id	sid	tid	token	lemma	pos
fcis-6914	1	1	frontiers	frontier	NOUN
fcis-6914	1	2	in	in	ADP
fcis-6914	1	3	computing	computing	NOUN
fcis-6914	1	4	and	and	CCONJ
fcis-6914	1	5	intelligent	intelligent	ADJ
fcis-6914	1	6	systems	system	NOUN
fcis-6914	1	7	issn	issn	VERB
fcis-6914	1	8	:	:	PUNCT
fcis-6914	1	9	2832	2832	NUM
fcis-6914	1	10	-	-	SYM
fcis-6914	1	11	6024	6024	NUM
fcis-6914	1	12	|	|	NOUN
fcis-6914	1	13	vol	vol	NOUN
fcis-6914	1	14	.	.	PROPN
fcis-6914	2	1	3	3	NUM
fcis-6914	2	2	,	,	PUNCT
fcis-6914	2	3	no	no	INTJ
fcis-6914	2	4	.	.	NOUN
fcis-6914	2	5	2	2	NUM
fcis-6914	2	6	,	,	PUNCT
fcis-6914	2	7	2023	2023	NUM
fcis-6914	2	8	25	25	NUM
fcis-6914	2	9	a	a	DET
fcis-6914	2	10	modified	modified	ADJ
fcis-6914	2	11	yolov5	yolov5	NOUN
fcis-6914	2	12	helmet	helmet	NOUN
fcis-6914	2	13	detection	detection	NOUN
fcis-6914	2	14	algorithm	algorithm	NOUN
fcis-6914	2	15	based	base	VERB
fcis-6914	2	16	on	on	ADP
fcis-6914	2	17	swin	swin	PROPN
fcis-6914	2	18	transformer	transformer	PROPN
fcis-6914	2	19	zhongling	zhongling	PROPN
fcis-6914	2	20	liu	liu	PROPN
fcis-6914	2	21	,	,	PUNCT
fcis-6914	2	22	yi	yi	PROPN
fcis-6914	2	23	luo	luo	PROPN
fcis-6914	2	24	,	,	PUNCT
fcis-6914	2	25	chanyi	chanyi	PROPN
fcis-6914	2	26	liu	liu	PROPN
fcis-6914	2	27	*	*	PROPN
fcis-6914	2	28	sichuan	sichuan	PROPN
fcis-6914	2	29	university	university	PROPN
fcis-6914	2	30	of	of	ADP
fcis-6914	2	31	science	science	PROPN
fcis-6914	2	32	&	&	CCONJ
fcis-6914	2	33	engineering	engineering	PROPN
fcis-6914	2	34	,	,	PUNCT
fcis-6914	2	35	yibin	yibin	PROPN
fcis-6914	2	36	633000	633000	NUM
fcis-6914	2	37	,	,	PUNCT
fcis-6914	2	38	china	china	PROPN
fcis-6914	2	39	*	*	PUNCT
fcis-6914	2	40	corresponding	correspond	VERB
fcis-6914	2	41	author	author	NOUN
fcis-6914	2	42	:	:	PUNCT
fcis-6914	2	43	chanyi	chanyi	PROPN
fcis-6914	2	44	liu	liu	PROPN
fcis-6914	2	45	(	(	PUNCT
fcis-6914	2	46	email	email	NOUN
fcis-6914	2	47	:	:	PUNCT
fcis-6914	2	48	1084400802@qq.com	1084400802@qq.com	NUM
fcis-6914	2	49	)	)	PUNCT
fcis-6914	2	50	abstract	abstract	NOUN
fcis-6914	2	51	:	:	PUNCT
fcis-6914	2	52	for	for	ADP
fcis-6914	2	53	the	the	DET
fcis-6914	2	54	current	current	ADJ
fcis-6914	2	55	stage	stage	NOUN
fcis-6914	2	56	of	of	ADP
fcis-6914	2	57	helmet	helmet	NOUN
fcis-6914	2	58	detection	detection	NOUN
fcis-6914	2	59	in	in	ADP
fcis-6914	2	60	complex	complex	ADJ
fcis-6914	2	61	environments	environment	NOUN
fcis-6914	2	62	with	with	ADP
fcis-6914	2	63	low	low	ADJ
fcis-6914	2	64	accuracy	accuracy	NOUN
fcis-6914	2	65	,	,	PUNCT
fcis-6914	2	66	missed	miss	VERB
fcis-6914	2	67	detection	detection	NOUN
fcis-6914	2	68	and	and	CCONJ
fcis-6914	2	69	not	not	PART
fcis-6914	2	70	easy	easy	ADJ
fcis-6914	2	71	to	to	PART
fcis-6914	2	72	manage	manage	VERB
fcis-6914	2	73	wearing	wear	VERB
fcis-6914	2	74	,	,	PUNCT
fcis-6914	2	75	this	this	DET
fcis-6914	2	76	paper	paper	NOUN
fcis-6914	2	77	proposes	propose	VERB
fcis-6914	2	78	a	a	DET
fcis-6914	2	79	yolov5	yolov5	NOUN
fcis-6914	2	80	face	face	NOUN
fcis-6914	2	81	helmet	helmet	NOUN
fcis-6914	2	82	detection	detection	NOUN
fcis-6914	2	83	algorithm	algorithm	NOUN
fcis-6914	2	84	based	base	VERB
fcis-6914	2	85	on	on	ADP
fcis-6914	2	86	swin	swin	PROPN
fcis-6914	2	87	transformer	transformer	PROPN
fcis-6914	2	88	improvement	improvement	NOUN
fcis-6914	2	89	from	from	ADP
fcis-6914	2	90	the	the	DET
fcis-6914	2	91	overall	overall	ADJ
fcis-6914	2	92	semantics	semantic	NOUN
fcis-6914	2	93	of	of	ADP
fcis-6914	2	94	the	the	DET
fcis-6914	2	95	image	image	NOUN
fcis-6914	2	96	.	.	PUNCT
fcis-6914	3	1	in	in	ADP
fcis-6914	3	2	this	this	DET
fcis-6914	3	3	paper	paper	NOUN
fcis-6914	3	4	,	,	PUNCT
fcis-6914	3	5	experiments	experiment	NOUN
fcis-6914	3	6	are	be	AUX
fcis-6914	3	7	conducted	conduct	VERB
fcis-6914	3	8	using	use	VERB
fcis-6914	3	9	a	a	DET
fcis-6914	3	10	self	self	NOUN
fcis-6914	3	11	-	-	PUNCT
fcis-6914	3	12	built	build	VERB
fcis-6914	3	13	dataset	dataset	NOUN
fcis-6914	3	14	to	to	PART
fcis-6914	3	15	further	far	ADV
fcis-6914	3	16	enhance	enhance	VERB
fcis-6914	3	17	the	the	DET
fcis-6914	3	18	performance	performance	NOUN
fcis-6914	3	19	of	of	ADP
fcis-6914	3	20	the	the	DET
fcis-6914	3	21	model	model	NOUN
fcis-6914	3	22	and	and	CCONJ
fcis-6914	3	23	improve	improve	VERB
fcis-6914	3	24	the	the	DET
fcis-6914	3	25	accuracy	accuracy	NOUN
fcis-6914	3	26	of	of	ADP
fcis-6914	3	27	face	face	NOUN
fcis-6914	3	28	helmet	helmet	NOUN
fcis-6914	3	29	detection	detection	NOUN
fcis-6914	3	30	through	through	ADP
fcis-6914	3	31	mosaic	mosaic	ADJ
fcis-6914	3	32	data	datum	NOUN
fcis-6914	3	33	enhancement	enhancement	NOUN
fcis-6914	3	34	,	,	PUNCT
fcis-6914	3	35	label	label	NOUN
fcis-6914	3	36	smoothing	smooth	VERB
fcis-6914	3	37	processing	processing	NOUN
fcis-6914	3	38	,	,	PUNCT
fcis-6914	3	39	adaptive	adaptive	ADJ
fcis-6914	3	40	weighted	weight	VERB
fcis-6914	3	41	features	feature	NOUN
fcis-6914	3	42	combined	combine	VERB
fcis-6914	3	43	with	with	ADP
fcis-6914	3	44	wconcat	wconcat	PROPN
fcis-6914	3	45	module	module	NOUN
fcis-6914	3	46	and	and	CCONJ
fcis-6914	3	47	the	the	DET
fcis-6914	3	48	application	application	NOUN
fcis-6914	3	49	of	of	ADP
fcis-6914	3	50	c3tr	c3tr	ADJ
fcis-6914	3	51	and	and	CCONJ
fcis-6914	3	52	c3str	c3str	NUM
fcis-6914	3	53	modules	module	NOUN
fcis-6914	3	54	to	to	PART
fcis-6914	3	55	fuse	fuse	VERB
fcis-6914	3	56	multi	multi	ADJ
fcis-6914	3	57	-	-	ADJ
fcis-6914	3	58	scale	scale	ADJ
fcis-6914	3	59	information	information	NOUN
fcis-6914	3	60	,	,	PUNCT
fcis-6914	3	61	enhance	enhance	VERB
fcis-6914	3	62	the	the	DET
fcis-6914	3	63	feature	feature	NOUN
fcis-6914	3	64	extraction	extraction	NOUN
fcis-6914	3	65	capability	capability	NOUN
fcis-6914	3	66	of	of	ADP
fcis-6914	3	67	the	the	DET
fcis-6914	3	68	network	network	NOUN
fcis-6914	3	69	,	,	PUNCT
fcis-6914	3	70	and	and	CCONJ
fcis-6914	3	71	improve	improve	VERB
fcis-6914	3	72	the	the	DET
fcis-6914	3	73	generalization	generalization	NOUN
fcis-6914	3	74	and	and	CCONJ
fcis-6914	3	75	robustness	robustness	NOUN
fcis-6914	3	76	of	of	ADP
fcis-6914	3	77	the	the	DET
fcis-6914	3	78	model	model	NOUN
fcis-6914	3	79	with	with	ADP
fcis-6914	3	80	a	a	DET
fcis-6914	3	81	self	self	NOUN
fcis-6914	3	82	-	-	PUNCT
fcis-6914	3	83	built	build	VERB
fcis-6914	3	84	dataset	dataset	NOUN
fcis-6914	3	85	.	.	PUNCT
fcis-6914	4	1	experiments	experiment	NOUN
fcis-6914	4	2	show	show	VERB
fcis-6914	4	3	that	that	SCONJ
fcis-6914	4	4	the	the	DET
fcis-6914	4	5	improved	improved	ADJ
fcis-6914	4	6	yolov5	yolov5	NOUN
fcis-6914	4	7	face	face	NOUN
fcis-6914	4	8	helmet	helmet	NOUN
fcis-6914	4	9	detection	detection	NOUN
fcis-6914	4	10	algorithm	algorithm	NOUN
fcis-6914	4	11	map	map	NOUN
fcis-6914	4	12	based	base	VERB
fcis-6914	4	13	on	on	ADP
fcis-6914	4	14	swin	swin	PROPN
fcis-6914	4	15	transformer	transformer	NOUN
fcis-6914	4	16	improves	improve	VERB
fcis-6914	4	17	5.7	5.7	NUM
fcis-6914	4	18	%	%	NOUN
fcis-6914	4	19	compared	compare	VERB
fcis-6914	4	20	with	with	ADP
fcis-6914	4	21	faster	fast	ADJ
fcis-6914	4	22	rcnn	rcnn	NOUN
fcis-6914	4	23	,	,	PUNCT
fcis-6914	4	24	6.1	6.1	NUM
fcis-6914	4	25	%	%	NOUN
fcis-6914	4	26	compared	compare	VERB
fcis-6914	4	27	with	with	ADP
fcis-6914	4	28	yolov3	yolov3	PROPN
fcis-6914	4	29	,	,	PUNCT
fcis-6914	4	30	5.3	5.3	NUM
fcis-6914	4	31	%	%	NOUN
fcis-6914	4	32	compared	compare	VERB
fcis-6914	4	33	with	with	ADP
fcis-6914	4	34	yolov4	yolov4	PROPN
fcis-6914	4	35	,	,	PUNCT
fcis-6914	4	36	and	and	CCONJ
fcis-6914	4	37	1.6	1.6	NUM
fcis-6914	4	38	%	%	NOUN
fcis-6914	4	39	compared	compare	VERB
fcis-6914	4	40	with	with	ADP
fcis-6914	4	41	the	the	DET
fcis-6914	4	42	original	original	ADJ
fcis-6914	4	43	algorithm	algorithm	NOUN
fcis-6914	4	44	.	.	PUNCT
fcis-6914	5	1	it	it	PRON
fcis-6914	5	2	performs	perform	VERB
fcis-6914	5	3	well	well	ADV
fcis-6914	5	4	in	in	ADP
fcis-6914	5	5	helmet	helmet	NOUN
fcis-6914	5	6	face	face	NOUN
fcis-6914	5	7	detection	detection	NOUN
fcis-6914	5	8	tasks	task	NOUN
fcis-6914	5	9	in	in	ADP
fcis-6914	5	10	complex	complex	ADJ
fcis-6914	5	11	environments	environment	NOUN
fcis-6914	5	12	,	,	PUNCT
fcis-6914	5	13	achieving	achieve	VERB
fcis-6914	5	14	real	real	ADJ
fcis-6914	5	15	-	-	PUNCT
fcis-6914	5	16	time	time	NOUN
fcis-6914	5	17	detection	detection	NOUN
fcis-6914	5	18	and	and	CCONJ
fcis-6914	5	19	higher	high	ADJ
fcis-6914	5	20	accuracy	accuracy	NOUN
fcis-6914	5	21	,	,	PUNCT
fcis-6914	5	22	while	while	SCONJ
fcis-6914	5	23	reducing	reduce	VERB
fcis-6914	5	24	missed	miss	VERB
fcis-6914	5	25	detections	detection	NOUN
fcis-6914	5	26	.	.	PUNCT
fcis-6914	6	1	keywords	keyword	NOUN
fcis-6914	6	2	:	:	PUNCT
fcis-6914	6	3	safety	safety	NOUN
fcis-6914	6	4	helmet	helmet	NOUN
fcis-6914	6	5	detection	detection	NOUN
fcis-6914	6	6	;	;	PUNCT
fcis-6914	6	7	yolov5	yolov5	NOUN
fcis-6914	6	8	;	;	PUNCT
fcis-6914	6	9	swin	swin	PROPN
fcis-6914	6	10	transformer	transformer	PROPN
fcis-6914	6	11	.	.	PUNCT
fcis-6914	7	1	1	1	X
fcis-6914	7	2	.	.	X
fcis-6914	7	3	introduction	introduction	NOUN
fcis-6914	7	4	in	in	ADP
fcis-6914	7	5	recent	recent	ADJ
fcis-6914	7	6	years	year	NOUN
fcis-6914	7	7	,	,	PUNCT
fcis-6914	7	8	the	the	DET
fcis-6914	7	9	phenomenon	phenomenon	NOUN
fcis-6914	7	10	that	that	SCONJ
fcis-6914	7	11	construction	construction	NOUN
fcis-6914	7	12	site	site	NOUN
fcis-6914	7	13	operators	operator	NOUN
fcis-6914	7	14	do	do	AUX
fcis-6914	7	15	not	not	PART
fcis-6914	7	16	wear	wear	VERB
fcis-6914	7	17	helmets	helmet	NOUN
fcis-6914	7	18	and	and	CCONJ
fcis-6914	7	19	cause	cause	VERB
fcis-6914	7	20	safety	safety	NOUN
fcis-6914	7	21	problems	problem	NOUN
fcis-6914	7	22	occurs	occur	VERB
fcis-6914	7	23	frequently	frequently	ADV
fcis-6914	7	24	[	[	X
fcis-6914	7	25	1	1	NUM
fcis-6914	7	26	]	]	PUNCT
fcis-6914	7	27	.	.	PUNCT
fcis-6914	8	1	the	the	DET
fcis-6914	8	2	safety	safety	NOUN
fcis-6914	8	3	awareness	awareness	NOUN
fcis-6914	8	4	of	of	ADP
fcis-6914	8	5	the	the	DET
fcis-6914	8	6	workers	worker	NOUN
fcis-6914	8	7	is	be	AUX
fcis-6914	8	8	weak	weak	ADJ
fcis-6914	8	9	,	,	PUNCT
fcis-6914	8	10	the	the	DET
fcis-6914	8	11	management	management	NOUN
fcis-6914	8	12	personnel	personnel	NOUN
fcis-6914	8	13	do	do	AUX
fcis-6914	8	14	not	not	PART
fcis-6914	8	15	do	do	VERB
fcis-6914	8	16	their	their	PRON
fcis-6914	8	17	duty	duty	NOUN
fcis-6914	8	18	of	of	ADP
fcis-6914	8	19	supervision	supervision	NOUN
fcis-6914	8	20	and	and	CCONJ
fcis-6914	8	21	supervision	supervision	NOUN
fcis-6914	8	22	in	in	ADP
fcis-6914	8	23	time	time	NOUN
fcis-6914	8	24	,	,	PUNCT
fcis-6914	8	25	and	and	CCONJ
fcis-6914	8	26	the	the	DET
fcis-6914	8	27	situation	situation	NOUN
fcis-6914	8	28	of	of	ADP
fcis-6914	8	29	not	not	PART
fcis-6914	8	30	wearing	wear	VERB
fcis-6914	8	31	helmets	helmet	NOUN
fcis-6914	8	32	as	as	ADP
fcis-6914	8	33	required	require	VERB
fcis-6914	8	34	is	be	AUX
fcis-6914	8	35	a	a	DET
fcis-6914	8	36	huge	huge	ADJ
fcis-6914	8	37	safety	safety	NOUN
fcis-6914	8	38	hazard	hazard	NOUN
fcis-6914	8	39	in	in	ADP
fcis-6914	8	40	construction	construction	NOUN
fcis-6914	8	41	.	.	PUNCT
fcis-6914	9	1	therefore	therefore	ADV
fcis-6914	9	2	,	,	PUNCT
fcis-6914	9	3	the	the	DET
fcis-6914	9	4	detection	detection	NOUN
fcis-6914	9	5	of	of	ADP
fcis-6914	9	6	the	the	DET
fcis-6914	9	7	helmet	helmet	NOUN
fcis-6914	9	8	wearing	wear	VERB
fcis-6914	9	9	situation	situation	NOUN
fcis-6914	9	10	of	of	ADP
fcis-6914	9	11	the	the	DET
fcis-6914	9	12	workers	worker	NOUN
fcis-6914	9	13	and	and	CCONJ
fcis-6914	9	14	the	the	DET
fcis-6914	9	15	identification	identification	NOUN
fcis-6914	9	16	of	of	ADP
fcis-6914	9	17	those	those	PRON
fcis-6914	9	18	who	who	PRON
fcis-6914	9	19	are	be	AUX
fcis-6914	9	20	not	not	PART
