id	sid	tid	token	lemma	pos
ajst-16323	1	1	academic	academic	ADJ
ajst-16323	1	2	journal	journal	NOUN
ajst-16323	1	3	of	of	ADP
ajst-16323	1	4	science	science	NOUN
ajst-16323	1	5	and	and	CCONJ
ajst-16323	1	6	technology	technology	NOUN
ajst-16323	1	7	issn	issn	NOUN
ajst-16323	1	8	:	:	PUNCT
ajst-16323	1	9	2771	2771	NUM
ajst-16323	1	10	-	-	SYM
ajst-16323	1	11	3032	3032	NUM
ajst-16323	1	12	|	|	NOUN
ajst-16323	1	13	vol	vol	NOUN
ajst-16323	1	14	.	.	PROPN
ajst-16323	2	1	9	9	NUM
ajst-16323	2	2	,	,	PUNCT
ajst-16323	2	3	no	no	INTJ
ajst-16323	2	4	.	.	NOUN
ajst-16323	2	5	1	1	NUM
ajst-16323	2	6	,	,	PUNCT
ajst-16323	2	7	2024	2024	NUM
ajst-16323	2	8	101	101	NUM
ajst-16323	2	9	research	research	NOUN
ajst-16323	2	10	on	on	ADP
ajst-16323	2	11	algorithm	algorithm	NOUN
ajst-16323	2	12	of	of	ADP
ajst-16323	2	13	light	light	ADJ
ajst-16323	2	14	strip	strip	PROPN
ajst-16323	2	15	center	center	NOUN
ajst-16323	2	16	extraction	extraction	NOUN
ajst-16323	2	17	based	base	VERB
ajst-16323	2	18	on	on	ADP
ajst-16323	2	19	deep	deep	ADJ
ajst-16323	2	20	learning	learning	NOUN
ajst-16323	2	21	xiangwen	xiangwen	PROPN
ajst-16323	3	1	zheng	zheng	PROPN
ajst-16323	3	2	,	,	PUNCT
ajst-16323	3	3	jun	jun	PROPN
ajst-16323	3	4	li	li	PROPN
ajst-16323	3	5	,	,	PUNCT
ajst-16323	3	6	chen	chen	PROPN
ajst-16323	3	7	he	he	PRON
ajst-16323	3	8	college	college	PROPN
ajst-16323	3	9	of	of	ADP
ajst-16323	3	10	mechanical	mechanical	ADJ
ajst-16323	3	11	engineering	engineering	NOUN
ajst-16323	3	12	,	,	PUNCT
ajst-16323	3	13	sichuan	sichuan	PROPN
ajst-16323	3	14	university	university	PROPN
ajst-16323	3	15	of	of	ADP
ajst-16323	3	16	science	science	PROPN
ajst-16323	3	17	&	&	CCONJ
ajst-16323	3	18	engineering	engineering	PROPN
ajst-16323	3	19	,	,	PUNCT
ajst-16323	3	20	zigong	zigong	PROPN
ajst-16323	3	21	64300	64300	NUM
ajst-16323	3	22	,	,	PUNCT
ajst-16323	3	23	sichuan	sichuan	PROPN
ajst-16323	3	24	,	,	PUNCT
ajst-16323	3	25	china	china	PROPN
ajst-16323	3	26	abstract	abstract	PROPN
ajst-16323	3	27	:	:	PUNCT
ajst-16323	3	28	in	in	ADP
ajst-16323	3	29	view	view	NOUN
ajst-16323	3	30	of	of	ADP
ajst-16323	3	31	the	the	DET
ajst-16323	3	32	phenomena	phenomenon	NOUN
ajst-16323	3	33	such	such	ADJ
ajst-16323	3	34	as	as	ADP
ajst-16323	3	35	under	under	NOUN
ajst-16323	3	36	-	-	PUNCT
ajst-16323	3	37	exposure	exposure	NOUN
ajst-16323	3	38	of	of	ADP
ajst-16323	3	39	light	light	ADJ
ajst-16323	3	40	strips	strip	NOUN
ajst-16323	3	41	and	and	CCONJ
ajst-16323	3	42	noise	noise	NOUN
ajst-16323	3	43	interference	interference	NOUN
ajst-16323	3	44	caused	cause	VERB
ajst-16323	3	45	by	by	ADP
ajst-16323	3	46	complex	complex	ADJ
ajst-16323	3	47	surfaces	surface	NOUN
ajst-16323	3	48	of	of	ADP
ajst-16323	3	49	objects	object	NOUN
ajst-16323	3	50	,	,	PUNCT
ajst-16323	3	51	it	it	PRON
ajst-16323	3	52	is	be	AUX
ajst-16323	3	53	difficult	difficult	ADJ
ajst-16323	3	54	for	for	SCONJ
ajst-16323	3	55	traditional	traditional	ADJ
ajst-16323	3	56	light	light	ADJ
ajst-16323	3	57	strip	strip	PROPN
ajst-16323	3	58	center	center	NOUN
ajst-16323	3	59	extraction	extraction	NOUN
ajst-16323	3	60	algorithm	algorithm	NOUN
ajst-16323	3	61	to	to	PART
ajst-16323	3	62	achieve	achieve	VERB
ajst-16323	3	63	light	light	ADJ
ajst-16323	3	64	strip	strip	NOUN
ajst-16323	3	65	center	center	NOUN
ajst-16323	3	66	extraction	extraction	NOUN
ajst-16323	3	67	.	.	PUNCT
ajst-16323	4	1	therefore	therefore	ADV
ajst-16323	4	2	,	,	PUNCT
ajst-16323	4	3	this	this	DET
ajst-16323	4	4	paper	paper	NOUN
ajst-16323	4	5	studies	study	VERB
ajst-16323	4	6	the	the	DET
ajst-16323	4	7	extraction	extraction	NOUN
ajst-16323	4	8	of	of	ADP
ajst-16323	4	9	light	light	ADJ
ajst-16323	4	10	strip	strip	PROPN
ajst-16323	4	11	center	center	NOUN
ajst-16323	4	12	line	line	NOUN
ajst-16323	4	13	based	base	VERB
ajst-16323	4	14	on	on	ADP
ajst-16323	4	15	semantic	semantic	ADJ
ajst-16323	4	16	segmentation	segmentation	NOUN
ajst-16323	4	17	network	network	NOUN
ajst-16323	4	18	algorithm	algorithm	NOUN
ajst-16323	4	19	based	base	VERB
ajst-16323	4	20	on	on	ADP
ajst-16323	4	21	deep	deep	ADJ
ajst-16323	4	22	learning	learning	NOUN
ajst-16323	4	23	,	,	PUNCT
ajst-16323	4	24	uses	use	VERB
ajst-16323	4	25	deep	deep	ADJ
ajst-16323	4	26	learning	learn	VERB
ajst-16323	4	27	algorithm	algorithm	NOUN
ajst-16323	4	28	to	to	ADP
ajst-16323	4	29	presegment	presegment	VERB
ajst-16323	4	30	light	light	ADJ
ajst-16323	4	31	strips	strip	NOUN
ajst-16323	4	32	,	,	PUNCT
ajst-16323	4	33	and	and	CCONJ
ajst-16323	4	34	then	then	ADV
ajst-16323	4	35	uses	use	VERB
ajst-16323	4	36	gray	gray	ADJ
ajst-16323	4	37	prime	prime	ADJ
ajst-16323	4	38	-	-	PUNCT
ajst-16323	4	39	core	core	NOUN
ajst-16323	4	40	method	method	NOUN
ajst-16323	4	41	to	to	PART
ajst-16323	4	42	extract	extract	VERB
ajst-16323	4	43	light	light	ADJ
ajst-16323	4	44	strips	strip	NOUN
ajst-16323	4	45	subpixel	subpixel	NOUN
ajst-16323	4	46	.	.	PUNCT
ajst-16323	5	1	improve	improve	VERB
ajst-16323	5	2	the	the	DET
ajst-16323	5	3	stability	stability	NOUN
ajst-16323	5	4	and	and	CCONJ
ajst-16323	5	5	accuracy	accuracy	NOUN
ajst-16323	5	6	of	of	ADP
ajst-16323	5	7	center	center	ADJ
ajst-16323	5	8	line	line	NOUN
ajst-16323	5	9	extraction	extraction	NOUN
ajst-16323	5	10	.	.	PUNCT
ajst-16323	6	1	keywords	keyword	NOUN
ajst-16323	6	2	:	:	PUNCT
ajst-16323	6	3	complex	complex	ADJ
ajst-16323	6	4	surface	surface	NOUN
ajst-16323	6	5	;	;	PUNCT
ajst-16323	6	6	strip	strip	NOUN
ajst-16323	6	7	center	center	NOUN
ajst-16323	6	8	;	;	PUNCT
ajst-16323	6	9	deep	deep	ADJ
ajst-16323	6	10	learning	learning	NOUN
ajst-16323	6	11	;	;	PUNCT
ajst-16323	6	12	pre	pre	ADJ
ajst-16323	6	13	-	-	VERB
ajst-16323	6	14	segmtioned	segmtioned	ADJ
ajst-16323	6	15	.	.	PUNCT
ajst-16323	7	1	1	1	X
ajst-16323	7	2	.	.	X
ajst-16323	7	3	introduction	introduction	NOUN
ajst-16323	7	4	as	as	ADP
ajst-16323	7	5	a	a	DET
ajst-16323	7	6	non	non	ADJ
ajst-16323	7	7	-	-	ADJ
ajst-16323	7	8	contact	contact	ADJ
ajst-16323	7	9	active	active	ADJ
ajst-16323	7	10	measurement	measurement	NOUN
ajst-16323	7	11	technology	technology	NOUN
ajst-16323	7	12	,	,	PUNCT
ajst-16323	7	13	the	the	DET
ajst-16323	7	14	line	line	NOUN
ajst-16323	7	15	structured	structure	VERB
ajst-16323	7	16	light	light	ADJ
ajst-16323	7	17	measurement	measurement	NOUN
ajst-16323	7	18	method	method	NOUN
ajst-16323	7	19	is	be	AUX
ajst-16323	7	20	widely	widely	ADV
ajst-16323	7	21	used	use	VERB
ajst-16323	7	22	in	in	ADP
ajst-16323	7	23	the	the	DET
ajst-16323	7	24	fields	field	NOUN
ajst-16323	7	25	of	of	ADP
ajst-16323	7	26	workpiece	workpiece	NOUN
ajst-16323	7	27	size	size	NOUN
ajst-16323	7	28	measurement	measurement	NOUN
ajst-16323	7	29	,	,	PUNCT
ajst-16323	7	30	defect	defect	ADJ
ajst-16323	7	31	detection	detection	NOUN
ajst-16323	7	32	,	,	PUNCT
ajst-16323	7	33	3d	3d	PROPN
ajst-16323	7	34	reconstruction	reconstruction	NOUN
ajst-16323	7	35	,	,	PUNCT
ajst-16323	7	36	etc[1	etc[1	X
ajst-16323	7	37	]	]	PUNCT
ajst-16323	7	38	.	.	PUNCT
ajst-16323	8	1	at	at	ADP
ajst-16323	8	2	present	present	ADJ
ajst-16323	8	3	,	,	PUNCT
ajst-16323	8	4	there	there	PRON
ajst-16323	8	5	are	be	VERB
ajst-16323	8	6	two	two	NUM
ajst-16323	8	7	main	main	ADJ
ajst-16323	8	8	methods	method	NOUN
ajst-16323	8	9	for	for	ADP
ajst-16323	8	10	extracting	extract	VERB
ajst-16323	8	11	linear	linear	ADJ
ajst-16323	8	12	structured	structure	VERB
ajst-16323	8	13	light	light	NOUN
ajst-16323	8	14	centers	center	NOUN
ajst-16323	8	15	,	,	PUNCT
ajst-16323	8	16	namely	namely	ADV
ajst-16323	8	17	,	,	PUNCT
ajst-16323	8	18	the	the	DET
ajst-16323	8	19	method	method	NOUN
ajst-16323	8	20	based	base	VERB
ajst-16323	8	21	on	on	ADP
ajst-16323	8	22	geometric	geometric	ADJ
ajst-16323	8	23	center	center	NOUN
ajst-16323	8	24	[	[	X
ajst-16323	8	25	1	1	X
ajst-16323	8	26	]	]	PUNCT
ajst-16323	8	27	and	and	CCONJ
ajst-16323	8	28	the	the	DET
ajst-16323	8	29	method	method	NOUN
ajst-16323	8	30	based	base	VERB
ajst-16323	8	31	on	on	ADP
ajst-16323	8	32	light	light	ADJ
ajst-16323	8	33	intensity	intensity	NOUN
ajst-16323	8	34	distribution	distribution	NOUN
ajst-16323	8	35	[	[	X
ajst-16323	8	36	2	2	NUM
ajst-16323	8	37	]	]	PUNCT
ajst-16323	8	38	.	.	PUNCT
ajst-16323	9	1	geometric	geometric	ADJ
ajst-16323	9	2	center	center	NOUN
ajst-16323	9	3	method	method	NOUN
ajst-16323	9	4	has	have	VERB
ajst-16323	9	5	a	a	DET
ajst-16323	9	6	fast	fast	ADJ
ajst-16323	9	7	operation	operation	NOUN
ajst-16323	9	8	speed	speed	NOUN
ajst-16323	9	9	,	,	PUNCT
ajst-16323	9	10	but	but	CCONJ
