id	sid	tid	token	lemma	pos
w6634171t41	1	1	we	we	PRON
w6634171t41	1	2	highlight	highlight	VERB
w6634171t41	1	3	the	the	DET
w6634171t41	1	4	contributions	contribution	NOUN
w6634171t41	1	5	described	describe	VERB
w6634171t41	1	6	in	in	ADP
w6634171t41	1	7	this	this	DET
w6634171t41	1	8	dissertation	dissertation	NOUN
w6634171t41	1	9	and	and	CCONJ
w6634171t41	1	10	envision	envision	NOUN
w6634171t41	1	11	the	the	DET
w6634171t41	1	12	future	future	ADJ
w6634171t41	1	13	works	work	NOUN
w6634171t41	1	14	.	.	PUNCT
w6634171t41	2	1	first	first	ADV
w6634171t41	2	2	,	,	PUNCT
w6634171t41	2	3	a	a	DET
w6634171t41	2	4	comprehensive	comprehensive	ADJ
w6634171t41	2	5	background	background	NOUN
w6634171t41	2	6	,	,	PUNCT
w6634171t41	2	7	related	relate	VERB
w6634171t41	2	8	works	work	NOUN
w6634171t41	2	9	,	,	PUNCT
w6634171t41	2	10	and	and	CCONJ
w6634171t41	2	11	dataset	dataset	NOUN
w6634171t41	2	12	information	information	NOUN
w6634171t41	2	13	are	be	AUX
w6634171t41	2	14	given	give	VERB
w6634171t41	2	15	at	at	ADP
w6634171t41	2	16	the	the	DET
w6634171t41	2	17	beginning	beginning	NOUN
w6634171t41	2	18	of	of	ADP
w6634171t41	2	19	this	this	DET
w6634171t41	2	20	dissertation	dissertation	NOUN
w6634171t41	2	21	.	.	PUNCT
w6634171t41	3	1	the	the	DET
w6634171t41	3	2	latter	latter	ADJ
w6634171t41	3	3	part	part	NOUN
w6634171t41	3	4	of	of	ADP
w6634171t41	3	5	the	the	DET
w6634171t41	3	6	dissertation	dissertation	NOUN
w6634171t41	3	7	focuses	focus	VERB
w6634171t41	3	8	on	on	ADP
w6634171t41	3	9	researching	research	VERB
w6634171t41	3	10	the	the	DET
w6634171t41	3	11	lrfr	lrfr	ADJ
w6634171t41	3	12	task	task	NOUN
w6634171t41	3	13	from	from	ADP
w6634171t41	3	14	several	several	ADJ
w6634171t41	3	15	aspects	aspect	NOUN
w6634171t41	3	16	.	.	PUNCT
w6634171t41	4	1	first	first	ADV
w6634171t41	4	2	,	,	PUNCT
w6634171t41	4	3	we	we	PRON
w6634171t41	4	4	curate	curate	VERB
w6634171t41	4	5	two	two	NUM
w6634171t41	4	6	face	face	NOUN
w6634171t41	4	7	datasets	dataset	NOUN
w6634171t41	4	8	,	,	PUNCT
w6634171t41	4	9	the	the	DET
w6634171t41	4	10	vbolo	vbolo	PROPN
w6634171t41	4	11	dataset	dataset	NOUN
w6634171t41	4	12	which	which	PRON
w6634171t41	4	13	was	be	AUX
w6634171t41	4	14	collected	collect	VERB
w6634171t41	4	15	from	from	ADP
w6634171t41	4	16	a	a	DET
w6634171t41	4	17	real	real	ADJ
w6634171t41	4	18	surveillance	surveillance	NOUN
w6634171t41	4	19	camera	camera	NOUN
w6634171t41	4	20	network	network	NOUN
w6634171t41	4	21	,	,	PUNCT
w6634171t41	4	22	and	and	CCONJ
w6634171t41	4	23	the	the	DET
w6634171t41	4	24	ndfacereid	ndfacereid	PROPN
w6634171t41	4	25	dataset	dataset	NOUN
w6634171t41	4	26	which	which	PRON
w6634171t41	4	27	was	be	AUX