fcis-6914	9	21	wearing	wear	VERB
fcis-6914	9	22	helmets	helmet	NOUN
fcis-6914	9	23	play	play	VERB
fcis-6914	9	24	an	an	DET
fcis-6914	9	25	important	important	ADJ
fcis-6914	9	26	role	role	NOUN
fcis-6914	9	27	in	in	ADP
fcis-6914	9	28	the	the	DET
fcis-6914	9	29	safety	safety	NOUN
fcis-6914	9	30	management	management	NOUN
fcis-6914	9	31	afterwards	afterwards	ADV
fcis-6914	9	32	[	[	X
fcis-6914	9	33	2	2	NUM
fcis-6914	9	34	]	]	PUNCT
fcis-6914	9	35	.	.	PUNCT
fcis-6914	10	1	target	target	NOUN
fcis-6914	10	2	detection	detection	NOUN
fcis-6914	10	3	is	be	AUX
fcis-6914	10	4	an	an	DET
fcis-6914	10	5	important	important	ADJ
fcis-6914	10	6	topic	topic	NOUN
fcis-6914	10	7	in	in	ADP
fcis-6914	10	8	computer	computer	NOUN
fcis-6914	10	9	vision	vision	NOUN
fcis-6914	10	10	.	.	PUNCT
fcis-6914	11	1	with	with	ADP
fcis-6914	11	2	the	the	DET
fcis-6914	11	3	rapid	rapid	ADJ
fcis-6914	11	4	development	development	NOUN
fcis-6914	11	5	of	of	ADP
fcis-6914	11	6	computer	computer	NOUN
fcis-6914	11	7	vision	vision	NOUN
fcis-6914	11	8	,	,	PUNCT
fcis-6914	11	9	deep	deep	ADJ
fcis-6914	11	10	learning	learning	NOUN
fcis-6914	11	11	methods	method	NOUN
fcis-6914	11	12	for	for	ADP
fcis-6914	11	13	target	target	NOUN
fcis-6914	11	14	detection	detection	NOUN
fcis-6914	11	15	are	be	AUX
fcis-6914	11	16	mainly	mainly	ADV
fcis-6914	11	17	divided	divide	VERB
fcis-6914	11	18	into	into	ADP
fcis-6914	11	19	two	two	NUM
fcis-6914	11	20	technical	technical	ADJ
fcis-6914	11	21	paths	path	NOUN
fcis-6914	11	22	:	:	PUNCT
fcis-6914	11	23	one	one	NUM
fcis-6914	11	24	stage	stage	NOUN
fcis-6914	11	25	detection	detection	NOUN
fcis-6914	11	26	algorithm	algorithm	NOUN
fcis-6914	11	27	and	and	CCONJ
fcis-6914	11	28	two	two	NUM
fcis-6914	11	29	stage	stage	NOUN
fcis-6914	11	30	detection	detection	NOUN
fcis-6914	11	31	algorithm	algorithm	NOUN
fcis-6914	11	32	.	.	PUNCT
fcis-6914	12	1	2014	2014	NUM
fcis-6914	12	2	ross	ross	PROPN
fcis-6914	12	3	b.	b.	PROPN
fcis-6914	12	4	girshick	girshick	PROPN
fcis-6914	12	5	of	of	ADP
fcis-6914	12	6	university	university	PROPN
fcis-6914	12	7	of	of	ADP
fcis-6914	12	8	california	california	PROPN
fcis-6914	12	9	,	,	PUNCT
fcis-6914	12	10	berkeley	berkeley	PROPN
fcis-6914	12	11	proposed	propose	VERB
fcis-6914	12	12	the	the	DET
fcis-6914	12	13	algorithm	algorithm	NOUN
fcis-6914	12	14	of	of	ADP
fcis-6914	12	15	r	r	NOUN
fcis-6914	12	16	-	-	PUNCT
fcis-6914	12	17	cnn	cnn	NOUN
fcis-6914	12	18	[	[	X
fcis-6914	12	19	3	3	X
fcis-6914	12	20	]	]	PUNCT
fcis-6914	12	21	to	to	PART
fcis-6914	12	22	combine	combine	VERB
fcis-6914	12	23	cnn	cnn	PROPN
fcis-6914	12	24	with	with	ADP
fcis-6914	12	25	the	the	DET
fcis-6914	12	26	field	field	NOUN
fcis-6914	12	27	of	of	ADP
fcis-6914	12	28	target	target	NOUN
fcis-6914	12	29	detection	detection	NOUN
fcis-6914	12	30	,	,	PUNCT
fcis-6914	12	31	which	which	PRON
fcis-6914	12	32	opened	open	VERB
fcis-6914	12	33	the	the	DET
fcis-6914	12	34	era	era	NOUN
fcis-6914	12	35	of	of	ADP
fcis-6914	12	36	two	two	NUM
fcis-6914	12	37	stage	stage	NOUN
fcis-6914	12	38	.	.	PUNCT
fcis-6914	13	1	in	in	ADP
fcis-6914	13	2	2015	2015	NUM
fcis-6914	13	3	,	,	PUNCT
fcis-6914	13	4	ross	ross	PROPN
fcis-6914	13	5	girshick	girshick	PROPN
fcis-6914	13	6	of	of	ADP
fcis-6914	13	7	microsoft	microsoft	PROPN
fcis-6914	13	8	research	research	PROPN
fcis-6914	13	9	proposed	propose	VERB
fcis-6914	13	10	an	an	DET
fcis-6914	13	11	improved	improved	ADJ
fcis-6914	13	12	fast	fast	ADJ
fcis-6914	13	13	r	r	NOUN
fcis-6914	13	14	-	-	PUNCT
fcis-6914	13	15	cnn	cnn	NOUN
fcis-6914	14	1	[	[	X
fcis-6914	14	2	4	4	NUM
fcis-6914	14	3	]	]	X
fcis-6914	14	4	algorithm	algorithm	NOUN
fcis-6914	14	5	,	,	PUNCT
fcis-6914	14	6	which	which	PRON
fcis-6914	14	7	borrowed	borrow	VERB
fcis-6914	14	8	the	the	DET
fcis-6914	14	9	structure	structure	NOUN
fcis-6914	14	10	of	of	ADP
fcis-6914	14	11	spp	spp	NOUN
fcis-6914	14	12	-	-	PUNCT
fcis-6914	14	13	net	net	ADJ
fcis-6914	14	14	algorithm	algorithm	NOUN
fcis-6914	14	15	to	to	PART
fcis-6914	14	16	accomplish	accomplish	VERB
fcis-6914	14	17	a	a	DET
fcis-6914	14	18	highefficiency	highefficiency	NOUN
fcis-6914	14	19	classification	classification	NOUN
fcis-6914	14	20	task	task	NOUN
fcis-6914	14	21	based	base	VERB
fcis-6914	14	22	on	on	ADP
fcis-6914	14	23	the	the	DET
fcis-6914	14	24	original	original	ADJ
fcis-6914	14	25	research	research	NOUN
fcis-6914	14	26	,	,	PUNCT
fcis-6914	14	27	and	and	CCONJ
fcis-6914	14	28	improved	improve	VERB
fcis-6914	14	29	the	the	DET
fcis-6914	14	30	training	training	NOUN
fcis-6914	14	31	speed	speed	NOUN
fcis-6914	14	32	and	and	CCONJ
fcis-6914	14	33	accuracy	accuracy	NOUN
fcis-6914	14	34	by	by	ADP
fcis-6914	14	35	using	use	VERB
fcis-6914	14	36	multitask	multitask	ADJ
fcis-6914	14	37	error	error	NOUN
fcis-6914	14	38	against	against	ADP
fcis-6914	14	39	the	the	DET
fcis-6914	14	40	shortcoming	shortcoming	NOUN
fcis-6914	14	41	of	of	ADP
fcis-6914	14	42	r	r	NOUN
fcis-6914	14	43	-	-	PUNCT
fcis-6914	14	44	cnn	cnn	PROPN
fcis-6914	14	45	training	training	NOUN
fcis-6914	14	46	into	into	ADP
fcis-6914	14	47	multiple	multiple	ADJ
fcis-6914	14	48	steps	step	NOUN
fcis-6914	14	49	.	.	PUNCT
fcis-6914	14	50	2015	2015	NUM
fcis-6914	14	51	,	,	PUNCT
fcis-6914	14	52	shaoqing	shaoqe	VERB
fcis-6914	14	53	ren	ren	PROPN
fcis-6914	14	54	and	and	CCONJ
fcis-6914	14	55	kai	kai	PROPN
fcis-6914	14	56	-	-	PUNCT
fcis-6914	14	57	ming	ming	PROPN
fcis-6914	14	58	he	he	PROPN
fcis-6914	14	59	of	of	ADP
fcis-6914	14	60	microsoft	microsoft	PROPN
fcis-6914	14	61	research	research	PROPN
fcis-6914	14	62	in	in	ADP
fcis-6914	14	63	2015	2015	NUM
fcis-6914	14	64	,	,	PUNCT
fcis-6914	14	65	shaoqing	shaoqe	VERB
fcis-6914	14	66	ren	ren	PROPN
fcis-6914	14	67	,	,	PUNCT
fcis-6914	14	68	kaiming	kaime	VERB
fcis-6914	14	69	he	he	PRON
fcis-6914	14	70	,	,	PUNCT
fcis-6914	14	71	and	and	CCONJ
fcis-6914	14	72	ross	ross	PROPN
fcis-6914	14	73	girshick	girshick	PROPN
fcis-6914	14	74	of	of	ADP
fcis-6914	14	75	microsoft	microsoft	PROPN
fcis-6914	14	76	research	research	PROPN
fcis-6914	14	77	used	use	VERB
fcis-6914	14	78	a	a	DET
fcis-6914	14	79	region	region	NOUN
fcis-6914	14	80	proposal	proposal	NOUN
fcis-6914	14	81	networks	network	NOUN
fcis-6914	14	82	(	(	PUNCT
fcis-6914	14	83	rpn	rpn	PROPN
fcis-6914	14	84	)	)	PUNCT
fcis-6914	14	85	to	to	PART
fcis-6914	14	86	compute	compute	VERB
fcis-6914	14	87	posals	posal	NOUN
fcis-6914	14	88	,	,	PUNCT
fcis-6914	14	89	and	and	CCONJ
fcis-6914	14	90	the	the	DET
fcis-6914	14	91	faster	fast	ADJ
fcis-6914	14	92	r	r	NOUN
fcis-6914	14	93	-	-	PUNCT
fcis-6914	14	94	cnn	cnn	NOUN
fcis-6914	14	95	[	[	X
fcis-6914	14	96	5	5	NUM
fcis-6914	14	97	]	]	PUNCT
fcis-6914	14	98	algorithm	algorithm	NOUN
fcis-6914	14	99	was	be	AUX
fcis-6914	14	100	proposed	propose	VERB
fcis-6914	14	101	by	by	ADP
fcis-6914	14	102	rpn	rpn	PROPN
fcis-6914	14	103	+	+	CCONJ
fcis-6914	14	104	fast	fast	ADJ
fcis-6914	14	105	r	r	NOUN
fcis-6914	14	106	-	-	PUNCT
fcis-6914	14	107	cnn	cnn	NOUN
fcis-6914	14	108	.	.	PROPN
fcis-6914	15	1	2017	2017	NUM
fcis-6914	15	2	tsungyi	tsungyi	NOUN
fcis-6914	15	3	lin	lin	PROPN
fcis-6914	15	4	of	of	ADP
fcis-6914	15	5	facebook	facebook	PROPN
fcis-6914	15	6	et	et	PROPN
fcis-6914	15	7	al	al	PROPN
fcis-6914	15	8	.	.	PROPN
fcis-6914	15	9	proposed	propose	VERB
fcis-6914	15	10	the	the	DET
fcis-6914	15	11	fpn	fpn	PROPN
fcis-6914	15	12	algorithm	algorithm	NOUN
fcis-6914	15	13	,	,	PUNCT
fcis-6914	15	14	which	which	PRON
fcis-6914	15	15	truly	truly	ADV
fcis-6914	15	16	incorporates	incorporate	VERB
fcis-6914	15	17	the	the	DET
fcis-6914	15	18	whole	whole	ADJ
fcis-6914	15	19	process	process	NOUN
fcis-6914	15	20	of	of	ADP
fcis-6914	15	21	object	object	NOUN
fcis-6914	15	22	detection	detection	NOUN
fcis-6914	15	23	into	into	ADP
fcis-6914	15	24	a	a	DET
fcis-6914	15	25	neural	neural	ADJ
fcis-6914	15	26	network	network	NOUN
fcis-6914	15	27	.	.	PUNCT
fcis-6914	16	1	the	the	DET
fcis-6914	16	2	two	two	NUM
fcis-6914	16	3	-	-	PUNCT
fcis-6914	16	4	stage	stage	NOUN
fcis-6914	16	5	approach	approach	NOUN
fcis-6914	16	6	represented	represent	VERB
fcis-6914	16	7	by	by	ADP
fcis-6914	16	8	r	r	PROPN
fcis-6914	16	9	-	-	PUNCT
fcis-6914	16	10	cnn	cnn	PROPN
fcis-6914	16	11	algorithm	algorithm	NOUN
fcis-6914	16	12	faster	fast	ADV
fcis-6914	16	13	almost	almost	ADV
fcis-6914	16	14	achieves	achieve	VERB
fcis-6914	16	15	the	the	DET
fcis-6914	16	16	optimal	optimal	ADJ
fcis-6914	16	17	effect	effect	NOUN
fcis-6914	16	18	,	,	PUNCT
fcis-6914	16	19	but	but	CCONJ
fcis-6914	16	20	encounters	encounter	VERB
fcis-6914	16	21	a	a	DET
fcis-6914	16	22	bottleneck	bottleneck	NOUN
fcis-6914	16	23	in	in	ADP
fcis-6914	16	24	speed	speed	NOUN
fcis-6914	16	25	,	,	PUNCT
fcis-6914	16	26	and	and	CCONJ
fcis-6914	16	27	there	there	PRON
fcis-6914	16	28	is	be	VERB
fcis-6914	16	29	great	great	ADJ
fcis-6914	16	30	room	room	NOUN
fcis-6914	16	31	for	for	ADP
fcis-6914	16	32	improvement	improvement	NOUN
fcis-6914	16	33	in	in	ADP
fcis-6914	16	34	real	real	ADJ
fcis-6914	16	35	-	-	PUNCT
fcis-6914	16	36	time	time	NOUN
fcis-6914	16	37	performance	performance	NOUN
fcis-6914	16	38	.	.	PUNCT
fcis-6914	17	1	the	the	DET
fcis-6914	17	2	yolo	yolo	ADJ
fcis-6914	17	3	algorithm	algorithm	NOUN
fcis-6914	17	4	proposed	propose	VERB
fcis-6914	17	5	by	by	ADP
fcis-6914	17	6	joseph	joseph	PROPN
fcis-6914	17	7	redmon	redmon	PROPN
fcis-6914	17	8	et	et	PROPN
fcis-6914	17	9	al	al	PROPN
fcis-6914	17	10	.	.	PROPN
fcis-6914	18	1	at	at	ADP
fcis-6914	18	2	the	the	DET
fcis-6914	18	3	university	university	PROPN
fcis-6914	18	4	of	of	ADP
fcis-6914	18	5	washington	washington	PROPN
fcis-6914	18	6	in	in	ADP
fcis-6914	18	7	2015	2015	NUM
fcis-6914	18	8	inherited	inherit	VERB
fcis-6914	18	9	the	the	DET
fcis-6914	18	10	overfeat	overfeat	ADJ
fcis-6914	18	11	algorithm	algorithm	NOUN
fcis-6914	18	12	,	,	PUNCT
fcis-6914	18	13	a	a	DET
fcis-6914	18	14	regression	regression	NOUN
fcis-6914	18	15	-	-	PUNCT
fcis-6914	18	16	based	base	VERB
fcis-6914	18	17	one	one	NUM
fcis-6914	18	18	-	-	PUNCT
fcis-6914	18	19	stage	stage	NOUN
fcis-6914	18	20	method	method	NOUN
fcis-6914	18	21	called	call	VERB
fcis-6914	18	22	yolo	yolo	ADJ
fcis-6914	18	23	(	(	PUNCT
fcis-6914	18	24	you	you	PRON
fcis-6914	18	25	only	only	ADV
fcis-6914	18	26	look	look	VERB
fcis-6914	18	27	once	once	ADV
fcis-6914	18	28	)	)	PUNCT
fcis-6914	18	29	,	,	PUNCT
fcis-6914	18	30	which	which	PRON
fcis-6914	18	31	abstracts	abstract	VERB
fcis-6914	18	32	detection	detection	NOUN
fcis-6914	18	33	as	as	ADP
fcis-6914	18	34	a	a	DET
fcis-6914	18	35	regression	regression	NOUN
fcis-6914	18	36	problem	problem	NOUN
fcis-6914	18	37	.	.	PUNCT
fcis-6914	19	1	after	after	ADP
fcis-6914	19	2	improvements	improvement	NOUN
fcis-6914	19	3	by	by	ADP
fcis-6914	19	4	joseph	joseph	PROPN
fcis-6914	19	5	redmon	redmon	PROPN
fcis-6914	19	6	et	et	PROPN
fcis-6914	19	7	al	al	PROPN
fcis-6914	19	8	.	.	PUNCT
fcis-6914	20	1	the	the	DET
fcis-6914	20	2	yolov2	yolov2	PROPN
fcis-6914	20	3	and	and	CCONJ
fcis-6914	20	4	yolo9000[6	yolo9000[6	PROPN
fcis-6914	20	5	]	]	X
fcis-6914	20	6	algorithms	algorithm	NOUN
fcis-6914	20	7	were	be	AUX
fcis-6914	20	8	presented	present	VERB
fcis-6914	20	9	at	at	ADP
fcis-6914	20	10	cvpr	cvpr	NOUN
fcis-6914	20	11	2017	2017	NUM
fcis-6914	20	12	,	,	PUNCT
fcis-6914	20	13	enabling	enable	VERB
fcis-6914	20	14	end	end	NOUN
fcis-6914	20	15	-	-	PUNCT
fcis-6914	20	16	to	to	ADP
fcis-6914	20	17	-	-	PUNCT
fcis-6914	20	18	end	end	NOUN
fcis-6914	20	19	training	training	NOUN
fcis-6914	20	20	.	.	PUNCT
fcis-6914	21	1	in	in	ADP
fcis-6914	21	2	2018	2018	NUM
fcis-6914	21	3	joseph	joseph	PROPN
fcis-6914	21	4	redmon	redmon	PROPN
fcis-6914	21	5	et	et	PROPN
fcis-6914	21	6	al	al	PROPN
