id	sid	tid	token	lemma	pos
ajst-5345	1	1	academic	academic	ADJ
ajst-5345	1	2	journal	journal	NOUN
ajst-5345	1	3	of	of	ADP
ajst-5345	1	4	science	science	NOUN
ajst-5345	1	5	and	and	CCONJ
ajst-5345	1	6	technology	technology	NOUN
ajst-5345	1	7	issn	issn	NOUN
ajst-5345	1	8	:	:	PUNCT
ajst-5345	1	9	2771	2771	NUM
ajst-5345	1	10	-	-	SYM
ajst-5345	1	11	3032	3032	NUM
ajst-5345	1	12	|	|	NOUN
ajst-5345	1	13	vol	vol	NOUN
ajst-5345	1	14	.	.	PROPN
ajst-5345	2	1	5	5	NUM
ajst-5345	2	2	,	,	PUNCT
ajst-5345	2	3	no	no	INTJ
ajst-5345	2	4	.	.	NOUN
ajst-5345	2	5	1	1	NUM
ajst-5345	2	6	,	,	PUNCT
ajst-5345	2	7	2023	2023	NUM
ajst-5345	2	8	46	46	NUM
ajst-5345	2	9	appearance	appearance	NOUN
ajst-5345	2	10	awared	aware	VERB
ajst-5345	2	11	detector	detector	NOUN
ajst-5345	2	12	for	for	ADP
ajst-5345	2	13	mot	mot	NOUN
ajst-5345	2	14	:	:	PUNCT
ajst-5345	2	15	an	an	DET
ajst-5345	2	16	enhanced	enhance	VERB
ajst-5345	2	17	reid	reid	PROPN
ajst-5345	2	18	branch	branch	NOUN
ajst-5345	2	19	for	for	ADP
ajst-5345	2	20	tracking	tracking	NOUN
ajst-5345	2	21	memorize	memorize	NOUN
ajst-5345	2	22	hongyu	hongyu	VERB
ajst-5345	2	23	chen	chen	PROPN
ajst-5345	2	24	college	college	PROPN
ajst-5345	2	25	of	of	ADP
ajst-5345	2	26	electrical	electrical	ADJ
ajst-5345	2	27	engineering	engineering	NOUN
ajst-5345	2	28	,	,	PUNCT
ajst-5345	2	29	southwest	southwest	PROPN
ajst-5345	2	30	minzu	minzu	PROPN
ajst-5345	2	31	university	university	PROPN
ajst-5345	2	32	,	,	PUNCT
ajst-5345	2	33	chengdu	chengdu	PROPN
ajst-5345	2	34	610206	610206	NUM
ajst-5345	2	35	,	,	PUNCT
ajst-5345	2	36	china	china	PROPN
ajst-5345	2	37	abstract	abstract	NOUN
ajst-5345	2	38	:	:	PUNCT
ajst-5345	2	39	the	the	DET
ajst-5345	2	40	traditional	traditional	ADJ
ajst-5345	2	41	bytetrack	bytetrack	NOUN
ajst-5345	2	42	approach	approach	NOUN
ajst-5345	2	43	to	to	ADP
ajst-5345	2	44	multi	multi	ADJ
ajst-5345	2	45	-	-	ADJ
ajst-5345	2	46	target	target	ADJ
ajst-5345	2	47	tracking	tracking	NOUN
ajst-5345	2	48	,	,	PUNCT
ajst-5345	2	49	focusing	focus	VERB
ajst-5345	2	50	on	on	ADP
ajst-5345	2	51	simple	simple	ADJ
ajst-5345	2	52	,	,	PUNCT
ajst-5345	2	53	effective	effective	ADJ
ajst-5345	2	54	algorithms	algorithm	NOUN
ajst-5345	2	55	.	.	PUNCT
ajst-5345	3	1	it	it	PRON
ajst-5345	3	2	performs	perform	VERB
ajst-5345	3	3	well	well	ADV
ajst-5345	3	4	in	in	ADP
ajst-5345	3	5	short	short	ADJ
ajst-5345	3	6	-	-	PUNCT
ajst-5345	3	7	term	term	NOUN
ajst-5345	3	8	multi	multi	ADJ
ajst-5345	3	9	-	-	ADJ
ajst-5345	3	10	target	target	ADJ
ajst-5345	3	11	tracking	tracking	NOUN
ajst-5345	3	12	tasks	task	NOUN
ajst-5345	3	13	with	with	ADP
ajst-5345	3	14	80.3	80.3	NUM
ajst-5345	3	15	mota	mota	NOUN
ajst-5345	3	16	,	,	PUNCT
ajst-5345	3	17	77.3	77.3	NUM
ajst-5345	3	18	idf1	idf1	NOUN
ajst-5345	3	19	and	and	CCONJ
ajst-5345	3	20	63.1	63.1	NUM
ajst-5345	3	21	hota	hota	NOUN
ajst-5345	3	22	on	on	ADP
ajst-5345	3	23	mot17	mot17	PROPN
ajst-5345	3	24	30	30	NUM
ajst-5345	3	25	fps	fps	NOUN
ajst-5345	3	26	and	and	CCONJ
ajst-5345	3	27	is	be	AUX
ajst-5345	3	28	currently	currently	ADV
ajst-5345	3	29	ranked	rank	VERB
ajst-5345	3	30	number	number	NOUN
ajst-5345	3	31	one	one	NUM
ajst-5345	3	32	in	in	ADP
ajst-5345	3	33	the	the	DET
ajst-5345	3	34	motchallenge	motchallenge	NOUN
ajst-5345	3	35	rankings	ranking	NOUN
ajst-5345	3	36	.	.	PUNCT
ajst-5345	4	1	but	but	CCONJ
ajst-5345	4	2	for	for	ADP
ajst-5345	4	3	situations	situation	NOUN
ajst-5345	4	4	where	where	SCONJ
ajst-5345	4	5	the	the	DET
ajst-5345	4	6	camera	camera	NOUN
ajst-5345	4	7	is	be	AUX
ajst-5345	4	8	moving	move	VERB
ajst-5345	4	9	,	,	PUNCT
ajst-5345	4	10	or	or	CCONJ
ajst-5345	4	11	for	for	ADP
ajst-5345	4	12	completing	complete	VERB
ajst-5345	4	13	target	target	NOUN
ajst-5345	4	14	recovery	recovery	NOUN
ajst-5345	4	15	tasks	task	NOUN
ajst-5345	4	16	after	after	ADP
ajst-5345	4	17	a	a	DET
ajst-5345	4	18	brief	brief	ADJ
ajst-5345	4	19	loss	loss	NOUN
ajst-5345	4	20	of	of	ADP
ajst-5345	4	21	position	position	NOUN
ajst-5345	4	22	and	and	CCONJ
ajst-5345	4	23	track	track	NOUN
ajst-5345	4	24	information	information	NOUN
ajst-5345	4	25	,	,	PUNCT
ajst-5345	4	26	conventional	conventional	ADJ
ajst-5345	4	27	mot	mot	NOUN
ajst-5345	4	28	modules	module	NOUN
ajst-5345	4	29	often	often	ADV
ajst-5345	4	30	do	do	AUX
ajst-5345	4	31	not	not	PART
ajst-5345	4	32	perform	perform	VERB
ajst-5345	4	33	as	as	ADV
ajst-5345	4	34	well	well	ADV
ajst-5345	4	35	as	as	SCONJ
ajst-5345	4	36	they	they	PRON
ajst-5345	4	37	should	should	AUX
ajst-5345	4	38	.	.	PUNCT
ajst-5345	5	1	to	to	PART
ajst-5345	5	2	make	make	VERB
ajst-5345	5	3	bytetrack	bytetrack	NOUN
ajst-5345	5	4	perform	perform	VERB
ajst-5345	5	5	better	well	ADV
ajst-5345	5	6	when	when	SCONJ
ajst-5345	5	7	the	the	DET
ajst-5345	5	8	motion	motion	NOUN
ajst-5345	5	9	background	background	NOUN
ajst-5345	5	10	pattern	pattern	NOUN
ajst-5345	5	11	is	be	AUX
ajst-5345	5	12	more	more	ADV
ajst-5345	5	13	complex	complex	ADJ
ajst-5345	5	14	,	,	PUNCT
ajst-5345	5	15	we	we	PRON
ajst-5345	5	16	used	use	VERB
ajst-5345	5	17	the	the	DET
ajst-5345	5	18	reidentification	reidentification	NOUN
ajst-5345	5	19	module	module	NOUN
ajst-5345	5	20	and	and	CCONJ
ajst-5345	5	21	added	add	VERB
ajst-5345	5	22	appearance	appearance	NOUN
ajst-5345	5	23	information	information	NOUN
ajst-5345	5	24	to	to	PART
ajst-5345	5	25	enhance	enhance	VERB
ajst-5345	5	26	bytetrack	bytetrack	NOUN
ajst-5345	5	27	’s	’s	PART
ajst-5345	5	28	mot	mot	NOUN
ajst-5345	5	29	process	process	NOUN
ajst-5345	5	30	.	.	PUNCT
ajst-5345	6	1	we	we	PRON
ajst-5345	6	2	verified	verify	VERB
ajst-5345	6	3	our	our	PRON
ajst-5345	6	4	improvement	improvement	NOUN
ajst-5345	6	5	on	on	ADP
ajst-5345	6	6	mot17	mot17	NOUN
ajst-5345	6	7	,	,	PUNCT
ajst-5345	6	8	which	which	PRON
ajst-5345	6	9	achieved	achieve	VERB
ajst-5345	6	10	higher	high	ADJ
ajst-5345	6	11	results	result	NOUN
ajst-5345	6	12	than	than	ADP
ajst-5345	6	13	the	the	DET
ajst-5345	6	14	original	original	NOUN
ajst-5345	6	15	on	on	ADP
ajst-5345	6	16	the	the	DET
ajst-5345	6	17	metrics	metric	NOUN
ajst-5345	6	18	precision	precision	NOUN
ajst-5345	6	19	and	and	CCONJ
ajst-5345	6	20	recall	recall	NOUN
ajst-5345	6	21	,	,	PUNCT
ajst-5345	6	22	and	and	CCONJ
ajst-5345	6	23	was	be	AUX
ajst-5345	6	24	also	also	ADV
ajst-5345	6	25	more	more	ADV
ajst-5345	6	26	suitable	suitable	ADJ
ajst-5345	6	27	for	for	ADP
ajst-5345	6	28	more	more	ADJ
ajst-5345	6	29	complex	complex	ADJ
ajst-5345	6	30	scenarios	scenario	NOUN
ajst-5345	6	31	.	.	PUNCT
ajst-5345	7	1	keywords	keyword	NOUN
ajst-5345	7	2	:	:	PUNCT
ajst-5345	7	3	bytetrack	bytetrack	NOUN
ajst-5345	7	4	approach	approach	NOUN
ajst-5345	7	5	,	,	PUNCT
ajst-5345	7	6	motchallenge	motchallenge	NOUN
ajst-5345	7	7	rankings	ranking	NOUN
ajst-5345	7	8	,	,	PUNCT
ajst-5345	7	9	multi	multi	ADJ
ajst-5345	7	10	-	-	ADJ
ajst-5345	7	11	target	target	ADJ
ajst-5345	7	12	tracking	tracking	NOUN
ajst-5345	7	13	,	,	PUNCT
ajst-5345	7	14	re	re	NOUN
ajst-5345	7	15	-	-	NOUN
ajst-5345	7	16	identification	identification	NOUN
ajst-5345	7	17	module	module	NOUN
ajst-5345	7	18	.	.	PUNCT
ajst-5345	8	1	1	1	X
ajst-5345	8	2	.	.	X
ajst-5345	8	3	introduction	introduction	NOUN
ajst-5345	8	4	bytetrack	bytetrack	NOUN
ajst-5345	8	5	[	[	X
ajst-5345	8	6	1	1	X
ajst-5345	8	7	]	]	PUNCT
ajst-5345	8	8	is	be	AUX
ajst-5345	8	9	a	a	DET
ajst-5345	8	10	state	state	NOUN
ajst-5345	8	11	of	of	ADP
ajst-5345	8	12	the	the	DET
ajst-5345	8	13	art	art	NOUN
ajst-5345	8	14	multi	multi	ADJ
ajst-5345	8	15	-	-	ADJ
ajst-5345	8	16	target	target	ADJ
ajst-5345	8	17	tracking	tracking	NOUN
ajst-5345	8	18	method	method	NOUN
ajst-5345	8	19	proposed	propose	VERB
ajst-5345	8	20	by	by	ADP
ajst-5345	8	21	yifu	yifu	PROPN
ajst-5345	8	22	zhang	zhang	PROPN
ajst-5345	8	23	from	from	ADP
ajst-5345	8	24	hust	hust	PROPN
ajst-5345	8	25	,	,	PUNCT
ajst-5345	8	26	based	base	VERB
ajst-5345	8	27	on	on	ADP
ajst-5345	8	28	tracking	tracking	NOUN
ajst-5345	8	29	-	-	PUNCT
ajst-5345	8	30	by	by	ADP
ajst-5345	8	31	-	-	PUNCT
ajst-5345	8	32	detection	detection	NOUN
ajst-5345	8	33	which	which	PRON
ajst-5345	8	34	is	be	AUX
ajst-5345	8	35	an	an	DET
ajst-5345	8	36	important	important	ADJ
ajst-5345	8	37	method	method	NOUN
ajst-5345	8	38	to	to	PART
ajst-5345	8	39	achieve	achieve	VERB
ajst-5345	8	40	multi	multi	ADJ
ajst-5345	8	41	-	-	ADJ
ajst-5345	8	42	target	target	ADJ
ajst-5345	8	43	tracking	tracking	NOUN
ajst-5345	8	44	.	.	PUNCT
ajst-5345	9	1	the	the	DET
ajst-5345	9	2	tracking	tracking	NOUN
ajst-5345	9	3	-	-	PUNCT
ajst-5345	9	4	by	by	ADP
ajst-5345	9	5	-	-	PUNCT
ajst-5345	9	6	detection	detection	NOUN
ajst-5345	9	7	method	method	NOUN
ajst-5345	9	8	uses	use	VERB
ajst-5345	9	9	the	the	DET
ajst-5345	9	10	similarity	similarity	NOUN
ajst-5345	9	11	(	(	PUNCT
ajst-5345	9	12	position	position	NOUN
ajst-5345	9	13	,	,	PUNCT
ajst-5345	9	14	appearance	appearance	NOUN
ajst-5345	9	15	,	,	PUNCT
ajst-5345	9	16	motion	motion	NOUN
ajst-5345	9	17	and	and	CCONJ
ajst-5345	9	18	other	other	ADJ
ajst-5345	9	19	information	information	NOUN
ajst-5345	9	20	)	)	PUNCT
ajst-5345	9	21	to	to	PART
ajst-5345	9	22	associate	associate	VERB
ajst-5345	9	23	the	the	DET
ajst-5345	9	24	detection	detection	NOUN
ajst-5345	9	25	frame	frame	NOUN
ajst-5345	9	26	between	between	ADP
ajst-5345	9	27	frames	frame	NOUN
ajst-5345	9	28	to	to	PART
ajst-5345	9	29	obtain	obtain	VERB
ajst-5345	9	30	the	the	DET
ajst-5345	9	31	tracking	tracking	NOUN
ajst-5345	9	32	trajectory	trajectory	NOUN
ajst-5345	9	33	.	.	PUNCT
ajst-5345	10	1	however	however	ADV
ajst-5345	10	2	due	due	ADP
ajst-5345	10	3	to	to	ADP
ajst-5345	10	4	the	the	DET
ajst-5345	10	5	imperfection	imperfection	NOUN
ajst-5345	10	6	of	of	ADP
ajst-5345	10	7	2d	2d	NOUN
ajst-5345	10	8	detectors	detector	NOUN
ajst-5345	10	9	in	in	ADP
ajst-5345	10	10	detection	detection	NOUN
