id	sid	tid	token	lemma	pos
ajst-7358	1	1	academic	academic	ADJ
ajst-7358	1	2	journal	journal	NOUN
ajst-7358	1	3	of	of	ADP
ajst-7358	1	4	science	science	NOUN
ajst-7358	1	5	and	and	CCONJ
ajst-7358	1	6	technology	technology	NOUN
ajst-7358	1	7	issn	issn	NOUN
ajst-7358	1	8	:	:	PUNCT
ajst-7358	1	9	2771	2771	NUM
ajst-7358	1	10	-	-	SYM
ajst-7358	1	11	3032	3032	NUM
ajst-7358	1	12	|	|	NOUN
ajst-7358	1	13	vol	vol	NOUN
ajst-7358	1	14	.	.	PROPN
ajst-7358	2	1	5	5	NUM
ajst-7358	2	2	,	,	PUNCT
ajst-7358	2	3	no	no	INTJ
ajst-7358	2	4	.	.	NOUN
ajst-7358	2	5	3	3	NUM
ajst-7358	2	6	,	,	PUNCT
ajst-7358	2	7	2023	2023	NUM
ajst-7358	2	8	41	41	NUM
ajst-7358	2	9	research	research	NOUN
ajst-7358	2	10	on	on	ADP
ajst-7358	2	11	method	method	NOUN
ajst-7358	2	12	of	of	ADP
ajst-7358	2	13	creating	create	VERB
ajst-7358	2	14	dynamic	dynamic	ADJ
ajst-7358	2	15	weld	weld	NOUN
ajst-7358	2	16	of	of	ADP
ajst-7358	2	17	roi	roi	NOUN
ajst-7358	2	18	region	region	NOUN
ajst-7358	2	19	based	base	VERB
ajst-7358	2	20	on	on	ADP
ajst-7358	2	21	faster‐rcnn	faster‐rcnn	PROPN
ajst-7358	2	22	qiang	qiang	PROPN
ajst-7358	2	23	song	song	PROPN
ajst-7358	2	24	,	,	PUNCT
ajst-7358	2	25	chenlei	chenlei	PROPN
ajst-7358	2	26	zhao	zhao	PROPN
ajst-7358	2	27	,	,	PUNCT
ajst-7358	2	28	shenghong	shenghong	PROPN
ajst-7358	2	29	wu	wu	PROPN
ajst-7358	2	30	,	,	PUNCT
ajst-7358	2	31	xi	xi	PROPN
ajst-7358	2	32	xu	xu	PROPN
ajst-7358	2	33	geely	geely	PROPN
ajst-7358	2	34	university	university	PROPN
ajst-7358	2	35	of	of	ADP
ajst-7358	2	36	china	china	PROPN
ajst-7358	2	37	,	,	PUNCT
ajst-7358	2	38	sichuan	sichuan	PROPN
ajst-7358	2	39	,	,	PUNCT
ajst-7358	2	40	641423	641423	NUM
ajst-7358	2	41	,	,	PUNCT
ajst-7358	2	42	china	china	PROPN
ajst-7358	2	43	abstract	abstract	NOUN
ajst-7358	2	44	:	:	PUNCT
ajst-7358	2	45	aiming	aim	VERB
ajst-7358	2	46	at	at	ADP
ajst-7358	2	47	the	the	DET
ajst-7358	2	48	issues	issue	NOUN
ajst-7358	2	49	of	of	ADP
ajst-7358	2	50	weld	weld	NOUN
ajst-7358	2	51	marking	marking	NOUN
ajst-7358	2	52	noise	noise	NOUN
ajst-7358	2	53	in	in	ADP
ajst-7358	2	54	welding	welding	NOUN
ajst-7358	2	55	path	path	NOUN
ajst-7358	2	56	planning	planning	NOUN
ajst-7358	2	57	of	of	ADP
ajst-7358	2	58	the	the	DET
ajst-7358	2	59	third	third	ADJ
ajst-7358	2	60	generation	generation	NOUN
ajst-7358	2	61	welding	welding	NOUN
ajst-7358	2	62	robot	robot	NOUN
ajst-7358	2	63	,	,	PUNCT
ajst-7358	2	64	that	that	SCONJ
ajst-7358	2	65	the	the	DET
ajst-7358	2	66	creation	creation	NOUN
ajst-7358	2	67	of	of	ADP
ajst-7358	2	68	roi	roi	NOUN
ajst-7358	2	69	region	region	NOUN
ajst-7358	2	70	is	be	AUX
ajst-7358	2	71	employed	employ	VERB
ajst-7358	2	72	as	as	ADP
ajst-7358	2	73	the	the	DET
ajst-7358	2	74	approach	approach	NOUN
ajst-7358	2	75	to	to	ADP
ajst-7358	2	76	noise	noise	NOUN
ajst-7358	2	77	suppression	suppression	NOUN
ajst-7358	2	78	.	.	PUNCT
ajst-7358	3	1	however	however	ADV
ajst-7358	3	2	,	,	PUNCT
ajst-7358	3	3	traditional	traditional	ADJ
ajst-7358	3	4	roi	roi	NOUN
ajst-7358	3	5	region	region	NOUN
ajst-7358	3	6	construction	construction	NOUN
ajst-7358	3	7	methods	method	NOUN
ajst-7358	3	8	can	can	AUX
ajst-7358	3	9	only	only	ADV
ajst-7358	3	10	create	create	VERB
ajst-7358	3	11	roi	roi	NOUN
ajst-7358	3	12	regions	region	NOUN
ajst-7358	3	13	at	at	ADP
ajst-7358	3	14	a	a	DET
ajst-7358	3	15	fixed	fix	VERB
ajst-7358	3	16	location	location	NOUN
ajst-7358	3	17	by	by	ADP
ajst-7358	3	18	presetting	presette	VERB
ajst-7358	3	19	parameters	parameter	NOUN
ajst-7358	3	20	in	in	ADP
ajst-7358	3	21	the	the	DET
ajst-7358	3	22	system	system	NOUN
ajst-7358	3	23	.	.	PUNCT
ajst-7358	4	1	the	the	DET
ajst-7358	4	2	welding	welding	NOUN
ajst-7358	4	3	target	target	NOUN
ajst-7358	4	4	position	position	NOUN
ajst-7358	4	5	usually	usually	ADV
ajst-7358	4	6	produces	produce	VERB
ajst-7358	4	7	displacement	displacement	NOUN
ajst-7358	4	8	in	in	ADP
ajst-7358	4	9	the	the	DET
ajst-7358	4	10	control	control	NOUN
ajst-7358	4	11	range	range	NOUN
ajst-7358	4	12	of	of	ADP
ajst-7358	4	13	the	the	DET
ajst-7358	4	14	tolerance	tolerance	NOUN
ajst-7358	4	15	due	due	ADP
ajst-7358	4	16	to	to	ADP
ajst-7358	4	17	an	an	DET
ajst-7358	4	18	important	important	ADJ
ajst-7358	4	19	tolerance	tolerance	NOUN
ajst-7358	4	20	concept	concept	NOUN
ajst-7358	4	21	in	in	ADP
ajst-7358	4	22	the	the	DET
ajst-7358	4	23	welding	welding	NOUN
ajst-7358	4	24	process	process	NOUN
ajst-7358	4	25	,	,	PUNCT
ajst-7358	4	26	which	which	PRON
ajst-7358	4	27	may	may	AUX
ajst-7358	4	28	result	result	VERB
ajst-7358	4	29	in	in	ADP
ajst-7358	4	30	an	an	DET
ajst-7358	4	31	roi	roi	NOUN
ajst-7358	4	32	region	region	NOUN
ajst-7358	4	33	created	create	VERB
ajst-7358	4	34	with	with	ADP
ajst-7358	4	35	traditional	traditional	ADJ
ajst-7358	4	36	methods	method	NOUN
ajst-7358	4	37	is	be	AUX
ajst-7358	4	38	not	not	PART
ajst-7358	4	39	able	able	ADJ
ajst-7358	4	40	to	to	PART
ajst-7358	4	41	coincide	coincide	VERB
ajst-7358	4	42	with	with	ADP
ajst-7358	4	43	the	the	DET
ajst-7358	4	44	roi	roi	NOUN
ajst-7358	4	45	region	region	NOUN
ajst-7358	4	46	required	require	VERB
ajst-7358	4	47	by	by	ADP
ajst-7358	4	48	the	the	DET
ajst-7358	4	49	system	system	NOUN
ajst-7358	4	50	,	,	PUNCT
ajst-7358	4	51	thereby	thereby	ADV
ajst-7358	4	52	affecting	affect	VERB
ajst-7358	4	53	the	the	DET
ajst-7358	4	54	quality	quality	NOUN
ajst-7358	4	55	of	of	ADP
ajst-7358	4	56	the	the	DET
ajst-7358	4	57	welding	welding	NOUN
ajst-7358	4	58	.	.	PUNCT
ajst-7358	5	1	to	to	PART
ajst-7358	5	2	improve	improve	VERB
ajst-7358	5	3	the	the	DET
ajst-7358	5	4	location	location	NOUN
ajst-7358	5	5	accuracy	accuracy	NOUN
ajst-7358	5	6	of	of	ADP
ajst-7358	5	7	the	the	DET
ajst-7358	5	8	created	create	VERB
ajst-7358	5	9	roi	roi	NOUN
ajst-7358	5	10	region	region	NOUN
ajst-7358	5	11	,	,	PUNCT
ajst-7358	5	12	a	a	DET
ajst-7358	5	13	dynamic	dynamic	ADJ
ajst-7358	5	14	roi	roi	NOUN
ajst-7358	5	15	region	region	NOUN
ajst-7358	5	16	creation	creation	NOUN
ajst-7358	5	17	method	method	NOUN
ajst-7358	5	18	based	base	VERB
ajst-7358	5	19	on	on	ADP
ajst-7358	5	20	faster	fast	ADJ
ajst-7358	5	21	-	-	PUNCT
ajst-7358	5	22	rcnn	rcnn	NOUN
ajst-7358	5	23	target	target	NOUN
ajst-7358	5	24	detection	detection	NOUN
ajst-7358	5	25	algorithm	algorithm	NOUN
ajst-7358	5	26	was	be	AUX
ajst-7358	5	27	proposed	propose	VERB
ajst-7358	5	28	.	.	PUNCT
ajst-7358	6	1	experimental	experimental	ADJ
ajst-7358	6	2	results	result	NOUN
ajst-7358	6	3	show	show	VERB
ajst-7358	6	4	that	that	SCONJ
ajst-7358	6	5	this	this	DET
ajst-7358	6	6	method	method	NOUN
ajst-7358	6	7	effectively	effectively	ADV
ajst-7358	6	8	reduce	reduce	VERB
ajst-7358	6	9	weld	weld	NOUN
ajst-7358	6	10	marking	marking	NOUN
ajst-7358	6	11	noise	noise	NOUN
ajst-7358	6	12	.	.	PUNCT
ajst-7358	7	1	keywords	keyword	NOUN
ajst-7358	7	2	:	:	PUNCT
ajst-7358	7	3	faster	fast	ADJ
ajst-7358	7	4	-	-	PUNCT
ajst-7358	7	5	rcnn	rcnn	NOUN
ajst-7358	7	6	,	,	PUNCT
ajst-7358	7	7	weld	weld	VERB
ajst-7358	7	8	dynamic	dynamic	ADJ
ajst-7358	7	9	roi	roi	NOUN
ajst-7358	7	10	area	area	NOUN
ajst-7358	7	11	,	,	PUNCT
ajst-7358	7	12	weld	weld	NOUN
ajst-7358	7	13	marking	marking	NOUN
ajst-7358	7	14	.	.	PUNCT
ajst-7358	8	1	1	1	X
ajst-7358	8	2	.	.	X
ajst-7358	8	3	introduction	introduction	NOUN
ajst-7358	8	4	in	in	ADP
ajst-7358	8	5	the	the	DET
ajst-7358	8	6	process	process	NOUN
ajst-7358	8	7	of	of	ADP
ajst-7358	8	8	using	use	VERB
ajst-7358	8	9	forstner	forstner	NOUN
ajst-7358	8	10	operator	operator	NOUN
ajst-7358	8	11	to	to	PART
ajst-7358	8	12	quickly	quickly	ADV
ajst-7358	8	13	mark	mark	VERB
ajst-7358	8	14	weld	weld	NOUN
ajst-7358	8	15	features	feature	NOUN
ajst-7358	8	16	,	,	PUNCT
ajst-7358	8	17	a	a	DET
ajst-7358	8	18	large	large	ADJ
ajst-7358	8	19	number	number	NOUN
ajst-7358	8	20	of	of	ADP
ajst-7358	8	21	feature	feature	NOUN
ajst-7358	8	22	points	point	NOUN
ajst-7358	8	23	in	in	ADP
ajst-7358	8	24	non	non	ADJ
ajst-7358	8	25	-	-	ADJ
ajst-7358	8	26	weld	weld	ADJ
ajst-7358	8	27	areas	area	NOUN
ajst-7358	8	28	will	will	AUX
ajst-7358	8	29	be	be	AUX
ajst-7358	8	30	extracted	extract	VERB
ajst-7358	8	31	by	by	ADP
ajst-7358	8	32	forstner	forstner	ADJ
ajst-7358	8	33	algorithm	algorithm	NOUN
ajst-7358	8	34	on	on	ADP
ajst-7358	8	35	account	account	NOUN
ajst-7358	8	36	of	of	ADP
ajst-7358	8	37	those	those	DET
ajst-7358	8	38	reasons	reason	NOUN
ajst-7358	8	39	including	include	VERB
ajst-7358	8	40	object	object	NOUN
ajst-7358	8	41	construction	construction	NOUN
ajst-7358	8	42	and	and	CCONJ
ajst-7358	8	43	camera	camera	NOUN
ajst-7358	8	44	shooting	shooting	NOUN
ajst-7358	8	45	environment	environment	NOUN
ajst-7358	8	46	,	,	PUNCT
ajst-7358	8	47	resulting	result	VERB
ajst-7358	8	48	in	in	ADP
ajst-7358	8	49	lot	lot	NOUN
ajst-7358	8	50	of	of	ADP
ajst-7358	8	51	noise	noise	NOUN
ajst-7358	8	52	interference	interference	NOUN
ajst-7358	8	53	to	to	ADP
ajst-7358	8	54	the	the	DET
ajst-7358	8	55	system	system	NOUN
ajst-7358	8	56	.	.	PUNCT
ajst-7358	9	1	as	as	SCONJ
ajst-7358	9	2	shown	show	VERB
ajst-7358	9	3	in	in	ADP
ajst-7358	9	4	fig	fig	NOUN
ajst-7358	9	5	.	.	PUNCT
ajst-7358	10	1	1a	1a	NOUN
ajst-7358	10	2	,	,	PUNCT
ajst-7358	10	3	the	the	DET
ajst-7358	10	4	yellow	yellow	ADJ
ajst-7358	10	5	part	part	NOUN
ajst-7358	10	6	refers	refer	VERB
ajst-7358	10	7	the	the	DET
ajst-7358	10	8	marked	mark	VERB
ajst-7358	10	9	feature	feature	NOUN
ajst-7358	10	10	points	point	NOUN
ajst-7358	10	11	,	,	PUNCT
ajst-7358	10	12	and	and	CCONJ
ajst-7358	10	13	a	a	DET
ajst-7358	10	14	large	large	ADJ
ajst-7358	10	15	number	number	NOUN
ajst-7358	10	16	of	of	ADP
ajst-7358	10	17	non	non	ADJ
ajst-7358	10	18	-	-	ADJ
ajst-7358	10	19	target	target	ADJ
ajst-7358	10	20	noise	noise	NOUN
ajst-7358	10	21	points	point	NOUN
ajst-7358	10	22	are	be	AUX
ajst-7358	10	23	marked	mark	VERB
ajst-7358	10	24	in	in	ADP
