id	sid	tid	token	lemma	pos
fcis-15825	1	1	frontiers	frontier	NOUN
fcis-15825	1	2	in	in	ADP
fcis-15825	1	3	computing	computing	NOUN
fcis-15825	1	4	and	and	CCONJ
fcis-15825	1	5	intelligent	intelligent	ADJ
fcis-15825	1	6	systems	system	NOUN
fcis-15825	1	7	issn	issn	VERB
fcis-15825	1	8	:	:	PUNCT
fcis-15825	1	9	2832	2832	NUM
fcis-15825	1	10	-	-	SYM
fcis-15825	1	11	6024	6024	NUM
fcis-15825	1	12	|	|	NOUN
fcis-15825	1	13	vol	vol	NOUN
fcis-15825	1	14	.	.	PROPN
fcis-15825	2	1	6	6	NUM
fcis-15825	2	2	,	,	PUNCT
fcis-15825	2	3	no	no	INTJ
fcis-15825	2	4	.	.	NOUN
fcis-15825	2	5	3	3	NUM
fcis-15825	2	6	,	,	PUNCT
fcis-15825	2	7	2023	2023	NUM
fcis-15825	2	8	113	113	NUM
fcis-15825	2	9	intelligent	intelligent	ADJ
fcis-15825	2	10	pavement	pavement	NOUN
fcis-15825	2	11	diseases	disease	NOUN
fcis-15825	2	12	detection	detection	NOUN
fcis-15825	2	13	model	model	NOUN
fcis-15825	2	14	based	base	VERB
fcis-15825	2	15	on	on	ADP
fcis-15825	2	16	yolov5s	yolov5s	PROPN
fcis-15825	2	17	weihua	weihua	PROPN
fcis-15825	2	18	pan	pan	PROPN
fcis-15825	2	19	,	,	PUNCT
fcis-15825	2	20	mengying	mengye	VERB
fcis-15825	2	21	nie	nie	PROPN
fcis-15825	2	22	north	north	PROPN
fcis-15825	2	23	china	china	PROPN
fcis-15825	2	24	electric	electric	PROPN
fcis-15825	2	25	power	power	PROPN
fcis-15825	2	26	university	university	PROPN
fcis-15825	2	27	department	department	PROPN
fcis-15825	2	28	of	of	ADP
fcis-15825	2	29	computer	computer	PROPN
fcis-15825	2	30	,	,	PUNCT
fcis-15825	2	31	baoding	baoding	PROPN
fcis-15825	2	32	,	,	PUNCT
fcis-15825	2	33	hebei	hebei	PROPN
fcis-15825	2	34	071003	071003	NUM
fcis-15825	2	35	,	,	PUNCT
fcis-15825	2	36	china	china	PROPN
fcis-15825	2	37	abstract	abstract	NOUN
fcis-15825	2	38	:	:	PUNCT
fcis-15825	2	39	there	there	PRON
fcis-15825	2	40	are	be	VERB
fcis-15825	2	41	many	many	ADJ
fcis-15825	2	42	kinds	kind	NOUN
fcis-15825	2	43	of	of	ADP
fcis-15825	2	44	pavement	pavement	NOUN
fcis-15825	2	45	diseases	disease	NOUN
fcis-15825	2	46	,	,	PUNCT
fcis-15825	2	47	and	and	CCONJ
fcis-15825	2	48	the	the	DET
fcis-15825	2	49	detection	detection	NOUN
fcis-15825	2	50	background	background	NOUN
fcis-15825	2	51	in	in	ADP
fcis-15825	2	52	the	the	DET
fcis-15825	2	53	actual	actual	ADJ
fcis-15825	2	54	scene	scene	NOUN
fcis-15825	2	55	is	be	AUX
fcis-15825	2	56	complex	complex	ADJ
fcis-15825	2	57	,	,	PUNCT
fcis-15825	2	58	so	so	CCONJ
fcis-15825	2	59	the	the	DET
fcis-15825	2	60	accuracy	accuracy	NOUN
fcis-15825	2	61	of	of	ADP
fcis-15825	2	62	the	the	DET
fcis-15825	2	63	detection	detection	NOUN
fcis-15825	2	64	model	model	NOUN
fcis-15825	2	65	is	be	AUX
fcis-15825	2	66	not	not	PART
fcis-15825	2	67	high	high	ADJ
fcis-15825	2	68	.	.	PUNCT
fcis-15825	3	1	based	base	VERB
fcis-15825	3	2	on	on	ADP
fcis-15825	3	3	yolov5s	yolov5s	PROPN
fcis-15825	3	4	,	,	PUNCT
fcis-15825	3	5	a	a	DET
fcis-15825	3	6	pavement	pavement	NOUN
fcis-15825	3	7	diseases	disease	NOUN
fcis-15825	3	8	detection	detection	NOUN
fcis-15825	3	9	model	model	NOUN
fcis-15825	3	10	yolov5s	yolov5s	PROPN
fcis-15825	3	11	-	-	PUNCT
fcis-15825	3	12	cbp	cbp	PROPN
fcis-15825	3	13	is	be	AUX
fcis-15825	3	14	proposed	propose	VERB
fcis-15825	3	15	,	,	PUNCT
fcis-15825	3	16	and	and	CCONJ
fcis-15825	3	17	the	the	DET
fcis-15825	3	18	cbam	cbam	NOUN
fcis-15825	3	19	(	(	PUNCT
fcis-15825	3	20	convolutional	convolutional	ADJ
fcis-15825	3	21	block	block	NOUN
fcis-15825	3	22	attention	attention	NOUN
fcis-15825	3	23	module	module	NOUN
fcis-15825	3	24	)	)	PUNCT
fcis-15825	3	25	attention	attention	NOUN
fcis-15825	3	26	module	module	NOUN
fcis-15825	3	27	was	be	AUX
fcis-15825	3	28	introduced	introduce	VERB
fcis-15825	3	29	into	into	ADP
fcis-15825	3	30	the	the	DET
fcis-15825	3	31	backbone	backbone	NOUN
fcis-15825	3	32	network	network	NOUN
fcis-15825	3	33	to	to	PART
fcis-15825	3	34	enhance	enhance	VERB
fcis-15825	3	35	the	the	DET
fcis-15825	3	36	model	model	NOUN
fcis-15825	3	37	's	's	PART
fcis-15825	3	38	attention	attention	NOUN
fcis-15825	3	39	to	to	ADP
fcis-15825	3	40	the	the	DET
fcis-15825	3	41	characteristics	characteristic	NOUN
fcis-15825	3	42	of	of	ADP
fcis-15825	3	43	pavement	pavement	NOUN
fcis-15825	3	44	diseases	disease	NOUN
fcis-15825	3	45	.	.	PUNCT
fcis-15825	4	1	the	the	DET
fcis-15825	4	2	bifpn	bifpn	PROPN
fcis-15825	4	3	network	network	NOUN
fcis-15825	4	4	was	be	AUX
fcis-15825	4	5	constructed	construct	VERB
fcis-15825	4	6	in	in	ADP
fcis-15825	4	7	the	the	DET
fcis-15825	4	8	neck	neck	NOUN
fcis-15825	4	9	,	,	PUNCT
fcis-15825	4	10	which	which	PRON
fcis-15825	4	11	realizes	realize	VERB
fcis-15825	4	12	the	the	DET
fcis-15825	4	13	cross	cross	ADJ
fcis-15825	4	14	-	-	ADJ
fcis-15825	4	15	layer	layer	ADJ
fcis-15825	4	16	exchange	exchange	NOUN
fcis-15825	4	17	and	and	CCONJ
fcis-15825	4	18	integration	integration	NOUN
fcis-15825	4	19	of	of	ADP
fcis-15825	4	20	information	information	NOUN
fcis-15825	4	21	and	and	CCONJ
fcis-15825	4	22	retains	retain	VERB
fcis-15825	4	23	more	more	ADV
fcis-15825	4	24	high	high	ADJ
fcis-15825	4	25	-	-	PUNCT
fcis-15825	4	26	level	level	NOUN
fcis-15825	4	27	semantic	semantic	ADJ
fcis-15825	4	28	information	information	NOUN
fcis-15825	4	29	.	.	PUNCT
fcis-15825	5	1	a	a	DET
fcis-15825	5	2	small	small	ADJ
fcis-15825	5	3	target	target	NOUN
fcis-15825	5	4	detection	detection	NOUN
fcis-15825	5	5	layer	layer	NOUN
fcis-15825	5	6	p2	p2	PROPN
fcis-15825	5	7	was	be	AUX
fcis-15825	5	8	added	add	VERB
fcis-15825	5	9	to	to	ADP
fcis-15825	5	10	the	the	DET
fcis-15825	5	11	detection	detection	NOUN
fcis-15825	5	12	head	head	NOUN
fcis-15825	5	13	to	to	PART
fcis-15825	5	14	improve	improve	VERB
fcis-15825	5	15	the	the	DET
fcis-15825	5	16	detection	detection	NOUN
fcis-15825	5	17	ability	ability	NOUN
fcis-15825	5	18	of	of	ADP
fcis-15825	5	19	the	the	DET
fcis-15825	5	20	model	model	NOUN
fcis-15825	5	21	for	for	ADP
fcis-15825	5	22	small	small	ADJ
fcis-15825	5	23	road	road	NOUN
fcis-15825	5	24	diseases	disease	NOUN
fcis-15825	5	25	.	.	PUNCT
fcis-15825	6	1	the	the	DET
fcis-15825	6	2	experimental	experimental	ADJ
fcis-15825	6	3	results	result	NOUN
fcis-15825	6	4	on	on	ADP
fcis-15825	6	5	data	datum	NOUN
fcis-15825	6	6	sets	set	NOUN
fcis-15825	6	7	show	show	VERB
fcis-15825	6	8	that	that	SCONJ
fcis-15825	6	9	the	the	DET
fcis-15825	6	10	precision	precision	NOUN
fcis-15825	6	11	and	and	CCONJ
fcis-15825	6	12	recall	recall	NOUN
fcis-15825	6	13	of	of	ADP
fcis-15825	6	14	the	the	DET
fcis-15825	6	15	model	model	NOUN
fcis-15825	6	16	are	be	AUX
fcis-15825	6	17	improved	improve	VERB
fcis-15825	6	18	by	by	ADP
fcis-15825	6	19	6.1	6.1	NUM
fcis-15825	6	20	%	%	NOUN
fcis-15825	6	21	and	and	CCONJ
fcis-15825	6	22	4.4	4.4	NUM
fcis-15825	6	23	%	%	NOUN
fcis-15825	6	24	respectively	respectively	ADV
fcis-15825	6	25	,	,	PUNCT
fcis-15825	6	26	and	and	CCONJ
fcis-15825	6	27	the	the	DET
fcis-15825	6	28	map	map	NOUN
fcis-15825	6	29	is	be	AUX
fcis-15825	6	30	improved	improve	VERB
fcis-15825	6	31	by	by	ADP
fcis-15825	6	32	5.7	5.7	NUM
fcis-15825	6	33	%	%	NOUN
fcis-15825	6	34	,	,	PUNCT
fcis-15825	6	35	which	which	PRON
fcis-15825	6	36	effectively	effectively	ADV
fcis-15825	6	37	improves	improve	VERB
fcis-15825	6	38	the	the	DET
fcis-15825	6	39	performance	performance	NOUN
fcis-15825	6	40	of	of	ADP
fcis-15825	6	41	the	the	DET
fcis-15825	6	42	detection	detection	NOUN
fcis-15825	6	43	model	model	NOUN
fcis-15825	6	44	.	.	PUNCT
fcis-15825	7	1	keywords	keyword	NOUN
fcis-15825	7	2	:	:	PUNCT
fcis-15825	7	3	target	target	VERB
fcis-15825	7	4	detection	detection	NOUN
fcis-15825	7	5	;	;	PUNCT
fcis-15825	7	6	pavement	pavement	NOUN
fcis-15825	7	7	diseases	disease	NOUN
fcis-15825	7	8	;	;	PUNCT
fcis-15825	7	9	neural	neural	ADJ
fcis-15825	7	10	network	network	NOUN
fcis-15825	7	11	;	;	PUNCT
fcis-15825	7	12	attention	attention	NOUN
fcis-15825	7	13	mechanism	mechanism	NOUN
fcis-15825	7	14	.	.	PUNCT
fcis-15825	8	1	1	1	X
fcis-15825	8	2	.	.	X
fcis-15825	8	3	introduction	introduction	NOUN
fcis-15825	8	4	pavement	pavement	NOUN
fcis-15825	8	5	diseases	disease	NOUN
fcis-15825	8	6	are	be	AUX
fcis-15825	8	7	the	the	DET
fcis-15825	8	8	result	result	NOUN
fcis-15825	8	9	of	of	ADP
fcis-15825	8	10	many	many	ADJ
fcis-15825	8	11	factors	factor	NOUN
fcis-15825	8	12	.	.	PUNCT
fcis-15825	9	1	longterm	longterm	PROPN
fcis-15825	9	2	road	road	NOUN
fcis-15825	9	3	load	load	NOUN
fcis-15825	9	4	,	,	PUNCT
fcis-15825	9	5	sun	sun	NOUN
fcis-15825	9	6	exposure	exposure	NOUN
fcis-15825	9	7	and	and	CCONJ
fcis-15825	9	8	snow	snow	NOUN
fcis-15825	9	9	immersion	immersion	NOUN
fcis-15825	9	10	have	have	AUX
fcis-15825	9	11	led	lead	VERB
fcis-15825	9	12	to	to	ADP
fcis-15825	9	13	frequent	frequent	ADJ
fcis-15825	9	14	road	road	NOUN
fcis-15825	9	15	diseases	disease	NOUN
fcis-15825	9	16	.	.	PUNCT
fcis-15825	10	1	the	the	DET
fcis-15825	10	2	increasingly	increasingly	ADV
fcis-15825	10	3	number	number	NOUN
fcis-15825	10	4	of	of	ADP
fcis-15825	10	5	pavement	pavement	NOUN
fcis-15825	10	6	diseases	disease	NOUN
fcis-15825	10	7	has	have	AUX
fcis-15825	10	8	greatly	greatly	ADV
fcis-15825	10	9	reduced	reduce	VERB
fcis-15825	10	10	the	the	DET
fcis-15825	10	11	service	service	NOUN
fcis-15825	10	12	life	life	NOUN
fcis-15825	10	13	of	of	ADP
fcis-15825	10	14	asphalt	asphalt	NOUN
fcis-15825	10	15	pavement	pavement	NOUN
fcis-15825	10	16	,	,	PUNCT
fcis-15825	10	17	increased	increase	VERB
fcis-15825	10	18	the	the	DET
fcis-15825	10	19	pressure	pressure	NOUN
fcis-15825	10	20	of	of	ADP
fcis-15825	10	21	road	road	NOUN
fcis-15825	10	22	maintenance	maintenance	NOUN
fcis-15825	10	23	,	,	PUNCT
fcis-15825	10	24	and	and	CCONJ
fcis-15825	10	25	had	have	VERB
fcis-15825	10	26	a	a	DET
fcis-15825	10	27	negative	negative	ADJ
fcis-15825	10	28	impact	impact	NOUN
fcis-15825	10	29	on	on	ADP
fcis-15825	10	30	road	road	NOUN
fcis-15825	10	31	traffic	traffic	NOUN
fcis-15825	10	32	.	.	PUNCT
fcis-15825	11	1	thus	thus	ADV
fcis-15825	11	2	,	,	PUNCT
fcis-15825	11	3	timely	timely	ADJ
fcis-15825	11	4	and	and	CCONJ
fcis-15825	11	5	effective	effective	ADJ
fcis-15825	11	6	detection	detection	NOUN
fcis-15825	11	7	and	and	CCONJ
fcis-15825	11	8	identification	identification	NOUN
fcis-15825	11	9	of	of	ADP
fcis-15825	11	10	road	road	NOUN
fcis-15825	11	11	diseases	disease	NOUN
fcis-15825	11	12	can	can	AUX
fcis-15825	11	13	eliminate	eliminate	VERB
fcis-15825	11	14	these	these	DET
fcis-15825	11	15	potential	potential	ADJ
fcis-15825	11	16	safety	safety	NOUN
fcis-15825	11	17	hazards	hazard	NOUN
fcis-15825	11	18	as	as	ADV
fcis-15825	11	19	soon	soon	ADV
fcis-15825	11	20	as	as	ADP
fcis-15825	11	21	possible	possible	ADJ
fcis-15825	11	22	and	and	CCONJ
fcis-15825	11	23	prolong	prolong	VERB
fcis-15825	11	24	the	the	DET
fcis-15825	11	25	service	service	NOUN
fcis-15825	11	26	life	life	NOUN
fcis-15825	11	27	of	of	ADP
fcis-15825	11	28	pavement	pavement	NOUN
fcis-15825	11	29	,	,	PUNCT
fcis-15825	11	30	which	which	PRON
fcis-15825	11	31	is	be	AUX
fcis-15825	11	32	very	very	ADV
fcis-15825	11	33	important	important	ADJ
fcis-15825	11	34	for	for	ADP
fcis-15825	11	35	road	road	NOUN
fcis-15825	11	36	maintenance	maintenance	NOUN
fcis-15825	11	37	and	and	CCONJ
fcis-15825	11	38	management	management	NOUN
fcis-15825	11	39	.	.	PUNCT
fcis-15825	12	1	at	at	ADP
fcis-15825	12	2	the	the	DET
fcis-15825	12	3	early	early	ADJ
fcis-15825	12	4	stage	stage	NOUN
fcis-15825	12	5	,	,	PUNCT
fcis-15825	12	6	pavement	pavement	NOUN
fcis-15825	12	7	diseases	disease	NOUN
fcis-15825	12	8	detection	detection	NOUN
fcis-15825	12	9	was	be	AUX
fcis-15825	12	10	mainly	mainly	ADV
fcis-15825	12	11	based	base	VERB
fcis-15825	12	12	on	on	ADP
fcis-15825	12	13	manual	manual	ADJ
fcis-15825	12	14	detection	detection	NOUN
fcis-15825	12	15	,	,	PUNCT
fcis-15825	12	16	which	which	PRON
fcis-15825	12	17	was	be	AUX
fcis-15825	12	18	time	time	NOUN
fcis-15825	12	19	-	-	PUNCT
fcis-15825	12	20	consuming	consume	VERB
fcis-15825	12	21	and	and	CCONJ
fcis-15825	12	22	laborious	laborious	ADJ
fcis-15825	12	23	,	,	PUNCT
fcis-15825	12	24	and	and	CCONJ
fcis-15825	12	25	the	the	DET
fcis-15825	12	26	detection	detection	NOUN
fcis-15825	12	27	efficiency	efficiency	NOUN
fcis-15825	12	28	was	be	AUX
fcis-15825	12	29	low	low	ADJ
fcis-15825	12	30	,	,	PUNCT
fcis-15825	12	31	so	so	SCONJ
fcis-15825	12	32	it	it	PRON
fcis-15825	12	33	could	could	AUX
fcis-15825	12	34	not	not	PART
fcis-15825	12	35	meet	meet	VERB
fcis-15825	12	36	the	the	DET
fcis-15825	12	37	large	large	ADJ
fcis-15825	12	38	-	-	PUNCT
fcis-15825	12	39	scale	scale	NOUN
fcis-15825	12	40	road	road	NOUN
fcis-15825	12	41	detection	detection	NOUN
fcis-15825	12	42	task	task	NOUN
fcis-15825	12	43	.	.	PUNCT
fcis-15825	13	1	with	with	ADP
fcis-15825	13	2	the	the	DET
fcis-15825	13	3	rapid	rapid	ADJ
fcis-15825	13	4	development	development	NOUN
fcis-15825	13	5	of	of	ADP
fcis-15825	13	6	computer	computer	NOUN
fcis-15825	13	7	software	software	NOUN
fcis-15825	13	8	and	and	CCONJ
fcis-15825	13	9	hardware	hardware	NOUN
fcis-15825	13	10	,	,	PUNCT
fcis-15825	13	11	the	the	DET
fcis-15825	13	12	automatic	automatic	ADJ
fcis-15825	13	13	detection	detection	NOUN
fcis-15825	13	14	methods	method	NOUN
fcis-15825	13	15	of	of	ADP
fcis-15825	13	16	pavement	pavement	NOUN
fcis-15825	13	17	diseases	disease	NOUN
fcis-15825	13	18	have	have	AUX
fcis-15825	13	19	been	be	AUX
fcis-15825	13	20	enriched	enrich	VERB
fcis-15825	13	21	.	.	PUNCT
fcis-15825	14	1	the	the	DET
fcis-15825	14	2	methods	method	NOUN
fcis-15825	14	3	of	of	ADP
fcis-15825	14	4	disease	disease	NOUN
fcis-15825	14	5	detection	detection	NOUN
fcis-15825	14	6	based	base	VERB
fcis-15825	14	7	on	on	ADP
fcis-15825	14	8	digital	digital	ADJ
fcis-15825	14	9	images	image	NOUN
fcis-15825	14	10	can	can	AUX
fcis-15825	14	11	be	be	AUX
fcis-15825	14	12	summarized	summarize	VERB
fcis-15825	14	13	as	as	ADP
fcis-15825	14	14	:	:	PUNCT
fcis-15825	14	15	threshold	threshold	NOUN
fcis-15825	14	16	segmentation	segmentation	NOUN
fcis-15825	14	17	method[2	method[2	PROPN
fcis-15825	14	18	-	-	PUNCT
fcis-15825	14	19	3	3	NUM
fcis-15825	14	20	]	]	PUNCT
fcis-15825	14	21	,	,	PUNCT
fcis-15825	14	22	edge	edge	NOUN
fcis-15825	14	23	detection	detection	NOUN
fcis-15825	14	24	method[4	method[4	NOUN
fcis-15825	14	25	-	-	SYM
fcis-15825	14	26	5	5	NUM
fcis-15825	14	27	]	]	PUNCT
fcis-15825	14	28	and	and	CCONJ
fcis-15825	14	29	region	region	NOUN
fcis-15825	14	30	growth	growth	NOUN
fcis-15825	14	31	method[6	method[6	PROPN
fcis-15825	14	32	-	-	SYM
fcis-15825	14	33	7	7	NUM
fcis-15825	14	34	]	]	PUNCT
fcis-15825	14	35	,	,	PUNCT
fcis-15825	14	36	but	but	CCONJ
fcis-15825	14	37	these	these	DET
fcis-15825	14	38	methods	method	NOUN
fcis-15825	14	39	have	have	VERB
fcis-15825	14	40	low	low	ADJ
fcis-15825	14	41	accuracy	accuracy	NOUN
fcis-15825	14	42	and	and	CCONJ
fcis-15825	14	43	recall	recall	NOUN
fcis-15825	14	44	rate	rate	NOUN
fcis-15825	14	45	and	and	CCONJ
fcis-15825	14	46	complicated	complicated	ADJ
fcis-15825	14	47	steps	step	NOUN
fcis-15825	14	48	.	.	PUNCT
fcis-15825	15	1	the	the	DET
fcis-15825	15	2	detection	detection	NOUN
fcis-15825	15	3	method	method	NOUN
fcis-15825	15	4	based	base	VERB
fcis-15825	15	5	on	on	ADP
fcis-15825	15	6	neural	neural	ADJ
fcis-15825	15	7	network	network	NOUN
fcis-15825	15	8	has	have	VERB
fcis-15825	15	9	two	two	NUM
fcis-15825	15	10	-	-	PUNCT
fcis-15825	15	11	stage	stage	NOUN
fcis-15825	15	12	algorithm	algorithm	NOUN
fcis-15825	15	13	[	[	X
fcis-15825	15	14	8	8	NUM
fcis-15825	15	15	]	]	PUNCT
fcis-15825	15	16	and	and	CCONJ
fcis-15825	15	17	one	one	NUM
fcis-15825	15	18	-	-	PUNCT
fcis-15825	15	19	stage	stage	NOUN
fcis-15825	15	20	algorithm[9	algorithm[9	NOUN
fcis-15825	15	21	]	]	PUNCT
fcis-15825	15	22	.	.	PUNCT
fcis-15825	16	1	in	in	ADP
fcis-15825	16	2	reference	reference	NOUN
fcis-15825	16	3	[	[	X
fcis-15825	16	4	10	10	NUM
fcis-15825	16	5	]	]	PUNCT
fcis-15825	16	6	,	,	PUNCT
fcis-15825	16	7	the	the	DET
fcis-15825	16	8	faster	fast	ADJ
fcis-15825	16	9	rcnn	rcnn	PROPN
fcis-15825	16	10	model	model	PROPN
fcis-15825	16	11	was	be	AUX
fcis-15825	16	12	used	use	VERB
fcis-15825	16	13	to	to	PART
fcis-15825	16	14	detect	detect	VERB
fcis-15825	16	15	pavement	pavement	NOUN
fcis-15825	16	16	diseases	disease	NOUN
fcis-15825	16	17	,	,	PUNCT
fcis-15825	16	18	and	and	CCONJ
fcis-15825	16	19	the	the	DET
fcis-15825	16	20	regional	regional	ADJ
fcis-15825	16	21	expert	expert	NOUN
fcis-15825	16	22	network	network	NOUN
fcis-15825	16	23	was	be	AUX
fcis-15825	16	24	introduced	introduce	VERB
fcis-15825	16	25	to	to	PART
fcis-15825	16	26	improve	improve	VERB
fcis-15825	16	27	the	the	DET
fcis-15825	16	28	network	network	NOUN
fcis-15825	16	29	performance	performance	NOUN
fcis-15825	16	30	.	.	PUNCT
fcis-15825	17	1	the	the	DET
fcis-15825	17	2	two	two	NUM
fcis-15825	17	3	-	-	PUNCT
fcis-15825	17	4	stage	stage	NOUN
fcis-15825	17	5	algorithm	algorithm	NOUN
fcis-15825	17	6	has	have	VERB
fcis-15825	17	7	high	high	ADJ
fcis-15825	17	8	accuracy	accuracy	NOUN
fcis-15825	17	9	,	,	PUNCT
fcis-15825	17	10	but	but	CCONJ
fcis-15825	17	11	the	the	DET
fcis-15825	17	12	detection	detection	NOUN
fcis-15825	17	13	speed	speed	NOUN
fcis-15825	17	14	is	be	AUX
fcis-15825	17	15	slow	slow	ADJ
fcis-15825	17	16	.	.	PUNCT
fcis-15825	18	1	reference	reference	NOUN
fcis-15825	19	1	[	[	X
fcis-15825	19	2	11	11	NUM
fcis-15825	19	3	]	]	PUNCT
fcis-15825	19	4	used	use	VERB
fcis-15825	19	5	yolov3	yolov3	PROPN
fcis-15825	19	6	model	model	PROPN
fcis-15825	19	7	to	to	PART
fcis-15825	19	8	identify	identify	VERB
fcis-15825	19	9	pavement	pavement	NOUN
fcis-15825	19	10	cracks	crack	NOUN
fcis-15825	19	11	.	.	PUNCT
fcis-15825	20	1	this	this	DET
fcis-15825	20	2	model	model	NOUN
fcis-15825	20	3	can	can	AUX
fcis-15825	20	4	meet	meet	VERB
fcis-15825	20	5	the	the	DET
fcis-15825	20	6	real	real	ADJ
fcis-15825	20	7	-	-	PUNCT
fcis-15825	20	8	time	time	NOUN
fcis-15825	20	9	requirements	requirement	NOUN
fcis-15825	20	10	with	with	ADP
fcis-15825	20	11	high	high	ADJ
fcis-15825	20	12	accuracy	accuracy	NOUN
fcis-15825	20	13	,	,	PUNCT
fcis-15825	20	14	but	but	CCONJ
fcis-15825	20	15	there	there	PRON
fcis-15825	20	16	has	have	VERB
fcis-15825	20	17	a	a	DET
fcis-15825	20	18	problem	problem	NOUN
fcis-15825	20	19	of	of	ADP
fcis-15825	20	20	missing	miss	VERB
fcis-15825	20	21	detection	detection	NOUN
fcis-15825	20	22	.	.	PUNCT
fcis-15825	21	1	yu	yu	PROPN
fcis-15825	22	1	lu[12	lu[12	PROPN
fcis-15825	22	2	]	]	X
fcis-15825	22	3	introduced	introduce	VERB
fcis-15825	22	4	senet	senet	NOUN
fcis-15825	22	5	attention	attention	NOUN
fcis-15825	22	6	and	and	CCONJ
fcis-15825	22	7	aspp	aspp	NOUN
fcis-15825	22	8	module	module	NOUN
fcis-15825	22	9	on	on	ADP
fcis-15825	22	10	the	the	DET
fcis-15825	22	11	basis	basis	NOUN
fcis-15825	22	12	of	of	ADP
fcis-15825	22	13	yolov5	yolov5	NOUN
fcis-15825	22	14	model	model	NOUN
fcis-15825	22	15	,	,	PUNCT
fcis-15825	22	16	which	which	PRON
fcis-15825	22	17	effectively	effectively	ADV
fcis-15825	22	18	improved	improve	VERB
fcis-15825	22	19	the	the	DET
fcis-15825	22	20	accuracy	accuracy	NOUN
fcis-15825	22	21	of	of	ADP
fcis-15825	22	22	pavement	pavement	NOUN
fcis-15825	22	23	diseases	disease	NOUN
fcis-15825	22	24	detection	detection	NOUN
fcis-15825	22	25	model	model	NOUN
fcis-15825	22	26	.	.	PUNCT
fcis-15825	23	1	yolov5	yolov5	NOUN
fcis-15825	23	2	model	model	NOUN
fcis-15825	23	3	can	can	AUX
fcis-15825	23	4	balance	balance	VERB
fcis-15825	23	5	precision	precision	NOUN
fcis-15825	23	6	and	and	CCONJ
fcis-15825	23	7	speed	speed	NOUN
fcis-15825	23	8	well	well	ADV
fcis-15825	23	9	,	,	PUNCT
fcis-15825	23	10	and	and	CCONJ
fcis-15825	23	11	it	it	PRON
fcis-15825	23	12	is	be	AUX
fcis-15825	23	13	a	a	DET
fcis-15825	23	14	mature	mature	ADJ
fcis-15825	23	15	detection	detection	NOUN
fcis-15825	23	16	model	model	NOUN
fcis-15825	23	17	with	with	ADP
fcis-15825	23	18	good	good	ADJ
fcis-15825	23	19	performance	performance	NOUN
