id	sid	tid	token	lemma	pos
fcis-18462	1	1	frontiers	frontier	NOUN
fcis-18462	1	2	in	in	ADP
fcis-18462	1	3	computing	computing	NOUN
fcis-18462	1	4	and	and	CCONJ
fcis-18462	1	5	intelligent	intelligent	ADJ
fcis-18462	1	6	systems	system	NOUN
fcis-18462	1	7	issn	issn	VERB
fcis-18462	1	8	:	:	PUNCT
fcis-18462	1	9	2832	2832	NUM
fcis-18462	1	10	-	-	SYM
fcis-18462	1	11	6024	6024	NUM
fcis-18462	1	12	|	|	NOUN
fcis-18462	1	13	vol	vol	NOUN
fcis-18462	1	14	.	.	PROPN
fcis-18462	2	1	7	7	NUM
fcis-18462	2	2	,	,	PUNCT
fcis-18462	2	3	no	no	INTJ
fcis-18462	2	4	.	.	NOUN
fcis-18462	2	5	2	2	NUM
fcis-18462	2	6	,	,	PUNCT
fcis-18462	2	7	2024	2024	NUM
fcis-18462	2	8	75	75	NUM
fcis-18462	2	9	research	research	NOUN
fcis-18462	2	10	on	on	ADP
fcis-18462	2	11	the	the	DET
fcis-18462	2	12	detection	detection	NOUN
fcis-18462	2	13	of	of	ADP
fcis-18462	2	14	traffic	traffic	NOUN
fcis-18462	2	15	flow	flow	NOUN
fcis-18462	2	16	based	base	VERB
fcis-18462	2	17	on	on	ADP
fcis-18462	2	18	video	video	NOUN
fcis-18462	2	19	images	image	NOUN
fcis-18462	2	20	jian	jian	PROPN
fcis-18462	2	21	he	he	PRON
fcis-18462	2	22	,	,	PUNCT
fcis-18462	2	23	wei	wei	PROPN
fcis-18462	2	24	teng	teng	PROPN
fcis-18462	2	25	*	*	PUNCT
fcis-18462	2	26	,	,	PUNCT
fcis-18462	2	27	zeyu	zeyu	PROPN
fcis-18462	2	28	zhao	zhao	NOUN
fcis-18462	2	29	,	,	PUNCT
fcis-18462	2	30	binche	binche	PROPN
fcis-18462	2	31	liu	liu	PROPN
fcis-18462	2	32	,	,	PUNCT
fcis-18462	2	33	bing	bing	VERB
fcis-18462	2	34	qin	qin	PROPN
fcis-18462	2	35	and	and	CCONJ
fcis-18462	2	36	jun	jun	PROPN
fcis-18462	2	37	jiang	jiang	PROPN
fcis-18462	2	38	school	school	PROPN
fcis-18462	2	39	of	of	ADP
fcis-18462	2	40	computer	computer	NOUN
fcis-18462	2	41	science	science	NOUN
fcis-18462	2	42	and	and	CCONJ
fcis-18462	2	43	software	software	NOUN
fcis-18462	2	44	engineering	engineering	NOUN
fcis-18462	2	45	,	,	PUNCT
fcis-18462	2	46	university	university	NOUN
fcis-18462	2	47	of	of	ADP
fcis-18462	2	48	science	science	NOUN
fcis-18462	2	49	and	and	CCONJ
fcis-18462	2	50	technology	technology	NOUN
fcis-18462	2	51	liaoning	liaoning	NOUN
fcis-18462	2	52	,	,	PUNCT
fcis-18462	2	53	anshan	anshan	PROPN
fcis-18462	2	54	liaoning	liaoning	NOUN
fcis-18462	2	55	,	,	PUNCT
fcis-18462	2	56	114051	114051	NUM
fcis-18462	2	57	,	,	PUNCT
fcis-18462	2	58	china	china	PROPN
fcis-18462	2	59	*	*	PUNCT
fcis-18462	2	60	corresponding	correspond	VERB
fcis-18462	2	61	author	author	NOUN
fcis-18462	2	62	:	:	PUNCT
fcis-18462	2	63	wei	wei	PROPN
fcis-18462	2	64	teng	teng	PROPN
fcis-18462	2	65	abstract	abstract	PROPN
fcis-18462	2	66	:	:	PUNCT
fcis-18462	2	67	based	base	VERB
fcis-18462	2	68	on	on	ADP
fcis-18462	2	69	the	the	DET
fcis-18462	2	70	current	current	ADJ
fcis-18462	2	71	level	level	NOUN
fcis-18462	2	72	of	of	ADP
fcis-18462	2	73	social	social	ADJ
fcis-18462	2	74	development	development	NOUN
fcis-18462	2	75	,	,	PUNCT
fcis-18462	2	76	everyone	everyone	PRON
fcis-18462	2	77	's	's	PART
fcis-18462	2	78	demand	demand	NOUN
fcis-18462	2	79	for	for	ADP
fcis-18462	2	80	cars	car	NOUN
fcis-18462	2	81	has	have	AUX
fcis-18462	2	82	increased	increase	VERB
fcis-18462	2	83	rapidly	rapidly	ADV
fcis-18462	2	84	.	.	PUNCT
fcis-18462	3	1	at	at	ADP
fcis-18462	3	2	present	present	ADJ
fcis-18462	3	3	,	,	PUNCT
fcis-18462	3	4	the	the	DET
fcis-18462	3	5	total	total	ADJ
fcis-18462	3	6	number	number	NOUN
fcis-18462	3	7	of	of	ADP
fcis-18462	3	8	motor	motor	NOUN
fcis-18462	3	9	vehicles	vehicle	NOUN
fcis-18462	3	10	and	and	CCONJ
fcis-18462	3	11	drivers	driver	NOUN
fcis-18462	3	12	in	in	ADP
fcis-18462	3	13	china	china	PROPN
fcis-18462	3	14	ranks	rank	VERB
fcis-18462	3	15	first	first	ADV
fcis-18462	3	16	in	in	ADP
fcis-18462	3	17	the	the	DET
fcis-18462	3	18	world	world	NOUN
fcis-18462	3	19	.	.	PUNCT
fcis-18462	4	1	with	with	ADP
fcis-18462	4	2	the	the	DET
fcis-18462	4	3	rapid	rapid	ADJ
fcis-18462	4	4	development	development	NOUN
fcis-18462	4	5	of	of	ADP
fcis-18462	4	6	deep	deep	ADJ
fcis-18462	4	7	learning	learning	NOUN
fcis-18462	4	8	,	,	PUNCT
fcis-18462	4	9	the	the	DET
fcis-18462	4	10	method	method	NOUN
fcis-18462	4	11	of	of	ADP
fcis-18462	4	12	vehicle	vehicle	NOUN
fcis-18462	4	13	flow	flow	NOUN
fcis-18462	4	14	statistics	statistic	NOUN
fcis-18462	4	15	based	base	VERB
fcis-18462	4	16	on	on	ADP
fcis-18462	4	17	video	video	NOUN
fcis-18462	4	18	can	can	AUX
fcis-18462	4	19	directly	directly	ADV
fcis-18462	4	20	use	use	VERB
fcis-18462	4	21	the	the	DET
fcis-18462	4	22	existing	exist	VERB
fcis-18462	4	23	traffic	traffic	NOUN
fcis-18462	4	24	monitoring	monitor	VERB
fcis-18462	4	25	camera	camera	NOUN
fcis-18462	4	26	to	to	PART
fcis-18462	4	27	realize	realize	VERB
fcis-18462	4	28	the	the	DET
fcis-18462	4	29	detection	detection	NOUN
fcis-18462	4	30	of	of	ADP
fcis-18462	4	31	vehicles	vehicle	NOUN
fcis-18462	4	32	,	,	PUNCT
fcis-18462	4	33	and	and	CCONJ
fcis-18462	4	34	some	some	DET
fcis-18462	4	35	traffic	traffic	NOUN
fcis-18462	4	36	flow	flow	NOUN
fcis-18462	4	37	detection	detection	NOUN
fcis-18462	4	38	based	base	VERB
fcis-18462	4	39	on	on	ADP
fcis-18462	4	40	yolov1	yolov1	PROPN
fcis-18462	4	41	,	,	PUNCT
fcis-18462	4	42	yolov2	yolov2	PROPN
fcis-18462	4	43	,	,	PUNCT
fcis-18462	4	44	yolov3	yolov3	PROPN
fcis-18462	4	45	,	,	PUNCT
fcis-18462	4	46	yolov4	yolov4	PROPN
fcis-18462	4	47	and	and	CCONJ
fcis-18462	4	48	other	other	ADJ
fcis-18462	4	49	algorithms	algorithm	NOUN
fcis-18462	4	50	have	have	VERB
fcis-18462	4	51	problems	problem	NOUN
fcis-18462	4	52	such	such	ADJ
fcis-18462	4	53	as	as	ADP
fcis-18462	4	54	insufficient	insufficient	ADJ
fcis-18462	4	55	accuracy	accuracy	NOUN
fcis-18462	4	56	and	and	CCONJ
fcis-18462	4	57	low	low	ADJ
fcis-18462	4	58	efficiency	efficiency	NOUN
fcis-18462	4	59	.	.	PUNCT
fcis-18462	5	1	therefore	therefore	ADV
fcis-18462	5	2	,	,	PUNCT
fcis-18462	5	3	this	this	DET
fcis-18462	5	4	paper	paper	NOUN
fcis-18462	5	5	proposes	propose	VERB
fcis-18462	5	6	to	to	PART
fcis-18462	5	7	use	use	VERB
fcis-18462	5	8	yolov5	yolov5	NOUN
fcis-18462	5	9	to	to	PART
fcis-18462	5	10	replace	replace	VERB
fcis-18462	5	11	the	the	DET
fcis-18462	5	12	original	original	ADJ
fcis-18462	5	13	algorithm	algorithm	NOUN
fcis-18462	5	14	to	to	PART
fcis-18462	5	15	achieve	achieve	VERB
fcis-18462	5	16	object	object	NOUN
fcis-18462	5	17	detection	detection	NOUN
fcis-18462	5	18	,	,	PUNCT
fcis-18462	5	19	tracking	tracking	NOUN
fcis-18462	5	20	,	,	PUNCT
fcis-18462	5	21	and	and	CCONJ
fcis-18462	5	22	processing	processing	NOUN
fcis-18462	5	23	.	.	PUNCT
fcis-18462	6	1	i	i	PRON
fcis-18462	6	2	improve	improve	VERB
fcis-18462	6	3	the	the	DET
fcis-18462	6	4	efficiency	efficiency	NOUN
fcis-18462	6	5	of	of	ADP
fcis-18462	6	6	the	the	DET
fcis-18462	6	7	statistics	statistic	NOUN
fcis-18462	6	8	of	of	ADP
fcis-18462	6	9	the	the	DET
fcis-18462	6	10	traffic	traffic	NOUN
fcis-18462	6	11	flow	flow	NOUN
fcis-18462	6	12	.	.	PUNCT
fcis-18462	7	1	keywords	keyword	NOUN
fcis-18462	7	2	:	:	PUNCT
fcis-18462	7	3	video	video	NOUN
fcis-18462	7	4	-	-	PUNCT
fcis-18462	7	5	based	base	VERB
fcis-18462	7	6	traffic	traffic	NOUN
fcis-18462	7	7	statistics	statistic	NOUN
fcis-18462	7	8	;	;	PUNCT
fcis-18462	7	9	yolov5	yolov5	NOUN
fcis-18462	7	10	;	;	PUNCT
fcis-18462	7	11	object	object	NOUN
fcis-18462	7	12	detection	detection	NOUN
fcis-18462	7	13	;	;	PUNCT
fcis-18462	7	14	tracking	tracking	NOUN
fcis-18462	7	15	;	;	PUNCT
fcis-18462	7	16	processing	processing	NOUN
fcis-18462	7	17	.	.	PUNCT
fcis-18462	8	1	1	1	X
fcis-18462	8	2	.	.	X
fcis-18462	8	3	introduction	introduction	NOUN
fcis-18462	8	4	the	the	DET
fcis-18462	8	5	total	total	ADJ
fcis-18462	8	6	number	number	NOUN
fcis-18462	8	7	of	of	ADP
fcis-18462	8	8	motor	motor	NOUN
fcis-18462	8	9	vehicles	vehicle	NOUN
fcis-18462	8	10	and	and	CCONJ
fcis-18462	8	11	drivers	driver	NOUN
fcis-18462	8	12	in	in	ADP
fcis-18462	8	13	china	china	PROPN
fcis-18462	8	14	ranks	rank	VERB
fcis-18462	8	15	first	first	ADV
fcis-18462	8	16	in	in	ADP
fcis-18462	8	17	the	the	DET
fcis-18462	8	18	world	world	NOUN
fcis-18462	8	19	,	,	PUNCT
fcis-18462	8	20	and	and	CCONJ
fcis-18462	8	21	the	the	DET
fcis-18462	8	22	detection	detection	NOUN
fcis-18462	8	23	of	of	ADP
fcis-18462	8	24	traffic	traffic	NOUN
fcis-18462	8	25	flow	flow	NOUN
fcis-18462	8	26	is	be	AUX
fcis-18462	8	27	a	a	DET
fcis-18462	8	28	crucial	crucial	ADJ
fcis-18462	8	29	study	study	NOUN
fcis-18462	8	30	in	in	ADP
fcis-18462	8	31	urban	urban	ADJ
fcis-18462	8	32	traffic	traffic	NOUN
fcis-18462	8	33	management	management	NOUN
fcis-18462	8	34	and	and	CCONJ
fcis-18462	8	35	planning	planning	NOUN
fcis-18462	8	36	.	.	PUNCT
fcis-18462	9	1	with	with	ADP
fcis-18462	9	2	the	the	DET
fcis-18462	9	3	increase	increase	NOUN
fcis-18462	9	4	of	of	ADP
fcis-18462	9	5	urban	urban	ADJ
fcis-18462	9	6	population	population	NOUN
fcis-18462	9	7	and	and	CCONJ
fcis-18462	9	8	the	the	DET
fcis-18462	9	9	continuous	continuous	ADJ
fcis-18462	9	10	growth	growth	NOUN
fcis-18462	9	11	of	of	ADP
fcis-18462	9	12	car	car	NOUN
fcis-18462	9	13	ownership	ownership	NOUN
fcis-18462	9	14	,	,	PUNCT
fcis-18462	9	15	the	the	DET
fcis-18462	9	16	problem	problem	NOUN
fcis-18462	9	17	of	of	ADP
fcis-18462	9	18	traffic	traffic	NOUN
fcis-18462	9	19	congestion	congestion	NOUN
fcis-18462	9	20	is	be	AUX
fcis-18462	9	21	becoming	become	VERB
fcis-18462	9	22	increasingly	increasingly	ADV
fcis-18462	9	23	prominent	prominent	ADJ
fcis-18462	9	24	,	,	PUNCT
fcis-18462	9	25	which	which	PRON
fcis-18462	9	26	brings	bring	VERB
fcis-18462	9	27	great	great	ADJ
fcis-18462	9	28	challenges	challenge	NOUN
fcis-18462	9	29	to	to	ADP
fcis-18462	9	30	the	the	DET
fcis-18462	9	31	sustainable	sustainable	ADJ
fcis-18462	9	32	development	development	NOUN
fcis-18462	9	33	of	of	ADP
fcis-18462	9	34	cities	city	NOUN
fcis-18462	9	35	.	.	PUNCT
fcis-18462	10	1	understanding	understanding	NOUN
fcis-18462	10	2	and	and	CCONJ
fcis-18462	10	3	mastering	master	VERB
fcis-18462	10	4	traffic	traffic	NOUN
fcis-18462	10	5	flow	flow	NOUN
fcis-18462	10	6	can	can	AUX
fcis-18462	10	7	provide	provide	VERB
fcis-18462	10	8	important	important	ADJ
fcis-18462	10	9	data	datum	NOUN
fcis-18462	10	10	support	support	NOUN
fcis-18462	10	11	for	for	ADP
fcis-18462	10	12	traffic	traffic	NOUN
fcis-18462	10	13	management	management	NOUN
fcis-18462	10	14	departments	department	NOUN
fcis-18462	10	15	,	,	PUNCT
fcis-18462	10	16	helping	help	VERB
fcis-18462	10	17	to	to	PART
fcis-18462	10	18	optimize	optimize	VERB
fcis-18462	10	19	decision	decision	NOUN
fcis-18462	10	20	-	-	PUNCT
fcis-18462	10	21	making	making	NOUN
fcis-18462	10	22	in	in	ADP
fcis-18462	10	23	road	road	NOUN
fcis-18462	10	24	planning	planning	NOUN
fcis-18462	10	25	,	,	PUNCT
fcis-18462	10	26	traffic	traffic	NOUN
fcis-18462	10	27	light	light	NOUN
fcis-18462	10	28	control	control	NOUN
fcis-18462	10	29	,	,	PUNCT
fcis-18462	10	30	traffic	traffic	NOUN
fcis-18462	10	31	congestion	congestion	NOUN
fcis-18462	10	32	mitigation	mitigation	NOUN
fcis-18462	10	33	,	,	PUNCT
fcis-18462	10	34	and	and	CCONJ
fcis-18462	10	35	more	more	ADJ
fcis-18462	10	36	.	.	PUNCT
fcis-18462	11	1	traffic	traffic	NOUN
fcis-18462	11	2	flow	flow	NOUN
fcis-18462	11	3	detection	detection	NOUN
fcis-18462	11	4	plays	play	VERB
fcis-18462	11	5	an	an	DET
fcis-18462	11	6	important	important	ADJ
fcis-18462	11	7	role	role	NOUN
fcis-18462	11	8	in	in	ADP
fcis-18462	11	9	urban	urban	ADJ
fcis-18462	11	10	traffic	traffic	NOUN
fcis-18462	11	11	management	management	NOUN
fcis-18462	11	12	and	and	CCONJ
fcis-18462	11	13	planning	planning	NOUN
fcis-18462	11	14	,	,	PUNCT
fcis-18462	11	15	which	which	PRON
fcis-18462	11	16	can	can	AUX
fcis-18462	11	17	provide	provide	VERB
fcis-18462	11	18	accurate	accurate	ADJ
fcis-18462	11	19	data	datum	NOUN
fcis-18462	11	20	support	support	NOUN
fcis-18462	11	21	for	for	ADP
fcis-18462	11	22	traffic	traffic	NOUN
fcis-18462	11	23	management	management	NOUN
fcis-18462	11	24	departments	department	NOUN
fcis-18462	11	25	,	,	PUNCT
fcis-18462	11	26	help	help	VERB
fcis-18462	11	27	optimize	optimize	VERB
fcis-18462	11	28	traffic	traffic	NOUN
fcis-18462	11	29	conditions	condition	NOUN
fcis-18462	11	30	,	,	PUNCT
fcis-18462	11	31	improve	improve	VERB
fcis-18462	11	32	road	road	NOUN
fcis-18462	11	33	traffic	traffic	NOUN
fcis-18462	11	34	efficiency	efficiency	NOUN
fcis-18462	11	35	,	,	PUNCT
fcis-18462	11	36	and	and	CCONJ
fcis-18462	11	37	promote	promote	VERB
fcis-18462	11	38	the	the	DET
fcis-18462	11	39	sustainable	sustainable	ADJ
fcis-18462	11	40	development	development	NOUN
fcis-18462	11	41	of	of	ADP
fcis-18462	11	42	cities	city	NOUN
fcis-18462	11	43	.	.	PUNCT
fcis-18462	12	1	however	however	ADV
fcis-18462	12	2	,	,	PUNCT
fcis-18462	12	3	many	many	ADJ
fcis-18462	12	4	algorithms	algorithm	NOUN
fcis-18462	12	5	have	have	VERB
fcis-18462	12	6	serious	serious	ADJ
fcis-18462	12	7	problems	problem	NOUN
fcis-18462	12	8	such	such	ADJ
fcis-18462	12	9	as	as	ADP
fcis-18462	12	10	low	low	ADJ
fcis-18462	12	11	detection	detection	NOUN
fcis-18462	12	12	accuracy	accuracy	NOUN
fcis-18462	12	13	and	and	CCONJ
fcis-18462	12	14	low	low	ADJ
fcis-18462	12	15	detection	detection	NOUN
fcis-18462	12	16	efficiency	efficiency	NOUN
fcis-18462	12	17	.	.	PUNCT
fcis-18462	13	1	to	to	PART
fcis-18462	13	2	better	well	ADV
fcis-18462	13	3	improve	improve	VERB
fcis-18462	13	4	the	the	DET
fcis-18462	13	5	accuracy	accuracy	NOUN
fcis-18462	13	6	and	and	CCONJ
fcis-18462	13	7	efficiency	efficiency	NOUN
fcis-18462	13	8	,	,	PUNCT
fcis-18462	13	9	we	we	PRON
fcis-18462	13	10	use	use	VERB
fcis-18462	13	11	the	the	DET
fcis-18462	13	12	yolov5	yolov5	NOUN
fcis-18462	13	13	algorithm	algorithm	NOUN
fcis-18462	13	14	to	to	PART
fcis-18462	13	15	improve	improve	VERB
fcis-18462	13	16	it	it	PRON
fcis-18462	13	17	to	to	PART
fcis-18462	13	18	achieve	achieve	VERB
fcis-18462	13	19	real	real	ADJ
fcis-18462	13	20	-	-	PUNCT
fcis-18462	13	21	time	time	NOUN
fcis-18462	13	22	monitoring	monitoring	NOUN
fcis-18462	13	23	[	[	X
fcis-18462	13	24	1	1	NUM
fcis-18462	13	25	]	]	PUNCT
fcis-18462	13	26	.	.	PUNCT
fcis-18462	14	1	2	2	X
fcis-18462	14	2	.	.	X
fcis-18462	14	3	introduction	introduction	NOUN
fcis-18462	14	4	of	of	ADP
fcis-18462	14	5	yolov	yolov	PROPN
fcis-18462	14	6	(	(	PUNCT
fcis-18462	14	7	1,2,3,4	1,2,3,4	NUM
fcis-18462	14	8	)	)	PUNCT
fcis-18462	14	9	and	and	CCONJ
fcis-18462	14	10	their	their	PRON
fcis-18462	14	11	shortcomings	shortcoming	NOUN
fcis-18462	14	12	the	the	DET
fcis-18462	14	13	hinton	hinton	PROPN
fcis-18462	14	14	team	team	NOUN
fcis-18462	14	15	designed	design	VERB
fcis-18462	14	16	alexne	alexne	NOUN
fcis-18462	14	17	by	by	ADP
fcis-18462	14	18	using	use	VERB
fcis-18462	14	19	a	a	DET
fcis-18462	14	20	convolutional	convolutional	ADJ
fcis-18462	14	21	neural	neural	ADJ
fcis-18462	14	22	network	network	NOUN
fcis-18462	14	23	,	,	PUNCT
fcis-18462	14	24	which	which	PRON
fcis-18462	14	25	works	work	VERB
fcis-18462	14	26	better	well	ADV
fcis-18462	14	27	on	on	ADP
fcis-18462	14	28	the	the	DET
fcis-18462	14	29	imagenet	imagenet	NOUN
fcis-18462	14	30	dataset	dataset	VERB
fcis-18462	14	31	than	than	ADP
fcis-18462	14	32	other	other	ADJ
fcis-18462	14	33	traditional	traditional	ADJ
fcis-18462	14	34	methods	method	NOUN
fcis-18462	14	35	.	.	PUNCT
fcis-18462	15	1	so	so	ADV
fcis-18462	15	2	far	far	ADV
fcis-18462	15	3	,	,	PUNCT
fcis-18462	15	4	convolutional	convolutional	ADJ
fcis-18462	15	5	neural	neural	ADJ
fcis-18462	15	6	networks	network	NOUN
fcis-18462	15	7	have	have	AUX
fcis-18462	15	8	gradually	gradually	ADV
fcis-18462	15	9	become	become	VERB
fcis-18462	15	10	popular	popular	ADJ
fcis-18462	15	11	in	in	ADP
fcis-18462	15	12	the	the	DET
fcis-18462	15	13	field	field	NOUN
fcis-18462	15	14	of	of	ADP
fcis-18462	15	15	computer	computer	NOUN
fcis-18462	15	16	vision	vision	NOUN
fcis-18462	15	17	,	,	PUNCT
fcis-18462	15	18	and	and	CCONJ
fcis-18462	15	19	more	more	ADJ
fcis-18462	15	20	algorithms	algorithm	NOUN
fcis-18462	15	21	for	for	ADP
fcis-18462	15	22	image	image	NOUN
fcis-18462	15	23	detection	detection	NOUN
fcis-18462	15	24	and	and	CCONJ
fcis-18462	15	25	processing	processing	NOUN
fcis-18462	15	26	have	have	AUX
fcis-18462	15	27	been	be	AUX
fcis-18462	15	28	proposed	propose	VERB
fcis-18462	15	29	.	.	PUNCT
fcis-18462	16	1	yolov1	yolov1	NOUN
fcis-18462	16	2	is	be	AUX
fcis-18462	16	3	the	the	DET
fcis-18462	16	4	first	first	ADJ
fcis-18462	16	5	version	version	NOUN
fcis-18462	16	6	of	of	ADP
fcis-18462	16	7	yolo	yolo	PROPN
fcis-18462	16	8	,	,	PUNCT
fcis-18462	16	9	presented	present	VERB
fcis-18462	16	10	in	in	ADP
fcis-18462	16	11	2015	2015	NUM
fcis-18462	16	12	.	.	PUNCT
fcis-18462	17	1	it	it	PRON
fcis-18462	17	2	takes	take	VERB
fcis-18462	17	3	an	an	DET
fcis-18462	17	4	input	input	NOUN
fcis-18462	17	5	size	size	NOUN
fcis-18462	17	6	of	of	ADP
fcis-18462	17	7	448x448	448x448	NUM
fcis-18462	17	8	and	and	CCONJ
fcis-18462	17	9	divides	divide	VERB
fcis-18462	17	10	the	the	DET
fcis-18462	17	11	image	image	NOUN
fcis-18462	17	12	into	into	ADP
fcis-18462	17	13	7x7	7x7	NUM
fcis-18462	17	14	grids	grid	NOUN
fcis-18462	17	15	,	,	PUNCT