ajst-7358	10	25	the	the	DET
ajst-7358	10	26	non	non	ADJ
ajst-7358	10	27	-	-	ADJ
ajst-7358	10	28	weld	weld	ADJ
ajst-7358	10	29	area	area	NOUN
ajst-7358	10	30	.	.	PUNCT
ajst-7358	11	1	these	these	DET
ajst-7358	11	2	noise	noise	NOUN
ajst-7358	11	3	feature	feature	NOUN
ajst-7358	11	4	points	point	NOUN
ajst-7358	11	5	will	will	AUX
ajst-7358	11	6	not	not	PART
ajst-7358	11	7	only	only	ADV
ajst-7358	11	8	increase	increase	VERB
ajst-7358	11	9	the	the	DET
ajst-7358	11	10	operating	operating	NOUN
ajst-7358	11	11	burden	burden	NOUN
ajst-7358	11	12	of	of	ADP
ajst-7358	11	13	the	the	DET
ajst-7358	11	14	system	system	NOUN
ajst-7358	11	15	,	,	PUNCT
ajst-7358	11	16	but	but	CCONJ
ajst-7358	11	17	also	also	ADV
ajst-7358	11	18	interfere	interfere	VERB
ajst-7358	11	19	with	with	ADP
ajst-7358	11	20	the	the	DET
ajst-7358	11	21	system	system	NOUN
ajst-7358	11	22	's	's	PART
ajst-7358	11	23	welding	welding	NOUN
ajst-7358	11	24	path	path	NOUN
ajst-7358	11	25	extraction	extraction	NOUN
ajst-7358	11	26	work	work	NOUN
ajst-7358	11	27	,	,	PUNCT
ajst-7358	11	28	and	and	CCONJ
ajst-7358	11	29	ultimately	ultimately	ADV
ajst-7358	11	30	affect	affect	VERB
ajst-7358	11	31	the	the	DET
ajst-7358	11	32	welding	welding	NOUN
ajst-7358	11	33	robot	robot	NOUN
ajst-7358	11	34	's	's	PART
ajst-7358	11	35	judgment	judgment	NOUN
ajst-7358	11	36	of	of	ADP
ajst-7358	11	37	the	the	DET
ajst-7358	11	38	welding	welding	NOUN
ajst-7358	11	39	path	path	NOUN
ajst-7358	11	40	.	.	PUNCT
ajst-7358	12	1	a	a	DET
ajst-7358	12	2	b	b	NUM
ajst-7358	12	3	figure	figure	NOUN
ajst-7358	12	4	1	1	NUM
ajst-7358	12	5	.	.	PUNCT
ajst-7358	12	6	comparison	comparison	NOUN
ajst-7358	12	7	of	of	ADP
ajst-7358	12	8	noise	noise	NOUN
ajst-7358	12	9	effects	effect	NOUN
ajst-7358	12	10	before	before	ADV
ajst-7358	12	11	and	and	CCONJ
ajst-7358	12	12	after	after	ADP
ajst-7358	12	13	the	the	DET
ajst-7358	12	14	construction	construction	NOUN
ajst-7358	12	15	of	of	ADP
ajst-7358	12	16	roi	roi	NOUN
ajst-7358	12	17	region	region	NOUN
ajst-7358	12	18	the	the	DET
ajst-7358	12	19	creation	creation	NOUN
ajst-7358	12	20	of	of	ADP
ajst-7358	12	21	roi	roi	NOUN
ajst-7358	12	22	area	area	NOUN
ajst-7358	12	23	of	of	ADP
ajst-7358	12	24	weld	weld	NOUN
ajst-7358	12	25	,	,	PUNCT
ajst-7358	12	26	narrowing	narrow	VERB
ajst-7358	12	27	the	the	DET
ajst-7358	12	28	target	target	NOUN
ajst-7358	12	29	area	area	NOUN
ajst-7358	12	30	of	of	ADP
ajst-7358	12	31	feature	feature	NOUN
ajst-7358	12	32	extraction	extraction	NOUN
ajst-7358	12	33	of	of	ADP
ajst-7358	12	34	forstner	forstner	NOUN
ajst-7358	12	35	operator	operator	NOUN
ajst-7358	12	36	is	be	AUX
ajst-7358	12	37	thus	thus	ADV
ajst-7358	12	38	used	use	VERB
ajst-7358	12	39	to	to	PART
ajst-7358	12	40	solve	solve	VERB
ajst-7358	12	41	the	the	DET
ajst-7358	12	42	problem	problem	NOUN
ajst-7358	12	43	of	of	ADP
ajst-7358	12	44	system	system	NOUN
ajst-7358	12	45	noise	noise	NOUN
ajst-7358	12	46	caused	cause	VERB
ajst-7358	12	47	by	by	ADP
ajst-7358	12	48	this	this	DET
ajst-7358	12	49	feature	feature	NOUN
ajst-7358	12	50	marking	mark	VERB
ajst-7358	12	51	process.as	process.as	PRON
ajst-7358	12	52	shown	show	VERB
ajst-7358	12	53	in	in	ADP
ajst-7358	12	54	the	the	DET
ajst-7358	12	55	red	red	PROPN
ajst-7358	12	56	box	box	PROPN
ajst-7358	12	57	in	in	ADP
ajst-7358	12	58	fig.1b	fig.1b	PROPN
ajst-7358	12	59	,	,	PUNCT
ajst-7358	12	60	the	the	DET
ajst-7358	12	61	area	area	NOUN
ajst-7358	12	62	where	where	SCONJ
ajst-7358	12	63	the	the	DET
ajst-7358	12	64	weld	weld	NOUN
ajst-7358	12	65	is	be	AUX
ajst-7358	12	66	located	locate	VERB
ajst-7358	12	67	is	be	AUX
ajst-7358	12	68	separated	separate	VERB
ajst-7358	12	69	and	and	CCONJ
ajst-7358	12	70	cut	cut	VERB
ajst-7358	12	71	out	out	ADP
ajst-7358	12	72	by	by	ADP
ajst-7358	12	73	the	the	DET
ajst-7358	12	74	target	target	NOUN
ajst-7358	12	75	detection	detection	NOUN
ajst-7358	12	76	algorithm	algorithm	NOUN
ajst-7358	12	77	and	and	CCONJ
ajst-7358	12	78	created	create	VERB
ajst-7358	12	79	as	as	ADP
ajst-7358	12	80	the	the	DET
ajst-7358	12	81	roi	roi	NOUN
ajst-7358	12	82	area	area	NOUN
ajst-7358	12	83	of	of	ADP
ajst-7358	12	84	the	the	DET
ajst-7358	12	85	weld	weld	NOUN
ajst-7358	12	86	,	,	PUNCT
ajst-7358	12	87	thus	thus	ADV
ajst-7358	12	88	further	far	ADV
ajst-7358	12	89	limiting	limit	VERB
ajst-7358	12	90	the	the	DET
ajst-7358	12	91	extraction	extraction	NOUN
ajst-7358	12	92	area	area	NOUN
ajst-7358	12	93	of	of	ADP
ajst-7358	12	94	feature	feature	NOUN
ajst-7358	12	95	points	point	NOUN
ajst-7358	12	96	and	and	CCONJ
ajst-7358	12	97	reducing	reduce	VERB
ajst-7358	12	98	noise	noise	NOUN
ajst-7358	12	99	.	.	PUNCT
ajst-7358	13	1	2	2	X
ajst-7358	13	2	.	.	X
ajst-7358	13	3	object	object	NOUN
ajst-7358	13	4	detection	detection	NOUN
ajst-7358	13	5	method	method	NOUN
ajst-7358	13	6	based	base	VERB
ajst-7358	13	7	on	on	ADP
ajst-7358	13	8	deep	deep	ADJ
ajst-7358	13	9	learning	learning	NOUN
ajst-7358	13	10	there	there	PRON
ajst-7358	13	11	are	be	VERB
ajst-7358	13	12	mainly	mainly	ADV
ajst-7358	13	13	two	two	NUM
ajst-7358	13	14	kinds	kind	NOUN
ajst-7358	13	15	of	of	ADP
ajst-7358	13	16	object	object	NOUN
ajst-7358	13	17	detection	detection	NOUN
ajst-7358	13	18	methods	method	NOUN
ajst-7358	13	19	based	base	VERB
ajst-7358	13	20	on	on	ADP
ajst-7358	13	21	deep	deep	ADJ
ajst-7358	13	22	learning	learning	NOUN
ajst-7358	13	23	:	:	PUNCT
ajst-7358	13	24	one	one	NUM
ajst-7358	13	25	is	be	AUX
ajst-7358	13	26	object	object	NOUN
ajst-7358	13	27	detection	detection	NOUN
ajst-7358	13	28	method	method	NOUN
ajst-7358	13	29	based	base	VERB
ajst-7358	13	30	on	on	ADP
ajst-7358	13	31	regression	regression	NOUN
ajst-7358	13	32	;	;	PUNCT
ajst-7358	13	33	and	and	CCONJ
ajst-7358	13	34	the	the	DET
ajst-7358	13	35	other	other	ADJ
ajst-7358	13	36	is	be	AUX
ajst-7358	13	37	target	target	NOUN
ajst-7358	13	38	detection	detection	NOUN
ajst-7358	13	39	algorithm	algorithm	NOUN
ajst-7358	13	40	based	base	VERB
ajst-7358	13	41	on	on	ADP
ajst-7358	13	42	candidate	candidate	NOUN
ajst-7358	13	43	region	region	NOUN
ajst-7358	13	44	.	.	PUNCT
ajst-7358	14	1	with	with	ADP
ajst-7358	14	2	the	the	DET
ajst-7358	14	3	former	former	ADJ
ajst-7358	14	4	is	be	AUX
ajst-7358	14	5	of	of	ADP
ajst-7358	14	6	excellent	excellent	ADJ
ajst-7358	14	7	performance	performance	NOUN
ajst-7358	14	8	in	in	ADP
ajst-7358	14	9	the	the	DET
ajst-7358	14	10	real	real	ADJ
ajst-7358	14	11	-	-	PUNCT
ajst-7358	14	12	time	time	NOUN
ajst-7358	14	13	detection	detection	NOUN
ajst-7358	14	14	of	of	ADP
ajst-7358	14	15	the	the	DET
ajst-7358	14	16	target	target	NOUN
ajst-7358	14	17	,	,	PUNCT
ajst-7358	14	18	though	though	SCONJ
ajst-7358	14	19	the	the	DET
ajst-7358	14	20	detection	detection	NOUN
ajst-7358	14	21	accuracy	accuracy	NOUN
ajst-7358	14	22	is	be	AUX
ajst-7358	14	23	not	not	PART
ajst-7358	14	24	as	as	ADV
ajst-7358	14	25	good	good	ADJ
ajst-7358	14	26	as	as	ADP
ajst-7358	14	27	the	the	DET
ajst-7358	14	28	latter	latter	ADJ
ajst-7358	14	29	,	,	PUNCT
ajst-7358	14	30	reflected	reflect	VERB
ajst-7358	14	31	in	in	ADP
ajst-7358	14	32	the	the	DET
ajst-7358	14	33	performance	performance	NOUN
ajst-7358	14	34	difference	difference	NOUN
ajst-7358	14	35	especially	especially	ADV
ajst-7358	14	36	for	for	ADP
ajst-7358	14	37	the	the	DET
ajst-7358	14	38	small	small	ADJ
ajst-7358	14	39	target	target	NOUN
ajst-7358	14	40	encountered	encounter	VERB
ajst-7358	14	41	in	in	ADP
ajst-7358	14	42	the	the	DET
ajst-7358	14	43	multi	multi	ADJ
ajst-7358	14	44	-	-	ADJ
ajst-7358	14	45	target	target	ADJ
ajst-7358	14	46	detection	detection	NOUN
ajst-7358	14	47	.	.	PUNCT
ajst-7358	15	1	the	the	DET
ajst-7358	15	2	deep	deep	ADJ
ajst-7358	15	3	learning	learning	NOUN
ajst-7358	15	4	object	object	NOUN
ajst-7358	15	5	detection	detection	NOUN
ajst-7358	15	6	algorithm	algorithm	NOUN
ajst-7358	15	7	based	base	VERB
ajst-7358	15	8	on	on	ADP
ajst-7358	15	9	the	the	DET
ajst-7358	15	10	candidate	candidate	NOUN
ajst-7358	15	11	region	region	NOUN
ajst-7358	15	12	is	be	AUX
ajst-7358	15	13	also	also	ADV
ajst-7358	15	14	known	know	VERB
ajst-7358	15	15	as	as	ADP
ajst-7358	15	16	the	the	DET
ajst-7358	15	17	two	two	NUM
ajst-7358	15	18	-	-	PUNCT
ajst-7358	15	19	step	step	NOUN
ajst-7358	15	20	method	method	NOUN
ajst-7358	15	21	.	.	PUNCT
ajst-7358	16	1	compared	compare	VERB
ajst-7358	16	2	with	with	ADP
ajst-7358	16	3	the	the	DET
ajst-7358	16	4	aforementioned	aforementioned	ADJ
ajst-7358	16	5	deep	deep	ADJ
ajst-7358	16	6	learning	learning	NOUN
ajst-7358	16	7	object	object	NOUN
ajst-7358	16	8	detection	detection	NOUN
ajst-7358	16	9	method	method	NOUN
ajst-7358	16	10	based	base	VERB
ajst-7358	16	11	on	on	ADP
ajst-7358	16	12	regression	regression	NOUN
ajst-7358	16	13	,	,	PUNCT
ajst-7358	16	14	the	the	DET
ajst-7358	16	15	construction	construction	NOUN
ajst-7358	16	16	process	process	NOUN
ajst-7358	16	17	of	of	ADP
ajst-7358	16	18	image	image	NOUN
ajst-7358	16	19	feature	feature	NOUN
ajst-7358	16	20	sample	sample	NOUN
ajst-7358	16	21	box	box	PROPN
ajst-7358	16	22	is	be	AUX
ajst-7358	16	23	increased	increase	VERB
ajst-7358	16	24	.	.	PUNCT
ajst-7358	17	1	therefore	therefore	ADV
ajst-7358	17	2	,	,	PUNCT
ajst-7358	17	3	the	the	DET
ajst-7358	17	4	detection	detection	NOUN
ajst-7358	17	5	accuracy	accuracy	NOUN
ajst-7358	17	6	of	of	ADP
ajst-7358	17	7	this	this	DET
ajst-7358	17	8	method	method	NOUN
ajst-7358	17	9	is	be	AUX
ajst-7358	17	10	obviously	obviously	ADV
ajst-7358	17	11	higher	high	ADJ
ajst-7358	17	12	,	,	PUNCT
ajst-7358	17	13	but	but	CCONJ
ajst-7358	17	14	it	it	PRON
ajst-7358	17	15	also	also	ADV
ajst-7358	17	16	makes	make	VERB
ajst-7358	17	17	the	the	DET
ajst-7358	17	18	algorithm	algorithm	NOUN
ajst-7358	17	19	more	more	ADV
ajst-7358	17	20	complex	complex	ADJ
ajst-7358	17	21	,	,	PUNCT