w6634171t41	4	28	collected	collect	VERB
w6634171t41	4	29	under	under	ADP
w6634171t41	4	30	a	a	DET
w6634171t41	4	31	near	near	ADJ
w6634171t41	4	32	-	-	PUNCT
w6634171t41	4	33	surveillance	surveillance	NOUN
w6634171t41	4	34	environment	environment	NOUN
w6634171t41	4	35	.	.	PUNCT
w6634171t41	5	1	second	second	ADV
w6634171t41	5	2	,	,	PUNCT
w6634171t41	5	3	we	we	PRON
w6634171t41	5	4	employ	employ	VERB
w6634171t41	5	5	deep	deep	ADJ
w6634171t41	5	6	neural	neural	ADJ
w6634171t41	5	7	networks	network	NOUN
w6634171t41	5	8	and	and	CCONJ
w6634171t41	5	9	research	research	NOUN
w6634171t41	5	10	on	on	ADP
w6634171t41	5	11	two	two	NUM
w6634171t41	5	12	major	major	ADJ
w6634171t41	5	13	face	face	NOUN
w6634171t41	5	14	recognition	recognition	NOUN
w6634171t41	5	15	tasks	task	NOUN
w6634171t41	5	16	in	in	ADP
w6634171t41	5	17	digital	digital	ADJ
w6634171t41	5	18	forensics	forensic	NOUN
w6634171t41	5	19	,	,	PUNCT
w6634171t41	5	20	face	face	NOUN
w6634171t41	5	21	re	re	NOUN
w6634171t41	5	22	-	-	NOUN
w6634171t41	5	23	identification	identification	NOUN
w6634171t41	5	24	,	,	PUNCT
w6634171t41	5	25	and	and	CCONJ
w6634171t41	5	26	face	face	NOUN
w6634171t41	5	27	identification	identification	NOUN
w6634171t41	5	28	.	.	PUNCT
w6634171t41	6	1	for	for	ADP
w6634171t41	6	2	each	each	DET
w6634171t41	6	3	task	task	NOUN
w6634171t41	6	4	,	,	PUNCT
w6634171t41	6	5	we	we	PRON
w6634171t41	6	6	propose	propose	VERB
w6634171t41	6	7	new	new	ADJ
w6634171t41	6	8	deep	deep	ADJ
w6634171t41	6	9	architectures	architecture	NOUN
w6634171t41	6	10	and	and	CCONJ
w6634171t41	6	11	training	training	NOUN
w6634171t41	6	12	strategies	strategy	NOUN
w6634171t41	6	13	to	to	PART
w6634171t41	6	14	improve	improve	VERB
w6634171t41	6	15	the	the	DET
w6634171t41	6	16	performance	performance	NOUN
w6634171t41	6	17	of	of	ADP
w6634171t41	6	18	the	the	DET
w6634171t41	6	19	real	real	ADJ
w6634171t41	6	20	surveillance	surveillance	NOUN
w6634171t41	6	21	datasets	dataset	NOUN
w6634171t41	6	22	.	.	PUNCT
w6634171t41	7	1	we	we	PRON
w6634171t41	7	2	also	also	ADV
w6634171t41	7	3	conduct	conduct	VERB
w6634171t41	7	4	studies	study	NOUN
w6634171t41	7	5	on	on	ADP
w6634171t41	7	6	the	the	DET
w6634171t41	7	7	capacity	capacity	NOUN
w6634171t41	7	8	of	of	ADP
w6634171t41	7	9	the	the	DET
w6634171t41	7	10	deep	deep	ADJ
w6634171t41	7	11	models	model	NOUN
w6634171t41	7	12	on	on	ADP
w6634171t41	7	13	the	the	DET
w6634171t41	7	14	lq	lq	ADJ
w6634171t41	7	15	face	face	NOUN
w6634171t41	7	16	images	image	NOUN
w6634171t41	7	17	from	from	ADP
w6634171t41	7	18	these	these	DET
w6634171t41	7	19	datasets	dataset	NOUN
w6634171t41	7	20	utilizing	utilize	VERB