fcis-15825	23	20	at	at	ADP
fcis-15825	23	21	present	present	ADJ
fcis-15825	23	22	.	.	PUNCT
fcis-15825	24	1	however	however	ADV
fcis-15825	24	2	,	,	PUNCT
fcis-15825	24	3	the	the	DET
fcis-15825	24	4	ability	ability	NOUN
fcis-15825	24	5	to	to	PART
fcis-15825	24	6	extract	extract	VERB
fcis-15825	24	7	the	the	DET
fcis-15825	24	8	characteristics	characteristic	NOUN
fcis-15825	24	9	of	of	ADP
fcis-15825	24	10	small	small	ADJ
fcis-15825	24	11	road	road	NOUN
fcis-15825	24	12	diseases	disease	NOUN
fcis-15825	24	13	under	under	ADP
fcis-15825	24	14	complex	complex	ADJ
fcis-15825	24	15	background	background	NOUN
fcis-15825	24	16	needs	need	VERB
fcis-15825	24	17	to	to	PART
fcis-15825	24	18	be	be	AUX
fcis-15825	24	19	improved	improve	VERB
fcis-15825	24	20	,	,	PUNCT
fcis-15825	24	21	and	and	CCONJ
fcis-15825	24	22	it	it	PRON
fcis-15825	24	23	is	be	AUX
fcis-15825	24	24	necessary	necessary	ADJ
fcis-15825	24	25	to	to	PART
fcis-15825	24	26	further	far	ADV
fcis-15825	24	27	analyze	analyze	VERB
fcis-15825	24	28	the	the	DET
fcis-15825	24	29	characteristics	characteristic	NOUN
fcis-15825	24	30	of	of	ADP
fcis-15825	24	31	pavement	pavement	NOUN
fcis-15825	24	32	diseases	disease	NOUN
fcis-15825	24	33	and	and	CCONJ
fcis-15825	24	34	put	put	VERB
fcis-15825	24	35	forward	forward	ADV
fcis-15825	24	36	targeted	target	VERB
fcis-15825	24	37	improvement	improvement	NOUN
fcis-15825	24	38	measures	measure	NOUN
fcis-15825	24	39	.	.	PUNCT
fcis-15825	25	1	therefore	therefore	ADV
fcis-15825	25	2	,	,	PUNCT
fcis-15825	25	3	based	base	VERB
fcis-15825	25	4	of	of	ADP
fcis-15825	25	5	yolov5s	yolov5s	PROPN
fcis-15825	25	6	model	model	NOUN
fcis-15825	25	7	,	,	PUNCT
fcis-15825	25	8	this	this	DET
fcis-15825	25	9	paper	paper	NOUN
fcis-15825	25	10	introduced	introduce	VERB
fcis-15825	25	11	cbam	cbam	NOUN
fcis-15825	25	12	attention	attention	NOUN
fcis-15825	25	13	mechanism	mechanism	NOUN
fcis-15825	25	14	to	to	PART
fcis-15825	25	15	enhance	enhance	VERB
fcis-15825	25	16	the	the	DET
fcis-15825	25	17	feature	feature	NOUN
fcis-15825	25	18	extraction	extraction	NOUN
fcis-15825	25	19	ability	ability	NOUN
fcis-15825	25	20	of	of	ADP
fcis-15825	25	21	the	the	DET
fcis-15825	25	22	model	model	NOUN
fcis-15825	25	23	,	,	PUNCT
fcis-15825	25	24	used	use	VERB
fcis-15825	25	25	bifpn	bifpn	PROPN
fcis-15825	25	26	structure	structure	NOUN
fcis-15825	25	27	to	to	PART
fcis-15825	25	28	realize	realize	VERB
fcis-15825	25	29	crosslayer	crosslayer	NOUN
fcis-15825	25	30	fusion	fusion	NOUN
fcis-15825	25	31	of	of	ADP
fcis-15825	25	32	characteristic	characteristic	ADJ
fcis-15825	25	33	information	information	NOUN
fcis-15825	25	34	,	,	PUNCT
fcis-15825	25	35	and	and	CCONJ
fcis-15825	25	36	added	add	VERB
fcis-15825	25	37	a	a	DET
fcis-15825	25	38	small	small	ADJ
fcis-15825	25	39	target	target	NOUN
fcis-15825	25	40	detection	detection	NOUN
fcis-15825	25	41	layer	layer	NOUN
fcis-15825	25	42	to	to	PART
fcis-15825	25	43	detect	detect	VERB
fcis-15825	25	44	small	small	ADJ
fcis-15825	25	45	target	target	NOUN
fcis-15825	25	46	road	road	NOUN
fcis-15825	25	47	diseases	disease	NOUN
fcis-15825	25	48	,	,	PUNCT
fcis-15825	25	49	enhancing	enhance	VERB
fcis-15825	25	50	the	the	DET
fcis-15825	25	51	application	application	NOUN
fcis-15825	25	52	ability	ability	NOUN
fcis-15825	25	53	of	of	ADP
fcis-15825	25	54	the	the	DET
fcis-15825	25	55	model	model	NOUN
fcis-15825	25	56	in	in	ADP
fcis-15825	25	57	complex	complex	ADJ
fcis-15825	25	58	background	background	NOUN
fcis-15825	25	59	.	.	PUNCT
fcis-15825	26	1	2	2	X
fcis-15825	26	2	.	.	X
fcis-15825	26	3	yolov5s	yolov5s	PROPN
fcis-15825	26	4	model	model	NOUN
fcis-15825	26	5	unlike	unlike	ADP
fcis-15825	26	6	other	other	ADJ
fcis-15825	26	7	target	target	NOUN
fcis-15825	26	8	detection	detection	NOUN
fcis-15825	26	9	objects	object	NOUN
fcis-15825	26	10	,	,	PUNCT
fcis-15825	26	11	pavement	pavement	NOUN
fcis-15825	26	12	diseases	disease	NOUN
fcis-15825	26	13	have	have	VERB
fcis-15825	26	14	their	their	PRON
fcis-15825	26	15	own	own	ADJ
fcis-15825	26	16	characteristics	characteristic	NOUN
fcis-15825	26	17	.	.	PUNCT
fcis-15825	27	1	firstly	firstly	ADV
fcis-15825	27	2	,	,	PUNCT
fcis-15825	27	3	the	the	DET
fcis-15825	27	4	background	background	NOUN
fcis-15825	27	5	of	of	ADP
fcis-15825	27	6	pavement	pavement	NOUN
fcis-15825	27	7	diseases	disease	NOUN
fcis-15825	27	8	images	image	NOUN
fcis-15825	27	9	in	in	ADP
fcis-15825	27	10	the	the	DET
fcis-15825	27	11	actual	actual	ADJ
fcis-15825	27	12	scene	scene	NOUN
fcis-15825	27	13	is	be	AUX
fcis-15825	27	14	complex	complex	ADJ
fcis-15825	27	15	and	and	CCONJ
fcis-15825	27	16	noisy	noisy	ADJ
fcis-15825	27	17	,	,	PUNCT
fcis-15825	27	18	which	which	PRON
fcis-15825	27	19	usually	usually	ADV
fcis-15825	27	20	contains	contain	VERB
fcis-15825	27	21	interference	interference	NOUN
fcis-15825	27	22	information	information	NOUN
fcis-15825	27	23	such	such	ADJ
fcis-15825	27	24	as	as	ADP
fcis-15825	27	25	pedestrians	pedestrian	NOUN
fcis-15825	27	26	,	,	PUNCT
fcis-15825	27	27	vehicles	vehicle	NOUN
fcis-15825	27	28	,	,	PUNCT
fcis-15825	27	29	lane	lane	NOUN
fcis-15825	27	30	lines	line	NOUN
fcis-15825	27	31	,	,	PUNCT
fcis-15825	27	32	road	road	NOUN
fcis-15825	27	33	garbage	garbage	NOUN
fcis-15825	27	34	,	,	PUNCT
fcis-15825	27	35	etc	etc	X
fcis-15825	27	36	.	.	X
fcis-15825	27	37	moreover	moreover	ADV
fcis-15825	27	38	,	,	PUNCT
fcis-15825	27	39	due	due	ADP
fcis-15825	27	40	to	to	ADP
fcis-15825	27	41	factors	factor	NOUN
fcis-15825	27	42	such	such	ADJ
fcis-15825	27	43	as	as	ADP
fcis-15825	27	44	light	light	ADJ
fcis-15825	27	45	intensity	intensity	NOUN
fcis-15825	27	46	and	and	CCONJ
fcis-15825	27	47	weather	weather	NOUN
fcis-15825	27	48	,	,	PUNCT
fcis-15825	27	49	road	road	NOUN
fcis-15825	27	50	images	image	NOUN
fcis-15825	27	51	often	often	ADV
fcis-15825	27	52	appear	appear	VERB
fcis-15825	27	53	problems	problem	NOUN
fcis-15825	27	54	such	such	ADJ
fcis-15825	27	55	as	as	ADP
fcis-15825	27	56	exposure	exposure	NOUN
fcis-15825	27	57	and	and	CCONJ
fcis-15825	27	58	blurring	blurring	NOUN
fcis-15825	27	59	,	,	PUNCT
fcis-15825	27	60	which	which	PRON
fcis-15825	27	61	greatly	greatly	ADV
fcis-15825	27	62	affects	affect	VERB
fcis-15825	27	63	the	the	DET
fcis-15825	27	64	quality	quality	NOUN
fcis-15825	27	65	of	of	ADP
fcis-15825	27	66	acquired	acquire	VERB
fcis-15825	27	67	images	image	NOUN
fcis-15825	27	68	.	.	PUNCT
fcis-15825	28	1	second	second	ADJ
fcis-15825	28	2	,	,	PUNCT
fcis-15825	28	3	the	the	DET
fcis-15825	28	4	types	type	NOUN
fcis-15825	28	5	of	of	ADP
fcis-15825	28	6	pavement	pavement	NOUN
fcis-15825	28	7	diseases	disease	NOUN
fcis-15825	28	8	are	be	AUX
fcis-15825	28	9	complex	complex	ADJ
fcis-15825	28	10	and	and	CCONJ
fcis-15825	28	11	diverse	diverse	ADJ
fcis-15825	28	12	,	,	PUNCT
fcis-15825	28	13	the	the	DET
fcis-15825	28	14	area	area	NOUN
fcis-15825	28	15	gap	gap	NOUN
fcis-15825	28	16	of	of	ADP
fcis-15825	28	17	pavement	pavement	NOUN
fcis-15825	28	18	diseases	disease	NOUN
fcis-15825	28	19	is	be	AUX
fcis-15825	28	20	large	large	ADJ
fcis-15825	28	21	,	,	PUNCT
fcis-15825	28	22	and	and	CCONJ
fcis-15825	28	23	the	the	DET
fcis-15825	28	24	proportion	proportion	NOUN
fcis-15825	28	25	of	of	ADP
fcis-15825	28	26	road	road	NOUN
fcis-15825	28	27	diseases	disease	NOUN
fcis-15825	28	28	in	in	ADP
fcis-15825	28	29	the	the	DET
fcis-15825	28	30	image	image	NOUN
fcis-15825	28	31	is	be	AUX
fcis-15825	28	32	uneven	uneven	ADJ
fcis-15825	28	33	,	,	PUNCT
fcis-15825	28	34	which	which	PRON
fcis-15825	28	35	makes	make	VERB
fcis-15825	28	36	the	the	DET
fcis-15825	28	37	width	width	ADJ
fcis-15825	28	38	-	-	PUNCT
fcis-15825	28	39	height	height	NOUN
fcis-15825	28	40	ratio	ratio	NOUN
fcis-15825	28	41	of	of	ADP
fcis-15825	28	42	the	the	DET
fcis-15825	28	43	anchor	anchor	NOUN
fcis-15825	28	44	frame	frame	NOUN
fcis-15825	28	45	different	different	ADJ
fcis-15825	28	46	from	from	ADP
fcis-15825	28	47	other	other	ADJ
fcis-15825	28	48	data	data	NOUN
fcis-15825	28	49	sets	set	NOUN
fcis-15825	28	50	,	,	PUNCT
fcis-15825	28	51	so	so	SCONJ
fcis-15825	28	52	it	it	PRON
fcis-15825	28	53	is	be	AUX
fcis-15825	28	54	necessary	necessary	ADJ
fcis-15825	28	55	to	to	PART
fcis-15825	28	56	retrain	retrain	VERB
fcis-15825	28	57	the	the	DET
fcis-15825	28	58	anchor	anchor	NOUN
fcis-15825	28	59	frame	frame	NOUN
fcis-15825	28	60	according	accord	VERB
fcis-15825	28	61	to	to	ADP
fcis-15825	28	62	the	the	DET
fcis-15825	28	63	road	road	NOUN
fcis-15825	28	64	datasets	dataset	NOUN
fcis-15825	28	65	;	;	PUNCT
fcis-15825	28	66	in	in	ADP
fcis-15825	28	67	the	the	DET
fcis-15825	28	68	end	end	NOUN
fcis-15825	28	69	,	,	PUNCT
fcis-15825	28	70	the	the	DET
fcis-15825	28	71	existing	exist	VERB
fcis-15825	28	72	open	open	ADJ
fcis-15825	28	73	source	source	NOUN
fcis-15825	28	74	datasets	dataset	NOUN
fcis-15825	28	75	of	of	ADP
fcis-15825	28	76	pavement	pavement	NOUN
fcis-15825	28	77	diseases	disease	NOUN
fcis-15825	28	78	contains	contain	VERB
fcis-15825	28	79	limited	limited	ADJ
fcis-15825	28	80	types	type	NOUN
fcis-15825	28	81	of	of	ADP
fcis-15825	28	82	pavement	pavement	NOUN
fcis-15825	28	83	diseases	disease	NOUN
fcis-15825	28	84	,	,	PUNCT
fcis-15825	28	85	and	and	CCONJ
fcis-15825	28	86	the	the	DET
fcis-15825	28	87	labeling	labeling	NOUN
fcis-15825	28	88	of	of	ADP
fcis-15825	28	89	pavement	pavement	NOUN
fcis-15825	28	90	diseases	disease	NOUN
fcis-15825	28	91	is	be	AUX
fcis-15825	28	92	not	not	PART
fcis-15825	28	93	ideal	ideal	ADJ
fcis-15825	28	94	,	,	PUNCT
fcis-15825	28	95	and	and	CCONJ
fcis-15825	28	96	there	there	PRON
fcis-15825	28	97	are	be	VERB
fcis-15825	28	98	some	some	DET
fcis-15825	28	99	problems	problem	NOUN
fcis-15825	28	100	such	such	ADJ
fcis-15825	28	101	as	as	ADP
fcis-15825	28	102	labeling	labeling	NOUN
fcis-15825	28	103	errors	error	NOUN
fcis-15825	28	104	,	,	PUNCT
fcis-15825	28	105	omissions	omission	NOUN
fcis-15825	28	106	and	and	CCONJ
fcis-15825	28	107	the	the	DET
fcis-15825	28	108	position	position	NOUN
fcis-15825	28	109	deviation	deviation	NOUN
fcis-15825	28	110	of	of	ADP
fcis-15825	28	111	labeling	labeling	NOUN
fcis-15825	28	112	boxes	box	NOUN
fcis-15825	28	113	.	.	PUNCT
fcis-15825	29	1	as	as	ADP
fcis-15825	29	2	a	a	DET
fcis-15825	29	3	result	result	NOUN
fcis-15825	29	4	,	,	PUNCT
fcis-15825	29	5	although	although	SCONJ
fcis-15825	29	6	yolov5s	yolov5s	PROPN
fcis-15825	29	7	model	model	NOUN
fcis-15825	29	8	has	have	AUX
fcis-15825	29	9	been	be	AUX
fcis-15825	29	10	widely	widely	ADV
fcis-15825	29	11	used	use	VERB
fcis-15825	29	12	in	in	ADP
fcis-15825	29	13	pavement	pavement	NOUN
fcis-15825	29	14	diseases	disease	NOUN
fcis-15825	29	15	detection	detection	NOUN
fcis-15825	29	16	,	,	PUNCT
fcis-15825	29	17	there	there	PRON
fcis-15825	29	18	are	be	VERB
fcis-15825	29	19	still	still	ADV
fcis-15825	29	20	some	some	DET
fcis-15825	29	21	problems	problem	NOUN
fcis-15825	29	22	,	,	PUNCT
fcis-15825	29	23	such	such	ADJ
fcis-15825	29	24	as	as	ADP
fcis-15825	29	25	false	false	ADJ
fcis-15825	29	26	detection	detection	NOUN
fcis-15825	29	27	,	,	PUNCT
fcis-15825	29	28	missed	miss	VERB
fcis-15825	29	29	detection	detection	NOUN
fcis-15825	29	30	and	and	CCONJ
fcis-15825	29	31	low	low	ADJ
fcis-15825	29	32	accuracy	accuracy	NOUN
fcis-15825	29	33	of	of	ADP
fcis-15825	29	34	small	small	ADJ
fcis-15825	29	35	target	target	NOUN
fcis-15825	29	36	detection	detection	NOUN
fcis-15825	29	37	.	.	PUNCT
fcis-15825	30	1	therefore	therefore	ADV
fcis-15825	30	2	,	,	PUNCT
fcis-15825	30	3	the	the	DET
fcis-15825	30	4	road	road	NOUN
fcis-15825	30	5	disease	disease	NOUN
fcis-15825	30	6	detection	detection	NOUN
fcis-15825	30	7	model	model	NOUN
fcis-15825	30	8	in	in	ADP
fcis-15825	30	9	actual	actual	ADJ
fcis-15825	30	10	scenes	scene	NOUN
fcis-15825	30	11	needs	need	VERB
fcis-15825	30	12	further	further	ADJ
fcis-15825	30	13	adjustment	adjustment	NOUN
fcis-15825	30	14	and	and	CCONJ
fcis-15825	30	15	improvement	improvement	NOUN
fcis-15825	30	16	.	.	PUNCT
fcis-15825	31	1	2.1	2.1	NUM
fcis-15825	31	2	.	.	PUNCT
fcis-15825	32	1	yolov5s	yolov5s	NOUN
fcis-15825	32	2	model	model	NOUN
fcis-15825	32	3	yolov5	yolov5	NOUN
fcis-15825	32	4	is	be	AUX
fcis-15825	32	5	a	a	DET
fcis-15825	32	6	representative	representative	ADJ
fcis-15825	32	7	model	model	NOUN
fcis-15825	32	8	of	of	ADP
fcis-15825	32	9	one	one	NUM
fcis-15825	32	10	-	-	PUNCT
fcis-15825	32	11	stage	stage	NOUN
fcis-15825	32	12	detection	detection	NOUN
fcis-15825	32	13	algorithm	algorithm	NOUN
fcis-15825	32	14	,	,	PUNCT
fcis-15825	32	15	which	which	PRON
fcis-15825	32	16	is	be	AUX
fcis-15825	32	17	widely	widely	ADV
fcis-15825	32	18	used	use	VERB
fcis-15825	32	19	in	in	ADP
fcis-15825	32	20	various	various	ADJ
fcis-15825	32	21	fields	field	NOUN
fcis-15825	32	22	and	and	CCONJ
fcis-15825	32	23	has	have	VERB
fcis-15825	32	24	good	good	ADJ
fcis-15825	32	25	detection	detection	NOUN
fcis-15825	32	26	performance	performance	NOUN
fcis-15825	32	27	.	.	PUNCT
fcis-15825	33	1	yolov5	yolov5	NOUN
fcis-15825	33	2	has	have	VERB
fcis-15825	33	3	four	four	NUM
fcis-15825	33	4	different	different	ADJ
fcis-15825	33	5	versions	version	NOUN
fcis-15825	33	6	,	,	PUNCT
fcis-15825	33	7	namely	namely	ADV
fcis-15825	33	8	yolov5s	yolov5s	PROPN
fcis-15825	33	9	,	,	PUNCT
fcis-15825	33	10	yolov5	yolov5	PROPN
fcis-15825	33	11	m	m	PROPN
fcis-15825	33	12	,	,	PUNCT
fcis-15825	33	13	yolov5l	yolov5l	PROPN
fcis-15825	33	14	and	and	CCONJ
fcis-15825	33	15	yolov5x	yolov5x	PROPN
fcis-15825	33	16	.	.	PUNCT
fcis-15825	34	1	the	the	DET
fcis-15825	34	2	difference	difference	NOUN
fcis-15825	34	3	between	between	ADP
fcis-15825	34	4	these	these	DET
fcis-15825	34	5	four	four	NUM
fcis-15825	34	6	versions	version	NOUN
fcis-15825	34	7	is	be	AUX
fcis-15825	34	8	that	that	SCONJ
fcis-15825	34	9	the	the	DET
fcis-15825	34	10	depth	depth	NOUN
fcis-15825	34	11	and	and	CCONJ
fcis-15825	34	12	width	width	NOUN
fcis-15825	34	13	of	of	ADP
fcis-15825	34	14	the	the	DET
fcis-15825	34	15	network	network	NOUN
fcis-15825	34	16	increase	increase	NOUN
fcis-15825	34	17	in	in	ADP
fcis-15825	34	18	turn	turn	NOUN
fcis-15825	34	19	.	.	PUNCT
fcis-15825	35	1	as	as	SCONJ
fcis-15825	35	2	shown	show	VERB
fcis-15825	35	3	in	in	ADP
fcis-15825	35	4	figure	figure	NOUN
fcis-15825	35	5	1	1	NUM
fcis-15825	35	6	,	,	PUNCT
fcis-15825	35	7	114	114	NUM
fcis-15825	35	8	yolov5s	yolov5s	PROPN
fcis-15825	35	9	model	model	NOUN
fcis-15825	35	10	is	be	AUX
fcis-15825	35	11	the	the	DET
fcis-15825	35	12	lightest	light	ADJ
fcis-15825	35	13	network	network	NOUN
fcis-15825	35	14	in	in	ADP
fcis-15825	35	15	yolov5	yolov5	NOUN
fcis-15825	35	16	model	model	PROPN
fcis-15825	35	17	,	,	PUNCT
fcis-15825	35	18	which	which	PRON
fcis-15825	35	19	includes	include	VERB
fcis-15825	35	20	four	four	NUM
fcis-15825	35	21	parts	part	NOUN
fcis-15825	35	22	:	:	PUNCT
fcis-15825	35	23	input	input	NOUN
fcis-15825	35	24	,	,	PUNCT
fcis-15825	35	25	backbone	backbone	NOUN
fcis-15825	35	26	,	,	PUNCT
fcis-15825	35	27	neck	neck	NOUN
fcis-15825	35	28	and	and	CCONJ
fcis-15825	35	29	prediction	prediction	NOUN
fcis-15825	35	30	head	head	NOUN
fcis-15825	35	31	.	.	PUNCT
fcis-15825	36	1	at	at	ADP
fcis-15825	36	2	the	the	DET
fcis-15825	36	3	input	input	NOUN
fcis-15825	36	4	,	,	PUNCT
fcis-15825	36	5	the	the	DET
fcis-15825	36	6	image	image	NOUN
fcis-15825	36	7	data	datum	NOUN
fcis-15825	36	8	was	be	AUX
fcis-15825	36	9	adaptively	adaptively	ADV
fcis-15825	36	10	scaled	scale	VERB
fcis-15825	36	11	to	to	ADP
fcis-15825	36	12	a	a	DET
fcis-15825	36	13	uniform	uniform	ADJ
fcis-15825	36	14	size	size	NOUN
fcis-15825	36	15	:	:	PUNCT
fcis-15825	36	16	608	608	NUM
fcis-15825	36	17	*	*	NUM
fcis-15825	36	18	608	608	NUM
fcis-15825	36	19	,	,	PUNCT
fcis-15825	36	20	and	and	CCONJ
fcis-15825	36	21	the	the	DET
fcis-15825	36	22	mosaic	mosaic	ADJ
fcis-15825	36	23	data	data	NOUN
fcis-15825	36	24	enhancement	enhancement	NOUN
fcis-15825	36	25	was	be	AUX
fcis-15825	36	26	used	use	VERB
fcis-15825	36	27	to	to	PART
fcis-15825	36	28	expand	expand	VERB
fcis-15825	36	29	the	the	DET
fcis-15825	36	30	datasets	dataset	NOUN
fcis-15825	36	31	.	.	PUNCT
fcis-15825	37	1	backbone	backbone	NOUN
fcis-15825	37	2	network	network	NOUN
fcis-15825	37	3	was	be	AUX
fcis-15825	37	4	responsible	responsible	ADJ
fcis-15825	37	5	for	for	ADP
fcis-15825	37	6	feature	feature	NOUN
fcis-15825	37	7	extraction	extraction	NOUN
fcis-15825	37	8	of	of	ADP
fcis-15825	37	9	image	image	NOUN
fcis-15825	37	10	data	datum	NOUN
fcis-15825	37	11	to	to	PART
fcis-15825	37	12	obtain	obtain	VERB
fcis-15825	37	13	high	high	ADJ
fcis-15825	37	14	-	-	PUNCT
fcis-15825	37	15	level	level	NOUN
fcis-15825	37	16	semantic	semantic	ADJ
fcis-15825	37	17	and	and	CCONJ
fcis-15825	37	18	shallow	shallow	ADJ
fcis-15825	37	19	features	feature	NOUN
fcis-15825	37	20	of	of	ADP
fcis-15825	37	21	image	image	NOUN
fcis-15825	37	22	data	datum	NOUN
fcis-15825	37	23	;	;	PUNCT
fcis-15825	37	24	the	the	DET
fcis-15825	37	25	neck	neck	NOUN
fcis-15825	37	26	network	network	NOUN
fcis-15825	37	27	realizes	realize	VERB
fcis-15825	37	28	the	the	DET
fcis-15825	37	29	exchange	exchange	NOUN
fcis-15825	37	30	and	and	CCONJ
fcis-15825	37	31	fusion	fusion	NOUN
fcis-15825	37	32	of	of	ADP
fcis-15825	37	33	different	different	ADJ
fcis-15825	37	34	levels	level	NOUN
fcis-15825	37	35	of	of	ADP
fcis-15825	37	36	characteristic	characteristic	ADJ
fcis-15825	37	37	information	information	NOUN
fcis-15825	37	38	through	through	ADP
fcis-15825	37	39	upsampling	upsampling	NOUN
fcis-15825	37	40	;	;	PUNCT
fcis-15825	37	41	the	the	DET
fcis-15825	37	42	prediction	prediction	NOUN
fcis-15825	37	43	head	head	NOUN
fcis-15825	37	44	includes	include	VERB
fcis-15825	37	45	three	three	NUM
fcis-15825	37	46	prediction	prediction	NOUN
fcis-15825	37	47	layers	layer	NOUN
fcis-15825	37	48	for	for	ADP
fcis-15825	37	49	feature	feature	NOUN
fcis-15825	37	50	map	map	NOUN
fcis-15825	37	51	prediction	prediction	NOUN
fcis-15825	37	52	.	.	PUNCT
fcis-15825	38	1	fig	fig	NOUN
fcis-15825	38	2	1	1	NUM
fcis-15825	38	3	.	.	PUNCT
fcis-15825	39	1	yolov5s	yolov5s	NOUN
fcis-15825	39	2	network	network	NOUN
fcis-15825	39	3	structure	structure	NOUN
fcis-15825	39	4	2.2	2.2	NUM
fcis-15825	39	5	.	.	PUNCT
fcis-15825	39	6	small	small	ADJ
fcis-15825	39	7	target	target	NOUN
fcis-15825	39	8	detection	detection	NOUN
fcis-15825	39	9	layer	layer	NOUN
fcis-15825	39	10	in	in	ADP
fcis-15825	39	11	the	the	DET
fcis-15825	39	12	original	original	ADJ
fcis-15825	39	13	yolov5s	yolov5s	PROPN
fcis-15825	39	14	model	model	NOUN
fcis-15825	39	15	,	,	PUNCT
fcis-15825	39	16	the	the	DET
fcis-15825	39	17	detector	detector	NOUN
fcis-15825	39	18	includes	include	VERB
fcis-15825	39	19	three	three	NUM
fcis-15825	39	20	detection	detection	NOUN
fcis-15825	39	21	layers	layer	NOUN
fcis-15825	39	22	:	:	PUNCT
fcis-15825	39	23	p3	p3	PROPN
fcis-15825	39	24	,	,	PUNCT
fcis-15825	39	25	p4	p4	ADJ
fcis-15825	39	26	and	and	CCONJ
fcis-15825	39	27	p5	p5	PROPN
fcis-15825	39	28	,	,	PUNCT
fcis-15825	39	29	which	which	PRON
fcis-15825	39	30	detect	detect	VERB
fcis-15825	39	31	small	small	ADJ
fcis-15825	39	32	,	,	PUNCT
fcis-15825	39	33	medium	medium	ADJ
fcis-15825	39	34	and	and	CCONJ
fcis-15825	39	35	large	large	ADJ
fcis-15825	39	36	targets	target	NOUN
fcis-15825	39	37	by	by	ADP
fcis-15825	39	38	8	8	NUM
fcis-15825	39	39	,	,	PUNCT
fcis-15825	39	40	16	16	NUM
fcis-15825	39	41	and	and	CCONJ
fcis-15825	39	42	32	32	NUM
fcis-15825	39	43	times	time	NOUN
fcis-15825	39	44	down	down	ADV
fcis-15825	39	45	sampling	sample	VERB
fcis-15825	39	46	respectively	respectively	ADV
fcis-15825	39	47	,	,	PUNCT
fcis-15825	39	48	and	and	CCONJ
fcis-15825	39	49	the	the	DET
fcis-15825	39	50	detection	detection	NOUN
fcis-15825	39	51	effect	effect	NOUN
fcis-15825	39	52	of	of	ADP
fcis-15825	39	53	large	large	ADJ
fcis-15825	39	54	targets	target	NOUN