ajst-7358	17	22	directly	directly	ADV
ajst-7358	17	23	affects	affect	VERB
ajst-7358	17	24	its	its	PRON
ajst-7358	17	25	detection	detection	NOUN
ajst-7358	17	26	efficiency	efficiency	NOUN
ajst-7358	17	27	,	,	PUNCT
ajst-7358	17	28	and	and	CCONJ
ajst-7358	17	29	the	the	DET
ajst-7358	17	30	real	real	ADJ
ajst-7358	17	31	-	-	PUNCT
ajst-7358	17	32	time	time	NOUN
ajst-7358	17	33	performance	performance	NOUN
ajst-7358	17	34	of	of	ADP
ajst-7358	17	35	the	the	DET
ajst-7358	17	36	algorithm	algorithm	NOUN
ajst-7358	17	37	is	be	AUX
ajst-7358	17	38	not	not	PART
ajst-7358	17	39	strong	strong	ADJ
ajst-7358	17	40	.	.	PUNCT
ajst-7358	18	1	at	at	ADP
ajst-7358	18	2	present	present	ADJ
ajst-7358	18	3	,	,	PUNCT
ajst-7358	18	4	typical	typical	ADJ
ajst-7358	18	5	representatives	representative	NOUN
ajst-7358	18	6	of	of	ADP
ajst-7358	18	7	this	this	DET
ajst-7358	18	8	method	method	NOUN
ajst-7358	18	9	include	include	VERB
ajst-7358	18	10	faster	fast	ADJ
ajst-7358	18	11	-	-	PUNCT
ajst-7358	18	12	rcnn	rcnn	NOUN
ajst-7358	18	13	,	,	PUNCT
ajst-7358	18	14	mask	mask	NOUN
ajst-7358	18	15	-	-	PUNCT
ajst-7358	18	16	rcnn	rcnn	PROPN
ajst-7358	18	17	and	and	CCONJ
ajst-7358	18	18	other	other	ADJ
ajst-7358	18	19	algorithms	algorithm	NOUN
ajst-7358	18	20	,	,	PUNCT
ajst-7358	18	21	among	among	ADP
ajst-7358	18	22	which	which	PRON
ajst-7358	18	23	the	the	DET
ajst-7358	18	24	use	use	NOUN
ajst-7358	18	25	of	of	ADP
ajst-7358	18	26	faster	fast	ADJ
ajst-7358	18	27	-	-	PUNCT
ajst-7358	18	28	rcnn	rcnn	NOUN
ajst-7358	18	29	algorithm	algorithm	NOUN
ajst-7358	18	30	to	to	PART
ajst-7358	18	31	solve	solve	VERB
ajst-7358	18	32	practical	practical	ADJ
ajst-7358	18	33	engineering	engineering	NOUN
ajst-7358	18	34	problems	problem	NOUN
ajst-7358	18	35	is	be	AUX
ajst-7358	18	36	a	a	DET
ajst-7358	18	37	hot	hot	ADJ
ajst-7358	18	38	research	research	NOUN
ajst-7358	18	39	and	and	CCONJ
ajst-7358	18	40	application	application	NOUN
ajst-7358	18	41	direction	direction	NOUN
ajst-7358	18	42	in	in	ADP
ajst-7358	18	43	the	the	DET
ajst-7358	18	44	field	field	NOUN
ajst-7358	18	45	of	of	ADP
ajst-7358	18	46	target	target	NOUN
ajst-7358	18	47	detection	detection	NOUN
ajst-7358	18	48	.	.	PUNCT
ajst-7358	19	1	the	the	DET
ajst-7358	19	2	faster	fast	ADJ
ajst-7358	19	3	-	-	PUNCT
ajst-7358	19	4	rcnn	rcnn	NOUN
ajst-7358	19	5	algorithm	algorithm	NOUN
ajst-7358	19	6	extracts	extract	VERB
ajst-7358	19	7	the	the	DET
ajst-7358	19	8	necessary	necessary	ADJ
ajst-7358	19	9	features	feature	NOUN
ajst-7358	19	10	by	by	ADP
ajst-7358	19	11	inputting	inputte	VERB
ajst-7358	19	12	the	the	DET
ajst-7358	19	13	image	image	NOUN
ajst-7358	19	14	into	into	ADP
ajst-7358	19	15	the	the	DET
ajst-7358	19	16	appropriate	appropriate	ADJ
ajst-7358	19	17	cnn	cnn	PROPN
ajst-7358	19	18	network	network	NOUN
ajst-7358	19	19	model	model	PROPN
ajst-7358	19	20	.	.	PUNCT
ajst-7358	20	1	the	the	DET
ajst-7358	20	2	feature	feature	NOUN
ajst-7358	20	3	is	be	AUX
ajst-7358	20	4	input	input	ADJ
ajst-7358	20	5	into	into	ADP
ajst-7358	20	6	rpn	rpn	PROPN
ajst-7358	20	7	network	network	NOUN
ajst-7358	20	8	to	to	PART
ajst-7358	20	9	generate	generate	VERB
ajst-7358	20	10	the	the	DET
ajst-7358	20	11	recommended	recommend	VERB
ajst-7358	20	12	candidate	candidate	NOUN
ajst-7358	20	13	region	region	NOUN
ajst-7358	20	14	,	,	PUNCT
ajst-7358	20	15	and	and	CCONJ
ajst-7358	20	16	the	the	DET
ajst-7358	20	17	feature	feature	NOUN
ajst-7358	20	18	region	region	NOUN
ajst-7358	20	19	extracted	extract	VERB
ajst-7358	20	20	from	from	ADP
ajst-7358	20	21	cnn	cnn	PROPN
ajst-7358	20	22	network	network	NOUN
ajst-7358	20	23	model	model	NOUN
ajst-7358	20	24	is	be	AUX
ajst-7358	20	25	input	input	NOUN
ajst-7358	20	26	into	into	ADP
ajst-7358	20	27	pooling	pool	VERB
ajst-7358	20	28	layer	layer	NOUN
ajst-7358	20	29	to	to	PART
ajst-7358	20	30	generate	generate	VERB
ajst-7358	20	31	the	the	DET
ajst-7358	20	32	feature	feature	NOUN
ajst-7358	20	33	region	region	NOUN
ajst-7358	20	34	with	with	ADP
ajst-7358	20	35	a	a	DET
ajst-7358	20	36	fixed	fix	VERB
ajst-7358	20	37	size	size	NOUN
ajst-7358	20	38	.	.	PUNCT
ajst-7358	21	1	finally	finally	ADV
ajst-7358	21	2	,	,	PUNCT
ajst-7358	21	3	the	the	DET
ajst-7358	21	4	targets	target	NOUN
ajst-7358	21	5	are	be	AUX
ajst-7358	21	6	differentiated	differentiate	VERB
ajst-7358	21	7	and	and	CCONJ
ajst-7358	21	8	detected	detect	VERB
ajst-7358	21	9	at	at	ADP
ajst-7358	21	10	the	the	DET
ajst-7358	21	11	full	full	ADJ
ajst-7358	21	12	connection	connection	NOUN
ajst-7358	21	13	layer	layer	NOUN
ajst-7358	21	14	,	,	PUNCT
ajst-7358	21	15	which	which	PRON
ajst-7358	21	16	is	be	AUX
ajst-7358	21	17	used	use	VERB
ajst-7358	21	18	to	to	PART
ajst-7358	21	19	classify	classify	VERB
ajst-7358	21	20	and	and	CCONJ
ajst-7358	21	21	regression	regression	VERB
ajst-7358	21	22	the	the	DET
ajst-7358	21	23	detected	detect	VERB
ajst-7358	21	24	targets	target	NOUN
ajst-7358	21	25	.	.	PUNCT
ajst-7358	22	1	therefore	therefore	ADV
ajst-7358	22	2	,	,	PUNCT
ajst-7358	22	3	the	the	DET
ajst-7358	22	4	faster	fast	ADJ
ajst-7358	22	5	-	-	PUNCT
ajst-7358	22	6	rcnn	rcnn	NOUN
ajst-7358	22	7	algorithm	algorithm	PROPN
ajst-7358	22	8	has	have	VERB
ajst-7358	22	9	two	two	NUM
ajst-7358	22	10	key	key	ADJ
ajst-7358	22	11	points	point	NOUN
ajst-7358	22	12	:	:	PUNCT
ajst-7358	22	13	cnn	cnn	PROPN
ajst-7358	22	14	network	network	NOUN
ajst-7358	22	15	model	model	PROPN
ajst-7358	22	16	and	and	CCONJ
ajst-7358	22	17	rpn	rpn	PROPN
ajst-7358	22	18	network	network	PROPN
ajst-7358	22	19	.	.	PUNCT
ajst-7358	23	1	cnn	cnn	PROPN
ajst-7358	23	2	network	network	PROPN
ajst-7358	23	3	model	model	PROPN
ajst-7358	23	4	:	:	PUNCT
ajst-7358	23	5	cnn	cnn	PROPN
ajst-7358	23	6	network	network	PROPN
ajst-7358	23	7	model	model	NOUN
ajst-7358	23	8	is	be	AUX
ajst-7358	23	9	one	one	NUM
ajst-7358	23	10	of	of	ADP
ajst-7358	23	11	the	the	DET
ajst-7358	23	12	key	key	ADJ
ajst-7358	23	13	factors	factor	NOUN
ajst-7358	23	14	that	that	PRON
ajst-7358	23	15	determine	determine	VERB
ajst-7358	23	16	the	the	DET
ajst-7358	23	17	operation	operation	NOUN
ajst-7358	23	18	effect	effect	NOUN
ajst-7358	23	19	of	of	ADP
ajst-7358	23	20	deep	deep	ADJ
ajst-7358	23	21	learning	learning	NOUN
ajst-7358	23	22	object	object	NOUN
ajst-7358	23	23	detection	detection	NOUN
ajst-7358	23	24	algorithm	algorithm	NOUN
ajst-7358	23	25	.	.	PUNCT
ajst-7358	24	1	from	from	ADP
ajst-7358	24	2	the	the	DET
ajst-7358	24	3	earliest	early	ADJ
ajst-7358	24	4	lenet5	lenet5	NOUN
ajst-7358	24	5	network	network	NOUN
ajst-7358	24	6	model	model	NOUN
ajst-7358	24	7	to	to	ADP
ajst-7358	24	8	the	the	DET
ajst-7358	24	9	revolutionary	revolutionary	ADJ
ajst-7358	24	10	alexnet	alexnet	ADJ
ajst-7358	24	11	network	network	NOUN
ajst-7358	24	12	model	model	NOUN
ajst-7358	24	13	,	,	PUNCT
ajst-7358	24	14	the	the	DET
ajst-7358	24	15	detection	detection	NOUN
ajst-7358	24	16	accuracy	accuracy	NOUN
ajst-7358	24	17	of	of	ADP
ajst-7358	24	18	cnn	cnn	PROPN
ajst-7358	24	19	network	network	NOUN
ajst-7358	24	20	model	model	NOUN
ajst-7358	24	21	has	have	AUX
ajst-7358	24	22	been	be	AUX
ajst-7358	24	23	continuously	continuously	ADV
ajst-7358	24	24	improved	improve	VERB
ajst-7358	24	25	by	by	ADP
ajst-7358	24	26	deepening	deepen	VERB
ajst-7358	24	27	the	the	DET
ajst-7358	24	28	depth	depth	NOUN
ajst-7358	24	29	and	and	CCONJ
ajst-7358	24	30	complexity	complexity	NOUN
ajst-7358	24	31	of	of	ADP
ajst-7358	24	32	network	network	NOUN
ajst-7358	24	33	model	model	NOUN
ajst-7358	24	34	.	.	PUNCT
ajst-7358	25	1	so	so	SCONJ
ajst-7358	25	2	that	that	SCONJ
ajst-7358	25	3	when	when	SCONJ
ajst-7358	25	4	resnet	resnet	NOUN
ajst-7358	25	5	network	network	NOUN
ajst-7358	25	6	model	model	NOUN
ajst-7358	25	7	was	be	AUX
ajst-7358	25	8	developed	develop	VERB
ajst-7358	25	9	,	,	PUNCT
ajst-7358	25	10	its	its	PRON
ajst-7358	25	11	detection	detection	NOUN
ajst-7358	25	12	accuracy	accuracy	NOUN
ajst-7358	25	13	even	even	ADV
ajst-7358	25	14	exceeded	exceed	VERB
ajst-7358	25	15	the	the	DET
ajst-7358	25	16	level	level	NOUN
ajst-7358	25	17	of	of	ADP
ajst-7358	25	18	human	human	ADJ
ajst-7358	25	19	brain	brain	NOUN
ajst-7358	25	20	cognition	cognition	NOUN
ajst-7358	25	21	,	,	PUNCT
ajst-7358	25	22	reaching	reach	VERB
ajst-7358	25	23	96.43	96.43	NUM
ajst-7358	25	24	%	%	NOUN
ajst-7358	25	25	.	.	PUNCT
ajst-7358	26	1	the	the	DET
ajst-7358	26	2	faster	fast	ADJ
ajst-7358	26	3	-	-	PUNCT
ajst-7358	26	4	rcnn	rcnn	NOUN
ajst-7358	26	5	algorithm	algorithm	NOUN
ajst-7358	26	6	in	in	ADP
ajst-7358	26	7	this	this	DET
ajst-7358	26	8	paper	paper	NOUN
ajst-7358	26	9	is	be	AUX
ajst-7358	26	10	based	base	VERB
ajst-7358	26	11	on	on	ADP
ajst-7358	26	12	alexnet	alexnet	ADJ
ajst-7358	26	13	network	network	NOUN
ajst-7358	26	14	model	model	NOUN
ajst-7358	26	15	,	,	PUNCT
ajst-7358	26	16	which	which	PRON
ajst-7358	26	17	is	be	AUX
ajst-7358	26	18	an	an	DET
ajst-7358	26	19	efficient	efficient	ADJ
ajst-7358	26	20	and	and	CCONJ
ajst-7358	26	21	reliable	reliable	ADJ
ajst-7358	26	22	one	one	NUM
ajst-7358	26	23	.	.	PUNCT
ajst-7358	27	1	after	after	ADP
ajst-7358	27	2	its	its	PRON
ajst-7358	27	3	appearance	appearance	NOUN
ajst-7358	27	4	,	,	PUNCT
ajst-7358	27	5	the	the	DET
ajst-7358	27	6	traditional	traditional	ADJ
ajst-7358	27	7	target	target	NOUN
ajst-7358	27	8	detection	detection	NOUN
ajst-7358	27	9	algorithm	algorithm	NOUN
ajst-7358	27	10	is	be	AUX
ajst-7358	27	11	no	no	ADV
ajst-7358	27	12	longer	long	ADV
ajst-7358	27	13	applicable	applicable	ADJ
ajst-7358	27	14	.	.	PUNCT
ajst-7358	28	1	new	new	ADJ
ajst-7358	28	2	object	object	NOUN
ajst-7358	28	3	detection	detection	NOUN
ajst-7358	28	4	algorithms	algorithm	NOUN
ajst-7358	28	5	with	with	ADP
ajst-7358	28	6	high	high	ADJ
ajst-7358	28	7	efficiency	efficiency	NOUN
ajst-7358	28	8	and	and	CCONJ
