id	sid	tid	token	lemma	pos
ajst-2981	1	1	academic	academic	ADJ
ajst-2981	1	2	journal	journal	NOUN
ajst-2981	1	3	of	of	ADP
ajst-2981	1	4	science	science	NOUN
ajst-2981	1	5	and	and	CCONJ
ajst-2981	1	6	technology	technology	NOUN
ajst-2981	1	7	issn	issn	NOUN
ajst-2981	1	8	:	:	PUNCT
ajst-2981	1	9	2771	2771	NUM
ajst-2981	1	10	-	-	SYM
ajst-2981	1	11	3032	3032	NUM
ajst-2981	1	12	|	|	NOUN
ajst-2981	1	13	vol	vol	NOUN
ajst-2981	1	14	.	.	PROPN
ajst-2981	2	1	3	3	NUM
ajst-2981	2	2	,	,	PUNCT
ajst-2981	2	3	no	no	INTJ
ajst-2981	2	4	.	.	NOUN
ajst-2981	2	5	3	3	NUM
ajst-2981	2	6	,	,	PUNCT
ajst-2981	2	7	2022	2022	NUM
ajst-2981	2	8	198	198	NUM
ajst-2981	2	9	aerial	aerial	ADJ
ajst-2981	2	10	traffic	traffic	NOUN
ajst-2981	2	11	statistics	statistic	NOUN
ajst-2981	2	12	based	base	VERB
ajst-2981	2	13	on	on	ADP
ajst-2981	2	14	yolov5+deepsort	yolov5+deepsort	PROPN
ajst-2981	2	15	wei	wei	PROPN
ajst-2981	2	16	liu1	liu1	PROPN
ajst-2981	2	17	,	,	PUNCT
ajst-2981	2	18	lin	lin	PROPN
ajst-2981	2	19	zhang1	zhang1	PROPN
ajst-2981	2	20	,	,	PUNCT
ajst-2981	2	21	2	2	NUM
ajst-2981	2	22	,	,	PUNCT
ajst-2981	2	23	*	*	PROPN
ajst-2981	2	24	1college	1college	NUM
ajst-2981	2	25	of	of	ADP
ajst-2981	2	26	civil	civil	ADJ
ajst-2981	2	27	and	and	CCONJ
ajst-2981	2	28	architectural	architectural	ADJ
ajst-2981	2	29	engineering	engineering	NOUN
ajst-2981	2	30	,	,	PUNCT
ajst-2981	2	31	north	north	PROPN
ajst-2981	2	32	china	china	PROPN
ajst-2981	2	33	university	university	PROPN
ajst-2981	2	34	of	of	ADP
ajst-2981	2	35	science	science	NOUN
ajst-2981	2	36	and	and	CCONJ
ajst-2981	2	37	technology	technology	NOUN
ajst-2981	2	38	,	,	PUNCT
ajst-2981	2	39	tangshan	tangshan	PROPN
ajst-2981	2	40	hebei	hebei	PROPN
ajst-2981	2	41	063210	063210	NUM
ajst-2981	2	42	,	,	PUNCT
ajst-2981	2	43	china	china	PROPN
ajst-2981	2	44	2tangshan	2tangshan	PROPN
ajst-2981	2	45	city	city	NOUN
ajst-2981	2	46	air	air	NOUN
ajst-2981	2	47	and	and	CCONJ
ajst-2981	2	48	ground	grind	VERB
ajst-2981	2	49	smart	smart	ADJ
ajst-2981	2	50	transportation	transportation	NOUN
ajst-2981	2	51	key	key	ADJ
ajst-2981	2	52	laboratory	laboratory	NOUN
ajst-2981	2	53	,	,	PUNCT
ajst-2981	2	54	tangshan	tangshan	ADP
ajst-2981	2	55	hebei	hebei	PROPN
ajst-2981	2	56	063210	063210	NUM
ajst-2981	2	57	,	,	PUNCT
ajst-2981	2	58	china	china	PROPN
ajst-2981	2	59	abstract	abstract	NOUN
ajst-2981	2	60	:	:	PUNCT
ajst-2981	2	61	traffic	traffic	NOUN
ajst-2981	2	62	flow	flow	NOUN
ajst-2981	2	63	statistics	statistic	NOUN
ajst-2981	2	64	,	,	PUNCT
ajst-2981	2	65	as	as	ADP
ajst-2981	2	66	an	an	DET
ajst-2981	2	67	important	important	ADJ
ajst-2981	2	68	part	part	NOUN
ajst-2981	2	69	of	of	ADP
ajst-2981	2	70	intelligent	intelligent	ADJ
ajst-2981	2	71	transportation	transportation	NOUN
ajst-2981	2	72	system	system	NOUN
ajst-2981	2	73	,	,	PUNCT
ajst-2981	2	74	usually	usually	ADV
ajst-2981	2	75	requires	require	VERB
ajst-2981	2	76	manual	manual	ADJ
ajst-2981	2	77	statistics	statistic	NOUN
ajst-2981	2	78	,	,	PUNCT
ajst-2981	2	79	which	which	PRON
ajst-2981	2	80	is	be	AUX
ajst-2981	2	81	time	time	NOUN
ajst-2981	2	82	-	-	PUNCT
ajst-2981	2	83	consuming	consume	VERB
ajst-2981	2	84	and	and	CCONJ
ajst-2981	2	85	labor	labor	NOUN
ajst-2981	2	86	-	-	PUNCT
ajst-2981	2	87	intensive	intensive	ADJ
ajst-2981	2	88	.	.	PUNCT
ajst-2981	3	1	in	in	ADP
ajst-2981	3	2	order	order	NOUN
ajst-2981	3	3	to	to	PART
ajst-2981	3	4	save	save	VERB
ajst-2981	3	5	manual	manual	ADJ
ajst-2981	3	6	labor	labor	NOUN
ajst-2981	3	7	and	and	CCONJ
ajst-2981	3	8	improve	improve	VERB
ajst-2981	3	9	the	the	DET
ajst-2981	3	10	statistical	statistical	ADJ
ajst-2981	3	11	efficiency	efficiency	NOUN
ajst-2981	3	12	,	,	PUNCT
ajst-2981	3	13	this	this	DET
ajst-2981	3	14	paper	paper	NOUN
ajst-2981	3	15	is	be	AUX
ajst-2981	3	16	based	base	VERB
ajst-2981	3	17	on	on	ADP
ajst-2981	3	18	the	the	DET
ajst-2981	3	19	strategy	strategy	NOUN
ajst-2981	3	20	of	of	ADP
ajst-2981	3	21	yolov5+deepsort	yolov5+deepsort	NOUN
ajst-2981	3	22	to	to	PART
ajst-2981	3	23	count	count	VERB
ajst-2981	3	24	the	the	DET
ajst-2981	3	25	aerial	aerial	ADJ
ajst-2981	3	26	traffic	traffic	NOUN
ajst-2981	3	27	flow	flow	NOUN
ajst-2981	3	28	by	by	ADP
ajst-2981	3	29	uav	uav	PROPN
ajst-2981	3	30	,	,	PUNCT
ajst-2981	3	31	and	and	CCONJ
ajst-2981	3	32	the	the	DET
ajst-2981	3	33	results	result	NOUN
ajst-2981	3	34	show	show	VERB
ajst-2981	3	35	that	that	SCONJ
ajst-2981	3	36	the	the	DET
ajst-2981	3	37	statistical	statistical	ADJ
ajst-2981	3	38	accuracy	accuracy	NOUN
ajst-2981	3	39	of	of	ADP
ajst-2981	3	40	this	this	DET
ajst-2981	3	41	method	method	NOUN
ajst-2981	3	42	is	be	AUX
ajst-2981	3	43	close	close	ADJ
ajst-2981	3	44	to	to	ADP
ajst-2981	3	45	that	that	PRON
ajst-2981	3	46	of	of	ADP
ajst-2981	3	47	manual	manual	ADJ
ajst-2981	3	48	statistics	statistic	NOUN
ajst-2981	3	49	,	,	PUNCT
ajst-2981	3	50	which	which	PRON
ajst-2981	3	51	has	have	VERB
ajst-2981	3	52	high	high	ADJ
ajst-2981	3	53	practical	practical	ADJ
ajst-2981	3	54	value	value	NOUN
ajst-2981	3	55	.	.	PUNCT
ajst-2981	4	1	keywords	keyword	NOUN
ajst-2981	4	2	:	:	PUNCT
ajst-2981	4	3	traffic	traffic	NOUN
ajst-2981	4	4	flow	flow	NOUN
ajst-2981	4	5	statistics	statistic	NOUN
ajst-2981	4	6	,	,	PUNCT
ajst-2981	4	7	intelligent	intelligent	ADJ
ajst-2981	4	8	traffic	traffic	NOUN
ajst-2981	4	9	system	system	NOUN
ajst-2981	4	10	,	,	PUNCT
ajst-2981	4	11	yolov5	yolov5	NOUN
ajst-2981	4	12	,	,	PUNCT
ajst-2981	4	13	deepsort	deepsort	NOUN
ajst-2981	4	14	,	,	PUNCT
ajst-2981	4	15	aerial	aerial	ADJ
ajst-2981	4	16	drone	drone	NOUN
ajst-2981	4	17	photography	photography	NOUN
ajst-2981	4	18	.	.	PUNCT
ajst-2981	5	1	1	1	X
ajst-2981	5	2	.	.	X
ajst-2981	5	3	introduction	introduction	NOUN
ajst-2981	5	4	as	as	ADP
ajst-2981	5	5	a	a	DET
ajst-2981	5	6	crucial	crucial	ADJ
ajst-2981	5	7	part	part	NOUN
ajst-2981	5	8	of	of	ADP
ajst-2981	5	9	intelligent	intelligent	ADJ
ajst-2981	5	10	transportation	transportation	NOUN
ajst-2981	5	11	system	system	NOUN
ajst-2981	5	12	[	[	X
ajst-2981	5	13	1	1	NUM
ajst-2981	5	14	]	]	PUNCT
ajst-2981	5	15	,	,	PUNCT
ajst-2981	5	16	the	the	DET
ajst-2981	5	17	use	use	NOUN
ajst-2981	5	18	of	of	ADP
ajst-2981	5	19	video	video	NOUN
ajst-2981	5	20	for	for	ADP
ajst-2981	5	21	traffic	traffic	NOUN
ajst-2981	5	22	statistics	statistic	NOUN
ajst-2981	5	23	is	be	AUX
ajst-2981	5	24	a	a	DET
ajst-2981	5	25	research	research	NOUN
ajst-2981	5	26	hotspot	hotspot	NOUN
ajst-2981	5	27	.	.	PUNCT
ajst-2981	6	1	as	as	ADP
ajst-2981	6	2	a	a	DET
ajst-2981	6	3	new	new	ADJ
ajst-2981	6	4	data	data	NOUN
ajst-2981	6	5	acquisition	acquisition	NOUN
ajst-2981	6	6	source	source	NOUN
ajst-2981	6	7	,	,	PUNCT
ajst-2981	6	8	uav	uav	PROPN
ajst-2981	6	9	has	have	VERB
ajst-2981	6	10	unique	unique	ADJ
ajst-2981	6	11	advantages	advantage	NOUN
ajst-2981	6	12	in	in	ADP
ajst-2981	6	13	monitoring	monitor	VERB
ajst-2981	6	14	road	road	NOUN
ajst-2981	6	15	traffic	traffic	NOUN
ajst-2981	6	16	because	because	SCONJ
ajst-2981	6	17	of	of	ADP
ajst-2981	6	18	its	its	PRON
ajst-2981	6	19	characteristics	characteristic	NOUN
ajst-2981	6	20	of	of	ADP
ajst-2981	6	21	high	high	ADJ
ajst-2981	6	22	mobility	mobility	NOUN
ajst-2981	6	23	,	,	PUNCT
ajst-2981	6	24	large	large	ADJ
ajst-2981	6	25	range	range	NOUN
ajst-2981	6	26	,	,	PUNCT
ajst-2981	6	27	high	high	ADJ
ajst-2981	6	28	resolution	resolution	NOUN
ajst-2981	6	29	and	and	CCONJ
ajst-2981	6	30	strong	strong	ADJ
ajst-2981	6	31	adaptability	adaptability	NOUN
ajst-2981	6	32	,	,	PUNCT
ajst-2981	6	33	and	and	CCONJ
ajst-2981	6	34	plays	play	VERB
ajst-2981	6	35	a	a	DET
ajst-2981	6	36	significant	significant	ADJ
ajst-2981	6	37	role	role	NOUN
ajst-2981	6	38	in	in	ADP
ajst-2981	6	39	the	the	DET
ajst-2981	6	40	development	development	NOUN
ajst-2981	6	41	of	of	ADP
ajst-2981	6	42	transportation	transportation	NOUN
ajst-2981	6	43	strategy	strategy	NOUN
ajst-2981	6	44	.	.	PUNCT
ajst-2981	7	1	therefore	therefore	ADV
ajst-2981	7	2	,	,	PUNCT
ajst-2981	7	3	traffic	traffic	NOUN
ajst-2981	7	4	statistics	statistic	NOUN
ajst-2981	7	5	based	base	VERB
ajst-2981	7	6	on	on	ADP
ajst-2981	7	7	uav	uav	PROPN
ajst-2981	7	8	aerial	aerial	ADJ
ajst-2981	7	9	video	video	NOUN
ajst-2981	7	10	has	have	AUX
ajst-2981	7	11	become	become	VERB
ajst-2981	7	12	the	the	DET
ajst-2981	7	13	development	development	NOUN
ajst-2981	7	14	direction	direction	NOUN
ajst-2981	7	15	of	of	ADP
ajst-2981	7	16	intelligent	intelligent	ADJ
ajst-2981	7	17	transportation	transportation	NOUN
ajst-2981	7	18	technology	technology	NOUN
ajst-2981	7	19	in	in	ADP
ajst-2981	7	20	the	the	DET
ajst-2981	7	21	future	future	NOUN
ajst-2981	7	22	,	,	PUNCT
ajst-2981	7	23	and	and	CCONJ
ajst-2981	7	24	has	have	VERB
ajst-2981	7	25	great	great	ADJ
ajst-2981	7	26	development	development	NOUN
ajst-2981	7	27	and	and	CCONJ
ajst-2981	7	28	application	application	NOUN
ajst-2981	7	29	potential	potential	NOUN
ajst-2981	7	30	.	.	PUNCT
ajst-2981	8	1	traffic	traffic	NOUN
ajst-2981	8	2	statistics	statistic	NOUN
ajst-2981	8	3	based	base	VERB
ajst-2981	8	4	on	on	ADP
ajst-2981	8	5	drone	drone	NOUN
ajst-2981	8	6	aerial	aerial	ADJ
ajst-2981	8	7	videos	video	NOUN
ajst-2981	8	8	usually	usually	ADV
ajst-2981	8	9	consist	consist	VERB
ajst-2981	8	10	of	of	ADP
ajst-2981	8	11	two	two	NUM
ajst-2981	8	12	parts	part	NOUN
ajst-2981	8	13	,	,	PUNCT
ajst-2981	8	14	namely	namely	ADV
ajst-2981	8	15	vehicle	vehicle	NOUN
ajst-2981	8	16	detection	detection	NOUN
ajst-2981	8	17	and	and	CCONJ
ajst-2981	8	18	vehicle	vehicle	NOUN
ajst-2981	8	19	tracking	tracking	NOUN
ajst-2981	8	20	.	.	PUNCT
ajst-2981	9	1	traditional	traditional	ADJ
ajst-2981	9	2	object	object	NOUN
ajst-2981	9	3	detection	detection	NOUN
ajst-2981	9	4	algorithms	algorithm	NOUN
ajst-2981	9	5	such	such	ADJ
ajst-2981	9	6	as	as	ADP
ajst-2981	9	7	haar+adaboost[2	haar+adaboost[2	PROPN
ajst-2981	9	8	]	]	X
ajst-2981	9	9	,	,	PUNCT
ajst-2981	9	10	hog+svm[3	hog+svm[3	NOUN
ajst-2981	9	11	]	]	PUNCT
ajst-2981	9	12	,	,	PUNCT
ajst-2981	9	13	etc	etc	X
ajst-2981	9	14	.	.	X
ajst-2981	9	15	manually	manually	ADV
ajst-2981	9	16	extract	extract	VERB
ajst-2981	9	17	target	target	NOUN
ajst-2981	9	18	features	feature	NOUN
ajst-2981	9	19	,	,	PUNCT
ajst-2981	9	20	and	and	CCONJ
ajst-2981	9	21	then	then	ADV
ajst-2981	9	22	use	use	VERB
ajst-2981	9	23	classification	classification	NOUN
ajst-2981	9	24	algorithms	algorithm	NOUN
ajst-2981	9	25	for	for	ADP
ajst-2981	9	26	classification	classification	NOUN
ajst-2981	9	27	and	and	CCONJ
ajst-2981	9	28	discrimination	discrimination	NOUN
ajst-2981	9	29	,	,	PUNCT
ajst-2981	9	30	these	these	DET
ajst-2981	9	31	algorithms	algorithm	NOUN
ajst-2981	9	32	have	have	VERB
ajst-2981	9	33	slow	slow	ADJ
ajst-2981	9	34	detection	detection	NOUN
ajst-2981	9	35	speed	speed	NOUN
ajst-2981	9	36	,	,	PUNCT
ajst-2981	9	37	poor	poor	ADJ
ajst-2981	9	38	detection	detection	NOUN
ajst-2981	9	39	effect	effect	NOUN
ajst-2981	9	40	and	and	CCONJ
ajst-2981	9	41	high	high	ADJ
ajst-2981	9	42	resource	resource	NOUN
ajst-2981	9	43	consumption	consumption	NOUN
ajst-2981	9	44	,	,	PUNCT
ajst-2981	9	45	and	and	CCONJ
ajst-2981	9	46	manual	manual	ADJ
