id	sid	tid	token	lemma	pos
easat-8918	1	1	edelweiss	edelweiss	PROPN
easat-8918	1	2	applied	apply	VERB
easat-8918	1	3	science	science	NOUN
easat-8918	1	4	and	and	CCONJ
easat-8918	1	5	technology	technology	NOUN
easat-8918	1	6	issn	issn	PROPN
easat-8918	1	7	:	:	PUNCT
easat-8918	1	8	2576	2576	NUM
easat-8918	1	9	-	-	SYM
easat-8918	1	10	8484	8484	NUM
easat-8918	1	11	vol	vol	NOUN
easat-8918	1	12	.	.	PROPN
easat-8918	2	1	9	9	NUM
easat-8918	2	2	,	,	PUNCT
easat-8918	2	3	no	no	INTJ
easat-8918	2	4	.	.	NOUN
easat-8918	2	5	7	7	NUM
easat-8918	2	6	,	,	PUNCT
easat-8918	2	7	1345	1345	NUM
easat-8918	2	8	-	-	SYM
easat-8918	2	9	1359	1359	NUM
easat-8918	2	10	2025	2025	NUM
easat-8918	2	11	publisher	publisher	NOUN
easat-8918	2	12	:	:	PUNCT
easat-8918	2	13	learning	learn	VERB
easat-8918	2	14	gate	gate	NOUN
easat-8918	2	15	doi	doi	PROPN
easat-8918	2	16	:	:	PUNCT
easat-8918	2	17	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	2	18	©	©	PROPN
easat-8918	2	19	2025	2025	NUM
easat-8918	2	20	by	by	ADP
easat-8918	2	21	the	the	DET
easat-8918	2	22	author	author	NOUN
easat-8918	2	23	;	;	PUNCT
easat-8918	2	24	licensee	licensee	PROPN
easat-8918	2	25	learning	learning	NOUN
easat-8918	2	26	gate	gate	NOUN
easat-8918	3	1	©	©	PROPN
easat-8918	3	2	2025	2025	NUM
easat-8918	3	3	by	by	ADP
easat-8918	3	4	the	the	DET
easat-8918	3	5	author	author	NOUN
easat-8918	3	6	;	;	PUNCT
easat-8918	3	7	licensee	licensee	PROPN
easat-8918	3	8	learning	learning	NOUN
easat-8918	3	9	gate	gate	NOUN
easat-8918	3	10	history	history	NOUN
easat-8918	3	11	:	:	PUNCT
easat-8918	3	12	received	receive	VERB
easat-8918	3	13	:	:	PUNCT
easat-8918	3	14	6	6	NUM
easat-8918	3	15	may	may	PROPN
easat-8918	3	16	2025	2025	NUM
easat-8918	3	17	;	;	PUNCT
easat-8918	3	18	revised	revise	VERB
easat-8918	3	19	:	:	PUNCT
easat-8918	3	20	7	7	NUM
easat-8918	3	21	june	june	PROPN
easat-8918	3	22	2025	2025	NUM
easat-8918	3	23	;	;	PUNCT
easat-8918	3	24	accepted	accept	VERB
easat-8918	3	25	:	:	PUNCT
easat-8918	3	26	11	11	NUM
easat-8918	3	27	june	june	PROPN
easat-8918	3	28	2025	2025	NUM
easat-8918	3	29	;	;	PUNCT
easat-8918	3	30	published	publish	VERB
easat-8918	3	31	:	:	PUNCT
easat-8918	4	1	17	17	NUM
easat-8918	4	2	july	july	PROPN
easat-8918	4	3	2025	2025	NUM
easat-8918	4	4	*	*	PUNCT
easat-8918	4	5	correspondence	correspondence	NOUN
easat-8918	4	6	:	:	PUNCT
easat-8918	4	7	gulden.eleyan@acm.edu.kw	gulden.eleyan@acm.edu.kw	NOUN
easat-8918	4	8	image	image	NOUN
easat-8918	4	9	denoising	denoising	NOUN
easat-8918	4	10	using	use	VERB
easat-8918	4	11	deep	deep	ADJ
easat-8918	4	12	learning	learning	NOUN
easat-8918	4	13	:	:	PUNCT
easat-8918	4	14	comparative	comparative	ADJ
easat-8918	4	15	study	study	NOUN
easat-8918	4	16	gülden	gülden	ADJ
easat-8918	5	1	eleyan1	eleyan1	PROPN
easat-8918	5	2	*	*	SYM
easat-8918	5	3	1department	1department	NUM
easat-8918	5	4	of	of	ADP
easat-8918	5	5	engineering	engineering	NOUN
easat-8918	5	6	and	and	CCONJ
easat-8918	5	7	technology	technology	NOUN
easat-8918	5	8	,	,	PUNCT
easat-8918	5	9	american	american	PROPN
easat-8918	5	10	college	college	PROPN
easat-8918	5	11	of	of	ADP
easat-8918	5	12	the	the	DET
easat-8918	5	13	middle	middle	PROPN
easat-8918	5	14	east	east	PROPN
easat-8918	5	15	,	,	PUNCT
easat-8918	5	16	egaila	egaila	PROPN
easat-8918	5	17	54200	54200	NUM
easat-8918	5	18	,	,	PUNCT
easat-8918	5	19	kuwait	kuwait	PROPN
easat-8918	5	20	;	;	PUNCT
easat-8918	5	21	gulden.eleyan@acm.edu.kw	gulden.eleyan@acm.edu.kw	X
easat-8918	5	22	(	(	PUNCT
easat-8918	5	23	g.e	g.e	PROPN
easat-8918	5	24	.	.	PUNCT
easat-8918	5	25	)	)	PUNCT
easat-8918	5	26	.	.	PUNCT
easat-8918	6	1	abstract	abstract	ADJ
easat-8918	6	2	:	:	PUNCT
easat-8918	6	3	deep	deep	ADJ
easat-8918	6	4	learning	learning	NOUN
easat-8918	6	5	offers	offer	VERB
easat-8918	6	6	a	a	DET
easat-8918	6	7	flexible	flexible	ADJ
easat-8918	6	8	and	and	CCONJ
easat-8918	6	9	effective	effective	ADJ
easat-8918	6	10	approach	approach	NOUN
easat-8918	6	11	to	to	ADP
easat-8918	6	12	automated	automate	VERB
easat-8918	6	13	image	image	NOUN
easat-8918	6	14	denoising	denoising	NOUN
easat-8918	6	15	.	.	PUNCT
easat-8918	7	1	this	this	DET
easat-8918	7	2	study	study	NOUN
easat-8918	7	3	investigated	investigate	VERB
easat-8918	7	4	the	the	DET
easat-8918	7	5	residual	residual	ADJ
easat-8918	7	6	learning	learning	NOUN
easat-8918	7	7	capabilities	capability	NOUN
easat-8918	7	8	of	of	ADP
easat-8918	7	9	two	two	NUM
easat-8918	7	10	recent	recent	ADJ
easat-8918	7	11	deep	deep	ADJ
easat-8918	7	12	learning	learning	NOUN
easat-8918	7	13	networks	network	NOUN
easat-8918	7	14	:	:	PUNCT
easat-8918	7	15	the	the	DET