fcis-18462	17	16	each	each	PRON
fcis-18462	17	17	predicting	predict	VERB
fcis-18462	17	18	two	two	NUM
fcis-18462	17	19	bounding	bounding	NOUN
fcis-18462	17	20	boxes	box	NOUN
fcis-18462	17	21	,	,	PUNCT
fcis-18462	17	22	each	each	PRON
fcis-18462	17	23	containing	contain	VERB
fcis-18462	17	24	a	a	DET
fcis-18462	17	25	target	target	NOUN
fcis-18462	17	26	.	.	PUNCT
fcis-18462	18	1	each	each	DET
fcis-18462	18	2	bounding	bounding	NOUN
fcis-18462	18	3	box	box	NOUN
fcis-18462	18	4	also	also	ADV
fcis-18462	18	5	predicts	predict	VERB
fcis-18462	18	6	the	the	DET
fcis-18462	18	7	category	category	NOUN
fcis-18462	18	8	probability	probability	NOUN
fcis-18462	18	9	of	of	ADP
fcis-18462	18	10	the	the	DET
fcis-18462	18	11	target	target	NOUN
fcis-18462	18	12	.	.	PUNCT
fcis-18462	19	1	yolov1	yolov1	NOUN
fcis-18462	19	2	was	be	AUX
fcis-18462	19	3	released	release	VERB
fcis-18462	19	4	by	by	ADP
fcis-18462	19	5	joseph	joseph	PROPN
fcis-18462	19	6	et	et	PROPN
fcis-18462	19	7	al	al	PROPN
fcis-18462	19	8	.	.	PROPN
fcis-18462	19	9	,	,	PUNCT
fcis-18462	19	10	which	which	PRON
fcis-18462	19	11	not	not	PART
fcis-18462	19	12	only	only	ADV
fcis-18462	19	13	has	have	VERB
fcis-18462	19	14	a	a	DET
fcis-18462	19	15	high	high	ADJ
fcis-18462	19	16	degree	degree	NOUN
fcis-18462	19	17	of	of	ADP
fcis-18462	19	18	generalization	generalization	NOUN
fcis-18462	19	19	ability	ability	NOUN
fcis-18462	19	20	,	,	PUNCT
fcis-18462	19	21	but	but	CCONJ
fcis-18462	19	22	also	also	ADV
fcis-18462	19	23	can	can	AUX
fcis-18462	19	24	meet	meet	VERB
fcis-18462	19	25	the	the	DET
fcis-18462	19	26	requirements	requirement	NOUN
fcis-18462	19	27	of	of	ADP
fcis-18462	19	28	real	real	ADJ
fcis-18462	19	29	-	-	PUNCT
fcis-18462	19	30	time	time	NOUN
fcis-18462	19	31	detection	detection	NOUN
fcis-18462	19	32	.	.	PUNCT
fcis-18462	20	1	however	however	ADV
fcis-18462	20	2	,	,	PUNCT
fcis-18462	20	3	because	because	SCONJ
fcis-18462	20	4	yolov1	yolov1	NOUN
fcis-18462	20	5	does	do	AUX
fcis-18462	20	6	not	not	PART
fcis-18462	20	7	use	use	VERB
fcis-18462	20	8	a	a	DET
fcis-18462	20	9	full	full	ADJ
fcis-18462	20	10	convolutional	convolutional	ADJ
fcis-18462	20	11	network	network	NOUN
fcis-18462	20	12	,	,	PUNCT
fcis-18462	20	13	the	the	DET
fcis-18462	20	14	size	size	NOUN
fcis-18462	20	15	of	of	ADP
fcis-18462	20	16	the	the	DET
fcis-18462	20	17	input	input	NOUN
fcis-18462	20	18	is	be	AUX
fcis-18462	20	19	fixed	fix	VERB
fcis-18462	20	20	,	,	PUNCT
fcis-18462	20	21	and	and	CCONJ
fcis-18462	20	22	the	the	DET
fcis-18462	20	23	image	image	NOUN
fcis-18462	20	24	often	often	ADV
fcis-18462	20	25	needs	need	VERB
fcis-18462	20	26	to	to	PART
fcis-18462	20	27	be	be	AUX
fcis-18462	20	28	cropped	crop	VERB
fcis-18462	20	29	,	,	PUNCT
fcis-18462	20	30	which	which	PRON
fcis-18462	20	31	has	have	VERB
fcis-18462	20	32	a	a	DET
fcis-18462	20	33	worse	bad	ADJ
fcis-18462	20	34	prediction	prediction	NOUN
fcis-18462	20	35	effect	effect	NOUN
fcis-18462	20	36	on	on	ADP
fcis-18462	20	37	small	small	ADJ
fcis-18462	20	38	targets	target	NOUN
fcis-18462	20	39	in	in	ADP
fcis-18462	20	40	large	large	ADJ
fcis-18462	20	41	images	image	NOUN
fcis-18462	20	42	,	,	PUNCT
fcis-18462	20	43	and	and	CCONJ
fcis-18462	20	44	also	also	ADV
fcis-18462	20	45	has	have	VERB
fcis-18462	20	46	a	a	DET
fcis-18462	20	47	certain	certain	ADJ
fcis-18462	20	48	impact	impact	NOUN
fcis-18462	20	49	on	on	ADP
fcis-18462	20	50	comprehensive	comprehensive	ADJ
fcis-18462	20	51	object	object	NOUN
fcis-18462	20	52	recognition	recognition	NOUN
fcis-18462	20	53	.	.	PUNCT
fcis-18462	21	1	fcn	fcn	PROPN
fcis-18462	21	2	network	network	NOUN
fcis-18462	21	3	structure	structure	NOUN
fcis-18462	21	4	is	be	AUX
fcis-18462	21	5	divided	divide	VERB
fcis-18462	21	6	into	into	ADP
fcis-18462	21	7	two	two	NUM
fcis-18462	21	8	parts	part	NOUN
fcis-18462	21	9	:	:	PUNCT
fcis-18462	21	10	full	full	ADJ
fcis-18462	21	11	convolution	convolution	NOUN
fcis-18462	21	12	part	part	NOUN
fcis-18462	21	13	and	and	CCONJ
fcis-18462	21	14	deconvolution	deconvolution	NOUN
fcis-18462	21	15	part	part	NOUN
fcis-18462	21	16	.	.	PUNCT
fcis-18462	22	1	the	the	DET
fcis-18462	22	2	full	full	ADJ
fcis-18462	22	3	convolution	convolution	NOUN
fcis-18462	22	4	part	part	NOUN
fcis-18462	22	5	is	be	AUX
fcis-18462	22	6	some	some	DET
fcis-18462	22	7	classic	classic	ADJ
fcis-18462	22	8	cnn	cnn	NOUN
fcis-18462	22	9	networks	network	NOUN
fcis-18462	22	10	(	(	PUNCT
fcis-18462	22	11	such	such	ADJ
fcis-18462	22	12	as	as	ADP
fcis-18462	22	13	vgg	vgg	NOUN
fcis-18462	22	14	,	,	PUNCT
fcis-18462	22	15	resnet	resnet	NOUN
fcis-18462	22	16	,	,	PUNCT
fcis-18462	22	17	etc	etc	X
fcis-18462	22	18	.	.	X
fcis-18462	22	19	)	)	PUNCT
fcis-18462	22	20	,	,	PUNCT
fcis-18462	22	21	which	which	PRON
fcis-18462	22	22	is	be	AUX
fcis-18462	22	23	used	use	VERB
fcis-18462	22	24	to	to	PART
fcis-18462	22	25	extract	extract	VERB
fcis-18462	22	26	features	feature	NOUN
fcis-18462	22	27	;	;	PUNCT
fcis-18462	22	28	the	the	DET
fcis-18462	22	29	deconvolution	deconvolution	NOUN
fcis-18462	22	30	part	part	NOUN
fcis-18462	22	31	is	be	AUX
fcis-18462	22	32	to	to	PART
fcis-18462	22	33	get	get	VERB
fcis-18462	22	34	the	the	DET
fcis-18462	22	35	original	original	ADJ
fcis-18462	22	36	size	size	NOUN
fcis-18462	22	37	of	of	ADP
fcis-18462	22	38	the	the	DET
fcis-18462	22	39	semantic	semantic	ADJ
fcis-18462	22	40	segmentation	segmentation	NOUN
fcis-18462	22	41	image	image	NOUN
fcis-18462	22	42	through	through	ADP
fcis-18462	22	43	upsampling	upsample	VERB
fcis-18462	22	44	.	.	PUNCT
fcis-18462	23	1	the	the	DET
fcis-18462	23	2	input	input	NOUN
fcis-18462	23	3	of	of	ADP
fcis-18462	23	4	fcn	fcn	NOUN
fcis-18462	23	5	can	can	AUX
fcis-18462	23	6	be	be	AUX
fcis-18462	23	7	a	a	DET
fcis-18462	23	8	color	color	NOUN
fcis-18462	23	9	image	image	NOUN
fcis-18462	23	10	of	of	ADP
fcis-18462	23	11	any	any	DET
fcis-18462	23	12	size	size	NOUN
fcis-18462	23	13	,	,	PUNCT
fcis-18462	23	14	and	and	CCONJ
fcis-18462	23	15	the	the	DET
fcis-18462	23	16	output	output	NOUN
fcis-18462	23	17	size	size	NOUN
fcis-18462	23	18	is	be	AUX
fcis-18462	23	19	the	the	DET
fcis-18462	23	20	same	same	ADJ
fcis-18462	23	21	as	as	ADP
fcis-18462	23	22	the	the	DET
fcis-18462	23	23	input	input	NOUN
fcis-18462	23	24	size	size	NOUN
fcis-18462	23	25	,	,	PUNCT
fcis-18462	23	26	and	and	CCONJ
fcis-18462	23	27	the	the	DET
fcis-18462	23	28	number	number	NOUN
fcis-18462	23	29	of	of	ADP
fcis-18462	23	30	channels	channel	NOUN
fcis-18462	23	31	is	be	AUX
fcis-18462	23	32	n	n	PRON
fcis-18462	23	33	(	(	PUNCT
fcis-18462	23	34	number	number	NOUN
fcis-18462	23	35	of	of	ADP
fcis-18462	23	36	target	target	NOUN
fcis-18462	23	37	categories	category	NOUN
fcis-18462	23	38	)	)	PUNCT
fcis-18462	24	1	+	+	CCONJ
fcis-18462	24	2	1	1	NUM
fcis-18462	24	3	(	(	PUNCT
fcis-18462	24	4	background	background	NOUN
fcis-18462	24	5	)	)	PUNCT
fcis-18462	24	6	.	.	PUNCT
fcis-18462	25	1	the	the	DET
fcis-18462	25	2	fcn	fcn	PROPN
fcis-18462	25	3	network	network	NOUN
fcis-18462	25	4	structure	structure	NOUN
fcis-18462	25	5	is	be	AUX
fcis-18462	25	6	as	as	SCONJ
fcis-18462	25	7	follows	follow	VERB
fcis-18462	25	8	:	:	PUNCT
fcis-18462	25	9	in	in	ADP
fcis-18462	25	10	order	order	NOUN
fcis-18462	25	11	to	to	PART
fcis-18462	25	12	improve	improve	VERB
fcis-18462	25	13	the	the	DET
fcis-18462	25	14	problem	problem	NOUN
fcis-18462	25	15	of	of	ADP
fcis-18462	25	16	yolov1	yolov1	NOUN
fcis-18462	25	17	,	,	PUNCT
fcis-18462	25	18	yolov2	yolov2	PROPN
fcis-18462	25	19	was	be	AUX
fcis-18462	25	20	proposed	propose	VERB
fcis-18462	25	21	in	in	ADP
fcis-18462	25	22	2016	2016	NUM
fcis-18462	25	23	.	.	PUNCT
fcis-18462	26	1	yolov2	yolov2	PROPN
fcis-18462	26	2	uses	use	VERB
fcis-18462	26	3	the	the	DET
fcis-18462	26	4	darknet-19	darknet-19	PROPN
fcis-18462	26	5	network	network	NOUN
fcis-18462	26	6	as	as	SCONJ
fcis-18462	26	7	the	the	DET
fcis-18462	26	8	backbone	backbone	NOUN
fcis-18462	26	9	network	network	NOUN
fcis-18462	26	10	and	and	CCONJ
fcis-18462	26	11	introduces	introduce	VERB
fcis-18462	26	12	anchor	anchor	NOUN
fcis-18462	26	13	boxes	box	NOUN
fcis-18462	26	14	to	to	PART
fcis-18462	26	15	better	well	ADV
fcis-18462	26	16	handle	handle	VERB
fcis-18462	26	17	targets	target	NOUN
fcis-18462	26	18	at	at	ADP
fcis-18462	26	19	different	different	ADJ
fcis-18462	26	20	scales	scale	NOUN
fcis-18462	26	21	.	.	PUNCT
fcis-18462	27	1	in	in	ADP
fcis-18462	27	2	addition	addition	NOUN
fcis-18462	27	3	,	,	PUNCT
fcis-18462	27	4	yolov2	yolov2	PROPN
fcis-18462	27	5	uses	use	VERB
fcis-18462	27	6	multi	multi	ADJ
fcis-18462	27	7	-	-	ADJ
fcis-18462	27	8	scale	scale	ADJ
fcis-18462	27	9	training	training	NOUN
fcis-18462	27	10	and	and	CCONJ
fcis-18462	27	11	testing	testing	NOUN
fcis-18462	27	12	to	to	PART
fcis-18462	27	13	improve	improve	VERB
fcis-18462	27	14	the	the	DET
fcis-18462	27	15	detection	detection	NOUN
fcis-18462	27	16	ability	ability	NOUN
fcis-18462	27	17	of	of	ADP
fcis-18462	27	18	small	small	ADJ
fcis-18462	27	19	targets	target	NOUN
fcis-18462	27	20	.	.	PUNCT
fcis-18462	28	1	compared	compare	VERB
fcis-18462	28	2	with	with	ADP
fcis-18462	28	3	yolov1	yolov1	NOUN
fcis-18462	28	4	,	,	PUNCT
fcis-18462	28	5	yolov2	yolov2	PROPN
fcis-18462	28	6	has	have	AUX
fcis-18462	28	7	improved	improve	VERB
fcis-18462	28	8	in	in	ADP
fcis-18462	28	9	terms	term	NOUN
fcis-18462	28	10	of	of	ADP
fcis-18462	28	11	accuracy	accuracy	NOUN
fcis-18462	28	12	and	and	CCONJ
fcis-18462	28	13	speed	speed	NOUN
fcis-18462	28	14	.	.	PUNCT
fcis-18462	29	1	however	however	ADV
fcis-18462	29	2	,	,	PUNCT
fcis-18462	29	3	due	due	ADP
fcis-18462	29	4	to	to	ADP
fcis-18462	29	5	the	the	DET
fcis-18462	29	6	small	small	ADJ
fcis-18462	29	7	number	number	NOUN
fcis-18462	29	8	of	of	ADP
fcis-18462	29	9	grids	grid	NOUN
fcis-18462	29	10	,	,	PUNCT
fcis-18462	29	11	the	the	DET
fcis-18462	29	12	positioning	positioning	NOUN
fcis-18462	29	13	accuracy	accuracy	NOUN
fcis-18462	29	14	for	for	ADP
fcis-18462	29	15	large	large	ADJ
fcis-18462	29	16	-	-	PUNCT
fcis-18462	29	17	scale	scale	NOUN
fcis-18462	29	18	targets	target	NOUN
fcis-18462	29	19	is	be	AUX
fcis-18462	29	20	not	not	PART
fcis-18462	29	21	high	high	ADJ
fcis-18462	29	22	.	.	PUNCT
fcis-18462	30	1	yolov3	yolov3	PROPN
fcis-18462	30	2	was	be	AUX
fcis-18462	30	3	introduced	introduce	VERB
fcis-18462	30	4	in	in	ADP
fcis-18462	30	5	2018	2018	NUM
fcis-18462	30	6	and	and	CCONJ
fcis-18462	30	7	is	be	AUX
fcis-18462	30	8	the	the	DET
fcis-18462	30	9	largest	large	ADJ
fcis-18462	30	10	version	version	NOUN
fcis-18462	30	11	of	of	ADP
fcis-18462	30	12	the	the	DET
fcis-18462	30	13	yolo	yolo	ADJ
fcis-18462	30	14	series	series	NOUN
fcis-18462	30	15	.	.	PUNCT
fcis-18462	31	1	yolov3	yolov3	PROPN
fcis-18462	31	2	uses	use	VERB
fcis-18462	31	3	a	a	DET
fcis-18462	31	4	deeper	deep	ADJ
fcis-18462	31	5	darknet-53	darknet-53	ADJ
fcis-18462	31	6	backbone	backbone	NOUN
fcis-18462	31	7	network	network	NOUN
fcis-18462	31	8	and	and	CCONJ
fcis-18462	31	9	adds	add	VERB
fcis-18462	31	10	detection	detection	NOUN
fcis-18462	31	11	at	at	ADP
fcis-18462	31	12	multiple	multiple	ADJ
fcis-18462	31	13	scales	scale	NOUN
fcis-18462	31	14	.	.	PUNCT
fcis-18462	32	1	yolov3	yolov3	PROPN
fcis-18462	32	2	also	also	ADV
fcis-18462	32	3	introduces	introduce	VERB
fcis-18462	32	4	three	three	NUM
fcis-18462	32	5	target	target	NOUN
fcis-18462	32	6	prediction	prediction	NOUN
fcis-18462	32	7	layers	layer	NOUN
fcis-18462	32	8	of	of	ADP
fcis-18462	32	9	different	different	ADJ
fcis-18462	32	10	sizes	size	NOUN
fcis-18462	32	11	for	for	ADP
fcis-18462	32	12	detecting	detect	VERB
fcis-18462	32	13	targets	target	NOUN
fcis-18462	32	14	of	of	ADP
fcis-18462	32	15	different	different	ADJ
fcis-18462	32	16	sizes	size	NOUN
fcis-18462	32	17	.	.	PUNCT
fcis-18462	33	1	in	in	ADP
fcis-18462	33	2	addition	addition	NOUN
fcis-18462	33	3	,	,	PUNCT
fcis-18462	33	4	yolov3	yolov3	PROPN
fcis-18462	33	5	also	also	ADV
fcis-18462	33	6	uses	use	VERB
fcis-18462	33	7	fpn	fpn	NOUN
fcis-18462	33	8	(	(	PUNCT
fcis-18462	33	9	feature	feature	NOUN
fcis-18462	33	10	pyramid	pyramid	NOUN
fcis-18462	33	11	network	network	NOUN
fcis-18462	33	12	)	)	PUNCT
fcis-18462	33	13	to	to	PART
fcis-18462	33	14	extract	extract	VERB
fcis-18462	33	15	multi	multi	ADJ
fcis-18462	33	16	-	-	ADJ
fcis-18462	33	17	scale	scale	ADJ
fcis-18462	33	18	features	feature	NOUN
fcis-18462	33	19	.	.	PUNCT
fcis-18462	34	1	fpn	fpn	VERB
fcis-18462	34	2	is	be	AUX
fcis-18462	34	3	a	a	DET
fcis-18462	34	4	top	top	ADJ
fcis-18462	34	5	-	-	PUNCT
fcis-18462	34	6	down	down	ADP
fcis-18462	34	7	feature	feature	NOUN
fcis-18462	34	8	fusion	fusion	NOUN
fcis-18462	34	9	method	method	NOUN
fcis-18462	34	10	,	,	PUNCT
fcis-18462	34	11	but	but	CCONJ
fcis-18462	34	12	fpn	fpn	NOUN
fcis-18462	34	13	is	be	AUX
fcis-18462	34	14	a	a	DET
fcis-18462	34	15	multi	multi	ADJ
fcis-18462	34	16	-	-	ADJ
fcis-18462	34	17	scale	scale	ADJ
fcis-18462	34	18	object	object	NOUN
fcis-18462	34	19	detection	detection	NOUN
fcis-18462	34	20	algorithm	algorithm	NOUN
fcis-18462	34	21	,	,	PUNCT
fcis-18462	34	22	that	that	ADV
fcis-18462	34	23	is	is	ADV
fcis-18462	34	24	,	,	PUNCT
fcis-18462	34	25	there	there	PRON
fcis-18462	34	26	is	be	VERB
fcis-18462	34	27	not	not	PART
fcis-18462	34	28	only	only	ADV
fcis-18462	34	29	one	one	NUM
fcis-18462	34	30	feature	feature	NOUN
fcis-18462	34	31	prediction	prediction	NOUN
fcis-18462	34	32	layer	layer	NOUN
fcis-18462	34	33	.	.	PUNCT
fcis-18462	35	1	although	although	SCONJ
fcis-18462	35	2	some	some	DET
fcis-18462	35	3	algorithms	algorithm	NOUN
fcis-18462	35	4	also	also	ADV
fcis-18462	35	5	adopt	adopt	VERB
fcis-18462	35	6	multi	multi	ADJ
fcis-18462	35	7	-	-	ADJ
fcis-18462	35	8	scale	scale	ADJ
fcis-18462	35	9	feature	feature	NOUN
fcis-18462	35	10	fusion	fusion	NOUN
fcis-18462	35	11	to	to	PART
fcis-18462	35	12	carry	carry	VERB
fcis-18462	35	13	out	out	ADP
fcis-18462	35	14	target	target	NOUN
fcis-18462	35	15	detection	detection	NOUN
fcis-18462	35	16	,	,	PUNCT
fcis-18462	35	17	they	they	PRON
fcis-18462	35	18	often	often	ADV
fcis-18462	35	19	only	only	ADV
fcis-18462	35	20	use	use	VERB
fcis-18462	35	21	the	the	DET
fcis-18462	35	22	features	feature	NOUN
fcis-18462	35	23	of	of	ADP
fcis-18462	35	24	one	one	NUM
fcis-18462	35	25	scale	scale	NOUN
fcis-18462	35	26	obtained	obtain	VERB
fcis-18462	35	27	after	after	ADP
fcis-18462	35	28	fusion	fusion	NOUN
fcis-18462	35	29	.	.	PUNCT
fcis-18462	36	1	although	although	SCONJ
fcis-18462	36	2	this	this	DET
fcis-18462	36	3	approach	approach	NOUN
fcis-18462	36	4	can	can	AUX
fcis-18462	36	5	combine	combine	VERB
fcis-18462	36	6	the	the	DET
fcis-18462	36	7	semantic	semantic	ADJ
fcis-18462	36	8	information	information	NOUN
fcis-18462	36	9	of	of	ADP
fcis-18462	36	10	top	top	ADJ
fcis-18462	36	11	-	-	PUNCT
fcis-18462	36	12	level	level	NOUN
fcis-18462	36	13	features	feature	NOUN
fcis-18462	36	14	with	with	ADP
fcis-18462	36	15	the	the	DET
fcis-18462	36	16	details	detail	NOUN
fcis-18462	36	17	of	of	ADP
fcis-18462	36	18	low	low	ADJ
fcis-18462	36	19	-	-	PUNCT
fcis-18462	36	20	level	level	NOUN
fcis-18462	36	21	features	feature	NOUN
fcis-18462	36	22	,	,	PUNCT
fcis-18462	36	23	it	it	PRON
fcis-18462	36	24	will	will	AUX
fcis-18462	36	25	cause	cause	VERB
fcis-18462	36	26	some	some	DET
fcis-18462	36	27	deviations	deviation	NOUN
fcis-18462	36	28	in	in	ADP
fcis-18462	36	29	the	the	DET
fcis-18462	36	30	process	process	NOUN
fcis-18462	36	31	of	of	ADP
fcis-18462	36	32	feature	feature	NOUN
fcis-18462	36	33	deconvolution	deconvolution	NOUN
fcis-18462	36	34	.	.	PUNCT
fcis-18462	37	1	the	the	DET
fcis-18462	37	2	prediction	prediction	NOUN
fcis-18462	37	3	using	use	VERB
fcis-18462	37	4	only	only	ADV
fcis-18462	37	5	the	the	DET
fcis-18462	37	6	features	feature	NOUN
fcis-18462	37	7	obtained	obtain	VERB
fcis-18462	37	8	after	after	ADP
fcis-18462	37	9	fusion	fusion	NOUN
fcis-18462	37	10	will	will	AUX
fcis-18462	37	11	have	have	VERB
fcis-18462	37	12	a	a	DET
fcis-18462	37	13	bad	bad	ADJ
fcis-18462	37	14	effect	effect	NOUN
fcis-18462	37	15	on	on	ADP
fcis-18462	37	16	the	the	DET
fcis-18462	37	17	detection	detection	NOUN
fcis-18462	37	18	accuracy	accuracy	NOUN
fcis-18462	37	19	.	.	PUNCT
fcis-18462	38	1	based	base	VERB
fcis-18462	38	2	on	on	ADP
fcis-18462	38	3	the	the	DET
fcis-18462	38	4	above	above	ADJ
fcis-18462	38	5	problems	problem	NOUN
fcis-18462	38	6	,	,	PUNCT
fcis-18462	38	7	fpn	fpn	ADJ
fcis-18462	38	8	method	method	NOUN
fcis-18462	38	9	can	can	AUX
fcis-18462	38	10	make	make	VERB
fcis-18462	38	11	prediction	prediction	NOUN
fcis-18462	38	12	on	on	ADP
fcis-18462	38	13	multiple	multiple	ADJ
fcis-18462	38	14	fusion	fusion	NOUN
fcis-18462	38	15	features	feature	NOUN
fcis-18462	38	16	of	of	ADP
fcis-18462	38	17	different	different	ADJ
fcis-18462	38	18	scales	scale	NOUN
fcis-18462	38	19	to	to	PART
fcis-18462	38	20	maximize	maximize	VERB
fcis-18462	38	21	the	the	DET
fcis-18462	38	22	detection	detection	NOUN
fcis-18462	38	23	accuracy	accuracy	NOUN
fcis-18462	38	24	.	.	PUNCT
fcis-18462	39	1	76	76	NUM
fcis-18462	39	2	figure	figure	NOUN