fcis-15825	39	55	is	be	AUX
fcis-15825	39	56	better	well	ADJ
fcis-15825	39	57	than	than	ADP
fcis-15825	39	58	that	that	PRON
fcis-15825	39	59	of	of	ADP
fcis-15825	39	60	small	small	ADJ
fcis-15825	39	61	targets	target	NOUN
fcis-15825	39	62	.	.	PUNCT
fcis-15825	40	1	in	in	ADP
fcis-15825	40	2	the	the	DET
fcis-15825	40	3	datasets	dataset	NOUN
fcis-15825	40	4	of	of	ADP
fcis-15825	40	5	this	this	DET
fcis-15825	40	6	paper	paper	NOUN
fcis-15825	40	7	,	,	PUNCT
fcis-15825	40	8	there	there	PRON
fcis-15825	40	9	are	be	VERB
fcis-15825	40	10	not	not	PART
fcis-15825	40	11	only	only	ADV
fcis-15825	40	12	large	large	ADJ
fcis-15825	40	13	cracks	crack	NOUN
fcis-15825	40	14	and	and	CCONJ
fcis-15825	40	15	net	net	ADJ
fcis-15825	40	16	cracks	crack	NOUN
fcis-15825	40	17	,	,	PUNCT
fcis-15825	40	18	but	but	CCONJ
fcis-15825	40	19	also	also	ADV
fcis-15825	40	20	small	small	ADJ
fcis-15825	40	21	pits	pit	NOUN
fcis-15825	40	22	and	and	CCONJ
fcis-15825	40	23	cracks	crack	NOUN
fcis-15825	40	24	that	that	PRON
fcis-15825	40	25	are	be	AUX
fcis-15825	40	26	difficult	difficult	ADJ
fcis-15825	40	27	to	to	PART
fcis-15825	40	28	detect	detect	VERB
fcis-15825	40	29	.	.	PUNCT
fcis-15825	41	1	there	there	PRON
fcis-15825	41	2	is	be	VERB
fcis-15825	41	3	a	a	DET
fcis-15825	41	4	big	big	ADJ
fcis-15825	41	5	difference	difference	NOUN
fcis-15825	41	6	in	in	ADP
fcis-15825	41	7	the	the	DET
fcis-15825	41	8	proportion	proportion	NOUN
fcis-15825	41	9	of	of	ADP
fcis-15825	41	10	road	road	NOUN
fcis-15825	41	11	diseases	disease	NOUN
fcis-15825	41	12	,	,	PUNCT
fcis-15825	41	13	and	and	CCONJ
fcis-15825	41	14	the	the	DET
fcis-15825	41	15	original	original	ADJ
fcis-15825	41	16	three	three	NUM
fcis-15825	41	17	detection	detection	NOUN
fcis-15825	41	18	layers	layer	NOUN
fcis-15825	41	19	can	can	AUX
fcis-15825	41	20	not	not	PART
fcis-15825	41	21	meet	meet	VERB
fcis-15825	41	22	the	the	DET
fcis-15825	41	23	detection	detection	NOUN
fcis-15825	41	24	requirements	requirement	NOUN
fcis-15825	41	25	,	,	PUNCT
fcis-15825	41	26	which	which	PRON
fcis-15825	41	27	leads	lead	VERB
fcis-15825	41	28	to	to	ADP
fcis-15825	41	29	the	the	DET
fcis-15825	41	30	problem	problem	NOUN
fcis-15825	41	31	of	of	ADP
fcis-15825	41	32	missed	miss	VERB
fcis-15825	41	33	detection	detection	NOUN
fcis-15825	41	34	of	of	ADP
fcis-15825	41	35	small	small	ADJ
fcis-15825	41	36	target	target	NOUN
fcis-15825	41	37	diseases	disease	NOUN
fcis-15825	41	38	in	in	ADP
fcis-15825	41	39	the	the	DET
fcis-15825	41	40	detection	detection	NOUN
fcis-15825	41	41	process	process	NOUN
fcis-15825	41	42	.	.	PUNCT
fcis-15825	42	1	in	in	ADP
fcis-15825	42	2	this	this	DET
fcis-15825	42	3	paper	paper	NOUN
fcis-15825	42	4	,	,	PUNCT
fcis-15825	42	5	on	on	ADP
fcis-15825	42	6	the	the	DET
fcis-15825	42	7	basis	basis	NOUN
fcis-15825	42	8	of	of	ADP
fcis-15825	42	9	the	the	DET
fcis-15825	42	10	original	original	ADJ
fcis-15825	42	11	three	three	NUM
fcis-15825	42	12	detection	detection	NOUN
fcis-15825	42	13	layers	layer	NOUN
fcis-15825	42	14	,	,	PUNCT
fcis-15825	42	15	the	the	DET
fcis-15825	42	16	p2	p2	NOUN
fcis-15825	42	17	detection	detection	NOUN
fcis-15825	42	18	layer	layer	NOUN
fcis-15825	42	19	with	with	ADP
fcis-15825	42	20	the	the	DET
fcis-15825	42	21	scale	scale	NOUN
fcis-15825	42	22	of	of	ADP
fcis-15825	42	23	152	152	NUM
fcis-15825	42	24	*	*	SYM
fcis-15825	42	25	152	152	NUM
fcis-15825	42	26	is	be	AUX
fcis-15825	42	27	added	add	VERB
fcis-15825	42	28	,	,	PUNCT
fcis-15825	42	29	which	which	PRON
fcis-15825	42	30	mainly	mainly	ADV
fcis-15825	42	31	detects	detect	VERB
fcis-15825	42	32	small	small	ADJ
fcis-15825	42	33	pavement	pavement	NOUN
fcis-15825	42	34	diseases	disease	NOUN
fcis-15825	42	35	.	.	PUNCT
fcis-15825	43	1	detection	detection	NOUN
fcis-15825	43	2	layer	layer	NOUN
fcis-15825	43	3	p2	p2	PROPN
fcis-15825	43	4	has	have	VERB
fcis-15825	43	5	a	a	DET
fcis-15825	43	6	smaller	small	ADJ
fcis-15825	43	7	receptive	receptive	ADJ
fcis-15825	43	8	field	field	NOUN
fcis-15825	43	9	,	,	PUNCT
fcis-15825	43	10	which	which	PRON
fcis-15825	43	11	can	can	AUX
fcis-15825	43	12	retain	retain	VERB
fcis-15825	43	13	more	more	ADJ
fcis-15825	43	14	position	position	NOUN
fcis-15825	43	15	information	information	NOUN
fcis-15825	43	16	and	and	CCONJ
fcis-15825	43	17	contour	contour	NOUN
fcis-15825	43	18	features	feature	NOUN
fcis-15825	43	19	of	of	ADP
fcis-15825	43	20	small	small	ADJ
fcis-15825	43	21	targets	target	NOUN
fcis-15825	43	22	.	.	PUNCT
fcis-15825	44	1	it	it	PRON
fcis-15825	44	2	can	can	AUX
fcis-15825	44	3	significantly	significantly	ADV
fcis-15825	44	4	improve	improve	VERB
fcis-15825	44	5	the	the	DET
fcis-15825	44	6	network	network	NOUN
fcis-15825	44	7	's	's	PART
fcis-15825	44	8	ability	ability	NOUN
fcis-15825	44	9	to	to	PART
fcis-15825	44	10	detect	detect	VERB
fcis-15825	44	11	small	small	ADJ
fcis-15825	44	12	target	target	NOUN
fcis-15825	44	13	diseases	disease	NOUN
fcis-15825	44	14	and	and	CCONJ
fcis-15825	44	15	effectively	effectively	ADV
fcis-15825	44	16	solve	solve	VERB
fcis-15825	44	17	the	the	DET
fcis-15825	44	18	negative	negative	ADJ
fcis-15825	44	19	impact	impact	NOUN
fcis-15825	44	20	of	of	ADP
fcis-15825	44	21	target	target	NOUN
fcis-15825	44	22	scale	scale	NOUN
fcis-15825	44	23	differences	difference	NOUN
fcis-15825	44	24	on	on	ADP
fcis-15825	44	25	the	the	DET
fcis-15825	44	26	model	model	NOUN
fcis-15825	44	27	.	.	PUNCT
fcis-15825	45	1	in	in	ADP
fcis-15825	45	2	addition	addition	NOUN
fcis-15825	45	3	,	,	PUNCT
fcis-15825	45	4	the	the	DET
fcis-15825	45	5	anchor	anchor	NOUN
fcis-15825	45	6	frame	frame	NOUN
fcis-15825	45	7	mechanism	mechanism	NOUN
fcis-15825	45	8	of	of	ADP
fcis-15825	45	9	the	the	DET
fcis-15825	45	10	original	original	ADJ
fcis-15825	45	11	yolov5s	yolov5s	PROPN
fcis-15825	45	12	model	model	NOUN
fcis-15825	45	13	was	be	AUX
fcis-15825	45	14	trained	train	VERB
fcis-15825	45	15	based	base	VERB
fcis-15825	45	16	on	on	ADP
fcis-15825	45	17	coco	coco	PROPN
fcis-15825	45	18	data	data	PROPN
fcis-15825	45	19	set	set	PROPN
fcis-15825	45	20	,	,	PUNCT
fcis-15825	45	21	which	which	PRON
fcis-15825	45	22	has	have	VERB
fcis-15825	45	23	different	different	ADJ
fcis-15825	45	24	detection	detection	NOUN
fcis-15825	45	25	target	target	NOUN
fcis-15825	45	26	area	area	NOUN
fcis-15825	45	27	with	with	ADP
fcis-15825	45	28	the	the	DET
fcis-15825	45	29	pavement	pavement	NOUN
fcis-15825	45	30	diseases	disease	NOUN
fcis-15825	45	31	detection	detection	NOUN
fcis-15825	45	32	in	in	ADP
fcis-15825	45	33	this	this	DET
fcis-15825	45	34	study	study	NOUN
fcis-15825	45	35	,	,	PUNCT
fcis-15825	45	36	which	which	PRON
fcis-15825	45	37	is	be	AUX
fcis-15825	45	38	not	not	PART
fcis-15825	45	39	suitable	suitable	ADJ
fcis-15825	45	40	for	for	ADP
fcis-15825	45	41	the	the	DET
fcis-15825	45	42	actual	actual	ADJ
fcis-15825	45	43	needs	need	NOUN
fcis-15825	45	44	of	of	ADP
fcis-15825	45	45	pavement	pavement	NOUN
fcis-15825	45	46	diseases	disease	NOUN
fcis-15825	45	47	detection	detection	NOUN
fcis-15825	45	48	in	in	ADP
fcis-15825	45	49	this	this	DET
fcis-15825	45	50	study	study	NOUN
fcis-15825	45	51	.	.	PUNCT
fcis-15825	46	1	therefore	therefore	ADV
fcis-15825	46	2	,	,	PUNCT
fcis-15825	46	3	it	it	PRON
fcis-15825	46	4	is	be	AUX
fcis-15825	46	5	necessary	necessary	ADJ
fcis-15825	46	6	to	to	PART
fcis-15825	46	7	design	design	VERB
fcis-15825	46	8	the	the	DET
fcis-15825	46	9	target	target	NOUN
fcis-15825	46	10	anchor	anchor	NOUN
fcis-15825	46	11	frame	frame	NOUN
fcis-15825	46	12	size	size	NOUN
fcis-15825	46	13	by	by	ADP
fcis-15825	46	14	reclustering	reclustere	VERB
fcis-15825	46	15	.	.	PUNCT
fcis-15825	47	1	on	on	ADP
fcis-15825	47	2	the	the	DET
fcis-15825	47	3	data	datum	NOUN
fcis-15825	47	4	set	set	VERB
fcis-15825	47	5	constructed	construct	VERB
fcis-15825	47	6	in	in	ADP
fcis-15825	47	7	this	this	DET
fcis-15825	47	8	paper	paper	NOUN
fcis-15825	47	9	,	,	PUNCT
fcis-15825	47	10	the	the	DET
fcis-15825	47	11	kmeans++	kmeans++	PROPN
fcis-15825	47	12	method	method	NOUN
fcis-15825	47	13	was	be	AUX
fcis-15825	47	14	used	use	VERB
fcis-15825	47	15	to	to	PART
fcis-15825	47	16	recombine	recombine	VERB
fcis-15825	47	17	the	the	DET
fcis-15825	47	18	classes	class	NOUN
fcis-15825	47	19	,	,	PUNCT
fcis-15825	47	20	and	and	CCONJ
fcis-15825	47	21	the	the	DET
fcis-15825	47	22	anchor	anchor	NOUN
fcis-15825	47	23	frame	frame	NOUN
fcis-15825	47	24	size	size	NOUN
fcis-15825	47	25	that	that	PRON
fcis-15825	47	26	can	can	AUX
fcis-15825	47	27	better	well	ADV
fcis-15825	47	28	match	match	VERB
fcis-15825	47	29	the	the	DET
fcis-15825	47	30	pavement	pavement	NOUN
fcis-15825	47	31	diseases	disease	NOUN
fcis-15825	47	32	target	target	NOUN
fcis-15825	47	33	is	be	AUX
fcis-15825	47	34	obtained	obtain	VERB
fcis-15825	47	35	.	.	PUNCT
fcis-15825	48	1	2.3	2.3	NUM
fcis-15825	48	2	.	.	PUNCT
fcis-15825	49	1	cbam	cbam	NOUN
fcis-15825	49	2	attention	attention	NOUN
fcis-15825	49	3	mechanism	mechanism	NOUN
fcis-15825	49	4	the	the	DET
fcis-15825	49	5	attention	attention	NOUN
fcis-15825	49	6	mechanism	mechanism	NOUN
fcis-15825	49	7	is	be	AUX
fcis-15825	49	8	an	an	DET
fcis-15825	49	9	imitation	imitation	NOUN
fcis-15825	49	10	of	of	ADP
fcis-15825	49	11	human	human	ADJ
fcis-15825	49	12	visual	visual	ADJ
fcis-15825	49	13	system	system	NOUN
fcis-15825	49	14	,	,	PUNCT
fcis-15825	49	15	which	which	PRON
fcis-15825	49	16	can	can	AUX
fcis-15825	49	17	focus	focus	VERB
fcis-15825	49	18	on	on	ADP
fcis-15825	49	19	pavement	pavement	NOUN
fcis-15825	49	20	diseases	disease	NOUN
fcis-15825	49	21	information	information	NOUN
fcis-15825	49	22	in	in	ADP
fcis-15825	49	23	the	the	DET
fcis-15825	49	24	process	process	NOUN
fcis-15825	49	25	of	of	ADP
fcis-15825	49	26	feature	feature	NOUN
fcis-15825	49	27	extraction	extraction	NOUN
fcis-15825	49	28	and	and	CCONJ
fcis-15825	49	29	improve	improve	VERB
fcis-15825	49	30	the	the	DET
fcis-15825	49	31	feature	feature	NOUN
fcis-15825	49	32	extraction	extraction	NOUN
fcis-15825	49	33	ability	ability	NOUN
fcis-15825	49	34	of	of	ADP
fcis-15825	49	35	the	the	DET
fcis-15825	49	36	model	model	NOUN
fcis-15825	49	37	.	.	PUNCT
fcis-15825	50	1	on	on	ADP
fcis-15825	50	2	the	the	DET
fcis-15825	50	3	road	road	NOUN
fcis-15825	50	4	disease	disease	NOUN
fcis-15825	50	5	datasets	dataset	NOUN
fcis-15825	50	6	,	,	PUNCT
fcis-15825	50	7	the	the	DET
fcis-15825	50	8	background	background	NOUN
fcis-15825	50	9	is	be	AUX
fcis-15825	50	10	complex	complex	ADJ
fcis-15825	50	11	and	and	CCONJ
fcis-15825	50	12	the	the	DET
fcis-15825	50	13	colors	color	NOUN
fcis-15825	50	14	are	be	AUX
fcis-15825	50	15	diverse	diverse	ADJ
fcis-15825	50	16	,	,	PUNCT
fcis-15825	50	17	and	and	CCONJ
fcis-15825	50	18	the	the	DET
fcis-15825	50	19	image	image	NOUN
fcis-15825	50	20	data	datum	NOUN
fcis-15825	50	21	containing	contain	VERB
fcis-15825	50	22	a	a	DET
fcis-15825	50	23	lot	lot	NOUN
fcis-15825	50	24	of	of	ADP
fcis-15825	50	25	unimportant	unimportant	ADJ
fcis-15825	50	26	semantic	semantic	ADJ
fcis-15825	50	27	information	information	NOUN
fcis-15825	50	28	accounts	account	VERB
fcis-15825	50	29	for	for	ADP
fcis-15825	50	30	a	a	DET
fcis-15825	50	31	large	large	ADJ
fcis-15825	50	32	proportion	proportion	NOUN
fcis-15825	50	33	,	,	PUNCT
fcis-15825	50	34	which	which	PRON
fcis-15825	50	35	increases	increase	VERB
fcis-15825	50	36	the	the	DET
fcis-15825	50	37	difficulty	difficulty	NOUN
fcis-15825	50	38	of	of	ADP
fcis-15825	50	39	feature	feature	NOUN
fcis-15825	50	40	extraction	extraction	NOUN
fcis-15825	50	41	of	of	ADP
fcis-15825	50	42	the	the	DET
fcis-15825	50	43	detection	detection	NOUN
fcis-15825	50	44	model	model	NOUN
fcis-15825	50	45	.	.	PUNCT
fcis-15825	51	1	the	the	DET
fcis-15825	51	2	attention	attention	NOUN
fcis-15825	51	3	mechanism	mechanism	NOUN
fcis-15825	51	4	can	can	AUX
fcis-15825	51	5	help	help	VERB
fcis-15825	51	6	the	the	DET
fcis-15825	51	7	model	model	NOUN
fcis-15825	51	8	pay	pay	VERB
fcis-15825	51	9	more	more	ADJ
fcis-15825	51	10	attention	attention	NOUN
fcis-15825	51	11	to	to	ADP
fcis-15825	51	12	useful	useful	ADJ
fcis-15825	51	13	information	information	NOUN
fcis-15825	51	14	in	in	ADP
fcis-15825	51	15	the	the	DET
fcis-15825	51	16	training	training	NOUN
fcis-15825	51	17	process	process	NOUN
fcis-15825	51	18	and	and	CCONJ
fcis-15825	51	19	ignore	ignore	VERB
fcis-15825	51	20	the	the	DET
fcis-15825	51	21	influence	influence	NOUN
fcis-15825	51	22	of	of	ADP
fcis-15825	51	23	redundant	redundant	ADJ
fcis-15825	51	24	background	background	NOUN
fcis-15825	51	25	information	information	NOUN
fcis-15825	51	26	to	to	ADP
fcis-15825	51	27	some	some	DET
fcis-15825	51	28	extent	extent	NOUN
fcis-15825	51	29	.	.	PUNCT
fcis-15825	52	1	cbam	cbam	NOUN
fcis-15825	52	2	(	(	PUNCT
fcis-15825	52	3	convolutional	convolutional	ADJ
fcis-15825	52	4	block	block	NOUN
fcis-15825	52	5	attention	attention	NOUN
fcis-15825	52	6	module)[14	module)[14	PROPN
fcis-15825	52	7	]	]	PUNCT
fcis-15825	52	8	is	be	AUX
fcis-15825	52	9	a	a	DET
fcis-15825	52	10	widely	widely	ADV
fcis-15825	52	11	used	use	VERB
fcis-15825	52	12	convolution	convolution	NOUN
fcis-15825	52	13	attention	attention	NOUN
fcis-15825	52	14	module	module	NOUN
fcis-15825	52	15	,	,	PUNCT
fcis-15825	52	16	which	which	PRON
fcis-15825	52	17	is	be	AUX
fcis-15825	52	18	plug	plug	NOUN
fcis-15825	52	19	and	and	CCONJ
fcis-15825	52	20	play	play	VERB
fcis-15825	52	21	,	,	PUNCT
fcis-15825	52	22	simple	simple	ADJ
fcis-15825	52	23	and	and	CCONJ
fcis-15825	52	24	efficient	efficient	ADJ
fcis-15825	52	25	,	,	PUNCT
fcis-15825	52	26	different	different	ADJ
fcis-15825	52	27	from	from	ADP
fcis-15825	52	28	channel	channel	NOUN
fcis-15825	52	29	attention	attention	NOUN
fcis-15825	52	30	mechanism	mechanism	NOUN
fcis-15825	52	31	or	or	CCONJ
fcis-15825	52	32	spatial	spatial	ADJ
fcis-15825	52	33	attention	attention	NOUN
fcis-15825	52	34	mechanism	mechanism	NOUN
fcis-15825	52	35	.	.	PUNCT
fcis-15825	53	1	cbam	cbam	NOUN
fcis-15825	53	2	module	module	NOUN
fcis-15825	53	3	realizes	realize	VERB
fcis-15825	53	4	the	the	DET
fcis-15825	53	5	fusion	fusion	NOUN
fcis-15825	53	6	of	of	ADP
fcis-15825	53	7	cross	cross	ADJ
fcis-15825	53	8	-	-	ADJ
fcis-15825	53	9	channel	channel	ADJ
fcis-15825	53	10	features	feature	NOUN
fcis-15825	53	11	and	and	CCONJ
fcis-15825	53	12	spatial	spatial	ADJ
fcis-15825	53	13	information	information	NOUN
fcis-15825	53	14	by	by	ADP
fcis-15825	53	15	convolution	convolution	NOUN
fcis-15825	53	16	operation	operation	NOUN
fcis-15825	53	17	.	.	PUNCT
fcis-15825	54	1	as	as	SCONJ
fcis-15825	54	2	shown	show	VERB
fcis-15825	54	3	in	in	ADP
fcis-15825	54	4	fig.2	fig.2	PROPN
fcis-15825	54	5	,	,	PUNCT
fcis-15825	54	6	the	the	DET
fcis-15825	54	7	input	input	NOUN
fcis-15825	54	8	feature	feature	NOUN
fcis-15825	54	9	map	map	NOUN
fcis-15825	54	10	f(h*w*c	f(h*w*c	PROPN
fcis-15825	54	11	)	)	PUNCT
fcis-15825	54	12	is	be	AUX
fcis-15825	54	13	subjected	subject	VERB
fcis-15825	54	14	to	to	ADP
fcis-15825	54	15	global	global	ADJ
fcis-15825	54	16	maximum	maximum	ADJ
fcis-15825	54	17	pooling	pooling	NOUN
fcis-15825	54	18	and	and	CCONJ
fcis-15825	54	19	global	global	ADJ
fcis-15825	54	20	average	average	ADJ
fcis-15825	54	21	pooling	pool	VERB
fcis-15825	54	22	operations	operation	NOUN
fcis-15825	54	23	,	,	PUNCT
fcis-15825	54	24	respectively	respectively	ADV
fcis-15825	54	25	,	,	PUNCT
fcis-15825	54	26	to	to	PART
fcis-15825	54	27	obtain	obtain	VERB
fcis-15825	54	28	one	one	NUM
fcis-15825	54	29	-	-	PUNCT
fcis-15825	54	30	dimensional	dimensional	ADJ
fcis-15825	54	31	channel	channel	NOUN
fcis-15825	54	32	features	feature	NOUN
fcis-15825	54	33	(	(	PUNCT
fcis-15825	54	34	1	1	NUM
fcis-15825	54	35	*	*	NUM
fcis-15825	54	36	1*c	1*c	NUM
fcis-15825	54	37	)	)	PUNCT
fcis-15825	54	38	,	,	PUNCT
fcis-15825	54	39	and	and	CCONJ
fcis-15825	54	40	the	the	DET
fcis-15825	54	41	feature	feature	NOUN
fcis-15825	54	42	maps	map	NOUN
fcis-15825	54	43	of	of	ADP
fcis-15825	54	44	these	these	DET
fcis-15825	54	45	two	two	NUM
fcis-15825	54	46	branches	branch	NOUN
fcis-15825	54	47	are	be	AUX
fcis-15825	54	48	respectively	respectively	ADV
fcis-15825	54	49	connected	connect	VERB
fcis-15825	54	50	to	to	ADP
fcis-15825	54	51	the	the	DET
fcis-15825	54	52	two	two	NUM
fcis-15825	54	53	-	-	PUNCT
fcis-15825	54	54	layer	layer	NOUN
fcis-15825	54	55	shared	share	VERB
fcis-15825	54	56	neural	neural	ADJ
fcis-15825	54	57	network	network	NOUN
fcis-15825	54	58	(	(	PUNCT
fcis-15825	54	59	mlp	mlp	PROPN
fcis-15825	54	60	)	)	PUNCT
fcis-15825	54	61	for	for	ADP
fcis-15825	54	62	training	training	NOUN
fcis-15825	54	63	,	,	PUNCT
fcis-15825	54	64	and	and	CCONJ
fcis-15825	54	65	then	then	ADV
fcis-15825	54	66	the	the	DET
fcis-15825	54	67	obtained	obtain	VERB
fcis-15825	54	68	output	output	NOUN
fcis-15825	54	69	results	result	NOUN
fcis-15825	54	70	are	be	AUX
fcis-15825	54	71	added	add	VERB
fcis-15825	54	72	by	by	ADP
fcis-15825	54	73	elements	element	NOUN
fcis-15825	54	74	.	.	PUNCT
fcis-15825	55	1	finally	finally	ADV
fcis-15825	55	2	,	,	PUNCT
fcis-15825	55	3	the	the	DET
fcis-15825	55	4	channel	channel	NOUN
fcis-15825	55	5	attention	attention	NOUN
fcis-15825	55	6	diagram	diagram	NOUN
fcis-15825	55	7	mc	mc	PROPN
fcis-15825	55	8	can	can	AUX
fcis-15825	55	9	be	be	AUX
fcis-15825	55	10	obtained	obtain	VERB
fcis-15825	55	11	through	through	ADP
fcis-15825	55	12	sigmoid	sigmoid	NOUN
fcis-15825	55	13	activation	activation	NOUN
fcis-15825	55	14	function	function	NOUN
fcis-15825	55	15	:	:	PUNCT
fcis-15825	55	16	mlp	mlp	PROPN
fcis-15825	55	17	avgpool	avgpool	PROPN
fcis-15825	55	18	mlp	mlp	PROPN
fcis-15825	55	19	maxpool	maxpool	PROPN
fcis-15825	55	20	(	(	PUNCT
fcis-15825	55	21	1	1	X
fcis-15825	55	22	)	)	PUNCT
fcis-15825	55	23	the	the	DET
fcis-15825	55	24	spatial	spatial	ADJ
fcis-15825	55	25	attention	attention	NOUN
fcis-15825	55	26	module	module	NOUN
fcis-15825	55	27	(	(	PUNCT
fcis-15825	55	28	sam	sam	PROPN
fcis-15825	55	29	)	)	PUNCT
fcis-15825	55	30	compresses	compress	VERB
fcis-15825	55	31	the	the	DET
fcis-15825	55	32	input	input	NOUN
fcis-15825	55	33	features	feature	VERB
fcis-15825	55	34	into	into	ADP
fcis-15825	55	35	one	one	NUM
fcis-15825	55	36	dimension	dimension	NOUN
fcis-15825	55	37	by	by	ADP
fcis-15825	55	38	maximum	maximum	ADJ
fcis-15825	55	39	pooling	pooling	NOUN
fcis-15825	55	40	and	and	CCONJ
fcis-15825	55	41	average	average	ADJ
fcis-15825	55	42	pooling	pooling	NOUN
fcis-15825	55	43	,	,	PUNCT
fcis-15825	55	44	and	and	CCONJ
fcis-15825	55	45	obtains	obtain	VERB
fcis-15825	55	46	two	two	NUM
fcis-15825	55	47	feature	feature	NOUN
fcis-15825	55	48	maps	map	NOUN
fcis-15825	55	49	(	(	PUNCT
fcis-15825	55	50	h*w*1	h*w*1	X
fcis-15825	55	51	)	)	PUNCT
fcis-15825	55	52	,	,	PUNCT
fcis-15825	55	53	then	then	ADV
fcis-15825	55	54	obtains	obtain	VERB
fcis-15825	55	55	the	the	DET
fcis-15825	55	56	single	single	ADJ
fcis-15825	55	57	-	-	PUNCT
fcis-15825	55	58	channel	channel	NOUN
fcis-15825	55	59	features	feature	NOUN
fcis-15825	55	60	by	by	ADP
fcis-15825	55	61	concat	concat	NOUN
fcis-15825	55	62	and	and	CCONJ
fcis-15825	55	63	convolution	convolution	NOUN
fcis-15825	55	64	operations	operation	NOUN
fcis-15825	55	65	,	,	PUNCT
fcis-15825	55	66	and	and	CCONJ
fcis-15825	55	67	finally	finally	ADV
fcis-15825	55	68	obtains	obtain	VERB
fcis-15825	55	69	the	the	DET
fcis-15825	55	70	spatial	spatial	ADJ
fcis-15825	55	71	attention	attention	NOUN
fcis-15825	55	72	map	map	NOUN
fcis-15825	55	73	ms	ms	NOUN
fcis-15825	55	74	by	by	ADP
fcis-15825	55	75	sigmoid	sigmoid	NOUN
fcis-15825	55	76	activation	activation	NOUN
fcis-15825	55	77	function	function	NOUN
fcis-15825	55	78	:	:	PUNCT
fcis-15825	55	79	ms	ms	PROPN
fcis-15825	55	80	avgpool	avgpool	PROPN
fcis-15825	55	81	;	;	PUNCT
fcis-15825	55	82	maxpool	maxpool	NOUN
fcis-15825	55	83	(	(	PUNCT
fcis-15825	55	84	2	2	X
fcis-15825	55	85	)	)	PUNCT
