id	sid	tid	token	lemma	pos
fcis-9730	1	1	frontiers	frontier	NOUN
fcis-9730	1	2	in	in	ADP
fcis-9730	1	3	computing	computing	NOUN
fcis-9730	1	4	and	and	CCONJ
fcis-9730	1	5	intelligent	intelligent	ADJ
fcis-9730	1	6	systems	system	NOUN
fcis-9730	1	7	issn	issn	VERB
fcis-9730	1	8	:	:	PUNCT
fcis-9730	1	9	2832	2832	NUM
fcis-9730	1	10	-	-	SYM
fcis-9730	1	11	6024	6024	NUM
fcis-9730	1	12	|	|	NOUN
fcis-9730	1	13	vol	vol	NOUN
fcis-9730	1	14	.	.	PROPN
fcis-9730	2	1	4	4	NUM
fcis-9730	2	2	,	,	PUNCT
fcis-9730	2	3	no	no	INTJ
fcis-9730	2	4	.	.	NOUN
fcis-9730	2	5	2	2	NUM
fcis-9730	2	6	,	,	PUNCT
fcis-9730	2	7	2023	2023	NUM
fcis-9730	2	8	17	17	NUM
fcis-9730	2	9	a	a	DET
fcis-9730	2	10	review	review	NOUN
fcis-9730	2	11	of	of	ADP
fcis-9730	2	12	yolo	yolo	ADJ
fcis-9730	2	13	object	object	NOUN
fcis-9730	2	14	detection	detection	NOUN
fcis-9730	2	15	algorithms	algorithm	NOUN
fcis-9730	2	16	based	base	VERB
fcis-9730	2	17	on	on	ADP
fcis-9730	2	18	deep	deep	ADJ
fcis-9730	2	19	learning	learning	NOUN
fcis-9730	2	20	xiaohan	xiaohan	PROPN
fcis-9730	2	21	cong	cong	PROPN
fcis-9730	2	22	*	*	PUNCT
fcis-9730	2	23	,	,	PUNCT
fcis-9730	2	24	shixin	shixin	PROPN
fcis-9730	2	25	li	li	PROPN
fcis-9730	2	26	,	,	PUNCT
fcis-9730	2	27	fankai	fankai	PROPN
fcis-9730	2	28	chen	chen	PROPN
fcis-9730	2	29	,	,	PUNCT
fcis-9730	2	30	chen	chen	PROPN
fcis-9730	2	31	liu	liu	PROPN
fcis-9730	2	32	,	,	PUNCT
fcis-9730	2	33	yue	yue	PROPN
fcis-9730	2	34	meng	meng	PROPN
fcis-9730	2	35	college	college	PROPN
fcis-9730	2	36	of	of	ADP
fcis-9730	2	37	electronic	electronic	ADJ
fcis-9730	2	38	engineering	engineering	NOUN
fcis-9730	2	39	,	,	PUNCT
fcis-9730	2	40	tianjin	tianjin	PROPN
fcis-9730	2	41	university	university	PROPN
fcis-9730	2	42	of	of	ADP
fcis-9730	2	43	technology	technology	NOUN
fcis-9730	2	44	and	and	CCONJ
fcis-9730	2	45	education	education	NOUN
fcis-9730	2	46	,	,	PUNCT
fcis-9730	2	47	tianjin	tianjin	PROPN
fcis-9730	2	48	300222	300222	NUM
fcis-9730	2	49	,	,	PUNCT
fcis-9730	2	50	china	china	PROPN
fcis-9730	2	51	*	*	PUNCT
fcis-9730	2	52	corresponding	correspond	VERB
fcis-9730	2	53	author	author	NOUN
fcis-9730	2	54	:	:	PUNCT
fcis-9730	2	55	xiaohan	xiaohan	PROPN
fcis-9730	2	56	cong	cong	PROPN
fcis-9730	2	57	(	(	PUNCT
fcis-9730	2	58	email	email	NOUN
fcis-9730	2	59	:	:	PUNCT
fcis-9730	2	60	1136094608@qq.com	1136094608@qq.com	NUM
fcis-9730	2	61	)	)	PUNCT
fcis-9730	3	1	abstract	abstract	NOUN
fcis-9730	3	2	:	:	PUNCT
fcis-9730	3	3	object	object	NOUN
fcis-9730	3	4	detection	detection	NOUN
fcis-9730	3	5	is	be	AUX
fcis-9730	3	6	a	a	DET
fcis-9730	3	7	research	research	NOUN
fcis-9730	3	8	hotspot	hotspot	NOUN
fcis-9730	3	9	in	in	ADP
fcis-9730	3	10	the	the	DET
fcis-9730	3	11	field	field	NOUN
fcis-9730	3	12	of	of	ADP
fcis-9730	3	13	computer	computer	NOUN
fcis-9730	3	14	vision	vision	NOUN
fcis-9730	3	15	,	,	PUNCT
fcis-9730	3	16	and	and	CCONJ
fcis-9730	3	17	yolo	yolo	PROPN
fcis-9730	3	18	series	series	PROPN
fcis-9730	3	19	shows	show	VERB
fcis-9730	3	20	good	good	ADJ
fcis-9730	3	21	performance	performance	NOUN
fcis-9730	3	22	in	in	ADP
fcis-9730	3	23	object	object	NOUN
fcis-9730	3	24	detection	detection	NOUN
fcis-9730	3	25	,	,	PUNCT
fcis-9730	3	26	and	and	CCONJ
fcis-9730	3	27	has	have	AUX
fcis-9730	3	28	been	be	AUX
fcis-9730	3	29	widely	widely	ADV
fcis-9730	3	30	used	use	VERB
fcis-9730	3	31	in	in	ADP
fcis-9730	3	32	robot	robot	NOUN
fcis-9730	3	33	vision	vision	NOUN
fcis-9730	3	34	,	,	PUNCT
fcis-9730	3	35	unmanned	unmanned	ADJ
fcis-9730	3	36	driving	driving	NOUN
fcis-9730	3	37	and	and	CCONJ
fcis-9730	3	38	other	other	ADJ
fcis-9730	3	39	fields	field	NOUN
fcis-9730	3	40	in	in	ADP
fcis-9730	3	41	recent	recent	ADJ
fcis-9730	3	42	years	year	NOUN
fcis-9730	3	43	.	.	PUNCT
fcis-9730	4	1	this	this	DET
fcis-9730	4	2	paper	paper	NOUN
fcis-9730	4	3	first	first	ADV
fcis-9730	4	4	introduces	introduce	VERB
fcis-9730	4	5	the	the	DET
fcis-9730	4	6	yolo	yolo	ADJ
fcis-9730	4	7	series	series	PROPN
fcis-9730	4	8	algorithm	algorithm	PROPN
fcis-9730	4	9	,	,	PUNCT
fcis-9730	4	10	including	include	VERB
fcis-9730	4	11	the	the	DET
fcis-9730	4	12	principle	principle	NOUN
fcis-9730	4	13	,	,	PUNCT
fcis-9730	4	14	innovation	innovation	NOUN
fcis-9730	4	15	points	point	NOUN
fcis-9730	4	16	,	,	PUNCT
fcis-9730	4	17	advantages	advantage	NOUN
fcis-9730	4	18	and	and	CCONJ
fcis-9730	4	19	disadvantages	disadvantage	NOUN
fcis-9730	4	20	of	of	ADP
fcis-9730	4	21	various	various	ADJ
fcis-9730	4	22	algorithms	algorithm	NOUN
fcis-9730	4	23	,	,	PUNCT
fcis-9730	4	24	then	then	ADV
fcis-9730	4	25	introduces	introduce	VERB
fcis-9730	4	26	the	the	DET
fcis-9730	4	27	application	application	NOUN
fcis-9730	4	28	field	field	NOUN
fcis-9730	4	29	of	of	ADP
fcis-9730	4	30	yolo	yolo	ADJ
fcis-9730	4	31	series	series	NOUN
fcis-9730	4	32	,	,	PUNCT
fcis-9730	4	33	and	and	CCONJ
fcis-9730	4	34	finally	finally	ADV
fcis-9730	4	35	analyzes	analyze	VERB
fcis-9730	4	36	its	its	PRON
fcis-9730	4	37	future	future	ADJ
fcis-9730	4	38	development	development	NOUN
fcis-9730	4	39	trend	trend	NOUN
fcis-9730	4	40	to	to	PART
fcis-9730	4	41	provide	provide	VERB
fcis-9730	4	42	reference	reference	NOUN
fcis-9730	4	43	for	for	ADP
fcis-9730	4	44	the	the	DET
fcis-9730	4	45	topic	topic	NOUN
fcis-9730	4	46	research	research	NOUN
fcis-9730	4	47	.	.	PUNCT
fcis-9730	5	1	keywords	keyword	NOUN
fcis-9730	5	2	:	:	PUNCT
fcis-9730	5	3	object	object	VERB
fcis-9730	5	4	detection	detection	NOUN
fcis-9730	5	5	;	;	PUNCT
fcis-9730	5	6	yolo	yolo	ADJ
fcis-9730	5	7	;	;	PUNCT
fcis-9730	5	8	deep	deep	ADJ
fcis-9730	5	9	learning	learning	NOUN
fcis-9730	5	10	.	.	PUNCT
fcis-9730	6	1	1	1	X
fcis-9730	6	2	.	.	X
fcis-9730	6	3	introduction	introduction	NOUN
fcis-9730	6	4	object	object	NOUN
fcis-9730	6	5	detection	detection	NOUN
fcis-9730	6	6	is	be	AUX
fcis-9730	6	7	a	a	DET
fcis-9730	6	8	kind	kind	NOUN
fcis-9730	6	9	of	of	ADP
fcis-9730	6	10	image	image	NOUN
fcis-9730	6	11	segmentation	segmentation	NOUN
fcis-9730	6	12	based	base	VERB
fcis-9730	6	13	on	on	ADP
fcis-9730	6	14	object	object	NOUN
fcis-9730	6	15	geometric	geometric	ADJ
fcis-9730	6	16	and	and	CCONJ
fcis-9730	6	17	statistical	statistical	ADJ
fcis-9730	6	18	features	feature	NOUN
fcis-9730	6	19	,	,	PUNCT
fcis-9730	6	20	also	also	ADV
fcis-9730	6	21	known	know	VERB
fcis-9730	6	22	as	as	ADP
fcis-9730	6	23	object	object	NOUN
fcis-9730	6	24	extraction	extraction	NOUN
fcis-9730	6	25	.	.	PUNCT
fcis-9730	7	1	with	with	ADP
fcis-9730	7	2	the	the	DET
fcis-9730	7	3	development	development	NOUN
fcis-9730	7	4	of	of	ADP
fcis-9730	7	5	computer	computer	NOUN
fcis-9730	7	6	and	and	CCONJ
fcis-9730	7	7	object	object	VERB
fcis-9730	7	8	detection	detection	NOUN
fcis-9730	7	9	theory	theory	NOUN
fcis-9730	7	10	,	,	PUNCT
fcis-9730	7	11	the	the	DET
fcis-9730	7	12	object	object	NOUN
fcis-9730	7	13	detection	detection	NOUN
fcis-9730	7	14	field	field	NOUN
fcis-9730	7	15	of	of	ADP
fcis-9730	7	16	deep	deep	ADJ
fcis-9730	7	17	learning	learning	NOUN
fcis-9730	7	18	should	should	AUX
fcis-9730	7	19	be	be	AUX
fcis-9730	7	20	in	in	ADP
fcis-9730	7	21	intelligent	intelligent	ADJ
fcis-9730	7	22	security	security	NOUN
fcis-9730	7	23	,	,	PUNCT
fcis-9730	7	24	security	security	NOUN
fcis-9730	7	25	monitoring	monitoring	NOUN
fcis-9730	7	26	,	,	PUNCT
fcis-9730	7	27	intelligent	intelligent	ADJ
fcis-9730	7	28	medical	medical	ADJ
fcis-9730	7	29	treatment	treatment	NOUN
fcis-9730	7	30	,	,	PUNCT
fcis-9730	7	31	traffic	traffic	NOUN
fcis-9730	7	32	detection	detection	NOUN
fcis-9730	7	33	,	,	PUNCT
fcis-9730	7	34	robot	robot	NOUN
fcis-9730	7	35	vision	vision	NOUN
fcis-9730	7	36	,	,	PUNCT
fcis-9730	7	37	unmanned	unmanned	ADJ
fcis-9730	7	38	driving	driving	NOUN
fcis-9730	7	39	and	and	CCONJ
fcis-9730	7	40	other	other	ADJ
fcis-9730	7	41	fields	field	NOUN
fcis-9730	7	42	.	.	PUNCT
fcis-9730	8	1	[	[	X
fcis-9730	8	2	1]-[3	1]-[3	X
fcis-9730	8	3	]	]	PUNCT
fcis-9730	8	4	based	base	VERB
fcis-9730	8	5	on	on	ADP
fcis-9730	8	6	the	the	DET
fcis-9730	8	7	convolutional	convolutional	ADJ
fcis-9730	8	8	neural	neural	ADJ
fcis-9730	8	9	network	network	NOUN
fcis-9730	8	10	(	(	PUNCT
fcis-9730	8	11	cnn	cnn	PROPN
fcis-9730	8	12	)	)	PUNCT
fcis-9730	8	13	,	,	PUNCT
fcis-9730	8	14	in	in	ADP
fcis-9730	8	15	2012	2012	NUM
fcis-9730	8	16	alexnet	alexnet	NOUN
fcis-9730	8	17	[	[	X
fcis-9730	8	18	4	4	X
fcis-9730	8	19	]	]	PUNCT
fcis-9730	8	20	won	win	VERB
fcis-9730	8	21	the	the	DET
fcis-9730	8	22	champion	champion	NOUN
fcis-9730	8	23	of	of	ADP
fcis-9730	8	24	imagenet	imagenet	ADJ
fcis-9730	8	25	image	image	NOUN
fcis-9730	8	26	recognition	recognition	NOUN
fcis-9730	8	27	competition	competition	NOUN
fcis-9730	8	28	by	by	ADP
fcis-9730	8	29	a	a	DET
fcis-9730	8	30	significant	significant	ADJ
fcis-9730	8	31	margin	margin	NOUN
fcis-9730	8	32	,	,	PUNCT
fcis-9730	8	33	and	and	CCONJ
fcis-9730	8	34	since	since	SCONJ
fcis-9730	8	35	then	then	ADV
fcis-9730	8	36	deep	deep	ADJ
fcis-9730	8	37	learning	learning	NOUN
fcis-9730	8	38	has	have	AUX
fcis-9730	8	39	received	receive	VERB
fcis-9730	8	40	wide	wide	ADJ
fcis-9730	8	41	attention	attention	NOUN
fcis-9730	8	42	.	.	PUNCT
fcis-9730	9	1	overfeat	overfeat	NOUN
fcis-9730	9	2	was	be	AUX
fcis-9730	9	3	born	bear	VERB
fcis-9730	9	4	in	in	ADP
fcis-9730	9	5	2013	2013	NUM
fcis-9730	9	6	and	and	CCONJ
fcis-9730	9	7	is	be	AUX
fcis-9730	9	8	regarded	regard	VERB
fcis-9730	9	9	as	as	ADP
fcis-9730	9	10	the	the	DET
fcis-9730	9	11	pioneer	pioneer	NOUN
fcis-9730	9	12	of	of	ADP
fcis-9730	9	13	single	single	ADJ
fcis-9730	9	14	-	-	PUNCT
fcis-9730	9	15	stage	stage	NOUN
fcis-9730	9	16	object	object	NOUN
fcis-9730	9	17	detection	detection	NOUN
fcis-9730	9	18	due	due	ADP
fcis-9730	9	19	to	to	ADP
fcis-9730	9	20	its	its	PRON
fcis-9730	9	21	integration	integration	NOUN
fcis-9730	9	22	of	of	ADP
fcis-9730	9	23	location	location	NOUN
fcis-9730	9	24	and	and	CCONJ
fcis-9730	9	25	detection	detection	NOUN
fcis-9730	9	26	technology	technology	NOUN
fcis-9730	9	27	.	.	PUNCT
fcis-9730	10	1	the	the	DET
fcis-9730	10	2	proposal	proposal	NOUN
fcis-9730	10	3	of	of	ADP
fcis-9730	10	4	r	r	NOUN
fcis-9730	10	5	-	-	PUNCT
fcis-9730	10	6	cnn	cnn	PROPN
fcis-9730	10	7	in	in	ADP
fcis-9730	10	8	2014	2014	NUM
fcis-9730	10	9	has	have	AUX
fcis-9730	10	10	achieved	achieve	VERB
fcis-9730	10	11	a	a	DET
fcis-9730	10	12	very	very	ADV
fcis-9730	10	13	big	big	ADJ
fcis-9730	10	14	breakthrough	breakthrough	NOUN
fcis-9730	10	15	in	in	ADP
fcis-9730	10	16	object	object	NOUN
fcis-9730	10	17	detection	detection	NOUN
fcis-9730	10	18	tasks	task	NOUN
fcis-9730	10	19	.	.	PUNCT
fcis-9730	11	1	after	after	ADP
fcis-9730	11	2	that	that	PRON
fcis-9730	11	3	,	,	PUNCT
fcis-9730	11	4	the	the	DET
fcis-9730	11	5	object	object	NOUN
fcis-9730	11	6	detection	detection	NOUN
fcis-9730	11	7	algorithm	algorithm	NOUN
fcis-9730	11	8	based	base	VERB
fcis-9730	11	9	on	on	ADP
fcis-9730	11	10	deep	deep	ADJ
fcis-9730	11	11	learning	learning	NOUN
fcis-9730	11	12	began	begin	VERB
fcis-9730	11	13	to	to	PART
fcis-9730	11	14	emerge	emerge	VERB
fcis-9730	11	15	in	in	ADP
fcis-9730	11	16	the	the	DET
fcis-9730	11	17	field	field	NOUN
fcis-9730	11	18	of	of	ADP
fcis-9730	11	19	natural	natural	ADJ
fcis-9730	11	20	image	image	NOUN
fcis-9730	11	21	recognition	recognition	NOUN
fcis-9730	11	22	,	,	PUNCT
fcis-9730	11	23	and	and	CCONJ
fcis-9730	11	24	the	the	DET
fcis-9730	11	25	two	two	NUM
fcis-9730	11	26	detection	detection	NOUN
fcis-9730	11	27	algorithms	algorithm	NOUN
fcis-9730	11	28	of	of	ADP
fcis-9730	11	29	single	single	ADJ
fcis-9730	11	30	-	-	PUNCT
fcis-9730	11	31	stage	stage	NOUN
fcis-9730	11	32	object	object	NOUN
fcis-9730	11	33	detection	detection	NOUN
fcis-9730	11	34	and	and	CCONJ
fcis-9730	11	35	two	two	NUM
fcis-9730	11	36	-	-	PUNCT
fcis-9730	11	37	stage	stage	NOUN
fcis-9730	11	38	object	object	NOUN
fcis-9730	11	39	detection	detection	NOUN
fcis-9730	11	40	also	also	ADV
fcis-9730	11	41	competed	compete	VERB
fcis-9730	11	42	with	with	ADP
fcis-9730	11	43	each	each	DET
fcis-9730	11	44	other	other	ADJ
fcis-9730	11	45	in	in	ADP
fcis-9730	11	46	their	their	PRON
fcis-9730	11	47	fields	field	NOUN
fcis-9730	11	48	and	and	CCONJ
fcis-9730	11	49	excelled	excel	VERB
fcis-9730	11	50	.	.	PUNCT
fcis-9730	12	1	the	the	DET
fcis-9730	12	2	performance	performance	NOUN
fcis-9730	12	3	of	of	ADP
fcis-9730	12	4	r	r	NOUN
fcis-9730	12	5	-	-	PUNCT
fcis-9730	12	6	cnn	cnn	NOUN
fcis-9730	12	7	on	on	ADP
fcis-9730	12	8	voc2007	voc2007	NOUN
fcis-9730	12	9	has	have	AUX
fcis-9730	12	10	been	be	AUX
fcis-9730	12	11	significantly	significantly	ADV
fcis-9730	12	12	improved	improve	VERB
fcis-9730	12	13	,	,	PUNCT
fcis-9730	12	14	and	and	CCONJ
fcis-9730	12	15	the	the	DET
fcis-9730	12	16	average	average	ADJ
fcis-9730	12	17	detection	detection	NOUN
fcis-9730	12	18	accuracy	accuracy	NOUN
fcis-9730	12	19	(	(	PUNCT
fcis-9730	12	20	map	map	NOUN
fcis-9730	12	21	)	)	PUNCT
fcis-9730	12	22	has	have	AUX
fcis-9730	12	23	been	be	AUX
fcis-9730	12	24	improved	improve	VERB
fcis-9730	12	25	by	by	ADP
fcis-9730	12	26	nearly	nearly	ADV
fcis-9730	12	27	25	25	NUM
fcis-9730	12	28	%	%	NOUN
fcis-9730	12	29	compared	compare	VERB
fcis-9730	12	30	with	with	ADP
fcis-9730	12	31	the	the	DET
fcis-9730	12	32	traditional	traditional	ADJ
fcis-9730	12	33	method	method	NOUN
fcis-9730	12	34	.	.	PUNCT
fcis-9730	13	1	but	but	CCONJ
fcis-9730	13	2	the	the	DET
fcis-9730	13	3	regression	regression	NOUN
fcis-9730	13	4	process	process	NOUN
fcis-9730	13	5	of	of	ADP
fcis-9730	13	6	this	this	DET
fcis-9730	13	7	method	method	NOUN
fcis-9730	13	8	is	be	AUX
fcis-9730	13	9	extremely	extremely	ADV
fcis-9730	13	10	time	time	NOUN
fcis-9730	13	11	and	and	CCONJ
fcis-9730	13	12	memory	memory	NOUN
fcis-9730	13	13	-	-	PUNCT
fcis-9730	13	14	intensive	intensive	ADJ
fcis-9730	13	15	.	.	PUNCT
fcis-9730	14	1	in	in	ADP
fcis-9730	14	2	the	the	DET
fcis-9730	14	3	same	same	ADJ
fcis-9730	14	4	year	year	NOUN
fcis-9730	14	5	,	,	PUNCT
fcis-9730	14	6	in	in	ADP
fcis-9730	14	7	order	order	NOUN
fcis-9730	14	8	to	to	PART
fcis-9730	14	9	improve	improve	VERB
fcis-9730	14	10	the	the	DET
fcis-9730	14	11	complex	complex	ADJ
fcis-9730	14	12	training	training	NOUN
fcis-9730	14	13	process	process	NOUN
fcis-9730	14	14	of	of	ADP
fcis-9730	14	15	r	r	NOUN
fcis-9730	14	16	-	-	PUNCT
fcis-9730	14	17	cnn	cnn	PROPN
fcis-9730	14	18	and	and	CCONJ
fcis-9730	14	19	improve	improve	VERB
fcis-9730	14	20	the	the	DET
fcis-9730	14	21	detection	detection	NOUN
fcis-9730	14	22	speed	speed	NOUN
fcis-9730	14	23	,	,	PUNCT
fcis-9730	14	24	he	he	PRON
fcis-9730	14	25	proposed	propose	VERB
fcis-9730	14	26	the	the	DET
fcis-9730	14	27	spatial	spatial	ADJ
fcis-9730	14	28	pyramid	pyramid	NOUN
fcis-9730	14	29	pool	pool	NOUN
fcis-9730	14	30	network	network	NOUN
fcis-9730	14	31	(	(	PUNCT
fcis-9730	14	32	sppnet	sppnet	PROPN
fcis-9730	14	33	)	)	PUNCT
fcis-9730	15	1	[	[	X
fcis-9730	15	2	5	5	NUM
fcis-9730	15	3	]	]	PUNCT
fcis-9730	15	4	,	,	PUNCT
fcis-9730	15	5	which	which	PRON
fcis-9730	15	6	can	can	AUX
fcis-9730	15	7	effectively	effectively	ADV
fcis-9730	15	8	reduce	reduce	VERB
fcis-9730	15	9	the	the	DET
fcis-9730	15	10	information	information	NOUN
fcis-9730	15	11	redundancy	redundancy	NOUN
fcis-9730	15	12	caused	cause	VERB
fcis-9730	15	13	by	by	ADP
fcis-9730	15	14	repeated	repeat	VERB