ajst-7358	28	9	high	high	ADJ
ajst-7358	28	10	precision	precision	NOUN
ajst-7358	28	11	based	base	VERB
ajst-7358	28	12	on	on	ADP
ajst-7358	28	13	deep	deep	ADJ
ajst-7358	28	14	learning	learning	NOUN
ajst-7358	28	15	and	and	CCONJ
ajst-7358	28	16	42	42	NUM
ajst-7358	28	17	convolution	convolution	NOUN
ajst-7358	28	18	training	training	NOUN
ajst-7358	28	19	by	by	ADP
ajst-7358	28	20	a	a	DET
ajst-7358	28	21	large	large	ADJ
ajst-7358	28	22	number	number	NOUN
ajst-7358	28	23	of	of	ADP
ajst-7358	28	24	images	image	NOUN
ajst-7358	28	25	are	be	AUX
ajst-7358	28	26	emerging	emerge	VERB
ajst-7358	28	27	constantly	constantly	ADV
ajst-7358	28	28	.	.	PUNCT
ajst-7358	29	1	alexnet	alexnet	ADJ
ajst-7358	29	2	network	network	NOUN
ajst-7358	29	3	model	model	NOUN
ajst-7358	29	4	structure	structure	NOUN
ajst-7358	29	5	is	be	AUX
ajst-7358	29	6	as	as	SCONJ
ajst-7358	29	7	follows	follow	VERB
ajst-7358	29	8	:	:	PUNCT
ajst-7358	29	9	figure	figure	NOUN
ajst-7358	29	10	2	2	NUM
ajst-7358	29	11	.	.	PUNCT
ajst-7358	29	12	alexnet	alexnet	ADJ
ajst-7358	29	13	network	network	NOUN
ajst-7358	29	14	model	model	NOUN
ajst-7358	29	15	structure	structure	NOUN
ajst-7358	29	16	it	it	PRON
ajst-7358	29	17	can	can	AUX
ajst-7358	29	18	be	be	AUX
ajst-7358	29	19	seen	see	VERB
ajst-7358	29	20	from	from	ADP
ajst-7358	29	21	the	the	DET
ajst-7358	29	22	figure	figure	NOUN
ajst-7358	29	23	that	that	SCONJ
ajst-7358	29	24	alexnet	alexnet	ADJ
ajst-7358	29	25	network	network	NOUN
ajst-7358	29	26	model	model	NOUN
ajst-7358	29	27	is	be	AUX
ajst-7358	29	28	deeper	deep	ADJ
ajst-7358	29	29	and	and	CCONJ
ajst-7358	29	30	more	more	ADV
ajst-7358	29	31	complex	complex	ADJ
ajst-7358	29	32	than	than	ADP
ajst-7358	29	33	lenet5	lenet5	NOUN
ajst-7358	29	34	network	network	NOUN
ajst-7358	29	35	model	model	NOUN
ajst-7358	29	36	,	,	PUNCT
ajst-7358	29	37	so	so	SCONJ
ajst-7358	29	38	it	it	PRON
ajst-7358	29	39	is	be	AUX
ajst-7358	29	40	more	more	ADV
ajst-7358	29	41	accurate	accurate	ADJ
ajst-7358	29	42	.	.	PUNCT
ajst-7358	30	1	image	image	NOUN
ajst-7358	30	2	feature	feature	NOUN
ajst-7358	30	3	detection	detection	NOUN
ajst-7358	30	4	can	can	AUX
ajst-7358	30	5	be	be	AUX
ajst-7358	30	6	completed	complete	VERB
ajst-7358	30	7	after	after	SCONJ
ajst-7358	30	8	the	the	DET
ajst-7358	30	9	input	input	NOUN
ajst-7358	30	10	image	image	NOUN
ajst-7358	30	11	passes	pass	VERB
ajst-7358	30	12	through	through	ADP
ajst-7358	30	13	5	5	NUM
ajst-7358	30	14	convolution	convolution	NOUN
ajst-7358	30	15	layers	layer	NOUN
ajst-7358	30	16	and	and	CCONJ
ajst-7358	30	17	3	3	NUM
ajst-7358	30	18	fully	fully	ADV
ajst-7358	30	19	connected	connected	ADJ
ajst-7358	30	20	layers	layer	NOUN
ajst-7358	30	21	.	.	PUNCT
ajst-7358	31	1	at	at	ADP
ajst-7358	31	2	the	the	DET
ajst-7358	31	3	same	same	ADJ
ajst-7358	31	4	time	time	NOUN
ajst-7358	31	5	,	,	PUNCT
ajst-7358	31	6	due	due	ADP
ajst-7358	31	7	to	to	ADP
ajst-7358	31	8	the	the	DET
ajst-7358	31	9	improvement	improvement	NOUN
ajst-7358	31	10	of	of	ADP
ajst-7358	31	11	the	the	DET
ajst-7358	31	12	traditional	traditional	ADJ
ajst-7358	31	13	activation	activation	NOUN
ajst-7358	31	14	function	function	NOUN
ajst-7358	31	15	and	and	CCONJ
ajst-7358	31	16	the	the	DET
ajst-7358	31	17	introduction	introduction	NOUN
ajst-7358	31	18	of	of	ADP
ajst-7358	31	19	relu	relu	NOUN
ajst-7358	31	20	activation	activation	NOUN
ajst-7358	31	21	function	function	NOUN
ajst-7358	31	22	,	,	PUNCT
ajst-7358	31	23	the	the	DET
ajst-7358	31	24	convergence	convergence	NOUN
ajst-7358	31	25	and	and	CCONJ
ajst-7358	31	26	implementation	implementation	NOUN
ajst-7358	31	27	speed	speed	NOUN
ajst-7358	31	28	of	of	ADP
ajst-7358	31	29	the	the	DET
ajst-7358	31	30	network	network	NOUN
ajst-7358	31	31	model	model	NOUN
ajst-7358	31	32	have	have	AUX
ajst-7358	31	33	been	be	AUX
ajst-7358	31	34	greatly	greatly	ADV
ajst-7358	31	35	improved	improve	VERB
ajst-7358	31	36	.	.	PUNCT
ajst-7358	32	1	rpn	rpn	PROPN
ajst-7358	32	2	network	network	PROPN
ajst-7358	32	3	(	(	PUNCT
ajst-7358	32	4	regional	regional	ADJ
ajst-7358	32	5	suggestion	suggestion	NOUN
ajst-7358	32	6	network	network	NOUN
ajst-7358	32	7	)	)	PUNCT
ajst-7358	32	8	:	:	PUNCT
ajst-7358	32	9	fasterrcnn	fasterrcnn	PROPN
ajst-7358	32	10	algorithm	algorithm	PROPN
ajst-7358	32	11	can	can	AUX
ajst-7358	32	12	be	be	AUX
ajst-7358	32	13	regarded	regard	VERB
ajst-7358	32	14	as	as	ADP
ajst-7358	32	15	the	the	DET
ajst-7358	32	16	construction	construction	NOUN
ajst-7358	32	17	of	of	ADP
ajst-7358	32	18	rpn	rpn	PROPN
ajst-7358	32	19	network	network	PROPN
ajst-7358	32	20	(	(	PUNCT
ajst-7358	32	21	fig.3	fig.3	PROPN
ajst-7358	32	22	)	)	PUNCT
ajst-7358	32	23	improved	improve	VERB
ajst-7358	32	24	and	and	CCONJ
ajst-7358	32	25	optimized	optimize	VERB
ajst-7358	32	26	on	on	ADP
ajst-7358	32	27	the	the	DET
ajst-7358	32	28	basis	basis	NOUN
ajst-7358	32	29	of	of	ADP
ajst-7358	32	30	fastrcnn	fastrcnn	NOUN
ajst-7358	32	31	algorithm	algorithm	NOUN
ajst-7358	32	32	,	,	PUNCT
ajst-7358	32	33	with	with	SCONJ
ajst-7358	32	34	two	two	NUM
ajst-7358	32	35	main	main	ADJ
ajst-7358	32	36	functions	function	NOUN
ajst-7358	32	37	included	include	VERB
ajst-7358	32	38	:	:	PUNCT
ajst-7358	32	39	training	train	VERB
ajst-7358	32	40	fast	fast	ADJ
ajst-7358	32	41	-	-	PUNCT
ajst-7358	32	42	rcnn	rcnn	NOUN
ajst-7358	32	43	network	network	NOUN
ajst-7358	32	44	and	and	CCONJ
ajst-7358	32	45	optimizing	optimize	VERB
ajst-7358	32	46	configuration	configuration	NOUN
ajst-7358	32	47	of	of	ADP
ajst-7358	32	48	fastrcnn	fastrcnn	NOUN
ajst-7358	32	49	network	network	NOUN
ajst-7358	32	50	by	by	ADP
ajst-7358	32	51	sharing	share	VERB
ajst-7358	32	52	weights	weight	NOUN
ajst-7358	32	53	.	.	PUNCT
ajst-7358	33	1	the	the	DET
ajst-7358	33	2	rpn	rpn	PROPN
ajst-7358	33	3	network	network	PROPN
ajst-7358	33	4	model	model	PROPN
ajst-7358	33	5	inputs	input	VERB
ajst-7358	33	6	the	the	DET
ajst-7358	33	7	features	feature	NOUN
ajst-7358	33	8	processed	process	VERB
ajst-7358	33	9	by	by	ADP
ajst-7358	33	10	cnn	cnn	PROPN
ajst-7358	33	11	model	model	NOUN
ajst-7358	33	12	into	into	ADP
ajst-7358	33	13	sliding	slide	VERB
ajst-7358	33	14	window	window	NOUN
ajst-7358	33	15	and	and	CCONJ
ajst-7358	33	16	performing	perform	VERB
ajst-7358	33	17	convolution	convolution	NOUN
ajst-7358	33	18	to	to	PART
ajst-7358	33	19	obtain	obtain	VERB
ajst-7358	33	20	256dimensional	256dimensional	PROPN
ajst-7358	33	21	feature	feature	NOUN
ajst-7358	33	22	vectors	vector	NOUN
ajst-7358	33	23	,	,	PUNCT
ajst-7358	33	24	which	which	PRON
ajst-7358	33	25	are	be	AUX
ajst-7358	33	26	entered	enter	VERB
ajst-7358	33	27	into	into	ADP
ajst-7358	33	28	the	the	DET
ajst-7358	33	29	full	full	ADJ
ajst-7358	33	30	connection	connection	NOUN
ajst-7358	33	31	layer	layer	NOUN
ajst-7358	33	32	to	to	PART
ajst-7358	33	33	obtain	obtain	VERB
ajst-7358	33	34	scores	score	NOUN
ajst-7358	33	35	(	(	PUNCT
ajst-7358	33	36	target	target	NOUN
ajst-7358	33	37	score	score	NOUN
ajst-7358	33	38	and	and	CCONJ
ajst-7358	33	39	background	background	NOUN
ajst-7358	33	40	score	score	NOUN
ajst-7358	33	41	)	)	PUNCT
ajst-7358	33	42	and	and	CCONJ
ajst-7358	33	43	displacement	displacement	NOUN
ajst-7358	33	44	(	(	PUNCT
ajst-7358	33	45	4	4	NUM
ajst-7358	33	46	offset	offset	NOUN
ajst-7358	33	47	relative	relative	ADJ
ajst-7358	33	48	to	to	ADP
ajst-7358	33	49	the	the	DET
ajst-7358	33	50	original	original	ADJ
ajst-7358	33	51	coordinate	coordinate	NOUN
ajst-7358	33	52	)	)	PUNCT
ajst-7358	33	53	.	.	PUNCT
ajst-7358	34	1	finally	finally	ADV
ajst-7358	34	2	,	,	PUNCT
ajst-7358	34	3	the	the	DET
ajst-7358	34	4	candidate	candidate	NOUN
ajst-7358	34	5	areas	area	NOUN
ajst-7358	34	6	are	be	AUX
ajst-7358	34	7	solved	solve	VERB
ajst-7358	34	8	based	base	VERB
ajst-7358	34	9	on	on	ADP
ajst-7358	34	10	these	these	DET
ajst-7358	34	11	results	result	NOUN
ajst-7358	34	12	.	.	PUNCT
ajst-7358	35	1	figure	figure	VERB
ajst-7358	35	2	3	3	NUM
ajst-7358	35	3	.	.	PUNCT
ajst-7358	36	1	rpn	rpn	PROPN
ajst-7358	36	2	network	network	NOUN
ajst-7358	36	3	model	model	PROPN
ajst-7358	36	4	structure	structure	NOUN
ajst-7358	36	5	3	3	X
ajst-7358	36	6	.	.	PUNCT
ajst-7358	36	7	construct	construct	VERB
ajst-7358	36	8	the	the	DET
ajst-7358	36	9	experimental	experimental	ADJ
ajst-7358	36	10	platform	platform	NOUN
ajst-7358	36	11	considering	consider	VERB
ajst-7358	36	12	that	that	SCONJ
ajst-7358	36	13	the	the	DET
ajst-7358	36	14	research	research	NOUN
ajst-7358	36	15	object	object	NOUN
ajst-7358	36	16	is	be	AUX
ajst-7358	36	17	in	in	ADP
ajst-7358	36	18	an	an	DET
ajst-7358	36	19	absolute	absolute	ADJ
ajst-7358	36	20	static	static	ADJ
ajst-7358	36	21	state	state	NOUN
ajst-7358	36	22	during	during	ADP
ajst-7358	36	23	the	the	DET
ajst-7358	36	24	working	working	NOUN
ajst-7358	36	25	process	process	NOUN
ajst-7358	36	26	,	,	PUNCT
ajst-7358	36	27	the	the	DET
ajst-7358	36	28	requirement	requirement	NOUN
ajst-7358	36	29	for	for	ADP
ajst-7358	36	30	realtime	realtime	ADJ
ajst-7358	36	31	performance	performance	NOUN
ajst-7358	36	32	of	of	ADP
ajst-7358	36	33	the	the	DET
ajst-7358	36	34	system	system	NOUN
ajst-7358	36	35	is	be	AUX
ajst-7358	36	36	not	not	PART
ajst-7358	36	37	high;therefore	high;therefore	NOUN
ajst-7358	36	38	,	,	PUNCT
ajst-7358	36	39	the	the	DET
ajst-7358	36	40	faster	fast	ADJ
ajst-7358	36	41	-	-	PUNCT
ajst-7358	36	42	rcnn	rcnn	NOUN
ajst-7358	36	43	algorithm	algorithm	NOUN
ajst-7358	36	44	,	,	PUNCT
ajst-7358	36	45	an	an	DET
ajst-7358	36	46	approach	approach	NOUN
ajst-7358	36	47	to	to	PART
ajst-7358	36	48	guarantee	guarantee	VERB
ajst-7358	36	49	the	the	DET
ajst-7358	36	50	welding	welding	NOUN
ajst-7358	36	51	accuracy	accuracy	NOUN
ajst-7358	36	52	,	,	PUNCT
ajst-7358	36	53	was	be	AUX
ajst-7358	36	54	selected	select	VERB
ajst-7358	36	55	to	to	PART