ajst-16323	9	11	poor	poor	ADJ
ajst-16323	9	12	extraction	extraction	NOUN
ajst-16323	9	13	accuracy	accuracy	NOUN
ajst-16323	9	14	,	,	PUNCT
ajst-16323	9	15	which	which	PRON
ajst-16323	9	16	can	can	AUX
ajst-16323	9	17	only	only	ADV
ajst-16323	9	18	extract	extract	VERB
ajst-16323	9	19	pixel	pixel	PROPN
ajst-16323	9	20	accuracy	accuracy	NOUN
ajst-16323	9	21	,	,	PUNCT
ajst-16323	9	22	including	include	VERB
ajst-16323	9	23	skeleton	skeleton	NOUN
ajst-16323	9	24	thinning	thin	VERB
ajst-16323	9	25	algorithm	algorithm	NOUN
ajst-16323	9	26	[	[	X
ajst-16323	9	27	3	3	NUM
ajst-16323	9	28	]	]	PUNCT
ajst-16323	9	29	and	and	CCONJ
ajst-16323	9	30	threshold	threshold	NOUN
ajst-16323	9	31	algorithm	algorithm	NOUN
ajst-16323	9	32	.	.	PUNCT
ajst-16323	10	1	methods	method	NOUN
ajst-16323	10	2	based	base	VERB
ajst-16323	10	3	on	on	ADP
ajst-16323	10	4	light	light	ADJ
ajst-16323	10	5	intensity	intensity	NOUN
ajst-16323	10	6	distribution	distribution	NOUN
ajst-16323	10	7	,	,	PUNCT
ajst-16323	10	8	including	include	VERB
ajst-16323	10	9	curve	curve	NOUN
ajst-16323	10	10	fitting	fitting	ADJ
ajst-16323	10	11	,	,	PUNCT
ajst-16323	10	12	gray	gray	ADJ
ajst-16323	10	13	gravity	gravity	NOUN
ajst-16323	10	14	center	center	NOUN
ajst-16323	10	15	and	and	CCONJ
ajst-16323	10	16	steger	steger	PROPN
ajst-16323	10	17	method	method	NOUN
ajst-16323	10	18	[	[	X
ajst-16323	10	19	4	4	NUM
ajst-16323	10	20	]	]	PUNCT
ajst-16323	10	21	,	,	PUNCT
ajst-16323	10	22	can	can	AUX
ajst-16323	10	23	achieve	achieve	VERB
ajst-16323	10	24	subpixel	subpixel	ADJ
ajst-16323	10	25	accuracy	accuracy	NOUN
ajst-16323	10	26	extraction.but	extraction.but	CCONJ
ajst-16323	10	27	these	these	DET
ajst-16323	10	28	methods	method	NOUN
ajst-16323	10	29	often	often	ADV
ajst-16323	10	30	have	have	VERB
ajst-16323	10	31	problems	problem	NOUN
ajst-16323	10	32	such	such	ADJ
ajst-16323	10	33	as	as	ADP
ajst-16323	10	34	low	low	ADJ
ajst-16323	10	35	extraction	extraction	NOUN
ajst-16323	10	36	accuracy	accuracy	NOUN
ajst-16323	10	37	and	and	CCONJ
ajst-16323	10	38	poor	poor	ADJ
ajst-16323	10	39	antiinterference	antiinterference	NOUN
ajst-16323	10	40	ability	ability	NOUN
ajst-16323	10	41	when	when	SCONJ
ajst-16323	10	42	dealing	deal	VERB
ajst-16323	10	43	with	with	ADP
ajst-16323	10	44	complex	complex	ADJ
ajst-16323	10	45	,	,	PUNCT
ajst-16323	10	46	noisy	noisy	ADJ
ajst-16323	10	47	and	and	CCONJ
ajst-16323	10	48	unequal	unequal	ADJ
ajst-16323	10	49	lighting	lighting	NOUN
ajst-16323	10	50	scenes	scene	NOUN
ajst-16323	10	51	.	.	PUNCT
ajst-16323	11	1	in	in	ADP
ajst-16323	11	2	recent	recent	ADJ
ajst-16323	11	3	years	year	NOUN
ajst-16323	11	4	,	,	PUNCT
ajst-16323	11	5	the	the	DET
ajst-16323	11	6	rapid	rapid	ADJ
ajst-16323	11	7	development	development	NOUN
ajst-16323	11	8	of	of	ADP
ajst-16323	11	9	deep	deep	ADJ
ajst-16323	11	10	learning	learning	NOUN
ajst-16323	11	11	technology	technology	NOUN
ajst-16323	11	12	provides	provide	VERB
ajst-16323	11	13	a	a	DET
ajst-16323	11	14	new	new	ADJ
ajst-16323	11	15	solution	solution	NOUN
ajst-16323	11	16	for	for	ADP
ajst-16323	11	17	optical	optical	ADJ
ajst-16323	11	18	strip	strip	NOUN
ajst-16323	11	19	center	center	NOUN
ajst-16323	11	20	extraction	extraction	NOUN
ajst-16323	11	21	.	.	PUNCT
ajst-16323	12	1	the	the	DET
ajst-16323	12	2	light	light	PROPN
ajst-16323	12	3	strip	strip	PROPN
ajst-16323	12	4	center	center	NOUN
ajst-16323	12	5	extraction	extraction	NOUN
ajst-16323	12	6	algorithm	algorithm	NOUN
ajst-16323	12	7	based	base	VERB
ajst-16323	12	8	on	on	ADP
ajst-16323	12	9	deep	deep	ADJ
ajst-16323	12	10	learning	learning	NOUN
ajst-16323	12	11	has	have	VERB
ajst-16323	12	12	strong	strong	ADJ
ajst-16323	12	13	feature	feature	NOUN
ajst-16323	12	14	learning	learning	NOUN
ajst-16323	12	15	and	and	CCONJ
ajst-16323	12	16	classification	classification	NOUN
ajst-16323	12	17	ability	ability	NOUN
ajst-16323	12	18	,	,	PUNCT
ajst-16323	12	19	and	and	CCONJ
ajst-16323	12	20	can	can	AUX
ajst-16323	12	21	automatically	automatically	ADV
ajst-16323	12	22	learn	learn	VERB
ajst-16323	12	23	the	the	DET
ajst-16323	12	24	difference	difference	NOUN
ajst-16323	12	25	between	between	ADP
ajst-16323	12	26	the	the	DET
ajst-16323	12	27	feature	feature	NOUN
ajst-16323	12	28	of	of	ADP
ajst-16323	12	29	the	the	DET
ajst-16323	12	30	light	light	ADJ
ajst-16323	12	31	strip	strip	NOUN
ajst-16323	12	32	and	and	CCONJ
ajst-16323	12	33	the	the	DET
ajst-16323	12	34	background	background	NOUN
ajst-16323	12	35	noise	noise	NOUN
ajst-16323	12	36	,	,	PUNCT
ajst-16323	12	37	and	and	CCONJ
ajst-16323	12	38	realize	realize	VERB
ajst-16323	12	39	the	the	DET
ajst-16323	12	40	light	light	ADJ
ajst-16323	12	41	strip	strip	PROPN
ajst-16323	12	42	center	center	NOUN
ajst-16323	12	43	extraction	extraction	NOUN
ajst-16323	12	44	with	with	ADP
ajst-16323	12	45	high	high	ADJ
ajst-16323	12	46	precision	precision	NOUN
ajst-16323	12	47	and	and	CCONJ
ajst-16323	12	48	high	high	ADJ
ajst-16323	12	49	anti	anti	ADJ
ajst-16323	12	50	-	-	NOUN
ajst-16323	12	51	interference	interference	NOUN
ajst-16323	12	52	.	.	PUNCT
ajst-16323	13	1	in	in	ADP
ajst-16323	13	2	this	this	DET
ajst-16323	13	3	study	study	NOUN
ajst-16323	13	4	,	,	PUNCT
ajst-16323	13	5	based	base	VERB
ajst-16323	13	6	on	on	ADP
ajst-16323	13	7	deep	deep	ADJ
ajst-16323	13	8	learning	learning	NOUN
ajst-16323	13	9	,	,	PUNCT
ajst-16323	13	10	the	the	DET
ajst-16323	13	11	center	center	NOUN
ajst-16323	13	12	of	of	ADP
ajst-16323	13	13	the	the	DET
ajst-16323	13	14	light	light	ADJ
ajst-16323	13	15	strip	strip	NOUN
ajst-16323	13	16	image	image	NOUN
ajst-16323	13	17	on	on	ADP
ajst-16323	13	18	the	the	DET
ajst-16323	13	19	surface	surface	NOUN
ajst-16323	13	20	of	of	ADP
ajst-16323	13	21	complex	complex	ADJ
ajst-16323	13	22	objects	object	NOUN
ajst-16323	13	23	is	be	AUX
ajst-16323	13	24	extracted	extract	VERB
ajst-16323	13	25	to	to	PART
ajst-16323	13	26	ensure	ensure	VERB
ajst-16323	13	27	the	the	DET
ajst-16323	13	28	accuracy	accuracy	NOUN
ajst-16323	13	29	and	and	CCONJ
ajst-16323	13	30	stability	stability	NOUN
ajst-16323	13	31	of	of	ADP
ajst-16323	13	32	the	the	DET
ajst-16323	13	33	light	light	ADJ
ajst-16323	13	34	strip	strip	PROPN
ajst-16323	13	35	center	center	NOUN
ajst-16323	13	36	.	.	PUNCT
ajst-16323	14	1	2	2	X
ajst-16323	14	2	.	.	X
ajst-16323	14	3	experimental	experimental	ADJ
ajst-16323	14	4	design	design	NOUN
ajst-16323	14	5	in	in	ADP
ajst-16323	14	6	this	this	DET
ajst-16323	14	7	study	study	NOUN
ajst-16323	14	8	,	,	PUNCT
ajst-16323	14	9	industrial	industrial	ADJ
ajst-16323	14	10	cameras	camera	NOUN
ajst-16323	14	11	were	be	AUX
ajst-16323	14	12	used	use	VERB
ajst-16323	14	13	to	to	PART
ajst-16323	14	14	shoot	shoot	VERB
ajst-16323	14	15	200	200	NUM
ajst-16323	14	16	images	image	NOUN
ajst-16323	14	17	of	of	ADP
ajst-16323	14	18	underexposure	underexposure	NOUN
ajst-16323	14	19	,	,	PUNCT
ajst-16323	14	20	noise	noise	NOUN
ajst-16323	14	21	interference	interference	NOUN
ajst-16323	14	22	and	and	CCONJ
ajst-16323	14	23	normal	normal	ADJ
ajst-16323	14	24	exposure	exposure	NOUN
ajst-16323	14	25	light	light	NOUN
ajst-16323	14	26	strips	strip	NOUN
ajst-16323	14	27	at	at	ADP
ajst-16323	14	28	different	different	ADJ
ajst-16323	14	29	locations	location	NOUN
ajst-16323	14	30	on	on	ADP
ajst-16323	14	31	the	the	DET
ajst-16323	14	32	girth	girth	NOUN
ajst-16323	14	33	weld	weld	NOUN
ajst-16323	14	34	,	,	PUNCT
ajst-16323	14	35	a	a	DET
ajst-16323	14	36	total	total	NOUN
ajst-16323	14	37	of	of	ADP
ajst-16323	14	38	600	600	NUM
ajst-16323	14	39	images	image	NOUN
ajst-16323	14	40	were	be	AUX
ajst-16323	14	41	taken	take	VERB
ajst-16323	14	42	,	,	PUNCT
ajst-16323	14	43	and	and	CCONJ
ajst-16323	14	44	the	the	DET
ajst-16323	14	45	images	image	NOUN
ajst-16323	14	46	were	be	AUX
ajst-16323	14	47	preprocessed	preprocesse	VERB
ajst-16323	14	48	.	.	PUNCT
ajst-16323	15	1	unet	unet	NOUN
ajst-16323	15	2	deep	deep	ADJ
ajst-16323	15	3	learning	learning	NOUN
ajst-16323	15	4	algorithm	algorithm	NOUN
ajst-16323	15	5	was	be	AUX
ajst-16323	15	6	used	use	VERB
ajst-16323	15	7	to	to	PART
ajst-16323	15	8	presegment	presegment	VERB
ajst-16323	15	9	the	the	DET
ajst-16323	15	10	light	light	ADJ
ajst-16323	15	11	strip	strip	NOUN
ajst-16323	15	12	images	image	NOUN
ajst-16323	15	13	,	,	PUNCT
ajst-16323	15	14	improve	improve	VERB
ajst-16323	15	15	the	the	DET
ajst-16323	15	16	resolution	resolution	NOUN
ajst-16323	15	17	of	of	ADP