fcis-18462	39	3	1	1	NUM
fcis-18462	39	4	.	.	PUNCT
fcis-18462	40	1	the	the	DET
fcis-18462	40	2	fcn	fcn	PROPN
fcis-18462	40	3	network	network	NOUN
fcis-18462	40	4	structure	structure	PROPN
fcis-18462	40	5	yolov4	yolov4	PROPN
fcis-18462	40	6	was	be	AUX
fcis-18462	40	7	proposed	propose	VERB
fcis-18462	40	8	in	in	ADP
fcis-18462	40	9	2020	2020	NUM
fcis-18462	40	10	.	.	PUNCT
fcis-18462	41	1	yolov4	yolov4	PROPN
fcis-18462	41	2	has	have	AUX
fcis-18462	41	3	made	make	VERB
fcis-18462	41	4	a	a	DET
fcis-18462	41	5	series	series	NOUN
fcis-18462	41	6	of	of	ADP
fcis-18462	41	7	improvements	improvement	NOUN
fcis-18462	41	8	in	in	ADP
fcis-18462	41	9	network	network	NOUN
fcis-18462	41	10	structure	structure	NOUN
fcis-18462	41	11	and	and	CCONJ
fcis-18462	41	12	training	training	NOUN
fcis-18462	41	13	strategy	strategy	NOUN
fcis-18462	41	14	,	,	PUNCT
fcis-18462	41	15	including	include	VERB
fcis-18462	41	16	using	use	VERB
fcis-18462	41	17	cspdarknet53	cspdarknet53	NOUN
fcis-18462	41	18	as	as	ADP
fcis-18462	41	19	the	the	DET
fcis-18462	41	20	backbone	backbone	NOUN
fcis-18462	41	21	network	network	NOUN
fcis-18462	41	22	,	,	PUNCT
fcis-18462	41	23	using	use	VERB
fcis-18462	41	24	the	the	DET
fcis-18462	41	25	spp	spp	NOUN
fcis-18462	41	26	(	(	PUNCT
fcis-18462	41	27	spatial	spatial	ADJ
fcis-18462	41	28	pyramid	pyramid	NOUN
fcis-18462	41	29	pooling	pooling	NOUN
fcis-18462	41	30	)	)	PUNCT
fcis-18462	41	31	module	module	NOUN
fcis-18462	41	32	to	to	PART
fcis-18462	41	33	process	process	VERB
fcis-18462	41	34	multi	multi	ADJ
fcis-18462	41	35	-	-	ADJ
fcis-18462	41	36	scale	scale	ADJ
fcis-18462	41	37	information	information	NOUN
fcis-18462	41	38	,	,	PUNCT
fcis-18462	41	39	and	and	CCONJ
fcis-18462	41	40	using	use	VERB
fcis-18462	41	41	panet	panet	NOUN
fcis-18462	41	42	to	to	PART
fcis-18462	41	43	fuse	fuse	VERB
fcis-18462	41	44	different	different	ADJ
fcis-18462	41	45	hierarchical	hierarchical	ADJ
fcis-18462	41	46	features	feature	NOUN
fcis-18462	41	47	.	.	PUNCT
fcis-18462	42	1	the	the	DET
fcis-18462	42	2	network	network	NOUN
fcis-18462	42	3	structure	structure	NOUN
fcis-18462	42	4	of	of	ADP
fcis-18462	42	5	panet	panet	NOUN
fcis-18462	42	6	is	be	AUX
fcis-18462	42	7	shown	show	VERB
fcis-18462	42	8	in	in	ADP
fcis-18462	42	9	figure	figure	NOUN
fcis-18462	42	10	2	2	NUM
fcis-18462	42	11	,	,	PUNCT
fcis-18462	42	12	which	which	PRON
fcis-18462	42	13	consists	consist	VERB
fcis-18462	42	14	of	of	ADP
fcis-18462	42	15	five	five	NUM
fcis-18462	42	16	core	core	NOUN
fcis-18462	42	17	modules	module	NOUN
fcis-18462	42	18	.	.	PUNCT
fcis-18462	43	1	among	among	ADP
fcis-18462	43	2	them	they	PRON
fcis-18462	43	3	,	,	PUNCT
fcis-18462	43	4	(	(	PUNCT
fcis-18462	43	5	a	a	X
fcis-18462	43	6	)	)	PUNCT
fcis-18462	43	7	is	be	AUX
fcis-18462	43	8	an	an	DET
fcis-18462	43	9	fpn[2	fpn[2	PROPN
fcis-18462	43	10	]	]	PUNCT
fcis-18462	43	11	,	,	PUNCT
fcis-18462	43	12	(	(	PUNCT
fcis-18462	43	13	b	b	X
fcis-18462	43	14	)	)	PUNCT
fcis-18462	43	15	is	be	AUX
fcis-18462	43	16	a	a	DET
fcis-18462	43	17	bottom	bottom	ADJ
fcis-18462	43	18	-	-	PUNCT
fcis-18462	43	19	up	up	ADP
fcis-18462	43	20	feature	feature	NOUN
fcis-18462	43	21	fusion	fusion	NOUN
fcis-18462	43	22	layer	layer	NOUN
fcis-18462	43	23	added	add	VERB
fcis-18462	43	24	by	by	ADP
fcis-18462	43	25	pan	pan	PROPN
fcis-18462	43	26	,	,	PUNCT
fcis-18462	43	27	(	(	PUNCT
fcis-18462	43	28	c	c	X
fcis-18462	43	29	)	)	PUNCT
fcis-18462	43	30	is	be	AUX
fcis-18462	43	31	an	an	DET
fcis-18462	43	32	adaptive	adaptive	ADJ
fcis-18462	43	33	feature	feature	NOUN
fcis-18462	43	34	pooling	pool	VERB
fcis-18462	43	35	layer	layer	NOUN
fcis-18462	43	36	,	,	PUNCT
fcis-18462	43	37	(	(	PUNCT
fcis-18462	43	38	d	d	X
fcis-18462	43	39	)	)	PUNCT
fcis-18462	43	40	is	be	AUX
fcis-18462	43	41	a	a	DET
fcis-18462	43	42	bounding	bounding	NOUN
fcis-18462	43	43	box	box	NOUN
fcis-18462	43	44	prediction	prediction	NOUN
fcis-18462	43	45	head	head	NOUN
fcis-18462	43	46	of	of	ADP
fcis-18462	43	47	panet	panet	NOUN
fcis-18462	43	48	,	,	PUNCT
fcis-18462	43	49	and	and	CCONJ
fcis-18462	43	50	(	(	PUNCT
fcis-18462	43	51	e	e	NOUN
fcis-18462	43	52	)	)	PUNCT
fcis-18462	43	53	is	be	AUX
fcis-18462	43	54	a	a	DET
fcis-18462	43	55	fully	fully	ADV
fcis-18462	43	56	connected	connect	VERB
fcis-18462	43	57	fusion	fusion	NOUN
fcis-18462	43	58	layer	layer	NOUN
fcis-18462	43	59	for	for	ADP
fcis-18462	43	60	predicting	predict	VERB
fcis-18462	43	61	masks.yolov4	masks.yolov4	NOUN
fcis-18462	43	62	has	have	AUX
fcis-18462	43	63	been	be	AUX
fcis-18462	43	64	significantly	significantly	ADV
fcis-18462	43	65	improved	improve	VERB
fcis-18462	43	66	in	in	ADP
fcis-18462	43	67	terms	term	NOUN
fcis-18462	43	68	of	of	ADP
fcis-18462	43	69	accuracy	accuracy	NOUN
fcis-18462	43	70	and	and	CCONJ
fcis-18462	43	71	speed	speed	NOUN
fcis-18462	43	72	.	.	PUNCT
fcis-18462	44	1	however	however	ADV
fcis-18462	44	2	,	,	PUNCT
fcis-18462	44	3	there	there	PRON
fcis-18462	44	4	is	be	VERB
fcis-18462	44	5	still	still	ADV
fcis-18462	44	6	the	the	DET
fcis-18462	44	7	problem	problem	NOUN
fcis-18462	44	8	of	of	ADP
fcis-18462	44	9	poor	poor	ADJ
fcis-18462	44	10	handling	handling	NOUN
fcis-18462	44	11	of	of	ADP
fcis-18462	44	12	overlapping	overlap	VERB
fcis-18462	44	13	targets	target	NOUN
fcis-18462	44	14	,	,	PUNCT
fcis-18462	44	15	identifying	identify	VERB
fcis-18462	44	16	them	they	PRON
fcis-18462	44	17	as	as	ADP
fcis-18462	44	18	one	one	NUM
fcis-18462	44	19	target	target	NOUN
fcis-18462	44	20	or	or	CCONJ
fcis-18462	44	21	missing	miss	VERB
fcis-18462	44	22	some	some	PRON
fcis-18462	44	23	of	of	ADP
fcis-18462	44	24	them	they	PRON
fcis-18462	44	25	.	.	PUNCT
fcis-18462	45	1	figure	figure	VERB
fcis-18462	45	2	2	2	NUM
fcis-18462	45	3	.	.	PUNCT
fcis-18462	45	4	network	network	NOUN
fcis-18462	45	5	structure	structure	NOUN
fcis-18462	45	6	of	of	ADP
fcis-18462	45	7	panet	panet	PROPN
fcis-18462	45	8	3	3	NUM
fcis-18462	45	9	.	.	PUNCT
fcis-18462	45	10	introduction	introduction	NOUN
fcis-18462	45	11	of	of	ADP
fcis-18462	45	12	yolov5	yolov5	NOUN
fcis-18462	45	13	in	in	ADP
fcis-18462	45	14	order	order	NOUN
fcis-18462	45	15	to	to	PART
fcis-18462	45	16	make	make	VERB
fcis-18462	45	17	the	the	DET
fcis-18462	45	18	tracking	tracking	NOUN
fcis-18462	45	19	,	,	PUNCT
fcis-18462	45	20	detection	detection	NOUN
fcis-18462	45	21	,	,	PUNCT
fcis-18462	45	22	and	and	CCONJ
fcis-18462	45	23	processing	processing	NOUN
fcis-18462	45	24	of	of	ADP
fcis-18462	45	25	images	image	NOUN
fcis-18462	45	26	better	well	ADV
fcis-18462	45	27	realized	realize	VERB
fcis-18462	45	28	,	,	PUNCT
fcis-18462	45	29	and	and	CCONJ
fcis-18462	45	30	the	the	DET
fcis-18462	45	31	efficiency	efficiency	NOUN
fcis-18462	45	32	of	of	ADP
fcis-18462	45	33	traffic	traffic	NOUN
fcis-18462	45	34	flow	flow	NOUN
fcis-18462	45	35	statistics	statistic	NOUN
fcis-18462	45	36	is	be	AUX
fcis-18462	45	37	greatly	greatly	ADV
fcis-18462	45	38	improved	improve	VERB
fcis-18462	45	39	,	,	PUNCT
fcis-18462	45	40	we	we	PRON
fcis-18462	45	41	use	use	VERB
fcis-18462	45	42	the	the	DET
fcis-18462	45	43	yolov5	yolov5	NOUN
fcis-18462	45	44	algorithm	algorithm	NOUN
fcis-18462	45	45	to	to	PART
fcis-18462	45	46	implement	implement	VERB
fcis-18462	45	47	it	it	PRON
fcis-18462	45	48	,	,	PUNCT
fcis-18462	45	49	which	which	PRON
fcis-18462	45	50	is	be	AUX
fcis-18462	45	51	compared	compare	VERB
fcis-18462	45	52	with	with	ADP
fcis-18462	45	53	other	other	ADJ
fcis-18462	45	54	yolo	yolo	ADJ
fcis-18462	45	55	algorithms	algorithm	NOUN
fcis-18462	45	56	,	,	PUNCT
fcis-18462	45	57	yolov5	yolov5	NOUN
fcis-18462	45	58	uses	use	VERB
fcis-18462	45	59	cspnet	cspnet	NOUN
fcis-18462	45	60	as	as	ADP
fcis-18462	45	61	the	the	DET
fcis-18462	45	62	backbone	backbone	NOUN
fcis-18462	45	63	network	network	NOUN
fcis-18462	45	64	.	.	PUNCT
fcis-18462	46	1	cspnet	cspnet	NOUN
fcis-18462	46	2	is	be	AUX
fcis-18462	46	3	a	a	DET
fcis-18462	46	4	new	new	ADJ
fcis-18462	46	5	network	network	NOUN
fcis-18462	46	6	structure	structure	NOUN
fcis-18462	46	7	,	,	PUNCT
fcis-18462	46	8	which	which	PRON
fcis-18462	46	9	can	can	AUX
fcis-18462	46	10	effectively	effectively	ADV
fcis-18462	46	11	reduce	reduce	VERB
fcis-18462	46	12	the	the	DET
fcis-18462	46	13	redundancy	redundancy	NOUN
fcis-18462	46	14	of	of	ADP
fcis-18462	46	15	feature	feature	NOUN
fcis-18462	46	16	maps	map	NOUN
fcis-18462	46	17	and	and	CCONJ
fcis-18462	46	18	improve	improve	VERB
fcis-18462	46	19	the	the	DET
fcis-18462	46	20	computational	computational	ADJ
fcis-18462	46	21	efficiency	efficiency	NOUN
fcis-18462	46	22	and	and	CCONJ
fcis-18462	46	23	accuracy	accuracy	NOUN
fcis-18462	46	24	,	,	PUNCT
fcis-18462	46	25	using	use	VERB
fcis-18462	46	26	feature	feature	NOUN
fcis-18462	46	27	maps	map	NOUN
fcis-18462	46	28	of	of	ADP
fcis-18462	46	29	four	four	NUM
fcis-18462	46	30	scales	scale	NOUN
fcis-18462	46	31	,	,	PUNCT
fcis-18462	46	32	which	which	PRON
fcis-18462	46	33	are	be	AUX
fcis-18462	46	34	40x40	40x40	NUM
fcis-18462	46	35	,	,	PUNCT
fcis-18462	46	36	80x80	80x80	NUM
fcis-18462	46	37	,	,	PUNCT
fcis-18462	46	38	160x160	160x160	NUM
fcis-18462	46	39	,	,	PUNCT
fcis-18462	46	40	and	and	CCONJ
fcis-18462	46	41	320x320	320x320	NUM
fcis-18462	46	42	.	.	PUNCT
fcis-18462	47	1	(	(	PUNCT
fcis-18462	47	2	cspnet	cspnet	NOUN
fcis-18462	47	3	structure	structure	NOUN
fcis-18462	47	4	by	by	ADP
fcis-18462	47	5	splitting	split	VERB
fcis-18462	47	6	the	the	DET
fcis-18462	47	7	shallow	shallow	ADJ
fcis-18462	47	8	feature	feature	NOUN
fcis-18462	47	9	map	map	NOUN
fcis-18462	47	10	in	in	ADP
fcis-18462	47	11	two	two	NUM
fcis-18462	47	12	at	at	ADP
fcis-18462	47	13	the	the	DET
fcis-18462	47	14	channel	channel	NOUN
fcis-18462	47	15	dimension	dimension	NOUN
fcis-18462	47	16	,	,	PUNCT
fcis-18462	47	17	a	a	DET
fcis-18462	47	18	portion	portion	NOUN
fcis-18462	47	19	propagates	propagate	VERB
fcis-18462	47	20	backwards	backwards	ADV
fcis-18462	47	21	via	via	ADP
fcis-18462	47	22	a	a	DET
fcis-18462	47	23	feature	feature	NOUN
fcis-18462	47	24	extraction	extraction	NOUN
fcis-18462	47	25	module	module	NOUN
fcis-18462	47	26	such	such	ADJ
fcis-18462	47	27	as	as	ADP
fcis-18462	47	28	a	a	DET
fcis-18462	47	29	residual	residual	ADJ
fcis-18462	47	30	block	block	NOUN
fcis-18462	47	31	,	,	PUNCT
fcis-18462	47	32	the	the	DET
fcis-18462	47	33	other	other	ADJ
fcis-18462	47	34	part	part	NOUN
fcis-18462	47	35	is	be	AUX
fcis-18462	47	36	directly	directly	ADV
fcis-18462	47	37	merged	merge	VERB
fcis-18462	47	38	with	with	ADP
fcis-18462	47	39	the	the	DET
fcis-18462	47	40	output	output	NOUN
fcis-18462	47	41	of	of	ADP
fcis-18462	47	42	the	the	DET
fcis-18462	47	43	feature	feature	NOUN
fcis-18462	47	44	extraction	extraction	NOUN
fcis-18462	47	45	module	module	NOUN
fcis-18462	47	46	through	through	ADP
fcis-18462	47	47	the	the	DET
fcis-18462	47	48	cross	cross	ADJ
fcis-18462	47	49	-	-	ADJ
fcis-18462	47	50	stage	stage	ADJ
fcis-18462	47	51	hierarchy	hierarchy	NOUN
fcis-18462	47	52	.	.	PUNCT
fcis-18462	48	1	a	a	DET
fcis-18462	48	2	richer	rich	ADJ
fcis-18462	48	3	gradient	gradient	ADJ
fcis-18462	48	4	combination	combination	NOUN
fcis-18462	48	5	is	be	AUX
fcis-18462	48	6	achieved	achieve	VERB
fcis-18462	48	7	,	,	PUNCT
fcis-18462	48	8	and	and	CCONJ
fcis-18462	48	9	can	can	AUX
fcis-18462	48	10	be	be	AUX
fcis-18462	48	11	reduced	reduce	VERB
fcis-18462	48	12	on	on	ADP
fcis-18462	48	13	the	the	DET
fcis-18462	48	14	basis	basis	NOUN
fcis-18462	48	15	of	of	ADP
fcis-18462	48	16	the	the	DET
fcis-18462	48	17	same	same	ADJ
fcis-18462	48	18	or	or	CCONJ
fcis-18462	48	19	improved	improve	VERB
fcis-18462	48	20	accuracy	accuracy	NOUN
fcis-18462	48	21	10	10	NUM
fcis-18462	48	22	%	%	NOUN
fcis-18462	48	23	~	~	PUNCT
fcis-18462	48	24	20	20	NUM
fcis-18462	48	25	%	%	NOUN
fcis-18462	48	26	of	of	ADP
fcis-18462	48	27	network	network	NOUN
fcis-18462	48	28	parameters	parameter	NOUN
fcis-18462	48	29	.	.	PUNCT
fcis-18462	49	1	）this	）this	PRON
fcis-18462	49	2	made	make	VERB
fcis-18462	49	3	it	it	PRON
fcis-18462	49	4	possible	possible	ADJ
fcis-18462	49	5	to	to	PART
fcis-18462	49	6	increase	increase	VERB
fcis-18462	49	7	the	the	DET
fcis-18462	49	8	ability	ability	NOUN
fcis-18462	49	9	to	to	PART
fcis-18462	49	10	detect	detect	VERB
fcis-18462	49	11	small	small	ADJ
fcis-18462	49	12	targets	target	NOUN
fcis-18462	49	13	,	,	PUNCT
fcis-18462	49	14	and	and	CCONJ
fcis-18462	49	15	at	at	ADP
fcis-18462	49	16	the	the	DET
fcis-18462	49	17	same	same	ADJ
fcis-18462	49	18	time	time	NOUN
fcis-18462	49	19	,	,	PUNCT
fcis-18462	49	20	a	a	DET
fcis-18462	49	21	variable	variable	ADJ
fcis-18462	49	22	input	input	NOUN
fcis-18462	49	23	size	size	NOUN
fcis-18462	49	24	of	of	ADP
fcis-18462	49	25	640x640	640x640	NUM
fcis-18462	49	26	was	be	AUX
fcis-18462	49	27	used	use	VERB
fcis-18462	49	28	.	.	PUNCT
fcis-18462	50	1	this	this	PRON
fcis-18462	50	2	allows	allow	VERB
fcis-18462	50	3	for	for	ADP
fcis-18462	50	4	adaptation	adaptation	NOUN
fcis-18462	50	5	to	to	ADP
fcis-18462	50	6	different	different	ADJ
fcis-18462	50	7	scenarios	scenario	NOUN
fcis-18462	50	8	and	and	CCONJ
fcis-18462	50	9	needs	need	NOUN
fcis-18462	50	10	.	.	PUNCT
fcis-18462	51	1	and	and	CCONJ
fcis-18462	51	2	there	there	PRON
fcis-18462	51	3	are	be	VERB
fcis-18462	51	4	876	876	NUM
fcis-18462	51	5	million	million	NUM
fcis-18462	51	6	parameters	parameter	NOUN
fcis-18462	51	7	.	.	PUNCT
fcis-18462	52	1	yolov5	yolov5	NOUN
fcis-18462	52	2	increased	increase	VERB
fcis-18462	52	3	the	the	DET
fcis-18462	52	4	number	number	NOUN
fcis-18462	52	5	of	of	ADP
fcis-18462	52	6	parameters	parameter	NOUN
fcis-18462	52	7	to	to	PART
fcis-18462	52	8	improve	improve	VERB
fcis-18462	52	9	the	the	DET
fcis-18462	52	10	expressiveness	expressiveness	NOUN
fcis-18462	52	11	and	and	CCONJ
fcis-18462	52	12	generalization	generalization	NOUN
fcis-18462	52	13	ability	ability	NOUN
fcis-18462	52	14	of	of	ADP
fcis-18462	52	15	the	the	DET
fcis-18462	52	16	model	model	NOUN
fcis-18462	52	17	,	,	PUNCT
fcis-18462	52	18	and	and	CCONJ
fcis-18462	52	19	the	the	DET
fcis-18462	52	20	map	map	NOUN
fcis-18462	52	21	and	and	CCONJ
fcis-18462	52	22	fps	fps	NOUN
fcis-18462	52	23	of	of	ADP
fcis-18462	52	24	yolov5	yolov5	NOUN
fcis-18462	52	25	on	on	ADP
fcis-18462	52	26	the	the	DET
fcis-18462	52	27	coco	coco	PROPN
fcis-18462	52	28	dataset	dataset	PROPN
fcis-18462	52	29	were	be	AUX
fcis-18462	52	30	36.2	36.2	NUM
fcis-18462	52	31	%	%	NOUN
fcis-18462	52	32	and	and	CCONJ
fcis-18462	52	33	140.0	140.0	NUM
fcis-18462	52	34	.	.	PUNCT
fcis-18462	53	1	yolov5	yolov5	NOUN
fcis-18462	53	2	is	be	AUX
fcis-18462	53	3	superior	superior	ADJ
fcis-18462	53	4	to	to	ADP
fcis-18462	53	5	the	the	DET
fcis-18462	53	6	rest	rest	NOUN
fcis-18462	53	7	in	in	ADP
fcis-18462	53	8	terms	term	NOUN
fcis-18462	53	9	of	of	ADP
fcis-18462	53	10	accuracy	accuracy	NOUN
fcis-18462	53	11	and	and	CCONJ
fcis-18462	53	12	speed	speed	NOUN
fcis-18462	53	13	.	.	PUNCT
fcis-18462	54	1	therefore	therefore	ADV
fcis-18462	54	2	,	,	PUNCT
fcis-18462	54	3	we	we	PRON
fcis-18462	54	4	use	use	VERB
fcis-18462	54	5	yolov5	yolov5	NOUN
fcis-18462	54	6	to	to	PART
fcis-18462	54	7	improve	improve	VERB
fcis-18462	54	8	the	the	DET
fcis-18462	54	9	accuracy	accuracy	NOUN
fcis-18462	54	10	and	and	CCONJ
fcis-18462	54	11	efficiency	efficiency	NOUN
fcis-18462	54	12	of	of	ADP
fcis-18462	54	13	traffic	traffic	NOUN
fcis-18462	54	14	flow	flow	NOUN
fcis-18462	54	15	measurement[2	measurement[2	PROPN
fcis-18462	54	16	]	]	PUNCT
fcis-18462	54	17	.	.	PUNCT
fcis-18462	55	1	yolov5	yolov5	NOUN
fcis-18462	55	2	is	be	AUX
fcis-18462	55	3	a	a	DET
fcis-18462	55	4	pre	pre	ADJ
fcis-18462	55	5	-	-	ADJ
fcis-18462	55	6	trained	train	VERB
fcis-18462	55	7	object	object	NOUN
fcis-18462	55	8	detection	detection	NOUN
fcis-18462	55	9	architecture	architecture	NOUN
fcis-18462	55	10	and	and	CCONJ
fcis-18462	55	11	model	model	NOUN
fcis-18462	55	12	series	series	PROPN
fcis-18462	55	13	on	on	ADP
fcis-18462	55	14	the	the	DET
fcis-18462	55	15	coco	coco	PROPN
fcis-18462	55	16	dataset	dataset	PROPN
fcis-18462	55	17	,	,	PUNCT
fcis-18462	55	18	which	which	PRON
fcis-18462	55	19	is	be	AUX
fcis-18462	55	20	an	an	DET
fcis-18462	55	21	extension	extension	NOUN
fcis-18462	55	22	of	of	ADP
fcis-18462	55	23	the	the	DET
fcis-18462	55	24	yolo	yolo	ADJ
fcis-18462	55	25	series	series	NOUN
fcis-18462	55	26	.	.	PUNCT
fcis-18462	56	1	neck	neck	NOUN
fcis-18462	56	2	and	and	CCONJ
fcis-18462	56	3	head	head	NOUN
fcis-18462	56	4	,	,	PUNCT
fcis-18462	56	5	among	among	ADP
fcis-18462	56	6	them	they	PRON
fcis-18462	56	7	,	,	PUNCT
fcis-18462	56	8	backbone	backbone	NOUN
fcis-18462	56	9	is	be	AUX
fcis-18462	56	10	responsible	responsible	ADJ
fcis-18462	56	11	for	for	ADP
fcis-18462	56	12	feature	feature	NOUN
fcis-18462	56	13	extraction	extraction	NOUN
fcis-18462	56	14	,	,	PUNCT
fcis-18462	56	15	neck	neck	NOUN
fcis-18462	56	16	is	be	AUX
fcis-18462	56	17	responsible	responsible	ADJ
fcis-18462	56	18	for	for	ADP
fcis-18462	56	19	feature	feature	NOUN
fcis-18462	56	20	fusion	fusion	NOUN
fcis-18462	56	21	,	,	PUNCT
fcis-18462	56	22	and	and	CCONJ
