id	sid	tid	token	lemma	pos
tb09j389k3v	1	1	development	development	NOUN
tb09j389k3v	1	2	in	in	ADP
tb09j389k3v	1	3	deep	deep	ADJ
tb09j389k3v	1	4	learning	learning	NOUN
tb09j389k3v	1	5	algorithms	algorithm	NOUN
tb09j389k3v	1	6	significantly	significantly	ADV
tb09j389k3v	1	7	boosts	boost	VERB
tb09j389k3v	1	8	the	the	DET
tb09j389k3v	1	9	performance	performance	NOUN
tb09j389k3v	1	10	of	of	ADP
tb09j389k3v	1	11	digital	digital	ADJ
tb09j389k3v	1	12	forensics	forensic	NOUN
tb09j389k3v	1	13	applications	application	NOUN
tb09j389k3v	1	14	,	,	PUNCT
tb09j389k3v	1	15	such	such	ADJ
tb09j389k3v	1	16	as	as	ADP
tb09j389k3v	1	17	media	medium	NOUN
tb09j389k3v	1	18	phylogeny	phylogeny	NOUN
tb09j389k3v	1	19	,	,	PUNCT
tb09j389k3v	1	20	optical	optical	ADJ
tb09j389k3v	1	21	character	character	NOUN
tb09j389k3v	1	22	recognition(ocr	recognition(ocr	PROPN
tb09j389k3v	1	23	)	)	PUNCT
tb09j389k3v	1	24	,	,	PUNCT
tb09j389k3v	1	25	face	face	VERB
tb09j389k3v	1	26	recognition	recognition	NOUN
tb09j389k3v	1	27	,	,	PUNCT
tb09j389k3v	1	28	and	and	CCONJ
tb09j389k3v	1	29	scene	scene	NOUN
tb09j389k3v	1	30	text	text	NOUN
tb09j389k3v	1	31	spotting	spot	VERB
tb09j389k3v	1	32	.	.	PUNCT
tb09j389k3v	2	1	as	as	SCONJ
tb09j389k3v	2	2	the	the	DET
tb09j389k3v	2	3	neural	neural	ADJ
tb09j389k3v	2	4	networks	network	NOUN
tb09j389k3v	2	5	achieved	achieve	VERB
tb09j389k3v	2	6	advanced	advanced	ADJ
tb09j389k3v	2	7	results	result	NOUN
tb09j389k3v	2	8	,	,	PUNCT
tb09j389k3v	2	9	the	the	DET
tb09j389k3v	2	10	depth	depth	NOUN
tb09j389k3v	2	11	,	,	PUNCT
tb09j389k3v	2	12	and	and	CCONJ
tb09j389k3v	2	13	width	width	NOUN
tb09j389k3v	2	14	of	of	ADP
tb09j389k3v	2	15	the	the	DET
tb09j389k3v	2	16	architecture	architecture	NOUN
tb09j389k3v	2	17	inevitably	inevitably	ADV
tb09j389k3v	2	18	increase	increase	VERB
tb09j389k3v	2	19	.	.	PUNCT
tb09j389k3v	3	1	neural	neural	ADJ
tb09j389k3v	3	2	networks	network	NOUN
tb09j389k3v	3	3	are	be	AUX
tb09j389k3v	3	4	data	data	NOUN
tb09j389k3v	3	5	-	-	PUNCT
tb09j389k3v	3	6	driven	drive	VERB
tb09j389k3v	3	7	which	which	PRON
tb09j389k3v	3	8	indicates	indicate	VERB
tb09j389k3v	3	9	that	that	SCONJ
tb09j389k3v	3	10	the	the	DET
tb09j389k3v	3	11	more	more	ADV
tb09j389k3v	3	12	complex	complex	ADJ
tb09j389k3v	3	13	a	a	DET
tb09j389k3v	3	14	neural	neural	ADJ
tb09j389k3v	3	15	network	network	NOUN
tb09j389k3v	3	16	is	be	AUX