ajst-5345	10	11	results	result	NOUN
ajst-5345	10	12	,	,	PUNCT
ajst-5345	10	13	in	in	ADP
ajst-5345	10	14	order	order	NOUN
ajst-5345	10	15	to	to	PART
ajst-5345	10	16	balance	balance	VERB
ajst-5345	10	17	the	the	DET
ajst-5345	10	18	true	true	ADJ
ajst-5345	10	19	and	and	CCONJ
ajst-5345	10	20	false	false	ADJ
ajst-5345	10	21	positive	positive	ADJ
ajst-5345	10	22	examples	example	NOUN
ajst-5345	10	23	to	to	PART
ajst-5345	10	24	be	be	AUX
ajst-5345	10	25	suitable	suitable	ADJ
ajst-5345	10	26	for	for	ADP
ajst-5345	10	27	complex	complex	ADJ
ajst-5345	10	28	actual	actual	ADJ
ajst-5345	10	29	scenarios	scenario	NOUN
ajst-5345	10	30	,	,	PUNCT
ajst-5345	10	31	most	most	ADJ
ajst-5345	10	32	of	of	ADP
ajst-5345	10	33	the	the	DET
ajst-5345	10	34	current	current	ADJ
ajst-5345	10	35	multi	multi	ADJ
ajst-5345	10	36	-	-	ADJ
ajst-5345	10	37	target	target	ADJ
ajst-5345	10	38	tracking	tracking	NOUN
ajst-5345	10	39	methods	method	NOUN
ajst-5345	10	40	will	will	AUX
ajst-5345	10	41	select	select	VERB
ajst-5345	10	42	a	a	DET
ajst-5345	10	43	threshold	threshold	NOUN
ajst-5345	10	44	,	,	PUNCT
ajst-5345	10	45	and	and	CCONJ
ajst-5345	10	46	only	only	ADV
ajst-5345	10	47	keep	keep	VERB
ajst-5345	10	48	the	the	DET
ajst-5345	10	49	detection	detection	NOUN
ajst-5345	10	50	results	result	VERB
ajst-5345	10	51	higher	high	ADJ
ajst-5345	10	52	than	than	ADP
ajst-5345	10	53	this	this	DET
ajst-5345	10	54	threshold	threshold	NOUN
ajst-5345	10	55	for	for	ADP
ajst-5345	10	56	correlation	correlation	NOUN
ajst-5345	10	57	,	,	PUNCT
ajst-5345	10	58	and	and	CCONJ
ajst-5345	10	59	then	then	ADV
ajst-5345	10	60	obtain	obtain	VERB
ajst-5345	10	61	the	the	DET
ajst-5345	10	62	tracking	tracking	NOUN
ajst-5345	10	63	results	result	NOUN
ajst-5345	10	64	.	.	PUNCT
ajst-5345	11	1	the	the	DET
ajst-5345	11	2	detection	detection	NOUN
ajst-5345	11	3	frame	frame	NOUN
ajst-5345	11	4	is	be	AUX
ajst-5345	11	5	discarded	discard	VERB
ajst-5345	11	6	directly	directly	ADV
ajst-5345	11	7	.	.	PUNCT
ajst-5345	12	1	bytetrack	bytetrack	NOUN
ajst-5345	12	2	,	,	PUNCT
ajst-5345	12	3	on	on	ADP
ajst-5345	12	4	the	the	DET
ajst-5345	12	5	other	other	ADJ
ajst-5345	12	6	hand	hand	NOUN
ajst-5345	12	7	,	,	PUNCT
ajst-5345	12	8	takes	take	VERB
ajst-5345	12	9	full	full	ADJ
ajst-5345	12	10	advantage	advantage	NOUN
ajst-5345	12	11	of	of	ADP
ajst-5345	12	12	low	low	ADJ
ajst-5345	12	13	-	-	PUNCT
ajst-5345	12	14	scoring	score	VERB
ajst-5345	12	15	detection	detection	NOUN
ajst-5345	12	16	results	result	NOUN
ajst-5345	12	17	.	.	PUNCT
ajst-5345	13	1	using	use	VERB
ajst-5345	13	2	the	the	DET
ajst-5345	13	3	similarity	similarity	NOUN
ajst-5345	13	4	between	between	ADP
ajst-5345	13	5	the	the	DET
ajst-5345	13	6	detection	detection	NOUN
ajst-5345	13	7	frame	frame	NOUN
ajst-5345	13	8	and	and	CCONJ
ajst-5345	13	9	the	the	DET
ajst-5345	13	10	tracking	tracking	NOUN
ajst-5345	13	11	trajectory	trajectory	NOUN
ajst-5345	13	12	,	,	PUNCT
ajst-5345	13	13	while	while	SCONJ
ajst-5345	13	14	retaining	retain	VERB
ajst-5345	13	15	the	the	DET
ajst-5345	13	16	high	high	ADJ
ajst-5345	13	17	-	-	PUNCT
ajst-5345	13	18	scoring	score	VERB
ajst-5345	13	19	detection	detection	NOUN
ajst-5345	13	20	results	result	NOUN
ajst-5345	13	21	,	,	PUNCT
ajst-5345	13	22	the	the	DET
ajst-5345	13	23	background	background	NOUN
ajst-5345	13	24	is	be	AUX
ajst-5345	13	25	removed	remove	VERB
ajst-5345	13	26	from	from	ADP
ajst-5345	13	27	the	the	DET
ajst-5345	13	28	low	low	ADJ
ajst-5345	13	29	-	-	PUNCT
ajst-5345	13	30	scoring	score	VERB
ajst-5345	13	31	detection	detection	NOUN
ajst-5345	13	32	results	result	NOUN
ajst-5345	13	33	,	,	PUNCT
ajst-5345	13	34	and	and	CCONJ
ajst-5345	13	35	the	the	DET
ajst-5345	13	36	real	real	ADJ
ajst-5345	13	37	objects	object	NOUN
ajst-5345	13	38	(	(	PUNCT
ajst-5345	13	39	difficult	difficult	ADJ
ajst-5345	13	40	samples	sample	NOUN
ajst-5345	13	41	such	such	ADJ
ajst-5345	13	42	as	as	ADP
ajst-5345	13	43	occlusion	occlusion	NOUN
ajst-5345	13	44	,	,	PUNCT
ajst-5345	13	45	blur	blur	NOUN
ajst-5345	13	46	,	,	PUNCT
ajst-5345	13	47	etc	etc	X
ajst-5345	13	48	.	.	X
ajst-5345	13	49	)	)	PUNCT
ajst-5345	13	50	are	be	AUX
ajst-5345	13	51	mined	mine	VERB
ajst-5345	13	52	.	.	PUNCT
ajst-5345	14	1	specifically	specifically	ADV
ajst-5345	14	2	,	,	PUNCT
ajst-5345	14	3	byte	byte	NOUN
ajst-5345	14	4	will	will	AUX
ajst-5345	14	5	each	each	DET
ajst-5345	14	6	detection	detection	NOUN
ajst-5345	14	7	box	box	NOUN
ajst-5345	14	8	is	be	AUX
ajst-5345	14	9	divided	divide	VERB
ajst-5345	14	10	into	into	ADP
ajst-5345	14	11	two	two	NUM
ajst-5345	14	12	categories	category	NOUN
ajst-5345	14	13	:	:	PUNCT
ajst-5345	14	14	high	high	ADJ
ajst-5345	14	15	-	-	PUNCT
ajst-5345	14	16	scoring	score	VERB
ajst-5345	14	17	box	box	NOUN
ajst-5345	14	18	and	and	CCONJ
ajst-5345	14	19	low	low	ADJ
ajst-5345	14	20	-	-	PUNCT
ajst-5345	14	21	scoring	scoring	NOUN
ajst-5345	14	22	box	box	NOUN
ajst-5345	14	23	according	accord	VERB
ajst-5345	14	24	to	to	ADP
ajst-5345	14	25	the	the	DET
ajst-5345	14	26	score	score	NOUN
ajst-5345	14	27	,	,	PUNCT
ajst-5345	14	28	and	and	CCONJ
ajst-5345	14	29	a	a	DET
ajst-5345	14	30	total	total	NOUN
ajst-5345	14	31	of	of	ADP
ajst-5345	14	32	two	two	NUM
ajst-5345	14	33	matches	match	NOUN
ajst-5345	14	34	are	be	AUX
ajst-5345	14	35	performed	perform	VERB
ajst-5345	14	36	.	.	PUNCT
ajst-5345	15	1	for	for	ADP
ajst-5345	15	2	the	the	DET
ajst-5345	15	3	first	first	ADJ
ajst-5345	15	4	time	time	NOUN
ajst-5345	15	5	,	,	PUNCT
ajst-5345	15	6	the	the	DET
ajst-5345	15	7	high	high	ADJ
ajst-5345	15	8	score	score	NOUN
ajst-5345	15	9	box	box	PROPN
ajst-5345	15	10	is	be	AUX
ajst-5345	15	11	used	use	VERB
ajst-5345	15	12	to	to	PART
ajst-5345	15	13	match	match	VERB
ajst-5345	15	14	the	the	DET
ajst-5345	15	15	previous	previous	ADJ
ajst-5345	15	16	tracking	tracking	NOUN
ajst-5345	15	17	trajectory	trajectory	NOUN
ajst-5345	15	18	.	.	PUNCT
ajst-5345	16	1	the	the	DET
ajst-5345	16	2	second	second	ADJ
ajst-5345	16	3	time	time	NOUN
ajst-5345	16	4	uses	use	VERB
ajst-5345	16	5	the	the	DET
ajst-5345	16	6	low	low	ADJ
ajst-5345	16	7	-	-	PUNCT
ajst-5345	16	8	scoring	scoring	NOUN
ajst-5345	16	9	box	box	NOUN
ajst-5345	16	10	and	and	CCONJ
ajst-5345	16	11	the	the	DET
ajst-5345	16	12	tracked	track	VERB
ajst-5345	16	13	trajectories	trajectory	NOUN
ajst-5345	16	14	that	that	PRON
ajst-5345	16	15	are	be	AUX
ajst-5345	16	16	not	not	PART
ajst-5345	16	17	matched	match	VERB
ajst-5345	16	18	to	to	ADP
ajst-5345	16	19	the	the	DET
ajst-5345	16	20	high	high	ADJ
ajst-5345	16	21	-	-	PUNCT
ajst-5345	16	22	scoring	scoring	NOUN
ajst-5345	16	23	box	box	NOUN
ajst-5345	16	24	in	in	ADP
ajst-5345	16	25	the	the	DET
ajst-5345	16	26	first	first	ADJ
ajst-5345	16	27	time	time	NOUN
ajst-5345	16	28	(	(	PUNCT
ajst-5345	16	29	for	for	ADP
ajst-5345	16	30	example	example	NOUN
ajst-5345	16	31	,	,	PUNCT
ajst-5345	16	32	objects	object	VERB
ajst-5345	16	33	that	that	PRON
ajst-5345	16	34	are	be	AUX
ajst-5345	16	35	severely	severely	ADV
ajst-5345	16	36	occluded	occlude	VERB
ajst-5345	16	37	in	in	ADP
ajst-5345	16	38	the	the	DET
ajst-5345	16	39	current	current	ADJ
ajst-5345	16	40	frame	frame	NOUN
ajst-5345	16	41	and	and	CCONJ
ajst-5345	16	42	cause	cause	VERB
ajst-5345	16	43	the	the	DET
ajst-5345	16	44	score	score	NOUN
ajst-5345	16	45	to	to	PART
ajst-5345	16	46	drop	drop	VERB
ajst-5345	16	47	)	)	PUNCT
ajst-5345	16	48	for	for	ADP
ajst-5345	16	49	matching	match	VERB
ajst-5345	16	50	.	.	PUNCT
ajst-5345	17	1	for	for	ADP
ajst-5345	17	2	the	the	DET
ajst-5345	17	3	detection	detection	NOUN
ajst-5345	17	4	box	box	NOUN
ajst-5345	17	5	that	that	PRON
ajst-5345	17	6	does	do	AUX
ajst-5345	17	7	not	not	PART
ajst-5345	17	8	match	match	VERB
ajst-5345	17	9	the	the	DET
ajst-5345	17	10	tracking	tracking	NOUN
ajst-5345	17	11	track	track	NOUN
ajst-5345	17	12	and	and	CCONJ
ajst-5345	17	13	the	the	DET
ajst-5345	17	14	score	score	NOUN
ajst-5345	17	15	is	be	AUX
ajst-5345	17	16	high	high	ADJ
ajst-5345	17	17	enough	enough	ADV
ajst-5345	17	18	,	,	PUNCT
ajst-5345	17	19	a	a	DET
ajst-5345	17	20	new	new	ADJ
ajst-5345	17	21	tracking	tracking	NOUN
ajst-5345	17	22	track	track	NOUN
ajst-5345	17	23	is	be	AUX
ajst-5345	17	24	created	create	VERB
ajst-5345	17	25	for	for	ADP
ajst-5345	17	26	it	it	PRON
ajst-5345	17	27	.	.	PUNCT
ajst-5345	18	1	for	for	ADP
ajst-5345	18	2	the	the	DET
ajst-5345	18	3	tracking	tracking	NOUN
ajst-5345	18	4	trajectories	trajectory	NOUN
ajst-5345	18	5	that	that	PRON
ajst-5345	18	6	do	do	AUX
ajst-5345	18	7	not	not	PART
ajst-5345	18	8	match	match	VERB
ajst-5345	18	9	the	the	DET
ajst-5345	18	10	detection	detection	NOUN
ajst-5345	18	11	frame	frame	NOUN
ajst-5345	18	12	,	,	PUNCT
ajst-5345	18	13	keep	keep	VERB
ajst-5345	18	14	30	30	NUM
ajst-5345	18	15	frames	frame	NOUN
ajst-5345	18	16	,	,	PUNCT
ajst-5345	18	17	and	and	CCONJ
ajst-5345	18	18	perform	perform	VERB
ajst-5345	18	19	matching	match	VERB
ajst-5345	18	20	when	when	SCONJ
ajst-5345	18	21	they	they	PRON
ajst-5345	18	22	appear	appear	VERB
ajst-5345	18	23	again	again	ADV
ajst-5345	18	24	,	,	PUNCT
ajst-5345	18	25	thereby	thereby	ADV
ajst-5345	18	26	reducing	reduce	VERB
ajst-5345	18	27	missed	miss	VERB
ajst-5345	18	28	detection	detection	NOUN
ajst-5345	18	29	and	and	CCONJ
ajst-5345	18	30	improving	improve	VERB
ajst-5345	18	31	the	the	DET
ajst-5345	18	32	continuity	continuity	NOUN
ajst-5345	18	33	of	of	ADP
ajst-5345	18	34	the	the	DET
ajst-5345	18	35	trajectory	trajectory	NOUN
ajst-5345	18	36	.	.	PUNCT
ajst-5345	19	1	re	re	NOUN
ajst-5345	20	1	-	-	NOUN
ajst-5345	20	2	identification	identification	NOUN
ajst-5345	20	3	(	(	PUNCT
ajst-5345	20	4	reid	reid	NOUN
ajst-5345	20	5	)	)	PUNCT
ajst-5345	20	6	is	be	AUX
ajst-5345	20	7	a	a	DET
ajst-5345	20	8	technique	technique	NOUN