fcis-6914	21	7	.	.	PROPN
fcis-6914	21	8	proposed	propose	VERB
fcis-6914	21	9	darknet-53	darknet-53	PROPN
fcis-6914	21	10	again	again	ADV
fcis-6914	21	11	,	,	PUNCT
fcis-6914	21	12	using	use	VERB
fcis-6914	21	13	three	three	NUM
fcis-6914	21	14	feature	feature	NOUN
fcis-6914	21	15	pyramids	pyramid	NOUN
fcis-6914	21	16	of	of	ADP
fcis-6914	21	17	different	different	ADJ
fcis-6914	21	18	sizes	size	NOUN
fcis-6914	21	19	to	to	PART
fcis-6914	21	20	fuse	fuse	VERB
fcis-6914	21	21	features	feature	NOUN
fcis-6914	21	22	(	(	PUNCT
fcis-6914	21	23	feature	feature	NOUN
fcis-6914	21	24	map	map	NOUN
fcis-6914	21	25	)	)	PUNCT
fcis-6914	21	26	making	make	VERB
fcis-6914	21	27	yolov3	yolov3	PROPN
fcis-6914	22	1	[	[	X
fcis-6914	22	2	7	7	X
fcis-6914	22	3	]	]	PUNCT
fcis-6914	22	4	to	to	PART
fcis-6914	22	5	achieve	achieve	VERB
fcis-6914	22	6	better	well	ADJ
fcis-6914	22	7	classification	classification	NOUN
fcis-6914	22	8	performance	performance	NOUN
fcis-6914	22	9	.	.	PUNCT
fcis-6914	23	1	in	in	ADP
fcis-6914	23	2	april	april	PROPN
fcis-6914	23	3	2020	2020	NUM
fcis-6914	23	4	,	,	PUNCT
fcis-6914	23	5	yolov4	yolov4	PROPN
fcis-6914	24	1	[	[	X
fcis-6914	24	2	8	8	NUM
fcis-6914	24	3	]	]	PUNCT
fcis-6914	24	4	using	use	VERB
fcis-6914	24	5	csp	csp	PROPN
fcis-6914	24	6	(	(	PUNCT
fcis-6914	24	7	cross	cross	NOUN
fcis-6914	24	8	stage	stage	NOUN
fcis-6914	24	9	partial	partial	NOUN
fcis-6914	24	10	)	)	PUNCT
fcis-6914	24	11	darknet-53	darknet-53	PROPN
fcis-6914	24	12	as	as	SCONJ
fcis-6914	24	13	backbone	backbone	NOUN
fcis-6914	24	14	and	and	CCONJ
fcis-6914	24	15	replacing	replace	VERB
fcis-6914	24	16	the	the	DET
fcis-6914	24	17	feature	feature	NOUN
fcis-6914	24	18	pyramid	pyramid	NOUN
fcis-6914	24	19	network	network	NOUN
fcis-6914	24	20	with	with	ADP
fcis-6914	24	21	panet	panet	PROPN
fcis-6914	24	22	(	(	PUNCT
fcis-6914	24	23	path	path	NOUN
fcis-6914	24	24	aggregation	aggregation	NOUN
fcis-6914	24	25	network	network	NOUN
fcis-6914	24	26	)	)	PUNCT
fcis-6914	24	27	was	be	AUX
fcis-6914	24	28	proposed	propose	VERB
fcis-6914	24	29	by	by	ADP
fcis-6914	24	30	alexey	alexey	PROPN
fcis-6914	24	31	bochkovskiy	bochkovskiy	PROPN
fcis-6914	24	32	et	et	PROPN
fcis-6914	24	33	al	al	PROPN
fcis-6914	24	34	.	.	PROPN
fcis-6914	24	35	proposed	propose	VERB
fcis-6914	24	36	.	.	PUNCT
fcis-6914	25	1	in	in	ADP
fcis-6914	25	2	june	june	PROPN
fcis-6914	25	3	of	of	ADP
fcis-6914	25	4	the	the	DET
fcis-6914	25	5	same	same	ADJ
fcis-6914	25	6	year	year	NOUN
fcis-6914	25	7	,	,	PUNCT
fcis-6914	25	8	the	the	DET
fcis-6914	25	9	yolov5	yolov5	NOUN
fcis-6914	25	10	[	[	X
fcis-6914	25	11	9	9	NUM
fcis-6914	25	12	]	]	PUNCT
fcis-6914	25	13	algorithm	algorithm	NOUN
fcis-6914	25	14	with	with	ADP
fcis-6914	25	15	improvements	improvement	NOUN
fcis-6914	25	16	such	such	ADJ
fcis-6914	25	17	as	as	ADP
fcis-6914	25	18	focus	focus	NOUN
fcis-6914	25	19	structure	structure	NOUN
fcis-6914	25	20	and	and	CCONJ
fcis-6914	25	21	mosaic	mosaic	ADJ
fcis-6914	25	22	data	datum	NOUN
fcis-6914	25	23	augmentation	augmentation	NOUN
fcis-6914	25	24	has	have	AUX
fcis-6914	25	25	greatly	greatly	ADV
fcis-6914	25	26	improved	improve	VERB
fcis-6914	25	27	its	its	PRON
fcis-6914	25	28	performance	performance	NOUN
fcis-6914	25	29	in	in	ADP
fcis-6914	25	30	terms	term	NOUN
fcis-6914	25	31	of	of	ADP
fcis-6914	25	32	speed	speed	NOUN
fcis-6914	25	33	and	and	CCONJ
fcis-6914	25	34	accuracy	accuracy	NOUN
fcis-6914	25	35	,	,	PUNCT
fcis-6914	25	36	and	and	CCONJ
fcis-6914	25	37	so	so	ADV
fcis-6914	25	38	far	far	ADV
fcis-6914	25	39	yolov5	yolov5	NOUN
fcis-6914	25	40	has	have	AUX
fcis-6914	25	41	been	be	AUX
fcis-6914	25	42	released	release	VERB
fcis-6914	25	43	in	in	ADP
fcis-6914	25	44	v6.0	v6.0	NOUN
fcis-6914	25	45	with	with	ADP
fcis-6914	25	46	many	many	ADJ
fcis-6914	25	47	necks	neck	NOUN
fcis-6914	25	48	.	.	PUNCT
fcis-6914	26	1	recent	recent	ADJ
fcis-6914	26	2	years	year	NOUN
fcis-6914	26	3	,	,	PUNCT
fcis-6914	26	4	more	more	ADJ
fcis-6914	26	5	and	and	CCONJ
fcis-6914	26	6	more	more	ADJ
fcis-6914	26	7	network	network	NOUN
fcis-6914	26	8	models	model	NOUN
fcis-6914	26	9	have	have	AUX
fcis-6914	26	10	been	be	AUX
fcis-6914	26	11	proposed	propose	VERB
fcis-6914	26	12	and	and	CCONJ
fcis-6914	26	13	promoted	promote	VERB
fcis-6914	26	14	to	to	ADP
fcis-6914	26	15	various	various	ADJ
fcis-6914	26	16	detection	detection	NOUN
fcis-6914	26	17	tasks	task	NOUN
fcis-6914	26	18	,	,	PUNCT
fcis-6914	26	19	among	among	ADP
fcis-6914	26	20	which	which	PRON
fcis-6914	26	21	the	the	DET
fcis-6914	26	22	yolo	yolo	ADJ
fcis-6914	26	23	series	series	PROPN
fcis-6914	26	24	algorithms	algorithm	NOUN
fcis-6914	26	25	have	have	VERB
fcis-6914	26	26	many	many	ADJ
fcis-6914	26	27	advantages	advantage	NOUN
fcis-6914	26	28	such	such	ADJ
fcis-6914	26	29	as	as	ADP
fcis-6914	26	30	simple	simple	ADJ
fcis-6914	26	31	structure	structure	NOUN
fcis-6914	26	32	,	,	PUNCT
fcis-6914	26	33	high	high	ADJ
fcis-6914	26	34	versatility	versatility	NOUN
fcis-6914	26	35	,	,	PUNCT
fcis-6914	26	36	fast	fast	ADJ
fcis-6914	26	37	speed	speed	NOUN
fcis-6914	26	38	,	,	PUNCT
fcis-6914	26	39	applicable	applicable	ADJ
fcis-6914	26	40	to	to	ADP
fcis-6914	26	41	multiple	multiple	ADJ
fcis-6914	26	42	scenes	scene	NOUN
fcis-6914	26	43	,	,	PUNCT
fcis-6914	26	44	low	low	ADJ
fcis-6914	26	45	background	background	NOUN
fcis-6914	26	46	misidentification	misidentification	NOUN
fcis-6914	26	47	rate	rate	NOUN
fcis-6914	26	48	,	,	PUNCT
fcis-6914	26	49	etc	etc	X
fcis-6914	26	50	.	.	X
fcis-6914	27	1	with	with	ADP
fcis-6914	27	2	the	the	DET
fcis-6914	27	3	continuous	continuous	ADJ
fcis-6914	27	4	iterative	iterative	NOUN
fcis-6914	27	5	update	update	NOUN
fcis-6914	27	6	of	of	ADP
fcis-6914	27	7	the	the	DET
fcis-6914	27	8	algorithm	algorithm	NOUN
fcis-6914	27	9	,	,	PUNCT
fcis-6914	27	10	the	the	DET
fcis-6914	27	11	algorithm	algorithm	NOUN
fcis-6914	27	12	performance	performance	NOUN
fcis-6914	27	13	and	and	CCONJ
fcis-6914	27	14	possibilities	possibility	NOUN
fcis-6914	27	15	are	be	AUX
fcis-6914	27	16	being	be	AUX
fcis-6914	27	17	explored	explore	VERB
fcis-6914	27	18	.	.	PUNCT
fcis-6914	28	1	ding	de	VERB
fcis-6914	28	2	[	[	X
fcis-6914	28	3	10	10	NUM
fcis-6914	28	4	]	]	PUNCT
fcis-6914	28	5	et	et	PROPN
fcis-6914	28	6	al	al	PROPN
fcis-6914	28	7	.	.	PROPN
fcis-6914	28	8	proposed	propose	VERB
fcis-6914	28	9	a	a	DET
fcis-6914	28	10	safety	safety	NOUN
fcis-6914	28	11	helmet	helmet	NOUN
fcis-6914	28	12	detection	detection	NOUN
fcis-6914	28	13	method	method	NOUN
fcis-6914	28	14	based	base	VERB
fcis-6914	28	15	on	on	ADP
fcis-6914	28	16	improved	improved	ADJ
fcis-6914	28	17	yolov3	yolov3	PROPN
fcis-6914	28	18	.	.	PUNCT
fcis-6914	29	1	for	for	ADP
fcis-6914	29	2	the	the	DET
fcis-6914	29	3	problems	problem	NOUN
fcis-6914	29	4	of	of	ADP
fcis-6914	29	5	weak	weak	ADJ
fcis-6914	29	6	real	real	ADJ
fcis-6914	29	7	-	-	PUNCT
fcis-6914	29	8	time	time	NOUN
fcis-6914	29	9	performance	performance	NOUN
fcis-6914	29	10	and	and	CCONJ
fcis-6914	29	11	low	low	ADJ
fcis-6914	29	12	detection	detection	NOUN
fcis-6914	29	13	accuracy	accuracy	NOUN
fcis-6914	29	14	in	in	ADP
fcis-6914	29	15	industrial	industrial	ADJ
fcis-6914	29	16	site	site	NOUN
fcis-6914	29	17	helmet	helmet	NOUN
fcis-6914	29	18	wear	wear	NOUN
fcis-6914	29	19	detection	detection	NOUN
fcis-6914	29	20	,	,	PUNCT
fcis-6914	29	21	kmeans++	kmeans++	PROPN
fcis-6914	29	22	clustering	cluster	VERB
fcis-6914	29	23	algorithm	algorithm	NOUN
fcis-6914	29	24	is	be	AUX
fcis-6914	29	25	used	use	VERB
fcis-6914	29	26	to	to	PART
fcis-6914	29	27	optimize	optimize	VERB
fcis-6914	29	28	the	the	DET
fcis-6914	29	29	acquisition	acquisition	NOUN
fcis-6914	29	30	of	of	ADP
fcis-6914	29	31	helmet	helmet	NOUN
fcis-6914	29	32	wear	wear	VERB
fcis-6914	29	33	prior	prior	ADJ
fcis-6914	29	34	frame	frame	NOUN
fcis-6914	29	35	and	and	CCONJ
fcis-6914	29	36	introduce	introduce	VERB
fcis-6914	29	37	residual	residual	ADJ
fcis-6914	29	38	module	module	NOUN
fcis-6914	29	39	in	in	ADP
fcis-6914	29	40	the	the	DET
fcis-6914	29	41	network	network	NOUN
fcis-6914	29	42	prediction	prediction	NOUN
fcis-6914	29	43	stage	stage	NOUN
fcis-6914	29	44	to	to	PART
fcis-6914	29	45	further	far	ADV
fcis-6914	29	46	improve	improve	VERB
fcis-6914	29	47	the	the	DET
fcis-6914	29	48	average	average	ADJ
fcis-6914	29	49	accuracy	accuracy	NOUN
fcis-6914	29	50	.	.	PUNCT
fcis-6914	30	1	jing	je	VERB
fcis-6914	31	1	[	[	X
fcis-6914	31	2	11	11	NUM
fcis-6914	31	3	]	]	PUNCT
fcis-6914	31	4	et	et	PROPN
fcis-6914	31	5	al	al	PROPN
fcis-6914	31	6	.	.	PROPN
fcis-6914	31	7	proposed	propose	VERB
fcis-6914	31	8	a	a	DET
fcis-6914	31	9	helmet	helmet	NOUN
fcis-6914	31	10	wear	wear	NOUN
fcis-6914	31	11	detection	detection	NOUN
fcis-6914	31	12	algorithm	algorithm	NOUN
fcis-6914	31	13	based	base	VERB
fcis-6914	31	14	on	on	ADP
fcis-6914	31	15	the	the	DET
fcis-6914	31	16	improved	improved	ADJ
fcis-6914	31	17	yolov4	yolov4	NOUN
fcis-6914	31	18	based	base	VERB
fcis-6914	31	19	on	on	ADP
fcis-6914	31	20	three	three	NUM
fcis-6914	31	21	feature	feature	NOUN
fcis-6914	31	22	map	map	NOUN
fcis-6914	32	1	outputs	output	NOUN
fcis-6914	32	2	yue	yue	PROPN
fcis-6914	32	3	[	[	X
fcis-6914	32	4	12	12	NUM
fcis-6914	32	5	]	]	PUNCT
fcis-6914	32	6	et	et	PROPN
fcis-6914	32	7	al	al	PROPN
fcis-6914	32	8	.	.	PROPN
fcis-6914	32	9	proposed	propose	VERB
fcis-6914	32	10	an	an	DET
fcis-6914	32	11	improved	improved	ADJ
fcis-6914	32	12	yolov5	yolov5	NOUN
fcis-6914	32	13	based	base	VERB
fcis-6914	32	14	helmet	helmet	NOUN
fcis-6914	32	15	wear	wear	NOUN
fcis-6914	32	16	detection	detection	NOUN
fcis-6914	32	17	algorithm	algorithm	NOUN
fcis-6914	32	18	based	base	VERB
fcis-6914	32	19	on	on	ADP
fcis-6914	32	20	a	a	DET
fcis-6914	32	21	128×128	128×128	NUM
fcis-6914	32	22	feature	feature	NOUN
fcis-6914	32	23	map	map	NOUN
fcis-6914	32	24	output	output	NOUN
fcis-6914	32	25	with	with	ADP
fcis-6914	32	26	8	8	ADJ
fcis-6914	32	27	-	-	ADJ
fcis-6914	32	28	fold	fold	ADJ
fcis-6914	32	29	sampling	sample	VERB
fcis-6914	32	30	to	to	ADP
fcis-6914	32	31	4	4	NUM
fcis-6914	32	32	-	-	ADJ
fcis-6914	32	33	fold	fold	ADJ
fcis-6914	32	34	fusion	fusion	NOUN
fcis-6914	32	35	of	of	ADP
fcis-6914	32	36	more	more	ADJ
fcis-6914	32	37	small	small	ADJ
fcis-6914	32	38	target	target	NOUN
fcis-6914	32	39	features	feature	NOUN
fcis-6914	32	40	,	,	PUNCT
fcis-6914	32	41	to	to	PART
fcis-6914	32	42	address	address	VERB
fcis-6914	32	43	the	the	DET
fcis-6914	32	44	26	26	NUM
fcis-6914	32	45	drawback	drawback	NOUN
fcis-6914	32	46	that	that	PRON
fcis-6914	32	47	yolov5	yolov5	NOUN
fcis-6914	32	48	can	can	AUX
fcis-6914	32	49	not	not	PART
fcis-6914	32	50	get	get	VERB
fcis-6914	32	51	effective	effective	ADJ
fcis-6914	32	52	features	feature	NOUN
fcis-6914	32	53	by	by	ADP
fcis-6914	32	54	weight	weight	NOUN
fcis-6914	32	55	focusing	focus	VERB
fcis-6914	32	56	,	,	PUNCT
fcis-6914	32	57	using	use	VERB
fcis-6914	32	58	the	the	DET
fcis-6914	32	59	attention	attention	NOUN
fcis-6914	32	60	module	module	NOUN
fcis-6914	32	61	and	and	CCONJ
fcis-6914	32	62	retaining	retain	VERB
fcis-6914	32	63	more	more	ADJ
fcis-6914	32	64	prediction	prediction	NOUN
fcis-6914	32	65	frames	frame	NOUN
fcis-6914	32	66	using	use	VERB
fcis-6914	32	67	the	the	DET
fcis-6914	32	68	soft	soft	ADJ
fcis-6914	32	69	-	-	PUNCT
fcis-6914	32	70	nms	nms	NOUN
fcis-6914	32	71	algorithm	algorithm	NOUN
fcis-6914	32	72	to	to	PART
fcis-6914	32	73	improve	improve	VERB
fcis-6914	32	74	the	the	DET
fcis-6914	32	75	accuracy	accuracy	NOUN
fcis-6914	32	76	of	of	ADP
fcis-6914	32	77	prediction	prediction	NOUN
fcis-6914	32	78	frames	frame	NOUN
fcis-6914	32	79	.	.	PUNCT
fcis-6914	33	1	facility	facility	NOUN