easat-8918	7	16	restoration	restoration	NOUN
easat-8918	7	17	transformer	transformer	NOUN
easat-8918	7	18	(	(	PUNCT
easat-8918	7	19	restormer	restormer	ADJ
easat-8918	7	20	)	)	PUNCT
easat-8918	7	21	network	network	NOUN
easat-8918	7	22	and	and	CCONJ
easat-8918	7	23	the	the	DET
easat-8918	7	24	deep	deep	ADJ
easat-8918	7	25	cnn	cnn	PROPN
easat-8918	7	26	(	(	PUNCT
easat-8918	7	27	dncnn	dncnn	NOUN
easat-8918	7	28	)	)	PUNCT
easat-8918	7	29	network	network	NOUN
easat-8918	7	30	.	.	PUNCT
easat-8918	8	1	we	we	PRON
easat-8918	8	2	compared	compare	VERB
easat-8918	8	3	their	their	PRON
easat-8918	8	4	denoising	denoising	NOUN
easat-8918	8	5	performance	performance	NOUN
easat-8918	8	6	on	on	ADP
easat-8918	8	7	the	the	DET
easat-8918	8	8	bsd68	bsd68	PROPN
easat-8918	8	9	dataset	dataset	VERB
easat-8918	8	10	under	under	ADP
easat-8918	8	11	varying	vary	VERB
easat-8918	8	12	levels	level	NOUN
easat-8918	8	13	of	of	ADP
easat-8918	8	14	gaussian	gaussian	ADJ
easat-8918	8	15	noise	noise	NOUN
easat-8918	8	16	against	against	ADP
easat-8918	8	17	established	establish	VERB
easat-8918	8	18	algorithms	algorithm	NOUN
easat-8918	8	19	such	such	ADJ
easat-8918	8	20	as	as	ADP
easat-8918	8	21	block	block	NOUN
easat-8918	8	22	-	-	PUNCT
easat-8918	8	23	matching	matching	NOUN
easat-8918	8	24	and	and	CCONJ
easat-8918	8	25	3d	3d	NUM
easat-8918	8	26	filtering	filtering	NOUN
easat-8918	8	27	(	(	PUNCT
easat-8918	8	28	bm3d	bm3d	PROPN
easat-8918	8	29	)	)	PUNCT
easat-8918	8	30	and	and	CCONJ
easat-8918	8	31	trainable	trainable	ADJ
easat-8918	8	32	nonlinear	nonlinear	ADJ
easat-8918	8	33	reaction	reaction	NOUN
easat-8918	8	34	diffusion	diffusion	NOUN
easat-8918	8	35	(	(	PUNCT
easat-8918	8	36	tnrd	tnrd	NOUN
easat-8918	8	37	)	)	PUNCT
easat-8918	8	38	.	.	PUNCT
easat-8918	9	1	our	our	PRON
easat-8918	9	2	findings	finding	NOUN
easat-8918	9	3	demonstrate	demonstrate	VERB
easat-8918	9	4	that	that	SCONJ
easat-8918	9	5	the	the	DET
easat-8918	9	6	restormer	restormer	ADJ
easat-8918	9	7	algorithm	algorithm	NOUN
easat-8918	9	8	excels	excel	VERB
easat-8918	9	9	in	in	ADP
easat-8918	9	10	noise	noise	NOUN
easat-8918	9	11	removal	removal	NOUN
easat-8918	9	12	.	.	PUNCT
easat-8918	10	1	this	this	PRON
easat-8918	10	2	highlights	highlight	VERB
easat-8918	10	3	the	the	DET
easat-8918	10	4	potential	potential	NOUN
easat-8918	10	5	of	of	ADP
easat-8918	10	6	transformer	transformer	NOUN
easat-8918	10	7	-	-	PUNCT
easat-8918	10	8	based	base	VERB
easat-8918	10	9	architectures	architecture	NOUN
easat-8918	10	10	in	in	ADP
easat-8918	10	11	image	image	NOUN
easat-8918	10	12	restoration	restoration	NOUN
easat-8918	10	13	tasks	task	NOUN
easat-8918	10	14	,	,	PUNCT
easat-8918	10	15	surpassing	surpass	VERB
easat-8918	10	16	traditional	traditional	ADJ
easat-8918	10	17	methods	method	NOUN
easat-8918	10	18	in	in	ADP
easat-8918	10	19	achieving	achieve	VERB
easat-8918	10	20	superior	superior	ADJ
easat-8918	10	21	denoising	denoising	NOUN
easat-8918	10	22	quality	quality	NOUN
easat-8918	10	23	.	.	PUNCT
easat-8918	11	1	further	further	ADJ
easat-8918	11	2	research	research	NOUN
easat-8918	11	3	can	can	AUX
easat-8918	11	4	explore	explore	VERB
easat-8918	11	5	the	the	DET
easat-8918	11	6	application	application	NOUN
easat-8918	11	7	of	of	ADP
easat-8918	11	8	restormer	restormer	NOUN
easat-8918	11	9	to	to	ADP
easat-8918	11	10	other	other	ADJ
easat-8918	11	11	noise	noise	NOUN
easat-8918	11	12	types	type	NOUN
easat-8918	11	13	and	and	CCONJ
easat-8918	11	14	datasets	dataset	NOUN
easat-8918	11	15	.	.	PUNCT
easat-8918	12	1	keywords	keyword	NOUN
easat-8918	12	2	:	:	PUNCT
easat-8918	12	3	convolutional	convolutional	ADJ
easat-8918	12	4	neural	neural	ADJ
easat-8918	12	5	networks	network	NOUN
easat-8918	12	6	,	,	PUNCT
easat-8918	12	7	deep	deep	ADJ
easat-8918	12	8	learning	learning	NOUN
easat-8918	12	9	,	,	PUNCT
easat-8918	12	10	image	image	NOUN
easat-8918	12	11	denoising	denoising	NOUN
easat-8918	12	12	,	,	PUNCT
easat-8918	12	13	transformers	transformer	NOUN
easat-8918	12	14	.	.	PUNCT
easat-8918	13	1	1	1	X
easat-8918	13	2	.	.	X
easat-8918	13	3	introduction	introduction	NOUN
easat-8918	13	4	deep	deep	ADJ
easat-8918	13	5	learning	learning	NOUN
easat-8918	13	6	is	be	AUX
easat-8918	13	7	a	a	DET
easat-8918	13	8	powerful	powerful	ADJ
easat-8918	13	9	tool	tool	NOUN
easat-8918	13	10	that	that	PRON
easat-8918	13	11	enables	enable	VERB
easat-8918	13	12	the	the	DET
easat-8918	13	13	automatic	automatic	ADJ
easat-8918	13	14	removal	removal	NOUN
easat-8918	13	15	of	of	ADP
easat-8918	13	16	noise	noise	NOUN
easat-8918	13	17	from	from	ADP
easat-8918	13	18	images	image	NOUN
easat-8918	13	19	in	in	ADP