w6634171t41	7	21	an	an	DET
w6634171t41	7	22	efficient	efficient	ADJ
w6634171t41	7	23	model	model	NOUN
w6634171t41	7	24	searching	search	VERB
w6634171t41	7	25	engine	engine	NOUN
w6634171t41	7	26	to	to	PART
w6634171t41	7	27	get	get	VERB
w6634171t41	7	28	the	the	DET
w6634171t41	7	29	insight	insight	NOUN
w6634171t41	7	30	on	on	ADP
w6634171t41	7	31	performance	performance	NOUN
w6634171t41	7	32	change	change	NOUN
w6634171t41	7	33	as	as	SCONJ
w6634171t41	7	34	gradually	gradually	ADV
w6634171t41	7	35	change	change	VERB
w6634171t41	7	36	the	the	DET
w6634171t41	7	37	complexity	complexity	NOUN
w6634171t41	7	38	of	of	ADP
w6634171t41	7	39	the	the	DET
w6634171t41	7	40	deep	deep	ADJ
w6634171t41	7	41	network	network	NOUN
w6634171t41	7	42	and	and	CCONJ
w6634171t41	7	43	the	the	DET
w6634171t41	7	44	input	input	NOUN
w6634171t41	7	45	size	size	NOUN
w6634171t41	7	46	.	.	PUNCT
w6634171t41	8	1	third	third	ADV
w6634171t41	8	2	,	,	PUNCT
w6634171t41	8	3	considering	consider	VERB
w6634171t41	8	4	the	the	DET
w6634171t41	8	5	fact	fact	NOUN
w6634171t41	8	6	of	of	ADP
w6634171t41	8	7	lacking	lack	VERB
w6634171t41	8	8	training	training	NOUN
w6634171t41	8	9	data	datum	NOUN
w6634171t41	8	10	at	at	ADP
w6634171t41	8	11	scale	scale	NOUN
w6634171t41	8	12	,	,	PUNCT
w6634171t41	8	13	we	we	PRON
w6634171t41	8	14	develop	develop	VERB
w6634171t41	8	15	a	a	DET
w6634171t41	8	16	practical	practical	ADJ
w6634171t41	8	17	way	way	NOUN
w6634171t41	8	18	to	to	PART
w6634171t41	8	19	solve	solve	VERB
w6634171t41	8	20	the	the	DET
w6634171t41	8	21	face	face	NOUN
w6634171t41	8	22	recognition	recognition	NOUN
w6634171t41	8	23	task	task	NOUN
w6634171t41	8	24	under	under	ADP
w6634171t41	8	25	this	this	DET
w6634171t41	8	26	consideration	consideration	NOUN
w6634171t41	8	27	.	.	PUNCT
w6634171t41	9	1	we	we	PRON
w6634171t41	9	2	address	address	VERB
w6634171t41	9	3	weakly	weakly	ADV
w6634171t41	9	4	-	-	PUNCT
w6634171t41	9	5	supervised	supervised	ADJ
w6634171t41	9	6	lrfr	lrfr	NOUN
w6634171t41	9	7	in	in	ADP
w6634171t41	9	8	videos	video	NOUN
w6634171t41	9	9	by	by	ADP
w6634171t41	9	10	incorporating	incorporate	VERB
w6634171t41	9	11	pre	pre	ADJ
w6634171t41	9	12	-	-	ADJ
w6634171t41	9	13	trained	train	VERB
w6634171t41	9	14	general	general	ADJ
w6634171t41	9	15	face	face	NOUN
w6634171t41	9	16	recognition	recognition	NOUN
w6634171t41	9	17	models	model	NOUN
w6634171t41	9	18	as	as	ADP
w6634171t41	9	19	prior	prior	ADV
w6634171t41	9	20	and	and	CCONJ
w6634171t41	9	21	propose	propose	VERB
w6634171t41	9	22	an	an	DET
w6634171t41	9	23	approach	approach	NOUN
w6634171t41	9	24	to	to	PART
w6634171t41	9	25	gradually	gradually	ADV
w6634171t41	9	26	adapting	adapt	VERB