fcis-6914	33	2	cost	cost	NOUN
fcis-6914	33	3	problem	problem	NOUN
fcis-6914	33	4	proposed	propose	VERB
fcis-6914	33	5	a	a	DET
fcis-6914	33	6	safety	safety	NOUN
fcis-6914	33	7	helmet	helmet	NOUN
fcis-6914	33	8	detection	detection	NOUN
fcis-6914	33	9	method	method	NOUN
fcis-6914	33	10	based	base	VERB
fcis-6914	33	11	on	on	ADP
fcis-6914	33	12	lightweight	lightweight	ADJ
fcis-6914	33	13	deep	deep	ADJ
fcis-6914	33	14	learning	learning	NOUN
fcis-6914	33	15	model	model	NOUN
fcis-6914	33	16	using	use	VERB
fcis-6914	33	17	a	a	DET
fcis-6914	33	18	lightweight	lightweight	ADJ
fcis-6914	33	19	improved	improve	VERB
fcis-6914	33	20	version	version	NOUN
fcis-6914	33	21	of	of	ADP
fcis-6914	33	22	tiny	tiny	ADJ
fcis-6914	33	23	-	-	PUNCT
fcis-6914	33	24	yolov3	yolov3	PROPN
fcis-6914	33	25	,	,	PUNCT
fcis-6914	33	26	lt	lt	PROPN
fcis-6914	33	27	-	-	PUNCT
fcis-6914	33	28	yolo	yolo	NOUN
fcis-6914	33	29	,	,	PUNCT
fcis-6914	33	30	which	which	PRON
fcis-6914	33	31	greatly	greatly	ADV
fcis-6914	33	32	reduces	reduce	VERB
fcis-6914	33	33	the	the	DET
fcis-6914	33	34	complexity	complexity	NOUN
fcis-6914	33	35	of	of	ADP
fcis-6914	33	36	the	the	DET
fcis-6914	33	37	model	model	NOUN
fcis-6914	33	38	.	.	PUNCT
fcis-6914	34	1	in	in	ADP
fcis-6914	34	2	summary	summary	NOUN
fcis-6914	34	3	,	,	PUNCT
fcis-6914	34	4	the	the	DET
fcis-6914	34	5	helmet	helmet	NOUN
fcis-6914	34	6	detection	detection	NOUN
fcis-6914	34	7	wearing	wear	VERB
fcis-6914	34	8	problem	problem	NOUN
fcis-6914	34	9	has	have	AUX
fcis-6914	34	10	been	be	AUX
fcis-6914	34	11	widely	widely	ADV
fcis-6914	34	12	studied	study	VERB
fcis-6914	34	13	,	,	PUNCT
fcis-6914	34	14	however	however	ADV
fcis-6914	34	15	,	,	PUNCT
fcis-6914	34	16	although	although	SCONJ
fcis-6914	34	17	the	the	DET
fcis-6914	34	18	above	above	ADJ
fcis-6914	34	19	yolo	yolo	ADJ
fcis-6914	34	20	helmet	helmet	NOUN
fcis-6914	34	21	detection	detection	NOUN
fcis-6914	34	22	algorithm	algorithm	NOUN
fcis-6914	34	23	has	have	AUX
fcis-6914	34	24	achieved	achieve	VERB
fcis-6914	34	25	performance	performance	NOUN
fcis-6914	34	26	optimization	optimization	NOUN
fcis-6914	34	27	in	in	ADP
fcis-6914	34	28	different	different	ADJ
fcis-6914	34	29	aspects	aspect	NOUN
fcis-6914	34	30	,	,	PUNCT
fcis-6914	34	31	it	it	PRON
fcis-6914	34	32	is	be	AUX
fcis-6914	34	33	still	still	ADV
fcis-6914	34	34	difficult	difficult	ADJ
fcis-6914	34	35	to	to	PART
fcis-6914	34	36	maintain	maintain	VERB
fcis-6914	34	37	high	high	ADJ
fcis-6914	34	38	accuracy	accuracy	NOUN
fcis-6914	34	39	in	in	ADP
fcis-6914	34	40	the	the	DET
fcis-6914	34	41	complex	complex	ADJ
fcis-6914	34	42	background	background	NOUN
fcis-6914	34	43	of	of	ADP
fcis-6914	34	44	the	the	DET
fcis-6914	34	45	construction	construction	NOUN
fcis-6914	34	46	environment	environment	NOUN
fcis-6914	34	47	with	with	ADP
fcis-6914	34	48	overlapping	overlap	VERB
fcis-6914	34	49	crowds	crowd	NOUN
fcis-6914	34	50	of	of	ADP
fcis-6914	34	51	multiple	multiple	ADJ
fcis-6914	34	52	small	small	ADJ
fcis-6914	34	53	and	and	CCONJ
fcis-6914	34	54	medium	medium	ADJ
fcis-6914	34	55	-	-	PUNCT
fcis-6914	34	56	sized	sized	ADJ
fcis-6914	34	57	targets	target	NOUN
fcis-6914	34	58	.	.	PUNCT
fcis-6914	35	1	the	the	DET
fcis-6914	35	2	reason	reason	NOUN
fcis-6914	35	3	is	be	AUX
fcis-6914	35	4	that	that	SCONJ
fcis-6914	35	5	the	the	DET
fcis-6914	35	6	complex	complex	ADJ
fcis-6914	35	7	construction	construction	NOUN
fcis-6914	35	8	environment	environment	NOUN
fcis-6914	35	9	,	,	PUNCT
fcis-6914	35	10	target	target	VERB
fcis-6914	35	11	occlusion	occlusion	NOUN
fcis-6914	35	12	,	,	PUNCT
fcis-6914	35	13	light	light	NOUN
fcis-6914	35	14	and	and	CCONJ
fcis-6914	35	15	other	other	ADJ
fcis-6914	35	16	factors	factor	NOUN
fcis-6914	35	17	are	be	AUX
fcis-6914	35	18	very	very	ADV
fcis-6914	35	19	easy	easy	ADJ
fcis-6914	35	20	to	to	PART
fcis-6914	35	21	affect	affect	VERB
fcis-6914	35	22	the	the	DET
fcis-6914	35	23	detection	detection	NOUN
fcis-6914	35	24	results	result	NOUN
fcis-6914	35	25	,	,	PUNCT
fcis-6914	35	26	and	and	CCONJ
fcis-6914	35	27	the	the	DET
fcis-6914	35	28	helmet	helmet	NOUN
fcis-6914	35	29	target	target	NOUN
fcis-6914	35	30	in	in	ADP
fcis-6914	35	31	the	the	DET
fcis-6914	35	32	original	original	ADJ
fcis-6914	35	33	size	size	NOUN
fcis-6914	35	34	of	of	ADP
fcis-6914	35	35	the	the	DET
fcis-6914	35	36	image	image	NOUN
fcis-6914	35	37	is	be	AUX
fcis-6914	35	38	relatively	relatively	ADV
fcis-6914	35	39	low	low	ADJ
fcis-6914	35	40	to	to	PART
fcis-6914	35	41	cause	cause	VERB
fcis-6914	35	42	target	target	NOUN
fcis-6914	35	43	miss	miss	NOUN
fcis-6914	35	44	detection	detection	NOUN
fcis-6914	35	45	and	and	CCONJ
fcis-6914	35	46	background	background	NOUN
fcis-6914	35	47	false	false	ADJ
fcis-6914	35	48	detection	detection	NOUN
fcis-6914	35	49	.	.	PUNCT
fcis-6914	36	1	inspired	inspire	VERB
fcis-6914	36	2	by	by	ADP
fcis-6914	36	3	the	the	DET
fcis-6914	36	4	design	design	NOUN
fcis-6914	36	5	ideas	idea	NOUN
fcis-6914	36	6	of	of	ADP
fcis-6914	36	7	transformer	transformer	NOUN
fcis-6914	36	8	[	[	X
fcis-6914	36	9	14	14	NUM
fcis-6914	36	10	]	]	X
fcis-6914	36	11	,	,	PUNCT
fcis-6914	36	12	which	which	PRON
fcis-6914	36	13	has	have	AUX
fcis-6914	36	14	achieved	achieve	VERB
fcis-6914	36	15	great	great	ADJ
fcis-6914	36	16	success	success	NOUN
fcis-6914	36	17	in	in	ADP
fcis-6914	36	18	cv	cv	PROPN
fcis-6914	36	19	field	field	NOUN
fcis-6914	36	20	in	in	ADP
fcis-6914	36	21	2020	2020	NUM
fcis-6914	36	22	(	(	PUNCT
fcis-6914	36	23	e.g.	e.g.	ADV
fcis-6914	36	24	,	,	PUNCT
fcis-6914	36	25	vit	vit	NOUN
fcis-6914	36	26	for	for	ADP
fcis-6914	36	27	image	image	NOUN
fcis-6914	36	28	classification	classification	NOUN
fcis-6914	36	29	,	,	PUNCT
fcis-6914	36	30	detr	detr	NOUN
fcis-6914	36	31	model	model	NOUN
fcis-6914	36	32	for	for	ADP
fcis-6914	36	33	target	target	NOUN
fcis-6914	36	34	detection	detection	NOUN
fcis-6914	36	35	)	)	PUNCT
fcis-6914	36	36	,	,	PUNCT
fcis-6914	36	37	transformer	transformer	NOUN
fcis-6914	36	38	has	have	AUX
fcis-6914	36	39	achieved	achieve	VERB
fcis-6914	36	40	optimal	optimal	ADJ
fcis-6914	36	41	results	result	NOUN
fcis-6914	36	42	in	in	ADP
fcis-6914	36	43	detection	detection	NOUN
fcis-6914	36	44	,	,	PUNCT
fcis-6914	36	45	classification	classification	NOUN
fcis-6914	36	46	and	and	CCONJ
fcis-6914	36	47	segmentation	segmentation	NOUN
fcis-6914	36	48	in	in	ADP
fcis-6914	36	49	successive	successive	ADJ
fcis-6914	36	50	fields	field	NOUN
fcis-6914	36	51	,	,	PUNCT
fcis-6914	36	52	with	with	ADP
fcis-6914	36	53	the	the	DET
fcis-6914	36	54	features	feature	NOUN
fcis-6914	36	55	of	of	ADP
fcis-6914	36	56	strong	strong	ADJ
fcis-6914	36	57	data	datum	NOUN
fcis-6914	36	58	adaptation	adaptation	NOUN
fcis-6914	36	59	ability	ability	NOUN
fcis-6914	36	60	and	and	CCONJ
fcis-6914	36	61	more	more	ADV
fcis-6914	36	62	robust	robust	ADJ
fcis-6914	36	63	global	global	ADJ
fcis-6914	36	64	information	information	NOUN
fcis-6914	36	65	extraction	extraction	NOUN
fcis-6914	36	66	.	.	PUNCT
fcis-6914	37	1	to	to	PART
fcis-6914	37	2	address	address	VERB
fcis-6914	37	3	the	the	DET
fcis-6914	37	4	problems	problem	NOUN
fcis-6914	37	5	of	of	ADP
fcis-6914	37	6	low	low	ADJ
fcis-6914	37	7	accuracy	accuracy	NOUN
fcis-6914	37	8	of	of	ADP
fcis-6914	37	9	helmet	helmet	NOUN
fcis-6914	37	10	detection	detection	NOUN
fcis-6914	37	11	in	in	ADP
fcis-6914	37	12	complex	complex	ADJ
fcis-6914	37	13	environments	environment	NOUN
fcis-6914	37	14	and	and	CCONJ
fcis-6914	37	15	the	the	DET
fcis-6914	37	16	difficulty	difficulty	NOUN
fcis-6914	37	17	of	of	ADP
fcis-6914	37	18	construction	construction	NOUN
fcis-6914	37	19	parties	party	NOUN
fcis-6914	37	20	to	to	PART
fcis-6914	37	21	manage	manage	VERB
fcis-6914	37	22	helmet	helmet	NOUN
fcis-6914	37	23	wearing	wear	VERB
fcis-6914	37	24	,	,	PUNCT
fcis-6914	37	25	this	this	DET
fcis-6914	37	26	paper	paper	NOUN
fcis-6914	37	27	proposes	propose	VERB
fcis-6914	37	28	a	a	DET
fcis-6914	37	29	yolov5	yolov5	NOUN
fcis-6914	37	30	face	face	NOUN
fcis-6914	37	31	helmet	helmet	NOUN
fcis-6914	37	32	detection	detection	NOUN
fcis-6914	37	33	algorithm	algorithm	NOUN
fcis-6914	37	34	based	base	VERB
fcis-6914	37	35	on	on	ADP
fcis-6914	37	36	the	the	DET
fcis-6914	37	37	improved	improved	ADJ
fcis-6914	37	38	swin	swin	PROPN
fcis-6914	37	39	transformer	transformer	NOUN
fcis-6914	37	40	from	from	ADP
fcis-6914	37	41	the	the	DET
fcis-6914	37	42	overall	overall	ADJ
fcis-6914	37	43	semantics	semantic	NOUN
fcis-6914	37	44	of	of	ADP
fcis-6914	37	45	images	image	NOUN
fcis-6914	37	46	and	and	CCONJ
fcis-6914	37	47	collects	collect	VERB
fcis-6914	37	48	3000	3000	NUM
fcis-6914	37	49	images	image	NOUN
fcis-6914	37	50	of	of	ADP
fcis-6914	37	51	faces	face	NOUN
fcis-6914	37	52	and	and	CCONJ
fcis-6914	37	53	helmets	helmet	NOUN
fcis-6914	37	54	in	in	ADP
fcis-6914	37	55	various	various	ADJ
fcis-6914	37	56	environments	environment	NOUN
fcis-6914	37	57	from	from	ADP
fcis-6914	37	58	the	the	DET
fcis-6914	37	59	web	web	NOUN
fcis-6914	37	60	to	to	PART
fcis-6914	37	61	build	build	VERB
fcis-6914	37	62	a	a	DET
fcis-6914	37	63	helmet	helmet	NOUN
fcis-6914	37	64	face	face	NOUN
fcis-6914	37	65	detection	detection	NOUN
fcis-6914	37	66	dataset	dataset	NOUN
fcis-6914	37	67	.	.	PUNCT
fcis-6914	38	1	while	while	SCONJ
fcis-6914	38	2	accurately	accurately	ADV
fcis-6914	38	3	distinguishing	distinguish	VERB
fcis-6914	38	4	helmetwearing	helmetwearing	NOUN
fcis-6914	38	5	and	and	CCONJ
fcis-6914	38	6	non	non	ADJ
fcis-6914	38	7	-	-	ADJ
fcis-6914	38	8	helmet	helmet	NOUN
fcis-6914	38	9	-	-	PUNCT
fcis-6914	38	10	wearing	wear	VERB
fcis-6914	38	11	people	people	NOUN
fcis-6914	38	12	,	,	PUNCT
fcis-6914	38	13	the	the	DET
fcis-6914	38	14	network	network	NOUN
fcis-6914	38	15	's	's	PART
fcis-6914	38	16	feature	feature	NOUN
fcis-6914	38	17	extraction	extraction	NOUN
fcis-6914	38	18	capability	capability	NOUN
fcis-6914	38	19	is	be	AUX
fcis-6914	38	20	enhanced	enhance	VERB
fcis-6914	38	21	by	by	ADP
fcis-6914	38	22	fusing	fuse	VERB
fcis-6914	38	23	multi	multi	ADJ
fcis-6914	38	24	-	-	ADJ
fcis-6914	38	25	scale	scale	ADJ
fcis-6914	38	26	information	information	NOUN
fcis-6914	38	27	based	base	VERB
fcis-6914	38	28	on	on	ADP
fcis-6914	38	29	transformer	transformer	ADJ
fcis-6914	38	30	block	block	NOUN
fcis-6914	38	31	and	and	CCONJ
fcis-6914	38	32	sliding	slide	VERB
fcis-6914	38	33	window	window	NOUN
fcis-6914	38	34	mechanism	mechanism	NOUN
fcis-6914	38	35	,	,	PUNCT
fcis-6914	38	36	and	and	CCONJ
fcis-6914	38	37	the	the	DET
fcis-6914	38	38	accuracy	accuracy	NOUN
fcis-6914	38	39	of	of	ADP
fcis-6914	38	40	face	face	NOUN
fcis-6914	38	41	helmet	helmet	NOUN
fcis-6914	38	42	detection	detection	NOUN
fcis-6914	38	43	is	be	AUX
fcis-6914	38	44	improved	improve	VERB
fcis-6914	38	45	.	.	PUNCT
fcis-6914	39	1	experiments	experiment	NOUN
fcis-6914	39	2	show	show	VERB
fcis-6914	39	3	that	that	SCONJ
fcis-6914	39	4	the	the	DET
fcis-6914	39	5	improved	improved	ADJ
fcis-6914	39	6	yolov5	yolov5	NOUN
fcis-6914	39	7	face	face	NOUN
fcis-6914	39	8	helmet	helmet	NOUN
fcis-6914	39	9	detection	detection	NOUN
fcis-6914	39	10	algorithm	algorithm	NOUN
fcis-6914	39	11	map	map	NOUN
fcis-6914	39	12	based	base	VERB
fcis-6914	39	13	on	on	ADP
fcis-6914	39	14	swin	swin	PROPN
fcis-6914	39	15	transformer	transformer	NOUN
fcis-6914	39	16	improves	improve	VERB
fcis-6914	39	17	5.7	5.7	NUM
fcis-6914	39	18	%	%	NOUN
fcis-6914	39	19	compared	compare	VERB
fcis-6914	39	20	to	to	ADP
fcis-6914	39	21	faster	fast	ADJ
fcis-6914	39	22	rcnn	rcnn	PROPN
fcis-6914	39	23	,	,	PUNCT
fcis-6914	39	24	6.1	6.1	NUM
fcis-6914	39	25	%	%	NOUN
fcis-6914	39	26	compared	compare	VERB
fcis-6914	39	27	to	to	ADP
