	DET
ajst-21816	65	40	offset	offset	ADJ
ajst-21816	65	41	term	term	NOUN
ajst-21816	65	42	is	be	AUX
ajst-21816	65	43	omitted	omit	VERB
ajst-21816	65	44	here	here	ADV
ajst-21816	65	45	for	for	ADP
ajst-21816	65	46	convenience	convenience	NOUN
ajst-21816	65	47	.	.	PUNCT
ajst-21816	66	1	s	s	PART
ajst-21816	67	1	cv	cv	PROPN
ajst-21816	67	2	is	be	AUX
ajst-21816	67	3	a	a	DET
ajst-21816	67	4	2d	2d	NUM
ajst-21816	67	5	space	space	NOUN
ajst-21816	67	6	kernel	kernel	NOUN
ajst-21816	67	7	,	,	PUNCT
ajst-21816	67	8	and	and	CCONJ
ajst-21816	67	9	cv	cv	PROPN
ajst-21816	67	10	represents	represent	VERB
ajst-21816	67	11	some	some	DET
ajst-21816	67	12	channel	channel	NOUN
ajst-21816	67	13	acting	act	VERB
ajst-21816	67	14	on	on	ADP
ajst-21816	67	15	the	the	DET
ajst-21816	67	16	corresponding	correspond	VERB
ajst-21816	67	17	channel	channel	NOUN
ajst-21816	67	18	in	in	ADP
ajst-21816	67	19	x	x	X
ajst-21816	67	20	.	.	PUNCT
ajst-21816	67	21	'	'	PUNCT
ajst-21816	68	1	1	1	NUM
ajst-21816	68	2	u	u	NOUN
ajst-21816	68	3	*	*	PUNCT
ajst-21816	68	4	*	*	PUNCT
ajst-21816	68	5	c	c	X
ajst-21816	68	6	s	s	X
ajst-21816	68	7	s	s	X
ajst-21816	68	8	c	c	NOUN
ajst-21816	68	9	c	c	NOUN
ajst-21816	68	10	c	c	NOUN
ajst-21816	68	11	s	s	PROPN
ajst-21816	68	12	v	v	NOUN
ajst-21816	68	13	x	x	SYM
ajst-21816	68	14	v	v	NOUN
ajst-21816	68	15	x	x	NOUN
ajst-21816	68	16			NUM
ajst-21816	68	17			NUM
ajst-21816	68	18			X
ajst-21816	68	19	(	(	PUNCT
ajst-21816	68	20	1	1	NUM
ajst-21816	68	21	)	)	PUNCT
ajst-21816	68	22	se	se	X
ajst-21816	68	23	block	block	NOUN
ajst-21816	68	24	proposes	propose	VERB
ajst-21816	68	25	a	a	DET
ajst-21816	68	26	branching	branch	VERB
ajst-21816	68	27	structure	structure	NOUN
ajst-21816	68	28	with	with	ADP
ajst-21816	68	29	two	two	NUM
ajst-21816	68	30	parts	part	NOUN
ajst-21816	68	31	:	:	PUNCT
ajst-21816	68	32	squeeze	squeeze	NOUN
ajst-21816	68	33	and	and	CCONJ
ajst-21816	68	34	excitation	excitation	NOUN
ajst-21816	68	35	.	.	PUNCT
ajst-21816	69	1	the	the	DET
ajst-21816	69	2	squeeze	squeeze	NOUN
ajst-21816	69	3	operation	operation	NOUN
ajst-21816	69	4	corresponds	correspond	VERB
ajst-21816	69	5	to	to	ADP
ajst-21816	69	6	the	the	DET
ajst-21816	69	7	function	function	NOUN
ajst-21816	69	8	in	in	ADP
ajst-21816	69	9	the	the	DET
ajst-21816	69	10	figure	figure	NOUN
ajst-21816	69	11	.	.	PUNCT
ajst-21816	70	1	in	in	ADP
ajst-21816	70	2	order	order	NOUN
ajst-21816	70	3	to	to	PART
ajst-21816	70	4	make	make	VERB
ajst-21816	70	5	full	full	ADJ
ajst-21816	70	6	use	use	NOUN
ajst-21816	70	7	of	of	ADP
ajst-21816	70	8	the	the	DET
ajst-21816	70	9	relationship	relationship	NOUN
ajst-21816	70	10	between	between	ADP
ajst-21816	70	11	channels	channel	NOUN
ajst-21816	70	12	,	,	PUNCT
ajst-21816	70	13	each	each	DET
ajst-21816	70	14	channel	channel	NOUN
ajst-21816	70	15	is	be	AUX
ajst-21816	70	16	compressed	compress	VERB
ajst-21816	70	17	,	,	PUNCT
ajst-21816	70	18	and	and	CCONJ
ajst-21816	70	19	its	its	PRON
ajst-21816	70	20	formula	formula	NOUN
ajst-21816	70	21	is	be	AUX
ajst-21816	70	22	shown	show	VERB
ajst-21816	70	23	as	as	ADP
ajst-21816	70	24	equation	equation	NOUN
ajst-21816	70	25	(	(	PUNCT
ajst-21816	70	26	2	2	NUM
ajst-21816	70	27	)	)	PUNCT
ajst-21816	70	28	.	.	PUNCT
ajst-21816	71	1	where	where	SCONJ
ajst-21816	71	2	is	be	AUX
ajst-21816	71	3	the	the	DET
ajst-21816	71	4	c	c	PROPN
ajst-21816	71	5	channel	channel	NOUN
ajst-21816	71	6	in	in	ADP
ajst-21816	71	7	feature	feature	NOUN
ajst-21816	71	8	u	u	NOUN
ajst-21816	71	9	,	,	PUNCT
ajst-21816	71	10	representing	represent	VERB
ajst-21816	71	11	the	the	DET
ajst-21816	71	12	c	c	PROPN
ajst-21816	71	13	vector	vector	NOUN
ajst-21816	71	14	after	after	ADP
ajst-21816	71	15	compression	compression	NOUN
ajst-21816	71	16	.	.	PUNCT
ajst-21816	72	1	it	it	PRON
ajst-21816	72	2	is	be	AUX
ajst-21816	72	3	not	not	PART
ajst-21816	72	4	difficult	difficult	ADJ
ajst-21816	72	5	to	to	PART
ajst-21816	72	6	see	see	VERB
ajst-21816	72	7	from	from	ADP
ajst-21816	72	8	the	the	DET
ajst-21816	72	9	formula	formula	NOUN
ajst-21816	72	10	that	that	PRON
ajst-21816	72	11	each	each	DET
ajst-21816	72	12	feature	feature	NOUN
ajst-21816	72	13	mapping	mapping	NOUN
ajst-21816	72	14	channel	channel	NOUN
ajst-21816	72	15	with	with	ADP
ajst-21816	72	16	size	size	NOUN
ajst-21816	72	17	h×w	h×w	PROPN
ajst-21816	72	18	is	be	AUX
ajst-21816	72	19	compressed	compress	VERB
ajst-21816	72	20	into	into	ADP
ajst-21816	72	21	a	a	DET
ajst-21816	72	22	pixel	pixel	NOUN
ajst-21816	72	23	point	point	NOUN
ajst-21816	72	24	.	.	PUNCT
ajst-21816	73	1	thus	thus	ADV
ajst-21816	73	2	,	,	PUNCT
ajst-21816	73	3	a	a	DET
ajst-21816	73	4	feature	feature	NOUN
ajst-21816	73	5	map	map	NOUN
ajst-21816	73	6	of	of	ADP
ajst-21816	73	7	size	size	NOUN
ajst-21816	73	8	h×w×c	h×w×c	NOUN
ajst-21816	73	9	,	,	PUNCT
ajst-21816	73	10	after	after	ADP
ajst-21816	73	11	passing	pass	VERB
ajst-21816	73	12	through	through	ADV
ajst-21816	73	13	,	,	PUNCT
ajst-21816	73	14	outputs	output	VERB
ajst-21816	73	15	a	a	DET
ajst-21816	73	16	vector	vector	NOUN
ajst-21816	73	17	of	of	ADP
ajst-21816	73	18	1×1×c	1×1×c	NUM
ajst-21816	73	19	.	.	PUNCT
ajst-21816	74	1	i	i	PRON
ajst-21816	74	2	1	1	NUM
ajst-21816	74	3	1	1	NUM
ajst-21816	74	4	1	1	NUM
ajst-21816	74	5	(	(	PUNCT
ajst-21816	74	6	)	)	PUNCT
ajst-21816	74	7	(	(	PUNCT
ajst-21816	74	8	,	,	PUNCT
ajst-21816	74	9	)	)	PUNCT
ajst-21816	74	10	×w	×w	NOUN
ajst-21816	74	11	h	h	PROPN
ajst-21816	74	12	w	w	PROPN
ajst-21816	74	13	c	c	NOUN
ajst-21816	74	14	sq	sq	INTJ
ajst-21816	74	15	c	c	PROPN
ajst-21816	74	16	c	c	PROPN
ajst-21816	74	17	j	j	PROPN
ajst-21816	75	1	z	z	NOUN
ajst-21816	75	2	f	f	PROPN
ajst-21816	75	3	u	u	X
ajst-21816	75	4	u	u	NOUN
ajst-21816	76	1	i	i	NOUN
ajst-21816	76	2	j	j	PROPN
ajst-21816	77	1	h	h	NOUN
ajst-21816	78	1			NOUN
ajst-21816	79	1			PRON
ajst-21816	80	1			ADJ
ajst-21816	81	1			NOUN
ajst-21816	81	2			PRON
ajst-21816	81	3	(	(	PUNCT
ajst-21816	81	4	2	2	X
ajst-21816	81	5	)	)	PUNCT
ajst-21816	81	6	the	the	DET
ajst-21816	81	7	excitation	excitation	NOUN
ajst-21816	81	8	operation	operation	NOUN
ajst-21816	81	9	corresponds	correspond	VERB
ajst-21816	81	10	to	to	ADP
ajst-21816	81	11	the	the	DET
ajst-21816	81	12	exf	exf	NOUN
ajst-21816	81	13	function	function	NOUN
ajst-21816	81	14	in	in	ADP
ajst-21816	81	15	the	the	DET
ajst-21816	81	16	figure	figure	NOUN
ajst-21816	81	17	.	.	PUNCT
ajst-21816	82	1	after	after	SCONJ
ajst-21816	82	2	the	the	DET
ajst-21816	82	3	channel	channel	NOUN
ajst-21816	82	4	features	feature	NOUN
ajst-21816	82	5	are	be	AUX
ajst-21816	82	6	compressed	compress	VERB
ajst-21816	82	7	by	by	ADP
ajst-21816	82	8	squeeze	squeeze	NOUN
ajst-21816	82	9	operation	operation	NOUN
ajst-21816	82	10	,	,	PUNCT
ajst-21816	82	11	the	the	DET
ajst-21816	82	12	excitation	excitation	NOUN
ajst-21816	82	13	operation	operation	NOUN
ajst-21816	82	14	is	be	AUX
ajst-21816	82	15	needed	need	VERB
ajst-21816	82	16	to	to	PART
ajst-21816	82	17	capture	capture	VERB
ajst-21816	82	18	the	the	DET
ajst-21816	82	19	dependency	dependency	NOUN
ajst-21816	82	20	between	between	ADP
ajst-21816	82	21	channels	channel	NOUN
ajst-21816	82	22	.	.	PUNCT
ajst-21816	83	1	in	in	ADP
ajst-21816	83	2	view	view	NOUN
ajst-21816	83	3	of	of	ADP
ajst-21816	83	4	the	the	DET
ajst-21816	83	5	above	above	ADJ
ajst-21816	83	6	,	,	PUNCT
ajst-21816	83	7	the	the	DET
ajst-21816	83	8	function	function	NOUN
ajst-21816	83	9	needs	need	VERB
ajst-21816	83	10	to	to	PART
ajst-21816	83	11	satisfy	satisfy	VERB
ajst-21816	83	12	two	two	NUM
ajst-21816	83	13	conditions	condition	NOUN
ajst-21816	83	14	:	:	PUNCT
ajst-21816	83	15	first	first	ADV
ajst-21816	83	16	,	,	PUNCT
ajst-21816	83	17	the	the	DET
ajst-21816	83	18	function	function	NOUN
ajst-21816	83	19	must	must	AUX
ajst-21816	83	20	be	be	AUX
ajst-21816	83	21	able	able	ADJ
ajst-21816	83	22	to	to	PART
ajst-21816	83	23	learn	learn	VERB
ajst-21816	83	24	nonlinear	nonlinear	ADJ
ajst-21816	83	25	relationships	relationship	NOUN
ajst-21816	83	26	between	between	ADP
ajst-21816	83	27	channels	channel	NOUN
ajst-21816	83	28	;	;	PUNCT
ajst-21816	83	29	second	second	X
ajst-21816	83	30	,	,	PUNCT
ajst-21816	83	31	the	the	DET
ajst-21816	83	32	function	function	NOUN
ajst-21816	83	33	must	must	AUX
ajst-21816	83	34	be	be	AUX
ajst-21816	83	35	able	able	ADJ
ajst-21816	83	36	to	to	PART
ajst-21816	83	37	learn	learn	VERB
ajst-21816	83	38	mutually	mutually	ADV
ajst-21816	83	39	exclusive	exclusive	ADJ
ajst-21816	83	40	relationships	relationship	NOUN
ajst-21816	83	41	between	between	ADP
ajst-21816	83	42	channels	channel	NOUN
ajst-21816	83	43	.	.	PUNCT
ajst-21816	84	1	the	the	DET
ajst-21816	84	2	function	function	NOUN
ajst-21816	84	3	formula	formula	NOUN
ajst-21816	84	4	is	be	AUX
ajst-21816	84	5	shown	show	VERB
ajst-21816	84	6	in	in	ADP
ajst-21816	84	7	equation	equation	NOUN
ajst-21816	84	8	(	(	PUNCT
ajst-21816	84	9	3	3	NUM
ajst-21816	84	10	)	)	PUNCT
ajst-21816	84	11	.	.	PUNCT
ajst-21816	85	1	where	where	SCONJ
ajst-21816	85	2	δ	δ	PROPN
ajst-21816	85	3	represents	represent	VERB
ajst-21816	85	4	the	the	DET
ajst-21816	85	5	relu	relu	NOUN
ajst-21816	85	6	function	function	NOUN
ajst-21816	85	7	,	,	PUNCT
ajst-21816	85	8	1w	1w	NUM
ajst-21816	85	9	and	and	CCONJ
ajst-21816	85	10	2w	2w	NUM
ajst-21816	85	11	are	be	AUX
ajst-21816	85	12	the	the	DET
ajst-21816	85	13	weight	weight	NOUN
ajst-21816	85	14	matrix	matrix	NOUN
ajst-21816	85	15	,	,	PUNCT
ajst-21816	85	16	(	(	PUNCT
ajst-21816	85	17	/	/	SYM
ajst-21816	85	18	)	)	PUNCT
ajst-21816	85	19	1	1	NUM
ajst-21816	85	20	c	c	NOUN
ajst-21816	85	21	r	r	NOUN
ajst-21816	85	22	cw	cw	NOUN
ajst-21816	85	23	r	r	NOUN
ajst-21816	85	24			NOUN
ajst-21816	85	25	,	,	PUNCT
ajst-21816	85	26	(	(	PUNCT
ajst-21816	85	27	/	/	SYM
ajst-21816	85	28	)	)	PUNCT
ajst-21816	85	29	2	2	NUM
ajst-21816	85	30	c	c	NOUN
ajst-21816	85	31	c	c	NOUN
ajst-21816	85	32	rw	rw	PROPN
ajst-21816	85	33	r	r	NOUN
ajst-21816	85	34			NOUN
ajst-21816	85	35	.in	.in	PUNCT
ajst-21816	85	36	simple	simple	ADJ
ajst-21816	85	37	terms	term	NOUN
ajst-21816	85	38	,	,	PUNCT
ajst-21816	85	39	excitation	excitation	NOUN
ajst-21816	85	40	is	be	AUX
ajst-21816	85	41	the	the	DET
ajst-21816	85	42	two	two	NUM
ajst-21816	85	43	fully	fully	ADV
ajst-21816	85	44	connected	connected	ADJ
ajst-21816	85	45	layers	layer	NOUN
ajst-21816	85	46	.	.	PUNCT
ajst-21816	86	1	the	the	DET
ajst-21816	86	2	first	first	ADJ
ajst-21816	86	3	layer	layer	NOUN
ajst-21816	86	4	is	be	AUX
ajst-21816	86	5	the	the	DET
ajst-21816	86	6	dimensionality	dimensionality	NOUN
ajst-21816	86	7	reduction	reduction	NOUN
ajst-21816	86	8	layer	layer	NOUN
ajst-21816	86	9	containing	contain	VERB
ajst-21816	86	10	parameters	parameter	NOUN
ajst-21816	86	11	,	,	PUNCT
ajst-21816	86	12	transforming	transform	VERB
ajst-21816	86	13	a	a	DET
ajst-21816	86	14	1×1×c	1×1×c	NUM
ajst-21816	86	15	vector	vector	NOUN
ajst-21816	86	16	into	into	ADP
ajst-21816	86	17	a	a	DET
ajst-21816	86	18	1×1×c	1×1×c	NUM
ajst-21816	86	19	/	/	SYM
ajst-21816	86	20	r	r	NOUN
ajst-21816	86	21	vector	vector	NOUN
ajst-21816	86	22	.	.	PUNCT
ajst-21816	87	1	the	the	DET
ajst-21816	87	2	excitation	excitation	NOUN
ajst-21816	87	3	layer	layer	NOUN
ajst-21816	87	4	is	be	AUX
ajst-21816	87	5	then	then	ADV
ajst-21816	87	6	connected	connect	VERB
ajst-21816	87	7	to	to	ADP
ajst-21816	87	8	the	the	DET
ajst-21816	87	9	relu	relu	NOUN
ajst-21816	87	10	.	.	PUNCT
ajst-21816	88	1	the	the	DET
ajst-21816	88	2	second	second	ADJ
ajst-21816	88	3	layer	layer	NOUN
ajst-21816	88	4	is	be	AUX
ajst-21816	88	5	an	an	DET
ajst-21816	88	6	ascending	ascend	VERB
ajst-21816	88	7	layer	layer	NOUN
ajst-21816	88	8	containing	contain	VERB
ajst-21816	88	9	parameters	parameter	NOUN
ajst-21816	88	10	,	,	PUNCT
ajst-21816	88	11	which	which	PRON
ajst-21816	88	12	restores	restore	VERB
ajst-21816	88	13	the	the	DET
ajst-21816	88	14	1×1×c	1×1×c	PROPN
ajst-21816	88	15	/	/	SYM
ajst-21816	88	16	r	r	NOUN
ajst-21816	88	17	vector	vector	NOUN
ajst-21816	88	18	to	to	ADP
ajst-21816	88	19	1×1×c	1×1×c	NUM
ajst-21816	88	20	vector	vector	NOUN
ajst-21816	88	21	,	,	PUNCT
ajst-21816	88	22	followed	follow	VERB
ajst-21816	88	23	by	by	ADP
ajst-21816	88	24	the	the	DET
ajst-21816	88	25	activation	activation	NOUN
ajst-21816	88	26	layer	layer	NOUN
ajst-21816	88	27	of	of	ADP
ajst-21816	88	28	sigmoid	sigmoid	NOUN
ajst-21816	88	29	.	.	PUNCT
ajst-21816	89	1	2	2	NUM
ajst-21816	89	2	1	1	NUM
ajst-21816	89	3	(	(	PUNCT
ajst-21816	89	4	,	,	PUNCT
ajst-21816	89	5	)	)	PUNCT
ajst-21816	89	6	(	(	PUNCT
ajst-21816	89	7	,	,	PUNCT
ajst-21816	89	8	)	)	PUNCT
ajst-21816	89	9	(	(	PUNCT
ajst-21816	89	10	)	)	PUNCT
ajst-21816	89	11	exs	exs	PROPN
ajst-21816	90	1	f	f	PROPN
ajst-21816	90	2	z	z	PROPN
ajst-21816	90	3	w	w	PROPN
ajst-21816	90	4	g	g	PROPN
ajst-21816	90	5	z	z	PROPN
ajst-21816	90	6	w	w	PROPN
ajst-21816	90	7	w	w	PROPN
ajst-21816	90	8	w	w	PROPN
ajst-21816	90	9	z	z	PROPN
ajst-21816	90	10			NUM
ajst-21816	90	11			PROPN
ajst-21816	90	12			PROPN
ajst-21816	90	13			NUM
ajst-21816	90	14			NOUN
ajst-21816	90	15			NUM
ajst-21816	90	16			PROPN
ajst-21816	90	17	(	(	PUNCT
ajst-21816	90	18	3	3	NUM
ajst-21816	90	19	)	)	PUNCT
ajst-21816	90	20	the	the	DET
ajst-21816	90	21	results	result	NOUN
ajst-21816	90	22	obtained	obtain	VERB
ajst-21816	90	23	by	by	ADP
ajst-21816	90	24	sqf	sqf	NOUN
ajst-21816	90	25	and	and	CCONJ
ajst-21816	90	26	exf	exf	NOUN
ajst-21816	90	27	functions	function	NOUN
ajst-21816	90	28	need	need	VERB
ajst-21816	90	29	to	to	PART
ajst-21816	90	30	be	be	AUX
ajst-21816	90	31	merged	merge	VERB
ajst-21816	90	32	with	with	ADP
ajst-21816	90	33	the	the	DET
ajst-21816	90	34	original	original	ADJ
ajst-21816	90	35	feature	feature	NOUN
ajst-21816	90	36	u	u	NOUN
ajst-21816	90	37	,	,	PUNCT
ajst-21816	90	38	which	which	PRON
ajst-21816	90	39	corresponds	correspond	VERB
ajst-21816	90	40	to	to	ADP
ajst-21816	90	41	the	the	DET
ajst-21816	90	42	scalef	scalef	NOUN
ajst-21816	90	43	function	function	VERB
ajst-21816	90	44	in	in	ADP
ajst-21816	90	45	figure	figure	NOUN
ajst-21816	90	46	3	3	NUM
ajst-21816	90	47	,	,	PUNCT
ajst-21816	90	48	whose	whose	DET
ajst-21816	90	49	formula	formula	NOUN
ajst-21816	90	50	is	be	AUX
ajst-21816	90	51	shown	show	VERB
ajst-21816	90	52	in	in	ADP
ajst-21816	90	53	equation	equation	NOUN
ajst-21816	90	54	(	(	PUNCT
ajst-21816	90	55	4	4	NUM
ajst-21816	90	56	)	)	PUNCT
ajst-21816	90	57	.	.	PUNCT
ajst-21816	91	1	let	let	VERB
ajst-21816	91	2	1	1	NUM
ajst-21816	91	3	2	2	NUM
ajst-21816	91	4	[	[	PUNCT
ajst-21816	91	5	,	,	PUNCT
ajst-21816	91	6	,	,	PUNCT
ajst-21816	91	7	...	...	PUNCT
ajst-21816	91	8	,	,	PUNCT
ajst-21816	91	9	]	]	X
ajst-21816	91	10	cx	cx	X
ajst-21816	91	11	x	x	PUNCT
ajst-21816	91	12	x	x	X
ajst-21816	91	13	x	x	PROPN
ajst-21816	91	14			PROPN
ajst-21816	91	15			PROPN
ajst-21816	91	16			NOUN
ajst-21816	91	17	be	be	AUX
ajst-21816	91	18	the	the	DET
ajst-21816	91	19	final	final	ADJ
ajst-21816	91	20	output	output	NOUN
ajst-21816	91	21	of	of	ADP
ajst-21816	91	22	the	the	DET
ajst-21816	91	23	se	se	PROPN
ajst-21816	91	24	block	block	NOUN
ajst-21816	91	25	and	and	CCONJ
ajst-21816	91	26	s	s	AUX
ajst-21816	91	27	be	be	AUX
ajst-21816	91	28	the	the	DET
ajst-21816	91	29	weight	weight	NOUN
ajst-21816	91	30	vector	vector	NOUN
ajst-21816	91	31	of	of	ADP
ajst-21816	91	32	the	the	DET
ajst-21816	91	33	channel	channel	NOUN
ajst-21816	91	34	.	.	PUNCT
ajst-21816	92	1	cx	cx	NOUN
ajst-21816	92	2	represents	represent	VERB
ajst-21816	92	3	the	the	DET
ajst-21816	92	4	output	output	NOUN
ajst-21816	92	5	of	of	ADP
ajst-21816	92	6	channel	channel	PROPN
ajst-21816	92	7	c.	c.	PROPN
ajst-21816	92	8	scalef	scalef	PROPN
ajst-21816	92	9	represents	represent	VERB
ajst-21816	92	10	the	the	DET
ajst-21816	92	11	channel	channel	NOUN
ajst-21816	92	12	-	-	PUNCT
ajst-21816	92	13	level	level	NOUN
ajst-21816	92	14	product	product	NOUN
ajst-21816	92	15	of	of	ADP
ajst-21816	92	16	feature	feature	NOUN
ajst-21816	92	17	cu	cu	NOUN
ajst-21816	92	18	and	and	CCONJ
ajst-21816	92	19	scalar	scalar	ADJ
ajst-21816	92	20	cs	cs	PROPN
ajst-21816	92	21	.	.	PUNCT
ajst-21816	93	1	in	in	ADP
ajst-21816	93	2	popular	popular	ADJ
ajst-21816	93	3	terms	term	NOUN
ajst-21816	93	4	,	,	PUNCT
ajst-21816	93	5	it	it	PRON
ajst-21816	93	6	is	be	AUX
ajst-21816	93	7	to	to	PART
ajst-21816	93	8	multiply	multiply	VERB
ajst-21816	93	9	each	each	DET
ajst-21816	93	10	element	element	NOUN
ajst-21816	93	11	in	in	ADP
ajst-21816	93	12	s	s	PRON
ajst-21816	93	13	with	with	ADP
ajst-21816	93	14	the	the	DET
ajst-21816	93	15	channel	channel	NOUN
ajst-21816	93	16	corresponding	correspond	VERB
ajst-21816	93	17	to	to	ADP
ajst-21816	93	18	the	the	DET
ajst-21816	93	19	feature	feature	NOUN
ajst-21816	93	20	u	u	NOUN
ajst-21816	93	21	by	by	ADP
ajst-21816	93	22	the	the	DET
ajst-21816	93	23	number	number	NOUN
ajst-21816	93	24	of	of	ADP
ajst-21816	93	25	channels	channel	NOUN
ajst-21816	93	26	.	.	PUNCT
ajst-21816	94	1	(	(	PUNCT
ajst-21816	94	2	,	,	PUNCT
ajst-21816	94	3	)	)	PUNCT
ajst-21816	94	4	c	c	NOUN
ajst-21816	94	5	scale	scale	NOUN
ajst-21816	94	6	c	c	NOUN
ajst-21816	94	7	c	c	NOUN
ajst-21816	94	8	c	c	NOUN
ajst-21816	94	9	cx	cx	PROPN
ajst-21816	95	1	f	f	PROPN
ajst-21816	95	2	u	u	PROPN
ajst-21816	95	3	s	s	PROPN
ajst-21816	95	4	s	s	X
ajst-21816	95	5	u	u	X
ajst-21816	95	6			PROPN
ajst-21816	95	7			NOUN
ajst-21816	95	8	(	(	PUNCT
ajst-21816	95	9	4	4	NUM
ajst-21816	95	10	)	)	PUNCT
ajst-21816	95	11	in	in	ADP
ajst-21816	95	12	this	this	DET
ajst-21816	95	13	study	study	NOUN
ajst-21816	95	14	,	,	PUNCT
ajst-21816	95	15	in	in	ADP
ajst-21816	95	16	order	order	NOUN
ajst-21816	95	17	to	to	PART
ajst-21816	95	18	better	well	ADV
ajst-21816	95	19	control	control	VERB
ajst-21816	95	20	the	the	DET
ajst-21816	95	21	size	size	NOUN
ajst-21816	95	22	of	of	ADP
ajst-21816	95	23	feature	feature	NOUN
ajst-21816	95	24	maps	map	NOUN
ajst-21816	95	25	,	,	PUNCT
ajst-21816	95	26	adaptive	adaptive	ADJ
ajst-21816	95	27	pooling	pool	VERB
ajst-21816	95	28	[	[	X
ajst-21816	95	29	]	]	X
ajst-21816	95	30	is	be	AUX
ajst-21816	95	31	adopted	adopt	VERB
ajst-21816	95	32	to	to	PART
ajst-21816	95	33	achieve	achieve	VERB
ajst-21816	95	34	squeeze	squeeze	NOUN
ajst-21816	95	35	operation	operation	NOUN
ajst-21816	95	36	.	.	PUNCT
ajst-21816	96	1	the	the	DET
ajst-21816	96	2	core	core	ADJ
ajst-21816	96	3	idea	idea	NOUN
ajst-21816	96	4	of	of	ADP
ajst-21816	96	5	this	this	DET
ajst-21816	96	6	strategy	strategy	NOUN
ajst-21816	96	7	is	be	AUX
ajst-21816	96	8	that	that	SCONJ
ajst-21816	96	9	,	,	PUNCT
ajst-21816	96	10	for	for	ADP
ajst-21816	96	11	inputs	input	NOUN
ajst-21816	96	12	of	of	ADP
ajst-21816	96	13	different	different	ADJ
ajst-21816	96	14	sizes	size	NOUN
ajst-21816	96	15	,	,	PUNCT
ajst-21816	96	16	adaptive	adaptive	ADJ
ajst-21816	96	17	pooling	pooling	NOUN
ajst-21816	96	18	can	can	AUX
ajst-21816	96	19	automatically	automatically	ADV
ajst-21816	96	20	average	average	VERB
ajst-21816	96	21	289	289	NUM
ajst-21816	96	22	each	each	DET
ajst-21816	96	23	position	position	NOUN
ajst-21816	96	24	,	,	PUNCT
ajst-21816	96	25	so	so	SCONJ
ajst-21816	96	26	as	as	SCONJ
ajst-21816	96	27	to	to	PART
ajst-21816	96	28	achieve	achieve	VERB
ajst-21816	96	29	accurate	accurate	ADJ
ajst-21816	96	30	processing	processing	NOUN
ajst-21816	96	31	of	of	ADP
ajst-21816	96	32	feature	feature	NOUN
ajst-21816	96	33	maps	map	NOUN
ajst-21816	96	34	.	.	PUNCT