tb09j389k3v	3	17	,	,	PUNCT
tb09j389k3v	3	18	the	the	PRON
tb09j389k3v	3	19	more	more	ADJ
tb09j389k3v	3	20	data	datum	NOUN
tb09j389k3v	3	21	it	it	PRON
tb09j389k3v	3	22	requires	require	VERB
tb09j389k3v	3	23	to	to	PART
tb09j389k3v	3	24	train	train	VERB
tb09j389k3v	3	25	to	to	PART
tb09j389k3v	3	26	converge	converge	VERB
tb09j389k3v	3	27	.	.	PUNCT
tb09j389k3v	4	1	therefore	therefore	ADV
tb09j389k3v	4	2	it	it	PRON
tb09j389k3v	4	3	is	be	AUX
tb09j389k3v	4	4	of	of	ADP
tb09j389k3v	4	5	great	great	ADJ
tb09j389k3v	4	6	importance	importance	NOUN
tb09j389k3v	4	7	to	to	PART
tb09j389k3v	4	8	generate	generate	VERB
tb09j389k3v	4	9	and	and	CCONJ
tb09j389k3v	4	10	provide	provide	VERB
tb09j389k3v	4	11	enough	enough	ADJ
tb09j389k3v	4	12	training	training	NOUN
tb09j389k3v	4	13	data	datum	NOUN
tb09j389k3v	4	14	with	with	ADP
tb09j389k3v	4	15	good	good	ADJ
tb09j389k3v	4	16	quality	quality	NOUN
tb09j389k3v	4	17	to	to	ADP
tb09j389k3v	4	18	models.in	models.in	NUM
tb09j389k3v	4	19	this	this	DET
tb09j389k3v	4	20	work	work	NOUN
tb09j389k3v	4	21	,	,	PUNCT
tb09j389k3v	4	22	i	i	PRON
tb09j389k3v	4	23	studied	study	VERB
tb09j389k3v	4	24	various	various	ADJ
tb09j389k3v	4	25	approaches	approach	NOUN
tb09j389k3v	4	26	to	to	PART
tb09j389k3v	4	27	collect	collect	VERB
tb09j389k3v	4	28	,	,	PUNCT
tb09j389k3v	4	29	augment	augment	VERB
tb09j389k3v	4	30	and	and	CCONJ
tb09j389k3v	4	31	generate	generate	VERB
tb09j389k3v	4	32	training	training	NOUN
tb09j389k3v	4	33	data	datum	NOUN
tb09j389k3v	4	34	to	to	PART
tb09j389k3v	4	35	help	help	VERB
tb09j389k3v	4	36	to	to	PART
tb09j389k3v	4	37	solve	solve	VERB
tb09j389k3v	4	38	forensics	forensic	NOUN
tb09j389k3v	4	39	tasks	task	NOUN
tb09j389k3v	4	40	using	use	VERB
tb09j389k3v	4	41	deep	deep	ADJ
tb09j389k3v	4	42	learning	learning	NOUN
tb09j389k3v	4	43	algorithms	algorithm	NOUN
tb09j389k3v	4	44	.	.	PUNCT
tb09j389k3v	5	1	i	i	PRON
tb09j389k3v	5	2	firstly	firstly	ADV
tb09j389k3v	5	3	explored	explore	VERB
tb09j389k3v	5	4	traditional	traditional	ADJ
tb09j389k3v	5	5	and	and	CCONJ
tb09j389k3v	5	6	state	state	NOUN
tb09j389k3v	5	7	-	-	PUNCT
tb09j389k3v	5	8	of	of	ADP
tb09j389k3v	5	9	-	-	PUNCT
tb09j389k3v	5	10	the	the	DET
tb09j389k3v	5	11	-	-	PUNCT
tb09j389k3v	5	12	art	art	NOUN
tb09j389k3v	5	13	natural	natural	ADJ
tb09j389k3v	5	14	language	language	NOUN
tb09j389k3v	5	15	processing	processing	NOUN