ajst-2981	9	47	extraction	extraction	NOUN
ajst-2981	9	48	of	of	ADP
ajst-2981	9	49	features	feature	NOUN
ajst-2981	9	50	leads	lead	VERB
ajst-2981	9	51	to	to	ADP
ajst-2981	9	52	frequent	frequent	ADJ
ajst-2981	9	53	false	false	ADJ
ajst-2981	9	54	detection	detection	NOUN
ajst-2981	9	55	and	and	CCONJ
ajst-2981	9	56	missed	miss	VERB
ajst-2981	9	57	detection	detection	NOUN
ajst-2981	9	58	.	.	PUNCT
ajst-2981	10	1	with	with	ADP
ajst-2981	10	2	the	the	DET
ajst-2981	10	3	continuous	continuous	ADJ
ajst-2981	10	4	development	development	NOUN
ajst-2981	10	5	of	of	ADP
ajst-2981	10	6	machine	machine	NOUN
ajst-2981	10	7	learning	learning	NOUN
ajst-2981	10	8	and	and	CCONJ
ajst-2981	10	9	computer	computer	NOUN
ajst-2981	10	10	computing	computing	NOUN
ajst-2981	10	11	power	power	NOUN
ajst-2981	10	12	,	,	PUNCT
ajst-2981	10	13	traditional	traditional	ADJ
ajst-2981	10	14	algorithms	algorithm	NOUN
ajst-2981	10	15	have	have	AUX
ajst-2981	10	16	been	be	AUX
ajst-2981	10	17	difficult	difficult	ADJ
ajst-2981	10	18	to	to	PART
ajst-2981	10	19	meet	meet	VERB
ajst-2981	10	20	actual	actual	ADJ
ajst-2981	10	21	needs	need	NOUN
ajst-2981	10	22	,	,	PUNCT
ajst-2981	10	23	so	so	ADV
ajst-2981	10	24	object	object	NOUN
ajst-2981	10	25	detection	detection	NOUN
ajst-2981	10	26	has	have	AUX
ajst-2981	10	27	gradually	gradually	ADV
ajst-2981	10	28	shifted	shift	VERB
ajst-2981	10	29	from	from	ADP
ajst-2981	10	30	traditional	traditional	ADJ
ajst-2981	10	31	algorithms	algorithm	NOUN
ajst-2981	10	32	to	to	ADP
ajst-2981	10	33	methods	method	NOUN
ajst-2981	10	34	based	base	VERB
ajst-2981	10	35	on	on	ADP
ajst-2981	10	36	deep	deep	ADJ
ajst-2981	10	37	learning	learning	NOUN
ajst-2981	10	38	.	.	PUNCT
ajst-2981	11	1	there	there	PRON
ajst-2981	11	2	are	be	VERB
ajst-2981	11	3	two	two	NUM
ajst-2981	11	4	types	type	NOUN
ajst-2981	11	5	of	of	ADP
ajst-2981	11	6	methods	method	NOUN
ajst-2981	11	7	based	base	VERB
ajst-2981	11	8	on	on	ADP
ajst-2981	11	9	deep	deep	ADJ
ajst-2981	11	10	learning	learning	NOUN
ajst-2981	11	11	:	:	PUNCT
ajst-2981	11	12	one	one	NUM
ajst-2981	11	13	-	-	PUNCT
ajst-2981	11	14	stage	stage	NOUN
ajst-2981	11	15	and	and	CCONJ
ajst-2981	11	16	two	two	NUM
ajst-2981	11	17	-	-	PUNCT
ajst-2981	11	18	stage	stage	NOUN
ajst-2981	11	19	.	.	PUNCT
ajst-2981	12	1	two	two	NUM
ajst-2981	12	2	-	-	PUNCT
ajst-2981	12	3	stage	stage	NOUN
ajst-2981	12	4	is	be	AUX
ajst-2981	12	5	characterized	characterize	VERB
ajst-2981	12	6	by	by	ADP
ajst-2981	12	7	obtaining	obtain	VERB
ajst-2981	12	8	candidate	candidate	NOUN
ajst-2981	12	9	regions	region	NOUN
ajst-2981	12	10	first	first	ADV
ajst-2981	12	11	,	,	PUNCT
ajst-2981	12	12	and	and	CCONJ
ajst-2981	12	13	then	then	ADV
ajst-2981	12	14	performing	perform	VERB
ajst-2981	12	15	object	object	NOUN
ajst-2981	12	16	detection	detection	NOUN
ajst-2981	12	17	,	,	PUNCT
ajst-2981	12	18	with	with	ADP
ajst-2981	12	19	high	high	ADJ
ajst-2981	12	20	accuracy	accuracy	NOUN
ajst-2981	12	21	,	,	PUNCT
ajst-2981	12	22	the	the	DET
ajst-2981	12	23	most	most	ADV
ajst-2981	12	24	representative	representative	NOUN
ajst-2981	12	25	of	of	ADP
ajst-2981	12	26	which	which	PRON
ajst-2981	12	27	is	be	AUX
ajst-2981	12	28	the	the	DET
ajst-2981	12	29	r	r	PROPN
ajst-2981	12	30	-	-	PUNCT
ajst-2981	12	31	cnn	cnn	PROPN
ajst-2981	12	32	algorithm	algorithm	NOUN
ajst-2981	12	33	formed	form	VERB
ajst-2981	12	34	by	by	ADP
ajst-2981	12	35	combining	combine	VERB
ajst-2981	12	36	region	region	NOUN
ajst-2981	12	37	proposal	proposal	NOUN
ajst-2981	12	38	and	and	CCONJ
ajst-2981	12	39	cnn	cnn	PROPN
ajst-2981	12	40	in	in	ADP
ajst-2981	12	41	2015	2015	NUM
ajst-2981	12	42	by	by	ADP
ajst-2981	12	43	glrshick	glrshick	PROPN
ajst-2981	12	44	et	et	PROPN
ajst-2981	12	45	al	al	PROPN
ajst-2981	12	46	.	.	PUNCT
ajst-2981	13	1	[	[	X
ajst-2981	13	2	4	4	NUM
ajst-2981	13	3	]	]	PUNCT
ajst-2981	13	4	.	.	PUNCT
ajst-2981	14	1	subsequent	subsequent	ADJ
ajst-2981	14	2	fast	fast	ADJ
ajst-2981	14	3	-	-	PUNCT
ajst-2981	14	4	rcnn	rcnn	NOUN
ajst-2981	15	1	[	[	X
ajst-2981	15	2	5	5	NUM
ajst-2981	15	3	]	]	PUNCT
ajst-2981	15	4	and	and	CCONJ
ajst-2981	15	5	faster	fast	ADV
ajst-2981	15	6	-	-	PUNCT
ajst-2981	15	7	rcnn	rcnn	NOUN
ajst-2981	15	8	[	[	X
ajst-2981	15	9	6	6	NUM
ajst-2981	15	10	]	]	PUNCT
ajst-2981	15	11	were	be	AUX
ajst-2981	15	12	developed	develop	VERB
ajst-2981	15	13	on	on	ADP
ajst-2981	15	14	the	the	DET
ajst-2981	15	15	basis	basis	NOUN
ajst-2981	15	16	of	of	ADP
ajst-2981	15	17	the	the	DET
ajst-2981	15	18	r	r	PROPN
ajst-2981	15	19	-	-	PUNCT
ajst-2981	15	20	cnn	cnn	PROPN
ajst-2981	15	21	algorithm	algorithm	NOUN
ajst-2981	15	22	.	.	PUNCT
ajst-2981	16	1	one	one	NUM
ajst-2981	16	2	-	-	PUNCT
ajst-2981	16	3	stage	stage	NOUN
ajst-2981	16	4	is	be	AUX
ajst-2981	16	5	characterized	characterize	VERB
ajst-2981	16	6	by	by	ADP
ajst-2981	16	7	direct	direct	ADJ
ajst-2981	16	8	classification	classification	NOUN
ajst-2981	16	9	and	and	CCONJ
ajst-2981	16	10	regression	regression	NOUN
ajst-2981	16	11	of	of	ADP
ajst-2981	16	12	input	input	NOUN
ajst-2981	16	13	targets	target	NOUN
ajst-2981	16	14	,	,	PUNCT
ajst-2981	16	15	representing	represent	VERB
ajst-2981	16	16	models	model	NOUN
ajst-2981	16	17	such	such	ADJ
ajst-2981	16	18	as	as	ADP
ajst-2981	16	19	ssd	ssd	NOUN
ajst-2981	16	20	[	[	X
ajst-2981	16	21	7	7	NUM
ajst-2981	16	22	]	]	PUNCT
ajst-2981	16	23	and	and	CCONJ
ajst-2981	16	24	yolo	yolo	ADJ
ajst-2981	16	25	[	[	X
ajst-2981	16	26	810	810	NUM
ajst-2981	16	27	]	]	PUNCT
ajst-2981	16	28	series	series	NOUN
ajst-2981	16	29	.	.	PUNCT
ajst-2981	17	1	at	at	ADP
ajst-2981	17	2	present	present	ADJ
ajst-2981	17	3	,	,	PUNCT
ajst-2981	17	4	the	the	DET
ajst-2981	17	5	yolo	yolo	ADJ
ajst-2981	17	6	model	model	NOUN
ajst-2981	17	7	has	have	AUX
ajst-2981	17	8	been	be	AUX
ajst-2981	17	9	iterated	iterate	VERB
ajst-2981	17	10	to	to	ADP
ajst-2981	17	11	yolov5	yolov5	NOUN
ajst-2981	17	12	,	,	PUNCT
ajst-2981	17	13	released	release	VERB
ajst-2981	17	14	by	by	ADP
ajst-2981	17	15	ultralytics	ultralytic	NOUN
ajst-2981	17	16	,	,	PUNCT
ajst-2981	17	17	yolov5	yolov5	NOUN
ajst-2981	17	18	contains	contain	VERB
ajst-2981	17	19	a	a	DET
ajst-2981	17	20	variety	variety	NOUN
ajst-2981	17	21	of	of	ADP
ajst-2981	17	22	models	model	NOUN
ajst-2981	17	23	,	,	PUNCT
ajst-2981	17	24	such	such	ADJ
ajst-2981	17	25	as	as	ADP
ajst-2981	17	26	yolov5s	yolov5s	PROPN
ajst-2981	17	27	and	and	CCONJ
ajst-2981	17	28	yolov5	yolov5	PROPN
ajst-2981	17	29	m	m	PROPN
ajst-2981	17	30	models	model	NOUN
ajst-2981	17	31	,	,	PUNCT
ajst-2981	17	32	the	the	DET
ajst-2981	17	33	main	main	ADJ
ajst-2981	17	34	difference	difference	NOUN
ajst-2981	17	35	is	be	AUX
ajst-2981	17	36	that	that	SCONJ
ajst-2981	17	37	the	the	DET
ajst-2981	17	38	depth	depth	NOUN
ajst-2981	17	39	and	and	CCONJ
ajst-2981	17	40	width	width	NOUN
ajst-2981	17	41	of	of	ADP
ajst-2981	17	42	the	the	DET
ajst-2981	17	43	network	network	NOUN
ajst-2981	17	44	can	can	AUX
ajst-2981	17	45	be	be	AUX
ajst-2981	17	46	controlled	control	VERB
ajst-2981	17	47	to	to	PART
ajst-2981	17	48	obtain	obtain	VERB
ajst-2981	17	49	models	model	NOUN
ajst-2981	17	50	of	of	ADP
ajst-2981	17	51	different	different	ADJ
ajst-2981	17	52	sizes	size	NOUN
ajst-2981	17	53	.	.	PUNCT
ajst-2981	18	1	in	in	ADP
ajst-2981	18	2	order	order	NOUN
ajst-2981	18	3	to	to	PART
ajst-2981	18	4	balance	balance	VERB
ajst-2981	18	5	accuracy	accuracy	NOUN
ajst-2981	18	6	with	with	ADP
ajst-2981	18	7	computational	computational	ADJ
ajst-2981	18	8	burden	burden	NOUN
ajst-2981	18	9	,	,	PUNCT
ajst-2981	18	10	this	this	DET
ajst-2981	18	11	paper	paper	NOUN
ajst-2981	18	12	uses	use	VERB
ajst-2981	18	13	yolov5x	yolov5x	PROPN
ajst-2981	18	14	as	as	ADP
ajst-2981	18	15	the	the	DET
ajst-2981	18	16	benchmark	benchmark	NOUN
ajst-2981	18	17	model	model	NOUN
ajst-2981	18	18	.	.	PUNCT
ajst-2981	19	1	the	the	DET
ajst-2981	19	2	target	target	NOUN
ajst-2981	19	3	tracking	tracking	NOUN
ajst-2981	19	4	algorithm	algorithm	NOUN
ajst-2981	19	5	can	can	AUX
ajst-2981	19	6	be	be	AUX
ajst-2981	19	7	divided	divide	VERB
ajst-2981	19	8	into	into	ADP
ajst-2981	19	9	detectionbased	detectionbase	VERB
ajst-2981	19	10	tracking	tracking	NOUN
ajst-2981	19	11	and	and	CCONJ
ajst-2981	19	12	detection	detection	NOUN
ajst-2981	19	13	-	-	PUNCT
ajst-2981	19	14	free	free	ADJ
ajst-2981	19	15	tracking	tracking	NOUN
ajst-2981	19	16	according	accord	VERB
ajst-2981	19	17	to	to	ADP
ajst-2981	19	18	the	the	DET
ajst-2981	19	19	initialization	initialization	NOUN
ajst-2981	19	20	method	method	NOUN
ajst-2981	19	21	.	.	PUNCT
ajst-2981	20	1	detection	detection	NOUN
ajst-2981	20	2	-	-	PUNCT
ajst-2981	20	3	based	base	VERB
ajst-2981	20	4	tracking	tracking	NOUN
ajst-2981	20	5	requires	require	VERB
ajst-2981	20	6	a	a	DET
ajst-2981	20	7	detector	detector	NOUN
ajst-2981	20	8	to	to	PART
ajst-2981	20	9	detect	detect	VERB
ajst-2981	20	10	the	the	DET
ajst-2981	20	11	object	object	NOUN
ajst-2981	20	12	in	in	ADP
ajst-2981	20	13	the	the	DET
ajst-2981	20	14	image	image	NOUN
ajst-2981	20	15	in	in	ADP
ajst-2981	20	16	advance	advance	NOUN
ajst-2981	20	17	,	,	PUNCT
ajst-2981	20	18	and	and	CCONJ
ajst-2981	20	19	then	then	ADV
ajst-2981	20	20	track	track	VERB
ajst-2981	20	21	the	the	DET
ajst-2981	20	22	detected	detect	VERB
ajst-2981	20	23	target	target	NOUN
ajst-2981	20	24	,	,	PUNCT
ajst-2981	20	25	so	so	CCONJ
ajst-2981	20	26	the	the	DET
ajst-2981	20	27	good	good	ADJ
ajst-2981	20	28	or	or	CCONJ
ajst-2981	20	29	bad	bad	ADJ
ajst-2981	20	30	detection	detection	NOUN
ajst-2981	20	31	effect	effect	NOUN
ajst-2981	20	32	has	have	VERB
ajst-2981	20	33	a	a	DET
ajst-2981	20	34	great	great	ADJ
ajst-2981	20	35	impact	impact	NOUN
ajst-2981	20	36	on	on	ADP
ajst-2981	20	37	the	the	DET
ajst-2981	20	38	tracking	tracking	NOUN
ajst-2981	20	39	effect	effect	NOUN
ajst-2981	20	40	;	;	PUNCT
ajst-2981	20	41	tracking	track	VERB
ajst-2981	20	42	without	without	ADP
ajst-2981	20	43	detection	detection	NOUN
ajst-2981	20	44	only	only	ADV
ajst-2981	20	45	requires	require	VERB
ajst-2981	20	46	manual	manual	ADJ
ajst-2981	20	47	annotation	annotation	NOUN
ajst-2981	20	48	of	of	ADP
ajst-2981	20	49	tracked	track	VERB
ajst-2981	20	50	targets	target	NOUN
ajst-2981	20	51	in	in	ADP
ajst-2981	20	52	the	the	DET
ajst-2981	20	53	initial	initial	ADJ
ajst-2981	20	54	frame	frame	NOUN
ajst-2981	20	55	,	,	PUNCT
ajst-2981	20	56	and	and	CCONJ
ajst-2981	20	57	the	the	DET
ajst-2981	20	58	algorithm	algorithm	NOUN
ajst-2981	20	59	is	be	AUX
ajst-2981	20	60	not	not	PART
ajst-2981	20	61	flexible	flexible	ADJ
ajst-2981	20	62	,	,	PUNCT
ajst-2981	20	63	and	and	CCONJ
ajst-2981	20	64	it	it	PRON
ajst-2981	20	65	can	can	AUX
ajst-2981	20	66	not	not	PART
ajst-2981	20	67	track	track	VERB
ajst-2981	20	68	different	different	ADJ
ajst-2981	20	69	targets	target	NOUN
ajst-2981	20	70	in	in	ADP
ajst-2981	20	71	subsequent	subsequent	ADJ
ajst-2981	20	72	frames	frame	NOUN
ajst-2981	20	73	.	.	PUNCT
ajst-2981	21	1	object	object	NOUN
ajst-2981	21	2	tracking	tracking	NOUN
ajst-2981	21	3	can	can	AUX
ajst-2981	21	4	also	also	ADV
ajst-2981	21	5	be	be	AUX
ajst-2981	21	6	divided	divide	VERB