ajst-21816	97	1	compared	compare	VERB
ajst-21816	97	2	with	with	ADP
ajst-21816	97	3	the	the	DET
ajst-21816	97	4	traditional	traditional	ADJ
ajst-21816	97	5	average	average	ADJ
ajst-21816	97	6	pooling	pooling	NOUN
ajst-21816	97	7	method	method	NOUN
ajst-21816	97	8	,	,	PUNCT
ajst-21816	97	9	this	this	DET
ajst-21816	97	10	technique	technique	NOUN
ajst-21816	97	11	does	do	AUX
ajst-21816	97	12	not	not	PART
ajst-21816	97	13	need	need	VERB
ajst-21816	97	14	to	to	PART
ajst-21816	97	15	specify	specify	VERB
ajst-21816	97	16	the	the	DET
ajst-21816	97	17	size	size	NOUN
ajst-21816	97	18	of	of	ADP
ajst-21816	97	19	the	the	DET
ajst-21816	97	20	pooling	pool	VERB
ajst-21816	97	21	kernel	kernel	NOUN
ajst-21816	97	22	,	,	PUNCT
ajst-21816	97	23	but	but	CCONJ
ajst-21816	97	24	adaptively	adaptively	ADV
ajst-21816	97	25	adjusts	adjust	VERB
ajst-21816	97	26	according	accord	VERB
ajst-21816	97	27	to	to	ADP
ajst-21816	97	28	the	the	DET
ajst-21816	97	29	size	size	NOUN
ajst-21816	97	30	of	of	ADP
ajst-21816	97	31	the	the	DET
ajst-21816	97	32	required	require	VERB
ajst-21816	97	33	output	output	NOUN
ajst-21816	97	34	,	,	PUNCT
ajst-21816	97	35	and	and	CCONJ
ajst-21816	97	36	the	the	DET
ajst-21816	97	37	output	output	NOUN
ajst-21816	97	38	size	size	NOUN
ajst-21816	97	39	can	can	AUX
ajst-21816	97	40	be	be	AUX
ajst-21816	97	41	manually	manually	ADV
ajst-21816	97	42	specified	specify	VERB
ajst-21816	97	43	according	accord	VERB
ajst-21816	97	44	to	to	ADP
ajst-21816	97	45	the	the	DET
ajst-21816	97	46	demand	demand	NOUN
ajst-21816	97	47	.	.	PUNCT
ajst-21816	98	1	when	when	SCONJ
ajst-21816	98	2	adaptive	adaptive	ADJ
ajst-21816	98	3	pooling	pooling	NOUN
ajst-21816	98	4	is	be	AUX
ajst-21816	98	5	adopted	adopt	VERB
ajst-21816	98	6	,	,	PUNCT
ajst-21816	98	7	the	the	DET
ajst-21816	98	8	specific	specific	ADJ
ajst-21816	98	9	application	application	NOUN
ajst-21816	98	10	scheme	scheme	NOUN
ajst-21816	98	11	of	of	ADP
ajst-21816	98	12	se	se	X
ajst-21816	98	13	block	block	NOUN
ajst-21816	98	14	in	in	ADP
ajst-21816	98	15	this	this	DET
ajst-21816	98	16	paper	paper	NOUN
ajst-21816	98	17	is	be	AUX
ajst-21816	98	18	shown	show	VERB
ajst-21816	98	19	in	in	ADP
ajst-21816	98	20	figure	figure	NOUN
ajst-21816	98	21	4	4	NUM
ajst-21816	98	22	.	.	PUNCT
ajst-21816	98	23	figure	figure	VERB
ajst-21816	98	24	4	4	NUM
ajst-21816	98	25	.	.	PUNCT
ajst-21816	98	26	se	se	X
ajst-21816	98	27	usage	usage	NOUN
ajst-21816	98	28	scheme	scheme	NOUN
ajst-21816	98	29	lfe	lfe	PROPN
ajst-21816	98	30	extracts	extract	VERB
ajst-21816	98	31	the	the	DET
ajst-21816	98	32	gait	gait	ADJ
ajst-21816	98	33	features	feature	NOUN
ajst-21816	98	34	of	of	ADP
ajst-21816	98	35	independent	independent	ADJ
ajst-21816	98	36	frames	frame	NOUN
ajst-21816	98	37	,	,	PUNCT
ajst-21816	98	38	and	and	CCONJ
ajst-21816	98	39	sp	sp	ADP
ajst-21816	98	40	aggregates	aggregate	NOUN
ajst-21816	98	41	these	these	DET
ajst-21816	98	42	feature	feature	NOUN
ajst-21816	98	43	information	information	NOUN
ajst-21816	98	44	into	into	ADP
ajst-21816	98	45	the	the	DET
ajst-21816	98	46	collectionlevel	collectionlevel	NOUN
ajst-21816	98	47	information	information	NOUN
ajst-21816	98	48	,	,	PUNCT
ajst-21816	98	49	the	the	DET
ajst-21816	98	50	formula	formula	NOUN
ajst-21816	98	51	is	be	AUX
ajst-21816	98	52	z	z	NOUN
ajst-21816	98	53	=	=	SYM
ajst-21816	98	54	s	s	X
ajst-21816	98	55	(	(	PUNCT
ajst-21816	98	56	v	v	NOUN
ajst-21816	98	57	)	)	PUNCT
ajst-21816	98	58	,	,	PUNCT
ajst-21816	98	59	where	where	SCONJ
ajst-21816	98	60	z	z	NOUN
ajst-21816	98	61	represents	represent	VERB
ajst-21816	98	62	the	the	DET
ajst-21816	98	63	collection	collection	NOUN
ajst-21816	98	64	-	-	PUNCT
ajst-21816	98	65	level	level	NOUN
ajst-21816	98	66	feature	feature	NOUN
ajst-21816	98	67	,	,	PUNCT
ajst-21816	98	68	v	v	NOUN
ajst-21816	98	69	represents	represent	VERB
ajst-21816	98	70	the	the	DET
ajst-21816	98	71	frame	frame	NOUN
ajst-21816	98	72	-	-	PUNCT
ajst-21816	98	73	level	level	NOUN
ajst-21816	98	74	feature	feature	NOUN
ajst-21816	98	75	,	,	PUNCT
ajst-21816	98	76	and	and	CCONJ
ajst-21816	98	77	the	the	DET
ajst-21816	98	78	formula	formula	NOUN
ajst-21816	98	79	is	be	AUX
ajst-21816	98	80	{	{	PUNCT
ajst-21816	98	81	|	|	ADV
ajst-21816	98	82	1	1	NUM
ajst-21816	98	83	,	,	PUNCT
ajst-21816	98	84	2	2	NUM
ajst-21816	98	85	,	,	PUNCT
ajst-21816	98	86	...	...	PUNCT
ajst-21816	98	87	,	,	PUNCT
ajst-21816	98	88	}	}	PUNCT
ajst-21816	98	89	jv	jv	NOUN
ajst-21816	98	90	v	v	X
ajst-21816	98	91	j	j	PROPN
ajst-21816	99	1	n	n	NOUN
ajst-21816	99	2			PROPN
ajst-21816	99	3	,	,	PUNCT
ajst-21816	99	4	n	n	PRON
ajst-21816	99	5	represents	represent	VERB
ajst-21816	99	6	the	the	DET
ajst-21816	99	7	number	number	NOUN
ajst-21816	99	8	of	of	ADP
ajst-21816	99	9	gait	gait	NOUN
ajst-21816	99	10	frames	frame	NOUN
ajst-21816	99	11	contained	contain	VERB
ajst-21816	99	12	in	in	ADP
ajst-21816	99	13	the	the	DET
ajst-21816	99	14	current	current	ADJ
ajst-21816	99	15	set	set	NOUN
ajst-21816	99	16	,	,	PUNCT
ajst-21816	99	17	and	and	CCONJ
ajst-21816	99	18	jv	jv	NOUN
ajst-21816	99	19	represents	represent	VERB
ajst-21816	99	20	the	the	DET
ajst-21816	99	21	feature	feature	NOUN
ajst-21816	99	22	map	map	NOUN
ajst-21816	99	23	of	of	ADP
ajst-21816	99	24	frame	frame	NOUN
ajst-21816	100	1	j.	j.	PROPN
ajst-21816	100	2	it	it	PRON
ajst-21816	100	3	should	should	AUX
ajst-21816	100	4	be	be	AUX
ajst-21816	100	5	noted	note	VERB
ajst-21816	100	6	that	that	SCONJ
ajst-21816	100	7	the	the	DET
ajst-21816	100	8	length	length	NOUN
ajst-21816	100	9	of	of	ADP
ajst-21816	100	10	a	a	DET
ajst-21816	100	11	video	video	NOUN
ajst-21816	100	12	is	be	AUX
ajst-21816	100	13	not	not	PART
ajst-21816	100	14	fixed	fix	VERB
ajst-21816	100	15	,	,	PUNCT
ajst-21816	100	16	so	so	CCONJ
ajst-21816	100	17	the	the	DET
ajst-21816	100	18	input	input	NOUN
ajst-21816	100	19	set	set	NOUN
ajst-21816	100	20	contains	contain	VERB
ajst-21816	100	21	no	no	DET
ajst-21816	100	22	fixed	fixed	ADJ
ajst-21816	100	23	number	number	NOUN
ajst-21816	100	24	of	of	ADP
ajst-21816	100	25	frames	frame	NOUN
ajst-21816	100	26	,	,	PUNCT
ajst-21816	100	27	which	which	PRON
ajst-21816	100	28	means	mean	VERB
ajst-21816	100	29	that	that	SCONJ
ajst-21816	100	30	the	the	DET
ajst-21816	100	31	function	function	NOUN
ajst-21816	100	32	s	s	VERB
ajst-21816	100	33	should	should	AUX
ajst-21816	100	34	be	be	AUX
ajst-21816	100	35	able	able	ADJ
ajst-21816	100	36	to	to	PART
ajst-21816	100	37	take	take	VERB
ajst-21816	100	38	any	any	DET
ajst-21816	100	39	set	set	NOUN
ajst-21816	100	40	of	of	ADP
ajst-21816	100	41	frames	frame	NOUN
ajst-21816	100	42	.	.	PUNCT
ajst-21816	101	1	under	under	ADP
ajst-21816	101	2	this	this	DET
ajst-21816	101	3	premise	premise	NOUN
ajst-21816	101	4	,	,	PUNCT
ajst-21816	101	5	two	two	NUM
ajst-21816	101	6	statistical	statistical	ADJ
ajst-21816	101	7	functions	function	NOUN
ajst-21816	101	8	,	,	PUNCT
ajst-21816	101	9	max	max	PROPN
ajst-21816	101	10	and	and	CCONJ
ajst-21816	101	11	median	median	ADJ
ajst-21816	101	12	,	,	PUNCT
ajst-21816	101	13	are	be	AUX
ajst-21816	101	14	considered	consider	VERB
ajst-21816	101	15	,	,	PUNCT
ajst-21816	101	16	and	and	CCONJ
ajst-21816	101	17	a	a	DET
ajst-21816	101	18	joint	joint	ADJ
ajst-21816	101	19	function	function	NOUN
ajst-21816	101	20	composed	compose	VERB
ajst-21816	101	21	of	of	ADP
ajst-21816	101	22	max	max	PROPN
ajst-21816	101	23	and	and	CCONJ
ajst-21816	101	24	median	median	NOUN
ajst-21816	101	25	is	be	AUX
ajst-21816	101	26	constructed	construct	VERB
ajst-21816	101	27	as	as	ADP
ajst-21816	101	28	an	an	DET
ajst-21816	101	29	instance	instance	NOUN
ajst-21816	101	30	of	of	ADP
ajst-21816	101	31	function	function	NOUN
ajst-21816	101	32	s	s	PROPN
ajst-21816	101	33	,	,	PUNCT
ajst-21816	101	34	as	as	SCONJ
ajst-21816	101	35	shown	show	VERB
ajst-21816	101	36	in	in	ADP
ajst-21816	101	37	equation	equation	NOUN
ajst-21816	101	38	(	(	PUNCT
ajst-21816	101	39	5	5	NUM
ajst-21816	101	40	)	)	PUNCT
ajst-21816	101	41	.	.	PUNCT
ajst-21816	102	1	in	in	ADP
ajst-21816	102	2	this	this	DET
ajst-21816	102	3	article	article	NOUN
ajst-21816	102	4	,	,	PUNCT
ajst-21816	102	5	we	we	PRON
ajst-21816	102	6	will	will	AUX
ajst-21816	102	7	compare	compare	VERB
ajst-21816	102	8	these	these	DET
ajst-21816	102	9	functions	function	NOUN
ajst-21816	102	10	in	in	ADP
ajst-21816	102	11	section	section	NOUN
ajst-21816	102	12	3	3	NUM
ajst-21816	102	13	to	to	PART
ajst-21816	102	14	get	get	VERB
ajst-21816	102	15	more	more	ADV
ajst-21816	102	16	useful	useful	ADJ
ajst-21816	102	17	examples	example	NOUN
ajst-21816	102	18	.	.	PUNCT
ajst-21816	103	1	1_1	1_1	NUM
ajst-21816	103	2	(	(	PUNCT
ajst-21816	103	3	(	(	PUNCT
ajst-21816	103	4	max	max	PROPN
ajst-21816	103	5	,	,	PUNCT
ajst-21816	103	6	)	)	PUNCT
ajst-21816	103	7	)	)	PUNCT
ajst-21816	103	8	unions	union	NOUN
ajst-21816	103	9	c	c	PROPN
ajst-21816	103	10	cat	cat	PROPN
ajst-21816	103	11	median	median	PROPN
ajst-21816	103	12	(	(	PUNCT
ajst-21816	103	13	5	5	NUM
ajst-21816	103	14	)	)	PUNCT
ajst-21816	103	15	where	where	SCONJ
ajst-21816	103	16	cat	cat	NOUN
ajst-21816	103	17	represents	represent	VERB
ajst-21816	103	18	the	the	DET
ajst-21816	103	19	connection	connection	NOUN
ajst-21816	103	20	of	of	ADP
ajst-21816	103	21	the	the	DET
ajst-21816	103	22	channel	channel	NOUN
ajst-21816	103	23	dimension	dimension	NOUN
ajst-21816	103	24	,	,	PUNCT
ajst-21816	103	25	max	max	PROPN
ajst-21816	103	26	statistical	statistical	ADJ
ajst-21816	103	27	function	function	NOUN
ajst-21816	103	28	and	and	CCONJ
ajst-21816	103	29	median	median	ADJ
ajst-21816	103	30	statistical	statistical	ADJ
ajst-21816	103	31	function	function	NOUN
ajst-21816	103	32	are	be	AUX
ajst-21816	103	33	applied	apply	VERB
ajst-21816	103	34	to	to	ADP
ajst-21816	103	35	the	the	DET
ajst-21816	103	36	set	set	NOUN
ajst-21816	103	37	dimension	dimension	NOUN
ajst-21816	103	38	,	,	PUNCT
ajst-21816	103	39	and	and	CCONJ
ajst-21816	103	40	the	the	DET
ajst-21816	103	41	1×1	1×1	ADJ
ajst-21816	103	42	convolution	convolution	NOUN
ajst-21816	103	43	layer	layer	NOUN
ajst-21816	103	44	represented	represent	VERB
ajst-21816	103	45	by	by	ADP
ajst-21816	103	46	1_1c	1_1c	PROPN
ajst-21816	103	47	can	can	AUX
ajst-21816	103	48	learn	learn	VERB
ajst-21816	103	49	appropriate	appropriate	ADJ
ajst-21816	103	50	weights	weight	NOUN
ajst-21816	103	51	to	to	PART
ajst-21816	103	52	fuse	fuse	VERB
ajst-21816	103	53	the	the	DET
ajst-21816	103	54	information	information	NOUN
ajst-21816	103	55	extracted	extract	VERB
ajst-21816	103	56	by	by	ADP
ajst-21816	103	57	the	the	DET
ajst-21816	103	58	two	two	NUM
ajst-21816	103	59	statistical	statistical	ADJ
ajst-21816	103	60	functions	function	NOUN
ajst-21816	103	61	.	.	PUNCT
ajst-21816	104	1	2.3	2.3	NUM
ajst-21816	104	2	.	.	PUNCT
ajst-21816	105	1	local	local	ADJ
ajst-21816	105	2	feature	feature	NOUN
ajst-21816	105	3	extractor	extractor	NOUN
ajst-21816	105	4	(	(	PUNCT
ajst-21816	105	5	lfe	lfe	PROPN
ajst-21816	105	6	)	)	PUNCT
ajst-21816	105	7	the	the	DET
ajst-21816	105	8	frame	frame	NOUN
ajst-21816	105	9	diagram	diagram	NOUN
ajst-21816	105	10	of	of	ADP
ajst-21816	105	11	the	the	DET
ajst-21816	105	12	local	local	ADJ
ajst-21816	105	13	feature	feature	NOUN
ajst-21816	105	14	extractor	extractor	NOUN
ajst-21816	105	15	is	be	AUX
ajst-21816	105	16	shown	show	VERB
ajst-21816	105	17	in	in	ADP
ajst-21816	105	18	figure	figure	NOUN
ajst-21816	105	19	5	5	NUM
ajst-21816	105	20	.	.	PUNCT
ajst-21816	106	1	the	the	DET
ajst-21816	106	2	input	input	NOUN
ajst-21816	106	3	is	be	AUX
ajst-21816	106	4	a	a	DET
ajst-21816	106	5	set	set	NOUN
ajst-21816	106	6	of	of	ADP
ajst-21816	106	7	disordered	disordered	ADJ
ajst-21816	106	8	gait	gait	ADJ
ajst-21816	106	9	profiles	profile	NOUN
ajst-21816	106	10	.	.	PUNCT
ajst-21816	107	1	lfe	lfe	PROPN
ajst-21816	107	2	extracts	extract	VERB
ajst-21816	107	3	the	the	DET
ajst-21816	107	4	gait	gait	ADJ
ajst-21816	107	5	features	feature	NOUN
ajst-21816	107	6	of	of	ADP
ajst-21816	107	7	individual	individual	ADJ
ajst-21816	107	8	frames	frame	NOUN
ajst-21816	107	9	from	from	ADP
ajst-21816	107	10	each	each	DET
ajst-21816	107	11	profile	profile	NOUN
ajst-21816	107	12	,	,	PUNCT
ajst-21816	107	13	and	and	CCONJ
ajst-21816	107	14	then	then	ADV
ajst-21816	107	15	aggregates	aggregate	VERB
ajst-21816	107	16	the	the	DET
ajst-21816	107	17	frame	frame	NOUN
ajst-21816	107	18	level	level	NOUN
ajst-21816	107	19	features	feature	VERB
ajst-21816	107	20	into	into	ADP
ajst-21816	107	21	set	set	ADJ
ajst-21816	107	22	level	level	NOUN
ajst-21816	107	23	features	feature	NOUN
ajst-21816	107	24	through	through	ADP
ajst-21816	107	25	sp	sp	ADP
ajst-21816	107	26	operation	operation	NOUN
ajst-21816	107	27	.	.	PUNCT
ajst-21816	108	1	figure	figure	NOUN
ajst-21816	108	2	5	5	NUM
ajst-21816	108	3	.	.	PUNCT
ajst-21816	109	1	lfe	lfe	PROPN
ajst-21816	109	2	structure	structure	PROPN
ajst-21816	109	3	diagram	diagram	PROPN
ajst-21816	109	4	290	290	NUM
ajst-21816	109	5	the	the	DET
ajst-21816	109	6	focal	focal	ADJ
ajst-21816	109	7	convolution	convolution	NOUN
ajst-21816	109	8	layer	layer	NOUN
ajst-21816	109	9	(	(	PUNCT
ajst-21816	109	10	fconv	fconv	INTJ
ajst-21816	109	11	)	)	PUNCT
ajst-21816	109	12	in	in	ADP
ajst-21816	109	13	the	the	DET
ajst-21816	109	14	gaitpart	gaitpart	NOUN
ajst-21816	109	15	model	model	NOUN
ajst-21816	109	16	divides	divide	VERB
ajst-21816	109	17	the	the	DET
ajst-21816	109	18	feature	feature	NOUN
ajst-21816	109	19	map	map	NOUN
ajst-21816	109	20	horizontally	horizontally	ADV
ajst-21816	109	21	,	,	PUNCT
ajst-21816	109	22	as	as	SCONJ
ajst-21816	109	23	shown	show	VERB
ajst-21816	109	24	in	in	ADP
ajst-21816	109	25	figure	figure	NOUN
ajst-21816	109	26	6(a	6(a	NUM
ajst-21816	109	27	)	)	PUNCT
ajst-21816	109	28	.	.	PUNCT
ajst-21816	110	1	inspired	inspire	VERB
ajst-21816	110	2	by	by	ADP
ajst-21816	110	3	his	his	PRON
ajst-21816	110	4	ideas	idea	NOUN
ajst-21816	110	5	,	,	PUNCT
ajst-21816	110	6	this	this	DET
ajst-21816	110	7	paper	paper	NOUN
ajst-21816	110	8	proposes	propose	VERB
ajst-21816	110	9	sconv	sconv	NOUN
ajst-21816	110	10	,	,	PUNCT
ajst-21816	110	11	which	which	PRON
ajst-21816	110	12	splits	split	VERB
ajst-21816	110	13	the	the	DET
ajst-21816	110	14	input	input	NOUN
ajst-21816	110	15	gait	gait	NOUN
ajst-21816	110	16	feature	feature	NOUN
ajst-21816	110	17	map	map	NOUN
ajst-21816	110	18	horizontally	horizontally	ADV
ajst-21816	110	19	in	in	ADP
ajst-21816	110	20	a	a	DET
ajst-21816	110	21	specific	specific	ADJ
ajst-21816	110	22	proportion	proportion	NOUN
ajst-21816	110	23	,	,	PUNCT
ajst-21816	110	24	as	as	SCONJ
ajst-21816	110	25	shown	show	VERB
ajst-21816	110	26	in	in	ADP
ajst-21816	110	27	figure	figure	NOUN
ajst-21816	110	28	6	6	NUM
ajst-21816	110	29	(	(	PUNCT
ajst-21816	110	30	b	b	NOUN
ajst-21816	110	31	)	)	PUNCT
ajst-21816	110	32	.	.	PUNCT
ajst-21816	111	1	(	(	PUNCT
ajst-21816	111	2	a	a	X
ajst-21816	111	3	)	)	PUNCT
ajst-21816	111	4	average	average	ADJ
ajst-21816	111	5	level	level	NOUN
ajst-21816	111	6	segmentation	segmentation	NOUN
ajst-21816	111	7	(	(	PUNCT
ajst-21816	111	8	b	b	NOUN
ajst-21816	111	9	)	)	PUNCT
ajst-21816	111	10	specific	specific	ADJ
ajst-21816	111	11	proportional	proportional	ADJ
ajst-21816	111	12	segmentation	segmentation	NOUN
ajst-21816	111	13	figure	figure	NOUN
ajst-21816	111	14	6	6	NUM
ajst-21816	111	15	.	.	PUNCT
ajst-21816	111	16	gait	gait	PROPN
ajst-21816	111	17	silhouette	silhouette	NOUN
ajst-21816	111	18	segmentation	segmentation	NOUN
ajst-21816	111	19	method	method	NOUN
ajst-21816	111	20	when	when	SCONJ
ajst-21816	111	21	people	people	NOUN
ajst-21816	111	22	walk	walk	VERB
ajst-21816	111	23	,	,	PUNCT
ajst-21816	111	24	the	the	DET
ajst-21816	111	25	movements	movement	NOUN
ajst-21816	111	26	of	of	ADP
ajst-21816	111	27	different	different	ADJ
ajst-21816	111	28	parts	part	NOUN
ajst-21816	111	29	of	of	ADP
ajst-21816	111	30	the	the	DET
ajst-21816	111	31	body	body	NOUN
ajst-21816	111	32	show	show	VERB
ajst-21816	111	33	obvious	obvious	ADJ
ajst-21816	111	34	differences	difference	NOUN
ajst-21816	111	35	.	.	PUNCT
ajst-21816	112	1	for	for	ADP
ajst-21816	112	2	example	example	NOUN
ajst-21816	112	3	,	,	PUNCT
ajst-21816	112	4	the	the	DET
ajst-21816	112	5	movement	movement	NOUN
ajst-21816	112	6	of	of	ADP
ajst-21816	112	7	the	the	DET
ajst-21816	112	8	head	head	NOUN
ajst-21816	112	9	and	and	CCONJ
ajst-21816	112	10	the	the	DET
ajst-21816	112	11	center	center	NOUN
ajst-21816	112	12	of	of	ADP
ajst-21816	112	13	gravity	gravity	NOUN
ajst-21816	112	14	of	of	ADP
ajst-21816	112	15	the	the	DET
ajst-21816	112	16	human	human	ADJ
ajst-21816	112	17	body	body	NOUN
ajst-21816	112	18	is	be	AUX
ajst-21816	112	19	relatively	relatively	ADV
ajst-21816	112	20	small	small	ADJ
ajst-21816	112	21	,	,	PUNCT
ajst-21816	112	22	while	while	SCONJ
ajst-21816	112	23	the	the	DET
ajst-21816	112	24	arms	arm	NOUN
ajst-21816	112	25	and	and	CCONJ
ajst-21816	112	26	legs	leg	NOUN
ajst-21816	112	27	have	have	VERB
ajst-21816	112	28	a	a	DET
ajst-21816	112	29	large	large	ADJ
ajst-21816	112	30	swing	swing	NOUN
ajst-21816	112	31	amplitude	amplitude	NOUN
ajst-21816	112	32	,	,	PUNCT
ajst-21816	112	33	showing	show	VERB
ajst-21816	112	34	a	a	DET
ajst-21816	112	35	periodic	periodic	ADJ
ajst-21816	112	36	pendulum	pendulum	NOUN
ajst-21816	112	37	movement	movement	NOUN
ajst-21816	112	38	.	.	PUNCT
ajst-21816	113	1	this	this	DET
ajst-21816	113	2	difference	difference	NOUN
ajst-21816	113	3	is	be	AUX
ajst-21816	113	4	of	of	ADP
ajst-21816	113	5	great	great	ADJ
ajst-21816	113	6	significance	significance	NOUN
ajst-21816	113	7	in	in	ADP
ajst-21816	113	8	gait	gait	PROPN
ajst-21816	113	9	recognition	recognition	PROPN
ajst-21816	113	10	.	.	PUNCT
ajst-21816	114	1	through	through	ADP
ajst-21816	114	2	detailed	detailed	ADJ
ajst-21816	114	3	analysis	analysis	NOUN
ajst-21816	114	4	and	and	CCONJ
ajst-21816	114	5	modeling	modeling	NOUN
ajst-21816	114	6	of	of	ADP
ajst-21816	114	7	the	the	DET
ajst-21816	114	8	movement	movement	NOUN
ajst-21816	114	9	characteristics	characteristic	NOUN
ajst-21816	114	10	of	of	ADP
ajst-21816	114	11	these	these	DET
ajst-21816	114	12	different	different	ADJ
ajst-21816	114	13	parts	part	NOUN
ajst-21816	114	14	,	,	PUNCT
ajst-21816	114	15	individuals	individual	NOUN
ajst-21816	114	16	'	'	PART
ajst-21816	114	17	gait	gait	PROPN
ajst-21816	114	18	patterns	pattern	NOUN
ajst-21816	114	19	can	can	AUX
ajst-21816	114	20	be	be	AUX
ajst-21816	114	21	identified	identify	VERB
ajst-21816	114	22	more	more	ADV
ajst-21816	114	23	accurately	accurately	ADV
ajst-21816	114	24	.	.	PUNCT
ajst-21816	115	1	among	among	ADP
ajst-21816	115	2	them	they	PRON
ajst-21816	115	3	,	,	PUNCT
ajst-21816	115	4	the	the	DET
ajst-21816	115	5	movement	movement	NOUN
ajst-21816	115	6	pattern	pattern	NOUN
ajst-21816	115	7	of	of	ADP
ajst-21816	115	8	the	the	DET
ajst-21816	115	9	head	head	NOUN
ajst-21816	115	10	and	and	CCONJ
ajst-21816	115	11	the	the	DET
ajst-21816	115	12	center	center	NOUN
ajst-21816	115	13	of	of	ADP
ajst-21816	115	14	gravity	gravity	NOUN
ajst-21816	115	15	of	of	ADP
ajst-21816	115	16	the	the	DET
ajst-21816	115	17	human	human	ADJ
ajst-21816	115	18	body	body	NOUN
ajst-21816	115	19	may	may	AUX
ajst-21816	115	20	be	be	AUX
ajst-21816	115	21	more	more	ADV
ajst-21816	115	22	suitable	suitable	ADJ
ajst-21816	115	23	for	for	ADP