tb09j389k3v	5	16	algorithms	algorithm	NOUN
tb09j389k3v	5	17	and	and	CCONJ
tb09j389k3v	5	18	applied	apply	VERB
tb09j389k3v	5	19	them	they	PRON
tb09j389k3v	5	20	to	to	ADP
tb09j389k3v	5	21	forensics	forensic	NOUN
tb09j389k3v	5	22	tasks	task	NOUN
tb09j389k3v	5	23	in	in	ADP
tb09j389k3v	5	24	real	real	ADJ
tb09j389k3v	5	25	-	-	PUNCT
tb09j389k3v	5	26	world	world	NOUN
tb09j389k3v	5	27	settings	setting	NOUN
tb09j389k3v	5	28	.	.	PUNCT
tb09j389k3v	6	1	then	then	ADV
tb09j389k3v	6	2	i	i	PRON
tb09j389k3v	6	3	integrated	integrate	VERB
tb09j389k3v	6	4	a	a	DET
tb09j389k3v	6	5	language	language	NOUN
tb09j389k3v	6	6	model	model	NOUN
tb09j389k3v	6	7	into	into	ADP
tb09j389k3v	6	8	a	a	DET
tb09j389k3v	6	9	handwritten	handwritten	ADJ
tb09j389k3v	6	10	text	text	NOUN
tb09j389k3v	6	11	recognition	recognition	NOUN
tb09j389k3v	6	12	task	task	NOUN
tb09j389k3v	6	13	to	to	PART
tb09j389k3v	6	14	correct	correct	ADJ
tb09j389k3v	6	15	texts	text	NOUN
tb09j389k3v	6	16	recognized	recognize	VERB
tb09j389k3v	6	17	by	by	ADP
tb09j389k3v	6	18	ocr	ocr	ADJ
tb09j389k3v	6	19	algorithms	algorithm	NOUN
tb09j389k3v	6	20	.	.	PUNCT
tb09j389k3v	7	1	the	the	DET
tb09j389k3v	7	2	training	training	NOUN
tb09j389k3v	7	3	process	process	NOUN
tb09j389k3v	7	4	is	be	AUX
tb09j389k3v	7	5	benefited	benefit	VERB
tb09j389k3v	7	6	from	from	ADP
tb09j389k3v	7	7	generating	generate	VERB
tb09j389k3v	7	8	synthetic	synthetic	ADJ
tb09j389k3v	7	9	texts	text	NOUN
tb09j389k3v	7	10	with	with	ADP
tb09j389k3v	7	11	spelling	spelling	NOUN
tb09j389k3v	7	12	errors	error	NOUN
tb09j389k3v	7	13	.	.	PUNCT
tb09j389k3v	8	1	moving	move	VERB
tb09j389k3v	8	2	from	from	ADP
tb09j389k3v	8	3	language	language	NOUN
tb09j389k3v	8	4	-	-	PUNCT
tb09j389k3v	8	5	related	relate	VERB
tb09j389k3v	8	6	forensics	forensic	NOUN
tb09j389k3v	8	7	tasks	task	NOUN
tb09j389k3v	8	8	to	to	ADP
tb09j389k3v	8	9	forensics	forensic	NOUN
tb09j389k3v	8	10	in	in	ADP
tb09j389k3v	8	11	the	the	DET
tb09j389k3v	8	12	computer	computer	NOUN
tb09j389k3v	8	13	vision	vision	NOUN
tb09j389k3v	8	14	area	area	NOUN
tb09j389k3v	8	15	,	,	PUNCT
tb09j389k3v	8	16	i	i	PRON
tb09j389k3v	8	17	attempted	attempt	VERB
tb09j389k3v	8	18	to	to	PART
tb09j389k3v	8	19	propose	propose	VERB
tb09j389k3v	8	20	methods	method	NOUN
tb09j389k3v	8	21	to	to	PART
tb09j389k3v	8	22	generate	generate	VERB