ajst-2981	21	7	into	into	ADP
ajst-2981	21	8	online	online	ADJ
ajst-2981	21	9	tracking	tracking	NOUN
ajst-2981	21	10	and	and	CCONJ
ajst-2981	21	11	offline	offline	ADJ
ajst-2981	21	12	tracking	tracking	NOUN
ajst-2981	21	13	according	accord	VERB
ajst-2981	21	14	to	to	ADP
ajst-2981	21	15	how	how	SCONJ
ajst-2981	21	16	video	video	NOUN
ajst-2981	21	17	frames	frame	NOUN
ajst-2981	21	18	are	be	AUX
ajst-2981	21	19	processed	process	VERB
ajst-2981	21	20	.	.	PUNCT
ajst-2981	22	1	online	online	ADJ
ajst-2981	22	2	tracking	tracking	NOUN
ajst-2981	22	3	can	can	AUX
ajst-2981	22	4	not	not	PART
ajst-2981	22	5	process	process	VERB
ajst-2981	22	6	previous	previous	ADJ
ajst-2981	22	7	frame	frame	NOUN
ajst-2981	22	8	tracking	tracking	NOUN
ajst-2981	22	9	results	result	NOUN
ajst-2981	22	10	based	base	VERB
ajst-2981	22	11	on	on	ADP
ajst-2981	22	12	current	current	ADJ
ajst-2981	22	13	frame	frame	NOUN
ajst-2981	22	14	information	information	NOUN
ajst-2981	22	15	;	;	PUNCT
ajst-2981	22	16	offline	offline	ADJ
ajst-2981	22	17	tracking	tracking	NOUN
ajst-2981	22	18	can	can	AUX
ajst-2981	22	19	use	use	VERB
ajst-2981	22	20	the	the	DET
ajst-2981	22	21	data	datum	NOUN
ajst-2981	22	22	before	before	ADV
ajst-2981	22	23	and	and	CCONJ
ajst-2981	22	24	after	after	ADP
ajst-2981	22	25	the	the	DET
ajst-2981	22	26	current	current	ADJ
ajst-2981	22	27	frame	frame	NOUN
ajst-2981	22	28	to	to	PART
ajst-2981	22	29	obtain	obtain	VERB
ajst-2981	22	30	the	the	DET
ajst-2981	22	31	global	global	ADJ
ajst-2981	22	32	optimal	optimal	ADJ
ajst-2981	22	33	solution	solution	NOUN
ajst-2981	22	34	,	,	PUNCT
ajst-2981	22	35	but	but	CCONJ
ajst-2981	22	36	it	it	PRON
ajst-2981	22	37	is	be	AUX
ajst-2981	22	38	not	not	PART
ajst-2981	22	39	suitable	suitable	ADJ
ajst-2981	22	40	for	for	ADP
ajst-2981	22	41	practical	practical	ADJ
ajst-2981	22	42	applications	application	NOUN
ajst-2981	22	43	.	.	PUNCT
ajst-2981	23	1	in	in	ADP
ajst-2981	23	2	the	the	DET
ajst-2981	23	3	vehicle	vehicle	NOUN
ajst-2981	23	4	tracking	tracking	NOUN
ajst-2981	23	5	task	task	NOUN
ajst-2981	23	6	of	of	ADP
ajst-2981	23	7	uav	uav	PROPN
ajst-2981	23	8	aerial	aerial	ADJ
ajst-2981	23	9	video	video	NOUN
ajst-2981	23	10	,	,	PUNCT
ajst-2981	23	11	considering	consider	VERB
ajst-2981	23	12	real	real	ADJ
ajst-2981	23	13	-	-	PUNCT
ajst-2981	23	14	time	time	NOUN
ajst-2981	23	15	and	and	CCONJ
ajst-2981	23	16	flexibility	flexibility	NOUN
ajst-2981	23	17	,	,	PUNCT
ajst-2981	23	18	detection	detection	NOUN
ajst-2981	23	19	-	-	PUNCT
ajst-2981	23	20	based	base	VERB
ajst-2981	23	21	online	online	ADJ
ajst-2981	23	22	tracking	tracking	NOUN
ajst-2981	23	23	is	be	AUX
ajst-2981	23	24	the	the	DET
ajst-2981	23	25	closest	close	ADJ
ajst-2981	23	26	method	method	NOUN
ajst-2981	23	27	to	to	ADP
ajst-2981	23	28	practical	practical	ADJ
ajst-2981	23	29	application	application	NOUN
ajst-2981	23	30	.	.	PUNCT
ajst-2981	24	1	in	in	ADP
ajst-2981	24	2	this	this	DET
ajst-2981	24	3	paper	paper	NOUN
ajst-2981	24	4	,	,	PUNCT
ajst-2981	24	5	the	the	DET
ajst-2981	24	6	yolov5	yolov5	NOUN
ajst-2981	24	7	object	object	NOUN
ajst-2981	24	8	detection	detection	NOUN
ajst-2981	24	9	algorithm	algorithm	NOUN
ajst-2981	24	10	based	base	VERB
ajst-2981	24	11	on	on	ADP
ajst-2981	24	12	deep	deep	ADJ
ajst-2981	24	13	learning	learning	NOUN
ajst-2981	24	14	is	be	AUX
ajst-2981	24	15	used	use	VERB
ajst-2981	24	16	to	to	PART
ajst-2981	24	17	detect	detect	VERB
ajst-2981	24	18	vehicles	vehicle	NOUN
ajst-2981	24	19	,	,	PUNCT
ajst-2981	24	20	combined	combine	VERB
ajst-2981	24	21	with	with	ADP
ajst-2981	24	22	the	the	DET
ajst-2981	24	23	detection	detection	NOUN
ajst-2981	24	24	-	-	PUNCT
ajst-2981	24	25	based	base	VERB
ajst-2981	24	26	deepsort	deepsort	NOUN
ajst-2981	24	27	target	target	NOUN
ajst-2981	24	28	tracking	tracking	NOUN
ajst-2981	24	29	algorithm	algorithm	NOUN
ajst-2981	24	30	of	of	ADP
ajst-2981	24	31	online	online	ADJ
ajst-2981	24	32	tracking	tracking	NOUN
ajst-2981	24	33	,	,	PUNCT
ajst-2981	24	34	and	and	CCONJ
ajst-2981	24	35	the	the	DET
ajst-2981	24	36	number	number	NOUN
ajst-2981	24	37	of	of	ADP
ajst-2981	24	38	vehicles	vehicle	NOUN
ajst-2981	24	39	in	in	ADP
ajst-2981	24	40	uav	uav	PROPN
ajst-2981	24	41	aerial	aerial	ADJ
ajst-2981	24	42	video	video	NOUN
ajst-2981	24	43	is	be	AUX
ajst-2981	24	44	completed	complete	VERB
ajst-2981	24	45	by	by	ADP
ajst-2981	24	46	using	use	VERB
ajst-2981	24	47	virtual	virtual	ADJ
ajst-2981	24	48	detection	detection	NOUN
ajst-2981	24	49	lines	line	NOUN
ajst-2981	24	50	.	.	PUNCT
ajst-2981	25	1	2	2	X
ajst-2981	25	2	.	.	X
ajst-2981	25	3	yolov5	yolov5	NOUN
ajst-2981	25	4	vehicle	vehicle	NOUN
ajst-2981	25	5	detection	detection	NOUN
ajst-2981	25	6	algorithm	algorithm	NOUN
ajst-2981	25	7	yolov5	yolov5	NOUN
ajst-2981	25	8	consists	consist	VERB
ajst-2981	25	9	of	of	ADP
ajst-2981	25	10	four	four	NUM
ajst-2981	25	11	parts	part	NOUN
ajst-2981	25	12	:	:	PUNCT
ajst-2981	25	13	input	input	NOUN
ajst-2981	25	14	,	,	PUNCT
ajst-2981	25	15	backbone	backbone	NOUN
ajst-2981	25	16	network	network	NOUN
ajst-2981	25	17	,	,	PUNCT
ajst-2981	25	18	neck	neck	NOUN
ajst-2981	25	19	network	network	NOUN
ajst-2981	25	20	and	and	CCONJ
ajst-2981	25	21	prediction	prediction	NOUN
ajst-2981	25	22	.	.	PUNCT
ajst-2981	26	1	the	the	DET
ajst-2981	26	2	network	network	NOUN
ajst-2981	26	3	structure	structure	NOUN
ajst-2981	26	4	is	be	AUX
ajst-2981	26	5	shown	show	VERB
ajst-2981	26	6	in	in	ADP
ajst-2981	26	7	figure	figure	NOUN
ajst-2981	26	8	1	1	NUM
ajst-2981	26	9	.	.	NOUN
ajst-2981	26	10	199	199	NUM
ajst-2981	26	11	figure	figure	NOUN
ajst-2981	26	12	1	1	NUM
ajst-2981	26	13	.	.	PUNCT
ajst-2981	27	1	yolov5	yolov5	NOUN
ajst-2981	27	2	network	network	NOUN
ajst-2981	27	3	structure	structure	NOUN
ajst-2981	27	4	2.1	2.1	NUM
ajst-2981	27	5	.	.	PUNCT
ajst-2981	28	1	input	input	NOUN
ajst-2981	28	2	terminal	terminal	VERB
ajst-2981	28	3	the	the	DET
ajst-2981	28	4	input	input	NOUN
ajst-2981	28	5	end	end	NOUN
ajst-2981	28	6	consists	consist	VERB
ajst-2981	28	7	of	of	ADP
ajst-2981	28	8	three	three	NUM
ajst-2981	28	9	parts	part	NOUN
ajst-2981	28	10	:	:	PUNCT
ajst-2981	28	11	mosaic	mosaic	ADJ
ajst-2981	28	12	data	datum	NOUN
ajst-2981	28	13	reinforcement	reinforcement	NOUN
ajst-2981	28	14	,	,	PUNCT
ajst-2981	28	15	adaptive	adaptive	ADJ
ajst-2981	28	16	anchor	anchor	NOUN
ajst-2981	28	17	frame	frame	NOUN
ajst-2981	28	18	calculation	calculation	NOUN
ajst-2981	28	19	and	and	CCONJ
ajst-2981	28	20	adaptive	adaptive	ADJ
ajst-2981	28	21	image	image	NOUN
ajst-2981	28	22	scaling	scaling	NOUN
ajst-2981	28	23	.	.	PUNCT
ajst-2981	29	1	mosaic	mosaic	ADJ
ajst-2981	29	2	data	datum	NOUN
ajst-2981	29	3	enhancement	enhancement	NOUN
ajst-2981	29	4	is	be	AUX
ajst-2981	29	5	to	to	PART
ajst-2981	29	6	randomly	randomly	ADV
ajst-2981	29	7	zoom	zoom	VERB
ajst-2981	29	8	,	,	PUNCT
ajst-2981	29	9	randomly	randomly	ADV
ajst-2981	29	10	crop	crop	NOUN
ajst-2981	29	11	,	,	PUNCT
ajst-2981	29	12	randomly	randomly	ADV
ajst-2981	29	13	arrange	arrange	NOUN
ajst-2981	29	14	and	and	CCONJ
ajst-2981	29	15	finally	finally	ADV
ajst-2981	29	16	stitch	stitch	VERB
ajst-2981	29	17	and	and	CCONJ
ajst-2981	29	18	combine	combine	VERB
ajst-2981	29	19	the	the	DET
ajst-2981	29	20	4	4	NUM
ajst-2981	29	21	randomly	randomly	ADV
ajst-2981	29	22	selected	select	VERB
ajst-2981	29	23	pictures	picture	NOUN
ajst-2981	29	24	,	,	PUNCT
ajst-2981	29	25	which	which	PRON
ajst-2981	29	26	has	have	VERB
ajst-2981	29	27	the	the	DET
ajst-2981	29	28	advantage	advantage	NOUN
ajst-2981	29	29	of	of	ADP
ajst-2981	29	30	enriching	enrich	VERB
ajst-2981	29	31	the	the	DET
ajst-2981	29	32	dataset	dataset	NOUN
ajst-2981	29	33	and	and	CCONJ
ajst-2981	29	34	reducing	reduce	VERB
ajst-2981	29	35	the	the	DET
ajst-2981	29	36	gpu	gpu	PROPN
ajst-2981	29	37	.	.	PROPN
ajst-2981	29	38	adaptive	adaptive	PROPN
ajst-2981	29	39	anchor	anchor	PROPN
ajst-2981	29	40	box	box	NOUN
ajst-2981	29	41	calculation	calculation	NOUN
ajst-2981	29	42	is	be	AUX
ajst-2981	29	43	mainly	mainly	ADV
ajst-2981	29	44	used	use	VERB
ajst-2981	29	45	to	to	PART
ajst-2981	29	46	output	output	VERB
ajst-2981	29	47	prediction	prediction	NOUN
ajst-2981	29	48	boxes	box	NOUN
ajst-2981	29	49	on	on	ADP
ajst-2981	29	50	the	the	DET
ajst-2981	29	51	basis	basis	NOUN
ajst-2981	29	52	of	of	ADP
ajst-2981	29	53	the	the	DET
ajst-2981	29	54	initial	initial	ADJ
ajst-2981	29	55	anchor	anchor	NOUN
ajst-2981	29	56	box	box	NOUN
ajst-2981	29	57	when	when	SCONJ
ajst-2981	29	58	training	train	VERB
ajst-2981	29	59	for	for	ADP
ajst-2981	29	60	different	different	ADJ
ajst-2981	29	61	datasets	dataset	NOUN
ajst-2981	29	62	,	,	PUNCT
ajst-2981	29	63	and	and	CCONJ
ajst-2981	29	64	then	then	ADV
ajst-2981	29	65	compare	compare	VERB
ajst-2981	29	66	with	with	ADP
ajst-2981	29	67	the	the	DET
ajst-2981	29	68	real	real	ADJ
ajst-2981	29	69	box	box	NOUN
ajst-2981	29	70	to	to	PART
ajst-2981	29	71	calculate	calculate	VERB
ajst-2981	29	72	the	the	DET
ajst-2981	29	73	gap	gap	NOUN
ajst-2981	29	74	between	between	ADP
ajst-2981	29	75	the	the	DET
ajst-2981	29	76	two	two	NUM
ajst-2981	29	77	.	.	PUNCT
ajst-2981	30	1	adaptive	adaptive	ADJ
ajst-2981	30	2	image	image	NOUN
ajst-2981	30	3	scaling	scale	VERB
ajst-2981	30	4	adaptively	adaptively	ADV
ajst-2981	30	5	zooms	zoom	VERB
ajst-2981	30	6	and	and	CCONJ
ajst-2981	30	7	fills	fill	VERB
ajst-2981	30	8	many	many	ADJ
ajst-2981	30	9	images	image	NOUN
ajst-2981	30	10	with	with	ADP
ajst-2981	30	11	different	different	ADJ
ajst-2981	30	12	aspect	aspect	NOUN
ajst-2981	30	13	ratios	ratio	NOUN
ajst-2981	30	14	according	accord	VERB
ajst-2981	30	15	to	to	ADP
ajst-2981	30	16	the	the	DET
ajst-2981	30	17	standard	standard	ADJ
ajst-2981	30	18	size	size	NOUN
ajst-2981	30	19	to	to	PART
ajst-2981	30	20	meet	meet	VERB
ajst-2981	30	21	training	training	NOUN
ajst-2981	30	22	requirements	requirement	NOUN
ajst-2981	30	23	,	,	PUNCT
ajst-2981	30	24	reduce	reduce	VERB
ajst-2981	30	25	the	the	DET
ajst-2981	30	26	amount	amount	NOUN
ajst-2981	30	27	of	of	ADP
ajst-2981	30	28	calculation	calculation	NOUN
ajst-2981	30	29	,	,	PUNCT
ajst-2981	30	30	and	and	CCONJ
ajst-2981	30	31	improve	improve	VERB
ajst-2981	30	32	the	the	DET
ajst-2981	30	33	detection	detection	NOUN
ajst-2981	30	34	speed	speed	NOUN
ajst-2981	30	35	.	.	PUNCT
ajst-2981	31	1	2.2	2.2	NUM
ajst-2981	31	2	.	.	PUNCT
ajst-2981	31	3	backbone	backbone	NOUN
ajst-2981	31	4	network	network	NOUN
ajst-2981	31	5	the	the	DET
ajst-2981	31	6	backbone	backbone	NOUN
ajst-2981	31	7	network	network	NOUN
ajst-2981	31	8	mainly	mainly	ADV
ajst-2981	31	9	contains	contain	VERB
ajst-2981	31	10	cbs	cbs	PROPN
ajst-2981	31	11	,	,	PUNCT
ajst-2981	31	12	csp	csp	PROPN
ajst-2981	31	13	and	and	CCONJ
ajst-2981	31	14	sppf	sppf	ADJ
ajst-2981	31	15	three	three	NUM
ajst-2981	31	16	structures	structure	NOUN
ajst-2981	31	17	,	,	PUNCT
ajst-2981	31	18	cbs	cbs	PROPN
ajst-2981	31	19	consists	consist	VERB
ajst-2981	31	20	of	of	ADP
ajst-2981	31	21	convolution	convolution	NOUN
ajst-2981	31	22	,	,	PUNCT
ajst-2981	31	23	batch	batch	NOUN
ajst-2981	31	24	normalization	normalization	NOUN