ajst-21816	115	24	recognizing	recognize	VERB
ajst-21816	115	25	specific	specific	ADJ
ajst-21816	115	26	identity	identity	NOUN
ajst-21816	115	27	information	information	NOUN
ajst-21816	115	28	or	or	CCONJ
ajst-21816	115	29	emotional	emotional	ADJ
ajst-21816	115	30	states	state	NOUN
ajst-21816	115	31	,	,	PUNCT
ajst-21816	115	32	while	while	SCONJ
ajst-21816	115	33	the	the	DET
ajst-21816	115	34	swing	swing	NOUN
ajst-21816	115	35	of	of	ADP
ajst-21816	115	36	the	the	DET
ajst-21816	115	37	arms	arm	NOUN
ajst-21816	115	38	and	and	CCONJ
ajst-21816	115	39	legs	leg	NOUN
ajst-21816	115	40	is	be	AUX
ajst-21816	115	41	more	more	ADV
ajst-21816	115	42	conducive	conducive	ADJ
ajst-21816	115	43	to	to	ADP
ajst-21816	115	44	judging	judge	VERB
ajst-21816	115	45	the	the	DET
ajst-21816	115	46	stability	stability	NOUN
ajst-21816	115	47	and	and	CCONJ
ajst-21816	115	48	speed	speed	NOUN
ajst-21816	115	49	of	of	ADP
ajst-21816	115	50	gait	gait	ADJ
ajst-21816	115	51	parameters	parameter	NOUN
ajst-21816	115	52	.	.	PUNCT
ajst-21816	116	1	therefore	therefore	ADV
ajst-21816	116	2	,	,	PUNCT
ajst-21816	116	3	the	the	DET
ajst-21816	116	4	in	in	ADP
ajst-21816	116	5	-	-	PUNCT
ajst-21816	116	6	depth	depth	NOUN
ajst-21816	116	7	study	study	NOUN
ajst-21816	116	8	of	of	ADP
ajst-21816	116	9	the	the	DET
ajst-21816	116	10	motion	motion	NOUN
ajst-21816	116	11	characteristics	characteristic	NOUN
ajst-21816	116	12	of	of	ADP
ajst-21816	116	13	various	various	ADJ
ajst-21816	116	14	parts	part	NOUN
ajst-21816	116	15	of	of	ADP
ajst-21816	116	16	the	the	DET
ajst-21816	116	17	human	human	ADJ
ajst-21816	116	18	body	body	NOUN
ajst-21816	116	19	and	and	CCONJ
ajst-21816	116	20	the	the	DET
ajst-21816	116	21	correlation	correlation	NOUN
ajst-21816	116	22	between	between	ADP
ajst-21816	116	23	these	these	DET
ajst-21816	116	24	characteristics	characteristic	NOUN
ajst-21816	116	25	and	and	CCONJ
ajst-21816	116	26	gait	gait	PROPN
ajst-21816	116	27	recognition	recognition	NOUN
ajst-21816	116	28	will	will	AUX
ajst-21816	116	29	help	help	VERB
ajst-21816	116	30	to	to	PART
ajst-21816	116	31	improve	improve	VERB
ajst-21816	116	32	the	the	DET
ajst-21816	116	33	accuracy	accuracy	NOUN
ajst-21816	116	34	and	and	CCONJ
ajst-21816	116	35	applicability	applicability	NOUN
ajst-21816	116	36	of	of	ADP
ajst-21816	116	37	gait	gait	ADJ
ajst-21816	116	38	recognition	recognition	NOUN
ajst-21816	116	39	technology	technology	NOUN
ajst-21816	116	40	.	.	PUNCT
ajst-21816	117	1	in	in	ADP
ajst-21816	117	2	view	view	NOUN
ajst-21816	117	3	of	of	ADP
ajst-21816	117	4	the	the	DET
ajst-21816	117	5	above	above	ADJ
ajst-21816	117	6	,	,	PUNCT
ajst-21816	117	7	this	this	DET
ajst-21816	117	8	paper	paper	NOUN
ajst-21816	117	9	divides	divide	VERB
ajst-21816	117	10	the	the	DET
ajst-21816	117	11	gait	gait	ADJ
ajst-21816	117	12	feature	feature	NOUN
ajst-21816	117	13	map	map	NOUN
ajst-21816	117	14	horizontally	horizontally	ADV
ajst-21816	117	15	into	into	ADP
ajst-21816	117	16	four	four	NUM
ajst-21816	117	17	parts	part	NOUN
ajst-21816	117	18	:	:	PUNCT
ajst-21816	117	19	head	head	NOUN
ajst-21816	117	20	,	,	PUNCT
ajst-21816	117	21	trunk	trunk	NOUN
ajst-21816	117	22	,	,	PUNCT
ajst-21816	117	23	center	center	NOUN
ajst-21816	117	24	of	of	ADP
ajst-21816	117	25	gravity	gravity	NOUN
ajst-21816	117	26	and	and	CCONJ
ajst-21816	117	27	lower	low	ADJ
ajst-21816	117	28	limb	limb	NOUN
ajst-21816	117	29	,	,	PUNCT
ajst-21816	117	30	and	and	CCONJ
ajst-21816	117	31	sets	set	VERB
ajst-21816	117	32	the	the	DET
ajst-21816	117	33	ratio	ratio	NOUN
ajst-21816	117	34	of	of	ADP
ajst-21816	117	35	these	these	DET
ajst-21816	117	36	four	four	NUM
ajst-21816	117	37	parts	part	NOUN
ajst-21816	117	38	as	as	ADP
ajst-21816	117	39	1:2:1:2	1:2:1:2	NUM
ajst-21816	117	40	.	.	PUNCT
ajst-21816	118	1	the	the	DET
ajst-21816	118	2	lfe	lfe	PROPN
ajst-21816	118	3	contains	contain	VERB
ajst-21816	118	4	three	three	NUM
ajst-21816	118	5	sconvs	sconvs	NOUN
ajst-21816	118	6	,	,	PUNCT
ajst-21816	118	7	each	each	PRON
ajst-21816	118	8	with	with	ADP
ajst-21816	118	9	the	the	DET
ajst-21816	118	10	same	same	ADJ
ajst-21816	118	11	structure	structure	NOUN
ajst-21816	118	12	,	,	PUNCT
ajst-21816	118	13	as	as	SCONJ
ajst-21816	118	14	shown	show	VERB
ajst-21816	118	15	in	in	ADP
ajst-21816	118	16	figure	figure	NOUN
ajst-21816	118	17	7	7	NUM
ajst-21816	118	18	.	.	PUNCT
ajst-21816	119	1	the	the	DET
ajst-21816	119	2	input	input	NOUN
ajst-21816	119	3	gait	gait	PROPN
ajst-21816	119	4	contour	contour	NOUN
ajst-21816	119	5	is	be	AUX
ajst-21816	119	6	divided	divide	VERB
ajst-21816	119	7	into	into	ADP
ajst-21816	119	8	four	four	NUM
ajst-21816	119	9	parts	part	NOUN
ajst-21816	119	10	in	in	ADP
ajst-21816	119	11	the	the	DET
ajst-21816	119	12	height	height	NOUN
ajst-21816	119	13	dimension	dimension	NOUN
ajst-21816	119	14	,	,	PUNCT
ajst-21816	119	15	and	and	CCONJ
ajst-21816	119	16	the	the	DET
ajst-21816	119	17	four	four	NUM
ajst-21816	119	18	parts	part	NOUN
ajst-21816	119	19	are	be	AUX
ajst-21816	119	20	convolution	convolution	NOUN
ajst-21816	119	21	operation	operation	NOUN
ajst-21816	119	22	respectively	respectively	ADV
ajst-21816	119	23	,	,	PUNCT
ajst-21816	119	24	and	and	CCONJ
ajst-21816	119	25	then	then	ADV
ajst-21816	119	26	the	the	DET
ajst-21816	119	27	height	height	NOUN
ajst-21816	119	28	dimension	dimension	NOUN
ajst-21816	119	29	is	be	AUX
ajst-21816	119	30	spliced	splice	VERB
ajst-21816	119	31	.	.	PUNCT
ajst-21816	120	1	suppose	suppose	VERB
ajst-21816	120	2	(	(	PUNCT
ajst-21816	120	3	)	)	PUNCT
ajst-21816	120	4	{	{	PUNCT
ajst-21816	120	5	(	(	PUNCT
ajst-21816	120	6	)	)	PUNCT
ajst-21816	120	7	|	|	ADV
ajst-21816	120	8	1	1	NUM
ajst-21816	120	9	,	,	PUNCT
ajst-21816	120	10	2,3	2,3	NUM
ajst-21816	120	11	,	,	PUNCT
ajst-21816	120	12	4}i	4}i	NUM
ajst-21816	120	13	sc	sc	PROPN
ajst-21816	120	14	scf	scf	PROPN
ajst-21816	120	15	input	input	NOUN
ajst-21816	120	16	f	f	PROPN
ajst-21816	120	17	input	input	NOUN
ajst-21816	120	18	i	i	NOUN
ajst-21816	120	19			ADV
ajst-21816	120	20	,	,	PUNCT
ajst-21816	120	21	sconv	sconv	NOUN
ajst-21816	120	22	can	can	AUX
ajst-21816	120	23	be	be	AUX
ajst-21816	120	24	expressed	express	VERB
ajst-21816	120	25	as	as	ADP
ajst-21816	120	26	formula	formula	NOUN
ajst-21816	120	27	(	(	PUNCT
ajst-21816	120	28	3	3	NUM
ajst-21816	120	29	-	-	SYM
ajst-21816	120	30	6	6	NUM
ajst-21816	120	31	)	)	PUNCT
ajst-21816	120	32	.	.	PUNCT
ajst-21816	121	1	1	1	NUM
ajst-21816	121	2	2	2	NUM
ajst-21816	121	3	3	3	NUM
ajst-21816	121	4	4	4	NUM
ajst-21816	121	5	3	3	NUM
ajst-21816	121	6	_	_	NOUN
ajst-21816	121	7	3	3	NUM
ajst-21816	121	8	(	(	PUNCT
ajst-21816	121	9	(	(	PUNCT
ajst-21816	121	10	)	)	PUNCT
ajst-21816	121	11	)	)	PUNCT
ajst-21816	121	12	3	3	NUM
ajst-21816	121	13	_	_	SYM
ajst-21816	121	14	3	3	NUM
ajst-21816	121	15	(	(	PUNCT
ajst-21816	121	16	(	(	PUNCT
ajst-21816	121	17	)	)	PUNCT
ajst-21816	121	18	)	)	PUNCT
ajst-21816	121	19	(	(	PUNCT
ajst-21816	121	20	)	)	PUNCT
ajst-21816	121	21	3	3	NUM
ajst-21816	121	22	_	_	SYM
ajst-21816	121	23	3	3	NUM
ajst-21816	121	24	(	(	PUNCT
ajst-21816	121	25	(	(	PUNCT
ajst-21816	121	26	)	)	PUNCT
ajst-21816	121	27	)	)	PUNCT
ajst-21816	121	28	3	3	NUM
ajst-21816	121	29	_	_	SYM
ajst-21816	121	30	3	3	NUM
ajst-21816	121	31	(	(	PUNCT
ajst-21816	121	32	(	(	PUNCT
ajst-21816	121	33	)	)	PUNCT
ajst-21816	121	34	)	)	PUNCT
ajst-21816	121	35	sc	sc	PROPN
ajst-21816	121	36	sc	sc	PROPN
ajst-21816	121	37	sc	sc	PROPN
ajst-21816	121	38	sc	sc	PROPN
ajst-21816	121	39	sc	sc	PROPN
ajst-21816	121	40	c	c	PROPN
ajst-21816	121	41	f	f	PROPN
ajst-21816	121	42	input	input	NOUN
ajst-21816	121	43	c	c	PROPN
ajst-21816	121	44	f	f	PROPN
ajst-21816	121	45	input	input	NOUN
ajst-21816	121	46	f	f	PROPN
ajst-21816	121	47	output	output	NOUN
ajst-21816	121	48	cat	cat	PROPN
ajst-21816	121	49	c	c	PROPN
ajst-21816	121	50	f	f	PROPN
ajst-21816	121	51	input	input	NOUN
ajst-21816	121	52	c	c	PROPN
ajst-21816	121	53	f	f	PROPN
ajst-21816	121	54	input	input	PROPN
ajst-21816	121	55			NOUN
ajst-21816	121	56			PROPN
ajst-21816	122	1			PROPN
ajst-21816	122	2			PROPN
ajst-21816	123	1			PROPN
ajst-21816	123	2			NOUN
ajst-21816	123	3			PROPN
ajst-21816	123	4			PROPN
ajst-21816	123	5			PROPN
ajst-21816	123	6			NOUN
ajst-21816	124	1			PROPN
ajst-21816	124	2			ADJ
ajst-21816	124	3			NOUN
ajst-21816	124	4	(	(	PUNCT
ajst-21816	124	5	6	6	NUM
ajst-21816	124	6	)	)	PUNCT
ajst-21816	124	7	where	where	SCONJ
ajst-21816	124	8	3_3c	3_3c	PROPN
ajst-21816	124	9	represents	represent	VERB
ajst-21816	124	10	a	a	DET
ajst-21816	124	11	3×3	3×3	NUM
ajst-21816	124	12	convolution	convolution	NOUN
ajst-21816	124	13	and	and	CCONJ
ajst-21816	124	14	cat	cat	NOUN
ajst-21816	124	15	represents	represent	VERB
ajst-21816	124	16	a	a	DET
ajst-21816	124	17	concatenation	concatenation	NOUN
ajst-21816	124	18	of	of	ADP
ajst-21816	124	19	height	height	NOUN
ajst-21816	124	20	dimensions	dimension	NOUN
ajst-21816	124	21	.	.	PUNCT
ajst-21816	125	1	figure	figure	VERB
ajst-21816	125	2	7	7	NUM
ajst-21816	125	3	.	.	PUNCT
ajst-21816	125	4	sconv	sconv	NOUN
ajst-21816	125	5	structure	structure	NOUN
ajst-21816	125	6	diagram	diagram	NOUN
ajst-21816	125	7	291	291	NUM
ajst-21816	125	8	given	give	VERB
ajst-21816	125	9	a	a	DET
ajst-21816	125	10	dataset	dataset	NOUN
ajst-21816	125	11	containing	contain	VERB
ajst-21816	125	12	n	n	PRON
ajst-21816	125	13	individuals	individual	NOUN
ajst-21816	125	14	'	'	PART
ajst-21816	125	15	gait	gait	PROPN
ajst-21816	125	16	profiles	profile	NOUN
ajst-21816	125	17	,	,	PUNCT
ajst-21816	125	18	all	all	DET
ajst-21816	125	19	profiles	profile	NOUN
ajst-21816	125	20	in	in	ADP
ajst-21816	125	21	one	one	NUM
ajst-21816	125	22	or	or	CCONJ
ajst-21816	125	23	more	more	ADJ
ajst-21816	125	24	sequences	sequence	NOUN
ajst-21816	125	25	for	for	ADP
ajst-21816	125	26	each	each	DET
ajst-21816	125	27	individual	individual	NOUN
ajst-21816	125	28	can	can	AUX
ajst-21816	125	29	be	be	AUX
ajst-21816	125	30	viewed	view	VERB
ajst-21816	125	31	as	as	ADP
ajst-21816	125	32	a	a	DET
ajst-21816	125	33	set	set	NOUN
ajst-21816	125	34	of	of	ADP
ajst-21816	125	35	n	n	DET
ajst-21816	125	36	profiles	profile	NOUN
ajst-21816	125	37	,	,	PUNCT
ajst-21816	125	38	expressed	express	VERB
ajst-21816	125	39	as	as	ADP
ajst-21816	125	40	x	x	X
ajst-21816	125	41	{	{	PUNCT
ajst-21816	125	42	|	|	ADV
ajst-21816	125	43	1	1	NUM
ajst-21816	125	44	,	,	PUNCT
ajst-21816	125	45	2	2	NUM
ajst-21816	125	46	,	,	PUNCT
ajst-21816	125	47	,	,	PUNCT
ajst-21816	125	48	}	}	PUNCT
ajst-21816	125	49	j	j	PROPN
ajst-21816	126	1	i	i	PRON
ajst-21816	126	2	ix	ix	PROPN
ajst-21816	126	3	j	j	PROPN
ajst-21816	126	4	n	n	PROPN
ajst-21816	127	1			PROPN
ajst-21816	127	2			NOUN
ajst-21816	127	3	,	,	PUNCT
ajst-21816	127	4	where	where	SCONJ
ajst-21816	127	5	1,2	1,2	NUM
ajst-21816	127	6	,	,	PUNCT
ajst-21816	127	7	,	,	PUNCT
ajst-21816	127	8	i	i	PRON
ajst-21816	127	9	n	n	VERB
ajst-21816	127	10			NOUN
ajst-21816	127	11	.	.	PUNCT
ajst-21816	128	1	with	with	ADP
ajst-21816	128	2	x	x	PRON
ajst-21816	128	3	as	as	ADP
ajst-21816	128	4	the	the	DET
ajst-21816	128	5	input	input	NOUN
ajst-21816	128	6	,	,	PUNCT
ajst-21816	128	7	the	the	DET
ajst-21816	128	8	characteristics	characteristic	NOUN
ajst-21816	128	9	of	of	ADP
ajst-21816	128	10	the	the	DET
ajst-21816	128	11	lfe	lfe	PROPN
ajst-21816	128	12	output	output	NOUN
ajst-21816	128	13	are	be	AUX
ajst-21816	128	14	expressed	express	VERB
ajst-21816	128	15	by	by	ADP
ajst-21816	128	16	(	(	PUNCT
ajst-21816	128	17	)	)	PUNCT
ajst-21816	128	18	lf	lf	ADP
ajst-21816	128	19	out	out	ADP
ajst-21816	128	20	,	,	PUNCT
ajst-21816	128	21	which	which	PRON
ajst-21816	128	22	can	can	AUX
ajst-21816	128	23	be	be	AUX
ajst-21816	128	24	expressed	express	VERB
ajst-21816	128	25	as	as	ADP
ajst-21816	128	26	formula	formula	NOUN
ajst-21816	128	27	(	(	PUNCT
ajst-21816	128	28	3	3	NUM
ajst-21816	128	29	-	-	SYM
ajst-21816	128	30	7	7	NUM
ajst-21816	128	31	)	)	PUNCT
ajst-21816	128	32	.	.	PUNCT
ajst-21816	129	1	1	1	NUM
ajst-21816	129	2	2	2	NUM
ajst-21816	129	3	(	(	PUNCT
ajst-21816	129	4	)	)	PUNCT
ajst-21816	129	5	(	(	PUNCT
ajst-21816	129	6	(	(	PUNCT
ajst-21816	129	7	)	)	PUNCT
ajst-21816	129	8	,	,	PUNCT
ajst-21816	129	9	(	(	PUNCT
ajst-21816	129	10	)	)	PUNCT
ajst-21816	129	11	,	,	PUNCT
ajst-21816	129	12	,	,	PUNCT
ajst-21816	129	13	(	(	PUNCT
ajst-21816	129	14	)	)	PUNCT
ajst-21816	129	15	)	)	PUNCT
ajst-21816	129	16	n	n	CCONJ
ajst-21816	129	17	l	l	NOUN
ajst-21816	129	18	s	s	VERB
ajst-21816	129	19	i	i	PRON
ajst-21816	129	20	s	s	VERB
ajst-21816	130	1	i	i	PRON
ajst-21816	130	2	s	s	X
ajst-21816	130	3	if	if	SCONJ
ajst-21816	130	4	out	out	ADP
ajst-21816	131	1	s	s	VERB
ajst-21816	131	2	f	f	NOUN
ajst-21816	131	3	x	x	SYM
ajst-21816	131	4	f	f	PROPN
ajst-21816	131	5	x	x	X
ajst-21816	131	6	f	f	X
ajst-21816	131	7	x	x	PROPN
ajst-21816	131	8			NOUN
ajst-21816	131	9	(	(	PUNCT
ajst-21816	131	10	7	7	NUM
ajst-21816	131	11	)	)	PUNCT
ajst-21816	131	12	where	where	SCONJ
ajst-21816	131	13	sf	sf	PROPN
ajst-21816	131	14	is	be	AUX
ajst-21816	131	15	a	a	DET
ajst-21816	131	16	convolutional	convolutional	ADJ
ajst-21816	131	17	network	network	NOUN
ajst-21816	131	18	consisting	consist	VERB
ajst-21816	131	19	of	of	ADP
ajst-21816	131	20	three	three	NUM
ajst-21816	131	21	sconvs	sconvs	NOUN
ajst-21816	131	22	and	and	CCONJ
ajst-21816	131	23	two	two	NUM
ajst-21816	131	24	pooling	pool	VERB
ajst-21816	131	25	operations	operation	NOUN
ajst-21816	131	26	designed	design	VERB
ajst-21816	131	27	to	to	PART
ajst-21816	131	28	extract	extract	VERB
ajst-21816	131	29	frame	frame	NOUN
ajst-21816	131	30	-	-	PUNCT
ajst-21816	131	31	level	level	NOUN
ajst-21816	131	32	features	feature	NOUN
ajst-21816	131	33	from	from	ADP
ajst-21816	131	34	each	each	DET
ajst-21816	131	35	gait	gait	ADJ
ajst-21816	131	36	profile	profile	NOUN
ajst-21816	131	37	.	.	PUNCT
ajst-21816	132	1	s	s	PART
ajst-21816	132	2	is	be	AUX
ajst-21816	132	3	used	use	VERB
ajst-21816	132	4	to	to	PART
ajst-21816	132	5	map	map	VERB
ajst-21816	132	6	a	a	DET
ajst-21816	132	7	set	set	NOUN
ajst-21816	132	8	of	of	ADP
ajst-21816	132	9	frame	frame	NOUN
ajst-21816	132	10	-	-	PUNCT
ajst-21816	132	11	level	level	NOUN
ajst-21816	132	12	features	feature	NOUN
ajst-21816	132	13	to	to	ADP
ajst-21816	132	14	a	a	DET
ajst-21816	132	15	set	set	VERB
ajst-21816	132	16	level	level	NOUN
ajst-21816	132	17	feature	feature	NOUN
ajst-21816	132	18	based	base	VERB
ajst-21816	132	19	on	on	ADP
ajst-21816	132	20	set	set	NOUN
ajst-21816	132	21	pooling	pooling	NOUN
ajst-21816	132	22	(	(	PUNCT
ajst-21816	132	23	sp	sp	NOUN
ajst-21816	132	24	)	)	PUNCT
ajst-21816	132	25	.	.	PUNCT
ajst-21816	133	1	2.4	2.4	NUM
ajst-21816	133	2	.	.	PUNCT
ajst-21816	134	1	joint	joint	ADJ
ajst-21816	134	2	loss	loss	NOUN
ajst-21816	134	3	function	function	NOUN
ajst-21816	134	4	(	(	PUNCT
ajst-21816	134	5	ul	ul	INTJ
ajst-21816	134	6	)	)	PUNCT
ajst-21816	134	7	the	the	DET
ajst-21816	134	8	joint	joint	ADJ
ajst-21816	134	9	loss	loss	NOUN
ajst-21816	134	10	function	function	NOUN
ajst-21816	134	11	used	use	VERB
ajst-21816	134	12	in	in	ADP
ajst-21816	134	13	this	this	DET
ajst-21816	134	14	paper	paper	NOUN
ajst-21816	134	15	is	be	AUX
ajst-21816	134	16	derived	derive	VERB
ajst-21816	134	17	from	from	ADP
ajst-21816	134	18	the	the	DET
ajst-21816	134	19	combination	combination	NOUN
ajst-21816	134	20	of	of	ADP
ajst-21816	134	21	cross	cross	PROPN
ajst-21816	134	22	entropy	entropy	PROPN
ajst-21816	134	23	loss	loss	NOUN
ajst-21816	134	24	and	and	CCONJ
ajst-21816	134	25	center	center	NOUN
ajst-21816	134	26	loss	loss	NOUN
ajst-21816	134	27	.	.	PUNCT
ajst-21816	135	1	cross	cross	PROPN
ajst-21816	135	2	entropy	entropy	PROPN
ajst-21816	135	3	is	be	AUX
ajst-21816	135	4	an	an	DET
ajst-21816	135	5	effective	effective	ADJ
ajst-21816	135	6	measure	measure	NOUN
ajst-21816	135	7	used	use	VERB
ajst-21816	135	8	to	to	PART
ajst-21816	135	9	assess	assess	VERB
ajst-21816	135	10	the	the	DET
ajst-21816	135	11	degree	degree	NOUN
ajst-21816	135	12	of	of	ADP
ajst-21816	135	13	difference	difference	NOUN
ajst-21816	135	14	between	between	ADP
ajst-21816	135	15	the	the	DET
ajst-21816	135	16	true	true	ADJ
ajst-21816	135	17	probability	probability	NOUN
ajst-21816	135	18	distribution	distribution	NOUN
ajst-21816	135	19	and	and	CCONJ
ajst-21816	135	20	the	the	DET
ajst-21816	135	21	predicted	predict	VERB
ajst-21816	135	22	probability	probability	NOUN
ajst-21816	135	23	distribution	distribution	NOUN
ajst-21816	135	24	.	.	PUNCT
ajst-21816	136	1	it	it	PRON
ajst-21816	136	2	quantifies	quantify	VERB
ajst-21816	136	3	the	the	DET
ajst-21816	136	4	similarity	similarity	NOUN
ajst-21816	136	5	or	or	CCONJ
ajst-21816	136	6	difference	difference	NOUN
ajst-21816	136	7	between	between	ADP
ajst-21816	136	8	two	two	NUM
ajst-21816	136	9	distributions	distribution	NOUN
ajst-21816	136	10	by	by	ADP
ajst-21816	136	11	calculating	calculate	VERB
ajst-21816	136	12	the	the	DET
ajst-21816	136	13	amount	amount	NOUN
ajst-21816	136	14	of	of	ADP
ajst-21816	136	15	cross	cross	NOUN
ajst-21816	136	16	-	-	NOUN
ajst-21816	136	17	information	information	NOUN
ajst-21816	136	18	between	between	ADP
ajst-21816	136	19	them	they	PRON
ajst-21816	136	20	,	,	PUNCT
ajst-21816	136	21	thereby	thereby	ADV
ajst-21816	136	22	helping	help	VERB
ajst-21816	136	23	to	to	PART
ajst-21816	136	24	understand	understand	VERB
ajst-21816	136	25	the	the	DET
ajst-21816	136	26	predictive	predictive	ADJ
ajst-21816	136	27	effect	effect	NOUN
ajst-21816	136	28	of	of	ADP
ajst-21816	136	29	the	the	DET
ajst-21816	136	30	model	model	NOUN
ajst-21816	136	31	.	.	PUNCT
ajst-21816	137	1	in	in	ADP
ajst-21816	137	2	machine	machine	NOUN
ajst-21816	137	3	learning	learning	NOUN
ajst-21816	137	4	and	and	CCONJ
ajst-21816	137	5	deep	deep	ADJ
ajst-21816	137	6	learning	learning	NOUN
ajst-21816	137	7	,	,	PUNCT
ajst-21816	137	8	cross	cross	NOUN
ajst-21816	137	9	entropy	entropy	PROPN
ajst-21816	137	10	is	be	AUX
ajst-21816	137	11	often	often	ADV
ajst-21816	137	12	used	use	VERB
ajst-21816	137	13	as	as	ADP
ajst-21816	137	14	a	a	DET
ajst-21816	137	15	loss	loss	NOUN
ajst-21816	137	16	function	function	NOUN
ajst-21816	137	17	to	to	PART
ajst-21816	137	18	optimize	optimize	VERB
ajst-21816	137	19	the	the	DET
ajst-21816	137	20	training	training	NOUN
ajst-21816	137	21	process	process	NOUN
ajst-21816	137	22	of	of	ADP
ajst-21816	137	23	the	the	DET
ajst-21816	137	24	model	model	NOUN
ajst-21816	137	25	so	so	SCONJ
ajst-21816	137	26	that	that	SCONJ
ajst-21816	137	27	it	it	PRON
ajst-21816	137	28	better	well	ADV
ajst-21816	137	29	fits	fit	VERB
ajst-21816	137	30	the	the	DET
ajst-21816	137	31	real	real	ADJ
ajst-21816	137	32	data	datum	NOUN
ajst-21816	137	33	distribution	distribution	NOUN
ajst-21816	137	34	.	.	PUNCT
ajst-21816	138	1	the	the	DET
ajst-21816	138	2	formula	formula	NOUN
ajst-21816	138	3	is	be	AUX
ajst-21816	138	4	as	as	SCONJ
ajst-21816	138	5	follows	follow	VERB
ajst-21816	138	6	:	:	PUNCT
ajst-21816	138	7	1	1	NUM
ajst-21816	138	8	1	1	NUM
ajst-21816	138	9	1	1	NUM
ajst-21816	138	10	(	(	PUNCT
ajst-21816	138	11	)	)	PUNCT
ajst-21816	138	12	log	log	NOUN
ajst-21816	138	13	(	(	PUNCT
ajst-21816	138	14	(	(	PUNCT
ajst-21816	138	15	)	)	PUNCT
ajst-21816	138	16	)	)	PUNCT
ajst-21816	139	1	m	m	VERB
ajst-21816	139	2	n	n	VERB
ajst-21816	139	3	ce	ce	X
ajst-21816	139	4	ij	ij	INTJ
ajst-21816	139	5	ij	ij	INTJ