w6634171t41	9	27	a	a	DET
w6634171t41	9	28	general	general	ADJ
w6634171t41	9	29	face	face	NOUN
w6634171t41	9	30	recognition	recognition	NOUN
w6634171t41	9	31	model	model	NOUN
w6634171t41	9	32	into	into	ADP
w6634171t41	9	33	a	a	DET
w6634171t41	9	34	more	more	ADV
w6634171t41	9	35	robust	robust	ADJ
w6634171t41	9	36	lq	lq	ADP
w6634171t41	9	37	face	face	NOUN
w6634171t41	9	38	recognition	recognition	NOUN
w6634171t41	9	39	model	model	NOUN
w6634171t41	9	40	.	.	PUNCT
w6634171t41	10	1	finally	finally	ADV
w6634171t41	10	2	,	,	PUNCT
w6634171t41	10	3	we	we	PRON
w6634171t41	10	4	explore	explore	VERB
w6634171t41	10	5	the	the	DET
w6634171t41	10	6	face	face	NOUN
w6634171t41	10	7	redaction	redaction	NOUN
w6634171t41	10	8	task	task	NOUN
w6634171t41	10	9	in	in	ADP
w6634171t41	10	10	surveillance	surveillance	NOUN
w6634171t41	10	11	quality	quality	NOUN
w6634171t41	10	12	videos	video	NOUN
w6634171t41	10	13	.	.	PUNCT
w6634171t41	11	1	we	we	PRON
w6634171t41	11	2	address	address	VERB
w6634171t41	11	3	the	the	DET
w6634171t41	11	4	challenges	challenge	NOUN
w6634171t41	11	5	compared	compare	VERB
w6634171t41	11	6	to	to	ADP
w6634171t41	11	7	a	a	DET
w6634171t41	11	8	redaction	redaction	NOUN
w6634171t41	11	9	system	system	NOUN
w6634171t41	11	10	for	for	ADP
w6634171t41	11	11	hq	hq	NOUN
w6634171t41	11	12	face	face	NOUN
w6634171t41	11	13	images	image	NOUN
w6634171t41	11	14	in	in	ADP
w6634171t41	11	15	video	video	NOUN
w6634171t41	11	16	frames	frame	NOUN
w6634171t41	11	17	.	.	PUNCT
w6634171t41	12	1	by	by	ADP
w6634171t41	12	2	employing	employ	VERB
w6634171t41	12	3	a	a	DET
w6634171t41	12	4	state	state	NOUN
w6634171t41	12	5	-	-	PUNCT
w6634171t41	12	6	of	of	ADP
w6634171t41	12	7	-	-	PUNCT
w6634171t41	12	8	the	the	DET
w6634171t41	12	9	-	-	PUNCT
w6634171t41	12	10	art	art	NOUN
w6634171t41	12	11	stacked	stack	VERB
w6634171t41	12	12	hour	hour	NOUN
w6634171t41	12	13	-	-	PUNCT
w6634171t41	12	14	glass	glass	NOUN
w6634171t41	12	15	as	as	ADV
w6634171t41	12	16	well	well	ADV
w6634171t41	12	17	as	as	ADP
w6634171t41	12	18	intermediate	intermediate	ADJ
w6634171t41	12	19	supervision	supervision	NOUN
w6634171t41	12	20	during	during	ADP
w6634171t41	12	21	training	training	NOUN
w6634171t41	12	22	,	,	PUNCT
w6634171t41	12	23	we	we	PRON
w6634171t41	12	24	are	be	AUX
w6634171t41	12	25	able	able	ADJ
w6634171t41	12	26	to	to	PART
w6634171t41	12	27	evaluate	evaluate	VERB
w6634171t41	12	28	the	the	DET
w6634171t41	12	29	performance	performance	NOUN
w6634171t41	12	30	of	of	ADP
w6634171t41	12	31	a	a	DET
w6634171t41	12	32	face	face	NOUN
w6634171t41	12	33	redaction	redaction	NOUN
w6634171t41	12	34	system	system	NOUN