fcis-15825	55	86	multiply	multiply	NOUN
fcis-15825	55	87	and	and	CCONJ
fcis-15825	55	88	weight	weight	VERB
fcis-15825	55	89	the	the	DET
fcis-15825	55	90	channel	channel	NOUN
fcis-15825	55	91	attention	attention	NOUN
fcis-15825	55	92	diagram	diagram	NOUN
fcis-15825	55	93	mc	mc	PROPN
fcis-15825	55	94	and	and	CCONJ
fcis-15825	55	95	the	the	DET
fcis-15825	55	96	spatial	spatial	ADJ
fcis-15825	55	97	attention	attention	NOUN
fcis-15825	55	98	diagram	diagram	NOUN
fcis-15825	55	99	ms	ms	NOUN
fcis-15825	55	100	to	to	PART
fcis-15825	55	101	get	get	VERB
fcis-15825	55	102	the	the	DET
fcis-15825	55	103	final	final	ADJ
fcis-15825	55	104	feature	feature	NOUN
fcis-15825	55	105	map	map	NOUN
fcis-15825	55	106	.	.	PUNCT
fcis-15825	56	1	by	by	ADP
fcis-15825	56	2	adding	add	VERB
fcis-15825	56	3	cbam	cbam	NOUN
fcis-15825	56	4	module	module	NOUN
fcis-15825	56	5	,	,	PUNCT
fcis-15825	56	6	the	the	DET
fcis-15825	56	7	flow	flow	NOUN
fcis-15825	56	8	of	of	ADP
fcis-15825	56	9	feature	feature	NOUN
fcis-15825	56	10	information	information	NOUN
fcis-15825	56	11	is	be	AUX
fcis-15825	56	12	promoted	promote	VERB
fcis-15825	56	13	,	,	PUNCT
fcis-15825	56	14	the	the	DET
fcis-15825	56	15	ability	ability	NOUN
fcis-15825	56	16	of	of	ADP
fcis-15825	56	17	model	model	NOUN
fcis-15825	56	18	feature	feature	NOUN
fcis-15825	56	19	extraction	extraction	NOUN
fcis-15825	56	20	is	be	AUX
fcis-15825	56	21	enhanced	enhance	VERB
fcis-15825	56	22	,	,	PUNCT
fcis-15825	56	23	and	and	CCONJ
fcis-15825	56	24	the	the	DET
fcis-15825	56	25	detection	detection	NOUN
fcis-15825	56	26	performance	performance	NOUN
fcis-15825	56	27	and	and	CCONJ
fcis-15825	56	28	accuracy	accuracy	NOUN
fcis-15825	56	29	of	of	ADP
fcis-15825	56	30	the	the	DET
fcis-15825	56	31	network	network	NOUN
fcis-15825	56	32	are	be	AUX
fcis-15825	56	33	improved	improve	VERB
fcis-15825	56	34	.	.	PUNCT
fcis-15825	57	1	fig	fig	NOUN
fcis-15825	57	2	2	2	NUM
fcis-15825	57	3	.	.	PUNCT
fcis-15825	57	4	cbam	cbam	NOUN
fcis-15825	57	5	attention	attention	NOUN
fcis-15825	57	6	mechanism	mechanism	NOUN
fcis-15825	57	7	module	module	NOUN
fcis-15825	57	8	2.4	2.4	NUM
fcis-15825	57	9	.	.	PUNCT
fcis-15825	58	1	bifpn	bifpn	PROPN
fcis-15825	58	2	cross	cross	ADJ
fcis-15825	58	3	-	-	ADJ
fcis-15825	58	4	scale	scale	ADJ
fcis-15825	58	5	feature	feature	NOUN
fcis-15825	58	6	fusion	fusion	NOUN
fcis-15825	58	7	yolov5s	yolov5s	NOUN
fcis-15825	58	8	model	model	NOUN
fcis-15825	58	9	uses	uses	AUX
fcis-15825	58	10	fpn	fpn	VERB
fcis-15825	58	11	-	-	PUNCT
fcis-15825	58	12	pan	pan	NOUN
fcis-15825	58	13	structure	structure	NOUN
fcis-15825	58	14	to	to	ADP
fcis-15825	58	15	fuse	fuse	NOUN
fcis-15825	58	16	features	feature	NOUN
fcis-15825	58	17	.	.	PUNCT
fcis-15825	59	1	fpn	fpn	VERB
fcis-15825	59	2	uses	uses	AUX
fcis-15825	59	3	up	up	ADV
fcis-15825	59	4	-	-	PUNCT
fcis-15825	59	5	sampling	sample	VERB
fcis-15825	59	6	to	to	PART
fcis-15825	59	7	construct	construct	VERB
fcis-15825	59	8	a	a	DET
fcis-15825	59	9	one	one	NUM
fcis-15825	59	10	-	-	PUNCT
fcis-15825	59	11	way	way	NOUN
fcis-15825	59	12	path	path	NOUN
fcis-15825	59	13	,	,	PUNCT
fcis-15825	59	14	which	which	PRON
fcis-15825	59	15	fuses	fuse	VERB
fcis-15825	59	16	high	high	ADJ
fcis-15825	59	17	-	-	PUNCT
fcis-15825	59	18	level	level	NOUN
fcis-15825	59	19	semantics	semantic	NOUN
fcis-15825	59	20	and	and	CCONJ
fcis-15825	59	21	low	low	ADJ
fcis-15825	59	22	-	-	PUNCT
fcis-15825	59	23	level	level	NOUN
fcis-15825	59	24	features	feature	NOUN
fcis-15825	59	25	.	.	PUNCT
fcis-15825	60	1	however	however	ADV
fcis-15825	60	2	,	,	PUNCT
fcis-15825	60	3	shallow	shallow	ADJ
fcis-15825	60	4	features	feature	NOUN
fcis-15825	60	5	are	be	AUX
fcis-15825	60	6	seriously	seriously	ADV
fcis-15825	60	7	lost	lose	VERB
fcis-15825	60	8	after	after	ADP
fcis-15825	60	9	many	many	ADJ
fcis-15825	60	10	up	up	ADV
fcis-15825	60	11	-	-	PUNCT
fcis-15825	60	12	sampling	sample	VERB
fcis-15825	60	13	operations	operation	NOUN
fcis-15825	60	14	,	,	PUNCT
fcis-15825	60	15	which	which	PRON
fcis-15825	60	16	directly	directly	ADV
fcis-15825	60	17	affects	affect	VERB
fcis-15825	60	18	the	the	DET
fcis-15825	60	19	prediction	prediction	NOUN
fcis-15825	60	20	effect	effect	NOUN
fcis-15825	60	21	of	of	ADP
fcis-15825	60	22	the	the	DET
fcis-15825	60	23	detection	detection	NOUN
fcis-15825	60	24	head	head	NOUN
fcis-15825	60	25	.	.	PUNCT
fcis-15825	61	1	as	as	SCONJ
fcis-15825	61	2	shown	show	VERB
fcis-15825	61	3	in	in	ADP
fcis-15825	61	4	fig.3	fig.3	PROPN
fcis-15825	61	5	,	,	PUNCT
fcis-15825	61	6	pan	pan	NOUN
fcis-15825	61	7	module	module	NOUN
fcis-15825	61	8	adds	add	VERB
fcis-15825	61	9	a	a	DET
fcis-15825	61	10	bottom	bottom	ADJ
fcis-15825	61	11	-	-	PUNCT
fcis-15825	61	12	up	up	ADP
fcis-15825	61	13	path	path	NOUN
fcis-15825	61	14	to	to	PART
fcis-15825	61	15	fpn	fpn	VERB
fcis-15825	61	16	,	,	PUNCT
fcis-15825	61	17	constructs	construct	VERB
fcis-15825	61	18	a	a	DET
fcis-15825	61	19	bidirectional	bidirectional	ADJ
fcis-15825	61	20	fusion	fusion	NOUN
fcis-15825	61	21	path	path	NOUN
fcis-15825	61	22	,	,	PUNCT
fcis-15825	61	23	and	and	CCONJ
fcis-15825	61	24	realizes	realize	VERB
fcis-15825	61	25	the	the	DET
fcis-15825	61	26	fusion	fusion	NOUN
fcis-15825	61	27	of	of	ADP
fcis-15825	61	28	different	different	ADJ
fcis-15825	61	29	levels	level	NOUN
fcis-15825	61	30	of	of	ADP
fcis-15825	61	31	features	feature	NOUN
fcis-15825	61	32	,	,	PUNCT
fcis-15825	61	33	but	but	CCONJ
fcis-15825	61	34	frequent	frequent	ADJ
fcis-15825	61	35	up	up	ADV
fcis-15825	61	36	-	-	PUNCT
fcis-15825	61	37	sampling	sample	VERB
fcis-15825	61	38	and	and	CCONJ
fcis-15825	61	39	down	down	ADV
fcis-15825	61	40	-	-	PUNCT
fcis-15825	61	41	sampling	sample	VERB
fcis-15825	61	42	operations	operation	NOUN
fcis-15825	61	43	also	also	ADV
fcis-15825	61	44	increase	increase	VERB
fcis-15825	61	45	the	the	DET
fcis-15825	61	46	parameters	parameter	NOUN
fcis-15825	61	47	and	and	CCONJ
fcis-15825	61	48	calculation	calculation	NOUN
fcis-15825	61	49	of	of	ADP
fcis-15825	61	50	the	the	DET
fcis-15825	61	51	model	model	NOUN
fcis-15825	61	52	.	.	PUNCT
fcis-15825	62	1	therefore	therefore	ADV
fcis-15825	62	2	,	,	PUNCT
fcis-15825	62	3	this	this	DET
fcis-15825	62	4	paper	paper	NOUN
fcis-15825	62	5	uses	use	VERB
fcis-15825	62	6	the	the	DET
fcis-15825	62	7	bifpn	bifpn	PROPN
fcis-15825	62	8	structure	structure	NOUN
fcis-15825	62	9	as	as	ADP
fcis-15825	62	10	the	the	DET
fcis-15825	62	11	neck	neck	NOUN
fcis-15825	62	12	network	network	NOUN
fcis-15825	62	13	,	,	PUNCT
fcis-15825	62	14	and	and	CCONJ
fcis-15825	62	15	bifpn[15	bifpn[15	NOUN
fcis-15825	62	16	]	]	PUNCT
fcis-15825	62	17	keeps	keep	VERB
fcis-15825	62	18	the	the	DET
fcis-15825	62	19	bidirectional	bidirectional	ADJ
fcis-15825	62	20	feature	feature	NOUN
fcis-15825	62	21	fusion	fusion	NOUN
fcis-15825	62	22	path	path	NOUN
fcis-15825	62	23	in	in	ADP
fcis-15825	62	24	panet	panet	NOUN
fcis-15825	62	25	,	,	PUNCT
fcis-15825	62	26	and	and	CCONJ
fcis-15825	62	27	makes	make	VERB
fcis-15825	62	28	the	the	DET
fcis-15825	62	29	following	follow	VERB
fcis-15825	62	30	improvements	improvement	NOUN
fcis-15825	62	31	:	:	PUNCT
fcis-15825	62	32	deleting	delete	VERB
fcis-15825	62	33	a	a	DET
fcis-15825	62	34	single	single	ADJ
fcis-15825	62	35	input	input	NOUN
fcis-15825	62	36	node	node	NOUN
fcis-15825	62	37	,	,	PUNCT
fcis-15825	62	38	and	and	CCONJ
fcis-15825	62	39	the	the	DET
fcis-15825	62	40	node	node	NOUN
fcis-15825	62	41	with	with	ADP
fcis-15825	62	42	only	only	ADV
fcis-15825	62	43	one	one	NUM
fcis-15825	62	44	input	input	NOUN
fcis-15825	62	45	115	115	NUM
fcis-15825	62	46	edge	edge	NOUN
fcis-15825	62	47	contains	contain	VERB
fcis-15825	62	48	the	the	DET
fcis-15825	62	49	same	same	ADJ
fcis-15825	62	50	information	information	NOUN
fcis-15825	62	51	as	as	ADP
fcis-15825	62	52	the	the	DET
fcis-15825	62	53	previous	previous	ADJ
fcis-15825	62	54	node	node	NOUN
fcis-15825	62	55	,	,	PUNCT
fcis-15825	62	56	which	which	PRON
fcis-15825	62	57	is	be	AUX
fcis-15825	62	58	a	a	DET
fcis-15825	62	59	structural	structural	ADJ
fcis-15825	62	60	redundancy	redundancy	NOUN
fcis-15825	62	61	,	,	PUNCT
fcis-15825	62	62	and	and	CCONJ
fcis-15825	62	63	the	the	DET
fcis-15825	62	64	calculation	calculation	NOUN
fcis-15825	62	65	amount	amount	NOUN
fcis-15825	62	66	is	be	AUX
fcis-15825	62	67	simplified	simplify	VERB
fcis-15825	62	68	through	through	ADP
fcis-15825	62	69	this	this	DET
fcis-15825	62	70	improvement	improvement	NOUN
fcis-15825	62	71	;	;	PUNCT
fcis-15825	62	72	jumping	jump	VERB
fcis-15825	62	73	connection	connection	NOUN
fcis-15825	62	74	is	be	AUX
fcis-15825	62	75	added	add	VERB
fcis-15825	62	76	between	between	ADP
fcis-15825	62	77	the	the	DET
fcis-15825	62	78	input	input	NOUN
fcis-15825	62	79	and	and	CCONJ
fcis-15825	62	80	output	output	NOUN
fcis-15825	62	81	nodes	node	NOUN
fcis-15825	62	82	at	at	ADP
fcis-15825	62	83	the	the	DET
fcis-15825	62	84	same	same	ADJ
fcis-15825	62	85	level	level	NOUN
fcis-15825	62	86	,	,	PUNCT
fcis-15825	62	87	and	and	CCONJ
fcis-15825	62	88	the	the	DET
fcis-15825	62	89	input	input	NOUN
fcis-15825	62	90	node	node	NOUN
fcis-15825	62	91	is	be	AUX
fcis-15825	62	92	directly	directly	ADV
fcis-15825	62	93	connected	connect	VERB
fcis-15825	62	94	to	to	ADP
fcis-15825	62	95	the	the	DET
fcis-15825	62	96	output	output	NOUN
fcis-15825	62	97	node	node	NOUN
fcis-15825	62	98	without	without	ADP
fcis-15825	62	99	up	up	ADV
fcis-15825	62	100	-	-	PUNCT
fcis-15825	62	101	sampling	sample	VERB
fcis-15825	62	102	,	,	PUNCT
fcis-15825	62	103	thus	thus	ADV
fcis-15825	62	104	retaining	retain	VERB
fcis-15825	62	105	more	more	ADJ
fcis-15825	62	106	original	original	ADJ
fcis-15825	62	107	information	information	NOUN
fcis-15825	62	108	;	;	PUNCT
fcis-15825	62	109	weighted	weight	VERB
fcis-15825	62	110	fusion	fusion	NOUN
fcis-15825	62	111	of	of	ADP
fcis-15825	62	112	different	different	ADJ
fcis-15825	62	113	levels	level	NOUN
fcis-15825	62	114	of	of	ADP
fcis-15825	62	115	information	information	NOUN
fcis-15825	62	116	is	be	AUX
fcis-15825	62	117	carried	carry	VERB
fcis-15825	62	118	out	out	ADP
fcis-15825	62	119	,	,	PUNCT
fcis-15825	62	120	and	and	CCONJ
fcis-15825	62	121	the	the	DET
fcis-15825	62	122	two	two	NUM
fcis-15825	62	123	paths	path	NOUN
fcis-15825	62	124	are	be	AUX
fcis-15825	62	125	integrated	integrate	VERB
fcis-15825	62	126	into	into	ADP
fcis-15825	62	127	a	a	DET
fcis-15825	62	128	whole	whole	NOUN
fcis-15825	62	129	,	,	PUNCT
fcis-15825	62	130	which	which	PRON
fcis-15825	62	131	enhances	enhance	VERB
fcis-15825	62	132	information	information	NOUN
fcis-15825	62	133	fusion	fusion	NOUN
fcis-15825	62	134	.	.	PUNCT
fcis-15825	63	1	in	in	ADP
fcis-15825	63	2	summary	summary	NOUN
fcis-15825	63	3	,	,	PUNCT
fcis-15825	63	4	bifpn	bifpn	PROPN
fcis-15825	63	5	structure	structure	NOUN
fcis-15825	63	6	simplifies	simplify	VERB
fcis-15825	63	7	the	the	DET
fcis-15825	63	8	neck	neck	NOUN
fcis-15825	63	9	network	network	NOUN
fcis-15825	63	10	,	,	PUNCT
fcis-15825	63	11	improves	improve	VERB
fcis-15825	63	12	the	the	DET
fcis-15825	63	13	cross	cross	ADJ
fcis-15825	63	14	-	-	ADJ
fcis-15825	63	15	layer	layer	ADJ
fcis-15825	63	16	information	information	NOUN
fcis-15825	63	17	exchange	exchange	NOUN
fcis-15825	63	18	and	and	CCONJ
fcis-15825	63	19	fusion	fusion	NOUN
fcis-15825	63	20	between	between	ADP
fcis-15825	63	21	different	different	ADJ
fcis-15825	63	22	levels	level	NOUN
fcis-15825	63	23	of	of	ADP
fcis-15825	63	24	feature	feature	NOUN
fcis-15825	63	25	maps	map	NOUN
fcis-15825	63	26	,	,	PUNCT
fcis-15825	63	27	improves	improve	VERB
fcis-15825	63	28	the	the	DET
fcis-15825	63	29	information	information	NOUN
fcis-15825	63	30	loss	loss	NOUN
fcis-15825	63	31	caused	cause	VERB
fcis-15825	63	32	by	by	ADP
fcis-15825	63	33	up	up	ADV
fcis-15825	63	34	-	-	PUNCT
fcis-15825	63	35	sampling	sampling	NOUN
fcis-15825	63	36	,	,	PUNCT
fcis-15825	63	37	and	and	CCONJ
fcis-15825	63	38	improves	improve	VERB
fcis-15825	63	39	the	the	DET
fcis-15825	63	40	detection	detection	NOUN
fcis-15825	63	41	performance	performance	NOUN
fcis-15825	63	42	of	of	ADP
fcis-15825	63	43	the	the	DET
fcis-15825	63	44	model	model	NOUN
fcis-15825	63	45	.	.	PUNCT
fcis-15825	64	1	by	by	ADP
fcis-15825	64	2	improving	improve	VERB
fcis-15825	64	3	the	the	DET
fcis-15825	64	4	neck	neck	NOUN
fcis-15825	64	5	network	network	NOUN
fcis-15825	64	6	,	,	PUNCT
fcis-15825	64	7	the	the	DET
fcis-15825	64	8	cross	cross	ADJ
fcis-15825	64	9	-	-	ADJ
fcis-15825	64	10	layer	layer	ADJ
fcis-15825	64	11	integration	integration	NOUN
fcis-15825	64	12	of	of	ADP
fcis-15825	64	13	information	information	NOUN
fcis-15825	64	14	is	be	AUX
fcis-15825	64	15	strengthened	strengthen	VERB
fcis-15825	64	16	,	,	PUNCT
fcis-15825	64	17	and	and	CCONJ
fcis-15825	64	18	the	the	DET
fcis-15825	64	19	attention	attention	NOUN
fcis-15825	64	20	of	of	ADP
fcis-15825	64	21	the	the	DET
fcis-15825	64	22	network	network	NOUN
fcis-15825	64	23	to	to	ADP
fcis-15825	64	24	small	small	ADJ
fcis-15825	64	25	targets	target	NOUN
fcis-15825	64	26	is	be	AUX
fcis-15825	64	27	enhanced	enhance	VERB
fcis-15825	64	28	.	.	PUNCT
fcis-15825	65	1	fig	fig	NOUN
fcis-15825	65	2	3	3	NUM
fcis-15825	65	3	.	.	PUNCT
fcis-15825	65	4	feature	feature	NOUN
fcis-15825	65	5	fusion	fusion	NOUN
fcis-15825	65	6	module	module	NOUN
fcis-15825	65	7	fig	fig	NOUN
fcis-15825	65	8	4	4	NUM
fcis-15825	65	9	.	.	PUNCT
fcis-15825	65	10	improved	improve	VERB
fcis-15825	65	11	yolov5s	yolov5s	PROPN
fcis-15825	65	12	model	model	NOUN
fcis-15825	65	13	as	as	SCONJ
fcis-15825	65	14	shown	show	VERB
fcis-15825	65	15	in	in	ADP
fcis-15825	65	16	fig.4	fig.4	PROPN
fcis-15825	65	17	,	,	PUNCT
fcis-15825	65	18	the	the	DET
fcis-15825	65	19	network	network	NOUN
fcis-15825	65	20	structure	structure	NOUN
fcis-15825	65	21	improved	improve	VERB
fcis-15825	65	22	by	by	ADP
fcis-15825	65	23	the	the	DET
fcis-15825	65	24	above	above	ADJ
fcis-15825	65	25	measures	measure	NOUN
fcis-15825	65	26	.	.	PUNCT
fcis-15825	66	1	to	to	PART
fcis-15825	66	2	solve	solve	VERB
fcis-15825	66	3	the	the	DET
fcis-15825	66	4	problems	problem	NOUN
fcis-15825	66	5	of	of	ADP
fcis-15825	66	6	complex	complex	ADJ
fcis-15825	66	7	road	road	NOUN
fcis-15825	66	8	image	image	NOUN
fcis-15825	66	9	detection	detection	NOUN
fcis-15825	66	10	background	background	NOUN
fcis-15825	66	11	,	,	PUNCT
fcis-15825	66	12	great	great	ADJ
fcis-15825	66	13	difference	difference	NOUN
fcis-15825	66	14	in	in	ADP
fcis-15825	66	15	the	the	DET
fcis-15825	66	16	proportion	proportion	NOUN
fcis-15825	66	17	of	of	ADP
fcis-15825	66	18	road	road	NOUN
fcis-15825	66	19	diseases	disease	NOUN
fcis-15825	66	20	,	,	PUNCT
fcis-15825	66	21	and	and	CCONJ
fcis-15825	66	22	difficulty	difficulty	NOUN
fcis-15825	66	23	in	in	ADP
fcis-15825	66	24	detecting	detect	VERB
fcis-15825	66	25	small	small	ADJ
fcis-15825	66	26	target	target	NOUN
fcis-15825	66	27	pavement	pavement	NOUN
fcis-15825	66	28	diseases	disease	NOUN
fcis-15825	66	29	in	in	ADP
fcis-15825	66	30	actual	actual	ADJ
fcis-15825	66	31	scenes	scene	NOUN
fcis-15825	66	32	,	,	PUNCT
fcis-15825	66	33	yolov5s	yolov5s	PROPN
fcis-15825	66	34	-	-	PUNCT
fcis-15825	66	35	cbp	cbp	PROPN
fcis-15825	66	36	proposed	propose	VERB
fcis-15825	66	37	targeted	target	VERB
fcis-15825	66	38	improvement	improvement	NOUN
fcis-15825	66	39	measures	measure	NOUN
fcis-15825	66	40	:	:	PUNCT
fcis-15825	66	41	adding	add	VERB
fcis-15825	66	42	cbam	cbam	NOUN
fcis-15825	66	43	attention	attention	NOUN
fcis-15825	66	44	mechanism	mechanism	NOUN
fcis-15825	66	45	to	to	ADP
fcis-15825	66	46	the	the	DET
fcis-15825	66	47	backbone	backbone	NOUN
fcis-15825	66	48	network	network	NOUN
fcis-15825	66	49	,	,	PUNCT
fcis-15825	66	50	which	which	PRON
fcis-15825	66	51	enhanced	enhance	VERB
fcis-15825	66	52	the	the	DET
fcis-15825	66	53	model	model	NOUN
fcis-15825	66	54	's	's	PART
fcis-15825	66	55	attention	attention	NOUN
fcis-15825	66	56	to	to	ADP
fcis-15825	66	57	the	the	DET
fcis-15825	66	58	data	datum	NOUN
fcis-15825	66	59	image	image	NOUN
fcis-15825	66	60	channel	channel	NOUN
fcis-15825	66	61	and	and	CCONJ
fcis-15825	66	62	spatial	spatial	ADJ
fcis-15825	66	63	dimension	dimension	NOUN
fcis-15825	66	64	characteristics	characteristic	NOUN
fcis-15825	66	65	,	,	PUNCT
fcis-15825	66	66	and	and	CCONJ
fcis-15825	66	67	then	then	ADV
fcis-15825	66	68	ignored	ignore	VERB
fcis-15825	66	69	the	the	DET
fcis-15825	66	70	background	background	NOUN
fcis-15825	66	71	semantic	semantic	ADJ
fcis-15825	66	72	information	information	NOUN
fcis-15825	66	73	;	;	PUNCT
fcis-15825	66	74	the	the	DET
fcis-15825	66	75	cross	cross	ADJ
fcis-15825	66	76	-	-	ADJ
fcis-15825	66	77	layer	layer	ADJ
fcis-15825	66	78	fusion	fusion	NOUN
fcis-15825	66	79	of	of	ADP
fcis-15825	66	80	feature	feature	NOUN
fcis-15825	66	81	information	information	NOUN
fcis-15825	66	82	in	in	ADP
fcis-15825	66	83	the	the	DET
fcis-15825	66	84	neck	neck	NOUN
fcis-15825	66	85	network	network	NOUN
fcis-15825	66	86	adopts	adopt	VERB
fcis-15825	66	87	bifpn	bifpn	PROPN
fcis-15825	66	88	structure	structure	NOUN
fcis-15825	66	89	,	,	PUNCT
fcis-15825	66	90	which	which	PRON
fcis-15825	66	91	solves	solve	VERB
fcis-15825	66	92	the	the	DET
fcis-15825	66	93	problem	problem	NOUN
fcis-15825	66	94	of	of	ADP
fcis-15825	66	95	high	high	ADJ
fcis-15825	66	96	-	-	PUNCT
fcis-15825	66	97	level	level	NOUN
fcis-15825	66	98	semantic	semantic	ADJ
fcis-15825	66	99	feature	feature	NOUN
fcis-15825	66	100	loss	loss	NOUN
fcis-15825	66	101	caused	cause	VERB
fcis-15825	66	102	by	by	ADP
fcis-15825	66	103	up	up	ADV
fcis-15825	66	104	-	-	PUNCT
fcis-15825	66	105	sampling	sampling	NOUN
fcis-15825	66	106	.	.	PUNCT
fcis-15825	67	1	a	a	DET
fcis-15825	67	2	small	small	ADJ
fcis-15825	67	3	target	target	NOUN
fcis-15825	67	4	detection	detection	NOUN
fcis-15825	67	5	layer	layer	NOUN
fcis-15825	67	6	with	with	ADP
fcis-15825	67	7	smaller	small	ADJ
fcis-15825	67	8	receptive	receptive	ADJ
fcis-15825	67	9	field	field	NOUN
fcis-15825	67	10	was	be	AUX
fcis-15825	67	11	added	add	VERB
fcis-15825	67	12	to	to	ADP
fcis-15825	67	13	the	the	DET
fcis-15825	67	14	prediction	prediction	NOUN
fcis-15825	67	15	head	head	NOUN
fcis-15825	67	16	,	,	PUNCT
fcis-15825	67	17	which	which	PRON
fcis-15825	67	18	is	be	AUX
fcis-15825	67	19	responsible	responsible	ADJ
fcis-15825	67	20	for	for	ADP
fcis-15825	67	21	detecting	detect	VERB
fcis-15825	67	22	tiny	tiny	ADJ
fcis-15825	67	23	cracks	crack	NOUN
fcis-15825	67	24	and	and	CCONJ
fcis-15825	67	25	pits	pit	NOUN
fcis-15825	67	26	in	in	ADP
fcis-15825	67	27	road	road	NOUN
fcis-15825	67	28	images	image	NOUN
fcis-15825	67	29	,	,	PUNCT
fcis-15825	67	30	thus	thus	ADV
fcis-15825	67	31	improving	improve	VERB
fcis-15825	67	32	the	the	DET
fcis-15825	67	33	detection	detection	NOUN
fcis-15825	67	34	ability	ability	NOUN
fcis-15825	67	35	of	of	ADP
fcis-15825	67	36	the	the	DET
fcis-15825	67	37	model	model	NOUN
fcis-15825	67	38	for	for	ADP
fcis-15825	67	39	small	small	ADJ
fcis-15825	67	40	target	target	NOUN
fcis-15825	67	41	road	road	NOUN
fcis-15825	67	42	diseases	disease	NOUN
fcis-15825	67	43	.	.	PUNCT
fcis-15825	68	1	besides	besides	SCONJ
fcis-15825	68	2	,	,	PUNCT
fcis-15825	68	3	the	the	DET
fcis-15825	68	4	k	k	NOUN
fcis-15825	68	5	-	-	ADJ
fcis-15825	68	6	means++	means++	ADV
fcis-15825	68	7	clustering	clustering	NOUN
fcis-15825	68	8	method	method	NOUN
fcis-15825	68	9	was	be	AUX
fcis-15825	68	10	used	use	VERB
fcis-15825	68	11	to	to	PART
fcis-15825	68	12	redesign	redesign	VERB
fcis-15825	68	13	the	the	DET
fcis-15825	68	14	anchor	anchor	NOUN
fcis-15825	68	15	frames	frame	NOUN
fcis-15825	68	16	,	,	PUNCT
fcis-15825	68	17	and	and	CCONJ
fcis-15825	68	18	12	12	NUM
fcis-15825	68	19	anchor	anchor	NOUN
fcis-15825	68	20	frames	frame	NOUN
fcis-15825	68	21	which	which	PRON
fcis-15825	68	22	are	be	AUX