ajst-21816	140	1	i	i	INTJ
ajst-21816	140	2	j	j	PROPN
ajst-21816	141	1	l	l	NOUN
ajst-21816	141	2	p	p	X
ajst-21816	141	3	x	x	X
ajst-21816	141	4	q	q	PUNCT
ajst-21816	141	5	x	x	PUNCT
ajst-21816	141	6	m	m	VERB
ajst-21816	141	7			VERB
ajst-21816	141	8			NUM
ajst-21816	141	9			PROPN
ajst-21816	141	10			NOUN
ajst-21816	141	11			PRON
ajst-21816	141	12	(	(	PUNCT
ajst-21816	141	13	8)	8)	NUM
ajst-21816	141	14	where	where	SCONJ
ajst-21816	141	15	m	m	NOUN
ajst-21816	141	16	is	be	AUX
ajst-21816	141	17	the	the	DET
ajst-21816	141	18	number	number	NOUN
ajst-21816	141	19	of	of	ADP
ajst-21816	141	20	samples	sample	NOUN
ajst-21816	141	21	,	,	PUNCT
ajst-21816	141	22	n	n	X
ajst-21816	141	23	is	be	AUX
ajst-21816	141	24	the	the	DET
ajst-21816	141	25	number	number	NOUN
ajst-21816	141	26	of	of	ADP
ajst-21816	141	27	classes	class	NOUN
ajst-21816	141	28	,	,	PUNCT
ajst-21816	141	29	and	and	CCONJ
ajst-21816	141	30	(	(	PUNCT
ajst-21816	141	31	)	)	PUNCT
ajst-21816	141	32	ijp	ijp	NOUN
ajst-21816	141	33	x	x	VERB
ajst-21816	141	34	is	be	AUX
ajst-21816	141	35	used	use	VERB
ajst-21816	141	36	to	to	PART
ajst-21816	141	37	indicate	indicate	VERB
ajst-21816	141	38	whether	whether	SCONJ
ajst-21816	141	39	sample	sample	NOUN
ajst-21816	141	40	i	i	PRON
ajst-21816	141	41	belongs	belong	VERB
ajst-21816	141	42	to	to	ADP
ajst-21816	141	43	class	class	PROPN
ajst-21816	141	44	j	j	PROPN
ajst-21816	141	45	in	in	ADP
ajst-21816	141	46	the	the	DET
ajst-21816	141	47	real	real	ADJ
ajst-21816	141	48	distribution	distribution	NOUN
ajst-21816	141	49	.	.	PUNCT
ajst-21816	142	1	if	if	SCONJ
ajst-21816	142	2	sample	sample	NOUN
ajst-21816	142	3	i	i	PRON
ajst-21816	142	4	belongs	belong	VERB
ajst-21816	142	5	to	to	ADP
ajst-21816	142	6	class	class	PROPN
ajst-21816	142	7	j	j	PROPN
ajst-21816	142	8	in	in	ADP
ajst-21816	142	9	the	the	DET
ajst-21816	142	10	true	true	ADJ
ajst-21816	142	11	distribution	distribution	NOUN
ajst-21816	142	12	,	,	PUNCT
ajst-21816	142	13	then	then	ADV
ajst-21816	142	14	(	(	PUNCT
ajst-21816	142	15	)	)	PUNCT
ajst-21816	142	16	1ijp	1ijp	NUM
ajst-21816	142	17	x	x	SYM
ajst-21816	142	18			NOUN
ajst-21816	142	19	;	;	PUNCT
ajst-21816	142	20	if	if	SCONJ
ajst-21816	142	21	not	not	PART
ajst-21816	142	22	,	,	PUNCT
ajst-21816	142	23	then	then	ADV
ajst-21816	142	24	(	(	PUNCT
ajst-21816	142	25	)	)	PUNCT
ajst-21816	143	1	0ijp	0ijp	NUM
ajst-21816	143	2	x	x	X
ajst-21816	143	3			NOUN
ajst-21816	143	4	.	.	PUNCT
ajst-21816	144	1	(	(	PUNCT
ajst-21816	144	2	)	)	PUNCT
ajst-21816	144	3	ijq	ijq	PROPN
ajst-21816	144	4	x	x	VERB
ajst-21816	144	5	is	be	AUX
ajst-21816	144	6	used	use	VERB
ajst-21816	144	7	to	to	PART
ajst-21816	144	8	represent	represent	VERB
ajst-21816	144	9	the	the	DET
ajst-21816	144	10	probability	probability	NOUN
ajst-21816	144	11	that	that	SCONJ
ajst-21816	144	12	sample	sample	NOUN
ajst-21816	144	13	i	i	PRON
ajst-21816	144	14	belongs	belong	VERB
ajst-21816	144	15	to	to	ADP
ajst-21816	144	16	class	class	PROPN
ajst-21816	144	17	j	j	PROPN
ajst-21816	144	18	in	in	ADP
ajst-21816	144	19	the	the	DET
ajst-21816	144	20	prediction	prediction	NOUN
ajst-21816	144	21	distribution	distribution	NOUN
ajst-21816	144	22	.	.	PUNCT
ajst-21816	145	1	the	the	DET
ajst-21816	145	2	cross	cross	ADJ
ajst-21816	145	3	-	-	ADJ
ajst-21816	145	4	entropy	entropy	ADJ
ajst-21816	145	5	loss	loss	NOUN
ajst-21816	145	6	function	function	NOUN
ajst-21816	145	7	adopts	adopt	VERB
ajst-21816	145	8	the	the	DET
ajst-21816	145	9	inter	inter	ADJ
ajst-21816	145	10	-	-	ADJ
ajst-21816	145	11	class	class	ADJ
ajst-21816	145	12	competition	competition	NOUN
ajst-21816	145	13	mechanism	mechanism	NOUN
ajst-21816	145	14	and	and	CCONJ
ajst-21816	145	15	is	be	AUX
ajst-21816	145	16	good	good	ADJ
ajst-21816	145	17	at	at	ADP
ajst-21816	145	18	learning	learn	VERB
ajst-21816	145	19	inter	inter	ADJ
ajst-21816	145	20	-	-	ADJ
ajst-21816	145	21	class	class	ADJ
ajst-21816	145	22	information	information	NOUN
ajst-21816	145	23	,	,	PUNCT
ajst-21816	145	24	while	while	SCONJ
ajst-21816	145	25	the	the	DET
ajst-21816	145	26	central	central	ADJ
ajst-21816	145	27	loss	loss	NOUN
ajst-21816	145	28	function	function	NOUN
ajst-21816	145	29	is	be	AUX
ajst-21816	145	30	dedicated	dedicate	VERB
ajst-21816	145	31	to	to	ADP
ajst-21816	145	32	constraining	constrain	VERB
ajst-21816	145	33	intra	intra	ADJ
ajst-21816	145	34	-	-	ADJ
ajst-21816	145	35	class	class	ADJ
ajst-21816	145	36	compactness	compactness	NOUN
ajst-21816	145	37	.	.	PUNCT
ajst-21816	146	1	its	its	PRON
ajst-21816	146	2	formula	formula	NOUN
ajst-21816	146	3	is	be	AUX
ajst-21816	146	4	as	as	SCONJ
ajst-21816	146	5	follows	follow	VERB
ajst-21816	146	6	:	:	PUNCT
ajst-21816	146	7	2	2	NUM
ajst-21816	146	8	2	2	NUM
ajst-21816	146	9	0	0	NUM
ajst-21816	146	10	1	1	NUM
ajst-21816	146	11	||	||	NOUN
ajst-21816	146	12	||	||	NOUN
ajst-21816	146	13	2	2	NUM
ajst-21816	147	1	i	i	PRON
ajst-21816	147	2	n	n	VERB
ajst-21816	147	3	c	c	VERB
ajst-21816	148	1	i	i	PRON
ajst-21816	148	2	y	y	VERB
ajst-21816	149	1	i	i	NOUN
ajst-21816	149	2	l	l	NOUN
ajst-21816	150	1	x	x	PUNCT
ajst-21816	150	2	c	c	X
ajst-21816	150	3			NOUN
ajst-21816	150	4			NUM
ajst-21816	151	1			X
ajst-21816	151	2	(	(	PUNCT
ajst-21816	151	3	9	9	NUM
ajst-21816	151	4	)	)	PUNCT
ajst-21816	151	5	where	where	SCONJ
ajst-21816	151	6	,	,	PUNCT
ajst-21816	151	7	n	n	PRON
ajst-21816	151	8	represents	represent	VERB
ajst-21816	151	9	the	the	DET
ajst-21816	151	10	number	number	NOUN
ajst-21816	151	11	of	of	ADP
ajst-21816	151	12	training	training	NOUN
ajst-21816	151	13	samples	sample	NOUN
ajst-21816	151	14	,	,	PUNCT
ajst-21816	151	15	ix	ix	ADV
ajst-21816	151	16	represents	represent	VERB
ajst-21816	151	17	the	the	DET
ajst-21816	151	18	output	output	NOUN
ajst-21816	151	19	value	value	NOUN
ajst-21816	151	20	of	of	ADP
ajst-21816	151	21	the	the	DET
ajst-21816	151	22	iy	iy	PROPN
ajst-21816	151	23	pedestrian	pedestrian	NOUN
ajst-21816	151	24	passing	pass	VERB
ajst-21816	151	25	through	through	ADP
ajst-21816	151	26	the	the	DET
ajst-21816	151	27	network	network	NOUN
ajst-21816	151	28	,	,	PUNCT
ajst-21816	151	29	and	and	CCONJ
ajst-21816	151	30	iy	iy	PROPN
ajst-21816	151	31	c	c	PROPN
ajst-21816	151	32	represents	represent	VERB
ajst-21816	151	33	the	the	DET
ajst-21816	151	34	feature	feature	NOUN
ajst-21816	151	35	center	center	NOUN
ajst-21816	151	36	of	of	ADP
ajst-21816	151	37	the	the	DET
ajst-21816	151	38	iy	iy	PROPN
ajst-21816	151	39	category	category	NOUN
ajst-21816	151	40	.	.	PUNCT
ajst-21816	152	1	the	the	DET
ajst-21816	152	2	center	center	NOUN
ajst-21816	152	3	loss	loss	NOUN
ajst-21816	152	4	function	function	NOUN
ajst-21816	152	5	has	have	VERB
ajst-21816	152	6	no	no	DET
ajst-21816	152	7	obvious	obvious	ADJ
ajst-21816	152	8	improvement	improvement	NOUN
ajst-21816	152	9	on	on	ADP
ajst-21816	152	10	the	the	DET
ajst-21816	152	11	model	model	NOUN
ajst-21816	152	12	with	with	ADP
ajst-21816	152	13	a	a	DET
ajst-21816	152	14	small	small	ADJ
ajst-21816	152	15	number	number	NOUN
ajst-21816	152	16	of	of	ADP
ajst-21816	152	17	classifications	classification	NOUN
ajst-21816	152	18	,	,	PUNCT
ajst-21816	152	19	and	and	CCONJ
ajst-21816	152	20	it	it	PRON
ajst-21816	152	21	is	be	AUX
ajst-21816	152	22	more	more	ADV
ajst-21816	152	23	suitable	suitable	ADJ
ajst-21816	152	24	for	for	ADP
ajst-21816	152	25	the	the	DET
ajst-21816	152	26	model	model	NOUN
ajst-21816	152	27	with	with	ADP
ajst-21816	152	28	a	a	DET
ajst-21816	152	29	large	large	ADJ
ajst-21816	152	30	number	number	NOUN
ajst-21816	152	31	of	of	ADP
ajst-21816	152	32	classifications	classification	NOUN
ajst-21816	152	33	.	.	PUNCT
ajst-21816	153	1	after	after	ADP
ajst-21816	153	2	using	use	VERB
ajst-21816	153	3	this	this	DET
ajst-21816	153	4	loss	loss	NOUN
ajst-21816	153	5	function	function	NOUN
ajst-21816	153	6	,	,	PUNCT
ajst-21816	153	7	the	the	DET
ajst-21816	153	8	samples	sample	NOUN
ajst-21816	153	9	of	of	ADP
ajst-21816	153	10	each	each	DET
ajst-21816	153	11	class	class	NOUN
ajst-21816	153	12	become	become	VERB
ajst-21816	153	13	more	more	ADV
ajst-21816	153	14	aggregated	aggregated	ADJ
ajst-21816	153	15	,	,	PUNCT
ajst-21816	153	16	so	so	CCONJ
ajst-21816	153	17	the	the	DET
ajst-21816	153	18	distance	distance	NOUN
ajst-21816	153	19	between	between	ADP
ajst-21816	153	20	classes	class	NOUN
ajst-21816	153	21	increases	increase	NOUN
ajst-21816	153	22	.	.	PUNCT
ajst-21816	154	1	with	with	ADP
ajst-21816	154	2	the	the	DET
ajst-21816	154	3	increasing	increase	VERB
ajst-21816	154	4	loss	loss	NOUN
ajst-21816	154	5	weight	weight	NOUN
ajst-21816	154	6	value	value	NOUN
ajst-21816	154	7	,	,	PUNCT
ajst-21816	154	8	the	the	DET
ajst-21816	154	9	class	class	NOUN
ajst-21816	154	10	spacing	spacing	NOUN
ajst-21816	154	11	becomes	become	VERB
ajst-21816	154	12	larger	large	ADJ
ajst-21816	154	13	and	and	CCONJ
ajst-21816	154	14	larger	large	ADJ
ajst-21816	154	15	.	.	PUNCT
ajst-21816	155	1	it	it	PRON
ajst-21816	155	2	achieves	achieve	VERB
ajst-21816	155	3	the	the	DET
ajst-21816	155	4	effect	effect	NOUN
ajst-21816	155	5	of	of	ADP
ajst-21816	155	6	increasing	increase	VERB
ajst-21816	155	7	the	the	DET
ajst-21816	155	8	class	class	NOUN
ajst-21816	155	9	distance	distance	NOUN
ajst-21816	155	10	and	and	CCONJ
ajst-21816	155	11	narrowing	narrow	VERB
ajst-21816	155	12	the	the	DET
ajst-21816	155	13	distance	distance	NOUN
ajst-21816	155	14	within	within	ADP
ajst-21816	155	15	the	the	DET
ajst-21816	155	16	class	class	NOUN
ajst-21816	155	17	.	.	PUNCT
ajst-21816	156	1	the	the	DET
ajst-21816	156	2	specific	specific	ADJ
ajst-21816	156	3	formula	formula	NOUN
ajst-21816	156	4	of	of	ADP
ajst-21816	156	5	the	the	DET
ajst-21816	156	6	joint	joint	ADJ
ajst-21816	156	7	loss	loss	NOUN
ajst-21816	156	8	function	function	NOUN
ajst-21816	156	9	ul	ul	INTJ
ajst-21816	156	10	,	,	PUNCT
ajst-21816	156	11	which	which	PRON
ajst-21816	156	12	consists	consist	VERB
ajst-21816	156	13	of	of	ADP
ajst-21816	156	14	the	the	DET
ajst-21816	156	15	cross	cross	PROPN
ajst-21816	156	16	entropy	entropy	PROPN
ajst-21816	156	17	loss	loss	NOUN
ajst-21816	156	18	function	function	NOUN
ajst-21816	156	19	and	and	CCONJ
ajst-21816	156	20	the	the	DET
ajst-21816	156	21	central	central	ADJ
ajst-21816	156	22	loss	loss	NOUN
ajst-21816	156	23	function	function	NOUN
ajst-21816	156	24	,	,	PUNCT
ajst-21816	156	25	is	be	AUX
ajst-21816	156	26	shown	show	VERB
ajst-21816	156	27	in	in	ADP
ajst-21816	156	28	equation	equation	NOUN
ajst-21816	156	29	(	(	PUNCT
ajst-21816	156	30	10	10	NUM
ajst-21816	156	31	)	)	PUNCT
ajst-21816	156	32	,	,	PUNCT
ajst-21816	156	33	where	where	SCONJ
ajst-21816	156	34			NOUN
ajst-21816	156	35	is	be	AUX
ajst-21816	156	36	the	the	DET
ajst-21816	156	37	weight	weight	NOUN
ajst-21816	156	38	parameter	parameter	NOUN
ajst-21816	156	39	.	.	PUNCT
ajst-21816	157	1	u	u	PROPN
ajst-21816	157	2	ce	ce	PROPN
ajst-21816	157	3	cl	cl	PROPN
ajst-21816	157	4	l	l	PROPN
ajst-21816	157	5	l	l	NUM
ajst-21816	157	6			PUNCT
ajst-21816	157	7	(	(	PUNCT
ajst-21816	157	8	10	10	NUM
ajst-21816	157	9	)	)	PUNCT
ajst-21816	157	10	3	3	NUM
ajst-21816	157	11	.	.	PUNCT
ajst-21816	157	12	experimental	experimental	ADJ
ajst-21816	157	13	results	result	NOUN
ajst-21816	157	14	and	and	CCONJ
ajst-21816	157	15	analysis	analysis	NOUN
ajst-21816	157	16	3.1	3.1	NUM
ajst-21816	157	17	.	.	PUNCT
ajst-21816	158	1	casia	casia	NOUN
ajst-21816	158	2	-	-	PUNCT
ajst-21816	158	3	b	b	PROPN
ajst-21816	158	4	data	datum	NOUN
ajst-21816	158	5	set	set	VERB
ajst-21816	158	6	figure	figure	NOUN
ajst-21816	158	7	8	8	NUM
ajst-21816	158	8	.	.	PUNCT
ajst-21816	158	9	silhouette	silhouette	NOUN
ajst-21816	158	10	at	at	ADP
ajst-21816	158	11	90	90	NUM
ajst-21816	158	12	°	°	NOUN
ajst-21816	158	13	of	of	ADP
ajst-21816	158	14	partial	partial	ADJ
ajst-21816	158	15	gait	gait	NOUN
ajst-21816	158	16	the	the	DET
ajst-21816	158	17	casia	casia	NOUN
ajst-21816	158	18	-	-	PUNCT
ajst-21816	158	19	b	b	NOUN
ajst-21816	158	20	dataset	dataset	NOUN
ajst-21816	158	21	is	be	AUX
ajst-21816	158	22	a	a	DET
ajst-21816	158	23	large	large	ADJ
ajst-21816	158	24	-	-	PUNCT
ajst-21816	158	25	scale	scale	NOUN
ajst-21816	158	26	,	,	PUNCT
ajst-21816	158	27	multi	multi	ADJ
ajst-21816	158	28	-	-	ADJ
ajst-21816	158	29	view	view	ADJ
ajst-21816	158	30	gait	gait	NOUN
ajst-21816	158	31	library	library	NOUN
ajst-21816	158	32	provided	provide	VERB
ajst-21816	158	33	by	by	ADP
ajst-21816	158	34	the	the	DET
ajst-21816	158	35	institute	institute	PROPN
ajst-21816	158	36	of	of	ADP
ajst-21816	158	37	automation	automation	NOUN
ajst-21816	158	38	,	,	PUNCT
ajst-21816	158	39	chinese	chinese	PROPN
ajst-21816	158	40	academy	academy	PROPN
ajst-21816	158	41	of	of	ADP
ajst-21816	158	42	sciences	sciences	PROPN
ajst-21816	158	43	.	.	PUNCT
ajst-21816	159	1	it	it	PRON
ajst-21816	159	2	contains	contain	VERB
ajst-21816	159	3	raw	raw	ADJ
ajst-21816	159	4	video	video	NOUN
ajst-21816	159	5	data	datum	NOUN
ajst-21816	159	6	and	and	CCONJ
ajst-21816	159	7	contours	contours	NOUN
ajst-21816	159	8	of	of	ADP
ajst-21816	159	9	124	124	NUM
ajst-21816	159	10	subjects	subject	NOUN
ajst-21816	159	11	.	.	PUNCT
ajst-21816	160	1	each	each	DET
ajst-21816	160	2	subject	subject	NOUN
ajst-21816	160	3	had	have	VERB
ajst-21816	160	4	11	11	NUM
ajst-21816	160	5	viewing	view	VERB
ajst-21816	160	6	angles	angle	NOUN
ajst-21816	160	7	and	and	CCONJ
ajst-21816	160	8	three	three	NUM
ajst-21816	160	9	walking	walk	VERB
ajst-21816	160	10	conditions	condition	NOUN
ajst-21816	160	11	,	,	PUNCT
ajst-21816	160	12	namely	namely	ADV
ajst-21816	160	13	normal	normal	ADJ
ajst-21816	160	14	walking	walking	NOUN
ajst-21816	160	15	(	(	PUNCT
ajst-21816	160	16	nm	nm	NOUN
ajst-21816	160	17	)	)	PUNCT
ajst-21816	160	18	,	,	PUNCT
ajst-21816	160	19	backpack	backpack	NOUN
ajst-21816	160	20	walking	walk	VERB
ajst-21816	160	21	(	(	PUNCT
ajst-21816	160	22	bg	bg	NOUN
ajst-21816	160	23	)	)	PUNCT
ajst-21816	160	24	,	,	PUNCT
ajst-21816	160	25	and	and	CCONJ
ajst-21816	160	26	coat	coat	NOUN
ajst-21816	160	27	walking	walk	VERB
ajst-21816	160	28	(	(	PUNCT
ajst-21816	160	29	cl	cl	NOUN
ajst-21816	160	30	)	)	PUNCT
ajst-21816	160	31	.	.	PUNCT
ajst-21816	161	1	each	each	DET
ajst-21816	161	2	subject	subject	NOUN
ajst-21816	161	3	had	have	VERB
ajst-21816	161	4	6	6	NUM
ajst-21816	161	5	video	video	NOUN
ajst-21816	161	6	sequences	sequence	NOUN
ajst-21816	161	7	in	in	ADP
ajst-21816	161	8	nm	nm	DET
ajst-21816	161	9	state	state	NOUN
ajst-21816	161	10	,	,	PUNCT
ajst-21816	161	11	2	2	NUM
ajst-21816	161	12	video	video	NOUN
ajst-21816	161	13	sequences	sequence	NOUN
ajst-21816	161	14	in	in	ADP
ajst-21816	161	15	bg	bg	PROPN
ajst-21816	161	16	state	state	NOUN
ajst-21816	161	17	,	,	PUNCT
ajst-21816	161	18	and	and	CCONJ
ajst-21816	161	19	2	2	NUM
ajst-21816	161	20	video	video	NOUN
ajst-21816	161	21	sequences	sequence	NOUN
ajst-21816	161	22	in	in	ADP
ajst-21816	161	23	cl	cl	NOUN
ajst-21816	161	24	state	state	NOUN
ajst-21816	161	25	.	.	PUNCT
ajst-21816	162	1	therefore	therefore	ADV
ajst-21816	162	2	,	,	PUNCT
ajst-21816	162	3	each	each	DET
ajst-21816	162	4	subject	subject	NOUN
ajst-21816	162	5	has	have	VERB
ajst-21816	162	6	11×(6	11×(6	NUM
ajst-21816	162	7	+	+	NOUN
ajst-21816	162	8	2	2	NUM
ajst-21816	162	9	+	+	ADJ
ajst-21816	162	10	2	2	NUM
ajst-21816	162	11	)	)	PUNCT
ajst-21816	162	12	=	=	SYM
ajst-21816	162	13	110	110	NUM
ajst-21816	162	14	gait	gait	ADJ
ajst-21816	162	15	sequences	sequence	NOUN
ajst-21816	162	16	.	.	PUNCT
ajst-21816	163	1	the	the	DET
ajst-21816	163	2	partial	partial	ADJ
ajst-21816	163	3	contours	contours	NOUN
ajst-21816	163	4	of	of	ADP
ajst-21816	163	5	the	the	DET
ajst-21816	163	6	three	three	NUM
ajst-21816	163	7	walking	walk	VERB
ajst-21816	163	8	states	state	NOUN
ajst-21816	163	9	of	of	ADP
ajst-21816	163	10	the	the	DET
ajst-21816	163	11	casia	casia	NOUN
ajst-21816	163	12	-	-	PUNCT
ajst-21816	163	13	b	b	NOUN
ajst-21816	163	14	dataset	dataset	NOUN
ajst-21816	163	15	are	be	AUX
ajst-21816	163	16	shown	show	VERB
ajst-21816	163	17	in	in	ADP
ajst-21816	163	18	figure	figure	NOUN
ajst-21816	163	19	8	8	NUM
ajst-21816	163	20	.	.	PUNCT
ajst-21816	163	21	292	292	NUM
ajst-21816	163	22	3.2	3.2	NUM
ajst-21816	163	23	.	.	PUNCT
ajst-21816	164	1	experimental	experimental	ADJ
ajst-21816	164	2	setup	setup	NOUN
ajst-21816	164	3	in	in	ADP
ajst-21816	164	4	this	this	DET
ajst-21816	164	5	paper	paper	NOUN
ajst-21816	164	6	,	,	PUNCT
ajst-21816	164	7	the	the	DET
ajst-21816	164	8	samples	sample	NOUN
ajst-21816	164	9	in	in	ADP
ajst-21816	164	10	casia	casia	NOUN
ajst-21816	164	11	-	-	PUNCT
ajst-21816	164	12	b	b	NOUN
ajst-21816	164	13	dataset	dataset	NOUN
ajst-21816	164	14	were	be	AUX
ajst-21816	164	15	preprocessed	preprocesse	VERB
ajst-21816	164	16	,	,	PUNCT
ajst-21816	164	17	and	and	CCONJ
ajst-21816	164	18	the	the	DET
ajst-21816	164	19	gait	gait	ADJ
ajst-21816	164	20	profile	profile	NOUN
ajst-21816	164	21	with	with	ADP
ajst-21816	164	22	size	size	NOUN
ajst-21816	164	23	128×88	128×88	NUM
ajst-21816	164	24	was	be	AUX
ajst-21816	164	25	obtained	obtain	VERB
ajst-21816	164	26	after	after	ADP
ajst-21816	164	27	normalization	normalization	NOUN
ajst-21816	164	28	.	.	PUNCT
ajst-21816	165	1	during	during	ADP
ajst-21816	165	2	training	training	NOUN
ajst-21816	165	3	,	,	PUNCT
ajst-21816	165	4	30	30	NUM
ajst-21816	165	5	consecutive	consecutive	ADJ
ajst-21816	165	6	gait	gait	NOUN
ajst-21816	165	7	contour	contour	NOUN
ajst-21816	165	8	diagrams	diagram	NOUN
ajst-21816	165	9	were	be	AUX
ajst-21816	165	10	randomly	randomly	ADV
ajst-21816	165	11	selected	select	VERB
ajst-21816	165	12	from	from	ADP
ajst-21816	165	13	the	the	DET
ajst-21816	165	14	sequence	sequence	NOUN
ajst-21816	165	15	as	as	ADP
ajst-21816	165	16	input	input	NOUN
ajst-21816	165	17	,	,	PUNCT
ajst-21816	165	18	the	the	DET
ajst-21816	165	19	learning	learning	NOUN
ajst-21816	165	20	rate	rate	NOUN
ajst-21816	165	21	was	be	AUX
ajst-21816	165	22	set	set	VERB
ajst-21816	165	23	to	to	ADP
ajst-21816	165	24	1e-4	1e-4	NUM
ajst-21816	165	25	,	,	PUNCT
ajst-21816	165	26	and	and	CCONJ
ajst-21816	165	27	the	the	DET
ajst-21816	165	28	batch	batch	NOUN
ajst-21816	165	29	size	size	NOUN
ajst-21816	165	30	was	be	AUX
ajst-21816	165	31	set	set	VERB
ajst-21816	165	32	as	as	ADP
ajst-21816	165	33	:	:	PUNCT
ajst-21816	165	34	p×k	p×k	PROPN
ajst-21816	165	35	,	,	PUNCT
ajst-21816	165	36	so	so	SCONJ
ajst-21816	165	37	that	that	SCONJ
ajst-21816	165	38	p=8	p=8	VERB
ajst-21816	165	39	,	,	PUNCT
ajst-21816	165	40	k=16	k=16	PROPN
ajst-21816	165	41	,	,	PUNCT
ajst-21816	165	42	where	where	SCONJ
ajst-21816	165	43	p	p	NOUN
ajst-21816	165	44	represented	represent	VERB
ajst-21816	165	45	the	the	DET
ajst-21816	165	46	number	number	NOUN
ajst-21816	165	47	of	of	ADP
ajst-21816	165	48	classes	class	NOUN
ajst-21816	165	49	,	,	PUNCT
ajst-21816	165	50	and	and	CCONJ
ajst-21816	165	51	k	k	PROPN
ajst-21816	165	52	represented	represent	VERB
ajst-21816	165	53	the	the	DET
ajst-21816	165	54	number	number	NOUN
ajst-21816	165	55	of	of	ADP
ajst-21816	165	56	training	training	NOUN
ajst-21816	165	57	samples	sample	NOUN
ajst-21816	165	58	for	for	ADP
ajst-21816	165	59	each	each	DET
ajst-21816	165	60	class	class	NOUN
ajst-21816	165	61	in	in	ADP
ajst-21816	165	62	the	the	DET
ajst-21816	165	63	batch	batch	NOUN
ajst-21816	165	64	.	.	PUNCT
ajst-21816	166	1	in	in	ADP
ajst-21816	166	2	the	the	DET