tb09j389k3v	8	23	data	datum	NOUN
tb09j389k3v	8	24	that	that	PRON
tb09j389k3v	8	25	can	can	AUX
tb09j389k3v	8	26	help	help	VERB
tb09j389k3v	8	27	the	the	DET
tb09j389k3v	8	28	training	training	NOUN
tb09j389k3v	8	29	process	process	NOUN
tb09j389k3v	8	30	of	of	ADP
tb09j389k3v	8	31	various	various	ADJ
tb09j389k3v	8	32	deep	deep	ADJ
tb09j389k3v	8	33	learning	learning	NOUN
tb09j389k3v	8	34	algorithms	algorithm	NOUN
tb09j389k3v	8	35	.	.	PUNCT
tb09j389k3v	9	1	i	i	PRON
tb09j389k3v	9	2	developed	develop	VERB
tb09j389k3v	9	3	a	a	DET
tb09j389k3v	9	4	3d	3d	NOUN
tb09j389k3v	9	5	engine	engine	NOUN
tb09j389k3v	9	6	-	-	PUNCT
tb09j389k3v	9	7	powered	power	VERB
tb09j389k3v	9	8	virtual	virtual	ADJ
tb09j389k3v	9	9	city	city	NOUN
tb09j389k3v	9	10	generator	generator	NOUN
tb09j389k3v	9	11	for	for	ADP
tb09j389k3v	9	12	synthetic	synthetic	ADJ
tb09j389k3v	9	13	data	datum	NOUN
tb09j389k3v	9	14	collection	collection	NOUN
tb09j389k3v	9	15	.	.	PUNCT
tb09j389k3v	10	1	this	this	DET
tb09j389k3v	10	2	virtual	virtual	ADJ
tb09j389k3v	10	3	city	city	NOUN
tb09j389k3v	10	4	generator	generator	NOUN
tb09j389k3v	10	5	can	can	AUX
tb09j389k3v	10	6	create	create	VERB
tb09j389k3v	10	7	an	an	DET
tb09j389k3v	10	8	unlimited	unlimited	ADJ
tb09j389k3v	10	9	amount	amount	NOUN
tb09j389k3v	10	10	of	of	ADP
tb09j389k3v	10	11	synthetic	synthetic	ADJ
tb09j389k3v	10	12	data	datum	NOUN
tb09j389k3v	10	13	for	for	ADP
tb09j389k3v	10	14	various	various	ADJ
tb09j389k3v	10	15	city	city	NOUN
tb09j389k3v	10	16	scene	scene	NOUN
tb09j389k3v	10	17	-	-	PUNCT
tb09j389k3v	10	18	related	relate	VERB
tb09j389k3v	10	19	computer	computer	NOUN
tb09j389k3v	10	20	vision	vision	NOUN
tb09j389k3v	10	21	tasks	task	NOUN
tb09j389k3v	10	22	.	.	PUNCT
tb09j389k3v	11	1	i	i	PRON
tb09j389k3v	11	2	then	then	ADV
tb09j389k3v	11	3	applied	apply	VERB
tb09j389k3v	11	4	it	it	PRON
tb09j389k3v	11	5	to	to	ADP
tb09j389k3v	11	6	a	a	DET
tb09j389k3v	11	7	forensics	forensic	NOUN
tb09j389k3v	11	8	study	study	NOUN
tb09j389k3v	11	9	relates	relate	VERB
tb09j389k3v	11	10	to	to	ADP
tb09j389k3v	11	11	the	the	DET
tb09j389k3v	11	12	human	human	ADJ
tb09j389k3v	11	13	perception	perception	NOUN
tb09j389k3v	11	14	of	of	ADP
tb09j389k3v	11	15	synthetic	synthetic	ADJ
tb09j389k3v	11	16	faces	face	NOUN
tb09j389k3v	11	17	.	.	PUNCT
tb09j389k3v	12	1	finally	finally	ADV
tb09j389k3v	12	2	,	,	PUNCT