fcis-18462	56	23	head	head	NOUN
fcis-18462	56	24	contains	contain	VERB
fcis-18462	56	25	three	three	NUM
fcis-18462	56	26	detection	detection	NOUN
fcis-18462	56	27	heads	head	NOUN
fcis-18462	56	28	,	,	PUNCT
fcis-18462	56	29	77	77	NUM
fcis-18462	56	30	which	which	PRON
fcis-18462	56	31	are	be	AUX
fcis-18462	56	32	responsible	responsible	ADJ
fcis-18462	56	33	for	for	ADP
fcis-18462	56	34	outputting	output	VERB
fcis-18462	56	35	detection	detection	NOUN
fcis-18462	56	36	information	information	NOUN
fcis-18462	56	37	,	,	PUNCT
fcis-18462	56	38	yolov5	yolov5	NOUN
fcis-18462	56	39	has	have	AUX
fcis-18462	56	40	modified	modify	VERB
fcis-18462	56	41	the	the	DET
fcis-18462	56	42	four	four	NUM
fcis-18462	56	43	parts	part	NOUN
fcis-18462	56	44	of	of	ADP
fcis-18462	56	45	the	the	DET
fcis-18462	56	46	yolov4	yolov4	PROPN
fcis-18462	56	47	network	network	NOUN
fcis-18462	56	48	,	,	PUNCT
fcis-18462	56	49	and	and	CCONJ
fcis-18462	56	50	has	have	AUX
fcis-18462	56	51	achieved	achieve	VERB
fcis-18462	56	52	great	great	ADJ
fcis-18462	56	53	improvements	improvement	NOUN
fcis-18462	56	54	,	,	PUNCT
fcis-18462	56	55	using	use	VERB
fcis-18462	56	56	mosaic	mosaic	ADJ
fcis-18462	56	57	data	datum	NOUN
fcis-18462	56	58	enhancement	enhancement	NOUN
fcis-18462	56	59	,	,	PUNCT
fcis-18462	56	60	adaptive	adaptive	ADJ
fcis-18462	56	61	anchor	anchor	NOUN
fcis-18462	56	62	frame	frame	NOUN
fcis-18462	56	63	calculation	calculation	NOUN
fcis-18462	56	64	,	,	PUNCT
fcis-18462	56	65	and	and	CCONJ
fcis-18462	56	66	adaptive	adaptive	ADJ
fcis-18462	56	67	image	image	NOUN
fcis-18462	56	68	scaling	scale	VERB
fcis-18462	56	69	on	on	ADP
fcis-18462	56	70	the	the	DET
fcis-18462	56	71	input	input	NOUN
fcis-18462	56	72	side	side	NOUN
fcis-18462	56	73	.	.	PUNCT
fcis-18462	57	1	cbs	cbs	PROPN
fcis-18462	57	2	module	module	NOUN
fcis-18462	57	3	consists	consist	VERB
fcis-18462	57	4	of	of	ADP
fcis-18462	57	5	a	a	DET
fcis-18462	57	6	convolution	convolution	NOUN
fcis-18462	57	7	layer	layer	NOUN
fcis-18462	57	8	conv	conv	NOUN
fcis-18462	57	9	,	,	PUNCT
fcis-18462	57	10	a	a	DET
fcis-18462	57	11	batch	batch	NOUN
fcis-18462	57	12	normalization	normalization	NOUN
fcis-18462	57	13	layer	layer	NOUN
fcis-18462	57	14	batchnorm	batchnorm	NOUN
fcis-18462	57	15	and	and	CCONJ
fcis-18462	57	16	silu	silu	ADJ
fcis-18462	57	17	activation	activation	NOUN
fcis-18462	57	18	function	function	NOUN
fcis-18462	57	19	layer	layer	NOUN
fcis-18462	57	20	.	.	PUNCT
fcis-18462	58	1	the	the	DET
fcis-18462	58	2	specific	specific	ADJ
fcis-18462	58	3	structure	structure	NOUN
fcis-18462	58	4	diagram	diagram	NOUN
fcis-18462	58	5	is	be	AUX
fcis-18462	58	6	shown	show	VERB
fcis-18462	58	7	below	below	ADV
fcis-18462	58	8	.	.	PUNCT
fcis-18462	59	1	figure	figure	VERB
fcis-18462	59	2	3	3	NUM
fcis-18462	59	3	.	.	PUNCT
fcis-18462	60	1	cbs	cbs	PROPN
fcis-18462	60	2	module	module	NOUN
fcis-18462	60	3	structure	structure	NOUN
fcis-18462	60	4	diagram	diagram	NOUN
fcis-18462	60	5	cbs	cbs	PROPN
fcis-18462	60	6	module	module	NOUN
fcis-18462	60	7	can	can	AUX
fcis-18462	60	8	extract	extract	VERB
fcis-18462	60	9	the	the	DET
fcis-18462	60	10	features	feature	NOUN
fcis-18462	60	11	in	in	ADP
fcis-18462	60	12	the	the	DET
fcis-18462	60	13	image	image	NOUN
fcis-18462	60	14	and	and	CCONJ
fcis-18462	60	15	sort	sort	VERB
fcis-18462	60	16	out	out	ADP
fcis-18462	60	17	the	the	DET
fcis-18462	60	18	extracted	extract	VERB
fcis-18462	60	19	feature	feature	NOUN
fcis-18462	60	20	map	map	NOUN
fcis-18462	60	21	.	.	PUNCT
fcis-18462	61	1	conv	conv	ADJ
fcis-18462	61	2	of	of	ADP
fcis-18462	61	3	the	the	DET
fcis-18462	61	4	convolution	convolution	NOUN
fcis-18462	61	5	layer	layer	NOUN
fcis-18462	61	6	in	in	ADP
fcis-18462	61	7	cbs	cbs	PROPN
fcis-18462	61	8	module	module	NOUN
fcis-18462	61	9	changes	change	VERB
fcis-18462	61	10	the	the	DET
fcis-18462	61	11	multiple	multiple	NOUN
fcis-18462	61	12	of	of	ADP
fcis-18462	61	13	the	the	DET
fcis-18462	61	14	feature	feature	NOUN
fcis-18462	61	15	map	map	NOUN
fcis-18462	61	16	reduction	reduction	NOUN
fcis-18462	61	17	by	by	ADP
fcis-18462	61	18	modifying	modify	VERB
fcis-18462	61	19	the	the	DET
fcis-18462	61	20	step	step	NOUN
fcis-18462	61	21	size	size	NOUN
fcis-18462	61	22	,	,	PUNCT
fcis-18462	61	23	and	and	CCONJ
fcis-18462	61	24	the	the	DET
fcis-18462	61	25	image	image	NOUN
fcis-18462	61	26	filling	fill	VERB
fcis-18462	61	27	operation	operation	NOUN
fcis-18462	61	28	is	be	AUX
fcis-18462	61	29	calculated	calculate	VERB
fcis-18462	61	30	automatically	automatically	ADV
fcis-18462	61	31	by	by	ADP
fcis-18462	61	32	the	the	DET
fcis-18462	61	33	model	model	NOUN
fcis-18462	61	34	.	.	PUNCT
fcis-18462	62	1	in	in	ADP
fcis-18462	62	2	backbone	backbone	NOUN
fcis-18462	62	3	,	,	PUNCT
fcis-18462	62	4	the	the	DET
fcis-18462	62	5	step	step	NOUN
fcis-18462	62	6	size	size	NOUN
fcis-18462	62	7	of	of	ADP
fcis-18462	62	8	conv	conv	NOUN
fcis-18462	62	9	of	of	ADP
fcis-18462	62	10	convolution	convolution	NOUN
fcis-18462	62	11	layer	layer	NOUN
fcis-18462	62	12	is	be	AUX
fcis-18462	62	13	2	2	NUM
fcis-18462	62	14	,	,	PUNCT
fcis-18462	62	15	and	and	CCONJ
fcis-18462	62	16	the	the	DET
fcis-18462	62	17	size	size	NOUN
fcis-18462	62	18	of	of	ADP
fcis-18462	62	19	convolution	convolution	NOUN
fcis-18462	62	20	kernel	kernel	NOUN
fcis-18462	62	21	is	be	AUX
fcis-18462	62	22	3	3	NUM
fcis-18462	62	23	x3	x3	ADJ
fcis-18462	62	24	.	.	PUNCT
fcis-18462	63	1	through	through	ADP
fcis-18462	63	2	this	this	DET
fcis-18462	63	3	convolution	convolution	NOUN
fcis-18462	63	4	layer	layer	NOUN
fcis-18462	63	5	,	,	PUNCT
fcis-18462	63	6	the	the	DET
fcis-18462	63	7	width	width	NOUN
fcis-18462	63	8	and	and	CCONJ
fcis-18462	63	9	height	height	NOUN
fcis-18462	63	10	of	of	ADP
fcis-18462	63	11	feature	feature	NOUN
fcis-18462	63	12	map	map	NOUN
fcis-18462	63	13	will	will	AUX
fcis-18462	63	14	be	be	AUX
fcis-18462	63	15	halved	halve	VERB
fcis-18462	63	16	,	,	PUNCT
fcis-18462	63	17	and	and	CCONJ
fcis-18462	63	18	feature	feature	NOUN
fcis-18462	63	19	extraction	extraction	NOUN
fcis-18462	63	20	of	of	ADP
fcis-18462	63	21	image	image	NOUN
fcis-18462	63	22	is	be	AUX
fcis-18462	63	23	carried	carry	VERB
fcis-18462	63	24	out	out	ADP
fcis-18462	63	25	at	at	ADP
fcis-18462	63	26	the	the	DET
fcis-18462	63	27	same	same	ADJ
fcis-18462	63	28	time	time	NOUN
fcis-18462	63	29	.	.	PUNCT
fcis-18462	64	1	in	in	ADP
fcis-18462	64	2	neck	neck	NOUN
fcis-18462	64	3	,	,	PUNCT
fcis-18462	64	4	convolution	convolution	NOUN
fcis-18462	64	5	check	check	NOUN
fcis-18462	64	6	feature	feature	NOUN
fcis-18462	64	7	maps	map	NOUN
fcis-18462	64	8	with	with	ADP
fcis-18462	64	9	a	a	DET
fcis-18462	64	10	size	size	NOUN
fcis-18462	64	11	of	of	ADP
fcis-18462	64	12	1x1	1x1	NUM
fcis-18462	64	13	are	be	AUX
fcis-18462	64	14	mostly	mostly	ADV
fcis-18462	64	15	used	use	VERB
fcis-18462	64	16	to	to	PART
fcis-18462	64	17	raise	raise	VERB
fcis-18462	64	18	or	or	CCONJ
fcis-18462	64	19	reduce	reduce	VERB
fcis-18462	64	20	dimension	dimension	NOUN
fcis-18462	64	21	.	.	PUNCT
fcis-18462	65	1	the	the	DET
fcis-18462	65	2	batchnorm	batchnorm	NOUN
fcis-18462	65	3	layer	layer	NOUN
fcis-18462	65	4	is	be	AUX
fcis-18462	65	5	added	add	VERB
fcis-18462	65	6	after	after	SCONJ
fcis-18462	65	7	the	the	DET
fcis-18462	65	8	convolution	convolution	NOUN
fcis-18462	65	9	layer	layer	NOUN
fcis-18462	65	10	to	to	PART
fcis-18462	65	11	normalize	normalize	VERB
fcis-18462	65	12	the	the	DET
fcis-18462	65	13	extracted	extract	VERB
fcis-18462	65	14	features	feature	NOUN
fcis-18462	65	15	,	,	PUNCT
fcis-18462	65	16	so	so	SCONJ
fcis-18462	65	17	that	that	SCONJ
fcis-18462	65	18	the	the	DET
fcis-18462	65	19	network	network	NOUN
fcis-18462	65	20	will	will	AUX
fcis-18462	65	21	not	not	PART
fcis-18462	65	22	be	be	AUX
fcis-18462	65	23	unstable	unstable	ADJ
fcis-18462	65	24	due	due	ADP
fcis-18462	65	25	to	to	ADP
fcis-18462	65	26	large	large	ADJ
fcis-18462	65	27	data	datum	NOUN
fcis-18462	65	28	before	before	ADP
fcis-18462	65	29	the	the	DET
fcis-18462	65	30	activation	activation	NOUN
fcis-18462	65	31	function	function	NOUN
fcis-18462	65	32	,	,	PUNCT
fcis-18462	65	33	and	and	CCONJ
fcis-18462	65	34	to	to	PART
fcis-18462	65	35	prevent	prevent	VERB
fcis-18462	65	36	the	the	DET
fcis-18462	65	37	gradient	gradient	NOUN
fcis-18462	65	38	explosion	explosion	NOUN
fcis-18462	65	39	,	,	PUNCT
fcis-18462	65	40	disappearance	disappearance	NOUN
fcis-18462	65	41	and	and	CCONJ
fcis-18462	65	42	overfitting	overfitte	VERB
fcis-18462	65	43	problems	problem	NOUN
fcis-18462	65	44	.	.	PUNCT
fcis-18462	66	1	at	at	ADP
fcis-18462	66	2	the	the	DET
fcis-18462	66	3	very	very	ADJ
fcis-18462	66	4	end	end	NOUN
fcis-18462	66	5	,	,	PUNCT
fcis-18462	66	6	the	the	DET
fcis-18462	66	7	silu	silu	ADJ
fcis-18462	66	8	activation	activation	NOUN
fcis-18462	66	9	function	function	NOUN
fcis-18462	66	10	is	be	AUX
fcis-18462	66	11	used	use	VERB
fcis-18462	66	12	,	,	PUNCT
fcis-18462	66	13	adding	add	VERB
fcis-18462	66	14	smooth	smooth	ADJ
fcis-18462	66	15	and	and	CCONJ
fcis-18462	66	16	non	non	ADJ
fcis-18462	66	17	-	-	ADJ
fcis-18462	66	18	monotonic	monotonic	ADJ
fcis-18462	66	19	features	feature	NOUN
fcis-18462	66	20	to	to	PART
fcis-18462	66	21	increase	increase	VERB
fcis-18462	66	22	the	the	DET
fcis-18462	66	23	nonlinearity	nonlinearity	NOUN
fcis-18462	66	24	of	of	ADP
fcis-18462	66	25	the	the	DET
fcis-18462	66	26	data	datum	NOUN
fcis-18462	66	27	.	.	PUNCT
fcis-18462	67	1	focus	focus	NOUN
fcis-18462	67	2	structure	structure	NOUN
fcis-18462	67	3	is	be	AUX
fcis-18462	67	4	used	use	VERB
fcis-18462	67	5	on	on	ADP
fcis-18462	67	6	the	the	DET
fcis-18462	67	7	backbone	backbone	NOUN
fcis-18462	67	8	side	side	NOUN
fcis-18462	67	9	.	.	PUNCT
fcis-18462	68	1	(	(	PUNCT
fcis-18462	68	2	focus	focus	NOUN
fcis-18462	68	3	structure	structure	NOUN
fcis-18462	68	4	is	be	AUX
fcis-18462	68	5	a	a	DET
fcis-18462	68	6	special	special	ADJ
fcis-18462	68	7	layer	layer	NOUN
fcis-18462	68	8	in	in	ADP
fcis-18462	68	9	yolov5	yolov5	NOUN
fcis-18462	68	10	,	,	PUNCT
fcis-18462	68	11	mainly	mainly	ADV
fcis-18462	68	12	used	use	VERB
fcis-18462	68	13	to	to	PART
fcis-18462	68	14	improve	improve	VERB
fcis-18462	68	15	the	the	DET
fcis-18462	68	16	detection	detection	NOUN
fcis-18462	68	17	ability	ability	NOUN
fcis-18462	68	18	of	of	ADP
fcis-18462	68	19	the	the	DET
fcis-18462	68	20	model	model	NOUN
fcis-18462	68	21	on	on	ADP
fcis-18462	68	22	small	small	ADJ
fcis-18462	68	23	targets	target	NOUN
fcis-18462	68	24	.	.	PUNCT
fcis-18462	69	1	this	this	DET
fcis-18462	69	2	design	design	NOUN
fcis-18462	69	3	can	can	AUX
fcis-18462	69	4	make	make	VERB
fcis-18462	69	5	the	the	DET
fcis-18462	69	6	model	model	NOUN
fcis-18462	69	7	better	well	ADJ
fcis-18462	69	8	focus	focus	NOUN
fcis-18462	69	9	on	on	ADP
fcis-18462	69	10	the	the	DET
fcis-18462	69	11	features	feature	NOUN
fcis-18462	69	12	of	of	ADP
fcis-18462	69	13	small	small	ADJ
fcis-18462	69	14	targets	target	NOUN
fcis-18462	69	15	,	,	PUNCT
fcis-18462	69	16	thus	thus	ADV
fcis-18462	69	17	improving	improve	VERB
fcis-18462	69	18	the	the	DET
fcis-18462	69	19	accuracy	accuracy	NOUN
fcis-18462	69	20	of	of	ADP
fcis-18462	69	21	target	target	NOUN
fcis-18462	69	22	detection	detection	NOUN
fcis-18462	69	23	and	and	CCONJ
fcis-18462	69	24	recall	recall	NOUN
fcis-18462	69	25	rate	rate	NOUN
fcis-18462	69	26	.	.	PUNCT
fcis-18462	70	1	especially	especially	ADV
fcis-18462	70	2	for	for	ADP
fcis-18462	70	3	small	small	ADJ
fcis-18462	70	4	target	target	NOUN
fcis-18462	70	5	detection	detection	NOUN
fcis-18462	70	6	,	,	PUNCT
fcis-18462	70	7	the	the	DET
fcis-18462	70	8	focus	focus	NOUN
fcis-18462	70	9	structure	structure	NOUN
fcis-18462	70	10	can	can	AUX
fcis-18462	70	11	bring	bring	VERB
fcis-18462	70	12	obvious	obvious	ADJ
fcis-18462	70	13	advantages	advantage	NOUN
fcis-18462	70	14	.	.	PUNCT
fcis-18462	71	1	in	in	ADP
fcis-18462	71	2	addition	addition	NOUN
fcis-18462	71	3	,	,	PUNCT
fcis-18462	71	4	the	the	DET
fcis-18462	71	5	focus	focus	NOUN
fcis-18462	71	6	structure	structure	NOUN
fcis-18462	71	7	can	can	AUX
fcis-18462	71	8	reduce	reduce	VERB
fcis-18462	71	9	the	the	DET
fcis-18462	71	10	computational	computational	ADJ
fcis-18462	71	11	effort	effort	NOUN
fcis-18462	71	12	and	and	CCONJ
fcis-18462	71	13	memory	memory	NOUN
fcis-18462	71	14	footprint	footprint	NOUN
fcis-18462	71	15	by	by	ADP
fcis-18462	71	16	breaking	break	VERB
fcis-18462	71	17	the	the	DET
fcis-18462	71	18	input	input	NOUN
fcis-18462	71	19	feature	feature	NOUN
fcis-18462	71	20	graph	graph	NOUN
fcis-18462	71	21	into	into	ADP
fcis-18462	71	22	multiple	multiple	ADJ
fcis-18462	71	23	subgraphs	subgraph	NOUN
fcis-18462	71	24	and	and	CCONJ
fcis-18462	71	25	concatenating	concatenate	VERB
fcis-18462	71	26	the	the	DET
fcis-18462	71	27	channel	channel	NOUN
fcis-18462	71	28	dimensions	dimension	NOUN
fcis-18462	71	29	instead	instead	ADV
fcis-18462	71	30	of	of	ADP
fcis-18462	71	31	directly	directly	ADV
fcis-18462	71	32	copying	copy	VERB
fcis-18462	71	33	the	the	DET
fcis-18462	71	34	spatial	spatial	ADJ
fcis-18462	71	35	dimensions	dimension	NOUN
fcis-18462	71	36	of	of	ADP
fcis-18462	71	37	the	the	DET
fcis-18462	71	38	entire	entire	ADJ
fcis-18462	71	39	feature	feature	NOUN
fcis-18462	71	40	graph	graph	NOUN
fcis-18462	71	41	.	.	PUNCT
fcis-18462	72	1	such	such	ADJ
fcis-18462	72	2	operations	operation	NOUN
fcis-18462	72	3	can	can	AUX
fcis-18462	72	4	improve	improve	VERB
fcis-18462	72	5	the	the	DET
fcis-18462	72	6	overall	overall	ADJ
fcis-18462	72	7	performance	performance	NOUN
fcis-18462	72	8	of	of	ADP
fcis-18462	72	9	the	the	DET
fcis-18462	72	10	model	model	NOUN
fcis-18462	72	11	by	by	ADP
fcis-18462	72	12	reducing	reduce	VERB
fcis-18462	72	13	the	the	DET
fcis-18462	72	14	number	number	NOUN
fcis-18462	72	15	of	of	ADP
fcis-18462	72	16	channels	channel	NOUN
fcis-18462	72	17	in	in	ADP
fcis-18462	72	18	the	the	DET
fcis-18462	72	19	feature	feature	NOUN
fcis-18462	72	20	graph	graph	NOUN
fcis-18462	72	21	while	while	SCONJ
fcis-18462	72	22	maintaining	maintain	VERB
fcis-18462	72	23	critical	critical	ADJ
fcis-18462	72	24	information	information	NOUN
fcis-18462	72	25	)	)	PUNCT
fcis-18462	72	26	and	and	CCONJ
fcis-18462	72	27	csp	csp	PROPN
fcis-18462	72	28	structures	structure	NOUN
fcis-18462	72	29	(	(	PUNCT
fcis-18462	72	30	csp	csp	PROPN
fcis-18462	72	31	structures	structure	NOUN
fcis-18462	72	32	improve	improve	VERB
fcis-18462	72	33	the	the	DET
fcis-18462	72	34	performance	performance	NOUN
fcis-18462	72	35	of	of	ADP
fcis-18462	72	36	deep	deep	ADJ
fcis-18462	72	37	neural	neural	ADJ
fcis-18462	72	38	networks	network	NOUN
fcis-18462	72	39	by	by	ADP
fcis-18462	72	40	dividing	divide	VERB
fcis-18462	72	41	input	input	NOUN
fcis-18462	72	42	features	feature	NOUN
fcis-18462	72	43	into	into	ADP
fcis-18462	72	44	two	two	NUM
fcis-18462	72	45	parts	part	NOUN
fcis-18462	72	46	and	and	CCONJ
fcis-18462	72	47	then	then	ADV
fcis-18462	72	48	cross	cross	ADJ
fcis-18462	72	49	-	-	ADJ
fcis-18462	72	50	connecting	connect	VERB
fcis-18462	72	51	between	between	ADP
fcis-18462	72	52	the	the	DET
fcis-18462	72	53	two	two	NUM
fcis-18462	72	54	parts	part	NOUN
fcis-18462	72	55	.	.	PUNCT
fcis-18462	73	1	csp	csp	PROPN
fcis-18462	73	2	structure	structure	NOUN
fcis-18462	73	3	can	can	AUX
fcis-18462	73	4	effectively	effectively	ADV
fcis-18462	73	5	improve	improve	VERB
fcis-18462	73	6	the	the	DET
fcis-18462	73	7	feature	feature	NOUN
fcis-18462	73	8	representation	representation	NOUN
fcis-18462	73	9	ability	ability	NOUN
fcis-18462	73	10	of	of	ADP
fcis-18462	73	11	the	the	DET
fcis-18462	73	12	model	model	NOUN
fcis-18462	73	13	,	,	PUNCT
fcis-18462	73	14	thus	thus	ADV
fcis-18462	73	15	improving	improve	VERB
fcis-18462	73	16	the	the	DET
fcis-18462	73	17	accuracy	accuracy	NOUN
fcis-18462	73	18	and	and	CCONJ
fcis-18462	73	19	generalization	generalization	NOUN
fcis-18462	73	20	ability	ability	NOUN
fcis-18462	73	21	of	of	ADP
fcis-18462	73	22	the	the	DET
fcis-18462	73	23	model	model	NOUN
fcis-18462	73	24	.	.	PUNCT
fcis-18462	74	1	fpn+pan	fpn+pan	NOUN
fcis-18462	74	2	structure	structure	NOUN
fcis-18462	74	3	was	be	AUX
fcis-18462	74	4	added	add	VERB
fcis-18462	74	5	to	to	ADP
fcis-18462	74	6	the	the	DET
fcis-18462	74	7	neck	neck	NOUN
fcis-18462	74	8	side	side	NOUN
fcis-18462	74	9	,	,	PUNCT
fcis-18462	74	10	the	the	DET
fcis-18462	74	11	loss	loss	NOUN
fcis-18462	74	12	function	function	NOUN
fcis-18462	74	13	during	during	ADP
fcis-18462	74	14	training	training	NOUN
fcis-18462	74	15	was	be	AUX
fcis-18462	74	16	improved	improve	VERB
fcis-18462	74	17	on	on	ADP
fcis-18462	74	18	the	the	DET
fcis-18462	74	19	head	head	NOUN
fcis-18462	74	20	side	side	NOUN
fcis-18462	74	21	,	,	PUNCT
fcis-18462	74	22	and	and	CCONJ
fcis-18462	74	23	giou_loss	giou_loss	PROPN
fcis-18462	74	24	was	be	AUX
fcis-18462	74	25	used	use	VERB
fcis-18462	74	26	to	to	PART
fcis-18462	74	27	improve	improve	VERB
fcis-18462	74	28	diou_nms	diou_nms	NOUN
fcis-18462	74	29	for	for	ADP
fcis-18462	74	30	prediction	prediction	NOUN
fcis-18462	74	31	box	box	NOUN