ajst-21816	166	3	experiment	experiment	NOUN
ajst-21816	166	4	,	,	PUNCT
ajst-21816	166	5	the	the	DET
ajst-21816	166	6	large	large	ADJ
ajst-21816	166	7	sample	sample	NOUN
ajst-21816	166	8	(	(	PUNCT
ajst-21816	166	9	lt	lt	NOUN
ajst-21816	166	10	)	)	PUNCT
ajst-21816	166	11	training	training	NOUN
ajst-21816	166	12	method	method	NOUN
ajst-21816	166	13	was	be	AUX
ajst-21816	166	14	used	use	VERB
ajst-21816	166	15	in	in	ADP
ajst-21816	166	16	this	this	DET
ajst-21816	166	17	paper	paper	NOUN
ajst-21816	166	18	.	.	PUNCT
ajst-21816	167	1	the	the	DET
ajst-21816	167	2	data	datum	NOUN
ajst-21816	167	3	of	of	ADP
ajst-21816	167	4	the	the	DET
ajst-21816	167	5	first	first	ADJ
ajst-21816	167	6	74	74	NUM
ajst-21816	167	7	subjects	subject	NOUN
ajst-21816	167	8	were	be	AUX
ajst-21816	167	9	used	use	VERB
ajst-21816	167	10	for	for	ADP
ajst-21816	167	11	training	training	NOUN
ajst-21816	167	12	,	,	PUNCT
ajst-21816	167	13	and	and	CCONJ
ajst-21816	167	14	the	the	DET
ajst-21816	167	15	data	datum	NOUN
ajst-21816	167	16	of	of	ADP
ajst-21816	167	17	the	the	DET
ajst-21816	167	18	remaining	remain	VERB
ajst-21816	167	19	50	50	NUM
ajst-21816	167	20	subjects	subject	NOUN
ajst-21816	167	21	were	be	AUX
ajst-21816	167	22	used	use	VERB
ajst-21816	167	23	for	for	ADP
ajst-21816	167	24	testing	testing	NOUN
ajst-21816	167	25	.	.	PUNCT
ajst-21816	168	1	in	in	ADP
ajst-21816	168	2	the	the	DET
ajst-21816	168	3	test	test	NOUN
ajst-21816	168	4	set	set	NOUN
ajst-21816	168	5	,	,	PUNCT
ajst-21816	168	6	the	the	DET
ajst-21816	168	7	first	first	ADJ
ajst-21816	168	8	4	4	NUM
ajst-21816	168	9	gait	gait	ADJ
ajst-21816	168	10	sequences	sequence	NOUN
ajst-21816	168	11	in	in	ADP
ajst-21816	168	12	the	the	DET
ajst-21816	168	13	gait	gait	ADJ
ajst-21816	168	14	data	datum	NOUN
ajst-21816	168	15	of	of	ADP
ajst-21816	168	16	50	50	NUM
ajst-21816	168	17	subjects	subject	NOUN
ajst-21816	168	18	in	in	ADP
ajst-21816	168	19	the	the	DET
ajst-21816	168	20	nm	nm	ADJ
ajst-21816	168	21	walking	walk	VERB
ajst-21816	168	22	state	state	NOUN
ajst-21816	168	23	were	be	AUX
ajst-21816	168	24	taken	take	VERB
ajst-21816	168	25	as	as	ADP
ajst-21816	168	26	the	the	DET
ajst-21816	168	27	registration	registration	NOUN
ajst-21816	168	28	set	set	NOUN
ajst-21816	168	29	,	,	PUNCT
ajst-21816	168	30	expressed	express	VERB
ajst-21816	168	31	as	as	ADP
ajst-21816	168	32	:	:	PUNCT
ajst-21816	168	33	nm#01	nm#01	PROPN
ajst-21816	168	34	-	-	PROPN
ajst-21816	168	35	nm#04	nm#04	PROPN
ajst-21816	168	36	.	.	PUNCT
ajst-21816	169	1	the	the	DET
ajst-21816	169	2	remaining	remain	VERB
ajst-21816	169	3	sequences	sequence	NOUN
ajst-21816	169	4	constitute	constitute	VERB
ajst-21816	169	5	the	the	DET
ajst-21816	169	6	verification	verification	NOUN
ajst-21816	169	7	set	set	NOUN
ajst-21816	169	8	,	,	PUNCT
ajst-21816	169	9	that	that	ADV
ajst-21816	169	10	is	is	ADV
ajst-21816	169	11	,	,	PUNCT
ajst-21816	169	12	the	the	DET
ajst-21816	169	13	verification	verification	NOUN
ajst-21816	169	14	set	set	NOUN
ajst-21816	169	15	includes	include	VERB
ajst-21816	169	16	the	the	DET
ajst-21816	169	17	remaining	remain	VERB
ajst-21816	169	18	2	2	NUM
ajst-21816	169	19	gait	gait	NOUN
ajst-21816	169	20	sequences	sequence	NOUN
ajst-21816	169	21	of	of	ADP
ajst-21816	169	22	the	the	DET
ajst-21816	169	23	subject	subject	NOUN
ajst-21816	169	24	in	in	ADP
ajst-21816	169	25	the	the	DET
ajst-21816	169	26	nm	nm	ADJ
ajst-21816	169	27	walking	walk	VERB
ajst-21816	169	28	state	state	NOUN
ajst-21816	169	29	,	,	PUNCT
ajst-21816	169	30	which	which	PRON
ajst-21816	169	31	are	be	AUX
ajst-21816	169	32	expressed	express	VERB
ajst-21816	169	33	as	as	ADP
ajst-21816	169	34	:	:	PUNCT
ajst-21816	169	35	nm#05	nm#05	NUM
ajst-21816	169	36	,	,	PUNCT
ajst-21816	169	37	nm#06	nm#06	PROPN
ajst-21816	169	38	;	;	PUNCT
ajst-21816	169	39	the	the	DET
ajst-21816	169	40	two	two	NUM
ajst-21816	169	41	gait	gait	ADJ
ajst-21816	169	42	sequences	sequence	NOUN
ajst-21816	169	43	in	in	ADP
ajst-21816	169	44	the	the	DET
ajst-21816	169	45	bg	bg	PROPN
ajst-21816	169	46	walking	walk	VERB
ajst-21816	169	47	state	state	NOUN
ajst-21816	169	48	are	be	AUX
ajst-21816	169	49	represented	represent	VERB
ajst-21816	169	50	as	as	ADP
ajst-21816	169	51	bg#01	bg#01	NOUN
ajst-21816	169	52	,	,	PUNCT
ajst-21816	169	53	bg#02	bg#02	PROPN
ajst-21816	169	54	;	;	PUNCT
ajst-21816	169	55	the	the	DET
ajst-21816	169	56	two	two	NUM
ajst-21816	169	57	gait	gait	ADJ
ajst-21816	169	58	sequences	sequence	NOUN
ajst-21816	169	59	in	in	ADP
ajst-21816	169	60	cl	cl	NOUN
ajst-21816	169	61	walking	walk	VERB
ajst-21816	169	62	state	state	NOUN
ajst-21816	169	63	are	be	AUX
ajst-21816	169	64	represented	represent	VERB
ajst-21816	169	65	as	as	ADP
ajst-21816	169	66	:	:	PUNCT
ajst-21816	169	67	cl#01	cl#01	PROPN
ajst-21816	169	68	,	,	PUNCT
ajst-21816	169	69	cl#02	cl#02	PROPN
ajst-21816	169	70	.	.	PUNCT
ajst-21816	170	1	each	each	DET
ajst-21816	170	2	gait	gait	ADJ
ajst-21816	170	3	sequence	sequence	NOUN
ajst-21816	170	4	mentioned	mention	VERB
ajst-21816	170	5	above	above	ADV
ajst-21816	170	6	contains	contain	VERB
ajst-21816	170	7	gait	gait	PROPN
ajst-21816	170	8	data	datum	NOUN
ajst-21816	170	9	from	from	ADP
ajst-21816	170	10	11	11	NUM
ajst-21816	170	11	angles	angle	NOUN
ajst-21816	170	12	,	,	PUNCT
ajst-21816	170	13	and	and	CCONJ
ajst-21816	170	14	the	the	DET
ajst-21816	170	15	specific	specific	ADJ
ajst-21816	170	16	settings	setting	NOUN
ajst-21816	170	17	of	of	ADP
ajst-21816	170	18	the	the	DET
ajst-21816	170	19	registration	registration	NOUN
ajst-21816	170	20	set	set	NOUN
ajst-21816	170	21	are	be	AUX
ajst-21816	170	22	shown	show	VERB
ajst-21816	170	23	in	in	ADP
ajst-21816	170	24	figure	figure	NOUN
ajst-21816	170	25	9	9	NUM
ajst-21816	170	26	.	.	PUNCT
ajst-21816	170	27	figure	figure	NOUN
ajst-21816	170	28	9	9	NUM
ajst-21816	170	29	.	.	PUNCT
ajst-21816	171	1	gallery	gallery	NOUN
ajst-21816	171	2	settings	setting	VERB
ajst-21816	171	3	3.3	3.3	NUM
ajst-21816	171	4	.	.	PUNCT
ajst-21816	172	1	contrast	contrast	NOUN
ajst-21816	172	2	test	test	NOUN
ajst-21816	172	3	table	table	NOUN
ajst-21816	172	4	3	3	NUM
ajst-21816	172	5	-	-	SYM
ajst-21816	172	6	1	1	NUM
ajst-21816	172	7	,	,	PUNCT
ajst-21816	172	8	3	3	NUM
ajst-21816	172	9	-	-	SYM
ajst-21816	172	10	2	2	NUM
ajst-21816	172	11	,	,	PUNCT
ajst-21816	172	12	and	and	CCONJ
ajst-21816	172	13	3	3	NUM
ajst-21816	172	14	-	-	SYM
ajst-21816	172	15	3	3	NUM
ajst-21816	172	16	show	show	VERB
ajst-21816	172	17	the	the	DET
ajst-21816	172	18	recognition	recognition	NOUN
ajst-21816	172	19	results	result	VERB
ajst-21816	172	20	between	between	ADP
ajst-21816	172	21	the	the	DET
ajst-21816	172	22	proposed	propose	VERB
ajst-21816	172	23	gait	gait	NOUN
ajst-21816	172	24	model	model	NOUN
ajst-21816	172	25	and	and	CCONJ
ajst-21816	172	26	four	four	NUM
ajst-21816	172	27	existing	exist	VERB
ajst-21816	172	28	gait	gait	NOUN
ajst-21816	172	29	models	model	NOUN
ajst-21816	172	30	(	(	PUNCT
ajst-21816	172	31	srn[17	srn[17	PROPN
ajst-21816	172	32	]	]	PUNCT
ajst-21816	172	33	,	,	PUNCT
ajst-21816	172	34	gaitset	gaitset	NOUN
ajst-21816	172	35	,	,	PUNCT
ajst-21816	172	36	capsnet	capsnet	NOUN
ajst-21816	172	37	,	,	PUNCT
ajst-21816	172	38	and	and	CCONJ
ajst-21816	172	39	gaitpart	gaitpart	NOUN
ajst-21816	172	40	)	)	PUNCT
ajst-21816	172	41	in	in	ADP
ajst-21816	172	42	nm	nm	PROPN
ajst-21816	172	43	,	,	PUNCT
ajst-21816	172	44	bg	bg	PROPN
ajst-21816	172	45	,	,	PUNCT
ajst-21816	172	46	and	and	CCONJ
ajst-21816	172	47	cl	cl	NOUN
ajst-21816	172	48	walking	walk	VERB
ajst-21816	172	49	states	state	NOUN
ajst-21816	172	50	on	on	ADP
ajst-21816	172	51	the	the	DET
ajst-21816	172	52	casia	casia	NOUN
ajst-21816	172	53	-	-	PUNCT
ajst-21816	172	54	b	b	NOUN
ajst-21816	172	55	data	datum	NOUN
ajst-21816	172	56	set	set	VERB
ajst-21816	172	57	under	under	ADP
ajst-21816	172	58	the	the	DET
ajst-21816	172	59	large	large	ADJ
ajst-21816	172	60	-	-	PUNCT
ajst-21816	172	61	sample	sample	NOUN
ajst-21816	172	62	(	(	PUNCT
ajst-21816	172	63	lt	lt	NOUN
ajst-21816	172	64	)	)	PUNCT
ajst-21816	172	65	training	training	NOUN
ajst-21816	172	66	mode	mode	NOUN
ajst-21816	172	67	.	.	PUNCT
ajst-21816	173	1	it	it	PRON
ajst-21816	173	2	should	should	AUX
ajst-21816	173	3	be	be	AUX
ajst-21816	173	4	noted	note	VERB
ajst-21816	173	5	that	that	SCONJ
ajst-21816	173	6	in	in	ADP
ajst-21816	173	7	the	the	DET
ajst-21816	173	8	table	table	NOUN
ajst-21816	173	9	,	,	PUNCT
ajst-21816	173	10	0	0	NUM
ajst-21816	173	11	°	°	NUM
ajst-21816	173	12	,	,	PUNCT
ajst-21816	173	13	18	18	NUM
ajst-21816	173	14	°	°	NUM
ajst-21816	173	15	...	...	PUNCT
ajst-21816	173	16	the	the	DET
ajst-21816	173	17	data	datum	NOUN
ajst-21816	173	18	in	in	ADP
ajst-21816	173	19	the	the	DET
ajst-21816	173	20	162	162	NUM
ajst-21816	173	21	°	°	NUM
ajst-21816	173	22	and	and	CCONJ
ajst-21816	173	23	180	180	NUM
ajst-21816	173	24	°	°	NOUN
ajst-21816	173	25	columns	column	NOUN
ajst-21816	173	26	are	be	AUX
ajst-21816	173	27	the	the	DET
ajst-21816	173	28	average	average	ADJ
ajst-21816	173	29	recognition	recognition	NOUN
ajst-21816	173	30	rate	rate	NOUN
ajst-21816	173	31	excluding	exclude	VERB
ajst-21816	173	32	the	the	DET
ajst-21816	173	33	same	same	ADJ
ajst-21816	173	34	viewing	view	VERB
ajst-21816	173	35	angle	angle	NOUN
ajst-21816	173	36	.	.	PUNCT
ajst-21816	174	1	for	for	ADP
ajst-21816	174	2	example	example	NOUN
ajst-21816	174	3	,	,	PUNCT
ajst-21816	174	4	the	the	DET
ajst-21816	174	5	recognition	recognition	NOUN
ajst-21816	174	6	rate	rate	NOUN
ajst-21816	174	7	of	of	ADP
ajst-21816	174	8	the	the	DET
ajst-21816	174	9	viewing	view	VERB
ajst-21816	174	10	angle	angle	NOUN
ajst-21816	174	11	108	108	NUM
ajst-21816	174	12	°	°	NUM
ajst-21816	174	13	is	be	AUX
ajst-21816	174	14	the	the	DET
ajst-21816	174	15	average	average	ADJ
ajst-21816	174	16	recognition	recognition	NOUN
ajst-21816	174	17	rate	rate	NOUN
ajst-21816	174	18	of	of	ADP
ajst-21816	174	19	the	the	DET
ajst-21816	174	20	other	other	ADJ
ajst-21816	174	21	10	10	NUM
ajst-21816	174	22	viewing	view	VERB
ajst-21816	174	23	angles	angle	NOUN
ajst-21816	174	24	except	except	SCONJ
ajst-21816	174	25	108	108	NUM
ajst-21816	174	26	°	°	NOUN
ajst-21816	174	27	.	.	PUNCT
ajst-21816	175	1	the	the	DET
ajst-21816	175	2	last	last	ADJ
ajst-21816	175	3	column	column	NOUN
ajst-21816	175	4	shows	show	VERB
ajst-21816	175	5	the	the	DET
ajst-21816	175	6	average	average	ADJ
ajst-21816	175	7	recognition	recognition	NOUN
ajst-21816	175	8	rate	rate	NOUN
ajst-21816	175	9	for	for	ADP
ajst-21816	175	10	11	11	NUM
ajst-21816	175	11	different	different	ADJ
ajst-21816	175	12	viewing	viewing	NOUN
ajst-21816	175	13	angles	angle	NOUN
ajst-21816	175	14	.	.	PUNCT
ajst-21816	176	1	through	through	ADP
ajst-21816	176	2	observation	observation	NOUN
ajst-21816	176	3	,	,	PUNCT
ajst-21816	176	4	it	it	PRON
ajst-21816	176	5	is	be	AUX
ajst-21816	176	6	found	find	VERB
ajst-21816	176	7	that	that	SCONJ
ajst-21816	176	8	the	the	DET
ajst-21816	176	9	proposed	propose	VERB
ajst-21816	176	10	method	method	NOUN
ajst-21816	176	11	has	have	VERB
ajst-21816	176	12	no	no	DET
ajst-21816	176	13	obvious	obvious	ADJ
ajst-21816	176	14	advantage	advantage	NOUN
ajst-21816	176	15	compared	compare	VERB
ajst-21816	176	16	with	with	ADP
ajst-21816	176	17	other	other	ADJ
ajst-21816	176	18	methods	method	NOUN
ajst-21816	176	19	when	when	SCONJ
ajst-21816	176	20	the	the	DET
ajst-21816	176	21	pedestrian	pedestrian	NOUN
ajst-21816	176	22	is	be	AUX
ajst-21816	176	23	in	in	ADP
ajst-21816	176	24	nm	nm	PRON
ajst-21816	176	25	state	state	NOUN
ajst-21816	176	26	.	.	PUNCT
ajst-21816	177	1	the	the	DET
ajst-21816	177	2	reason	reason	NOUN
ajst-21816	177	3	may	may	AUX
ajst-21816	177	4	be	be	AUX
ajst-21816	177	5	that	that	SCONJ
ajst-21816	177	6	the	the	DET
ajst-21816	177	7	information	information	NOUN
ajst-21816	177	8	contained	contain	VERB
ajst-21816	177	9	in	in	ADP
ajst-21816	177	10	the	the	DET
ajst-21816	177	11	input	input	NOUN
ajst-21816	177	12	gait	gait	PROPN
ajst-21816	177	13	contour	contour	NOUN
ajst-21816	177	14	diagram	diagram	NOUN
ajst-21816	177	15	of	of	ADP
ajst-21816	177	16	the	the	DET
ajst-21816	177	17	proposed	propose	VERB
ajst-21816	177	18	method	method	NOUN
ajst-21816	177	19	is	be	AUX
ajst-21816	177	20	limited	limit	VERB
ajst-21816	177	21	,	,	PUNCT
ajst-21816	177	22	or	or	CCONJ
ajst-21816	177	23	the	the	DET
ajst-21816	177	24	proposed	propose	VERB
ajst-21816	177	25	method	method	NOUN
ajst-21816	177	26	does	do	AUX
ajst-21816	177	27	not	not	PART
ajst-21816	177	28	extract	extract	VERB
ajst-21816	177	29	enough	enough	ADJ
ajst-21816	177	30	useful	useful	ADJ
ajst-21816	177	31	information	information	NOUN
ajst-21816	177	32	.	.	PUNCT
ajst-21816	178	1	as	as	SCONJ
ajst-21816	178	2	shown	show	VERB
ajst-21816	178	3	in	in	ADP
ajst-21816	178	4	table	table	NOUN
ajst-21816	178	5	3	3	NUM
ajst-21816	178	6	-	-	SYM
ajst-21816	178	7	2	2	NUM
ajst-21816	178	8	,	,	PUNCT
ajst-21816	178	9	when	when	SCONJ
ajst-21816	178	10	pedestrians	pedestrian	NOUN
ajst-21816	178	11	are	be	AUX
ajst-21816	178	12	in	in	ADP
ajst-21816	178	13	the	the	DET
ajst-21816	178	14	bg	bg	PROPN
ajst-21816	178	15	state	state	NOUN
ajst-21816	178	16	,	,	PUNCT
ajst-21816	178	17	the	the	DET
ajst-21816	178	18	proposed	propose	VERB
ajst-21816	178	19	method	method	NOUN
ajst-21816	178	20	has	have	VERB
ajst-21816	178	21	obvious	obvious	ADJ
ajst-21816	178	22	advantages	advantage	NOUN
ajst-21816	178	23	compared	compare	VERB
ajst-21816	178	24	with	with	ADP
ajst-21816	178	25	gaitset	gaitset	NOUN
ajst-21816	178	26	,	,	PUNCT
ajst-21816	178	27	capsnet	capsnet	NOUN
ajst-21816	178	28	and	and	CCONJ
ajst-21816	178	29	gaitpart	gaitpart	NOUN
ajst-21816	178	30	,	,	PUNCT
ajst-21816	178	31	while	while	SCONJ
ajst-21816	178	32	the	the	DET
ajst-21816	178	33	performance	performance	NOUN
ajst-21816	178	34	is	be	AUX
ajst-21816	178	35	not	not	PART
ajst-21816	178	36	significantly	significantly	ADV
ajst-21816	178	37	improved	improve	VERB
ajst-21816	178	38	compared	compare	VERB
ajst-21816	178	39	with	with	ADP
ajst-21816	178	40	srn	srn	NOUN
ajst-21816	178	41	.	.	PUNCT
ajst-21816	179	1	according	accord	VERB
ajst-21816	179	2	to	to	ADP
ajst-21816	179	3	table	table	NOUN
ajst-21816	179	4	3	3	NUM
ajst-21816	179	5	-	-	SYM
ajst-21816	179	6	3	3	NUM
ajst-21816	179	7	,	,	PUNCT
ajst-21816	179	8	when	when	SCONJ
ajst-21816	179	9	pedestrians	pedestrian	NOUN
ajst-21816	179	10	are	be	AUX
ajst-21816	179	11	in	in	ADP
ajst-21816	179	12	cl	cl	NOUN
ajst-21816	179	13	state	state	NOUN
ajst-21816	179	14	,	,	PUNCT
ajst-21816	179	15	the	the	DET
ajst-21816	179	16	recognition	recognition	NOUN
ajst-21816	179	17	rate	rate	NOUN
ajst-21816	179	18	of	of	ADP
ajst-21816	179	19	the	the	DET
ajst-21816	179	20	proposed	propose	VERB
ajst-21816	179	21	method	method	NOUN
ajst-21816	179	22	reaches	reach	VERB
ajst-21816	179	23	81.0	81.0	NUM
ajst-21816	179	24	%	%	NOUN
ajst-21816	179	25	.	.	PUNCT
ajst-21816	180	1	compared	compare	VERB
ajst-21816	180	2	with	with	ADP
ajst-21816	180	3	srn	srn	NOUN
ajst-21816	180	4	,	,	PUNCT
ajst-21816	180	5	gaitset	gaitset	NOUN
ajst-21816	180	6	,	,	PUNCT
ajst-21816	180	7	capsnet	capsnet	NOUN
ajst-21816	180	8	and	and	CCONJ
ajst-21816	180	9	gaitpart	gaitpart	NOUN
ajst-21816	180	10	,	,	PUNCT
ajst-21816	180	11	the	the	DET
ajst-21816	180	12	average	average	ADJ
ajst-21816	180	13	rank-1	rank-1	NUM
ajst-21816	180	14	recognition	recognition	NOUN
ajst-21816	180	15	rate	rate	NOUN
ajst-21816	180	16	has	have	AUX
ajst-21816	180	17	been	be	AUX
ajst-21816	180	18	significantly	significantly	ADV
ajst-21816	180	19	improved	improve	VERB
ajst-21816	180	20	by	by	ADP
ajst-21816	180	21	3.3	3.3	NUM
ajst-21816	180	22	,	,	PUNCT
ajst-21816	180	23	10.7	10.7	NUM
ajst-21816	180	24	,	,	PUNCT
ajst-21816	180	25	8.6	8.6	NUM
ajst-21816	180	26	and	and	CCONJ
ajst-21816	180	27	2.3	2.3	NUM
ajst-21816	180	28	percentage	percentage	NOUN
ajst-21816	180	29	points	point	NOUN
ajst-21816	180	30	respectively	respectively	ADV
ajst-21816	180	31	.	.	PUNCT
ajst-21816	181	1	except	except	SCONJ
ajst-21816	181	2	that	that	SCONJ
ajst-21816	181	3	the	the	DET
ajst-21816	181	4	average	average	ADJ
ajst-21816	181	5	recognition	recognition	NOUN
ajst-21816	181	6	rate	rate	NOUN
ajst-21816	181	7	at	at	ADP
ajst-21816	181	8	108	108	NUM
ajst-21816	181	9	°	°	NUM
ajst-21816	181	10	and	and	CCONJ
ajst-21816	181	11	126	126	NUM
ajst-21816	181	12	°	°	NOUN
ajst-21816	181	13	is	be	AUX
ajst-21816	181	14	slightly	slightly	ADV
ajst-21816	181	15	lower	low	ADJ
ajst-21816	181	16	than	than	ADP
ajst-21816	181	17	srn	srn	NOUN
ajst-21816	181	18	,	,	PUNCT
ajst-21816	181	19	the	the	DET
ajst-21816	181	20	average	average	ADJ
ajst-21816	181	21	recognition	recognition	NOUN
ajst-21816	181	22	rate	rate	NOUN
ajst-21816	181	23	at	at	ADP
ajst-21816	181	24	the	the	DET
ajst-21816	181	25	other	other	ADJ
ajst-21816	181	26	nine	nine	NUM
ajst-21816	181	27	angles	angle	NOUN
ajst-21816	181	28	is	be	AUX
ajst-21816	181	29	higher	high	ADJ
ajst-21816	181	30	than	than	ADP
ajst-21816	181	31	the	the	DET
ajst-21816	181	32	other	other	ADJ
ajst-21816	181	33	four	four	NUM
ajst-21816	181	34	methods	method	NOUN
ajst-21816	181	35	.	.	PUNCT
ajst-21816	182	1	table	table	NOUN
ajst-21816	182	2	1	1	NUM
ajst-21816	182	3	.	.	PUNCT
ajst-21816	183	1	the	the	DET
ajst-21816	183	2	average	average	ADJ
ajst-21816	183	3	rank-1	rank-1	NUM
ajst-21816	183	4	accuracy	accuracy	NOUN
ajst-21816	183	5	of	of	ADP
ajst-21816	183	6	different	different	ADJ
ajst-21816	183	7	algorithms	algorithm	NOUN
ajst-21816	183	8	on	on	ADP
ajst-21816	183	9	the	the	DET
ajst-21816	183	10	casia	casia	NOUN
ajst-21816	183	11	-	-	PUNCT
ajst-21816	183	12	b	b	NOUN
ajst-21816	183	13	data	datum	NOUN
ajst-21816	183	14	set	set	VERB
ajst-21816	183	15	at	at	ADP
ajst-21816	183	16	nm	nm	DET
ajst-21816	183	17	method	method	NOUN
ajst-21816	183	18	recognition	recognition	NOUN
ajst-21816	183	19	rate	rate	NOUN
ajst-21816	183	20	under	under	ADP
ajst-21816	183	21	different	different	ADJ
ajst-21816	183	22	perspectives	perspective	NOUN
ajst-21816	183	23	/（%	/（%	PUNCT
ajst-21816	183	24	）	）	PUNCT
ajst-21816	183	25	0	0	NUM
ajst-21816	183	26	°	°	NUM
ajst-21816	183	27	18	18	NUM
ajst-21816	183	28	°	°	NOUN
ajst-21816	183	29	36	36	NUM
ajst-21816	183	30	°	°	NOUN
ajst-21816	183	31	54	54	NUM
ajst-21816	183	32	°	°	NOUN
ajst-21816	183	33	72	72	NUM
ajst-21816	183	34	°	°	NOUN
ajst-21816	183	35	90	90	NUM
ajst-21816	183	36	°	°	NOUN
ajst-21816	183	37	108	108	NUM
ajst-21816	183	38	°	°	NOUN
ajst-21816	183	39	126	126	NUM
ajst-21816	183	40	°	°	NOUN
ajst-21816	183	41	144	144	NUM
ajst-21816	183	42	°	°	NOUN
ajst-21816	183	43	162	162	NUM
ajst-21816	183	44	°	°	NOUN
ajst-21816	183	45	180	180	NUM
ajst-21816	183	46	°	°	PRON
ajst-21816	183	47	average	average	ADJ
ajst-21816	183	48	srn	srn	NOUN
ajst-21816	183	49	94.4	94.4	NUM
ajst-21816	183	50	99.3	99.3	NUM
ajst-21816	183	51	99.4	99.4	NUM
ajst-21816	183	52	98.7	98.7	NUM
ajst-21816	183	53	96.8	96.8	NUM
ajst-21816	183	54	96.8	96.8	NUM
ajst-21816	183	55	97.5	97.5	NUM
ajst-21816	183	56	98.5	98.5	NUM
ajst-21816	183	57	99.5	99.5	NUM
ajst-21816	183	58	98.8	98.8	NUM
ajst-21816	183	59	92.3	92.3	NUM
ajst-21816	183	60	97.45	97.45	NUM
ajst-21816	183	61	gaitset	gaitset	NOUN
ajst-21816	183	62	91.1	91.1	NUM
ajst-21816	183	63	99.0	99.0	NUM
ajst-21816	183	64	99.9	99.9	NUM