ajst-16323	15	18	the	the	DET
ajst-16323	15	19	underexposure	underexposure	NOUN
ajst-16323	15	20	light	light	NOUN
ajst-16323	15	21	strips	strip	NOUN
ajst-16323	15	22	and	and	CCONJ
ajst-16323	15	23	the	the	DET
ajst-16323	15	24	environment	environment	NOUN
ajst-16323	15	25	,	,	PUNCT
ajst-16323	15	26	and	and	CCONJ
ajst-16323	15	27	reduce	reduce	VERB
ajst-16323	15	28	the	the	DET
ajst-16323	15	29	noise	noise	NOUN
ajst-16323	15	30	interference	interference	NOUN
ajst-16323	15	31	of	of	ADP
ajst-16323	15	32	the	the	DET
ajst-16323	15	33	images	image	NOUN
ajst-16323	15	34	.	.	PUNCT
ajst-16323	16	1	then	then	ADV
ajst-16323	16	2	the	the	DET
ajst-16323	16	3	gray	gray	ADJ
ajst-16323	16	4	-	-	PUNCT
ajst-16323	16	5	scale	scale	NOUN
ajst-16323	16	6	centroid	centroid	NOUN
ajst-16323	16	7	algorithm	algorithm	NOUN
ajst-16323	16	8	is	be	AUX
ajst-16323	16	9	used	use	VERB
ajst-16323	16	10	to	to	PART
ajst-16323	16	11	extract	extract	VERB
ajst-16323	16	12	the	the	DET
ajst-16323	16	13	center	center	ADJ
ajst-16323	16	14	point	point	NOUN
ajst-16323	16	15	of	of	ADP
ajst-16323	16	16	the	the	DET
ajst-16323	16	17	pre	pre	ADJ
ajst-16323	16	18	-	-	ADJ
ajst-16323	16	19	segmented	segmented	ADJ
ajst-16323	16	20	image	image	NOUN
ajst-16323	16	21	,	,	PUNCT
ajst-16323	16	22	improve	improve	VERB
ajst-16323	16	23	the	the	DET
ajst-16323	16	24	stability	stability	NOUN
ajst-16323	16	25	of	of	ADP
ajst-16323	16	26	the	the	DET
ajst-16323	16	27	center	center	ADJ
ajst-16323	16	28	line	line	NOUN
ajst-16323	16	29	of	of	ADP
ajst-16323	16	30	the	the	DET
ajst-16323	16	31	light	light	ADJ
ajst-16323	16	32	strip	strip	NOUN
ajst-16323	16	33	under	under	ADP
ajst-16323	16	34	three	three	NUM
ajst-16323	16	35	conditions	condition	NOUN
ajst-16323	16	36	,	,	PUNCT
ajst-16323	16	37	and	and	CCONJ
ajst-16323	16	38	realize	realize	VERB
ajst-16323	16	39	the	the	DET
ajst-16323	16	40	extraction	extraction	NOUN
ajst-16323	16	41	of	of	ADP
ajst-16323	16	42	the	the	DET
ajst-16323	16	43	light	light	ADJ
ajst-16323	16	44	strip	strip	PROPN
ajst-16323	16	45	center	center	NOUN
ajst-16323	16	46	under	under	ADP
ajst-16323	16	47	different	different	ADJ
ajst-16323	16	48	conditions	condition	NOUN
ajst-16323	16	49	.	.	PUNCT
ajst-16323	17	1	3	3	X
ajst-16323	17	2	.	.	X
ajst-16323	17	3	research	research	NOUN
ajst-16323	17	4	on	on	ADP
ajst-16323	17	5	light	light	ADJ
ajst-16323	17	6	strip	strip	PROPN
ajst-16323	17	7	center	center	NOUN
ajst-16323	17	8	algorithm	algorithm	NOUN
ajst-16323	17	9	based	base	VERB
ajst-16323	17	10	on	on	ADP
ajst-16323	17	11	deep	deep	ADJ
ajst-16323	17	12	learning	learn	VERB
ajst-16323	17	13	3.1	3.1	NUM
ajst-16323	17	14	.	.	PUNCT
ajst-16323	18	1	the	the	DET
ajst-16323	18	2	theory	theory	NOUN
ajst-16323	18	3	for	for	ADP
ajst-16323	18	4	under	under	NOUN
ajst-16323	18	5	-	-	PUNCT
ajst-16323	18	6	exposure	exposure	NOUN
ajst-16323	18	7	and	and	CCONJ
ajst-16323	18	8	noise	noise	NOUN
ajst-16323	18	9	interfered	interfere	VERB
ajst-16323	18	10	images	image	NOUN
ajst-16323	18	11	,	,	PUNCT
ajst-16323	18	12	the	the	DET
ajst-16323	18	13	geometric	geometric	ADJ
ajst-16323	18	14	center	center	NOUN
ajst-16323	18	15	and	and	CCONJ
ajst-16323	18	16	light	light	ADJ
ajst-16323	18	17	intensity	intensity	NOUN
ajst-16323	18	18	distribution	distribution	NOUN
ajst-16323	18	19	method	method	NOUN
ajst-16323	18	20	are	be	AUX
ajst-16323	18	21	not	not	PART
ajst-16323	18	22	easy	easy	ADJ
ajst-16323	18	23	to	to	PART
ajst-16323	18	24	achieve	achieve	VERB
ajst-16323	18	25	the	the	DET
ajst-16323	18	26	centerline	centerline	NOUN
ajst-16323	18	27	extraction	extraction	NOUN
ajst-16323	18	28	in	in	ADP
ajst-16323	18	29	both	both	DET
ajst-16323	18	30	cases	case	NOUN
ajst-16323	18	31	.	.	PUNCT
ajst-16323	19	1	in	in	ADP
ajst-16323	19	2	this	this	DET
ajst-16323	19	3	paper	paper	NOUN
ajst-16323	19	4	,	,	PUNCT
ajst-16323	19	5	unet	unet	NOUN
ajst-16323	19	6	deep	deep	ADJ
ajst-16323	19	7	learning	learning	NOUN
ajst-16323	19	8	algorithm	algorithm	NOUN
ajst-16323	19	9	is	be	AUX
ajst-16323	19	10	used	use	VERB
ajst-16323	19	11	to	to	PART
ajst-16323	19	12	segment	segment	VERB
ajst-16323	19	13	the	the	DET
ajst-16323	19	14	light	light	ADJ
ajst-16323	19	15	strip	strip	NOUN
ajst-16323	19	16	according	accord	VERB
ajst-16323	19	17	to	to	ADP
ajst-16323	19	18	its	its	PRON
ajst-16323	19	19	shape	shape	NOUN
ajst-16323	19	20	,	,	PUNCT
ajst-16323	19	21	and	and	CCONJ
ajst-16323	19	22	then	then	ADV
ajst-16323	19	23	the	the	DET
ajst-16323	19	24	light	light	ADJ
ajst-16323	19	25	intensity	intensity	NOUN
ajst-16323	19	26	distribution	distribution	NOUN
ajst-16323	19	27	algorithm	algorithm	NOUN
ajst-16323	19	28	is	be	AUX
ajst-16323	19	29	used	use	VERB
ajst-16323	19	30	to	to	PART
ajst-16323	19	31	extract	extract	VERB
ajst-16323	19	32	the	the	DET
ajst-16323	19	33	centerline	centerline	NOUN
ajst-16323	19	34	of	of	ADP
ajst-16323	19	35	the	the	DET
ajst-16323	19	36	light	light	NOUN
ajst-16323	19	37	strip.the	strip.the	DET
ajst-16323	19	38	unet	unet	NOUN
ajst-16323	19	39	network	network	NOUN
ajst-16323	19	40	structure	structure	NOUN
ajst-16323	19	41	is	be	AUX
ajst-16323	19	42	shown	show	VERB
ajst-16323	19	43	in	in	ADP
ajst-16323	19	44	fig.1	fig.1	PROPN
ajst-16323	19	45	.	.	PUNCT
ajst-16323	20	1	the	the	DET
ajst-16323	20	2	decoder	decoder	NOUN
ajst-16323	20	3	and	and	CCONJ
ajst-16323	20	4	encoder	encoder	NOUN
ajst-16323	20	5	are	be	AUX
ajst-16323	20	6	symmetric	symmetric	ADJ
ajst-16323	20	7	in	in	ADP
ajst-16323	20	8	unet	unet	NOUN
ajst-16323	20	9	network	network	NOUN
ajst-16323	20	10	,	,	PUNCT
ajst-16323	20	11	and	and	CCONJ
ajst-16323	20	12	the	the	DET
ajst-16323	20	13	network	network	NOUN
ajst-16323	20	14	structure	structure	NOUN
ajst-16323	20	15	is	be	AUX
ajst-16323	20	16	u	u	NOUN
ajst-16323	20	17	-	-	VERB
ajst-16323	20	18	shaped	shaped	ADJ
ajst-16323	20	19	.	.	PUNCT
ajst-16323	21	1	there	there	PRON
ajst-16323	21	2	is	be	VERB
ajst-16323	21	3	no	no	DET
ajst-16323	21	4	fully	fully	ADV
ajst-16323	21	5	connected	connect	VERB
ajst-16323	21	6	layer	layer	NOUN
ajst-16323	21	7	in	in	ADP
ajst-16323	21	8	the	the	DET
ajst-16323	21	9	network	network	NOUN
ajst-16323	21	10	,	,	PUNCT
ajst-16323	21	11	and	and	CCONJ
ajst-16323	21	12	it	it	PRON
ajst-16323	21	13	is	be	AUX
ajst-16323	21	14	a	a	DET
ajst-16323	21	15	typical	typical	ADJ
ajst-16323	21	16	structure	structure	NOUN
ajst-16323	21	17	of	of	ADP
ajst-16323	21	18	convolutional	convolutional	ADJ
ajst-16323	21	19	neural	neural	ADJ
ajst-16323	21	20	networks	network	NOUN
ajst-16323	21	21	in	in	ADP
ajst-16323	21	22	the	the	DET
ajst-16323	21	23	decoder	decoder	NOUN
ajst-16323	21	24	,	,	PUNCT
ajst-16323	21	25	which	which	PRON
ajst-16323	21	26	includes	include	VERB
ajst-16323	21	27	two	two	NUM
ajst-16323	21	28	repeated	repeat	VERB
ajst-16323	21	29	3	3	NUM
ajst-16323	21	30	*	*	SYM
ajst-16323	21	31	3	3	NUM
ajst-16323	21	32	convolutional	convolutional	ADJ
ajst-16323	21	33	operations	operation	NOUN
ajst-16323	21	34	.	.	PUNCT
ajst-16323	22	1	each	each	DET
ajst-16323	22	2	convolution	convolution	NOUN
ajst-16323	22	3	is	be	AUX
ajst-16323	22	4	followed	follow	VERB
ajst-16323	22	5	by	by	ADP
ajst-16323	22	6	a	a	DET
ajst-16323	22	7	relu(rectifier	relu(rectifier	ADJ
ajst-16323	22	8	linear	linear	ADJ
ajst-16323	22	9	units	unit	NOUN
ajst-16323	22	10	)	)	PUNCT
ajst-16323	22	11	activation	activation	NOUN
ajst-16323	22	12	function	function	NOUN
ajst-16323	22	13	and	and	CCONJ
ajst-16323	22	14	a	a	DET
ajst-16323	22	15	2	2	NUM
ajst-16323	22	16	*	*	SYM
ajst-16323	22	17	2	2	NUM
ajst-16323	22	18	maximum	maximum	ADJ
ajst-16323	22	19	pooling	pool	VERB
ajst-16323	22	20	operation	operation	NOUN
ajst-16323	22	21	.	.	PUNCT
ajst-16323	23	1	the	the	DET
ajst-16323	23	2	formula	formula	NOUN
ajst-16323	23	3	of	of	ADP
ajst-16323	23	4	relu	relu	NOUN
ajst-16323	23	5	activation	activation	NOUN
ajst-16323	23	6	function	function	NOUN
ajst-16323	23	7	is	be	AUX
ajst-16323	23	8	shown	show	VERB
ajst-16323	23	9	in	in	ADP
ajst-16323	23	10	eq.(1	eq.(1	ADJ
ajst-16323	23	11	)	)	PUNCT
ajst-16323	23	12	.	.	PUNCT
ajst-16323	24	1	for	for	ADP
ajst-16323	24	2	the	the	DET
ajst-16323	24	3	input	input	NOUN
ajst-16323	24	4	signal	signal	NOUN
ajst-16323	24	5	,	,	PUNCT
ajst-16323	24	6	the	the	DET
ajst-16323	24	7	relu	relu	NOUN
ajst-16323	24	8	function	function	NOUN
ajst-16323	24	9	takes	take	VERB
ajst-16323	24	10	0	0	PUNCT
ajst-16323	24	11	as	as	ADP
ajst-16323	24	12	the	the	DET