fcis-18462	74	32	screening	screening	NOUN
fcis-18462	74	33	.	.	PUNCT
fcis-18462	75	1	[	[	X
fcis-18462	75	2	3	3	NUM
fcis-18462	75	3	]	]	PUNCT
fcis-18462	75	4	.	.	PUNCT
fcis-18462	76	1	compared	compare	VERB
fcis-18462	76	2	with	with	ADP
fcis-18462	76	3	the	the	DET
fcis-18462	76	4	fixed	fix	VERB
fcis-18462	76	5	size	size	NOUN
fcis-18462	76	6	and	and	CCONJ
fcis-18462	76	7	proportion	proportion	NOUN
fcis-18462	76	8	used	use	VERB
fcis-18462	76	9	in	in	ADP
fcis-18462	76	10	the	the	DET
fcis-18462	76	11	traditional	traditional	ADJ
fcis-18462	76	12	method	method	NOUN
fcis-18462	76	13	,	,	PUNCT
fcis-18462	76	14	the	the	DET
fcis-18462	76	15	adaptive	adaptive	ADJ
fcis-18462	76	16	anchor	anchor	NOUN
fcis-18462	76	17	frame	frame	NOUN
fcis-18462	76	18	has	have	VERB
fcis-18462	76	19	many	many	ADJ
fcis-18462	76	20	advantages	advantage	NOUN
fcis-18462	76	21	,	,	PUNCT
fcis-18462	76	22	such	such	ADJ
fcis-18462	76	23	as	as	ADP
fcis-18462	76	24	enriching	enrich	VERB
fcis-18462	76	25	the	the	DET
fcis-18462	76	26	detection	detection	NOUN
fcis-18462	76	27	of	of	ADP
fcis-18462	76	28	the	the	DET
fcis-18462	76	29	object	object	NOUN
fcis-18462	76	30	's	's	PART
fcis-18462	76	31	background	background	NOUN
fcis-18462	76	32	and	and	CCONJ
fcis-18462	76	33	small	small	ADJ
fcis-18462	76	34	target	target	NOUN
fcis-18462	76	35	,	,	PUNCT
fcis-18462	76	36	and	and	CCONJ
fcis-18462	76	37	calculating	calculate	VERB
fcis-18462	76	38	the	the	DET
fcis-18462	76	39	batch	batch	NOUN
fcis-18462	76	40	as	as	SCONJ
fcis-18462	76	41	it	it	PRON
fcis-18462	76	42	uses	use	VERB
fcis-18462	76	43	normalization	normalization	NOUN
fcis-18462	76	44	,	,	PUNCT
fcis-18462	76	45	it	it	PRON
fcis-18462	76	46	computations	computation	VERB
fcis-18462	76	47	data	datum	NOUN
fcis-18462	76	48	on	on	ADP
fcis-18462	76	49	four	four	NUM
fcis-18462	76	50	images	image	NOUN
fcis-18462	76	51	at	at	ADP
fcis-18462	76	52	a	a	DET
fcis-18462	76	53	time	time	NOUN
fcis-18462	76	54	,	,	PUNCT
fcis-18462	76	55	so	so	SCONJ
fcis-18462	76	56	that	that	SCONJ
fcis-18462	76	57	the	the	DET
fcis-18462	76	58	mini	mini	NOUN
fcis-18462	76	59	-	-	NOUN
fcis-18462	76	60	batch	batch	ADJ
fcis-18462	76	61	size	size	NOUN
fcis-18462	76	62	does	do	AUX
fcis-18462	76	63	n't	not	PART
fcis-18462	76	64	have	have	VERB
fcis-18462	76	65	to	to	PART
fcis-18462	76	66	be	be	AUX
fcis-18462	76	67	as	as	ADV
fcis-18462	76	68	large	large	ADJ
fcis-18462	76	69	as	as	SCONJ
fcis-18462	76	70	it	it	PRON
fcis-18462	76	71	uses	use	VERB
fcis-18462	76	72	a	a	DET
fcis-18462	76	73	single	single	ADJ
fcis-18462	76	74	gpu	gpu	NOUN
fcis-18462	76	75	to	to	PART
fcis-18462	76	76	get	get	VERB
fcis-18462	76	77	good	good	ADJ
fcis-18462	76	78	results	result	NOUN
fcis-18462	76	79	.	.	PUNCT
fcis-18462	77	1	adaptive	adaptive	ADJ
fcis-18462	77	2	anchor	anchor	NOUN
fcis-18462	77	3	frame	frame	NOUN
fcis-18462	77	4	calculation	calculation	NOUN
fcis-18462	77	5	in	in	ADP
fcis-18462	77	6	both	both	DET
fcis-18462	77	7	yolov3	yolov3	PROPN
fcis-18462	77	8	and	and	CCONJ
fcis-18462	77	9	yolov4	yolov4	PROPN
fcis-18462	77	10	,	,	PUNCT
fcis-18462	77	11	anchor	anchor	PROPN
fcis-18462	77	12	needs	need	VERB
fcis-18462	77	13	to	to	PART
fcis-18462	77	14	be	be	AUX
fcis-18462	77	15	calculated	calculate	VERB
fcis-18462	77	16	by	by	ADP
fcis-18462	77	17	k	k	PROPN
fcis-18462	77	18	-	-	PUNCT
fcis-18462	77	19	means	mean	VERB
fcis-18462	77	20	clustering	clustering	NOUN
fcis-18462	77	21	method	method	NOUN
fcis-18462	77	22	in	in	ADP
fcis-18462	77	23	advance	advance	NOUN
fcis-18462	77	24	,	,	PUNCT
fcis-18462	77	25	and	and	CCONJ
fcis-18462	77	26	the	the	DET
fcis-18462	77	27	anchor	anchor	NOUN
fcis-18462	77	28	is	be	AUX
fcis-18462	77	29	fixed	fix	VERB
fcis-18462	77	30	.	.	PUNCT
fcis-18462	78	1	however	however	ADV
fcis-18462	78	2	,	,	PUNCT
fcis-18462	78	3	in	in	ADP
fcis-18462	78	4	yolov5	yolov5	NOUN
fcis-18462	78	5	,	,	PUNCT
fcis-18462	78	6	although	although	SCONJ
fcis-18462	78	7	the	the	DET
fcis-18462	78	8	anchor	anchor	NOUN
fcis-18462	78	9	is	be	AUX
fcis-18462	78	10	set	set	VERB
fcis-18462	78	11	in	in	ADP
fcis-18462	78	12	advance	advance	NOUN
fcis-18462	78	13	,	,	PUNCT
fcis-18462	78	14	the	the	DET
fcis-18462	78	15	optimal	optimal	ADJ
fcis-18462	78	16	anchor	anchor	NOUN
fcis-18462	78	17	in	in	ADP
fcis-18462	78	18	different	different	ADJ
fcis-18462	78	19	training	training	NOUN
fcis-18462	78	20	sets	set	NOUN
fcis-18462	78	21	can	can	AUX
fcis-18462	78	22	be	be	AUX
fcis-18462	78	23	calculated	calculate	VERB
fcis-18462	78	24	adaptively	adaptively	ADV
fcis-18462	78	25	during	during	ADP
fcis-18462	78	26	training	training	NOUN
fcis-18462	78	27	,	,	PUNCT
fcis-18462	78	28	so	so	SCONJ
fcis-18462	78	29	as	as	SCONJ
fcis-18462	78	30	to	to	PART
fcis-18462	78	31	update	update	VERB
fcis-18462	78	32	the	the	DET
fcis-18462	78	33	anchor	anchor	NOUN
fcis-18462	78	34	value	value	NOUN
fcis-18462	78	35	.	.	PUNCT
fcis-18462	79	1	this	this	DET
fcis-18462	79	2	feature	feature	NOUN
fcis-18462	79	3	can	can	AUX
fcis-18462	79	4	also	also	ADV
fcis-18462	79	5	be	be	AUX
fcis-18462	79	6	turned	turn	VERB
fcis-18462	79	7	off	off	ADP
fcis-18462	79	8	manually	manually	ADV
fcis-18462	79	9	,	,	PUNCT
fcis-18462	79	10	changed	change	VERB
fcis-18462	79	11	in	in	ADP
fcis-18462	79	12	train.py	train.py	NUM
fcis-18462	79	13	,	,	PUNCT
fcis-18462	79	14	set	set	VERB
fcis-18462	79	15	to	to	ADP
fcis-18462	79	16	false	false	ADJ
fcis-18462	79	17	.	.	PUNCT
fcis-18462	80	1	the	the	DET
fcis-18462	80	2	specific	specific	ADJ
fcis-18462	80	3	implementation	implementation	NOUN
fcis-18462	80	4	structure	structure	NOUN
fcis-18462	80	5	of	of	ADP
fcis-18462	80	6	yolov5	yolov5	NOUN
fcis-18462	80	7	is	be	AUX
fcis-18462	80	8	as	as	SCONJ
fcis-18462	80	9	follows	follow	VERB
fcis-18462	80	10	[	[	X
fcis-18462	80	11	4	4	NUM
fcis-18462	80	12	]	]	PUNCT
fcis-18462	80	13	:	:	PUNCT
fcis-18462	80	14	figure	figure	NOUN
fcis-18462	80	15	4	4	NUM
fcis-18462	80	16	.	.	PUNCT
fcis-18462	81	1	yolov5	yolov5	NOUN
fcis-18462	81	2	structure	structure	NOUN
fcis-18462	81	3	diagram	diagram	PROPN
fcis-18462	81	4	4	4	NUM
fcis-18462	81	5	.	.	PUNCT
fcis-18462	82	1	algorithm	algorithm	NOUN
fcis-18462	82	2	design	design	VERB
fcis-18462	82	3	the	the	DET
fcis-18462	82	4	yolov5	yolov5	NOUN
fcis-18462	82	5	algorithm	algorithm	NOUN
fcis-18462	82	6	is	be	AUX
fcis-18462	82	7	used	use	VERB
fcis-18462	82	8	to	to	PART
fcis-18462	82	9	realize	realize	VERB
fcis-18462	82	10	object	object	NOUN
fcis-18462	82	11	detection	detection	NOUN
fcis-18462	82	12	,	,	PUNCT
fcis-18462	82	13	image	image	NOUN
fcis-18462	82	14	segmentation	segmentation	NOUN
fcis-18462	82	15	,	,	PUNCT
fcis-18462	82	16	and	and	CCONJ
fcis-18462	82	17	target	target	VERB
fcis-18462	82	18	tracking	tracking	NOUN
fcis-18462	82	19	to	to	PART
fcis-18462	82	20	realize	realize	VERB
fcis-18462	82	21	the	the	DET
fcis-18462	82	22	statistics	statistic	NOUN
fcis-18462	82	23	of	of	ADP
fcis-18462	82	24	traffic	traffic	NOUN
fcis-18462	82	25	flow	flow	NOUN
fcis-18462	82	26	.	.	PUNCT
fcis-18462	83	1	there	there	PRON
fcis-18462	83	2	are	be	VERB
fcis-18462	83	3	five	five	NUM
fcis-18462	83	4	versions	version	NOUN
fcis-18462	83	5	of	of	ADP
fcis-18462	83	6	yolov5	yolov5	NOUN
fcis-18462	83	7	,	,	PUNCT
fcis-18462	83	8	namely	namely	ADV
fcis-18462	83	9	yolov5n	yolov5n	PROPN
fcis-18462	83	10	,	,	PUNCT
fcis-18462	83	11	yolov5s	yolov5s	PROPN
fcis-18462	83	12	,	,	PUNCT
fcis-18462	83	13	yolov5	yolov5	PROPN
fcis-18462	83	14	m	m	PROPN
fcis-18462	83	15	,	,	PUNCT
fcis-18462	83	16	yolov5l	yolov5l	PROPN
fcis-18462	83	17	and	and	CCONJ
fcis-18462	83	18	yolo5x	yolo5x	PROPN
fcis-18462	83	19	.	.	PUNCT
fcis-18462	84	1	the	the	DET
fcis-18462	84	2	network	network	NOUN
fcis-18462	84	3	depth	depth	NOUN
fcis-18462	84	4	,	,	PUNCT
fcis-18462	84	5	width	width	ADJ
fcis-18462	84	6	,	,	PUNCT
fcis-18462	84	7	parameter	parameter	NOUN
fcis-18462	84	8	size	size	NOUN
fcis-18462	84	9	and	and	CCONJ
fcis-18462	84	10	number	number	NOUN
fcis-18462	84	11	of	of	ADP
fcis-18462	84	12	convolutional	convolutional	ADJ
fcis-18462	84	13	nuclei	nucleus	NOUN
fcis-18462	84	14	in	in	ADP
fcis-18462	84	15	different	different	ADJ
fcis-18462	84	16	stages	stage	NOUN
fcis-18462	84	17	are	be	AUX
fcis-18462	84	18	different	different	ADJ
fcis-18462	84	19	1.yolov5s	1.yolov5s	NOUN
fcis-18462	84	20	:	:	PUNCT
fcis-18462	84	21	this	this	PRON
fcis-18462	84	22	is	be	AUX
fcis-18462	84	23	the	the	DET
fcis-18462	84	24	smallest	small	ADJ
fcis-18462	84	25	model	model	NOUN
fcis-18462	84	26	in	in	ADP
fcis-18462	84	27	the	the	DET
fcis-18462	84	28	yolov5	yolov5	NOUN
fcis-18462	84	29	series	series	NOUN
fcis-18462	84	30	.	.	PUNCT
fcis-18462	85	1	"	"	PUNCT
fcis-18462	85	2	s	s	X
fcis-18462	85	3	"	"	PUNCT
fcis-18462	85	4	stands	stand	VERB
fcis-18462	85	5	for	for	ADP
fcis-18462	85	6	"	"	PUNCT
fcis-18462	85	7	small	small	ADJ
fcis-18462	85	8	"	"	PUNCT
fcis-18462	85	9	.	.	PUNCT
fcis-18462	86	1	the	the	DET
fcis-18462	86	2	model	model	NOUN
fcis-18462	86	3	performs	perform	VERB
fcis-18462	86	4	best	good	ADJ
fcis-18462	86	5	on	on	ADP
fcis-18462	86	6	devices	device	NOUN
fcis-18462	86	7	with	with	ADP
fcis-18462	86	8	limited	limited	ADJ
fcis-18462	86	9	computing	computing	NOUN
fcis-18462	86	10	resources	resource	NOUN
fcis-18462	86	11	,	,	PUNCT
fcis-18462	86	12	such	such	ADJ
fcis-18462	86	13	as	as	ADP
fcis-18462	86	14	mobile	mobile	ADJ
fcis-18462	86	15	devices	device	NOUN
fcis-18462	86	16	or	or	CCONJ
fcis-18462	86	17	edge	edge	NOUN
fcis-18462	86	18	devices	device	NOUN
fcis-18462	86	19	.	.	PUNCT
fcis-18462	87	1	yolov5s	yolov5s	NOUN
fcis-18462	87	2	has	have	VERB
fcis-18462	87	3	the	the	DET
fcis-18462	87	4	fastest	fast	ADJ
fcis-18462	87	5	detection	detection	NOUN
fcis-18462	87	6	,	,	PUNCT
fcis-18462	87	7	but	but	CCONJ
fcis-18462	87	8	the	the	DET
fcis-18462	87	9	accuracy	accuracy	NOUN
fcis-18462	87	10	is	be	AUX
fcis-18462	87	11	relatively	relatively	ADV
fcis-18462	87	12	low	low	ADJ
fcis-18462	87	13	.	.	PUNCT
fcis-18462	88	1	2.yolov5	2.yolov5	NUM
fcis-18462	88	2	m	m	NOUN
fcis-18462	88	3	:	:	PUNCT
fcis-18462	88	4	this	this	PRON
fcis-18462	88	5	is	be	AUX
fcis-18462	88	6	a	a	DET
fcis-18462	88	7	medium	medium	ADJ
fcis-18462	88	8	sized	sized	ADJ
fcis-18462	88	9	model	model	NOUN
fcis-18462	88	10	in	in	ADP
fcis-18462	88	11	the	the	DET
fcis-18462	88	12	yolov5	yolov5	NOUN
fcis-18462	88	13	series	series	NOUN
fcis-18462	88	14	.	.	PUNCT
fcis-18462	89	1	"	"	PUNCT
fcis-18462	89	2	m	m	X
fcis-18462	89	3	"	"	PUNCT
fcis-18462	89	4	stands	stand	VERB
fcis-18462	89	5	for	for	ADP
fcis-18462	89	6	"	"	PUNCT
fcis-18462	89	7	medium	medium	ADJ
fcis-18462	89	8	"	"	PUNCT
fcis-18462	89	9	.	.	PUNCT
fcis-18462	90	1	the	the	DET
fcis-18462	90	2	yolov5	yolov5	PROPN
fcis-18462	90	3	m	m	VERB
fcis-18462	90	4	offers	offer	VERB
fcis-18462	90	5	a	a	DET
fcis-18462	90	6	good	good	ADJ
fcis-18462	90	7	balance	balance	NOUN
fcis-18462	90	8	between	between	ADP
fcis-18462	90	9	speed	speed	NOUN
fcis-18462	90	10	and	and	CCONJ
fcis-18462	90	11	accuracy	accuracy	NOUN
fcis-18462	90	12	and	and	CCONJ
fcis-18462	90	13	is	be	AUX
fcis-18462	90	14	suitable	suitable	ADJ
fcis-18462	90	15	for	for	ADP
fcis-18462	90	16	devices	device	NOUN
fcis-18462	90	17	with	with	ADP
fcis-18462	90	18	a	a	DET
fcis-18462	90	19	certain	certain	ADJ
fcis-18462	90	20	amount	amount	NOUN
fcis-18462	90	21	of	of	ADP
fcis-18462	90	22	computing	computing	NOUN
fcis-18462	90	23	power	power	NOUN
fcis-18462	90	24	.	.	PUNCT
fcis-18462	91	1	3.yolov5l	3.yolov5l	NUM
fcis-18462	91	2	:	:	PUNCT
fcis-18462	91	3	this	this	PRON
fcis-18462	91	4	is	be	AUX
fcis-18462	91	5	a	a	DET
fcis-18462	91	6	larger	large	ADJ
fcis-18462	91	7	model	model	NOUN
fcis-18462	91	8	in	in	ADP
fcis-18462	91	9	the	the	DET
fcis-18462	91	10	yolov5	yolov5	NOUN
fcis-18462	91	11	series	series	NOUN
fcis-18462	91	12	.	.	PUNCT
fcis-18462	92	1	"	"	PUNCT
fcis-18462	92	2	l	l	NOUN
fcis-18462	92	3	"	"	PUNCT
fcis-18462	92	4	stands	stand	VERB
fcis-18462	92	5	for	for	ADP
fcis-18462	92	6	"	"	PUNCT
fcis-18462	92	7	large	large	ADJ
fcis-18462	92	8	"	"	PUNCT
fcis-18462	92	9	.	.	PUNCT
fcis-18462	93	1	yolov5l	yolov5l	PROPN
fcis-18462	93	2	is	be	AUX
fcis-18462	93	3	relatively	relatively	ADV
fcis-18462	93	4	accurate	accurate	ADJ
fcis-18462	93	5	,	,	PUNCT
fcis-18462	93	6	but	but	CCONJ
fcis-18462	93	7	the	the	DET
fcis-18462	93	8	78	78	NUM
fcis-18462	93	9	detection	detection	NOUN
fcis-18462	93	10	speed	speed	NOUN
fcis-18462	93	11	is	be	AUX
fcis-18462	93	12	slower	slow	ADJ
fcis-18462	93	13	.	.	PUNCT
fcis-18462	94	1	it	it	PRON
fcis-18462	94	2	is	be	AUX
fcis-18462	94	3	suitable	suitable	ADJ
fcis-18462	94	4	for	for	ADP
fcis-18462	94	5	devices	device	NOUN
fcis-18462	94	6	that	that	PRON
fcis-18462	94	7	require	require	VERB
fcis-18462	94	8	high	high	ADJ
fcis-18462	94	9	accuracy	accuracy	NOUN
fcis-18462	94	10	and	and	CCONJ
fcis-18462	94	11	have	have	VERB
fcis-18462	94	12	strong	strong	ADJ
fcis-18462	94	13	computing	computing	NOUN
fcis-18462	94	14	power	power	NOUN
fcis-18462	94	15	.	.	PUNCT
fcis-18462	95	1	4.yolov5x	4.yolov5x	NOUN
fcis-18462	95	2	:	:	PUNCT
fcis-18462	95	3	this	this	PRON
fcis-18462	95	4	is	be	AUX
fcis-18462	95	5	the	the	DET
fcis-18462	95	6	largest	large	ADJ
fcis-18462	95	7	model	model	NOUN
fcis-18462	95	8	in	in	ADP
fcis-18462	95	9	the	the	DET
fcis-18462	95	10	yolov5	yolov5	NOUN
fcis-18462	95	11	series	series	NOUN
fcis-18462	95	12	.	.	PUNCT
fcis-18462	96	1	"	"	PUNCT
fcis-18462	96	2	x	x	X
fcis-18462	96	3	"	"	PUNCT
fcis-18462	96	4	stands	stand	VERB
fcis-18462	96	5	for	for	ADP
fcis-18462	96	6	"	"	PUNCT
fcis-18462	96	7	extra	extra	ADV
fcis-18462	96	8	large	large	ADJ
fcis-18462	96	9	.	.	PUNCT
fcis-18462	96	10	"	"	PUNCT
fcis-18462	97	1	yolov5x	yolov5x	PROPN
fcis-18462	97	2	performed	perform	VERB
fcis-18462	97	3	best	well	ADV
fcis-18462	97	4	in	in	ADP
fcis-18462	97	5	terms	term	NOUN
fcis-18462	97	6	of	of	ADP
fcis-18462	97	7	accuracy	accuracy	NOUN
fcis-18462	97	8	,	,	PUNCT
fcis-18462	97	9	but	but	CCONJ
fcis-18462	97	10	had	have	VERB
fcis-18462	97	11	the	the	DET
fcis-18462	97	12	slowest	slow	ADJ
fcis-18462	97	13	detection	detection	NOUN
fcis-18462	97	14	speed	speed	NOUN
fcis-18462	97	15	.	.	PUNCT
fcis-18462	98	1	ideal	ideal	ADJ
fcis-18462	98	2	for	for	ADP
fcis-18462	98	3	tasks	task	NOUN
fcis-18462	98	4	requiring	require	VERB
fcis-18462	98	5	extremely	extremely	ADV
fcis-18462	98	6	high	high	ADJ
fcis-18462	98	7	accuracy	accuracy	NOUN
fcis-18462	98	8	and	and	CCONJ
fcis-18462	98	9	devices	device	NOUN
fcis-18462	98	10	with	with	ADP
fcis-18462	98	11	powerful	powerful	ADJ
fcis-18462	98	12	computing	computing	NOUN
fcis-18462	98	13	power	power	NOUN
fcis-18462	98	14	,	,	PUNCT
fcis-18462	98	15	such	such	ADJ
fcis-18462	98	16	as	as	ADP
fcis-18462	98	17	gpus	gpu	NOUN
fcis-18462	98	18	.	.	PUNCT
fcis-18462	99	1	5.yolov5n	5.yolov5n	ADV
fcis-18462	99	2	:	:	PUNCT
fcis-18462	99	3	this	this	PRON
fcis-18462	99	4	is	be	AUX
fcis-18462	99	5	a	a	DET
fcis-18462	99	6	variant	variant	NOUN
fcis-18462	99	7	in	in	ADP
fcis-18462	99	8	the	the	DET
fcis-18462	99	9	yolov5	yolov5	NOUN
fcis-18462	99	10	family	family	NOUN
fcis-18462	99	11	,	,	PUNCT
fcis-18462	99	12	optimized	optimize	VERB
fcis-18462	99	13	for	for	ADP
fcis-18462	99	14	nano	nano	NOUN
fcis-18462	99	15	devices	device	NOUN
fcis-18462	99	16	such	such	ADJ
fcis-18462	99	17	as	as	ADP
fcis-18462	99	18	the	the	DET
fcis-18462	99	19	nvidia	nvidia	PROPN
fcis-18462	99	20	jetson	jetson	PROPN
fcis-18462	99	21	nano	nano	PROPN
fcis-18462	99	22	.	.	PUNCT
fcis-18462	100	1	yolov5n	yolov5n	NOUN
fcis-18462	100	2	delivers	deliver	VERB
fcis-18462	100	3	accuracy	accuracy	NOUN
fcis-18462	100	4	for	for	ADP
fcis-18462	100	5	edge	edge	NOUN
fcis-18462	100	6	devices	device	NOUN
fcis-18462	100	7	while	while	SCONJ
fcis-18462	100	8	maintaining	maintain	VERB
fcis-18462	100	9	high	high	ADJ
fcis-18462	100	10	speeds	speed	NOUN
fcis-18462	100	11	.	.	PUNCT
fcis-18462	101	1	we	we	PRON
fcis-18462	101	2	used	use	VERB
fcis-18462	101	3	yolov	yolov	NOUN
fcis-18462	101	4	5s	5s	NUM
fcis-18462	101	5	in	in	ADP
fcis-18462	101	6	this	this	DET
fcis-18462	101	7	experiment	experiment	NOUN
fcis-18462	101	8	.	.	PUNCT
fcis-18462	102	1	first	first	ADV
fcis-18462	102	2	,	,	PUNCT
fcis-18462	102	3	we	we	PRON
fcis-18462	102	4	built	build	VERB
fcis-18462	102	5	a	a	DET
fcis-18462	102	6	vehicle	vehicle	NOUN
fcis-18462	102	7	model	model	NOUN
fcis-18462	102	8	and	and	CCONJ
fcis-18462	102	9	improved	improve	VERB
fcis-18462	102	10	it	it	PRON
fcis-18462	102	11	by	by	ADP
fcis-18462	102	12	using	use	VERB
fcis-18462	102	13	the	the	DET
fcis-18462	102	14	yolov5	yolov5	NOUN
fcis-18462	102	15	network	network	NOUN
fcis-18462	102	16	,	,	PUNCT
fcis-18462	102	17	which	which	PRON
fcis-18462	102	18	effectively	effectively	ADV
fcis-18462	102	19	alleviated	alleviate	VERB
fcis-18462	102	20	the	the	DET
fcis-18462	102	21	problem	problem	NOUN
fcis-18462	102	22	of	of	ADP
fcis-18462	102	23	target	target	NOUN
fcis-18462	102	24	loss	loss	NOUN
fcis-18462	102	25	and	and	CCONJ
fcis-18462	102	26	target	target	NOUN
fcis-18462	102	27	location	location	NOUN