ajst-5345	20	9	designed	design	VERB
ajst-5345	20	10	to	to	PART
ajst-5345	20	11	search	search	VERB
ajst-5345	20	12	for	for	ADP
ajst-5345	20	13	objects	object	NOUN
ajst-5345	20	14	in	in	ADP
ajst-5345	20	15	non	non	ADJ
ajst-5345	20	16	-	-	ADJ
ajst-5345	20	17	overlapping	overlapping	ADJ
ajst-5345	20	18	camera	camera	NOUN
ajst-5345	20	19	views	view	NOUN
ajst-5345	20	20	at	at	ADP
ajst-5345	20	21	different	different	ADJ
ajst-5345	20	22	locations	location	NOUN
ajst-5345	20	23	by	by	ADP
ajst-5345	20	24	matching	match	VERB
ajst-5345	20	25	query	query	NOUN
ajst-5345	20	26	images	image	NOUN
ajst-5345	20	27	.	.	PUNCT
ajst-5345	21	1	reid	reid	PROPN
ajst-5345	21	2	can	can	AUX
ajst-5345	21	3	partially	partially	ADV
ajst-5345	21	4	replace	replace	VERB
ajst-5345	21	5	face	face	NOUN
ajst-5345	21	6	recognition	recognition	NOUN
ajst-5345	21	7	to	to	PART
ajst-5345	21	8	find	find	VERB
ajst-5345	21	9	target	target	NOUN
ajst-5345	21	10	objects	object	NOUN
ajst-5345	21	11	in	in	ADP
ajst-5345	21	12	video	video	NOUN
ajst-5345	21	13	sequences	sequence	NOUN
ajst-5345	21	14	,	,	PUNCT
ajst-5345	21	15	and	and	CCONJ
ajst-5345	21	16	can	can	AUX
ajst-5345	21	17	be	be	AUX
ajst-5345	21	18	widely	widely	ADV
ajst-5345	21	19	used	use	VERB
ajst-5345	21	20	in	in	ADP
ajst-5345	21	21	security	security	NOUN
ajst-5345	21	22	,	,	PUNCT
ajst-5345	21	23	personal	personal	ADJ
ajst-5345	21	24	positioning	positioning	NOUN
ajst-5345	21	25	,	,	PUNCT
ajst-5345	21	26	recommendation	recommendation	NOUN
ajst-5345	21	27	systems	system	NOUN
ajst-5345	21	28	in	in	ADP
ajst-5345	21	29	physical	physical	ADJ
ajst-5345	21	30	shopping	shopping	NOUN
ajst-5345	21	31	malls	mall	NOUN
ajst-5345	21	32	and	and	CCONJ
ajst-5345	21	33	other	other	ADJ
ajst-5345	21	34	fields	field	NOUN
ajst-5345	21	35	.	.	PUNCT
ajst-5345	22	1	the	the	DET
ajst-5345	22	2	reid	reid	PROPN
ajst-5345	22	3	task	task	PROPN
ajst-5345	22	4	was	be	AUX
ajst-5345	22	5	first	first	ADV
ajst-5345	22	6	proposed	propose	VERB
ajst-5345	22	7	at	at	ADP
ajst-5345	22	8	the	the	DET
ajst-5345	22	9	2006	2006	NUM
ajst-5345	22	10	cvpr	cvpr	NOUN
ajst-5345	22	11	conference	conference	NOUN
ajst-5345	22	12	.	.	PUNCT
ajst-5345	23	1	in	in	ADP
ajst-5345	23	2	2007	2007	NUM
ajst-5345	23	3	,	,	PUNCT
ajst-5345	23	4	the	the	DET
ajst-5345	23	5	first	first	PROPN
ajst-5345	23	6	reid	reid	PROPN
ajst-5345	23	7	dataset	dataset	VERB
ajst-5345	23	8	viper	viper	NOUN
ajst-5345	23	9	[	[	X
ajst-5345	23	10	2	2	NUM
ajst-5345	23	11	]	]	PUNCT
ajst-5345	23	12	,	,	PUNCT
ajst-5345	23	13	[	[	X
ajst-5345	23	14	3	3	NUM
ajst-5345	23	15	]	]	PUNCT
ajst-5345	23	16	was	be	AUX
ajst-5345	23	17	released	release	VERB
ajst-5345	23	18	,	,	PUNCT
ajst-5345	23	19	and	and	CCONJ
ajst-5345	23	20	then	then	ADV
ajst-5345	23	21	the	the	DET
ajst-5345	23	22	datasets	dataset	NOUN
ajst-5345	23	23	under	under	ADP
ajst-5345	23	24	different	different	ADJ
ajst-5345	23	25	scenarios	scenario	NOUN
ajst-5345	23	26	were	be	AUX
ajst-5345	23	27	open	open	ADJ
ajst-5345	23	28	sourced	source	VERB
ajst-5345	23	29	,	,	PUNCT
ajst-5345	23	30	which	which	PRON
ajst-5345	23	31	further	far	ADV
ajst-5345	23	32	promoted	promote	VERB
ajst-5345	23	33	the	the	DET
ajst-5345	23	34	development	development	NOUN
ajst-5345	23	35	of	of	ADP
ajst-5345	23	36	reid	reid	PROPN
ajst-5345	23	37	.	.	PUNCT
ajst-5345	24	1	around	around	ADP
ajst-5345	24	2	2015	2015	NUM
ajst-5345	24	3	,	,	PUNCT
ajst-5345	24	4	most	most	ADJ
ajst-5345	24	5	of	of	ADP
ajst-5345	24	6	the	the	DET
ajst-5345	24	7	research	research	NOUN
ajst-5345	24	8	on	on	ADP
ajst-5345	24	9	reid	reid	PROPN
ajst-5345	24	10	was	be	AUX
ajst-5345	24	11	based	base	VERB
ajst-5345	24	12	on	on	ADP
ajst-5345	24	13	features	feature	NOUN
ajst-5345	24	14	,	,	PUNCT
ajst-5345	24	15	such	such	ADJ
ajst-5345	24	16	as	as	ADP
ajst-5345	24	17	color	color	NOUN
ajst-5345	24	18	,	,	PUNCT
ajst-5345	24	19	hog	hog	NOUN
ajst-5345	24	20	features	feature	NOUN
ajst-5345	24	21	,	,	PUNCT
ajst-5345	24	22	etc	etc	X
ajst-5345	25	1	[	[	X
ajst-5345	25	2	4	4	NUM
ajst-5345	25	3	]	]	PUNCT
ajst-5345	25	4	,	,	PUNCT
ajst-5345	25	5	[	[	X
ajst-5345	25	6	5	5	NUM
ajst-5345	25	7	]	]	PUNCT
ajst-5345	25	8	.	.	PUNCT
ajst-5345	26	1	the	the	DET
ajst-5345	26	2	subsequent	subsequent	ADJ
ajst-5345	26	3	metric	metric	ADJ
ajst-5345	26	4	learning	learning	NOUN
ajst-5345	26	5	was	be	AUX
ajst-5345	26	6	to	to	PART
ajst-5345	26	7	find	find	VERB
ajst-5345	26	8	the	the	DET
ajst-5345	26	9	best	good	ADJ
ajst-5345	26	10	approximation	approximation	NOUN
ajst-5345	26	11	between	between	ADP
ajst-5345	26	12	features	feature	NOUN
ajst-5345	26	13	,	,	PUNCT
ajst-5345	26	14	but	but	CCONJ
ajst-5345	26	15	this	this	DET
ajst-5345	26	16	method	method	NOUN
ajst-5345	26	17	is	be	AUX
ajst-5345	26	18	suitable	suitable	ADJ
ajst-5345	26	19	for	for	ADP
ajst-5345	26	20	complex	complex	ADJ
ajst-5345	26	21	scene	scene	NOUN
ajst-5345	26	22	data	datum	NOUN
ajst-5345	26	23	,	,	PUNCT
ajst-5345	26	24	especially	especially	ADV
ajst-5345	26	25	when	when	SCONJ
ajst-5345	26	26	lighting	lighting	NOUN
ajst-5345	26	27	,	,	PUNCT
ajst-5345	26	28	occlusion	occlusion	NOUN
ajst-5345	26	29	,	,	PUNCT
ajst-5345	26	30	resolution	resolution	NOUN
ajst-5345	26	31	,	,	PUNCT
ajst-5345	26	32	human	human	NOUN
ajst-5345	26	33	pose	pose	NOUN
ajst-5345	26	34	,	,	PUNCT
ajst-5345	26	35	perspective	perspective	NOUN
ajst-5345	26	36	,	,	PUNCT
ajst-5345	26	37	clothing	clothing	NOUN
ajst-5345	26	38	,	,	PUNCT
ajst-5345	26	39	and	and	CCONJ
ajst-5345	26	40	background	background	NOUN
ajst-5345	26	41	with	with	ADP
ajst-5345	26	42	large	large	ADJ
ajst-5345	26	43	changes	change	NOUN
ajst-5345	26	44	.	.	PUNCT
ajst-5345	27	1	with	with	ADP
ajst-5345	27	2	the	the	DET
ajst-5345	27	3	further	further	ADJ
ajst-5345	27	4	development	development	NOUN
ajst-5345	27	5	and	and	CCONJ
ajst-5345	27	6	application	application	NOUN
ajst-5345	27	7	of	of	ADP
ajst-5345	27	8	deep	deep	ADJ
ajst-5345	27	9	learning	learning	NOUN
ajst-5345	27	10	,	,	PUNCT
ajst-5345	27	11	deep	deep	ADJ
ajst-5345	27	12	learning	learning	NOUN
ajst-5345	27	13	has	have	AUX
ajst-5345	27	14	been	be	AUX
ajst-5345	27	15	gradually	gradually	ADV
ajst-5345	27	16	applied	apply	VERB
ajst-5345	27	17	to	to	ADP
ajst-5345	27	18	reid	reid	PROPN
ajst-5345	27	19	and	and	CCONJ
ajst-5345	27	20	achieved	achieve	VERB
ajst-5345	27	21	a	a	DET
ajst-5345	27	22	higher	high	ADJ
ajst-5345	27	23	recognition	recognition	NOUN
ajst-5345	27	24	performance	performance	NOUN
ajst-5345	27	25	than	than	ADP
ajst-5345	27	26	previous	previous	ADJ
ajst-5345	27	27	methods	method	NOUN
ajst-5345	27	28	.	.	PUNCT
ajst-5345	28	1	[	[	X
ajst-5345	28	2	6]–[8	6]–[8	X
ajst-5345	28	3	]	]	X
ajst-5345	28	4	.	.	PUNCT
ajst-5345	29	1	however	however	ADV
ajst-5345	29	2	in	in	ADP
ajst-5345	29	3	the	the	DET
ajst-5345	29	4	data	data	PROPN
ajst-5345	29	5	association	association	PROPN
ajst-5345	29	6	part	part	NOUN
ajst-5345	29	7	,	,	PUNCT
ajst-5345	29	8	bytetrack	bytetrack	NOUN
ajst-5345	29	9	only	only	ADV
ajst-5345	29	10	uses	use	VERB
ajst-5345	29	11	kalman	kalman	NOUN
ajst-5345	29	12	filter	filter	NOUN
ajst-5345	29	13	to	to	PART
ajst-5345	29	14	predict	predict	VERB
ajst-5345	29	15	the	the	DET
ajst-5345	29	16	position	position	NOUN
ajst-5345	29	17	of	of	ADP
ajst-5345	29	18	the	the	DET
ajst-5345	29	19	tracking	tracking	NOUN
ajst-5345	29	20	trajectory	trajectory	NOUN
ajst-5345	29	21	of	of	ADP
ajst-5345	29	22	the	the	DET
ajst-5345	29	23	current	current	ADJ
ajst-5345	29	24	frame	frame	NOUN
ajst-5345	29	25	in	in	ADP
ajst-5345	29	26	the	the	DET
ajst-5345	29	27	next	next	ADJ
ajst-5345	29	28	frame	frame	NOUN
ajst-5345	29	29	,	,	PUNCT
ajst-5345	29	30	the	the	DET
ajst-5345	29	31	iou	iou	NOUN
ajst-5345	29	32	between	between	ADP
ajst-5345	29	33	the	the	DET
ajst-5345	29	34	predicted	predict	VERB
ajst-5345	29	35	frame	frame	NOUN
ajst-5345	29	36	and	and	CCONJ
ajst-5345	29	37	the	the	DET
ajst-5345	29	38	actual	actual	ADJ
ajst-5345	29	39	detection	detection	NOUN
ajst-5345	29	40	frame	frame	NOUN
ajst-5345	29	41	is	be	AUX
ajst-5345	29	42	used	use	VERB
ajst-5345	29	43	as	as	ADP
ajst-5345	29	44	the	the	DET
ajst-5345	29	45	similarity	similarity	NOUN
ajst-5345	29	46	between	between	ADP
ajst-5345	29	47	two	two	NUM
ajst-5345	29	48	matches	match	NOUN
ajst-5345	29	49	,	,	PUNCT
ajst-5345	29	50	and	and	CCONJ
ajst-5345	29	51	then	then	ADV
ajst-5345	29	52	passes	pass	VERB
ajst-5345	29	53	to	to	ADP
ajst-5345	29	54	hungary	hungary	PROPN
ajst-5345	29	55	algorithm	algorithm	NOUN
ajst-5345	29	56	to	to	PART
ajst-5345	29	57	complete	complete	VERB
ajst-5345	29	58	the	the	DET
ajst-5345	29	59	match	match	NOUN
ajst-5345	29	60	.	.	PUNCT
ajst-5345	30	1	but	but	CCONJ
ajst-5345	30	2	when	when	SCONJ
ajst-5345	30	3	the	the	DET
ajst-5345	30	4	motion	motion	NOUN
ajst-5345	30	5	pattern	pattern	NOUN
ajst-5345	30	6	is	be	AUX
ajst-5345	30	7	considered	consider	VERB
ajst-5345	30	8	complex	complex	ADJ
ajst-5345	30	9	,	,	PUNCT
ajst-5345	30	10	the	the	DET
ajst-5345	30	11	probability	probability	NOUN
ajst-5345	30	12	of	of	ADP
ajst-5345	30	13	misses	miss	VERB
ajst-5345	30	14	detection	detection	NOUN
ajst-5345	30	15	and	and	CCONJ
ajst-5345	30	16	track	track	NOUN
ajst-5345	30	17	breaks	break	NOUN
ajst-5345	30	18	will	will	AUX
ajst-5345	30	19	be	be	AUX
ajst-5345	30	20	higher	high	ADJ
ajst-5345	30	21	.	.	PUNCT
ajst-5345	31	1	at	at	ADP
ajst-5345	31	2	the	the	DET
ajst-5345	31	3	same	same	ADJ
ajst-5345	31	4	time	time	NOUN
ajst-5345	31	5	,	,	PUNCT
ajst-5345	31	6	we	we	PRON
ajst-5345	31	7	found	find	VERB
ajst-5345	31	8	that	that	SCONJ
ajst-5345	31	9	when	when	SCONJ
ajst-5345	31	10	there	there	PRON
ajst-5345	31	11	is	be	VERB
ajst-5345	31	12	severe	severe	ADJ
ajst-5345	31	13	occlusion	occlusion	NOUN
ajst-5345	31	14	between	between	ADP