ajst-16323	24	13	dividing	dividing	NOUN
ajst-16323	24	14	line	line	NOUN
ajst-16323	24	15	,	,	PUNCT
ajst-16323	24	16	and	and	CCONJ
ajst-16323	24	17	outputs	output	NOUN
ajst-16323	24	18	less	less	ADJ
ajst-16323	24	19	than	than	ADP
ajst-16323	24	20	0	0	NUM
ajst-16323	24	21	are	be	AUX
ajst-16323	24	22	0	0	NUM
ajst-16323	24	23	.	.	PUNCT
ajst-16323	25	1	when	when	SCONJ
ajst-16323	25	2	the	the	DET
ajst-16323	25	3	input	input	NOUN
ajst-16323	25	4	is	be	AUX
ajst-16323	25	5	greater	great	ADJ
ajst-16323	25	6	than	than	ADP
ajst-16323	25	7	0	0	NUM
ajst-16323	25	8	,	,	PUNCT
ajst-16323	25	9	the	the	DET
ajst-16323	25	10	output	output	NOUN
ajst-16323	25	11	value	value	NOUN
ajst-16323	25	12	is	be	AUX
ajst-16323	25	13	equal	equal	ADJ
ajst-16323	25	14	to	to	ADP
ajst-16323	25	15	the	the	DET
ajst-16323	25	16	input	input	NOUN
ajst-16323	25	17	value	value	NOUN
ajst-16323	25	18	.	.	PUNCT
ajst-16323	26	1	in	in	ADP
ajst-16323	26	2	the	the	DET
ajst-16323	26	3	downsampling	downsample	VERB
ajst-16323	26	4	process	process	NOUN
ajst-16323	26	5	,	,	PUNCT
ajst-16323	26	6	the	the	DET
ajst-16323	26	7	process	process	NOUN
ajst-16323	26	8	of	of	ADP
ajst-16323	26	9	pooling	pool	VERB
ajst-16323	26	10	is	be	AUX
ajst-16323	26	11	2	2	NUM
ajst-16323	26	12	steps	step	NOUN
ajst-16323	26	13	,	,	PUNCT
ajst-16323	26	14	and	and	CCONJ
ajst-16323	26	15	the	the	DET
ajst-16323	26	16	depth	depth	NOUN
ajst-16323	26	17	becomes	become	VERB
ajst-16323	26	18	twice	twice	DET
ajst-16323	26	19	the	the	DET
ajst-16323	26	20	original	original	ADJ
ajst-16323	26	21	after	after	ADP
ajst-16323	26	22	each	each	DET
ajst-16323	26	23	stage	stage	NOUN
ajst-16323	26	24	.	.	PUNCT
ajst-16323	27	1	in	in	ADP
ajst-16323	27	2	the	the	DET
ajst-16323	27	3	encoder	encoder	NOUN
ajst-16323	27	4	process	process	NOUN
ajst-16323	27	5	of	of	ADP
ajst-16323	27	6	the	the	DET
ajst-16323	27	7	network	network	NOUN
ajst-16323	27	8	,	,	PUNCT
ajst-16323	27	9	a	a	DET
ajst-16323	27	10	3	3	NUM
ajst-16323	27	11	*	*	SYM
ajst-16323	27	12	3	3	NUM
ajst-16323	27	13	transposed	transpose	VERB
ajst-16323	27	14	convolution	convolution	NOUN
ajst-16323	27	15	with	with	ADP
ajst-16323	27	16	step	step	NOUN
ajst-16323	27	17	size	size	NOUN
ajst-16323	27	18	2	2	NUM
ajst-16323	27	19	is	be	AUX
ajst-16323	27	20	carried	carry	VERB
ajst-16323	27	21	out	out	ADP
ajst-16323	27	22	.	.	PUNCT
ajst-16323	28	1	the	the	DET
ajst-16323	28	2	convolution	convolution	NOUN
ajst-16323	28	3	process	process	NOUN
ajst-16323	28	4	halves	halve	VERB
ajst-16323	28	5	the	the	DET
ajst-16323	28	6	feature	feature	NOUN
ajst-16323	28	7	depth	depth	NOUN
ajst-16323	28	8	,	,	PUNCT
ajst-16323	28	9	concatenates	concatenate	VERB
ajst-16323	28	10	the	the	DET
ajst-16323	28	11	upsampled	upsample	VERB
ajst-16323	28	12	image	image	NOUN
ajst-16323	28	13	with	with	ADP
ajst-16323	28	14	the	the	DET
ajst-16323	28	15	feature	feature	NOUN
ajst-16323	28	16	map	map	NOUN
ajst-16323	28	17	from	from	ADP
ajst-16323	28	18	the	the	DET
ajst-16323	28	19	corresponding	corresponding	ADJ
ajst-16323	28	20	position	position	NOUN
ajst-16323	28	21	in	in	ADP
ajst-16323	28	22	the	the	DET
ajst-16323	28	23	decoder	decoder	NOUN
ajst-16323	28	24	,	,	PUNCT
ajst-16323	28	25	and	and	CCONJ
ajst-16323	28	26	convolves	convolve	VERB
ajst-16323	28	27	the	the	DET
ajst-16323	28	28	concatenated	concatenate	VERB
ajst-16323	28	29	image	image	NOUN
ajst-16323	28	30	.	.	PUNCT
ajst-16323	29	1	each	each	DET
ajst-16323	29	2	convolution	convolution	NOUN
ajst-16323	29	3	is	be	AUX
ajst-16323	29	4	followed	follow	VERB
ajst-16323	29	5	by	by	ADP
ajst-16323	29	6	a	a	DET
ajst-16323	29	7	relu	relu	NOUN
ajst-16323	29	8	activation	activation	NOUN
ajst-16323	29	9	function	function	NOUN
ajst-16323	29	10	.	.	PUNCT
ajst-16323	30	1	because	because	SCONJ
ajst-16323	30	2	some	some	DET
ajst-16323	30	3	details	detail	NOUN
ajst-16323	30	4	will	will	AUX
ajst-16323	30	5	be	be	AUX
ajst-16323	30	6	lost	lose	VERB
ajst-16323	30	7	after	after	ADP
ajst-16323	30	8	each	each	DET
ajst-16323	30	9	operation	operation	NOUN
ajst-16323	30	10	,	,	PUNCT
ajst-16323	30	11	the	the	DET
ajst-16323	30	12	concatenation	concatenation	NOUN
ajst-16323	30	13	operation	operation	NOUN
ajst-16323	30	14	in	in	ADP
ajst-16323	30	15	the	the	DET
ajst-16323	30	16	corresponding	corresponding	ADJ
ajst-16323	30	17	stage	stage	NOUN
ajst-16323	30	18	in	in	ADP
ajst-16323	30	19	the	the	DET
ajst-16323	30	20	network	network	NOUN
ajst-16323	30	21	can	can	AUX
ajst-16323	30	22	effectively	effectively	ADV
ajst-16323	30	23	reduce	reduce	VERB
ajst-16323	30	24	the	the	DET
ajst-16323	30	25	loss	loss	NOUN
ajst-16323	30	26	of	of	ADP
ajst-16323	30	27	details	detail	NOUN
ajst-16323	30	28	in	in	ADP
ajst-16323	30	29	the	the	DET
ajst-16323	30	30	convolution	convolution	NOUN
ajst-16323	30	31	process	process	NOUN
ajst-16323	30	32	.	.	PUNCT
ajst-16323	31	1	102	102	NUM
ajst-16323	31	2	stage1	stage1	PROPN
ajst-16323	31	3	stage2	stage2	PROPN
ajst-16323	31	4	stage3	stage3	PROPN
ajst-16323	31	5	stage4	stage4	PROPN
ajst-16323	31	6	stage5	stage5	PROPN
ajst-16323	31	7	stage6	stage6	PROPN
ajst-16323	31	8	stage7	stage7	PROPN
ajst-16323	32	1	stage8	stage8	PROPN
ajst-16323	32	2	stage9	stage9	X
ajst-16323	32	3	input	input	VERB
ajst-16323	32	4	out	out	ADP
ajst-16323	32	5	figure	figure	NOUN
ajst-16323	32	6	1	1	NUM
ajst-16323	32	7	.	.	PUNCT
ajst-16323	32	8	unet	unet	NOUN
ajst-16323	32	9	deep	deep	ADJ
ajst-16323	32	10	learning	learning	NOUN
ajst-16323	32	11	algorithm	algorithm	NOUN
ajst-16323	32	12	network	network	NOUN
ajst-16323	32	13	structure	structure	NOUN
ajst-16323	32	14	0	0	NUM
ajst-16323	32	15	,	,	PUNCT
ajst-16323	32	16	0	0	NUM
ajst-16323	32	17	(	(	PUNCT
ajst-16323	32	18	)	)	PUNCT
ajst-16323	32	19	{	{	PUNCT
ajst-16323	32	20	,	,	PUNCT
ajst-16323	32	21	0	0	NUM
ajst-16323	33	1	x	x	SYM
ajst-16323	33	2	f	f	NOUN
ajst-16323	33	3	x	x	PUNCT
ajst-16323	33	4	x	x	PUNCT
ajst-16323	33	5	x	x	SYM
ajst-16323	33	6			NUM
ajst-16323	34	1			PRON
ajst-16323	34	2			INTJ
ajst-16323	34	3	(	(	PUNCT
ajst-16323	34	4	1	1	X
ajst-16323	34	5	)	)	PUNCT
ajst-16323	34	6	the	the	DET
ajst-16323	34	7	gray	gray	ADJ
ajst-16323	34	8	-	-	PUNCT
ajst-16323	34	9	level	level	NOUN
ajst-16323	34	10	barycentric	barycentric	ADJ
ajst-16323	34	11	method	method	NOUN
ajst-16323	34	12	can	can	AUX
ajst-16323	34	13	be	be	AUX
ajst-16323	34	14	used	use	VERB
ajst-16323	34	15	to	to	PART
ajst-16323	34	16	calculate	calculate	VERB
ajst-16323	34	17	the	the	DET
ajst-16323	34	18	light	light	ADJ
ajst-16323	34	19	power	power	NOUN
ajst-16323	34	20	gravity	gravity	NOUN
ajst-16323	34	21	centroid	centroid	NOUN
ajst-16323	34	22	coordinates	coordinate	NOUN
ajst-16323	34	23	for	for	ADP
ajst-16323	34	24	the	the	DET
ajst-16323	34	25	target	target	NOUN
ajst-16323	34	26	with	with	ADP
ajst-16323	34	27	uneven	uneven	ADJ
ajst-16323	34	28	brightness	brightness	NOUN
ajst-16323	34	29	according	accord	VERB
ajst-16323	34	30	to	to	ADP
ajst-16323	34	31	the	the	DET
ajst-16323	34	32	target	target	NOUN
ajst-16323	34	33	light	light	ADJ
ajst-16323	34	34	intensity	intensity	NOUN
ajst-16323	34	35	distribution	distribution	NOUN
ajst-16323	34	36	.	.	PUNCT
ajst-16323	35	1	for	for	ADP
ajst-16323	35	2	an	an	DET
ajst-16323	35	3	image	image	NOUN
ajst-16323	35	4	f	f	NOUN
ajst-16323	35	5	with	with	ADP
ajst-16323	35	6	a	a	DET
ajst-16323	35	7	pixel	pixel	ADJ
ajst-16323	35	8	size	size	NOUN
ajst-16323	35	9	of	of	ADP
ajst-16323	35	10	m*n	m*n	PROPN
ajst-16323	35	11	,	,	PUNCT
ajst-16323	35	12	if	if	SCONJ
ajst-16323	35	13	the	the	DET
ajst-16323	35	14	gray	gray	ADJ
ajst-16323	35	15	value	value	NOUN
ajst-16323	35	16	of	of	ADP
ajst-16323	35	17	a	a	DET
ajst-16323	35	18	pixel	pixel	NOUN
ajst-16323	35	19	exceeds	exceed	VERB
ajst-16323	35	20	the	the	DET
ajst-16323	35	21	threshold	threshold	NOUN
ajst-16323	35	22	value	value	NOUN
ajst-16323	35	23	t	t	PROPN
ajst-16323	35	24	,	,	PUNCT
ajst-16323	35	25	it	it	PRON
ajst-16323	35	26	is	be	AUX
ajst-16323	35	27	involved	involve	VERB
ajst-16323	35	28	in	in	ADP
ajst-16323	35	29	the	the	DET
ajst-16323	35	30	gravity	gravity	NOUN
ajst-16323	35	31	center	center	NOUN
ajst-16323	35	32	processing	processing	NOUN
ajst-16323	35	33	as	as	SCONJ
ajst-16323	35	34	shown	show	VERB
ajst-16323	35	35	in	in	ADP
ajst-16323	35	36	eq.(2	eq.(2	ADJ
ajst-16323	35	37	)	)	PUNCT
ajst-16323	35	38	,	,	PUNCT
ajst-16323	35	39	so	so	CCONJ
ajst-16323	35	40	the	the	DET
ajst-16323	35	41	gravity	gravity	NOUN
ajst-16323	35	42	center	center	NOUN