fcis-18462	102	28	.	.	PUNCT
fcis-18462	103	1	yolov5	yolov5	NOUN
fcis-18462	103	2	is	be	AUX
fcis-18462	103	3	used	use	VERB
fcis-18462	103	4	to	to	PART
fcis-18462	103	5	predict	predict	VERB
fcis-18462	103	6	the	the	DET
fcis-18462	103	7	position	position	NOUN
fcis-18462	103	8	and	and	CCONJ
fcis-18462	103	9	category	category	NOUN
fcis-18462	103	10	of	of	ADP
fcis-18462	103	11	the	the	DET
fcis-18462	103	12	target	target	NOUN
fcis-18462	103	13	in	in	ADP
fcis-18462	103	14	the	the	DET
fcis-18462	103	15	image	image	NOUN
fcis-18462	103	16	,	,	PUNCT
fcis-18462	103	17	and	and	CCONJ
fcis-18462	103	18	the	the	DET
fcis-18462	103	19	prediction	prediction	NOUN
fcis-18462	103	20	results	result	NOUN
fcis-18462	103	21	are	be	AUX
fcis-18462	103	22	processed	process	VERB
fcis-18462	103	23	by	by	ADP
fcis-18462	103	24	yolov5	yolov5	NOUN
fcis-18462	103	25	.	.	PUNCT
fcis-18462	104	1	in	in	ADP
fcis-18462	104	2	the	the	DET
fcis-18462	104	3	process	process	NOUN
fcis-18462	104	4	of	of	ADP
fcis-18462	104	5	target	target	NOUN
fcis-18462	104	6	tracking	tracking	NOUN
fcis-18462	104	7	and	and	CCONJ
fcis-18462	104	8	detection	detection	NOUN
fcis-18462	104	9	,	,	PUNCT
fcis-18462	104	10	we	we	PRON
fcis-18462	104	11	conduct	conduct	VERB
fcis-18462	104	12	target	target	NOUN
fcis-18462	104	13	tracking	tracking	NOUN
fcis-18462	104	14	by	by	ADP
fcis-18462	104	15	studying	study	VERB
fcis-18462	104	16	the	the	DET
fcis-18462	104	17	process	process	NOUN
fcis-18462	104	18	of	of	ADP
fcis-18462	104	19	target	target	NOUN
fcis-18462	104	20	positioning	positioning	NOUN
fcis-18462	104	21	moving	move	VERB
fcis-18462	104	22	target	target	NOUN
fcis-18462	104	23	in	in	ADP
fcis-18462	104	24	video	video	NOUN
fcis-18462	104	25	over	over	ADP
fcis-18462	104	26	time	time	NOUN
fcis-18462	104	27	.	.	PUNCT
fcis-18462	105	1	in	in	ADP
fcis-18462	105	2	the	the	DET
fcis-18462	105	3	target	target	NOUN
fcis-18462	105	4	tracking	tracking	NOUN
fcis-18462	105	5	task	task	NOUN
fcis-18462	105	6	in	in	ADP
fcis-18462	105	7	the	the	DET
fcis-18462	105	8	real	real	ADJ
fcis-18462	105	9	scene	scene	NOUN
fcis-18462	105	10	,	,	PUNCT
fcis-18462	105	11	considering	consider	VERB
fcis-18462	105	12	the	the	DET
fcis-18462	105	13	real	real	ADJ
fcis-18462	105	14	-	-	PUNCT
fcis-18462	105	15	time	time	NOUN
fcis-18462	105	16	and	and	CCONJ
fcis-18462	105	17	flexibility	flexibility	NOUN
fcis-18462	105	18	of	of	ADP
fcis-18462	105	19	the	the	DET
fcis-18462	105	20	algorithm	algorithm	NOUN
fcis-18462	105	21	,	,	PUNCT
fcis-18462	105	22	we	we	PRON
fcis-18462	105	23	intend	intend	VERB
fcis-18462	105	24	to	to	PART
fcis-18462	105	25	adopt	adopt	VERB
fcis-18462	105	26	the	the	DET
fcis-18462	105	27	online	online	ADJ
fcis-18462	105	28	tracking	tracking	NOUN
fcis-18462	105	29	method	method	NOUN
fcis-18462	105	30	based	base	VERB
fcis-18462	105	31	on	on	ADP
fcis-18462	105	32	detection	detection	NOUN
fcis-18462	105	33	.	.	PUNCT
fcis-18462	106	1	multi	multi	ADJ
fcis-18462	106	2	-	-	ADJ
fcis-18462	106	3	target	target	ADJ
fcis-18462	106	4	tracking	tracking	NOUN
fcis-18462	106	5	technology	technology	NOUN
fcis-18462	106	6	(	(	PUNCT
fcis-18462	106	7	mot	mot	NOUN
fcis-18462	106	8	)	)	PUNCT
fcis-18462	106	9	is	be	AUX
fcis-18462	106	10	designed	design	VERB
fcis-18462	106	11	to	to	PART
fcis-18462	106	12	detect	detect	VERB
fcis-18462	106	13	and	and	CCONJ
fcis-18462	106	14	obtain	obtain	VERB
fcis-18462	106	15	the	the	DET
fcis-18462	106	16	position	position	NOUN
fcis-18462	106	17	of	of	ADP
fcis-18462	106	18	the	the	DET
fcis-18462	106	19	target	target	NOUN
fcis-18462	106	20	that	that	PRON
fcis-18462	106	21	is	be	AUX
fcis-18462	106	22	interested	interested	ADJ
fcis-18462	106	23	or	or	CCONJ
fcis-18462	106	24	wants	want	VERB
fcis-18462	106	25	to	to	PART
fcis-18462	106	26	track	track	VERB
fcis-18462	106	27	in	in	ADP
fcis-18462	106	28	the	the	DET
fcis-18462	106	29	video	video	NOUN
fcis-18462	106	30	frame	frame	NOUN
fcis-18462	106	31	,	,	PUNCT
fcis-18462	106	32	and	and	CCONJ
fcis-18462	106	33	assign	assign	VERB
fcis-18462	106	34	an	an	DET
fcis-18462	106	35	i	i	PROPN
fcis-18462	106	36	d	d	PROPN
fcis-18462	106	37	to	to	ADP
fcis-18462	106	38	each	each	DET
fcis-18462	106	39	target	target	NOUN
fcis-18462	106	40	,	,	PUNCT
fcis-18462	106	41	and	and	CCONJ
fcis-18462	106	42	keep	keep	VERB
fcis-18462	106	43	each	each	DET
fcis-18462	106	44	target	target	NOUN
fcis-18462	106	45	i	i	NOUN
fcis-18462	106	46	d	d	PROPN
fcis-18462	106	47	unchanged	unchanged	ADJ
fcis-18462	106	48	during	during	ADP
fcis-18462	106	49	the	the	DET
fcis-18462	106	50	process	process	NOUN
fcis-18462	106	51	of	of	ADP
fcis-18462	106	52	target	target	NOUN
fcis-18462	106	53	movement	movement	NOUN
fcis-18462	106	54	.	.	PUNCT
fcis-18462	107	1	after	after	ADP
fcis-18462	107	2	consulting	consult	VERB
fcis-18462	107	3	a	a	DET
fcis-18462	107	4	lot	lot	NOUN
fcis-18462	107	5	of	of	ADP
fcis-18462	107	6	data	datum	NOUN
fcis-18462	107	7	,	,	PUNCT
fcis-18462	107	8	we	we	PRON
fcis-18462	107	9	were	be	AUX
fcis-18462	107	10	able	able	ADJ
fcis-18462	107	11	to	to	PART
fcis-18462	107	12	reconstruct	reconstruct	VERB
fcis-18462	107	13	the	the	DET
fcis-18462	107	14	feature	feature	NOUN
fcis-18462	107	15	extraction	extraction	NOUN
fcis-18462	107	16	network	network	NOUN
fcis-18462	107	17	of	of	ADP
fcis-18462	107	18	yolov5	yolov5	NOUN
fcis-18462	107	19	using	use	VERB
fcis-18462	107	20	the	the	DET
fcis-18462	107	21	improved	improved	ADJ
fcis-18462	107	22	efficientnetv2	efficientnetv2	NOUN
fcis-18462	107	23	,	,	PUNCT
fcis-18462	107	24	and	and	CCONJ
fcis-18462	107	25	the	the	DET
fcis-18462	107	26	association	association	NOUN
fcis-18462	107	27	fusion	fusion	NOUN
fcis-18462	107	28	network	network	NOUN
fcis-18462	107	29	solved	solve	VERB
fcis-18462	107	30	the	the	DET
fcis-18462	107	31	competition	competition	NOUN
fcis-18462	107	32	problem	problem	NOUN
fcis-18462	107	33	between	between	ADP
fcis-18462	107	34	multi	multi	ADJ
fcis-18462	107	35	-	-	NOUN
fcis-18462	107	36	task	task	ADJ
fcis-18462	107	37	learning	learning	NOUN
fcis-18462	107	38	in	in	ADP
fcis-18462	107	39	jde	jde	NOUN
fcis-18462	107	40	algorithm	algorithm	NOUN
fcis-18462	107	41	,	,	PUNCT
fcis-18462	107	42	greatly	greatly	ADV
fcis-18462	107	43	reducing	reduce	VERB
fcis-18462	107	44	the	the	DET
fcis-18462	107	45	number	number	NOUN
fcis-18462	107	46	of	of	ADP
fcis-18462	107	47	parameters	parameter	NOUN
fcis-18462	107	48	and	and	CCONJ
fcis-18462	107	49	calculation	calculation	NOUN
fcis-18462	107	50	amount	amount	NOUN
fcis-18462	107	51	of	of	ADP
fcis-18462	107	52	multi	multi	ADJ
fcis-18462	107	53	-	-	ADJ
fcis-18462	107	54	target	target	ADJ
fcis-18462	107	55	tracking	tracking	NOUN
fcis-18462	107	56	algorithm	algorithm	NOUN
fcis-18462	107	57	,	,	PUNCT
fcis-18462	107	58	and	and	CCONJ
fcis-18462	107	59	better	well	ADV
fcis-18462	107	60	meeting	meet	VERB
fcis-18462	107	61	the	the	DET
fcis-18462	107	62	real	real	ADJ
fcis-18462	107	63	-	-	PUNCT
fcis-18462	107	64	time	time	NOUN
fcis-18462	107	65	requirements	requirement	NOUN
fcis-18462	107	66	of	of	ADP
fcis-18462	107	67	mobile	mobile	ADJ
fcis-18462	107	68	devices	device	NOUN
fcis-18462	107	69	.	.	PUNCT
fcis-18462	108	1	at	at	ADP
fcis-18462	108	2	the	the	DET
fcis-18462	108	3	same	same	ADJ
fcis-18462	108	4	time	time	NOUN
fcis-18462	108	5	,	,	PUNCT
fcis-18462	108	6	a	a	DET
fcis-18462	108	7	fast	fast	ADJ
fcis-18462	108	8	multi	multi	ADJ
fcis-18462	108	9	-	-	ADJ
fcis-18462	108	10	object	object	ADJ
fcis-18462	108	11	based	base	VERB
fcis-18462	108	12	algorithm	algorithm	NOUN
fcis-18462	108	13	based	base	VERB
fcis-18462	108	14	on	on	ADP
fcis-18462	108	15	semi	semi	ADJ
fcis-18462	108	16	-	-	ADJ
fcis-18462	108	17	suppressed	suppressed	ADJ
fcis-18462	108	18	fuzzy	fuzzy	ADJ
fcis-18462	108	19	clustering	clustering	NOUN
fcis-18462	108	20	proposed	propose	VERB
fcis-18462	108	21	by	by	ADP
fcis-18462	108	22	zhang	zhang	PROPN
fcis-18462	108	23	junhui	junhui	PROPN
fcis-18462	108	24	and	and	CCONJ
fcis-18462	108	25	others	other	NOUN
fcis-18462	108	26	is	be	AUX
fcis-18462	108	27	used	use	VERB
fcis-18462	108	28	to	to	PART
fcis-18462	108	29	greatly	greatly	ADV
fcis-18462	108	30	improve	improve	VERB
fcis-18462	108	31	the	the	DET
fcis-18462	108	32	real	real	ADJ
fcis-18462	108	33	-	-	PUNCT
fcis-18462	108	34	time	time	NOUN
fcis-18462	108	35	and	and	CCONJ
fcis-18462	108	36	reliability	reliability	NOUN
fcis-18462	108	37	of	of	ADP
fcis-18462	108	38	multi	multi	ADJ
fcis-18462	108	39	-	-	ADJ
fcis-18462	108	40	object	object	ADJ
fcis-18462	108	41	data	datum	NOUN
fcis-18462	108	42	association	association	NOUN
fcis-18462	108	43	in	in	ADP
fcis-18462	108	44	complex	complex	ADJ
fcis-18462	108	45	environments	environment	NOUN
fcis-18462	108	46	.	.	PUNCT
fcis-18462	109	1	for	for	ADP
fcis-18462	109	2	image	image	NOUN
fcis-18462	109	3	segmentation	segmentation	NOUN
fcis-18462	109	4	,	,	PUNCT
fcis-18462	109	5	a	a	DET
fcis-18462	109	6	given	give	VERB
fcis-18462	109	7	image	image	NOUN
fcis-18462	109	8	data	datum	NOUN
fcis-18462	109	9	is	be	AUX
fcis-18462	109	10	treated	treat	VERB
fcis-18462	109	11	as	as	ADP
fcis-18462	109	12	a	a	DET
fcis-18462	109	13	collection	collection	NOUN
fcis-18462	109	14	of	of	ADP
fcis-18462	109	15	many	many	ADJ
fcis-18462	109	16	fuzzy	fuzzy	ADJ
fcis-18462	109	17	elements	element	NOUN
fcis-18462	109	18	belonging	belong	VERB
fcis-18462	109	19	to	to	ADP
fcis-18462	109	20	different	different	ADJ
fcis-18462	109	21	clusters	cluster	NOUN
fcis-18462	109	22	.	.	PUNCT
fcis-18462	110	1	the	the	DET
fcis-18462	110	2	membership	membership	NOUN
fcis-18462	110	3	degree	degree	NOUN
fcis-18462	110	4	of	of	ADP
fcis-18462	110	5	each	each	DET
fcis-18462	110	6	pixel	pixel	NOUN
fcis-18462	110	7	is	be	AUX
fcis-18462	110	8	blurred	blur	VERB
fcis-18462	110	9	to	to	ADP
fcis-18462	110	10	each	each	DET
fcis-18462	110	11	cluster	cluster	NOUN
fcis-18462	110	12	,	,	PUNCT
fcis-18462	110	13	and	and	CCONJ
fcis-18462	110	14	the	the	DET
fcis-18462	110	15	image	image	NOUN
fcis-18462	110	16	segmentation	segmentation	NOUN
fcis-18462	110	17	algorithm	algorithm	NOUN
fcis-18462	110	18	based	base	VERB
fcis-18462	110	19	on	on	ADP
fcis-18462	110	20	otsu	otsu	NOUN
fcis-18462	110	21	-	-	PUNCT
fcis-18462	110	22	ifcm	ifcm	PROPN
fcis-18462	110	23	is	be	AUX
fcis-18462	110	24	used	use	VERB
fcis-18462	110	25	.	.	PUNCT
fcis-18462	111	1	the	the	DET
fcis-18462	111	2	algorithm	algorithm	NOUN
fcis-18462	111	3	first	first	ADV
fcis-18462	111	4	utilizes	utilize	VERB
fcis-18462	111	5	the	the	DET
fcis-18462	111	6	otsu	otsu	ADJ
fcis-18462	111	7	algorithm	algorithm	NOUN
fcis-18462	111	8	(	(	PUNCT
fcis-18462	111	9	otsu	otsu	ADJ
fcis-18462	111	10	algorithm	algorithm	NOUN
fcis-18462	111	11	is	be	AUX
fcis-18462	111	12	an	an	DET
fcis-18462	111	13	adaptive	adaptive	ADJ
fcis-18462	111	14	threshold	threshold	NOUN
fcis-18462	111	15	segmentation	segmentation	NOUN
fcis-18462	111	16	algorithm	algorithm	NOUN
fcis-18462	111	17	based	base	VERB
fcis-18462	111	18	on	on	ADP
fcis-18462	111	19	image	image	NOUN
fcis-18462	111	20	gray	gray	ADJ
fcis-18462	111	21	histogram	histogram	NOUN
fcis-18462	111	22	.	.	PUNCT
fcis-18462	112	1	it	it	PRON
fcis-18462	112	2	determines	determine	VERB
fcis-18462	112	3	an	an	DET
fcis-18462	112	4	optimal	optimal	ADJ
fcis-18462	112	5	threshold	threshold	NOUN
fcis-18462	112	6	by	by	ADP
fcis-18462	112	7	maximizing	maximize	VERB
fcis-18462	112	8	the	the	DET
fcis-18462	112	9	interclass	interclass	NOUN
fcis-18462	112	10	variance	variance	NOUN
fcis-18462	112	11	and	and	CCONJ
fcis-18462	112	12	minimizing	minimize	VERB
fcis-18462	112	13	the	the	DET
fcis-18462	112	14	intra	intra	ADJ
fcis-18462	112	15	-	-	ADJ
fcis-18462	112	16	class	class	ADJ
fcis-18462	112	17	variance	variance	NOUN
fcis-18462	112	18	,	,	PUNCT
fcis-18462	112	19	which	which	PRON
fcis-18462	112	20	realizes	realize	VERB
fcis-18462	112	21	the	the	DET
fcis-18462	112	22	purpose	purpose	NOUN
fcis-18462	112	23	of	of	ADP
fcis-18462	112	24	automatic	automatic	ADJ
fcis-18462	112	25	image	image	NOUN
fcis-18462	112	26	segmentation	segmentation	NOUN
fcis-18462	112	27	.	.	PUNCT
fcis-18462	113	1	the	the	DET
fcis-18462	113	2	basic	basic	ADJ
fcis-18462	113	3	principle	principle	NOUN
fcis-18462	113	4	of	of	ADP
fcis-18462	113	5	otsu	otsu	ADJ
fcis-18462	113	6	algorithm	algorithm	NOUN
fcis-18462	113	7	is	be	AUX
fcis-18462	113	8	to	to	PART
fcis-18462	113	9	select	select	VERB
fcis-18462	113	10	the	the	DET
fcis-18462	113	11	optimal	optimal	ADJ
fcis-18462	113	12	threshold	threshold	NOUN
fcis-18462	113	13	by	by	ADP
fcis-18462	113	14	the	the	DET
fcis-18462	113	15	maximum	maximum	ADJ
fcis-18462	113	16	inter	inter	ADJ
fcis-18462	113	17	-	-	ADJ
fcis-18462	113	18	class	class	ADJ
fcis-18462	113	19	variance	variance	NOUN
fcis-18462	113	20	.	.	PUNCT
fcis-18462	114	1	in	in	ADP
fcis-18462	114	2	the	the	DET
fcis-18462	114	3	process	process	NOUN
fcis-18462	114	4	of	of	ADP
fcis-18462	114	5	image	image	NOUN
fcis-18462	114	6	segmentation	segmentation	NOUN
fcis-18462	114	7	,	,	PUNCT
fcis-18462	114	8	we	we	PRON
fcis-18462	114	9	want	want	VERB
fcis-18462	114	10	the	the	DET
fcis-18462	114	11	difference	difference	NOUN
fcis-18462	114	12	between	between	ADP
fcis-18462	114	13	the	the	DET
fcis-18462	114	14	two	two	NUM
fcis-18462	114	15	categories	category	NOUN
fcis-18462	114	16	after	after	SCONJ
fcis-18462	114	17	segmentation	segmentation	NOUN
fcis-18462	114	18	to	to	PART
fcis-18462	114	19	be	be	AUX
fcis-18462	114	20	as	as	ADV
fcis-18462	114	21	large	large	ADJ
fcis-18462	114	22	as	as	ADP
fcis-18462	114	23	possible	possible	ADJ
fcis-18462	114	24	,	,	PUNCT
fcis-18462	114	25	that	that	ADV
fcis-18462	114	26	is	is	ADV
fcis-18462	114	27	,	,	PUNCT
fcis-18462	114	28	to	to	PART
fcis-18462	114	29	maximize	maximize	VERB
fcis-18462	114	30	the	the	DET
fcis-18462	114	31	inter	inter	ADJ
fcis-18462	114	32	-	-	ADJ
fcis-18462	114	33	class	class	ADJ
fcis-18462	114	34	variance	variance	NOUN
fcis-18462	114	35	.	.	PUNCT
fcis-18462	115	1	otsu	otsu	ADJ
fcis-18462	115	2	algorithm	algorithm	PROPN
fcis-18462	115	3	calculates	calculate	VERB
fcis-18462	115	4	the	the	DET
fcis-18462	115	5	inter	inter	ADJ
fcis-18462	115	6	-	-	ADJ
fcis-18462	115	7	class	class	ADJ
fcis-18462	115	8	variance	variance	NOUN
fcis-18462	115	9	under	under	ADP
fcis-18462	115	10	different	different	ADJ
fcis-18462	115	11	thresholds	threshold	NOUN
fcis-18462	115	12	and	and	CCONJ
fcis-18462	115	13	selects	select	VERB
fcis-18462	115	14	the	the	DET
fcis-18462	115	15	threshold	threshold	NOUN
fcis-18462	115	16	with	with	ADP
fcis-18462	115	17	the	the	DET
fcis-18462	115	18	largest	large	ADJ
fcis-18462	115	19	inter	inter	ADJ
fcis-18462	115	20	-	-	ADJ
fcis-18462	115	21	class	class	ADJ
fcis-18462	115	22	variance	variance	NOUN
fcis-18462	115	23	as	as	ADP
fcis-18462	115	24	the	the	DET
fcis-18462	115	25	best	good	ADJ
fcis-18462	115	26	threshold	threshold	NOUN
fcis-18462	115	27	.	.	PUNCT
fcis-18462	116	1	to	to	PART
fcis-18462	116	2	find	find	VERB
fcis-18462	116	3	the	the	DET
fcis-18462	116	4	optimal	optimal	ADJ
fcis-18462	116	5	segmentation	segmentation	NOUN
fcis-18462	116	6	threshold	threshold	NOUN
fcis-18462	116	7	of	of	ADP
fcis-18462	116	8	image	image	NOUN
fcis-18462	116	9	foreground	foreground	NOUN
fcis-18462	116	10	and	and	CCONJ
fcis-18462	116	11	background	background	NOUN
fcis-18462	116	12	,	,	PUNCT
fcis-18462	116	13	an	an	DET
fcis-18462	116	14	image	image	NOUN
fcis-18462	116	15	segmentation	segmentation	NOUN
fcis-18462	116	16	algorithm	algorithm	NOUN
fcis-18462	116	17	based	base	VERB
fcis-18462	116	18	on	on	ADP
fcis-18462	116	19	otsu	otsu	NOUN
fcis-18462	116	20	-	-	PUNCT
fcis-18462	116	21	ifcm	ifcm	PROPN
fcis-18462	116	22	is	be	AUX
fcis-18462	116	23	used	use	VERB
fcis-18462	116	24	.	.	PUNCT
fcis-18462	117	1	the	the	DET
fcis-18462	117	2	algorithm	algorithm	NOUN
fcis-18462	117	3	firstly	firstly	ADV
fcis-18462	117	4	uses	use	VERB
fcis-18462	117	5	otsu	otsu	ADJ
fcis-18462	117	6	algorithm	algorithm	NOUN
fcis-18462	117	7	(	(	PUNCT
fcis-18462	117	8	otsu	otsu	ADJ
fcis-18462	117	9	algorithm	algorithm	NOUN
fcis-18462	117	10	is	be	AUX
fcis-18462	117	11	an	an	DET
fcis-18462	117	12	adaptive	adaptive	ADJ
fcis-18462	117	13	threshold	threshold	NOUN
fcis-18462	117	14	segmentation	segmentation	NOUN
fcis-18462	117	15	algorithm	algorithm	NOUN
fcis-18462	117	16	based	base	VERB
fcis-18462	117	17	on	on	ADP
fcis-18462	117	18	image	image	NOUN
fcis-18462	117	19	gray	gray	ADJ
fcis-18462	117	20	histogram	histogram	NOUN
fcis-18462	117	21	.	.	PUNCT
fcis-18462	118	1	it	it	PRON
fcis-18462	118	2	determines	determine	VERB
fcis-18462	118	3	an	an	DET
fcis-18462	118	4	optimal	optimal	ADJ
fcis-18462	118	5	threshold	threshold	NOUN
fcis-18462	118	6	by	by	ADP
fcis-18462	118	7	maximizing	maximize	VERB
fcis-18462	118	8	the	the	DET
fcis-18462	118	9	inter	inter	ADJ
fcis-18462	118	10	-	-	ADJ
fcis-18462	118	11	class	class	ADJ
fcis-18462	118	12	variance	variance	NOUN
fcis-18462	118	13	and	and	CCONJ
fcis-18462	118	14	minimizing	minimize	VERB