ajst-2981	31	25	and	and	CCONJ
ajst-2981	31	26	silu	silu	ADJ
ajst-2981	31	27	activation	activation	NOUN
ajst-2981	31	28	function	function	NOUN
ajst-2981	31	29	.	.	PUNCT
ajst-2981	32	1	the	the	DET
ajst-2981	32	2	yolov56.0	yolov56.0	PROPN
ajst-2981	32	3	version	version	NOUN
ajst-2981	32	4	has	have	VERB
ajst-2981	32	5	two	two	NUM
ajst-2981	32	6	csp	csp	PROPN
ajst-2981	32	7	structures	structure	NOUN
ajst-2981	32	8	,	,	PUNCT
ajst-2981	32	9	csp1_x	csp1_x	PROPN
ajst-2981	32	10	applied	apply	VERB
ajst-2981	32	11	to	to	ADP
ajst-2981	32	12	the	the	DET
ajst-2981	32	13	backbone	backbone	NOUN
ajst-2981	32	14	part	part	NOUN
ajst-2981	32	15	and	and	CCONJ
ajst-2981	32	16	csp2_x	csp2_x	ADJ
ajst-2981	32	17	to	to	ADP
ajst-2981	32	18	the	the	DET
ajst-2981	32	19	neck	neck	NOUN
ajst-2981	32	20	part	part	NOUN
ajst-2981	32	21	,	,	PUNCT
ajst-2981	32	22	x	x	PRON
ajst-2981	32	23	represents	represent	VERB
ajst-2981	32	24	the	the	DET
ajst-2981	32	25	number	number	NOUN
ajst-2981	32	26	of	of	ADP
ajst-2981	32	27	residual	residual	ADJ
ajst-2981	32	28	components	component	NOUN
ajst-2981	32	29	(	(	PUNCT
ajst-2981	32	30	resunit	resunit	NOUN
ajst-2981	32	31	)	)	PUNCT
ajst-2981	32	32	,	,	PUNCT
ajst-2981	32	33	and	and	CCONJ
ajst-2981	32	34	the	the	DET
ajst-2981	32	35	network	network	NOUN
ajst-2981	32	36	depth	depth	NOUN
ajst-2981	32	37	is	be	AUX
ajst-2981	32	38	controlled	control	VERB
ajst-2981	32	39	by	by	ADP
ajst-2981	32	40	the	the	DET
ajst-2981	32	41	number	number	NOUN
ajst-2981	32	42	of	of	ADP
ajst-2981	32	43	residual	residual	ADJ
ajst-2981	32	44	components	component	NOUN
ajst-2981	32	45	.	.	PUNCT
ajst-2981	33	1	sppf	sppf	PROPN
ajst-2981	33	2	connects	connect	VERB
ajst-2981	33	3	multiple	multiple	ADJ
ajst-2981	33	4	maxpool	maxpool	ADJ
ajst-2981	33	5	layers	layer	NOUN
ajst-2981	33	6	to	to	PART
ajst-2981	33	7	achieve	achieve	VERB
ajst-2981	33	8	feature	feature	NOUN
ajst-2981	33	9	fusion	fusion	NOUN
ajst-2981	33	10	at	at	ADP
ajst-2981	33	11	different	different	ADJ
ajst-2981	33	12	scales	scale	NOUN
ajst-2981	33	13	.	.	PUNCT
ajst-2981	34	1	2.3	2.3	NUM
ajst-2981	34	2	.	.	PUNCT
ajst-2981	34	3	neck	neck	VERB
ajst-2981	34	4	the	the	DET
ajst-2981	34	5	neck	neck	NOUN
ajst-2981	34	6	network	network	NOUN
ajst-2981	34	7	consists	consist	VERB
ajst-2981	34	8	of	of	ADP
ajst-2981	34	9	a	a	DET
ajst-2981	34	10	feature	feature	NOUN
ajst-2981	34	11	pyramid	pyramid	NOUN
ajst-2981	34	12	network	network	NOUN
ajst-2981	34	13	(	(	PUNCT
ajst-2981	34	14	fpn	fpn	PROPN
ajst-2981	34	15	)	)	PUNCT
ajst-2981	34	16	and	and	CCONJ
ajst-2981	34	17	a	a	DET
ajst-2981	34	18	path	path	NOUN
ajst-2981	34	19	aggregation	aggregation	NOUN
ajst-2981	34	20	network	network	NOUN
ajst-2981	34	21	(	(	PUNCT
ajst-2981	34	22	pan	pan	NOUN
ajst-2981	34	23	)	)	PUNCT
ajst-2981	34	24	.	.	PUNCT
ajst-2981	35	1	fpn	fpn	VERB
ajst-2981	35	2	upsamples	upsample	NOUN
ajst-2981	35	3	high	high	ADJ
ajst-2981	35	4	-	-	PUNCT
ajst-2981	35	5	level	level	NOUN
ajst-2981	35	6	features	feature	NOUN
ajst-2981	35	7	from	from	ADP
ajst-2981	35	8	bottom	bottom	NOUN
ajst-2981	35	9	to	to	ADP
ajst-2981	35	10	top	top	NOUN
ajst-2981	35	11	,	,	PUNCT
ajst-2981	35	12	and	and	CCONJ
ajst-2981	35	13	enhances	enhance	VERB
ajst-2981	35	14	semantic	semantic	ADJ
ajst-2981	35	15	information	information	NOUN
ajst-2981	35	16	by	by	ADP
ajst-2981	35	17	fusing	fuse	VERB
ajst-2981	35	18	with	with	ADP
ajst-2981	35	19	low	low	ADJ
ajst-2981	35	20	-	-	PUNCT
ajst-2981	35	21	level	level	NOUN
ajst-2981	35	22	feature	feature	NOUN
ajst-2981	35	23	information	information	NOUN
ajst-2981	35	24	.	.	PUNCT
ajst-2981	36	1	pan	pan	PROPN
ajst-2981	36	2	adds	add	VERB
ajst-2981	36	3	top	top	ADJ
ajst-2981	36	4	-	-	PUNCT
ajst-2981	36	5	to	to	ADP
ajst-2981	36	6	-	-	PUNCT
ajst-2981	36	7	bottom	bottom	NOUN
ajst-2981	36	8	feature	feature	NOUN
ajst-2981	36	9	fusion	fusion	NOUN
ajst-2981	36	10	on	on	ADP
ajst-2981	36	11	the	the	DET
ajst-2981	36	12	basis	basis	NOUN
ajst-2981	36	13	of	of	ADP
ajst-2981	36	14	fpn	fpn	PROPN
ajst-2981	36	15	,	,	PUNCT
ajst-2981	36	16	and	and	CCONJ
ajst-2981	36	17	uses	use	VERB
ajst-2981	36	18	the	the	DET
ajst-2981	36	19	downsampling	downsample	VERB
ajst-2981	36	20	method	method	NOUN
ajst-2981	36	21	to	to	PART
ajst-2981	36	22	transmit	transmit	VERB
ajst-2981	36	23	low	low	ADJ
ajst-2981	36	24	-	-	PUNCT
ajst-2981	36	25	level	level	NOUN
ajst-2981	36	26	features	feature	NOUN
ajst-2981	36	27	to	to	ADP
ajst-2981	36	28	high	high	ADJ
ajst-2981	36	29	-	-	PUNCT
ajst-2981	36	30	level	level	NOUN
ajst-2981	36	31	information	information	NOUN
ajst-2981	36	32	fusion	fusion	NOUN
ajst-2981	36	33	,	,	PUNCT
ajst-2981	36	34	which	which	PRON
ajst-2981	36	35	strengthens	strengthen	VERB
ajst-2981	36	36	the	the	DET
ajst-2981	36	37	underlying	underlie	VERB
ajst-2981	36	38	information	information	NOUN
ajst-2981	36	39	positioning	positioning	NOUN
ajst-2981	36	40	ability	ability	NOUN
ajst-2981	36	41	.	.	PUNCT
ajst-2981	37	1	the	the	DET
ajst-2981	37	2	combination	combination	NOUN
ajst-2981	37	3	of	of	ADP
ajst-2981	37	4	the	the	DET
ajst-2981	37	5	two	two	NUM
ajst-2981	37	6	is	be	AUX
ajst-2981	37	7	conducive	conducive	ADJ
ajst-2981	37	8	to	to	ADP
ajst-2981	37	9	the	the	DET
ajst-2981	37	10	model	model	NOUN
ajst-2981	37	11	to	to	PART
ajst-2981	37	12	better	well	ADV
ajst-2981	37	13	learn	learn	VERB
ajst-2981	37	14	features	feature	NOUN
ajst-2981	37	15	and	and	CCONJ
ajst-2981	37	16	enhance	enhance	VERB
ajst-2981	37	17	the	the	DET
ajst-2981	37	18	sensitivity	sensitivity	NOUN
ajst-2981	37	19	of	of	ADP
ajst-2981	37	20	the	the	DET
ajst-2981	37	21	model	model	NOUN
ajst-2981	37	22	to	to	ADP
ajst-2981	37	23	small	small	ADJ
ajst-2981	37	24	targets	target	NOUN
ajst-2981	37	25	.	.	PUNCT
ajst-2981	38	1	2.4	2.4	NUM
ajst-2981	38	2	.	.	PUNCT
ajst-2981	38	3	prediction	prediction	NOUN
ajst-2981	38	4	considering	consider	VERB
ajst-2981	38	5	the	the	DET
ajst-2981	38	6	overlapping	overlap	VERB
ajst-2981	38	7	area	area	NOUN
ajst-2981	38	8	,	,	PUNCT
ajst-2981	38	9	center	center	NOUN
ajst-2981	38	10	point	point	NOUN
ajst-2981	38	11	distance	distance	NOUN
ajst-2981	38	12	and	and	CCONJ
ajst-2981	38	13	aspect	aspect	VERB
ajst-2981	38	14	ratio	ratio	NOUN
ajst-2981	38	15	between	between	ADP
ajst-2981	38	16	the	the	DET
ajst-2981	38	17	prediction	prediction	NOUN
ajst-2981	38	18	box	box	NOUN
ajst-2981	38	19	and	and	CCONJ
ajst-2981	38	20	the	the	DET
ajst-2981	38	21	target	target	NOUN
ajst-2981	38	22	box	box	NOUN
ajst-2981	38	23	,	,	PUNCT
ajst-2981	38	24	yolov5	yolov5	NOUN
ajst-2981	38	25	-	-	PUNCT
ajst-2981	38	26	6.0	6.0	NUM
ajst-2981	38	27	prediction	prediction	NOUN
ajst-2981	38	28	uses	use	VERB
ajst-2981	38	29	ciou	ciou	NOUN
ajst-2981	38	30	-	-	PUNCT
ajst-2981	38	31	loss	loss	NOUN
ajst-2981	38	32	as	as	ADP
ajst-2981	38	33	the	the	DET
ajst-2981	38	34	target	target	NOUN
ajst-2981	38	35	box	box	NOUN
ajst-2981	38	36	loss	loss	NOUN
ajst-2981	38	37	function	function	NOUN
ajst-2981	38	38	,	,	PUNCT
ajst-2981	38	39	which	which	PRON
ajst-2981	38	40	makes	make	VERB
ajst-2981	38	41	the	the	DET
ajst-2981	38	42	prediction	prediction	NOUN
ajst-2981	38	43	box	box	NOUN
ajst-2981	38	44	regression	regression	NOUN
ajst-2981	38	45	faster	fast	ADV
ajst-2981	38	46	and	and	CCONJ
ajst-2981	38	47	more	more	ADV
ajst-2981	38	48	accurate	accurate	ADJ
ajst-2981	38	49	.	.	PUNCT
ajst-2981	39	1	at	at	ADP
ajst-2981	39	2	the	the	DET
ajst-2981	39	3	same	same	ADJ
ajst-2981	39	4	time	time	NOUN
ajst-2981	39	5	,	,	PUNCT
ajst-2981	39	6	the	the	DET
ajst-2981	39	7	prediction	prediction	NOUN
ajst-2981	39	8	box	box	NOUN
ajst-2981	39	9	is	be	AUX
ajst-2981	39	10	screened	screen	VERB
ajst-2981	39	11	by	by	ADP
ajst-2981	39	12	non	non	ADJ
ajst-2981	39	13	-	-	ADJ
ajst-2981	39	14	maximum	maximum	ADJ
ajst-2981	39	15	suppression	suppression	NOUN
ajst-2981	39	16	(	(	PUNCT
ajst-2981	39	17	nms	nms	NOUN
ajst-2981	39	18	)	)	PUNCT
ajst-2981	39	19	to	to	PART
ajst-2981	39	20	remove	remove	VERB
ajst-2981	39	21	the	the	DET
ajst-2981	39	22	redundant	redundant	ADJ
ajst-2981	39	23	prediction	prediction	NOUN
ajst-2981	39	24	box	box	NOUN
ajst-2981	39	25	.	.	PUNCT
ajst-2981	40	1	3	3	X
ajst-2981	40	2	.	.	X
ajst-2981	40	3	deepsort	deepsort	NOUN
ajst-2981	40	4	vehicle	vehicle	NOUN
ajst-2981	40	5	tracking	tracking	NOUN
ajst-2981	40	6	algorithm	algorithm	NOUN
ajst-2981	40	7	deepsort	deepsort	NOUN
ajst-2981	40	8	algorithm	algorithm	NOUN
ajst-2981	40	9	is	be	AUX
ajst-2981	40	10	improved	improve	VERB
ajst-2981	40	11	on	on	ADP
ajst-2981	40	12	the	the	DET
ajst-2981	40	13	basis	basis	NOUN
ajst-2981	40	14	of	of	ADP
ajst-2981	40	15	sort	sort	NOUN
ajst-2981	40	16	,	,	PUNCT
ajst-2981	40	17	although	although	SCONJ
ajst-2981	40	18	the	the	DET
ajst-2981	40	19	original	original	ADJ
ajst-2981	40	20	sort	sort	NOUN
ajst-2981	40	21	algorithm	algorithm	NOUN
ajst-2981	40	22	can	can	AUX
ajst-2981	40	23	effectively	effectively	ADV
ajst-2981	40	24	associate	associate	VERB
ajst-2981	40	25	targets	target	NOUN
ajst-2981	40	26	and	and	CCONJ
ajst-2981	40	27	can	can	AUX
ajst-2981	40	28	track	track	VERB
ajst-2981	40	29	in	in	ADP
ajst-2981	40	30	real	real	ADJ
ajst-2981	40	31	time	time	NOUN
ajst-2981	40	32	,	,	PUNCT
ajst-2981	40	33	but	but	CCONJ
ajst-2981	40	34	for	for	ADP
ajst-2981	40	35	multi	multi	ADJ
ajst-2981	40	36	-	-	ADJ
ajst-2981	40	37	target	target	ADJ
ajst-2981	40	38	tracking	tracking	NOUN
ajst-2981	40	39	,	,	PUNCT
ajst-2981	40	40	the	the	DET
ajst-2981	40	41	sort	sort	ADJ
ajst-2981	40	42	algorithm	algorithm	NOUN
ajst-2981	40	43	can	can	AUX
ajst-2981	40	44	not	not	PART
ajst-2981	40	45	take	take	VERB
ajst-2981	40	46	into	into	ADP
ajst-2981	40	47	account	account	NOUN
ajst-2981	40	48	the	the	DET
ajst-2981	40	49	content	content	NOUN
ajst-2981	40	50	in	in	ADP
ajst-2981	40	51	the	the	DET
ajst-2981	40	52	recognition	recognition	NOUN
ajst-2981	40	53	box	box	NOUN
ajst-2981	40	54	,	,	PUNCT
ajst-2981	40	55	can	can	AUX
ajst-2981	40	56	not	not	PART
ajst-2981	40	57	guarantee	guarantee	VERB
ajst-2981	40	58	the	the	DET
ajst-2981	40	59	target	target	NOUN
ajst-2981	40	60	identity	identity	NOUN
ajst-2981	40	61	switches	switch	NOUN
ajst-2981	40	62	(	(	PUNCT
ajst-2981	40	63	ids	id	NOUN
ajst-2981	40	64	)	)	PUNCT
ajst-2981	40	65	.	.	PUNCT
ajst-2981	41	1	the	the	DET
ajst-2981	41	2	deepsort	deepsort	NOUN
ajst-2981	41	3	algorithm	algorithm	NOUN
ajst-2981	41	4	adds	add	VERB
ajst-2981	41	5	cascade	cascade	NOUN
ajst-2981	41	6	matching	matching	NOUN
ajst-2981	41	7	and	and	CCONJ
ajst-2981	41	8	confirmation	confirmation	NOUN
ajst-2981	41	9	of	of	ADP
ajst-2981	41	10	new	new	ADJ
ajst-2981	41	11	trajectories	trajectory	NOUN
ajst-2981	41	12	on	on	ADP
ajst-2981	41	13	the	the	DET
ajst-2981	41	14	basis	basis	NOUN
ajst-2981	41	15	of	of	ADP
ajst-2981	41	16	sort	sort	NOUN
ajst-2981	41	17	,	,	PUNCT
ajst-2981	41	18	reduces	reduce	VERB
ajst-2981	41	19	the	the	DET
ajst-2981	41	20	number	number	NOUN
ajst-2981	41	21	of	of	ADP
ajst-2981	41	22	i	i	PROPN
ajst-2981	41	23	d	d	PROPN
ajst-2981	41	24	switching	switching	NOUN
ajst-2981	41	25	,	,	PUNCT
ajst-2981	41	26	and	and	CCONJ