ajst-16323	35	43	coordinates	coordinate	NOUN
ajst-16323	35	44	are	be	AUX
ajst-16323	35	45	shown	show	VERB
ajst-16323	35	46	in	in	ADP
ajst-16323	35	47	eq.(3	eq.(3	NOUN
ajst-16323	35	48	)	)	PUNCT
ajst-16323	35	49	,	,	PUNCT
ajst-16323	35	50	eq.(4	eq.(4	NUM
ajst-16323	35	51	)	)	PUNCT
ajst-16323	35	52	.	.	PUNCT
ajst-16323	36	1	pixel	pixel	PROPN
ajst-16323	36	2	gray	gray	ADJ
ajst-16323	36	3	value	value	NOUN
ajst-16323	36	4	pixel	pixel	PROPN
ajst-16323	36	5	gray	gray	ADJ
ajst-16323	36	6	val	val	PROPN
ajst-16323	36	7	0	0	NUM
ajst-16323	36	8	,	,	PUNCT
ajst-16323	36	9	t	t	PROPN
ajst-16323	36	10	{	{	PUNCT
ajst-16323	36	11	,	,	PUNCT
ajst-16323	36	12	tueij	tueij	PROPN
ajst-16323	37	1	ij	ij	INTJ
ajst-16323	37	2	f	f	PROPN
ajst-16323	37	3	f	f	PROPN
ajst-16323	37	4			PROPN
ajst-16323	37	5			NOUN
ajst-16323	37	6			NUM
ajst-16323	37	7	(	(	PUNCT
ajst-16323	37	8	2	2	NUM
ajst-16323	37	9	)	)	PUNCT
ajst-16323	37	10	1	1	NUM
ajst-16323	37	11	1	1	NUM
ajst-16323	37	12	0	0	NUM
ajst-16323	37	13	1	1	NUM
ajst-16323	37	14	1	1	NUM
ajst-16323	37	15	n	n	NOUN
ajst-16323	38	1	i	i	PRON
ajst-16323	38	2	ij	ij	INTJ
ajst-16323	38	3	i	i	INTJ
ajst-16323	38	4	j	j	PROPN
ajst-16323	38	5	m	m	VERB
ajst-16323	38	6	j	j	PROPN
ajst-16323	38	7	m	m	VERB
ajst-16323	39	1	n	n	ADV
ajst-16323	40	1	i	i	PRON
ajst-16323	41	1	i	i	INTJ
ajst-16323	41	2	j	j	NOUN
ajst-16323	42	1	x	x	X
ajst-16323	42	2	f	f	PROPN
ajst-16323	42	3	x	x	X
ajst-16323	42	4	f	f	PROPN
ajst-16323	43	1			NOUN
ajst-16323	43	2			NUM
ajst-16323	44	1			ADJ
ajst-16323	45	1			ADJ
ajst-16323	46	1			NOUN
ajst-16323	47	1			VERB
ajst-16323	47	2			PRON
ajst-16323	47	3	(	(	PUNCT
ajst-16323	47	4	3	3	NUM
ajst-16323	47	5	)	)	PUNCT
ajst-16323	47	6	1	1	NUM
ajst-16323	47	7	1	1	NUM
ajst-16323	47	8	0	0	NUM
ajst-16323	47	9	1	1	NUM
ajst-16323	47	10	1	1	NUM
ajst-16323	47	11	m	m	NOUN
ajst-16323	47	12	n	n	PRON
ajst-16323	47	13	j	j	NOUN
ajst-16323	48	1	ij	ij	INTJ
ajst-16323	49	1	i	i	PRON
ajst-16323	49	2	j	j	PROPN
ajst-16323	49	3	n	n	CCONJ
ajst-16323	49	4	n	n	ADV
ajst-16323	50	1	ij	ij	INTJ
ajst-16323	51	1	i	i	INTJ
ajst-16323	51	2	j	j	PROPN
ajst-16323	52	1	y	y	PROPN
ajst-16323	52	2	f	f	PROPN
ajst-16323	52	3	y	y	PROPN
ajst-16323	52	4	f	f	PROPN
ajst-16323	53	1			NUM
ajst-16323	54	1			ADJ
ajst-16323	55	1			ADJ
ajst-16323	56	1			ADJ
ajst-16323	57	1			NOUN
ajst-16323	58	1			VERB
ajst-16323	58	2			X
ajst-16323	58	3	(	(	PUNCT
ajst-16323	58	4	4	4	NUM
ajst-16323	58	5	)	)	PUNCT
ajst-16323	58	6	3.2	3.2	NUM
ajst-16323	58	7	.	.	PUNCT
ajst-16323	59	1	image	image	NOUN
ajst-16323	59	2	preprocessing	preprocessing	NOUN
ajst-16323	59	3	since	since	SCONJ
ajst-16323	59	4	the	the	DET
ajst-16323	59	5	optical	optical	ADJ
ajst-16323	59	6	strip	strip	NOUN
ajst-16323	59	7	area	area	NOUN
ajst-16323	59	8	only	only	ADV
ajst-16323	59	9	accounts	account	VERB
ajst-16323	59	10	for	for	ADP
ajst-16323	59	11	a	a	DET
ajst-16323	59	12	small	small	ADJ
ajst-16323	59	13	part	part	NOUN
ajst-16323	59	14	of	of	ADP
ajst-16323	59	15	the	the	DET
ajst-16323	59	16	entire	entire	ADJ
ajst-16323	59	17	image	image	NOUN
ajst-16323	59	18	area	area	NOUN
ajst-16323	59	19	,	,	PUNCT
ajst-16323	59	20	in	in	ADP
ajst-16323	59	21	order	order	NOUN
ajst-16323	59	22	to	to	PART
ajst-16323	59	23	improve	improve	VERB
ajst-16323	59	24	the	the	DET
ajst-16323	59	25	segmentation	segmentation	NOUN
ajst-16323	59	26	speed	speed	NOUN
ajst-16323	59	27	of	of	ADP
ajst-16323	59	28	the	the	DET
ajst-16323	59	29	deep	deep	ADJ
ajst-16323	59	30	learning	learning	NOUN
ajst-16323	59	31	algorithm	algorithm	NOUN
ajst-16323	59	32	,	,	PUNCT
ajst-16323	59	33	the	the	DET
ajst-16323	59	34	area	area	NOUN
ajst-16323	59	35	of	of	ADP
ajst-16323	59	36	interest	interest	NOUN
ajst-16323	59	37	in	in	ADP
ajst-16323	59	38	the	the	DET
ajst-16323	59	39	weld	weld	NOUN
ajst-16323	59	40	area	area	NOUN
ajst-16323	59	41	is	be	AUX
ajst-16323	59	42	clipped	clip	VERB
ajst-16323	59	43	.	.	PUNCT
ajst-16323	60	1	the	the	DET
ajst-16323	60	2	original	original	ADJ
ajst-16323	60	3	image	image	NOUN
ajst-16323	60	4	is	be	AUX
ajst-16323	60	5	4096	4096	NUM
ajst-16323	60	6	pixel×3000	pixel×3000	NOUN
ajst-16323	60	7	pixel	pixel	NOUN
ajst-16323	60	8	,	,	PUNCT
ajst-16323	60	9	and	and	CCONJ
ajst-16323	60	10	the	the	DET
ajst-16323	60	11	area	area	NOUN
ajst-16323	60	12	where	where	SCONJ
ajst-16323	60	13	the	the	DET
ajst-16323	60	14	optical	optical	ADJ
ajst-16323	60	15	strip	strip	NOUN
ajst-16323	60	16	is	be	AUX
ajst-16323	60	17	located	locate	VERB
ajst-16323	60	18	is	be	AUX
ajst-16323	60	19	clipped	clip	VERB
ajst-16323	60	20	,	,	PUNCT
ajst-16323	60	21	and	and	CCONJ
ajst-16323	60	22	the	the	DET
ajst-16323	60	23	image	image	NOUN
ajst-16323	60	24	after	after	ADP
ajst-16323	60	25	clipping	clip	VERB
ajst-16323	60	26	is	be	AUX
ajst-16323	60	27	1024	1024	NUM
ajst-16323	60	28	pixel×1024	pixel×1024	NOUN
ajst-16323	60	29	pixel	pixel	NOUN
ajst-16323	60	30	.	.	PUNCT
ajst-16323	61	1	the	the	DET
ajst-16323	61	2	cropped	cropped	ADJ
ajst-16323	61	3	image	image	NOUN
ajst-16323	61	4	is	be	AUX
ajst-16323	61	5	then	then	ADV
ajst-16323	61	6	reduced	reduce	VERB
ajst-16323	61	7	to	to	ADP
ajst-16323	61	8	512	512	NUM
ajst-16323	61	9	pixel	pixel	NOUN
ajst-16323	61	10	by	by	ADP
ajst-16323	61	11	512	512	NUM
ajst-16323	61	12	pixel	pixel	NOUN
ajst-16323	61	13	using	use	VERB
ajst-16323	61	14	a	a	DET
ajst-16323	61	15	two	two	NUM
ajst-16323	61	16	-	-	PUNCT
ajst-16323	61	17	line	line	NOUN
ajst-16323	61	18	interpolation	interpolation	NOUN
ajst-16323	61	19	algorithm	algorithm	NOUN
ajst-16323	61	20	.	.	PUNCT
ajst-16323	62	1	3.3	3.3	NUM
ajst-16323	62	2	.	.	PUNCT
ajst-16323	63	1	image	image	NOUN
ajst-16323	63	2	data	datum	NOUN
ajst-16323	63	3	set	set	VERB
ajst-16323	63	4	production	production	NOUN
ajst-16323	63	5	for	for	ADP
ajst-16323	63	6	under	under	NOUN
ajst-16323	63	7	-	-	PUNCT
ajst-16323	63	8	exposure	exposure	NOUN
ajst-16323	63	9	,	,	PUNCT
ajst-16323	63	10	normal	normal	ADJ
ajst-16323	63	11	exposure	exposure	NOUN
ajst-16323	63	12	and	and	CCONJ
ajst-16323	63	13	noise	noise	NOUN
ajst-16323	63	14	interference	interference	NOUN
ajst-16323	63	15	,	,	PUNCT
ajst-16323	63	16	the	the	DET
ajst-16323	63	17	image	image	NOUN
ajst-16323	63	18	data	datum	NOUN
ajst-16323	63	19	sets	set	NOUN
ajst-16323	63	20	were	be	AUX
ajst-16323	63	21	made	make	VERB
ajst-16323	63	22	respectively	respectively	ADV
ajst-16323	63	23	,	,	PUNCT
ajst-16323	63	24	and	and	CCONJ
ajst-16323	63	25	200	200	NUM
ajst-16323	63	26	data	datum	NOUN
ajst-16323	63	27	were	be	AUX
ajst-16323	63	28	made	make	VERB
ajst-16323	63	29	respectively	respectively	ADV
ajst-16323	63	30	for	for	ADP
ajst-16323	63	31	the	the	DET
ajst-16323	63	32	three	three	NUM
ajst-16323	63	33	cases	case	NOUN
ajst-16323	63	34	,	,	PUNCT
ajst-16323	63	35	a	a	DET
ajst-16323	63	36	total	total	NOUN
ajst-16323	63	37	of	of	ADP
ajst-16323	63	38	600	600	NUM
ajst-16323	63	39	data	datum	NOUN
ajst-16323	63	40	.	.	PUNCT
ajst-16323	64	1	the	the	DET
ajst-16323	64	2	training	training	NOUN
ajst-16323	64	3	set	set	NOUN
ajst-16323	64	4	and	and	CCONJ
ajst-16323	64	5	the	the	DET
ajst-16323	64	6	test	test	NOUN
ajst-16323	64	7	set	set	NOUN
ajst-16323	64	8	are	be	AUX
ajst-16323	64	9	divided	divide	VERB
ajst-16323	64	10	according	accord	VERB
ajst-16323	64	11	to	to	ADP
ajst-16323	64	12	9:1	9:1	NUM
ajst-16323	64	13	,	,	PUNCT
ajst-16323	64	14	and	and	CCONJ
ajst-16323	64	15	the	the	DET
ajst-16323	64	16	test	test	NOUN
ajst-16323	64	17	set	set	NOUN
ajst-16323	64	18	contains	contain	VERB
ajst-16323	64	19	60	60	NUM
ajst-16323	64	20	images	image	NOUN
ajst-16323	64	21	,	,	PUNCT
ajst-16323	64	22	20	20	NUM
ajst-16323	64	23	for	for	ADP
ajst-16323	64	24	each	each	PRON
ajst-16323	64	25	of	of	ADP
ajst-16323	64	26	the	the	DET
ajst-16323	64	27	three	three	NUM
ajst-16323	64	28	conditions	condition	NOUN
ajst-16323	64	29	.	.	PUNCT
ajst-16323	65	1	as	as	SCONJ
ajst-16323	65	2	shown	show	VERB
ajst-16323	65	3	in	in	ADP
ajst-16323	65	4	fig.2	fig.2	PROPN
ajst-16323	65	5	.	.	PUNCT
ajst-16323	66	1	the	the	DET
ajst-16323	66	2	data	datum	NOUN
ajst-16323	66	3	set	set	VERB
ajst-16323	66	4	annotation	annotation	NOUN
ajst-16323	66	5	software	software	NOUN
ajst-16323	66	6	uses	use	VERB
ajst-16323	66	7	labelme	labelme	PROPN
ajst-16323	66	8	plug	plug	NOUN
ajst-16323	66	9	-	-	PUNCT
ajst-16323	66	10	in	in	NOUN
ajst-16323	66	11	for	for	ADP
ajst-16323	66	12	annotation	annotation	NOUN
ajst-16323	66	13	to	to	PART