ajst-5345	31	15	two	two	NUM
ajst-5345	31	16	objects	object	NOUN
ajst-5345	31	17	or	or	CCONJ
ajst-5345	31	18	some	some	DET
ajst-5345	31	19	background	background	NOUN
ajst-5345	31	20	(	(	PUNCT
ajst-5345	31	21	such	such	ADJ
ajst-5345	31	22	as	as	ADP
ajst-5345	31	23	doors	door	NOUN
ajst-5345	31	24	,	,	PUNCT
ajst-5345	31	25	trees	tree	NOUN
ajst-5345	31	26	,	,	PUNCT
ajst-5345	31	27	etc	etc	X
ajst-5345	31	28	.	.	X
ajst-5345	31	29	)	)	PUNCT
ajst-5345	31	30	is	be	AUX
ajst-5345	31	31	detected	detect	VERB
ajst-5345	31	32	again	again	ADV
ajst-5345	31	33	after	after	ADP
ajst-5345	31	34	being	be	AUX
ajst-5345	31	35	occluded	occlude	VERB
ajst-5345	31	36	for	for	ADP
ajst-5345	31	37	several	several	ADJ
ajst-5345	31	38	frames	frame	NOUN
ajst-5345	31	39	,	,	PUNCT
ajst-5345	31	40	the	the	DET
ajst-5345	31	41	i	i	PROPN
ajst-5345	31	42	d	d	PROPN
ajst-5345	31	43	of	of	ADP
ajst-5345	31	44	the	the	DET
ajst-5345	31	45	detected	detect	VERB
ajst-5345	31	46	object	object	NOUN
ajst-5345	31	47	will	will	AUX
ajst-5345	31	48	change	change	VERB
ajst-5345	31	49	incorrectly	incorrectly	ADV
ajst-5345	31	50	.	.	PUNCT
ajst-5345	32	1	and	and	CCONJ
ajst-5345	32	2	because	because	SCONJ
ajst-5345	32	3	the	the	DET
ajst-5345	32	4	appearance	appearance	NOUN
ajst-5345	32	5	features	feature	NOUN
ajst-5345	32	6	are	be	AUX
ajst-5345	32	7	not	not	PART
ajst-5345	32	8	used	use	VERB
ajst-5345	32	9	,	,	PUNCT
ajst-5345	32	10	the	the	DET
ajst-5345	32	11	tracking	tracking	NOUN
ajst-5345	32	12	effect	effect	NOUN
ajst-5345	32	13	is	be	AUX
ajst-5345	32	14	very	very	ADV
ajst-5345	32	15	dependent	dependent	ADJ
ajst-5345	32	16	on	on	ADP
ajst-5345	32	17	the	the	DET
ajst-5345	32	18	detection	detection	NOUN
ajst-5345	32	19	effect	effect	NOUN
ajst-5345	32	20	.	.	PUNCT
ajst-5345	33	1	when	when	SCONJ
ajst-5345	33	2	the	the	DET
ajst-5345	33	3	detection	detection	NOUN
ajst-5345	33	4	effect	effect	NOUN
ajst-5345	33	5	is	be	AUX
ajst-5345	33	6	not	not	PART
ajst-5345	33	7	good	good	ADJ
ajst-5345	33	8	,	,	PUNCT
ajst-5345	33	9	it	it	PRON
ajst-5345	33	10	will	will	AUX
ajst-5345	33	11	seriously	seriously	ADV
ajst-5345	33	12	affect	affect	VERB
ajst-5345	33	13	the	the	DET
ajst-5345	33	14	tracking	track	VERB
ajst-5345	33	15	effect.2	effect.2	PROPN
ajst-5345	33	16	.	.	PUNCT
ajst-5345	34	1	data	datum	NOUN
ajst-5345	34	2	processing	processing	NOUN
ajst-5345	34	3	and	and	CCONJ
ajst-5345	34	4	model	model	NOUN
ajst-5345	34	5	design	design	NOUN
ajst-5345	34	6	2	2	NUM
ajst-5345	34	7	.	.	PUNCT
ajst-5345	34	8	data	datum	NOUN
ajst-5345	34	9	processing	processing	NOUN
ajst-5345	34	10	and	and	CCONJ
ajst-5345	34	11	model	model	NOUN
ajst-5345	34	12	design	design	NOUN
ajst-5345	34	13	2.1	2.1	NUM
ajst-5345	34	14	.	.	PUNCT
ajst-5345	35	1	data	datum	NOUN
ajst-5345	35	2	processing	processing	NOUN
ajst-5345	35	3	we	we	PRON
ajst-5345	35	4	evaluate	evaluate	VERB
ajst-5345	35	5	our	our	PRON
ajst-5345	35	6	proposed	propose	VERB
ajst-5345	35	7	improvements	improvement	NOUN
ajst-5345	35	8	on	on	ADP
ajst-5345	35	9	mot17	mot17	PROPN
ajst-5345	36	1	[	[	X
ajst-5345	36	2	9	9	NUM
ajst-5345	36	3	]	]	PUNCT
ajst-5345	36	4	and	and	CCONJ
ajst-5345	36	5	mot20	mot20	X
ajst-5345	36	6	[	[	X
ajst-5345	36	7	10	10	NUM
ajst-5345	36	8	]	]	PUNCT
ajst-5345	36	9	.	.	PUNCT
ajst-5345	37	1	mot17	mot17	PROPN
ajst-5345	37	2	and	and	CCONJ
ajst-5345	37	3	mot20	mot20	PROPN
ajst-5345	37	4	are	be	AUX
ajst-5345	37	5	two	two	NUM
ajst-5345	37	6	current	current	ADJ
ajst-5345	37	7	mainstream	mainstream	NOUN
ajst-5345	37	8	datasets	dataset	NOUN
ajst-5345	37	9	in	in	ADP
ajst-5345	37	10	mot	mot	NOUN
ajst-5345	37	11	task	task	NOUN
ajst-5345	37	12	,	,	PUNCT
ajst-5345	37	13	which	which	PRON
ajst-5345	37	14	mainly	mainly	ADV
ajst-5345	37	15	focus	focus	VERB
ajst-5345	37	16	on	on	ADP
ajst-5345	37	17	the	the	DET
ajst-5345	37	18	task	task	NOUN
ajst-5345	37	19	of	of	ADP
ajst-5345	37	20	pedestrian	pedestrian	NOUN
ajst-5345	37	21	tracking	tracking	NOUN
ajst-5345	37	22	in	in	ADP
ajst-5345	37	23	dense	dense	ADJ
ajst-5345	37	24	scenes	scene	NOUN
ajst-5345	37	25	.	.	PUNCT
ajst-5345	38	1	the	the	DET
ajst-5345	38	2	mot	mot	NOUN
ajst-5345	38	3	17	17	NUM
ajst-5345	38	4	dataset	dataset	NOUN
ajst-5345	38	5	contains	contain	VERB
ajst-5345	38	6	7	7	NUM
ajst-5345	38	7	training	training	NOUN
ajst-5345	38	8	set	set	VERB
ajst-5345	38	9	videos	video	NOUN
ajst-5345	38	10	and	and	CCONJ
ajst-5345	38	11	7	7	NUM
ajst-5345	38	12	test	test	NOUN
ajst-5345	38	13	set	set	VERB
ajst-5345	38	14	videos	video	NOUN
ajst-5345	38	15	.	.	PUNCT
ajst-5345	39	1	while	while	SCONJ
ajst-5345	39	2	in	in	ADP
ajst-5345	39	3	mot20	mot20	PROPN
ajst-5345	39	4	dataset	dataset	NOUN
ajst-5345	39	5	,	,	PUNCT
ajst-5345	39	6	8	8	NUM
ajst-5345	39	7	new	new	ADJ
ajst-5345	39	8	dense	dense	ADJ
ajst-5345	39	9	crowd	crowd	NOUN
ajst-5345	39	10	sequences	sequence	NOUN
ajst-5345	39	11	are	be	AUX
ajst-5345	39	12	47	47	NUM
ajst-5345	39	13	extracted	extract	VERB
ajst-5345	39	14	from	from	ADP
ajst-5345	39	15	3	3	NUM
ajst-5345	39	16	scenes	scene	NOUN
ajst-5345	39	17	,	,	PUNCT
ajst-5345	39	18	reaching	reach	VERB
ajst-5345	39	19	an	an	DET
ajst-5345	39	20	average	average	NOUN
ajst-5345	39	21	of	of	ADP
ajst-5345	39	22	246	246	NUM
ajst-5345	39	23	pedestrians	pedestrian	NOUN
ajst-5345	39	24	per	per	ADP
ajst-5345	39	25	frame	frame	NOUN
ajst-5345	39	26	.	.	PUNCT
ajst-5345	40	1	these	these	DET
ajst-5345	40	2	sequences	sequence	NOUN
ajst-5345	40	3	contain	contain	VERB
ajst-5345	40	4	indoor	indoor	ADJ
ajst-5345	40	5	,	,	PUNCT
ajst-5345	40	6	outdoor	outdoor	ADJ
ajst-5345	40	7	,	,	PUNCT
ajst-5345	40	8	day	day	NOUN
ajst-5345	40	9	and	and	CCONJ
ajst-5345	40	10	night	night	NOUN
ajst-5345	40	11	scenes	scene	NOUN
ajst-5345	40	12	.	.	PUNCT
ajst-5345	41	1	we	we	PRON
ajst-5345	41	2	use	use	VERB
ajst-5345	41	3	mosaic	mosaic	ADJ
ajst-5345	41	4	and	and	CCONJ
ajst-5345	41	5	mixup	mixup	NOUN
ajst-5345	41	6	data	datum	NOUN
ajst-5345	41	7	augmentation	augmentation	NOUN
ajst-5345	41	8	:	:	PUNCT
ajst-5345	41	9	1	1	X
ajst-5345	41	10	)	)	PUNCT
ajst-5345	41	11	mosaic	mosaic	PROPN
ajst-5345	41	12	:	:	PUNCT
ajst-5345	41	13	the	the	DET
ajst-5345	41	14	mosaic	mosaic	ADJ
ajst-5345	41	15	data	data	NOUN
ajst-5345	41	16	enhancement	enhancement	NOUN
ajst-5345	41	17	method	method	NOUN
ajst-5345	41	18	is	be	AUX
ajst-5345	41	19	proposed	propose	VERB
ajst-5345	41	20	in	in	ADP
ajst-5345	41	21	the	the	DET
ajst-5345	41	22	yolov4	yolov4	PROPN
ajst-5345	41	23	paper	paper	NOUN
ajst-5345	41	24	.	.	PUNCT
ajst-5345	42	1	the	the	DET
ajst-5345	42	2	main	main	ADJ
ajst-5345	42	3	idea	idea	NOUN
ajst-5345	42	4	is	be	AUX
ajst-5345	42	5	to	to	PART
ajst-5345	42	6	randomly	randomly	VERB
ajst-5345	42	7	crop	crop	NOUN
ajst-5345	42	8	four	four	NUM
ajst-5345	42	9	pictures	picture	NOUN
ajst-5345	42	10	,	,	PUNCT
ajst-5345	42	11	and	and	CCONJ
ajst-5345	42	12	then	then	ADV
ajst-5345	42	13	stitch	stitch	VERB
ajst-5345	42	14	them	they	PRON
ajst-5345	42	15	into	into	ADP
ajst-5345	42	16	one	one	NUM
ajst-5345	42	17	picture	picture	NOUN
ajst-5345	42	18	as	as	ADP
ajst-5345	42	19	training	training	NOUN
ajst-5345	42	20	data	datum	NOUN
ajst-5345	42	21	.	.	PUNCT
ajst-5345	43	1	this	this	PRON
ajst-5345	43	2	not	not	PART
ajst-5345	43	3	only	only	ADV
ajst-5345	43	4	enriches	enrich	VERB
ajst-5345	43	5	the	the	DET
ajst-5345	43	6	background	background	NOUN
ajst-5345	43	7	of	of	ADP
ajst-5345	43	8	the	the	DET
ajst-5345	43	9	picture	picture	NOUN
ajst-5345	43	10	,	,	PUNCT
ajst-5345	43	11	but	but	CCONJ
ajst-5345	43	12	also	also	ADV
ajst-5345	43	13	increases	increase	VERB
ajst-5345	43	14	the	the	DET
ajst-5345	43	15	batch	batch	NOUN
ajst-5345	43	16	size	size	NOUN
ajst-5345	43	17	in	in	ADP
ajst-5345	43	18	disguise	disguise	NOUN
ajst-5345	43	19	,	,	PUNCT
ajst-5345	43	20	reducing	reduce	VERB
ajst-5345	43	21	its	its	PRON
ajst-5345	43	22	dependence	dependence	NOUN
ajst-5345	43	23	on	on	ADP
ajst-5345	43	24	batch	batch	NOUN
ajst-5345	43	25	size	size	NOUN
ajst-5345	43	26	.	.	PUNCT
ajst-5345	44	1	2	2	NUM
ajst-5345	44	2	)	)	PUNCT
ajst-5345	44	3	mixup	mixup	NOUN
ajst-5345	44	4	:	:	PUNCT
ajst-5345	45	1	mixup	mixup	NOUN
ajst-5345	45	2	is	be	AUX
ajst-5345	45	3	a	a	DET
ajst-5345	45	4	widely	widely	ADV
ajst-5345	45	5	used	use	VERB
ajst-5345	45	6	algorithm	algorithm	NOUN
ajst-5345	45	7	for	for	ADP
ajst-5345	45	8	mixedclass	mixedclass	ADJ
ajst-5345	45	9	enhancement	enhancement	NOUN
ajst-5345	45	10	of	of	ADP
ajst-5345	45	11	images	image	NOUN
ajst-5345	45	12	,	,	PUNCT
ajst-5345	45	13	which	which	PRON
ajst-5345	45	14	can	can	AUX
ajst-5345	45	15	mix	mix	VERB
ajst-5345	45	16	images	image	NOUN
ajst-5345	45	17	between	between	ADP
ajst-5345	45	18	different	different	ADJ
ajst-5345	45	19	classes	class	NOUN
ajst-5345	45	20	to	to	PART
ajst-5345	45	21	expand	expand	VERB
ajst-5345	45	22	the	the	DET
ajst-5345	45	23	training	training	NOUN
ajst-5345	45	24	dataset	dataset	NOUN
ajst-5345	45	25	.	.	PUNCT
ajst-5345	46	1	some	some	PRON
ajst-5345	46	2	of	of	ADP
ajst-5345	46	3	the	the	DET
ajst-5345	46	4	in	in	ADP
ajst-5345	46	5	-	-	PUNCT
ajst-5345	46	6	process	process	NOUN
ajst-5345	46	7	samples	sample	NOUN
ajst-5345	46	8	after	after	SCONJ
ajst-5345	46	9	data	data	NOUN
ajst-5345	46	10	enhancement	enhancement	NOUN
ajst-5345	46	11	is	be	AUX
ajst-5345	46	12	shown	show	VERB
ajst-5345	46	13	in	in	ADP
ajst-5345	46	14	the	the	DET
ajst-5345	46	15	figure	figure	NOUN
ajst-5345	46	16	1	1	NUM
ajst-5345	46	17	:	:	PUNCT
ajst-5345	46	18	figure	figure	NOUN
ajst-5345	46	19	1	1	NUM
ajst-5345	46	20	.	.	PUNCT
ajst-5345	46	21	data	datum	NOUN
ajst-5345	46	22	augmentation	augmentation	NOUN
ajst-5345	46	23	samples	sample	NOUN
ajst-5345	46	24	2.2	2.2	NUM
ajst-5345	46	25	.	.	PUNCT
ajst-5345	47	1	model	model	NOUN
ajst-5345	47	2	design	design	NOUN
ajst-5345	47	3	we	we	PRON