ajst-2981	41	27	processes	process	VERB
ajst-2981	41	28	the	the	DET
ajst-2981	41	29	motion	motion	NOUN
ajst-2981	41	30	prediction	prediction	NOUN
ajst-2981	41	31	and	and	CCONJ
ajst-2981	41	32	data	datum	NOUN
ajst-2981	41	33	correlation	correlation	NOUN
ajst-2981	41	34	parts	part	NOUN
ajst-2981	41	35	of	of	ADP
ajst-2981	41	36	the	the	DET
ajst-2981	41	37	tracking	tracking	NOUN
ajst-2981	41	38	problem	problem	NOUN
ajst-2981	41	39	through	through	ADP
ajst-2981	41	40	the	the	DET
ajst-2981	41	41	kalman	kalman	PROPN
ajst-2981	41	42	filter	filter	NOUN
ajst-2981	41	43	and	and	CCONJ
ajst-2981	41	44	the	the	DET
ajst-2981	41	45	hungarian	hungarian	ADJ
ajst-2981	41	46	algorithm	algorithm	NOUN
ajst-2981	41	47	,	,	PUNCT
ajst-2981	41	48	respectively	respectively	ADV
ajst-2981	41	49	.	.	PUNCT
ajst-2981	42	1	to	to	PART
ajst-2981	42	2	track	track	VERB
ajst-2981	42	3	each	each	DET
ajst-2981	42	4	vehicle	vehicle	NOUN
ajst-2981	42	5	detected	detect	VERB
ajst-2981	42	6	by	by	ADP
ajst-2981	42	7	the	the	DET
ajst-2981	42	8	yolov5	yolov5	NOUN
ajst-2981	42	9	detector	detector	NOUN
ajst-2981	42	10	,	,	PUNCT
ajst-2981	42	11	define	define	VERB
ajst-2981	42	12	an	an	DET
ajst-2981	42	13	8	8	NUM
ajst-2981	42	14	-	-	PUNCT
ajst-2981	42	15	dimensional	dimensional	ADJ
ajst-2981	42	16	status	status	NOUN
ajst-2981	42	17	vector	vector	NOUN
ajst-2981	42	18	x	x	PUNCT
ajst-2981	42	19	representing	represent	VERB
ajst-2981	42	20	the	the	DET
ajst-2981	42	21	vehicle	vehicle	NOUN
ajst-2981	42	22	state	state	NOUN
ajst-2981	42	23	,	,	PUNCT
ajst-2981	42	24	see	see	VERB
ajst-2981	42	25	equation	equation	NOUN
ajst-2981	42	26	(	(	PUNCT
ajst-2981	42	27	1	1	NUM
ajst-2981	42	28	):	):	PUNCT
ajst-2981	42	29	'	'	PUNCT
ajst-2981	42	30	'	'	PUNCT
ajst-2981	42	31	'	'	PUNCT
ajst-2981	42	32	'	'	PUNCT
ajst-2981	42	33	(	(	PUNCT
ajst-2981	42	34	x	x	X
ajst-2981	42	35	,	,	PUNCT
ajst-2981	42	36	,	,	PUNCT
ajst-2981	42	37	,	,	PUNCT
ajst-2981	42	38	,	,	PUNCT
ajst-2981	42	39	,	,	PUNCT
ajst-2981	42	40	,	,	PUNCT
ajst-2981	42	41	,	,	PUNCT
ajst-2981	42	42	)	)	PUNCT
ajst-2981	42	43	x	x	X
ajst-2981	43	1	y	y	NOUN
ajst-2981	43	2	g	g	NOUN
ajst-2981	43	3	h	h	NOUN
ajst-2981	43	4	x	x	VERB
ajst-2981	44	1	y	y	NOUN
ajst-2981	44	2	g	g	PROPN
ajst-2981	44	3	h	h	NOUN
ajst-2981	44	4	(	(	PUNCT
ajst-2981	44	5	1	1	NUM
ajst-2981	44	6	)	)	PUNCT
ajst-2981	44	7	where	where	SCONJ
ajst-2981	44	8	(	(	PUNCT
ajst-2981	44	9	x	x	X
ajst-2981	44	10	,	,	PUNCT
ajst-2981	44	11	y	y	NOUN
ajst-2981	44	12	)	)	PUNCT
ajst-2981	44	13	represents	represent	VERB
ajst-2981	44	14	the	the	DET
ajst-2981	44	15	vehicle	vehicle	NOUN
ajst-2981	44	16	center	center	NOUN
ajst-2981	44	17	coordinates	coordinate	NOUN
ajst-2981	44	18	;	;	PUNCT
ajst-2981	44	19	g	g	PROPN
ajst-2981	44	20	and	and	CCONJ
ajst-2981	44	21	h	h	NOUN
ajst-2981	44	22	indicate	indicate	VERB
ajst-2981	44	23	the	the	DET
ajst-2981	44	24	vehicle	vehicle	NOUN
ajst-2981	44	25	bounding	bound	VERB
ajst-2981	44	26	box	box	PROPN
ajst-2981	44	27	aspect	aspect	NOUN
ajst-2981	44	28	ratio	ratio	NOUN
ajst-2981	44	29	and	and	CCONJ
ajst-2981	44	30	height	height	NOUN
ajst-2981	44	31	,	,	PUNCT
ajst-2981	44	32	respectively	respectively	ADV
ajst-2981	44	33	;	;	PUNCT
ajst-2981	44	34	(	(	PUNCT
ajst-2981	44	35	'	'	PUNCT
ajst-2981	44	36	'	'	PUNCT
ajst-2981	44	37	'	'	PUNCT
ajst-2981	44	38	'	'	PUNCT
ajst-2981	44	39	,	,	PUNCT
ajst-2981	44	40	,	,	PUNCT
ajst-2981	44	41	,	,	PUNCT
ajst-2981	44	42	x	x	PROPN
ajst-2981	44	43	y	y	NOUN
ajst-2981	44	44	g	g	PROPN
ajst-2981	44	45	h	h	PROPN
ajst-2981	44	46	)	)	PUNCT
ajst-2981	44	47	represents	represent	VERB
ajst-2981	44	48	the	the	DET
ajst-2981	44	49	speed	speed	NOUN
ajst-2981	44	50	information	information	NOUN
ajst-2981	44	51	corresponding	correspond	VERB
ajst-2981	44	52	to	to	ADP
ajst-2981	44	53	the	the	DET
ajst-2981	44	54	vehicle	vehicle	NOUN
ajst-2981	44	55	(	(	PUNCT
ajst-2981	44	56	x	x	X
ajst-2981	44	57	,	,	PUNCT
ajst-2981	44	58	,	,	PUNCT
ajst-2981	44	59	,	,	PUNCT
ajst-2981	44	60	y	y	PROPN
ajst-2981	44	61	g	g	PROPN
ajst-2981	44	62	h	h	PROPN
ajst-2981	44	63	)	)	PUNCT
ajst-2981	44	64	in	in	ADP
ajst-2981	44	65	the	the	DET
ajst-2981	44	66	image	image	NOUN
ajst-2981	44	67	coordinates	coordinate	NOUN
ajst-2981	44	68	.	.	PUNCT
ajst-2981	45	1	the	the	DET
ajst-2981	45	2	mahalanobis	mahalanobis	ADJ
ajst-2981	45	3	distance	distance	NOUN
ajst-2981	45	4	between	between	ADP
ajst-2981	45	5	the	the	DET
ajst-2981	45	6	predicted	predict	VERB
ajst-2981	45	7	target	target	NOUN
ajst-2981	45	8	frame	frame	NOUN
ajst-2981	45	9	obtained	obtain	VERB
ajst-2981	45	10	by	by	ADP
ajst-2981	45	11	the	the	DET
ajst-2981	45	12	kalman	kalman	PROPN
ajst-2981	45	13	filter	filter	NOUN
ajst-2981	45	14	and	and	CCONJ
ajst-2981	45	15	the	the	DET
ajst-2981	45	16	target	target	NOUN
ajst-2981	45	17	frame	frame	NOUN
ajst-2981	45	18	of	of	ADP
ajst-2981	45	19	the	the	DET
ajst-2981	45	20	current	current	ADJ
ajst-2981	45	21	frame	frame	NOUN
ajst-2981	45	22	is	be	AUX
ajst-2981	45	23	calculated	calculate	VERB
ajst-2981	45	24	,	,	PUNCT
ajst-2981	45	25	and	and	CCONJ
ajst-2981	45	26	when	when	SCONJ
ajst-2981	45	27	its	its	PRON
ajst-2981	45	28	value	value	NOUN
ajst-2981	45	29	is	be	AUX
ajst-2981	45	30	greater	great	ADJ
ajst-2981	45	31	than	than	ADP
ajst-2981	45	32	the	the	DET
ajst-2981	45	33	set	set	NOUN
ajst-2981	45	34	threshold	threshold	NOUN
ajst-2981	45	35	,	,	PUNCT
ajst-2981	45	36	it	it	PRON
ajst-2981	45	37	means	mean	VERB
ajst-2981	45	38	that	that	SCONJ
ajst-2981	45	39	the	the	DET
ajst-2981	45	40	predicted	predict	VERB
ajst-2981	45	41	target	target	NOUN
ajst-2981	45	42	frame	frame	NOUN
ajst-2981	45	43	200	200	NUM
ajst-2981	45	44	is	be	AUX
ajst-2981	45	45	unreasonable	unreasonable	ADJ
ajst-2981	45	46	,	,	PUNCT
ajst-2981	45	47	and	and	CCONJ
ajst-2981	45	48	it	it	PRON
ajst-2981	45	49	is	be	AUX
ajst-2981	45	50	reasonable	reasonable	ADJ
ajst-2981	45	51	and	and	CCONJ
ajst-2981	45	52	accurate	accurate	ADJ
ajst-2981	45	53	to	to	PART
ajst-2981	45	54	be	be	AUX
ajst-2981	45	55	less	less	ADJ
ajst-2981	45	56	than	than	ADP
ajst-2981	45	57	the	the	DET
ajst-2981	45	58	set	set	NOUN
ajst-2981	45	59	value	value	NOUN
ajst-2981	45	60	,	,	PUNCT
ajst-2981	45	61	so	so	SCONJ
ajst-2981	45	62	as	as	SCONJ
ajst-2981	45	63	to	to	PART
ajst-2981	45	64	complete	complete	VERB
ajst-2981	45	65	the	the	DET
ajst-2981	45	66	tracking	tracking	NOUN
ajst-2981	45	67	of	of	ADP
ajst-2981	45	68	the	the	DET
ajst-2981	45	69	target	target	NOUN
ajst-2981	45	70	.	.	PUNCT
ajst-2981	46	1	see	see	VERB
ajst-2981	46	2	equation	equation	NOUN
ajst-2981	46	3	(	(	PUNCT
ajst-2981	46	4	2	2	NUM
ajst-2981	46	5	):	):	PUNCT
ajst-2981	46	6	(	(	PUNCT
ajst-2981	46	7	1	1	X
ajst-2981	46	8	)	)	PUNCT
ajst-2981	46	9	1	1	NUM
ajst-2981	46	10	m	m	NOUN
ajst-2981	46	11	m	m	VERB
ajst-2981	46	12	(	(	PUNCT
ajst-2981	46	13	,	,	PUNCT
ajst-2981	46	14	)	)	PUNCT
ajst-2981	46	15	(	(	PUNCT
ajst-2981	46	16	)	)	PUNCT
ajst-2981	46	17	(	(	PUNCT
ajst-2981	46	18	)	)	PUNCT
ajst-2981	46	19	t	t	PROPN
ajst-2981	46	20	n	n	INTJ
ajst-2981	46	21	m	m	VERB
ajst-2981	46	22	nd	nd	ADV
ajst-2981	46	23	m	m	PROPN
ajst-2981	46	24	n	n	PROPN
ajst-2981	47	1	d	d	X
ajst-2981	47	2	y	y	PROPN
ajst-2981	47	3	s	s	AUX
ajst-2981	47	4	d	d	X
ajst-2981	47	5	y	y	ADJ
ajst-2981	47	6			PROPN
ajst-2981	47	7			NOUN
ajst-2981	47	8	(	(	PUNCT
ajst-2981	47	9	2	2	NUM
ajst-2981	47	10	)	)	PUNCT
ajst-2981	47	11	where	where	SCONJ
ajst-2981	47	12	indicates	indicate	VERB
ajst-2981	47	13	the	the	DET
ajst-2981	47	14	matching	matching	NOUN
ajst-2981	47	15	degree	degree	NOUN
ajst-2981	47	16	between	between	ADP
ajst-2981	47	17	the	the	DET
ajst-2981	47	18	nth	nth	NOUN
ajst-2981	47	19	detected	detect	VERB
ajst-2981	47	20	target	target	NOUN
ajst-2981	47	21	and	and	CCONJ
ajst-2981	47	22	the	the	DET
ajst-2981	47	23	m	m	NOUN
ajst-2981	47	24	-	-	PUNCT
ajst-2981	47	25	track	track	NOUN
ajst-2981	47	26	track	track	NOUN
ajst-2981	47	27	;	;	PUNCT
ajst-2981	47	28	ym	ym	PRON
ajst-2981	47	29	represents	represent	VERB
ajst-2981	47	30	the	the	DET
ajst-2981	47	31	target	target	NOUN
ajst-2981	47	32	prediction	prediction	NOUN
ajst-2981	47	33	position	position	NOUN
ajst-2981	47	34	of	of	ADP
ajst-2981	47	35	the	the	DET
ajst-2981	47	36	mth	mth	NOUN
ajst-2981	47	37	trajectory	trajectory	NOUN
ajst-2981	47	38	;	;	PUNCT
ajst-2981	47	39	dn	dn	PROPN
ajst-2981	47	40	represents	represent	VERB
ajst-2981	47	41	the	the	DET
ajst-2981	47	42	nth	nth	ADJ
ajst-2981	47	43	object	object	NOUN
ajst-2981	47	44	detection	detection	NOUN
ajst-2981	47	45	bounding	bound	VERB
ajst-2981	47	46	box	box	NOUN
ajst-2981	47	47	;	;	PUNCT
ajst-2981	47	48	sm	sm	PROPN
ajst-2981	47	49	represents	represent	VERB
ajst-2981	47	50	the	the	DET
ajst-2981	47	51	current	current	ADJ
ajst-2981	47	52	frame	frame	NOUN
ajst-2981	47	53	covariance	covariance	NOUN
ajst-2981	47	54	matrix	matrix	NOUN
ajst-2981	47	55	predicted	predict	VERB
ajst-2981	47	56	by	by	ADP
ajst-2981	47	57	the	the	DET
ajst-2981	47	58	kalman	kalman	PROPN
ajst-2981	47	59	filter	filter	PROPN
ajst-2981	47	60	,	,	PUNCT
ajst-2981	47	61	to	to	PART
ajst-2981	47	62	improve	improve	VERB
ajst-2981	47	63	ids	id	NOUN
ajst-2981	47	64	,	,	PUNCT
ajst-2981	47	65	the	the	DET
ajst-2981	47	66	prediction	prediction	NOUN
ajst-2981	47	67	accuracy	accuracy	NOUN
ajst-2981	47	68	is	be	AUX
ajst-2981	47	69	measured	measure	VERB
ajst-2981	47	70	by	by	ADP
ajst-2981	47	71	using	use	VERB
ajst-2981	47	72	the	the	DET
ajst-2981	47	73	minimum	minimum	ADJ
ajst-2981	47	74	cosine	cosine	NOUN
ajst-2981	47	75	between	between	ADP
ajst-2981	47	76	the	the	DET
ajst-2981	47	77	predicted	predict	VERB
ajst-2981	47	78	target	target	NOUN
ajst-2981	47	79	and	and	CCONJ
ajst-2981	47	80	the	the	DET
ajst-2981	47	81	feature	feature	NOUN
ajst-2981	47	82	vector	vector	NOUN
ajst-2981	47	83	of	of	ADP
ajst-2981	47	84	the	the	DET
ajst-2981	47	85	predicted	predict	VERB
ajst-2981	47	86	target	target	NOUN
ajst-2981	47	87	contained	contain	VERB
ajst-2981	47	88	in	in	ADP
ajst-2981	47	89	the	the	DET
ajst-2981	47	90	trajectory	trajectory	NOUN
ajst-2981	47	91	as	as	ADP
ajst-2981	47	92	the	the	DET
ajst-2981	47	93	degree	degree	NOUN
ajst-2981	47	94	of	of	ADP
ajst-2981	47	95	apparent	apparent	ADJ
ajst-2981	47	96	match	match	NOUN
ajst-2981	47	97	.	.	PUNCT
ajst-2981	48	1	see	see	VERB
ajst-2981	48	2	equation	equation	NOUN
ajst-2981	48	3	(	(	PUNCT
ajst-2981	48	4	3	3	NUM
ajst-2981	48	5	)	)	PUNCT
ajst-2981	48	6	for	for	ADP
ajst-2981	48	7	the	the	DET
ajst-2981	48	8	formula	formula	NOUN
ajst-2981	48	9	:	:	PUNCT
ajst-2981	48	10	(	(	PUNCT
ajst-2981	48	11	2	2	X
ajst-2981	48	12	)	)	PUNCT
ajst-2981	48	13	(	(	PUNCT
ajst-2981	48	14	)	)	PUNCT
ajst-2981	48	15	(	(	PUNCT
ajst-2981	48	16	)	)	PUNCT
ajst-2981	48	17	(	(	PUNCT
ajst-2981	48	18	,	,	PUNCT
ajst-2981	48	19	)	)	PUNCT
ajst-2981	48	20	min{1	min{1	NOUN
ajst-2981	49	1	|	|	ADV
ajst-2981	49	2	}	}	PUNCT
ajst-2981	49	3	t	t	PROPN
ajst-2981	49	4	m	m	VERB
ajst-2981	49	5	m	m	VERB
ajst-2981	49	6	n	n	PRON
ajst-2981	49	7	k	k	PROPN
ajst-2981	49	8	k	k	PROPN
ajst-2981	49	9	md	md	PROPN
ajst-2981	49	10	m	m	PROPN
ajst-2981	49	11	n	n	ADP