ajst-16323	66	14	realize	realize	VERB
ajst-16323	66	15	pixel	pixel	ADJ
ajst-16323	66	16	-	-	PUNCT
ajst-16323	66	17	level	level	NOUN
ajst-16323	66	18	annotation	annotation	NOUN
ajst-16323	66	19	of	of	ADP
ajst-16323	66	20	the	the	DET
ajst-16323	66	21	light	light	ADJ
ajst-16323	66	22	strip	strip	NOUN
ajst-16323	66	23	image	image	NOUN
ajst-16323	66	24	,	,	PUNCT
ajst-16323	66	25	and	and	CCONJ
ajst-16323	66	26	the	the	DET
ajst-16323	66	27	unlabeled	unlabeled	ADJ
ajst-16323	66	28	area	area	NOUN
ajst-16323	66	29	is	be	AUX
ajst-16323	66	30	the	the	DET
ajst-16323	66	31	background	background	NOUN
ajst-16323	66	32	area	area	NOUN
ajst-16323	66	33	.	.	PUNCT
ajst-16323	67	1	underexposed	underexpose	VERB
ajst-16323	67	2	light	light	PROPN
ajst-16323	67	3	strip	strip	PROPN
ajst-16323	67	4	normal	normal	ADJ
ajst-16323	67	5	strip	strip	NOUN
ajst-16323	67	6	noise	noise	NOUN
ajst-16323	67	7	interferes	interfere	VERB
ajst-16323	67	8	with	with	ADP
ajst-16323	67	9	the	the	DET
ajst-16323	67	10	light	light	ADJ
ajst-16323	67	11	strip	strip	PROPN
ajst-16323	67	12	underexposed	underexpose	VERB
ajst-16323	67	13	light	light	ADJ
ajst-16323	67	14	strip	strip	PROPN
ajst-16323	67	15	label	label	VERB
ajst-16323	67	16	normal	normal	ADJ
ajst-16323	67	17	optical	optical	ADJ
ajst-16323	67	18	strip	strip	NOUN
ajst-16323	67	19	label	label	NOUN
ajst-16323	67	20	noise	noise	NOUN
ajst-16323	67	21	interferes	interfere	VERB
ajst-16323	67	22	with	with	ADP
ajst-16323	67	23	the	the	DET
ajst-16323	67	24	light	light	ADJ
ajst-16323	67	25	strip	strip	NOUN
ajst-16323	67	26	label	label	NOUN
ajst-16323	67	27	(	(	PUNCT
ajst-16323	67	28	a	a	NOUN
ajst-16323	67	29	)	)	PUNCT
ajst-16323	67	30	(	(	PUNCT
ajst-16323	67	31	b	b	X
ajst-16323	67	32	)	)	PUNCT
ajst-16323	67	33	(	(	PUNCT
ajst-16323	67	34	c	c	X
ajst-16323	67	35	)	)	PUNCT
ajst-16323	67	36	figure	figure	NOUN
ajst-16323	67	37	2	2	NUM
ajst-16323	67	38	.	.	PUNCT
ajst-16323	68	1	under	under	ADP
ajst-16323	68	2	exposure	exposure	NOUN
ajst-16323	68	3	,	,	PUNCT
ajst-16323	68	4	normal	normal	ADJ
ajst-16323	68	5	,	,	PUNCT
ajst-16323	68	6	noise	noise	NOUN
ajst-16323	68	7	interference	interference	NOUN
ajst-16323	68	8	light	light	ADJ
ajst-16323	68	9	strips	strip	NOUN
ajst-16323	68	10	and	and	CCONJ
ajst-16323	68	11	labels	label	NOUN
ajst-16323	68	12	:	:	PUNCT
ajst-16323	68	13	(	(	PUNCT
ajst-16323	68	14	a	a	X
ajst-16323	68	15	)	)	PUNCT
ajst-16323	68	16	;	;	PUNCT
ajst-16323	68	17	(	(	PUNCT
ajst-16323	68	18	b	b	X
ajst-16323	68	19	)	)	PUNCT
ajst-16323	68	20	;	;	PUNCT
ajst-16323	68	21	(	(	PUNCT
ajst-16323	68	22	c	c	NOUN
ajst-16323	68	23	)	)	PUNCT
ajst-16323	68	24	.	.	PUNCT
ajst-16323	69	1	3.4	3.4	NUM
ajst-16323	69	2	.	.	PUNCT
ajst-16323	69	3	segmentation	segmentation	NOUN
ajst-16323	69	4	network	network	NOUN
ajst-16323	69	5	optimization	optimization	NOUN
ajst-16323	69	6	and	and	CCONJ
ajst-16323	69	7	metrics	metric	NOUN
ajst-16323	69	8	adaptive	adaptive	ADJ
ajst-16323	69	9	moment	moment	NOUN
ajst-16323	69	10	estimation	estimation	NOUN
ajst-16323	69	11	(	(	PUNCT
ajst-16323	69	12	adam	adam	PROPN
ajst-16323	69	13	)	)	PUNCT
ajst-16323	69	14	is	be	AUX
ajst-16323	69	15	used	use	VERB
ajst-16323	69	16	to	to	PART
ajst-16323	69	17	optimize	optimize	VERB
ajst-16323	69	18	the	the	DET
ajst-16323	69	19	network	network	NOUN
ajst-16323	69	20	model	model	NOUN
ajst-16323	69	21	during	during	ADP
ajst-16323	69	22	the	the	DET
ajst-16323	69	23	training	training	NOUN
ajst-16323	69	24	process	process	NOUN
ajst-16323	69	25	.	.	PUNCT
ajst-16323	70	1	adam	adam	PROPN
ajst-16323	70	2	algorithm	algorithm	PROPN
ajst-16323	70	3	is	be	AUX
ajst-16323	70	4	an	an	DET
ajst-16323	70	5	algorithm	algorithm	NOUN
ajst-16323	70	6	that	that	PRON
ajst-16323	70	7	performs	perform	VERB
ajst-16323	70	8	a	a	DET
ajst-16323	70	9	step	step	NOUN
ajst-16323	70	10	degree	degree	NOUN
ajst-16323	70	11	optimization	optimization	NOUN
ajst-16323	70	12	of	of	ADP
ajst-16323	70	13	random	random	ADJ
ajst-16323	70	14	objective	objective	ADJ
ajst-16323	70	15	functions	function	NOUN
ajst-16323	70	16	,	,	PUNCT
ajst-16323	70	17	and	and	CCONJ
ajst-16323	70	18	the	the	DET
ajst-16323	70	19	learning	learning	NOUN
ajst-16323	70	20	rate	rate	NOUN
ajst-16323	70	21	will	will	AUX
ajst-16323	70	22	change	change	VERB
ajst-16323	70	23	during	during	ADP
ajst-16323	70	24	the	the	DET
ajst-16323	70	25	training	training	NOUN
ajst-16323	70	26	process	process	NOUN
ajst-16323	70	27	.	.	PUNCT
ajst-16323	71	1	adam	adam	PROPN
ajst-16323	71	2	designs	design	VERB
ajst-16323	71	3	independent	independent	ADJ
ajst-16323	71	4	adaptive	adaptive	ADJ
ajst-16323	71	5	learning	learning	NOUN
ajst-16323	71	6	rates	rate	NOUN
ajst-16323	71	7	for	for	ADP
ajst-16323	71	8	different	different	ADJ
ajst-16323	71	9	parameters	parameter	NOUN
ajst-16323	71	10	by	by	ADP
ajst-16323	71	11	calculating	calculate	VERB
ajst-16323	71	12	the	the	DET
ajst-16323	71	13	first	first	ADJ
ajst-16323	71	14	and	and	CCONJ
ajst-16323	71	15	second	second	ADJ
ajst-16323	71	16	moment	moment	NOUN
ajst-16323	71	17	estimates	estimate	NOUN
ajst-16323	71	18	of	of	ADP
ajst-16323	71	19	the	the	DET
ajst-16323	71	20	gradient	gradient	NOUN
ajst-16323	71	21	.	.	PUNCT
ajst-16323	72	1	the	the	DET
ajst-16323	72	2	initial	initial	ADJ
ajst-16323	72	3	learning	learning	NOUN
ajst-16323	72	4	rate	rate	NOUN
ajst-16323	72	5	η=0.0001	η=0.0001	PROPN
ajst-16323	72	6	,	,	PUNCT
ajst-16323	72	7	the	the	DET
ajst-16323	72	8	exponential	exponential	ADJ
ajst-16323	72	9	decay	decay	NOUN
ajst-16323	72	10	rate	rate	NOUN
ajst-16323	72	11	η=0.9	η=0.9	ADJ
ajst-16323	72	12	for	for	ADP
ajst-16323	72	13	the	the	DET
ajst-16323	72	14	first	first	ADJ
ajst-16323	72	15	-	-	PUNCT
ajst-16323	72	16	order	order	NOUN
ajst-16323	72	17	moment	moment	NOUN
ajst-16323	72	18	estimation	estimation	NOUN
ajst-16323	72	19	,	,	PUNCT
ajst-16323	72	20	and	and	CCONJ
ajst-16323	72	21	the	the	DET
ajst-16323	72	22	exponential	exponential	ADJ
ajst-16323	72	23	decay	decay	NOUN
ajst-16323	72	24	rate	rate	NOUN
ajst-16323	72	25	η=0.999	η=0.999	PUNCT
ajst-16323	72	26	for	for	ADP
ajst-16323	72	27	the	the	DET
ajst-16323	72	28	second	second	ADJ
ajst-16323	72	29	-	-	PUNCT
ajst-16323	72	30	order	order	NOUN
ajst-16323	72	31	moment	moment	NOUN
ajst-16323	72	32	estimation	estimation	NOUN
ajst-16323	72	33	.	.	PUNCT
ajst-16323	73	1	4	4	NUM
ajst-16323	73	2	light	light	ADJ
ajst-16323	73	3	strip	strip	NOUN
ajst-16323	73	4	images	image	NOUN
ajst-16323	73	5	were	be	AUX
ajst-16323	73	6	randomly	randomly	ADV
ajst-16323	73	7	selected	select	VERB
ajst-16323	73	8	each	each	DET
ajst-16323	73	9	time	time	NOUN
ajst-16323	73	10	as	as	ADP
ajst-16323	73	11	a	a	DET
ajst-16323	73	12	small	small	ADJ
ajst-16323	73	13	batch	batch	NOUN
ajst-16323	73	14	for	for	ADP
ajst-16323	73	15	training	training	NOUN
ajst-16323	73	16	.	.	PUNCT
ajst-16323	74	1	the	the	DET
ajst-16323	74	2	segmentation	segmentation	NOUN
ajst-16323	74	3	evaluation	evaluation	NOUN
ajst-16323	74	4	index	index	NOUN
ajst-16323	74	5	is	be	AUX
ajst-16323	74	6	dice	dice	NOUN
ajst-16323	74	7	coefficient	coefficient	NOUN
ajst-16323	74	8	,	,	PUNCT
ajst-16323	74	9	and	and	CCONJ
ajst-16323	74	10	its	its	PRON
ajst-16323	74	11	103	103	NUM
ajst-16323	74	12	calculation	calculation	NOUN
ajst-16323	74	13	formula	formula	NOUN
ajst-16323	74	14	is	be	AUX
ajst-16323	74	15	shown	show	VERB
ajst-16323	74	16	in	in	ADP
ajst-16323	74	17	eq.(5	eq.(5	NOUN
ajst-16323	74	18	)	)	PUNCT
ajst-16323	74	19	.	.	PUNCT
ajst-16323	75	1	2	2	NUM
ajst-16323	76	1	|	|	ADV
ajst-16323	76	2	|	|	ADV
ajst-16323	77	1	|	|	INTJ
ajst-16323	78	1	|	|	ADV
ajst-16323	78	2	|	|	ADV
ajst-16323	78	3	|	|	ADV
ajst-16323	78	4	mask	mask	VERB
ajst-16323	79	1	prediction	prediction	NOUN
ajst-16323	79	2	dice	dice	NOUN
ajst-16323	79	3	mask	mask	NOUN
ajst-16323	79	4	prediction	prediction	NOUN
ajst-16323	79	5			NOUN
ajst-16323	79	6			ADJ
ajst-16323	79	7	∩	∩	NOUN
ajst-16323	79	8	(	(	PUNCT
ajst-16323	79	9	5	5	NUM
ajst-16323	79	10	)	)	PUNCT
ajst-16323	79	11	where	where	SCONJ
ajst-16323	79	12	:	:	PUNCT
ajst-16323	79	13	mask	mask	NOUN
ajst-16323	79	14	as	as	ADP
ajst-16323	79	15	label	label	NOUN
ajst-16323	79	16	;	;	PUNCT
ajst-16323	79	17	prediction	prediction	NOUN
ajst-16323	79	18	is	be	AUX
ajst-16323	79	19	the	the	DET
ajst-16323	79	20	result	result	NOUN
ajst-16323	79	21	of	of	ADP
ajst-16323	79	22	prediction	prediction	NOUN
ajst-16323	79	23	.	.	PUNCT
ajst-16323	80	1	dice	dice	NOUN