fcis-15825	68	23	more	more	ADV
fcis-15825	68	24	suitable	suitable	ADJ
fcis-15825	68	25	for	for	SCONJ
fcis-15825	68	26	this	this	DET
fcis-15825	68	27	data	datum	NOUN
fcis-15825	68	28	set	set	VERB
fcis-15825	68	29	are	be	AUX
fcis-15825	68	30	obtained	obtain	VERB
fcis-15825	68	31	.	.	PUNCT
fcis-15825	69	1	meanwhile	meanwhile	ADV
fcis-15825	69	2	,	,	PUNCT
fcis-15825	69	3	in	in	ADP
fcis-15825	69	4	order	order	NOUN
fcis-15825	69	5	to	to	PART
fcis-15825	69	6	offset	offset	VERB
fcis-15825	69	7	the	the	DET
fcis-15825	69	8	overhead	overhead	NOUN
fcis-15825	69	9	caused	cause	VERB
fcis-15825	69	10	by	by	ADP
fcis-15825	69	11	increasing	increase	VERB
fcis-15825	69	12	the	the	DET
fcis-15825	69	13	small	small	ADJ
fcis-15825	69	14	target	target	NOUN
fcis-15825	69	15	detection	detection	NOUN
fcis-15825	69	16	layer	layer	NOUN
fcis-15825	69	17	,	,	PUNCT
fcis-15825	69	18	the	the	DET
fcis-15825	69	19	deep	deep	ADJ
fcis-15825	69	20	separable	separable	ADJ
fcis-15825	69	21	convolution	convolution	NOUN
fcis-15825	69	22	product	product	NOUN
fcis-15825	69	23	is	be	AUX
fcis-15825	69	24	used	use	VERB
fcis-15825	69	25	in	in	ADP
fcis-15825	69	26	the	the	DET
fcis-15825	69	27	backbone	backbone	NOUN
fcis-15825	69	28	network	network	NOUN
fcis-15825	69	29	instead	instead	ADV
fcis-15825	69	30	of	of	ADP
fcis-15825	69	31	the	the	DET
fcis-15825	69	32	ordinary	ordinary	ADJ
fcis-15825	69	33	convolution	convolution	NOUN
fcis-15825	69	34	,	,	PUNCT
fcis-15825	69	35	which	which	PRON
fcis-15825	69	36	reduces	reduce	VERB
fcis-15825	69	37	the	the	DET
fcis-15825	69	38	calculation	calculation	NOUN
fcis-15825	69	39	amount	amount	NOUN
fcis-15825	69	40	and	and	CCONJ
fcis-15825	69	41	parameters	parameter	NOUN
fcis-15825	69	42	of	of	ADP
fcis-15825	69	43	the	the	DET
fcis-15825	69	44	model	model	NOUN
fcis-15825	69	45	.	.	PUNCT
fcis-15825	70	1	3	3	X
fcis-15825	70	2	.	.	X
fcis-15825	70	3	experimental	experimental	ADJ
fcis-15825	70	4	results	result	NOUN
fcis-15825	70	5	and	and	CCONJ
fcis-15825	70	6	analysis	analysis	NOUN
fcis-15825	70	7	3.1	3.1	NUM
fcis-15825	70	8	.	.	PUNCT
fcis-15825	71	1	introduction	introduction	NOUN
fcis-15825	71	2	of	of	ADP
fcis-15825	71	3	datasets	dataset	NOUN
fcis-15825	71	4	in	in	ADP
fcis-15825	71	5	this	this	DET
fcis-15825	71	6	study	study	NOUN
fcis-15825	71	7	,	,	PUNCT
fcis-15825	71	8	the	the	DET
fcis-15825	71	9	datasets	dataset	NOUN
fcis-15825	71	10	consist	consist	VERB
fcis-15825	71	11	of	of	ADP
fcis-15825	71	12	three	three	NUM
fcis-15825	71	13	parts	part	NOUN
fcis-15825	71	14	,	,	PUNCT
fcis-15825	71	15	including	include	VERB
fcis-15825	71	16	two	two	NUM
fcis-15825	71	17	public	public	ADJ
fcis-15825	71	18	datasets	dataset	NOUN
fcis-15825	71	19	and	and	CCONJ
fcis-15825	71	20	local	local	ADJ
fcis-15825	71	21	datasets	dataset	NOUN
fcis-15825	71	22	.	.	PUNCT
fcis-15825	72	1	there	there	PRON
fcis-15825	72	2	are	be	VERB
fcis-15825	72	3	two	two	NUM
fcis-15825	72	4	public	public	ADJ
fcis-15825	72	5	datasets	dataset	NOUN
fcis-15825	72	6	,	,	PUNCT
fcis-15825	72	7	namely	namely	ADV
fcis-15825	72	8	,	,	PUNCT
fcis-15825	72	9	the	the	DET
fcis-15825	72	10	global	global	ADJ
fcis-15825	72	11	open	open	ADJ
fcis-15825	72	12	innovation	innovation	NOUN
fcis-15825	72	13	application	application	NOUN
fcis-15825	72	14	competition	competition	NOUN
fcis-15825	72	15	data	datum	NOUN
fcis-15825	72	16	set	set	VERB
fcis-15825	72	17	and	and	CCONJ
fcis-15825	72	18	the	the	DET
fcis-15825	72	19	rdd2022china	rdd2022china	PROPN
fcis-15825	72	20	data	datum	NOUN
fcis-15825	72	21	set	set	VERB
fcis-15825	72	22	.	.	PUNCT
fcis-15825	73	1	the	the	DET
fcis-15825	73	2	training	training	NOUN
fcis-15825	73	3	set	set	NOUN
fcis-15825	73	4	of	of	ADP
fcis-15825	73	5	innovative	innovative	ADJ
fcis-15825	73	6	application	application	NOUN
fcis-15825	73	7	competition	competition	NOUN
fcis-15825	73	8	data	datum	NOUN
fcis-15825	73	9	set	set	NOUN
fcis-15825	73	10	contains	contain	VERB
fcis-15825	73	11	6000	6000	NUM
fcis-15825	73	12	road	road	NOUN
fcis-15825	73	13	images	image	NOUN
fcis-15825	73	14	,	,	PUNCT
fcis-15825	73	15	and	and	CCONJ
fcis-15825	73	16	8	8	NUM
fcis-15825	73	17	kinds	kind	NOUN
fcis-15825	73	18	of	of	ADP
fcis-15825	73	19	road	road	NOUN
fcis-15825	73	20	diseases	disease	NOUN
fcis-15825	73	21	are	be	AUX
fcis-15825	73	22	marked	mark	VERB
fcis-15825	73	23	.	.	PUNCT
fcis-15825	74	1	rdd2022china	rdd2022china	PROPN
fcis-15825	74	2	data	data	PROPN
fcis-15825	74	3	set	set	PROPN
fcis-15825	74	4	contains	contain	VERB
fcis-15825	74	5	4,378	4,378	NUM
fcis-15825	74	6	training	training	NOUN
fcis-15825	74	7	pictures	picture	NOUN
fcis-15825	74	8	,	,	PUNCT
fcis-15825	74	9	and	and	CCONJ
fcis-15825	74	10	four	four	NUM
fcis-15825	74	11	kinds	kind	NOUN
fcis-15825	74	12	of	of	ADP
fcis-15825	74	13	road	road	NOUN
fcis-15825	74	14	diseases	disease	NOUN
fcis-15825	74	15	are	be	AUX
fcis-15825	74	16	marked	mark	VERB
fcis-15825	74	17	.	.	PUNCT
fcis-15825	75	1	the	the	DET
fcis-15825	75	2	local	local	ADJ
fcis-15825	75	3	data	datum	NOUN
fcis-15825	75	4	set	set	VERB
fcis-15825	75	5	includes	include	VERB
fcis-15825	75	6	2000	2000	NUM
fcis-15825	75	7	pictures	picture	NOUN
fcis-15825	75	8	.	.	PUNCT
fcis-15825	76	1	comparing	compare	VERB
fcis-15825	76	2	the	the	DET
fcis-15825	76	3	two	two	NUM
fcis-15825	76	4	public	public	ADJ
fcis-15825	76	5	data	datum	NOUN
fcis-15825	76	6	sets	set	NOUN
fcis-15825	76	7	,	,	PUNCT
fcis-15825	76	8	it	it	PRON
fcis-15825	76	9	is	be	AUX
fcis-15825	76	10	found	find	VERB
fcis-15825	76	11	that	that	SCONJ
fcis-15825	76	12	the	the	DET
fcis-15825	76	13	image	image	NOUN
fcis-15825	76	14	of	of	ADP
fcis-15825	76	15	the	the	DET
fcis-15825	76	16	innovation	innovation	NOUN
fcis-15825	76	17	application	application	NOUN
fcis-15825	76	18	contest	contest	NOUN
fcis-15825	76	19	data	datum	NOUN
fcis-15825	76	20	set	set	VERB
fcis-15825	76	21	contains	contain	VERB
fcis-15825	76	22	more	more	ADJ
fcis-15825	76	23	complex	complex	ADJ
fcis-15825	76	24	background	background	NOUN
fcis-15825	76	25	information	information	NOUN
fcis-15825	76	26	such	such	ADJ
fcis-15825	76	27	as	as	ADP
fcis-15825	76	28	pedestrians	pedestrian	NOUN
fcis-15825	76	29	and	and	CCONJ
fcis-15825	76	30	vehicles	vehicle	NOUN
fcis-15825	76	31	,	,	PUNCT
fcis-15825	76	32	and	and	CCONJ
fcis-15825	76	33	it	it	PRON
fcis-15825	76	34	is	be	AUX
fcis-15825	76	35	noisy	noisy	ADJ
fcis-15825	76	36	,	,	PUNCT
fcis-15825	76	37	with	with	ADP
fcis-15825	76	38	poor	poor	ADJ
fcis-15825	76	39	labeling	labeling	NOUN
fcis-15825	76	40	quality	quality	NOUN
fcis-15825	76	41	,	,	PUNCT
fcis-15825	76	42	and	and	CCONJ
fcis-15825	76	43	there	there	PRON
fcis-15825	76	44	are	be	VERB
fcis-15825	76	45	problems	problem	NOUN
fcis-15825	76	46	of	of	ADP
fcis-15825	76	47	wrong	wrong	ADJ
fcis-15825	76	48	labeling	labeling	NOUN
fcis-15825	76	49	and	and	CCONJ
fcis-15825	76	50	missing	miss	VERB
fcis-15825	76	51	labeling	labeling	NOUN
fcis-15825	76	52	.	.	PUNCT
fcis-15825	77	1	the	the	DET
fcis-15825	77	2	shooting	shooting	NOUN
fcis-15825	77	3	angles	angle	NOUN
fcis-15825	77	4	of	of	ADP
fcis-15825	77	5	rdd2022china	rdd2022china	NOUN
fcis-15825	77	6	data	datum	NOUN
fcis-15825	77	7	set	set	VERB
fcis-15825	77	8	are	be	AUX
fcis-15825	77	9	quite	quite	ADV
fcis-15825	77	10	different	different	ADJ
fcis-15825	77	11	,	,	PUNCT
fcis-15825	77	12	and	and	CCONJ
fcis-15825	77	13	there	there	PRON
fcis-15825	77	14	are	be	VERB
fcis-15825	77	15	few	few	ADJ
fcis-15825	77	16	kinds	kind	NOUN
fcis-15825	77	17	of	of	ADP
fcis-15825	77	18	pavement	pavement	NOUN
fcis-15825	77	19	diseases	disease	NOUN
fcis-15825	77	20	marked	mark	VERB
fcis-15825	77	21	.	.	PUNCT
fcis-15825	78	1	therefore	therefore	ADV
fcis-15825	78	2	,	,	PUNCT
fcis-15825	78	3	in	in	ADP
fcis-15825	78	4	order	order	NOUN
fcis-15825	78	5	to	to	PART
fcis-15825	78	6	further	far	ADV
fcis-15825	78	7	improve	improve	VERB
fcis-15825	78	8	the	the	DET
fcis-15825	78	9	quality	quality	NOUN
fcis-15825	78	10	of	of	ADP
fcis-15825	78	11	datasets	dataset	NOUN
fcis-15825	78	12	,	,	PUNCT
fcis-15825	78	13	the	the	DET
fcis-15825	78	14	public	public	ADJ
fcis-15825	78	15	datasets	dataset	NOUN
fcis-15825	78	16	are	be	AUX
fcis-15825	78	17	marked	mark	VERB
fcis-15825	78	18	twice	twice	ADV
fcis-15825	78	19	.	.	PUNCT
fcis-15825	79	1	firstly	firstly	ADV
fcis-15825	79	2	,	,	PUNCT
fcis-15825	79	3	the	the	DET
fcis-15825	79	4	initial	initial	ADJ
fcis-15825	79	5	road	road	NOUN
fcis-15825	79	6	disease	disease	NOUN
fcis-15825	79	7	datasets	dataset	NOUN
fcis-15825	79	8	containing	contain	VERB
fcis-15825	79	9	12,378	12,378	NUM
fcis-15825	79	10	pictures	picture	NOUN
fcis-15825	79	11	was	be	AUX
fcis-15825	79	12	obtained	obtain	VERB
fcis-15825	79	13	by	by	ADP
fcis-15825	79	14	integrating	integrate	VERB
fcis-15825	79	15	the	the	DET
fcis-15825	79	16	data	datum	NOUN
fcis-15825	79	17	of	of	ADP
fcis-15825	79	18	three	three	NUM
fcis-15825	79	19	data	datum	NOUN
fcis-15825	79	20	sets	set	NOUN
fcis-15825	79	21	,	,	PUNCT
fcis-15825	79	22	and	and	CCONJ
fcis-15825	79	23	the	the	DET
fcis-15825	79	24	disease	disease	NOUN
fcis-15825	79	25	was	be	AUX
fcis-15825	79	26	marked	mark	VERB
fcis-15825	79	27	with	with	ADP
fcis-15825	79	28	labelimg	labelimg	ADJ
fcis-15825	79	29	software	software	NOUN
fcis-15825	79	30	,	,	PUNCT
fcis-15825	79	31	which	which	PRON
fcis-15825	79	32	mainly	mainly	ADV
fcis-15825	79	33	marked	mark	VERB
fcis-15825	79	34	five	five	NUM
fcis-15825	79	35	common	common	ADJ
fcis-15825	79	36	road	road	NOUN
fcis-15825	79	37	diseases	disease	NOUN
fcis-15825	79	38	,	,	PUNCT
fcis-15825	79	39	namely	namely	ADV
fcis-15825	79	40	crack	crack	VERB
fcis-15825	79	41	,	,	PUNCT
fcis-15825	79	42	net	net	ADJ
fcis-15825	79	43	crack	crack	NOUN
fcis-15825	79	44	,	,	PUNCT
fcis-15825	79	45	pothole	pothole	NOUN
fcis-15825	79	46	,	,	PUNCT
fcis-15825	79	47	patch	patch	NOUN
fcis-15825	79	48	-	-	PUNCT
fcis-15825	79	49	crack	crack	NOUN
fcis-15825	79	50	and	and	CCONJ
fcis-15825	79	51	patch	patch	NOUN
fcis-15825	79	52	-	-	PUNCT
fcis-15825	79	53	net	net	NOUN
fcis-15825	79	54	.	.	PUNCT
fcis-15825	80	1	then	then	ADV
fcis-15825	80	2	,	,	PUNCT
fcis-15825	80	3	the	the	DET
fcis-15825	80	4	data	datum	NOUN
fcis-15825	80	5	set	set	VERB
fcis-15825	80	6	is	be	AUX
fcis-15825	80	7	expanded	expand	VERB
fcis-15825	80	8	by	by	ADP
fcis-15825	80	9	data	datum	NOUN
fcis-15825	80	10	enhancement	enhancement	NOUN
fcis-15825	80	11	method	method	NOUN
fcis-15825	80	12	,	,	PUNCT
fcis-15825	80	13	and	and	CCONJ
fcis-15825	80	14	the	the	DET
fcis-15825	80	15	road	road	NOUN
fcis-15825	80	16	disease	disease	NOUN
fcis-15825	80	17	images	image	VERB
fcis-15825	80	18	under	under	ADP
fcis-15825	80	19	different	different	ADJ
fcis-15825	80	20	lighting	lighting	NOUN
fcis-15825	80	21	conditions	condition	NOUN
fcis-15825	80	22	are	be	AUX
fcis-15825	80	23	simulated	simulate	VERB
fcis-15825	80	24	by	by	ADP
fcis-15825	80	25	randomly	randomly	ADV
fcis-15825	80	26	adjusting	adjust	VERB
fcis-15825	80	27	the	the	DET
fcis-15825	80	28	brightness	brightness	NOUN
fcis-15825	80	29	of	of	ADP
fcis-15825	80	30	the	the	DET
fcis-15825	80	31	images	image	NOUN
fcis-15825	80	32	.	.	PUNCT
fcis-15825	81	1	by	by	ADP
fcis-15825	81	2	simulating	simulate	VERB
fcis-15825	81	3	rainy	rainy	ADJ
fcis-15825	81	4	and	and	CCONJ
fcis-15825	81	5	foggy	foggy	ADJ
fcis-15825	81	6	days	day	NOUN
fcis-15825	81	7	,	,	PUNCT
fcis-15825	81	8	the	the	DET
fcis-15825	81	9	weather	weather	NOUN
fcis-15825	81	10	data	datum	NOUN
fcis-15825	81	11	that	that	PRON
fcis-15825	81	12	may	may	AUX
fcis-15825	81	13	be	be	AUX
fcis-15825	81	14	encountered	encounter	VERB
fcis-15825	81	15	in	in	ADP
fcis-15825	81	16	real	real	ADJ
fcis-15825	81	17	scenes	scene	NOUN
fcis-15825	81	18	are	be	AUX
fcis-15825	81	19	supplemented	supplement	VERB
fcis-15825	81	20	.	.	PUNCT
fcis-15825	82	1	the	the	DET
fcis-15825	82	2	effects	effect	NOUN
fcis-15825	82	3	of	of	ADP
fcis-15825	82	4	different	different	ADJ
fcis-15825	82	5	data	data	NOUN
fcis-15825	82	6	enhancement	enhancement	NOUN
fcis-15825	82	7	techniques	technique	NOUN
fcis-15825	82	8	are	be	AUX
fcis-15825	82	9	shown	show	VERB
fcis-15825	82	10	in	in	ADP
fcis-15825	82	11	figure	figure	NOUN
fcis-15825	82	12	5	5	NUM
fcis-15825	82	13	.	.	PUNCT
fcis-15825	83	1	the	the	DET
fcis-15825	83	2	expanded	expand	VERB
fcis-15825	83	3	data	data	NOUN
fcis-15825	83	4	set	set	NOUN
fcis-15825	83	5	includes	include	VERB
fcis-15825	83	6	49,512	49,512	NUM
fcis-15825	83	7	road	road	NOUN
fcis-15825	83	8	data	datum	NOUN
fcis-15825	83	9	.	.	PUNCT
fcis-15825	84	1	finally	finally	ADV
fcis-15825	84	2	,	,	PUNCT
fcis-15825	84	3	according	accord	VERB
fcis-15825	84	4	to	to	ADP
fcis-15825	84	5	the	the	DET
fcis-15825	84	6	ratio	ratio	NOUN
fcis-15825	84	7	of	of	ADP
fcis-15825	84	8	8	8	NUM
fcis-15825	84	9	:	:	SYM
fcis-15825	84	10	2	2	NUM
fcis-15825	84	11	,	,	PUNCT
fcis-15825	84	12	the	the	DET
fcis-15825	84	13	training	training	NOUN
fcis-15825	84	14	set	set	NOUN
fcis-15825	84	15	includes	include	VERB
fcis-15825	84	16	39,610	39,610	NUM
fcis-15825	84	17	image	image	NOUN
fcis-15825	84	18	data	datum	NOUN
fcis-15825	84	19	and	and	CCONJ
fcis-15825	84	20	the	the	DET
fcis-15825	84	21	verification	verification	NOUN
fcis-15825	84	22	set	set	NOUN
fcis-15825	84	23	includes	include	VERB
fcis-15825	84	24	9,902	9,902	NUM
fcis-15825	84	25	image	image	NOUN
fcis-15825	84	26	data	datum	NOUN
fcis-15825	84	27	.	.	PUNCT
fcis-15825	85	1	a	a	DET
fcis-15825	85	2	)	)	PUNCT
fcis-15825	85	3	original	original	ADJ
fcis-15825	85	4	picture	picture	NOUN
fcis-15825	85	5	b	b	NOUN
fcis-15825	85	6	)	)	PUNCT
fcis-15825	85	7	randomly	randomly	ADV
fcis-15825	85	8	adjusts	adjust	VERB
fcis-15825	85	9	brightness	brightness	NOUN
fcis-15825	85	10	c	c	NOUN
fcis-15825	85	11	)	)	PUNCT
fcis-15825	85	12	simulated	simulate	VERB
fcis-15825	85	13	foggy	foggy	ADJ
fcis-15825	85	14	days	day	NOUN
fcis-15825	85	15	d	d	NOUN
fcis-15825	85	16	)	)	PUNCT
fcis-15825	85	17	simulated	simulate	VERB
fcis-15825	85	18	rainy	rainy	ADJ
fcis-15825	85	19	day	day	NOUN
fcis-15825	85	20	fig	fig	NOUN
fcis-15825	85	21	5	5	NUM
fcis-15825	85	22	.	.	PUNCT
fcis-15825	86	1	data	datum	NOUN
fcis-15825	86	2	enhancement	enhancement	NOUN
fcis-15825	86	3	method	method	NOUN
fcis-15825	86	4	3.2	3.2	NUM
fcis-15825	86	5	.	.	PUNCT
fcis-15825	87	1	experimental	experimental	ADJ
fcis-15825	87	2	environment	environment	NOUN
fcis-15825	87	3	in	in	ADP
fcis-15825	87	4	this	this	DET
fcis-15825	87	5	paper	paper	NOUN
fcis-15825	87	6	,	,	PUNCT
fcis-15825	87	7	python3.7	python3.7	NOUN
fcis-15825	87	8	version	version	VERB
fcis-15825	87	9	3.7	3.7	NUM
fcis-15825	87	10	is	be	AUX
fcis-15825	87	11	selected	select	VERB
fcis-15825	87	12	as	as	ADP
fcis-15825	87	13	the	the	DET
fcis-15825	87	14	development	development	NOUN
fcis-15825	87	15	language	language	NOUN
fcis-15825	87	16	,	,	PUNCT
fcis-15825	87	17	and	and	CCONJ
fcis-15825	87	18	pytorch	pytorch	NOUN
fcis-15825	87	19	,	,	PUNCT
fcis-15825	87	20	a	a	DET
fcis-15825	87	21	deep	deep	ADJ
fcis-15825	87	22	learning	learn	VERB
fcis-15825	87	23	116	116	NUM
fcis-15825	87	24	framework	framework	NOUN
fcis-15825	87	25	,	,	PUNCT
fcis-15825	87	26	is	be	AUX
fcis-15825	87	27	used	use	VERB
fcis-15825	87	28	to	to	PART
fcis-15825	87	29	build	build	VERB
fcis-15825	87	30	a	a	DET
fcis-15825	87	31	neural	neural	ADJ
fcis-15825	87	32	network	network	NOUN
fcis-15825	87	33	.	.	PUNCT
fcis-15825	88	1	the	the	DET
fcis-15825	88	2	experiment	experiment	NOUN
fcis-15825	88	3	is	be	AUX
fcis-15825	88	4	conducted	conduct	VERB
fcis-15825	88	5	on	on	ADP
fcis-15825	88	6	windows10	windows10	NOUN
fcis-15825	88	7	operating	operating	NOUN
fcis-15825	88	8	system	system	NOUN
fcis-15825	88	9	,	,	PUNCT
fcis-15825	88	10	with	with	ADP
fcis-15825	88	11	intel(r	intel(r	PROPN
fcis-15825	88	12	)	)	PUNCT
fcis-15825	88	13	core	core	NOUN
fcis-15825	88	14	(	(	PUNCT
fcis-15825	88	15	tm)i7	tm)i7	PROPN
fcis-15825	88	16	-	-	PUNCT
fcis-15825	88	17	13700h	13700h	NUM
fcis-15825	88	18	as	as	ADP
fcis-15825	88	19	the	the	DET
fcis-15825	88	20	cpu	cpu	NOUN
fcis-15825	88	21	version	version	NOUN
fcis-15825	88	22	and	and	CCONJ
fcis-15825	88	23	nvidia	nvidia	PROPN
fcis-15825	88	24	geforce	geforce	NOUN
fcis-15825	88	25	rtx3050	rtx3050	VERB
fcis-15825	88	26	as	as	ADP
fcis-15825	88	27	the	the	DET
fcis-15825	88	28	gpu	gpu	PROPN
fcis-15825	88	29	version	version	NOUN
fcis-15825	88	30	.	.	PUNCT
fcis-15825	89	1	the	the	DET
fcis-15825	89	2	initial	initial	ADJ
fcis-15825	89	3	parameters	parameter	NOUN
fcis-15825	89	4	of	of	ADP
fcis-15825	89	5	the	the	DET
fcis-15825	89	6	experiment	experiment	NOUN
fcis-15825	89	7	are	be	AUX
fcis-15825	89	8	shown	show	VERB
fcis-15825	89	9	in	in	ADP
fcis-15825	89	10	table	table	NOUN
fcis-15825	89	11	1	1	NUM
fcis-15825	89	12	.	.	PUNCT
fcis-15825	89	13	table	table	NOUN
fcis-15825	89	14	1	1	NUM
fcis-15825	89	15	.	.	PUNCT
fcis-15825	89	16	parameter	parameter	NOUN
fcis-15825	89	17	configuration	configuration	NOUN
fcis-15825	89	18	experimental	experimental	ADJ
fcis-15825	89	19	hyperparameter	hyperparameter	NOUN
fcis-15825	89	20	parameter	parameter	NOUN
fcis-15825	89	21	setting	set	VERB
fcis-15825	89	22	input	input	NOUN
fcis-15825	89	23	picture	picture	NOUN
fcis-15825	89	24	608	608	NUM
fcis-15825	89	25	*	*	NUM
fcis-15825	89	26	608	608	NUM
fcis-15825	89	27	*	*	SYM
fcis-15825	89	28	3	3	NUM
fcis-15825	89	29	number	number	NOUN
fcis-15825	89	30	of	of	ADP
fcis-15825	89	31	categories	category	NOUN
fcis-15825	89	32	five	five	NUM
fcis-15825	89	33	batch	batch	NOUN
fcis-15825	89	34	size	size	NOUN
fcis-15825	89	35	eight	eight	NUM
fcis-15825	89	36	training	training	NOUN
fcis-15825	89	37	times	time	NOUN
fcis-15825	89	38	100	100	NUM
fcis-15825	89	39	learning	learning	NOUN
fcis-15825	89	40	rate	rate	NOUN
fcis-15825	89	41	0.01	0.01	NUM
fcis-15825	89	42	optimizer	optimizer	NOUN
fcis-15825	89	43	adam	adam	PROPN
fcis-15825	89	44	3.3	3.3	NUM
fcis-15825	89	45	.	.	PUNCT
fcis-15825	90	1	evaluation	evaluation	NOUN
fcis-15825	90	2	indicators	indicator	NOUN
fcis-15825	90	3	in	in	ADP
fcis-15825	90	4	order	order	NOUN
fcis-15825	90	5	to	to	PART
fcis-15825	90	6	verify	verify	VERB
fcis-15825	90	7	the	the	DET
fcis-15825	90	8	performance	performance	NOUN
fcis-15825	90	9	of	of	ADP
fcis-15825	90	10	the	the	DET
fcis-15825	90	11	improved	improved	ADJ
fcis-15825	90	12	detection	detection	NOUN
fcis-15825	90	13	model	model	NOUN
fcis-15825	90	14	,	,	PUNCT
fcis-15825	90	15	accuracy	accuracy	NOUN
fcis-15825	90	16	(	(	PUNCT
fcis-15825	90	17	pr	pr	NOUN
fcis-15825	90	18	)	)	PUNCT
fcis-15825	90	19	,	,	PUNCT
fcis-15825	90	20	recall	recall	INTJ
fcis-15825	90	21	(	(	PUNCT
fcis-15825	90	22	re	re	NOUN
fcis-15825	90	23	)	)	PUNCT
fcis-15825	90	24	and	and	CCONJ
fcis-15825	90	25	average	average	ADJ
fcis-15825	90	26	accuracy	accuracy	NOUN
fcis-15825	90	27	(	(	PUNCT
fcis-15825	90	28	map	map	NOUN
fcis-15825	90	29	)	)	PUNCT
fcis-15825	90	30	are	be	AUX
fcis-15825	90	31	used	use	VERB
fcis-15825	90	32	as	as	ADP
fcis-15825	90	33	evaluation	evaluation	NOUN
fcis-15825	90	34	indexes	index	NOUN
fcis-15825	90	35	.	.	PUNCT
fcis-15825	91	1	accuracy	accuracy	NOUN
fcis-15825	91	2	is	be	AUX
fcis-15825	91	3	the	the	DET
fcis-15825	91	4	ratio	ratio	NOUN
fcis-15825	91	5	of	of	ADP
fcis-15825	91	6	the	the	DET
fcis-15825	91	7	number	number	NOUN
fcis-15825	91	8	of	of	ADP
fcis-15825	91	9	correctly	correctly	ADV
fcis-15825	91	10	detected	detect	VERB
fcis-15825	91	11	diseases	disease	NOUN
fcis-15825	91	12	to	to	ADP
fcis-15825	91	13	the	the	DET
fcis-15825	91	14	number	number	NOUN
fcis-15825	91	15	of	of	ADP
fcis-15825	91	16	positive	positive	ADJ
fcis-15825	91	17	samples	sample	NOUN
fcis-15825	91	18	judged	judge	VERB
fcis-15825	91	19	by	by	ADP
fcis-15825	91	20	the	the	DET
fcis-15825	91	21	model	model	NOUN
fcis-15825	91	22	,	,	PUNCT