ajst-7358	36	56	carry	carry	VERB
ajst-7358	36	57	out	out	ADP
ajst-7358	36	58	the	the	DET
ajst-7358	36	59	verification	verification	NOUN
ajst-7358	36	60	test	test	NOUN
ajst-7358	36	61	based	base	VERB
ajst-7358	36	62	on	on	ADP
ajst-7358	36	63	alexnet	alexnet	ADJ
ajst-7358	36	64	network	network	NOUN
ajst-7358	36	65	model	model	NOUN
ajst-7358	36	66	on	on	ADP
ajst-7358	36	67	matlab2018b	matlab2018b	PROPN
ajst-7358	36	68	platform.the	platform.the	DET
ajst-7358	36	69	system	system	NOUN
ajst-7358	36	70	configuration	configuration	NOUN
ajst-7358	36	71	is	be	AUX
ajst-7358	36	72	as	as	SCONJ
ajst-7358	36	73	follows	follow	VERB
ajst-7358	36	74	:	:	PUNCT
ajst-7358	36	75	table	table	NOUN
ajst-7358	36	76	1	1	NUM
ajst-7358	36	77	.	.	PUNCT
ajst-7358	36	78	system	system	NOUN
ajst-7358	36	79	configuration	configuration	NOUN
ajst-7358	36	80	software	software	NOUN
ajst-7358	36	81	matlab2018b	matlab2018b	NOUN
ajst-7358	36	82	system	system	NOUN
ajst-7358	36	83	win10	win10	PROPN
ajst-7358	36	84	(	(	PUNCT
ajst-7358	36	85	64	64	NUM
ajst-7358	36	86	bit	bit	NOUN
ajst-7358	36	87	)	)	PUNCT
ajst-7358	36	88	cpu	cpu	VERB
ajst-7358	36	89	i5	i5	NOUN
ajst-7358	36	90	-	-	PUNCT
ajst-7358	36	91	6500	6500	NUM
ajst-7358	36	92	ram	ram	NOUN
ajst-7358	36	93	16	16	NUM
ajst-7358	36	94	g	g	NOUN
ajst-7358	36	95	(	(	PUNCT
ajst-7358	36	96	ddr4	ddr4	PROPN
ajst-7358	36	97	)	)	PUNCT
ajst-7358	37	1	gpu	gpu	PROPN
ajst-7358	37	2	gtx	gtx	ADJ
ajst-7358	37	3	950	950	NUM
ajst-7358	37	4	(	(	PUNCT
ajst-7358	37	5	2	2	NUM
ajst-7358	37	6	g	g	NOUN
ajst-7358	37	7	)	)	PUNCT
ajst-7358	37	8	nvidia	nvidia	PROPN
ajst-7358	37	9	gpu	gpu	PROPN
ajst-7358	37	10	computing	computing	NOUN
ajst-7358	37	11	toolkit	toolkit	NOUN
ajst-7358	37	12	cuda	cuda	NOUN
ajst-7358	37	13	10.2	10.2	NUM
ajst-7358	37	14	4	4	NUM
ajst-7358	37	15	.	.	PUNCT
ajst-7358	38	1	experimental	experimental	ADJ
ajst-7358	38	2	data	datum	NOUN
ajst-7358	38	3	set	set	VERB
ajst-7358	38	4	construction	construction	NOUN
ajst-7358	38	5	in	in	ADP
ajst-7358	38	6	the	the	DET
ajst-7358	38	7	actual	actual	ADJ
ajst-7358	38	8	welding	welding	NOUN
ajst-7358	38	9	operation	operation	NOUN
ajst-7358	38	10	environment	environment	NOUN
ajst-7358	38	11	(	(	PUNCT
ajst-7358	38	12	take	take	VERB
ajst-7358	38	13	automobile	automobile	NOUN
ajst-7358	38	14	welding	welding	NOUN
ajst-7358	38	15	line	line	NOUN
ajst-7358	38	16	as	as	ADP
ajst-7358	38	17	an	an	DET
ajst-7358	38	18	example	example	NOUN
ajst-7358	38	19	)	)	PUNCT
ajst-7358	38	20	,	,	PUNCT
ajst-7358	38	21	in	in	ADP
ajst-7358	38	22	order	order	NOUN
ajst-7358	38	23	to	to	PART
ajst-7358	38	24	ensure	ensure	VERB
ajst-7358	38	25	production	production	NOUN
ajst-7358	38	26	efficiency	efficiency	NOUN
ajst-7358	38	27	,	,	PUNCT
ajst-7358	38	28	the	the	DET
ajst-7358	38	29	working	work	VERB
ajst-7358	38	30	state	state	NOUN
ajst-7358	38	31	of	of	ADP
ajst-7358	38	32	each	each	DET
ajst-7358	38	33	welding	welding	NOUN
ajst-7358	38	34	robot	robot	NOUN
ajst-7358	38	35	unit	unit	NOUN
ajst-7358	38	36	is	be	AUX
ajst-7358	38	37	only	only	ADV
ajst-7358	38	38	to	to	PART
ajst-7358	38	39	weld	weld	VERB
ajst-7358	38	40	several	several	ADJ
ajst-7358	38	41	specific	specific	ADJ
ajst-7358	38	42	welding	welding	NOUN
ajst-7358	38	43	spots	spot	NOUN
ajst-7358	38	44	or	or	CCONJ
ajst-7358	38	45	welds	weld	NOUN
ajst-7358	38	46	,	,	PUNCT
ajst-7358	38	47	with	with	ADP
ajst-7358	38	48	single	single	ADJ
ajst-7358	38	49	welding	welding	NOUN
ajst-7358	38	50	type	type	NOUN
ajst-7358	38	51	and	and	CCONJ
ajst-7358	38	52	high	high	ADJ
ajst-7358	38	53	repetition	repetition	NOUN
ajst-7358	38	54	.	.	PUNCT
ajst-7358	39	1	it	it	PRON
ajst-7358	39	2	can	can	AUX
ajst-7358	39	3	be	be	AUX
ajst-7358	39	4	seen	see	VERB
ajst-7358	39	5	that	that	SCONJ
ajst-7358	39	6	in	in	ADP
ajst-7358	39	7	the	the	DET
ajst-7358	39	8	actual	actual	ADJ
ajst-7358	39	9	deployment	deployment	NOUN
ajst-7358	39	10	process	process	NOUN
ajst-7358	39	11	of	of	ADP
ajst-7358	39	12	welding	welding	NOUN
ajst-7358	39	13	robot	robot	NOUN
ajst-7358	39	14	,	,	PUNCT
ajst-7358	39	15	it	it	PRON
ajst-7358	39	16	is	be	AUX
ajst-7358	39	17	necessary	necessary	ADJ
ajst-7358	39	18	to	to	PART
ajst-7358	39	19	train	train	VERB
ajst-7358	39	20	the	the	DET
ajst-7358	39	21	welding	welding	NOUN
ajst-7358	39	22	roi	roi	NOUN
ajst-7358	39	23	region	region	NOUN
ajst-7358	39	24	detection	detection	NOUN
ajst-7358	39	25	system	system	NOUN
ajst-7358	39	26	.	.	PUNCT
ajst-7358	40	1	the	the	DET
ajst-7358	40	2	system	system	NOUN
ajst-7358	40	3	only	only	ADV
ajst-7358	40	4	needs	need	VERB
ajst-7358	40	5	to	to	PART
ajst-7358	40	6	realize	realize	VERB
ajst-7358	40	7	the	the	DET
ajst-7358	40	8	construction	construction	NOUN
ajst-7358	40	9	of	of	ADP
ajst-7358	40	10	training	training	NOUN
ajst-7358	40	11	data	datum	NOUN
ajst-7358	40	12	by	by	ADP
ajst-7358	40	13	changing	change	VERB
ajst-7358	40	14	the	the	DET
ajst-7358	40	15	camera	camera	NOUN
ajst-7358	40	16	shooting	shooting	NOUN
ajst-7358	40	17	angle	angle	NOUN
ajst-7358	40	18	and	and	CCONJ
ajst-7358	40	19	position	position	NOUN
ajst-7358	40	20	according	accord	VERB
ajst-7358	40	21	to	to	ADP
ajst-7358	40	22	the	the	DET
ajst-7358	40	23	actual	actual	ADJ
ajst-7358	40	24	conditions	condition	NOUN
ajst-7358	40	25	of	of	ADP
ajst-7358	40	26	welding	welding	NOUN
ajst-7358	40	27	parts	part	NOUN
ajst-7358	40	28	.	.	PUNCT
ajst-7358	41	1	the	the	DET
ajst-7358	41	2	advantage	advantage	NOUN
ajst-7358	41	3	of	of	ADP
ajst-7358	41	4	this	this	PRON
ajst-7358	41	5	is	be	AUX
ajst-7358	41	6	that	that	PRON
ajst-7358	41	7	high	high	ADJ
ajst-7358	41	8	detection	detection	NOUN
ajst-7358	41	9	rates	rate	NOUN
ajst-7358	41	10	can	can	AUX
ajst-7358	41	11	be	be	AUX
ajst-7358	41	12	achieved	achieve	VERB
ajst-7358	41	13	with	with	ADP
ajst-7358	41	14	a	a	DET
ajst-7358	41	15	small	small	ADJ
ajst-7358	41	16	amount	amount	NOUN
ajst-7358	41	17	of	of	ADP
ajst-7358	41	18	training	training	NOUN
ajst-7358	41	19	data	datum	NOUN
ajst-7358	41	20	.	.	PUNCT
ajst-7358	42	1	in	in	ADP
ajst-7358	42	2	short	short	ADJ
ajst-7358	42	3	,	,	PUNCT
ajst-7358	42	4	there	there	PRON
ajst-7358	42	5	are	be	VERB
ajst-7358	42	6	two	two	NUM
ajst-7358	42	7	main	main	ADJ
ajst-7358	42	8	reasons	reason	NOUN
ajst-7358	42	9	for	for	ADP
ajst-7358	42	10	this	this	PRON
ajst-7358	42	11	:	:	PUNCT
ajst-7358	42	12	(	(	PUNCT
ajst-7358	42	13	1	1	X
ajst-7358	42	14	)	)	PUNCT
ajst-7358	42	15	the	the	DET
ajst-7358	42	16	operating	operate	VERB
ajst-7358	42	17	environment	environment	NOUN
ajst-7358	42	18	of	of	ADP
ajst-7358	42	19	welding	welding	NOUN
ajst-7358	42	20	robot	robot	NOUN
ajst-7358	42	21	is	be	AUX
ajst-7358	42	22	stable	stable	ADJ
ajst-7358	42	23	and	and	CCONJ
ajst-7358	42	24	does	do	AUX
ajst-7358	42	25	not	not	PART
ajst-7358	42	26	involve	involve	VERB
ajst-7358	42	27	the	the	DET
ajst-7358	42	28	problem	problem	NOUN
ajst-7358	42	29	of	of	ADP
ajst-7358	42	30	target	target	NOUN
ajst-7358	42	31	detection	detection	NOUN
ajst-7358	42	32	in	in	ADP
ajst-7358	42	33	different	different	ADJ
ajst-7358	42	34	scenes	scene	NOUN
ajst-7358	42	35	.	.	PUNCT
ajst-7358	43	1	(	(	PUNCT
ajst-7358	43	2	2	2	X
ajst-7358	43	3	)	)	PUNCT
ajst-7358	43	4	detection	detection	NOUN
ajst-7358	43	5	image	image	NOUN
ajst-7358	43	6	and	and	CCONJ
ajst-7358	43	7	training	training	NOUN
ajst-7358	43	8	image	image	NOUN
ajst-7358	43	9	acquisition	acquisition	NOUN
ajst-7358	43	10	objects	object	VERB
ajst-7358	43	11	43	43	NUM
ajst-7358	43	12	are	be	AUX
ajst-7358	43	13	consistent	consistent	ADJ
ajst-7358	43	14	,	,	PUNCT
ajst-7358	43	15	which	which	PRON
ajst-7358	43	16	does	do	AUX
ajst-7358	43	17	not	not	PART
ajst-7358	43	18	involve	involve	VERB
ajst-7358	43	19	detection	detection	NOUN
ajst-7358	43	20	object	object	VERB
ajst-7358	43	21	classification	classification	NOUN
ajst-7358	43	22	.	.	PUNCT
ajst-7358	44	1	therefore	therefore	ADV
ajst-7358	44	2	,	,	PUNCT
ajst-7358	44	3	the	the	DET
ajst-7358	44	4	research	research	NOUN
ajst-7358	44	5	object	object	NOUN
ajst-7358	44	6	in	in	ADP
ajst-7358	44	7	this	this	DET
ajst-7358	44	8	paper	paper	NOUN
ajst-7358	44	9	is	be	AUX
ajst-7358	44	10	simpler	simple	ADJ
ajst-7358	44	11	and	and	CCONJ
ajst-7358	44	12	requires	require	VERB
ajst-7358	44	13	fewer	few	ADJ
ajst-7358	44	14	training	training	NOUN
ajst-7358	44	15	data	datum	NOUN
ajst-7358	44	16	set	set	VERB
ajst-7358	44	17	pictures	picture	NOUN
ajst-7358	44	18	.	.	PUNCT
ajst-7358	45	1	in	in	ADP
ajst-7358	45	2	this	this	DET
ajst-7358	45	3	paper	paper	NOUN
ajst-7358	45	4	,	,	PUNCT
ajst-7358	45	5	the	the	DET
ajst-7358	45	6	types	type	NOUN
ajst-7358	45	7	of	of	ADP
ajst-7358	45	8	welds	weld	NOUN
ajst-7358	45	9	are	be	AUX
ajst-7358	45	10	divided	divide	VERB
ajst-7358	45	11	into	into	ADP
ajst-7358	45	12	3	3	NUM
ajst-7358	45	13	groups	group	NOUN
ajst-7358	45	14	(	(	PUNCT
ajst-7358	45	15	straight	straight	ADJ
ajst-7358	45	16	line	line	NOUN
ajst-7358	45	17	,	,	PUNCT
ajst-7358	45	18	arc	arc	NOUN
ajst-7358	45	19	and	and	CCONJ
ajst-7358	45	20	composite	composite	ADJ
ajst-7358	45	21	weld	weld	NOUN
ajst-7358	45	22	are	be	AUX
ajst-7358	45	23	simulated	simulate	VERB
ajst-7358	45	24	respectively	respectively	ADV
ajst-7358	45	25	)	)	PUNCT
ajst-7358	45	26	,	,	PUNCT
ajst-7358	45	27	with	with	ADP
ajst-7358	45	28	10	10	NUM
ajst-7358	45	29	pieces	piece	NOUN
ajst-7358	45	30	in	in	ADP
ajst-7358	45	31	each	each	DET
ajst-7358	45	32	group	group	NOUN
ajst-7358	45	33	and	and	CCONJ
ajst-7358	45	34	a	a	DET
ajst-7358	45	35	total	total	NOUN
ajst-7358	45	36	of	of	ADP
ajst-7358	45	37	30	30	NUM
ajst-7358	45	38	pieces	piece	NOUN
ajst-7358	45	39	for	for	ADP
ajst-7358	45	40	training(fig.4	training(fig.4	PROPN
ajst-7358	45	41	a	a	PRON
ajst-7358	45	42	)	)	PUNCT
ajst-7358	45	43	.	.	PUNCT
ajst-7358	46	1	there	there	PRON