ajst-5345	47	4	add	add	VERB
ajst-5345	47	5	a	a	DET
ajst-5345	47	6	new	new	ADJ
ajst-5345	47	7	i	i	PROPN
ajst-5345	47	8	d	d	PROPN
ajst-5345	47	9	head	head	NOUN
ajst-5345	47	10	in	in	ADP
ajst-5345	47	11	yolox	yolox	NOUN
ajst-5345	47	12	,	,	PUNCT
ajst-5345	47	13	which	which	PRON
ajst-5345	47	14	is	be	AUX
ajst-5345	47	15	obtained	obtain	VERB
ajst-5345	47	16	from	from	ADP
ajst-5345	47	17	the	the	DET
ajst-5345	47	18	feature	feature	NOUN
ajst-5345	47	19	map	map	NOUN
ajst-5345	47	20	duplicate	duplicate	NOUN
ajst-5345	47	21	of	of	ADP
ajst-5345	47	22	the	the	DET
ajst-5345	47	23	earlier	early	ADJ
ajst-5345	47	24	conv2d	conv2d	ADJ
ajst-5345	47	25	.	.	PUNCT
ajst-5345	48	1	this	this	DET
ajst-5345	48	2	method	method	NOUN
ajst-5345	48	3	allows	allow	VERB
ajst-5345	48	4	for	for	ADP
ajst-5345	48	5	effective	effective	ADJ
ajst-5345	48	6	extraction	extraction	NOUN
ajst-5345	48	7	of	of	ADP
ajst-5345	48	8	the	the	DET
ajst-5345	48	9	reid	reid	PROPN
ajst-5345	48	10	features	feature	NOUN
ajst-5345	48	11	of	of	ADP
ajst-5345	48	12	the	the	DET
ajst-5345	48	13	target	target	NOUN
ajst-5345	48	14	.	.	PUNCT
ajst-5345	49	1	this	this	PRON
ajst-5345	49	2	allows	allow	VERB
ajst-5345	49	3	the	the	DET
ajst-5345	49	4	network	network	NOUN
ajst-5345	49	5	to	to	PART
ajst-5345	49	6	add	add	VERB
ajst-5345	49	7	a	a	DET
ajst-5345	49	8	focus	focus	NOUN
ajst-5345	49	9	on	on	ADP
ajst-5345	49	10	reid	reid	PROPN
ajst-5345	49	11	information	information	NOUN
ajst-5345	49	12	while	while	SCONJ
ajst-5345	49	13	maintaining	maintain	VERB
ajst-5345	49	14	the	the	DET
ajst-5345	49	15	original	original	ADJ
ajst-5345	49	16	detection	detection	NOUN
ajst-5345	49	17	performance	performance	NOUN
ajst-5345	49	18	,	,	PUNCT
ajst-5345	49	19	and	and	CCONJ
ajst-5345	49	20	we	we	PRON
ajst-5345	49	21	eventually	eventually	ADV
ajst-5345	49	22	add	add	VERB
ajst-5345	49	23	a	a	DET
ajst-5345	49	24	loss	loss	NOUN
ajst-5345	49	25	of	of	ADP
ajst-5345	49	26	i	i	PROPN
ajst-5345	49	27	d	d	PROPN
ajst-5345	49	28	to	to	ADP
ajst-5345	49	29	the	the	DET
ajst-5345	49	30	loss	loss	NOUN
ajst-5345	49	31	function	function	NOUN
ajst-5345	49	32	.	.	PUNCT
ajst-5345	50	1	further	far	ADV
ajst-5345	50	2	,	,	PUNCT
ajst-5345	50	3	the	the	DET
ajst-5345	50	4	modified	modify	VERB
ajst-5345	50	5	network	network	NOUN
ajst-5345	50	6	structure	structure	NOUN
ajst-5345	50	7	is	be	AUX
ajst-5345	50	8	shown	show	VERB
ajst-5345	50	9	in	in	ADP
ajst-5345	50	10	the	the	DET
ajst-5345	50	11	figure	figure	NOUN
ajst-5345	50	12	2	2	NUM
ajst-5345	50	13	.	.	PUNCT
ajst-5345	50	14	figure	figure	NOUN
ajst-5345	50	15	2	2	NUM
ajst-5345	50	16	.	.	PUNCT
ajst-5345	50	17	description	description	NOUN
ajst-5345	50	18	of	of	ADP
ajst-5345	50	19	the	the	DET
ajst-5345	50	20	i	i	PROPN
ajst-5345	50	21	d	d	PROPN
ajst-5345	50	22	head	head	NOUN
ajst-5345	50	23	we	we	PRON
ajst-5345	50	24	added	add	VERB
ajst-5345	50	25	to	to	ADP
ajst-5345	50	26	the	the	DET
ajst-5345	50	27	model	model	NOUN
ajst-5345	50	28	figure	figure	NOUN
ajst-5345	50	29	3	3	X
ajst-5345	50	30	.	.	PUNCT
ajst-5345	51	1	our	our	PRON
ajst-5345	51	2	proposed	propose	VERB
ajst-5345	51	3	design	design	NOUN
ajst-5345	51	4	for	for	ADP
ajst-5345	51	5	the	the	DET
ajst-5345	51	6	whole	whole	ADJ
ajst-5345	51	7	model	model	NOUN
ajst-5345	51	8	after	after	ADP
ajst-5345	51	9	that	that	PRON
ajst-5345	51	10	,	,	PUNCT
ajst-5345	51	11	we	we	PRON
ajst-5345	51	12	use	use	VERB
ajst-5345	51	13	sigmoid	sigmoid	NOUN
ajst-5345	51	14	focal	focal	ADJ
ajst-5345	51	15	loss	loss	NOUN
ajst-5345	51	16	as	as	ADP
ajst-5345	51	17	detection	detection	NOUN
ajst-5345	51	18	stage	stage	NOUN
ajst-5345	51	19	loss	loss	NOUN
ajst-5345	51	20	which	which	PRON
ajst-5345	51	21	remains	remain	VERB
ajst-5345	51	22	the	the	DET
ajst-5345	51	23	same	same	ADJ
ajst-5345	51	24	with	with	ADP
ajst-5345	51	25	the	the	DET
ajst-5345	51	26	origin	origin	NOUN
ajst-5345	51	27	bytetracker	bytetracker	NOUN
ajst-5345	51	28	.	.	PUNCT
ajst-5345	52	1	the	the	DET
ajst-5345	52	2	whole	whole	ADJ
ajst-5345	52	3	loss	loss	NOUN
ajst-5345	52	4	equation	equation	NOUN
ajst-5345	52	5	can	can	AUX
ajst-5345	52	6	be	be	AUX
ajst-5345	52	7	described	describe	VERB
ajst-5345	52	8	as	as	SCONJ
ajst-5345	52	9	follows	follow	VERB
ajst-5345	52	10	:	:	PUNCT
ajst-5345	52	11	reid2det1sum	reid2det1sum	NOUN
ajst-5345	52	12	losswlosswloss	losswlosswloss	NOUN
ajst-5345	53	1			PROPN
ajst-5345	53	2	(	(	PUNCT
ajst-5345	53	3	1	1	X
ajst-5345	53	4	)	)	PUNCT
ajst-5345	53	5	where	where	SCONJ
ajst-5345	53	6	w1	w1	NOUN
ajst-5345	53	7	and	and	CCONJ
ajst-5345	53	8	w2	w2	NOUN
ajst-5345	53	9	are	be	AUX
ajst-5345	53	10	weights	weight	NOUN
ajst-5345	53	11	for	for	ADP
ajst-5345	53	12	detection	detection	NOUN
ajst-5345	53	13	loss	loss	NOUN
ajst-5345	53	14	and	and	CCONJ
ajst-5345	53	15	i	i	NOUN
ajst-5345	53	16	d	d	PROPN
ajst-5345	53	17	loss	loss	NOUN
ajst-5345	53	18	,	,	PUNCT
ajst-5345	53	19	we	we	PRON
ajst-5345	53	20	choose	choose	VERB
ajst-5345	53	21	w1=0.638	w1=0.638	NOUN
ajst-5345	53	22	and	and	CCONJ
ajst-5345	53	23	w2=0.362	w2=0.362	PROPN
ajst-5345	53	24	as	as	ADP
ajst-5345	53	25	defaults	default	NOUN
ajst-5345	53	26	in	in	ADP
ajst-5345	53	27	the	the	DET
ajst-5345	53	28	following	follow	VERB
ajst-5345	53	29	experiments	experiment	NOUN
ajst-5345	53	30	.	.	PUNCT
ajst-5345	54	1	thus	thus	ADV
ajst-5345	54	2	,	,	PUNCT
ajst-5345	54	3	we	we	PRON
ajst-5345	54	4	have	have	VERB
ajst-5345	54	5	the	the	DET
ajst-5345	54	6	whole	whole	ADJ
ajst-5345	54	7	network	network	NOUN
ajst-5345	54	8	defined	define	VERB
ajst-5345	54	9	as	as	ADP
ajst-5345	54	10	figure	figure	NOUN
ajst-5345	54	11	3	3	NUM
ajst-5345	54	12	.	.	NOUN
ajst-5345	54	13	3	3	NUM
ajst-5345	54	14	.	.	X
ajst-5345	54	15	experiment	experiment	NOUN
ajst-5345	54	16	3.1	3.1	NUM
ajst-5345	54	17	.	.	PUNCT
ajst-5345	55	1	experiment	experiment	NOUN
ajst-5345	55	2	hyperparameters	hyperparameter	NOUN
ajst-5345	55	3	and	and	CCONJ
ajst-5345	55	4	settings	setting	NOUN
ajst-5345	55	5	we	we	PRON
ajst-5345	55	6	use	use	VERB
ajst-5345	55	7	pytorch	pytorch	NOUN
ajst-5345	55	8	as	as	ADP
ajst-5345	55	9	our	our	PRON
ajst-5345	55	10	deep	deep	ADJ
ajst-5345	55	11	learning	learning	NOUN
ajst-5345	55	12	framework	framework	NOUN
ajst-5345	55	13	and	and	CCONJ
ajst-5345	55	14	python	python	NOUN
ajst-5345	55	15	version	version	NOUN
ajst-5345	55	16	3.9.4	3.9.4	NUM
ajst-5345	55	17	.	.	PUNCT
ajst-5345	56	1	other	other	ADJ
ajst-5345	56	2	environment	environment	NOUN
ajst-5345	56	3	details	detail	NOUN
ajst-5345	56	4	includes	include	VERB
ajst-5345	56	5	64	64	NUM
ajst-5345	56	6	-	-	PUNCT
ajst-5345	56	7	bit	bit	NOUN
ajst-5345	56	8	ubuntu18.04	ubuntu18.04	PROPN
ajst-5345	56	9	os	os	NOUN
ajst-5345	56	10	and	and	CCONJ
ajst-5345	56	11	a	a	DET
ajst-5345	56	12	single	single	ADJ
ajst-5345	56	13	nvidia	nvidia	PROPN
ajst-5345	56	14	gtx	gtx	PROPN
ajst-5345	56	15	3090	3090	NUM
ajst-5345	56	16	gpu	gpu	NOUN
ajst-5345	56	17	.	.	PUNCT
ajst-5345	57	1	we	we	PRON
ajst-5345	57	2	train	train	VERB
ajst-5345	57	3	the	the	DET
ajst-5345	57	4	models	model	NOUN
ajst-5345	57	5	for	for	ADP
ajst-5345	57	6	a	a	DET
ajst-5345	57	7	total	total	NOUN
ajst-5345	57	8	of	of	ADP
ajst-5345	57	9	300	300	NUM
ajst-5345	57	10	epochs	epoch	NOUN
ajst-5345	57	11	with	with	ADP
ajst-5345	57	12	sgd	sgd	NOUN
ajst-5345	57	13	optimizer	optimizer	NOUN
ajst-5345	57	14	and	and	CCONJ
ajst-5345	57	15	first	first	ADJ
ajst-5345	57	16	5	5	NUM
ajst-5345	57	17	epochs	epoch	NOUN
ajst-5345	57	18	as	as	ADP
ajst-5345	57	19	warmup	warmup	NOUN
ajst-5345	57	20	on	on	ADP
ajst-5345	57	21	coco	coco	PROPN
ajst-5345	57	22	train2017	train2017	PROPN
ajst-5345	57	23	,	,	PUNCT
ajst-5345	57	24	where	where	SCONJ
ajst-5345	57	25	the	the	DET
ajst-5345	57	26	momentum	momentum	NOUN
ajst-5345	57	27	decay	decay	NOUN
ajst-5345	57	28	factor	factor	NOUN
ajst-5345	57	29	and	and	CCONJ
ajst-5345	57	30	weight	weight	NOUN
ajst-5345	57	31	decay	decay	NOUN
ajst-5345	57	32	factor	factor	NOUN
ajst-5345	57	33	are	be	AUX
ajst-5345	57	34	0.9	0.9	NUM
ajst-5345	57	35	and	and	CCONJ
ajst-5345	57	36	0.0005,our	0.0005,our	PRON
ajst-5345	57	37	initial	initial	ADJ
ajst-5345	57	38	learning	learning	NOUN
ajst-5345	57	39	rate	rate	NOUN
ajst-5345	57	40	is	be	AUX
ajst-5345	57	41	set	set	VERB
ajst-5345	57	42	to	to	ADP
ajst-5345	57	43	0.0004	0.0004	NUM
ajst-5345	57	44	and	and	CCONJ
ajst-5345	57	45	is	be	AUX
ajst-5345	57	46	set	set	VERB
ajst-5345	57	47	to	to	PART
ajst-5345	57	48	decrease	decrease	VERB
ajst-5345	57	49	for	for	ADP
ajst-5345	57	50	every	every	DET
ajst-5345	57	51	40	40	NUM
ajst-5345	57	52	iterations	iteration	NOUN
ajst-5345	57	53	.	.	PUNCT
ajst-5345	58	1	in	in	ADP
ajst-5345	58	2	the	the	DET
ajst-5345	58	3	experiment	experiment	NOUN
ajst-5345	58	4	,	,	PUNCT
ajst-5345	58	5	we	we	PRON
ajst-5345	58	6	set	set	VERB
ajst-5345	58	7	the	the	DET
ajst-5345	58	8	high	high	ADJ
ajst-5345	58	9	score	score	NOUN
ajst-5345	58	10	threshold	threshold	NOUN
ajst-5345	58	11	of	of	ADP
ajst-5345	58	12	detection	detection	NOUN
ajst-5345	58	13	results	result	NOUN
ajst-5345	58	14	as	as	ADP
ajst-5345	58	15	0.6	0.6	NUM
ajst-5345	58	16	,	,	PUNCT
ajst-5345	58	17	the	the	DET
ajst-5345	58	18	low	low	ADJ
ajst-5345	58	19	score	score	NOUN
ajst-5345	58	20	threshold	threshold	NOUN
ajst-5345	58	21	as	as	ADP
ajst-5345	58	22	0.1	0.1	NUM
ajst-5345	58	23	while	while	SCONJ
ajst-5345	58	24	the	the	DET
ajst-5345	58	25	tracking	tracking	NOUN
ajst-5345	58	26	threshold	threshold	NOUN
ajst-5345	58	27	as	as	ADP
ajst-5345	58	28	0.7	0.7	NUM
ajst-5345	58	29	.	.	PUNCT
ajst-5345	59	1	the	the	DET
ajst-5345	59	2	iou	iou	PROPN
ajst-5345	59	3	threshold	threshold	NOUN
ajst-5345	59	4	in	in	ADP
ajst-5345	59	5	the	the	DET
ajst-5345	59	6	matching	matching	NOUN
ajst-5345	59	7	stage	stage	NOUN