tb09j389k3v	12	3	i	i	PRON
tb09j389k3v	12	4	proposed	propose	VERB
tb09j389k3v	12	5	a	a	DET
tb09j389k3v	12	6	method	method	NOUN
tb09j389k3v	12	7	to	to	PART
tb09j389k3v	12	8	solve	solve	VERB
tb09j389k3v	12	9	the	the	DET
tb09j389k3v	12	10	arbitrarily	arbitrarily	ADV
tb09j389k3v	12	11	shaped	shape	VERB
tb09j389k3v	12	12	scene	scene	NOUN
tb09j389k3v	12	13	text	text	NOUN
tb09j389k3v	12	14	detection	detection	NOUN
tb09j389k3v	12	15	problem	problem	NOUN
tb09j389k3v	12	16	.	.	PUNCT
tb09j389k3v	13	1	scene	scene	NOUN
tb09j389k3v	13	2	texts	text	NOUN
tb09j389k3v	13	3	provide	provide	VERB
tb09j389k3v	13	4	important	important	ADJ
tb09j389k3v	13	5	forensics	forensic	NOUN
tb09j389k3v	13	6	information	information	NOUN
tb09j389k3v	13	7	and	and	CCONJ
tb09j389k3v	13	8	how	how	SCONJ
tb09j389k3v	13	9	to	to	PART
tb09j389k3v	13	10	extract	extract	VERB
tb09j389k3v	13	11	them	they	PRON
tb09j389k3v	13	12	accurately	accurately	ADV
tb09j389k3v	13	13	has	have	AUX
tb09j389k3v	13	14	been	be	AUX
tb09j389k3v	13	15	an	an	DET
tb09j389k3v	13	16	active	active	ADJ
tb09j389k3v	13	17	research	research	NOUN
tb09j389k3v	13	18	area	area	NOUN
tb09j389k3v	13	19	for	for	ADP
tb09j389k3v	13	20	years	year	NOUN
tb09j389k3v	13	21	.	.	PUNCT
tb09j389k3v	14	1	i	i	PRON
tb09j389k3v	14	2	proposed	propose	VERB
tb09j389k3v	14	3	a	a	DET
tb09j389k3v	14	4	modified	modify	VERB
tb09j389k3v	14	5	loss	loss	NOUN
tb09j389k3v	14	6	function	function	NOUN
tb09j389k3v	14	7	that	that	PRON
tb09j389k3v	14	8	can	can	AUX
tb09j389k3v	14	9	help	help	VERB
tb09j389k3v	14	10	the	the	DET
tb09j389k3v	14	11	regression	regression	NOUN
tb09j389k3v	14	12	of	of	ADP
tb09j389k3v	14	13	polygon	polygon	NOUN
tb09j389k3v	14	14	shapes	shape	NOUN
tb09j389k3v	14	15	which	which	PRON
tb09j389k3v	14	16	significantly	significantly	ADV
tb09j389k3v	14	17	simplifies	simplify	VERB
tb09j389k3v	14	18	the	the	DET
tb09j389k3v	14	19	model	model	NOUN
tb09j389k3v	14	20	architecture	architecture	NOUN
tb09j389k3v	14	21	and	and	CCONJ
tb09j389k3v	14	22	expedite	expedite	VERB
tb09j389k3v	14	23	the	the	DET
tb09j389k3v	14	24	training	training	NOUN
tb09j389k3v	14	25	process	process	NOUN
tb09j389k3v	14	26	.	.	PUNCT
tb09j389k3v	15	1	therefore	therefore	ADV
tb09j389k3v	15	2	we	we	PRON
tb09j389k3v	15	3	can	can	AUX
tb09j389k3v	15	4	achieve	achieve	VERB
tb09j389k3v	15	5	arbitrarily	arbitrarily	ADV