fcis-15825	91	23	recall	recall	PROPN
fcis-15825	91	24	is	be	AUX
fcis-15825	91	25	the	the	DET
fcis-15825	91	26	ratio	ratio	NOUN
fcis-15825	91	27	of	of	ADP
fcis-15825	91	28	correctly	correctly	ADV
fcis-15825	91	29	identified	identify	VERB
fcis-15825	91	30	samples	sample	NOUN
fcis-15825	91	31	to	to	ADP
fcis-15825	91	32	the	the	DET
fcis-15825	91	33	number	number	NOUN
fcis-15825	91	34	of	of	ADP
fcis-15825	91	35	positive	positive	ADJ
fcis-15825	91	36	samples	sample	NOUN
fcis-15825	91	37	,	,	PUNCT
fcis-15825	91	38	and	and	CCONJ
fcis-15825	91	39	map	map	NOUN
fcis-15825	91	40	is	be	AUX
fcis-15825	91	41	a	a	DET
fcis-15825	91	42	comprehensive	comprehensive	ADJ
fcis-15825	91	43	index	index	NOUN
fcis-15825	91	44	to	to	PART
fcis-15825	91	45	evaluate	evaluate	VERB
fcis-15825	91	46	the	the	DET
fcis-15825	91	47	overall	overall	ADJ
fcis-15825	91	48	performance	performance	NOUN
fcis-15825	91	49	of	of	ADP
fcis-15825	91	50	the	the	DET
fcis-15825	91	51	detection	detection	NOUN
fcis-15825	91	52	model	model	NOUN
fcis-15825	91	53	,	,	PUNCT
fcis-15825	91	54	which	which	PRON
fcis-15825	91	55	can	can	AUX
fcis-15825	91	56	measure	measure	VERB
fcis-15825	91	57	the	the	DET
fcis-15825	91	58	average	average	ADJ
fcis-15825	91	59	accuracy	accuracy	NOUN
fcis-15825	91	60	of	of	ADP
fcis-15825	91	61	detection	detection	NOUN
fcis-15825	91	62	of	of	ADP
fcis-15825	91	63	all	all	DET
fcis-15825	91	64	disease	disease	NOUN
fcis-15825	91	65	categories	category	NOUN
fcis-15825	91	66	in	in	ADP
fcis-15825	91	67	the	the	DET
fcis-15825	91	68	model	model	NOUN
fcis-15825	91	69	.	.	PUNCT
fcis-15825	92	1	tp	tp	PART
fcis-15825	92	2	pr=	pr=	VERB
fcis-15825	92	3	tp+fp	tp+fp	NUM
fcis-15825	92	4	(	(	PUNCT
fcis-15825	92	5	3	3	NUM
fcis-15825	92	6	)	)	PUNCT
fcis-15825	92	7	tp	tp	PART
fcis-15825	92	8	re=	re=	VERB
fcis-15825	92	9	tp+fn	tp+fn	X
fcis-15825	92	10	(	(	PUNCT
fcis-15825	92	11	4	4	NUM
fcis-15825	92	12	)	)	PUNCT
fcis-15825	92	13	1	1	NUM
fcis-15825	92	14	2	2	NUM
fcis-15825	92	15	np	np	NOUN
fcis-15825	92	16	+	+	NOUN
fcis-15825	92	17	p	p	ADJ
fcis-15825	92	18	+	+	PROPN
fcis-15825	92	19	l+p	l+p	PROPN
fcis-15825	92	20	ap=	ap=	PROPN
fcis-15825	92	21	n	n	X
fcis-15825	92	22	(	(	PUNCT
fcis-15825	92	23	5	5	NUM
fcis-15825	92	24	)	)	PUNCT
fcis-15825	92	25	i	i	PROPN
fcis-15825	92	26	=	=	PROPN
fcis-15825	92	27	n	n	SYM
fcis-15825	92	28	ii=0	ii=0	PROPN
fcis-15825	92	29	1	1	NUM
fcis-15825	92	30	map=	map=	PROPN
fcis-15825	92	31	ap	ap	PROPN
fcis-15825	92	32	n	n	NOUN
fcis-15825	92	33			X
fcis-15825	92	34	(	(	PUNCT
fcis-15825	92	35	6	6	NUM
fcis-15825	92	36	)	)	PUNCT
fcis-15825	92	37	among	among	ADP
fcis-15825	92	38	them	they	PRON
fcis-15825	92	39	,	,	PUNCT
fcis-15825	92	40	tp	tp	NOUN
fcis-15825	92	41	represents	represent	VERB
fcis-15825	92	42	the	the	DET
fcis-15825	92	43	number	number	NOUN
fcis-15825	92	44	of	of	ADP
fcis-15825	92	45	positive	positive	ADJ
fcis-15825	92	46	samples	sample	NOUN
fcis-15825	92	47	correctly	correctly	ADV
fcis-15825	92	48	identified	identify	VERB
fcis-15825	92	49	by	by	ADP
fcis-15825	92	50	the	the	DET
fcis-15825	92	51	model	model	NOUN
fcis-15825	92	52	,	,	PUNCT
fcis-15825	92	53	fn	fn	PROPN
fcis-15825	92	54	represents	represent	VERB
fcis-15825	92	55	the	the	DET
fcis-15825	92	56	number	number	NOUN
fcis-15825	92	57	of	of	ADP
fcis-15825	92	58	negative	negative	ADJ
fcis-15825	92	59	samples	sample	NOUN
fcis-15825	92	60	wrongly	wrongly	ADV
fcis-15825	92	61	detected	detect	VERB
fcis-15825	92	62	by	by	ADP
fcis-15825	92	63	the	the	DET
fcis-15825	92	64	model	model	NOUN
fcis-15825	92	65	,	,	PUNCT
fcis-15825	92	66	and	and	CCONJ
fcis-15825	92	67	fp	fp	PROPN
fcis-15825	92	68	is	be	AUX
fcis-15825	92	69	the	the	DET
fcis-15825	92	70	number	number	NOUN
fcis-15825	92	71	of	of	ADP
fcis-15825	92	72	positive	positive	ADJ
fcis-15825	92	73	samples	sample	NOUN
fcis-15825	92	74	wrongly	wrongly	ADV
fcis-15825	92	75	detected	detect	VERB
fcis-15825	92	76	by	by	ADP
fcis-15825	92	77	the	the	DET
fcis-15825	92	78	model	model	NOUN
fcis-15825	92	79	.	.	PUNCT
fcis-15825	93	1	3.4	3.4	NUM
fcis-15825	93	2	.	.	PUNCT
fcis-15825	93	3	ablation	ablation	NOUN
fcis-15825	93	4	experiment	experiment	NOUN
fcis-15825	93	5	the	the	DET
fcis-15825	93	6	influence	influence	NOUN
fcis-15825	93	7	of	of	ADP
fcis-15825	93	8	the	the	DET
fcis-15825	93	9	improved	improved	ADJ
fcis-15825	93	10	method	method	NOUN
fcis-15825	93	11	on	on	ADP
fcis-15825	93	12	the	the	DET
fcis-15825	93	13	model	model	NOUN
fcis-15825	93	14	detection	detection	NOUN
fcis-15825	93	15	performance	performance	NOUN
fcis-15825	93	16	is	be	AUX
fcis-15825	93	17	verified	verify	VERB
fcis-15825	93	18	by	by	ADP
fcis-15825	93	19	ablation	ablation	NOUN
fcis-15825	93	20	experiments	experiment	NOUN
fcis-15825	93	21	.	.	PUNCT
fcis-15825	94	1	based	base	VERB
fcis-15825	94	2	on	on	ADP
fcis-15825	94	3	yolov5s	yolov5s	PROPN
fcis-15825	94	4	network	network	NOUN
fcis-15825	94	5	,	,	PUNCT
fcis-15825	94	6	model	model	NOUN
fcis-15825	94	7	yolov5s	yolov5s	PROPN
fcis-15825	94	8	-	-	PUNCT
fcis-15825	94	9	p	p	PROPN
fcis-15825	94	10	is	be	AUX
fcis-15825	94	11	obtained	obtain	VERB
fcis-15825	94	12	by	by	ADP
fcis-15825	94	13	adding	add	VERB
fcis-15825	94	14	small	small	ADJ
fcis-15825	94	15	target	target	NOUN
fcis-15825	94	16	detection	detection	NOUN
fcis-15825	94	17	layer	layer	NOUN
fcis-15825	94	18	,	,	PUNCT
fcis-15825	94	19	model	model	NOUN
fcis-15825	94	20	yolov5s	yolov5s	PROPN
fcis-15825	94	21	-	-	PUNCT
fcis-15825	94	22	c	c	PROPN
fcis-15825	94	23	is	be	AUX
fcis-15825	94	24	obtained	obtain	VERB
fcis-15825	94	25	by	by	ADP
fcis-15825	94	26	adding	add	VERB
fcis-15825	94	27	cbam	cbam	NOUN
fcis-15825	94	28	attention	attention	NOUN
fcis-15825	94	29	mechanism	mechanism	NOUN
fcis-15825	94	30	,	,	PUNCT
fcis-15825	94	31	and	and	CCONJ
fcis-15825	94	32	model	model	PROPN
fcis-15825	94	33	yolov5s	yolov5s	PROPN
fcis-15825	94	34	-	-	PUNCT
fcis-15825	94	35	b	b	PROPN
fcis-15825	94	36	is	be	AUX
fcis-15825	94	37	obtained	obtain	VERB
fcis-15825	94	38	by	by	ADP
fcis-15825	94	39	adding	add	VERB
fcis-15825	94	40	bifpn	bifpn	PROPN
fcis-15825	94	41	module	module	NOUN
fcis-15825	94	42	.	.	PUNCT
fcis-15825	95	1	the	the	DET
fcis-15825	95	2	experimental	experimental	ADJ
fcis-15825	95	3	results	result	NOUN
fcis-15825	95	4	are	be	AUX
fcis-15825	95	5	shown	show	VERB
fcis-15825	95	6	in	in	ADP
fcis-15825	95	7	table	table	NOUN
fcis-15825	95	8	2	2	NUM
fcis-15825	95	9	.	.	PUNCT
fcis-15825	95	10	table	table	NOUN
fcis-15825	95	11	2	2	NUM
fcis-15825	95	12	.	.	PUNCT
fcis-15825	95	13	results	result	NOUN
fcis-15825	95	14	of	of	ADP
fcis-15825	95	15	ablation	ablation	NOUN
fcis-15825	95	16	experiment	experiment	NOUN
fcis-15825	95	17	model	model	NOUN
fcis-15825	95	18	p2	p2	PROPN
fcis-15825	95	19	detection	detection	NOUN
fcis-15825	95	20	layer	layer	NOUN
fcis-15825	95	21	cbam	cbam	NOUN
fcis-15825	95	22	bifpn	bifpn	PROPN
fcis-15825	95	23	pr	pr	X
fcis-15825	95	24	re	re	PUNCT
fcis-15825	95	25	map@0.5	map@0.5	PROPN
fcis-15825	95	26	yolov5s	yolov5s	PROPN
fcis-15825	95	27	×	×	ADJ
fcis-15825	95	28	×	×	NOUN
fcis-15825	95	29	×	×	NOUN
fcis-15825	95	30	82.1	82.1	NUM
fcis-15825	95	31	%	%	NOUN
fcis-15825	95	32	76.3	76.3	NUM
fcis-15825	95	33	%	%	NOUN
fcis-15825	95	34	79.8	79.8	NUM
fcis-15825	95	35	%	%	NOUN
fcis-15825	95	36	yolov5s	yolov5s	NOUN
fcis-15825	95	37	-	-	PUNCT
fcis-15825	95	38	p	p	NOUN
fcis-15825	95	39	√	√	NUM
fcis-15825	95	40	×	×	NOUN
fcis-15825	95	41	×	×	NOUN
fcis-15825	95	42	82.1%(+2.1	82.1%(+2.1	NUM
fcis-15825	95	43	%	%	NOUN
fcis-15825	95	44	)	)	PUNCT
fcis-15825	96	1	76.3%(+5.2	76.3%(+5.2	NUM
fcis-15825	96	2	%	%	NOUN
fcis-15825	96	3	)	)	PUNCT
fcis-15825	96	4	79.8%(+1.5	79.8%(+1.5	NUM
fcis-15825	96	5	%	%	NOUN
fcis-15825	96	6	)	)	PUNCT
fcis-15825	96	7	yolov5s	yolov5s	PROPN
fcis-15825	96	8	-	-	PUNCT
fcis-15825	96	9	c	c	PROPN
fcis-15825	96	10	×	×	NOUN
fcis-15825	96	11	√	√	NUM
fcis-15825	96	12	×	×	NOUN
fcis-15825	96	13	82.1%(+3.5	82.1%(+3.5	NUM
fcis-15825	96	14	%	%	NOUN
fcis-15825	96	15	)	)	PUNCT
fcis-15825	97	1	76.3%(+2.8	76.3%(+2.8	NUM
fcis-15825	97	2	%	%	NOUN
fcis-15825	97	3	)	)	PUNCT
fcis-15825	98	1	79.8%(+2.2	79.8%(+2.2	NUM
fcis-15825	98	2	%	%	NOUN
fcis-15825	98	3	)	)	PUNCT
fcis-15825	98	4	yolov5s	yolov5s	PROPN
fcis-15825	98	5	-	-	PUNCT
fcis-15825	98	6	b	b	PROPN
fcis-15825	98	7	×	×	NOUN
fcis-15825	98	8	×	×	NOUN
fcis-15825	98	9	√	√	NOUN
fcis-15825	98	10	82.1%(+2.9	82.1%(+2.9	NUM
fcis-15825	98	11	%	%	NOUN
fcis-15825	98	12	)	)	PUNCT
fcis-15825	98	13	76.3%(+1.6	76.3%(+1.6	NUM
fcis-15825	98	14	%	%	NOUN
fcis-15825	98	15	)	)	PUNCT
fcis-15825	98	16	79.8%(+1.7	79.8%(+1.7	NUM
fcis-15825	98	17	%	%	NOUN
fcis-15825	98	18	)	)	PUNCT
fcis-15825	99	1	yolov5scbp	yolov5scbp	PRON
fcis-15825	99	2	√	√	NOUN
fcis-15825	100	1	√	√	ADP
fcis-15825	101	1	√	√	NUM
fcis-15825	101	2	82.1%(+6.1	82.1%(+6.1	NUM
fcis-15825	101	3	%	%	NOUN
fcis-15825	101	4	)	)	PUNCT
fcis-15825	102	1	76.3%(+4.4	76.3%(+4.4	NUM
fcis-15825	102	2	%	%	NOUN
fcis-15825	102	3	)	)	PUNCT
fcis-15825	102	4	79.8%(+5.7	79.8%(+5.7	NUM
fcis-15825	102	5	%	%	NOUN
fcis-15825	102	6	)	)	PUNCT
fcis-15825	102	7	the	the	DET
fcis-15825	102	8	results	result	NOUN
fcis-15825	102	9	of	of	ADP
fcis-15825	102	10	ablation	ablation	NOUN
fcis-15825	102	11	experiments	experiment	NOUN
fcis-15825	102	12	show	show	VERB
fcis-15825	102	13	that	that	SCONJ
fcis-15825	102	14	the	the	DET
fcis-15825	102	15	accuracy	accuracy	NOUN
fcis-15825	102	16	and	and	CCONJ
fcis-15825	102	17	recall	recall	NOUN
fcis-15825	102	18	of	of	ADP
fcis-15825	102	19	the	the	DET
fcis-15825	102	20	original	original	ADJ
fcis-15825	102	21	model	model	NOUN
fcis-15825	102	22	are	be	AUX
fcis-15825	102	23	improved	improve	VERB
fcis-15825	102	24	by	by	ADP
fcis-15825	102	25	3.5	3.5	NUM
fcis-15825	102	26	%	%	NOUN
fcis-15825	102	27	and	and	CCONJ
fcis-15825	102	28	2.8	2.8	NUM
fcis-15825	102	29	%	%	NOUN
fcis-15825	102	30	respectively	respectively	ADV
fcis-15825	102	31	by	by	ADP
fcis-15825	102	32	adding	add	VERB
fcis-15825	102	33	cbam	cbam	NOUN
fcis-15825	102	34	attention	attention	NOUN
fcis-15825	102	35	mechanism	mechanism	NOUN
fcis-15825	102	36	to	to	ADP
fcis-15825	102	37	the	the	DET
fcis-15825	102	38	model	model	NOUN
fcis-15825	102	39	detector	detector	NOUN
fcis-15825	102	40	,	,	PUNCT
fcis-15825	102	41	which	which	PRON
fcis-15825	102	42	shows	show	VERB
fcis-15825	102	43	that	that	SCONJ
fcis-15825	102	44	the	the	DET
fcis-15825	102	45	introduction	introduction	NOUN
fcis-15825	102	46	of	of	ADP
fcis-15825	102	47	cbam	cbam	NOUN
fcis-15825	102	48	attention	attention	NOUN
fcis-15825	102	49	mechanism	mechanism	NOUN
fcis-15825	102	50	pays	pay	VERB
fcis-15825	102	51	more	more	ADJ
fcis-15825	102	52	attention	attention	NOUN
fcis-15825	102	53	to	to	ADP
fcis-15825	102	54	pavement	pavement	NOUN
fcis-15825	102	55	diseases	disease	NOUN
fcis-15825	102	56	information	information	NOUN
fcis-15825	102	57	,	,	PUNCT
fcis-15825	102	58	effectively	effectively	ADV
fcis-15825	102	59	improves	improve	VERB
fcis-15825	102	60	the	the	DET
fcis-15825	102	61	feature	feature	NOUN
fcis-15825	102	62	extraction	extraction	NOUN
fcis-15825	102	63	ability	ability	NOUN
fcis-15825	102	64	of	of	ADP
fcis-15825	102	65	the	the	DET
fcis-15825	102	66	model	model	NOUN
fcis-15825	102	67	and	and	CCONJ
fcis-15825	102	68	further	far	ADV
fcis-15825	102	69	improves	improve	VERB
fcis-15825	102	70	the	the	DET
fcis-15825	102	71	performance	performance	NOUN
fcis-15825	102	72	of	of	ADP
fcis-15825	102	73	the	the	DET
fcis-15825	102	74	model	model	NOUN
fcis-15825	102	75	.	.	PUNCT
fcis-15825	103	1	using	use	VERB
fcis-15825	103	2	bifpn	bifpn	PROPN
fcis-15825	103	3	instead	instead	ADV
fcis-15825	103	4	of	of	ADP
fcis-15825	103	5	panet	panet	NOUN
fcis-15825	103	6	in	in	ADP
fcis-15825	103	7	the	the	DET
fcis-15825	103	8	neck	neck	NOUN
fcis-15825	103	9	network	network	NOUN
fcis-15825	103	10	improves	improve	VERB
fcis-15825	103	11	the	the	DET
fcis-15825	103	12	accuracy	accuracy	NOUN
fcis-15825	103	13	and	and	CCONJ
fcis-15825	103	14	recall	recall	NOUN
fcis-15825	103	15	of	of	ADP
fcis-15825	103	16	the	the	DET
fcis-15825	103	17	original	original	ADJ
fcis-15825	103	18	model	model	NOUN
fcis-15825	103	19	to	to	ADP
fcis-15825	103	20	85	85	NUM
fcis-15825	103	21	%	%	NOUN
fcis-15825	103	22	and	and	CCONJ
fcis-15825	103	23	77.9	77.9	NUM
fcis-15825	103	24	%	%	NOUN
fcis-15825	103	25	,	,	PUNCT
fcis-15825	103	26	which	which	PRON
fcis-15825	103	27	shows	show	VERB
fcis-15825	103	28	that	that	SCONJ
fcis-15825	103	29	bifpn	bifpn	NOUN
fcis-15825	103	30	,	,	PUNCT
fcis-15825	103	31	as	as	ADP
fcis-15825	103	32	a	a	DET
fcis-15825	103	33	feature	feature	NOUN
fcis-15825	103	34	fusion	fusion	NOUN
fcis-15825	103	35	network	network	NOUN
fcis-15825	103	36	,	,	PUNCT
fcis-15825	103	37	can	can	AUX
fcis-15825	103	38	better	well	ADV
fcis-15825	103	39	realize	realize	VERB
fcis-15825	103	40	cross	cross	ADJ
fcis-15825	103	41	-	-	ADJ
fcis-15825	103	42	layer	layer	ADJ
fcis-15825	103	43	communication	communication	NOUN
fcis-15825	103	44	and	and	CCONJ
fcis-15825	103	45	weighted	weight	VERB
fcis-15825	103	46	fusion	fusion	NOUN
fcis-15825	103	47	of	of	ADP
fcis-15825	103	48	feature	feature	NOUN
fcis-15825	103	49	information	information	NOUN
fcis-15825	103	50	and	and	CCONJ
fcis-15825	103	51	effectively	effectively	ADV
fcis-15825	103	52	solve	solve	VERB
fcis-15825	103	53	the	the	DET
fcis-15825	103	54	problem	problem	NOUN
fcis-15825	103	55	of	of	ADP
fcis-15825	103	56	losing	lose	VERB
fcis-15825	103	57	high	high	ADJ
fcis-15825	103	58	-	-	PUNCT
fcis-15825	103	59	level	level	NOUN
fcis-15825	103	60	semantic	semantic	ADJ
fcis-15825	103	61	information	information	NOUN
fcis-15825	103	62	.	.	PUNCT
fcis-15825	104	1	after	after	ADP
fcis-15825	104	2	adding	add	VERB
fcis-15825	104	3	p2	p2	PROPN
fcis-15825	104	4	detection	detection	NOUN
fcis-15825	104	5	layer	layer	NOUN
fcis-15825	104	6	to	to	ADP
fcis-15825	104	7	the	the	DET
fcis-15825	104	8	network	network	NOUN
fcis-15825	104	9	prediction	prediction	NOUN
fcis-15825	104	10	head	head	NOUN
fcis-15825	104	11	,	,	PUNCT
fcis-15825	104	12	the	the	DET
fcis-15825	104	13	accuracy	accuracy	NOUN
fcis-15825	104	14	and	and	CCONJ
fcis-15825	104	15	recall	recall	NOUN
fcis-15825	104	16	of	of	ADP
fcis-15825	104	17	the	the	DET
fcis-15825	104	18	model	model	NOUN
fcis-15825	104	19	increased	increase	VERB
fcis-15825	104	20	by	by	ADP
fcis-15825	104	21	2.1	2.1	NUM
fcis-15825	104	22	and	and	CCONJ
fcis-15825	104	23	5.2	5.2	NUM
fcis-15825	104	24	percentage	percentage	NOUN
fcis-15825	104	25	points	point	NOUN
fcis-15825	104	26	respectively	respectively	ADV
fcis-15825	104	27	,	,	PUNCT
fcis-15825	104	28	which	which	PRON
fcis-15825	104	29	is	be	AUX
fcis-15825	104	30	a	a	DET
fcis-15825	104	31	targeted	target	VERB
fcis-15825	104	32	improvement	improvement	NOUN
fcis-15825	104	33	measure	measure	NOUN
fcis-15825	104	34	.	.	PUNCT
fcis-15825	105	1	the	the	DET
fcis-15825	105	2	newly	newly	ADV
fcis-15825	105	3	added	add	VERB
fcis-15825	105	4	small	small	ADJ
fcis-15825	105	5	target	target	NOUN
fcis-15825	105	6	detection	detection	NOUN
fcis-15825	105	7	layer	layer	NOUN
fcis-15825	105	8	has	have	VERB
fcis-15825	105	9	a	a	DET
fcis-15825	105	10	smaller	small	ADJ
fcis-15825	105	11	receptive	receptive	ADJ
fcis-15825	105	12	field	field	NOUN
fcis-15825	105	13	and	and	CCONJ
fcis-15825	105	14	is	be	AUX
fcis-15825	105	15	responsible	responsible	ADJ
fcis-15825	105	16	for	for	ADP
fcis-15825	105	17	detecting	detect	VERB
fcis-15825	105	18	small	small	ADJ
fcis-15825	105	19	target	target	NOUN
fcis-15825	105	20	pavement	pavement	NOUN
fcis-15825	105	21	diseases	disease	NOUN
fcis-15825	105	22	,	,	PUNCT
fcis-15825	105	23	which	which	PRON
fcis-15825	105	24	improves	improve	VERB
fcis-15825	105	25	the	the	DET
fcis-15825	105	26	detection	detection	NOUN
fcis-15825	105	27	efficiency	efficiency	NOUN
fcis-15825	105	28	and	and	CCONJ
fcis-15825	105	29	recall	recall	NOUN
fcis-15825	105	30	of	of	ADP
fcis-15825	105	31	small	small	ADJ
fcis-15825	105	32	targets	target	NOUN
fcis-15825	105	33	such	such	ADJ
fcis-15825	105	34	as	as	ADP
fcis-15825	105	35	small	small	ADJ
fcis-15825	105	36	cracks	crack	NOUN
fcis-15825	105	37	and	and	CCONJ
fcis-15825	105	38	pits	pit	NOUN
fcis-15825	105	39	.	.	PUNCT
fcis-15825	106	1	it	it	PRON
fcis-15825	106	2	can	can	AUX
fcis-15825	106	3	be	be	AUX
fcis-15825	106	4	seen	see	VERB
fcis-15825	106	5	that	that	SCONJ
fcis-15825	106	6	the	the	DET
fcis-15825	106	7	three	three	NUM
fcis-15825	106	8	model	model	NOUN
fcis-15825	106	9	improvement	improvement	NOUN
fcis-15825	106	10	approaches	approach	NOUN
fcis-15825	106	11	proposed	propose	VERB
fcis-15825	106	12	according	accord	VERB
fcis-15825	106	13	to	to	ADP
fcis-15825	106	14	the	the	DET
fcis-15825	106	15	characteristics	characteristic	NOUN
fcis-15825	106	16	of	of	ADP
fcis-15825	106	17	pavement	pavement	NOUN
fcis-15825	106	18	diseases	disease	NOUN
fcis-15825	106	19	are	be	AUX
fcis-15825	106	20	effective	effective	ADJ
fcis-15825	106	21	.	.	PUNCT
fcis-15825	107	1	3.5	3.5	NUM
fcis-15825	107	2	.	.	PUNCT
fcis-15825	107	3	comparative	comparative	ADJ
fcis-15825	107	4	experiment	experiment	NOUN
fcis-15825	107	5	in	in	ADP
fcis-15825	107	6	order	order	NOUN
fcis-15825	107	7	to	to	PART
fcis-15825	107	8	verify	verify	VERB
fcis-15825	107	9	the	the	DET
fcis-15825	107	10	detection	detection	NOUN
fcis-15825	107	11	performance	performance	NOUN
fcis-15825	107	12	of	of	ADP
fcis-15825	107	13	the	the	DET
fcis-15825	107	14	improved	improved	ADJ
fcis-15825	107	15	model	model	NOUN
fcis-15825	107	16	,	,	PUNCT
fcis-15825	107	17	the	the	DET
fcis-15825	107	18	yolov5s	yolov5s	PROPN
fcis-15825	107	19	-	-	PUNCT
fcis-15825	107	20	cbp	cbp	PROPN
fcis-15825	107	21	model	model	NOUN
fcis-15825	107	22	is	be	AUX
fcis-15825	107	23	compared	compare	VERB
fcis-15825	107	24	with	with	ADP
fcis-15825	107	25	the	the	DET
fcis-15825	107	26	mainstream	mainstream	NOUN
fcis-15825	107	27	target	target	NOUN
fcis-15825	107	28	detection	detection	NOUN
fcis-15825	107	29	algorithms	algorithm	NOUN
fcis-15825	107	30	faster	fast	ADV
fcis-15825	107	31	r	r	NOUN
fcis-15825	107	32	-	-	PUNCT
fcis-15825	107	33	cnn	cnn	PROPN
fcis-15825	107	34	,	,	PUNCT
fcis-15825	107	35	yolov3	yolov3	PROPN
fcis-15825	107	36	,	,	PUNCT
fcis-15825	107	37	yolov4	yolov4	PROPN
fcis-15825	107	38	and	and	CCONJ
fcis-15825	107	39	yolov5s	yolov5s	PROPN
fcis-15825	107	40	,	,	PUNCT
fcis-15825	107	41	and	and	CCONJ
fcis-15825	107	42	the	the	DET
fcis-15825	107	43	comparison	comparison	NOUN
fcis-15825	107	44	test	test	NOUN
fcis-15825	107	45	results	result	NOUN
fcis-15825	107	46	are	be	AUX
fcis-15825	107	47	shown	show	VERB
fcis-15825	107	48	in	in	ADP
fcis-15825	107	49	table	table	NOUN
fcis-15825	107	50	3	3	NUM
fcis-15825	107	51	.	.	PUNCT
fcis-15825	107	52	table	table	NOUN
fcis-15825	107	53	3	3	NUM
fcis-15825	107	54	.	.	PUNCT
fcis-15825	107	55	performance	performance	NOUN
fcis-15825	107	56	comparison	comparison	NOUN
fcis-15825	107	57	with	with	ADP
fcis-15825	107	58	other	other	ADJ
fcis-15825	107	59	detection	detection	NOUN
fcis-15825	107	60	models	model	NOUN
fcis-15825	107	61	model	model	VERB
fcis-15825	107	62	pr	pr	X
fcis-15825	107	63	re	re	PUNCT
fcis-15825	107	64	map@0.5	map@0.5	VERB
fcis-15825	107	65	faster	fast	ADV
fcis-15825	107	66	rcnn	rcnn	NOUN
fcis-15825	107	67	79.3	79.3	NUM
fcis-15825	107	68	%	%	NOUN
fcis-15825	107	69	73.1	73.1	NUM
fcis-15825	107	70	%	%	NOUN
fcis-15825	107	71	75.4	75.4	NUM
fcis-15825	107	72	%	%	NOUN
fcis-15825	107	73	yolov3	yolov3	NOUN
fcis-15825	107	74	76.1	76.1	NUM
fcis-15825	107	75	%	%	NOUN
fcis-15825	107	76	72.3	72.3	NUM
fcis-15825	107	77	%	%	NOUN
fcis-15825	107	78	74.7	74.7	NUM
fcis-15825	107	79	%	%	NOUN
fcis-15825	107	80	yolov4	yolov4	NOUN
fcis-15825	107	81	79.5	79.5	NUM
fcis-15825	107	82	%	%	NOUN
fcis-15825	107	83	74.0	74.0	NUM