easat-8918	13	20	a	a	DET
easat-8918	13	21	flexible	flexible	ADJ
easat-8918	13	22	and	and	CCONJ
easat-8918	13	23	effective	effective	ADJ
easat-8918	13	24	manner	manner	NOUN
easat-8918	13	25	.	.	PUNCT
easat-8918	14	1	one	one	NUM
easat-8918	14	2	of	of	ADP
easat-8918	14	3	the	the	DET
easat-8918	14	4	most	most	ADV
easat-8918	14	5	significant	significant	ADJ
easat-8918	14	6	advantages	advantage	NOUN
easat-8918	14	7	of	of	ADP
easat-8918	14	8	using	use	VERB
easat-8918	14	9	deep	deep	ADJ
easat-8918	14	10	learning	learning	NOUN
easat-8918	14	11	for	for	ADP
easat-8918	14	12	image	image	NOUN
easat-8918	14	13	denoising	denoising	NOUN
easat-8918	14	14	is	be	AUX
easat-8918	14	15	its	its	PRON
easat-8918	14	16	ability	ability	NOUN
easat-8918	14	17	to	to	PART
easat-8918	14	18	learn	learn	VERB
easat-8918	14	19	and	and	CCONJ
easat-8918	14	20	capture	capture	VERB
easat-8918	14	21	complex	complex	ADJ
easat-8918	14	22	patterns	pattern	NOUN
easat-8918	14	23	in	in	ADP
easat-8918	14	24	the	the	DET
easat-8918	14	25	data	datum	NOUN
easat-8918	14	26	,	,	PUNCT
easat-8918	14	27	unlike	unlike	ADP
easat-8918	14	28	traditional	traditional	ADJ
easat-8918	14	29	image	image	NOUN
easat-8918	14	30	denoising	denoising	NOUN
easat-8918	14	31	techniques	technique	NOUN
easat-8918	14	32	,	,	PUNCT
easat-8918	14	33	which	which	PRON
easat-8918	14	34	may	may	AUX
easat-8918	14	35	not	not	PART
easat-8918	14	36	effectively	effectively	ADV
easat-8918	14	37	remove	remove	VERB
easat-8918	14	38	noise	noise	NOUN
easat-8918	14	39	in	in	ADP
easat-8918	14	40	all	all	DET
easat-8918	14	41	cases	case	NOUN
easat-8918	14	42	.	.	PUNCT
easat-8918	15	1	deep	deep	ADJ
easat-8918	15	2	learning	learning	NOUN
easat-8918	15	3	-	-	PUNCT
easat-8918	15	4	based	base	VERB
easat-8918	15	5	approaches	approach	NOUN
easat-8918	15	6	have	have	AUX
easat-8918	15	7	been	be	AUX
easat-8918	15	8	successfully	successfully	ADV
easat-8918	15	9	applied	apply	VERB
easat-8918	15	10	in	in	ADP
easat-8918	15	11	various	various	ADJ
easat-8918	15	12	fields	field	NOUN
easat-8918	15	13	,	,	PUNCT
easat-8918	15	14	including	include	VERB
easat-8918	15	15	low	low	ADJ
easat-8918	15	16	-	-	PUNCT
easat-8918	15	17	light	light	NOUN
easat-8918	15	18	image	image	NOUN
easat-8918	15	19	enhancement	enhancement	NOUN
easat-8918	15	20	,	,	PUNCT
easat-8918	15	21	noise	noise	NOUN
easat-8918	15	22	reduction	reduction	NOUN
easat-8918	15	23	in	in	ADP
easat-8918	15	24	medical	medical	ADJ
easat-8918	15	25	imaging	imaging	NOUN
easat-8918	15	26	,	,	PUNCT
easat-8918	15	27	and	and	CCONJ
easat-8918	15	28	image	image	NOUN
easat-8918	15	29	restoration	restoration	NOUN
easat-8918	15	30	.	.	PUNCT
easat-8918	16	1	noise	noise	NOUN
easat-8918	16	2	reduction	reduction	NOUN
easat-8918	16	3	in	in	ADP
easat-8918	16	4	images	image	NOUN
easat-8918	16	5	is	be	AUX
easat-8918	16	6	the	the	DET
easat-8918	16	7	process	process	NOUN
easat-8918	16	8	of	of	ADP
easat-8918	16	9	removing	remove	VERB
easat-8918	16	10	noise	noise	NOUN
easat-8918	16	11	to	to	PART
easat-8918	16	12	improve	improve	VERB
easat-8918	16	13	the	the	DET
easat-8918	16	14	visual	visual	ADJ
easat-8918	16	15	quality	quality	NOUN
easat-8918	16	16	of	of	ADP
easat-8918	16	17	an	an	DET
easat-8918	16	18	image	image	NOUN
easat-8918	16	19	.	.	PUNCT
easat-8918	17	1	noise	noise	NOUN
easat-8918	17	2	can	can	AUX
easat-8918	17	3	be	be	AUX
easat-8918	17	4	introduced	introduce	VERB
easat-8918	17	5	into	into	ADP
easat-8918	17	6	an	an	DET
easat-8918	17	7	image	image	NOUN
easat-8918	17	8	through	through	ADP
easat-8918	17	9	various	various	ADJ
easat-8918	17	10	means	mean	NOUN
easat-8918	17	11	,	,	PUNCT
easat-8918	17	12	such	such	ADJ
easat-8918	17	13	as	as	ADP
easat-8918	17	14	capturing	capture	VERB
easat-8918	17	15	in	in	ADP
easat-8918	17	16	low	low	ADJ
easat-8918	17	17	-	-	PUNCT
easat-8918	17	18	light	light	NOUN
easat-8918	17	19	conditions	condition	NOUN
easat-8918	17	20	,	,	PUNCT
easat-8918	17	21	image	image	NOUN
easat-8918	17	22	compression	compression	NOUN
easat-8918	17	23	,	,	PUNCT
easat-8918	17	24	or	or	CCONJ
easat-8918	17	25	transmission	transmission	NOUN
easat-8918	17	26	over	over	ADP
easat-8918	17	27	a	a	DET
easat-8918	17	28	noisy	noisy	ADJ
easat-8918	17	29	channel	channel	NOUN
easat-8918	17	30	.	.	PUNCT
easat-8918	18	1	removing	remove	VERB
easat-8918	18	2	noise	noise	NOUN
easat-8918	18	3	from	from	ADP
easat-8918	18	4	an	an	DET
easat-8918	18	5	image	image	NOUN
easat-8918	18	6	helps	helps	AUX
easat-8918	18	7	restore	restore	VERB
easat-8918	18	8	its	its	PRON
easat-8918	18	9	original	original	ADJ
easat-8918	18	10	clarity	clarity	NOUN