ajst-2981	50	1	r	r	NOUN
ajst-2981	50	2	r	r	NOUN
ajst-2981	50	3	r	r	NOUN
ajst-2981	50	4	r	r	NOUN
ajst-2981	50	5			NOUN
ajst-2981	50	6			NOUN
ajst-2981	50	7	(	(	PUNCT
ajst-2981	50	8	3	3	NUM
ajst-2981	50	9	)	)	PUNCT
ajst-2981	50	10	in	in	ADP
ajst-2981	50	11	the	the	DET
ajst-2981	50	12	chinese	chinese	ADJ
ajst-2981	50	13	formula	formula	NOUN
ajst-2981	50	14	,	,	PUNCT
ajst-2981	50	15	rn	rn	PROPN
ajst-2981	50	16	represents	represent	VERB
ajst-2981	50	17	a	a	DET
ajst-2981	50	18	feature	feature	NOUN
ajst-2981	50	19	vector	vector	NOUN
ajst-2981	50	20	corresponding	correspond	VERB
ajst-2981	50	21	to	to	ADP
ajst-2981	50	22	each	each	DET
ajst-2981	50	23	detection	detection	NOUN
ajst-2981	50	24	block	block	NOUN
ajst-2981	50	25	dn	dn	NOUN
ajst-2981	50	26	,	,	PUNCT
ajst-2981	50	27	and	and	CCONJ
ajst-2981	50	28	rk	rk	NOUN
ajst-2981	50	29	represents	represent	VERB
ajst-2981	50	30	the	the	DET
ajst-2981	50	31	feature	feature	NOUN
ajst-2981	50	32	vector	vector	NOUN
ajst-2981	50	33	successfully	successfully	ADV
ajst-2981	50	34	associated	associate	VERB
ajst-2981	50	35	with	with	ADP
ajst-2981	50	36	the	the	DET
ajst-2981	50	37	last	last	ADJ
ajst-2981	50	38	100	100	NUM
ajst-2981	50	39	frames	frame	NOUN
ajst-2981	50	40	.	.	PUNCT
ajst-2981	51	1	weighted	weight	VERB
ajst-2981	51	2	fusion	fusion	NOUN
ajst-2981	51	3	of	of	ADP
ajst-2981	51	4	the	the	DET
ajst-2981	51	5	two	two	NUM
ajst-2981	51	6	as	as	ADP
ajst-2981	51	7	the	the	DET
ajst-2981	51	8	final	final	ADJ
ajst-2981	51	9	measure	measure	NOUN
ajst-2981	51	10	,	,	PUNCT
ajst-2981	51	11	see	see	VERB
ajst-2981	51	12	equation	equation	NOUN
ajst-2981	51	13	(	(	PUNCT
ajst-2981	51	14	4	4	NUM
ajst-2981	51	15	):	):	PUNCT
ajst-2981	51	16	(	(	PUNCT
ajst-2981	51	17	1	1	X
ajst-2981	51	18	)	)	PUNCT
ajst-2981	51	19	(	(	PUNCT
ajst-2981	51	20	2	2	NUM
ajst-2981	51	21	)	)	PUNCT
ajst-2981	51	22	(	(	PUNCT
ajst-2981	51	23	)	)	PUNCT
ajst-2981	51	24	(	(	PUNCT
ajst-2981	51	25	,	,	PUNCT
ajst-2981	51	26	)	)	PUNCT
ajst-2981	51	27	(	(	PUNCT
ajst-2981	51	28	1	1	X
ajst-2981	51	29	)	)	PUNCT
ajst-2981	51	30	(	(	PUNCT
ajst-2981	51	31	,	,	PUNCT
ajst-2981	51	32	)	)	PUNCT
ajst-2981	51	33	c	c	NOUN
ajst-2981	51	34	m	m	VERB
ajst-2981	51	35	n	n	VERB
ajst-2981	51	36	m	m	PROPN
ajst-2981	51	37	n	n	PROPN
ajst-2981	51	38			PROPN
ajst-2981	51	39	m	m	ADJ
ajst-2981	51	40	,	,	PUNCT
ajst-2981	51	41	n	n	NOUN
ajst-2981	51	42	=	=	SYM
ajst-2981	51	43	d	d	PROPN
ajst-2981	51	44	d	d	PROPN
ajst-2981	51	45	(	(	PUNCT
ajst-2981	51	46	4	4	NUM
ajst-2981	51	47	)	)	PUNCT
ajst-2981	51	48	finally	finally	ADV
ajst-2981	51	49	,	,	PUNCT
ajst-2981	51	50	the	the	DET
ajst-2981	51	51	hungarian	hungarian	ADJ
ajst-2981	51	52	algorithm	algorithm	NOUN
ajst-2981	51	53	is	be	AUX
ajst-2981	51	54	used	use	VERB
ajst-2981	51	55	to	to	PART
ajst-2981	51	56	detect	detect	VERB
ajst-2981	51	57	whether	whether	SCONJ
ajst-2981	51	58	a	a	DET
ajst-2981	51	59	target	target	NOUN
ajst-2981	51	60	of	of	ADP
ajst-2981	51	61	the	the	DET
ajst-2981	51	62	current	current	ADJ
ajst-2981	51	63	frame	frame	NOUN
ajst-2981	51	64	is	be	AUX
ajst-2981	51	65	a	a	DET
ajst-2981	51	66	target	target	NOUN
ajst-2981	51	67	of	of	ADP
ajst-2981	51	68	the	the	DET
ajst-2981	51	69	previous	previous	ADJ
ajst-2981	51	70	frame	frame	NOUN
ajst-2981	51	71	.	.	PUNCT
ajst-2981	52	1	the	the	DET
ajst-2981	52	2	cascade	cascade	NOUN
ajst-2981	52	3	matching	matching	NOUN
ajst-2981	52	4	idea	idea	NOUN
ajst-2981	52	5	is	be	AUX
ajst-2981	52	6	adopted	adopt	VERB
ajst-2981	52	7	to	to	PART
ajst-2981	52	8	avoid	avoid	VERB
ajst-2981	52	9	the	the	DET
ajst-2981	52	10	loss	loss	NOUN
ajst-2981	52	11	of	of	ADP
ajst-2981	52	12	trajectory	trajectory	NOUN
ajst-2981	52	13	caused	cause	VERB
ajst-2981	52	14	by	by	ADP
ajst-2981	52	15	the	the	DET
ajst-2981	52	16	vehicle	vehicle	NOUN
ajst-2981	52	17	being	be	AUX
ajst-2981	52	18	blocked	block	VERB
ajst-2981	52	19	for	for	ADP
ajst-2981	52	20	a	a	DET
ajst-2981	52	21	long	long	ADJ
ajst-2981	52	22	time	time	NOUN
ajst-2981	52	23	,	,	PUNCT
ajst-2981	52	24	and	and	CCONJ
ajst-2981	52	25	the	the	DET
ajst-2981	52	26	kalman	kalman	PROPN
ajst-2981	52	27	filter	filter	NOUN
ajst-2981	52	28	prediction	prediction	NOUN
ajst-2981	52	29	will	will	AUX
ajst-2981	52	30	lead	lead	VERB
ajst-2981	52	31	to	to	ADP
ajst-2981	52	32	the	the	DET
ajst-2981	52	33	problem	problem	NOUN
ajst-2981	52	34	of	of	ADP
ajst-2981	52	35	probability	probability	NOUN
ajst-2981	52	36	diffusion	diffusion	NOUN
ajst-2981	52	37	.	.	PUNCT
ajst-2981	53	1	the	the	DET
ajst-2981	53	2	deepsort	deepsort	NOUN
ajst-2981	53	3	algorithm	algorithm	NOUN
ajst-2981	53	4	uses	use	VERB
ajst-2981	53	5	the	the	DET
ajst-2981	53	6	above	above	ADJ
ajst-2981	53	7	idea	idea	NOUN
ajst-2981	53	8	,	,	PUNCT
ajst-2981	53	9	and	and	CCONJ
ajst-2981	53	10	after	after	ADP
ajst-2981	53	11	accurately	accurately	ADV
ajst-2981	53	12	detecting	detect	VERB
ajst-2981	53	13	the	the	DET
ajst-2981	53	14	target	target	NOUN
ajst-2981	53	15	,	,	PUNCT
ajst-2981	53	16	it	it	PRON
ajst-2981	53	17	accurately	accurately	ADV
ajst-2981	53	18	locates	locate	VERB
ajst-2981	53	19	the	the	DET
ajst-2981	53	20	target	target	NOUN
ajst-2981	53	21	position	position	NOUN
ajst-2981	53	22	in	in	ADP
ajst-2981	53	23	the	the	DET
ajst-2981	53	24	next	next	ADJ
ajst-2981	53	25	frame	frame	NOUN
ajst-2981	53	26	.	.	PUNCT
ajst-2981	54	1	4	4	X
ajst-2981	54	2	.	.	X
ajst-2981	54	3	experiments	experiment	NOUN
ajst-2981	54	4	and	and	CCONJ
ajst-2981	54	5	analysis	analysis	NOUN
ajst-2981	54	6	4.1	4.1	NUM
ajst-2981	54	7	.	.	PUNCT
ajst-2981	55	1	experimental	experimental	ADJ
ajst-2981	55	2	preparation	preparation	NOUN
ajst-2981	55	3	the	the	DET
ajst-2981	55	4	vehicle	vehicle	NOUN
ajst-2981	55	5	detection	detection	NOUN
ajst-2981	55	6	and	and	CCONJ
ajst-2981	55	7	vehicle	vehicle	NOUN
ajst-2981	55	8	tracking	tracking	NOUN
ajst-2981	55	9	dataset	dataset	NOUN
ajst-2981	55	10	is	be	AUX
ajst-2981	55	11	a	a	DET
ajst-2981	55	12	visdrone2019	visdrone2019	PROPN
ajst-2981	55	13	open	open	ADJ
ajst-2981	55	14	-	-	PUNCT
ajst-2981	55	15	source	source	NOUN
ajst-2981	55	16	aerial	aerial	ADJ
ajst-2981	55	17	photography	photography	NOUN
ajst-2981	55	18	dataset	dataset	VERB
ajst-2981	55	19	.	.	PUNCT
ajst-2981	56	1	the	the	DET
ajst-2981	56	2	dataset	dataset	NOUN
ajst-2981	56	3	contains	contain	VERB
ajst-2981	56	4	different	different	ADJ
ajst-2981	56	5	traffic	traffic	NOUN
ajst-2981	56	6	scenarios	scenario	NOUN
ajst-2981	56	7	,	,	PUNCT
ajst-2981	56	8	including	include	VERB
ajst-2981	56	9	highways	highway	NOUN
ajst-2981	56	10	,	,	PUNCT
ajst-2981	56	11	intersections	intersection	NOUN
ajst-2981	56	12	,	,	PUNCT
ajst-2981	56	13	t	t	NOUN
ajst-2981	56	14	-	-	PUNCT
ajst-2981	56	15	junctions	junction	NOUN
ajst-2981	56	16	,	,	PUNCT
ajst-2981	56	17	etc	etc	X
ajst-2981	56	18	.	.	X
ajst-2981	56	19	,	,	PUNCT
ajst-2981	56	20	including	include	VERB
ajst-2981	56	21	different	different	ADJ
ajst-2981	56	22	environmental	environmental	ADJ
ajst-2981	56	23	backgrounds	background	NOUN
ajst-2981	56	24	during	during	ADP
ajst-2981	56	25	the	the	DET
ajst-2981	56	26	day	day	NOUN
ajst-2981	56	27	and	and	CCONJ
ajst-2981	56	28	night	night	NOUN
ajst-2981	56	29	.	.	PUNCT
ajst-2981	57	1	in	in	ADP
ajst-2981	57	2	this	this	DET
ajst-2981	57	3	paper	paper	NOUN
ajst-2981	57	4	,	,	PUNCT
ajst-2981	57	5	a	a	DET
ajst-2981	57	6	total	total	NOUN
ajst-2981	57	7	of	of	ADP
ajst-2981	57	8	6100	6100	NUM
ajst-2981	57	9	car	car	NOUN
ajst-2981	57	10	,	,	PUNCT
ajst-2981	57	11	truck	truck	NOUN
ajst-2981	57	12	and	and	CCONJ
ajst-2981	57	13	bus	bus	NOUN
ajst-2981	57	14	sheets	sheet	NOUN
ajst-2981	57	15	were	be	AUX
ajst-2981	57	16	extracted	extract	VERB
ajst-2981	57	17	,	,	PUNCT
ajst-2981	57	18	including	include	VERB
ajst-2981	57	19	4800	4800	NUM
ajst-2981	57	20	training	training	NOUN
ajst-2981	57	21	sets	set	NOUN
ajst-2981	57	22	and	and	CCONJ
ajst-2981	57	23	1220	1220	NUM
ajst-2981	57	24	test	test	NOUN
ajst-2981	57	25	sets	set	NOUN
ajst-2981	57	26	.	.	PUNCT
ajst-2981	58	1	the	the	DET
ajst-2981	58	2	experimental	experimental	ADJ
ajst-2981	58	3	configuration	configuration	NOUN
ajst-2981	58	4	environment	environment	NOUN
ajst-2981	58	5	is	be	AUX
ajst-2981	58	6	shown	show	VERB
ajst-2981	58	7	in	in	ADP
ajst-2981	58	8	table	table	NOUN
ajst-2981	58	9	1	1	NUM
ajst-2981	58	10	,	,	PUNCT
ajst-2981	58	11	and	and	CCONJ
ajst-2981	58	12	the	the	DET
ajst-2981	58	13	verification	verification	NOUN
ajst-2981	58	14	data	data	NOUN
ajst-2981	58	15	is	be	AUX
ajst-2981	58	16	a	a	DET
ajst-2981	58	17	ten	ten	NUM
ajst-2981	58	18	-	-	PUNCT
ajst-2981	58	19	minute	minute	NOUN
ajst-2981	58	20	video	video	NOUN
ajst-2981	58	21	of	of	ADP
ajst-2981	58	22	the	the	DET
ajst-2981	58	23	intersection	intersection	NOUN
ajst-2981	58	24	of	of	ADP
ajst-2981	58	25	beixin	beixin	PROPN
ajst-2981	58	26	road	road	PROPN
ajst-2981	58	27	and	and	CCONJ
ajst-2981	58	28	xueyuan	xueyuan	PROPN
ajst-2981	58	29	road	road	PROPN
ajst-2981	58	30	in	in	ADP
ajst-2981	58	31	tangshan	tangshan	NOUN
ajst-2981	58	32	city	city	NOUN
ajst-2981	58	33	taken	take	VERB
ajst-2981	58	34	by	by	ADP
ajst-2981	58	35	uav	uav	PROPN
ajst-2981	58	36	,	,	PUNCT
ajst-2981	58	37	and	and	CCONJ
ajst-2981	58	38	the	the	DET
ajst-2981	58	39	intersection	intersection	NOUN
ajst-2981	58	40	of	of	ADP
ajst-2981	58	41	beixin	beixin	PROPN
ajst-2981	58	42	road	road	PROPN
ajst-2981	58	43	and	and	CCONJ
ajst-2981	58	44	xueyuan	xueyuan	PROPN
ajst-2981	58	45	road	road	NOUN
ajst-2981	58	46	is	be	AUX
ajst-2981	58	47	shown	show	VERB
ajst-2981	58	48	in	in	ADP
ajst-2981	58	49	figure	figure	NOUN
ajst-2981	58	50	2	2	NUM
ajst-2981	58	51	.	.	PUNCT
ajst-2981	58	52	table	table	NOUN
ajst-2981	58	53	1	1	NUM
ajst-2981	58	54	.	.	PUNCT
ajst-2981	58	55	experimental	experimental	ADJ
ajst-2981	58	56	configuration	configuration	NOUN
ajst-2981	58	57	environment	environment	NOUN
ajst-2981	58	58	experimental	experimental	ADJ
ajst-2981	58	59	environment	environment	NOUN
ajst-2981	58	60	parameter	parameter	PROPN
ajst-2981	58	61	deep	deep	ADJ
ajst-2981	58	62	learning	learning	NOUN
ajst-2981	58	63	framework	framework	NOUN
ajst-2981	58	64	pytorch	pytorch	NOUN
ajst-2981	58	65	operating	operating	NOUN
ajst-2981	58	66	system	system	NOUN
ajst-2981	58	67	windows	window	VERB
ajst-2981	58	68	10	10	NUM
ajst-2981	58	69	gpu	gpu	NOUN
ajst-2981	58	70	model	model	NOUN
ajst-2981	58	71	nvidia	nvidia	PROPN
ajst-2981	58	72	geforce	geforce	PROPN
ajst-2981	58	73	gtx	gtx	PROPN
ajst-2981	58	74	1080	1080	NUM
ajst-2981	58	75	programming	programming	NOUN
ajst-2981	58	76	tools	tool	NOUN
ajst-2981	58	77	pycharm	pycharm	VERB
ajst-2981	58	78	programming	programming	NOUN
ajst-2981	58	79	language	language	NOUN
ajst-2981	58	80	python3.8	python3.8	ADJ
ajst-2981	58	81	data	datum	NOUN
ajst-2981	58	82	set	set	VERB
ajst-2981	58	83	visdrone2019	visdrone2019	PROPN
ajst-2981	58	84	-	-	PUNCT
ajst-2981	58	85	det	det	NOUN
ajst-2981	58	86	figure	figure	NOUN
ajst-2981	58	87	2	2	NUM
ajst-2981	58	88	.	.	PUNCT
ajst-2981	58	89	intersection	intersection	NOUN
ajst-2981	58	90	of	of	ADP
ajst-2981	58	91	north	north	ADJ