fcis-18462	118	15	the	the	DET
fcis-18462	118	16	intraclass	intraclass	NOUN
fcis-18462	118	17	variance	variance	NOUN
fcis-18462	118	18	,	,	PUNCT
fcis-18462	118	19	which	which	PRON
fcis-18462	118	20	realizes	realize	VERB
fcis-18462	118	21	the	the	DET
fcis-18462	118	22	purpose	purpose	NOUN
fcis-18462	118	23	of	of	ADP
fcis-18462	118	24	automatic	automatic	ADJ
fcis-18462	118	25	image	image	NOUN
fcis-18462	118	26	segmentation	segmentation	NOUN
fcis-18462	118	27	.	.	PUNCT
fcis-18462	119	1	the	the	DET
fcis-18462	119	2	basic	basic	ADJ
fcis-18462	119	3	principle	principle	NOUN
fcis-18462	119	4	of	of	ADP
fcis-18462	119	5	otsu	otsu	ADJ
fcis-18462	119	6	algorithm	algorithm	NOUN
fcis-18462	119	7	is	be	AUX
fcis-18462	119	8	to	to	PART
fcis-18462	119	9	select	select	VERB
fcis-18462	119	10	the	the	DET
fcis-18462	119	11	optimal	optimal	ADJ
fcis-18462	119	12	threshold	threshold	NOUN
fcis-18462	119	13	by	by	ADP
fcis-18462	119	14	the	the	DET
fcis-18462	119	15	maximum	maximum	ADJ
fcis-18462	119	16	inter	inter	ADJ
fcis-18462	119	17	-	-	ADJ
fcis-18462	119	18	class	class	ADJ
fcis-18462	119	19	variance	variance	NOUN
fcis-18462	119	20	.	.	PUNCT
fcis-18462	120	1	in	in	ADP
fcis-18462	120	2	the	the	DET
fcis-18462	120	3	process	process	NOUN
fcis-18462	120	4	of	of	ADP
fcis-18462	120	5	image	image	NOUN
fcis-18462	120	6	segmentation	segmentation	NOUN
fcis-18462	120	7	,	,	PUNCT
fcis-18462	120	8	we	we	PRON
fcis-18462	120	9	want	want	VERB
fcis-18462	120	10	the	the	DET
fcis-18462	120	11	difference	difference	NOUN
fcis-18462	120	12	between	between	ADP
fcis-18462	120	13	the	the	DET
fcis-18462	120	14	two	two	NUM
fcis-18462	120	15	categories	category	NOUN
fcis-18462	120	16	after	after	SCONJ
fcis-18462	120	17	segmentation	segmentation	NOUN
fcis-18462	120	18	to	to	PART
fcis-18462	120	19	be	be	AUX
fcis-18462	120	20	as	as	ADV
fcis-18462	120	21	large	large	ADJ
fcis-18462	120	22	as	as	ADP
fcis-18462	120	23	possible	possible	ADJ
fcis-18462	120	24	,	,	PUNCT
fcis-18462	120	25	that	that	ADV
fcis-18462	120	26	is	is	ADV
fcis-18462	120	27	,	,	PUNCT
fcis-18462	120	28	to	to	PART
fcis-18462	120	29	maximize	maximize	VERB
fcis-18462	120	30	the	the	DET
fcis-18462	120	31	inter	inter	ADJ
fcis-18462	120	32	-	-	ADJ
fcis-18462	120	33	class	class	ADJ
fcis-18462	120	34	variance	variance	NOUN
fcis-18462	120	35	.	.	PUNCT
fcis-18462	121	1	otsu	otsu	ADJ
fcis-18462	121	2	algorithm	algorithm	PROPN
fcis-18462	121	3	calculates	calculate	VERB
fcis-18462	121	4	the	the	DET
fcis-18462	121	5	inter	inter	ADJ
fcis-18462	121	6	-	-	ADJ
fcis-18462	121	7	class	class	ADJ
fcis-18462	121	8	variance	variance	NOUN
fcis-18462	121	9	under	under	ADP
fcis-18462	121	10	different	different	ADJ
fcis-18462	121	11	thresholds	threshold	NOUN
fcis-18462	121	12	and	and	CCONJ
fcis-18462	121	13	selects	select	VERB
fcis-18462	121	14	the	the	DET
fcis-18462	121	15	threshold	threshold	NOUN
fcis-18462	121	16	with	with	ADP
fcis-18462	121	17	the	the	DET
fcis-18462	121	18	largest	large	ADJ
fcis-18462	121	19	inter	inter	ADJ
fcis-18462	121	20	-	-	ADJ
fcis-18462	121	21	class	class	ADJ
fcis-18462	121	22	variance	variance	NOUN
fcis-18462	121	23	as	as	ADP
fcis-18462	121	24	the	the	DET
fcis-18462	121	25	best	good	ADJ
fcis-18462	121	26	threshold	threshold	NOUN
fcis-18462	121	27	.	.	PUNCT
fcis-18462	122	1	the	the	DET
fcis-18462	122	2	optimal	optimal	ADJ
fcis-18462	122	3	segmentation	segmentation	NOUN
fcis-18462	122	4	threshold	threshold	NOUN
fcis-18462	122	5	of	of	ADP
fcis-18462	122	6	foreground	foreground	NOUN
fcis-18462	122	7	and	and	CCONJ
fcis-18462	122	8	background	background	NOUN
fcis-18462	122	9	of	of	ADP
fcis-18462	122	10	image	image	NOUN
fcis-18462	122	11	is	be	AUX
fcis-18462	122	12	found	find	VERB
fcis-18462	122	13	.	.	PUNCT
fcis-18462	123	1	by	by	ADP
fcis-18462	123	2	studying	study	VERB
fcis-18462	123	3	the	the	DET
fcis-18462	123	4	positioning	positioning	NOUN
fcis-18462	123	5	process	process	NOUN
fcis-18462	123	6	of	of	ADP
fcis-18462	123	7	moving	move	VERB
fcis-18462	123	8	targets	target	NOUN
fcis-18462	123	9	in	in	ADP
fcis-18462	123	10	video	video	NOUN
fcis-18462	123	11	over	over	ADP
fcis-18462	123	12	time	time	NOUN
fcis-18462	123	13	,	,	PUNCT
fcis-18462	123	14	the	the	DET
fcis-18462	123	15	detection	detection	NOUN
fcis-18462	123	16	-	-	PUNCT
fcis-18462	123	17	based	base	VERB
fcis-18462	123	18	online	online	ADJ
fcis-18462	123	19	tracking	tracking	NOUN
fcis-18462	123	20	method	method	NOUN
fcis-18462	123	21	is	be	AUX
fcis-18462	123	22	adopted	adopt	VERB
fcis-18462	123	23	in	in	ADP
fcis-18462	123	24	the	the	DET
fcis-18462	123	25	target	target	NOUN
fcis-18462	123	26	tracking	tracking	NOUN
fcis-18462	123	27	task	task	NOUN
fcis-18462	123	28	of	of	ADP
fcis-18462	123	29	real	real	ADJ
fcis-18462	123	30	scene	scene	NOUN
fcis-18462	123	31	,	,	PUNCT
fcis-18462	123	32	considering	consider	VERB
fcis-18462	123	33	the	the	DET
fcis-18462	123	34	real	real	ADJ
fcis-18462	123	35	-	-	PUNCT
fcis-18462	123	36	time	time	NOUN
fcis-18462	123	37	and	and	CCONJ
fcis-18462	123	38	flexibility	flexibility	NOUN
fcis-18462	123	39	of	of	ADP
fcis-18462	123	40	the	the	DET
fcis-18462	123	41	algorithm	algorithm	NOUN
fcis-18462	123	42	.	.	PUNCT
fcis-18462	124	1	then	then	ADV
fcis-18462	124	2	the	the	DET
fcis-18462	124	3	ifcm	ifcm	NOUN
fcis-18462	124	4	algorithm	algorithm	NOUN
fcis-18462	124	5	is	be	AUX
fcis-18462	124	6	used	use	VERB
fcis-18462	124	7	,	,	PUNCT
fcis-18462	124	8	which	which	PRON
fcis-18462	124	9	is	be	AUX
fcis-18462	124	10	a	a	DET
fcis-18462	124	11	variant	variant	NOUN
fcis-18462	124	12	of	of	ADP
fcis-18462	124	13	the	the	DET
fcis-18462	124	14	intuitive	intuitive	ADJ
fcis-18462	124	15	fuzzy	fuzzy	ADJ
fcis-18462	124	16	c	c	NOUN
fcis-18462	124	17	-	-	PUNCT
fcis-18462	124	18	means	mean	VERB
fcis-18462	124	19	clustering	clustering	ADJ
fcis-18462	124	20	algorithm	algorithm	NOUN
fcis-18462	124	21	.	.	PUNCT
fcis-18462	125	1	the	the	DET
fcis-18462	125	2	basic	basic	ADJ
fcis-18462	125	3	idea	idea	NOUN
fcis-18462	125	4	is	be	AUX
fcis-18462	125	5	to	to	PART
fcis-18462	125	6	treat	treat	VERB
fcis-18462	125	7	a	a	DET
fcis-18462	125	8	given	give	VERB
fcis-18462	125	9	image	image	NOUN
fcis-18462	125	10	data	datum	NOUN
fcis-18462	125	11	as	as	ADP
fcis-18462	125	12	a	a	DET
fcis-18462	125	13	set	set	NOUN
fcis-18462	125	14	of	of	ADP
fcis-18462	125	15	fuzzy	fuzzy	ADJ
fcis-18462	125	16	elements	element	NOUN
fcis-18462	125	17	belonging	belong	VERB
fcis-18462	125	18	to	to	ADP
fcis-18462	125	19	different	different	ADJ
fcis-18462	125	20	clusters	cluster	NOUN
fcis-18462	125	21	.	.	PUNCT
fcis-18462	126	1	by	by	ADP
fcis-18462	126	2	blurring	blur	VERB
fcis-18462	126	3	the	the	DET
fcis-18462	126	4	membership	membership	NOUN
fcis-18462	126	5	degree	degree	NOUN
fcis-18462	126	6	of	of	ADP
fcis-18462	126	7	each	each	DET
fcis-18462	126	8	pixel	pixel	NOUN
fcis-18462	126	9	to	to	ADP
fcis-18462	126	10	each	each	DET
fcis-18462	126	11	cluster	cluster	NOUN
fcis-18462	126	12	,	,	PUNCT
fcis-18462	126	13	and	and	CCONJ
fcis-18462	126	14	then	then	ADV
fcis-18462	126	15	iteratively	iteratively	ADV
fcis-18462	126	16	adjusting	adjust	VERB
fcis-18462	126	17	the	the	DET
fcis-18462	126	18	cluster	cluster	NOUN
fcis-18462	126	19	center	center	NOUN
fcis-18462	126	20	and	and	CCONJ
fcis-18462	126	21	membership	membership	NOUN
fcis-18462	126	22	degree	degree	NOUN
fcis-18462	126	23	according	accord	VERB
fcis-18462	126	24	to	to	ADP
fcis-18462	126	25	the	the	DET
fcis-18462	126	26	update	update	NOUN
fcis-18462	126	27	formula	formula	NOUN
fcis-18462	126	28	,	,	PUNCT
fcis-18462	126	29	the	the	DET
fcis-18462	126	30	image	image	NOUN
fcis-18462	126	31	segmentation	segmentation	NOUN
fcis-18462	126	32	is	be	AUX
fcis-18462	126	33	finally	finally	ADV
fcis-18462	126	34	realized	realize	VERB
fcis-18462	126	35	.	.	PUNCT
fcis-18462	127	1	multi	multi	ADJ
fcis-18462	127	2	-	-	ADJ
fcis-18462	127	3	target	target	ADJ
fcis-18462	127	4	tracking	tracking	NOUN
fcis-18462	127	5	technology	technology	NOUN
fcis-18462	127	6	(	(	PUNCT
fcis-18462	127	7	mot	mot	NOUN
fcis-18462	127	8	)	)	PUNCT
fcis-18462	127	9	is	be	AUX
fcis-18462	127	10	designed	design	VERB
fcis-18462	127	11	to	to	PART
fcis-18462	127	12	detect	detect	VERB
fcis-18462	127	13	and	and	CCONJ
fcis-18462	127	14	obtain	obtain	VERB
fcis-18462	127	15	the	the	DET
fcis-18462	127	16	position	position	NOUN
fcis-18462	127	17	of	of	ADP
fcis-18462	127	18	the	the	DET
fcis-18462	127	19	target	target	NOUN
fcis-18462	127	20	that	that	PRON
fcis-18462	127	21	is	be	AUX
fcis-18462	127	22	interested	interested	ADJ
fcis-18462	127	23	or	or	CCONJ
fcis-18462	127	24	wants	want	VERB
fcis-18462	127	25	to	to	PART
fcis-18462	127	26	track	track	VERB
fcis-18462	127	27	in	in	ADP
fcis-18462	127	28	the	the	DET
fcis-18462	127	29	video	video	NOUN
fcis-18462	127	30	frame	frame	NOUN
fcis-18462	127	31	,	,	PUNCT
fcis-18462	127	32	and	and	CCONJ
fcis-18462	127	33	assign	assign	VERB
fcis-18462	127	34	an	an	DET
fcis-18462	127	35	i	i	PROPN
fcis-18462	127	36	d	d	PROPN
fcis-18462	127	37	to	to	ADP
fcis-18462	127	38	each	each	DET
fcis-18462	127	39	target	target	NOUN
fcis-18462	127	40	,	,	PUNCT
fcis-18462	127	41	and	and	CCONJ
fcis-18462	127	42	keep	keep	VERB
fcis-18462	127	43	each	each	DET
fcis-18462	127	44	target	target	NOUN
fcis-18462	127	45	i	i	NOUN
fcis-18462	127	46	d	d	PROPN
fcis-18462	127	47	unchanged	unchanged	ADJ
fcis-18462	127	48	during	during	ADP
fcis-18462	127	49	the	the	DET
fcis-18462	127	50	process	process	NOUN
fcis-18462	127	51	of	of	ADP
fcis-18462	127	52	target	target	NOUN
fcis-18462	127	53	movement	movement	NOUN
fcis-18462	127	54	.	.	PUNCT
fcis-18462	128	1	after	after	ADP
fcis-18462	128	2	consulting	consult	VERB
fcis-18462	128	3	a	a	DET
fcis-18462	128	4	lot	lot	NOUN
fcis-18462	128	5	of	of	ADP
fcis-18462	128	6	data	datum	NOUN
fcis-18462	128	7	,	,	PUNCT
fcis-18462	128	8	we	we	PRON
fcis-18462	128	9	were	be	AUX
fcis-18462	128	10	able	able	ADJ
fcis-18462	128	11	to	to	PART
fcis-18462	128	12	reconstruct	reconstruct	VERB
fcis-18462	128	13	the	the	DET
fcis-18462	128	14	feature	feature	NOUN
fcis-18462	128	15	extraction	extraction	NOUN
fcis-18462	128	16	network	network	NOUN
fcis-18462	128	17	of	of	ADP
fcis-18462	128	18	yolov5	yolov5	NOUN
fcis-18462	128	19	using	use	VERB
fcis-18462	128	20	the	the	DET
fcis-18462	128	21	improved	improved	ADJ
fcis-18462	128	22	efficientnetv2	efficientnetv2	NOUN
fcis-18462	128	23	,	,	PUNCT
fcis-18462	128	24	and	and	CCONJ
fcis-18462	128	25	the	the	DET
fcis-18462	128	26	association	association	NOUN
fcis-18462	128	27	fusion	fusion	NOUN
fcis-18462	128	28	network	network	NOUN
fcis-18462	128	29	solved	solve	VERB
fcis-18462	128	30	the	the	DET
fcis-18462	128	31	competition	competition	NOUN
fcis-18462	128	32	problem	problem	NOUN
fcis-18462	128	33	between	between	ADP
fcis-18462	128	34	multi	multi	ADJ
fcis-18462	128	35	-	-	NOUN
fcis-18462	128	36	task	task	ADJ
fcis-18462	128	37	learning	learning	NOUN
fcis-18462	128	38	in	in	ADP
fcis-18462	128	39	jde	jde	NOUN
fcis-18462	128	40	algorithm	algorithm	NOUN
fcis-18462	128	41	,	,	PUNCT
fcis-18462	128	42	greatly	greatly	ADV
fcis-18462	128	43	reducing	reduce	VERB
fcis-18462	128	44	the	the	DET
fcis-18462	128	45	number	number	NOUN
fcis-18462	128	46	of	of	ADP
fcis-18462	128	47	parameters	parameter	NOUN
fcis-18462	128	48	and	and	CCONJ
fcis-18462	128	49	calculation	calculation	NOUN
fcis-18462	128	50	amount	amount	NOUN
fcis-18462	128	51	of	of	ADP
fcis-18462	128	52	multi	multi	ADJ
fcis-18462	128	53	-	-	ADJ
fcis-18462	128	54	target	target	ADJ
fcis-18462	128	55	tracking	tracking	NOUN
fcis-18462	128	56	algorithm	algorithm	NOUN
fcis-18462	128	57	,	,	PUNCT
fcis-18462	128	58	and	and	CCONJ
fcis-18462	128	59	better	well	ADV
fcis-18462	128	60	meeting	meet	VERB
fcis-18462	128	61	the	the	DET
fcis-18462	128	62	real	real	ADJ
fcis-18462	128	63	-	-	PUNCT
fcis-18462	128	64	time	time	NOUN
fcis-18462	128	65	requirements	requirement	NOUN
fcis-18462	128	66	of	of	ADP
fcis-18462	128	67	mobile	mobile	ADJ
fcis-18462	128	68	devices	device	NOUN
fcis-18462	128	69	.	.	PUNCT
fcis-18462	129	1	at	at	ADP
fcis-18462	129	2	the	the	DET
fcis-18462	129	3	same	same	ADJ
fcis-18462	129	4	time	time	NOUN
fcis-18462	129	5	,	,	PUNCT
fcis-18462	129	6	a	a	DET
fcis-18462	129	7	fast	fast	ADJ
fcis-18462	129	8	multi	multi	ADJ
fcis-18462	129	9	-	-	ADJ
fcis-18462	129	10	object	object	ADJ
fcis-18462	129	11	basis	basis	NOUN
fcis-18462	129	12	algorithm	algorithm	NOUN
fcis-18462	129	13	based	base	VERB
fcis-18462	129	14	on	on	ADP
fcis-18462	129	15	semi	semi	ADJ
fcis-18462	129	16	-	-	ADJ
fcis-18462	129	17	suppressed	suppressed	ADJ
fcis-18462	129	18	fuzzy	fuzzy	ADJ
fcis-18462	129	19	clustering	clustering	NOUN
fcis-18462	129	20	proposed	propose	VERB
fcis-18462	129	21	by	by	ADP
fcis-18462	129	22	zhang	zhang	PROPN
fcis-18462	129	23	junhui	junhui	PROPN
fcis-18462	129	24	et	et	PROPN
fcis-18462	129	25	al	al	PROPN
fcis-18462	129	26	is	be	AUX
fcis-18462	129	27	used	use	VERB
fcis-18462	129	28	to	to	PART
fcis-18462	129	29	greatly	greatly	ADV
fcis-18462	129	30	improve	improve	VERB
fcis-18462	129	31	the	the	DET
fcis-18462	129	32	real	real	ADJ
fcis-18462	129	33	-	-	PUNCT
fcis-18462	129	34	time	time	NOUN
fcis-18462	129	35	and	and	CCONJ
fcis-18462	129	36	reliability	reliability	NOUN
fcis-18462	129	37	of	of	ADP
fcis-18462	129	38	multi	multi	ADJ
fcis-18462	129	39	-	-	ADJ
fcis-18462	129	40	object	object	ADJ
fcis-18462	129	41	data	datum	NOUN
fcis-18462	129	42	association	association	NOUN
fcis-18462	129	43	in	in	ADP
fcis-18462	129	44	complex	complex	ADJ
fcis-18462	129	45	environments	environment	NOUN
fcis-18462	129	46	.	.	PUNCT
fcis-18462	130	1	the	the	DET
fcis-18462	130	2	specific	specific	ADJ
fcis-18462	130	3	realization	realization	NOUN
fcis-18462	130	4	flow	flow	NOUN
fcis-18462	130	5	chart	chart	NOUN
fcis-18462	130	6	is	be	AUX
fcis-18462	130	7	as	as	SCONJ
fcis-18462	130	8	follows	follow	VERB
fcis-18462	130	9	:	:	PUNCT
fcis-18462	130	10	the	the	DET
fcis-18462	130	11	specific	specific	ADJ
fcis-18462	130	12	implementation	implementation	NOUN
fcis-18462	130	13	flow	flow	NOUN
fcis-18462	130	14	chart	chart	NOUN
fcis-18462	130	15	is	be	AUX
fcis-18462	130	16	as	as	SCONJ
fcis-18462	130	17	follows	follow	VERB
fcis-18462	130	18	:	:	PUNCT
fcis-18462	130	19	figure	figure	NOUN
fcis-18462	130	20	5	5	NUM
fcis-18462	130	21	.	.	PUNCT
fcis-18462	130	22	implementation	implementation	NOUN
fcis-18462	130	23	flowchart	flowchart	NOUN
fcis-18462	130	24	79	79	NUM
fcis-18462	130	25	5	5	NUM
fcis-18462	130	26	.	.	PUNCT
fcis-18462	131	1	experimental	experimental	ADJ
fcis-18462	131	2	results	result	NOUN
fcis-18462	131	3	in	in	ADP
fcis-18462	131	4	the	the	DET
fcis-18462	131	5	area	area	NOUN
fcis-18462	131	6	of	of	ADP
fcis-18462	131	7	target	target	NOUN
fcis-18462	131	8	detection	detection	NOUN
fcis-18462	131	9	,	,	PUNCT
fcis-18462	131	10	there	there	PRON
fcis-18462	131	11	are	be	VERB
fcis-18462	131	12	four	four	NUM
fcis-18462	131	13	types	type	NOUN
fcis-18462	131	14	of	of	ADP
fcis-18462	131	15	samples	sample	NOUN
fcis-18462	131	16	:	:	PUNCT
fcis-18462	131	17	true	true	ADJ
fcis-18462	131	18	positive(tp	positive(tp	ADJ
fcis-18462	131	19	)	)	PUNCT
fcis-18462	131	20	,	,	PUNCT
fcis-18462	131	21	false	false	ADJ
fcis-18462	131	22	positive(fp	positive(fp	NOUN
fcis-18462	131	23	)	)	PUNCT
fcis-18462	131	24	,	,	PUNCT
fcis-18462	131	25	true	true	ADJ
fcis-18462	131	26	positive(tp	positive(tp	ADJ
fcis-18462	131	27	)	)	PUNCT
fcis-18462	131	28	,	,	PUNCT
fcis-18462	131	29	false	false	ADJ
fcis-18462	131	30	positive(fp	positive(fp	NOUN
fcis-18462	131	31	)	)	PUNCT
fcis-18462	131	32	,	,	PUNCT
fcis-18462	131	33	in	in	ADP
fcis-18462	131	34	order	order	NOUN
fcis-18462	131	35	to	to	PART
fcis-18462	131	36	quantify	quantify	VERB
fcis-18462	131	37	the	the	DET
fcis-18462	131	38	performance	performance	NOUN
fcis-18462	131	39	comparison	comparison	NOUN
fcis-18462	131	40	of	of	ADP
fcis-18462	131	41	the	the	DET
fcis-18462	131	42	proposed	propose	VERB
fcis-18462	131	43	network	network	NOUN
fcis-18462	131	44	with	with	ADP
fcis-18462	131	45	other	other	ADJ
fcis-18462	131	46	networks	network	NOUN
fcis-18462	131	47	,	,	PUNCT
fcis-18462	131	48	this	this	DET
fcis-18462	131	49	paper	paper	NOUN
fcis-18462	131	50	uses	use	VERB
fcis-18462	131	51	typical	typical	ADJ
fcis-18462	131	52	evaluation	evaluation	NOUN
fcis-18462	131	53	metrics	metric	NOUN
fcis-18462	131	54	such	such	ADJ
fcis-18462	131	55	as	as	ADP
fcis-18462	131	56	precision	precision	NOUN
fcis-18462	131	57	(	(	PUNCT
fcis-18462	131	58	p	p	NOUN
fcis-18462	131	59	)	)	PUNCT
fcis-18462	131	60	and	and	CCONJ
fcis-18462	131	61	recall	recall	NOUN
fcis-18462	131	62	(	(	PUNCT
fcis-18462	131	63	r	r	NOUN
fcis-18462	131	64	)	)	PUNCT
fcis-18462	131	65	,	,	PUNCT
fcis-18462	131	66	the	the	DET
fcis-18462	131	67	formula	formula	NOUN
fcis-18462	131	68	is	be	AUX
fcis-18462	131	69	as	as	SCONJ
fcis-18462	131	70	follows	follow	VERB
fcis-18462	131	71	:	:	PUNCT
fcis-18462	131	72	the	the	DET
fcis-18462	131	73	average	average	ADJ