easat-8918	18	11	and	and	CCONJ
easat-8918	18	12	enhances	enhance	VERB
easat-8918	18	13	its	its	PRON
easat-8918	18	14	overall	overall	ADJ
easat-8918	18	15	appearance	appearance	NOUN
easat-8918	18	16	[	[	X
easat-8918	18	17	1	1	NUM
easat-8918	18	18	]	]	PUNCT
easat-8918	18	19	.	.	PUNCT
easat-8918	19	1	several	several	ADJ
easat-8918	19	2	techniques	technique	NOUN
easat-8918	19	3	can	can	AUX
easat-8918	19	4	be	be	AUX
easat-8918	19	5	used	use	VERB
easat-8918	19	6	to	to	PART
easat-8918	19	7	remove	remove	VERB
easat-8918	19	8	noise	noise	NOUN
easat-8918	19	9	from	from	ADP
easat-8918	19	10	images	image	NOUN
easat-8918	19	11	,	,	PUNCT
easat-8918	19	12	including	include	VERB
easat-8918	19	13	adaptive	adaptive	ADJ
easat-8918	19	14	filters	filter	NOUN
easat-8918	19	15	[	[	PUNCT
easat-8918	19	16	2	2	NUM
easat-8918	19	17	-	-	SYM
easat-8918	19	18	5	5	NUM
easat-8918	19	19	]	]	PUNCT
easat-8918	19	20	and	and	CCONJ
easat-8918	19	21	wavelet	wavelet	NOUN
easat-8918	19	22	-	-	PUNCT
easat-8918	19	23	based	base	VERB
easat-8918	19	24	denoising	denoising	NOUN
easat-8918	19	25	[	[	X
easat-8918	19	26	6	6	NUM
easat-8918	19	27	-	-	SYM
easat-8918	19	28	8	8	NUM
easat-8918	19	29	]	]	PUNCT
easat-8918	19	30	.	.	PUNCT
easat-8918	20	1	smoothing	smooth	VERB
easat-8918	20	2	filters	filter	NOUN
easat-8918	20	3	work	work	VERB
easat-8918	20	4	by	by	ADP
easat-8918	20	5	averaging	average	VERB
easat-8918	20	6	the	the	DET
easat-8918	20	7	pixel	pixel	PROPN
easat-8918	20	8	values	value	NOUN
easat-8918	20	9	in	in	ADP
easat-8918	20	10	an	an	DET
easat-8918	20	11	image	image	NOUN
easat-8918	20	12	,	,	PUNCT
easat-8918	20	13	while	while	SCONJ
easat-8918	20	14	median	median	ADJ
easat-8918	20	15	filters	filter	NOUN
easat-8918	20	16	replace	replace	VERB
easat-8918	20	17	a	a	DET
easat-8918	20	18	pixel	pixel	NOUN
easat-8918	20	19	’s	’s	PART
easat-8918	20	20	value	value	NOUN
easat-8918	20	21	with	with	ADP
easat-8918	20	22	the	the	DET
easat-8918	20	23	median	median	ADJ
easat-8918	20	24	value	value	NOUN
easat-8918	20	25	of	of	ADP
easat-8918	20	26	its	its	PRON
easat-8918	20	27	surrounding	surround	VERB
easat-8918	20	28	pixels	pixel	NOUN
easat-8918	20	29	.	.	PUNCT
easat-8918	21	1	wavelet	wavelet	NOUN
easat-8918	21	2	-	-	PUNCT
easat-8918	21	3	based	base	VERB
easat-8918	21	4	denoising	denoising	NOUN
easat-8918	21	5	employs	employ	VERB
easat-8918	21	6	wavelet	wavelet	NOUN
easat-8918	21	7	transforms	transform	VERB
easat-8918	21	8	to	to	PART
easat-8918	21	9	separate	separate	VERB
easat-8918	21	10	an	an	DET
easat-8918	21	11	image	image	NOUN
easat-8918	21	12	into	into	ADP
easat-8918	21	13	different	different	ADJ
easat-8918	21	14	frequency	frequency	NOUN
easat-8918	21	15	bands	band	NOUN
easat-8918	21	16	,	,	PUNCT
easat-8918	21	17	allowing	allow	VERB
easat-8918	21	18	noise	noise	NOUN
easat-8918	21	19	to	to	PART
easat-8918	21	20	be	be	AUX
easat-8918	21	21	distinguished	distinguish	VERB
easat-8918	21	22	from	from	ADP
easat-8918	21	23	the	the	DET
easat-8918	21	24	signal	signal	NOUN
easat-8918	21	25	and	and	CCONJ
easat-8918	21	26	removed	remove	VERB
easat-8918	21	27	.	.	PUNCT
easat-8918	22	1	completely	completely	ADV
easat-8918	22	2	eliminating	eliminate	VERB
easat-8918	22	3	noise	noise	NOUN
easat-8918	22	4	often	often	ADV
easat-8918	22	5	requires	require	VERB
easat-8918	22	6	sacrificing	sacrifice	VERB
easat-8918	22	7	some	some	DET
easat-8918	22	8	level	level	NOUN
easat-8918	22	9	of	of	ADP
easat-8918	22	10	detail	detail	NOUN
easat-8918	22	11	,	,	PUNCT
easat-8918	22	12	but	but	CCONJ
easat-8918	22	13	with	with	ADP
easat-8918	22	14	careful	careful	ADJ
easat-8918	22	15	tuning	tuning	NOUN
easat-8918	22	16	,	,	PUNCT
easat-8918	22	17	it	it	PRON
easat-8918	22	18	is	be	AUX
easat-8918	22	19	possible	possible	ADJ
easat-8918	22	20	to	to	PART
easat-8918	22	21	minimize	minimize	VERB
easat-8918	22	22	detail	detail	NOUN
easat-8918	22	23	loss	loss	NOUN
easat-8918	22	24	while	while	SCONJ
easat-8918	22	25	achieving	achieve	VERB
easat-8918	22	26	effective	effective	ADJ
easat-8918	22	27	noise	noise	NOUN
easat-8918	22	28	reduction	reduction	NOUN
easat-8918	22	29	.	.	PUNCT
easat-8918	23	1	deep	deep	ADJ
easat-8918	23	2	learning	learning	NOUN
easat-8918	23	3	has	have	AUX
easat-8918	23	4	recently	recently	ADV
easat-8918	23	5	emerged	emerge	VERB
easat-8918	23	6	as	as	ADP
easat-8918	23	7	a	a	DET
easat-8918	23	8	powerful	powerful	ADJ
easat-8918	23	9	tool	tool	NOUN
easat-8918	23	10	for	for	ADP
easat-8918	23	11	image	image	NOUN
easat-8918	23	12	noise	noise	NOUN
easat-8918	23	13	reduction	reduction	NOUN
easat-8918	23	14	.	.	PUNCT
easat-8918	24	1	in	in	ADP