ajst-7358	46	2	are	be	VERB
ajst-7358	46	3	40	40	NUM
ajst-7358	46	4	images	image	NOUN
ajst-7358	46	5	in	in	ADP
ajst-7358	46	6	each	each	DET
ajst-7358	46	7	test	test	NOUN
ajst-7358	46	8	set	set	VERB
ajst-7358	46	9	and	and	CCONJ
ajst-7358	46	10	120	120	NUM
ajst-7358	46	11	images	image	NOUN
ajst-7358	46	12	in	in	ADP
ajst-7358	46	13	total	total	NOUN
ajst-7358	46	14	for	for	ADP
ajst-7358	46	15	testing	testing	NOUN
ajst-7358	46	16	.	.	PUNCT
ajst-7358	47	1	the	the	DET
ajst-7358	47	2	training	training	NOUN
ajst-7358	47	3	set	set	NOUN
ajst-7358	47	4	and	and	CCONJ
ajst-7358	47	5	part	part	NOUN
ajst-7358	47	6	of	of	ADP
ajst-7358	47	7	the	the	DET
ajst-7358	47	8	test	test	NOUN
ajst-7358	47	9	set	set	NOUN
ajst-7358	47	10	(	(	PUNCT
ajst-7358	47	11	fig.4	fig.4	PROPN
ajst-7358	47	12	b	b	NOUN
ajst-7358	47	13	)	)	PUNCT
ajst-7358	47	14	are	be	AUX
ajst-7358	47	15	enumerated	enumerate	VERB
ajst-7358	47	16	:	:	PUNCT
ajst-7358	47	17	a	a	DET
ajst-7358	47	18	b	b	NUM
ajst-7358	47	19	figure	figure	NOUN
ajst-7358	47	20	4	4	NUM
ajst-7358	47	21	.	.	PUNCT
ajst-7358	47	22	data	datum	NOUN
ajst-7358	47	23	set	set	VERB
ajst-7358	47	24	enumeration	enumeration	NOUN
ajst-7358	47	25	5	5	NUM
ajst-7358	47	26	.	.	PUNCT
ajst-7358	47	27	faster	fast	ADJ
ajst-7358	47	28	-	-	PUNCT
ajst-7358	47	29	rcnn	rcnn	NOUN
ajst-7358	47	30	data	datum	NOUN
ajst-7358	47	31	training	train	VERB
ajst-7358	47	32	the	the	DET
ajst-7358	47	33	completed	complete	VERB
ajst-7358	47	34	training	training	NOUN
ajst-7358	47	35	set	set	VERB
ajst-7358	47	36	image	image	NOUN
ajst-7358	47	37	collection	collection	NOUN
ajst-7358	47	38	is	be	AUX
ajst-7358	47	39	followed	follow	VERB
ajst-7358	47	40	by	by	ADP
ajst-7358	47	41	the	the	DET
ajst-7358	47	42	roi	roi	NOUN
ajst-7358	47	43	labeling	labeling	NOUN
ajst-7358	47	44	on	on	ADP
ajst-7358	47	45	the	the	DET
ajst-7358	47	46	welding	welding	NOUN
ajst-7358	47	47	target	target	NOUN
ajst-7358	47	48	area	area	NOUN
ajst-7358	47	49	individually	individually	ADV
ajst-7358	47	50	,	,	PUNCT
ajst-7358	47	51	and	and	CCONJ
ajst-7358	47	52	then	then	ADV
ajst-7358	47	53	save	save	VERB
ajst-7358	47	54	these	these	DET
ajst-7358	47	55	roi	roi	NOUN
ajst-7358	47	56	annotation	annotation	NOUN
ajst-7358	47	57	areas	area	NOUN
ajst-7358	47	58	to	to	ADP
ajst-7358	47	59	the	the	DET
ajst-7358	47	60	work	work	NOUN
ajst-7358	47	61	area	area	NOUN
ajst-7358	47	62	after	after	SCONJ
ajst-7358	47	63	the	the	DET
ajst-7358	47	64	annotation	annotation	NOUN
ajst-7358	47	65	is	be	AUX
ajst-7358	47	66	complete	complete	ADJ
ajst-7358	47	67	.	.	PUNCT
ajst-7358	48	1	after	after	SCONJ
ajst-7358	48	2	image	image	NOUN
ajst-7358	48	3	preprocessing	preprocessing	NOUN
ajst-7358	48	4	is	be	AUX
ajst-7358	48	5	completed	complete	VERB
ajst-7358	48	6	,	,	PUNCT
ajst-7358	48	7	training	training	NOUN
ajst-7358	48	8	parameters	parameter	NOUN
ajst-7358	48	9	of	of	ADP
ajst-7358	48	10	the	the	DET
ajst-7358	48	11	faster	fast	ADJ
ajst-7358	48	12	-	-	PUNCT
ajst-7358	48	13	rcnn	rcnn	NOUN
ajst-7358	48	14	algorithm	algorithm	PROPN
ajst-7358	48	15	need	need	VERB
ajst-7358	48	16	to	to	PART
ajst-7358	48	17	be	be	AUX
ajst-7358	48	18	preset	preset	ADJ
ajst-7358	48	19	.	.	PUNCT
ajst-7358	49	1	the	the	DET
ajst-7358	49	2	main	main	ADJ
ajst-7358	49	3	parameters	parameter	NOUN
ajst-7358	49	4	are	be	AUX
ajst-7358	49	5	shown	show	VERB
ajst-7358	49	6	in	in	ADP
ajst-7358	49	7	the	the	DET
ajst-7358	49	8	following	follow	VERB
ajst-7358	49	9	table	table	NOUN
ajst-7358	49	10	:	:	PUNCT
ajst-7358	49	11	table	table	NOUN
ajst-7358	49	12	2	2	NUM
ajst-7358	49	13	.	.	PUNCT
ajst-7358	49	14	preset	preset	ADJ
ajst-7358	49	15	parameters	parameter	NOUN
ajst-7358	49	16	of	of	ADP
ajst-7358	49	17	faster	fast	ADJ
ajst-7358	49	18	-	-	PUNCT
ajst-7358	49	19	rcnn	rcnn	NOUN
ajst-7358	49	20	algorithm	algorithm	PROPN
ajst-7358	49	21	minibatchsize	minibatchsize	VERB
ajst-7358	49	22	1	1	NUM
ajst-7358	49	23	maxepochs	maxepoch	NOUN
ajst-7358	49	24	20	20	NUM
ajst-7358	49	25	initiallearnrate	initiallearnrate	ADJ
ajst-7358	49	26	1e-4	1e-4	NUM
ajst-7358	49	27	solvername	solvername	ADJ
ajst-7358	49	28	sgdm	sgdm	NOUN
ajst-7358	49	29	learnrateschedule	learnrateschedule	NOUN
ajst-7358	49	30	piecewise	piecewise	NOUN
ajst-7358	49	31	6	6	NUM
ajst-7358	49	32	.	.	PUNCT
ajst-7358	49	33	experimental	experimental	ADJ
ajst-7358	49	34	result	result	NOUN
ajst-7358	49	35	further	further	ADJ
ajst-7358	49	36	test	test	NOUN
ajst-7358	49	37	set	set	NOUN
ajst-7358	49	38	is	be	AUX
ajst-7358	49	39	used	use	VERB
ajst-7358	49	40	to	to	PART
ajst-7358	49	41	test	test	VERB
ajst-7358	49	42	the	the	DET
ajst-7358	49	43	effectiveness	effectiveness	NOUN
ajst-7358	49	44	of	of	ADP
ajst-7358	49	45	this	this	DET
ajst-7358	49	46	method	method	NOUN
ajst-7358	49	47	for	for	ADP
ajst-7358	49	48	welding	weld	VERB
ajst-7358	49	49	targets	target	NOUN
ajst-7358	49	50	in	in	ADP
ajst-7358	49	51	the	the	DET
ajst-7358	49	52	completion	completion	NOUN
ajst-7358	49	53	training	training	NOUN
ajst-7358	49	54	of	of	ADP
ajst-7358	49	55	the	the	DET
ajst-7358	49	56	faster	fast	ADJ
ajst-7358	49	57	-	-	PUNCT
ajst-7358	49	58	rcnn	rcnn	NOUN
ajst-7358	49	59	algorithm	algorithm	PROPN
ajst-7358	49	60	model	model	PROPN
ajst-7358	50	1	.the	.the	DET
ajst-7358	50	2	current	current	ADJ
ajst-7358	50	3	evaluation	evaluation	NOUN
ajst-7358	50	4	of	of	ADP
ajst-7358	50	5	the	the	DET
ajst-7358	50	6	algorithm	algorithm	NOUN
ajst-7358	50	7	's	's	PART
ajst-7358	50	8	effectiveness	effectiveness	NOUN
ajst-7358	50	9	is	be	AUX
ajst-7358	50	10	mainly	mainly	ADV
ajst-7358	50	11	based	base	VERB
ajst-7358	50	12	on	on	ADP
ajst-7358	50	13	some	some	DET
ajst-7358	50	14	indicators	indicator	NOUN
ajst-7358	50	15	,	,	PUNCT
ajst-7358	50	16	including	include	VERB
ajst-7358	50	17	two	two	NUM
ajst-7358	50	18	indicators	indicator	NOUN
ajst-7358	50	19	:	:	PUNCT
ajst-7358	50	20	recall	recall	NOUN
ajst-7358	50	21	rate	rate	NOUN
ajst-7358	50	22	and	and	CCONJ
ajst-7358	50	23	accuracy	accuracy	NOUN
ajst-7358	50	24	rate	rate	NOUN
ajst-7358	50	25	;	;	PUNCT
ajst-7358	50	26	recall	recall	NOUN
ajst-7358	50	27	rate	rate	NOUN
ajst-7358	50	28	refers	refer	VERB
ajst-7358	50	29	to	to	ADP
ajst-7358	50	30	the	the	DET
ajst-7358	50	31	ratio	ratio	NOUN
ajst-7358	50	32	of	of	ADP
ajst-7358	50	33	the	the	DET
ajst-7358	50	34	number	number	NOUN
ajst-7358	50	35	of	of	ADP
ajst-7358	50	36	images	image	NOUN
ajst-7358	50	37	with	with	ADP
ajst-7358	50	38	welding	welding	NOUN
ajst-7358	50	39	target	target	NOUN
ajst-7358	50	40	detected	detect	VERB
ajst-7358	50	41	to	to	ADP
ajst-7358	50	42	the	the	DET
ajst-7358	50	43	number	number	NOUN
ajst-7358	50	44	of	of	ADP
ajst-7358	50	45	samples	sample	NOUN
ajst-7358	50	46	in	in	ADP
ajst-7358	50	47	the	the	DET
ajst-7358	50	48	test	test	NOUN
ajst-7358	50	49	set	set	NOUN
ajst-7358	50	50	images	image	NOUN
ajst-7358	50	51	.	.	PUNCT
ajst-7358	51	1	accuracy	accuracy	NOUN
ajst-7358	51	2	refers	refer	VERB
ajst-7358	51	3	to	to	ADP
ajst-7358	51	4	the	the	DET
ajst-7358	51	5	ratio	ratio	NOUN
ajst-7358	51	6	between	between	ADP
ajst-7358	51	7	the	the	DET
ajst-7358	51	8	number	number	NOUN
ajst-7358	51	9	of	of	ADP
ajst-7358	51	10	correct	correct	ADJ
ajst-7358	51	11	target	target	NOUN
ajst-7358	51	12	images	image	NOUN
ajst-7358	51	13	and	and	CCONJ
ajst-7358	51	14	the	the	DET
ajst-7358	51	15	total	total	ADJ
ajst-7358	51	16	number	number	NOUN
ajst-7358	51	17	of	of	ADP
ajst-7358	51	18	detected	detect	VERB
ajst-7358	51	19	target	target	NOUN
ajst-7358	51	20	images	image	NOUN
ajst-7358	51	21	.	.	PUNCT
ajst-7358	52	1	figure	figure	NOUN
ajst-7358	52	2	5	5	NUM
ajst-7358	52	3	lists	list	NOUN
ajst-7358	52	4	two	two	NUM
ajst-7358	52	5	typical	typical	ADJ
ajst-7358	52	6	detection	detection	NOUN
ajst-7358	52	7	defects	defect	NOUN
ajst-7358	52	8	that	that	PRON
ajst-7358	52	9	affect	affect	VERB
ajst-7358	52	10	accuracy	accuracy	NOUN
ajst-7358	52	11	:	:	PUNCT
ajst-7358	52	12	a.	a.	NOUN
ajst-7358	52	13	error	error	NOUN
ajst-7358	52	14	areas	area	NOUN
ajst-7358	52	15	are	be	AUX
ajst-7358	52	16	detected	detect	VERB
ajst-7358	52	17	;	;	PUNCT
ajst-7358	52	18	b.	b.	PROPN
ajst-7358	52	19	the	the	DET
ajst-7358	52	20	detection	detection	NOUN
ajst-7358	52	21	position	position	NOUN
ajst-7358	52	22	deviates	deviate	VERB
ajst-7358	52	23	from	from	ADP
ajst-7358	52	24	the	the	DET
ajst-7358	52	25	target	target	NOUN
ajst-7358	52	26	area	area	NOUN
ajst-7358	52	27	.	.	PUNCT
ajst-7358	53	1	a	a	DET
ajst-7358	53	2	b	b	NUM
ajst-7358	53	3	figure	figure	NOUN
ajst-7358	53	4	5	5	NUM
ajst-7358	53	5	.	.	PUNCT
ajst-7358	53	6	list	list	NOUN
ajst-7358	53	7	of	of	ADP
ajst-7358	53	8	detection	detection	NOUN
ajst-7358	53	9	defects	defect	NOUN
ajst-7358	53	10	that	that	PRON
ajst-7358	53	11	affect	affect	VERB
ajst-7358	53	12	detection	detection	NOUN
ajst-7358	53	13	accuracy	accuracy	NOUN
ajst-7358	53	14	table	table	NOUN
ajst-7358	53	15	3	3	NUM
ajst-7358	53	16	.	.	PUNCT
ajst-7358	53	17	faster	fast	ADJ
ajst-7358	53	18	-	-	PUNCT
ajst-7358	53	19	rcnn	rcnn	NOUN
ajst-7358	53	20	detection	detection	NOUN
ajst-7358	53	21	effect	effect	NOUN
ajst-7358	53	22	test	test	NOUN
ajst-7358	53	23	and	and	CCONJ
ajst-7358	53	24	analysis	analysis	NOUN
ajst-7358	53	25	results	result	NOUN
ajst-7358	53	26	recall	recall	VERB
ajst-7358	53	27	rate	rate	NOUN
ajst-7358	53	28	accuracy	accuracy	NOUN
ajst-7358	53	29	rate	rate	NOUN
ajst-7358	53	30	straight	straight	ADJ
ajst-7358	53	31	weld	weld	NOUN
ajst-7358	53	32	100	100	NUM
ajst-7358	53	33	%	%	NOUN
ajst-7358	53	34	100	100	NUM
ajst-7358	53	35	%	%	NOUN
ajst-7358	53	36	arc	arc	NOUN
ajst-7358	53	37	weld	weld	VERB
ajst-7358	53	38	100	100	NUM
ajst-7358	53	39	%	%	NOUN
ajst-7358	53	40	65	65	NUM
ajst-7358	53	41	%	%	NOUN
ajst-7358	53	42	composite	composite	ADJ
ajst-7358	53	43	weld	weld	NOUN