fcis-6914	39	28	yolov3	yolov3	PROPN
fcis-6914	39	29	,	,	PUNCT
fcis-6914	39	30	5.3	5.3	NUM
fcis-6914	39	31	%	%	NOUN
fcis-6914	39	32	compared	compare	VERB
fcis-6914	39	33	to	to	ADP
fcis-6914	39	34	yolov4	yolov4	PROPN
fcis-6914	39	35	,	,	PUNCT
fcis-6914	39	36	and	and	CCONJ
fcis-6914	39	37	1.6	1.6	NUM
fcis-6914	39	38	%	%	NOUN
fcis-6914	39	39	compared	compare	VERB
fcis-6914	39	40	to	to	ADP
fcis-6914	39	41	the	the	DET
fcis-6914	39	42	original	original	ADJ
fcis-6914	39	43	algorithm	algorithm	NOUN
fcis-6914	39	44	.	.	PUNCT
fcis-6914	40	1	it	it	PRON
fcis-6914	40	2	performs	perform	VERB
fcis-6914	40	3	well	well	ADV
fcis-6914	40	4	in	in	ADP
fcis-6914	40	5	helmet	helmet	NOUN
fcis-6914	40	6	face	face	NOUN
fcis-6914	40	7	detection	detection	NOUN
fcis-6914	40	8	tasks	task	NOUN
fcis-6914	40	9	in	in	ADP
fcis-6914	40	10	complex	complex	ADJ
fcis-6914	40	11	environments	environment	NOUN
fcis-6914	40	12	and	and	CCONJ
fcis-6914	40	13	achieving	achieve	VERB
fcis-6914	40	14	real	real	ADJ
fcis-6914	40	15	-	-	PUNCT
fcis-6914	40	16	time	time	NOUN
fcis-6914	40	17	detection	detection	NOUN
fcis-6914	40	18	with	with	ADP
fcis-6914	40	19	higher	high	ADJ
fcis-6914	40	20	accuracy	accuracy	NOUN
fcis-6914	40	21	.	.	PUNCT
fcis-6914	41	1	2	2	X
fcis-6914	41	2	.	.	NUM
fcis-6914	41	3	improved	improved	ADJ
fcis-6914	41	4	yolov5	yolov5	NOUN
fcis-6914	41	5	algorithms	algorithm	NOUN
fcis-6914	41	6	figure	figure	VERB
fcis-6914	41	7	1	1	NUM
fcis-6914	41	8	.	.	PUNCT
fcis-6914	42	1	the	the	DET
fcis-6914	42	2	improved	improved	ADJ
fcis-6914	42	3	yolov5	yolov5	NOUN
fcis-6914	42	4	network	network	NOUN
fcis-6914	42	5	the	the	DET
fcis-6914	42	6	neck	neck	NOUN
fcis-6914	42	7	introduces	introduce	NOUN
fcis-6914	42	8	c3tr	c3tr	ADJ
fcis-6914	42	9	and	and	CCONJ
fcis-6914	42	10	c3str	c3str	NUM
fcis-6914	42	11	modules	module	NOUN
fcis-6914	42	12	and	and	CCONJ
fcis-6914	42	13	location	location	NOUN
fcis-6914	42	14	codes	code	NOUN
fcis-6914	42	15	into	into	ADP
fcis-6914	42	16	the	the	DET
fcis-6914	42	17	original	original	ADJ
fcis-6914	42	18	network	network	NOUN
fcis-6914	42	19	of	of	ADP
fcis-6914	42	20	yolov5	yolov5	NOUN
fcis-6914	42	21	,	,	PUNCT
fcis-6914	42	22	and	and	CCONJ
fcis-6914	42	23	replaces	replace	VERB
fcis-6914	42	24	the	the	DET
fcis-6914	42	25	bottleneck	bottleneck	NOUN
fcis-6914	42	26	in	in	ADP
fcis-6914	42	27	the	the	DET
fcis-6914	42	28	original	original	ADJ
fcis-6914	42	29	c3	c3	NOUN
fcis-6914	42	30	module	module	NOUN
fcis-6914	42	31	with	with	ADP
fcis-6914	42	32	transformer	transformer	NOUN
fcis-6914	42	33	block	block	NOUN
fcis-6914	42	34	and	and	CCONJ
fcis-6914	42	35	swin	swin	PROPN
fcis-6914	42	36	transformer	transformer	NOUN
fcis-6914	42	37	block	block	NOUN
fcis-6914	42	38	modules	module	NOUN
fcis-6914	42	39	.	.	PUNCT
fcis-6914	43	1	the	the	DET
fcis-6914	43	2	retained	retain	VERB
fcis-6914	43	3	c3	c3	NOUN
fcis-6914	43	4	module	module	NOUN
fcis-6914	43	5	aims	aim	VERB
fcis-6914	43	6	to	to	PART
fcis-6914	43	7	streamline	streamline	VERB
fcis-6914	43	8	the	the	DET
fcis-6914	43	9	network	network	NOUN
fcis-6914	43	10	structure	structure	NOUN
fcis-6914	43	11	of	of	ADP
fcis-6914	43	12	yolov5	yolov5	NOUN
fcis-6914	43	13	,	,	PUNCT
fcis-6914	43	14	reduce	reduce	VERB
fcis-6914	43	15	the	the	DET
fcis-6914	43	16	computation	computation	NOUN
fcis-6914	43	17	and	and	CCONJ
fcis-6914	43	18	reduce	reduce	VERB
fcis-6914	43	19	the	the	DET
fcis-6914	43	20	reasoning	reasoning	NOUN
fcis-6914	43	21	time	time	NOUN
fcis-6914	43	22	of	of	ADP
fcis-6914	43	23	the	the	DET
fcis-6914	43	24	model	model	NOUN
fcis-6914	43	25	.	.	PUNCT
fcis-6914	44	1	on	on	ADP
fcis-6914	44	2	this	this	DET
fcis-6914	44	3	basis	basis	NOUN
fcis-6914	44	4	,	,	PUNCT
fcis-6914	44	5	c3tr	c3tr	ADJ
fcis-6914	44	6	and	and	CCONJ
fcis-6914	44	7	c3str	c3str	PRON
fcis-6914	44	8	have	have	VERB
fcis-6914	44	9	stronger	strong	ADJ
fcis-6914	44	10	contextual	contextual	ADJ
fcis-6914	44	11	learning	learning	NOUN
fcis-6914	44	12	ability	ability	NOUN
fcis-6914	44	13	.	.	PUNCT
fcis-6914	45	1	considering	consider	VERB
fcis-6914	45	2	that	that	PRON
fcis-6914	45	3	c3tr	c3tr	ADJ
fcis-6914	45	4	has	have	VERB
fcis-6914	45	5	a	a	DET
fcis-6914	45	6	strong	strong	ADJ
fcis-6914	45	7	ability	ability	NOUN
fcis-6914	45	8	to	to	PART
fcis-6914	45	9	capture	capture	VERB
fcis-6914	45	10	global	global	ADJ
fcis-6914	45	11	information	information	NOUN
fcis-6914	45	12	,	,	PUNCT
fcis-6914	45	13	but	but	CCONJ
fcis-6914	45	14	the	the	DET
fcis-6914	45	15	number	number	NOUN
fcis-6914	45	16	of	of	ADP
fcis-6914	45	17	parameters	parameter	NOUN
fcis-6914	45	18	is	be	AUX
fcis-6914	45	19	large	large	ADJ
fcis-6914	45	20	,	,	PUNCT
fcis-6914	45	21	it	it	PRON
fcis-6914	45	22	needs	need	VERB
fcis-6914	45	23	to	to	PART
fcis-6914	45	24	calculate	calculate	VERB
fcis-6914	45	25	the	the	DET
fcis-6914	45	26	self	self	NOUN
fcis-6914	45	27	-	-	PUNCT
fcis-6914	45	28	attention	attention	NOUN
fcis-6914	45	29	of	of	ADP
fcis-6914	45	30	each	each	DET
fcis-6914	45	31	region	region	NOUN
fcis-6914	45	32	,	,	PUNCT
fcis-6914	45	33	and	and	CCONJ
fcis-6914	45	34	the	the	DET
fcis-6914	45	35	speed	speed	NOUN
fcis-6914	45	36	is	be	AUX
fcis-6914	45	37	slow	slow	ADJ
fcis-6914	45	38	,	,	PUNCT
fcis-6914	45	39	while	while	SCONJ
fcis-6914	45	40	c3str	c3str	PRON
fcis-6914	45	41	has	have	VERB
fcis-6914	45	42	a	a	DET
fcis-6914	45	43	relatively	relatively	ADV
fcis-6914	45	44	weak	weak	ADJ
fcis-6914	45	45	ability	ability	NOUN
fcis-6914	45	46	to	to	PART
fcis-6914	45	47	capture	capture	VERB
fcis-6914	45	48	global	global	ADJ
fcis-6914	45	49	information	information	NOUN
fcis-6914	45	50	,	,	PUNCT
fcis-6914	45	51	but	but	CCONJ
fcis-6914	45	52	fewer	few	ADJ
fcis-6914	45	53	parameters	parameter	NOUN
fcis-6914	45	54	than	than	ADP
fcis-6914	45	55	c3tr	c3tr	ADJ
fcis-6914	45	56	,	,	PUNCT
fcis-6914	45	57	window	window	NOUN
fcis-6914	45	58	migration	migration	NOUN
fcis-6914	45	59	can	can	AUX
fcis-6914	45	60	make	make	VERB
fcis-6914	45	61	up	up	ADP
fcis-6914	45	62	for	for	ADP
fcis-6914	45	63	the	the	DET
fcis-6914	45	64	limitations	limitation	NOUN
fcis-6914	45	65	of	of	ADP
fcis-6914	45	66	c3tr	c3tr	ADJ
fcis-6914	45	67	in	in	ADP
fcis-6914	45	68	reducing	reduce	VERB
fcis-6914	45	69	the	the	DET
fcis-6914	45	70	global	global	ADJ
fcis-6914	45	71	field	field	NOUN
fcis-6914	45	72	of	of	ADP
fcis-6914	45	73	sensitivity	sensitivity	NOUN
fcis-6914	45	74	and	and	CCONJ
fcis-6914	45	75	capturing	capture	VERB
fcis-6914	45	76	more	more	ADJ
fcis-6914	45	77	comprehensive	comprehensive	ADJ
fcis-6914	45	78	features	feature	NOUN
fcis-6914	45	79	.	.	PUNCT
fcis-6914	46	1	figure	figure	VERB
fcis-6914	46	2	8	8	NUM
fcis-6914	46	3	shows	show	VERB
fcis-6914	46	4	the	the	DET
fcis-6914	46	5	specific	specific	ADJ
fcis-6914	46	6	replacement	replacement	NOUN
fcis-6914	46	7	method	method	NOUN
fcis-6914	46	8	.	.	PUNCT
fcis-6914	47	1	the	the	DET
fcis-6914	47	2	performance	performance	NOUN
fcis-6914	47	3	of	of	ADP
fcis-6914	47	4	the	the	DET
fcis-6914	47	5	model	model	NOUN
fcis-6914	47	6	is	be	AUX
fcis-6914	47	7	effectively	effectively	ADV
fcis-6914	47	8	improved	improve	VERB
fcis-6914	47	9	in	in	ADP
fcis-6914	47	10	the	the	DET
fcis-6914	47	11	complex	complex	ADJ
fcis-6914	47	12	environment	environment	NOUN
fcis-6914	47	13	where	where	SCONJ
fcis-6914	47	14	the	the	DET
fcis-6914	47	15	target	target	NOUN
fcis-6914	47	16	helmet	helmet	NOUN
fcis-6914	47	17	occupies	occupy	VERB
fcis-6914	47	18	a	a	DET
fcis-6914	47	19	relatively	relatively	ADV
fcis-6914	47	20	small	small	ADJ
fcis-6914	47	21	part	part	NOUN
fcis-6914	47	22	in	in	ADP
fcis-6914	47	23	the	the	DET
fcis-6914	47	24	original	original	ADJ
fcis-6914	47	25	drawing	drawing	NOUN
fcis-6914	47	26	.	.	PUNCT
fcis-6914	48	1	3	3	X
fcis-6914	48	2	.	.	X
fcis-6914	48	3	experimental	experimental	ADJ
fcis-6914	48	4	results	result	NOUN
fcis-6914	48	5	and	and	CCONJ
fcis-6914	48	6	analysis	analysis	NOUN
fcis-6914	48	7	in	in	ADP
fcis-6914	48	8	order	order	NOUN
fcis-6914	48	9	to	to	PART
fcis-6914	48	10	verify	verify	VERB
fcis-6914	48	11	the	the	DET
fcis-6914	48	12	detection	detection	NOUN
fcis-6914	48	13	performance	performance	NOUN
fcis-6914	48	14	of	of	ADP
fcis-6914	48	15	the	the	DET
fcis-6914	48	16	yolov5	yolov5	NOUN
fcis-6914	48	17	algorithm	algorithm	NOUN
fcis-6914	48	18	combined	combine	VERB
fcis-6914	48	19	with	with	ADP
fcis-6914	48	20	transformer	transformer	NOUN
fcis-6914	48	21	,	,	PUNCT
fcis-6914	48	22	this	this	DET
fcis-6914	48	23	paper	paper	NOUN
fcis-6914	48	24	compares	compare	VERB
fcis-6914	48	25	it	it	PRON
fcis-6914	48	26	with	with	ADP
fcis-6914	48	27	the	the	DET
fcis-6914	48	28	mainstream	mainstream	NOUN
fcis-6914	48	29	target	target	NOUN
fcis-6914	48	30	detection	detection	NOUN
fcis-6914	48	31	algorithms	algorithm	NOUN
fcis-6914	48	32	that	that	PRON
fcis-6914	48	33	have	have	AUX
fcis-6914	48	34	emerged	emerge	VERB
fcis-6914	48	35	in	in	ADP
fcis-6914	48	36	recent	recent	ADJ
fcis-6914	48	37	years	year	NOUN
fcis-6914	48	38	in	in	ADP
fcis-6914	48	39	the	the	DET
fcis-6914	48	40	same	same	ADJ
fcis-6914	48	41	data	datum	NOUN
fcis-6914	48	42	set	set	VERB
fcis-6914	48	43	and	and	CCONJ
fcis-6914	48	44	environment	environment	NOUN
fcis-6914	48	45	.	.	PUNCT
fcis-6914	49	1	the	the	DET
fcis-6914	49	2	experimental	experimental	ADJ
fcis-6914	49	3	results	result	NOUN
fcis-6914	49	4	are	be	AUX
fcis-6914	49	5	shown	show	VERB
fcis-6914	49	6	in	in	ADP
fcis-6914	49	7	table	table	NOUN
fcis-6914	49	8	1	1	NUM
fcis-6914	49	9	since	since	SCONJ
fcis-6914	49	10	there	there	PRON
fcis-6914	49	11	are	be	VERB
fcis-6914	49	12	many	many	ADJ
fcis-6914	49	13	small	small	ADJ
fcis-6914	49	14	and	and	CCONJ
fcis-6914	49	15	medium	medium	ADJ
fcis-6914	49	16	-	-	PUNCT
fcis-6914	49	17	sized	sized	ADJ
fcis-6914	49	18	targets	target	NOUN
fcis-6914	49	19	in	in	ADP
fcis-6914	49	20	the	the	DET
fcis-6914	49	21	helmet	helmet	NOUN
fcis-6914	49	22	wearing	wear	VERB
fcis-6914	49	23	detection	detection	NOUN
fcis-6914	49	24	task	task	NOUN
fcis-6914	49	25	,	,	PUNCT
fcis-6914	49	26	faster	fast	ADJ
fcis-6914	49	27	r	r	NOUN
fcis-6914	49	28	-	-	PUNCT
fcis-6914	49	29	cnn	cnn	NOUN
fcis-6914	49	30	as	as	ADP
fcis-6914	49	31	the	the	DET
fcis-6914	49	32	classical	classical	ADJ
fcis-6914	49	33	representative	representative	NOUN
fcis-6914	49	34	of	of	ADP
fcis-6914	49	35	two	two	NUM
fcis-6914	49	36	-	-	PUNCT
fcis-6914	49	37	stage	stage	NOUN
fcis-6914	49	38	target	target	NOUN
fcis-6914	49	39	detection	detection	NOUN
fcis-6914	49	40	model	model	NOUN
fcis-6914	49	41	is	be	AUX
fcis-6914	49	42	not	not	PART
fcis-6914	49	43	conducive	conducive	ADJ
fcis-6914	49	44	to	to	ADP
fcis-6914	49	45	the	the	DET
fcis-6914	49	46	detection	detection	NOUN
fcis-6914	49	47	of	of	ADP
fcis-6914	49	48	small	small	ADJ
fcis-6914	49	49	and	and	CCONJ
fcis-6914	49	50	multi	multi	ADJ
fcis-6914	49	51	-	-	ADJ
fcis-6914	49	52	scale	scale	ADJ
fcis-6914	49	53	objects	object	NOUN
fcis-6914	49	54	because	because	SCONJ
fcis-6914	49	55	its	its	PRON
fcis-6914	49	56	feature	feature	NOUN
fcis-6914	49	57	map	map	NOUN
fcis-6914	49	58	is	be	AUX
fcis-6914	49	59	only	only	ADV
fcis-6914	49	60	a	a	DET
fcis-6914	49	61	single	single	ADJ
fcis-6914	49	62	layer	layer	NOUN
fcis-6914	49	63	and	and	CCONJ
fcis-6914	49	64	the	the	DET
fcis-6914	49	65	two	two	NUM
fcis-6914	49	66	stages	stage	NOUN
fcis-6914	49	67	of	of	ADP
fcis-6914	49	68	rpn	rpn	PROPN
fcis-6914	49	69	and	and	CCONJ
fcis-6914	49	70	rcnn	rcnn	PROPN
fcis-6914	49	71	are	be	AUX
fcis-6914	49	72	divided	divide	VERB
fcis-6914	49	73	,	,	PUNCT
fcis-6914	49	74	and	and	CCONJ
fcis-6914	49	75	it	it	PRON
fcis-6914	49	76	takes	take	VERB