ajst-16323	80	2	is	be	AUX
ajst-16323	80	3	used	use	VERB
ajst-16323	80	4	to	to	PART
ajst-16323	80	5	describe	describe	VERB
ajst-16323	80	6	the	the	DET
ajst-16323	80	7	similarity	similarity	NOUN
ajst-16323	80	8	between	between	ADP
ajst-16323	80	9	the	the	DET
ajst-16323	80	10	result	result	NOUN
ajst-16323	80	11	of	of	ADP
ajst-16323	80	12	image	image	NOUN
ajst-16323	80	13	segmentation	segmentation	NOUN
ajst-16323	80	14	algorithm	algorithm	NOUN
ajst-16323	80	15	and	and	CCONJ
ajst-16323	80	16	its	its	PRON
ajst-16323	80	17	corresponding	corresponding	ADJ
ajst-16323	80	18	real	real	ADJ
ajst-16323	80	19	defect	defect	ADJ
ajst-16323	80	20	labeling	labeling	NOUN
ajst-16323	80	21	.	.	PUNCT
ajst-16323	81	1	training	training	NOUN
ajst-16323	81	2	with	with	ADP
ajst-16323	81	3	dice	dice	NOUN
ajst-16323	81	4	loss	loss	NOUN
ajst-16323	81	5	function	function	NOUN
ajst-16323	81	6	can	can	AUX
ajst-16323	81	7	achieve	achieve	VERB
ajst-16323	81	8	higher	high	ADJ
ajst-16323	81	9	dice	dice	NOUN
ajst-16323	81	10	value	value	NOUN
ajst-16323	81	11	more	more	ADV
ajst-16323	81	12	intuitively	intuitively	ADV
ajst-16323	81	13	.	.	PUNCT
ajst-16323	82	1	the	the	DET
ajst-16323	82	2	dice	dice	NOUN
ajst-16323	82	3	loss	loss	NOUN
ajst-16323	82	4	function	function	NOUN
ajst-16323	82	5	is	be	AUX
ajst-16323	82	6	shown	show	VERB
ajst-16323	82	7	in	in	ADP
ajst-16323	82	8	eq.(6	eq.(6	NOUN
ajst-16323	82	9	)	)	PUNCT
ajst-16323	82	10	.	.	PUNCT
ajst-16323	83	1	2	2	NUM
ajst-16323	84	1	|	|	ADV
ajst-16323	84	2	|	|	ADV
ajst-16323	84	3	1	1	NUM
ajst-16323	85	1	|	|	ADV
ajst-16323	85	2	|	|	ADV
ajst-16323	85	3	|	|	ADV
ajst-16323	85	4	|	|	ADV
ajst-16323	85	5	mask	mask	VERB
ajst-16323	85	6	prediction	prediction	NOUN
ajst-16323	85	7	dice	dice	NOUN
ajst-16323	85	8	loss	loss	NOUN
ajst-16323	85	9	mask	mask	NOUN
ajst-16323	85	10	prediction	prediction	NOUN
ajst-16323	85	11			PUNCT
ajst-16323	85	12			PROPN
ajst-16323	85	13			PROPN
ajst-16323	85	14			PUNCT
ajst-16323	85	15	∩	∩	NOUN
ajst-16323	85	16	(	(	PUNCT
ajst-16323	85	17	6	6	NUM
ajst-16323	85	18	)	)	SYM
ajst-16323	85	19	3.5	3.5	NUM
ajst-16323	85	20	.	.	PUNCT
ajst-16323	86	1	experimental	experimental	ADJ
ajst-16323	86	2	result	result	VERB
ajst-16323	86	3	the	the	DET
ajst-16323	86	4	images	image	NOUN
ajst-16323	86	5	of	of	ADP
ajst-16323	86	6	under	under	ADV
ajst-16323	86	7	-	-	PUNCT
ajst-16323	86	8	exposed	expose	VERB
ajst-16323	86	9	light	light	ADJ
ajst-16323	86	10	strips	strip	NOUN
ajst-16323	86	11	,	,	PUNCT
ajst-16323	86	12	noise	noise	NOUN
ajst-16323	86	13	-	-	PUNCT
ajst-16323	86	14	interfered	interfere	VERB
ajst-16323	86	15	light	light	ADJ
ajst-16323	86	16	strips	strip	NOUN
ajst-16323	86	17	and	and	CCONJ
ajst-16323	86	18	normal	normal	ADJ
ajst-16323	86	19	light	light	ADJ
ajst-16323	86	20	strips	strip	NOUN
ajst-16323	86	21	are	be	AUX
ajst-16323	86	22	pre	pre	ADJ
ajst-16323	86	23	-	-	VERB
ajst-16323	86	24	segmented	segmented	ADJ
ajst-16323	86	25	.	.	PUNCT
ajst-16323	87	1	the	the	DET
ajst-16323	87	2	segmentation	segmentation	NOUN
ajst-16323	87	3	index	index	NOUN
ajst-16323	87	4	results	result	NOUN
ajst-16323	87	5	are	be	AUX
ajst-16323	87	6	shown	show	VERB
ajst-16323	87	7	in	in	ADP
ajst-16323	87	8	table	table	NOUN
ajst-16323	87	9	1	1	NUM
ajst-16323	87	10	below	below	ADV
ajst-16323	87	11	.	.	PUNCT
ajst-16323	88	1	grayscale	grayscale	NOUN
ajst-16323	88	2	prime	prime	PROPN
ajst-16323	88	3	center	center	NOUN
ajst-16323	88	4	method	method	NOUN
ajst-16323	88	5	was	be	AUX
ajst-16323	88	6	used	use	VERB
ajst-16323	88	7	to	to	PART
ajst-16323	88	8	extract	extract	VERB
ajst-16323	88	9	the	the	DET
ajst-16323	88	10	light	light	ADJ
ajst-16323	88	11	strip	strip	NOUN
ajst-16323	88	12	results	result	NOUN
ajst-16323	88	13	for	for	ADP
ajst-16323	88	14	the	the	DET
ajst-16323	88	15	pre	pre	ADJ
ajst-16323	88	16	-	-	ADJ
ajst-16323	88	17	segmented	segmented	ADJ
ajst-16323	88	18	light	light	NOUN
ajst-16323	88	19	strip	strip	NOUN
ajst-16323	88	20	.	.	PUNCT
ajst-16323	89	1	the	the	DET
ajst-16323	89	2	extraction	extraction	NOUN
ajst-16323	89	3	results	result	NOUN
ajst-16323	89	4	are	be	AUX
ajst-16323	89	5	shown	show	VERB
ajst-16323	89	6	in	in	ADP
ajst-16323	89	7	fig.3	fig.3	PROPN
ajst-16323	89	8	.	.	PUNCT
ajst-16323	90	1	it	it	PRON
ajst-16323	90	2	can	can	AUX
ajst-16323	90	3	be	be	AUX
ajst-16323	90	4	seen	see	VERB
ajst-16323	90	5	that	that	SCONJ
ajst-16323	90	6	the	the	DET
ajst-16323	90	7	method	method	NOUN
ajst-16323	90	8	used	use	VERB
ajst-16323	90	9	in	in	ADP
ajst-16323	90	10	this	this	DET
ajst-16323	90	11	paper	paper	NOUN
ajst-16323	90	12	can	can	AUX
ajst-16323	90	13	realize	realize	VERB
ajst-16323	90	14	the	the	DET
ajst-16323	90	15	extraction	extraction	NOUN
ajst-16323	90	16	of	of	ADP
ajst-16323	90	17	the	the	DET
ajst-16323	90	18	centerline	centerline	NOUN
ajst-16323	90	19	of	of	ADP
ajst-16323	90	20	linear	linear	ADJ
ajst-16323	90	21	structured	structured	ADJ
ajst-16323	90	22	light	light	NOUN
ajst-16323	90	23	under	under	ADP
ajst-16323	90	24	three	three	NUM
ajst-16323	90	25	conditions	condition	NOUN
ajst-16323	90	26	.	.	PUNCT
ajst-16323	91	1	the	the	DET
ajst-16323	91	2	obtained	obtain	VERB
ajst-16323	91	3	light	light	PROPN
ajst-16323	91	4	strip	strip	PROPN
ajst-16323	91	5	center	center	NOUN
ajst-16323	91	6	line	line	NOUN
ajst-16323	91	7	trend	trend	NOUN
ajst-16323	91	8	is	be	AUX
ajst-16323	91	9	basically	basically	ADV
ajst-16323	91	10	the	the	DET
ajst-16323	91	11	same	same	ADJ
ajst-16323	91	12	as	as	ADP
ajst-16323	91	13	the	the	DET
ajst-16323	91	14	light	light	ADJ
ajst-16323	91	15	strip	strip	NOUN
ajst-16323	91	16	trend	trend	NOUN
ajst-16323	91	17	,	,	PUNCT
ajst-16323	91	18	which	which	PRON
ajst-16323	91	19	improves	improve	VERB
ajst-16323	91	20	the	the	DET
ajst-16323	91	21	stability	stability	NOUN
ajst-16323	91	22	of	of	ADP
ajst-16323	91	23	the	the	DET
ajst-16323	91	24	light	light	PROPN
ajst-16323	91	25	strip	strip	PROPN
ajst-16323	91	26	center	center	PROPN
ajst-16323	91	27	extraction	extraction	NOUN
ajst-16323	91	28	.	.	PUNCT
ajst-16323	92	1	table	table	NOUN
ajst-16323	92	2	1	1	NUM
ajst-16323	92	3	.	.	PUNCT
ajst-16323	92	4	segmentation	segmentation	NOUN
ajst-16323	92	5	index	index	NOUN
ajst-16323	92	6	dice	dice	NOUN
ajst-16323	92	7	training	training	NOUN
ajst-16323	92	8	set	set	NOUN
ajst-16323	92	9	dice	dice	NOUN
ajst-16323	92	10	loss	loss	NOUN
ajst-16323	92	11	validation	validation	NOUN
ajst-16323	92	12	set	set	VERB
ajst-16323	92	13	dice	dice	NOUN
ajst-16323	92	14	loss	loss	NOUN
ajst-16323	92	15	98.2	98.2	NUM
ajst-16323	92	16	0.027	0.027	NUM
ajst-16323	92	17	0.027	0.027	NUM
ajst-16323	92	18	underexposed	underexpose	VERB
ajst-16323	92	19	light	light	NOUN
ajst-16323	92	20	strip	strip	PROPN
ajst-16323	92	21	normal	normal	ADJ
ajst-16323	92	22	strip	strip	NOUN
ajst-16323	92	23	noise	noise	NOUN
ajst-16323	92	24	interferes	interfere	VERB
ajst-16323	92	25	with	with	ADP
ajst-16323	92	26	the	the	DET
ajst-16323	92	27	light	light	ADJ
ajst-16323	92	28	strip	strip	NOUN
ajst-16323	92	29	presegmentation	presegmentation	NOUN
ajst-16323	92	30	presegmentation	presegmentation	NOUN
ajst-16323	92	31	presegmentation	presegmentation	NOUN
ajst-16323	92	32	centerline	centerline	NOUN
ajst-16323	92	33	extraction	extraction	NOUN
ajst-16323	92	34	centerline	centerline	NOUN
ajst-16323	92	35	extraction	extraction	NOUN
ajst-16323	92	36	centerline	centerline	NOUN
ajst-16323	92	37	extraction	extraction	NOUN
ajst-16323	92	38	(	(	PUNCT
ajst-16323	92	39	a	a	NOUN
ajst-16323	92	40	)	)	PUNCT
ajst-16323	92	41	(	(	PUNCT
ajst-16323	92	42	b	b	X
ajst-16323	92	43	)	)	PUNCT
ajst-16323	92	44	(	(	PUNCT
ajst-16323	92	45	c	c	X
ajst-16323	92	46	)	)	PUNCT
ajst-16323	92	47	figure	figure	NOUN
ajst-16323	92	48	3	3	NUM
ajst-16323	92	49	.	.	PUNCT
ajst-16323	92	50	underexposure	underexposure	NOUN
ajst-16323	92	51	,	,	PUNCT
ajst-16323	92	52	normal	normal	ADJ
ajst-16323	92	53	,	,	PUNCT
ajst-16323	92	54	noise	noise	NOUN
ajst-16323	92	55	interference	interference	NOUN
ajst-16323	92	56	light	light	NOUN
ajst-16323	92	57	strip	strip	NOUN
ajst-16323	92	58	presegmentation	presegmentation	NOUN
ajst-16323	92	59	and	and	CCONJ
ajst-16323	92	60	center	center	NOUN
ajst-16323	92	61	line	line	NOUN
ajst-16323	92	62	extraction	extraction	NOUN
ajst-16323	92	63	results	result	NOUN
ajst-16323	92	64	:	:	PUNCT
ajst-16323	92	65	(	(	PUNCT
ajst-16323	92	66	a	a	X
ajst-16323	92	67	)	)	PUNCT
ajst-16323	92	68	;	;	PUNCT