easat-8918	24	2	deep	deep	ADJ
easat-8918	24	3	learning	learning	NOUN
easat-8918	24	4	,	,	PUNCT
easat-8918	24	5	a	a	DET
easat-8918	24	6	neural	neural	ADJ
easat-8918	24	7	network	network	NOUN
easat-8918	24	8	is	be	AUX
easat-8918	24	9	trained	train	VERB
easat-8918	24	10	to	to	PART
easat-8918	24	11	learn	learn	VERB
easat-8918	24	12	the	the	DET
easat-8918	24	13	underlying	underlie	VERB
easat-8918	24	14	structure	structure	NOUN
easat-8918	24	15	of	of	ADP
easat-8918	24	16	an	an	DET
easat-8918	24	17	image	image	NOUN
easat-8918	24	18	and	and	CCONJ
easat-8918	24	19	predict	predict	VERB
easat-8918	24	20	a	a	DET
easat-8918	24	21	noise	noise	NOUN
easat-8918	24	22	-	-	PUNCT
easat-8918	24	23	free	free	ADJ
easat-8918	24	24	version	version	NOUN
easat-8918	24	25	from	from	ADP
easat-8918	24	26	a	a	DET
easat-8918	24	27	noisy	noisy	ADJ
easat-8918	24	28	input	input	NOUN
easat-8918	25	1	[	[	X
easat-8918	25	2	1	1	NUM
easat-8918	25	3	]	]	PUNCT
easat-8918	25	4	.	.	PUNCT
easat-8918	26	1	this	this	PRON
easat-8918	26	2	is	be	AUX
easat-8918	26	3	typically	typically	ADV
easat-8918	26	4	achieved	achieve	VERB
easat-8918	26	5	using	use	VERB
easat-8918	26	6	a	a	DET
easat-8918	26	7	large	large	ADJ
easat-8918	26	8	image	image	NOUN
easat-8918	26	9	dataset	dataset	NOUN
easat-8918	26	10	containing	contain	VERB
easat-8918	26	11	both	both	CCONJ
easat-8918	26	12	noisy	noisy	ADJ
easat-8918	26	13	images	image	NOUN
easat-8918	26	14	and	and	CCONJ
easat-8918	26	15	their	their	PRON
easat-8918	26	16	corresponding	corresponding	ADJ
easat-8918	26	17	clean	clean	ADJ
easat-8918	26	18	versions	version	NOUN
easat-8918	26	19	,	,	PUNCT
easat-8918	26	20	which	which	PRON
easat-8918	26	21	is	be	AUX
easat-8918	26	22	used	use	VERB
easat-8918	26	23	to	to	PART
easat-8918	26	24	train	train	VERB
easat-8918	26	25	the	the	DET
easat-8918	26	26	network	network	NOUN
easat-8918	26	27	to	to	PART
easat-8918	26	28	learn	learn	VERB
easat-8918	26	29	the	the	DET
easat-8918	26	30	relationship	relationship	NOUN
easat-8918	26	31	between	between	ADP
easat-8918	26	32	noisy	noisy	ADJ
easat-8918	26	33	and	and	CCONJ
easat-8918	26	34	noise	noise	NOUN
easat-8918	26	35	-	-	PUNCT
easat-8918	26	36	free	free	ADJ
easat-8918	26	37	images	image	NOUN
easat-8918	26	38	.	.	PUNCT
easat-8918	27	1	one	one	NUM
easat-8918	27	2	of	of	ADP
easat-8918	27	3	the	the	DET
easat-8918	27	4	key	key	ADJ
easat-8918	27	5	advantages	advantage	NOUN
easat-8918	27	6	of	of	ADP
easat-8918	27	7	using	use	VERB
easat-8918	27	8	deep	deep	ADJ
easat-8918	27	9	learning	learning	NOUN
easat-8918	27	10	for	for	ADP
easat-8918	27	11	image	image	NOUN
easat-8918	27	12	denoising	denoising	NOUN
easat-8918	27	13	is	be	AUX
easat-8918	27	14	its	its	PRON
easat-8918	27	15	ability	ability	NOUN
easat-8918	27	16	to	to	PART
easat-8918	27	17	learn	learn	VERB
easat-8918	27	18	and	and	CCONJ
easat-8918	27	19	capture	capture	VERB
easat-8918	27	20	complex	complex	ADJ
easat-8918	27	21	structures	structure	NOUN
easat-8918	27	22	within	within	ADP
easat-8918	27	23	data	datum	NOUN
easat-8918	27	24	.	.	PUNCT
easat-8918	28	1	traditional	traditional	ADJ
easat-8918	28	2	image	image	NOUN
easat-8918	28	3	denoising	denoise	VERB
easat-8918	28	4	techniques	technique	NOUN
easat-8918	28	5	rely	rely	VERB
easat-8918	28	6	on	on	ADP
easat-8918	28	7	fixed	fix	VERB
easat-8918	28	8	filtering	filtering	NOUN
easat-8918	28	9	,	,	PUNCT
easat-8918	28	10	which	which	PRON
easat-8918	28	11	may	may	AUX
easat-8918	28	12	not	not	PART
easat-8918	28	13	effectively	effectively	ADV
easat-8918	28	14	remove	remove	VERB
easat-8918	28	15	noise	noise	NOUN
easat-8918	28	16	in	in	ADP
easat-8918	28	17	all	all	DET
easat-8918	28	18	situations	situation	NOUN
easat-8918	28	19	.	.	PUNCT
easat-8918	29	1	in	in	ADP
easat-8918	29	2	contrast	contrast	NOUN
easat-8918	29	3	,	,	PUNCT
easat-8918	29	4	deep	deep	ADJ
easat-8918	29	5	learning	learning	NOUN
easat-8918	29	6	-	-	PUNCT
easat-8918	29	7	based	base	VERB
easat-8918	29	8	approaches	approach	NOUN
easat-8918	29	9	can	can	AUX
easat-8918	29	10	1346	1346	NUM
easat-8918	29	11	edelweiss	edelweiss	PROPN
easat-8918	29	12	applied	apply	VERB
easat-8918	29	13	science	science	NOUN
easat-8918	29	14	and	and	CCONJ
easat-8918	29	15	technology	technology	NOUN
easat-8918	29	16	issn	issn	PROPN
easat-8918	29	17	:	:	PUNCT
easat-8918	29	18	2576	2576	NUM
easat-8918	29	19	-	-	SYM
easat-8918	29	20	8484	8484	NUM
easat-8918	29	21	vol	vol	NOUN
easat-8918	29	22	.	.	PROPN
easat-8918	30	1	9	9	NUM
easat-8918	30	2	,	,	PUNCT
easat-8918	30	3	no	no	INTJ
easat-8918	30	4	.	.	NOUN
easat-8918	30	5	7	7	NUM
easat-8918	30	6	:	:	SYM
easat-8918	30	7	1345	1345	NUM