ajst-2981	58	92	new	new	ADJ
ajst-2981	58	93	road	road	NOUN
ajst-2981	58	94	and	and	CCONJ
ajst-2981	58	95	college	college	NOUN
ajst-2981	58	96	road	road	NOUN
ajst-2981	58	97	4.2	4.2	NUM
ajst-2981	58	98	.	.	PUNCT
ajst-2981	59	1	analysis	analysis	NOUN
ajst-2981	59	2	of	of	ADP
ajst-2981	59	3	experimental	experimental	ADJ
ajst-2981	59	4	results	result	NOUN
ajst-2981	59	5	intersection	intersection	NOUN
ajst-2981	59	6	name	name	NOUN
ajst-2981	59	7	manual	manual	ADJ
ajst-2981	59	8	statistics	statistic	NOUN
ajst-2981	59	9	(	(	PUNCT
ajst-2981	59	10	units	unit	NOUN
ajst-2981	59	11	)	)	PUNCT
ajst-2981	59	12	algorithm	algorithm	PROPN
ajst-2981	59	13	statistics	statistic	NOUN
ajst-2981	59	14	(	(	PUNCT
ajst-2981	59	15	units	unit	NOUN
ajst-2981	59	16	)	)	PUNCT
ajst-2981	59	17	average	average	ADJ
ajst-2981	59	18	accuracy	accuracy	NOUN
ajst-2981	59	19	(	(	PUNCT
ajst-2981	59	20	%	%	INTJ
ajst-2981	59	21	)	)	PUNCT
ajst-2981	59	22	east	east	NOUN
ajst-2981	59	23	import	import	NOUN
ajst-2981	59	24	go	go	VERB
ajst-2981	59	25	straight	straight	ADV
ajst-2981	59	26	189	189	NUM
ajst-2981	59	27	186	186	NUM
ajst-2981	59	28	94.9	94.9	NUM
ajst-2981	59	29	left	leave	VERB
ajst-2981	59	30	69	69	NUM
ajst-2981	59	31	63	63	NUM
ajst-2981	59	32	turn	turn	NOUN
ajst-2981	59	33	right	right	ADJ
ajst-2981	59	34	75	75	NUM
ajst-2981	59	35	71	71	NUM
ajst-2981	59	36	west	west	ADJ
ajst-2981	59	37	import	import	NOUN
ajst-2981	59	38	go	go	VERB
ajst-2981	59	39	straight	straight	ADV
ajst-2981	59	40	137	137	NUM
ajst-2981	59	41	135	135	NUM
ajst-2981	59	42	94.7	94.7	NUM
ajst-2981	59	43	left	leave	VERB
ajst-2981	59	44	7	7	NUM
ajst-2981	59	45	6	6	NUM
ajst-2981	59	46	turn	turn	NOUN
ajst-2981	59	47	right	right	ADJ
ajst-2981	59	48	36	36	NUM
ajst-2981	59	49	36	36	NUM
ajst-2981	59	50	south	south	ADJ
ajst-2981	59	51	import	import	NOUN
ajst-2981	59	52	go	go	VERB
ajst-2981	59	53	straight	straight	ADV
ajst-2981	59	54	103	103	NUM
ajst-2981	59	55	102	102	NUM
ajst-2981	59	56	97.5	97.5	NUM
ajst-2981	59	57	left	leave	VERB
ajst-2981	59	58	72	72	NUM
ajst-2981	59	59	70	70	NUM
ajst-2981	59	60	turn	turn	NOUN
ajst-2981	59	61	right	right	ADJ
ajst-2981	59	62	54	54	NUM
ajst-2981	59	63	52	52	NUM
ajst-2981	59	64	north	north	NOUN
ajst-2981	59	65	import	import	NOUN
ajst-2981	59	66	go	go	VERB
ajst-2981	59	67	straight	straight	ADV
ajst-2981	59	68	72	72	NUM
ajst-2981	59	69	71	71	NUM
ajst-2981	59	70	97.7	97.7	NUM
ajst-2981	59	71	left	leave	VERB
ajst-2981	59	72	63	63	NUM
ajst-2981	59	73	63	63	NUM
ajst-2981	59	74	turn	turn	NOUN
ajst-2981	59	75	right	right	ADJ
ajst-2981	59	76	18	18	NUM
ajst-2981	59	77	17	17	NUM
ajst-2981	59	78	as	as	SCONJ
ajst-2981	59	79	can	can	AUX
ajst-2981	59	80	be	be	AUX
ajst-2981	59	81	seen	see	VERB
ajst-2981	59	82	from	from	ADP
ajst-2981	59	83	the	the	DET
ajst-2981	59	84	above	above	ADJ
ajst-2981	59	85	table	table	NOUN
ajst-2981	59	86	,	,	PUNCT
ajst-2981	59	87	the	the	DET
ajst-2981	59	88	average	average	ADJ
ajst-2981	59	89	accuracy	accuracy	NOUN
ajst-2981	59	90	of	of	ADP
ajst-2981	59	91	the	the	DET
ajst-2981	59	92	algorithm	algorithm	NOUN
ajst-2981	59	93	results	result	NOUN
ajst-2981	59	94	of	of	ADP
ajst-2981	59	95	the	the	DET
ajst-2981	59	96	east	east	PROPN
ajst-2981	59	97	import	import	NOUN
ajst-2981	59	98	algorithm	algorithm	NOUN
ajst-2981	59	99	at	at	ADP
ajst-2981	59	100	the	the	DET
ajst-2981	59	101	intersection	intersection	NOUN
ajst-2981	59	102	of	of	ADP
ajst-2981	59	103	beixin	beixin	PROPN
ajst-2981	59	104	road	road	PROPN
ajst-2981	59	105	and	and	CCONJ
ajst-2981	59	106	xueyuan	xueyuan	PROPN
ajst-2981	59	107	road	road	NOUN
ajst-2981	59	108	is	be	AUX
ajst-2981	59	109	94.9	94.9	NUM
ajst-2981	59	110	%	%	NOUN
ajst-2981	59	111	compared	compare	VERB
ajst-2981	59	112	with	with	ADP
ajst-2981	59	113	the	the	DET
ajst-2981	59	114	manual	manual	ADJ
ajst-2981	59	115	statistical	statistical	ADJ
ajst-2981	59	116	results	result	NOUN
ajst-2981	59	117	.	.	PUNCT
ajst-2981	60	1	compared	compare	VERB
ajst-2981	60	2	with	with	ADP
ajst-2981	60	3	the	the	DET
ajst-2981	60	4	manual	manual	ADJ
ajst-2981	60	5	statistical	statistical	ADJ
ajst-2981	60	6	results	result	NOUN
ajst-2981	60	7	,	,	PUNCT
ajst-2981	60	8	the	the	DET
ajst-2981	60	9	average	average	ADJ
ajst-2981	60	10	accuracy	accuracy	NOUN
ajst-2981	60	11	rate	rate	NOUN
ajst-2981	60	12	of	of	ADP
ajst-2981	60	13	the	the	DET
ajst-2981	60	14	western	western	ADJ
ajst-2981	60	15	import	import	NOUN
ajst-2981	60	16	algorithm	algorithm	NOUN
ajst-2981	60	17	reached	reach	VERB
ajst-2981	60	18	94.7	94.7	NUM
ajst-2981	60	19	%	%	NOUN
ajst-2981	60	20	;	;	PUNCT
ajst-2981	60	21	compared	compare	VERB
ajst-2981	60	22	with	with	ADP
ajst-2981	60	23	the	the	DET
ajst-2981	60	24	manual	manual	ADJ
ajst-2981	60	25	statistical	statistical	ADJ
ajst-2981	60	26	results	result	NOUN
ajst-2981	60	27	,	,	PUNCT
ajst-2981	60	28	the	the	DET
ajst-2981	60	29	average	average	ADJ
ajst-2981	60	30	accuracy	accuracy	NOUN
ajst-2981	60	31	rate	rate	NOUN
ajst-2981	60	32	of	of	ADP
ajst-2981	60	33	the	the	DET
ajst-2981	60	34	south	south	PROPN
ajst-2981	60	35	import	import	NOUN
ajst-2981	60	36	algorithm	algorithm	NOUN
ajst-2981	60	37	reached	reach	VERB
ajst-2981	60	38	97.5	97.5	NUM
ajst-2981	60	39	%	%	NOUN
ajst-2981	60	40	;	;	PUNCT
ajst-2981	60	41	compared	compare	VERB
ajst-2981	60	42	with	with	ADP
ajst-2981	60	43	the	the	DET
ajst-2981	60	44	manual	manual	ADJ
ajst-2981	60	45	statistical	statistical	ADJ
ajst-2981	60	46	results	result	NOUN
ajst-2981	60	47	,	,	PUNCT
ajst-2981	60	48	the	the	DET
ajst-2981	60	49	average	average	ADJ
ajst-2981	60	50	accuracy	accuracy	NOUN
ajst-2981	60	51	of	of	ADP
ajst-2981	60	52	the	the	DET
ajst-2981	60	53	north	north	NOUN
ajst-2981	60	54	import	import	NOUN
ajst-2981	60	55	algorithm	algorithm	NOUN
ajst-2981	60	56	reached	reach	VERB
ajst-2981	60	57	97.7	97.7	NUM
ajst-2981	60	58	%	%	NOUN
ajst-2981	60	59	.	.	PUNCT
ajst-2981	61	1	on	on	ADP
ajst-2981	61	2	the	the	DET
ajst-2981	61	3	whole	whole	NOUN
ajst-2981	61	4	,	,	PUNCT
ajst-2981	61	5	the	the	DET
ajst-2981	61	6	statistical	statistical	ADJ
ajst-2981	61	7	results	result	NOUN
ajst-2981	61	8	of	of	ADP
ajst-2981	61	9	the	the	DET
ajst-2981	61	10	algorithm	algorithm	NOUN
ajst-2981	61	11	in	in	ADP
ajst-2981	61	12	this	this	DET
ajst-2981	61	13	paper	paper	NOUN
ajst-2981	61	14	are	be	AUX
ajst-2981	61	15	close	close	ADJ
ajst-2981	61	16	to	to	ADP
ajst-2981	61	17	the	the	DET
ajst-2981	61	18	manual	manual	ADJ
ajst-2981	61	19	statistical	statistical	ADJ
ajst-2981	61	20	results	result	NOUN
ajst-2981	61	21	,	,	PUNCT
ajst-2981	61	22	which	which	PRON
ajst-2981	61	23	has	have	VERB
ajst-2981	61	24	high	high	ADJ
ajst-2981	61	25	practical	practical	ADJ
ajst-2981	61	26	value	value	NOUN
ajst-2981	61	27	.	.	PUNCT
ajst-2981	62	1	201	201	NUM
ajst-2981	62	2	5	5	NUM
ajst-2981	62	3	.	.	PUNCT
ajst-2981	62	4	conclusion	conclusion	NOUN
ajst-2981	62	5	in	in	ADP
ajst-2981	62	6	this	this	DET
ajst-2981	62	7	paper	paper	NOUN
ajst-2981	62	8	,	,	PUNCT
ajst-2981	62	9	the	the	DET
ajst-2981	62	10	strategy	strategy	NOUN
ajst-2981	62	11	of	of	ADP
ajst-2981	62	12	combining	combine	VERB
ajst-2981	62	13	yolov5	yolov5	NOUN
ajst-2981	62	14	object	object	NOUN
ajst-2981	62	15	detection	detection	NOUN
ajst-2981	62	16	algorithm	algorithm	NOUN
ajst-2981	62	17	and	and	CCONJ
ajst-2981	62	18	deepsort	deepsort	VERB
ajst-2981	62	19	multi	multi	ADJ
ajst-2981	62	20	-	-	NOUN
ajst-2981	62	21	target	target	ADJ
ajst-2981	62	22	tracking	tracking	NOUN
ajst-2981	62	23	algorithm	algorithm	NOUN
ajst-2981	62	24	is	be	AUX
ajst-2981	62	25	used	use	VERB
ajst-2981	62	26	to	to	PART
ajst-2981	62	27	realize	realize	VERB
ajst-2981	62	28	the	the	DET
ajst-2981	62	29	statistics	statistic	NOUN
ajst-2981	62	30	of	of	ADP
ajst-2981	62	31	traffic	traffic	NOUN
ajst-2981	62	32	flow	flow	NOUN
ajst-2981	62	33	at	at	ADP
ajst-2981	62	34	traffic	traffic	NOUN
ajst-2981	62	35	intersections	intersection	NOUN
ajst-2981	62	36	under	under	ADP
ajst-2981	62	37	uav	uav	PROPN
ajst-2981	62	38	aerial	aerial	ADJ
ajst-2981	62	39	video	video	NOUN
ajst-2981	62	40	,	,	PUNCT
ajst-2981	62	41	and	and	CCONJ
ajst-2981	62	42	the	the	DET
ajst-2981	62	43	experimental	experimental	ADJ
ajst-2981	62	44	results	result	NOUN
ajst-2981	62	45	show	show	VERB
ajst-2981	62	46	that	that	SCONJ
ajst-2981	62	47	the	the	DET
ajst-2981	62	48	proposed	propose	VERB
ajst-2981	62	49	algorithm	algorithm	NOUN
ajst-2981	62	50	shows	show	VERB
ajst-2981	62	51	high	high	ADJ
ajst-2981	62	52	statistical	statistical	ADJ
ajst-2981	62	53	accuracy	accuracy	NOUN
ajst-2981	62	54	and	and	CCONJ
ajst-2981	62	55	is	be	AUX
ajst-2981	62	56	practical	practical	ADJ
ajst-2981	62	57	.	.	PUNCT
ajst-2981	63	1	although	although	SCONJ
ajst-2981	63	2	the	the	DET
ajst-2981	63	3	model	model	NOUN
ajst-2981	63	4	can	can	AUX
ajst-2981	63	5	run	run	VERB
ajst-2981	63	6	smoothly	smoothly	ADV
ajst-2981	63	7	,	,	PUNCT
ajst-2981	63	8	it	it	PRON
ajst-2981	63	9	is	be	AUX
ajst-2981	63	10	still	still	ADV
ajst-2981	63	11	too	too	ADV
ajst-2981	63	12	burdensome	burdensome	ADJ
ajst-2981	63	13	for	for	ADP
ajst-2981	63	14	some	some	DET
ajst-2981	63	15	devices	device	NOUN
ajst-2981	63	16	with	with	ADP
ajst-2981	63	17	low	low	ADJ
ajst-2981	63	18	computing	computing	NOUN
ajst-2981	63	19	power	power	NOUN
ajst-2981	63	20	,	,	PUNCT
ajst-2981	63	21	and	and	CCONJ
ajst-2981	63	22	the	the	DET
ajst-2981	63	23	model	model	NOUN
ajst-2981	63	24	will	will	AUX
ajst-2981	63	25	be	be	AUX
ajst-2981	63	26	further	far	ADV
ajst-2981	63	27	compressed	compress	VERB
ajst-2981	63	28	and	and	CCONJ
ajst-2981	63	29	lightweight	lightweight	ADJ
ajst-2981	63	30	and	and	CCONJ
ajst-2981	63	31	transplanted	transplant	VERB
ajst-2981	63	32	to	to	ADP
ajst-2981	63	33	embedded	embed	VERB
ajst-2981	63	34	devices	device	NOUN
ajst-2981	63	35	.	.	PUNCT
ajst-2981	64	1	acknowledgment	acknowledgment	NOUN
ajst-2981	64	2	fund	fund	PROPN
ajst-2981	64	3	:	:	PUNCT
ajst-2981	64	4	natural	natural	ADJ
ajst-2981	64	5	science	science	PROPN
ajst-2981	64	6	foundation	foundation	PROPN
ajst-2981	64	7	of	of	ADP
ajst-2981	64	8	jiangsu	jiangsu	PROPN
ajst-2981	64	9	province	province	PROPN
ajst-2981	64	10	(	(	PUNCT
ajst-2981	64	11	20170157	20170157	NUM
ajst-2981	64	12	)	)	PUNCT
ajst-2981	64	13	.	.	PUNCT
ajst-2981	65	1	references	reference	NOUN
ajst-2981	65	2	[	[	X
ajst-2981	65	3	1	1	NUM
ajst-2981	65	4	]	]	X
ajst-2981	65	5	guan	guan	PROPN
ajst-2981	65	6	jizhen	jizhen	PROPN
ajst-2981	65	7	.	.	PUNCT
ajst-2981	66	1	the	the	DET
ajst-2981	66	2	development	development	NOUN
ajst-2981	66	3	and	and	CCONJ
ajst-2981	66	4	evolution	evolution	NOUN
ajst-2981	66	5	of	of	ADP
ajst-2981	66	6	intelligent	intelligent	ADJ
ajst-2981	66	7	transportation	transportation	NOUN
ajst-2981	66	8	system	system	NOUN
ajst-2981	66	9	and	and	CCONJ
ajst-2981	66	10	its	its	PRON
ajst-2981	66	11	intergenerational	intergenerational	ADJ
ajst-2981	66	12	characteristics[j].artificial	characteristics[j].artificial	ADJ
ajst-2981	66	13	intelligence,2022(04):4049.doi:10.16453	intelligence,2022(04):4049.doi:10.16453	NOUN
ajst-2981	66	14	/	/	SYM