ajst-5345	59	8	is	be	AUX
ajst-5345	59	9	set	set	VERB
ajst-5345	59	10	to	to	ADP
ajst-5345	59	11	0.2	0.2	NUM
ajst-5345	59	12	.	.	PUNCT
ajst-5345	60	1	the	the	DET
ajst-5345	60	2	temporary	temporary	ADJ
ajst-5345	60	3	storage	storage	NOUN
ajst-5345	60	4	period	period	NOUN
ajst-5345	60	5	of	of	ADP
ajst-5345	60	6	lost	lost	ADJ
ajst-5345	60	7	objects	object	NOUN
ajst-5345	60	8	is	be	AUX
ajst-5345	60	9	45	45	NUM
ajst-5345	60	10	frames	frame	NOUN
ajst-5345	60	11	.	.	PUNCT
ajst-5345	61	1	the	the	DET
ajst-5345	61	2	detector	detector	NOUN
ajst-5345	61	3	uses	use	VERB
ajst-5345	61	4	yolox	yolox	NOUN
ajst-5345	61	5	and	and	CCONJ
ajst-5345	61	6	is	be	AUX
ajst-5345	61	7	pretrained	pretraine	VERB
ajst-5345	61	8	on	on	ADP
ajst-5345	61	9	the	the	DET
ajst-5345	61	10	coco	coco	PROPN
ajst-5345	61	11	dataset.the	dataset.the	DET
ajst-5345	61	12	input	input	NOUN
ajst-5345	61	13	image	image	NOUN
ajst-5345	61	14	resolution	resolution	NOUN
ajst-5345	61	15	is	be	AUX
ajst-5345	61	16	1440	1440	NUM
ajst-5345	61	17	×	×	NOUN
ajst-5345	61	18	800	800	NUM
ajst-5345	61	19	and	and	CCONJ
ajst-5345	61	20	the	the	DET
ajst-5345	61	21	shortest	short	ADJ
ajst-5345	61	22	side	side	NOUN
ajst-5345	61	23	ranges	range	VERB
ajst-5345	61	24	from	from	ADP
ajst-5345	61	25	576	576	NUM
ajst-5345	61	26	to	to	ADP
ajst-5345	61	27	1024	1024	NUM
ajst-5345	61	28	during	during	ADP
ajst-5345	61	29	multiscale	multiscale	ADJ
ajst-5345	61	30	training	training	NOUN
ajst-5345	61	31	,	,	PUNCT
ajst-5345	61	32	the	the	DET
ajst-5345	61	33	data	datum	NOUN
ajst-5345	61	34	was	be	AUX
ajst-5345	61	35	then	then	ADV
ajst-5345	61	36	augmented	augment	VERB
ajst-5345	61	37	using	use	VERB
ajst-5345	61	38	methods	method	NOUN
ajst-5345	61	39	shown	show	VERB
ajst-5345	61	40	in	in	ADP
ajst-5345	61	41	the	the	DET
ajst-5345	61	42	previous	previous	ADJ
ajst-5345	61	43	section	section	NOUN
ajst-5345	61	44	in	in	ADP
ajst-5345	61	45	this	this	DET
ajst-5345	61	46	paper	paper	NOUN
ajst-5345	61	47	.	.	PUNCT
ajst-5345	62	1	48	48	NUM
ajst-5345	62	2	table	table	NOUN
ajst-5345	62	3	1	1	NUM
ajst-5345	62	4	.	.	PUNCT
ajst-5345	62	5	comparison	comparison	NOUN
ajst-5345	62	6	of	of	ADP
ajst-5345	62	7	experimental	experimental	ADJ
ajst-5345	62	8	results	result	NOUN
ajst-5345	62	9	on	on	ADP
ajst-5345	62	10	mot17	mot17	PROPN
ajst-5345	62	11	and	and	CCONJ
ajst-5345	62	12	mot20	mot20	PROPN
ajst-5345	62	13	data	datum	NOUN
ajst-5345	62	14	sets	set	VERB
ajst-5345	62	15	mot17	mot17	PROPN
ajst-5345	62	16	with	with	ADP
ajst-5345	62	17	reid	reid	PROPN
ajst-5345	62	18	fusion	fusion	NOUN
ajst-5345	62	19	without	without	ADP
ajst-5345	62	20	reid	reid	PROPN
ajst-5345	62	21	fusion	fusion	PROPN
ajst-5345	62	22	precision	precision	NOUN
ajst-5345	62	23	/	/	SYM
ajst-5345	62	24	recall	recall	NOUN
ajst-5345	62	25	iou=0.50:0.95	iou=0.50:0.95	VERB
ajst-5345	63	1	iou=0.50	iou=0.50	PROPN
ajst-5345	63	2	iou=0.75	iou=0.75	PROPN
ajst-5345	63	3	iou=0.50:0.95	iou=0.50:0.95	NOUN
ajst-5345	63	4	iou=0.50	iou=0.50	PROPN
ajst-5345	63	5	iou=0.75	iou=0.75	PROPN
ajst-5345	63	6	all	all	DET
ajst-5345	63	7	0.651	0.651	NUM
ajst-5345	63	8	0.851	0.851	NUM
ajst-5345	63	9	0.757	0.757	NUM
ajst-5345	63	10	0.635	0.635	NUM
ajst-5345	63	11	0.827	0.827	NUM
ajst-5345	63	12	0.764	0.764	NUM
ajst-5345	63	13	precision	precision	NOUN
ajst-5345	63	14	all	all	DET
ajst-5345	63	15	0.679	0.679	NUM
ajst-5345	63	16	0.874	0.874	NUM
ajst-5345	63	17	0.793	0.793	NUM
ajst-5345	63	18	0.643	0.643	NUM
ajst-5345	63	19	0.852	0.852	NUM
ajst-5345	63	20	0.782	0.782	NUM
ajst-5345	63	21	recall	recall	NOUN
ajst-5345	63	22	area	area	NOUN
ajst-5345	63	23	with	with	ADP
ajst-5345	63	24	reid	reid	PROPN
ajst-5345	63	25	fusion	fusion	NOUN
ajst-5345	63	26	without	without	ADP
ajst-5345	63	27	reid	reid	PROPN
ajst-5345	63	28	fusion	fusion	NOUN
ajst-5345	63	29	small	small	ADJ
ajst-5345	63	30	medium	medium	ADJ
ajst-5345	63	31	large	large	ADJ
ajst-5345	63	32	small	small	ADJ
ajst-5345	63	33	medium	medium	ADJ
ajst-5345	63	34	large	large	ADJ
ajst-5345	63	35	precision	precision	NOUN
ajst-5345	63	36	n	n	CCONJ
ajst-5345	63	37	/	/	SYM
ajst-5345	63	38	a	a	DET
ajst-5345	63	39	0.671	0.671	NUM
ajst-5345	63	40	0.610	0.610	NUM
ajst-5345	63	41	0.712	0.712	NUM
ajst-5345	63	42	0.408	0.408	NUM
ajst-5345	63	43	0.599	0.599	NUM
ajst-5345	63	44	0.690	0.690	NUM
ajst-5345	63	45	precision	precision	NOUN
ajst-5345	63	46	n	n	CCONJ
ajst-5345	63	47	/	/	SYM
ajst-5345	63	48	a	a	DET
ajst-5345	63	49	0.579	0.579	NUM
ajst-5345	63	50	0.654	0.654	NUM
ajst-5345	63	51	0.744	0.744	NUM
ajst-5345	63	52	0.451	0.451	NUM
ajst-5345	63	53	0.643	0.643	NUM
ajst-5345	63	54	0.737	0.737	NUM
ajst-5345	63	55	recall	recall	NOUN
ajst-5345	63	56	mot20	mot20	PROPN
ajst-5345	63	57	with	with	ADP
ajst-5345	63	58	reid	reid	PROPN
ajst-5345	63	59	fusion	fusion	NOUN
ajst-5345	63	60	without	without	ADP
ajst-5345	63	61	reid	reid	PROPN
ajst-5345	63	62	fusion	fusion	PROPN
ajst-5345	63	63	precision	precision	NOUN
ajst-5345	63	64	/	/	SYM
ajst-5345	63	65	recall	recall	PROPN
ajst-5345	63	66	b	b	PROPN
ajst-5345	63	67	iou=0.50:0.95	iou=0.50:0.95	PUNCT
ajst-5345	63	68	iou=0.50	iou=0.50	PROPN
ajst-5345	63	69	iou=0.75	iou=0.75	PROPN
ajst-5345	63	70	iou=0.50:0.95	iou=0.50:0.95	NOUN
ajst-5345	63	71	iou=0.50	iou=0.50	PROPN
ajst-5345	63	72	iou=0.75	iou=0.75	PROPN
ajst-5345	63	73	all	all	DET
ajst-5345	63	74	0.626	0.626	NUM
ajst-5345	63	75	0.877	0.877	NUM
ajst-5345	63	76	0.741	0.741	NUM
ajst-5345	63	77	0.564	0.564	NUM
ajst-5345	63	78	0.828	0.828	NUM
ajst-5345	63	79	0.672	0.672	NUM
ajst-5345	63	80	precision	precision	NOUN
ajst-5345	63	81	all	all	PRON
ajst-5345	63	82	0.684	0.684	NUM
ajst-5345	63	83	0.821	0.821	NUM
ajst-5345	63	84	0.774	0.774	NUM
ajst-5345	63	85	0.643	0.643	NUM
ajst-5345	63	86	0.786	0.786	NUM
ajst-5345	63	87	0.692	0.692	NUM
ajst-5345	63	88	recall	recall	NOUN
ajst-5345	63	89	area	area	NOUN
ajst-5345	63	90	with	with	ADP
ajst-5345	63	91	reid	reid	PROPN
ajst-5345	63	92	fusion	fusion	NOUN
ajst-5345	63	93	without	without	ADP
ajst-5345	63	94	reid	reid	PROPN
ajst-5345	63	95	fusion	fusion	NOUN
ajst-5345	63	96	small	small	ADJ
ajst-5345	63	97	medium	medium	ADJ
ajst-5345	63	98	large	large	ADJ
ajst-5345	63	99	small	small	ADJ
ajst-5345	63	100	medium	medium	ADJ
ajst-5345	63	101	large	large	ADJ
ajst-5345	63	102	precision	precision	NOUN
ajst-5345	63	103	n	n	CCONJ
ajst-5345	63	104	/	/	SYM
ajst-5345	63	105	a	a	DET
ajst-5345	63	106	0.341	0.341	NUM
ajst-5345	63	107	0.621	0.621	NUM
ajst-5345	63	108	0.706	0.706	NUM
ajst-5345	63	109	0.319	0.319	NUM
ajst-5345	63	110	0.603	0.603	NUM
ajst-5345	63	111	0.701	0.701	NUM
ajst-5345	63	112	precision	precision	NOUN
ajst-5345	63	113	n	n	CCONJ
ajst-5345	63	114	/	/	SYM
ajst-5345	63	115	a	a	DET
ajst-5345	63	116	0.441	0.441	NUM
ajst-5345	63	117	0.549	0.549	NUM
ajst-5345	63	118	0.632	0.632	NUM
ajst-5345	63	119	0.423	0.423	NUM
ajst-5345	63	120	0.652	0.652	NUM
ajst-5345	63	121	0.628	0.628	NUM
ajst-5345	63	122	recall	recall	NOUN
ajst-5345	63	123	3.2	3.2	NUM
ajst-5345	63	124	.	.	PUNCT
ajst-5345	64	1	quantitative	quantitative	ADJ
ajst-5345	64	2	comparison	comparison	NOUN
ajst-5345	64	3	experiment	experiment	NOUN
ajst-5345	64	4	in	in	ADP
ajst-5345	64	5	this	this	DET
ajst-5345	64	6	section	section	NOUN
ajst-5345	64	7	we	we	PRON
ajst-5345	64	8	show	show	VERB
ajst-5345	64	9	comparison	comparison	NOUN
ajst-5345	64	10	with	with	ADP
ajst-5345	64	11	performance	performance	NOUN
ajst-5345	64	12	differences	difference	NOUN
ajst-5345	64	13	between	between	ADP
ajst-5345	64	14	our	our	PRON
ajst-5345	64	15	improved	improved	ADJ
ajst-5345	64	16	bytetrack	bytetrack	NOUN
ajst-5345	64	17	and	and	CCONJ
ajst-5345	64	18	origin	origin	NOUN
ajst-5345	64	19	bytetrack	bytetrack	NOUN
ajst-5345	64	20	on	on	ADP
ajst-5345	64	21	the	the	DET
ajst-5345	64	22	mot17	mot17	ADJ
ajst-5345	64	23	and	and	CCONJ
ajst-5345	64	24	mot20	mot20	ADJ
ajst-5345	64	25	datasets	dataset	NOUN
ajst-5345	64	26	at	at	ADP
ajst-5345	64	27	table	table	NOUN
ajst-5345	64	28	i.	i.	NOUN
ajst-5345	64	29	as	as	SCONJ
ajst-5345	64	30	shown	show	VERB
ajst-5345	64	31	in	in	ADP
ajst-5345	64	32	table1	table1	PROPN
ajst-5345	64	33	,	,	PUNCT
ajst-5345	64	34	the	the	DET
ajst-5345	64	35	improved	improved	ADJ
ajst-5345	64	36	method	method	NOUN
ajst-5345	64	37	proposed	propose	VERB
ajst-5345	64	38	in	in	ADP
ajst-5345	64	39	this	this	DET
ajst-5345	64	40	paper	paper	NOUN
ajst-5345	64	41	has	have	AUX
ajst-5345	64	42	been	be	AUX
ajst-5345	64	43	systematically	systematically	ADV
ajst-5345	64	44	enhanced	enhance	VERB
ajst-5345	64	45	on	on	ADP
ajst-5345	64	46	mot17	mot17	PROPN
ajst-5345	64	47	and	and	CCONJ
ajst-5345	64	48	mot20	mot20	PROPN
ajst-5345	64	49	.	.	PUNCT
ajst-5345	65	1	in	in	ADP
ajst-5345	65	2	mot17	mot17	PROPN
ajst-5345	65	3	the	the	DET
ajst-5345	65	4	improvement	improvement	NOUN
ajst-5345	65	5	averaged	average	VERB
ajst-5345	65	6	about	about	ADV
ajst-5345	65	7	2	2	NUM
ajst-5345	65	8	%	%	NOUN
ajst-5345	65	9	across	across	ADP
ajst-5345	65	10	the	the	DET
ajst-5345	65	11	iou	iou	NOUN
ajst-5345	65	12	intervals	interval	NOUN
ajst-5345	65	13	,	,	PUNCT
ajst-5345	65	14	while	while	SCONJ
ajst-5345	65	15	in	in	ADP
ajst-5345	65	16	mot20	mot20	PROPN
ajst-5345	65	17	we	we	PRON
ajst-5345	65	18	achieved	achieve	VERB
ajst-5345	65	19	close	close	ADV
ajst-5345	65	20	to	to	PART
ajst-5345	65	21	5	5	NUM
ajst-5345	65	22	%	%	NOUN
ajst-5345	65	23	.	.	PUNCT
ajst-5345	66	1	the	the	DET
ajst-5345	66	2	improvement	improvement	NOUN
ajst-5345	66	3	on	on	ADP
ajst-5345	66	4	recall	recall	NOUN
ajst-5345	66	5	is	be	AUX
ajst-5345	66	6	not	not	PART
ajst-5345	66	7	significant	significant	ADJ
ajst-5345	66	8	,	,	PUNCT
ajst-5345	66	9	probably	probably	ADV
ajst-5345	66	10	because	because	SCONJ
ajst-5345	66	11	the	the	DET
ajst-5345	66	12	original	original	ADJ