fcis-6914	49	77	the	the	DET
fcis-6914	49	78	longest	long	ADJ
fcis-6914	49	79	time	time	NOUN
fcis-6914	49	80	with	with	ADP
fcis-6914	49	81	the	the	DET
fcis-6914	49	82	lowest	low	ADJ
fcis-6914	49	83	accuracy	accuracy	NOUN
fcis-6914	49	84	.	.	PUNCT
fcis-6914	50	1	yolov5	yolov5	NOUN
fcis-6914	50	2	has	have	AUX
fcis-6914	50	3	added	add	VERB
fcis-6914	50	4	more	more	ADJ
fcis-6914	50	5	tricks	trick	NOUN
fcis-6914	50	6	,	,	PUNCT
fcis-6914	50	7	and	and	CCONJ
fcis-6914	50	8	the	the	DET
fcis-6914	50	9	yolov5	yolov5	NOUN
fcis-6914	50	10	v6.0	v6.0	PROPN
fcis-6914	50	11	version	version	NOUN
fcis-6914	50	12	algorithm	algorithm	NOUN
fcis-6914	50	13	used	use	VERB
fcis-6914	50	14	in	in	ADP
fcis-6914	50	15	this	this	DET
fcis-6914	50	16	paper	paper	NOUN
fcis-6914	50	17	has	have	VERB
fcis-6914	50	18	higher	high	ADJ
fcis-6914	50	19	accuracy	accuracy	NOUN
fcis-6914	50	20	before	before	ADP
fcis-6914	50	21	improvement	improvement	NOUN
fcis-6914	50	22	,	,	PUNCT
fcis-6914	50	23	but	but	CCONJ
fcis-6914	50	24	the	the	DET
fcis-6914	50	25	recall	recall	NOUN
fcis-6914	50	26	is	be	AUX
fcis-6914	50	27	lower	low	ADJ
fcis-6914	50	28	,	,	PUNCT
fcis-6914	50	29	which	which	PRON
fcis-6914	50	30	tends	tend	VERB
fcis-6914	50	31	to	to	PART
fcis-6914	50	32	affect	affect	VERB
fcis-6914	50	33	the	the	DET
fcis-6914	50	34	correct	correct	ADJ
fcis-6914	50	35	recognition	recognition	NOUN
fcis-6914	50	36	of	of	ADP
fcis-6914	50	37	detection	detection	NOUN
fcis-6914	50	38	targets	target	NOUN
fcis-6914	50	39	in	in	ADP
fcis-6914	50	40	complex	complex	ADJ
fcis-6914	50	41	situations	situation	NOUN
fcis-6914	50	42	.	.	PUNCT
fcis-6914	51	1	combined	combine	VERB
fcis-6914	51	2	with	with	ADP
fcis-6914	51	3	the	the	DET
fcis-6914	51	4	transformer	transformer	ADJ
fcis-6914	51	5	structure	structure	NOUN
fcis-6914	51	6	of	of	ADP
fcis-6914	51	7	yolov5	yolov5	NOUN
fcis-6914	51	8	-	-	PUNCT
fcis-6914	51	9	tr	tr	VERB
fcis-6914	51	10	although	although	SCONJ
fcis-6914	51	11	precision	precision	NOUN
fcis-6914	51	12	slightly	slightly	ADV
fcis-6914	51	13	decreased	decrease	VERB
fcis-6914	51	14	,	,	PUNCT
fcis-6914	51	15	but	but	CCONJ
fcis-6914	51	16	recall	recall	NOUN
fcis-6914	51	17	and	and	CCONJ
fcis-6914	51	18	map@0.5	map@0.5	PROPN
fcis-6914	51	19	increased	increase	VERB
fcis-6914	51	20	by	by	ADP
fcis-6914	51	21	4.9	4.9	NUM
fcis-6914	51	22	%	%	NOUN
fcis-6914	51	23	and	and	CCONJ
fcis-6914	51	24	1.6	1.6	NUM
fcis-6914	51	25	%	%	NOUN
fcis-6914	51	26	,	,	PUNCT
fcis-6914	51	27	respectively	respectively	ADV
fcis-6914	51	28	.	.	PUNCT
fcis-6914	52	1	in	in	ADP
fcis-6914	52	2	meeting	meet	VERB
fcis-6914	52	3	the	the	DET
fcis-6914	52	4	requirements	requirement	NOUN
fcis-6914	52	5	of	of	ADP
fcis-6914	52	6	realtime	realtime	ADJ
fcis-6914	52	7	detection	detection	NOUN
fcis-6914	52	8	to	to	PART
fcis-6914	52	9	obtain	obtain	VERB
fcis-6914	52	10	better	well	ADJ
fcis-6914	52	11	detection	detection	NOUN
fcis-6914	52	12	results	result	NOUN
fcis-6914	52	13	.	.	PUNCT
fcis-6914	53	1	table	table	NOUN
fcis-6914	53	2	1	1	NUM
fcis-6914	53	3	.	.	PUNCT
fcis-6914	54	1	three	three	NUM
fcis-6914	54	2	scheme	scheme	NOUN
fcis-6914	54	3	comparing	compare	VERB
fcis-6914	54	4	model	model	NOUN
fcis-6914	54	5	p	p	NOUN
fcis-6914	54	6	r	r	NOUN
fcis-6914	54	7	map@.5	map@.5	NOUN
fcis-6914	54	8	faster	fast	ADV
fcis-6914	54	9	r	r	VERB
fcis-6914	54	10	-	-	PUNCT
fcis-6914	54	11	cnn	cnn	PROPN
fcis-6914	54	12	0.940	0.940	NUM
fcis-6914	54	13	0.891	0.891	NUM
fcis-6914	54	14	0.925	0.925	NUM
fcis-6914	54	15	yolov3	yolov3	NOUN
fcis-6914	54	16	0.967	0.967	NUM
fcis-6914	54	17	0.868	0.868	NUM
fcis-6914	54	18	0.921	0.921	NUM
fcis-6914	54	19	yolov4	yolov4	NOUN
fcis-6914	54	20	0.944	0.944	NUM
fcis-6914	54	21	0.880	0.880	NUM
fcis-6914	54	22	0.929	0.929	NUM
fcis-6914	54	23	yolov5	yolov5	NOUN
fcis-6914	54	24	0.976	0.976	NUM
fcis-6914	54	25	0.929	0.929	NUM
fcis-6914	54	26	0.966	0.966	NUM
fcis-6914	54	27	yolov5	yolov5	NOUN
fcis-6914	54	28	-	-	PUNCT
fcis-6914	54	29	tr	tr	VERB
fcis-6914	54	30	0.968	0.968	NUM
fcis-6914	54	31	0.978	0.978	NUM
fcis-6914	54	32	0.982	0.982	NUM
fcis-6914	54	33	since	since	SCONJ
fcis-6914	54	34	there	there	PRON
fcis-6914	54	35	are	be	VERB
fcis-6914	54	36	many	many	ADJ
fcis-6914	54	37	small	small	ADJ
fcis-6914	54	38	and	and	CCONJ
fcis-6914	54	39	medium	medium	ADJ
fcis-6914	54	40	-	-	PUNCT
fcis-6914	54	41	sized	sized	ADJ
fcis-6914	54	42	targets	target	NOUN
fcis-6914	54	43	in	in	ADP
fcis-6914	54	44	the	the	DET
fcis-6914	54	45	helmet	helmet	NOUN
fcis-6914	54	46	wearing	wear	VERB
fcis-6914	54	47	detection	detection	NOUN
fcis-6914	54	48	task	task	NOUN
fcis-6914	54	49	,	,	PUNCT
fcis-6914	54	50	faster	fast	ADJ
fcis-6914	54	51	r	r	NOUN
fcis-6914	54	52	-	-	PUNCT
fcis-6914	54	53	cnn	cnn	NOUN
fcis-6914	54	54	as	as	ADP
fcis-6914	54	55	the	the	DET
fcis-6914	54	56	classical	classical	ADJ
fcis-6914	54	57	representative	representative	NOUN
fcis-6914	54	58	of	of	ADP
fcis-6914	54	59	two	two	NUM
fcis-6914	54	60	-	-	PUNCT
fcis-6914	54	61	stage	stage	NOUN
fcis-6914	54	62	target	target	NOUN
fcis-6914	54	63	detection	detection	NOUN
fcis-6914	54	64	model	model	NOUN
fcis-6914	54	65	is	be	AUX
fcis-6914	54	66	not	not	PART
fcis-6914	54	67	conducive	conducive	ADJ
fcis-6914	54	68	to	to	ADP
fcis-6914	54	69	the	the	DET
fcis-6914	54	70	detection	detection	NOUN
fcis-6914	54	71	of	of	ADP
fcis-6914	54	72	small	small	ADJ
fcis-6914	54	73	and	and	CCONJ
fcis-6914	54	74	multi	multi	ADJ
fcis-6914	54	75	-	-	ADJ
fcis-6914	54	76	scale	scale	ADJ
fcis-6914	54	77	objects	object	NOUN
fcis-6914	54	78	because	because	SCONJ
fcis-6914	54	79	its	its	PRON
fcis-6914	54	80	feature	feature	NOUN
fcis-6914	54	81	map	map	NOUN
fcis-6914	54	82	is	be	AUX
fcis-6914	54	83	only	only	ADV
fcis-6914	54	84	a	a	DET
fcis-6914	54	85	single	single	ADJ
fcis-6914	54	86	layer	layer	NOUN
fcis-6914	54	87	and	and	CCONJ
fcis-6914	54	88	the	the	DET
fcis-6914	54	89	two	two	NUM
fcis-6914	54	90	stages	stage	NOUN
fcis-6914	54	91	of	of	ADP
fcis-6914	54	92	rpn	rpn	PROPN
fcis-6914	54	93	and	and	CCONJ
fcis-6914	54	94	rcnn	rcnn	PROPN
fcis-6914	54	95	are	be	AUX
fcis-6914	54	96	divided	divide	VERB
fcis-6914	54	97	,	,	PUNCT
fcis-6914	54	98	and	and	CCONJ
fcis-6914	54	99	it	it	PRON
fcis-6914	54	100	takes	take	VERB
fcis-6914	54	101	the	the	DET
fcis-6914	54	102	longest	long	ADJ
fcis-6914	54	103	time	time	NOUN
fcis-6914	54	104	with	with	ADP
fcis-6914	54	105	the	the	DET
fcis-6914	54	106	lowest	low	ADJ
fcis-6914	54	107	accuracy	accuracy	NOUN
fcis-6914	54	108	.	.	PUNCT
fcis-6914	55	1	yolov5	yolov5	NOUN
fcis-6914	55	2	has	have	AUX
fcis-6914	55	3	added	add	VERB
fcis-6914	55	4	more	more	ADJ
fcis-6914	55	5	tricks	trick	NOUN
fcis-6914	55	6	,	,	PUNCT
fcis-6914	55	7	and	and	CCONJ
fcis-6914	55	8	the	the	DET
fcis-6914	55	9	yolov5	yolov5	NOUN
fcis-6914	55	10	v6.0	v6.0	PROPN
fcis-6914	55	11	version	version	NOUN
fcis-6914	55	12	algorithm	algorithm	NOUN
fcis-6914	55	13	used	use	VERB
fcis-6914	55	14	in	in	ADP
fcis-6914	55	15	this	this	DET
fcis-6914	55	16	paper	paper	NOUN
fcis-6914	55	17	has	have	VERB
fcis-6914	55	18	higher	high	ADJ
fcis-6914	55	19	accuracy	accuracy	NOUN
fcis-6914	55	20	before	before	ADP
fcis-6914	55	21	improvement	improvement	NOUN
fcis-6914	55	22	,	,	PUNCT
fcis-6914	55	23	but	but	CCONJ
fcis-6914	55	24	the	the	DET
fcis-6914	55	25	recall	recall	NOUN
fcis-6914	55	26	is	be	AUX
fcis-6914	55	27	lower	low	ADJ
fcis-6914	55	28	,	,	PUNCT
fcis-6914	55	29	which	which	PRON
fcis-6914	55	30	tends	tend	VERB
fcis-6914	55	31	to	to	PART
fcis-6914	55	32	affect	affect	VERB
fcis-6914	55	33	the	the	DET
fcis-6914	55	34	correct	correct	ADJ
fcis-6914	55	35	recognition	recognition	NOUN
fcis-6914	55	36	of	of	ADP
fcis-6914	55	37	detection	detection	NOUN
fcis-6914	55	38	targets	target	NOUN
fcis-6914	55	39	in	in	ADP
fcis-6914	55	40	complex	complex	ADJ
fcis-6914	55	41	situations	situation	NOUN
fcis-6914	55	42	.	.	PUNCT
fcis-6914	56	1	combined	combine	VERB
fcis-6914	56	2	with	with	ADP
fcis-6914	56	3	the	the	DET
fcis-6914	56	4	transformer	transformer	ADJ
fcis-6914	56	5	structure	structure	NOUN
fcis-6914	56	6	of	of	ADP
fcis-6914	56	7	yolov5	yolov5	NOUN
fcis-6914	56	8	-	-	PUNCT
fcis-6914	56	9	tr	tr	VERB
fcis-6914	56	10	although	although	SCONJ
fcis-6914	56	11	precision	precision	NOUN
fcis-6914	56	12	slightly	slightly	ADV
fcis-6914	56	13	decreased	decrease	VERB
fcis-6914	56	14	,	,	PUNCT
fcis-6914	56	15	but	but	CCONJ
fcis-6914	56	16	recalland	recalland	PROPN
fcis-6914	56	17	map@0.5	map@0.5	PROPN
fcis-6914	56	18	increased	increase	VERB
fcis-6914	56	19	by	by	ADP
fcis-6914	56	20	4.9	4.9	NUM
fcis-6914	56	21	%	%	NOUN
fcis-6914	56	22	and	and	CCONJ
fcis-6914	56	23	1.6	1.6	NUM
fcis-6914	56	24	%	%	NOUN
fcis-6914	56	25	,	,	PUNCT
fcis-6914	56	26	respectively	respectively	ADV
fcis-6914	56	27	.	.	PUNCT
fcis-6914	57	1	in	in	ADP
fcis-6914	57	2	meeting	meet	VERB
fcis-6914	57	3	the	the	DET
fcis-6914	57	4	requirements	requirement	NOUN
fcis-6914	57	5	of	of	ADP
fcis-6914	57	6	realtime	realtime	ADJ
fcis-6914	57	7	detection	detection	NOUN
fcis-6914	57	8	to	to	PART
fcis-6914	57	9	obtain	obtain	VERB
fcis-6914	57	10	better	well	ADJ
fcis-6914	57	11	detection	detection	NOUN
fcis-6914	57	12	results	result	NOUN
fcis-6914	57	13	.	.	PUNCT
fcis-6914	58	1	27	27	NUM
fcis-6914	58	2	4	4	NUM
fcis-6914	58	3	.	.	PUNCT
fcis-6914	58	4	conclusion	conclusion	NOUN
fcis-6914	58	5	in	in	ADP
fcis-6914	58	6	this	this	DET
fcis-6914	58	7	paper	paper	NOUN
fcis-6914	58	8	,	,	PUNCT
fcis-6914	58	9	we	we	PRON
fcis-6914	58	10	propose	propose	VERB
fcis-6914	58	11	the	the	DET
fcis-6914	58	12	yolov5	yolov5	NOUN
fcis-6914	58	13	-	-	PUNCT
fcis-6914	58	14	tr	tr	VERB
fcis-6914	58	15	algorithm	algorithm	NOUN
fcis-6914	58	16	combined	combine	VERB
fcis-6914	58	17	with	with	ADP
fcis-6914	58	18	transformer	transformer	NOUN
fcis-6914	58	19	module	module	NOUN
fcis-6914	58	20	on	on	ADP
fcis-6914	58	21	the	the	DET
fcis-6914	58	22	basis	basis	NOUN
fcis-6914	58	23	of	of	ADP
fcis-6914	58	24	yolov5	yolov5	NOUN
fcis-6914	58	25	,	,	PUNCT
fcis-6914	58	26	and	and	CCONJ
fcis-6914	58	27	conduct	conduct	VERB
fcis-6914	58	28	experiments	experiment	NOUN
fcis-6914	58	29	using	use	VERB
fcis-6914	58	30	the	the	DET
fcis-6914	58	31	home	home	NOUN
fcis-6914	58	32	-	-	PUNCT
fcis-6914	58	33	built	build	VERB
fcis-6914	58	34	dataset	dataset	NOUN
fcis-6914	58	35	to	to	PART
fcis-6914	58	36	increase	increase	VERB
fcis-6914	58	37	the	the	DET
fcis-6914	58	38	generalization	generalization	NOUN
fcis-6914	58	39	and	and	CCONJ
fcis-6914	58	40	robustness	robustness	NOUN
fcis-6914	58	41	of	of	ADP
fcis-6914	58	42	the	the	DET
fcis-6914	58	43	model	model	NOUN
fcis-6914	58	44	through	through	ADP
fcis-6914	58	45	mosaic	mosaic	ADJ
fcis-6914	58	46	data	datum	NOUN
fcis-6914	58	47	enhancement	enhancement	NOUN
fcis-6914	58	48	,	,	PUNCT
fcis-6914	58	49	label	label	NOUN
fcis-6914	58	50	smoothing	smooth	VERB
fcis-6914	58	51	processing	processing	NOUN
fcis-6914	58	52	,	,	PUNCT
fcis-6914	58	53	adaptive	adaptive	ADJ
fcis-6914	58	54	weighting	weighting	NOUN
fcis-6914	58	55	features	feature	NOUN
fcis-6914	58	56	combined	combine	VERB
fcis-6914	58	57	with	with	ADP
fcis-6914	58	58	wconcat	wconcat	PROPN
fcis-6914	58	59	module	module	NOUN
fcis-6914	58	60	and	and	CCONJ
fcis-6914	58	61	the	the	DET
fcis-6914	58	62	application	application	NOUN
fcis-6914	58	63	of	of	ADP
fcis-6914	58	64	c3tr	c3tr	ADJ
fcis-6914	58	65	and	and	CCONJ
fcis-6914	58	66	c3str.experiments	c3str.experiment	NOUN
fcis-6914	58	67	show	show	VERB
fcis-6914	58	68	that	that	SCONJ
fcis-6914	58	69	the	the	DET
fcis-6914	58	70	improved	improved	ADJ
fcis-6914	58	71	yolov5	yolov5	NOUN