tb09j389k3v	15	6	shaped	shape	VERB
tb09j389k3v	15	7	scene	scene	NOUN
tb09j389k3v	15	8	text	text	NOUN
tb09j389k3v	15	9	detection	detection	NOUN
tb09j389k3v	15	10	with	with	ADP
tb09j389k3v	15	11	a	a	DET
tb09j389k3v	15	12	comparable	comparable	ADJ
tb09j389k3v	15	13	performance	performance	NOUN
tb09j389k3v	15	14	of	of	ADP
tb09j389k3v	15	15	state	state	NOUN
tb09j389k3v	15	16	-	-	PUNCT
tb09j389k3v	15	17	of	of	ADP
tb09j389k3v	15	18	-	-	PUNCT
tb09j389k3v	15	19	the	the	DET
tb09j389k3v	15	20	-	-	PUNCT
tb09j389k3v	15	21	art	art	NOUN
tb09j389k3v	15	22	algorithms	algorithm	NOUN
tb09j389k3v	15	23	at	at	ADP
tb09j389k3v	15	24	a	a	DET
tb09j389k3v	15	25	real	real	ADJ
tb09j389k3v	15	26	-	-	PUNCT
tb09j389k3v	15	27	time	time	NOUN
tb09j389k3v	15	28	speed	speed	NOUN
tb09j389k3v	15	29	.	.	PUNCT
tb09j389k3v	16	1	the	the	DET
tb09j389k3v	16	2	training	training	NOUN
tb09j389k3v	16	3	process	process	NOUN
tb09j389k3v	16	4	is	be	AUX
tb09j389k3v	16	5	supported	support	VERB
tb09j389k3v	16	6	by	by	ADP
tb09j389k3v	16	7	a	a	DET
tb09j389k3v	16	8	new	new	ADJ
tb09j389k3v	16	9	synthetic	synthetic	ADJ
tb09j389k3v	16	10	data	datum	NOUN
tb09j389k3v	16	11	generation	generation	NOUN
tb09j389k3v	16	12	approach	approach	NOUN
tb09j389k3v	16	13	that	that	PRON
tb09j389k3v	16	14	focused	focus	VERB
tb09j389k3v	16	15	on	on	ADP
tb09j389k3v	16	16	curved	curved	ADJ
tb09j389k3v	16	17	and	and	CCONJ
tb09j389k3v	16	18	irregular	irregular	ADJ
tb09j389k3v	16	19	-	-	PUNCT
tb09j389k3v	16	20	shaped	shape	VERB
tb09j389k3v	16	21	text	text	NOUN
tb09j389k3v	16	22	instances	instance	NOUN
tb09j389k3v	16	23	.	.	PUNCT
tb09j389k3v	17	1	in	in	ADP
tb09j389k3v	17	2	the	the	DET
tb09j389k3v	17	3	end	end	NOUN
tb09j389k3v	17	4	,	,	PUNCT
tb09j389k3v	17	5	i	i	PRON
tb09j389k3v	17	6	created	create	VERB
tb09j389k3v	17	7	a	a	DET
tb09j389k3v	17	8	new	new	ADJ
tb09j389k3v	17	9	large	large	ADJ
tb09j389k3v	17	10	-	-	PUNCT
tb09j389k3v	17	11	scale	scale	NOUN
tb09j389k3v	17	12	synthetic	synthetic	ADJ
tb09j389k3v	17	13	scene	scene	NOUN
tb09j389k3v	17	14	texts	text	NOUN
tb09j389k3v	17	15	dataset	dataset	NOUN
tb09j389k3v	17	16	that	that	PRON
tb09j389k3v	17	17	contributes	contribute	VERB
tb09j389k3v	17	18	to	to	ADP
tb09j389k3v	17	19	the	the	DET
tb09j389k3v	17	20	research	research	NOUN
tb09j389k3v	17	21	community	community	NOUN
tb09j389k3v	17	22	.	.	PUNCT