ajst-21816	183	65	97.8	97.8	NUM
ajst-21816	183	66	95.1	95.1	NUM
ajst-21816	183	67	94.5	94.5	NUM
ajst-21816	183	68	96.1	96.1	NUM
ajst-21816	183	69	98.3	98.3	NUM
ajst-21816	183	70	99.2	99.2	NUM
ajst-21816	183	71	98.1	98.1	NUM
ajst-21816	183	72	88.0	88.0	NUM
ajst-21816	183	73	96.10	96.10	NUM
ajst-21816	183	74	capsnet	capsnet	NOUN
ajst-21816	183	75	91.8	91.8	NUM
ajst-21816	183	76	98.3	98.3	NUM
ajst-21816	183	77	99.0	99.0	NUM
ajst-21816	183	78	98.0	98.0	NUM
ajst-21816	183	79	94.1	94.1	NUM
ajst-21816	183	80	92.8	92.8	NUM
ajst-21816	183	81	96.3	96.3	NUM
ajst-21816	183	82	98.1	98.1	NUM
ajst-21816	183	83	98.4	98.4	NUM
ajst-21816	183	84	96.2	96.2	NUM
ajst-21816	183	85	89.2	89.2	NUM
ajst-21816	183	86	95.65	95.65	NUM
ajst-21816	183	87	gaitpart	gaitpart	NOUN
ajst-21816	183	88	94.1	94.1	NUM
ajst-21816	183	89	98.6	98.6	NUM
ajst-21816	183	90	99.3	99.3	NUM
ajst-21816	183	91	98.5	98.5	NUM
ajst-21816	183	92	94.0	94.0	NUM
ajst-21816	183	93	92.3	92.3	NUM
ajst-21816	183	94	95.9	95.9	NUM
ajst-21816	183	95	98.4	98.4	NUM
ajst-21816	183	96	99.2	99.2	NUM
ajst-21816	183	97	97.8	97.8	NUM
ajst-21816	183	98	90.4	90.4	NUM
ajst-21816	183	99	96.22	96.22	NUM
ajst-21816	183	100	ours	our	NOUN
ajst-21816	183	101	94.8	94.8	NUM
ajst-21816	183	102	99.2	99.2	NUM
ajst-21816	183	103	99.8	99.8	NUM
ajst-21816	183	104	98.8	98.8	NUM
ajst-21816	183	105	96.8	96.8	NUM
ajst-21816	183	106	96.9	96.9	NUM
ajst-21816	183	107	97.1	97.1	NUM
ajst-21816	183	108	98.6	98.6	NUM
ajst-21816	183	109	99.4	99.4	NUM
ajst-21816	183	110	98.7	98.7	NUM
ajst-21816	183	111	92.7	92.7	NUM
ajst-21816	183	112	97.53	97.53	NUM
ajst-21816	183	113	table	table	NOUN
ajst-21816	183	114	2	2	NUM
ajst-21816	183	115	.	.	PUNCT
ajst-21816	184	1	the	the	DET
ajst-21816	184	2	average	average	ADJ
ajst-21816	184	3	rank-1	rank-1	NUM
ajst-21816	184	4	accuracy	accuracy	NOUN
ajst-21816	184	5	of	of	ADP
ajst-21816	184	6	different	different	ADJ
ajst-21816	184	7	algorithms	algorithm	NOUN
ajst-21816	184	8	on	on	ADP
ajst-21816	184	9	the	the	DET
ajst-21816	184	10	casia	casia	NOUN
ajst-21816	184	11	-	-	PUNCT
ajst-21816	184	12	b	b	NOUN
ajst-21816	184	13	data	datum	NOUN
ajst-21816	184	14	set	set	VERB
ajst-21816	184	15	at	at	ADP
ajst-21816	184	16	bg	bg	PROPN
ajst-21816	184	17	method	method	NOUN
ajst-21816	184	18	recognition	recognition	NOUN
ajst-21816	184	19	rate	rate	NOUN
ajst-21816	184	20	under	under	ADP
ajst-21816	184	21	different	different	ADJ
ajst-21816	184	22	perspectives	perspective	NOUN
ajst-21816	184	23	/（%	/（%	PUNCT
ajst-21816	184	24	）	）	PUNCT
ajst-21816	184	25	0	0	NUM
ajst-21816	185	1	°	°	NUM
ajst-21816	185	2	18	18	NUM
ajst-21816	185	3	°	°	NOUN
ajst-21816	185	4	36	36	NUM
ajst-21816	185	5	°	°	NOUN
ajst-21816	185	6	54	54	NUM
ajst-21816	185	7	°	°	NOUN
ajst-21816	185	8	72	72	NUM
ajst-21816	185	9	°	°	NOUN
ajst-21816	185	10	90	90	NUM
ajst-21816	185	11	°	°	NOUN
ajst-21816	185	12	108	108	NUM
ajst-21816	185	13	°	°	NOUN
ajst-21816	185	14	126	126	NUM
ajst-21816	185	15	°	°	NOUN
ajst-21816	185	16	144	144	NUM
ajst-21816	185	17	°	°	NOUN
ajst-21816	185	18	162	162	NUM
ajst-21816	185	19	°	°	NOUN
ajst-21816	185	20	180	180	NUM
ajst-21816	185	21	°	°	PRON
ajst-21816	185	22	average	average	ADJ
ajst-21816	185	23	srn	srn	NOUN
ajst-21816	185	24	91.5	91.5	NUM
ajst-21816	185	25	97.4	97.4	NUM
ajst-21816	185	26	98.4	98.4	NUM
ajst-21816	185	27	97.1	97.1	NUM
ajst-21816	185	28	92.2	92.2	NUM
ajst-21816	185	29	89.7	89.7	NUM
ajst-21816	185	30	93.1	93.1	NUM
ajst-21816	185	31	96.2	96.2	NUM
ajst-21816	185	32	97.5	97.5	NUM
ajst-21816	185	33	96.5	96.5	NUM
ajst-21816	185	34	88.0	88.0	NUM
ajst-21816	185	35	94.32	94.32	NUM
ajst-21816	185	36	gaitset	gaitset	NOUN
ajst-21816	185	37	86.7	86.7	NUM
ajst-21816	185	38	94.2	94.2	NUM
ajst-21816	185	39	95.7	95.7	NUM
ajst-21816	185	40	93.4	93.4	NUM
ajst-21816	185	41	88.9	88.9	NUM
ajst-21816	185	42	85.5	85.5	NUM
ajst-21816	185	43	89.0	89.0	NUM
ajst-21816	185	44	91.7	91.7	NUM
ajst-21816	185	45	94.5	94.5	NUM
ajst-21816	185	46	95.9	95.9	NUM
ajst-21816	185	47	83.3	83.3	NUM
ajst-21816	185	48	90.80	90.80	NUM
ajst-21816	185	49	capsnet	capsnet	NOUN
ajst-21816	185	50	87.3	87.3	NUM
ajst-21816	185	51	93.7	93.7	NUM
ajst-21816	185	52	94.8	94.8	NUM
ajst-21816	185	53	93.1	93.1	NUM
ajst-21816	185	54	88.1	88.1	NUM
ajst-21816	185	55	84.5	84.5	NUM
ajst-21816	185	56	88.8	88.8	NUM
ajst-21816	185	57	93.5	93.5	NUM
ajst-21816	185	58	96.3	96.3	NUM
ajst-21816	185	59	93.3	93.3	NUM
ajst-21816	185	60	83.9	83.9	NUM
ajst-21816	185	61	90.66	90.66	NUM
ajst-21816	185	62	gaitpart	gaitpart	NOUN
ajst-21816	185	63	89.1	89.1	NUM
ajst-21816	185	64	94.8	94.8	NUM
ajst-21816	185	65	96.7	96.7	NUM
ajst-21816	185	66	95.1	95.1	NUM
ajst-21816	185	67	88.3	88.3	NUM
ajst-21816	185	68	94.9	94.9	NUM
ajst-21816	185	69	89.0	89.0	NUM
ajst-21816	185	70	93.5	93.5	NUM
ajst-21816	185	71	96.1	96.1	NUM
ajst-21816	185	72	93.8	93.8	NUM
ajst-21816	185	73	85.8	85.8	NUM
ajst-21816	185	74	92.46	92.46	NUM
ajst-21816	185	75	ours	ours	PRON
ajst-21816	185	76	92.9	92.9	NUM
ajst-21816	185	77	98.5	98.5	NUM
ajst-21816	185	78	99.7	99.7	NUM
ajst-21816	185	79	97.0	97.0	NUM
ajst-21816	185	80	91.9	91.9	NUM
ajst-21816	185	81	86.1	86.1	NUM
ajst-21816	185	82	90.0	90.0	NUM
ajst-21816	185	83	96.6	96.6	NUM
ajst-21816	185	84	99.7	99.7	NUM
ajst-21816	185	85	98.7	98.7	NUM
ajst-21816	185	86	88.8	88.8	NUM
ajst-21816	185	87	94.54	94.54	NUM
ajst-21816	185	88	293	293	NUM
ajst-21816	185	89	table	table	NOUN
ajst-21816	185	90	3	3	NUM
ajst-21816	185	91	.	.	PUNCT
ajst-21816	186	1	the	the	DET
ajst-21816	186	2	average	average	ADJ
ajst-21816	186	3	rank-1	rank-1	NUM
ajst-21816	186	4	accuracy	accuracy	NOUN
ajst-21816	186	5	of	of	ADP
ajst-21816	186	6	different	different	ADJ
ajst-21816	186	7	algorithms	algorithm	NOUN
ajst-21816	186	8	on	on	ADP
ajst-21816	186	9	the	the	DET
ajst-21816	186	10	casia	casia	NOUN
ajst-21816	186	11	-	-	PUNCT
ajst-21816	186	12	b	b	NOUN
ajst-21816	186	13	data	datum	NOUN
ajst-21816	186	14	set	set	VERB
ajst-21816	186	15	at	at	ADP
ajst-21816	186	16	cl	cl	NOUN
ajst-21816	186	17	method	method	NOUN
ajst-21816	186	18	recognition	recognition	NOUN
ajst-21816	186	19	rate	rate	NOUN
ajst-21816	186	20	under	under	ADP
ajst-21816	186	21	different	different	ADJ
ajst-21816	186	22	perspectives	perspective	NOUN
ajst-21816	186	23	/（%	/（%	PUNCT
ajst-21816	186	24	）	）	PUNCT
ajst-21816	186	25	0	0	NUM
ajst-21816	186	26	°	°	NUM
ajst-21816	186	27	18	18	NUM
ajst-21816	186	28	°	°	NOUN
ajst-21816	186	29	36	36	NUM
ajst-21816	186	30	°	°	NOUN
ajst-21816	186	31	54	54	NUM
ajst-21816	186	32	°	°	NOUN
ajst-21816	186	33	72	72	NUM
ajst-21816	186	34	°	°	NOUN
ajst-21816	186	35	90	90	NUM
ajst-21816	186	36	°	°	NOUN
ajst-21816	186	37	108	108	NUM
ajst-21816	186	38	°	°	NOUN
ajst-21816	186	39	126	126	NUM
ajst-21816	186	40	°	°	NOUN
ajst-21816	186	41	144	144	NUM
ajst-21816	186	42	°	°	NOUN
ajst-21816	186	43	162	162	NUM
ajst-21816	186	44	°	°	NOUN
ajst-21816	186	45	180	180	NUM
ajst-21816	186	46	°	°	PRON
ajst-21816	186	47	average	average	ADJ
ajst-21816	186	48	srn	srn	NOUN
ajst-21816	186	49	69.2	69.2	NUM
ajst-21816	186	50	82.5	82.5	NUM
ajst-21816	186	51	84.0	84.0	NUM
ajst-21816	186	52	81.0	81.0	NUM
ajst-21816	186	53	78.6	78.6	NUM
ajst-21816	186	54	76.3	76.3	NUM
ajst-21816	186	55	78.6	78.6	NUM
ajst-21816	186	56	82.8	82.8	NUM
ajst-21816	186	57	80.5	80.5	NUM
ajst-21816	186	58	76.8	76.8	NUM
ajst-21816	186	59	64.7	64.7	NUM
ajst-21816	186	60	77.72	77.72	NUM
ajst-21816	186	61	gaitset	gaitset	VERB
ajst-21816	186	62	59.5	59.5	NUM
ajst-21816	186	63	75.0	75.0	NUM
ajst-21816	186	64	78.3	78.3	NUM
ajst-21816	186	65	74.6	74.6	NUM
ajst-21816	186	66	71.4	71.4	NUM
ajst-21816	186	67	71.3	71.3	NUM
ajst-21816	186	68	70.8	70.8	NUM
ajst-21816	186	69	74.1	74.1	NUM
ajst-21816	186	70	74.6	74.6	NUM
ajst-21816	186	71	69.4	69.4	NUM
ajst-21816	186	72	54.1	54.1	NUM
ajst-21816	186	73	70.28	70.28	NUM
ajst-21816	186	74	capsnet	capsnet	NOUN
ajst-21816	186	75	63.4	63.4	NUM
ajst-21816	186	76	77.3	77.3	NUM
ajst-21816	186	77	80.1	80.1	NUM
ajst-21816	186	78	79.4	79.4	NUM
ajst-21816	186	79	72.4	72.4	NUM
ajst-21816	186	80	69.8	69.8	NUM
ajst-21816	186	81	71.2	71.2	NUM
ajst-21816	186	82	73.8	73.8	NUM
ajst-21816	186	83	75.5	75.5	NUM
ajst-21816	186	84	71.7	71.7	NUM
ajst-21816	186	85	62.0	62.0	NUM
ajst-21816	186	86	72.41	72.41	NUM
ajst-21816	186	87	gaitpart	gaitpart	NOUN
ajst-21816	186	88	70.7	70.7	NUM
ajst-21816	186	89	85.5	85.5	NUM
ajst-21816	186	90	86.9	86.9	NUM
ajst-21816	186	91	83.3	83.3	NUM
ajst-21816	186	92	77.1	77.1	NUM
ajst-21816	186	93	72.5	72.5	NUM
ajst-21816	186	94	76.9	76.9	NUM
ajst-21816	186	95	82.2	82.2	NUM
ajst-21816	186	96	83.8	83.8	NUM
ajst-21816	186	97	80.2	80.2	NUM
ajst-21816	186	98	66.5	66.5	NUM
ajst-21816	186	99	78.69	78.69	NUM
ajst-21816	186	100	ours	ours	PRON
ajst-21816	187	1	74.8	74.8	NUM
ajst-21816	187	2	88.5	88.5	NUM
ajst-21816	187	3	90.8	90.8	NUM
ajst-21816	187	4	84.9	84.9	NUM
ajst-21816	187	5	79.1	79.1	NUM
ajst-21816	187	6	77.6	77.6	NUM
ajst-21816	187	7	78.5	78.5	NUM
ajst-21816	187	8	82.4	82.4	NUM
ajst-21816	187	9	84.9	84.9	NUM
ajst-21816	187	10	80.9	80.9	NUM
ajst-21816	187	11	69.3	69.3	NUM
ajst-21816	187	12	81.06	81.06	NUM
ajst-21816	187	13	by	by	ADP
ajst-21816	187	14	comprehensively	comprehensively	ADV
ajst-21816	187	15	observing	observe	VERB
ajst-21816	187	16	these	these	DET
ajst-21816	187	17	three	three	NUM
ajst-21816	187	18	tables	table	NOUN
ajst-21816	188	1	,	,	PUNCT
ajst-21816	188	2	it	it	PRON
ajst-21816	188	3	is	be	AUX
ajst-21816	188	4	not	not	PART
ajst-21816	188	5	difficult	difficult	ADJ
ajst-21816	188	6	to	to	PART
ajst-21816	188	7	find	find	VERB
ajst-21816	188	8	that	that	SCONJ
ajst-21816	188	9	the	the	DET
ajst-21816	188	10	proposed	propose	VERB
ajst-21816	188	11	method	method	NOUN
ajst-21816	188	12	has	have	VERB
ajst-21816	188	13	better	well	ADJ
ajst-21816	188	14	robustness	robustness	NOUN
ajst-21816	188	15	compared	compare	VERB
ajst-21816	188	16	with	with	ADP
ajst-21816	188	17	srn	srn	NOUN
ajst-21816	188	18	,	,	PUNCT
ajst-21816	188	19	gaitset	gaitset	NOUN
ajst-21816	188	20	,	,	PUNCT
ajst-21816	188	21	capsnet	capsnet	NOUN
ajst-21816	188	22	and	and	CCONJ
ajst-21816	188	23	gaitpart	gaitpart	NOUN
ajst-21816	188	24	when	when	SCONJ
ajst-21816	188	25	the	the	DET
ajst-21816	188	26	walking	walk	VERB
ajst-21816	188	27	state	state	NOUN
ajst-21816	188	28	of	of	ADP
ajst-21816	188	29	pedestrians	pedestrian	NOUN
ajst-21816	188	30	changes	change	NOUN
ajst-21816	188	31	.	.	PUNCT
ajst-21816	189	1	especially	especially	ADV
ajst-21816	189	2	in	in	ADP
ajst-21816	189	3	the	the	DET
ajst-21816	189	4	cl	cl	NOUN
ajst-21816	189	5	state	state	NOUN
ajst-21816	189	6	,	,	PUNCT
ajst-21816	189	7	the	the	DET
ajst-21816	189	8	method	method	NOUN
ajst-21816	189	9	presented	present	VERB
ajst-21816	189	10	in	in	ADP
ajst-21816	189	11	this	this	DET
ajst-21816	189	12	paper	paper	NOUN
ajst-21816	189	13	shows	show	VERB
ajst-21816	189	14	greater	great	ADJ
ajst-21816	189	15	advantages	advantage	NOUN
ajst-21816	189	16	.	.	PUNCT
ajst-21816	190	1	3.4	3.4	NUM
ajst-21816	190	2	.	.	PUNCT
ajst-21816	190	3	ablation	ablation	NOUN
ajst-21816	190	4	experiment	experiment	NOUN
ajst-21816	190	5	3.4.1	3.4.1	NUM
ajst-21816	190	6	.	.	PUNCT
ajst-21816	191	1	local	local	ADJ
ajst-21816	191	2	feature	feature	NOUN
ajst-21816	191	3	extractor	extractor	NOUN
ajst-21816	191	4	(	(	PUNCT
ajst-21816	191	5	lfe	lfe	PROPN
ajst-21816	191	6	)	)	PUNCT
ajst-21816	191	7	ablation	ablation	NOUN
ajst-21816	191	8	experiment	experiment	NOUN
ajst-21816	191	9	in	in	ADP
ajst-21816	191	10	order	order	NOUN
ajst-21816	191	11	to	to	PART
ajst-21816	191	12	verify	verify	VERB
ajst-21816	191	13	the	the	DET
ajst-21816	191	14	advantages	advantage	NOUN
ajst-21816	191	15	of	of	ADP
ajst-21816	191	16	sconv	sconv	NOUN
ajst-21816	191	17	over	over	ADP
ajst-21816	191	18	fconv	fconv	NOUN
ajst-21816	191	19	and	and	CCONJ
ajst-21816	191	20	find	find	VERB
ajst-21816	191	21	out	out	ADP
ajst-21816	191	22	the	the	DET
ajst-21816	191	23	function	function	NOUN
ajst-21816	191	24	most	most	ADV
ajst-21816	191	25	suitable	suitable	ADJ
ajst-21816	191	26	for	for	ADP
ajst-21816	191	27	sp	sp	NOUN
ajst-21816	191	28	,	,	PUNCT
ajst-21816	191	29	ablation	ablation	NOUN
ajst-21816	191	30	experiments	experiment	NOUN
ajst-21816	191	31	were	be	AUX
ajst-21816	191	32	conducted	conduct	VERB
ajst-21816	191	33	on	on	ADP
ajst-21816	191	34	the	the	DET
ajst-21816	191	35	casia	casia	NOUN
ajst-21816	191	36	-	-	PUNCT
ajst-21816	191	37	b	b	NOUN
ajst-21816	191	38	dataset	dataset	NOUN
ajst-21816	191	39	,	,	PUNCT
ajst-21816	191	40	and	and	CCONJ
ajst-21816	191	41	the	the	DET
ajst-21816	191	42	experimental	experimental	ADJ
ajst-21816	191	43	results	result	NOUN
ajst-21816	191	44	were	be	AUX
ajst-21816	191	45	shown	show	VERB
ajst-21816	191	46	in	in	ADP
ajst-21816	191	47	table	table	NOUN
ajst-21816	191	48	4	4	NUM
ajst-21816	191	49	.	.	PUNCT
ajst-21816	191	50	√	√	PROPN
ajst-21816	191	51	means	mean	VERB
ajst-21816	191	52	use	use	VERB
ajst-21816	191	53	the	the	DET
ajst-21816	191	54	current	current	ADJ
ajst-21816	191	55	convolution	convolution	NOUN
ajst-21816	191	56	or	or	CCONJ
ajst-21816	191	57	function	function	NOUN
ajst-21816	191	58	.	.	PUNCT
ajst-21816	192	1	conv1	conv1	PROPN
ajst-21816	192	2	in	in	ADP
ajst-21816	192	3	conv	conv	NOUN
ajst-21816	192	4	represents	represent	VERB
ajst-21816	192	5	ordinary	ordinary	ADJ
ajst-21816	192	6	convolution	convolution	NOUN
ajst-21816	192	7	;	;	PUNCT
ajst-21816	192	8	fconv	fconv	INTJ
ajst-21816	192	9	stands	stand	VERB
ajst-21816	192	10	for	for	ADP
ajst-21816	192	11	focal	focal	ADJ
ajst-21816	192	12	convolution	convolution	NOUN
ajst-21816	192	13	in	in	ADP
ajst-21816	192	14	gaitpart	gaitpart	NOUN
ajst-21816	192	15	;	;	PUNCT
ajst-21816	192	16	sconv	sconv	NOUN
ajst-21816	192	17	represents	represent	VERB
ajst-21816	192	18	a	a	DET
ajst-21816	192	19	specific	specific	ADJ
ajst-21816	192	20	proportional	proportional	ADJ
ajst-21816	192	21	partition	partition	NOUN
ajst-21816	192	22	convolution	convolution	NOUN
ajst-21816	192	23	proposed	propose	VERB
ajst-21816	192	24	in	in	ADP
ajst-21816	192	25	this	this	DET
ajst-21816	192	26	paper	paper	NOUN
ajst-21816	192	27	.	.	PUNCT
ajst-21816	193	1	max	max	PROPN
ajst-21816	193	2	,	,	PUNCT
ajst-21816	193	3	median	median	NOUN
ajst-21816	193	4	,	,	PUNCT
ajst-21816	193	5	and	and	CCONJ
ajst-21816	193	6	union	union	NOUN
ajst-21816	193	7	in	in	ADP
ajst-21816	193	8	set	set	ADJ
ajst-21816	193	9	poling	poling	NOUN
ajst-21816	193	10	represent	represent	VERB
ajst-21816	193	11	the	the	DET
ajst-21816	193	12	max	max	PROPN
ajst-21816	193	13	function	function	NOUN
ajst-21816	193	14	,	,	PUNCT
ajst-21816	193	15	median	median	ADJ
ajst-21816	193	16	function	function	NOUN
ajst-21816	193	17	,	,	PUNCT
ajst-21816	193	18	and	and	CCONJ
ajst-21816	193	19	union	union	NOUN
ajst-21816	193	20	function	function	NOUN
ajst-21816	193	21	respectively	respectively	ADV
ajst-21816	193	22	.	.	PUNCT
ajst-21816	194	1	by	by	ADP
ajst-21816	194	2	comparing	compare	VERB
ajst-21816	194	3	the	the	DET
ajst-21816	194	4	results	result	NOUN
ajst-21816	194	5	of	of	ADP
ajst-21816	194	6	the	the	DET
ajst-21816	194	7	first	first	ADJ
ajst-21816	194	8	three	three	NUM
ajst-21816	194	9	groups	group	NOUN
ajst-21816	194	10	of	of	ADP
ajst-21816	194	11	experiments	experiment	NOUN
ajst-21816	194	12	,	,	PUNCT
ajst-21816	194	13	it	it	PRON
ajst-21816	194	14	is	be	AUX
ajst-21816	194	15	found	find	VERB
ajst-21816	194	16	that	that	SCONJ
ajst-21816	194	17	the	the	DET
ajst-21816	194	18	sconv	sconv	NOUN
ajst-21816	194	19	proposed	propose	VERB
ajst-21816	194	20	in	in	ADP
ajst-21816	194	21	this	this	DET
ajst-21816	194	22	paper	paper	NOUN
ajst-21816	194	23	can	can	AUX
ajst-21816	194	24	extract	extract	VERB
ajst-21816	194	25	more	more	ADV
ajst-21816	194	26	useful	useful	ADJ
ajst-21816	194	27	gait	gait	NOUN
ajst-21816	194	28	features	feature	NOUN
ajst-21816	194	29	and	and	CCONJ
ajst-21816	194	30	obtain	obtain	VERB
ajst-21816	194	31	better	well	ADJ
ajst-21816	194	32	experimental	experimental	ADJ
ajst-21816	194	33	results	result	NOUN
ajst-21816	194	34	,	,	PUNCT
ajst-21816	194	35	indicating	indicate	VERB
ajst-21816	194	36	the	the	DET
ajst-21816	194	37	effectiveness	effectiveness	NOUN
ajst-21816	194	38	of	of	ADP
ajst-21816	194	39	sconv	sconv	NOUN
ajst-21816	194	40	.	.	PUNCT
ajst-21816	195	1	after	after	ADP
ajst-21816	195	2	comparing	compare	VERB
ajst-21816	195	3	the	the	DET
ajst-21816	195	4	results	result	NOUN
ajst-21816	195	5	of	of	ADP
ajst-21816	195	6	the	the	DET
ajst-21816	195	7	first	first	ADJ
ajst-21816	195	8	set	set	NOUN
ajst-21816	195	9	of	of	ADP
ajst-21816	195	10	experiments	experiment	NOUN
ajst-21816	195	11	with	with	ADP
ajst-21816	195	12	the	the	DET
ajst-21816	195	13	latter	latter	ADJ
ajst-21816	195	14	two	two	NUM
ajst-21816	195	15	sets	set	NOUN
ajst-21816	195	16	,	,	PUNCT
ajst-21816	195	17	it	it	PRON
ajst-21816	195	18	is	be	AUX
ajst-21816	195	19	found	find	VERB
ajst-21816	195	20	that	that	SCONJ
ajst-21816	195	21	neither	neither	CCONJ
ajst-21816	195	22	the	the	DET
ajst-21816	195	23	median	median	ADJ
ajst-21816	195	24	function	function	NOUN
ajst-21816	195	25	nor	nor	CCONJ
ajst-21816	195	26	the	the	DET
ajst-21816	195	27	joint	joint	ADJ
ajst-21816	195	28	function	function	NOUN
ajst-21816	195	29	composed	compose	VERB
ajst-21816	195	30	of	of	ADP
ajst-21816	195	31	1×1	1×1	NUM
ajst-21816	195	32	convolution	convolution	NOUN
ajst-21816	195	33	has	have	VERB
ajst-21816	195	34	a	a	DET
ajst-21816	195	35	better	well	ADJ
ajst-21816	195	36	effect	effect	NOUN
ajst-21816	195	37	than	than	ADP
ajst-21816	195	38	the	the	DET
ajst-21816	195	39	max	max	PROPN
ajst-21816	195	40	function	function	NOUN
ajst-21816	195	41	.	.	PUNCT
ajst-21816	196	1	therefore	therefore	ADV
ajst-21816	196	2	,	,	PUNCT
ajst-21816	196	3	max	max	PROPN
ajst-21816	196	4	function	function	NOUN
ajst-21816	196	5	is	be	AUX
ajst-21816	196	6	finally	finally	ADV
ajst-21816	196	7	selected	select	VERB
ajst-21816	196	8	as	as	ADP
ajst-21816	196	9	the	the	DET
ajst-21816	196	10	example	example	NOUN
ajst-21816	196	11	of	of	ADP
ajst-21816	196	12	set	set	ADJ
ajst-21816	196	13	pooling	pooling	NOUN
ajst-21816	196	14	.	.	PUNCT
ajst-21816	197	1	table	table	NOUN
ajst-21816	197	2	4	4	NUM
ajst-21816	197	3	.	.	PUNCT
ajst-21816	197	4	ablation	ablation	NOUN
ajst-21816	197	5	experiments	experiment	NOUN
ajst-21816	197	6	of	of	ADP
ajst-21816	197	7	lfe	lfe	PROPN
ajst-21816	197	8	on	on	ADP
ajst-21816	197	9	casia	casia	PROPN
ajst-21816	197	10	-	-	PUNCT
ajst-21816	197	11	b	b	NOUN
ajst-21816	197	12	dataset	dataset	VERB
ajst-21816	197	13	conv	conv	NOUN
ajst-21816	197	14	set	set	NOUN
ajst-21816	197	15	pooling	pool	VERB
ajst-21816	197	16	nm	nm	ADJ
ajst-21816	197	17	bg	bg	PROPN
ajst-21816	197	18	cl	cl	NOUN
ajst-21816	197	19	conv1	conv1	PROPN
ajst-21816	197	20	fconv	fconv	PROPN
ajst-21816	197	21	sconv	sconv	PROPN
ajst-21816	197	22	max	max	PROPN
ajst-21816	197	23	median	median	PROPN
ajst-21816	197	24	union	union	PROPN
ajst-21816	197	25	√	√	VERB
ajst-21816	198	1	√	√	NUM
ajst-21816	198	2	97.5	97.5	NUM
ajst-21816	198	3	94.5	94.5	NUM
ajst-21816	198	4	81.0	81.0	NUM
ajst-21816	198	5	√	√	NUM
ajst-21816	199	1	√	√	ADP
ajst-21816	200	1	96.1	96.1	NUM
ajst-21816	200	2	90.9	90.9	NUM
ajst-21816	200	3	77.3	77.3	NUM
ajst-21816	200	4	√	√	NUM