fcis-18462	131	74	precision	precision	NOUN
fcis-18462	131	75	(	(	PUNCT
fcis-18462	131	76	ap	ap	PROPN
fcis-18462	131	77	)	)	PUNCT
fcis-18462	131	78	is	be	AUX
fcis-18462	131	79	the	the	DET
fcis-18462	131	80	area	area	NOUN
fcis-18462	131	81	below	below	ADP
fcis-18462	131	82	the	the	DET
fcis-18462	131	83	p	p	NOUN
fcis-18462	131	84	-	-	PUNCT
fcis-18462	131	85	r	r	NOUN
fcis-18462	131	86	curve	curve	NOUN
fcis-18462	131	87	,	,	PUNCT
fcis-18462	131	88	which	which	PRON
fcis-18462	131	89	is	be	AUX
fcis-18462	131	90	the	the	DET
fcis-18462	131	91	average	average	NOUN
fcis-18462	131	92	of	of	ADP
fcis-18462	131	93	the	the	DET
fcis-18462	131	94	precision	precision	NOUN
fcis-18462	131	95	values	value	NOUN
fcis-18462	131	96	of	of	ADP
fcis-18462	131	97	the	the	DET
fcis-18462	131	98	p	p	NOUN
fcis-18462	131	99	-	-	PUNCT
fcis-18462	131	100	r	r	NOUN
fcis-18462	131	101	curve	curve	NOUN
fcis-18462	131	102	.	.	PUNCT
fcis-18462	132	1	for	for	ADP
fcis-18462	132	2	p	p	NOUN
fcis-18462	132	3	-	-	PUNCT
fcis-18462	132	4	r	r	NOUN
fcis-18462	132	5	curve	curve	NOUN
fcis-18462	132	6	,	,	PUNCT
fcis-18462	132	7	a	a	DET
fcis-18462	132	8	definite	definite	ADJ
fcis-18462	132	9	integral	integral	NOUN
fcis-18462	132	10	is	be	AUX
fcis-18462	132	11	used	use	VERB
fcis-18462	132	12	for	for	ADP
fcis-18462	132	13	the	the	DET
fcis-18462	132	14	calculation	calculation	NOUN
fcis-18462	132	15	:	:	PUNCT
fcis-18462	132	16	in	in	ADP
fcis-18462	132	17	the	the	DET
fcis-18462	132	18	actual	actual	ADJ
fcis-18462	132	19	calculation	calculation	NOUN
fcis-18462	132	20	of	of	ADP
fcis-18462	132	21	target	target	NOUN
fcis-18462	132	22	detection	detection	NOUN
fcis-18462	132	23	performance	performance	NOUN
fcis-18462	132	24	test	test	NOUN
fcis-18462	132	25	,	,	PUNCT
fcis-18462	132	26	p	p	X
fcis-18462	132	27	-	-	PUNCT
fcis-18462	132	28	r	r	NOUN
fcis-18462	132	29	curve	curve	NOUN
fcis-18462	132	30	should	should	AUX
fcis-18462	132	31	be	be	AUX
fcis-18462	132	32	smoothed	smooth	VERB
fcis-18462	132	33	.	.	PUNCT
fcis-18462	133	1	for	for	ADP
fcis-18462	133	2	each	each	DET
fcis-18462	133	3	point	point	NOUN
fcis-18462	133	4	on	on	ADP
fcis-18462	133	5	the	the	DET
fcis-18462	133	6	pr	pr	NOUN
fcis-18462	133	7	curve	curve	NOUN
fcis-18462	133	8	,	,	PUNCT
fcis-18462	133	9	the	the	DET
fcis-18462	133	10	precision	precision	NOUN
fcis-18462	133	11	value	value	NOUN
fcis-18462	133	12	is	be	AUX
fcis-18462	133	13	calculated	calculate	VERB
fcis-18462	133	14	as	as	SCONJ
fcis-18462	133	15	follows	follow	VERB
fcis-18462	133	16	:	:	PUNCT
fcis-18462	133	17	map	map	NOUN
fcis-18462	133	18	is	be	AUX
fcis-18462	133	19	an	an	DET
fcis-18462	133	20	average	average	ADJ
fcis-18462	133	21	value	value	NOUN
fcis-18462	133	22	of	of	ADP
fcis-18462	133	23	a	a	DET
fcis-18462	133	24	plurality	plurality	NOUN
fcis-18462	133	25	of	of	ADP
fcis-18462	133	26	classes	class	NOUN
fcis-18462	133	27	map	map	NOUN
fcis-18462	133	28	.	.	PUNCT
fcis-18462	134	1	because	because	SCONJ
fcis-18462	134	2	the	the	DET
fcis-18462	134	3	method	method	NOUN
fcis-18462	134	4	in	in	ADP
fcis-18462	134	5	this	this	DET
fcis-18462	134	6	paper	paper	NOUN
fcis-18462	134	7	only	only	ADV
fcis-18462	134	8	detects	detect	VERB
fcis-18462	134	9	pavement	pavement	NOUN
fcis-18462	134	10	objects	object	NOUN
fcis-18462	134	11	,	,	PUNCT
fcis-18462	134	12	the	the	DET
fcis-18462	134	13	meaning	meaning	NOUN
fcis-18462	134	14	of	of	ADP
fcis-18462	134	15	map	map	NOUN
fcis-18462	134	16	and	and	CCONJ
fcis-18462	134	17	ap	ap	X
fcis-18462	134	18	in	in	ADP
fcis-18462	134	19	this	this	DET
fcis-18462	134	20	paper	paper	NOUN
fcis-18462	134	21	is	be	AUX
fcis-18462	134	22	the	the	DET
fcis-18462	134	23	same	same	ADJ
fcis-18462	134	24	.	.	PUNCT
fcis-18462	135	1	map	map	NOUN
fcis-18462	135	2	is	be	AUX
fcis-18462	135	3	the	the	DET
fcis-18462	135	4	most	most	ADV
fcis-18462	135	5	important	important	ADJ
fcis-18462	135	6	index	index	NOUN
fcis-18462	135	7	to	to	PART
fcis-18462	135	8	measure	measure	VERB
fcis-18462	135	9	the	the	DET
fcis-18462	135	10	performance	performance	NOUN
fcis-18462	135	11	of	of	ADP
fcis-18462	135	12	target	target	NOUN
fcis-18462	135	13	detection	detection	NOUN
fcis-18462	135	14	.	.	PUNCT
fcis-18462	136	1	in	in	ADP
fcis-18462	136	2	this	this	DET
fcis-18462	136	3	experiment	experiment	NOUN
fcis-18462	136	4	,	,	PUNCT
fcis-18462	136	5	we	we	PRON
fcis-18462	136	6	collected	collect	VERB
fcis-18462	136	7	a	a	DET
fcis-18462	136	8	lot	lot	NOUN
fcis-18462	136	9	of	of	ADP
fcis-18462	136	10	videos	video	NOUN
fcis-18462	136	11	on	on	ADP
fcis-18462	136	12	the	the	DET
fcis-18462	136	13	network	network	NOUN
fcis-18462	136	14	and	and	CCONJ
fcis-18462	136	15	took	take	VERB
fcis-18462	136	16	some	some	DET
fcis-18462	136	17	videos	video	NOUN
fcis-18462	136	18	for	for	ADP
fcis-18462	136	19	testing	testing	NOUN
fcis-18462	136	20	from	from	ADP
fcis-18462	136	21	different	different	ADJ
fcis-18462	136	22	angles	angle	NOUN
fcis-18462	136	23	,	,	PUNCT
fcis-18462	136	24	and	and	CCONJ
fcis-18462	136	25	a	a	DET
fcis-18462	136	26	large	large	ADJ
fcis-18462	136	27	number	number	NOUN
fcis-18462	136	28	of	of	ADP
fcis-18462	136	29	videos	video	NOUN
fcis-18462	136	30	were	be	AUX
fcis-18462	136	31	recognized	recognize	VERB
fcis-18462	136	32	well	well	ADV
fcis-18462	136	33	under	under	ADP
fcis-18462	136	34	the	the	DET
fcis-18462	136	35	model	model	NOUN
fcis-18462	136	36	and	and	CCONJ
fcis-18462	136	37	algorithm	algorithm	NOUN
fcis-18462	136	38	,	,	PUNCT
fcis-18462	136	39	and	and	CCONJ
fcis-18462	136	40	the	the	DET
fcis-18462	136	41	accuracy	accuracy	NOUN
fcis-18462	136	42	was	be	AUX
fcis-18462	136	43	very	very	ADV
fcis-18462	136	44	high	high	ADJ
fcis-18462	136	45	.	.	PUNCT
fcis-18462	137	1	the	the	DET
fcis-18462	137	2	accuracy	accuracy	NOUN
fcis-18462	137	3	and	and	CCONJ
fcis-18462	137	4	efficiency	efficiency	NOUN
fcis-18462	137	5	of	of	ADP
fcis-18462	137	6	traffic	traffic	NOUN
fcis-18462	137	7	flow	flow	NOUN
fcis-18462	137	8	statistics	statistic	NOUN
fcis-18462	137	9	are	be	AUX
fcis-18462	137	10	fully	fully	ADV
fcis-18462	137	11	improved	improve	VERB
fcis-18462	137	12	,	,	PUNCT
fcis-18462	137	13	and	and	CCONJ
fcis-18462	137	14	the	the	DET
fcis-18462	137	15	algorithm	algorithm	NOUN
fcis-18462	137	16	mentioned	mention	VERB
fcis-18462	137	17	in	in	ADP
fcis-18462	137	18	this	this	DET
fcis-18462	137	19	paper	paper	NOUN
fcis-18462	137	20	can	can	AUX
fcis-18462	137	21	have	have	VERB
fcis-18462	137	22	a	a	DET
fcis-18462	137	23	good	good	ADJ
fcis-18462	137	24	detection	detection	NOUN
fcis-18462	137	25	effect	effect	NOUN
fcis-18462	137	26	on	on	ADP
fcis-18462	137	27	traffic	traffic	NOUN
fcis-18462	137	28	flow	flow	NOUN
fcis-18462	137	29	statistics	statistic	NOUN
fcis-18462	137	30	.	.	PUNCT
fcis-18462	138	1	the	the	DET
fcis-18462	138	2	yolov5	yolov5	NOUN
fcis-18462	138	3	algorithm	algorithm	NOUN
fcis-18462	138	4	can	can	AUX
fcis-18462	138	5	be	be	AUX
fcis-18462	138	6	applied	apply	VERB
fcis-18462	138	7	to	to	ADP
fcis-18462	138	8	a	a	DET
fcis-18462	138	9	large	large	ADJ
fcis-18462	138	10	number	number	NOUN
fcis-18462	138	11	of	of	ADP
fcis-18462	138	12	object	object	NOUN
fcis-18462	138	13	detection	detection	NOUN
fcis-18462	138	14	projects	project	NOUN
fcis-18462	138	15	to	to	PART
fcis-18462	138	16	improve	improve	VERB
fcis-18462	138	17	the	the	DET
fcis-18462	138	18	accuracy	accuracy	NOUN
fcis-18462	138	19	and	and	CCONJ
fcis-18462	138	20	efficiency	efficiency	NOUN
fcis-18462	138	21	of	of	ADP
fcis-18462	138	22	the	the	DET
fcis-18462	138	23	project	project	NOUN
fcis-18462	138	24	.	.	PUNCT
fcis-18462	139	1	6	6	X
fcis-18462	139	2	.	.	X
fcis-18462	139	3	conclusion	conclusion	NOUN
fcis-18462	139	4	in	in	ADP
fcis-18462	139	5	this	this	DET
fcis-18462	139	6	paper	paper	NOUN
fcis-18462	139	7	,	,	PUNCT
fcis-18462	139	8	the	the	DET
fcis-18462	139	9	yolov5	yolov5	NOUN
fcis-18462	139	10	algorithm	algorithm	NOUN
fcis-18462	139	11	is	be	AUX
fcis-18462	139	12	used	use	VERB
fcis-18462	139	13	to	to	PART
fcis-18462	139	14	improve	improve	VERB
fcis-18462	139	15	the	the	DET
fcis-18462	139	16	accuracy	accuracy	NOUN
fcis-18462	139	17	and	and	CCONJ
fcis-18462	139	18	efficiency	efficiency	NOUN
fcis-18462	139	19	of	of	ADP
fcis-18462	139	20	traffic	traffic	NOUN
fcis-18462	139	21	flow	flow	NOUN
fcis-18462	139	22	statistics	statistic	NOUN
fcis-18462	139	23	,	,	PUNCT
fcis-18462	139	24	and	and	CCONJ
fcis-18462	139	25	improve	improve	VERB
fcis-18462	139	26	the	the	DET
fcis-18462	139	27	accuracy	accuracy	NOUN
fcis-18462	139	28	of	of	ADP
fcis-18462	139	29	actual	actual	ADJ
fcis-18462	139	30	detection	detection	NOUN
fcis-18462	139	31	.	.	PUNCT
fcis-18462	140	1	in	in	ADP
fcis-18462	140	2	the	the	DET
fcis-18462	140	3	future	future	NOUN
fcis-18462	140	4	,	,	PUNCT
fcis-18462	140	5	we	we	PRON
fcis-18462	140	6	will	will	AUX
fcis-18462	140	7	consider	consider	VERB
fcis-18462	140	8	adapting	adapt	VERB
fcis-18462	140	9	the	the	DET
fcis-18462	140	10	model	model	NOUN
fcis-18462	140	11	algorithm	algorithm	NOUN
fcis-18462	140	12	to	to	ADP
fcis-18462	140	13	the	the	DET
fcis-18462	140	14	traffic	traffic	NOUN
fcis-18462	140	15	camera	camera	NOUN
fcis-18462	140	16	to	to	PART
fcis-18462	140	17	truly	truly	ADV
fcis-18462	140	18	achieve	achieve	VERB
fcis-18462	140	19	the	the	DET
fcis-18462	140	20	statistics	statistic	NOUN
fcis-18462	140	21	of	of	ADP
fcis-18462	140	22	the	the	DET
fcis-18462	140	23	actual	actual	ADJ
fcis-18462	140	24	application	application	NOUN
fcis-18462	140	25	of	of	ADP
fcis-18462	140	26	traffic	traffic	NOUN
fcis-18462	140	27	flow	flow	NOUN
fcis-18462	140	28	.	.	PUNCT
fcis-18462	141	1	acknowledgments	acknowledgment	NOUN
fcis-18462	141	2	college	college	PROPN
fcis-18462	141	3	students	student	NOUN
fcis-18462	141	4	'	'	PART
fcis-18462	141	5	innovation	innovation	NOUN
fcis-18462	141	6	and	and	CCONJ
fcis-18462	141	7	entrepreneurship	entrepreneurship	NOUN
fcis-18462	141	8	training	training	NOUN
fcis-18462	141	9	plan	plan	NOUN
fcis-18462	141	10	of	of	ADP
fcis-18462	141	11	university	university	NOUN
fcis-18462	141	12	of	of	ADP
fcis-18462	141	13	science	science	NOUN
fcis-18462	141	14	and	and	CCONJ
fcis-18462	141	15	technology	technology	NOUN
fcis-18462	141	16	liaoning	liaoning	NOUN
fcis-18462	141	17	in	in	ADP
fcis-18462	141	18	2024	2024	NUM
fcis-18462	142	1	.	.	PUNCT
fcis-18462	143	1	references	reference	NOUN
fcis-18462	143	2	[	[	X
fcis-18462	143	3	1	1	NUM
fcis-18462	143	4	]	]	X
fcis-18462	143	5	zhang	zhang	PROPN
fcis-18462	143	6	yingwei	yingwei	PROPN
fcis-18462	143	7	.	.	PUNCT
fcis-18462	144	1	research	research	NOUN
fcis-18462	144	2	on	on	ADP
fcis-18462	144	3	traffic	traffic	NOUN
fcis-18462	144	4	flow	flow	NOUN
fcis-18462	144	5	and	and	CCONJ
fcis-18462	144	6	violation	violation	NOUN
fcis-18462	144	7	detection	detection	NOUN
fcis-18462	144	8	technology	technology	NOUN
fcis-18462	144	9	in	in	ADP
fcis-18462	144	10	intelligent	intelligent	ADJ
fcis-18462	144	11	transportation[d].liaoning	transportation[d].liaoning	NOUN
fcis-18462	144	12	university	university	NOUN
fcis-18462	144	13	of	of	ADP
fcis-18462	144	14	science	science	NOUN
fcis-18462	144	15	and	and	CCONJ
fcis-18462	144	16	technology	technology	NOUN
fcis-18462	144	17	,	,	PUNCT
fcis-18462	144	18	2023.doi	2023.doi	NUM
fcis-18462	144	19	:	:	PUNCT
fcis-18462	144	20	10	10	NUM
fcis-18462	144	21	.	.	NOUN
fcis-18462	144	22	26923/	26923/	NUM
fcis-18462	144	23	d.	d.	PROPN
fcis-18462	144	24	cnki.gasgc.2023.000474	cnki.gasgc.2023.000474	NOUN
fcis-18462	144	25	.	.	PUNCT
fcis-18462	145	1	[	[	X
fcis-18462	145	2	2	2	NUM
fcis-18462	145	3	]	]	X
fcis-18462	145	4	research	research	NOUN
fcis-18462	145	5	and	and	CCONJ
fcis-18462	145	6	application	application	NOUN
fcis-18462	145	7	of	of	ADP
fcis-18462	145	8	multi	multi	ADJ
fcis-18462	145	9	-	-	ADJ
fcis-18462	145	10	target	target	ADJ
fcis-18462	145	11	tracking	tracking	NOUN
fcis-18462	145	12	algorithm	algorithm	NOUN
fcis-18462	145	13	based	base	VERB
fcis-18462	145	14	on	on	ADP
fcis-18462	145	15	yolov5	yolov5	NOUN
fcis-18462	145	16	and	and	CCONJ
fcis-18462	145	17	deepsort[d	deepsort[d	NOUN
fcis-18462	145	18	]	]	PUNCT
fcis-18462	145	19	.	.	PUNCT
fcis-18462	146	1	wang	wang	PROPN
fcis-18462	146	2	jialin	jialin	PROPN
fcis-18462	146	3	shandong	shandong	PROPN
fcis-18462	146	4	university	university	PROPN
fcis-18462	146	5	,	,	PUNCT
fcis-18462	146	6	2021	2021	NUM
fcis-18462	146	7	.	.	PUNCT
fcis-18462	147	1	[	[	X
fcis-18462	147	2	3	3	X
fcis-18462	147	3	]	]	X
fcis-18462	147	4	liu	liu	PROPN
fcis-18462	147	5	bei	bei	PROPN
fcis-18462	147	6	.	.	PUNCT
fcis-18462	148	1	research	research	NOUN
fcis-18462	148	2	on	on	ADP
fcis-18462	148	3	traffic	traffic	NOUN
fcis-18462	148	4	flow	flow	NOUN
fcis-18462	148	5	detection	detection	NOUN
fcis-18462	148	6	algorithm	algorithm	NOUN
fcis-18462	148	7	based	base	VERB
fcis-18462	148	8	on	on	ADP
fcis-18462	148	9	object	object	NOUN
fcis-18462	148	10	detection	detection	NOUN
fcis-18462	148	11	and	and	CCONJ
fcis-18462	148	12	tracking[d	tracking[d	PROPN
fcis-18462	148	13	]	]	PUNCT
fcis-18462	148	14	.	.	PUNCT
fcis-18462	149	1	inner	inner	PROPN
fcis-18462	149	2	mongolia	mongolia	PROPN
fcis-18462	149	3	:	:	PUNCT
fcis-18462	149	4	inner	inner	PROPN
fcis-18462	149	5	mongolia	mongolia	PROPN
fcis-18462	149	6	university,2022	university,2022	NOUN
fcis-18462	149	7	.	.	PUNCT
fcis-18462	150	1	[	[	X
fcis-18462	150	2	4	4	NUM
fcis-18462	150	3	]	]	X
fcis-18462	150	4	editorial	editorial	ADJ
fcis-18462	150	5	department	department	NOUN
fcis-18462	150	6	of	of	ADP
fcis-18462	150	7	chinese	chinese	PROPN
fcis-18462	150	8	journal	journal	PROPN
fcis-18462	150	9	of	of	ADP
fcis-18462	150	10	highway	highway	NOUN
fcis-18462	150	11	and	and	CCONJ
fcis-18462	150	12	transportation	transportation	NOUN
fcis-18462	150	13	.	.	PUNCT
fcis-18462	151	1	review	review	NOUN
fcis-18462	151	2	of	of	ADP
fcis-18462	151	3	china	china	PROPN
fcis-18462	151	4	automotive	automotive	PROPN
fcis-18462	151	5	engineering	engineering	PROPN
fcis-18462	151	6	research	research	NOUN
fcis-18462	151	7	progress	progress	NOUN
fcis-18462	151	8	in	in	ADP
fcis-18462	151	9	2017	2017	NUM
fcis-18462	151	10	.	.	PUNCT
fcis-18462	152	1	china	china	PROPN
fcis-18462	152	2	soumnal	soumnal	PROPN
fcis-18462	152	3	of	of	ADP
fcis-18462	152	4	highway	highway	NOUN
fcis-18462	152	5	and	and	CCONJ
fcis-18462	152	6	transpont,2017	transpont,2017	PRON
fcis-18462	152	7	,	,	PUNCT
fcis-18462	152	8	30(6):1	30(6):1	PROPN
fcis-18462	152	9	-	-	PUNCT
fcis-18462	152	10	197	197	NUM
fcis-18462	152	11	.	.	PUNCT
fcis-18462	153	1	editorial	editorial	ADJ
fcis-18462	153	2	department	department	PROPN
fcis-18462	153	3	of	of	ADP
fcis-18462	153	4	chinese	chinese	PROPN
fcis-18462	153	5	journal	journal	PROPN
fcis-18462	153	6	of	of	ADP
fcis-18462	153	7	highway	highway	NOUN
fcis-18462	153	8	engineering	engineering	PROPN
fcis-18462	153	9	,	,	PUNCT
fcis-18462	153	10	review	review	NOUN
fcis-18462	153	11	of	of	ADP
fcis-18462	153	12	chinese	chinese	ADJ
fcis-18462	153	13	automotive	automotive	ADJ
fcis-18462	153	14	engineering	engineering	PROPN
fcis-18462	153	15	academic	academic	ADJ
fcis-18462	153	16	research	research	NOUN
fcis-18462	153	17	,	,	PUNCT
fcis-18462	153	18	2017	2017	NUM
fcis-18462	154	1	[	[	X
fcis-18462	154	2	j	j	X
fcis-18462	154	3	]	]	X
fcis-18462	154	4	.	.	PUNCT
fcis-18462	155	1	chinese	chinese	ADJ
fcis-18462	155	2	journal	journal	PROPN
fcis-18462	155	3	of	of	ADP
fcis-18462	155	4	highway	highway	NOUN
fcis-18462	155	5	,	,	PUNCT
fcis-18462	155	6	2017,30	2017,30	NUM
fcis-18462	155	7	(	(	PUNCT
fcis-18462	155	8	6	6	NUM
fcis-18462	155	9	):	):	SYM
fcis-18462	155	10	1	1	NUM
fcis-18462	155	11	-	-	SYM
fcis-18462	155	12	197	197	NUM
fcis-18462	155	13	.	.	PUNCT
fcis-18462	156	1	[	[	X
fcis-18462	156	2	5	5	NUM
fcis-18462	156	3	]	]	PUNCT
fcis-18462	156	4	lin	lin	PROPN
fcis-18462	156	5	shu	shu	PROPN
fcis-18462	156	6	,	,	PUNCT
fcis-18462	156	7	liu	liu	PROPN
fcis-18462	156	8	mingying	mingying	PROPN
fcis-18462	156	9	,	,	PUNCT
fcis-18462	156	10	tao	tao	PROPN
fcis-18462	156	11	zhiying	zhiying	NOUN
fcis-18462	156	12	,	,	PUNCT
fcis-18462	156	13	underwater	underwater	NOUN
fcis-18462	156	14	treasure	treasure	NOUN
fcis-18462	156	15	detection	detection	NOUN
fcis-18462	156	16	based	base	VERB
fcis-18462	156	17	on	on	ADP
fcis-18462	156	18	attention	attention	NOUN
fcis-18462	156	19	mechanism	mechanism	NOUN
fcis-18462	156	20	and	and	CCONJ
fcis-18462	156	21	improved	improve	VERB
fcis-18462	156	22	yolov5	yolov5	NOUN
fcis-18462	156	23	y!.transactions	y!.transaction	NOUN
fcis-18462	156	24	of	of	ADP
fcis-18462	156	25	the	the	DET
fcis-18462	156	26	chinese	chinese	ADJ
fcis-18462	156	27	society	society	NOUN
fcis-18462	156	28	of	of	ADP
fcis-18462	156	29	agricutural	agricutural	ADJ
fcis-18462	156	30	engineerng	engineerng	NOUN
fcis-18462	156	31	,	,	PUNCT
fcis-18462	156	32	2021	2021	NUM
fcis-18462	156	33	,	,	PUNCT
fcis-18462	156	34	37(18):307	37(18):307	NUM
fcis-18462	156	35	-	-	SYM
fcis-18462	156	36	314	314	NUM
fcis-18462	156	37	.	.	PUNCT