ajst-2981	66	15	j.cnki.issn2096	j.cnki.issn2096	NUM
ajst-2981	66	16	-	-	PUNCT
ajst-2981	66	17	5036.2022.04.004	5036.2022.04.004	PROPN
ajst-2981	66	18	.	.	PUNCT
ajst-2981	66	19	)	)	PUNCT
ajst-2981	67	1	[	[	X
ajst-2981	67	2	2	2	X
ajst-2981	67	3	]	]	X
ajst-2981	67	4	viola	viola	PROPN
ajst-2981	67	5	p，jones	p，jones	PROPN
ajst-2981	67	6	m.rapid	m.rapid	ADJ
ajst-2981	67	7	object	object	NOUN
ajst-2981	67	8	detection	detection	NOUN
ajst-2981	67	9	using	use	VERB
ajst-2981	67	10	aboosted	abooste	VERB
ajst-2981	67	11	cascade	cascade	NOUN
ajst-2981	67	12	of	of	ADP
ajst-2981	67	13	simple	simple	ADJ
ajst-2981	67	14	features[c]//proceedings	features[c]//proceeding	NOUN
ajst-2981	67	15	of	of	ADP
ajst-2981	67	16	the2001	the2001	ADJ
ajst-2981	67	17	ieee	ieee	NOUN
ajst-2981	67	18	computer	computer	NOUN
ajst-2981	67	19	society	society	PROPN
ajst-2981	67	20	conference	conference	NOUN
ajst-2981	67	21	on	on	ADP
ajst-2981	67	22	computer	computer	NOUN
ajst-2981	67	23	vision	vision	NOUN
ajst-2981	67	24	and	and	CCONJ
ajst-2981	67	25	pattern	pattern	NOUN
ajst-2981	67	26	recognition	recognition	NOUN
ajst-2981	67	27	,	,	PUNCT
ajst-2981	67	28	2001	2001	NUM
ajst-2981	67	29	.	.	PUNCT
ajst-2981	68	1	[	[	X
ajst-2981	68	2	3	3	X
ajst-2981	68	3	]	]	X
ajst-2981	68	4	dalal	dalal	PROPN
ajst-2981	68	5	n，triggs	n，trigg	NOUN
ajst-2981	68	6	b.histograms	b.histogram	NOUN
ajst-2981	68	7	of	of	ADP
ajst-2981	68	8	oriented	orient	VERB
ajst-2981	68	9	gradientsfor	gradientsfor	NOUN
ajst-2981	68	10	human	human	ADJ
ajst-2981	68	11	detection[c]//2005	detection[c]//2005	PROPN
ajst-2981	68	12	ieee	ieee	NOUN
ajst-2981	68	13	computer	computer	NOUN
ajst-2981	68	14	societyconference	societyconference	NOUN
ajst-2981	68	15	on	on	ADP
ajst-2981	68	16	computer	computer	NOUN
ajst-2981	68	17	vision	vision	NOUN
ajst-2981	68	18	and	and	CCONJ
ajst-2981	68	19	pattern	pattern	NOUN
ajst-2981	68	20	recognition	recognition	NOUN
ajst-2981	68	21	(	(	PUNCT
ajst-2981	68	22	cvpr’05	cvpr’05	PROPN
ajst-2981	68	23	)	)	PUNCT
ajst-2981	68	24	,	,	PUNCT
ajst-2981	68	25	2005	2005	NUM
ajst-2981	68	26	:	:	PUNCT
ajst-2981	68	27	886	886	NUM
ajst-2981	68	28	-	-	SYM
ajst-2981	68	29	893	893	NUM
ajst-2981	68	30	.	.	PUNCT
ajst-2981	69	1	[	[	X
ajst-2981	69	2	4	4	X
ajst-2981	69	3	]	]	PUNCT
ajst-2981	69	4	glrshick	glrshick	NOUN
ajst-2981	69	5	r	r	PROPN
ajst-2981	69	6	,	,	PUNCT
ajst-2981	69	7	donahue	donahue	PROPN
ajst-2981	69	8	j	j	PROPN
ajst-2981	69	9	,	,	PUNCT
ajst-2981	69	10	darrell	darrell	PROPN
ajst-2981	69	11	t	t	PROPN
ajst-2981	69	12	,	,	PUNCT
ajst-2981	69	13	et	et	PROPN
ajst-2981	69	14	al	al	PROPN
ajst-2981	69	15	.	.	PROPN
ajst-2981	70	1	rich	rich	ADJ
ajst-2981	70	2	feature	feature	NOUN
ajst-2981	70	3	hierarchies	hierarchy	NOUN
ajst-2981	70	4	for	for	ADP
ajst-2981	70	5	accurate	accurate	ADJ
ajst-2981	70	6	object	object	NOUN
ajst-2981	70	7	detec	detec	ADJ
ajst-2981	70	8	-	-	PUNCT
ajst-2981	70	9	tion	tion	NOUN
ajst-2981	70	10	and	and	CCONJ
ajst-2981	70	11	semantic	semantic	ADJ
ajst-2981	70	12	segmentation	segmentation	NOUN
ajst-2981	70	13	[	[	X
ajst-2981	70	14	cj	cj	X
ajst-2981	70	15	/	/	SYM
ajst-2981	70	16	/	/	SYM
ajst-2981	70	17	ieee	ieee	NOUN
ajst-2981	70	18	.	.	PUNCT
ajst-2981	71	1	pro	pro	ADJ
ajst-2981	71	2	-	-	NOUN
ajst-2981	71	3	ceedings	ceeding	NOUN
ajst-2981	71	4	of	of	ADP
ajst-2981	71	5	the	the	DET
ajst-2981	71	6	leee	leee	PROPN
ajst-2981	71	7	conference	conference	PROPN
ajst-2981	71	8	on	on	ADP
ajst-2981	71	9	computer	computer	NOUN
ajst-2981	71	10	v1s1on	v1s1on	NOUN
ajst-2981	71	11	and	and	CCONJ
ajst-2981	71	12	pattern	pattern	NOUN
ajst-2981	71	13	recognition	recognition	NOUN
ajst-2981	71	14	.	.	PUNCT
ajst-2981	72	1	new	new	PROPN
ajst-2981	72	2	york	york	PROPN
ajst-2981	72	3	:	:	PUNCT
ajst-2981	72	4	ieee	ieee	NOUN
ajst-2981	72	5	,	,	PUNCT
ajst-2981	72	6	2014:580	2014:580	NUM
ajst-2981	72	7	-	-	SYM
ajst-2981	72	8	587	587	NUM
ajst-2981	72	9	.	.	PUNCT
ajst-2981	72	10	]	]	X
ajst-2981	73	1	[	[	X
ajst-2981	73	2	5	5	NUM
ajst-2981	73	3	]	]	X
ajst-2981	73	4	girshick	girshick	PROPN
ajst-2981	73	5	r.	r.	PROPN
ajst-2981	73	6	fast	fast	PROPN
ajst-2981	73	7	rcnn[j]in	rcnn[j]in	PROPN
ajst-2981	73	8	proceedings	proceeding	NOUN
ajst-2981	73	9	of	of	ADP
ajst-2981	73	10	the	the	DET
ajst-2981	73	11	ieee	ieee	NOUN
ajst-2981	73	12	international	international	PROPN
ajst-2981	73	13	conference	conference	NOUN
ajst-2981	73	14	on	on	ADP
ajst-2981	73	15	computer	computer	NOUN
ajst-2981	73	16	vision,2015,7	vision,2015,7	NOUN
ajst-2981	73	17	:	:	PUNCT
ajst-2981	73	18	14401448	14401448	NUM
ajst-2981	73	19	.	.	PUNCT
ajst-2981	74	1	[	[	X
ajst-2981	74	2	6	6	NUM
ajst-2981	74	3	]	]	X
ajst-2981	74	4	ren	ren	PROPN
ajst-2981	74	5	s	s	PROPN
ajst-2981	74	6	,	,	PUNCT
ajst-2981	74	7	he	he	PRON
ajst-2981	74	8	k	k	NOUN
ajst-2981	74	9	,	,	PUNCT
ajst-2981	74	10	gir	gir	VERB
ajst-2981	74	11	shick	shick	NOUN
ajst-2981	74	12	r	r	NOUN
ajst-2981	74	13	,	,	PUNCT
ajst-2981	74	14	et	et	PROPN
ajst-2981	74	15	al．faster	al．faster	PROPN
ajst-2981	74	16	rcnn	rcnn	PROPN
ajst-2981	74	17	:	:	PUNCT
ajst-2981	74	18	towards	towards	ADP
ajst-2981	74	19	real	real	ADJ
ajst-2981	74	20	-	-	PUNCT
ajst-2981	74	21	time	time	NOUN
ajst-2981	74	22	object	object	NOUN
ajst-2981	74	23	detection	detection	NOUN
ajst-2981	74	24	with	with	ADP
ajst-2981	74	25	region	region	NOUN
ajst-2981	74	26	proposal	proposal	NOUN
ajst-2981	74	27	network	network	NOUN
ajst-2981	74	28	[	[	X
ajst-2981	74	29	j].ieee	j].ieee	NOUN
ajst-2981	74	30	transactions	transaction	NOUN
ajst-2981	74	31	on	on	ADP
ajst-2981	74	32	pattern	pattern	NOUN
ajst-2981	74	33	analysis	analysis	NOUN
ajst-2981	74	34	and	and	CCONJ
ajst-2981	74	35	machine	machine	NOUN
ajst-2981	74	36	intelligence,2017,39	intelligence,2017,39	NOUN
ajst-2981	74	37	:	:	PUNCT
ajst-2981	74	38	1137	1137	NUM
ajst-2981	74	39	-	-	SYM
ajst-2981	74	40	1149	1149	NUM
ajst-2981	74	41	.	.	PUNCT
ajst-2981	75	1	[	[	X
ajst-2981	75	2	7	7	X
ajst-2981	75	3	]	]	X
ajst-2981	75	4	liu	liu	PROPN
ajst-2981	75	5	w	w	PROPN
ajst-2981	75	6	,	,	PUNCT
ajst-2981	75	7	anguelov	anguelov	NOUN
ajst-2981	75	8	d	d	NOUN
ajst-2981	75	9	,	,	PUNCT
ajst-2981	75	10	erhan	erhan	SCONJ
ajst-2981	75	11	d	d	PROPN
ajst-2981	75	12	,	,	PUNCT
ajst-2981	75	13	et	et	PROPN
ajst-2981	75	14	al.ssd	al.ssd	PROPN
ajst-2981	75	15	:	:	PUNCT
ajst-2981	75	16	single	single	ADJ
ajst-2981	75	17	shot	shot	NOUN
ajst-2981	75	18	multibox	multibox	PROPN
ajst-2981	75	19	detector[c]//european	detector[c]//european	PROPN
ajst-2981	75	20	conference	conference	NOUN
ajst-2981	75	21	on	on	ADP
ajst-2981	75	22	computer	computer	NOUN
ajst-2981	75	23	vision	vision	NOUN
ajst-2981	75	24	,	,	PUNCT
ajst-2981	75	25	2016	2016	NUM
ajst-2981	75	26	:	:	PUNCT
ajst-2981	75	27	21	21	NUM
ajst-2981	75	28	-	-	SYM
ajst-2981	75	29	37	37	NUM
ajst-2981	75	30	.	.	PUNCT
ajst-2981	76	1	[	[	X
ajst-2981	76	2	8	8	NUM
ajst-2981	76	3	]	]	X
ajst-2981	76	4	redmon	redmon	PROPN
ajst-2981	76	5	j	j	PROPN
ajst-2981	76	6	,	,	PUNCT
ajst-2981	76	7	farhadi	farhadi	PROPN
ajst-2981	76	8	a.yolov3	a.yolov3	PROPN
ajst-2981	76	9	:	:	PUNCT
ajst-2981	76	10	an	an	DET
ajst-2981	76	11	incremental	incremental	ADJ
ajst-2981	76	12	improvement[j].arxiv	improvement[j].arxiv	NOUN
ajst-2981	76	13	:	:	PUNCT
ajst-2981	76	14	1804.02767	1804.02767	NUM
ajst-2981	76	15	,	,	PUNCT
ajst-2981	76	16	2018	2018	NUM
ajst-2981	76	17	.	.	PUNCT
ajst-2981	77	1	[	[	X
ajst-2981	77	2	9	9	NUM
ajst-2981	77	3	]	]	PUNCT
ajst-2981	77	4	bochkovskiy	bochkovskiy	X
ajst-2981	77	5	a	a	PRON
ajst-2981	77	6	,	,	PUNCT
ajst-2981	77	7	wang	wang	PROPN
ajst-2981	77	8	c	c	PROPN
ajst-2981	77	9	y	y	PROPN
ajst-2981	77	10	,	,	PUNCT
ajst-2981	77	11	liao	liao	PROPN
ajst-2981	77	12	h	h	PROPN
ajst-2981	77	13	y	y	PROPN
ajst-2981	77	14	m.yolov4	m.yolov4	PROPN
ajst-2981	77	15	:	:	PUNCT
ajst-2981	77	16	optimal	optimal	ADJ
ajst-2981	77	17	speed	speed	NOUN
ajst-2981	77	18	and	and	CCONJ
ajst-2981	77	19	accuracy	accuracy	NOUN
ajst-2981	77	20	of	of	ADP
ajst-2981	77	21	object	object	NOUN
ajst-2981	77	22	detection[j].arxiv	detection[j].arxiv	NOUN
ajst-2981	77	23	:	:	PUNCT
ajst-2981	77	24	2004.10934	2004.10934	NOUN
ajst-2981	77	25	,	,	PUNCT
ajst-2981	77	26	2020	2020	NUM
ajst-2981	77	27	.	.	PUNCT
ajst-2981	78	1	[	[	X
ajst-2981	78	2	10	10	NUM
ajst-2981	78	3	]	]	SYM
ajst-2981	78	4	ultralytics.yolov5[eb	ultralytics.yolov5[eb	NOUN
ajst-2981	78	5	/	/	SYM
ajst-2981	78	6	ol].[2021].https	ol].[2021].https	PROPN
ajst-2981	78	7	:	:	PUNCT
ajst-2981	78	8	//github.com	//github.com	SYM
ajst-2981	78	9	/	/	SYM
ajst-2981	78	10	ultralytics	ultralytic	NOUN
ajst-2981	78	11	/	/	SYM
ajst-2981	78	12	yolov5	yolov5	NOUN
ajst-2981	78	13	.	.	PUNCT
ajst-2981	79	1	[	[	X
ajst-2981	79	2	11	11	NUM
ajst-2981	79	3	]	]	X
ajst-2981	79	4	liu	liu	PROPN
ajst-2981	79	5	lei	lei	PROPN
ajst-2981	79	6	.	.	PROPN
ajst-2981	79	7	statistical	statistical	ADJ
ajst-2981	79	8	research	research	NOUN
ajst-2981	79	9	on	on	ADP
ajst-2981	79	10	intelligent	intelligent	ADJ
ajst-2981	79	11	traffic	traffic	NOUN
ajst-2981	79	12	flow	flow	NOUN
ajst-2981	79	13	based	base	VERB
ajst-2981	79	14	on	on	ADP
ajst-2981	79	15	yolo	yolo	PROPN
ajst-2981	79	16	network[d].xi'an	network[d].xi'an	PROPN
ajst-2981	79	17	university	university	PROPN
ajst-2981	79	18	of	of	ADP
ajst-2981	79	19	science	science	NOUN
ajst-2981	79	20	and	and	CCONJ
ajst-2981	79	21	technology,2019	technology,2019	NOUN
ajst-2981	79	22	.	.	PUNCT
ajst-2981	79	23	)	)	PUNCT
ajst-2981	80	1	[	[	X
ajst-2981	80	2	12	12	NUM
ajst-2981	80	3	]	]	X
ajst-2981	80	4	zhao	zhao	PROPN
ajst-2981	80	5	kaidi	kaidi	PROPN
ajst-2981	80	6	.	.	PUNCT
ajst-2981	81	1	design	design	NOUN
ajst-2981	81	2	and	and	CCONJ
ajst-2981	81	3	implementation	implementation	NOUN
ajst-2981	81	4	of	of	ADP
ajst-2981	81	5	vehicle	vehicle	NOUN
ajst-2981	81	6	and	and	CCONJ
ajst-2981	81	7	lane	lane	NOUN
ajst-2981	81	8	detection	detection	NOUN
ajst-2981	81	9	system	system	NOUN
ajst-2981	81	10	based	base	VERB
ajst-2981	81	11	on	on	ADP
ajst-2981	81	12	uav[d].xidian	uav[d].xidian	ADJ
ajst-2981	81	13	university,2018	university,2018	NOUN
ajst-2981	81	14	.	.	PUNCT
ajst-2981	81	15	)	)	PUNCT
ajst-2981	82	1	[	[	X
ajst-2981	82	2	13	13	NUM
ajst-2981	82	3	]	]	SYM
ajst-2981	82	4	hu	hu	PROPN
ajst-2981	82	5	zhicheng	zhicheng	PROPN
ajst-2981	82	6	,	,	PUNCT
ajst-2981	82	7	xie	xie	PROPN
ajst-2981	82	8	wei	wei	PROPN
ajst-2981	82	9	,	,	PUNCT
ajst-2981	82	10	wang	wang	PROPN
ajst-2981	82	11	jinsong	jinsong	PROPN
ajst-2981	82	12	,	,	PUNCT
ajst-2981	82	13	cui	cui	PROPN
ajst-2981	82	14	zhen	zhen	PROPN
ajst-2981	82	15	.	.	PUNCT
ajst-2981	83	1	research	research	NOUN
ajst-2981	83	2	and	and	CCONJ
ajst-2981	83	3	application	application	NOUN
ajst-2981	83	4	of	of	ADP
ajst-2981	83	5	video	video	NOUN
ajst-2981	83	6	detection	detection	NOUN
ajst-2981	83	7	technology	technology	NOUN
ajst-2981	83	8	on	on	ADP
ajst-2981	83	9	vehicle	vehicle	NOUN
ajst-2981	83	10	flow[j].digital	flow[j].digital	PROPN
ajst-2981	83	11	world,2017(12):89	world,2017(12):89	NOUN
ajst-2981	83	12	-	-	PUNCT
ajst-2981	83	13	91	91	NUM
ajst-2981	83	14	.	.	PUNCT
ajst-2981	83	15	)	)	PUNCT