easat-8918	30	8	-	-	SYM
easat-8918	30	9	1359	1359	NUM
easat-8918	30	10	,	,	PUNCT
easat-8918	30	11	2025	2025	NUM
easat-8918	30	12	doi	doi	NOUN
easat-8918	30	13	:	:	PUNCT
easat-8918	30	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	30	15	©	©	PROPN
easat-8918	30	16	2025	2025	NUM
easat-8918	30	17	by	by	ADP
easat-8918	30	18	the	the	DET
easat-8918	30	19	author	author	NOUN
easat-8918	30	20	;	;	PUNCT
easat-8918	30	21	licensee	licensee	NOUN
easat-8918	30	22	learning	learn	VERB
easat-8918	30	23	gate	gate	NOUN
easat-8918	30	24	adapt	adapt	VERB
easat-8918	30	25	to	to	ADP
easat-8918	30	26	different	different	ADJ
easat-8918	30	27	types	type	NOUN
easat-8918	30	28	of	of	ADP
easat-8918	30	29	noise	noise	NOUN
easat-8918	30	30	and	and	CCONJ
easat-8918	30	31	denoise	denoise	NOUN
easat-8918	30	32	images	image	NOUN
easat-8918	30	33	more	more	ADV
easat-8918	30	34	flexibly	flexibly	ADV
easat-8918	30	35	and	and	CCONJ
easat-8918	30	36	effectively	effectively	ADV
easat-8918	30	37	.	.	PUNCT
easat-8918	31	1	additionally	additionally	ADV
easat-8918	31	2	,	,	PUNCT
easat-8918	31	3	deep	deep	ADJ
easat-8918	31	4	learning	learning	NOUN
easat-8918	31	5	models	model	NOUN
easat-8918	31	6	can	can	AUX
easat-8918	31	7	be	be	AUX
easat-8918	31	8	trained	train	VERB
easat-8918	31	9	in	in	ADP
easat-8918	31	10	an	an	DET
easat-8918	31	11	end	end	NOUN
easat-8918	31	12	-	-	PUNCT
easat-8918	31	13	to	to	ADP
easat-8918	31	14	-	-	PUNCT
easat-8918	31	15	end	end	NOUN
easat-8918	31	16	manner	manner	NOUN
easat-8918	31	17	,	,	PUNCT
easat-8918	31	18	allowing	allow	VERB
easat-8918	31	19	them	they	PRON
easat-8918	31	20	to	to	PART
easat-8918	31	21	learn	learn	VERB
easat-8918	31	22	from	from	ADP
easat-8918	31	23	large	large	ADJ
easat-8918	31	24	datasets	dataset	NOUN
easat-8918	31	25	and	and	CCONJ
easat-8918	31	26	automatically	automatically	ADV
easat-8918	31	27	determine	determine	VERB
easat-8918	31	28	the	the	DET
easat-8918	31	29	most	most	ADV
easat-8918	31	30	suitable	suitable	ADJ
easat-8918	31	31	noise	noise	NOUN
easat-8918	31	32	reduction	reduction	NOUN
easat-8918	31	33	technique	technique	NOUN
easat-8918	31	34	for	for	ADP
easat-8918	31	35	a	a	DET
easat-8918	31	36	given	give	VERB
easat-8918	31	37	task	task	NOUN
easat-8918	31	38	.	.	PUNCT
easat-8918	32	1	as	as	ADP
easat-8918	32	2	a	a	DET
easat-8918	32	3	result	result	NOUN
easat-8918	32	4	,	,	PUNCT
easat-8918	32	5	deep	deep	ADJ
easat-8918	32	6	learning	learning	NOUN
easat-8918	32	7	has	have	VERB
easat-8918	32	8	the	the	DET
easat-8918	32	9	potential	potential	NOUN
easat-8918	32	10	to	to	PART
easat-8918	32	11	enhance	enhance	VERB
easat-8918	32	12	the	the	DET
easat-8918	32	13	performance	performance	NOUN
easat-8918	32	14	of	of	ADP
easat-8918	32	15	image	image	NOUN
easat-8918	32	16	denoising	denoising	NOUN
easat-8918	32	17	techniques	technique	NOUN
easat-8918	32	18	and	and	CCONJ
easat-8918	32	19	has	have	AUX
easat-8918	32	20	already	already	ADV
easat-8918	32	21	been	be	AUX
easat-8918	32	22	successfully	successfully	ADV
easat-8918	32	23	applied	apply	VERB
easat-8918	32	24	in	in	ADP
easat-8918	32	25	various	various	ADJ
easat-8918	32	26	scenarios	scenario	NOUN
easat-8918	32	27	.	.	PUNCT
easat-8918	33	1	the	the	DET
easat-8918	33	2	rest	rest	NOUN
easat-8918	33	3	of	of	ADP
easat-8918	33	4	this	this	DET
easat-8918	33	5	work	work	NOUN
easat-8918	33	6	is	be	AUX
easat-8918	33	7	structured	structure	VERB
easat-8918	33	8	as	as	SCONJ
easat-8918	33	9	follows	follow	VERB
easat-8918	33	10	:	:	PUNCT
easat-8918	33	11	the	the	DET
easat-8918	33	12	history	history	NOUN
easat-8918	33	13	and	and	CCONJ
easat-8918	33	14	related	relate	VERB
easat-8918	33	15	work	work	NOUN
easat-8918	33	16	on	on	ADP
easat-8918	33	17	image	image	NOUN
easat-8918	33	18	denoising	denoising	NOUN
easat-8918	33	19	are	be	AUX
easat-8918	33	20	summarized	summarize	VERB
easat-8918	33	21	in	in	ADP
easat-8918	33	22	section	section	NOUN
easat-8918	33	23	2	2	NUM
easat-8918	33	24	.	.	PUNCT
easat-8918	33	25	section	section	NOUN
easat-8918	33	26	3	3	NUM
easat-8918	33	27	discusses	discusse	NOUN
easat-8918	33	28	and	and	CCONJ
easat-8918	33	29	presents	present	VERB
easat-8918	33	30	two	two	NUM
easat-8918	33	31	recently	recently	ADV
easat-8918	33	32	introduced	introduce	VERB
easat-8918	33	33	image	image	NOUN
easat-8918	33	34	denoising	denoise	VERB
easat-8918	33	35	algorithms	algorithm	NOUN
easat-8918	33	36	explored	explore	VERB
easat-8918	33	37	in	in	ADP
easat-8918	33	38	this	this	DET
easat-8918	33	39	project	project	NOUN
easat-8918	33	40	—	—	PUNCT
easat-8918	33	41	the	the	DET
easat-8918	33	42	restoration	restoration	NOUN