fcis-9730	15	15	calculation	calculation	NOUN
fcis-9730	15	16	.	.	PUNCT
fcis-9730	16	1	the	the	DET
fcis-9730	16	2	detection	detection	NOUN
fcis-9730	16	3	speed	speed	NOUN
fcis-9730	16	4	is	be	AUX
fcis-9730	16	5	more	more	ADJ
fcis-9730	16	6	than	than	ADP
fcis-9730	16	7	20	20	NUM
fcis-9730	16	8	times	time	NOUN
fcis-9730	16	9	that	that	PRON
fcis-9730	16	10	of	of	ADP
fcis-9730	16	11	r	r	NOUN
fcis-9730	16	12	-	-	PUNCT
fcis-9730	16	13	cnn	cnn	PROPN
fcis-9730	16	14	,	,	PUNCT
fcis-9730	16	15	and	and	CCONJ
fcis-9730	16	16	the	the	DET
fcis-9730	16	17	map	map	NOUN
fcis-9730	16	18	can	can	AUX
fcis-9730	16	19	reach	reach	VERB
fcis-9730	16	20	66	66	NUM
fcis-9730	16	21	%	%	NOUN
fcis-9730	16	22	.	.	PUNCT
fcis-9730	17	1	the	the	DET
fcis-9730	17	2	network	network	NOUN
fcis-9730	17	3	only	only	ADV
fcis-9730	17	4	fine	fine	ADV
fcis-9730	17	5	-	-	PUNCT
fcis-9730	17	6	tuned	tune	VERB
fcis-9730	17	7	the	the	DET
fcis-9730	17	8	full	full	ADJ
fcis-9730	17	9	connection	connection	NOUN
fcis-9730	17	10	layer	layer	NOUN
fcis-9730	17	11	,	,	PUNCT
fcis-9730	17	12	but	but	CCONJ
fcis-9730	17	13	did	do	AUX
fcis-9730	17	14	not	not	PART
fcis-9730	17	15	deal	deal	VERB
fcis-9730	17	16	with	with	ADP
fcis-9730	17	17	other	other	ADJ
fcis-9730	17	18	feature	feature	NOUN
fcis-9730	17	19	layers	layer	NOUN
fcis-9730	17	20	in	in	ADP
fcis-9730	17	21	the	the	DET
fcis-9730	17	22	training	training	NOUN
fcis-9730	17	23	process	process	NOUN
fcis-9730	17	24	.	.	PUNCT
fcis-9730	18	1	in	in	ADP
fcis-9730	18	2	2015	2015	NUM
fcis-9730	18	3	,	,	PUNCT
fcis-9730	18	4	girshick	girshick	VERB
fcis-9730	18	5	once	once	ADV
fcis-9730	18	6	again	again	ADV
fcis-9730	18	7	proposed	propose	VERB
fcis-9730	18	8	the	the	DET
fcis-9730	18	9	fast	fast	ADJ
fcis-9730	18	10	r	r	NOUN
fcis-9730	18	11	-	-	PUNCT
fcis-9730	18	12	cnn	cnn	NOUN
fcis-9730	18	13	detector	detector	NOUN
fcis-9730	18	14	[	[	X
fcis-9730	18	15	6	6	NUM
fcis-9730	18	16	]	]	PUNCT
fcis-9730	18	17	,	,	PUNCT
fcis-9730	18	18	which	which	PRON
fcis-9730	18	19	integrated	integrate	VERB
fcis-9730	18	20	the	the	DET
fcis-9730	18	21	r	r	NOUN
fcis-9730	18	22	-	-	PUNCT
fcis-9730	18	23	cnn	cnn	PROPN
fcis-9730	18	24	and	and	CCONJ
fcis-9730	18	25	sppnet	sppnet	PROPN
fcis-9730	18	26	network	network	NOUN
fcis-9730	18	27	structure	structure	NOUN
fcis-9730	18	28	,	,	PUNCT
fcis-9730	18	29	and	and	CCONJ
fcis-9730	18	30	was	be	AUX
fcis-9730	18	31	able	able	ADJ
fcis-9730	18	32	to	to	PART
fcis-9730	18	33	train	train	VERB
fcis-9730	18	34	both	both	DET
fcis-9730	18	35	the	the	DET
fcis-9730	18	36	target	target	NOUN
fcis-9730	18	37	category	category	NOUN
fcis-9730	18	38	detector	detector	NOUN
fcis-9730	18	39	and	and	CCONJ
fcis-9730	18	40	the	the	DET
fcis-9730	18	41	bounding	bounding	NOUN
fcis-9730	18	42	box	box	NOUN
fcis-9730	18	43	regresser	regresser	NOUN
fcis-9730	18	44	during	during	ADP
fcis-9730	18	45	detection	detection	NOUN
fcis-9730	18	46	.	.	PUNCT
fcis-9730	19	1	it	it	PRON
fcis-9730	19	2	was	be	AUX
fcis-9730	19	3	found	find	VERB
fcis-9730	19	4	that	that	SCONJ
fcis-9730	19	5	the	the	DET
fcis-9730	19	6	fast	fast	ADJ
fcis-9730	19	7	r	r	NOUN
fcis-9730	19	8	-	-	PUNCT
fcis-9730	19	9	cnn	cnn	PROPN
fcis-9730	19	10	detector	detector	NOUN
fcis-9730	19	11	significantly	significantly	ADV
fcis-9730	19	12	accelerated	accelerate	VERB
fcis-9730	19	13	the	the	DET
fcis-9730	19	14	training	training	NOUN
fcis-9730	19	15	process	process	NOUN
fcis-9730	19	16	and	and	CCONJ
fcis-9730	19	17	test	test	NOUN
fcis-9730	19	18	speed	speed	NOUN
fcis-9730	19	19	,	,	PUNCT
fcis-9730	19	20	and	and	CCONJ
fcis-9730	19	21	further	far	ADV
fcis-9730	19	22	improved	improve	VERB
fcis-9730	19	23	the	the	DET
fcis-9730	19	24	detection	detection	NOUN
fcis-9730	19	25	accuracy	accuracy	NOUN
fcis-9730	19	26	.	.	PUNCT
fcis-9730	20	1	with	with	ADP
fcis-9730	20	2	map	map	NOUN
fcis-9730	20	3	reaching	reach	VERB
fcis-9730	20	4	66.9	66.9	NUM
fcis-9730	20	5	%	%	NOUN
fcis-9730	20	6	.	.	PUNCT
fcis-9730	21	1	soon	soon	ADV
fcis-9730	21	2	after	after	ADV
fcis-9730	21	3	,	,	PUNCT
fcis-9730	21	4	ren	ren	PROPN
fcis-9730	21	5	proposed	propose	VERB
fcis-9730	21	6	faster	fast	ADV
fcis-9730	21	7	r	r	NOUN
fcis-9730	21	8	-	-	PUNCT
fcis-9730	21	9	cnn	cnn	NOUN
fcis-9730	22	1	[	[	X
fcis-9730	22	2	7	7	NUM
fcis-9730	22	3	]	]	PUNCT
fcis-9730	22	4	,	,	PUNCT
fcis-9730	22	5	which	which	PRON
fcis-9730	22	6	replaces	replace	VERB
fcis-9730	22	7	slow	slow	ADJ
fcis-9730	22	8	selective	selective	ADJ
fcis-9730	22	9	search	search	NOUN
fcis-9730	22	10	with	with	ADP
fcis-9730	22	11	efficient	efficient	ADJ
fcis-9730	22	12	regional	regional	ADJ
fcis-9730	22	13	candidate	candidate	NOUN
fcis-9730	22	14	network	network	NOUN
fcis-9730	22	15	(	(	PUNCT
fcis-9730	22	16	rpn	rpn	PROPN
fcis-9730	22	17	)	)	PUNCT
fcis-9730	22	18	,	,	PUNCT
fcis-9730	22	19	overcomes	overcome	VERB
fcis-9730	22	20	the	the	DET
fcis-9730	22	21	speed	speed	NOUN
fcis-9730	22	22	bottleneck	bottleneck	NOUN
fcis-9730	22	23	of	of	ADP
fcis-9730	22	24	fast	fast	ADJ
fcis-9730	22	25	r	r	NOUN
fcis-9730	22	26	-	-	PUNCT
fcis-9730	22	27	cnn	cnn	PROPN
fcis-9730	22	28	,	,	PUNCT
fcis-9730	22	29	improves	improve	VERB
fcis-9730	22	30	detection	detection	NOUN
fcis-9730	22	31	accuracy	accuracy	NOUN
fcis-9730	22	32	and	and	CCONJ
fcis-9730	22	33	operational	operational	ADJ
fcis-9730	22	34	efficiency	efficiency	NOUN
fcis-9730	22	35	,	,	PUNCT
fcis-9730	22	36	and	and	CCONJ
fcis-9730	22	37	achieves	achieve	VERB
fcis-9730	22	38	end	end	NOUN
fcis-9730	22	39	-	-	PUNCT
fcis-9730	22	40	to	to	ADP
fcis-9730	22	41	-	-	PUNCT
fcis-9730	22	42	end	end	NOUN
fcis-9730	22	43	target	target	NOUN
fcis-9730	22	44	detection	detection	NOUN
fcis-9730	22	45	.	.	PUNCT
fcis-9730	23	1	map	map	NOUN
fcis-9730	23	2	achieved	achieve	VERB
fcis-9730	23	3	69.9	69.9	NUM
fcis-9730	23	4	%	%	NOUN
fcis-9730	23	5	on	on	ADP
fcis-9730	23	6	the	the	DET
fcis-9730	23	7	voc2007	voc2007	NOUN
fcis-9730	23	8	dataset	dataset	NOUN
fcis-9730	23	9	.	.	PUNCT
fcis-9730	24	1	one	one	NUM
fcis-9730	24	2	of	of	ADP
fcis-9730	24	3	the	the	DET
fcis-9730	24	4	fast	fast	ADJ
fcis-9730	24	5	versions	version	NOUN
fcis-9730	24	6	,	,	PUNCT
fcis-9730	24	7	map	map	NOUN
fcis-9730	24	8	,	,	PUNCT
fcis-9730	24	9	achieved	achieve	VERB
fcis-9730	24	10	59.9	59.9	NUM
fcis-9730	24	11	%	%	NOUN
fcis-9730	24	12	and	and	CCONJ
fcis-9730	24	13	is	be	AUX
fcis-9730	24	14	the	the	DET
fcis-9730	24	15	first	first	ADJ
fcis-9730	24	16	ever	ever	ADV
fcis-9730	24	17	near	near	ADV
fcis-9730	24	18	realtime	realtime	ADJ
fcis-9730	24	19	deep	deep	ADJ
fcis-9730	24	20	learning	learning	NOUN
fcis-9730	24	21	object	object	NOUN
fcis-9730	24	22	detector	detector	NOUN
fcis-9730	24	23	.	.	PUNCT
fcis-9730	25	1	at	at	ADP
fcis-9730	25	2	this	this	DET
fcis-9730	25	3	point	point	NOUN
fcis-9730	25	4	,	,	PUNCT
fcis-9730	25	5	the	the	DET
fcis-9730	25	6	basic	basic	ADJ
fcis-9730	25	7	architecture	architecture	NOUN
fcis-9730	25	8	of	of	ADP
fcis-9730	25	9	the	the	DET
fcis-9730	25	10	two	two	NUM
fcis-9730	25	11	-	-	PUNCT
fcis-9730	25	12	stage	stage	NOUN
fcis-9730	25	13	detector	detector	NOUN
fcis-9730	25	14	is	be	AUX
fcis-9730	25	15	determined	determine	VERB
fcis-9730	25	16	.	.	PUNCT
fcis-9730	26	1	this	this	DET
fcis-9730	26	2	paper	paper	NOUN
fcis-9730	26	3	first	first	ADV
fcis-9730	26	4	introduces	introduce	VERB
fcis-9730	26	5	the	the	DET
fcis-9730	26	6	yolo	yolo	ADJ
fcis-9730	26	7	series	series	PROPN
fcis-9730	26	8	algorithm	algorithm	PROPN
fcis-9730	26	9	,	,	PUNCT
fcis-9730	26	10	including	include	VERB
fcis-9730	26	11	the	the	DET
fcis-9730	26	12	principle	principle	ADJ
fcis-9730	26	13	,	,	PUNCT
fcis-9730	26	14	innovation	innovation	NOUN
fcis-9730	26	15	and	and	CCONJ
fcis-9730	26	16	advantages	advantage	NOUN
fcis-9730	26	17	and	and	CCONJ
fcis-9730	26	18	disadvantages	disadvantage	NOUN
fcis-9730	26	19	of	of	ADP
fcis-9730	26	20	various	various	ADJ
fcis-9730	26	21	algorithms	algorithm	NOUN
fcis-9730	26	22	,	,	PUNCT
fcis-9730	26	23	then	then	ADV
fcis-9730	26	24	introduces	introduce	VERB
fcis-9730	26	25	the	the	DET
fcis-9730	26	26	application	application	NOUN
fcis-9730	26	27	field	field	NOUN
fcis-9730	26	28	of	of	ADP
fcis-9730	26	29	yolo	yolo	ADJ
fcis-9730	26	30	series	series	NOUN
fcis-9730	26	31	,	,	PUNCT
fcis-9730	26	32	and	and	CCONJ
fcis-9730	26	33	finally	finally	ADV
fcis-9730	26	34	analyzes	analyze	VERB
fcis-9730	26	35	its	its	PRON
fcis-9730	26	36	future	future	ADJ
fcis-9730	26	37	development	development	NOUN
fcis-9730	26	38	trend	trend	NOUN
fcis-9730	26	39	.	.	PUNCT
fcis-9730	27	1	[	[	X
fcis-9730	27	2	7	7	NUM
fcis-9730	27	3	]	]	SYM
fcis-9730	27	4	2	2	NUM
fcis-9730	27	5	.	.	PUNCT
fcis-9730	27	6	yolo	yolo	PROPN
fcis-9730	27	7	series	series	PROPN
fcis-9730	27	8	algorithm	algorithm	PROPN
fcis-9730	27	9	development	development	NOUN
fcis-9730	27	10	process	process	NOUN
fcis-9730	27	11	2.1	2.1	NUM
fcis-9730	27	12	.	.	PUNCT
fcis-9730	28	1	introduction	introduction	NOUN
fcis-9730	28	2	to	to	ADP
fcis-9730	28	3	yolo	yolo	NOUN
fcis-9730	28	4	in	in	ADP
fcis-9730	28	5	2016	2016	NUM
fcis-9730	28	6	,	,	PUNCT
fcis-9730	28	7	the	the	DET
fcis-9730	28	8	one	one	NUM
fcis-9730	28	9	-	-	PUNCT
fcis-9730	28	10	stage	stage	NOUN
fcis-9730	28	11	object	object	NOUN
fcis-9730	28	12	detection	detection	NOUN
fcis-9730	28	13	network	network	NOUN
fcis-9730	28	14	was	be	AUX
fcis-9730	28	15	proposed	propose	VERB
fcis-9730	28	16	.	.	PUNCT
fcis-9730	29	1	when	when	SCONJ
fcis-9730	29	2	tested	test	VERB
fcis-9730	29	3	in	in	ADP
fcis-9730	29	4	the	the	DET
fcis-9730	29	5	same	same	ADJ
fcis-9730	29	6	configuration	configuration	NOUN
fcis-9730	29	7	,	,	PUNCT
fcis-9730	29	8	it	it	PRON
fcis-9730	29	9	was	be	AUX
fcis-9730	29	10	found	find	VERB
fcis-9730	29	11	to	to	PART
fcis-9730	29	12	be	be	AUX
fcis-9730	29	13	able	able	ADJ
fcis-9730	29	14	to	to	PART
fcis-9730	29	15	process	process	VERB
fcis-9730	29	16	45fps	45fps	ADJ
fcis-9730	29	17	images	image	NOUN
fcis-9730	29	18	per	per	ADP
fcis-9730	29	19	second	second	ADJ
fcis-9730	29	20	while	while	SCONJ
fcis-9730	29	21	easily	easily	ADV
fcis-9730	29	22	running	run	VERB
fcis-9730	29	23	the	the	DET
fcis-9730	29	24	detection	detection	NOUN
fcis-9730	29	25	in	in	ADP
fcis-9730	29	26	real	real	ADJ
fcis-9730	29	27	time	time	NOUN
fcis-9730	29	28	.	.	PUNCT
fcis-9730	30	1	because	because	SCONJ
fcis-9730	30	2	of	of	ADP
fcis-9730	30	3	its	its	PRON
fcis-9730	30	4	speed	speed	NOUN
fcis-9730	30	5	and	and	CCONJ
fcis-9730	30	6	its	its	PRON
fcis-9730	30	7	special	special	ADJ
fcis-9730	30	8	use	use	NOUN
fcis-9730	30	9	method	method	NOUN
fcis-9730	30	10	,	,	PUNCT
fcis-9730	30	11	the	the	DET
fcis-9730	30	12	authors	author	NOUN
fcis-9730	30	13	give	give	VERB
fcis-9730	30	14	it	it	PRON
fcis-9730	30	15	the	the	DET
fcis-9730	30	16	name	name	NOUN
fcis-9730	30	17	yolo	yolo	ADJ
fcis-9730	30	18	(	(	PUNCT
fcis-9730	30	19	you	you	PRON
fcis-9730	30	20	only	only	ADV
fcis-9730	30	21	look	look	VERB
fcis-9730	30	22	once	once	ADV
fcis-9730	30	23	)	)	PUNCT
fcis-9730	30	24	.	.	PUNCT
fcis-9730	31	1	[	[	X
fcis-9730	31	2	8	8	X
fcis-9730	31	3	]	]	PUNCT
fcis-9730	31	4	this	this	DET
fcis-9730	31	5	method	method	NOUN
fcis-9730	31	6	completely	completely	ADV
fcis-9730	31	7	abandons	abandon	VERB
fcis-9730	31	8	the	the	DET
fcis-9730	31	9	detection	detection	NOUN
fcis-9730	31	10	mode	mode	NOUN
fcis-9730	31	11	of	of	ADP
fcis-9730	31	12	"	"	PUNCT
fcis-9730	31	13	region	region	NOUN
fcis-9730	31	14	candidate	candidate	NOUN
fcis-9730	31	15	+	+	CCONJ
fcis-9730	31	16	regression	regression	NOUN
fcis-9730	31	17	"	"	PUNCT
fcis-9730	31	18	in	in	ADP
fcis-9730	31	19	the	the	DET
fcis-9730	31	20	two	two	NUM
fcis-9730	31	21	stages	stage	NOUN
fcis-9730	31	22	,	,	PUNCT
fcis-9730	31	23	scales	scale	VERB
fcis-9730	31	24	the	the	DET
fcis-9730	31	25	image	image	NOUN
fcis-9730	31	26	to	to	PART
fcis-9730	31	27	be	be	AUX
fcis-9730	31	28	measured	measure	VERB
fcis-9730	31	29	into	into	ADP
fcis-9730	31	30	a	a	DET
fcis-9730	31	31	uniform	uniform	ADJ
fcis-9730	31	32	size	size	NOUN
fcis-9730	31	33	and	and	CCONJ
fcis-9730	31	34	divides	divide	VERB
fcis-9730	31	35	it	it	PRON
fcis-9730	31	36	into	into	ADP
fcis-9730	31	37	multiple	multiple	ADJ
fcis-9730	31	38	grids	grid	NOUN
fcis-9730	31	39	,	,	PUNCT
fcis-9730	31	40	then	then	ADV
fcis-9730	31	41	predicts	predict	VERB
fcis-9730	31	42	the	the	DET
fcis-9730	31	43	target	target	NOUN
fcis-9730	31	44	category	category	NOUN
fcis-9730	31	45	according	accord	VERB
fcis-9730	31	46	to	to	ADP
fcis-9730	31	47	the	the	DET
fcis-9730	31	48	grid	grid	NOUN
fcis-9730	31	49	where	where	SCONJ
fcis-9730	31	50	the	the	DET
fcis-9730	31	51	target	target	NOUN
fcis-9730	31	52	center	center	NOUN
fcis-9730	31	53	is	be	AUX
fcis-9730	31	54	located	locate	VERB
fcis-9730	31	55	,	,	PUNCT
fcis-9730	31	56	and	and	CCONJ
fcis-9730	31	57	outputs	output	VERB
fcis-9730	31	58	the	the	DET
fcis-9730	31	59	detection	detection	NOUN
fcis-9730	31	60	results	result	VERB
fcis-9730	31	61	on	on	ADP
fcis-9730	31	62	the	the	DET
fcis-9730	31	63	last	last	ADJ
fcis-9730	31	64	convolution	convolution	NOUN
fcis-9730	31	65	layer	layer	NOUN
fcis-9730	31	66	.	.	PUNCT
fcis-9730	32	1	the	the	DET
fcis-9730	32	2	core	core	ADJ
fcis-9730	32	3	idea	idea	NOUN
fcis-9730	32	4	of	of	ADP
fcis-9730	32	5	yolo	yolo	PROPN
fcis-9730	32	6	is	be	AUX
fcis-9730	32	7	to	to	PART
fcis-9730	32	8	transform	transform	VERB
fcis-9730	32	9	target	target	NOUN
fcis-9730	32	10	detection	detection	NOUN
fcis-9730	32	11	into	into	ADP
fcis-9730	32	12	a	a	DET
fcis-9730	32	13	regression	regression	NOUN
fcis-9730	32	14	problem	problem	NOUN
fcis-9730	32	15	,	,	PUNCT
fcis-9730	32	16	using	use	VERB
fcis-9730	32	17	the	the	DET
fcis-9730	32	18	whole	whole	ADJ
fcis-9730	32	19	image	image	NOUN
fcis-9730	32	20	as	as	ADP
fcis-9730	32	21	the	the	DET
fcis-9730	32	22	input	input	NOUN
fcis-9730	32	23	of	of	ADP
fcis-9730	32	24	the	the	DET
fcis-9730	32	25	network	network	NOUN
fcis-9730	32	26	,	,	PUNCT
fcis-9730	32	27	and	and	CCONJ
fcis-9730	32	28	just	just	ADV
fcis-9730	32	29	through	through	ADP
fcis-9730	32	30	a	a	DET
fcis-9730	32	31	neural	neural	ADJ
fcis-9730	32	32	network	network	NOUN
fcis-9730	32	33	,	,	PUNCT
fcis-9730	32	34	the	the	DET
fcis-9730	32	35	location	location	NOUN
fcis-9730	32	36	of	of	ADP
fcis-9730	32	37	the	the	DET
fcis-9730	32	38	bounding	bounding	NOUN
fcis-9730	32	39	box	box	NOUN
fcis-9730	32	40	and	and	CCONJ
fcis-9730	32	41	its	its	PRON
fcis-9730	32	42	category	category	NOUN
fcis-9730	32	43	can	can	AUX
fcis-9730	32	44	be	be	AUX
fcis-9730	32	45	obtained	obtain	VERB
fcis-9730	32	46	.	.	PUNCT
fcis-9730	33	1	[	[	X
fcis-9730	33	2	9	9	NUM
fcis-9730	33	3	]	]	X
fcis-9730	33	4	first	first	ADV
fcis-9730	33	5	,	,	PUNCT
fcis-9730	33	6	the	the	DET
fcis-9730	33	7	input	input	NOUN
fcis-9730	33	8	image	image	NOUN
fcis-9730	33	9	is	be	AUX
fcis-9730	33	10	uniformly	uniformly	ADV
fcis-9730	33	11	adjusted	adjust	VERB
fcis-9730	33	12	to	to	ADP
fcis-9730	33	13	448×448	448×448	NUM
fcis-9730	33	14	pixels	pixel	NOUN
fcis-9730	33	15	,	,	PUNCT
fcis-9730	33	16	and	and	CCONJ
fcis-9730	33	17	then	then	ADV
fcis-9730	33	18	the	the	DET
fcis-9730	33	19	image	image	NOUN
fcis-9730	33	20	is	be	AUX
fcis-9730	33	21	divided	divide	VERB