fcis-15825	107	84	%	%	NOUN
fcis-15825	107	85	76.5	76.5	NUM
fcis-15825	107	86	%	%	NOUN
fcis-15825	107	87	yolov5s	yolov5s	NOUN
fcis-15825	107	88	82.1	82.1	NUM
fcis-15825	107	89	%	%	NOUN
fcis-15825	107	90	76.3	76.3	NUM
fcis-15825	107	91	%	%	NOUN
fcis-15825	107	92	79.8	79.8	NUM
fcis-15825	107	93	%	%	NOUN
fcis-15825	107	94	yolov5s	yolov5s	PROPN
fcis-15825	107	95	-	-	PUNCT
fcis-15825	107	96	cbp	cbp	PROPN
fcis-15825	107	97	88.2	88.2	NUM
fcis-15825	107	98	%	%	NOUN
fcis-15825	107	99	80.7	80.7	NUM
fcis-15825	107	100	%	%	NOUN
fcis-15825	107	101	85.5	85.5	NUM
fcis-15825	107	102	%	%	NOUN
fcis-15825	107	103	as	as	SCONJ
fcis-15825	107	104	can	can	AUX
fcis-15825	107	105	be	be	AUX
fcis-15825	107	106	seen	see	VERB
fcis-15825	107	107	from	from	ADP
fcis-15825	107	108	table	table	NOUN
fcis-15825	107	109	3	3	NUM
fcis-15825	107	110	,	,	PUNCT
fcis-15825	107	111	the	the	DET
fcis-15825	107	112	improved	improved	ADJ
fcis-15825	107	113	yolov5s	yolov5s	PROPN
fcis-15825	107	114	-	-	PUNCT
fcis-15825	107	115	cbp	cbp	PROPN
fcis-15825	107	116	model	model	NOUN
fcis-15825	107	117	has	have	VERB
fcis-15825	107	118	the	the	DET
fcis-15825	107	119	highest	high	ADJ
fcis-15825	107	120	detection	detection	NOUN
fcis-15825	107	121	accuracy	accuracy	NOUN
fcis-15825	107	122	(	(	PUNCT
fcis-15825	107	123	88.2	88.2	NUM
fcis-15825	107	124	%	%	NOUN
fcis-15825	107	125	)	)	PUNCT
fcis-15825	107	126	and	and	CCONJ
fcis-15825	107	127	recall	recall	NOUN
fcis-15825	107	128	rate	rate	NOUN
fcis-15825	107	129	(	(	PUNCT
fcis-15825	107	130	80.7	80.7	NUM
fcis-15825	107	131	%	%	NOUN
fcis-15825	107	132	)	)	PUNCT
fcis-15825	107	133	.	.	PUNCT
fcis-15825	108	1	compared	compare	VERB
fcis-15825	108	2	with	with	ADP
fcis-15825	108	3	the	the	DET
fcis-15825	108	4	original	original	ADJ
fcis-15825	108	5	yolov5s	yolov5s	PROPN
fcis-15825	108	6	,	,	PUNCT
fcis-15825	108	7	the	the	DET
fcis-15825	108	8	accuracy	accuracy	NOUN
fcis-15825	108	9	has	have	AUX
fcis-15825	108	10	increased	increase	VERB
fcis-15825	108	11	by	by	ADP
fcis-15825	108	12	6.1	6.1	NUM
fcis-15825	108	13	percentage	percentage	NOUN
fcis-15825	108	14	points	point	NOUN
fcis-15825	108	15	,	,	PUNCT
fcis-15825	108	16	and	and	CCONJ
fcis-15825	108	17	the	the	DET
fcis-15825	108	18	recall	recall	NOUN
fcis-15825	108	19	rate	rate	NOUN
fcis-15825	108	20	has	have	AUX
fcis-15825	108	21	increased	increase	VERB
fcis-15825	108	22	by	by	ADP
fcis-15825	108	23	4.4	4.4	NUM
fcis-15825	108	24	%	%	NOUN
fcis-15825	108	25	.	.	PUNCT
fcis-15825	109	1	the	the	DET
fcis-15825	109	2	overall	overall	ADJ
fcis-15825	109	3	performance	performance	NOUN
fcis-15825	109	4	is	be	AUX
fcis-15825	109	5	superior	superior	ADJ
fcis-15825	109	6	to	to	ADP
fcis-15825	109	7	other	other	ADJ
fcis-15825	109	8	models	model	NOUN
fcis-15825	109	9	,	,	PUNCT
fcis-15825	109	10	which	which	PRON
fcis-15825	109	11	shows	show	VERB
fcis-15825	109	12	that	that	SCONJ
fcis-15825	109	13	the	the	DET
fcis-15825	109	14	improved	improved	ADJ
fcis-15825	109	15	model	model	NOUN
fcis-15825	109	16	method	method	NOUN
fcis-15825	109	17	used	use	VERB
fcis-15825	109	18	in	in	ADP
fcis-15825	109	19	this	this	DET
fcis-15825	109	20	paper	paper	NOUN
fcis-15825	109	21	is	be	AUX
fcis-15825	109	22	a	a	DET
fcis-15825	109	23	practical	practical	ADJ
fcis-15825	109	24	and	and	CCONJ
fcis-15825	109	25	effective	effective	ADJ
fcis-15825	109	26	measure	measure	NOUN
fcis-15825	109	27	to	to	PART
fcis-15825	109	28	improve	improve	VERB
fcis-15825	109	29	the	the	DET
fcis-15825	109	30	model	model	NOUN
fcis-15825	109	31	performance	performance	NOUN
fcis-15825	109	32	.	.	PUNCT
fcis-15825	110	1	the	the	DET
fcis-15825	110	2	detection	detection	NOUN
fcis-15825	110	3	117	117	NUM
fcis-15825	110	4	results	result	NOUN
fcis-15825	110	5	of	of	ADP
fcis-15825	110	6	each	each	DET
fcis-15825	110	7	model	model	NOUN
fcis-15825	110	8	are	be	AUX
fcis-15825	110	9	shown	show	VERB
fcis-15825	110	10	in	in	ADP
fcis-15825	110	11	figure	figure	NOUN
fcis-15825	110	12	6	6	NUM
fcis-15825	110	13	.	.	PUNCT
fcis-15825	111	1	after	after	ADP
fcis-15825	111	2	careful	careful	ADJ
fcis-15825	111	3	observation	observation	NOUN
fcis-15825	111	4	,	,	PUNCT
fcis-15825	111	5	it	it	PRON
fcis-15825	111	6	can	can	AUX
fcis-15825	111	7	be	be	AUX
fcis-15825	111	8	found	find	VERB
fcis-15825	111	9	that	that	SCONJ
fcis-15825	111	10	there	there	PRON
fcis-15825	111	11	is	be	VERB
fcis-15825	111	12	little	little	ADJ
fcis-15825	111	13	difference	difference	NOUN
fcis-15825	111	14	in	in	ADP
fcis-15825	111	15	the	the	DET
fcis-15825	111	16	recall	recall	NOUN
fcis-15825	111	17	rate	rate	NOUN
fcis-15825	111	18	among	among	ADP
fcis-15825	111	19	the	the	DET
fcis-15825	111	20	models	model	NOUN
fcis-15825	111	21	except	except	SCONJ
fcis-15825	111	22	yolov3	yolov3	PROPN
fcis-15825	111	23	model	model	PROPN
fcis-15825	111	24	,	,	PUNCT
fcis-15825	111	25	and	and	CCONJ
fcis-15825	111	26	they	they	PRON
fcis-15825	111	27	all	all	PRON
fcis-15825	111	28	have	have	VERB
fcis-15825	111	29	high	high	ADJ
fcis-15825	111	30	recall	recall	NOUN
fcis-15825	111	31	rate	rate	NOUN
fcis-15825	111	32	,	,	PUNCT
fcis-15825	111	33	but	but	CCONJ
fcis-15825	111	34	the	the	DET
fcis-15825	111	35	confidence	confidence	NOUN
fcis-15825	111	36	of	of	ADP
fcis-15825	111	37	the	the	DET
fcis-15825	111	38	improved	improved	ADJ
fcis-15825	111	39	v5	v5	NOUN
fcis-15825	111	40	model	model	NOUN
fcis-15825	111	41	is	be	AUX
fcis-15825	111	42	as	as	ADV
fcis-15825	111	43	high	high	ADJ
fcis-15825	111	44	as	as	ADP
fcis-15825	111	45	94	94	NUM
fcis-15825	111	46	%	%	NOUN
fcis-15825	111	47	,	,	PUNCT
fcis-15825	111	48	which	which	PRON
fcis-15825	111	49	can	can	AUX
fcis-15825	111	50	realize	realize	VERB
fcis-15825	111	51	the	the	DET
fcis-15825	111	52	accurate	accurate	ADJ
fcis-15825	111	53	detection	detection	NOUN
fcis-15825	111	54	of	of	ADP
fcis-15825	111	55	the	the	DET
fcis-15825	111	56	disease	disease	NOUN
fcis-15825	111	57	.	.	PUNCT
fcis-15825	112	1	for	for	ADP
fcis-15825	112	2	minor	minor	ADJ
fcis-15825	112	3	diseases	disease	NOUN
fcis-15825	112	4	(	(	PUNCT
fcis-15825	112	5	potholes	pothole	NOUN
fcis-15825	112	6	)	)	PUNCT
fcis-15825	112	7	,	,	PUNCT
fcis-15825	112	8	both	both	CCONJ
fcis-15825	112	9	v3	v3	PROPN
fcis-15825	112	10	and	and	CCONJ
fcis-15825	112	11	v4	v4	NOUN
fcis-15825	112	12	versions	version	NOUN
fcis-15825	112	13	ca	can	AUX
fcis-15825	112	14	n't	not	PART
fcis-15825	112	15	identify	identify	VERB
fcis-15825	112	16	them	they	PRON
fcis-15825	112	17	,	,	PUNCT
fcis-15825	112	18	but	but	CCONJ
fcis-15825	112	19	the	the	DET
fcis-15825	112	20	improved	improved	ADJ
fcis-15825	112	21	yolov5s	yolov5s	PROPN
fcis-15825	112	22	model	model	NOUN
fcis-15825	112	23	can	can	AUX
fcis-15825	112	24	accurately	accurately	ADV
fcis-15825	112	25	identify	identify	VERB
fcis-15825	112	26	them	they	PRON
fcis-15825	112	27	,	,	PUNCT
fcis-15825	112	28	which	which	PRON
fcis-15825	112	29	shows	show	VERB
fcis-15825	112	30	that	that	SCONJ
fcis-15825	112	31	the	the	DET
fcis-15825	112	32	improved	improved	ADJ
fcis-15825	112	33	measures	measure	NOUN
fcis-15825	112	34	can	can	AUX
fcis-15825	112	35	strengthen	strengthen	VERB
fcis-15825	112	36	the	the	DET
fcis-15825	112	37	model	model	NOUN
fcis-15825	112	38	's	's	PART
fcis-15825	112	39	attention	attention	NOUN
fcis-15825	112	40	to	to	ADP
fcis-15825	112	41	minor	minor	ADJ
fcis-15825	112	42	road	road	NOUN
fcis-15825	112	43	diseases	disease	NOUN
fcis-15825	112	44	and	and	CCONJ
fcis-15825	112	45	optimize	optimize	VERB
fcis-15825	112	46	the	the	DET
fcis-15825	112	47	model	model	NOUN
fcis-15825	112	48	's	's	PART
fcis-15825	112	49	ability	ability	NOUN
fcis-15825	112	50	to	to	PART
fcis-15825	112	51	detect	detect	VERB
fcis-15825	112	52	small	small	ADJ
fcis-15825	112	53	targets	target	NOUN
fcis-15825	112	54	.	.	PUNCT
fcis-15825	113	1	judging	judge	VERB
fcis-15825	113	2	from	from	ADP
fcis-15825	113	3	the	the	DET
fcis-15825	113	4	detection	detection	NOUN
fcis-15825	113	5	effect	effect	NOUN
fcis-15825	113	6	,	,	PUNCT
fcis-15825	113	7	yolov5s	yolov5s	PROPN
fcis-15825	113	8	-	-	PUNCT
fcis-15825	113	9	cbp	cbp	PROPN
fcis-15825	113	10	model	model	NOUN
fcis-15825	113	11	not	not	PART
fcis-15825	113	12	only	only	ADV
fcis-15825	113	13	can	can	AUX
fcis-15825	113	14	accurately	accurately	ADV
fcis-15825	113	15	identify	identify	VERB
fcis-15825	113	16	road	road	NOUN
fcis-15825	113	17	diseases	disease	NOUN
fcis-15825	113	18	,	,	PUNCT
fcis-15825	113	19	but	but	CCONJ
fcis-15825	113	20	also	also	ADV
fcis-15825	113	21	has	have	VERB
fcis-15825	113	22	high	high	ADJ
fcis-15825	113	23	confidence	confidence	NOUN
fcis-15825	113	24	and	and	CCONJ
fcis-15825	113	25	better	well	ADJ
fcis-15825	113	26	small	small	ADJ
fcis-15825	113	27	target	target	NOUN
fcis-15825	113	28	detection	detection	NOUN
fcis-15825	113	29	ability	ability	NOUN
fcis-15825	113	30	,	,	PUNCT
fcis-15825	113	31	which	which	PRON
fcis-15825	113	32	can	can	AUX
fcis-15825	113	33	better	well	ADV
fcis-15825	113	34	cope	cope	VERB
fcis-15825	113	35	with	with	ADP
fcis-15825	113	36	complex	complex	ADJ
fcis-15825	113	37	background	background	NOUN
fcis-15825	113	38	problems	problem	NOUN
fcis-15825	113	39	in	in	ADP
fcis-15825	113	40	actual	actual	ADJ
fcis-15825	113	41	scenes	scene	NOUN
fcis-15825	113	42	.	.	PUNCT
fcis-15825	114	1	fig	fig	NOUN
fcis-15825	114	2	6	6	NUM
fcis-15825	114	3	.	.	PUNCT
fcis-15825	115	1	comparison	comparison	NOUN
fcis-15825	115	2	of	of	ADP
fcis-15825	115	3	test	test	NOUN
fcis-15825	115	4	results	result	NOUN
fcis-15825	115	5	4	4	NUM
fcis-15825	115	6	.	.	PUNCT
fcis-15825	115	7	conclusion	conclusion	NOUN
fcis-15825	115	8	in	in	ADP
fcis-15825	115	9	order	order	NOUN
fcis-15825	115	10	to	to	PART
fcis-15825	115	11	improve	improve	VERB
fcis-15825	115	12	the	the	DET
fcis-15825	115	13	detection	detection	NOUN
fcis-15825	115	14	ability	ability	NOUN
fcis-15825	115	15	of	of	ADP
fcis-15825	115	16	the	the	DET
fcis-15825	115	17	road	road	NOUN
fcis-15825	115	18	disease	disease	NOUN
fcis-15825	115	19	detection	detection	NOUN
fcis-15825	115	20	model	model	NOUN
fcis-15825	115	21	in	in	ADP
fcis-15825	115	22	the	the	DET
fcis-15825	115	23	actual	actual	ADJ
fcis-15825	115	24	scene	scene	NOUN
fcis-15825	115	25	,	,	PUNCT
fcis-15825	115	26	this	this	DET
fcis-15825	115	27	paper	paper	NOUN
fcis-15825	115	28	obtains	obtain	VERB
fcis-15825	115	29	a	a	DET
fcis-15825	115	30	neural	neural	ADJ
fcis-15825	115	31	network	network	NOUN
fcis-15825	115	32	model	model	NOUN
fcis-15825	115	33	with	with	ADP
fcis-15825	115	34	higher	high	ADJ
fcis-15825	115	35	detection	detection	NOUN
fcis-15825	115	36	accuracy	accuracy	NOUN
fcis-15825	115	37	from	from	ADP
fcis-15825	115	38	two	two	NUM
fcis-15825	115	39	aspects	aspect	NOUN
fcis-15825	115	40	:	:	PUNCT
fcis-15825	115	41	data	datum	NOUN
fcis-15825	115	42	set	set	VERB
fcis-15825	115	43	construction	construction	NOUN
fcis-15825	115	44	and	and	CCONJ
fcis-15825	115	45	model	model	NOUN
fcis-15825	115	46	improvement	improvement	NOUN
fcis-15825	115	47	.	.	PUNCT
fcis-15825	116	1	through	through	ADP
fcis-15825	116	2	the	the	DET
fcis-15825	116	3	analysis	analysis	NOUN
fcis-15825	116	4	of	of	ADP
fcis-15825	116	5	the	the	DET
fcis-15825	116	6	characteristics	characteristic	NOUN
fcis-15825	116	7	of	of	ADP
fcis-15825	116	8	road	road	NOUN
fcis-15825	116	9	diseases	disease	NOUN
fcis-15825	116	10	,	,	PUNCT
fcis-15825	116	11	the	the	DET
fcis-15825	116	12	open	open	ADJ
fcis-15825	116	13	source	source	NOUN
fcis-15825	116	14	road	road	NOUN
fcis-15825	116	15	disease	disease	NOUN
fcis-15825	116	16	data	datum	NOUN
fcis-15825	116	17	set	set	VERB
fcis-15825	116	18	is	be	AUX
fcis-15825	116	19	sorted	sort	VERB
fcis-15825	116	20	out	out	ADP
fcis-15825	116	21	and	and	CCONJ
fcis-15825	116	22	marked	mark	VERB
fcis-15825	116	23	twice	twice	ADV
fcis-15825	116	24	,	,	PUNCT
fcis-15825	116	25	and	and	CCONJ
fcis-15825	116	26	the	the	DET
fcis-15825	116	27	local	local	ADJ
fcis-15825	116	28	road	road	NOUN
fcis-15825	116	29	image	image	NOUN
fcis-15825	116	30	data	datum	NOUN
fcis-15825	116	31	is	be	AUX
fcis-15825	116	32	obtained	obtain	VERB
fcis-15825	116	33	by	by	ADP
fcis-15825	116	34	field	field	NOUN
fcis-15825	116	35	shooting	shooting	NOUN
fcis-15825	116	36	.	.	PUNCT
fcis-15825	117	1	at	at	ADP
fcis-15825	117	2	the	the	DET
fcis-15825	117	3	same	same	ADJ
fcis-15825	117	4	time	time	NOUN
fcis-15825	117	5	,	,	PUNCT
fcis-15825	117	6	the	the	DET
fcis-15825	117	7	data	datum	NOUN
fcis-15825	117	8	set	set	VERB
fcis-15825	117	9	is	be	AUX
fcis-15825	117	10	expanded	expand	VERB
fcis-15825	117	11	by	by	ADP
fcis-15825	117	12	data	datum	NOUN
fcis-15825	117	13	enhancement	enhancement	NOUN
fcis-15825	117	14	method	method	NOUN
fcis-15825	117	15	to	to	PART
fcis-15825	117	16	obtain	obtain	VERB
fcis-15825	117	17	high	high	ADJ
fcis-15825	117	18	-	-	PUNCT
fcis-15825	117	19	quality	quality	NOUN
fcis-15825	117	20	road	road	NOUN
fcis-15825	117	21	disease	disease	NOUN
fcis-15825	117	22	data	datum	NOUN
fcis-15825	117	23	set	set	VERB
fcis-15825	117	24	.	.	PUNCT
fcis-15825	118	1	based	base	VERB
fcis-15825	118	2	on	on	ADP
fcis-15825	118	3	yolov5s	yolov5s	PROPN
fcis-15825	118	4	model	model	NOUN
fcis-15825	118	5	,	,	PUNCT
fcis-15825	118	6	small	small	ADJ
fcis-15825	118	7	target	target	NOUN
fcis-15825	118	8	detection	detection	NOUN
fcis-15825	118	9	layer	layer	NOUN
fcis-15825	118	10	p2	p2	NOUN
fcis-15825	118	11	is	be	AUX
fcis-15825	118	12	added	add	VERB
fcis-15825	118	13	to	to	PART
fcis-15825	118	14	detect	detect	VERB
fcis-15825	118	15	smaller	small	ADJ
fcis-15825	118	16	road	road	NOUN
fcis-15825	118	17	diseases	disease	NOUN
fcis-15825	118	18	,	,	PUNCT
fcis-15825	118	19	which	which	PRON
fcis-15825	118	20	effectively	effectively	ADV
fcis-15825	118	21	improves	improve	VERB
fcis-15825	118	22	the	the	DET
fcis-15825	118	23	small	small	ADJ
fcis-15825	118	24	target	target	NOUN
fcis-15825	118	25	detection	detection	NOUN
fcis-15825	118	26	performance	performance	NOUN
fcis-15825	118	27	of	of	ADP
fcis-15825	118	28	the	the	DET
fcis-15825	118	29	model	model	NOUN
fcis-15825	118	30	.	.	PUNCT
fcis-15825	119	1	add	add	VERB
fcis-15825	119	2	cbam	cbam	NOUN
fcis-15825	119	3	attention	attention	NOUN
fcis-15825	119	4	to	to	ADP
fcis-15825	119	5	the	the	DET
fcis-15825	119	6	backbone	backbone	NOUN
fcis-15825	119	7	network	network	NOUN
fcis-15825	119	8	,	,	PUNCT
fcis-15825	119	9	optimize	optimize	VERB
fcis-15825	119	10	feature	feature	NOUN
fcis-15825	119	11	extraction	extraction	NOUN
fcis-15825	119	12	,	,	PUNCT
fcis-15825	119	13	focus	focus	VERB
fcis-15825	119	14	on	on	ADP
fcis-15825	119	15	important	important	ADJ
fcis-15825	119	16	feature	feature	NOUN
fcis-15825	119	17	information	information	NOUN
fcis-15825	119	18	,	,	PUNCT
fcis-15825	119	19	suppress	suppress	VERB
fcis-15825	119	20	unimportant	unimportant	ADJ
fcis-15825	119	21	features	feature	NOUN
fcis-15825	119	22	,	,	PUNCT
fcis-15825	119	23	and	and	CCONJ
fcis-15825	119	24	promote	promote	VERB
fcis-15825	119	25	information	information	NOUN
fcis-15825	119	26	flow	flow	NOUN
fcis-15825	119	27	in	in	ADP
fcis-15825	119	28	both	both	DET
fcis-15825	119	29	channel	channel	NOUN
fcis-15825	119	30	and	and	CCONJ
fcis-15825	119	31	space	space	NOUN
fcis-15825	119	32	dimensions	dimension	NOUN
fcis-15825	119	33	;	;	PUNCT
fcis-15825	119	34	the	the	DET
fcis-15825	119	35	bifpn	bifpn	PROPN
fcis-15825	119	36	bidirectional	bidirectional	ADJ
fcis-15825	119	37	connection	connection	NOUN
fcis-15825	119	38	feature	feature	NOUN
fcis-15825	119	39	fusion	fusion	NOUN
fcis-15825	119	40	network	network	NOUN
fcis-15825	119	41	is	be	AUX
fcis-15825	119	42	constructed	construct	VERB
fcis-15825	119	43	in	in	ADP
fcis-15825	119	44	the	the	DET
fcis-15825	119	45	neck	neck	NOUN
fcis-15825	119	46	network	network	NOUN
fcis-15825	119	47	,	,	PUNCT
fcis-15825	119	48	which	which	PRON
fcis-15825	119	49	simplifies	simplify	VERB
fcis-15825	119	50	the	the	DET
fcis-15825	119	51	bidirectional	bidirectional	ADJ
fcis-15825	119	52	fusion	fusion	NOUN
fcis-15825	119	53	path	path	NOUN
fcis-15825	119	54	,	,	PUNCT
fcis-15825	119	55	adds	add	VERB
fcis-15825	119	56	jumping	jumping	NOUN
fcis-15825	119	57	connection	connection	NOUN
fcis-15825	119	58	,	,	PUNCT
fcis-15825	119	59	retains	retain	VERB
fcis-15825	119	60	the	the	DET
fcis-15825	119	61	semantic	semantic	ADJ
fcis-15825	119	62	features	feature	NOUN
fcis-15825	119	63	of	of	ADP
fcis-15825	119	64	higher	high	ADJ
fcis-15825	119	65	layers	layer	NOUN
fcis-15825	119	66	,	,	PUNCT
fcis-15825	119	67	and	and	CCONJ
fcis-15825	119	68	effectively	effectively	ADV
fcis-15825	119	69	prevents	prevent	VERB
fcis-15825	119	70	information	information	NOUN
fcis-15825	119	71	loss	loss	NOUN
fcis-15825	119	72	caused	cause	VERB
fcis-15825	119	73	by	by	ADP
fcis-15825	119	74	up	up	ADV
fcis-15825	119	75	-	-	PUNCT
fcis-15825	119	76	sampling	sampling	NOUN
fcis-15825	119	77	.	.	PUNCT
fcis-15825	120	1	compared	compare	VERB
fcis-15825	120	2	with	with	ADP
fcis-15825	120	3	other	other	ADJ
fcis-15825	120	4	algorithms	algorithm	NOUN
fcis-15825	120	5	,	,	PUNCT
fcis-15825	120	6	yolo	yolo	ADJ
fcis-15825	120	7	-	-	PUNCT
fcis-15825	120	8	pcb	pcb	NOUN
fcis-15825	120	9	model	model	NOUN
fcis-15825	120	10	can	can	AUX
fcis-15825	120	11	better	well	ADV
fcis-15825	120	12	detect	detect	VERB
fcis-15825	120	13	small	small	ADJ
fcis-15825	120	14	target	target	NOUN
fcis-15825	120	15	road	road	NOUN
fcis-15825	120	16	diseases	disease	NOUN
fcis-15825	120	17	in	in	ADP
fcis-15825	120	18	complex	complex	ADJ
fcis-15825	120	19	background	background	NOUN
fcis-15825	120	20	and	and	CCONJ
fcis-15825	120	21	has	have	VERB
fcis-15825	120	22	stronger	strong	ADJ
fcis-15825	120	23	practical	practical	ADJ
fcis-15825	120	24	application	application	NOUN
fcis-15825	120	25	ability	ability	NOUN
fcis-15825	120	26	.	.	PUNCT
fcis-15825	121	1	at	at	ADP
fcis-15825	121	2	the	the	DET
fcis-15825	121	3	same	same	ADJ
fcis-15825	121	4	time	time	NOUN
fcis-15825	121	5	,	,	PUNCT
fcis-15825	121	6	the	the	DET
fcis-15825	121	7	actual	actual	ADJ
fcis-15825	121	8	detection	detection	NOUN
fcis-15825	121	9	results	result	NOUN
fcis-15825	121	10	also	also	ADV
fcis-15825	121	11	show	show	VERB
fcis-15825	121	12	that	that	SCONJ
fcis-15825	121	13	for	for	ADP
fcis-15825	121	14	some	some	DET
fcis-15825	121	15	uncommon	uncommon	ADJ
fcis-15825	121	16	diseases	disease	NOUN
fcis-15825	121	17	,	,	PUNCT
fcis-15825	121	18	such	such	ADJ
fcis-15825	121	19	as	as	ADP
fcis-15825	121	20	potholes	pothole	NOUN
fcis-15825	121	21	,	,	PUNCT
fcis-15825	121	22	the	the	DET
fcis-15825	121	23	detection	detection	NOUN
fcis-15825	121	24	effect	effect	NOUN
fcis-15825	121	25	of	of	ADP
fcis-15825	121	26	the	the	DET
fcis-15825	121	27	model	model	NOUN
fcis-15825	121	28	needs	need	VERB
fcis-15825	121	29	to	to	PART
fcis-15825	121	30	be	be	AUX
fcis-15825	121	31	further	far	ADV
fcis-15825	121	32	improved	improve	VERB
fcis-15825	121	33	.	.	PUNCT
fcis-15825	122	1	the	the	DET
fcis-15825	122	2	next	next	ADJ
fcis-15825	122	3	step	step	NOUN
fcis-15825	122	4	will	will	AUX
fcis-15825	122	5	mainly	mainly	ADV
fcis-15825	122	6	focus	focus	VERB
fcis-15825	122	7	on	on	ADP
fcis-15825	122	8	optimizing	optimize	VERB
fcis-15825	122	9	the	the	DET
fcis-15825	122	10	data	datum	NOUN
fcis-15825	122	11	set	set	VERB
fcis-15825	122	12	construction	construction	NOUN
fcis-15825	122	13	,	,	PUNCT
fcis-15825	122	14	increasing	increase	VERB
fcis-15825	122	15	the	the	DET
fcis-15825	122	16	number	number	NOUN
fcis-15825	122	17	of	of	ADP
fcis-15825	122	18	uncommon	uncommon	ADJ
fcis-15825	122	19	road	road	NOUN
fcis-15825	122	20	diseases	disease	NOUN
fcis-15825	122	21	through	through	ADP
fcis-15825	122	22	measures	measure	NOUN
fcis-15825	122	23	such	such	ADJ
fcis-15825	122	24	as	as	ADP
fcis-15825	122	25	gan	gan	PROPN