fcis-6914	58	72	-	-	PUNCT
fcis-6914	58	73	tr	tr	VERB
fcis-6914	58	74	has	have	VERB
fcis-6914	58	75	high	high	ADJ
fcis-6914	58	76	accuracy	accuracy	NOUN
fcis-6914	58	77	for	for	ADP
fcis-6914	58	78	multiple	multiple	ADJ
fcis-6914	58	79	occlusion	occlusion	NOUN
fcis-6914	58	80	detection	detection	NOUN
fcis-6914	58	81	of	of	ADP
fcis-6914	58	82	small	small	ADJ
fcis-6914	58	83	and	and	CCONJ
fcis-6914	58	84	medium	medium	ADJ
fcis-6914	58	85	-	-	PUNCT
fcis-6914	58	86	sized	sized	ADJ
fcis-6914	58	87	targets	target	NOUN
fcis-6914	58	88	in	in	ADP
fcis-6914	58	89	complex	complex	ADJ
fcis-6914	58	90	situations	situation	NOUN
fcis-6914	58	91	,	,	PUNCT
fcis-6914	58	92	and	and	CCONJ
fcis-6914	58	93	reduces	reduce	VERB
fcis-6914	58	94	the	the	DET
fcis-6914	58	95	leakage	leakage	NOUN
fcis-6914	58	96	detection	detection	NOUN
fcis-6914	58	97	while	while	SCONJ
fcis-6914	58	98	satisfying	satisfy	VERB
fcis-6914	58	99	real	real	ADJ
fcis-6914	58	100	-	-	PUNCT
fcis-6914	58	101	time	time	NOUN
fcis-6914	58	102	detection	detection	NOUN
fcis-6914	58	103	.	.	PUNCT
fcis-6914	59	1	transformer	transformer	NOUN
fcis-6914	59	2	undoubtedly	undoubtedly	ADV
fcis-6914	59	3	has	have	VERB
fcis-6914	59	4	superior	superior	ADJ
fcis-6914	59	5	performance	performance	NOUN
fcis-6914	59	6	and	and	CCONJ
fcis-6914	59	7	powerful	powerful	ADJ
fcis-6914	59	8	potential	potential	NOUN
fcis-6914	59	9	compared	compare	VERB
fcis-6914	59	10	with	with	ADP
fcis-6914	59	11	convolutional	convolutional	ADJ
fcis-6914	59	12	neural	neural	ADJ
fcis-6914	59	13	networks	network	NOUN
fcis-6914	59	14	.	.	PUNCT
fcis-6914	60	1	the	the	DET
fcis-6914	60	2	limitation	limitation	NOUN
fcis-6914	60	3	of	of	ADP
fcis-6914	60	4	the	the	DET
fcis-6914	60	5	improvement	improvement	NOUN
fcis-6914	60	6	based	base	VERB
fcis-6914	60	7	on	on	ADP
fcis-6914	60	8	transformer	transformer	NOUN
fcis-6914	60	9	in	in	ADP
fcis-6914	60	10	this	this	DET
fcis-6914	60	11	paper	paper	NOUN
fcis-6914	60	12	is	be	AUX
fcis-6914	60	13	that	that	SCONJ
fcis-6914	60	14	the	the	DET
fcis-6914	60	15	accuracy	accuracy	NOUN
fcis-6914	60	16	is	be	AUX
fcis-6914	60	17	higher	high	ADJ
fcis-6914	60	18	while	while	SCONJ
fcis-6914	60	19	the	the	DET
fcis-6914	60	20	model	model	NOUN
fcis-6914	60	21	is	be	AUX
fcis-6914	60	22	theoretically	theoretically	ADV
fcis-6914	60	23	more	more	ADV
fcis-6914	60	24	computationally	computationally	ADV
fcis-6914	60	25	intensive	intensive	ADJ
fcis-6914	60	26	,	,	PUNCT
fcis-6914	60	27	and	and	CCONJ
fcis-6914	60	28	the	the	DET
fcis-6914	60	29	hardware	hardware	NOUN
fcis-6914	60	30	deployment	deployment	NOUN
fcis-6914	60	31	conditions	condition	NOUN
fcis-6914	60	32	are	be	AUX
fcis-6914	60	33	more	more	ADV
fcis-6914	60	34	demanding	demanding	ADJ
fcis-6914	60	35	.	.	PUNCT
fcis-6914	61	1	in	in	ADP
fcis-6914	61	2	future	future	ADJ
fcis-6914	61	3	work	work	NOUN
fcis-6914	61	4	,	,	PUNCT
fcis-6914	61	5	it	it	PRON
fcis-6914	61	6	is	be	AUX
fcis-6914	61	7	expected	expect	VERB
fcis-6914	61	8	to	to	PART
fcis-6914	61	9	discuss	discuss	VERB
fcis-6914	61	10	the	the	DET
fcis-6914	61	11	impact	impact	NOUN
fcis-6914	61	12	of	of	ADP
fcis-6914	61	13	modules	module	NOUN
fcis-6914	61	14	in	in	ADP
fcis-6914	61	15	different	different	ADJ
fcis-6914	61	16	locations	location	NOUN
fcis-6914	61	17	on	on	ADP
fcis-6914	61	18	the	the	DET
fcis-6914	61	19	detection	detection	NOUN
fcis-6914	61	20	results	result	NOUN
fcis-6914	61	21	and	and	CCONJ
fcis-6914	61	22	computational	computational	ADJ
fcis-6914	61	23	effort	effort	NOUN
fcis-6914	61	24	,	,	PUNCT
fcis-6914	61	25	and	and	CCONJ
fcis-6914	61	26	to	to	PART
fcis-6914	61	27	establish	establish	VERB
fcis-6914	61	28	an	an	DET
fcis-6914	61	29	identification	identification	NOUN
fcis-6914	61	30	system	system	NOUN
fcis-6914	61	31	to	to	PART
fcis-6914	61	32	determine	determine	VERB
fcis-6914	61	33	the	the	DET
fcis-6914	61	34	information	information	NOUN
fcis-6914	61	35	related	relate	VERB
fcis-6914	61	36	to	to	ADP
fcis-6914	61	37	people	people	NOUN
fcis-6914	61	38	not	not	PART
fcis-6914	61	39	wearing	wear	VERB
fcis-6914	61	40	helmets	helmet	NOUN
fcis-6914	61	41	.	.	PUNCT
fcis-6914	62	1	it	it	PRON
fcis-6914	62	2	is	be	AUX
fcis-6914	62	3	hoped	hope	VERB
fcis-6914	62	4	that	that	SCONJ
fcis-6914	62	5	better	well	ADJ
fcis-6914	62	6	algorithms	algorithm	NOUN
fcis-6914	62	7	can	can	AUX
fcis-6914	62	8	be	be	AUX
fcis-6914	62	9	borrowed	borrow	VERB
fcis-6914	62	10	for	for	ADP
fcis-6914	62	11	future	future	ADJ
fcis-6914	62	12	experiments	experiment	NOUN
fcis-6914	62	13	and	and	CCONJ
fcis-6914	62	14	exploration	exploration	NOUN
fcis-6914	62	15	.	.	PUNCT
fcis-6914	63	1	references	reference	NOUN
fcis-6914	63	2	[	[	X
fcis-6914	63	3	1	1	X
fcis-6914	63	4	]	]	X
fcis-6914	63	5	shi	shi	PROPN
fcis-6914	63	6	hui	hui	PROPN
fcis-6914	63	7	.	.	PUNCT
fcis-6914	64	1	research	research	NOUN
fcis-6914	64	2	on	on	ADP
fcis-6914	64	3	deep	deep	ADJ
fcis-6914	64	4	learning	learning	NOUN
fcis-6914	64	5	based	base	VERB
fcis-6914	64	6	construction	construction	NOUN
fcis-6914	64	7	site	site	NOUN
fcis-6914	64	8	helmet	helmet	NOUN
fcis-6914	64	9	wearing	wear	VERB
fcis-6914	64	10	detection	detection	NOUN
fcis-6914	64	11	algorithm	algorithm	NOUN
fcis-6914	65	1	[	[	X
fcis-6914	65	2	d	d	X
fcis-6914	65	3	]	]	X
fcis-6914	65	4	.	.	PUNCT
fcis-6914	66	1	wuhan	wuhan	PROPN
fcis-6914	66	2	institute	institute	PROPN
fcis-6914	66	3	of	of	ADP
fcis-6914	66	4	technology,2019.doi:10.27381	technology,2019.doi:10.27381	PROPN
fcis-6914	66	5	/	/	SYM
fcis-6914	66	6	d.cnki.gwlgu.2019.001638	d.cnki.gwlgu.2019.001638	NOUN
fcis-6914	66	7	.	.	PUNCT
fcis-6914	67	1	[	[	X
fcis-6914	67	2	2	2	NUM
fcis-6914	67	3	]	]	PUNCT
fcis-6914	67	4	yu	yu	PROPN
fcis-6914	67	5	bo	bo	PROPN
fcis-6914	67	6	.	.	PUNCT
fcis-6914	67	7	safety	safety	NOUN
fcis-6914	67	8	helmet	helmet	NOUN
fcis-6914	67	9	detection	detection	NOUN
fcis-6914	67	10	based	base	VERB
fcis-6914	67	11	on	on	ADP
fcis-6914	67	12	intelligent	intelligent	ADJ
fcis-6914	67	13	video	video	NOUN
fcis-6914	67	14	surveillance	surveillance	NOUN
fcis-6914	68	1	[	[	X
fcis-6914	68	2	d	d	X
fcis-6914	68	3	]	]	X
fcis-6914	68	4	.	.	PUNCT
fcis-6914	69	1	hebei	hebei	PROPN
fcis-6914	69	2	university	university	PROPN
fcis-6914	69	3	of	of	ADP
fcis-6914	69	4	technology	technology	NOUN
fcis-6914	69	5	,	,	PUNCT
fcis-6914	69	6	2011	2011	NUM
fcis-6914	69	7	.	.	PUNCT
fcis-6914	70	1	[	[	X
fcis-6914	70	2	3	3	X
fcis-6914	70	3	]	]	PUNCT
fcis-6914	70	4	gkioxari	gkioxari	NOUN
fcis-6914	70	5	g	g	PROPN
fcis-6914	70	6	,	,	PUNCT
fcis-6914	70	7	hariharan	hariharan	PROPN
fcis-6914	70	8	b	b	PROPN
fcis-6914	70	9	,	,	PUNCT
fcis-6914	70	10	girshick	girshick	ADJ
fcis-6914	70	11	r	r	NOUN
fcis-6914	70	12	,	,	PUNCT
fcis-6914	70	13	et	et	PROPN
fcis-6914	70	14	al	al	PROPN
fcis-6914	70	15	.	.	PUNCT
fcis-6914	71	1	r	r	X
fcis-6914	71	2	-	-	PUNCT
fcis-6914	71	3	cnns	cnns	PROPN
fcis-6914	71	4	for	for	ADP
fcis-6914	71	5	pose	pose	NOUN
fcis-6914	71	6	estimation	estimation	NOUN
fcis-6914	71	7	and	and	CCONJ
fcis-6914	71	8	action	action	NOUN
fcis-6914	71	9	detection[j	detection[j	PROPN
fcis-6914	71	10	]	]	PUNCT
fcis-6914	71	11	.	.	PUNCT
fcis-6914	72	1	computer	computer	NOUN
fcis-6914	72	2	ence	ence	PROPN
fcis-6914	72	3	,	,	PUNCT
fcis-6914	72	4	2014.j.-m	2014.j.-m	PROPN
fcis-6914	72	5	.	.	PUNCT
fcis-6914	72	6	chang	chang	PROPN
fcis-6914	72	7	,	,	PUNCT
fcis-6914	72	8	w.-t	w.-t	PROPN
fcis-6914	72	9	.	.	PUNCT
fcis-6914	73	1	hsiao	hsiao	PROPN
fcis-6914	73	2	,	,	PUNCT
fcis-6914	73	3	j.-l	j.-l	PROPN
fcis-6914	73	4	.	.	PUNCT
fcis-6914	74	1	chen	chen	PROPN
fcis-6914	74	2	,	,	PUNCT
fcis-6914	74	3	h.-c	h.-c	PROPN
fcis-6914	74	4	.	.	PUNCT
fcis-6914	75	1	chao	chao	PROPN
fcis-6914	75	2	,	,	PUNCT
fcis-6914	75	3	mobile	mobile	ADJ
fcis-6914	75	4	relay	relay	NOUN
fcis-6914	75	5	stations	station	VERB
fcis-6914	75	6	navigation	navigation	NOUN
fcis-6914	75	7	-	-	PUNCT
fcis-6914	75	8	based	base	VERB
fcis-6914	75	9	self	self	NOUN
fcis-6914	75	10	-	-	PUNCT
fcis-6914	75	11	optimization	optimization	NOUN
fcis-6914	75	12	handover	handover	NOUN
fcis-6914	75	13	mechanism	mechanism	NOUN
fcis-6914	75	14	in	in	ADP
fcis-6914	75	15	wimax	wimax	PROPN
fcis-6914	75	16	networks	network	NOUN
fcis-6914	75	17	,	,	PUNCT
fcis-6914	75	18	in	in	ADP
fcis-6914	75	19	:	:	PUNCT
fcis-6914	75	20	proc	proc	NOUN
fcis-6914	75	21	.	.	PUNCT
fcis-6914	76	1	2009	2009	NUM
fcis-6914	76	2	international	international	ADJ
fcis-6914	76	3	conference	conference	NOUN
fcis-6914	76	4	on	on	ADP
fcis-6914	76	5	ubiquitous	ubiquitous	ADJ
fcis-6914	76	6	information	information	NOUN
fcis-6914	76	7	technologies	technology	NOUN
fcis-6914	76	8	&	&	CCONJ
fcis-6914	76	9	applications	application	NOUN
fcis-6914	76	10	,	,	PUNCT
fcis-6914	76	11	2009.b	2009.b	PROPN
fcis-6914	76	12	.	.	PUNCT
fcis-6914	76	13	smith	smith	PROPN
fcis-6914	76	14	,	,	PUNCT
fcis-6914	76	15	“	"	PUNCT
fcis-6914	76	16	an	an	DET
fcis-6914	76	17	approach	approach	NOUN
fcis-6914	76	18	to	to	ADP
fcis-6914	76	19	graphs	graph	NOUN
fcis-6914	76	20	of	of	ADP
fcis-6914	76	21	linear	linear	ADJ
fcis-6914	76	22	forms	form	NOUN
fcis-6914	76	23	(	(	PUNCT
fcis-6914	76	24	unpublished	unpublished	ADJ
fcis-6914	76	25	work	work	NOUN
fcis-6914	76	26	style	style	NOUN
fcis-6914	76	27	)	)	PUNCT
fcis-6914	76	28	,	,	PUNCT
fcis-6914	76	29	”	"	PUNCT
fcis-6914	76	30	unpublished	unpublished	ADJ
fcis-6914	76	31	.	.	PUNCT
fcis-6914	77	1	[	[	X
fcis-6914	77	2	4	4	X
fcis-6914	77	3	]	]	X
fcis-6914	77	4	girshick	girshick	PROPN
fcis-6914	77	5	r.	r.	PROPN
fcis-6914	77	6	fast	fast	VERB
fcis-6914	77	7	r	r	PROPN
fcis-6914	77	8	-	-	PUNCT
fcis-6914	77	9	cnn[c]//	cnn[c]//	ADJ
fcis-6914	77	10	international	international	ADJ
fcis-6914	77	11	conference	conference	NOUN
fcis-6914	77	12	on	on	ADP
fcis-6914	77	13	computer	computer	NOUN
fcis-6914	77	14	vision	vision	NOUN
fcis-6914	77	15	.	.	PUNCT
fcis-6914	78	1	ieee	ieee	NOUN
fcis-6914	78	2	computer	computer	NOUN
fcis-6914	78	3	society	society	NOUN
fcis-6914	78	4	,	,	PUNCT
fcis-6914	78	5	2015	2015	NUM
fcis-6914	78	6	.	.	PUNCT
fcis-6914	79	1	j.g	j.g	PROPN
fcis-6914	79	2	.	.	PROPN
fcis-6914	79	3	wilson	wilson	PROPN
fcis-6914	79	4	,	,	PUNCT
fcis-6914	79	5	f.c	f.c	PROPN
fcis-6914	79	6	.	.	PROPN
fcis-6914	79	7	fraser	fraser	PROPN
fcis-6914	79	8	(	(	PUNCT
fcis-6914	79	9	eds	eds	PROPN
fcis-6914	79	10	.	.	PUNCT
fcis-6914	79	11	)	)	PUNCT
fcis-6914	79	12	,	,	PUNCT
fcis-6914	79	13	handbook	handbook	NOUN
fcis-6914	79	14	of	of	ADP
fcis-6914	79	15	teratology	teratology	NOUN
fcis-6914	79	16	,	,	PUNCT
fcis-6914	79	17	vols	vol	NOUN
fcis-6914	79	18	.	.	PUNCT
fcis-6914	80	1	1	1	NUM
fcis-6914	80	2	-	-	SYM
fcis-6914	80	3	4	4	NUM
fcis-6914	80	4	,	,	PUNCT
fcis-6914	80	5	plenum	plenum	PROPN
fcis-6914	80	6	press	press	PROPN
fcis-6914	80	7	,	,	PUNCT
fcis-6914	80	8	new	new	PROPN
fcis-6914	80	9	york	york	PROPN
fcis-6914	80	10	,	,	PUNCT
fcis-6914	80	11	1977	1977	NUM
fcis-6914	80	12	-	-	SYM
fcis-6914	80	13	1978	1978	NUM
fcis-6914	80	14	.	.	PUNCT
fcis-6914	81	1	[	[	X
fcis-6914	81	2	5	5	X
fcis-6914	81	3	]	]	X
fcis-6914	81	4	ren	ren	PROPN
fcis-6914	81	5	s	s	PART
fcis-6914	81	6	,	,	PUNCT
fcis-6914	81	7	he	he	PRON
fcis-6914	81	8	k	k	X
fcis-6914	81	9	,	,	PUNCT
fcis-6914	81	10	girshick	girshick	PROPN
fcis-6914	81	11	r	r	NOUN
fcis-6914	81	12	,	,	PUNCT
fcis-6914	81	13	et	et	PROPN
fcis-6914	81	14	al	al	PROPN
fcis-6914	81	15	.	.	PUNCT
fcis-6914	82	1	faster	fast	ADJ
fcis-6914	82	2	r	r	NOUN
fcis-6914	82	3	-	-	PUNCT
fcis-6914	82	4	cnn	cnn	NOUN
fcis-6914	82	5	:	:	PUNCT
fcis-6914	82	6	towards	towards	ADP
fcis-6914	82	7	real	real	ADJ
fcis-6914	82	8	-	-	PUNCT
fcis-6914	82	9	time	time	NOUN