fcis-9730	33	22	into	into	ADP
fcis-9730	33	23	s×s	s×s	ADJ
fcis-9730	33	24	grid	grid	NOUN
fcis-9730	33	25	.	.	PUNCT
fcis-9730	34	1	if	if	SCONJ
fcis-9730	34	2	the	the	DET
fcis-9730	34	3	center	center	NOUN
fcis-9730	34	4	of	of	ADP
fcis-9730	34	5	the	the	DET
fcis-9730	34	6	detection	detection	NOUN
fcis-9730	34	7	object	object	NOUN
fcis-9730	34	8	falls	fall	VERB
fcis-9730	34	9	into	into	ADP
fcis-9730	34	10	a	a	DET
fcis-9730	34	11	grid	grid	NOUN
fcis-9730	34	12	cell	cell	NOUN
fcis-9730	34	13	,	,	PUNCT
fcis-9730	34	14	the	the	DET
fcis-9730	34	15	grid	grid	NOUN
fcis-9730	34	16	cell	cell	NOUN
fcis-9730	34	17	is	be	AUX
fcis-9730	34	18	responsible	responsible	ADJ
fcis-9730	34	19	for	for	ADP
fcis-9730	34	20	detecting	detect	VERB
fcis-9730	34	21	the	the	DET
fcis-9730	34	22	object	object	NOUN
fcis-9730	34	23	.	.	PUNCT
fcis-9730	35	1	each	each	DET
fcis-9730	35	2	grid	grid	NOUN
fcis-9730	35	3	cell	cell	NOUN
fcis-9730	35	4	scores	score	NOUN
fcis-9730	35	5	its	its	PRON
fcis-9730	35	6	bounding	bounding	NOUN
fcis-9730	35	7	boxes	box	NOUN
fcis-9730	35	8	to	to	PART
fcis-9730	35	9	predict	predict	VERB
fcis-9730	35	10	the	the	DET
fcis-9730	35	11	likelihood	likelihood	NOUN
fcis-9730	35	12	of	of	ADP
fcis-9730	35	13	detection	detection	NOUN
fcis-9730	35	14	objects	object	NOUN
fcis-9730	35	15	in	in	ADP
fcis-9730	35	16	the	the	DET
fcis-9730	35	17	grid	grid	NOUN
fcis-9730	35	18	cells	cell	NOUN
fcis-9730	35	19	.	.	PUNCT
fcis-9730	36	1	if	if	SCONJ
fcis-9730	36	2	the	the	DET
fcis-9730	36	3	predicted	predict	VERB
fcis-9730	36	4	value	value	NOUN
fcis-9730	36	5	is	be	AUX
fcis-9730	36	6	0	0	NUM
fcis-9730	36	7	,	,	PUNCT
fcis-9730	36	8	it	it	PRON
fcis-9730	36	9	means	mean	VERB
fcis-9730	36	10	no	no	DET
fcis-9730	36	11	detection	detection	NOUN
fcis-9730	36	12	objects	object	NOUN
fcis-9730	36	13	are	be	AUX
fcis-9730	36	14	present	present	ADJ
fcis-9730	36	15	in	in	ADP
fcis-9730	36	16	the	the	DET
fcis-9730	36	17	grid	grid	NOUN
fcis-9730	36	18	cells	cell	NOUN
fcis-9730	36	19	.	.	PUNCT
fcis-9730	37	1	each	each	PRON
fcis-9730	37	2	of	of	ADP
fcis-9730	37	3	yolo	yolo	PROPN
fcis-9730	37	4	's	's	PART
fcis-9730	37	5	grid	grid	NOUN
fcis-9730	37	6	cells	cell	NOUN
fcis-9730	37	7	can	can	AUX
fcis-9730	37	8	provide	provide	VERB
fcis-9730	37	9	an	an	DET
fcis-9730	37	10	envelope	envelope	NOUN
fcis-9730	37	11	of	of	ADP
fcis-9730	37	12	complex	complex	ADJ
fcis-9730	37	13	numbers	number	NOUN
fcis-9730	37	14	,	,	PUNCT
fcis-9730	37	15	but	but	CCONJ
fcis-9730	37	16	an	an	DET
fcis-9730	37	17	envelope	envelope	NOUN
fcis-9730	37	18	only	only	ADV
fcis-9730	37	19	selects	select	VERB
fcis-9730	37	20	the	the	DET
fcis-9730	37	21	highest	high	ADJ
fcis-9730	37	22	scoring	scoring	NOUN
fcis-9730	37	23	,	,	PUNCT
fcis-9730	37	24	most	most	ADV
fcis-9730	37	25	likely	likely	ADJ
fcis-9730	37	26	object	object	VERB
fcis-9730	37	27	for	for	ADP
fcis-9730	37	28	prediction	prediction	NOUN
fcis-9730	37	29	.	.	PUNCT
fcis-9730	38	1	since	since	SCONJ
fcis-9730	38	2	each	each	DET
fcis-9730	38	3	grid	grid	NOUN
fcis-9730	38	4	cell	cell	NOUN
fcis-9730	38	5	can	can	AUX
fcis-9730	38	6	only	only	ADV
fcis-9730	38	7	18	18	NUM
fcis-9730	38	8	make	make	VERB
fcis-9730	38	9	local	local	ADJ
fcis-9730	38	10	predictions	prediction	NOUN
fcis-9730	38	11	,	,	PUNCT
fcis-9730	38	12	it	it	PRON
fcis-9730	38	13	can	can	AUX
fcis-9730	38	14	avoid	avoid	VERB
fcis-9730	38	15	falling	fall	VERB
fcis-9730	38	16	between	between	ADP
fcis-9730	38	17	several	several	ADJ
fcis-9730	38	18	cells	cell	NOUN
fcis-9730	38	19	with	with	ADP
fcis-9730	38	20	good	good	ADJ
fcis-9730	38	21	confidence	confidence	NOUN
fcis-9730	38	22	at	at	ADP
fcis-9730	38	23	the	the	DET
fcis-9730	38	24	same	same	ADJ
fcis-9730	38	25	time	time	NOUN
fcis-9730	38	26	,	,	PUNCT
fcis-9730	38	27	but	but	CCONJ
fcis-9730	38	28	it	it	PRON
fcis-9730	38	29	is	be	AUX
fcis-9730	38	30	not	not	PART
fcis-9730	38	31	a	a	DET
fcis-9730	38	32	matter	matter	NOUN
fcis-9730	38	33	of	of	ADP
fcis-9730	38	34	grabbing	grab	VERB
fcis-9730	38	35	the	the	DET
fcis-9730	38	36	things	thing	NOUN
fcis-9730	38	37	needed	need	VERB
fcis-9730	38	38	.	.	PUNCT
fcis-9730	39	1	[	[	X
fcis-9730	39	2	10	10	NUM
fcis-9730	39	3	]	]	PUNCT
fcis-9730	39	4	with	with	ADP
fcis-9730	39	5	the	the	DET
fcis-9730	39	6	further	further	ADJ
fcis-9730	39	7	development	development	NOUN
fcis-9730	39	8	of	of	ADP
fcis-9730	39	9	neural	neural	ADJ
fcis-9730	39	10	networks	network	NOUN
fcis-9730	39	11	,	,	PUNCT
fcis-9730	39	12	the	the	DET
fcis-9730	39	13	yolo	yolo	ADJ
fcis-9730	39	14	series	series	NOUN
fcis-9730	39	15	of	of	ADP
fcis-9730	39	16	object	object	NOUN
fcis-9730	39	17	detection	detection	NOUN
fcis-9730	39	18	shows	show	VERB
fcis-9730	39	19	good	good	ADJ
fcis-9730	39	20	performance	performance	NOUN
fcis-9730	39	21	,	,	PUNCT
fcis-9730	39	22	and	and	CCONJ
fcis-9730	39	23	many	many	ADJ
fcis-9730	39	24	branches	branch	NOUN
fcis-9730	39	25	are	be	AUX
fcis-9730	39	26	derived	derive	VERB
fcis-9730	39	27	,	,	PUNCT
fcis-9730	39	28	as	as	SCONJ
fcis-9730	39	29	shown	show	VERB
fcis-9730	39	30	in	in	ADP
fcis-9730	39	31	figure	figure	NOUN
fcis-9730	39	32	1	1	NUM
fcis-9730	39	33	below	below	ADV
fcis-9730	39	34	.	.	PUNCT
fcis-9730	40	1	figure	figure	NOUN
fcis-9730	40	2	1	1	NUM
fcis-9730	40	3	.	.	PUNCT
fcis-9730	40	4	part	part	NOUN
fcis-9730	40	5	of	of	ADP
fcis-9730	40	6	the	the	DET
fcis-9730	40	7	yolo	yolo	PROPN
fcis-9730	40	8	series	series	PROPN
fcis-9730	40	9	products	product	NOUN
fcis-9730	40	10	2.2	2.2	NUM
fcis-9730	40	11	.	.	PUNCT
fcis-9730	41	1	yolov2	yolov2	PROPN
fcis-9730	41	2	compared	compare	VERB
fcis-9730	41	3	with	with	ADP
fcis-9730	41	4	yolov1	yolov1	NOUN
fcis-9730	41	5	algorithm	algorithm	NOUN
fcis-9730	41	6	,	,	PUNCT
fcis-9730	41	7	yolov2	yolov2	PROPN
fcis-9730	41	8	adopts	adopt	VERB
fcis-9730	41	9	darknet19	darknet19	NOUN
fcis-9730	41	10	network	network	NOUN
fcis-9730	41	11	,	,	PUNCT
fcis-9730	41	12	which	which	PRON
fcis-9730	41	13	includes	include	VERB
fcis-9730	41	14	19	19	NUM
fcis-9730	41	15	convolution	convolution	NOUN
fcis-9730	41	16	layers	layer	NOUN
fcis-9730	41	17	and	and	CCONJ
fcis-9730	41	18	5	5	NUM
fcis-9730	41	19	max	max	NOUN
fcis-9730	41	20	pooling	pool	VERB
fcis-9730	41	21	layers	layer	NOUN
fcis-9730	41	22	.	.	PUNCT
fcis-9730	42	1	3x3	3x3	NUM
fcis-9730	42	2	and	and	CCONJ
fcis-9730	42	3	1x1	1x1	NUM
fcis-9730	42	4	convolution	convolution	NOUN
fcis-9730	42	5	are	be	AUX
fcis-9730	42	6	mainly	mainly	ADV
fcis-9730	42	7	adopted	adopt	VERB
fcis-9730	42	8	,	,	PUNCT
fcis-9730	42	9	which	which	PRON
fcis-9730	42	10	are	be	AUX
fcis-9730	42	11	two	two	NUM
fcis-9730	42	12	kinds	kind	NOUN
fcis-9730	42	13	of	of	ADP
fcis-9730	42	14	convolution	convolution	NOUN
fcis-9730	42	15	layers	layer	NOUN
fcis-9730	42	16	.	.	PUNCT
fcis-9730	43	1	the	the	DET
fcis-9730	43	2	1x1	1x1	NUM
fcis-9730	43	3	convolution	convolution	NOUN
fcis-9730	43	4	here	here	ADV
fcis-9730	43	5	can	can	AUX
fcis-9730	43	6	compress	compress	VERB
fcis-9730	43	7	the	the	DET
fcis-9730	43	8	channel	channel	NOUN
fcis-9730	43	9	number	number	NOUN
fcis-9730	43	10	of	of	ADP
fcis-9730	43	11	feature	feature	NOUN
fcis-9730	43	12	map	map	NOUN
fcis-9730	43	13	.	.	PUNCT
fcis-9730	44	1	in	in	ADP
fcis-9730	44	2	this	this	DET
fcis-9730	44	3	way	way	NOUN
fcis-9730	44	4	,	,	PUNCT
fcis-9730	44	5	model	model	NOUN
fcis-9730	44	6	computation	computation	NOUN
fcis-9730	44	7	and	and	CCONJ
fcis-9730	44	8	parameters	parameter	NOUN
fcis-9730	44	9	can	can	AUX
fcis-9730	44	10	be	be	AUX
fcis-9730	44	11	reduced	reduce	VERB
fcis-9730	44	12	;	;	PUNCT
fcis-9730	44	13	the	the	DET
fcis-9730	44	14	introduction	introduction	NOUN
fcis-9730	44	15	of	of	ADP
fcis-9730	44	16	anchor	anchor	NOUN
fcis-9730	44	17	and	and	CCONJ
fcis-9730	44	18	k	k	NOUN
fcis-9730	44	19	-	-	PUNCT
fcis-9730	44	20	means	means	NOUN
fcis-9730	44	21	clustering	clustering	NOUN
fcis-9730	44	22	improved	improve	VERB
fcis-9730	44	23	the	the	DET
fcis-9730	44	24	recall	recall	NOUN
fcis-9730	44	25	rate	rate	NOUN
fcis-9730	44	26	;	;	PUNCT
fcis-9730	44	27	in	in	ADP
fcis-9730	44	28	order	order	NOUN
fcis-9730	44	29	to	to	PART
fcis-9730	44	30	prevent	prevent	VERB
fcis-9730	44	31	overfitting	overfitting	NOUN
fcis-9730	44	32	,	,	PUNCT
fcis-9730	44	33	bn	bn	ADP
fcis-9730	44	34	layer	layer	NOUN
fcis-9730	44	35	is	be	AUX
fcis-9730	44	36	used	use	VERB
fcis-9730	44	37	after	after	ADP
fcis-9730	44	38	each	each	DET
fcis-9730	44	39	convolutional	convolutional	ADJ
fcis-9730	44	40	layer	layer	NOUN
fcis-9730	44	41	to	to	PART
fcis-9730	44	42	speed	speed	VERB
fcis-9730	44	43	up	up	ADP
fcis-9730	44	44	model	model	NOUN
fcis-9730	44	45	convergence	convergence	NOUN
fcis-9730	44	46	.	.	PUNCT
fcis-9730	45	1	finally	finally	ADV
fcis-9730	45	2	,	,	PUNCT
fcis-9730	45	3	global	global	ADJ
fcis-9730	45	4	avg	avg	NOUN
fcis-9730	45	5	pool	pool	PROPN
fcis-9730	45	6	is	be	AUX
fcis-9730	45	7	used	use	VERB
fcis-9730	45	8	for	for	ADP
fcis-9730	45	9	prediction	prediction	NOUN
fcis-9730	45	10	.	.	PUNCT
fcis-9730	46	1	the	the	DET
fcis-9730	46	2	feature	feature	NOUN
fcis-9730	46	3	fusion	fusion	NOUN
fcis-9730	46	4	module	module	NOUN
fcis-9730	46	5	(	(	PUNCT
fcis-9730	46	6	passthrough	passthrough	NOUN
fcis-9730	46	7	)	)	PUNCT
fcis-9730	46	8	is	be	AUX
fcis-9730	46	9	introduced	introduce	VERB
fcis-9730	46	10	to	to	PART
fcis-9730	46	11	fuse	fuse	VERB
fcis-9730	46	12	fine	fine	ADV
fcis-9730	46	13	-	-	PUNCT
fcis-9730	46	14	grained	grain	VERB
fcis-9730	46	15	features	feature	NOUN
fcis-9730	46	16	.	.	PUNCT
fcis-9730	47	1	by	by	ADP
fcis-9730	47	2	using	use	VERB
fcis-9730	47	3	yolov2	yolov2	PROPN
fcis-9730	47	4	,	,	PUNCT
fcis-9730	47	5	the	the	DET
fcis-9730	47	6	map	map	NOUN
fcis-9730	47	7	value	value	NOUN
fcis-9730	47	8	of	of	ADP
fcis-9730	47	9	the	the	DET
fcis-9730	47	10	model	model	NOUN
fcis-9730	47	11	is	be	AUX
fcis-9730	47	12	not	not	PART
fcis-9730	47	13	significantly	significantly	ADV
fcis-9730	47	14	improved	improve	VERB
fcis-9730	47	15	,	,	PUNCT
fcis-9730	47	16	but	but	CCONJ
fcis-9730	47	17	the	the	DET
fcis-9730	47	18	calculation	calculation	NOUN
fcis-9730	47	19	amount	amount	NOUN
fcis-9730	47	20	is	be	AUX
fcis-9730	47	21	reduced	reduce	VERB
fcis-9730	47	22	.	.	PUNCT
fcis-9730	48	1	however	however	ADV
fcis-9730	48	2	,	,	PUNCT
fcis-9730	48	3	yolov2	yolov2	PROPN
fcis-9730	48	4	will	will	AUX
fcis-9730	48	5	lose	lose	VERB
fcis-9730	48	6	small	small	ADJ
fcis-9730	48	7	targets	target	NOUN
fcis-9730	48	8	,	,	PUNCT
fcis-9730	48	9	because	because	SCONJ
fcis-9730	48	10	the	the	DET
fcis-9730	48	11	resolution	resolution	NOUN
fcis-9730	48	12	of	of	ADP
fcis-9730	48	13	the	the	DET
fcis-9730	48	14	feature	feature	NOUN
fcis-9730	48	15	map	map	NOUN
fcis-9730	48	16	is	be	AUX
fcis-9730	48	17	subsampled	subsample	VERB
fcis-9730	48	18	during	during	ADP
fcis-9730	48	19	the	the	DET
fcis-9730	48	20	design	design	NOUN
fcis-9730	48	21	,	,	PUNCT
fcis-9730	48	22	yolov2	yolov2	PROPN
fcis-9730	48	23	can	can	AUX
fcis-9730	48	24	not	not	PART
fcis-9730	48	25	detect	detect	VERB
fcis-9730	48	26	small	small	ADJ
fcis-9730	48	27	targets	target	NOUN
fcis-9730	48	28	well	well	ADV
fcis-9730	48	29	.	.	PUNCT
fcis-9730	49	1	[	[	X
fcis-9730	49	2	11	11	NUM
fcis-9730	49	3	]	]	PUNCT
fcis-9730	49	4	the	the	DET
fcis-9730	49	5	cnn	cnn	PROPN
fcis-9730	49	6	infrastructure	infrastructure	NOUN
fcis-9730	49	7	used	use	VERB
fcis-9730	49	8	in	in	ADP
fcis-9730	49	9	yolov2	yolov2	PROPN
fcis-9730	49	10	is	be	AUX
fcis-9730	49	11	relatively	relatively	ADV
fcis-9730	49	12	simple	simple	ADJ
fcis-9730	49	13	and	and	CCONJ
fcis-9730	49	14	does	do	AUX
fcis-9730	49	15	not	not	PART
fcis-9730	49	16	use	use	VERB
fcis-9730	49	17	rpn	rpn	NOUN
fcis-9730	49	18	network	network	NOUN
fcis-9730	49	19	,	,	PUNCT
fcis-9730	49	20	which	which	PRON
fcis-9730	49	21	leads	lead	VERB
fcis-9730	49	22	to	to	ADP
fcis-9730	49	23	its	its	PRON
fcis-9730	49	24	accuracy	accuracy	NOUN
fcis-9730	49	25	gap	gap	NOUN
fcis-9730	49	26	compared	compare	VERB
fcis-9730	49	27	with	with	ADP
fcis-9730	49	28	some	some	DET
fcis-9730	49	29	advanced	advanced	ADJ
fcis-9730	49	30	object	object	NOUN
fcis-9730	49	31	detection	detection	NOUN
fcis-9730	49	32	algorithms	algorithm	NOUN
fcis-9730	49	33	.	.	PUNCT
fcis-9730	50	1	2.3	2.3	NUM
fcis-9730	50	2	.	.	PUNCT
fcis-9730	51	1	yolov3	yolov3	PROPN
fcis-9730	51	2	compared	compare	VERB
fcis-9730	51	3	with	with	ADP
fcis-9730	51	4	yolov2	yolov2	PROPN
fcis-9730	51	5	algorithm	algorithm	NOUN
fcis-9730	51	6	,	,	PUNCT
fcis-9730	51	7	yolov3	yolov3	PROPN
fcis-9730	51	8	adopts	adopt	VERB
fcis-9730	51	9	darknet-53	darknet-53	PROPN
fcis-9730	51	10	as	as	ADP
fcis-9730	51	11	the	the	DET
fcis-9730	51	12	backbone	backbone	NOUN
fcis-9730	51	13	network	network	NOUN
fcis-9730	51	14	.	.	PUNCT
fcis-9730	52	1	compared	compare	VERB
fcis-9730	52	2	with	with	ADP
fcis-9730	52	3	darknet-19	darknet-19	PROPN
fcis-9730	52	4	,	,	PUNCT
fcis-9730	52	5	yolov3	yolov3	PROPN
fcis-9730	52	6	has	have	VERB
fcis-9730	52	7	deeper	deep	ADJ
fcis-9730	52	8	network	network	NOUN
fcis-9730	52	9	,	,	PUNCT
fcis-9730	52	10	more	more	ADJ
fcis-9730	52	11	parameters	parameter	NOUN
fcis-9730	52	12	and	and	CCONJ
fcis-9730	52	13	more	more	ADV
fcis-9730	52	14	adequate	adequate	ADJ
fcis-9730	52	15	training	training	NOUN
fcis-9730	52	16	,	,	PUNCT
fcis-9730	52	17	so	so	CCONJ
fcis-9730	52	18	the	the	DET
fcis-9730	52	19	detection	detection	NOUN
fcis-9730	52	20	performance	performance	NOUN
fcis-9730	52	21	is	be	AUX
fcis-9730	52	22	better	well	ADJ
fcis-9730	52	23	.	.	PUNCT
fcis-9730	53	1	yolov3	yolov3	PROPN
fcis-9730	53	2	adopts	adopt	VERB
fcis-9730	53	3	multi	multi	ADJ
fcis-9730	53	4	-	-	ADJ
fcis-9730	53	5	scale	scale	ADJ
fcis-9730	53	6	prediction	prediction	NOUN
fcis-9730	53	7	,	,	PUNCT
fcis-9730	53	8	which	which	PRON
fcis-9730	53	9	can	can	AUX
fcis-9730	53	10	improve	improve	VERB
fcis-9730	53	11	the	the	DET
fcis-9730	53	12	detection	detection	NOUN
fcis-9730	53	13	accuracy	accuracy	NOUN
fcis-9730	53	14	of	of	ADP
fcis-9730	53	15	small	small	ADJ
fcis-9730	53	16	targets	target	NOUN
fcis-9730	53	17	by	by	ADP
fcis-9730	53	18	detecting	detect	VERB
fcis-9730	53	19	objects	object	NOUN
fcis-9730	53	20	on	on	ADP
fcis-9730	53	21	different	different	ADJ
fcis-9730	53	22	scales	scale	NOUN
fcis-9730	53	23	.	.	PUNCT
fcis-9730	54	1	feature	feature	NOUN
fcis-9730	54	2	pyramids	pyramid	NOUN
fcis-9730	54	3	with	with	ADP
fcis-9730	54	4	different	different	ADJ
fcis-9730	54	5	convolution	convolution	NOUN
fcis-9730	54	6	kernel	kernel	NOUN
fcis-9730	54	7	sizes	size	NOUN
fcis-9730	54	8	are	be	AUX
fcis-9730	54	9	used	use	VERB
fcis-9730	54	10	to	to	PART
fcis-9730	54	11	detect	detect	VERB
fcis-9730	54	12	objects	object	NOUN
fcis-9730	54	13	of	of	ADP
fcis-9730	54	14	different	different	ADJ
fcis-9730	54	15	sizes	size	NOUN
fcis-9730	54	16	,	,	PUNCT
fcis-9730	54	17	which	which	PRON
fcis-9730	54	18	overcomes	overcome	VERB
fcis-9730	54	19	the	the	DET
fcis-9730	54	20	shortcomings	shortcoming	NOUN
fcis-9730	54	21	of	of	ADP
fcis-9730	54	22	yolov2	yolov2	PROPN
fcis-9730	54	23	in	in	ADP
fcis-9730	54	24	small	small	ADJ
fcis-9730	54	25	target	target	NOUN
fcis-9730	54	26	detection	detection	NOUN
fcis-9730	54	27	.	.	PUNCT
fcis-9730	55	1	yolov3	yolov3	PROPN