w6634171t41	12	35	with	with	ADP
w6634171t41	12	36	very	very	ADV
w6634171t41	12	37	lq	lq	NOUN
w6634171t41	12	38	surveillance	surveillance	NOUN
w6634171t41	12	39	videos	video	NOUN
w6634171t41	12	40	.	.	PUNCT
w6634171t41	13	1	by	by	ADP
w6634171t41	13	2	providing	provide	VERB
w6634171t41	13	3	a	a	DET
w6634171t41	13	4	qualitative	qualitative	ADJ
w6634171t41	13	5	result	result	NOUN
w6634171t41	13	6	,	,	PUNCT
w6634171t41	13	7	we	we	PRON
w6634171t41	13	8	are	be	AUX
w6634171t41	13	9	able	able	ADJ
w6634171t41	13	10	to	to	PART
w6634171t41	13	11	foresee	foresee	VERB
w6634171t41	13	12	the	the	DET
w6634171t41	13	13	potential	potential	ADJ
w6634171t41	13	14	directions	direction	NOUN
w6634171t41	13	15	to	to	PART
w6634171t41	13	16	improve	improve	VERB
w6634171t41	13	17	the	the	DET
w6634171t41	13	18	existing	exist	VERB
w6634171t41	13	19	method	method	NOUN
w6634171t41	13	20	.	.	PUNCT
w6634171t41	14	1	first	first	ADV
w6634171t41	14	2	,	,	PUNCT
w6634171t41	14	3	a	a	DET
w6634171t41	14	4	more	more	ADV
w6634171t41	14	5	reliable	reliable	ADJ
w6634171t41	14	6	facial	facial	ADJ
w6634171t41	14	7	keypoint	keypoint	NOUN
w6634171t41	14	8	detector	detector	NOUN
w6634171t41	14	9	needs	need	VERB
w6634171t41	14	10	to	to	PART
w6634171t41	14	11	specifically	specifically	ADV
w6634171t41	14	12	curated	curate	VERB
w6634171t41	14	13	for	for	ADP
w6634171t41	14	14	lq	lq	INTJ
w6634171t41	14	15	face	face	NOUN
w6634171t41	14	16	images	image	NOUN
w6634171t41	14	17	,	,	PUNCT
w6634171t41	14	18	contour	contour	NOUN
w6634171t41	14	19	or	or	CCONJ
w6634171t41	14	20	template	template	NOUN
w6634171t41	14	21	matching	matching	NOUN
w6634171t41	14	22	could	could	AUX
w6634171t41	14	23	be	be	AUX
w6634171t41	14	24	applied	apply	VERB
w6634171t41	14	25	when	when	SCONJ
w6634171t41	14	26	not	not	PART
w6634171t41	14	27	detailed	detailed	ADJ
w6634171t41	14	28	information	information	NOUN
w6634171t41	14	29	on	on	ADP
w6634171t41	14	30	the	the	DET
w6634171t41	14	31	face	face	NOUN
w6634171t41	14	32	could	could	AUX
w6634171t41	14	33	be	be	AUX
w6634171t41	14	34	catching	catch	VERB
w6634171t41	14	35	due	due	ADP
w6634171t41	14	36	to	to	ADP
w6634171t41	14	37	the	the	DET
w6634171t41	14	38	image	image	NOUN
w6634171t41	14	39	degradation	degradation	NOUN
w6634171t41	14	40	.	.	PUNCT
w6634171t41	15	1	this	this	PRON
w6634171t41	15	2	could	could	AUX
w6634171t41	15	3	be	be	AUX
w6634171t41	15	4	fused	fuse	VERB
w6634171t41	15	5	with	with	ADP
w6634171t41	15	6	the	the	DET
w6634171t41	15	7	sota	sota	PROPN
w6634171t41	15	8	face	face	PROPN
w6634171t41	15	9	landmark	landmark	PROPN
w6634171t41	15	10	detectors	detectors	PROPN
w6634171t41	15	11	.	.	PUNCT