fcis-15825	122	26	,	,	PUNCT
fcis-15825	122	27	ensuring	ensure	VERB
fcis-15825	122	28	that	that	SCONJ
fcis-15825	122	29	the	the	DET
fcis-15825	122	30	number	number	NOUN
fcis-15825	122	31	of	of	ADP
fcis-15825	122	32	various	various	ADJ
fcis-15825	122	33	road	road	NOUN
fcis-15825	122	34	diseases	disease	NOUN
fcis-15825	122	35	in	in	ADP
fcis-15825	122	36	the	the	DET
fcis-15825	122	37	data	datum	NOUN
fcis-15825	122	38	set	set	VERB
fcis-15825	122	39	is	be	AUX
fcis-15825	122	40	basically	basically	ADV
fcis-15825	122	41	the	the	DET
fcis-15825	122	42	same	same	ADJ
fcis-15825	122	43	,	,	PUNCT
fcis-15825	122	44	reducing	reduce	VERB
fcis-15825	122	45	the	the	DET
fcis-15825	122	46	difference	difference	NOUN
fcis-15825	122	47	of	of	ADP
fcis-15825	122	48	road	road	NOUN
fcis-15825	122	49	disease	disease	NOUN
fcis-15825	122	50	categories	category	NOUN
fcis-15825	122	51	,	,	PUNCT
fcis-15825	122	52	and	and	CCONJ
fcis-15825	122	53	further	far	ADV
fcis-15825	122	54	improving	improve	VERB
fcis-15825	122	55	the	the	DET
fcis-15825	122	56	detection	detection	NOUN
fcis-15825	122	57	performance	performance	NOUN
fcis-15825	122	58	of	of	ADP
fcis-15825	122	59	the	the	DET
fcis-15825	122	60	model	model	NOUN
fcis-15825	122	61	in	in	ADP
fcis-15825	122	62	practical	practical	ADJ
fcis-15825	122	63	application	application	NOUN
fcis-15825	122	64	.	.	PUNCT
fcis-15825	123	1	references	reference	NOUN
fcis-15825	123	2	[	[	X
fcis-15825	123	3	1	1	NUM
fcis-15825	123	4	]	]	X
fcis-15825	123	5	ning	ning	NOUN
fcis-15825	123	6	tingting	tingting	PROPN
fcis-15825	123	7	.	.	PUNCT
fcis-15825	124	1	research	research	NOUN
fcis-15825	124	2	on	on	ADP
fcis-15825	124	3	pavement	pavement	NOUN
fcis-15825	124	4	crack	crack	NOUN
fcis-15825	124	5	detection	detection	NOUN
fcis-15825	124	6	algorithm	algorithm	NOUN
fcis-15825	124	7	based	base	VERB
fcis-15825	124	8	on	on	ADP
fcis-15825	124	9	digital	digital	ADJ
fcis-15825	124	10	image	image	NOUN
fcis-15825	124	11	processing	processing	NOUN
fcis-15825	124	12	[	[	X
fcis-15825	124	13	d	d	X
fcis-15825	124	14	]	]	X
fcis-15825	124	15	.	.	PUNCT
fcis-15825	125	1	liaoning	liaoning	PROPN
fcis-15825	125	2	normal	normal	ADJ
fcis-15825	125	3	university	university	NOUN
fcis-15825	125	4	,	,	PUNCT
fcis-15825	125	5	2023	2023	NUM
fcis-15825	125	6	.	.	PUNCT
fcis-15825	126	1	118	118	NUM
fcis-15825	127	1	[	[	X
fcis-15825	127	2	2	2	NUM
fcis-15825	127	3	]	]	PUNCT
fcis-15825	127	4	chu	chu	NOUN
fcis-15825	127	5	xiumin	xiumin	PROPN
fcis-15825	127	6	,	,	PUNCT
fcis-15825	127	7	yan	yan	PROPN
fcis-15825	127	8	xinping	xinping	PROPN
fcis-15825	127	9	,	,	PUNCT
fcis-15825	127	10	chen	chen	PROPN
fcis-15825	127	11	xianqiao	xianqiao	PROPN
fcis-15825	127	12	.	.	PUNCT
fcis-15825	128	1	research	research	NOUN
fcis-15825	128	2	on	on	ADP
fcis-15825	128	3	binarization	binarization	NOUN
fcis-15825	128	4	method	method	NOUN
fcis-15825	128	5	of	of	ADP
fcis-15825	128	6	pavement	pavement	NOUN
fcis-15825	128	7	damage	damage	NOUN
fcis-15825	128	8	image	image	NOUN
fcis-15825	129	1	[	[	X
fcis-15825	129	2	j	j	X
fcis-15825	129	3	]	]	X
fcis-15825	129	4	.	.	PUNCT
fcis-15825	130	1	computer	computer	NOUN
fcis-15825	130	2	engineering	engineering	NOUN
fcis-15825	130	3	and	and	CCONJ
fcis-15825	130	4	application	application	NOUN
fcis-15825	130	5	,	,	PUNCT
fcis-15825	130	6	2008(28):161	2008(28):161	NOUN
fcis-15825	130	7	-	-	SYM
fcis-15825	130	8	165	165	NUM
fcis-15825	130	9	.	.	PUNCT
fcis-15825	131	1	[	[	X
fcis-15825	131	2	3	3	X
fcis-15825	131	3	]	]	X
fcis-15825	131	4	jiang	jiang	PROPN
fcis-15825	131	5	wei	wei	PROPN
fcis-15825	131	6	.	.	PUNCT
fcis-15825	131	7	research	research	NOUN
fcis-15825	131	8	and	and	CCONJ
fcis-15825	131	9	implementation	implementation	NOUN
fcis-15825	131	10	of	of	ADP
fcis-15825	131	11	adaptive	adaptive	ADJ
fcis-15825	131	12	threshold	threshold	NOUN
fcis-15825	131	13	segmentation	segmentation	NOUN
fcis-15825	131	14	algorithm	algorithm	NOUN
fcis-15825	131	15	for	for	ADP
fcis-15825	131	16	pavement	pavement	NOUN
fcis-15825	131	17	crack	crack	NOUN
fcis-15825	131	18	diseases	disease	NOUN
fcis-15825	131	19	[	[	X
fcis-15825	131	20	d	d	X
fcis-15825	131	21	]	]	X
fcis-15825	131	22	.	.	PUNCT
fcis-15825	132	1	nanjing	nanjing	PROPN
fcis-15825	132	2	university	university	PROPN
fcis-15825	132	3	of	of	ADP
fcis-15825	132	4	information	information	NOUN
fcis-15825	132	5	science	science	NOUN
fcis-15825	132	6	and	and	CCONJ
fcis-15825	132	7	technology	technology	NOUN
fcis-15825	132	8	,	,	PUNCT
fcis-15825	132	9	2020	2020	NUM
fcis-15825	132	10	.	.	PUNCT
fcis-15825	133	1	[	[	X
fcis-15825	133	2	4	4	X
fcis-15825	133	3	]	]	PUNCT
fcis-15825	133	4	wang	wang	PROPN
fcis-15825	133	5	g	g	PROPN
fcis-15825	133	6	,	,	PUNCT
fcis-15825	133	7	peter	peter	PROPN
fcis-15825	133	8	w	w	PROPN
fcis-15825	133	9	t	t	PROPN
fcis-15825	133	10	,	,	PUNCT
fcis-15825	133	11	yuan	yuan	NOUN
fcis-15825	133	12	m.	m.	NOUN
fcis-15825	133	13	automatic	automatic	ADJ
fcis-15825	133	14	internal	internal	ADJ
fcis-15825	133	15	crack	crack	NOUN
fcis-15825	133	16	detection	detection	NOUN
fcis-15825	133	17	from	from	ADP
fcis-15825	133	18	a	a	DET
fcis-15825	133	19	sequence	sequence	NOUN
fcis-15825	133	20	of	of	ADP
fcis-15825	133	21	infrared	infrared	ADJ
fcis-15825	133	22	images	image	NOUN
fcis-15825	133	23	with	with	ADP
fcis-15825	133	24	a	a	DET
fcis-15825	133	25	triplethreshold	triplethreshold	ADJ
fcis-15825	133	26	canny	canny	ADJ
fcis-15825	133	27	edge	edge	NOUN
fcis-15825	133	28	detector[j	detector[j	NOUN
fcis-15825	133	29	]	]	PUNCT
fcis-15825	133	30	.	.	PUNCT
fcis-15825	134	1	measurement	measurement	NOUN
fcis-15825	134	2	science	science	NOUN
fcis-15825	134	3	and	and	CCONJ
fcis-15825	134	4	technology	technology	NOUN
fcis-15825	134	5	,	,	PUNCT
fcis-15825	134	6	2018	2018	NUM
fcis-15825	134	7	,	,	PUNCT
fcis-15825	134	8	29(2	29(2	NUM
fcis-15825	134	9	):	):	PUNCT
fcis-15825	134	10	025403	025403	NUM
fcis-15825	134	11	.	.	PUNCT
fcis-15825	135	1	[	[	X
fcis-15825	135	2	5	5	NUM
fcis-15825	135	3	]	]	X
fcis-15825	135	4	rolling	rolling	NOUN
fcis-15825	135	5	,	,	PUNCT
fcis-15825	135	6	meng	meng	PROPN
fcis-15825	135	7	hua	hua	PROPN
fcis-15825	135	8	,	,	PUNCT
fcis-15825	135	9	liu	liu	PROPN
fcis-15825	135	10	tangzhi	tangzhi	PROPN
fcis-15825	135	11	.	.	PUNCT
fcis-15825	136	1	research	research	NOUN
fcis-15825	136	2	on	on	ADP
fcis-15825	136	3	edge	edge	NOUN
fcis-15825	136	4	detection	detection	NOUN
fcis-15825	136	5	algorithm	algorithm	NOUN
fcis-15825	136	6	of	of	ADP
fcis-15825	136	7	asphalt	asphalt	NOUN
fcis-15825	136	8	pavement	pavement	NOUN
fcis-15825	136	9	crack	crack	NOUN
fcis-15825	136	10	image	image	NOUN
fcis-15825	137	1	[	[	X
fcis-15825	137	2	j	j	X
fcis-15825	137	3	]	]	X
fcis-15825	137	4	.	.	PUNCT
fcis-15825	138	1	northern	northern	ADJ
fcis-15825	138	2	transportation	transportation	NOUN
fcis-15825	138	3	,	,	PUNCT
fcis-15825	138	4	2015(03):56	2015(03):56	NUM
fcis-15825	138	5	-	-	SYM
fcis-15825	138	6	59	59	NUM
fcis-15825	138	7	.	.	PUNCT
fcis-15825	139	1	[	[	X
fcis-15825	139	2	6	6	NUM
fcis-15825	139	3	]	]	X
fcis-15825	139	4	wang	wang	PROPN
fcis-15825	139	5	wei	wei	PROPN
fcis-15825	139	6	,	,	PUNCT
fcis-15825	139	7	meng	meng	PROPN
fcis-15825	139	8	lina	lina	PROPN
fcis-15825	139	9	,	,	PUNCT
fcis-15825	139	10	yan	yan	PROPN
fcis-15825	139	11	zhaojin	zhaojin	PROPN
fcis-15825	139	12	.	.	PUNCT
fcis-15825	140	1	pavement	pavement	NOUN
fcis-15825	140	2	damage	damage	NOUN
fcis-15825	140	3	detection	detection	NOUN
fcis-15825	140	4	based	base	VERB
fcis-15825	140	5	on	on	ADP
fcis-15825	140	6	regional	regional	ADJ
fcis-15825	140	7	growth	growth	NOUN
fcis-15825	140	8	and	and	CCONJ
fcis-15825	140	9	outermost	outermost	ADJ
fcis-15825	140	10	edge	edge	NOUN
fcis-15825	140	11	tracking	track	VERB
fcis-15825	140	12	[	[	X
fcis-15825	140	13	j	j	X
fcis-15825	140	14	]	]	X
fcis-15825	140	15	.	.	PUNCT
fcis-15825	141	1	journal	journal	PROPN
fcis-15825	141	2	of	of	ADP
fcis-15825	141	3	civil	civil	ADJ
fcis-15825	141	4	aviation	aviation	NOUN
fcis-15825	141	5	university	university	PROPN
fcis-15825	141	6	of	of	ADP
fcis-15825	141	7	china	china	PROPN
fcis-15825	141	8	,	,	PUNCT
fcis-15825	141	9	2016	2016	NUM
fcis-15825	141	10	,	,	PUNCT
fcis-15825	141	11	34(06):51	34(06):51	NUM
fcis-15825	141	12	-	-	SYM
fcis-15825	141	13	54	54	NUM
fcis-15825	141	14	.	.	PUNCT
fcis-15825	142	1	[	[	X
fcis-15825	142	2	7	7	X
fcis-15825	142	3	]	]	X
fcis-15825	142	4	xiao	xiao	PROPN
fcis-15825	142	5	xiaoming	xiaoming	PROPN
fcis-15825	142	6	,	,	PUNCT
fcis-15825	142	7	ma	ma	PROPN
fcis-15825	142	8	zhi	zhi	PROPN
fcis-15825	142	9	,	,	PUNCT
fcis-15825	142	10	cai	cai	PROPN
fcis-15825	142	11	zixing	zixing	PROPN
fcis-15825	142	12	et	et	PROPN
fcis-15825	142	13	al	al	PROPN
fcis-15825	142	14	.	.	PUNCT
fcis-15825	143	1	an	an	DET
fcis-15825	143	2	adaptive	adaptive	ADJ
fcis-15825	143	3	region	region	NOUN
fcis-15825	143	4	growing	grow	VERB
fcis-15825	143	5	algorithm	algorithm	NOUN
fcis-15825	143	6	for	for	ADP
fcis-15825	143	7	road	road	NOUN
fcis-15825	143	8	segmentation	segmentation	NOUN
fcis-15825	144	1	[	[	X
fcis-15825	144	2	j	j	X
fcis-15825	144	3	]	]	X
fcis-15825	144	4	.	.	PUNCT
fcis-15825	145	1	control	control	PROPN
fcis-15825	145	2	engineering	engineering	PROPN
fcis-15825	145	3	,	,	PUNCT
fcis-15825	145	4	2011,18(03):364	2011,18(03):364	NUM
fcis-15825	145	5	-	-	SYM
fcis-15825	145	6	368	368	NUM
fcis-15825	145	7	.	.	PUNCT
fcis-15825	146	1	[	[	X
fcis-15825	146	2	8	8	X
fcis-15825	146	3	]	]	X
fcis-15825	146	4	chen	chen	PROPN
fcis-15825	146	5	yimin	yimin	PROPN
fcis-15825	146	6	,	,	PUNCT
fcis-15825	146	7	li	li	PROPN
fcis-15825	146	8	wanyi	wanyi	PROPN
fcis-15825	146	9	,	,	PUNCT
fcis-15825	146	10	weng	weng	PROPN
fcis-15825	146	11	hanrui	hanrui	PROPN
fcis-15825	146	12	,	,	PUNCT
fcis-15825	146	13	et	et	PROPN
fcis-15825	146	14	al	al	PROPN
fcis-15825	146	15	.	.	PROPN
fcis-15825	146	16	overview	overview	NOUN
fcis-15825	146	17	of	of	ADP
fcis-15825	146	18	twostage	twostage	NOUN
fcis-15825	146	19	target	target	NOUN
fcis-15825	146	20	detection	detection	NOUN
fcis-15825	146	21	algorithms	algorithm	NOUN
fcis-15825	146	22	based	base	VERB
fcis-15825	146	23	on	on	ADP
fcis-15825	146	24	deep	deep	ADJ
fcis-15825	146	25	learning	learning	NOUN
fcis-15825	146	26	[	[	X
fcis-15825	146	27	j	j	X
fcis-15825	146	28	]	]	X
fcis-15825	146	29	.	.	PUNCT
fcis-15825	147	1	information	information	NOUN
fcis-15825	147	2	and	and	CCONJ
fcis-15825	147	3	computer	computer	NOUN
fcis-15825	147	4	(	(	PUNCT
fcis-15825	147	5	theoretical	theoretical	ADJ
fcis-15825	147	6	edition	edition	NOUN
fcis-15825	147	7	)	)	PUNCT
fcis-15825	147	8	,	,	PUNCT
fcis-15825	147	9	2023,35	2023,35	NUM
fcis-15825	147	10	(	(	PUNCT
fcis-15825	147	11	14	14	NUM
fcis-15825	147	12	):	):	PUNCT
fcis-15825	147	13	112	112	NUM
fcis-15825	147	14	-	-	SYM
fcis-15825	147	15	114	114	NUM
fcis-15825	147	16	.	.	PUNCT
fcis-15825	148	1	[	[	X
fcis-15825	148	2	9	9	NUM
fcis-15825	148	3	]	]	X
fcis-15825	148	4	zhu	zhu	PROPN
fcis-15825	148	5	hao	hao	PROPN
fcis-15825	148	6	,	,	PUNCT
fcis-15825	148	7	zhou	zhou	PROPN
fcis-15825	148	8	shunyong	shunyong	PROPN
fcis-15825	148	9	,	,	PUNCT
fcis-15825	148	10	liu	liu	PROPN
fcis-15825	148	11	xue	xue	PROPN
fcis-15825	148	12	,	,	PUNCT
fcis-15825	148	13	et	et	PROPN
fcis-15825	148	14	al	al	PROPN
fcis-15825	148	15	.	.	PROPN
fcis-15825	148	16	overview	overview	NOUN
fcis-15825	148	17	of	of	ADP
fcis-15825	148	18	singlestage	singlestage	NOUN
fcis-15825	148	19	target	target	NOUN
fcis-15825	148	20	detection	detection	NOUN
fcis-15825	148	21	algorithm	algorithm	NOUN
fcis-15825	148	22	based	base	VERB
fcis-15825	148	23	on	on	ADP
fcis-15825	148	24	deep	deep	ADJ
fcis-15825	148	25	learning	learning	NOUN
fcis-15825	148	26	[	[	X
fcis-15825	148	27	j	j	X
fcis-15825	148	28	]	]	X
fcis-15825	148	29	.	.	PUNCT
fcis-15825	149	1	industrial	industrial	PROPN
fcis-15825	149	2	control	control	PROPN
fcis-15825	149	3	computer	computer	NOUN
fcis-15825	149	4	,	,	PUNCT
fcis-15825	149	5	2023,36(04):101	2023,36(04):101	PROPN
fcis-15825	149	6	-	-	SYM
fcis-15825	149	7	103	103	NUM
fcis-15825	149	8	.	.	PUNCT
fcis-15825	150	1	[	[	X
fcis-15825	150	2	10	10	NUM
fcis-15825	150	3	]	]	X
fcis-15825	150	4	kortmann	kortmann	PROPN
fcis-15825	150	5	f	f	PROPN
fcis-15825	150	6	,	,	PUNCT
fcis-15825	150	7	talits	talit	VERB
fcis-15825	150	8	k	k	PROPN
fcis-15825	150	9	,	,	PUNCT
fcis-15825	150	10	fassmeyer	fassmeyer	NOUN
fcis-15825	150	11	p	p	NOUN
fcis-15825	150	12	,	,	PUNCT
fcis-15825	150	13	et	et	PROPN
fcis-15825	150	14	al	al	PROPN
fcis-15825	150	15	.	.	PUNCT
fcis-15825	151	1	detecting	detect	VERB
fcis-15825	151	2	various	various	ADJ
fcis-15825	151	3	road	road	NOUN
fcis-15825	151	4	damage	damage	NOUN
fcis-15825	151	5	types	type	NOUN
fcis-15825	151	6	in	in	ADP
fcis-15825	151	7	global	global	ADJ
fcis-15825	151	8	countries	country	NOUN
fcis-15825	151	9	utilizing	utilize	VERB
fcis-15825	151	10	faster	fast	ADV
fcis-15825	151	11	rcnn[c]//2020	rcnn[c]//2020	ADP
fcis-15825	151	12	ieee	ieee	PROPN
fcis-15825	151	13	international	international	PROPN
fcis-15825	151	14	conference	conference	NOUN
fcis-15825	151	15	on	on	ADP
fcis-15825	151	16	big	big	ADJ
fcis-15825	151	17	data	datum	NOUN
fcis-15825	151	18	(	(	PUNCT
fcis-15825	151	19	big	big	ADJ
fcis-15825	151	20	data	datum	NOUN
fcis-15825	151	21	)	)	PUNCT
fcis-15825	151	22	.	.	PUNCT
fcis-15825	152	1	ieee	ieee	PROPN
fcis-15825	152	2	,	,	PUNCT
fcis-15825	152	3	2020	2020	NUM
fcis-15825	152	4	:	:	PUNCT
fcis-15825	152	5	5563	5563	NUM
fcis-15825	152	6	-	-	SYM
fcis-15825	152	7	5571	5571	NUM
fcis-15825	152	8	.	.	PUNCT
fcis-15825	153	1	[	[	X
fcis-15825	153	2	11	11	NUM
fcis-15825	153	3	]	]	X
fcis-15825	153	4	nie	nie	PROPN
fcis-15825	153	5	m	m	PROPN
fcis-15825	153	6	,	,	PUNCT
fcis-15825	153	7	wang	wang	PROPN
fcis-15825	153	8	c.	c.	PROPN
fcis-15825	153	9	pavement	pavement	NOUN
fcis-15825	153	10	crack	crack	NOUN
fcis-15825	153	11	detection	detection	NOUN
fcis-15825	153	12	based	base	VERB
fcis-15825	153	13	on	on	ADP
fcis-15825	153	14	yolo	yolo	PROPN
fcis-15825	153	15	v3[c]//2019	v3[c]//2019	PROPN
fcis-15825	153	16	2nd	2nd	ADJ
fcis-15825	153	17	international	international	ADJ
fcis-15825	153	18	conference	conference	NOUN
fcis-15825	153	19	on	on	ADP
fcis-15825	153	20	safety	safety	NOUN
fcis-15825	153	21	produce	produce	NOUN
fcis-15825	153	22	informatization	informatization	NOUN
fcis-15825	153	23	(	(	PUNCT
fcis-15825	153	24	iicspi	iicspi	NOUN
fcis-15825	153	25	)	)	PUNCT
fcis-15825	153	26	.	.	PUNCT
fcis-15825	154	1	ieee	ieee	NOUN
fcis-15825	154	2	,	,	PUNCT
fcis-15825	154	3	2019	2019	NUM
fcis-15825	154	4	:	:	PUNCT
fcis-15825	154	5	327	327	NUM
fcis-15825	154	6	-	-	SYM
fcis-15825	154	7	330	330	NUM
fcis-15825	154	8	.	.	PUNCT
fcis-15825	155	1	[	[	X
fcis-15825	155	2	12	12	NUM
fcis-15825	155	3	]	]	PUNCT
fcis-15825	155	4	yu	yu	PROPN
fcis-15825	155	5	lu	lu	PROPN
fcis-15825	155	6	,	,	PUNCT
fcis-15825	155	7	dai	dai	PROPN
fcis-15825	155	8	tianjie	tianjie	PROPN
fcis-15825	155	9	,	,	PUNCT
fcis-15825	155	10	yu	yu	PROPN
fcis-15825	155	11	lihua	lihua	PROPN
fcis-15825	155	12	.	.	PUNCT
fcis-15825	156	1	intelligent	intelligent	ADJ
fcis-15825	156	2	detection	detection	NOUN
fcis-15825	156	3	of	of	ADP
fcis-15825	156	4	road	road	NOUN
fcis-15825	156	5	diseases	disease	NOUN
fcis-15825	156	6	based	base	VERB
fcis-15825	156	7	on	on	ADP
fcis-15825	156	8	improved	improved	ADJ
fcis-15825	156	9	yolov5	yolov5	NOUN
fcis-15825	157	1	[	[	X
fcis-15825	157	2	j	j	X
fcis-15825	157	3	]	]	X
fcis-15825	157	4	.	.	PUNCT
fcis-15825	158	1	journal	journal	PROPN
fcis-15825	158	2	of	of	ADP
fcis-15825	158	3	fujian	fujian	PROPN
fcis-15825	158	4	university	university	PROPN
fcis-15825	158	5	of	of	ADP
fcis-15825	158	6	technology	technology	NOUN
fcis-15825	158	7	,	,	PUNCT
fcis-15825	158	8	2023,21(04):332	2023,21(04):332	NOUN
fcis-15825	158	9	-	-	SYM
fcis-15825	158	10	337	337	NUM
fcis-15825	158	11	.	.	PUNCT
fcis-15825	159	1	[	[	X
fcis-15825	159	2	13	13	NUM
fcis-15825	159	3	]	]	X
fcis-15825	159	4	wang	wang	PROPN
fcis-15825	159	5	linyi	linyi	PROPN
fcis-15825	159	6	,	,	PUNCT
fcis-15825	159	7	bai	bai	PROPN
fcis-15825	159	8	jing	jing	PROPN
fcis-15825	159	9	,	,	PUNCT
fcis-15825	159	10	li	li	PROPN
fcis-15825	159	11	wenjing	wenje	VERB
fcis-15825	159	12	et	et	PROPN
fcis-15825	159	13	al	al	PROPN
fcis-15825	159	14	.	.	PROPN
fcis-15825	159	15	research	research	NOUN
fcis-15825	159	16	progress	progress	NOUN
fcis-15825	159	17	of	of	ADP
fcis-15825	159	18	yolo	yolo	ADJ
fcis-15825	159	19	series	series	NOUN
fcis-15825	159	20	target	target	NOUN
fcis-15825	159	21	detection	detection	NOUN
fcis-15825	159	22	algorithms	algorithm	NOUN
fcis-15825	160	1	[	[	X
fcis-15825	160	2	j	j	X
fcis-15825	160	3	]	]	X
fcis-15825	160	4	.	.	PUNCT
fcis-15825	161	1	computer	computer	NOUN
fcis-15825	161	2	engineering	engineering	NOUN
fcis-15825	161	3	and	and	CCONJ
fcis-15825	161	4	application	application	NOUN
fcis-15825	161	5	,	,	PUNCT
fcis-15825	161	6	2023,59(14):15	2023,59(14):15	NUM
fcis-15825	161	7	-	-	SYM
fcis-15825	161	8	29	29	NUM
fcis-15825	161	9	.	.	PUNCT
fcis-15825	162	1	[	[	X
fcis-15825	162	2	14	14	NUM
fcis-15825	162	3	]	]	X
fcis-15825	162	4	qi	qi	PROPN
fcis-15825	162	5	xuanhao	xuanhao	PROPN
fcis-15825	162	6	,	,	PUNCT
fcis-15825	162	7	zhi	zhi	PROPN
fcis-15825	162	8	min	min	PROPN
fcis-15825	162	9	.	.	PROPN
fcis-15825	163	1	a	a	DET
fcis-15825	163	2	summary	summary	NOUN
fcis-15825	163	3	of	of	ADP
fcis-15825	163	4	attention	attention	NOUN
fcis-15825	163	5	mechanism	mechanism	NOUN
fcis-15825	163	6	in	in	ADP
fcis-15825	163	7	image	image	NOUN
fcis-15825	163	8	processing	processing	NOUN
fcis-15825	163	9	[	[	X
fcis-15825	163	10	j	j	X
fcis-15825	163	11	/	/	SYM
fcis-15825	163	12	ol	ol	PROPN
fcis-15825	163	13	]	]	PUNCT
fcis-15825	163	14	.	.	PUNCT
fcis-15825	164	1	computer	computer	NOUN
fcis-15825	164	2	science	science	NOUN
fcis-15825	164	3	and	and	CCONJ
fcis-15825	164	4	exploration	exploration	NOUN
fcis-15825	164	5	:	:	PUNCT
fcis-15825	164	6	120	120	NUM
fcis-15825	165	1	[	[	PUNCT
fcis-15825	165	2	2023	2023	NUM
fcis-15825	165	3	-	-	SYM
fcis-15825	165	4	10	10	NUM
fcis-15825	165	5	-	-	NUM
fcis-15825	165	6	16	16	NUM
fcis-15825	165	7	]	]	PUNCT
fcis-15825	165	8	.	.	PUNCT
fcis-15825	166	1	http://kns	http://kns	PROPN
fcis-15825	166	2	.	.	PROPN
fcis-15825	166	3	cnki.net/	cnki.net/	PROPN
fcis-15825	166	4	kcms	kcms	PROPN
fcis-15825	166	5	/	/	SYM
fcis-15825	166	6	detail/	detail/	NUM
fcis-15825	166	7	11	11	NUM
fcis-15825	166	8	.	.	PUNCT
fcis-15825	166	9	5602	5602	NUM
fcis-15825	166	10	.	.	PUNCT
fcis-15825	167	1	tp.20629.10606060606	tp.20629.10606060606	VERB
fcis-15825	167	2	[	[	X
fcis-15825	167	3	15	15	NUM
fcis-15825	167	4	]	]	X
fcis-15825	167	5	tan	tan	PROPN
fcis-15825	167	6	m	m	PROPN
fcis-15825	167	7	,	,	PUNCT
fcis-15825	167	8	pang	pang	NOUN
fcis-15825	167	9	r	r	NOUN
fcis-15825	167	10	,	,	PUNCT
fcis-15825	167	11	le	le	X
fcis-15825	167	12	q	q	PROPN
fcis-15825	167	13	v.	v.	ADP
fcis-15825	167	14	efficientdet	efficientdet	PROPN
fcis-15825	167	15	:	:	PUNCT
fcis-15825	167	16	scalable	scalable	ADJ
fcis-15825	167	17	and	and	CCONJ
fcis-15825	167	18	efficient	efficient	ADJ
fcis-15825	167	19	object	object	NOUN
fcis-15825	167	20	detection[c]//proceedings	detection[c]//proceeding	NOUN
fcis-15825	167	21	of	of	ADP
fcis-15825	167	22	the	the	DET
fcis-15825	167	23	ieee	ieee	NOUN
fcis-15825	167	24	/	/	SYM
fcis-15825	167	25	cvf	cvf	NOUN
fcis-15825	167	26	conference	conference	NOUN
fcis-15825	167	27	on	on	ADP
fcis-15825	167	28	computer	computer	NOUN
fcis-15825	167	29	vision	vision	NOUN
fcis-15825	167	30	and	and	CCONJ
fcis-15825	167	31	pattern	pattern	NOUN
fcis-15825	167	32	recognition	recognition	NOUN
fcis-15825	167	33	.	.	PUNCT
fcis-15825	168	1	2020	2020	NUM
fcis-15825	168	2	:	:	PUNCT
fcis-15825	168	3	1078110790	1078110790	NUM
fcis-15825	168	4	.	.	PUNCT