fcis-9730	55	2	uses	use	VERB
fcis-9730	55	3	three	three	NUM
fcis-9730	55	4	anchor	anchor	NOUN
fcis-9730	55	5	frames	frame	NOUN
fcis-9730	55	6	of	of	ADP
fcis-9730	55	7	different	different	ADJ
fcis-9730	55	8	sizes	size	NOUN
fcis-9730	55	9	,	,	PUNCT
fcis-9730	55	10	which	which	PRON
fcis-9730	55	11	can	can	AUX
fcis-9730	55	12	better	well	ADV
fcis-9730	55	13	adapt	adapt	VERB
fcis-9730	55	14	to	to	ADP
fcis-9730	55	15	objects	object	NOUN
fcis-9730	55	16	of	of	ADP
fcis-9730	55	17	different	different	ADJ
fcis-9730	55	18	sizes	size	NOUN
fcis-9730	55	19	and	and	CCONJ
fcis-9730	55	20	aspect	aspect	NOUN
fcis-9730	55	21	ratios	ratio	NOUN
fcis-9730	55	22	;	;	PUNCT
fcis-9730	55	23	and	and	CCONJ
fcis-9730	55	24	the	the	DET
fcis-9730	55	25	introduction	introduction	NOUN
fcis-9730	55	26	of	of	ADP
fcis-9730	55	27	a	a	DET
fcis-9730	55	28	residual	residual	ADJ
fcis-9730	55	29	network	network	NOUN
fcis-9730	55	30	structure	structure	NOUN
fcis-9730	55	31	,	,	PUNCT
fcis-9730	55	32	which	which	PRON
fcis-9730	55	33	is	be	AUX
fcis-9730	55	34	able	able	ADJ
fcis-9730	55	35	to	to	PART
fcis-9730	55	36	learn	learn	VERB
fcis-9730	55	37	features	feature	NOUN
fcis-9730	55	38	better	well	ADV
fcis-9730	55	39	and	and	CCONJ
fcis-9730	55	40	speed	speed	VERB
fcis-9730	55	41	up	up	ADP
fcis-9730	55	42	training	training	NOUN
fcis-9730	55	43	.	.	PUNCT
fcis-9730	56	1	[	[	X
fcis-9730	56	2	12	12	NUM
fcis-9730	56	3	]	]	PUNCT
fcis-9730	56	4	however	however	ADV
fcis-9730	56	5	,	,	PUNCT
fcis-9730	56	6	yolov3	yolov3	PROPN
fcis-9730	56	7	's	's	PART
fcis-9730	56	8	inference	inference	NOUN
fcis-9730	56	9	is	be	AUX
fcis-9730	56	10	slower	slow	ADJ
fcis-9730	56	11	because	because	SCONJ
fcis-9730	56	12	the	the	DET
fcis-9730	56	13	detection	detection	NOUN
fcis-9730	56	14	process	process	NOUN
fcis-9730	56	15	needs	need	VERB
fcis-9730	56	16	to	to	PART
fcis-9730	56	17	run	run	VERB
fcis-9730	56	18	multiple	multiple	ADJ
fcis-9730	56	19	convolutional	convolutional	ADJ
fcis-9730	56	20	layers	layer	NOUN
fcis-9730	56	21	;	;	PUNCT
fcis-9730	56	22	yolov3	yolov3	PROPN
fcis-9730	56	23	has	have	VERB
fcis-9730	56	24	problems	problem	NOUN
fcis-9730	56	25	with	with	ADP
fcis-9730	56	26	inaccurate	inaccurate	ADJ
fcis-9730	56	27	object	object	NOUN
fcis-9730	56	28	positions	position	NOUN
fcis-9730	56	29	.	.	PUNCT
fcis-9730	57	1	2.4	2.4	NUM
fcis-9730	57	2	.	.	PUNCT
fcis-9730	57	3	yolov4	yolov4	NOUN
fcis-9730	57	4	compared	compare	VERB
fcis-9730	57	5	with	with	ADP
fcis-9730	57	6	yolov3	yolov3	PROPN
fcis-9730	57	7	algorithm	algorithm	PROPN
fcis-9730	57	8	,	,	PUNCT
fcis-9730	57	9	yolov4	yolov4	PROPN
fcis-9730	57	10	combines	combine	VERB
fcis-9730	57	11	many	many	ADJ
fcis-9730	57	12	previous	previous	ADJ
fcis-9730	57	13	research	research	NOUN
fcis-9730	57	14	techniques	technique	NOUN
fcis-9730	57	15	,	,	PUNCT
fcis-9730	57	16	combines	combine	VERB
fcis-9730	57	17	them	they	PRON
fcis-9730	57	18	with	with	ADP
fcis-9730	57	19	appropriate	appropriate	ADJ
fcis-9730	57	20	innovative	innovative	ADJ
fcis-9730	57	21	algorithms	algorithm	NOUN
fcis-9730	57	22	,	,	PUNCT
fcis-9730	57	23	and	and	CCONJ
fcis-9730	57	24	achieves	achieve	VERB
fcis-9730	57	25	a	a	DET
fcis-9730	57	26	perfect	perfect	ADJ
fcis-9730	57	27	balance	balance	NOUN
fcis-9730	57	28	between	between	ADP
fcis-9730	57	29	speed	speed	NOUN
fcis-9730	57	30	and	and	CCONJ
fcis-9730	57	31	accuracy	accuracy	NOUN
fcis-9730	57	32	compared	compare	VERB
fcis-9730	57	33	with	with	ADP
fcis-9730	57	34	the	the	DET
fcis-9730	57	35	previous	previous	ADJ
fcis-9730	57	36	yolo	yolo	ADJ
fcis-9730	57	37	series	series	NOUN
fcis-9730	57	38	.	.	PUNCT
fcis-9730	58	1	however	however	ADV
fcis-9730	58	2	,	,	PUNCT
fcis-9730	58	3	yolov4	yolov4	PROPN
fcis-9730	58	4	is	be	AUX
fcis-9730	58	5	not	not	PART
fcis-9730	58	6	much	much	ADV
fcis-9730	58	7	different	different	ADJ
fcis-9730	58	8	from	from	ADP
fcis-9730	58	9	yolov3	yolov3	PROPN
fcis-9730	58	10	in	in	ADP
fcis-9730	58	11	essence	essence	NOUN
fcis-9730	58	12	.	.	PUNCT
fcis-9730	59	1	yolov4	yolov4	PROPN
fcis-9730	59	2	uses	use	VERB
fcis-9730	59	3	cspdarknet53	cspdarknet53	NOUN
fcis-9730	59	4	as	as	ADP
fcis-9730	59	5	the	the	DET
fcis-9730	59	6	backbone	backbone	NOUN
fcis-9730	59	7	network	network	NOUN
fcis-9730	59	8	for	for	ADP
fcis-9730	59	9	feature	feature	NOUN
fcis-9730	59	10	extraction	extraction	NOUN
fcis-9730	59	11	,	,	PUNCT
fcis-9730	59	12	which	which	PRON
fcis-9730	59	13	reduces	reduce	VERB
fcis-9730	59	14	the	the	DET
fcis-9730	59	15	consumption	consumption	NOUN
fcis-9730	59	16	of	of	ADP
fcis-9730	59	17	computing	compute	VERB
fcis-9730	59	18	resources	resource	NOUN
fcis-9730	59	19	.	.	PUNCT
fcis-9730	60	1	spp	spp	NOUN
fcis-9730	60	2	structure	structure	NOUN
fcis-9730	60	3	and	and	CCONJ
fcis-9730	60	4	residual	residual	ADJ
fcis-9730	60	5	connection	connection	NOUN
fcis-9730	60	6	network	network	NOUN
fcis-9730	60	7	are	be	AUX
fcis-9730	60	8	introduced	introduce	VERB
fcis-9730	60	9	to	to	PART
fcis-9730	60	10	accelerate	accelerate	VERB
fcis-9730	60	11	and	and	CCONJ
fcis-9730	60	12	optimize	optimize	VERB
fcis-9730	60	13	the	the	DET
fcis-9730	60	14	convolution	convolution	NOUN
fcis-9730	60	15	process	process	NOUN
fcis-9730	60	16	,	,	PUNCT
fcis-9730	60	17	thus	thus	ADV
fcis-9730	60	18	improving	improve	VERB
fcis-9730	60	19	the	the	DET
fcis-9730	60	20	prediction	prediction	NOUN
fcis-9730	60	21	performance	performance	NOUN
fcis-9730	60	22	and	and	CCONJ
fcis-9730	60	23	speed	speed	NOUN
fcis-9730	60	24	;	;	PUNCT
fcis-9730	60	25	postprocessing	postprocesse	VERB
fcis-9730	60	26	optimization	optimization	NOUN
fcis-9730	60	27	measures	measure	NOUN
fcis-9730	60	28	such	such	ADJ
fcis-9730	60	29	as	as	ADP
fcis-9730	60	30	dynamic	dynamic	ADJ
fcis-9730	60	31	adjustment	adjustment	NOUN
fcis-9730	60	32	of	of	ADP
fcis-9730	60	33	iou	iou	NOUN
fcis-9730	60	34	threshold	threshold	NOUN
fcis-9730	60	35	and	and	CCONJ
fcis-9730	60	36	logistic	logistic	ADJ
fcis-9730	60	37	activation	activation	NOUN
fcis-9730	60	38	instead	instead	ADV
fcis-9730	60	39	of	of	ADP
fcis-9730	60	40	softmax	softmax	NOUN
fcis-9730	60	41	were	be	AUX
fcis-9730	60	42	added	add	VERB
fcis-9730	60	43	to	to	PART
fcis-9730	60	44	optimize	optimize	VERB
fcis-9730	60	45	the	the	DET
fcis-9730	60	46	accuracy	accuracy	NOUN
fcis-9730	60	47	and	and	CCONJ
fcis-9730	60	48	robustness	robustness	NOUN
fcis-9730	60	49	of	of	ADP
fcis-9730	60	50	detection	detection	NOUN
fcis-9730	60	51	results	result	NOUN
fcis-9730	60	52	.	.	PUNCT
fcis-9730	61	1	the	the	DET
fcis-9730	61	2	opencv	opencv	PROPN
fcis-9730	61	3	dnn	dnn	PROPN
fcis-9730	61	4	module	module	NOUN
fcis-9730	61	5	is	be	AUX
fcis-9730	61	6	used	use	VERB
fcis-9730	61	7	to	to	PART
fcis-9730	61	8	optimize	optimize	VERB
fcis-9730	61	9	the	the	DET
fcis-9730	61	10	forward	forward	ADJ
fcis-9730	61	11	calculation	calculation	NOUN
fcis-9730	61	12	performance	performance	NOUN
fcis-9730	61	13	,	,	PUNCT
fcis-9730	61	14	and	and	CCONJ
fcis-9730	61	15	the	the	DET
fcis-9730	61	16	detection	detection	NOUN
fcis-9730	61	17	speed	speed	NOUN
fcis-9730	61	18	is	be	AUX
fcis-9730	61	19	greatly	greatly	ADV
fcis-9730	61	20	improved	improve	VERB
fcis-9730	61	21	on	on	ADP
fcis-9730	61	22	the	the	DET
fcis-9730	61	23	premise	premise	NOUN
fcis-9730	61	24	of	of	ADP
fcis-9730	61	25	ensuring	ensure	VERB
fcis-9730	61	26	the	the	DET
fcis-9730	61	27	detection	detection	NOUN
fcis-9730	61	28	accuracy	accuracy	NOUN
fcis-9730	61	29	.	.	PUNCT
fcis-9730	62	1	however	however	ADV
fcis-9730	62	2	,	,	PUNCT
fcis-9730	62	3	the	the	DET
fcis-9730	62	4	computing	compute	VERB
fcis-9730	62	5	resource	resource	NOUN
fcis-9730	62	6	consumption	consumption	NOUN
fcis-9730	62	7	is	be	AUX
fcis-9730	62	8	high	high	ADJ
fcis-9730	62	9	,	,	PUNCT
fcis-9730	62	10	because	because	SCONJ
fcis-9730	62	11	the	the	DET
fcis-9730	62	12	yolov4	yolov4	PROPN
fcis-9730	62	13	model	model	NOUN
fcis-9730	62	14	is	be	AUX
fcis-9730	62	15	relatively	relatively	ADV
fcis-9730	62	16	large	large	ADJ
fcis-9730	62	17	and	and	CCONJ
fcis-9730	62	18	requires	require	VERB
fcis-9730	62	19	a	a	DET
fcis-9730	62	20	lot	lot	NOUN
fcis-9730	62	21	of	of	ADP
fcis-9730	62	22	calculation	calculation	NOUN
fcis-9730	62	23	.	.	PUNCT
fcis-9730	63	1	compared	compare	VERB
fcis-9730	63	2	with	with	ADP
fcis-9730	63	3	some	some	DET
fcis-9730	63	4	other	other	ADJ
fcis-9730	63	5	networks	network	NOUN
fcis-9730	63	6	,	,	PUNCT
fcis-9730	63	7	yolov4	yolov4	PROPN
fcis-9730	63	8	is	be	AUX
fcis-9730	63	9	more	more	ADV
fcis-9730	63	10	complex	complex	ADJ
fcis-9730	63	11	.	.	PUNCT
fcis-9730	64	1	although	although	SCONJ
fcis-9730	64	2	yolov4	yolov4	NOUN
fcis-9730	64	3	has	have	VERB
fcis-9730	64	4	great	great	ADJ
fcis-9730	64	5	advantages	advantage	NOUN
fcis-9730	64	6	in	in	ADP
fcis-9730	64	7	performance	performance	NOUN
fcis-9730	64	8	,	,	PUNCT
fcis-9730	64	9	it	it	PRON
fcis-9730	64	10	is	be	AUX
fcis-9730	64	11	also	also	ADV
fcis-9730	64	12	more	more	ADV
fcis-9730	64	13	complex	complex	ADJ
fcis-9730	64	14	,	,	PUNCT
fcis-9730	64	15	requiring	require	VERB
fcis-9730	64	16	more	more	ADJ
fcis-9730	64	17	training	training	NOUN
fcis-9730	64	18	time	time	NOUN
fcis-9730	64	19	and	and	CCONJ
fcis-9730	64	20	more	more	ADJ
fcis-9730	64	21	network	network	NOUN
fcis-9730	64	22	parameters	parameter	NOUN
fcis-9730	64	23	.	.	PUNCT
fcis-9730	65	1	poor	poor	ADJ
fcis-9730	65	2	effect	effect	NOUN
fcis-9730	65	3	on	on	ADP
fcis-9730	65	4	s	s	NOUN
fcis-9730	65	5	mall	mall	NOUN
fcis-9730	65	6	target	target	NOUN
fcis-9730	65	7	detection	detection	NOUN
fcis-9730	65	8	:	:	PUNCT
fcis-9730	65	9	because	because	SCONJ
fcis-9730	65	10	yolov4	yolov4	PROPN
fcis-9730	65	11	uses	use	VERB
fcis-9730	65	12	a	a	DET
fcis-9730	65	13	larger	large	ADJ
fcis-9730	65	14	anchor	anchor	NOUN
fcis-9730	65	15	box	box	NOUN
fcis-9730	65	16	and	and	CCONJ
fcis-9730	65	17	higher	high	ADJ
fcis-9730	65	18	prediction	prediction	NOUN
fcis-9730	65	19	resolution	resolution	NOUN
fcis-9730	65	20	,	,	PUNCT
fcis-9730	65	21	its	its	PRON
fcis-9730	65	22	effect	effect	NOUN
fcis-9730	65	23	on	on	ADP
fcis-9730	65	24	small	small	ADJ
fcis-9730	65	25	target	target	NOUN
fcis-9730	65	26	detection	detection	NOUN
fcis-9730	65	27	is	be	AUX
fcis-9730	65	28	worse	bad	ADJ
fcis-9730	65	29	than	than	ADP
fcis-9730	65	30	that	that	PRON
fcis-9730	65	31	of	of	ADP
fcis-9730	65	32	other	other	ADJ
fcis-9730	65	33	algorithms	algorithm	NOUN
fcis-9730	65	34	.	.	PUNCT
fcis-9730	66	1	2.5	2.5	NUM
fcis-9730	66	2	.	.	PUNCT
fcis-9730	67	1	yolov5	yolov5	NOUN
fcis-9730	67	2	compared	compare	VERB
fcis-9730	67	3	with	with	ADP
fcis-9730	67	4	yolov4	yolov4	PROPN
fcis-9730	67	5	algorithm	algorithm	PROPN
fcis-9730	67	6	,	,	PUNCT
fcis-9730	67	7	yolov5	yolov5	NOUN
fcis-9730	67	8	is	be	AUX
fcis-9730	67	9	smaller	small	ADJ
fcis-9730	67	10	than	than	ADP
fcis-9730	67	11	yolov4	yolov4	PROPN
fcis-9730	67	12	model	model	NOUN
fcis-9730	67	13	,	,	PUNCT
fcis-9730	67	14	so	so	SCONJ
fcis-9730	67	15	it	it	PRON
fcis-9730	67	16	has	have	VERB
fcis-9730	67	17	higher	high	ADJ
fcis-9730	67	18	computing	computing	NOUN
fcis-9730	67	19	efficiency	efficiency	NOUN
fcis-9730	67	20	and	and	CCONJ
fcis-9730	67	21	lower	low	ADJ
fcis-9730	67	22	video	video	NOUN
fcis-9730	67	23	memory	memory	NOUN
fcis-9730	67	24	occupation	occupation	NOUN
fcis-9730	67	25	.	.	PUNCT
fcis-9730	68	1	some	some	DET
fcis-9730	68	2	new	new	ADJ
fcis-9730	68	3	evaluation	evaluation	NOUN
fcis-9730	68	4	indexes	index	NOUN
fcis-9730	68	5	are	be	AUX
fcis-9730	68	6	added	add	VERB
fcis-9730	68	7	to	to	ADP
fcis-9730	68	8	yolov5	yolov5	NOUN
fcis-9730	68	9	,	,	PUNCT
fcis-9730	68	10	such	such	ADJ
fcis-9730	68	11	as	as	ADP
fcis-9730	68	12	map@.5	map@.5	PROPN
fcis-9730	68	13	,	,	PUNCT
fcis-9730	68	14	map@.75	map@.75	NOUN
fcis-9730	68	15	,	,	PUNCT
fcis-9730	68	16	etc	etc	X
fcis-9730	68	17	.	.	X
fcis-9730	68	18	,	,	PUNCT
fcis-9730	68	19	which	which	PRON
fcis-9730	68	20	help	help	VERB
fcis-9730	68	21	to	to	PART
fcis-9730	68	22	evaluate	evaluate	VERB
fcis-9730	68	23	the	the	DET
fcis-9730	68	24	model	model	NOUN
fcis-9730	68	25	performance	performance	NOUN
fcis-9730	68	26	in	in	ADP
fcis-9730	68	27	a	a	DET
fcis-9730	68	28	more	more	ADV
fcis-9730	68	29	comprehensive	comprehensive	ADJ
fcis-9730	68	30	way	way	NOUN
fcis-9730	68	31	.	.	PUNCT
fcis-9730	69	1	yolov5	yolov5	NOUN
fcis-9730	69	2	enhances	enhance	VERB
fcis-9730	69	3	the	the	DET
fcis-9730	69	4	learning	learning	NOUN
fcis-9730	69	5	ability	ability	NOUN
fcis-9730	69	6	of	of	ADP
fcis-9730	69	7	cnn	cnn	PROPN
fcis-9730	69	8	,	,	PUNCT
fcis-9730	69	9	which	which	PRON
fcis-9730	69	10	makes	make	VERB
fcis-9730	69	11	it	it	PRON
fcis-9730	69	12	lightweight	lightweight	ADJ
fcis-9730	69	13	while	while	SCONJ
fcis-9730	69	14	maintaining	maintain	VERB
fcis-9730	69	15	its	its	PRON
fcis-9730	69	16	accuracy	accuracy	NOUN
fcis-9730	69	17	in	in	ADP
fcis-9730	69	18	the	the	DET
fcis-9730	69	19	detection	detection	NOUN
fcis-9730	69	20	process	process	NOUN
fcis-9730	69	21	.	.	PUNCT
fcis-9730	70	1	yolov5	yolov5	NOUN
fcis-9730	70	2	can	can	AUX
fcis-9730	70	3	deduce	deduce	VERB
fcis-9730	70	4	effectively	effectively	ADV
fcis-9730	70	5	from	from	ADP
fcis-9730	70	6	single	single	ADJ
fcis-9730	70	7	image	image	NOUN
fcis-9730	70	8	,	,	PUNCT
fcis-9730	70	9	batch	batch	NOUN
fcis-9730	70	10	image	image	NOUN
fcis-9730	70	11	,	,	PUNCT
fcis-9730	70	12	video	video	NOUN
fcis-9730	70	13	and	and	CCONJ
fcis-9730	70	14	even	even	ADV
fcis-9730	70	15	webcam	webcam	AUX
fcis-9730	70	16	port	port	NOUN
fcis-9730	70	17	input	input	NOUN
fcis-9730	70	18	directly	directly	ADV
fcis-9730	70	19	.	.	PUNCT
fcis-9730	71	1	yolov5s	yolov5s	PROPN
fcis-9730	71	2	's	's	PART
fcis-9730	71	3	object	object	NOUN
fcis-9730	71	4	recognition	recognition	NOUN
fcis-9730	71	5	speed	speed	NOUN
fcis-9730	71	6	of	of	ADP
fcis-9730	71	7	up	up	ADP
fcis-9730	71	8	to	to	PART
fcis-9730	71	9	140fps	140fps	NUM
fcis-9730	71	10	is	be	AUX
fcis-9730	71	11	impressive	impressive	ADJ
fcis-9730	71	12	.	.	PUNCT
fcis-9730	72	1	from	from	ADP
fcis-9730	72	2	yolov5n	yolov5n	PROPN
fcis-9730	72	3	to	to	ADP
fcis-9730	72	4	yolov5x	yolov5x	PROPN
fcis-9730	72	5	,	,	PUNCT
fcis-9730	72	6	the	the	DET
fcis-9730	72	7	detection	detection	NOUN
fcis-9730	72	8	accuracy	accuracy	NOUN
fcis-9730	72	9	of	of	ADP
fcis-9730	72	10	these	these	DET
fcis-9730	72	11	five	five	NUM
fcis-9730	72	12	yolov5s	yolov5s	NOUN
fcis-9730	72	13	models	model	NOUN
fcis-9730	72	14	gradually	gradually	ADV
fcis-9730	72	15	increased	increase	VERB
fcis-9730	72	16	,	,	PUNCT
fcis-9730	72	17	while	while	SCONJ
fcis-9730	72	18	the	the	DET
fcis-9730	72	19	detection	detection	NOUN
fcis-9730	72	20	speed	speed	NOUN
fcis-9730	72	21	gradually	gradually	ADV
fcis-9730	72	22	decreased	decrease	VERB
fcis-9730	72	23	.	.	PUNCT
fcis-9730	73	1	[	[	X
fcis-9730	73	2	13	13	NUM
fcis-9730	73	3	]	]	PUNCT
fcis-9730	73	4	however	however	ADV
fcis-9730	73	5	,	,	PUNCT
fcis-9730	73	6	the	the	DET
fcis-9730	73	7	biggest	big	ADJ
fcis-9730	73	8	drawback	drawback	NOUN
fcis-9730	73	9	of	of	ADP
fcis-9730	73	10	yolov5	yolov5	NOUN
fcis-9730	73	11	is	be	AUX
fcis-9730	73	12	its	its	PRON
fcis-9730	73	13	weak	weak	ADJ
fcis-9730	73	14	target	target	NOUN
fcis-9730	73	15	detection	detection	NOUN