ajst-16323	92	69	(	(	PUNCT
ajst-16323	92	70	b	b	X
ajst-16323	92	71	)	)	PUNCT
ajst-16323	92	72	;	;	PUNCT
ajst-16323	92	73	(	(	PUNCT
ajst-16323	92	74	c	c	NOUN
ajst-16323	92	75	)	)	PUNCT
ajst-16323	92	76	.	.	PUNCT
ajst-16323	93	1	4	4	X
ajst-16323	93	2	.	.	X
ajst-16323	93	3	conclusion	conclusion	NOUN
ajst-16323	93	4	(	(	PUNCT
ajst-16323	93	5	1	1	NUM
ajst-16323	93	6	)	)	PUNCT
ajst-16323	93	7	in	in	ADP
ajst-16323	93	8	this	this	DET
ajst-16323	93	9	paper	paper	NOUN
ajst-16323	93	10	,	,	PUNCT
ajst-16323	93	11	the	the	DET
ajst-16323	93	12	deep	deep	ADJ
ajst-16323	93	13	learning	learning	NOUN
ajst-16323	93	14	algorithm	algorithm	NOUN
ajst-16323	93	15	is	be	AUX
ajst-16323	93	16	used	use	VERB
ajst-16323	93	17	to	to	PART
ajst-16323	93	18	presegment	presegment	VERB
ajst-16323	93	19	the	the	DET
ajst-16323	93	20	light	light	ADJ
ajst-16323	93	21	strip	strip	NOUN
ajst-16323	93	22	,	,	PUNCT
ajst-16323	93	23	which	which	PRON
ajst-16323	93	24	greatly	greatly	ADV
ajst-16323	93	25	reduces	reduce	VERB
ajst-16323	93	26	the	the	DET
ajst-16323	93	27	interference	interference	NOUN
ajst-16323	93	28	of	of	ADP
ajst-16323	93	29	noise	noise	NOUN
ajst-16323	93	30	on	on	ADP
ajst-16323	93	31	the	the	DET
ajst-16323	93	32	light	light	ADJ
ajst-16323	93	33	strip	strip	NOUN
ajst-16323	93	34	.	.	PUNCT
ajst-16323	94	1	(	(	PUNCT
ajst-16323	94	2	2	2	X
ajst-16323	94	3	)	)	PUNCT
ajst-16323	94	4	for	for	ADP
ajst-16323	94	5	the	the	DET
ajst-16323	94	6	light	light	ADJ
ajst-16323	94	7	strip	strip	NOUN
ajst-16323	94	8	image	image	NOUN
ajst-16323	94	9	under	under	ADP
ajst-16323	94	10	under	under	NOUN
ajst-16323	94	11	-	-	PUNCT
ajst-16323	94	12	exposure	exposure	NOUN
ajst-16323	94	13	and	and	CCONJ
ajst-16323	94	14	noise	noise	NOUN
ajst-16323	94	15	interference	interference	NOUN
ajst-16323	94	16	,	,	PUNCT
ajst-16323	94	17	the	the	DET
ajst-16323	94	18	center	center	NOUN
ajst-16323	94	19	line	line	NOUN
ajst-16323	94	20	trend	trend	NOUN
ajst-16323	94	21	of	of	ADP
ajst-16323	94	22	the	the	DET
ajst-16323	94	23	light	light	ADJ
ajst-16323	94	24	strip	strip	NOUN
ajst-16323	94	25	is	be	AUX
ajst-16323	94	26	basically	basically	ADV
ajst-16323	94	27	the	the	DET
ajst-16323	94	28	same	same	ADJ
ajst-16323	94	29	as	as	ADP
ajst-16323	94	30	that	that	PRON
ajst-16323	94	31	of	of	ADP
ajst-16323	94	32	the	the	DET
ajst-16323	94	33	light	light	NOUN
ajst-16323	94	34	strip	strip	NOUN
ajst-16323	94	35	,	,	PUNCT
ajst-16323	94	36	and	and	CCONJ
ajst-16323	94	37	the	the	DET
ajst-16323	94	38	stability	stability	NOUN
ajst-16323	94	39	of	of	ADP
ajst-16323	94	40	the	the	DET
ajst-16323	94	41	light	light	PROPN
ajst-16323	94	42	strip	strip	PROPN
ajst-16323	94	43	center	center	NOUN
ajst-16323	94	44	extraction	extraction	NOUN
ajst-16323	94	45	is	be	AUX
ajst-16323	94	46	improved	improve	VERB
ajst-16323	94	47	references	reference	NOUN
ajst-16323	94	48	[	[	X
ajst-16323	94	49	1	1	NUM
ajst-16323	94	50	]	]	PUNCT
ajst-16323	94	51	a.a.al	a.a.al	NOUN
ajst-16323	94	52	-	-	PUNCT
ajst-16323	94	53	temeemy	temeemy	NOUN
ajst-16323	94	54	,	,	PUNCT
ajst-16323	94	55	s.a	s.a	PROPN
ajst-16323	94	56	.	.	PROPN
ajst-16323	94	57	al	al	PROPN
ajst-16323	94	58	-	-	PUNCT
ajst-16323	94	59	saqal	saqal	PROPN
ajst-16323	94	60	laser	laser	NOUN
ajst-16323	94	61	-	-	PUNCT
ajst-16323	94	62	based	base	VERB
ajst-16323	94	63	structured	structured	ADJ
ajst-16323	94	64	light	light	NOUN
ajst-16323	94	65	technique	technique	NOUN
ajst-16323	94	66	for	for	ADP
ajst-16323	94	67	3d	3d	NOUN
ajst-16323	94	68	reconstruction	reconstruction	NOUN
ajst-16323	94	69	using	use	VERB
ajst-16323	94	70	extreme	extreme	ADJ
ajst-16323	94	71	laser	laser	NOUN
ajst-16323	94	72	stripes	stripe	NOUN
ajst-16323	94	73	extraction	extraction	NOUN
ajst-16323	94	74	method	method	NOUN
ajst-16323	94	75	with	with	ADP
ajst-16323	94	76	global	global	ADJ
ajst-16323	94	77	information	information	NOUN
ajst-16323	94	78	extraction	extraction	NOUN
ajst-16323	94	79	opt	opt	NOUN
ajst-16323	94	80	.	.	PUNCT
ajst-16323	95	1	laser	laser	NOUN
ajst-16323	95	2	technol	technol	NOUN
ajst-16323	95	3	.	.	PROPN
ajst-16323	95	4	,	,	PUNCT
ajst-16323	95	5	138	138	NUM
ajst-16323	95	6	(	(	PUNCT
ajst-16323	95	7	2021	2021	NUM
ajst-16323	95	8	)	)	PUNCT
ajst-16323	95	9	,	,	PUNCT
ajst-16323	95	10	article	article	NOUN
ajst-16323	95	11	106897	106897	NUM
ajst-16323	95	12	,	,	PUNCT
ajst-16323	95	13	10.1016	10.1016	NUM
ajst-16323	95	14	/	/	SYM
ajst-16323	95	15	j.optlastec.2020.106897	j.optlastec.2020.106897	PROPN
ajst-16323	96	1	[	[	X
ajst-16323	96	2	2	2	X
ajst-16323	96	3	]	]	PUNCT
ajst-16323	96	4	s.	s.	PROPN
ajst-16323	96	5	pang	pang	PROPN
ajst-16323	96	6	,	,	PUNCT
ajst-16323	96	7	h.	h.	PROPN
ajst-16323	96	8	yang	yang	PROPN
ajst-16323	96	9	,	,	PUNCT
ajst-16323	96	10	an	an	DET
ajst-16323	96	11	algorithm	algorithm	NOUN
ajst-16323	96	12	for	for	ADP
ajst-16323	96	13	extracting	extract	VERB
ajst-16323	96	14	the	the	DET
ajst-16323	96	15	center	center	NOUN
ajst-16323	96	16	of	of	ADP
ajst-16323	96	17	linear	linear	PROPN
ajst-16323	96	18	structured	structure	VERB
ajst-16323	96	19	light	light	ADJ
ajst-16323	96	20	fringe	fringe	NOUN
ajst-16323	96	21	based	base	VERB
ajst-16323	96	22	on	on	ADP
ajst-16323	96	23	directional	directional	ADJ
ajst-16323	96	24	template	template	NOUN
ajst-16323	96	25	,	,	PUNCT
ajst-16323	96	26	in	in	ADP
ajst-16323	96	27	:	:	PUNCT
ajst-16323	96	28	2021	2021	NUM
ajst-16323	96	29	4th	4th	ADJ
ajst-16323	96	30	international	international	ADJ
ajst-16323	96	31	conference	conference	NOUN
ajst-16323	96	32	on	on	ADP
ajst-16323	96	33	advanced	advanced	ADJ
ajst-16323	96	34	electronic	electronic	ADJ
ajst-16323	96	35	materials	material	NOUN
ajst-16323	96	36	,	,	PUNCT
ajst-16323	96	37	computers	computer	NOUN
ajst-16323	96	38	and	and	CCONJ
ajst-16323	96	39	software	software	NOUN
ajst-16323	96	40	engineering	engineering	NOUN
ajst-16323	96	41	(	(	PUNCT
ajst-16323	96	42	aemcse	aemcse	NOUN
ajst-16323	96	43	)	)	PUNCT
ajst-16323	96	44	.	.	PUNCT
ajst-16323	97	1	ieee	ieee	NOUN
ajst-16323	97	2	,	,	PUNCT
ajst-16323	97	3	2021	2021	NUM
ajst-16323	97	4	,	,	PUNCT
ajst-16323	97	5	pp	pp	ADV
ajst-16323	97	6	.	.	PUNCT
ajst-16323	98	1	203	203	NUM
ajst-16323	98	2	-	-	SYM
ajst-16323	98	3	207	207	NUM
ajst-16323	98	4	.	.	PUNCT
ajst-16323	99	1	[	[	X
ajst-16323	99	2	3	3	X
ajst-16323	99	3	]	]	X
ajst-16323	99	4	t.y	t.y	PROPN
ajst-16323	99	5	.	.	PROPN
ajst-16323	99	6	zhang	zhang	PROPN
ajst-16323	99	7	,	,	PUNCT
ajst-16323	99	8	c.y	c.y	PROPN
ajst-16323	99	9	.	.	PROPN
ajst-16323	99	10	suen	suen	PROPN
ajst-16323	99	11	a	a	DET
ajst-16323	99	12	fast	fast	ADJ
ajst-16323	99	13	parallel	parallel	ADJ
ajst-16323	99	14	algorithm	algorithm	NOUN
ajst-16323	99	15	for	for	ADP
ajst-16323	99	16	thinning	thin	VERB
ajst-16323	99	17	digital	digital	ADJ
ajst-16323	99	18	patterns	pattern	NOUN
ajst-16323	99	19	commun	commun	PROPN
ajst-16323	99	20	.	.	PUNCT
ajst-16323	99	21	acm	acm	PROPN
ajst-16323	99	22	,	,	PUNCT
ajst-16323	99	23	27	27	NUM
ajst-16323	99	24	(	(	PUNCT
ajst-16323	99	25	3	3	NUM
ajst-16323	99	26	)	)	PUNCT
ajst-16323	99	27	(	(	PUNCT
ajst-16323	99	28	1984	1984	NUM
ajst-16323	99	29	)	)	PUNCT
ajst-16323	99	30	,	,	PUNCT
ajst-16323	99	31	pp	pp	PROPN
ajst-16323	99	32	.	.	PUNCT
ajst-16323	100	1	236	236	NUM
ajst-16323	100	2	-	-	SYM
ajst-16323	100	3	239	239	NUM
ajst-16323	100	4	.	.	PUNCT
ajst-16323	101	1	[	[	X
ajst-16323	101	2	4	4	X
ajst-16323	101	3	]	]	PUNCT
ajst-16323	101	4	c.	c.	PROPN
ajst-16323	101	5	steger	steger	PROPN
ajst-16323	101	6	an	an	DET
ajst-16323	101	7	unbiased	unbiased	ADJ
ajst-16323	101	8	detector	detector	NOUN
ajst-16323	101	9	of	of	ADP
ajst-16323	101	10	curvilinear	curvilinear	PROPN
ajst-16323	101	11	structures	structure	NOUN
ajst-16323	101	12	ieee	ieee	PROPN
ajst-16323	101	13	trans	tran	NOUN
ajst-16323	101	14	.	.	PUNCT
ajst-16323	102	1	pattern	pattern	PROPN
ajst-16323	102	2	anal	anal	PROPN
ajst-16323	102	3	.	.	PUNCT
ajst-16323	103	1	mach	mach	PROPN
ajst-16323	103	2	.	.	PUNCT
ajst-16323	104	1	intell	intell	PROPN
ajst-16323	104	2	.	.	PUNCT
ajst-16323	105	1	,	,	PUNCT
ajst-16323	105	2	20	20	NUM
ajst-16323	105	3	(	(	PUNCT
ajst-16323	105	4	2	2	NUM
ajst-16323	105	5	)	)	PUNCT
ajst-16323	105	6	(	(	PUNCT
ajst-16323	105	7	1998	1998	NUM
ajst-16323	105	8	)	)	PUNCT
ajst-16323	105	9	,	,	PUNCT
ajst-16323	105	10	pp	pp	PROPN
ajst-16323	105	11	.	.	PUNCT
ajst-16323	106	1	113	113	NUM
ajst-16323	106	2	-	-	SYM
ajst-16323	106	3	125	125	NUM
ajst-16323	106	4	.	.	PUNCT