easat-8918	33	43	transformer	transformer	NOUN
easat-8918	33	44	(	(	PUNCT
easat-8918	33	45	restormer	restormer	ADJ
easat-8918	33	46	)	)	PUNCT
easat-8918	33	47	network	network	NOUN
easat-8918	33	48	and	and	CCONJ
easat-8918	33	49	deep	deep	ADJ
easat-8918	33	50	convolutional	convolutional	ADJ
easat-8918	33	51	neural	neural	ADJ
easat-8918	33	52	networks	network	NOUN
easat-8918	33	53	(	(	PUNCT
easat-8918	33	54	dncnn	dncnn	PROPN
easat-8918	33	55	)	)	PUNCT
easat-8918	33	56	residual	residual	ADJ
easat-8918	33	57	learning	learning	NOUN
easat-8918	33	58	.	.	PUNCT
easat-8918	34	1	the	the	DET
easat-8918	34	2	results	result	NOUN
easat-8918	34	3	and	and	CCONJ
easat-8918	34	4	comparisons	comparison	NOUN
easat-8918	34	5	are	be	AUX
easat-8918	34	6	presented	present	VERB
easat-8918	34	7	and	and	CCONJ
easat-8918	34	8	analyzed	analyze	VERB
easat-8918	34	9	in	in	ADP
easat-8918	34	10	section	section	NOUN
easat-8918	34	11	4	4	NUM
easat-8918	34	12	.	.	PUNCT
easat-8918	35	1	finally	finally	ADV
easat-8918	35	2	,	,	PUNCT
easat-8918	35	3	section	section	NOUN
easat-8918	35	4	5	5	NUM
easat-8918	35	5	concludes	conclude	VERB
easat-8918	35	6	the	the	DET
easat-8918	35	7	project	project	NOUN
easat-8918	35	8	and	and	CCONJ
easat-8918	35	9	discusses	discuss	VERB
easat-8918	35	10	future	future	ADJ
easat-8918	35	11	work	work	NOUN
easat-8918	35	12	.	.	PUNCT
easat-8918	36	1	2	2	X
easat-8918	36	2	.	.	X
easat-8918	36	3	history	history	NOUN
easat-8918	36	4	and	and	CCONJ
easat-8918	36	5	related	relate	VERB
easat-8918	36	6	work	work	NOUN
easat-8918	36	7	2.1	2.1	NUM
easat-8918	36	8	.	.	PUNCT
easat-8918	37	1	convolutional	convolutional	ADJ
easat-8918	37	2	neural	neural	ADJ
easat-8918	37	3	networks	network	NOUN
easat-8918	37	4	(	(	PUNCT
easat-8918	37	5	cnns	cnns	PROPN
easat-8918	37	6	)	)	PUNCT
easat-8918	37	7	cnns	cnn	NOUN
easat-8918	37	8	have	have	AUX
easat-8918	37	9	achieved	achieve	VERB
easat-8918	37	10	remarkable	remarkable	ADJ
easat-8918	37	11	success	success	NOUN
easat-8918	37	12	in	in	ADP
easat-8918	37	13	image	image	NOUN
easat-8918	37	14	processing	processing	NOUN
easat-8918	37	15	due	due	ADP
easat-8918	37	16	to	to	ADP
easat-8918	37	17	their	their	PRON
easat-8918	37	18	plug	plug	VERB
easat-8918	37	19	-	-	PUNCT
easat-8918	37	20	and	and	CCONJ
easat-8918	37	21	-	-	PUNCT
easat-8918	37	22	play	play	NOUN
easat-8918	37	23	network	network	NOUN
easat-8918	37	24	architectures	architecture	NOUN
easat-8918	37	25	[	[	X
easat-8918	37	26	9	9	NUM
easat-8918	37	27	-	-	SYM
easat-8918	37	28	14	14	NUM
easat-8918	37	29	]	]	PUNCT
easat-8918	37	30	.	.	PUNCT
easat-8918	38	1	as	as	ADP
easat-8918	38	2	a	a	DET
easat-8918	38	3	pioneer	pioneer	NOUN
easat-8918	38	4	in	in	ADP
easat-8918	38	5	cnn	cnn	PROPN
easat-8918	38	6	technology	technology	NOUN
easat-8918	38	7	,	,	PUNCT
easat-8918	38	8	lenet	lenet	PROPN
easat-8918	38	9	[	[	X
easat-8918	38	10	15	15	NUM
easat-8918	38	11	]	]	X
easat-8918	38	12	utilized	utilize	VERB
easat-8918	38	13	convolutional	convolutional	ADJ
easat-8918	38	14	kernels	kernel	NOUN
easat-8918	38	15	of	of	ADP
easat-8918	38	16	different	different	ADJ
easat-8918	38	17	sizes	size	NOUN
easat-8918	38	18	to	to	PART
easat-8918	38	19	extract	extract	VERB
easat-8918	38	20	features	feature	NOUN
easat-8918	38	21	and	and	CCONJ
easat-8918	38	22	achieve	achieve	VERB
easat-8918	38	23	effective	effective	ADJ
easat-8918	38	24	image	image	NOUN
easat-8918	38	25	classification	classification	NOUN
easat-8918	38	26	.	.	PUNCT
easat-8918	39	1	however	however	ADV
easat-8918	39	2	,	,	PUNCT
easat-8918	39	3	due	due	ADP
easat-8918	39	4	to	to	ADP
easat-8918	39	5	the	the	DET
easat-8918	39	6	use	use	NOUN
easat-8918	39	7	of	of	ADP
easat-8918	39	8	the	the	DET
easat-8918	39	9	sigmoid	sigmoid	NOUN
easat-8918	39	10	activation	activation	NOUN
easat-8918	39	11	function	function	NOUN
easat-8918	39	12	,	,	PUNCT
easat-8918	39	13	lenet	lenet	PROPN
easat-8918	39	14	had	have	VERB
easat-8918	39	15	a	a	DET
easat-8918	39	16	slow	slow	ADJ
easat-8918	39	17	convergence	convergence	NOUN
easat-8918	39	18	rate	rate	NOUN
easat-8918	39	19	,	,	PUNCT
easat-8918	39	20	which	which	PRON
easat-8918	39	21	posed	pose	VERB
easat-8918	39	22	a	a	DET
easat-8918	39	23	limitation	limitation	NOUN
easat-8918	39	24	in	in	ADP
easat-8918	39	25	realworld	realworld	PROPN
easat-8918	39	26	applications	application	NOUN
easat-8918	39	27	.	.	PUNCT
easat-8918	40	1	following	follow	VERB
easat-8918	40	2	lenet	lenet	NOUN
easat-8918	40	3	,	,	PUNCT
easat-8918	40	4	alexnet	alexnet	ADJ
easat-8918	40	5	krizhevsky	krizhevsky	NOUN
easat-8918	40	6	,	,	PUNCT
easat-8918	40	7	et	et	PROPN