fcis-9730	73	16	capability	capability	NOUN
fcis-9730	73	17	.	.	PUNCT
fcis-9730	74	1	in	in	ADP
fcis-9730	74	2	some	some	DET
fcis-9730	74	3	complex	complex	ADJ
fcis-9730	74	4	scenes	scene	NOUN
fcis-9730	74	5	,	,	PUNCT
fcis-9730	74	6	the	the	DET
fcis-9730	74	7	effect	effect	NOUN
fcis-9730	74	8	will	will	AUX
fcis-9730	74	9	be	be	AUX
fcis-9730	74	10	weak	weak	ADJ
fcis-9730	74	11	.	.	PUNCT
fcis-9730	75	1	for	for	ADP
fcis-9730	75	2	non	non	ADJ
fcis-9730	75	3	-	-	ADJ
fcis-9730	75	4	standard	standard	ADJ
fcis-9730	75	5	scenes	scene	NOUN
fcis-9730	75	6	,	,	PUNCT
fcis-9730	75	7	the	the	DET
fcis-9730	75	8	performance	performance	NOUN
fcis-9730	75	9	is	be	AUX
fcis-9730	75	10	poor	poor	ADJ
fcis-9730	75	11	.	.	PUNCT
fcis-9730	76	1	if	if	SCONJ
fcis-9730	76	2	the	the	DET
fcis-9730	76	3	scene	scene	NOUN
fcis-9730	76	4	is	be	AUX
fcis-9730	76	5	special	special	ADJ
fcis-9730	76	6	or	or	CCONJ
fcis-9730	76	7	the	the	DET
fcis-9730	76	8	target	target	NOUN
fcis-9730	76	9	distance	distance	NOUN
fcis-9730	76	10	is	be	AUX
fcis-9730	76	11	relatively	relatively	ADV
fcis-9730	76	12	far	far	ADV
fcis-9730	76	13	,	,	PUNCT
fcis-9730	76	14	the	the	DET
fcis-9730	76	15	accuracy	accuracy	NOUN
fcis-9730	76	16	of	of	ADP
fcis-9730	76	17	yolov5	yolov5	NOUN
fcis-9730	76	18	will	will	AUX
fcis-9730	76	19	be	be	AUX
fcis-9730	76	20	decreased	decrease	VERB
fcis-9730	76	21	and	and	CCONJ
fcis-9730	76	22	the	the	DET
fcis-9730	76	23	performance	performance	NOUN
fcis-9730	76	24	is	be	AUX
fcis-9730	76	25	poor	poor	ADJ
fcis-9730	76	26	.	.	PUNCT
fcis-9730	77	1	shallow	shallow	ADJ
fcis-9730	77	2	network	network	NOUN
fcis-9730	77	3	depth	depth	NOUN
fcis-9730	77	4	.	.	PUNCT
fcis-9730	78	1	compared	compare	VERB
fcis-9730	78	2	with	with	ADP
fcis-9730	78	3	some	some	DET
fcis-9730	78	4	other	other	ADJ
fcis-9730	78	5	target	target	NOUN
fcis-9730	78	6	detection	detection	NOUN
fcis-9730	78	7	algorithms	algorithm	NOUN
fcis-9730	78	8	,	,	PUNCT
fcis-9730	78	9	yolov5	yolov5	NOUN
fcis-9730	78	10	has	have	VERB
fcis-9730	78	11	a	a	DET
fcis-9730	78	12	shallow	shallow	ADJ
fcis-9730	78	13	network	network	NOUN
fcis-9730	78	14	depth	depth	NOUN
fcis-9730	78	15	,	,	PUNCT
fcis-9730	78	16	which	which	PRON
fcis-9730	78	17	will	will	AUX
fcis-9730	78	18	affect	affect	VERB
fcis-9730	78	19	its	its	PRON
fcis-9730	78	20	performance	performance	NOUN
fcis-9730	78	21	.	.	PUNCT
fcis-9730	79	1	2.6	2.6	NUM
fcis-9730	79	2	.	.	PUNCT
fcis-9730	80	1	yolov7	yolov7	PROPN
fcis-9730	80	2	at	at	ADP
fcis-9730	80	3	present	present	ADJ
fcis-9730	80	4	,	,	PUNCT
fcis-9730	80	5	the	the	DET
fcis-9730	80	6	model	model	NOUN
fcis-9730	80	7	accuracy	accuracy	NOUN
fcis-9730	80	8	and	and	CCONJ
fcis-9730	80	9	inference	inference	NOUN
fcis-9730	80	10	performance	performance	NOUN
fcis-9730	80	11	are	be	AUX
fcis-9730	80	12	more	more	ADV
fcis-9730	80	13	balanced	balanced	ADJ
fcis-9730	80	14	is	be	AUX
fcis-9730	80	15	yolov7	yolov7	PROPN
fcis-9730	80	16	model	model	NOUN
fcis-9730	80	17	(	(	PUNCT
fcis-9730	80	18	corresponding	correspond	VERB
fcis-9730	80	19	to	to	ADP
fcis-9730	80	20	the	the	DET
fcis-9730	80	21	open	open	ADJ
fcis-9730	80	22	-	-	PUNCT
fcis-9730	80	23	source	source	NOUN
fcis-9730	80	24	git	git	NOUN
fcis-9730	80	25	version	version	NOUN
fcis-9730	80	26	0.1	0.1	NUM
fcis-9730	80	27	)	)	PUNCT
fcis-9730	80	28	.	.	PUNCT
fcis-9730	81	1	yolov7	yolov7	PROPN
fcis-9730	81	2	is	be	AUX
fcis-9730	81	3	the	the	DET
fcis-9730	81	4	most	most	ADV
fcis-9730	81	5	advanced	advanced	ADJ
fcis-9730	81	6	algorithm	algorithm	NOUN
fcis-9730	81	7	of	of	ADP
fcis-9730	81	8	yolo	yolo	ADJ
fcis-9730	81	9	series	series	NOUN
fcis-9730	81	10	at	at	ADP
fcis-9730	81	11	present	present	NOUN
fcis-9730	81	12	,	,	PUNCT
fcis-9730	81	13	surpassing	surpass	VERB
fcis-9730	81	14	the	the	DET
fcis-9730	81	15	previous	previous	ADJ
fcis-9730	81	16	yolo	yolo	ADJ
fcis-9730	81	17	series	series	NOUN
fcis-9730	81	18	in	in	ADP
fcis-9730	81	19	accuracy	accuracy	NOUN
fcis-9730	81	20	and	and	CCONJ
fcis-9730	81	21	speed	speed	NOUN
fcis-9730	81	22	.	.	PUNCT
fcis-9730	82	1	yolov7	yolov7	PROPN
fcis-9730	82	2	introduces	introduce	VERB
fcis-9730	82	3	the	the	DET
fcis-9730	82	4	resnet50	resnet50	NOUN
fcis-9730	82	5	deep	deep	ADJ
fcis-9730	82	6	residual	residual	ADJ
fcis-9730	82	7	network	network	NOUN
fcis-9730	82	8	,	,	PUNCT
fcis-9730	82	9	replacing	replace	VERB
fcis-9730	82	10	the	the	DET
fcis-9730	82	11	cspdarknet53	cspdarknet53	NOUN
fcis-9730	82	12	of	of	ADP
fcis-9730	82	13	yolov5	yolov5	NOUN
fcis-9730	82	14	;	;	PUNCT
fcis-9730	82	15	to	to	PART
fcis-9730	82	16	improve	improve	VERB
fcis-9730	82	17	the	the	DET
fcis-9730	82	18	generation	generation	NOUN
fcis-9730	82	19	of	of	ADP
fcis-9730	82	20	anchor	anchor	NOUN
fcis-9730	82	21	frame	frame	NOUN
fcis-9730	82	22	,	,	PUNCT
fcis-9730	82	23	yolov7	yolov7	NOUN
fcis-9730	82	24	introduces	introduce	VERB
fcis-9730	82	25	anchor	anchor	NOUN
fcis-9730	82	26	free	free	ADJ
fcis-9730	82	27	frame	frame	NOUN
fcis-9730	82	28	,	,	PUNCT
fcis-9730	82	29	which	which	PRON
fcis-9730	82	30	does	do	AUX
fcis-9730	82	31	not	not	PART
fcis-9730	82	32	require	require	VERB
fcis-9730	82	33	preset	preset	ADJ
fcis-9730	82	34	anchor	anchor	NOUN
fcis-9730	82	35	frame	frame	NOUN
fcis-9730	82	36	,	,	PUNCT
fcis-9730	82	37	and	and	CCONJ
fcis-9730	82	38	can	can	AUX
fcis-9730	82	39	adapt	adapt	VERB
fcis-9730	82	40	to	to	ADP
fcis-9730	82	41	various	various	ADJ
fcis-9730	82	42	scales	scale	NOUN
fcis-9730	82	43	and	and	CCONJ
fcis-9730	82	44	aspect	aspect	VERB
fcis-9730	82	45	ratio	ratio	NOUN
fcis-9730	82	46	targets	target	NOUN
fcis-9730	82	47	;	;	PUNCT
fcis-9730	82	48	the	the	DET
fcis-9730	82	49	multi	multi	ADJ
fcis-9730	82	50	-	-	ADJ
fcis-9730	82	51	scale	scale	ADJ
fcis-9730	82	52	19	19	NUM
fcis-9730	82	53	training	training	NOUN
fcis-9730	82	54	strategy	strategy	NOUN
fcis-9730	82	55	can	can	AUX
fcis-9730	82	56	improve	improve	VERB
fcis-9730	82	57	the	the	DET
fcis-9730	82	58	detection	detection	NOUN
fcis-9730	82	59	ability	ability	NOUN
fcis-9730	82	60	of	of	ADP
fcis-9730	82	61	the	the	DET
fcis-9730	82	62	model	model	NOUN
fcis-9730	82	63	for	for	ADP
fcis-9730	82	64	s	s	NOUN
fcis-9730	82	65	mall	mall	NOUN
fcis-9730	82	66	targets	target	NOUN
fcis-9730	82	67	and	and	CCONJ
fcis-9730	82	68	better	well	ADV
fcis-9730	82	69	adapt	adapt	VERB
fcis-9730	82	70	to	to	ADP
fcis-9730	82	71	the	the	DET
fcis-9730	82	72	change	change	NOUN
fcis-9730	82	73	of	of	ADP
fcis-9730	82	74	image	image	NOUN
fcis-9730	82	75	resolution	resolution	NOUN
fcis-9730	82	76	.	.	PUNCT
fcis-9730	83	1	mosaic	mosaic	ADJ
fcis-9730	83	2	data	datum	NOUN
fcis-9730	83	3	enhancement	enhancement	NOUN
fcis-9730	83	4	method	method	NOUN
fcis-9730	83	5	is	be	AUX
fcis-9730	83	6	introduced	introduce	VERB
fcis-9730	83	7	,	,	PUNCT
fcis-9730	83	8	which	which	PRON
fcis-9730	83	9	can	can	AUX
fcis-9730	83	10	randomly	randomly	ADV
fcis-9730	83	11	combine	combine	VERB
fcis-9730	83	12	images	image	NOUN
fcis-9730	83	13	to	to	PART
fcis-9730	83	14	increase	increase	VERB
fcis-9730	83	15	the	the	DET
fcis-9730	83	16	diversity	diversity	NOUN
fcis-9730	83	17	and	and	CCONJ
fcis-9730	83	18	quantity	quantity	NOUN
fcis-9730	83	19	of	of	ADP
fcis-9730	83	20	training	training	NOUN
fcis-9730	83	21	samples	sample	NOUN
fcis-9730	83	22	.	.	PUNCT
fcis-9730	84	1	the	the	DET
fcis-9730	84	2	cudnn	cudnn	NOUN
fcis-9730	84	3	library	library	NOUN
fcis-9730	84	4	is	be	AUX
fcis-9730	84	5	used	use	VERB
fcis-9730	84	6	for	for	ADP
fcis-9730	84	7	deep	deep	ADJ
fcis-9730	84	8	learning	learning	NOUN
fcis-9730	84	9	calculation	calculation	NOUN
fcis-9730	84	10	,	,	PUNCT
fcis-9730	84	11	which	which	PRON
fcis-9730	84	12	improves	improve	VERB
fcis-9730	84	13	the	the	DET
fcis-9730	84	14	training	training	NOUN
fcis-9730	84	15	and	and	CCONJ
fcis-9730	84	16	reasoning	reasoning	NOUN
fcis-9730	84	17	speed	speed	NOUN
fcis-9730	84	18	of	of	ADP
fcis-9730	84	19	the	the	DET
fcis-9730	84	20	model	model	NOUN
fcis-9730	84	21	;	;	PUNCT
fcis-9730	84	22	yolov7	yolov7	SYM
fcis-9730	84	23	algorithm	algorithm	NOUN
fcis-9730	84	24	is	be	AUX
fcis-9730	84	25	built	build	VERB
fcis-9730	84	26	based	base	VERB
fcis-9730	84	27	on	on	ADP
fcis-9730	84	28	deep	deep	ADJ
fcis-9730	84	29	learning	learning	NOUN
fcis-9730	84	30	technology	technology	NOUN
fcis-9730	84	31	,	,	PUNCT
fcis-9730	84	32	which	which	PRON
fcis-9730	84	33	can	can	AUX
fcis-9730	84	34	improve	improve	VERB
fcis-9730	84	35	its	its	PRON
fcis-9730	84	36	detection	detection	NOUN
fcis-9730	84	37	performance	performance	NOUN
fcis-9730	84	38	through	through	ADP
fcis-9730	84	39	continuous	continuous	ADJ
fcis-9730	84	40	learning	learning	NOUN
fcis-9730	84	41	and	and	CCONJ
fcis-9730	84	42	support	support	NOUN
fcis-9730	84	43	transfer	transfer	NOUN
fcis-9730	84	44	learning	learning	NOUN
fcis-9730	84	45	of	of	ADP
fcis-9730	84	46	deep	deep	ADJ
fcis-9730	84	47	models	model	NOUN
fcis-9730	84	48	.	.	PUNCT
fcis-9730	85	1	[	[	X
fcis-9730	85	2	14	14	NUM
fcis-9730	85	3	]	]	PUNCT
fcis-9730	85	4	however	however	ADV
fcis-9730	85	5	,	,	PUNCT
fcis-9730	85	6	higher	high	ADJ
fcis-9730	85	7	computing	computing	NOUN
fcis-9730	85	8	resources	resource	NOUN
fcis-9730	85	9	,	,	PUNCT
fcis-9730	85	10	more	more	ADJ
fcis-9730	85	11	layers	layer	NOUN
fcis-9730	85	12	and	and	CCONJ
fcis-9730	85	13	parameters	parameter	NOUN
fcis-9730	85	14	are	be	AUX
fcis-9730	85	15	required	require	VERB
fcis-9730	85	16	,	,	PUNCT
fcis-9730	85	17	so	so	ADV
fcis-9730	85	18	higher	high	ADJ
fcis-9730	85	19	computing	compute	VERB
fcis-9730	85	20	resources	resource	NOUN
fcis-9730	85	21	are	be	AUX
fcis-9730	85	22	required	require	VERB
fcis-9730	85	23	,	,	PUNCT
fcis-9730	85	24	and	and	CCONJ
fcis-9730	85	25	more	more	ADV
fcis-9730	85	26	efficient	efficient	ADJ
fcis-9730	85	27	gpus	gpu	NOUN
fcis-9730	85	28	are	be	AUX
fcis-9730	85	29	required	require	VERB
fcis-9730	85	30	for	for	ADP
fcis-9730	85	31	training	training	NOUN
fcis-9730	85	32	and	and	CCONJ
fcis-9730	85	33	reasoning	reasoning	NOUN
fcis-9730	85	34	.	.	PUNCT
fcis-9730	86	1	the	the	DET
fcis-9730	86	2	detection	detection	NOUN
fcis-9730	86	3	effect	effect	NOUN
fcis-9730	86	4	of	of	ADP
fcis-9730	86	5	small	small	ADJ
fcis-9730	86	6	size	size	NOUN
fcis-9730	86	7	and	and	CCONJ
fcis-9730	86	8	crowded	crowded	ADJ
fcis-9730	86	9	scenes	scene	NOUN
fcis-9730	86	10	is	be	AUX
fcis-9730	86	11	not	not	PART
fcis-9730	86	12	good	good	ADJ
fcis-9730	86	13	.	.	PUNCT
fcis-9730	87	1	although	although	SCONJ
fcis-9730	87	2	yolov7	yolov7	NOUN
fcis-9730	87	3	has	have	AUX
fcis-9730	87	4	improved	improve	VERB
fcis-9730	87	5	and	and	CCONJ
fcis-9730	87	6	optimized	optimize	VERB
fcis-9730	87	7	it	it	PRON
fcis-9730	87	8	,	,	PUNCT
fcis-9730	87	9	the	the	DET
fcis-9730	87	10	detection	detection	NOUN
fcis-9730	87	11	effect	effect	NOUN
fcis-9730	87	12	is	be	AUX
fcis-9730	87	13	still	still	ADV
fcis-9730	87	14	poor	poor	ADJ
fcis-9730	87	15	.	.	PUNCT
fcis-9730	88	1	3	3	X
fcis-9730	88	2	.	.	X
fcis-9730	88	3	yolo	yolo	ADJ
fcis-9730	88	4	algorithm	algorithm	PROPN
fcis-9730	88	5	application	application	NOUN
fcis-9730	88	6	field	field	NOUN
fcis-9730	88	7	based	base	VERB
fcis-9730	88	8	on	on	ADP
fcis-9730	88	9	deep	deep	ADJ
fcis-9730	88	10	learning	learn	VERB
fcis-9730	88	11	yolo	yolo	ADJ
fcis-9730	88	12	series	series	PROPN
fcis-9730	88	13	involves	involve	VERB
fcis-9730	88	14	many	many	ADJ
fcis-9730	88	15	application	application	NOUN
fcis-9730	88	16	fields	field	NOUN
fcis-9730	88	17	.	.	PUNCT
fcis-9730	89	1	for	for	ADP
fcis-9730	89	2	yolov7	yolov7	NOUN
fcis-9730	89	3	,	,	PUNCT
fcis-9730	89	4	which	which	PRON
fcis-9730	89	5	has	have	VERB
fcis-9730	89	6	the	the	DET
fcis-9730	89	7	best	good	ADJ
fcis-9730	89	8	performance	performance	NOUN
fcis-9730	89	9	at	at	ADP
fcis-9730	89	10	present	present	ADJ
fcis-9730	89	11	,	,	PUNCT
fcis-9730	89	12	the	the	DET
fcis-9730	89	13	following	follow	VERB
fcis-9730	89	14	application	application	NOUN
fcis-9730	89	15	examples	example	NOUN
fcis-9730	89	16	are	be	AUX
fcis-9730	89	17	introduced	introduce	VERB
fcis-9730	89	18	:	:	PUNCT
fcis-9730	89	19	1	1	X
fcis-9730	89	20	.	.	X
fcis-9730	89	21	real	real	ADJ
fcis-9730	89	22	-	-	PUNCT
fcis-9730	89	23	time	time	NOUN
fcis-9730	89	24	target	target	NOUN
fcis-9730	89	25	detection	detection	NOUN
fcis-9730	89	26	:	:	PUNCT
fcis-9730	89	27	yolov7	yolov7	PRON
fcis-9730	89	28	can	can	AUX
fcis-9730	89	29	be	be	AUX
fcis-9730	89	30	used	use	VERB
fcis-9730	89	31	for	for	ADP
fcis-9730	89	32	real	real	ADJ
fcis-9730	89	33	-	-	PUNCT
fcis-9730	89	34	time	time	NOUN
fcis-9730	89	35	target	target	NOUN
fcis-9730	89	36	detection	detection	NOUN
fcis-9730	89	37	,	,	PUNCT
fcis-9730	89	38	such	such	ADJ
fcis-9730	89	39	as	as	ADP
fcis-9730	89	40	automatic	automatic	ADJ
fcis-9730	89	41	driving	driving	NOUN
fcis-9730	89	42	,	,	PUNCT
fcis-9730	89	43	video	video	NOUN
fcis-9730	89	44	surveillance	surveillance	NOUN
fcis-9730	89	45	and	and	CCONJ
fcis-9730	89	46	security	security	NOUN
fcis-9730	89	47	systems	system	NOUN
fcis-9730	89	48	.	.	PUNCT
fcis-9730	90	1	it	it	PRON
fcis-9730	90	2	needs	need	VERB
fcis-9730	90	3	to	to	PART
fcis-9730	90	4	process	process	VERB
fcis-9730	90	5	highresolution	highresolution	NOUN
fcis-9730	90	6	images	image	NOUN
fcis-9730	90	7	and	and	CCONJ
fcis-9730	90	8	high	high	ADJ
fcis-9730	90	9	-	-	PUNCT
fcis-9730	90	10	speed	speed	NOUN
fcis-9730	90	11	video	video	NOUN
fcis-9730	90	12	,	,	PUNCT
fcis-9730	90	13	and	and	CCONJ
fcis-9730	90	14	can	can	AUX
fcis-9730	90	15	correctly	correctly	ADV
fcis-9730	90	16	detect	detect	VERB
fcis-9730	90	17	targets	target	NOUN
fcis-9730	90	18	,	,	PUNCT
fcis-9730	90	19	thus	thus	ADV
fcis-9730	90	20	helping	help	VERB
fcis-9730	90	21	monitors	monitor	NOUN
fcis-9730	90	22	and	and	CCONJ
fcis-9730	90	23	decision	decision	NOUN
fcis-9730	90	24	makers	maker	NOUN
fcis-9730	90	25	to	to	PART
fcis-9730	90	26	react	react	VERB
fcis-9730	90	27	quickly	quickly	ADV
fcis-9730	90	28	.	.	PUNCT
fcis-9730	91	1	[	[	X
fcis-9730	91	2	15][16	15][16	X
fcis-9730	91	3	]	]	X
fcis-9730	91	4	2	2	NUM
fcis-9730	91	5	.	.	NOUN
fcis-9730	91	6	object	object	NOUN
fcis-9730	91	7	recognition	recognition	NOUN
fcis-9730	91	8	:	:	PUNCT
fcis-9730	91	9	yolov7	yolov7	PROPN
fcis-9730	91	10	can	can	AUX
fcis-9730	91	11	also	also	ADV
fcis-9730	91	12	implement	implement	VERB
fcis-9730	91	13	object	object	NOUN
fcis-9730	91	14	recognition	recognition	NOUN
fcis-9730	91	15	.	.	PUNCT
fcis-9730	92	1	for	for	ADP
fcis-9730	92	2	example	example	NOUN
fcis-9730	92	3	,	,	PUNCT
fcis-9730	92	4	in	in	ADP
fcis-9730	92	5	a	a	DET
fcis-9730	92	6	large	large	ADJ
fcis-9730	92	7	warehouse	warehouse	NOUN
fcis-9730	92	8	,	,	PUNCT
fcis-9730	92	9	there	there	PRON
fcis-9730	92	10	are	be	VERB
fcis-9730	92	11	many	many	ADJ
fcis-9730	92	12	different	different	ADJ
fcis-9730	92	13	types	type	NOUN
fcis-9730	92	14	of	of	ADP
fcis-9730	92	15	products	product	NOUN
fcis-9730	92	16	and	and	CCONJ
fcis-9730	92	17	equipment	equipment	NOUN
fcis-9730	92	18	.	.	PUNCT
fcis-9730	93	1	each	each	DET
fcis-9730	93	2	object	object	NOUN
fcis-9730	93	3	can	can	AUX
fcis-9730	93	4	be	be	AUX
fcis-9730	93	5	accurately	accurately	ADV
fcis-9730	93	6	identified	identify	VERB
fcis-9730	93	7	and	and	CCONJ