ajst-5345	66	13	model	model	NOUN
ajst-5345	66	14	is	be	AUX
ajst-5345	66	15	already	already	ADV
ajst-5345	66	16	good	good	ADJ
ajst-5345	66	17	enough	enough	ADV
ajst-5345	66	18	in	in	ADP
ajst-5345	66	19	terms	term	NOUN
ajst-5345	66	20	of	of	ADP
ajst-5345	66	21	recall	recall	NOUN
ajst-5345	66	22	performance	performance	NOUN
ajst-5345	66	23	,	,	PUNCT
ajst-5345	66	24	but	but	CCONJ
ajst-5345	66	25	the	the	DET
ajst-5345	66	26	added	add	VERB
ajst-5345	66	27	reid	reid	PROPN
ajst-5345	66	28	information	information	NOUN
ajst-5345	66	29	can	can	AUX
ajst-5345	66	30	significantly	significantly	ADV
ajst-5345	66	31	improve	improve	VERB
ajst-5345	66	32	precision	precision	NOUN
ajst-5345	66	33	performance	performance	NOUN
ajst-5345	66	34	.	.	PUNCT
ajst-5345	67	1	4	4	X
ajst-5345	67	2	.	.	X
ajst-5345	67	3	conclusion	conclusion	NOUN
ajst-5345	67	4	in	in	ADP
ajst-5345	67	5	this	this	DET
ajst-5345	67	6	paper	paper	NOUN
ajst-5345	67	7	,	,	PUNCT
ajst-5345	67	8	we	we	PRON
ajst-5345	67	9	propose	propose	VERB
ajst-5345	67	10	a	a	DET
ajst-5345	67	11	method	method	NOUN
ajst-5345	67	12	to	to	PART
ajst-5345	67	13	optimize	optimize	VERB
ajst-5345	67	14	bytetrack	bytetrack	NOUN
ajst-5345	67	15	by	by	ADP
ajst-5345	67	16	using	use	VERB
ajst-5345	67	17	reid	reid	PROPN
ajst-5345	67	18	to	to	PART
ajst-5345	67	19	extract	extract	VERB
ajst-5345	67	20	appearance	appearance	NOUN
ajst-5345	67	21	feature	feature	NOUN
ajst-5345	67	22	information	information	NOUN
ajst-5345	67	23	.	.	PUNCT
ajst-5345	68	1	we	we	PRON
ajst-5345	68	2	make	make	VERB
ajst-5345	68	3	better	well	ADJ
ajst-5345	68	4	use	use	NOUN
ajst-5345	68	5	of	of	ADP
ajst-5345	68	6	the	the	DET
ajst-5345	68	7	appearance	appearance	NOUN
ajst-5345	68	8	feature	feature	NOUN
ajst-5345	68	9	information	information	NOUN
ajst-5345	68	10	by	by	ADP
ajst-5345	68	11	designing	design	VERB
ajst-5345	68	12	the	the	DET
ajst-5345	68	13	appearance	appearance	NOUN
ajst-5345	68	14	feature	feature	NOUN
ajst-5345	68	15	extraction	extraction	NOUN
ajst-5345	68	16	network	network	NOUN
ajst-5345	68	17	.	.	PUNCT
ajst-5345	69	1	the	the	DET
ajst-5345	69	2	network	network	NOUN
ajst-5345	69	3	structure	structure	NOUN
ajst-5345	69	4	of	of	ADP
ajst-5345	69	5	this	this	DET
ajst-5345	69	6	paper	paper	NOUN
ajst-5345	69	7	is	be	AUX
ajst-5345	69	8	to	to	PART
ajst-5345	69	9	add	add	VERB
ajst-5345	69	10	a	a	DET
ajst-5345	69	11	reid	reid	NOUN
ajst-5345	69	12	branch	branch	NOUN
ajst-5345	69	13	based	base	VERB
ajst-5345	69	14	on	on	ADP
ajst-5345	69	15	the	the	DET
ajst-5345	69	16	original	original	ADJ
ajst-5345	69	17	yolox	yolox	NOUN
ajst-5345	69	18	decoupled	decouple	VERB
ajst-5345	69	19	head	head	NOUN
ajst-5345	69	20	.	.	PUNCT
ajst-5345	70	1	the	the	DET
ajst-5345	70	2	reid	reid	PROPN
ajst-5345	70	3	branch	branch	PROPN
ajst-5345	70	4	network	network	NOUN
ajst-5345	70	5	extracts	extract	VERB
ajst-5345	70	6	the	the	DET
ajst-5345	70	7	appearance	appearance	NOUN
ajst-5345	70	8	feature	feature	NOUN
ajst-5345	70	9	information	information	NOUN
ajst-5345	70	10	,	,	PUNCT
ajst-5345	70	11	and	and	CCONJ
ajst-5345	70	12	at	at	ADP
ajst-5345	70	13	the	the	DET
ajst-5345	70	14	same	same	ADJ
ajst-5345	70	15	time	time	NOUN
ajst-5345	70	16	uses	use	VERB
ajst-5345	70	17	the	the	DET
ajst-5345	70	18	reid	reid	PROPN
ajst-5345	70	19	loss	loss	NOUN
ajst-5345	70	20	function	function	NOUN
ajst-5345	70	21	.	.	PUNCT
ajst-5345	71	1	the	the	DET
ajst-5345	71	2	experimental	experimental	ADJ
ajst-5345	71	3	results	result	NOUN
ajst-5345	71	4	show	show	VERB
ajst-5345	71	5	that	that	SCONJ
ajst-5345	71	6	the	the	DET
ajst-5345	71	7	performance	performance	NOUN
ajst-5345	71	8	of	of	ADP
ajst-5345	71	9	the	the	DET
ajst-5345	71	10	proposed	propose	VERB
ajst-5345	71	11	method	method	NOUN
ajst-5345	71	12	on	on	ADP
ajst-5345	71	13	the	the	DET
ajst-5345	71	14	mot17	mot17	NOUN
ajst-5345	71	15	,	,	PUNCT
ajst-5345	71	16	mot20	mot20	PROPN
ajst-5345	71	17	has	have	AUX
ajst-5345	71	18	been	be	AUX
ajst-5345	71	19	improved	improve	VERB
ajst-5345	71	20	to	to	ADP
ajst-5345	71	21	a	a	DET
ajst-5345	71	22	certain	certain	ADJ
ajst-5345	71	23	extent	extent	NOUN
ajst-5345	71	24	compared	compare	VERB
ajst-5345	71	25	with	with	ADP
ajst-5345	71	26	the	the	DET
ajst-5345	71	27	original	original	ADJ
ajst-5345	71	28	bytetrack	bytetrack	NOUN
ajst-5345	71	29	,	,	PUNCT
ajst-5345	71	30	indicating	indicate	VERB
ajst-5345	71	31	that	that	SCONJ
ajst-5345	71	32	the	the	DET
ajst-5345	71	33	method	method	NOUN
ajst-5345	71	34	in	in	ADP
ajst-5345	71	35	this	this	DET
ajst-5345	71	36	paper	paper	NOUN
ajst-5345	71	37	can	can	AUX
ajst-5345	71	38	work	work	VERB
ajst-5345	71	39	better	well	ADV
ajst-5345	71	40	in	in	ADP
ajst-5345	71	41	complex	complex	ADJ
ajst-5345	71	42	motion	motion	NOUN
ajst-5345	71	43	and	and	CCONJ
ajst-5345	71	44	diverse	diverse	ADJ
ajst-5345	71	45	motion	motion	NOUN
ajst-5345	71	46	patterns	pattern	NOUN
ajst-5345	71	47	.	.	PUNCT
ajst-5345	72	1	acknowledgment	acknowledgment	NOUN
ajst-5345	72	2	this	this	DET
ajst-5345	72	3	work	work	NOUN
ajst-5345	72	4	was	be	AUX
ajst-5345	72	5	supported	support	VERB
ajst-5345	72	6	by	by	ADP
ajst-5345	72	7	the	the	DET
ajst-5345	72	8	innovative	innovative	ADJ
ajst-5345	72	9	research	research	NOUN
ajst-5345	72	10	project	project	NOUN
ajst-5345	72	11	for	for	ADP
ajst-5345	72	12	postgraduates	postgraduate	NOUN
ajst-5345	72	13	of	of	ADP
ajst-5345	72	14	southwest	southwest	PROPN
ajst-5345	72	15	minzu	minzu	PROPN
ajst-5345	72	16	university	university	PROPN
ajst-5345	72	17	(	(	PUNCT
ajst-5345	72	18	cx2021061	cx2021061	PROPN
ajst-5345	72	19	)	)	PUNCT
ajst-5345	72	20	.	.	PUNCT
ajst-5345	73	1	references	reference	NOUN
ajst-5345	73	2	[	[	X
ajst-5345	73	3	1	1	NUM
ajst-5345	73	4	]	]	X
ajst-5345	73	5	y.	y.	PROPN
ajst-5345	73	6	zhang	zhang	PROPN
ajst-5345	73	7	,	,	PUNCT
ajst-5345	73	8	p.	p.	PROPN
ajst-5345	73	9	sun	sun	PROPN
ajst-5345	73	10	,	,	PUNCT
ajst-5345	73	11	y.	y.	PROPN
ajst-5345	73	12	jiang	jiang	PROPN
ajst-5345	73	13	,	,	PUNCT
ajst-5345	73	14	et	et	PROPN
ajst-5345	73	15	al	al	PROPN
ajst-5345	73	16	.	.	PUNCT
ajst-5345	74	1	“	"	PUNCT
ajst-5345	74	2	bytetrack	bytetrack	NOUN
ajst-5345	74	3	:	:	PUNCT
ajst-5345	74	4	multi	multi	ADJ
ajst-5345	74	5	-	-	ADJ
ajst-5345	74	6	object	object	ADJ
ajst-5345	74	7	tracking	tracking	NOUN
ajst-5345	74	8	by	by	ADP
ajst-5345	74	9	associating	associate	VERB
ajst-5345	74	10	every	every	DET
ajst-5345	74	11	detection	detection	NOUN
ajst-5345	74	12	box	box	NOUN
ajst-5345	74	13	,	,	PUNCT
ajst-5345	74	14	”	"	PUNCT
ajst-5345	74	15	corr	corr	NOUN
ajst-5345	74	16	,	,	PUNCT
ajst-5345	74	17	vol	vol	NOUN
ajst-5345	74	18	.	.	PUNCT
ajst-5345	75	1	abs/2110.06864	abs/2110.06864	PROPN
ajst-5345	75	2	,	,	PUNCT
ajst-5345	75	3	2021	2021	NUM
ajst-5345	75	4	.	.	PUNCT
ajst-5345	76	1	[	[	X
ajst-5345	76	2	online	online	X
ajst-5345	76	3	]	]	X
ajst-5345	76	4	.	.	PUNCT
ajst-5345	77	1	available	available	ADJ
ajst-5345	77	2	:	:	PUNCT
ajst-5345	77	3	https://arxiv.org/abs/2110.06864	https://arxiv.org/abs/2110.06864	PROPN
ajst-5345	78	1	[	[	X
ajst-5345	78	2	2	2	NUM
ajst-5345	78	3	]	]	PUNCT
ajst-5345	78	4	d.	d.	PROPN
ajst-5345	78	5	gray	gray	PROPN
ajst-5345	78	6	,	,	PUNCT
ajst-5345	78	7	and	and	CCONJ
ajst-5345	78	8	h.	h.	PROPN
ajst-5345	78	9	tao.viewpoint	tao.viewpoint	PROPN
ajst-5345	78	10	invariant	invariant	ADJ
ajst-5345	78	11	pedestrian	pedestrian	NOUN
ajst-5345	78	12	recognition	recognition	NOUN
ajst-5345	78	13	with	with	ADP
ajst-5345	78	14	an	an	DET
ajst-5345	78	15	ensemble	ensemble	NOUN
ajst-5345	78	16	of	of	ADP
ajst-5345	78	17	localized	localized	ADJ
ajst-5345	78	18	features	feature	NOUN
ajst-5345	78	19	,	,	PUNCT
ajst-5345	78	20	in	in	ADP
ajst-5345	78	21	proc	proc	NOUN
ajst-5345	78	22	.	.	PUNCT
ajst-5345	79	1	european	european	ADJ
ajst-5345	79	2	conference	conference	PROPN
ajst-5345	79	3	on	on	ADP
ajst-5345	79	4	computer	computer	NOUN
ajst-5345	79	5	vision	vision	NOUN
ajst-5345	79	6	(	(	PUNCT
ajst-5345	79	7	eccv	eccv	ADV
ajst-5345	79	8	)	)	PUNCT
ajst-5345	79	9	,	,	PUNCT
ajst-5345	79	10	2008	2008	NUM
ajst-5345	79	11	.	.	PUNCT
ajst-5345	80	1	[	[	X
ajst-5345	80	2	3	3	X
ajst-5345	80	3	]	]	X
ajst-5345	80	4	d.	d.	PROPN
ajst-5345	80	5	gray	gray	PROPN
ajst-5345	80	6	,	,	PUNCT
ajst-5345	80	7	s.	s.	PROPN
ajst-5345	80	8	brennan	brennan	PROPN
ajst-5345	80	9	,	,	PUNCT
ajst-5345	80	10	and	and	CCONJ
ajst-5345	80	11	h.	h.	PROPN
ajst-5345	80	12	tao	tao	PROPN
ajst-5345	80	13	,	,	PUNCT
ajst-5345	80	14	“	"	PUNCT
ajst-5345	80	15	evaluating	evaluate	VERB
ajst-5345	80	16	appearance	appearance	NOUN
ajst-5345	80	17	models	model	NOUN
ajst-5345	80	18	for	for	ADP
ajst-5345	80	19	recognition	recognition	NOUN
ajst-5345	80	20	,	,	PUNCT
ajst-5345	80	21	reacquisition	reacquisition	NOUN
ajst-5345	80	22	,	,	PUNCT
ajst-5345	80	23	and	and	CCONJ
ajst-5345	80	24	tracking	tracking	NOUN
ajst-5345	80	25	,	,	PUNCT
ajst-5345	80	26	”	"	PUNCT
ajst-5345	80	27	performance	performance	NOUN
ajst-5345	80	28	evaluation	evaluation	NOUN
ajst-5345	80	29	of	of	ADP
ajst-5345	80	30	tracking	tracking	NOUN
ajst-5345	80	31	and	and	CCONJ
ajst-5345	80	32	surveillance	surveillance	NOUN
ajst-5345	80	33	(	(	PUNCT
ajst-5345	80	34	pets	pet	NOUN
ajst-5345	80	35	)	)	PUNCT
ajst-5345	80	36	.	.	PUNCT
ajst-5345	81	1	ieee	ieee	PROPN
ajst-5345	81	2	international	international	ADJ
ajst-5345	81	3	workshop	workshop	NOUN
ajst-5345	81	4	on	on	ADP
ajst-5345	81	5	,	,	PUNCT
ajst-5345	81	6	2007	2007	NUM
ajst-5345	81	7	.	.	PUNCT
ajst-5345	82	1	[	[	X
ajst-5345	82	2	4	4	X
ajst-5345	82	3	]	]	PUNCT
ajst-5345	82	4	x.	x.	NOUN
ajst-5345	82	5	zhou	zhou	PROPN
ajst-5345	82	6	,	,	PUNCT
ajst-5345	82	7	v.	v.	PROPN
ajst-5345	82	8	koltun	koltun	PROPN
ajst-5345	82	9	,	,	PUNCT
ajst-5345	82	10	and	and	CCONJ