ajst-7358	53	44	100	100	NUM
ajst-7358	53	45	%	%	NOUN
ajst-7358	53	46	78	78	NUM
ajst-7358	53	47	%	%	NOUN
ajst-7358	53	48	mean	mean	NOUN
ajst-7358	53	49	value	value	NOUN
ajst-7358	53	50	100	100	NUM
ajst-7358	53	51	%	%	NOUN
ajst-7358	53	52	81	81	NUM
ajst-7358	53	53	%	%	NOUN
ajst-7358	53	54	it	it	PRON
ajst-7358	53	55	can	can	AUX
ajst-7358	53	56	be	be	AUX
ajst-7358	53	57	seen	see	VERB
ajst-7358	53	58	from	from	ADP
ajst-7358	53	59	table	table	NOUN
ajst-7358	53	60	3	3	NUM
ajst-7358	53	61	that	that	SCONJ
ajst-7358	53	62	the	the	DET
ajst-7358	53	63	average	average	ADJ
ajst-7358	53	64	recall	recall	NOUN
ajst-7358	53	65	rate	rate	NOUN
ajst-7358	53	66	of	of	ADP
ajst-7358	53	67	the	the	DET
ajst-7358	53	68	research	research	NOUN
ajst-7358	53	69	object	object	NOUN
ajst-7358	53	70	mentioned	mention	VERB
ajst-7358	53	71	in	in	ADP
ajst-7358	53	72	this	this	DET
ajst-7358	53	73	paper	paper	NOUN
ajst-7358	53	74	is	be	AUX
ajst-7358	53	75	100	100	NUM
ajst-7358	53	76	%	%	NOUN
ajst-7358	53	77	,	,	PUNCT
ajst-7358	53	78	which	which	PRON
ajst-7358	53	79	results	result	VERB
ajst-7358	53	80	in	in	ADP
ajst-7358	53	81	the	the	DET
ajst-7358	53	82	realization	realization	NOUN
ajst-7358	53	83	of	of	ADP
ajst-7358	53	84	the	the	DET
ajst-7358	53	85	100	100	NUM
ajst-7358	53	86	%	%	NOUN
ajst-7358	53	87	detection	detection	NOUN
ajst-7358	53	88	ration	ration	NOUN
ajst-7358	53	89	of	of	ADP
ajst-7358	53	90	the	the	DET
ajst-7358	53	91	roi	roi	NOUN
ajst-7358	53	92	area	area	NOUN
ajst-7358	53	93	of	of	ADP
ajst-7358	53	94	the	the	DET
ajst-7358	53	95	weld	weld	NOUN
ajst-7358	53	96	in	in	ADP
ajst-7358	53	97	the	the	DET
ajst-7358	53	98	detection	detection	NOUN
ajst-7358	53	99	set	set	VERB
ajst-7358	53	100	constructed	construct	VERB
ajst-7358	53	101	in	in	ADP
ajst-7358	53	102	this	this	DET
ajst-7358	53	103	paper	paper	NOUN
ajst-7358	53	104	.	.	PUNCT
ajst-7358	54	1	however	however	ADV
ajst-7358	54	2	,	,	PUNCT
ajst-7358	54	3	the	the	DET
ajst-7358	54	4	accuracy	accuracy	NOUN
ajst-7358	54	5	of	of	ADP
ajst-7358	54	6	weld	weld	NOUN
ajst-7358	54	7	roi	roi	NOUN
ajst-7358	54	8	region	region	NOUN
ajst-7358	54	9	detection	detection	NOUN
ajst-7358	54	10	was	be	AUX
ajst-7358	54	11	low	low	ADJ
ajst-7358	54	12	,	,	PUNCT
ajst-7358	54	13	with	with	ADP
ajst-7358	54	14	an	an	DET
ajst-7358	54	15	average	average	NOUN
ajst-7358	54	16	of	of	ADP
ajst-7358	54	17	81	81	NUM
ajst-7358	54	18	%	%	NOUN
ajst-7358	54	19	(	(	PUNCT
ajst-7358	54	20	especially	especially	ADV
ajst-7358	54	21	for	for	ADP
ajst-7358	54	22	arc	arc	NOUN
ajst-7358	54	23	welds	weld	NOUN
ajst-7358	54	24	)	)	PUNCT
ajst-7358	54	25	.	.	PUNCT
ajst-7358	55	1	based	base	VERB
ajst-7358	55	2	on	on	ADP
ajst-7358	55	3	the	the	DET
ajst-7358	55	4	alexnet	alexnet	ADJ
ajst-7358	55	5	network	network	NOUN
ajst-7358	55	6	model	model	NOUN
ajst-7358	55	7	,	,	PUNCT
ajst-7358	55	8	its	its	PRON
ajst-7358	55	9	detection	detection	NOUN
ajst-7358	55	10	accuracy	accuracy	NOUN
ajst-7358	55	11	is	be	AUX
ajst-7358	55	12	generally	generally	ADV
ajst-7358	55	13	57.1	57.1	NUM
ajst-7358	55	14	%	%	NOUN
ajst-7358	55	15	,	,	PUNCT
ajst-7358	55	16	and	and	CCONJ
ajst-7358	55	17	the	the	DET
ajst-7358	55	18	average	average	ADJ
ajst-7358	55	19	accuracy	accuracy	NOUN
ajst-7358	55	20	in	in	ADP
ajst-7358	55	21	the	the	DET
ajst-7358	55	22	roi	roi	NOUN
ajst-7358	55	23	area	area	NOUN
ajst-7358	55	24	of	of	ADP
ajst-7358	55	25	the	the	DET
ajst-7358	55	26	weld	weld	NOUN
ajst-7358	55	27	studied	study	VERB
ajst-7358	55	28	in	in	ADP
ajst-7358	55	29	this	this	DET
ajst-7358	55	30	paper	paper	NOUN
ajst-7358	55	31	reaches	reach	VERB
ajst-7358	55	32	a	a	DET
ajst-7358	55	33	relatively	relatively	ADV
ajst-7358	55	34	high	high	ADJ
ajst-7358	55	35	level	level	NOUN
ajst-7358	55	36	of	of	ADP
ajst-7358	55	37	81	81	NUM
ajst-7358	55	38	%	%	NOUN
ajst-7358	55	39	.	.	PUNCT
ajst-7358	56	1	therefore	therefore	ADV
ajst-7358	56	2	,	,	PUNCT
ajst-7358	56	3	for	for	ADP
ajst-7358	56	4	the	the	DET
ajst-7358	56	5	research	research	NOUN
ajst-7358	56	6	object	object	NOUN
ajst-7358	56	7	described	describe	VERB
ajst-7358	56	8	in	in	ADP
ajst-7358	56	9	this	this	DET
ajst-7358	56	10	paper	paper	NOUN
ajst-7358	56	11	,	,	PUNCT
ajst-7358	56	12	the	the	DET
ajst-7358	56	13	detection	detection	NOUN
ajst-7358	56	14	effect	effect	NOUN
ajst-7358	56	15	is	be	AUX
ajst-7358	56	16	good	good	ADJ
ajst-7358	56	17	,	,	PUNCT
ajst-7358	56	18	and	and	CCONJ
ajst-7358	56	19	can	can	AUX
ajst-7358	56	20	effectively	effectively	ADV
ajst-7358	56	21	reduce	reduce	VERB
ajst-7358	56	22	the	the	DET
ajst-7358	56	23	noise	noise	NOUN
ajst-7358	56	24	level	level	NOUN
ajst-7358	56	25	of	of	ADP
ajst-7358	56	26	welding	welding	NOUN
ajst-7358	56	27	system	system	NOUN
ajst-7358	56	28	.	.	PUNCT
ajst-7358	57	1	acknowledgment	acknowledgment	NOUN
ajst-7358	57	2	supported	support	VERB
ajst-7358	57	3	by	by	ADP
ajst-7358	57	4	university	university	NOUN
ajst-7358	57	5	-	-	PUNCT
ajst-7358	57	6	level	level	NOUN
ajst-7358	57	7	scientific	scientific	ADJ
ajst-7358	57	8	research	research	NOUN
ajst-7358	57	9	project	project	NOUN
ajst-7358	57	10	of	of	ADP
ajst-7358	57	11	geely	geely	PROPN
ajst-7358	57	12	university	university	PROPN
ajst-7358	57	13	of	of	ADP
ajst-7358	57	14	china	china	PROPN
ajst-7358	57	15	"	"	PUNCT
ajst-7358	57	16	research	research	NOUN
ajst-7358	57	17	on	on	ADP
ajst-7358	57	18	welding	welding	NOUN
ajst-7358	57	19	path	path	NOUN
ajst-7358	57	20	planning	planning	NOUN
ajst-7358	57	21	method	method	NOUN
ajst-7358	57	22	based	base	VERB
ajst-7358	57	23	on	on	ADP
ajst-7358	57	24	binocular	binocular	ADJ
ajst-7358	57	25	machine	machine	NOUN
ajst-7358	57	26	vision	vision	NOUN
ajst-7358	57	27	"	"	PUNCT
ajst-7358	57	28	(	(	PUNCT
ajst-7358	57	29	2022xxsy003	2022xxsy003	NUM
ajst-7358	57	30	)	)	PUNCT
ajst-7358	57	31	.	.	PUNCT
ajst-7358	58	1	references	reference	NOUN
ajst-7358	58	2	[	[	X
ajst-7358	58	3	1	1	NUM
ajst-7358	58	4	]	]	X
ajst-7358	58	5	zhang	zhang	PROPN
ajst-7358	58	6	j	j	PROPN
ajst-7358	58	7	,	,	PUNCT
ajst-7358	58	8	tao	tao	PROPN
ajst-7358	58	9	d	d	PROPN
ajst-7358	58	10	.	.	PUNCT
ajst-7358	59	1	famed	fame	VERB
ajst-7358	59	2	-	-	PUNCT
ajst-7358	59	3	net	net	NOUN
ajst-7358	59	4	:	:	PUNCT
ajst-7358	59	5	a	a	DET
ajst-7358	59	6	fast	fast	ADJ
ajst-7358	59	7	and	and	CCONJ
ajst-7358	59	8	accurate	accurate	ADJ
ajst-7358	59	9	multiscale	multiscale	ADJ
ajst-7358	59	10	end	end	NOUN
ajst-7358	59	11	-	-	PUNCT
ajst-7358	59	12	to	to	ADP
ajst-7358	59	13	-	-	PUNCT
ajst-7358	59	14	end	end	NOUN
ajst-7358	59	15	dehazing	dehaze	VERB
ajst-7358	59	16	network[j	network[j	NOUN
ajst-7358	59	17	]	]	PUNCT
ajst-7358	59	18	.	.	PUNCT
ajst-7358	60	1	ieee	ieee	NOUN
ajst-7358	60	2	transactions	transaction	NOUN
ajst-7358	60	3	on	on	ADP
ajst-7358	60	4	image	image	NOUN
ajst-7358	60	5	processing	processing	NOUN
ajst-7358	60	6	,	,	PUNCT
ajst-7358	60	7	2019	2019	NUM
ajst-7358	60	8	.	.	PUNCT
ajst-7358	61	1	[	[	X
ajst-7358	61	2	2	2	NUM
ajst-7358	61	3	]	]	PUNCT
ajst-7358	61	4	chenfan	chenfan	NOUN
ajst-7358	61	5	liu	liu	PROPN
ajst-7358	61	6	,	,	PUNCT
ajst-7358	61	7	junqi	junqi	PROPN
ajst-7358	61	8	shen	shen	PROPN
ajst-7358	61	9	,	,	PUNCT
ajst-7358	61	10	shengsun	shengsun	PROPN
ajst-7358	61	11	hu	hu	PROPN
ajst-7358	61	12	,	,	PUNCT
ajst-7358	61	13	dingyong	dingyong	PROPN
ajst-7358	61	14	wu	wu	PROPN
ajst-7358	61	15	,	,	PUNCT
ajst-7358	61	16	chao	chao	PROPN
ajst-7358	61	17	zhang	zhang	PROPN
ajst-7358	61	18	,	,	PUNCT
ajst-7358	61	19	hui	hui	PROPN
ajst-7358	61	20	yang	yang	PROPN
ajst-7358	61	21	,	,	PUNCT
ajst-7358	61	22	seam	seam	NOUN
ajst-7358	61	23	tracking	tracking	NOUN
ajst-7358	61	24	system	system	NOUN
ajst-7358	61	25	based	base	VERB
ajst-7358	61	26	on	on	ADP
ajst-7358	61	27	laser	laser	NOUN
ajst-7358	61	28	vision	vision	NOUN
ajst-7358	61	29	and	and	CCONJ
ajst-7358	61	30	cgan	cgan	VERB
ajst-7358	61	31	for	for	ADP
ajst-7358	61	32	robotic	robotic	ADJ
ajst-7358	61	33	multi	multi	ADJ
ajst-7358	61	34	-	-	ADJ
ajst-7358	61	35	layer	layer	NOUN
ajst-7358	61	36	and	and	CCONJ
ajst-7358	61	37	multi	multi	ADJ
ajst-7358	61	38	-	-	ADJ
ajst-7358	61	39	pass	pass	ADJ
ajst-7358	61	40	mag	mag	ADJ
ajst-7358	61	41	welding[j].engineering	welding[j].engineere	VERB
ajst-7358	61	42	applications	application	NOUN
ajst-7358	61	43	of	of	ADP
ajst-7358	61	44	artificial	artificial	ADJ
ajst-7358	61	45	intelligence	intelligence	NOUN
ajst-7358	61	46	,	,	PUNCT
ajst-7358	61	47	volume	volume	NOUN
ajst-7358	61	48	116,2022	116,2022	NUM
ajst-7358	61	49	.	.	PUNCT
ajst-7358	62	1	[	[	X
ajst-7358	62	2	3	3	X
ajst-7358	62	3	]	]	X
ajst-7358	62	4	shengfeng	shengfeng	PROPN
ajst-7358	62	5	chen	chen	PROPN
ajst-7358	62	6	,	,	PUNCT
ajst-7358	62	7	dezhi	dezhi	PROPN
ajst-7358	62	8	yang	yang	PROPN
ajst-7358	62	9	,	,	PUNCT
ajst-7358	62	10	jian	jian	PROPN
ajst-7358	62	11	liu	liu	PROPN
ajst-7358	62	12	,	,	PUNCT
ajst-7358	62	13	qi	qi	PROPN
ajst-7358	62	14	tian	tian	PROPN
ajst-7358	62	15	,	,	PUNCT
ajst-7358	62	16	feitao	feitao	ADJ
ajst-7358	62	17	zhou.automatic	zhou.automatic	ADJ
ajst-7358	62	18	weld	weld	NOUN
ajst-7358	62	19	type	type	NOUN
ajst-7358	62	20	classification	classification	NOUN
ajst-7358	62	21	,	,	PUNCT
ajst-7358	62	22	tacked	tack	VERB
ajst-7358	62	23	spot	spot	NOUN
ajst-7358	62	24	recognition	recognition	NOUN
ajst-7358	62	25	and	and	CCONJ
ajst-7358	62	26	weld	weld	NOUN
ajst-7358	62	27	roi	roi	NOUN
ajst-7358	62	28	determination	determination	NOUN
ajst-7358	62	29	for	for	ADP
ajst-7358	62	30	robotic	robotic	ADJ
ajst-7358	62	31	welding	welding	NOUN
ajst-7358	62	32	based	base	VERB
ajst-7358	62	33	on	on	ADP
ajst-7358	62	34	modified	modify	VERB
ajst-7358	62	35	yolov5[j].robotics	yolov5[j].robotic	NOUN
ajst-7358	62	36	and	and	CCONJ
ajst-7358	62	37	computerintegrated	computerintegrate	VERB