fcis-9730	93	8	classified	classify	VERB
fcis-9730	93	9	using	use	VERB
fcis-9730	93	10	yolov7	yolov7	NOUN
fcis-9730	93	11	,	,	PUNCT
fcis-9730	93	12	allowing	allow	VERB
fcis-9730	93	13	for	for	ADP
fcis-9730	93	14	better	well	ADJ
fcis-9730	93	15	management	management	NOUN
fcis-9730	93	16	and	and	CCONJ
fcis-9730	93	17	control	control	NOUN
fcis-9730	93	18	of	of	ADP
fcis-9730	93	19	inventory	inventory	NOUN
fcis-9730	93	20	and	and	CCONJ
fcis-9730	93	21	production	production	NOUN
fcis-9730	93	22	processes	process	NOUN
fcis-9730	93	23	.	.	PUNCT
fcis-9730	94	1	3	3	X
fcis-9730	94	2	.	.	X
fcis-9730	94	3	face	face	NOUN
fcis-9730	94	4	recognition	recognition	NOUN
fcis-9730	94	5	:	:	PUNCT
fcis-9730	94	6	yolov7	yolov7	PROPN
fcis-9730	94	7	's	's	PART
fcis-9730	94	8	face	face	NOUN
fcis-9730	94	9	detection	detection	NOUN
fcis-9730	94	10	function	function	NOUN
fcis-9730	94	11	can	can	AUX
fcis-9730	94	12	be	be	AUX
fcis-9730	94	13	used	use	VERB
fcis-9730	94	14	for	for	ADP
fcis-9730	94	15	applications	application	NOUN
fcis-9730	94	16	such	such	ADJ
fcis-9730	94	17	as	as	ADP
fcis-9730	94	18	security	security	NOUN
fcis-9730	94	19	access	access	NOUN
fcis-9730	94	20	control	control	NOUN
fcis-9730	94	21	,	,	PUNCT
fcis-9730	94	22	intelligent	intelligent	ADJ
fcis-9730	94	23	gate	gate	NOUN
fcis-9730	94	24	,	,	PUNCT
fcis-9730	94	25	party	party	NOUN
fcis-9730	94	26	check	check	NOUN
fcis-9730	94	27	-	-	PUNCT
fcis-9730	94	28	in	in	NOUN
fcis-9730	94	29	,	,	PUNCT
fcis-9730	94	30	etc	etc	X
fcis-9730	94	31	.	.	X
fcis-9730	94	32	by	by	ADP
fcis-9730	94	33	using	use	VERB
fcis-9730	94	34	yolov7	yolov7	NOUN
fcis-9730	94	35	for	for	ADP
fcis-9730	94	36	face	face	NOUN
fcis-9730	94	37	detection	detection	NOUN
fcis-9730	94	38	and	and	CCONJ
fcis-9730	94	39	recognition	recognition	NOUN
fcis-9730	94	40	,	,	PUNCT
fcis-9730	94	41	it	it	PRON
fcis-9730	94	42	can	can	AUX
fcis-9730	94	43	ensure	ensure	VERB
fcis-9730	94	44	that	that	SCONJ
fcis-9730	94	45	only	only	ADV
fcis-9730	94	46	authorized	authorize	VERB
fcis-9730	94	47	personnel	personnel	NOUN
fcis-9730	94	48	can	can	AUX
fcis-9730	94	49	enter	enter	VERB
fcis-9730	94	50	the	the	DET
fcis-9730	94	51	area	area	NOUN
fcis-9730	94	52	,	,	PUNCT
fcis-9730	94	53	and	and	CCONJ
fcis-9730	94	54	the	the	DET
fcis-9730	94	55	activities	activity	NOUN
fcis-9730	94	56	of	of	ADP
fcis-9730	94	57	participants	participant	NOUN
fcis-9730	94	58	can	can	AUX
fcis-9730	94	59	be	be	AUX
fcis-9730	94	60	tracked	track	VERB
fcis-9730	94	61	.	.	PUNCT
fcis-9730	95	1	4	4	X
fcis-9730	95	2	.	.	NOUN
fcis-9730	95	3	natural	natural	ADJ
fcis-9730	95	4	language	language	NOUN
fcis-9730	95	5	processing	processing	NOUN
fcis-9730	95	6	:	:	PUNCT
fcis-9730	95	7	yolov7	yolov7	PROPN
fcis-9730	95	8	can	can	AUX
fcis-9730	95	9	also	also	ADV
fcis-9730	95	10	be	be	AUX
fcis-9730	95	11	used	use	VERB
fcis-9730	95	12	in	in	ADP
fcis-9730	95	13	conjunction	conjunction	NOUN
fcis-9730	95	14	with	with	ADP
fcis-9730	95	15	natural	natural	ADJ
fcis-9730	95	16	language	language	NOUN
fcis-9730	95	17	processing	processing	NOUN
fcis-9730	95	18	(	(	PUNCT
fcis-9730	95	19	nlp	nlp	NOUN
fcis-9730	95	20	)	)	PUNCT
fcis-9730	95	21	technology	technology	NOUN
fcis-9730	95	22	to	to	PART
fcis-9730	95	23	enable	enable	VERB
fcis-9730	95	24	automatic	automatic	ADJ
fcis-9730	95	25	text	text	NOUN
fcis-9730	95	26	extraction	extraction	NOUN
fcis-9730	95	27	and	and	CCONJ
fcis-9730	95	28	classification	classification	NOUN
fcis-9730	95	29	.	.	PUNCT
fcis-9730	96	1	for	for	ADP
fcis-9730	96	2	example	example	NOUN
fcis-9730	96	3	,	,	PUNCT
fcis-9730	96	4	monitoring	monitor	VERB
fcis-9730	96	5	specific	specific	ADJ
fcis-9730	96	6	keywords	keyword	NOUN
fcis-9730	96	7	on	on	ADP
fcis-9730	96	8	social	social	ADJ
fcis-9730	96	9	media	medium	NOUN
fcis-9730	96	10	to	to	PART
fcis-9730	96	11	see	see	VERB
fcis-9730	96	12	what	what	PRON
fcis-9730	96	13	people	people	NOUN
fcis-9730	96	14	are	be	AUX
fcis-9730	96	15	saying	say	VERB
fcis-9730	96	16	about	about	ADP
fcis-9730	96	17	an	an	DET
fcis-9730	96	18	event	event	NOUN
fcis-9730	96	19	.	.	PUNCT
fcis-9730	97	1	this	this	PRON
fcis-9730	97	2	can	can	AUX
fcis-9730	97	3	help	help	VERB
fcis-9730	97	4	businesses	business	NOUN
fcis-9730	97	5	and	and	CCONJ
fcis-9730	97	6	institutions	institution	NOUN
fcis-9730	97	7	understand	understand	VERB
fcis-9730	97	8	the	the	DET
fcis-9730	97	9	views	view	NOUN
fcis-9730	97	10	of	of	ADP
fcis-9730	97	11	opinion	opinion	NOUN
fcis-9730	97	12	leaders	leader	NOUN
fcis-9730	97	13	and	and	CCONJ
fcis-9730	97	14	consumers	consumer	NOUN
fcis-9730	97	15	to	to	PART
fcis-9730	97	16	make	make	VERB
fcis-9730	97	17	better	well	ADJ
fcis-9730	97	18	marketing	marketing	NOUN
fcis-9730	97	19	strategies	strategy	NOUN
fcis-9730	97	20	and	and	CCONJ
fcis-9730	97	21	decisions	decision	NOUN
fcis-9730	97	22	.	.	PUNCT
fcis-9730	98	1	5	5	X
fcis-9730	98	2	.	.	X
fcis-9730	98	3	robot	robot	NOUN
fcis-9730	98	4	control	control	NOUN
fcis-9730	98	5	:	:	PUNCT
fcis-9730	98	6	yolov7	yolov7	PRON
fcis-9730	98	7	can	can	AUX
fcis-9730	98	8	be	be	AUX
fcis-9730	98	9	used	use	VERB
fcis-9730	98	10	as	as	ADP
fcis-9730	98	11	a	a	DET
fcis-9730	98	12	vision	vision	NOUN
fcis-9730	98	13	engine	engine	NOUN
fcis-9730	98	14	in	in	ADP
fcis-9730	98	15	robot	robot	NOUN
fcis-9730	98	16	control	control	NOUN
fcis-9730	98	17	.	.	PUNCT
fcis-9730	99	1	robots	robot	NOUN
fcis-9730	99	2	can	can	AUX
fcis-9730	99	3	use	use	VERB
fcis-9730	99	4	yolov7	yolov7	NOUN
fcis-9730	99	5	to	to	PART
fcis-9730	99	6	detect	detect	VERB
fcis-9730	99	7	and	and	CCONJ
fcis-9730	99	8	track	track	NOUN
fcis-9730	99	9	objects	object	NOUN
fcis-9730	99	10	,	,	PUNCT
fcis-9730	99	11	such	such	ADJ
fcis-9730	99	12	as	as	ADP
fcis-9730	99	13	humans	human	NOUN
fcis-9730	99	14	,	,	PUNCT
fcis-9730	99	15	other	other	ADJ
fcis-9730	99	16	robots	robot	NOUN
fcis-9730	99	17	or	or	CCONJ
fcis-9730	99	18	obstacles	obstacle	NOUN
fcis-9730	99	19	.	.	PUNCT
fcis-9730	100	1	this	this	PRON
fcis-9730	100	2	can	can	AUX
fcis-9730	100	3	help	help	VERB
fcis-9730	100	4	the	the	DET
fcis-9730	100	5	robot	robot	NOUN
fcis-9730	100	6	avoid	avoid	VERB
fcis-9730	100	7	collisions	collision	NOUN
fcis-9730	100	8	,	,	PUNCT
fcis-9730	100	9	locate	locate	VERB
fcis-9730	100	10	the	the	DET
fcis-9730	100	11	target	target	NOUN
fcis-9730	100	12	correctly	correctly	ADV
fcis-9730	100	13	,	,	PUNCT
fcis-9730	100	14	and	and	CCONJ
fcis-9730	100	15	react	react	VERB
fcis-9730	100	16	quickly	quickly	ADV
fcis-9730	100	17	in	in	ADP
fcis-9730	100	18	case	case	NOUN
fcis-9730	100	19	of	of	ADP
fcis-9730	100	20	an	an	DET
fcis-9730	100	21	emergency	emergency	NOUN
fcis-9730	100	22	.	.	PUNCT
fcis-9730	101	1	4	4	X
fcis-9730	101	2	.	.	X
fcis-9730	101	3	look	look	VERB
fcis-9730	101	4	to	to	ADP
fcis-9730	101	5	the	the	DET
fcis-9730	101	6	future	future	NOUN
fcis-9730	101	7	the	the	DET
fcis-9730	101	8	current	current	ADJ
fcis-9730	101	9	yolo	yolo	ADJ
fcis-9730	101	10	algorithm	algorithm	NOUN
fcis-9730	101	11	based	base	VERB
fcis-9730	101	12	on	on	ADP
fcis-9730	101	13	deep	deep	ADJ
fcis-9730	101	14	learning	learning	NOUN
fcis-9730	101	15	has	have	VERB
fcis-9730	101	16	a	a	DET
fcis-9730	101	17	great	great	ADJ
fcis-9730	101	18	improvement	improvement	NOUN
fcis-9730	101	19	in	in	ADP
fcis-9730	101	20	performance	performance	NOUN
fcis-9730	101	21	compared	compare	VERB
fcis-9730	101	22	to	to	ADP
fcis-9730	101	23	traditional	traditional	ADJ
fcis-9730	101	24	detection	detection	NOUN
fcis-9730	101	25	algorithms	algorithm	NOUN
fcis-9730	101	26	.	.	PUNCT
fcis-9730	102	1	although	although	SCONJ
fcis-9730	102	2	the	the	DET
fcis-9730	102	3	detection	detection	NOUN
fcis-9730	102	4	speed	speed	NOUN
fcis-9730	102	5	and	and	CCONJ
fcis-9730	102	6	accuracy	accuracy	NOUN
fcis-9730	102	7	are	be	AUX
fcis-9730	102	8	greatly	greatly	ADV
fcis-9730	102	9	improved	improve	VERB
fcis-9730	102	10	,	,	PUNCT
fcis-9730	102	11	there	there	PRON
fcis-9730	102	12	are	be	VERB
fcis-9730	102	13	still	still	ADV
fcis-9730	102	14	some	some	DET
fcis-9730	102	15	unsolved	unsolved	ADJ
fcis-9730	102	16	problems	problem	NOUN
fcis-9730	102	17	,	,	PUNCT
fcis-9730	102	18	such	such	ADJ
fcis-9730	102	19	as	as	ADP
fcis-9730	102	20	incomplete	incomplete	ADJ
fcis-9730	102	21	data	datum	NOUN
fcis-9730	102	22	set	set	VERB
fcis-9730	102	23	,	,	PUNCT
fcis-9730	102	24	less	less	ADJ
fcis-9730	102	25	data	datum	NOUN
fcis-9730	102	26	for	for	ADP
fcis-9730	102	27	small	small	ADJ
fcis-9730	102	28	target	target	NOUN
fcis-9730	102	29	detection	detection	NOUN
fcis-9730	102	30	,	,	PUNCT
fcis-9730	102	31	low	low	ADJ
fcis-9730	102	32	detection	detection	NOUN
fcis-9730	102	33	performance	performance	NOUN
fcis-9730	102	34	,	,	PUNCT
fcis-9730	102	35	and	and	CCONJ
fcis-9730	102	36	failure	failure	NOUN
fcis-9730	102	37	to	to	PART
fcis-9730	102	38	achieve	achieve	VERB
fcis-9730	102	39	real	real	ADJ
fcis-9730	102	40	-	-	PUNCT
fcis-9730	102	41	time	time	NOUN
fcis-9730	102	42	online	online	ADJ
fcis-9730	102	43	detection	detection	NOUN
fcis-9730	102	44	without	without	ADP
fcis-9730	102	45	reducing	reduce	VERB
fcis-9730	102	46	detection	detection	NOUN
fcis-9730	102	47	accuracy	accuracy	NOUN
fcis-9730	102	48	.	.	PUNCT
fcis-9730	103	1	at	at	ADP
fcis-9730	103	2	present	present	ADJ
fcis-9730	103	3	,	,	PUNCT
fcis-9730	103	4	in	in	ADP
fcis-9730	103	5	the	the	DET
fcis-9730	103	6	development	development	NOUN
fcis-9730	103	7	of	of	ADP
fcis-9730	103	8	society	society	NOUN
fcis-9730	103	9	,	,	PUNCT
fcis-9730	103	10	country	country	NOUN
fcis-9730	103	11	and	and	CCONJ
fcis-9730	103	12	other	other	ADJ
fcis-9730	103	13	aspects	aspect	NOUN
fcis-9730	103	14	,	,	PUNCT
fcis-9730	103	15	target	target	NOUN
fcis-9730	103	16	detection	detection	NOUN
fcis-9730	103	17	is	be	AUX
fcis-9730	103	18	a	a	DET
fcis-9730	103	19	big	big	ADJ
fcis-9730	103	20	topic	topic	NOUN
fcis-9730	103	21	,	,	PUNCT
fcis-9730	103	22	and	and	CCONJ
fcis-9730	103	23	real	real	ADJ
fcis-9730	103	24	-	-	PUNCT
fcis-9730	103	25	time	time	NOUN
fcis-9730	103	26	online	online	ADJ
fcis-9730	103	27	detection	detection	NOUN
fcis-9730	103	28	is	be	AUX
fcis-9730	103	29	the	the	DET
fcis-9730	103	30	top	top	ADJ
fcis-9730	103	31	priority	priority	NOUN
fcis-9730	103	32	,	,	PUNCT
fcis-9730	103	33	which	which	PRON
fcis-9730	103	34	is	be	AUX
fcis-9730	103	35	worthy	worthy	ADJ
fcis-9730	103	36	of	of	ADP
fcis-9730	103	37	continued	continue	VERB
fcis-9730	103	38	research	research	NOUN
fcis-9730	103	39	and	and	CCONJ
fcis-9730	103	40	development	development	NOUN
fcis-9730	103	41	.	.	PUNCT
fcis-9730	104	1	this	this	DET
fcis-9730	104	2	paper	paper	NOUN
fcis-9730	104	3	puts	put	VERB
fcis-9730	104	4	forward	forward	ADV
fcis-9730	104	5	several	several	ADJ
fcis-9730	104	6	prospects	prospect	NOUN
fcis-9730	104	7	for	for	ADP
fcis-9730	104	8	the	the	DET
fcis-9730	104	9	future	future	ADJ
fcis-9730	104	10	development	development	NOUN
fcis-9730	104	11	direction	direction	NOUN
fcis-9730	104	12	of	of	ADP
fcis-9730	104	13	yolo	yolo	NOUN
fcis-9730	104	14	:	:	PUNCT
fcis-9730	104	15	1	1	X
fcis-9730	104	16	.	.	PUNCT
fcis-9730	104	17	data	datum	NOUN
fcis-9730	104	18	enhancement	enhancement	NOUN
fcis-9730	104	19	:	:	PUNCT
fcis-9730	104	20	adding	add	VERB
fcis-9730	104	21	more	more	ADJ
fcis-9730	104	22	data	datum	NOUN
fcis-9730	104	23	can	can	AUX
fcis-9730	104	24	improve	improve	VERB
fcis-9730	104	25	the	the	DET
fcis-9730	104	26	accuracy	accuracy	NOUN
fcis-9730	104	27	of	of	ADP
fcis-9730	104	28	the	the	DET
fcis-9730	104	29	model	model	NOUN
fcis-9730	104	30	.	.	PUNCT
fcis-9730	105	1	data	datum	NOUN
fcis-9730	105	2	enhancement	enhancement	NOUN
fcis-9730	105	3	can	can	AUX
fcis-9730	105	4	be	be	AUX
fcis-9730	105	5	achieved	achieve	VERB
fcis-9730	105	6	by	by	ADP
fcis-9730	105	7	rotating	rotate	VERB
fcis-9730	105	8	,	,	PUNCT
fcis-9730	105	9	scaling	scaling	NOUN
fcis-9730	105	10	,	,	PUNCT
fcis-9730	105	11	cutting	cut	VERB
fcis-9730	105	12	,	,	PUNCT
fcis-9730	105	13	flipping	flip	VERB
fcis-9730	105	14	,	,	PUNCT
fcis-9730	105	15	color	color	NOUN
fcis-9730	105	16	transformation	transformation	NOUN
fcis-9730	105	17	,	,	PUNCT
fcis-9730	105	18	etc	etc	X
fcis-9730	105	19	.	.	X
fcis-9730	105	20	,	,	PUNCT
fcis-9730	105	21	on	on	ADP
fcis-9730	105	22	the	the	DET
fcis-9730	105	23	training	training	NOUN
fcis-9730	105	24	image	image	NOUN
fcis-9730	105	25	.	.	PUNCT
fcis-9730	106	1	consider	consider	VERB
fcis-9730	106	2	using	use	VERB
fcis-9730	106	3	more	more	ADJ
fcis-9730	106	4	data	datum	NOUN
fcis-9730	106	5	enhancement	enhancement	NOUN
fcis-9730	106	6	techniques	technique	NOUN
fcis-9730	106	7	to	to	PART
fcis-9730	106	8	augment	augment	VERB
fcis-9730	106	9	the	the	DET
fcis-9730	106	10	data	datum	NOUN
fcis-9730	106	11	set	set	VERB
fcis-9730	106	12	.	.	PUNCT
fcis-9730	107	1	2	2	X
fcis-9730	107	2	.	.	X
fcis-9730	107	3	better	well	ADJ
fcis-9730	107	4	pre	pre	ADJ
fcis-9730	107	5	-	-	ADJ
fcis-9730	107	6	trained	train	VERB
fcis-9730	107	7	model	model	NOUN
fcis-9730	107	8	:	:	PUNCT
fcis-9730	107	9	consider	consider	VERB
fcis-9730	107	10	using	use	VERB
fcis-9730	107	11	stronger	strong	ADJ
fcis-9730	107	12	pretrained	pretraine	VERB
fcis-9730	107	13	models	model	NOUN
fcis-9730	107	14	to	to	PART
fcis-9730	107	15	initialize	initialize	VERB
fcis-9730	107	16	the	the	DET
fcis-9730	107	17	yolov7	yolov7	NOUN
fcis-9730	107	18	network	network	NOUN
fcis-9730	107	19	to	to	PART
fcis-9730	107	20	improve	improve	VERB
fcis-9730	107	21	the	the	DET
fcis-9730	107	22	accuracy	accuracy	NOUN
fcis-9730	107	23	and	and	CCONJ
fcis-9730	107	24	robustness	robustness	NOUN
fcis-9730	107	25	of	of	ADP
fcis-9730	107	26	the	the	DET
fcis-9730	107	27	model	model	NOUN
fcis-9730	107	28	.	.	PUNCT
fcis-9730	108	1	for	for	ADP
fcis-9730	108	2	example	example	NOUN
fcis-9730	108	3	,	,	PUNCT
fcis-9730	108	4	larger	large	ADJ
fcis-9730	108	5	and	and	CCONJ
fcis-9730	108	6	deeper	deep	ADJ
fcis-9730	108	7	models	model	NOUN
fcis-9730	108	8	such	such	ADJ
fcis-9730	108	9	as	as	ADP
fcis-9730	108	10	resnet	resnet	NOUN
fcis-9730	108	11	,	,	PUNCT
fcis-9730	108	12	efficientnet	efficientnet	NOUN
fcis-9730	108	13	,	,	PUNCT
fcis-9730	108	14	etc	etc	X
fcis-9730	108	15	.	.	X
fcis-9730	108	16	,	,	PUNCT
fcis-9730	108	17	can	can	AUX
fcis-9730	108	18	be	be	AUX
fcis-9730	108	19	used	use	VERB
fcis-9730	108	20	.	.	PUNCT
fcis-9730	109	1	3	3	NUM
fcis-9730	109	2	adaptive	adaptive	ADJ
fcis-9730	109	3	learning	learning	NOUN
fcis-9730	109	4	rates	rate	NOUN
fcis-9730	109	5	:	:	PUNCT
fcis-9730	109	6	to	to	PART
fcis-9730	109	7	better	well	ADV
fcis-9730	109	8	optimize	optimize	VERB
fcis-9730	109	9	the	the	DET
fcis-9730	109	10	model	model	NOUN
fcis-9730	109	11	,	,	PUNCT
fcis-9730	109	12	you	you	PRON
fcis-9730	109	13	can	can	AUX
fcis-9730	109	14	use	use	VERB
fcis-9730	109	15	the	the	DET
fcis-9730	109	16	adaptive	adaptive	ADJ
fcis-9730	109	17	learning	learning	NOUN
fcis-9730	109	18	rate	rate	NOUN
fcis-9730	109	19	to	to	PART
fcis-9730	109	20	adjust	adjust	VERB
fcis-9730	109	21	the	the	DET
fcis-9730	109	22	learning	learning	NOUN
fcis-9730	109	23	rate	rate	NOUN
fcis-9730	109	24	of	of	ADP
fcis-9730	109	25	each	each	DET
fcis-9730	109	26	layer	layer	NOUN
fcis-9730	109	27	to	to	ADP