ajst-21816	201	1	√	√	ADP
ajst-21816	201	2	96.0	96.0	NUM
ajst-21816	201	3	89.8	89.8	NUM
ajst-21816	201	4	70.8	70.8	NUM
ajst-21816	201	5	√	√	NUM
ajst-21816	202	1	√	√	ADP
ajst-21816	203	1	96.9	96.9	NUM
ajst-21816	203	2	93.7	93.7	NUM
ajst-21816	203	3	80.3	80.3	NUM
ajst-21816	203	4	√	√	NUM
ajst-21816	204	1	√	√	ADP
ajst-21816	204	2	96.8	96.8	NUM
ajst-21816	204	3	91.9	91.9	NUM
ajst-21816	204	4	78.7	78.7	NUM
ajst-21816	204	5	3.4.2	3.4.2	NUM
ajst-21816	204	6	.	.	PUNCT
ajst-21816	205	1	global	global	ADJ
ajst-21816	205	2	model	model	NOUN
ajst-21816	205	3	ablation	ablation	NOUN
ajst-21816	205	4	experiment	experiment	NOUN
ajst-21816	205	5	in	in	ADP
ajst-21816	205	6	order	order	NOUN
ajst-21816	205	7	to	to	PART
ajst-21816	205	8	verify	verify	VERB
ajst-21816	205	9	the	the	DET
ajst-21816	205	10	effectiveness	effectiveness	NOUN
ajst-21816	205	11	of	of	ADP
ajst-21816	205	12	each	each	DET
ajst-21816	205	13	component	component	NOUN
ajst-21816	205	14	in	in	ADP
ajst-21816	205	15	this	this	DET
ajst-21816	205	16	method	method	NOUN
ajst-21816	205	17	,	,	PUNCT
ajst-21816	205	18	ablation	ablation	NOUN
ajst-21816	205	19	experiments	experiment	NOUN
ajst-21816	205	20	were	be	AUX
ajst-21816	205	21	conducted	conduct	VERB
ajst-21816	205	22	on	on	ADP
ajst-21816	205	23	these	these	DET
ajst-21816	205	24	components	component	NOUN
ajst-21816	205	25	on	on	ADP
ajst-21816	205	26	the	the	DET
ajst-21816	205	27	casia	casia	NOUN
ajst-21816	205	28	-	-	PUNCT
ajst-21816	205	29	b	b	NOUN
ajst-21816	205	30	dataset	dataset	NOUN
ajst-21816	205	31	,	,	PUNCT
ajst-21816	205	32	and	and	CCONJ
ajst-21816	205	33	the	the	DET
ajst-21816	205	34	experimental	experimental	ADJ
ajst-21816	205	35	results	result	NOUN
ajst-21816	205	36	are	be	AUX
ajst-21816	205	37	shown	show	VERB
ajst-21816	205	38	in	in	ADP
ajst-21816	205	39	table	table	NOUN
ajst-21816	205	40	5	5	NUM
ajst-21816	205	41	.	.	PUNCT
ajst-21816	206	1	√	√	PROPN
ajst-21816	206	2	indicates	indicate	VERB
ajst-21816	206	3	the	the	DET
ajst-21816	206	4	use	use	NOUN
ajst-21816	206	5	of	of	ADP
ajst-21816	206	6	this	this	DET
ajst-21816	206	7	component	component	NOUN
ajst-21816	206	8	.	.	PUNCT
ajst-21816	207	1	in	in	ADP
ajst-21816	207	2	the	the	DET
ajst-21816	207	3	table	table	NOUN
ajst-21816	207	4	,	,	PUNCT
ajst-21816	207	5	conv	conv	NOUN
ajst-21816	207	6	stands	stand	VERB
ajst-21816	207	7	for	for	ADP
ajst-21816	207	8	common	common	ADJ
ajst-21816	207	9	convolution	convolution	NOUN
ajst-21816	207	10	,	,	PUNCT
ajst-21816	207	11	sconv	sconv	NOUN
ajst-21816	207	12	stands	stand	VERB
ajst-21816	207	13	for	for	ADP
ajst-21816	207	14	the	the	DET
ajst-21816	207	15	specific	specific	ADJ
ajst-21816	207	16	proportional	proportional	ADJ
ajst-21816	207	17	segmentation	segmentation	NOUN
ajst-21816	207	18	convolution	convolution	NOUN
ajst-21816	207	19	proposed	propose	VERB
ajst-21816	207	20	in	in	ADP
ajst-21816	207	21	this	this	DET
ajst-21816	207	22	paper	paper	NOUN
ajst-21816	207	23	,	,	PUNCT
ajst-21816	207	24	mgp	mgp	PROPN
ajst-21816	207	25	stands	stand	VERB
ajst-21816	207	26	for	for	ADP
ajst-21816	207	27	multilayer	multilayer	ADJ
ajst-21816	207	28	global	global	ADJ
ajst-21816	207	29	pipeline	pipeline	NOUN
ajst-21816	207	30	,	,	PUNCT
ajst-21816	207	31	se	se	PROPN
ajst-21816	207	32	stands	stand	VERB
ajst-21816	207	33	for	for	ADP
ajst-21816	207	34	se	se	X
ajst-21816	207	35	block	block	NOUN
ajst-21816	207	36	,	,	PUNCT
ajst-21816	207	37	mgp+se	mgp+se	PROPN
ajst-21816	207	38	stands	stand	VERB
ajst-21816	207	39	for	for	ADP
ajst-21816	207	40	mgpa	mgpa	NOUN
ajst-21816	207	41	module	module	NOUN
ajst-21816	207	42	.	.	PUNCT
ajst-21816	208	1	by	by	ADP
ajst-21816	208	2	dividing	divide	VERB
ajst-21816	208	3	the	the	DET
ajst-21816	208	4	six	six	NUM
ajst-21816	208	5	groups	group	NOUN
ajst-21816	208	6	of	of	ADP
ajst-21816	208	7	experimental	experimental	ADJ
ajst-21816	208	8	results	result	NOUN
ajst-21816	208	9	into	into	ADP
ajst-21816	208	10	the	the	DET
ajst-21816	208	11	first	first	ADJ
ajst-21816	208	12	three	three	NUM
ajst-21816	208	13	groups	group	NOUN
ajst-21816	208	14	and	and	CCONJ
ajst-21816	208	15	the	the	DET
ajst-21816	208	16	last	last	ADJ
ajst-21816	208	17	three	three	NUM
ajst-21816	208	18	groups	group	NOUN
ajst-21816	208	19	for	for	ADP
ajst-21816	208	20	analysis	analysis	NOUN
ajst-21816	208	21	,	,	PUNCT
ajst-21816	208	22	it	it	PRON
ajst-21816	208	23	can	can	AUX
ajst-21816	208	24	be	be	AUX
ajst-21816	208	25	seen	see	VERB
ajst-21816	208	26	that	that	SCONJ
ajst-21816	208	27	whether	whether	SCONJ
ajst-21816	208	28	using	use	VERB
ajst-21816	208	29	conv	conv	NOUN
ajst-21816	208	30	or	or	CCONJ
ajst-21816	208	31	sconv	sconv	NOUN
ajst-21816	208	32	,	,	PUNCT
ajst-21816	208	33	the	the	DET
ajst-21816	208	34	recognition	recognition	NOUN
ajst-21816	208	35	accuracy	accuracy	NOUN
ajst-21816	208	36	of	of	ADP
ajst-21816	208	37	our	our	PRON
ajst-21816	208	38	model	model	NOUN
ajst-21816	208	39	is	be	AUX
ajst-21816	208	40	slightly	slightly	ADV
ajst-21816	208	41	improved	improve	VERB
ajst-21816	208	42	by	by	ADP
ajst-21816	208	43	using	use	VERB
ajst-21816	208	44	mgp	mgp	NOUN
ajst-21816	208	45	alone	alone	ADV
ajst-21816	208	46	.	.	PUNCT
ajst-21816	209	1	however	however	ADV
ajst-21816	209	2	,	,	PUNCT
ajst-21816	209	3	the	the	DET
ajst-21816	209	4	performance	performance	NOUN
ajst-21816	209	5	of	of	ADP
ajst-21816	209	6	mgp	mgp	NOUN
ajst-21816	209	7	module	module	NOUN
ajst-21816	209	8	is	be	AUX
ajst-21816	209	9	not	not	PART
ajst-21816	209	10	good	good	ADJ
ajst-21816	209	11	in	in	ADP
ajst-21816	209	12	cl	cl	NOUN
ajst-21816	209	13	walking	walk	VERB
ajst-21816	209	14	state	state	NOUN
ajst-21816	209	15	.	.	PUNCT
ajst-21816	210	1	with	with	ADP
ajst-21816	210	2	the	the	DET
ajst-21816	210	3	addition	addition	NOUN
ajst-21816	210	4	of	of	ADP
ajst-21816	210	5	se	se	PROPN
ajst-21816	210	6	to	to	ADP
ajst-21816	210	7	mgp	mgp	PROPN
ajst-21816	210	8	,	,	PUNCT
ajst-21816	210	9	the	the	DET
ajst-21816	210	10	performance	performance	NOUN
ajst-21816	210	11	of	of	ADP
ajst-21816	210	12	our	our	PRON
ajst-21816	210	13	models	model	NOUN
ajst-21816	210	14	has	have	AUX
ajst-21816	210	15	been	be	AUX
ajst-21816	210	16	significantly	significantly	ADV
ajst-21816	210	17	improved	improve	VERB
ajst-21816	210	18	.	.	PUNCT
ajst-21816	211	1	comparing	compare	VERB
ajst-21816	211	2	the	the	DET
ajst-21816	211	3	first	first	ADJ
ajst-21816	211	4	and	and	CCONJ
ajst-21816	211	5	fourth	fourth	ADJ
ajst-21816	211	6	sets	set	NOUN
ajst-21816	211	7	of	of	ADP
ajst-21816	211	8	data	datum	NOUN
ajst-21816	211	9	,	,	PUNCT
ajst-21816	211	10	the	the	DET
ajst-21816	211	11	second	second	ADJ
ajst-21816	211	12	and	and	CCONJ
ajst-21816	211	13	fifth	fifth	ADJ
ajst-21816	211	14	sets	set	NOUN
ajst-21816	211	15	of	of	ADP
ajst-21816	211	16	data	datum	NOUN
ajst-21816	211	17	,	,	PUNCT
ajst-21816	211	18	and	and	CCONJ
ajst-21816	211	19	the	the	DET
ajst-21816	211	20	third	third	ADJ
ajst-21816	211	21	and	and	CCONJ
ajst-21816	211	22	sixth	sixth	ADJ
ajst-21816	211	23	sets	set	NOUN
ajst-21816	211	24	of	of	ADP
ajst-21816	211	25	data	datum	NOUN
ajst-21816	211	26	,	,	PUNCT
ajst-21816	211	27	it	it	PRON
ajst-21816	211	28	is	be	AUX
ajst-21816	211	29	obvious	obvious	ADJ
ajst-21816	211	30	that	that	SCONJ
ajst-21816	211	31	sconv	sconv	NOUN
ajst-21816	211	32	is	be	AUX
ajst-21816	211	33	more	more	ADV
ajst-21816	211	34	effective	effective	ADJ
ajst-21816	211	35	than	than	ADP
ajst-21816	211	36	conv	conv	ADJ
ajst-21816	211	37	in	in	ADP
ajst-21816	211	38	this	this	DET
ajst-21816	211	39	model	model	NOUN
ajst-21816	211	40	.	.	PUNCT
ajst-21816	212	1	table	table	NOUN
ajst-21816	212	2	5	5	NUM
ajst-21816	212	3	.	.	PUNCT
ajst-21816	212	4	ablation	ablation	NOUN
ajst-21816	212	5	experiments	experiment	NOUN
ajst-21816	212	6	of	of	ADP
ajst-21816	212	7	lfe	lfe	PROPN
ajst-21816	212	8	-	-	PUNCT
ajst-21816	212	9	mgpa	mgpa	PROPN
ajst-21816	212	10	on	on	ADP
ajst-21816	212	11	casia	casia	NOUN
ajst-21816	212	12	-	-	PUNCT
ajst-21816	212	13	b	b	NOUN
ajst-21816	212	14	dataset	dataset	VERB
ajst-21816	212	15	conv	conv	PROPN
ajst-21816	212	16	sconv	sconv	PROPN
ajst-21816	212	17	mgp	mgp	PROPN
ajst-21816	212	18	se	se	PROPN
ajst-21816	212	19	nm	nm	PROPN
ajst-21816	212	20	bg	bg	PROPN
ajst-21816	212	21	cl	cl	NOUN
ajst-21816	212	22	√	√	NOUN
ajst-21816	212	23	95.3	95.3	NUM
ajst-21816	212	24	88.4	88.4	NUM
ajst-21816	212	25	73.8	73.8	NUM
ajst-21816	212	26	√	√	NUM
ajst-21816	213	1	√	√	ADP
ajst-21816	213	2	95.3	95.3	NUM
ajst-21816	213	3	88.5	88.5	NUM
ajst-21816	213	4	73.9	73.9	NUM
ajst-21816	213	5	√	√	NOUN
ajst-21816	213	6	√	√	ADP
ajst-21816	214	1	√	√	NUM
ajst-21816	215	1	96.2	96.2	NUM
ajst-21816	215	2	91.2	91.2	NUM
ajst-21816	215	3	75.3	75.3	NUM
ajst-21816	215	4	√	√	PROPN
ajst-21816	215	5	96.0	96.0	NUM
ajst-21816	215	6	90.3	90.3	NUM
ajst-21816	215	7	79.6	79.6	NUM
ajst-21816	215	8	√	√	NUM
ajst-21816	215	9	√	√	ADP
ajst-21816	215	10	96.2	96.2	NUM
ajst-21816	215	11	91.0	91.0	NUM
ajst-21816	215	12	79.5	79.5	NUM
ajst-21816	215	13	√	√	NUM
ajst-21816	215	14	√	√	ADP
ajst-21816	215	15	√	√	ADP
ajst-21816	216	1	97.5	97.5	NUM
ajst-21816	216	2	94.5	94.5	NUM
ajst-21816	216	3	81.1	81.1	NUM
ajst-21816	216	4	4	4	NUM
ajst-21816	216	5	.	.	PUNCT
ajst-21816	217	1	summary	summary	NOUN
ajst-21816	217	2	to	to	PART
ajst-21816	217	3	solve	solve	VERB
ajst-21816	217	4	the	the	DET
ajst-21816	217	5	problem	problem	NOUN
ajst-21816	217	6	that	that	PRON
ajst-21816	217	7	gait	gait	ADJ
ajst-21816	217	8	recognition	recognition	NOUN
ajst-21816	217	9	is	be	AUX
ajst-21816	217	10	greatly	greatly	ADV
ajst-21816	217	11	affected	affect	VERB
ajst-21816	217	12	by	by	ADP
ajst-21816	217	13	external	external	ADJ
ajst-21816	217	14	factors	factor	NOUN
ajst-21816	217	15	such	such	ADJ
ajst-21816	217	16	as	as	ADP
ajst-21816	217	17	wearing	wear	VERB
ajst-21816	217	18	and	and	CCONJ
ajst-21816	217	19	carrying	carry	VERB
ajst-21816	217	20	,	,	PUNCT
ajst-21816	217	21	the	the	DET
ajst-21816	217	22	lfe	lfe	PROPN
ajst-21816	217	23	-	-	PUNCT
ajst-21816	217	24	mgpa	mgpa	PROPN
ajst-21816	217	25	network	network	NOUN
ajst-21816	217	26	model	model	NOUN
ajst-21816	217	27	is	be	AUX
ajst-21816	217	28	proposed	propose	VERB
ajst-21816	217	29	in	in	ADP
ajst-21816	217	30	this	this	DET
ajst-21816	217	31	paper	paper	NOUN
ajst-21816	217	32	.	.	PUNCT
ajst-21816	218	1	firstly	firstly	ADV
ajst-21816	218	2	,	,	PUNCT
ajst-21816	218	3	an	an	DET
ajst-21816	218	4	lfe	lfe	PROPN
ajst-21816	218	5	containing	contain	VERB
ajst-21816	218	6	multiple	multiple	ADJ
ajst-21816	218	7	sconvs	sconvs	NOUN
ajst-21816	218	8	is	be	AUX
ajst-21816	218	9	constructed	construct	VERB
ajst-21816	218	10	in	in	ADP
ajst-21816	218	11	the	the	DET
ajst-21816	218	12	model	model	NOUN
ajst-21816	218	13	.	.	PUNCT
ajst-21816	219	1	the	the	DET
ajst-21816	219	2	sconv	sconv	NOUN
ajst-21816	219	3	splits	split	VERB
ajst-21816	219	4	the	the	DET
ajst-21816	219	5	input	input	NOUN
ajst-21816	219	6	gait	gait	PROPN
ajst-21816	219	7	contour	contour	NOUN
ajst-21816	219	8	to	to	PART
ajst-21816	219	9	extract	extract	VERB
ajst-21816	219	10	finegrained	finegraine	VERB
ajst-21816	219	11	gait	gait	NOUN
ajst-21816	219	12	features	feature	NOUN
ajst-21816	219	13	.	.	PUNCT
ajst-21816	220	1	secondly	secondly	ADV
ajst-21816	220	2	,	,	PUNCT
ajst-21816	220	3	the	the	DET
ajst-21816	220	4	mgp	mgp	NOUN
ajst-21816	220	5	module	module	NOUN
ajst-21816	220	6	is	be	AUX
ajst-21816	220	7	integrated	integrate	VERB
ajst-21816	220	8	,	,	PUNCT
ajst-21816	220	9	and	and	CCONJ
ajst-21816	220	10	the	the	DET
ajst-21816	220	11	channel	channel	NOUN
ajst-21816	220	12	attention	attention	NOUN
ajst-21816	220	13	mechanism	mechanism	NOUN
ajst-21816	220	14	is	be	AUX
ajst-21816	220	15	added	add	VERB
ajst-21816	220	16	to	to	ADP
ajst-21816	220	17	the	the	DET
ajst-21816	220	18	module	module	NOUN
ajst-21816	220	19	.	.	PUNCT
ajst-21816	221	1	finally	finally	ADV
ajst-21816	221	2	,	,	PUNCT
ajst-21816	221	3	a	a	DET
ajst-21816	221	4	joint	joint	ADJ
ajst-21816	221	5	loss	loss	NOUN
ajst-21816	221	6	function	function	NOUN
ajst-21816	221	7	is	be	AUX
ajst-21816	221	8	constructed	construct	VERB
ajst-21816	221	9	to	to	PART
ajst-21816	221	10	improve	improve	VERB
ajst-21816	221	11	the	the	DET
ajst-21816	221	12	recognition	recognition	NOUN
ajst-21816	221	13	rate	rate	NOUN
ajst-21816	221	14	.	.	PUNCT
ajst-21816	222	1	experimental	experimental	ADJ
ajst-21816	222	2	results	result	NOUN
ajst-21816	222	3	on	on	ADP
ajst-21816	222	4	the	the	DET
ajst-21816	222	5	casia	casia	NOUN
ajst-21816	222	6	-	-	PUNCT
ajst-21816	222	7	b	b	NOUN
ajst-21816	222	8	dataset	dataset	NOUN
ajst-21816	222	9	show	show	NOUN
ajst-21816	222	10	that	that	SCONJ
ajst-21816	222	11	compared	compare	VERB
ajst-21816	222	12	with	with	ADP
ajst-21816	222	13	the	the	DET
ajst-21816	222	14	current	current	ADJ
ajst-21816	222	15	mainstream	mainstream	NOUN
ajst-21816	222	16	294	294	NUM
ajst-21816	222	17	algorithms	algorithm	NOUN
ajst-21816	222	18	,	,	PUNCT
ajst-21816	222	19	the	the	DET
ajst-21816	222	20	proposed	propose	VERB
ajst-21816	222	21	method	method	NOUN
ajst-21816	222	22	achieves	achieve	VERB
ajst-21816	222	23	a	a	DET
ajst-21816	222	24	higher	high	ADJ
ajst-21816	222	25	recognition	recognition	NOUN
ajst-21816	222	26	rate	rate	NOUN
ajst-21816	222	27	,	,	PUNCT
ajst-21816	222	28	especially	especially	ADV
ajst-21816	222	29	in	in	ADP
ajst-21816	222	30	the	the	DET
ajst-21816	222	31	disturbed	disturbed	ADJ
ajst-21816	222	32	state	state	NOUN
ajst-21816	222	33	,	,	PUNCT
ajst-21816	222	34	the	the	DET
ajst-21816	222	35	recognition	recognition	NOUN
ajst-21816	222	36	rate	rate	NOUN
ajst-21816	222	37	is	be	AUX
ajst-21816	222	38	significantly	significantly	ADV
ajst-21816	222	39	improved	improve	VERB
ajst-21816	222	40	,	,	PUNCT
ajst-21816	222	41	indicating	indicate	VERB
ajst-21816	222	42	that	that	SCONJ
ajst-21816	222	43	the	the	DET
ajst-21816	222	44	method	method	NOUN
ajst-21816	222	45	has	have	VERB
ajst-21816	222	46	good	good	ADJ
ajst-21816	222	47	robustness	robustness	NOUN
ajst-21816	222	48	in	in	ADP
ajst-21816	222	49	complex	complex	ADJ
ajst-21816	222	50	scenarios	scenario	NOUN
ajst-21816	222	51	.	.	PUNCT
ajst-21816	223	1	in	in	ADP
ajst-21816	223	2	spite	spite	NOUN
ajst-21816	223	3	of	of	ADP
ajst-21816	223	4	this	this	PRON
ajst-21816	223	5	,	,	PUNCT
ajst-21816	223	6	the	the	DET
ajst-21816	223	7	influence	influence	NOUN
ajst-21816	223	8	of	of	ADP
ajst-21816	223	9	the	the	DET
ajst-21816	223	10	shape	shape	NOUN
ajst-21816	223	11	change	change	NOUN
ajst-21816	223	12	on	on	ADP
ajst-21816	223	13	the	the	DET
ajst-21816	223	14	gait	gait	ADJ
ajst-21816	223	15	profile	profile	NOUN
ajst-21816	223	16	can	can	AUX
ajst-21816	223	17	not	not	PART
ajst-21816	223	18	be	be	AUX
ajst-21816	223	19	ignored	ignore	VERB
ajst-21816	223	20	.	.	PUNCT
ajst-21816	224	1	in	in	ADP
ajst-21816	224	2	the	the	DET
ajst-21816	224	3	next	next	ADJ
ajst-21816	224	4	work	work	NOUN
ajst-21816	224	5	,	,	PUNCT
ajst-21816	224	6	the	the	DET
ajst-21816	224	7	fusion	fusion	NOUN
ajst-21816	224	8	with	with	ADP
ajst-21816	224	9	model	model	NOUN
ajst-21816	224	10	-	-	PUNCT
ajst-21816	224	11	based	base	VERB
ajst-21816	224	12	methods	method	NOUN
ajst-21816	224	13	such	such	ADJ
ajst-21816	224	14	as	as	ADP
ajst-21816	224	15	human	human	ADJ
ajst-21816	224	16	posture	posture	NOUN
ajst-21816	224	17	will	will	AUX
ajst-21816	224	18	be	be	AUX
ajst-21816	224	19	considered	consider	VERB
ajst-21816	224	20	to	to	PART
ajst-21816	224	21	obtain	obtain	VERB
ajst-21816	224	22	a	a	DET
ajst-21816	224	23	more	more	ADV
ajst-21816	224	24	accurate	accurate	ADJ
ajst-21816	224	25	representation	representation	NOUN
ajst-21816	224	26	of	of	ADP
ajst-21816	224	27	gait	gait	NOUN
ajst-21816	224	28	.	.	PUNCT
ajst-21816	225	1	references	reference	NOUN
ajst-21816	225	2	[	[	X
ajst-21816	225	3	1	1	X
ajst-21816	225	4	]	]	PUNCT
ajst-21816	225	5	zeng	zeng	PROPN
ajst-21816	225	6	f	f	PROPN
ajst-21816	225	7	,	,	PUNCT
ajst-21816	225	8	hu	hu	PROPN
ajst-21816	225	9	s	s	PROPN
ajst-21816	225	10	,	,	PUNCT
ajst-21816	225	11	xiao	xiao	PROPN
ajst-21816	225	12	,	,	PUNCT
ajst-21816	225	13	k.	k.	PROPN
ajst-21816	225	14	research	research	PROPN
ajst-21816	225	15	on	on	ADP
ajst-21816	225	16	partial	partial	ADJ
ajst-21816	225	17	fingerprint	fingerprint	NOUN
ajst-21816	225	18	recognition	recognition	NOUN
ajst-21816	225	19	algorithm	algorithm	NOUN
ajst-21816	225	20	based	base	VERB
ajst-21816	225	21	on	on	ADP
ajst-21816	225	22	deep	deep	ADJ
ajst-21816	225	23	learning[j	learning[j	NOUN
ajst-21816	225	24	]	]	PUNCT
ajst-21816	225	25	.	.	PUNCT
ajst-21816	226	1	neural	neural	PROPN
ajst-21816	226	2	computing	computing	PROPN
ajst-21816	226	3	&	&	CCONJ
ajst-21816	226	4	applications,2019,31(9):4789	applications,2019,31(9):4789	ADV
ajst-21816	226	5	-	-	PUNCT
ajst-21816	226	6	4798	4798	NUM
ajst-21816	226	7	.	.	PUNCT
ajst-21816	227	1	[	[	X
ajst-21816	227	2	2	2	X
ajst-21816	227	3	]	]	PUNCT
ajst-21816	227	4	sepasmoghaddam	sepasmoghaddam	NOUN
ajst-21816	227	5	a	a	PRON
ajst-21816	227	6	,	,	PUNCT
ajst-21816	227	7	pereira	pereira	PROPN
ajst-21816	227	8	f	f	PROPN
ajst-21816	227	9	m	m	PROPN
ajst-21816	227	10	,	,	PUNCT
ajst-21816	227	11	correia	correia	PROPN
ajst-21816	227	12	p	p	PROPN
ajst-21816	227	13	l	l	PROPN
ajst-21816	227	14	.face	.face	NOUN
ajst-21816	227	15	recognition	recognition	NOUN
ajst-21816	227	16	:	:	PUNCT
ajst-21816	227	17	a	a	DET
ajst-21816	227	18	novel	novel	ADJ
ajst-21816	227	19	multi	multi	ADJ
ajst-21816	227	20	-	-	ADJ
ajst-21816	227	21	level	level	ADJ
ajst-21816	227	22	taxonomy	taxonomy	NOUN
ajst-21816	227	23	based	base	VERB
ajst-21816	227	24	survey[j].iet	survey[j].iet	PROPN
ajst-21816	227	25	biometrics	biometric	NOUN
ajst-21816	227	26	,	,	PUNCT
ajst-21816	227	27	2019	2019	NUM
ajst-21816	227	28	,	,	PUNCT
ajst-21816	227	29	9(2	9(2	NUM
ajst-21816	227	30	)	)	PUNCT
ajst-21816	227	31	.	.	PUNCT
ajst-21816	228	1	[	[	X
ajst-21816	228	2	3	3	X
ajst-21816	228	3	]	]	X
ajst-21816	228	4	sun	sun	PROPN
ajst-21816	228	5	y	y	PROPN
ajst-21816	228	6	,	,	PUNCT
ajst-21816	228	7	wang	wang	PROPN
ajst-21816	228	8	x	x	PROPN
ajst-21816	228	9	,	,	PUNCT
ajst-21816	228	10	tang	tang	PROPN
ajst-21816	228	11	x.	x.	AUX
ajst-21816	228	12	deep	deep	ADJ
ajst-21816	228	13	learning	learning	NOUN
ajst-21816	228	14	face	face	VERB
ajst-21816	228	15	representation	representation	NOUN
ajst-21816	228	16	from	from	ADP
ajst-21816	228	17	predicting	predict	VERB
ajst-21816	228	18	10,000	10,000	NUM
ajst-21816	228	19	classes[j	classes[j	NOUN
ajst-21816	228	20	]	]	PUNCT
ajst-21816	228	21	.	.	PUNCT
ajst-21816	229	1	2014	2014	NUM
ajst-21816	229	2	ieee	ieee	NOUN
ajst-21816	229	3	conference	conference	NOUN
ajst-21816	229	4	on	on	ADP
ajst-21816	229	5	computer	computer	NOUN
ajst-21816	229	6	vision	vision	NOUN
ajst-21816	229	7	and	and	CCONJ
ajst-21816	229	8	pattern	pattern	NOUN
ajst-21816	229	9	recognition	recognition	NOUN
ajst-21816	229	10	,	,	PUNCT
ajst-21816	229	11	columbus	columbus	PROPN
ajst-21816	229	12	,	,	PUNCT
ajst-21816	229	13	oh	oh	INTJ
ajst-21816	229	14	,	,	PUNCT
ajst-21816	229	15	usa	usa	PROPN
ajst-21816	229	16	,	,	PUNCT
ajst-21816	229	17	2014	2014	NUM
ajst-21816	229	18	:	:	PUNCT
ajst-21816	229	19	1891–1898	1891–1898	NUM
ajst-21816	229	20	.	.	PUNCT
ajst-21816	230	1	[	[	X
ajst-21816	230	2	4	4	X
ajst-21816	230	3	]	]	X