ajst-7358	62	38	manufacturing	manufacturing	NOUN
ajst-7358	62	39	,	,	PUNCT
ajst-7358	62	40	volume	volume	NOUN
ajst-7358	62	41	81,2023	81,2023	NUM
ajst-7358	62	42	.	.	PUNCT
ajst-7358	63	1	[	[	X
ajst-7358	63	2	4	4	NUM
ajst-7358	63	3	]	]	X
ajst-7358	63	4	yusen	yusen	PROPN
ajst-7358	63	5	geng	geng	PROPN
ajst-7358	63	6	,	,	PUNCT
ajst-7358	63	7	min	min	PROPN
ajst-7358	63	8	lai	lai	PROPN
ajst-7358	63	9	,	,	PUNCT
ajst-7358	63	10	xincheng	xincheng	PROPN
ajst-7358	63	11	tian	tian	PROPN
ajst-7358	63	12	,	,	PUNCT
ajst-7358	63	13	xiaolong	xiaolong	PROPN
ajst-7358	63	14	xu	xu	PROPN
ajst-7358	63	15	,	,	PUNCT
ajst-7358	63	16	yong	yong	PROPN
ajst-7358	63	17	jiang	jiang	PROPN
ajst-7358	63	18	,	,	PUNCT
ajst-7358	63	19	yuankai	yuankai	PROPN
ajst-7358	63	20	zhang	zhang	PROPN
ajst-7358	63	21	,	,	PUNCT
ajst-7358	63	22	a	a	DET
ajst-7358	63	23	novel	novel	ADJ
ajst-7358	63	24	seam	seam	NOUN
ajst-7358	63	25	extraction	extraction	NOUN
ajst-7358	63	26	and	and	CCONJ
ajst-7358	63	27	path	path	NOUN
ajst-7358	63	28	planning	planning	NOUN
ajst-7358	63	29	method	method	NOUN
ajst-7358	63	30	for	for	ADP
ajst-7358	63	31	robotic	robotic	ADJ
ajst-7358	63	32	welding	welding	NOUN
ajst-7358	63	33	of	of	ADP
ajst-7358	63	34	medium	medium	ADJ
ajst-7358	63	35	-	-	PUNCT
ajst-7358	63	36	thickness	thickness	NOUN
ajst-7358	63	37	44	44	NUM
ajst-7358	63	38	plate	plate	NOUN
ajst-7358	63	39	structural	structural	ADJ
ajst-7358	63	40	parts	part	NOUN
ajst-7358	63	41	based	base	VERB
ajst-7358	63	42	on	on	ADP
ajst-7358	63	43	3d	3d	PROPN
ajst-7358	63	44	vision[j].robotics	vision[j].robotic	NOUN
ajst-7358	63	45	and	and	CCONJ
ajst-7358	63	46	computer	computer	NOUN
ajst-7358	63	47	-	-	PUNCT
ajst-7358	63	48	integrated	integrate	VERB
ajst-7358	63	49	manufacturing	manufacturing	NOUN
ajst-7358	63	50	,	,	PUNCT
ajst-7358	63	51	volume	volume	NOUN
ajst-7358	63	52	79,2023	79,2023	NUM
ajst-7358	63	53	.	.	PUNCT
ajst-7358	64	1	[	[	X
ajst-7358	64	2	5	5	X
ajst-7358	64	3	]	]	SYM
ajst-7358	64	4	yanbiao	yanbiao	X
ajst-7358	64	5	zou	zou	PROPN
ajst-7358	64	6	,	,	PUNCT
ajst-7358	64	7	guohao	guohao	PROPN
ajst-7358	64	8	zeng	zeng	PROPN
ajst-7358	64	9	,	,	PUNCT
ajst-7358	64	10	light	light	ADJ
ajst-7358	64	11	-	-	PUNCT
ajst-7358	64	12	weight	weight	NOUN
ajst-7358	64	13	segmentation	segmentation	NOUN
ajst-7358	64	14	network	network	NOUN
ajst-7358	64	15	based	base	VERB
ajst-7358	64	16	on	on	ADP
ajst-7358	64	17	solov2	solov2	NOUN
ajst-7358	64	18	for	for	ADP
ajst-7358	64	19	weld	weld	NOUN
ajst-7358	64	20	seam	seam	NOUN
ajst-7358	64	21	feature	feature	NOUN
ajst-7358	64	22	extraction[j	extraction[j	PROPN
ajst-7358	64	23	]	]	PUNCT
ajst-7358	64	24	.	.	PUNCT
ajst-7358	65	1	measureent	measureent	PROPN
ajst-7358	65	2	,	,	PUNCT
ajst-7358	65	3	volume	volume	NOUN
ajst-7358	65	4	208,2023	208,2023	NUM
ajst-7358	65	5	.	.	PUNCT
ajst-7358	66	1	[	[	X
ajst-7358	66	2	6	6	NUM
ajst-7358	66	3	]	]	X
ajst-7358	66	4	liangyuan	liangyuan	PROPN
ajst-7358	66	5	deng	deng	PROPN
ajst-7358	66	6	,	,	PUNCT
ajst-7358	66	7	ting	ting	PROPN
ajst-7358	66	8	lei	lei	PROPN
ajst-7358	66	9	,	,	PUNCT
ajst-7358	66	10	chaoqun	chaoqun	PROPN
ajst-7358	66	11	wu	wu	PROPN
ajst-7358	66	12	,	,	PUNCT
ajst-7358	66	13	yibo	yibo	PROPN
ajst-7358	66	14	liu	liu	PROPN
ajst-7358	66	15	,	,	PUNCT
ajst-7358	66	16	shiyu	shiyu	PROPN
ajst-7358	66	17	cao	cao	PROPN
ajst-7358	66	18	,	,	PUNCT
ajst-7358	66	19	song	song	NOUN
ajst-7358	66	20	zhao	zhao	PROPN
ajst-7358	66	21	,	,	PUNCT
ajst-7358	66	22	a	a	DET
ajst-7358	66	23	weld	weld	NOUN
ajst-7358	66	24	seam	seam	NOUN
ajst-7358	66	25	feature	feature	VERB
ajst-7358	66	26	real	real	ADJ
ajst-7358	66	27	-	-	PUNCT
ajst-7358	66	28	time	time	NOUN
ajst-7358	66	29	extraction	extraction	NOUN
ajst-7358	66	30	method	method	NOUN
ajst-7358	66	31	of	of	ADP
ajst-7358	66	32	three	three	NUM
ajst-7358	66	33	typical	typical	ADJ
ajst-7358	66	34	welds	weld	NOUN
ajst-7358	66	35	based	base	VERB
ajst-7358	66	36	on	on	ADP
ajst-7358	66	37	target	target	NOUN
ajst-7358	66	38	detection[j	detection[j	PROPN
ajst-7358	66	39	]	]	PUNCT
ajst-7358	66	40	.	.	PUNCT
ajst-7358	67	1	measurement	measurement	PROPN
ajst-7358	67	2	,	,	PUNCT
ajst-7358	67	3	volume	volume	NOUN
ajst-7358	67	4	207,2023	207,2023	NUM
ajst-7358	67	5	.	.	PUNCT
ajst-7358	68	1	[	[	X
ajst-7358	68	2	7	7	X
ajst-7358	68	3	]	]	X
ajst-7358	68	4	runquan	runquan	PROPN
ajst-7358	68	5	xiao	xiao	PROPN
ajst-7358	68	6	,	,	PUNCT
ajst-7358	68	7	yanling	yanling	PROPN
ajst-7358	68	8	xu	xu	PROPN
ajst-7358	68	9	,	,	PUNCT
ajst-7358	68	10	zhen	zhen	PROPN
ajst-7358	68	11	hou	hou	PROPN
ajst-7358	68	12	,	,	PUNCT
ajst-7358	68	13	chao	chao	PROPN
ajst-7358	68	14	chen	chen	PROPN
ajst-7358	68	15	,	,	PUNCT
ajst-7358	68	16	shanben	shanben	PROPN
ajst-7358	68	17	chen	chen	PROPN
ajst-7358	68	18	,	,	PUNCT
ajst-7358	68	19	a	a	DET
ajst-7358	68	20	feature	feature	NOUN
ajst-7358	68	21	extraction	extraction	NOUN
ajst-7358	68	22	algorithm	algorithm	NOUN
ajst-7358	68	23	based	base	VERB
ajst-7358	68	24	on	on	ADP
ajst-7358	68	25	improved	improved	ADJ
ajst-7358	68	26	snake	snake	NOUN
ajst-7358	68	27	model	model	NOUN
ajst-7358	68	28	for	for	ADP
ajst-7358	68	29	multi	multi	ADJ
ajst-7358	68	30	-	-	ADJ
ajst-7358	68	31	pass	pass	ADJ
ajst-7358	68	32	seam	seam	NOUN
ajst-7358	68	33	tracking	tracking	NOUN
ajst-7358	68	34	in	in	ADP
ajst-7358	68	35	robotic	robotic	ADJ
ajst-7358	68	36	arc	arc	NOUN
ajst-7358	68	37	welding	weld	VERB
ajst-7358	68	38	[	[	X
ajst-7358	68	39	j	j	X
ajst-7358	68	40	]	]	X
ajst-7358	68	41	.	.	PUNCT
ajst-7358	69	1	journal	journal	PROPN
ajst-7358	69	2	of	of	ADP
ajst-7358	69	3	manufacturing	manufacturing	NOUN
ajst-7358	69	4	processes	process	NOUN
ajst-7358	69	5	,	,	PUNCT
ajst-7358	69	6	volume	volume	NOUN
ajst-7358	69	7	72,2021	72,2021	NUM
ajst-7358	69	8	.	.	PUNCT
ajst-7358	70	1	[	[	X
ajst-7358	70	2	8	8	NUM
ajst-7358	70	3	]	]	X
ajst-7358	70	4	runquan	runquan	PROPN
ajst-7358	70	5	xiao	xiao	PROPN
ajst-7358	70	6	,	,	PUNCT
ajst-7358	70	7	yanling	yanling	PROPN
ajst-7358	70	8	xu	xu	PROPN
ajst-7358	70	9	,	,	PUNCT
ajst-7358	70	10	zhen	zhen	PROPN
ajst-7358	70	11	hou	hou	PROPN
ajst-7358	70	12	,	,	PUNCT
ajst-7358	70	13	chao	chao	PROPN
ajst-7358	70	14	chen	chen	PROPN
ajst-7358	70	15	,	,	PUNCT
ajst-7358	70	16	shanben	shanben	PROPN
ajst-7358	70	17	chen	chen	PROPN
ajst-7358	70	18	,	,	PUNCT
ajst-7358	70	19	an	an	DET
ajst-7358	70	20	adaptive	adaptive	ADJ
ajst-7358	70	21	feature	feature	NOUN
ajst-7358	70	22	extraction	extraction	NOUN
ajst-7358	70	23	algorithm	algorithm	NOUN
ajst-7358	70	24	for	for	ADP
ajst-7358	70	25	multiple	multiple	ADJ
ajst-7358	70	26	typical	typical	ADJ
ajst-7358	70	27	seam	seam	NOUN
ajst-7358	70	28	tracking	tracking	NOUN
ajst-7358	70	29	based	base	VERB
ajst-7358	70	30	on	on	ADP
ajst-7358	70	31	vision	vision	NOUN
ajst-7358	70	32	sensor	sensor	NOUN
ajst-7358	70	33	in	in	ADP
ajst-7358	70	34	robotic	robotic	ADJ
ajst-7358	70	35	arc	arc	NOUN
ajst-7358	70	36	welding[j].sensors	welding[j].sensor	NOUN
ajst-7358	70	37	and	and	CCONJ
ajst-7358	70	38	actuators	actuator	NOUN
ajst-7358	70	39	a	a	DET
ajst-7358	70	40	:	:	PUNCT
ajst-7358	70	41	physical	physical	ADJ
ajst-7358	70	42	,	,	PUNCT
ajst-7358	70	43	volume	volume	NOUN
ajst-7358	70	44	297	297	NUM
ajst-7358	70	45	,	,	PUNCT
ajst-7358	70	46	2019	2019	NUM
ajst-7358	70	47	.	.	PUNCT
ajst-7358	71	1	[	[	X
ajst-7358	71	2	9	9	NUM
ajst-7358	71	3	]	]	X
ajst-7358	71	4	hui	hui	PROPN
ajst-7358	71	5	wang	wang	PROPN
ajst-7358	71	6	,	,	PUNCT
ajst-7358	71	7	yu	yu	PROPN
ajst-7358	71	8	huang	huang	PROPN
ajst-7358	71	9	,	,	PUNCT
ajst-7358	71	10	guojun	guojun	PROPN
ajst-7358	71	11	zhang	zhang	PROPN
ajst-7358	71	12	,	,	PUNCT
ajst-7358	71	13	youmin	youmin	PROPN
ajst-7358	71	14	rong	rong	PROPN
ajst-7358	71	15	,	,	PUNCT
ajst-7358	71	16	a	a	DET
ajst-7358	71	17	novel	novel	ADJ
ajst-7358	71	18	method	method	NOUN
ajst-7358	71	19	for	for	ADP
ajst-7358	71	20	dense	dense	ADJ
ajst-7358	71	21	point	point	NOUN
ajst-7358	71	22	cloud	cloud	NOUN
ajst-7358	71	23	reconstruction	reconstruction	NOUN
ajst-7358	71	24	and	and	CCONJ
ajst-7358	71	25	weld	weld	NOUN
ajst-7358	71	26	seam	seam	NOUN
ajst-7358	71	27	detection	detection	NOUN
ajst-7358	71	28	for	for	ADP
ajst-7358	71	29	tubesheet	tubesheet	NOUN
ajst-7358	71	30	welding	welding	NOUN
ajst-7358	71	31	robot[j].optics	robot[j].optics	PROPN
ajst-7358	71	32	&	&	CCONJ
ajst-7358	71	33	laser	laser	PROPN
ajst-7358	71	34	technology	technology	NOUN
ajst-7358	71	35	,	,	PUNCT
ajst-7358	71	36	volume	volume	NOUN
ajst-7358	71	37	163,2023	163,2023	NUM
ajst-7358	71	38	.	.	PUNCT
ajst-7358	72	1	[	[	X
ajst-7358	72	2	10	10	NUM
ajst-7358	72	3	]	]	X
ajst-7358	72	4	jianyong	jianyong	PROPN
ajst-7358	72	5	wang	wang	PROPN
ajst-7358	72	6	,	,	PUNCT
ajst-7358	72	7	chunyang	chunyang	PROPN
ajst-7358	72	8	mu	mu	PROPN
ajst-7358	72	9	,	,	PUNCT
ajst-7358	72	10	song	song	NOUN
ajst-7358	72	11	mu	mu	PROPN
ajst-7358	72	12	,	,	PUNCT
ajst-7358	72	13	rui	rui	PROPN
ajst-7358	72	14	zhu	zhu	PROPN
ajst-7358	72	15	,	,	PUNCT
ajst-7358	72	16	hua	hua	PROPN
ajst-7358	72	17	yu	yu	PROPN
ajst-7358	72	18	,	,	PUNCT
ajst-7358	72	19	welding	weld	VERB
ajst-7358	72	20	seam	seam	NOUN
ajst-7358	72	21	detection	detection	NOUN
ajst-7358	72	22	and	and	CCONJ
ajst-7358	72	23	location	location	NOUN
ajst-7358	72	24	:	:	PUNCT
ajst-7358	72	25	deep	deep	ADJ
ajst-7358	72	26	learning	learn	VERB
ajst-7358	72	27	network	network	NOUN
ajst-7358	72	28	-	-	PUNCT
ajst-7358	72	29	based	base	VERB
ajst-7358	72	30	approach[j].international	approach[j].international	ADJ
ajst-7358	72	31	journal	journal	NOUN
ajst-7358	72	32	of	of	ADP
ajst-7358	72	33	pressure	pressure	NOUN
ajst-7358	72	34	vessels	vessel	NOUN
ajst-7358	72	35	and	and	CCONJ
ajst-7358	72	36	piping	piping	NOUN
ajst-7358	72	37	,	,	PUNCT
ajst-7358	72	38	volume	volume	NOUN
ajst-7358	72	39	202,202	202,202	NUM