fcis-9730	109	28	better	well	ADJ
fcis-9730	109	29	balance	balance	VERB
fcis-9730	109	30	speed	speed	NOUN
fcis-9730	109	31	and	and	CCONJ
fcis-9730	109	32	precision	precision	NOUN
fcis-9730	109	33	during	during	ADP
fcis-9730	109	34	training	training	NOUN
fcis-9730	109	35	.	.	PUNCT
fcis-9730	110	1	4	4	X
fcis-9730	110	2	.	.	X
fcis-9730	110	3	improvement	improvement	NOUN
fcis-9730	110	4	of	of	ADP
fcis-9730	110	5	activation	activation	NOUN
fcis-9730	110	6	function	function	NOUN
fcis-9730	110	7	:	:	PUNCT
fcis-9730	110	8	it	it	PRON
fcis-9730	110	9	is	be	AUX
fcis-9730	110	10	possible	possible	ADJ
fcis-9730	110	11	to	to	PART
fcis-9730	110	12	consider	consider	VERB
fcis-9730	110	13	using	use	VERB
fcis-9730	110	14	other	other	ADJ
fcis-9730	110	15	activation	activation	NOUN
fcis-9730	110	16	functions	function	NOUN
fcis-9730	110	17	such	such	ADJ
fcis-9730	110	18	as	as	ADP
fcis-9730	110	19	swish	swish	ADJ
fcis-9730	110	20	,	,	PUNCT
fcis-9730	110	21	mish	mish	PROPN
fcis-9730	110	22	,	,	PUNCT
fcis-9730	110	23	etc	etc	X
fcis-9730	110	24	.	.	X
fcis-9730	110	25	to	to	PART
fcis-9730	110	26	replace	replace	VERB
fcis-9730	110	27	leakyrelu	leakyrelu	NOUN
fcis-9730	110	28	used	use	VERB
fcis-9730	110	29	in	in	ADP
fcis-9730	110	30	yolov7	yolov7	NOUN
fcis-9730	110	31	to	to	PART
fcis-9730	110	32	improve	improve	VERB
fcis-9730	110	33	the	the	DET
fcis-9730	110	34	performance	performance	NOUN
fcis-9730	110	35	of	of	ADP
fcis-9730	110	36	the	the	DET
fcis-9730	110	37	model	model	NOUN
fcis-9730	110	38	.	.	PUNCT
fcis-9730	111	1	5	5	X
fcis-9730	111	2	.	.	X
fcis-9730	111	3	improvement	improvement	NOUN
fcis-9730	111	4	of	of	ADP
fcis-9730	111	5	target	target	NOUN
fcis-9730	111	6	detection	detection	NOUN
fcis-9730	111	7	loss	loss	NOUN
fcis-9730	111	8	function	function	NOUN
fcis-9730	111	9	:	:	PUNCT
fcis-9730	111	10	the	the	DET
fcis-9730	111	11	cross	cross	ADJ
fcis-9730	111	12	-	-	ADJ
fcis-9730	111	13	entropy	entropy	ADJ
fcis-9730	111	14	loss	loss	NOUN
fcis-9730	111	15	function	function	NOUN
fcis-9730	111	16	used	use	VERB
fcis-9730	111	17	by	by	ADP
fcis-9730	111	18	yolov7	yolov7	NOUN
fcis-9730	111	19	is	be	AUX
fcis-9730	111	20	one	one	NUM
fcis-9730	111	21	of	of	ADP
fcis-9730	111	22	the	the	DET
fcis-9730	111	23	common	common	ADJ
fcis-9730	111	24	target	target	NOUN
fcis-9730	111	25	detection	detection	NOUN
fcis-9730	111	26	loss	loss	NOUN
fcis-9730	111	27	functions	function	NOUN
fcis-9730	111	28	,	,	PUNCT
fcis-9730	111	29	but	but	CCONJ
fcis-9730	111	30	it	it	PRON
fcis-9730	111	31	may	may	AUX
fcis-9730	111	32	not	not	PART
fcis-9730	111	33	be	be	AUX
fcis-9730	111	34	optimal	optimal	ADJ
fcis-9730	111	35	in	in	ADP
fcis-9730	111	36	some	some	DET
fcis-9730	111	37	cases	case	NOUN
fcis-9730	111	38	.	.	PUNCT
fcis-9730	112	1	you	you	PRON
fcis-9730	112	2	can	can	AUX
fcis-9730	112	3	try	try	VERB
fcis-9730	112	4	to	to	PART
fcis-9730	112	5	use	use	VERB
fcis-9730	112	6	other	other	ADJ
fcis-9730	112	7	target	target	NOUN
fcis-9730	112	8	detection	detection	NOUN
fcis-9730	112	9	loss	loss	NOUN
fcis-9730	112	10	functions	function	NOUN
fcis-9730	112	11	,	,	PUNCT
fcis-9730	112	12	such	such	ADJ
fcis-9730	112	13	as	as	ADP
fcis-9730	112	14	focal	focal	ADJ
fcis-9730	112	15	loss	loss	NOUN
fcis-9730	112	16	,	,	PUNCT
fcis-9730	112	17	iou	iou	NOUN
fcis-9730	112	18	loss	loss	NOUN
fcis-9730	112	19	,	,	PUNCT
fcis-9730	112	20	etc	etc	X
fcis-9730	112	21	.	.	X
fcis-9730	113	1	6	6	X
fcis-9730	113	2	.	.	X
fcis-9730	113	3	multi	multi	ADJ
fcis-9730	113	4	-	-	NOUN
fcis-9730	113	5	task	task	ADJ
fcis-9730	113	6	learning	learning	NOUN
fcis-9730	113	7	:	:	PUNCT
fcis-9730	113	8	in	in	ADP
fcis-9730	113	9	addition	addition	NOUN
fcis-9730	113	10	to	to	AUX
fcis-9730	113	11	target	target	VERB
fcis-9730	113	12	detection	detection	NOUN
fcis-9730	113	13	,	,	PUNCT
fcis-9730	113	14	other	other	ADJ
fcis-9730	113	15	tasks	task	NOUN
fcis-9730	113	16	such	such	ADJ
fcis-9730	113	17	as	as	ADP
fcis-9730	113	18	semantic	semantic	ADJ
fcis-9730	113	19	segmentation	segmentation	NOUN
fcis-9730	113	20	and	and	CCONJ
fcis-9730	113	21	instance	instance	NOUN
fcis-9730	113	22	segmentation	segmentation	NOUN
fcis-9730	113	23	can	can	AUX
fcis-9730	113	24	be	be	AUX
fcis-9730	113	25	included	include	VERB
fcis-9730	113	26	in	in	ADP
fcis-9730	113	27	the	the	DET
fcis-9730	113	28	training	training	NOUN
fcis-9730	113	29	to	to	PART
fcis-9730	113	30	improve	improve	VERB
fcis-9730	113	31	the	the	DET
fcis-9730	113	32	comprehensive	comprehensive	ADJ
fcis-9730	113	33	performance	performance	NOUN
fcis-9730	113	34	of	of	ADP
fcis-9730	113	35	the	the	DET
fcis-9730	113	36	model	model	NOUN
fcis-9730	113	37	.	.	PUNCT
fcis-9730	114	1	7.model	7.model	NUM
fcis-9730	114	2	compression	compression	NOUN
fcis-9730	114	3	:	:	PUNCT
fcis-9730	114	4	some	some	DET
fcis-9730	114	5	model	model	NOUN
fcis-9730	114	6	compression	compression	NOUN
fcis-9730	114	7	techniques	technique	NOUN
fcis-9730	114	8	,	,	PUNCT
fcis-9730	114	9	such	such	ADJ
fcis-9730	114	10	as	as	ADP
fcis-9730	114	11	pruning	prune	VERB
fcis-9730	114	12	and	and	CCONJ
fcis-9730	114	13	quantization	quantization	NOUN
fcis-9730	114	14	,	,	PUNCT
fcis-9730	114	15	can	can	AUX
fcis-9730	114	16	be	be	AUX
fcis-9730	114	17	considered	consider	VERB
fcis-9730	114	18	to	to	PART
fcis-9730	114	19	reduce	reduce	VERB
fcis-9730	114	20	the	the	DET
fcis-9730	114	21	model	model	NOUN
fcis-9730	114	22	size	size	NOUN
fcis-9730	114	23	and	and	CCONJ
fcis-9730	114	24	inference	inference	NOUN
fcis-9730	114	25	time	time	NOUN
fcis-9730	114	26	of	of	ADP
fcis-9730	114	27	yolov7	yolov7	NOUN
fcis-9730	114	28	,	,	PUNCT
fcis-9730	114	29	so	so	SCONJ
fcis-9730	114	30	as	as	SCONJ
fcis-9730	114	31	to	to	PART
fcis-9730	114	32	adapt	adapt	VERB
fcis-9730	114	33	to	to	ADP
fcis-9730	114	34	the	the	DET
fcis-9730	114	35	deployment	deployment	NOUN
fcis-9730	114	36	on	on	ADP
fcis-9730	114	37	low	low	ADJ
fcis-9730	114	38	-	-	PUNCT
fcis-9730	114	39	power	power	NOUN
fcis-9730	114	40	devices	device	NOUN
fcis-9730	114	41	.	.	PUNCT
fcis-9730	115	1	these	these	PRON
fcis-9730	115	2	are	be	AUX
fcis-9730	115	3	just	just	ADV
fcis-9730	115	4	some	some	DET
fcis-9730	115	5	ideas	idea	NOUN
fcis-9730	115	6	,	,	PUNCT
fcis-9730	115	7	and	and	CCONJ
fcis-9730	115	8	how	how	SCONJ
fcis-9730	115	9	to	to	PART
fcis-9730	115	10	combine	combine	VERB
fcis-9730	115	11	and	and	CCONJ
fcis-9730	115	12	implement	implement	VERB
fcis-9730	115	13	them	they	PRON
fcis-9730	115	14	still	still	ADV
fcis-9730	115	15	needs	need	VERB
fcis-9730	115	16	specific	specific	ADJ
fcis-9730	115	17	experiments	experiment	NOUN
fcis-9730	115	18	and	and	CCONJ
fcis-9730	115	19	debugging	debugging	NOUN
fcis-9730	115	20	.	.	PUNCT
fcis-9730	116	1	how	how	SCONJ
fcis-9730	116	2	to	to	PART
fcis-9730	116	3	achieve	achieve	VERB
fcis-9730	116	4	these	these	DET
fcis-9730	116	5	improvements	improvement	NOUN
fcis-9730	116	6	while	while	SCONJ
fcis-9730	116	7	maintaining	maintain	VERB
fcis-9730	116	8	the	the	DET
fcis-9730	116	9	performance	performance	NOUN
fcis-9730	116	10	of	of	ADP
fcis-9730	116	11	the	the	DET
fcis-9730	116	12	model	model	NOUN
fcis-9730	116	13	also	also	ADV
fcis-9730	116	14	needs	need	VERB
fcis-9730	116	15	to	to	PART
fcis-9730	116	16	be	be	AUX
fcis-9730	116	17	considered	consider	VERB
fcis-9730	116	18	.	.	PUNCT
fcis-9730	117	1	5	5	X
fcis-9730	117	2	.	.	X
fcis-9730	117	3	conclusion	conclusion	NOUN
fcis-9730	117	4	based	base	VERB
fcis-9730	117	5	on	on	ADP
fcis-9730	117	6	the	the	DET
fcis-9730	117	7	development	development	NOUN
fcis-9730	117	8	history	history	NOUN
fcis-9730	117	9	of	of	ADP
fcis-9730	117	10	yolo	yolo	PROPN
fcis-9730	117	11	,	,	PUNCT
fcis-9730	117	12	this	this	DET
fcis-9730	117	13	paper	paper	NOUN
fcis-9730	117	14	introduces	introduce	VERB
fcis-9730	117	15	the	the	DET
fcis-9730	117	16	principle	principle	NOUN
fcis-9730	117	17	,	,	PUNCT
fcis-9730	117	18	innovation	innovation	NOUN
fcis-9730	117	19	points	point	NOUN
fcis-9730	117	20	,	,	PUNCT
fcis-9730	117	21	advantages	advantage	NOUN
fcis-9730	117	22	and	and	CCONJ
fcis-9730	117	23	disadvantages	disadvantage	NOUN
fcis-9730	117	24	,	,	PUNCT
fcis-9730	117	25	application	application	NOUN
fcis-9730	117	26	fields	field	NOUN
fcis-9730	117	27	and	and	CCONJ
fcis-9730	117	28	future	future	ADJ
fcis-9730	117	29	development	development	NOUN
fcis-9730	117	30	trend	trend	NOUN
fcis-9730	117	31	of	of	ADP
fcis-9730	117	32	yolo	yolo	ADJ
fcis-9730	117	33	series	series	PROPN
fcis-9730	117	34	algorithms	algorithm	NOUN
fcis-9730	117	35	,	,	PUNCT
fcis-9730	117	36	and	and	CCONJ
fcis-9730	117	37	shows	show	VERB
fcis-9730	117	38	the	the	DET
fcis-9730	117	39	effect	effect	NOUN
fcis-9730	117	40	that	that	SCONJ
fcis-9730	117	41	yolo	yolo	ADJ
fcis-9730	117	42	series	series	NOUN
fcis-9730	117	43	algorithms	algorithm	NOUN
fcis-9730	117	44	can	can	AUX
fcis-9730	117	45	achieve	achieve	VERB
fcis-9730	117	46	at	at	ADP
fcis-9730	117	47	present	present	NOUN
fcis-9730	117	48	,	,	PUNCT
fcis-9730	117	49	which	which	PRON
fcis-9730	117	50	can	can	AUX
fcis-9730	117	51	greatly	greatly	ADV
fcis-9730	117	52	improve	improve	VERB
fcis-9730	117	53	the	the	DET
fcis-9730	117	54	efficiency	efficiency	NOUN
fcis-9730	117	55	of	of	ADP
fcis-9730	117	56	production	production	NOUN
fcis-9730	117	57	and	and	CCONJ
fcis-9730	117	58	life	life	NOUN
fcis-9730	117	59	in	in	ADP
fcis-9730	117	60	the	the	DET
fcis-9730	117	61	application	application	NOUN
fcis-9730	117	62	field	field	NOUN
fcis-9730	117	63	.	.	PUNCT
fcis-9730	118	1	it	it	PRON
fcis-9730	118	2	has	have	VERB
fcis-9730	118	3	great	great	ADJ
fcis-9730	118	4	application	application	NOUN
fcis-9730	118	5	prospects	prospect	NOUN
fcis-9730	118	6	in	in	ADP
fcis-9730	118	7	the	the	DET
fcis-9730	118	8	field	field	NOUN
fcis-9730	118	9	of	of	ADP
fcis-9730	118	10	safety	safety	NOUN
fcis-9730	118	11	monitoring	monitoring	NOUN
fcis-9730	118	12	with	with	ADP
fcis-9730	118	13	high	high	ADJ
fcis-9730	118	14	flexibility	flexibility	NOUN
fcis-9730	118	15	and	and	CCONJ
fcis-9730	118	16	timeliness	timeliness	NOUN
fcis-9730	118	17	.	.	PUNCT
fcis-9730	119	1	this	this	DET
fcis-9730	119	2	paper	paper	NOUN
fcis-9730	119	3	looks	look	VERB
fcis-9730	119	4	forward	forward	ADV
fcis-9730	119	5	to	to	ADP
fcis-9730	119	6	yolo	yolo	ADJ
fcis-9730	119	7	series	series	NOUN
fcis-9730	119	8	and	and	CCONJ
fcis-9730	119	9	provides	provide	VERB
fcis-9730	119	10	ideas	idea	NOUN
fcis-9730	119	11	for	for	ADP
fcis-9730	119	12	the	the	DET
fcis-9730	119	13	future	future	ADJ
fcis-9730	119	14	development	development	NOUN
fcis-9730	119	15	of	of	ADP
fcis-9730	119	16	yolo	yolo	ADJ
fcis-9730	119	17	series	series	NOUN
fcis-9730	119	18	.	.	PUNCT
fcis-9730	120	1	references	reference	NOUN
fcis-9730	120	2	[	[	X
fcis-9730	120	3	1	1	X
fcis-9730	120	4	]	]	X
fcis-9730	120	5	liu	liu	PROPN
fcis-9730	120	6	li	li	PROPN
fcis-9730	120	7	,	,	PUNCT
fcis-9730	120	8	ouyang	ouyang	PROPN
fcis-9730	120	9	wanli	wanli	PROPN
fcis-9730	120	10	,	,	PUNCT
fcis-9730	120	11	wang	wang	PROPN
fcis-9730	120	12	xiaogang	xiaogang	PROPN
fcis-9730	120	13	,	,	PUNCT
fcis-9730	120	14	et	et	PROPN
fcis-9730	120	15	al	al	PROPN
fcis-9730	120	16	.	.	PUNCT
fcis-9730	121	1	deep	deep	ADJ
fcis-9730	121	2	learning	learning	NOUN
fcis-9730	121	3	for	for	ADP
fcis-9730	121	4	generic	generic	ADJ
fcis-9730	121	5	object	object	NOUN
fcis-9730	121	6	detection	detection	NOUN
fcis-9730	121	7	:	:	PUNCT
fcis-9730	121	8	a	a	DET
fcis-9730	121	9	survey[j	survey[j	PROPN
fcis-9730	121	10	]	]	PUNCT
fcis-9730	121	11	.	.	PUNCT
fcis-9730	122	1	international	international	ADJ
fcis-9730	122	2	journal	journal	PROPN
fcis-9730	122	3	of	of	ADP
fcis-9730	122	4	computer	computer	NOUN
fcis-9730	122	5	vision	vision	NOUN
fcis-9730	122	6	,	,	PUNCT
fcis-9730	122	7	2020	2020	NUM
fcis-9730	122	8	,	,	PUNCT
fcis-9730	122	9	128(2	128(2	NUM
fcis-9730	122	10	):	):	PUNCT
fcis-9730	122	11	261	261	NUM
fcis-9730	122	12	-318	-318	PROPN
fcis-9730	122	13	.	.	PUNCT
fcis-9730	123	1	doi	doi	NOUN
fcis-9730	123	2	:	:	PUNCT
fcis-9730	123	3	10	10	NUM
fcis-9730	123	4	.	.	X
fcis-9730	124	1	1007/	1007/	NUM
fcis-9730	124	2	s11263	s11263	NUM
fcis-9730	124	3	-	-	PUNCT
fcis-9730	124	4	019	019	NUM
fcis-9730	124	5	-	-	PUNCT
fcis-9730	124	6	01247	01247	NUM
fcis-9730	124	7	-	-	PUNCT
fcis-9730	124	8	4	4	NUM
fcis-9730	124	9	.	.	PUNCT
fcis-9730	125	1	[	[	X
fcis-9730	125	2	2	2	NUM
fcis-9730	125	3	]	]	PUNCT
fcis-9730	125	4	zou	zou	PROPN
fcis-9730	125	5	zhengxia	zhengxia	PROPN
fcis-9730	125	6	,	,	PUNCT
fcis-9730	125	7	shi	shi	PROPN
fcis-9730	125	8	zhenwei	zhenwei	PROPN
fcis-9730	125	9	,	,	PUNCT
fcis-9730	125	10	guo	guo	PROPN
fcis-9730	125	11	yuhong	yuhong	PROPN
fcis-9730	125	12	,	,	PUNCT
fcis-9730	125	13	et	et	PROPN
fcis-9730	125	14	al	al	PROPN
fcis-9730	125	15	.	.	PROPN
fcis-9730	125	16	object	object	PROPN
fcis-9730	125	17	detection	detection	NOUN
fcis-9730	125	18	in	in	ADP
fcis-9730	125	19	20	20	NUM
fcis-9730	125	20	years	year	NOUN
fcis-9730	125	21	:	:	PUNCT
fcis-9730	125	22	a	a	DET
fcis-9730	125	23	survey[j	survey[j	PROPN
fcis-9730	125	24	]	]	PUNCT
fcis-9730	125	25	.	.	PUNCT
fcis-9730	126	1	arxiv	arxiv	PROPN
fcis-9730	126	2	preprint	preprint	PROPN
fcis-9730	126	3	arxiv	arxiv	PROPN
fcis-9730	126	4	:	:	PUNCT
fcis-9730	126	5	1905.05055	1905.05055	NOUN
fcis-9730	126	6	,	,	PUNCT
fcis-9730	126	7	2019	2019	NUM
fcis-9730	126	8	.	.	PUNCT
fcis-9730	126	9	20	20	NUM
fcis-9730	127	1	[	[	SYM
fcis-9730	127	2	3	3	X
fcis-9730	127	3	]	]	X
fcis-9730	127	4	dalal	dalal	PROPN
fcis-9730	127	5	n	n	PROPN
fcis-9730	127	6	and	and	CCONJ
fcis-9730	127	7	triggs	triggs	PROPN
fcis-9730	127	8	b.	b.	PROPN
fcis-9730	127	9	histograms	histograms	PROPN
fcis-9730	127	10	of	of	ADP
fcis-9730	127	11	oriented	orient	VERB
fcis-9730	127	12	gradients	gradient	NOUN
fcis-9730	127	13	for	for	ADP
fcis-9730	127	14	human	human	NOUN
fcis-9730	127	15	detection[c	detection[c	PROPN
fcis-9730	127	16	]	]	PUNCT
fcis-9730	127	17	.	.	PUNCT
fcis-9730	128	1	2005	2005	NUM
fcis-9730	128	2	ieee	ieee	PROPN
fcis-9730	128	3	computer	computer	NOUN
fcis-9730	128	4	society	society	PROPN
fcis-9730	128	5	conference	conference	NOUN
fcis-9730	128	6	on	on	ADP
fcis-9730	128	7	computer	computer	NOUN
fcis-9730	128	8	vision	vision	NOUN
fcis-9730	128	9	and	and	CCONJ
fcis-9730	128	10	pattern	pattern	NOUN
fcis-9730	128	11	recognition	recognition	NOUN
fcis-9730	128	12	,	,	PUNCT
fcis-9730	128	13	san	san	PROPN
fcis-9730	128	14	diego	diego	PROPN
fcis-9730	128	15	,	,	PUNCT
fcis-9730	128	16	usa	usa	PROPN
fcis-9730	128	17	,	,	PUNCT
fcis-9730	128	18	2005:886	2005:886	NUM
fcis-9730	128	19	-	-	PUNCT
fcis-9730	128	20	893	893	NUM
fcis-9730	128	21	.	.	PUNCT
fcis-9730	129	1	doi	doi	NOUN
fcis-9730	129	2	:	:	PUNCT
fcis-9730	129	3	10.1109/	10.1109/	NUM
fcis-9730	129	4	cvpr.2005.177	cvpr.2005.177	NOUN
fcis-9730	129	5	.	.	PUNCT
fcis-9730	130	1	[	[	X
fcis-9730	130	2	4	4	X
fcis-9730	130	3	]	]	X
fcis-9730	130	4	liu	liu	PROPN
fcis-9730	130	5	pu	pu	PROPN
fcis-9730	130	6	,	,	PUNCT
fcis-9730	130	7	zhang	zhang	PROPN
fcis-9730	130	8	xing	xing	PROPN
fcis-9730	130	9	-	-	PUNCT
fcis-9730	130	10	hui	hui	PROPN
fcis-9730	130	11	,	,	PUNCT
fcis-9730	130	12	zhang	zhang	PROPN
fcis-9730	130	13	zhi	zhi	PROPN
fcis-9730	130	14	-	-	PUNCT