fcis-6914	82	10	object	object	NOUN
fcis-6914	82	11	detection	detection	NOUN
fcis-6914	82	12	with	with	ADP
fcis-6914	82	13	region	region	NOUN
fcis-6914	82	14	proposal	proposal	NOUN
fcis-6914	82	15	networks[c]//	networks[c]//	ADJ
fcis-6914	82	16	nips	nip	NOUN
fcis-6914	82	17	.	.	PUNCT
fcis-6914	83	1	2016	2016	NUM
fcis-6914	83	2	.	.	PUNCT
fcis-6914	84	1	w.	w.	PROPN
fcis-6914	84	2	strunk	strunk	PROPN
fcis-6914	84	3	jr	jr	PROPN
fcis-6914	84	4	.	.	PROPN
fcis-6914	84	5	,	,	PUNCT
fcis-6914	84	6	e.b	e.b	PROPN
fcis-6914	84	7	.	.	PROPN
fcis-6914	84	8	white	white	PROPN
fcis-6914	84	9	,	,	PUNCT
fcis-6914	84	10	the	the	DET
fcis-6914	84	11	elements	element	NOUN
fcis-6914	84	12	of	of	ADP
fcis-6914	84	13	style	style	NOUN
fcis-6914	84	14	,	,	PUNCT
fcis-6914	84	15	third	third	ADJ
fcis-6914	84	16	ed	ed	NOUN
fcis-6914	84	17	.	.	PROPN
fcis-6914	84	18	,	,	PUNCT
fcis-6914	84	19	macmillan	macmillan	PROPN
fcis-6914	84	20	,	,	PUNCT
fcis-6914	84	21	new	new	PROPN
fcis-6914	84	22	york	york	PROPN
fcis-6914	84	23	,	,	PUNCT
fcis-6914	84	24	1979	1979	NUM
fcis-6914	84	25	(	(	PUNCT
fcis-6914	84	26	chapter	chapter	NOUN
fcis-6914	84	27	4	4	NUM
fcis-6914	84	28	)	)	PUNCT
fcis-6914	84	29	.	.	PUNCT
fcis-6914	85	1	[	[	X
fcis-6914	85	2	6	6	NUM
fcis-6914	85	3	]	]	X
fcis-6914	85	4	redmon	redmon	PROPN
fcis-6914	85	5	j	j	PROPN
fcis-6914	85	6	,	,	PUNCT
fcis-6914	85	7	farhadi	farhadi	PROPN
fcis-6914	85	8	a	a	PRON
fcis-6914	85	9	.	.	PUNCT
fcis-6914	86	1	yolo9000	yolo9000	PROPN
fcis-6914	86	2	:	:	PUNCT
fcis-6914	86	3	better	well	ADJ
fcis-6914	86	4	,	,	PUNCT
fcis-6914	86	5	faster	fast	ADV
fcis-6914	86	6	,	,	PUNCT
fcis-6914	86	7	stronger[c]//	stronger[c]//	PROPN
fcis-6914	86	8	ieee	ieee	PROPN
fcis-6914	86	9	conference	conference	NOUN
fcis-6914	86	10	on	on	ADP
fcis-6914	86	11	computer	computer	NOUN
fcis-6914	86	12	vision	vision	NOUN
fcis-6914	86	13	&	&	CCONJ
fcis-6914	86	14	pattern	pattern	NOUN
fcis-6914	86	15	recognition	recognition	NOUN
fcis-6914	86	16	.	.	PUNCT
fcis-6914	87	1	ieee	ieee	PROPN
fcis-6914	87	2	,	,	PUNCT
fcis-6914	87	3	2017:6517	2017:6517	NUM
fcis-6914	87	4	-	-	SYM
fcis-6914	87	5	6525	6525	NUM
fcis-6914	87	6	.	.	PUNCT
fcis-6914	88	1	cancer	cancer	NOUN
fcis-6914	88	2	research	research	PROPN
fcis-6914	88	3	uk	uk	PROPN
fcis-6914	88	4	,	,	PUNCT
fcis-6914	88	5	cancer	cancer	NOUN
fcis-6914	88	6	statistics	statistic	NOUN
fcis-6914	88	7	reports	report	VERB
fcis-6914	88	8	for	for	ADP
fcis-6914	88	9	the	the	DET
fcis-6914	88	10	uk2003	uk2003	PROPN
fcis-6914	88	11	(	(	PUNCT
fcis-6914	88	12	accessed	access	VERB
fcis-6914	88	13	13.03.03	13.03.03	ADJ
fcis-6914	88	14	)	)	PUNCT
fcis-6914	88	15	.	.	PUNCT
fcis-6914	89	1	http://www.cancerresearchuk.org/aboutcancer/statistics/cance	http://www.cancerresearchuk.org/aboutcancer/statistics/cance	NOUN
fcis-6914	89	2	rstatsreport/	rstatsreport/	NUM
fcis-6914	90	1	[	[	X
fcis-6914	90	2	7	7	X
fcis-6914	90	3	]	]	X
fcis-6914	90	4	redmon	redmon	PROPN
fcis-6914	90	5	j	j	PROPN
fcis-6914	90	6	,	,	PUNCT
fcis-6914	90	7	farhadi	farhadi	PROPN
fcis-6914	90	8	a.yolov3	a.yolov3	PROPN
fcis-6914	90	9	:	:	PUNCT
fcis-6914	90	10	an	an	DET
fcis-6914	90	11	incremental	incremental	ADJ
fcis-6914	90	12	improvement[j	improvement[j	NOUN
fcis-6914	90	13	]	]	PUNCT
fcis-6914	90	14	.	.	PUNCT
fcis-6914	91	1	arxiv	arxiv	PROPN
fcis-6914	91	2	e	e	PROPN
fcis-6914	91	3	-	-	NOUN
fcis-6914	91	4	prints	print	NOUN
fcis-6914	91	5	,	,	PUNCT
fcis-6914	91	6	2018.m	2018.m	NUM
fcis-6914	91	7	.	.	PUNCT
fcis-6914	91	8	young	young	ADJ
fcis-6914	91	9	,	,	PUNCT
fcis-6914	91	10	the	the	DET
fcis-6914	91	11	techincal	techincal	ADJ
fcis-6914	91	12	writers	writer	NOUN
fcis-6914	91	13	handbook	handbook	NOUN
fcis-6914	91	14	.	.	PUNCT
fcis-6914	92	1	mill	mill	NOUN
fcis-6914	92	2	valley	valley	PROPN
fcis-6914	92	3	,	,	PUNCT
fcis-6914	92	4	ca	ca	PROPN
fcis-6914	92	5	:	:	PUNCT
fcis-6914	92	6	university	university	NOUN
fcis-6914	92	7	science	science	NOUN
fcis-6914	92	8	,	,	PUNCT
fcis-6914	92	9	1989	1989	NUM
fcis-6914	92	10	.	.	PUNCT
fcis-6914	93	1	[	[	X
fcis-6914	93	2	8	8	NUM
fcis-6914	93	3	]	]	X
fcis-6914	93	4	bochkovskiy	bochkovskiy	X
fcis-6914	93	5	a	a	PRON
fcis-6914	93	6	,	,	PUNCT
fcis-6914	93	7	wang	wang	PROPN
fcis-6914	93	8	c	c	PROPN
fcis-6914	93	9	y	y	PROPN
fcis-6914	93	10	,	,	PUNCT
fcis-6914	93	11	liao	liao	PROPN
fcis-6914	93	12	h	h	PROPN
fcis-6914	93	13	.	.	PUNCT
fcis-6914	94	1	yolov4	yolov4	NOUN
fcis-6914	94	2	:	:	PUNCT
fcis-6914	94	3	optimal	optimal	ADJ
fcis-6914	94	4	speed	speed	NOUN
fcis-6914	94	5	and	and	CCONJ
fcis-6914	94	6	accuracy	accuracy	NOUN
fcis-6914	94	7	of	of	ADP
fcis-6914	94	8	object	object	NOUN
fcis-6914	94	9	detection[j	detection[j	PROPN
fcis-6914	94	10	]	]	PUNCT
fcis-6914	94	11	.	.	PUNCT
fcis-6914	95	1	2020	2020	NUM
fcis-6914	95	2	.	.	PUNCT
fcis-6914	96	1	[	[	X
fcis-6914	96	2	9	9	NUM
fcis-6914	96	3	]	]	X
fcis-6914	96	4	wang	wang	PROPN
fcis-6914	96	5	j	j	PROPN
fcis-6914	96	6	,	,	PUNCT
fcis-6914	96	7	chen	chen	PROPN
fcis-6914	96	8	y	y	PROPN
fcis-6914	96	9	,	,	PUNCT
fcis-6914	96	10	gao	gao	PROPN
fcis-6914	96	11	m	m	PROPN
fcis-6914	96	12	,	,	PUNCT
fcis-6914	96	13	et	et	PROPN
fcis-6914	96	14	al	al	PROPN
fcis-6914	96	15	.	.	PROPN
fcis-6914	96	16	improved	improve	VERB
fcis-6914	96	17	yolov5	yolov5	NOUN
fcis-6914	96	18	network	network	NOUN
fcis-6914	96	19	for	for	ADP
fcis-6914	96	20	real	real	ADJ
fcis-6914	96	21	-	-	PUNCT
fcis-6914	96	22	time	time	NOUN
fcis-6914	96	23	multi	multi	ADJ
fcis-6914	96	24	-	-	ADJ
fcis-6914	96	25	scale	scale	ADJ
fcis-6914	96	26	traffic	traffic	NOUN
fcis-6914	96	27	sign	sign	NOUN
fcis-6914	96	28	detection[j	detection[j	PROPN
fcis-6914	96	29	]	]	PUNCT
fcis-6914	96	30	.	.	PUNCT
fcis-6914	96	31	2021	2021	NUM
fcis-6914	96	32	.	.	PUNCT
fcis-6914	97	1	[	[	X
fcis-6914	97	2	10	10	NUM
fcis-6914	97	3	]	]	X
fcis-6914	97	4	ding	de	VERB
fcis-6914	97	5	w	w	PROPN
fcis-6914	97	6	-	-	PUNCT
fcis-6914	97	7	l	l	NOUN
fcis-6914	97	8	,	,	PUNCT
fcis-6914	97	9	fei	fei	PROPN
fcis-6914	97	10	sh	sh	PROPN
fcis-6914	97	11	-	-	PUNCT
fcis-6914	97	12	min	min	PROPN
fcis-6914	97	13	.	.	PROPN
fcis-6914	97	14	research	research	NOUN
fcis-6914	97	15	on	on	ADP
fcis-6914	97	16	helmet	helmet	NOUN
fcis-6914	97	17	detection	detection	NOUN
fcis-6914	97	18	method	method	NOUN
fcis-6914	97	19	based	base	VERB
fcis-6914	97	20	on	on	ADP
fcis-6914	97	21	improved	improved	ADJ
fcis-6914	97	22	yolov3	yolov3	PROPN
fcis-6914	98	1	[	[	X
fcis-6914	98	2	j	j	X
fcis-6914	98	3	]	]	X
fcis-6914	98	4	.	.	PUNCT
fcis-6914	99	1	electronic	electronic	ADJ
fcis-6914	99	2	testing	testing	NOUN
fcis-6914	99	3	,	,	PUNCT
fcis-6914	99	4	2022	2022	NUM
fcis-6914	99	5	.	.	PUNCT
fcis-6914	100	1	[	[	X
fcis-6914	100	2	11	11	NUM
fcis-6914	100	3	]	]	PUNCT
fcis-6914	100	4	jin	jin	PROPN
fcis-6914	100	5	yufang	yufang	PROPN
fcis-6914	100	6	,	,	PUNCT
fcis-6914	100	7	wu	wu	PROPN
fcis-6914	100	8	xiang	xiang	PROPN
fcis-6914	100	9	,	,	PUNCT
fcis-6914	100	10	dong	dong	PROPN
fcis-6914	100	11	hui	hui	PROPN
fcis-6914	100	12	,	,	PUNCT
fcis-6914	100	13	et	et	PROPN
fcis-6914	100	14	al	al	PROPN
fcis-6914	100	15	.	.	PROPN
fcis-6914	100	16	improved	improve	VERB
fcis-6914	100	17	helmet	helmet	NOUN
fcis-6914	100	18	wearing	wear	VERB
fcis-6914	100	19	detection	detection	NOUN
fcis-6914	100	20	algorithm	algorithm	NOUN
fcis-6914	100	21	based	base	VERB
fcis-6914	100	22	on	on	ADP
fcis-6914	100	23	yolo	yolo	PROPN
fcis-6914	100	24	v4	v4	PROPN
fcis-6914	101	1	[	[	X
fcis-6914	101	2	j	j	X
fcis-6914	101	3	]	]	X
fcis-6914	101	4	.	.	PUNCT
fcis-6914	102	1	computer	computer	NOUN
fcis-6914	102	2	science	science	NOUN
fcis-6914	102	3	,	,	PUNCT
fcis-6914	102	4	2021	2021	NUM
fcis-6914	102	5	.	.	PUNCT
fcis-6914	103	1	[	[	X
fcis-6914	103	2	12	12	NUM
fcis-6914	103	3	]	]	X
fcis-6914	103	4	yue	yue	PROPN
fcis-6914	103	5	h	h	PROPN
fcis-6914	103	6	,	,	PUNCT
fcis-6914	103	7	huang	huang	PROPN
fcis-6914	103	8	h	h	PROPN
fcis-6914	103	9	,	,	PUNCT
fcis-6914	103	10	lin	lin	PROPN
fcis-6914	103	11	mh	mh	PROPN
fcis-6914	103	12	,	,	PUNCT
fcis-6914	103	13	gao	gao	PROPN
fcis-6914	103	14	m	m	PROPN
fcis-6914	103	15	,	,	PUNCT
fcis-6914	103	16	li	li	PROPN
fcis-6914	103	17	y	y	PROPN
fcis-6914	103	18	,	,	PUNCT
fcis-6914	103	19	chen	chen	PROPN
fcis-6914	103	20	l.	l.	PROPN
fcis-6914	103	21	safety	safety	PROPN
fcis-6914	103	22	helmet	helmet	NOUN
fcis-6914	103	23	wearing	wear	VERB
fcis-6914	103	24	detection	detection	NOUN
fcis-6914	103	25	based	base	VERB
fcis-6914	103	26	on	on	ADP
fcis-6914	103	27	improved	improved	ADJ
fcis-6914	103	28	yolov5	yolov5	NOUN
fcis-6914	104	1	[	[	X
fcis-6914	104	2	j	j	X
fcis-6914	104	3	]	]	X
fcis-6914	104	4	.	.	PUNCT
fcis-6914	105	1	computers	computer	NOUN
fcis-6914	105	2	and	and	CCONJ
fcis-6914	105	3	modernization,2022(06):104	modernization,2022(06):104	PROPN
fcis-6914	105	4	-	-	PUNCT
fcis-6914	105	5	108	108	NUM
fcis-6914	105	6	+	+	NOUN
fcis-6914	105	7	126	126	NUM
fcis-6914	105	8	.	.	PUNCT
fcis-6914	106	1	[	[	X
fcis-6914	106	2	13	13	NUM
fcis-6914	106	3	]	]	X
fcis-6914	106	4	qin	qin	X
fcis-6914	106	5	zh	zh	PROPN
fcis-6914	106	6	,	,	PUNCT
fcis-6914	106	7	lei	lei	PROPN
fcis-6914	106	8	m	m	PROPN
fcis-6914	106	9	,	,	PUNCT
fcis-6914	106	10	song	song	PROPN
fcis-6914	106	11	wg	wg	PROPN
fcis-6914	106	12	,	,	PUNCT
fcis-6914	106	13	zhang	zhang	PROPN
fcis-6914	106	14	w.	w.	PROPN
fcis-6914	106	15	safety	safety	PROPN
fcis-6914	106	16	helmet	helmet	NOUN
fcis-6914	106	17	detection	detection	NOUN
fcis-6914	106	18	method	method	NOUN
fcis-6914	106	19	based	base	VERB
fcis-6914	106	20	on	on	ADP
fcis-6914	106	21	lightweight	lightweight	ADJ
fcis-6914	106	22	deep	deep	ADJ
fcis-6914	106	23	learning	learning	NOUN
fcis-6914	106	24	model[j	model[j	PROPN
fcis-6914	106	25	]	]	PUNCT
fcis-6914	106	26	.	.	PUNCT
fcis-6914	107	1	science	science	NOUN
fcis-6914	107	2	technology	technology	NOUN
fcis-6914	107	3	and	and	CCONJ
fcis-6914	107	4	engineering,2022,22(14):5659	engineering,2022,22(14):5659	NOUN
fcis-6914	107	5	-	-	SYM
fcis-6914	107	6	5665	5665	NUM
fcis-6914	107	7	.	.	PUNCT
fcis-6914	108	1	[	[	X
fcis-6914	108	2	14	14	NUM
fcis-6914	108	3	]	]	X
fcis-6914	108	4	vaswani	vaswani	NOUN
fcis-6914	108	5	a	a	PRON
fcis-6914	108	6	,	,	PUNCT
fcis-6914	108	7	shazeer	shazeer	NOUN
fcis-6914	108	8	n	n	NOUN
fcis-6914	108	9	,	,	PUNCT
fcis-6914	108	10	parmar	parmar	PROPN
fcis-6914	108	11	n	n	PROPN
fcis-6914	108	12	,	,	PUNCT
fcis-6914	108	13	et	et	PROPN
fcis-6914	108	14	al	al	PROPN
fcis-6914	108	15	.	.	PUNCT
fcis-6914	108	16	attention	attention	NOUN
fcis-6914	108	17	is	be	AUX
fcis-6914	108	18	all	all	PRON
fcis-6914	108	19	you	you	PRON
fcis-6914	108	20	need[c]//	need[c]//	PROPN
fcis-6914	108	21	arxiv	arxiv	PROPN
fcis-6914	108	22	.	.	PUNCT
fcis-6914	109	1	arxiv	arxiv	PROPN
fcis-6914	109	2	,	,	PUNCT
fcis-6914	109	3	2017	2017	NUM
fcis-6914	109	4	.	.	PUNCT
fcis-6914	110	1	[	[	X
fcis-6914	110	2	15	15	NUM
fcis-6914	110	3	]	]	X
fcis-6914	110	4	liu	liu	PROPN
fcis-6914	110	5	z	z	PROPN
fcis-6914	110	6	,	,	PUNCT
fcis-6914	110	7	lin	lin	PROPN
fcis-6914	110	8	y	y	PROPN
fcis-6914	110	9	,	,	PUNCT
fcis-6914	110	10	cao	cao	PROPN
fcis-6914	110	11	y	y	PROPN
fcis-6914	110	12	,	,	PUNCT
fcis-6914	110	13	et	et	PROPN
fcis-6914	110	14	al	al	PROPN
fcis-6914	110	15	.	.	PROPN
fcis-6914	110	16	swin	swin	PROPN
fcis-6914	110	17	transformer	transformer	PROPN
fcis-6914	110	18	:	:	PUNCT
fcis-6914	110	19	hierarchical	hierarchical	ADJ
fcis-6914	110	20	vision	vision	NOUN
fcis-6914	110	21	transformer	transformer	NOUN
fcis-6914	110	22	using	use	VERB
fcis-6914	110	23	shifted	shift	VERB
fcis-6914	110	24	windows[j	windows[j	PROPN
fcis-6914	110	25	]	]	PUNCT
fcis-6914	110	26	.	.	PUNCT
fcis-6914	110	27	2021	2021	NUM
fcis-6914	110	28	.	.	PUNCT