ajst-5345	82	11	p.	p.	NOUN
ajst-5345	82	12	krahenb	krahenb	PROPN
ajst-5345	82	13	uhl	uhl	PROPN
ajst-5345	82	14	.	.	PUNCT
ajst-5345	83	1	tracking	track	VERB
ajst-5345	83	2	objects	object	NOUN
ajst-5345	83	3	as	as	ADP
ajst-5345	83	4	points	point	NOUN
ajst-5345	83	5	.	.	PUNCT
ajst-5345	84	1	in	in	ADP
ajst-5345	84	2	european	european	ADJ
ajst-5345	84	3	conference	conference	PROPN
ajst-5345	84	4	on	on	ADP
ajst-5345	84	5	computer	computer	NOUN
ajst-5345	84	6	vision	vision	NOUN
ajst-5345	84	7	,	,	PUNCT
ajst-5345	84	8	pages	page	NOUN
ajst-5345	84	9	474–490	474–490	NUM
ajst-5345	84	10	.	.	PUNCT
ajst-5345	84	11	springer	springer	NOUN
ajst-5345	84	12	,	,	PUNCT
ajst-5345	84	13	2020	2020	NUM
ajst-5345	84	14	.	.	PUNCT
ajst-5345	85	1	[	[	X
ajst-5345	85	2	5	5	NUM
ajst-5345	85	3	]	]	X
ajst-5345	85	4	guanshuo	guanshuo	NOUN
ajst-5345	85	5	wang	wang	PROPN
ajst-5345	85	6	,	,	PUNCT
ajst-5345	85	7	yufeng	yufeng	PROPN
ajst-5345	85	8	yuan	yuan	PROPN
ajst-5345	85	9	,	,	PUNCT
ajst-5345	85	10	xiong	xiong	PROPN
ajst-5345	85	11	chen	chen	PROPN
ajst-5345	85	12	,	,	PUNCT
ajst-5345	85	13	jiwei	jiwei	PROPN
ajst-5345	85	14	li	li	PROPN
ajst-5345	85	15	,	,	PUNCT
ajst-5345	85	16	and	and	CCONJ
ajst-5345	85	17	xi	xi	PROPN
ajst-5345	85	18	zhou	zhou	PROPN
ajst-5345	85	19	.	.	PUNCT
ajst-5345	85	20	2018	2018	NUM
ajst-5345	85	21	.	.	PUNCT
ajst-5345	86	1	learning	learn	VERB
ajst-5345	86	2	discriminative	discriminative	NOUN
ajst-5345	86	3	features	feature	NOUN
ajst-5345	86	4	with	with	ADP
ajst-5345	86	5	multiple	multiple	ADJ
ajst-5345	86	6	granularities	granularity	NOUN
ajst-5345	86	7	for	for	ADP
ajst-5345	86	8	person	person	NOUN
ajst-5345	86	9	re	re	NOUN
ajst-5345	86	10	-	-	NOUN
ajst-5345	86	11	identification	identification	NOUN
ajst-5345	86	12	.	.	PUNCT
ajst-5345	87	1	in	in	ADP
ajst-5345	87	2	proceedings	proceeding	NOUN
ajst-5345	87	3	of	of	ADP
ajst-5345	87	4	the	the	DET
ajst-5345	87	5	26th	26th	ADJ
ajst-5345	87	6	acm	acm	PROPN
ajst-5345	87	7	international	international	ADJ
ajst-5345	87	8	conference	conference	NOUN
ajst-5345	87	9	on	on	ADP
ajst-5345	87	10	multimedia	multimedia	NOUN
ajst-5345	87	11	.	.	PUNCT
ajst-5345	88	1	association	association	NOUN
ajst-5345	88	2	for	for	ADP
ajst-5345	88	3	computing	compute	VERB
ajst-5345	88	4	machinery	machinery	NOUN
ajst-5345	88	5	,	,	PUNCT
ajst-5345	88	6	new	new	PROPN
ajst-5345	88	7	york	york	PROPN
ajst-5345	88	8	,	,	PUNCT
ajst-5345	88	9	ny	ny	PROPN
ajst-5345	88	10	,	,	PUNCT
ajst-5345	88	11	usa	usa	PROPN
ajst-5345	88	12	,	,	PUNCT
ajst-5345	88	13	274–282	274–282	NUM
ajst-5345	88	14	.	.	PUNCT
ajst-5345	89	1	[	[	X
ajst-5345	89	2	6	6	NUM
ajst-5345	89	3	]	]	PUNCT
ajst-5345	89	4	he	he	PRON
ajst-5345	89	5	s	s	PROPN
ajst-5345	89	6	,	,	PUNCT
ajst-5345	89	7	luo	luo	PROPN
ajst-5345	89	8	h	h	PROPN
ajst-5345	89	9	,	,	PUNCT
ajst-5345	89	10	wang	wang	PROPN
ajst-5345	89	11	p	p	PROPN
ajst-5345	89	12	,	,	PUNCT
ajst-5345	89	13	et	et	PROPN
ajst-5345	89	14	al	al	PROPN
ajst-5345	89	15	.	.	PROPN
ajst-5345	89	16	transreid	transreid	PROPN
ajst-5345	89	17	:	:	PUNCT
ajst-5345	89	18	transformer	transformer	NOUN
ajst-5345	89	19	-	-	PUNCT
ajst-5345	89	20	based	base	VERB
ajst-5345	89	21	object	object	NOUN
ajst-5345	89	22	re	re	NOUN
ajst-5345	89	23	-	-	NOUN
ajst-5345	89	24	identification[c]//proceedings	identification[c]//proceeding	NOUN
ajst-5345	89	25	of	of	ADP
ajst-5345	89	26	the	the	DET
ajst-5345	89	27	ieee	ieee	NOUN
ajst-5345	89	28	/	/	SYM
ajst-5345	89	29	cvf	cvf	NOUN
ajst-5345	89	30	international	international	ADJ
ajst-5345	89	31	conference	conference	NOUN
ajst-5345	89	32	on	on	ADP
ajst-5345	89	33	computer	computer	NOUN
ajst-5345	89	34	vision	vision	NOUN
ajst-5345	89	35	.	.	PUNCT
ajst-5345	90	1	2021	2021	NUM
ajst-5345	90	2	:	:	PUNCT
ajst-5345	90	3	1501315022	1501315022	NUM
ajst-5345	90	4	.	.	PUNCT
ajst-5345	91	1	[	[	X
ajst-5345	91	2	7	7	X
ajst-5345	91	3	]	]	X
ajst-5345	91	4	zou	zou	PROPN
ajst-5345	91	5	y	y	PROPN
ajst-5345	91	6	,	,	PUNCT
ajst-5345	91	7	yang	yang	PROPN
ajst-5345	91	8	x	x	PROPN
ajst-5345	91	9	,	,	PUNCT
ajst-5345	91	10	yu	yu	PROPN
ajst-5345	91	11	z	z	PROPN
ajst-5345	91	12	,	,	PUNCT
ajst-5345	91	13	et	et	PROPN
ajst-5345	91	14	al	al	PROPN
ajst-5345	91	15	.	.	PUNCT
ajst-5345	91	16	joint	joint	ADJ
ajst-5345	91	17	disentangling	disentangling	NOUN
ajst-5345	91	18	and	and	CCONJ
ajst-5345	91	19	adaptation	adaptation	NOUN
ajst-5345	91	20	for	for	ADP
ajst-5345	91	21	cross	cross	ADJ
ajst-5345	91	22	-	-	ADJ
ajst-5345	91	23	domain	domain	ADJ
ajst-5345	91	24	person	person	NOUN
ajst-5345	91	25	re	re	NOUN
ajst-5345	91	26	-	-	PROPN
ajst-5345	91	27	identification[c]//computer	identification[c]//computer	ADJ
ajst-5345	91	28	vision	vision	NOUN
ajst-5345	91	29	–	–	PUNCT
ajst-5345	91	30	eccv	eccv	ADJ
ajst-5345	91	31	2020	2020	NUM
ajst-5345	91	32	:	:	PUNCT
ajst-5345	91	33	16th	16th	ADJ
ajst-5345	91	34	european	european	ADJ
ajst-5345	91	35	conference	conference	PROPN
ajst-5345	91	36	,	,	PUNCT
ajst-5345	91	37	glasgow	glasgow	PROPN
ajst-5345	91	38	,	,	PUNCT
ajst-5345	91	39	uk	uk	PROPN
ajst-5345	91	40	,	,	PUNCT
ajst-5345	91	41	august	august	PROPN
ajst-5345	91	42	23–28	23–28	NUM
ajst-5345	91	43	,	,	PUNCT
ajst-5345	91	44	2020	2020	NUM
ajst-5345	91	45	,	,	PUNCT
ajst-5345	91	46	proceedings	proceeding	NOUN
ajst-5345	91	47	,	,	PUNCT
ajst-5345	91	48	part	part	NOUN
ajst-5345	91	49	ii	ii	PROPN
ajst-5345	91	50	16	16	NUM
ajst-5345	91	51	.	.	PUNCT
ajst-5345	92	1	springer	springer	NOUN
ajst-5345	92	2	international	international	ADJ
ajst-5345	92	3	publishing	publishing	NOUN
ajst-5345	92	4	,	,	PUNCT
ajst-5345	92	5	2020	2020	NUM
ajst-5345	92	6	:	:	PUNCT
ajst-5345	92	7	87	87	NUM
ajst-5345	92	8	-	-	SYM
ajst-5345	92	9	104	104	NUM
ajst-5345	92	10	.	.	PUNCT
ajst-5345	93	1	[	[	X
ajst-5345	93	2	8	8	X
ajst-5345	93	3	]	]	PUNCT
ajst-5345	93	4	zhai	zhai	PROPN
ajst-5345	93	5	y	y	PROPN
ajst-5345	93	6	,	,	PUNCT
ajst-5345	93	7	ye	ye	PRON
ajst-5345	93	8	q	q	ADJ
ajst-5345	93	9	,	,	PUNCT
ajst-5345	93	10	lu	lu	PROPN
ajst-5345	93	11	s	s	PROPN
ajst-5345	93	12	,	,	PUNCT
ajst-5345	93	13	et	et	PROPN
ajst-5345	93	14	al	al	PROPN
ajst-5345	93	15	.	.	PUNCT
ajst-5345	93	16	multiple	multiple	ADJ
ajst-5345	93	17	expert	expert	NOUN
ajst-5345	93	18	brainstorming	brainstorm	VERB
ajst-5345	93	19	for	for	ADP
ajst-5345	93	20	domain	domain	NOUN
ajst-5345	93	21	adaptive	adaptive	ADJ
ajst-5345	93	22	person	person	NOUN
ajst-5345	93	23	re	re	NOUN
ajst-5345	93	24	-	-	PROPN
ajst-5345	93	25	identification[c]//computer	identification[c]//computer	ADJ
ajst-5345	93	26	vision	vision	NOUN
ajst-5345	93	27	–	–	PUNCT
ajst-5345	93	28	eccv	eccv	ADJ
ajst-5345	93	29	2020	2020	NUM
ajst-5345	93	30	:	:	PUNCT
ajst-5345	93	31	16th	16th	ADJ
ajst-5345	93	32	european	european	ADJ
ajst-5345	93	33	conference	conference	PROPN
ajst-5345	93	34	,	,	PUNCT
ajst-5345	93	35	glasgow	glasgow	PROPN
ajst-5345	93	36	,	,	PUNCT
ajst-5345	93	37	uk	uk	PROPN
ajst-5345	93	38	,	,	PUNCT
ajst-5345	93	39	august	august	PROPN
ajst-5345	93	40	23–28	23–28	NUM
ajst-5345	93	41	,	,	PUNCT
ajst-5345	93	42	2020	2020	NUM
ajst-5345	93	43	,	,	PUNCT
ajst-5345	93	44	proceedings	proceeding	NOUN
ajst-5345	93	45	,	,	PUNCT
ajst-5345	93	46	part	part	NOUN
ajst-5345	93	47	vii	vii	PROPN
ajst-5345	93	48	16	16	NUM
ajst-5345	93	49	.	.	PUNCT
ajst-5345	94	1	springer	springer	NOUN
ajst-5345	94	2	international	international	ADJ
ajst-5345	94	3	publishing	publishing	NOUN
ajst-5345	94	4	,	,	PUNCT
ajst-5345	94	5	2020	2020	NUM
ajst-5345	94	6	:	:	PUNCT
ajst-5345	95	1	594	594	NUM
ajst-5345	95	2	-	-	SYM
ajst-5345	95	3	611	611	NUM
ajst-5345	95	4	.	.	PUNCT
ajst-5345	96	1	[	[	X
ajst-5345	96	2	9	9	NUM
ajst-5345	96	3	]	]	PUNCT
ajst-5345	96	4	a.	a.	PROPN
ajst-5345	96	5	milan	milan	PROPN
ajst-5345	96	6	,	,	PUNCT
ajst-5345	96	7	l.	l.	PROPN
ajst-5345	96	8	leal	leal	PROPN
ajst-5345	96	9	-	-	PUNCT
ajst-5345	96	10	taix´e	taix´e	NOUN
ajst-5345	96	11	,	,	PUNCT
ajst-5345	96	12	i.	i.	PROPN
ajst-5345	96	13	reid	reid	PROPN
ajst-5345	96	14	,	,	PUNCT
ajst-5345	96	15	s.	s.	PROPN
ajst-5345	96	16	roth	roth	PROPN
ajst-5345	96	17	,	,	PUNCT
ajst-5345	96	18	and	and	CCONJ
ajst-5345	96	19	k.	k.	PROPN
ajst-5345	96	20	schindler.mot16	schindler.mot16	NOUN
ajst-5345	96	21	:	:	PUNCT
ajst-5345	96	22	a	a	DET
ajst-5345	96	23	benchmark	benchmark	NOUN
ajst-5345	96	24	for	for	ADP
ajst-5345	96	25	multi	multi	ADJ
ajst-5345	96	26	-	-	ADJ
ajst-5345	96	27	object	object	ADJ
ajst-5345	96	28	tracking	tracking	NOUN
ajst-5345	96	29	.	.	PUNCT
ajst-5345	97	1	arxiv	arxiv	PROPN
ajst-5345	97	2	preprint	preprint	PROPN
ajst-5345	97	3	arxiv:1603.00831	arxiv:1603.00831	PROPN
ajst-5345	97	4	,	,	PUNCT
ajst-5345	97	5	2016	2016	NUM
ajst-5345	97	6	.	.	PUNCT
ajst-5345	98	1	[	[	X
ajst-5345	98	2	10	10	NUM
ajst-5345	98	3	]	]	PUNCT
ajst-5345	98	4	dendorfer	dendorfer	NOUN
ajst-5345	98	5	p	p	NOUN
ajst-5345	98	6	,	,	PUNCT
ajst-5345	98	7	rezatofighi	rezatofighi	PROPN
ajst-5345	98	8	h	h	PROPN
ajst-5345	98	9	,	,	PUNCT
ajst-5345	98	10	milan	milan	PROPN
ajst-5345	98	11	a	a	PRON
ajst-5345	98	12	,	,	PUNCT
ajst-5345	98	13	et	et	PROPN
ajst-5345	98	14	al	al	PROPN
ajst-5345	98	15	.	.	PUNCT
ajst-5345	99	1	mot20	mot20	PROPN
ajst-5345	99	2	:	:	PUNCT
ajst-5345	99	3	a	a	DET
ajst-5345	99	4	benchmark	benchmark	NOUN
ajst-5345	99	5	for	for	ADP
ajst-5345	99	6	multi	multi	ADJ
ajst-5345	99	7	object	object	NOUN
ajst-5345	99	8	tracking	tracking	NOUN
ajst-5345	99	9	in	in	ADP
ajst-5345	99	10	crowded	crowded	ADJ
ajst-5345	99	11	scenes	scene	NOUN
ajst-5345	99	12	:	:	PUNCT
ajst-5345	99	13	arxiv	arxiv	PROPN
ajst-5345	99	14	,	,	PUNCT
ajst-5345	99	15	10.48550	10.48550	NUM
ajst-5345	99	16	/	/	SYM
ajst-5345	99	17	arxiv.2003.09003[p	arxiv.2003.09003[p	NUM
ajst-5345	99	18	]	]	PUNCT
ajst-5345	99	19	.	.	PUNCT
ajst-5345	100	1	2020	2020	NUM
ajst-5345	100	2	.	.	PUNCT