fcis-9730	130	15	li	li	PROPN
fcis-9730	130	16	et	et	PROPN
fcis-9730	130	17	al	al	PROPN
fcis-9730	130	18	.	.	PROPN
fcis-9730	130	19	overview	overview	NOUN
fcis-9730	130	20	of	of	ADP
fcis-9730	130	21	object	object	NOUN
fcis-9730	130	22	detection	detection	NOUN
fcis-9730	130	23	from	from	ADP
fcis-9730	130	24	rcnn	rcnn	PROPN
fcis-9730	130	25	to	to	ADP
fcis-9730	130	26	yolo	yolo	PRON
fcis-9730	130	27	[	[	X
fcis-9730	130	28	c]//	c]//	PROPN
fcis-9730	130	29	china	china	PROPN
fcis-9730	130	30	high	high	ADJ
fcis-9730	130	31	-	-	PUNCT
fcis-9730	130	32	tech	tech	NOUN
fcis-9730	130	33	industrialization	industrialization	NOUN
fcis-9730	130	34	research	research	NOUN
fcis-9730	130	35	association	association	NOUN
fcis-9730	130	36	intelligent	intelligent	ADJ
fcis-9730	130	37	information	information	NOUN
fcis-9730	130	38	processing	process	VERB
fcis-9730	130	39	industrialization	industrialization	NOUN
fcis-9730	130	40	branch	branch	NOUN
fcis-9730	130	41	.	.	PUNCT
fcis-9730	131	1	the	the	DET
fcis-9730	131	2	sixteenth	sixteenth	ADJ
fcis-9730	131	3	national	national	ADJ
fcis-9730	131	4	signal	signal	NOUN
fcis-9730	131	5	and	and	CCONJ
fcis-9730	131	6	intelligent	intelligent	ADJ
fcis-9730	131	7	information	information	NOUN
fcis-9730	131	8	processing	processing	NOUN
fcis-9730	131	9	and	and	CCONJ
fcis-9730	131	10	application	application	NOUN
fcis-9730	131	11	of	of	ADP
fcis-9730	131	12	academic	academic	ADJ
fcis-9730	131	13	conference	conference	NOUN
fcis-9730	131	14	proceedings	proceeding	NOUN
fcis-9730	131	15	.	.	PUNCT
fcis-9730	132	1	[	[	X
fcis-9730	132	2	publisher	publisher	NOUN
fcis-9730	132	3	unknown	unknown	ADJ
fcis-9730	132	4	]	]	PUNCT
fcis-9730	132	5	,	,	PUNCT
fcis-9730	132	6	2022:8	2022:8	NUM
fcis-9730	132	7	.	.	PUNCT
fcis-9730	133	1	doi	doi	NOUN
fcis-9730	133	2	:	:	PUNCT
fcis-9730	133	3	10.26914	10.26914	NUM
fcis-9730	133	4	/	/	SYM
fcis-9730	133	5	arthur	arthur	PROPN
fcis-9730	133	6	c.	c.	PROPN
fcis-9730	133	7	nkihy	nkihy	PROPN
fcis-9730	133	8	.	.	PUNCT
fcis-9730	134	1	2022.053359	2022.053359	INTJ
fcis-9730	134	2	.	.	PUNCT
fcis-9730	135	1	[	[	X
fcis-9730	135	2	5	5	X
fcis-9730	135	3	]	]	X
fcis-9730	135	4	liu	liu	PROPN
fcis-9730	135	5	yang	yang	PROPN
fcis-9730	135	6	.	.	PUNCT
fcis-9730	136	1	masks	mask	NOUN
fcis-9730	136	2	worn	wear	VERB
fcis-9730	136	3	under	under	ADP
fcis-9730	136	4	the	the	DET
fcis-9730	136	5	dense	dense	ADJ
fcis-9730	136	6	crowd	crowd	NOUN
fcis-9730	136	7	scene	scene	NOUN
fcis-9730	136	8	detection	detection	NOUN
fcis-9730	136	9	algorithm	algorithm	NOUN
fcis-9730	136	10	research	research	NOUN
fcis-9730	137	1	[	[	X
fcis-9730	137	2	d	d	X
fcis-9730	137	3	]	]	X
fcis-9730	137	4	.	.	PUNCT
fcis-9730	138	1	hebei	hebei	PROPN
fcis-9730	138	2	university	university	PROPN
fcis-9730	138	3	of	of	ADP
fcis-9730	138	4	engineering	engineering	NOUN
fcis-9730	138	5	,	,	PUNCT
fcis-9730	138	6	2022	2022	NUM
fcis-9730	138	7	.	.	PUNCT
fcis-9730	139	1	the	the	DET
fcis-9730	139	2	doi	doi	NOUN
fcis-9730	139	3	:	:	PUNCT
fcis-9730	139	4	10.27104	10.27104	NUM
fcis-9730	139	5	/	/	SYM
fcis-9730	139	6	,	,	PUNCT
fcis-9730	139	7	dc	dc	PROPN
fcis-9730	139	8	nki	nki	PROPN
fcis-9730	139	9	.	.	PROPN
fcis-9730	139	10	ghbjy	ghbjy	PROPN
fcis-9730	139	11	.	.	PUNCT
fcis-9730	140	1	2022.000365	2022.000365	X
fcis-9730	140	2	.	.	PUNCT
fcis-9730	141	1	[	[	X
fcis-9730	141	2	6	6	NUM
fcis-9730	141	3	]	]	X
fcis-9730	141	4	shi	shi	PROPN
fcis-9730	141	5	duan	duan	PROPN
fcis-9730	141	6	-	-	PUNCT
fcis-9730	141	7	yang	yang	PROPN
fcis-9730	141	8	,	,	PUNCT
fcis-9730	141	9	lin	lin	PROPN
fcis-9730	141	10	qiang	qiang	PROPN
fcis-9730	141	11	,	,	PUNCT
fcis-9730	141	12	hu	hu	PROPN
fcis-9730	141	13	bing	bing	PROPN
fcis-9730	141	14	,	,	PUNCT
fcis-9730	141	15	zhang	zhang	PROPN
fcis-9730	141	16	xin	xin	PROPN
fcis-9730	141	17	-	-	PUNCT
fcis-9730	141	18	yu	yu	PROPN
fcis-9730	141	19	.	.	PUNCT
fcis-9730	141	20	radar	radar	NOUN
fcis-9730	141	21	science	science	NOUN
fcis-9730	141	22	and	and	CCONJ
fcis-9730	141	23	technology,2022,20(06):589	technology,2022,20(06):589	NUM
fcis-9730	141	24	-	-	SYM
fcis-9730	141	25	605	605	NUM
fcis-9730	141	26	.	.	PUNCT
fcis-9730	142	1	[	[	X
fcis-9730	142	2	7	7	X
fcis-9730	142	3	]	]	X
fcis-9730	142	4	zhou	zhou	PROPN
fcis-9730	142	5	shujuan	shujuan	PROPN
fcis-9730	142	6	,	,	PUNCT
fcis-9730	142	7	zhang	zhang	PROPN
fcis-9730	142	8	yu	yu	PROPN
fcis-9730	142	9	,	,	PUNCT
fcis-9730	142	10	zhang	zhang	PROPN
fcis-9730	142	11	heng	heng	PROPN
fcis-9730	142	12	,	,	PUNCT
fcis-9730	142	13	wang	wang	PROPN
fcis-9730	142	14	qi	qi	PROPN
fcis-9730	142	15	,	,	PUNCT
fcis-9730	142	16	liu	liu	PROPN
fcis-9730	142	17	guangjie	guangjie	PROPN
fcis-9730	142	18	,	,	PUNCT
fcis-9730	142	19	zhu	zhu	PROPN
fcis-9730	142	20	jinlong	jinlong	PROPN
fcis-9730	142	21	.	.	PUNCT
fcis-9730	143	1	object	object	NOUN
fcis-9730	143	2	recognition	recognition	NOUN
fcis-9730	143	3	of	of	ADP
fcis-9730	143	4	tem	tem	PROPN
fcis-9730	143	5	nanoparticle	nanoparticle	NOUN
fcis-9730	143	6	structures	structure	NOUN
fcis-9730	143	7	based	base	VERB
fcis-9730	143	8	on	on	ADP
fcis-9730	143	9	deep	deep	ADJ
fcis-9730	143	10	learning	learning	NOUN
fcis-9730	144	1	[	[	X
fcis-9730	144	2	j	j	X
fcis-9730	144	3	]	]	X
fcis-9730	144	4	.	.	PUNCT
fcis-9730	145	1	journal	journal	PROPN
fcis-9730	145	2	of	of	ADP
fcis-9730	145	3	changchun	changchun	PROPN
fcis-9730	145	4	normal	normal	ADJ
fcis-9730	145	5	university,2022,41(12):35	university,2022,41(12):35	NOUN
fcis-9730	145	6	-	-	PUNCT
fcis-9730	145	7	40	40	NUM
fcis-9730	145	8	.	.	PUNCT
fcis-9730	146	1	[	[	X
fcis-9730	146	2	8	8	NUM
fcis-9730	146	3	]	]	X
fcis-9730	146	4	zhang	zhang	PROPN
fcis-9730	146	5	shan	shan	PROPN
fcis-9730	146	6	,	,	PUNCT
fcis-9730	146	7	lu	lu	PROPN
fcis-9730	146	8	yujiao	yujiao	NOUN
fcis-9730	146	9	,	,	PUNCT
fcis-9730	146	10	luo	luo	PROPN
fcis-9730	146	11	dawei	dawei	PROPN
fcis-9730	146	12	.	.	PUNCT
fcis-9730	147	1	overview	overview	NOUN
fcis-9730	147	2	of	of	ADP
fcis-9730	147	3	object	object	NOUN
fcis-9730	147	4	detection	detection	NOUN
fcis-9730	147	5	algorithms	algorithm	NOUN
fcis-9730	147	6	based	base	VERB
fcis-9730	147	7	on	on	ADP
fcis-9730	147	8	deep	deep	ADJ
fcis-9730	147	9	learning	learning	NOUN
fcis-9730	147	10	[	[	PUNCT
fcis-9730	147	11	c]//	c]//	ADJ
fcis-9730	147	12	proceedings	proceeding	NOUN
fcis-9730	147	13	of	of	ADP
fcis-9730	147	14	the	the	DET
fcis-9730	147	15	12th	12th	ADJ
fcis-9730	147	16	annual	annual	ADJ
fcis-9730	147	17	conference	conference	NOUN
fcis-9730	147	18	on	on	ADP
fcis-9730	147	19	new	new	ADJ
fcis-9730	147	20	network	network	NOUN
fcis-9730	147	21	technologies	technology	NOUN
fcis-9730	147	22	and	and	CCONJ
fcis-9730	147	23	applications	application	NOUN
fcis-9730	147	24	,	,	PUNCT
fcis-9730	147	25	network	network	NOUN
fcis-9730	147	26	application	application	NOUN
fcis-9730	147	27	branch	branch	NOUN
fcis-9730	147	28	,	,	PUNCT
fcis-9730	147	29	china	china	PROPN
fcis-9730	147	30	computer	computer	PROPN
fcis-9730	147	31	users	users	PROPN
fcis-9730	147	32	association	association	PROPN
fcis-9730	147	33	,	,	PUNCT
fcis-9730	147	34	2018	2018	NUM
fcis-9730	147	35	.	.	PUNCT
fcis-9730	148	1	[	[	X
fcis-9730	148	2	9	9	NUM
fcis-9730	148	3	]	]	X
fcis-9730	148	4	zhou	zhou	NOUN
fcis-9730	148	5	xiaoyan	xiaoyan	PROPN
fcis-9730	148	6	,	,	PUNCT
fcis-9730	148	7	wang	wang	PROPN
fcis-9730	148	8	ke	ke	PROPN
fcis-9730	148	9	,	,	PUNCT
fcis-9730	148	10	li	li	PROPN
fcis-9730	148	11	lingyan	lingyan	PROPN
fcis-9730	148	12	.	.	PUNCT
fcis-9730	149	1	an	an	DET
fcis-9730	149	2	overview	overview	NOUN
fcis-9730	149	3	of	of	ADP
fcis-9730	149	4	object	object	NOUN
fcis-9730	149	5	detection	detection	NOUN
fcis-9730	149	6	algorithms	algorithm	NOUN
fcis-9730	149	7	based	base	VERB
fcis-9730	149	8	on	on	ADP
fcis-9730	149	9	deep	deep	ADJ
fcis-9730	149	10	learning	learning	NOUN
fcis-9730	149	11	[	[	X
fcis-9730	149	12	j	j	X
fcis-9730	149	13	]	]	X
fcis-9730	149	14	.	.	PUNCT
fcis-9730	150	1	electronic	electronic	ADJ
fcis-9730	150	2	measurement	measurement	NOUN
fcis-9730	150	3	technology	technology	NOUN
fcis-9730	150	4	,	,	PUNCT
fcis-9730	150	5	2017	2017	NUM
fcis-9730	150	6	,	,	PUNCT
fcis-9730	150	7	40(11):5	40(11):5	NOUN
fcis-9730	150	8	.	.	PUNCT
fcis-9730	151	1	[	[	X
fcis-9730	151	2	10	10	NUM
fcis-9730	151	3	]	]	X
fcis-9730	151	4	zhang	zhang	PROPN
fcis-9730	151	5	qi	qi	PROPN
fcis-9730	151	6	,	,	PUNCT
fcis-9730	151	7	zhang	zhang	PROPN
fcis-9730	151	8	rongmei	rongmei	PROPN
fcis-9730	151	9	,	,	PUNCT
fcis-9730	151	10	chen	chen	PROPN
fcis-9730	151	11	bin	bin	PROPN
fcis-9730	151	12	.	.	PUNCT
fcis-9730	152	1	a	a	DET
fcis-9730	152	2	review	review	NOUN
fcis-9730	152	3	of	of	ADP
fcis-9730	152	4	image	image	NOUN
fcis-9730	152	5	recognition	recognition	NOUN
fcis-9730	152	6	technology	technology	NOUN
fcis-9730	152	7	based	base	VERB
fcis-9730	152	8	on	on	ADP
fcis-9730	152	9	deep	deep	ADJ
fcis-9730	152	10	learning	learning	NOUN
fcis-9730	152	11	[	[	X
fcis-9730	152	12	j	j	X
fcis-9730	152	13	]	]	X
fcis-9730	152	14	.	.	PUNCT
fcis-9730	153	1	2019	2019	NUM
fcis-9730	153	2	.	.	PUNCT
fcis-9730	154	1	[	[	X
fcis-9730	154	2	11	11	NUM
fcis-9730	154	3	]	]	X
fcis-9730	154	4	zheng	zheng	PROPN
fcis-9730	154	5	wei	wei	PROPN
fcis-9730	154	6	-	-	PROPN
fcis-9730	154	7	cheng	cheng	PROPN
fcis-9730	154	8	,	,	PUNCT
fcis-9730	154	9	li	li	PROPN
fcis-9730	154	10	xue	xue	PROPN
fcis-9730	154	11	-	-	PUNCT
fcis-9730	154	12	wei	wei	PROPN
fcis-9730	154	13	,	,	PUNCT
fcis-9730	154	14	liu	liu	PROPN
fcis-9730	154	15	hong	hong	PROPN
fcis-9730	154	16	-	-	PUNCT
fcis-9730	154	17	zhe	zhe	PROPN
fcis-9730	154	18	.	.	PROPN
fcis-9730	155	1	overview	overview	NOUN
fcis-9730	155	2	of	of	ADP
fcis-9730	155	3	object	object	NOUN
fcis-9730	155	4	detection	detection	NOUN
fcis-9730	155	5	algorithms	algorithm	NOUN
fcis-9730	155	6	based	base	VERB
fcis-9730	155	7	on	on	ADP
fcis-9730	155	8	deep	deep	ADJ
fcis-9730	155	9	learning	learning	NOUN
fcis-9730	155	10	[	[	AUX
fcis-9730	155	11	c]//	c]//	PROPN
fcis-9730	155	12	the	the	DET
fcis-9730	155	13	22nd	22nd	ADJ
fcis-9730	155	14	annual	annual	ADJ
fcis-9730	155	15	conference	conference	NOUN
fcis-9730	155	16	on	on	ADP
fcis-9730	155	17	new	new	ADJ
fcis-9730	155	18	network	network	NOUN
fcis-9730	155	19	technologies	technology	NOUN
fcis-9730	155	20	and	and	CCONJ
fcis-9730	155	21	applications	application	NOUN
fcis-9730	155	22	,	,	PUNCT
fcis-9730	155	23	2018	2018	NUM
fcis-9730	155	24	,	,	PUNCT
fcis-9730	155	25	network	network	NOUN
fcis-9730	155	26	application	application	NOUN
fcis-9730	155	27	branch	branch	NOUN
fcis-9730	155	28	of	of	ADP
fcis-9730	155	29	china	china	PROPN
fcis-9730	155	30	computer	computer	PROPN
fcis-9730	155	31	users	users	PROPN
fcis-9730	155	32	association	association	PROPN
fcis-9730	155	33	.	.	PUNCT
fcis-9730	156	1	0	0	PUNCT
fcis-9730	156	2	.	.	PUNCT
fcis-9730	157	1	[	[	X
fcis-9730	157	2	12	12	NUM
fcis-9730	157	3	]	]	X
fcis-9730	157	4	liu	liu	PROPN
fcis-9730	157	5	yan	yan	PROPN
fcis-9730	157	6	-	-	PUNCT
fcis-9730	157	7	qing	qing	PROPN
fcis-9730	157	8	.	.	PUNCT
fcis-9730	158	1	improved	improve	VERB
fcis-9730	158	2	object	object	NOUN
fcis-9730	158	3	detection	detection	NOUN
fcis-9730	158	4	algorithm	algorithm	NOUN
fcis-9730	158	5	based	base	VERB
fcis-9730	158	6	on	on	ADP
fcis-9730	158	7	yolo	yolo	ADJ
fcis-9730	158	8	series	series	NOUN
fcis-9730	159	1	[	[	X
fcis-9730	159	2	d	d	X
fcis-9730	159	3	]	]	X
fcis-9730	159	4	.	.	PUNCT
fcis-9730	160	1	jilin	jilin	PROPN
fcis-9730	160	2	university	university	PROPN
fcis-9730	160	3	.	.	PUNCT
fcis-9730	161	1	[	[	X
fcis-9730	161	2	13	13	NUM
fcis-9730	161	3	]	]	X
fcis-9730	161	4	zheng	zheng	PROPN
fcis-9730	161	5	weicheng	weicheng	PROPN
fcis-9730	161	6	,	,	PUNCT
fcis-9730	161	7	li	li	PROPN
fcis-9730	161	8	xuewei	xuewei	PROPN
fcis-9730	161	9	,	,	PUNCT
fcis-9730	161	10	liu	liu	PROPN
fcis-9730	161	11	hongzhe	hongzhe	PROPN
fcis-9730	161	12	.	.	PUNCT
fcis-9730	162	1	an	an	DET
fcis-9730	162	2	overview	overview	NOUN
fcis-9730	162	3	of	of	ADP
fcis-9730	162	4	object	object	NOUN
fcis-9730	162	5	detection	detection	NOUN
fcis-9730	162	6	algorithms	algorithm	NOUN
fcis-9730	162	7	based	base	VERB
fcis-9730	162	8	on	on	ADP
fcis-9730	162	9	deep	deep	ADJ
fcis-9730	162	10	learning	learning	NOUN
fcis-9730	162	11	[	[	X
fcis-9730	162	12	j	j	X
fcis-9730	162	13	]	]	X
fcis-9730	162	14	.	.	PUNCT
fcis-9730	163	1	china	china	PROPN
fcis-9730	163	2	broadband	broadband	PROPN
fcis-9730	163	3	,	,	PUNCT
fcis-9730	163	4	2022(3):3	2022(3):3	PROPN
fcis-9730	163	5	.	.	PUNCT
fcis-9730	164	1	[	[	X
fcis-9730	164	2	14	14	NUM
fcis-9730	164	3	]	]	X
fcis-9730	164	4	li	li	PROPN
fcis-9730	164	5	bingzhen	bingzhen	PROPN
fcis-9730	164	6	,	,	PUNCT
fcis-9730	164	7	jiang	jiang	PROPN
fcis-9730	164	8	wenzhi	wenzhi	PROPN
fcis-9730	164	9	,	,	PUNCT
fcis-9730	164	10	gu	gu	NOUN
fcis-9730	164	11	guide	guide	NOUN
fcis-9730	164	12	,	,	PUNCT
fcis-9730	164	13	et	et	PROPN
fcis-9730	164	14	al	al	PROPN
fcis-9730	164	15	.	.	PROPN
fcis-9730	165	1	review	review	NOUN
fcis-9730	165	2	of	of	ADP
fcis-9730	165	3	object	object	NOUN
fcis-9730	165	4	detection	detection	NOUN
fcis-9730	165	5	algorithms	algorithm	NOUN
fcis-9730	165	6	based	base	VERB
fcis-9730	165	7	on	on	ADP
fcis-9730	165	8	convolutional	convolutional	ADJ
fcis-9730	165	9	neural	neural	ADJ
fcis-9730	165	10	networks	network	NOUN
fcis-9730	166	1	[	[	X
fcis-9730	166	2	j	j	X
fcis-9730	166	3	]	]	X
fcis-9730	166	4	.	.	PUNCT
fcis-9730	167	1	computer	computer	NOUN
fcis-9730	167	2	and	and	CCONJ
fcis-9730	167	3	digital	digital	PROPN
fcis-9730	167	4	engineering	engineering	NOUN
fcis-9730	167	5	,	,	PUNCT
fcis-9730	167	6	2022(005):050	2022(005):050	NUM
fcis-9730	167	7	.	.	PUNCT
fcis-9730	168	1	[	[	X
fcis-9730	168	2	15	15	NUM
fcis-9730	168	3	]	]	X
fcis-9730	168	4	guo	guo	PROPN
fcis-9730	168	5	zefang	zefang	PROPN
fcis-9730	168	6	.	.	PUNCT
fcis-9730	169	1	overview	overview	NOUN
fcis-9730	169	2	of	of	ADP
fcis-9730	169	3	deep	deep	ADJ
fcis-9730	169	4	learning	learning	NOUN
fcis-9730	169	5	algorithms	algorithm	NOUN
fcis-9730	169	6	for	for	ADP
fcis-9730	169	7	image	image	NOUN
fcis-9730	169	8	object	object	NOUN
fcis-9730	169	9	detection	detection	NOUN
fcis-9730	170	1	[	[	X
fcis-9730	170	2	j	j	X
fcis-9730	170	3	]	]	X
fcis-9730	170	4	.	.	PUNCT
fcis-9730	171	1	mechanical	mechanical	ADJ
fcis-9730	171	2	engineering	engineering	PROPN
fcis-9730	171	3	&	&	CCONJ
fcis-9730	171	4	automation	automation	PROPN
fcis-9730	171	5	,	,	PUNCT
fcis-9730	171	6	2019	2019	NUM
fcis-9730	171	7	(	(	PUNCT
fcis-9730	171	8	1):4	1):4	NUM
fcis-9730	171	9	.	.	PUNCT
fcis-9730	172	1	[	[	X
fcis-9730	172	2	16	16	NUM
fcis-9730	172	3	]	]	X
fcis-9730	172	4	wang	wang	PROPN
fcis-9730	172	5	yan	yan	PROPN
fcis-9730	172	6	.	.	PUNCT
fcis-9730	173	1	research	research	NOUN
fcis-9730	173	2	on	on	ADP
fcis-9730	173	3	visual	visual	ADJ
fcis-9730	173	4	detection	detection	NOUN
fcis-9730	173	5	and	and	CCONJ
fcis-9730	173	6	tracking	tracking	NOUN
fcis-9730	173	7	technology	technology	NOUN
fcis-9730	173	8	of	of	ADP
fcis-9730	173	9	surgical	surgical	ADJ
fcis-9730	173	10	instruments	instrument	NOUN
fcis-9730	173	11	based	base	VERB
fcis-9730	173	12	on	on	ADP
fcis-9730	173	13	deep	deep	ADJ
fcis-9730	173	14	learning	learning	NOUN
fcis-9730	173	15	.	.	PUNCT