ajst-21816	230	4	wang	wang	PROPN
ajst-21816	230	5	k	k	PROPN
ajst-21816	230	6	,	,	PUNCT
ajst-21816	230	7	kumar	kumar	PROPN
ajst-21816	230	8	a	a	DET
ajst-21816	230	9	.periocular	.periocular	ADV
ajst-21816	230	10	-	-	PUNCT
ajst-21816	230	11	assisted	assist	VERB
ajst-21816	230	12	multi	multi	ADJ
ajst-21816	230	13	-	-	ADJ
ajst-21816	230	14	feature	feature	ADJ
ajst-21816	230	15	collaboration	collaboration	NOUN
ajst-21816	230	16	for	for	ADP
ajst-21816	230	17	dynamic	dynamic	ADJ
ajst-21816	230	18	iris	iris	NOUN
ajst-21816	230	19	recognition[j].ieee	recognition[j].ieee	NOUN
ajst-21816	230	20	transactions	transaction	NOUN
ajst-21816	230	21	on	on	ADP
ajst-21816	230	22	information	information	NOUN
ajst-21816	230	23	forensics	forensic	NOUN
ajst-21816	230	24	and	and	CCONJ
ajst-21816	230	25	security	security	NOUN
ajst-21816	230	26	,	,	PUNCT
ajst-21816	230	27	2021	2021	NUM
ajst-21816	230	28	,	,	PUNCT
ajst-21816	230	29	16:866	16:866	PROPN
ajst-21816	230	30	-	-	PUNCT
ajst-21816	230	31	879	879	NUM
ajst-21816	230	32	.	.	PUNCT
ajst-21816	231	1	[	[	X
ajst-21816	231	2	5	5	NUM
ajst-21816	231	3	]	]	X
ajst-21816	231	4	han	han	PROPN
ajst-21816	231	5	j	j	PROPN
ajst-21816	231	6	,	,	PUNCT
ajst-21816	231	7	bhanu	bhanu	PROPN
ajst-21816	231	8	b.	b.	PROPN
ajst-21816	231	9	individual	individual	PROPN
ajst-21816	231	10	recognition	recognition	PROPN
ajst-21816	231	11	using	use	VERB
ajst-21816	231	12	gait	gait	PROPN
ajst-21816	231	13	energy	energy	NOUN
ajst-21816	231	14	image[j	image[j	PROPN
ajst-21816	231	15	]	]	PUNCT
ajst-21816	231	16	.	.	PUNCT
ajst-21816	232	1	ieee	ieee	NOUN
ajst-21816	232	2	transactions	transaction	NOUN
ajst-21816	232	3	on	on	ADP
ajst-21816	232	4	pattern	pattern	NOUN
ajst-21816	232	5	analysis	analysis	NOUN
ajst-21816	232	6	and	and	CCONJ
ajst-21816	232	7	machine	machine	NOUN
ajst-21816	232	8	intelligence	intelligence	NOUN
ajst-21816	232	9	,	,	PUNCT
ajst-21816	232	10	2006	2006	NUM
ajst-21816	232	11	,	,	PUNCT
ajst-21816	232	12	28(2	28(2	NUM
ajst-21816	232	13	):	):	PUNCT
ajst-21816	232	14	316	316	NUM
ajst-21816	232	15	-	-	SYM
ajst-21816	232	16	322	322	NUM
ajst-21816	232	17	.	.	PUNCT
ajst-21816	233	1	[	[	X
ajst-21816	233	2	6	6	NUM
ajst-21816	233	3	]	]	X
ajst-21816	233	4	wu	wu	PROPN
ajst-21816	233	5	z	z	PROPN
ajst-21816	233	6	,	,	PUNCT
ajst-21816	233	7	huang	huang	PROPN
ajst-21816	233	8	y	y	PROPN
ajst-21816	233	9	,	,	PUNCT
ajst-21816	233	10	wang	wang	PROPN
ajst-21816	233	11	l	l	PROPN
ajst-21816	233	12	,	,	PUNCT
ajst-21816	233	13	et	et	PROPN
ajst-21816	233	14	al	al	PROPN
ajst-21816	233	15	.	.	PUNCT
ajst-21816	234	1	a	a	DET
ajst-21816	234	2	comprehensive	comprehensive	ADJ
ajst-21816	234	3	study	study	NOUN
ajst-21816	234	4	on	on	ADP
ajst-21816	234	5	cross	cross	ADJ
ajst-21816	234	6	-	-	ADJ
ajst-21816	234	7	view	view	ADJ
ajst-21816	234	8	gait	gait	NOUN
ajst-21816	234	9	based	base	VERB
ajst-21816	234	10	human	human	ADJ
ajst-21816	234	11	identification	identification	NOUN
ajst-21816	234	12	with	with	ADP
ajst-21816	234	13	deep	deep	ADJ
ajst-21816	234	14	cnns[j	cnns[j	NOUN
ajst-21816	234	15	]	]	X
ajst-21816	234	16	.	.	PUNCT
ajst-21816	235	1	ieee	ieee	NOUN
ajst-21816	235	2	transactions	transaction	NOUN
ajst-21816	235	3	on	on	ADP
ajst-21816	235	4	pattern	pattern	NOUN
ajst-21816	235	5	analysis	analysis	NOUN
ajst-21816	235	6	and	and	CCONJ
ajst-21816	235	7	machine	machine	NOUN
ajst-21816	235	8	intelligence	intelligence	NOUN
ajst-21816	235	9	,	,	PUNCT
ajst-21816	235	10	2017	2017	NUM
ajst-21816	235	11	,	,	PUNCT
ajst-21816	235	12	39(2	39(2	NUM
ajst-21816	235	13	):	):	PUNCT
ajst-21816	235	14	209	209	NUM
ajst-21816	235	15	-	-	SYM
ajst-21816	235	16	226	226	NUM
ajst-21816	235	17	.	.	PUNCT
ajst-21816	236	1	[	[	X
ajst-21816	236	2	7	7	X
ajst-21816	236	3	]	]	X
ajst-21816	236	4	wang	wang	PROPN
ajst-21816	236	5	c	c	PROPN
ajst-21816	236	6	,	,	PUNCT
ajst-21816	236	7	zhang	zhang	PROPN
ajst-21816	236	8	j	j	PROPN
ajst-21816	236	9	,	,	PUNCT
ajst-21816	236	10	wang	wang	PROPN
ajst-21816	236	11	l	l	PROPN
ajst-21816	236	12	,	,	PUNCT
ajst-21816	236	13	et	et	PROPN
ajst-21816	236	14	al	al	PROPN
ajst-21816	236	15	.	.	PUNCT
ajst-21816	237	1	human	human	ADJ
ajst-21816	237	2	identification	identification	NOUN
ajst-21816	237	3	using	use	VERB
ajst-21816	237	4	temporal	temporal	ADJ
ajst-21816	237	5	information	information	NOUN
ajst-21816	237	6	preserving	preserve	VERB
ajst-21816	237	7	gait	gait	NOUN
ajst-21816	237	8	template[j	template[j	NOUN
ajst-21816	237	9	]	]	PUNCT
ajst-21816	237	10	.	.	PUNCT
ajst-21816	238	1	ieee	ieee	NOUN
ajst-21816	238	2	transactions	transaction	NOUN
ajst-21816	238	3	on	on	ADP
ajst-21816	238	4	pattern	pattern	NOUN
ajst-21816	238	5	analysis	analysis	NOUN
ajst-21816	238	6	and	and	CCONJ
ajst-21816	238	7	machine	machine	NOUN
ajst-21816	238	8	intelligence	intelligence	NOUN
ajst-21816	238	9	,	,	PUNCT
ajst-21816	238	10	2012	2012	NUM
ajst-21816	238	11	,	,	PUNCT
ajst-21816	238	12	34(11	34(11	NUM
ajst-21816	238	13	):	):	PUNCT
ajst-21816	238	14	2164	2164	NUM
ajst-21816	238	15	-	-	SYM
ajst-21816	238	16	2176	2176	NUM
ajst-21816	238	17	.	.	PUNCT
ajst-21816	239	1	[	[	X
ajst-21816	239	2	8	8	NUM
ajst-21816	239	3	]	]	X
ajst-21816	239	4	alireza	alireza	PROPN
ajst-21816	239	5	s	s	PROPN
ajst-21816	239	6	,	,	PUNCT
ajst-21816	239	7	ali	ali	PROPN
ajst-21816	239	8	e.	e.	PROPN
ajst-21816	239	9	deep	deep	PROPN
ajst-21816	239	10	gait	gait	PROPN
ajst-21816	239	11	recognition	recognition	PROPN
ajst-21816	239	12	:	:	PUNCT
ajst-21816	239	13	a	a	DET
ajst-21816	239	14	survey[j	survey[j	PROPN
ajst-21816	239	15	]	]	PUNCT
ajst-21816	239	16	.	.	PUNCT
ajst-21816	240	1	ieee	ieee	NOUN
ajst-21816	240	2	transactions	transaction	NOUN
ajst-21816	240	3	on	on	ADP
ajst-21816	240	4	pattern	pattern	NOUN
ajst-21816	240	5	analysis	analysis	NOUN
ajst-21816	240	6	and	and	CCONJ
ajst-21816	240	7	machine	machine	NOUN
ajst-21816	240	8	intelligence	intelligence	NOUN
ajst-21816	240	9	,	,	PUNCT
ajst-21816	240	10	2022	2022	NUM
ajst-21816	240	11	,	,	PUNCT
ajst-21816	240	12	45(1	45(1	NOUN
ajst-21816	240	13	):	):	PUNCT
ajst-21816	240	14	264	264	NUM
ajst-21816	240	15	-	-	SYM
ajst-21816	240	16	284	284	NUM
ajst-21816	240	17	.	.	PUNCT
ajst-21816	241	1	[	[	X
ajst-21816	241	2	9	9	NUM
ajst-21816	241	3	]	]	SYM
ajst-21816	241	4	shiraga	shiraga	NOUN
ajst-21816	241	5	k	k	PROPN
ajst-21816	241	6	,	,	PUNCT
ajst-21816	241	7	makihara	makihara	ADJ
ajst-21816	241	8	y	y	NOUN
ajst-21816	241	9	,	,	PUNCT
ajst-21816	241	10	muramatsu	muramatsu	NOUN
ajst-21816	241	11	d	d	PROPN
ajst-21816	241	12	,	,	PUNCT
ajst-21816	241	13	et	et	PROPN
ajst-21816	241	14	al	al	PROPN
ajst-21816	241	15	.	.	PROPN
ajst-21816	241	16	geinet	geinet	PROPN
ajst-21816	241	17	:	:	PUNCT
ajst-21816	241	18	viewinvariant	viewinvariant	PROPN
ajst-21816	241	19	gait	gait	PROPN
ajst-21816	241	20	recognition	recognition	PROPN
ajst-21816	241	21	using	use	VERB
ajst-21816	241	22	a	a	DET
ajst-21816	241	23	convolutional	convolutional	ADJ
ajst-21816	241	24	neural	neural	NOUN
ajst-21816	241	25	network[c	network[c	PROPN
ajst-21816	241	26	]	]	PUNCT
ajst-21816	241	27	.	.	PUNCT
ajst-21816	242	1	proceedings	proceeding	NOUN
ajst-21816	242	2	of	of	ADP
ajst-21816	242	3	the	the	DET
ajst-21816	242	4	2016	2016	NUM
ajst-21816	242	5	international	international	ADJ
ajst-21816	242	6	conference	conference	NOUN
ajst-21816	242	7	on	on	ADP
ajst-21816	242	8	biometrics	biometric	NOUN
ajst-21816	242	9	(	(	PUNCT
ajst-21816	242	10	icb	icb	PROPN
ajst-21816	242	11	)	)	PUNCT
ajst-21816	242	12	,	,	PUNCT
ajst-21816	242	13	2016	2016	NUM
ajst-21816	242	14	:	:	PUNCT
ajst-21816	242	15	1	1	NUM
ajst-21816	242	16	-	-	SYM
ajst-21816	242	17	8	8	NUM
ajst-21816	242	18	.	.	PUNCT
ajst-21816	243	1	[	[	X
ajst-21816	243	2	10	10	NUM
ajst-21816	243	3	]	]	X
ajst-21816	243	4	fan	fan	NOUN
ajst-21816	243	5	c	c	NOUN
ajst-21816	243	6	,	,	PUNCT
ajst-21816	243	7	peng	peng	PROPN
ajst-21816	243	8	y	y	PROPN
ajst-21816	243	9	,	,	PUNCT
ajst-21816	243	10	cao	cao	PROPN
ajst-21816	243	11	c	c	X
ajst-21816	243	12	,	,	PUNCT
ajst-21816	244	1	et	et	PROPN
ajst-21816	244	2	al	al	PROPN
ajst-21816	244	3	.	.	PROPN
ajst-21816	244	4	gaitpart	gaitpart	NOUN
ajst-21816	244	5	:	:	PUNCT
ajst-21816	244	6	temporal	temporal	ADJ
ajst-21816	244	7	part	part	NOUN
ajst-21816	244	8	-	-	PUNCT
ajst-21816	244	9	based	base	VERB
ajst-21816	244	10	model	model	NOUN
ajst-21816	244	11	for	for	ADP
ajst-21816	244	12	gait	gait	ADJ
ajst-21816	244	13	recognition[j	recognition[j	PROPN
ajst-21816	244	14	]	]	PUNCT
ajst-21816	244	15	.	.	PUNCT
ajst-21816	245	1	proceedings	proceeding	NOUN
ajst-21816	245	2	of	of	ADP
ajst-21816	245	3	the	the	DET
ajst-21816	245	4	ieee	ieee	NOUN
ajst-21816	245	5	computer	computer	NOUN
ajst-21816	245	6	society	society	PROPN
ajst-21816	245	7	conference	conference	NOUN
ajst-21816	245	8	on	on	ADP
ajst-21816	245	9	computer	computer	NOUN
ajst-21816	245	10	vision	vision	NOUN
ajst-21816	245	11	and	and	CCONJ
ajst-21816	245	12	pattern	pattern	NOUN
ajst-21816	245	13	recognition	recognition	NOUN
ajst-21816	245	14	,	,	PUNCT
ajst-21816	245	15	2020	2020	NUM
ajst-21816	245	16	:	:	PUNCT
ajst-21816	245	17	14213	14213	NUM
ajst-21816	245	18	-	-	SYM
ajst-21816	245	19	14221	14221	NUM
ajst-21816	245	20	.	.	PUNCT
ajst-21816	246	1	[	[	X
ajst-21816	246	2	11	11	NUM
ajst-21816	246	3	]	]	X
ajst-21816	246	4	sepas	sepas	NOUN
ajst-21816	246	5	-	-	PUNCT
ajst-21816	246	6	moghaddam	moghaddam	NOUN
ajst-21816	246	7	a	a	DET
ajst-21816	246	8	,	,	PUNCT
ajst-21816	246	9	ghorbani	ghorbani	PROPN
ajst-21816	246	10	s	s	PART
ajst-21816	246	11	,	,	PUNCT
ajst-21816	246	12	troje	troje	NOUN
ajst-21816	246	13	nf	nf	PROPN
ajst-21816	246	14	,	,	PUNCT
ajst-21816	246	15	et	et	PROPN
ajst-21816	247	1	al	al	PROPN
ajst-21816	247	2	.	.	PROPN
ajst-21816	247	3	gait	gait	PROPN
ajst-21816	247	4	recognition	recognition	PROPN
ajst-21816	247	5	using	use	VERB
ajst-21816	247	6	multi	multi	ADJ
ajst-21816	247	7	-	-	ADJ
ajst-21816	247	8	scale	scale	ADJ
ajst-21816	247	9	partial	partial	ADJ
ajst-21816	247	10	representation	representation	NOUN
ajst-21816	247	11	transformation	transformation	NOUN
ajst-21816	247	12	with	with	ADP
ajst-21816	247	13	capsules[c	capsules[c	PROPN
ajst-21816	247	14	]	]	PUNCT
ajst-21816	247	15	.	.	PUNCT
ajst-21816	248	1	proceedings	proceeding	NOUN
ajst-21816	248	2	of	of	ADP
ajst-21816	248	3	the	the	DET
ajst-21816	248	4	2020	2020	NUM
ajst-21816	248	5	25th	25th	NOUN
ajst-21816	248	6	international	international	ADJ
ajst-21816	248	7	conference	conference	NOUN
ajst-21816	248	8	on	on	ADP
ajst-21816	248	9	pattern	pattern	NOUN
ajst-21816	248	10	recognition	recognition	NOUN
ajst-21816	248	11	(	(	PUNCT
ajst-21816	248	12	icpr	icpr	PROPN
ajst-21816	248	13	)	)	PUNCT
ajst-21816	248	14	,	,	PUNCT
ajst-21816	248	15	2021	2021	NUM
ajst-21816	248	16	,	,	PUNCT
ajst-21816	248	17	8045	8045	NUM
ajst-21816	248	18	-	-	SYM
ajst-21816	248	19	8052	8052	NUM
ajst-21816	248	20	.	.	PUNCT
ajst-21816	249	1	[	[	X
ajst-21816	249	2	12	12	NUM
ajst-21816	249	3	]	]	X
ajst-21816	249	4	chao	chao	PROPN
ajst-21816	249	5	h	h	PROPN
ajst-21816	249	6	,	,	PUNCT
ajst-21816	249	7	wang	wang	PROPN
ajst-21816	249	8	k	k	PROPN
ajst-21816	249	9	,	,	PUNCT
ajst-21816	249	10	he	he	PRON
ajst-21816	249	11	y	y	PROPN
ajst-21816	249	12	,	,	PUNCT
ajst-21816	249	13	et	et	PROPN
ajst-21816	249	14	al	al	PROPN
ajst-21816	249	15	.	.	PROPN
ajst-21816	249	16	gaitset	gaitset	PROPN
ajst-21816	249	17	:	:	PUNCT
ajst-21816	249	18	cross	cross	ADJ
ajst-21816	249	19	-	-	ADJ
ajst-21816	249	20	view	view	ADJ
ajst-21816	249	21	gait	gait	NOUN
ajst-21816	249	22	recognition	recognition	NOUN
ajst-21816	249	23	through	through	ADP
ajst-21816	249	24	utilizing	utilize	VERB
ajst-21816	249	25	gait	gait	NOUN
ajst-21816	249	26	as	as	ADP
ajst-21816	249	27	a	a	DET
ajst-21816	249	28	deep	deep	ADJ
ajst-21816	249	29	set[j	set[j	NOUN
ajst-21816	249	30	]	]	PUNCT
ajst-21816	249	31	.	.	PUNCT
ajst-21816	250	1	ieee	ieee	NOUN
ajst-21816	250	2	transactions	transaction	NOUN
ajst-21816	250	3	on	on	ADP
ajst-21816	250	4	pattern	pattern	NOUN
ajst-21816	250	5	analysis	analysis	NOUN
ajst-21816	250	6	and	and	CCONJ
ajst-21816	250	7	machine	machine	NOUN
ajst-21816	250	8	intelligence	intelligence	NOUN
ajst-21816	250	9	,	,	PUNCT
ajst-21816	250	10	2021	2021	NUM
ajst-21816	250	11	:	:	PUNCT
ajst-21816	250	12	1	1	NUM
ajst-21816	250	13	-	-	SYM
ajst-21816	250	14	12	12	NUM
ajst-21816	250	15	.	.	PUNCT
ajst-21816	251	1	[	[	X
ajst-21816	251	2	13	13	NUM
ajst-21816	251	3	]	]	X
ajst-21816	251	4	liu	liu	PROPN
ajst-21816	251	5	n	n	PROPN
ajst-21816	251	6	,	,	PUNCT
ajst-21816	251	7	lu	lu	PROPN
ajst-21816	251	8	j	j	PROPN
ajst-21816	251	9	,	,	PUNCT
ajst-21816	251	10	tan	tan	PROPN
ajst-21816	251	11	y.	y.	PROPN
ajst-21816	251	12	joint	joint	PROPN
ajst-21816	251	13	subspace	subspace	NOUN
ajst-21816	251	14	learning	learn	VERB
ajst-21816	251	15	for	for	ADP
ajst-21816	251	16	view	view	NOUN
ajst-21816	251	17	-	-	PUNCT
ajst-21816	251	18	invariant	invariant	ADJ
ajst-21816	251	19	gait	gait	ADJ
ajst-21816	251	20	recognition[j	recognition[j	PROPN
ajst-21816	251	21	]	]	PUNCT
ajst-21816	251	22	.	.	PUNCT
ajst-21816	252	1	ieee	ieee	PROPN
ajst-21816	252	2	signal	signal	NOUN
ajst-21816	252	3	process	process	NOUN
ajst-21816	252	4	.	.	PUNCT
ajst-21816	253	1	lett	lett	PROPN
ajst-21816	253	2	,	,	PUNCT
ajst-21816	253	3	2011	2011	NUM
ajst-21816	253	4	,	,	PUNCT
ajst-21816	253	5	18(7	18(7	NUM
ajst-21816	253	6	):	):	PUNCT
ajst-21816	253	7	431	431	NUM
ajst-21816	253	8	-	-	SYM
ajst-21816	253	9	434	434	NUM
ajst-21816	253	10	.	.	PUNCT
ajst-21816	254	1	[	[	X
ajst-21816	254	2	14	14	NUM
ajst-21816	254	3	]	]	X
ajst-21816	254	4	jean	jean	PROPN
ajst-21816	254	5	f	f	PROPN
ajst-21816	254	6	,	,	PUNCT
ajst-21816	254	7	albu	albu	PROPN
ajst-21816	254	8	ab	ab	PROPN
ajst-21816	254	9	,	,	PUNCT
ajst-21816	254	10	bergevin	bergevin	PROPN
ajst-21816	254	11	r.	r.	PROPN
ajst-21816	254	12	towards	towards	ADP
ajst-21816	254	13	view	view	NOUN
ajst-21816	254	14	-	-	PUNCT
ajst-21816	254	15	invariant	invariant	ADJ
ajst-21816	254	16	gait	gait	NOUN
ajst-21816	254	17	modeling	modeling	NOUN
ajst-21816	254	18	:	:	PUNCT
ajst-21816	254	19	computing	compute	VERB
ajst-21816	254	20	view	view	NOUN
ajst-21816	254	21	-	-	PUNCT
ajst-21816	254	22	normalized	normalize	VERB
ajst-21816	254	23	body	body	NOUN
ajst-21816	254	24	part	part	NOUN
ajst-21816	254	25	trajectories[j	trajectories[j	NOUN
ajst-21816	254	26	]	]	PUNCT
ajst-21816	254	27	.	.	PUNCT
ajst-21816	255	1	pattern	pattern	NOUN
ajst-21816	255	2	recognition	recognition	PROPN
ajst-21816	255	3	,	,	PUNCT
ajst-21816	255	4	2009	2009	NUM
ajst-21816	255	5	,	,	PUNCT
ajst-21816	255	6	42(11	42(11	NUM
ajst-21816	255	7	):	):	PUNCT
ajst-21816	255	8	2936	2936	NUM
ajst-21816	255	9	-	-	SYM
ajst-21816	255	10	2949	2949	NUM
ajst-21816	255	11	.	.	PUNCT
ajst-21816	256	1	[	[	X
ajst-21816	256	2	15	15	NUM
ajst-21816	256	3	]	]	X
ajst-21816	256	4	goffredo	goffredo	NOUN
ajst-21816	256	5	m	m	PROPN
ajst-21816	256	6	,	,	PUNCT
ajst-21816	256	7	bouchrika	bouchrika	PROPN
ajst-21816	256	8	i	i	PROPN
ajst-21816	256	9	,	,	PUNCT
ajst-21816	256	10	carter	carter	PROPN
ajst-21816	256	11	jn	jn	PROPN
ajst-21816	256	12	,	,	PUNCT
ajst-21816	256	13	et	et	PROPN
ajst-21816	256	14	al	al	PROPN
ajst-21816	256	15	.	.	PROPN
ajst-21816	256	16	self	self	NOUN
ajst-21816	256	17	-	-	PUNCT
ajst-21816	256	18	calibrating	calibrate	VERB
ajst-21816	256	19	view	view	NOUN
ajst-21816	256	20	-	-	PUNCT
ajst-21816	256	21	invariant	invariant	ADJ
ajst-21816	256	22	gait	gait	NOUN
ajst-21816	256	23	biometrics[j	biometrics[j	PROPN
ajst-21816	256	24	]	]	PUNCT
ajst-21816	256	25	.	.	PUNCT
ajst-21816	257	1	ieee	ieee	NOUN
ajst-21816	257	2	transactions	transaction	NOUN
ajst-21816	257	3	on	on	ADP
ajst-21816	257	4	systems	system	NOUN
ajst-21816	257	5	,	,	PUNCT
ajst-21816	257	6	man	man	NOUN
ajst-21816	257	7	,	,	PUNCT
ajst-21816	257	8	and	and	CCONJ
ajst-21816	257	9	cybernetics	cybernetic	NOUN
ajst-21816	257	10	,	,	PUNCT
ajst-21816	257	11	part	part	NOUN
ajst-21816	257	12	b	b	PROPN
ajst-21816	257	13	(	(	PUNCT
ajst-21816	257	14	cybernetics	cybernetic	NOUN
ajst-21816	257	15	)	)	PUNCT
ajst-21816	257	16	,	,	PUNCT
ajst-21816	257	17	2010	2010	NUM
ajst-21816	257	18	,	,	PUNCT
ajst-21816	257	19	40	40	NUM
ajst-21816	257	20	(	(	PUNCT
ajst-21816	257	21	4	4	NUM
ajst-21816	257	22	):	):	PUNCT
ajst-21816	257	23	997	997	NUM
ajst-21816	257	24	-	-	SYM
ajst-21816	257	25	1008	1008	NUM
ajst-21816	257	26	.	.	PUNCT
ajst-21816	258	1	[	[	X
ajst-21816	258	2	16	16	NUM
ajst-21816	258	3	]	]	PUNCT
ajst-21816	258	4	yu	yu	PROPN
ajst-21816	258	5	s	s	PROPN
ajst-21816	258	6	,	,	PUNCT
ajst-21816	258	7	tan	tan	PROPN
ajst-21816	258	8	d	d	PROPN
ajst-21816	258	9	,	,	PUNCT
ajst-21816	258	10	tan	tan	PROPN
ajst-21816	258	11	t.	t.	NOUN
ajst-21816	258	12	a	a	DET
ajst-21816	258	13	framework	framework	NOUN
ajst-21816	258	14	for	for	ADP
ajst-21816	258	15	evaluating	evaluate	VERB
ajst-21816	258	16	the	the	DET
ajst-21816	258	17	effect	effect	NOUN
ajst-21816	258	18	of	of	ADP
ajst-21816	258	19	view	view	NOUN
ajst-21816	258	20	angle	angle	NOUN
ajst-21816	258	21	,	,	PUNCT
ajst-21816	258	22	clothing	clothing	NOUN
ajst-21816	258	23	and	and	CCONJ
ajst-21816	258	24	carrying	carry	VERB
ajst-21816	258	25	condition	condition	NOUN
ajst-21816	258	26	on	on	ADP
ajst-21816	258	27	gait	gait	PROPN
ajst-21816	258	28	recognition[c	recognition[c	PROPN
ajst-21816	258	29	]	]	PUNCT
ajst-21816	258	30	.	.	PUNCT
ajst-21816	259	1	proceedings	proceeding	NOUN
ajst-21816	259	2	of	of	ADP
ajst-21816	259	3	the	the	DET
ajst-21816	259	4	18th	18th	ADJ
ajst-21816	259	5	international	international	ADJ
ajst-21816	259	6	conference	conference	NOUN
ajst-21816	259	7	on	on	ADP
ajst-21816	259	8	pattern	pattern	NOUN
ajst-21816	259	9	recognition	recognition	NOUN
ajst-21816	259	10	(	(	PUNCT
ajst-21816	259	11	icpr'06	icpr'06	PROPN
ajst-21816	259	12	)	)	PUNCT
ajst-21816	259	13	,	,	PUNCT
ajst-21816	259	14	2006	2006	NUM
ajst-21816	259	15	:	:	PUNCT
ajst-21816	259	16	441	441	NUM
ajst-21816	259	17	-	-	SYM
ajst-21816	259	18	444	444	NUM
ajst-21816	259	19	.	.	PUNCT
ajst-21816	260	1	[	[	X
ajst-21816	260	2	17	17	NUM
ajst-21816	260	3	]	]	X
ajst-21816	260	4	hou	hou	PROPN
ajst-21816	260	5	s	s	PROPN
ajst-21816	260	6	,	,	PUNCT
ajst-21816	260	7	liu	liu	PROPN
ajst-21816	260	8	x	x	PROPN
ajst-21816	260	9	,	,	PUNCT
ajst-21816	260	10	cao	cao	PROPN
ajst-21816	260	11	c	c	PROPN
ajst-21816	260	12	,	,	PUNCT
ajst-21816	260	13	et	et	PROPN
ajst-21816	260	14	al.set	al.set	VERB
ajst-21816	260	15	residual	residual	ADJ
ajst-21816	260	16	network	network	NOUN
ajst-21816	260	17	for	for	ADP
ajst-21816	260	18	silhouette	silhouette	NOUN
ajst-21816	260	19	-	-	PUNCT
ajst-21816	260	20	based	base	VERB
ajst-21816	260	21	gait	gait	ADJ
ajst-21816	260	22	recognition[j].ieee	recognition[j].ieee	NOUN
ajst-21816	260	23	transactions	transaction	NOUN
ajst-21816	260	24	on	on	ADP
ajst-21816	260	25	biometrics	biometric	NOUN
ajst-21816	260	26	,	,	PUNCT
ajst-21816	260	27	behavior	behavior	NOUN
ajst-21816	260	28	,	,	PUNCT
ajst-21816	260	29	and	and	CCONJ
ajst-21816	260	30	identity	identity	NOUN
ajst-21816	260	31	science	science	NOUN
ajst-21816	260	32	,	,	PUNCT
ajst-21816	260	33	2021(33).doi:10.1109	2021(33).doi:10.1109	NUM
ajst-21816	260	34	/	/	SYM
ajst-21816	260	35	tbiom.2021.3074963	tbiom.2021.3074963	NOUN
ajst-21816	260	36	.	.	PUNCT