easat-8918	40	8	al	al	PROPN
easat-8918	40	9	.	.	PUNCT
easat-8918	41	1	[	[	X
easat-8918	41	2	16	16	NUM
easat-8918	41	3	]	]	PUNCT
easat-8918	41	4	and	and	CCONJ
easat-8918	41	5	daalah	daalah	PROPN
easat-8918	41	6	,	,	PUNCT
easat-8918	41	7	et	et	PROPN
easat-8918	41	8	al	al	PROPN
easat-8918	41	9	.	.	PUNCT
easat-8918	42	1	[	[	X
easat-8918	42	2	17	17	NUM
easat-8918	42	3	]	]	PUNCT
easat-8918	42	4	became	become	VERB
easat-8918	42	5	a	a	DET
easat-8918	42	6	milestone	milestone	NOUN
easat-8918	42	7	in	in	ADP
easat-8918	42	8	deep	deep	ADJ
easat-8918	42	9	learning	learning	NOUN
easat-8918	42	10	(	(	PUNCT
easat-8918	42	11	see	see	VERB
easat-8918	42	12	figure	figure	NOUN
easat-8918	42	13	1	1	NUM
easat-8918	42	14	)	)	PUNCT
easat-8918	42	15	.	.	PUNCT
easat-8918	43	1	its	its	PRON
easat-8918	43	2	success	success	NOUN
easat-8918	43	3	stemmed	stem	VERB
easat-8918	43	4	from	from	ADP
easat-8918	43	5	its	its	PRON
easat-8918	43	6	ability	ability	NOUN
easat-8918	43	7	to	to	PART
easat-8918	43	8	address	address	VERB
easat-8918	43	9	the	the	DET
easat-8918	43	10	overfitting	overfitte	VERB
easat-8918	43	11	problem	problem	NOUN
easat-8918	43	12	and	and	CCONJ
easat-8918	43	13	improve	improve	VERB
easat-8918	43	14	the	the	DET
easat-8918	43	15	speed	speed	NOUN
easat-8918	43	16	of	of	ADP
easat-8918	43	17	stochastic	stochastic	ADJ
easat-8918	43	18	gradient	gradient	ADJ
easat-8918	43	19	descent	descent	NOUN
easat-8918	43	20	(	(	PUNCT
easat-8918	43	21	sgd	sgd	PROPN
easat-8918	43	22	)	)	PUNCT
easat-8918	43	23	instead	instead	ADV
easat-8918	43	24	of	of	ADP
easat-8918	43	25	using	use	VERB
easat-8918	43	26	the	the	DET
easat-8918	43	27	sigmoid	sigmoid	NOUN
easat-8918	43	28	activation	activation	NOUN
easat-8918	43	29	function	function	NOUN
easat-8918	43	30	[	[	X
easat-8918	43	31	18	18	NUM
easat-8918	43	32	]	]	PUNCT
easat-8918	43	33	.	.	PUNCT
easat-8918	44	1	while	while	SCONJ
easat-8918	44	2	alexnet	alexnet	PROPN
easat-8918	44	3	achieved	achieve	VERB
easat-8918	44	4	high	high	ADJ
easat-8918	44	5	performance	performance	NOUN
easat-8918	44	6	,	,	PUNCT
easat-8918	44	7	its	its	PRON
easat-8918	44	8	large	large	ADJ
easat-8918	44	9	convolutional	convolutional	ADJ
easat-8918	44	10	kernels	kernel	NOUN
easat-8918	44	11	required	require	VERB
easat-8918	44	12	significant	significant	ADJ
easat-8918	44	13	memory	memory	NOUN
easat-8918	44	14	usage	usage	NOUN
easat-8918	44	15	,	,	PUNCT
easat-8918	44	16	limiting	limit	VERB
easat-8918	44	17	its	its	PRON
easat-8918	44	18	application	application	NOUN
easat-8918	44	19	in	in	ADP
easat-8918	44	20	real	real	ADJ
easat-8918	44	21	-	-	PUNCT
easat-8918	44	22	world	world	NOUN
easat-8918	44	23	scenarios	scenario	NOUN
easat-8918	44	24	such	such	ADJ
easat-8918	44	25	as	as	ADP
easat-8918	44	26	smart	smart	ADJ
easat-8918	44	27	cameras	camera	NOUN
easat-8918	44	28	.	.	PUNCT
easat-8918	45	1	figure	figure	NOUN
easat-8918	45	2	1	1	NUM
easat-8918	45	3	.	.	PUNCT
easat-8918	46	1	lenet	lenet	NOUN
easat-8918	46	2	(	(	PUNCT
easat-8918	46	3	top	top	ADJ
easat-8918	46	4	)	)	PUNCT
easat-8918	46	5	and	and	CCONJ
easat-8918	46	6	alexnet	alexnet	PROPN
easat-8918	46	7	(	(	PUNCT
easat-8918	46	8	bottom	bottom	NOUN
easat-8918	46	9	)	)	PUNCT
easat-8918	46	10	.	.	PUNCT
easat-8918	47	1	1347	1347	NUM
easat-8918	47	2	edelweiss	edelweiss	PROPN
easat-8918	47	3	applied	apply	VERB
easat-8918	47	4	science	science	NOUN
easat-8918	47	5	and	and	CCONJ
easat-8918	47	6	technology	technology	NOUN
easat-8918	47	7	issn	issn	PROPN
easat-8918	47	8	:	:	PUNCT
easat-8918	47	9	2576	2576	NUM
easat-8918	47	10	-	-	SYM
easat-8918	47	11	8484	8484	NUM
easat-8918	47	12	vol	vol	NOUN
easat-8918	47	13	.	.	PROPN
easat-8918	48	1	9	9	NUM
easat-8918	48	2	,	,	PUNCT
easat-8918	48	3	no	no	INTJ
easat-8918	48	4	.	.	NOUN
easat-8918	48	5	7	7	NUM
easat-8918	48	6	:	:	SYM
easat-8918	48	7	1345	1345	NUM
easat-8918	48	8	-	-	SYM
easat-8918	48	9	1359	1359	NUM
easat-8918	48	10	,	,	PUNCT
easat-8918	48	11	2025	2025	NUM
easat-8918	48	12	doi	doi	NOUN
easat-8918	48	13	:	:	PUNCT
easat-8918	48	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	48	15	©	©	PROPN
easat-8918	48	16	2025	2025	NUM
easat-8918	48	17	by	by	ADP
easat-8918	48	18	the	the	DET
easat-8918	48	19	author	author	NOUN
easat-8918	48	20	;	;	PUNCT
easat-8918	48	21	licensee	licensee	PROPN
easat-8918	48	22	learning	learning	NOUN
easat-8918	48	23	gate	gate	NOUN
easat-8918	48	24	to	to	PART
easat-8918	48	25	enhance	enhance	VERB
easat-8918	48	26	performance	performance	NOUN
easat-8918	48	27	and	and	CCONJ
easat-8918	48	28	reduce	reduce	VERB
easat-8918	48	29	computational	computational	ADJ
easat-8918	48	30	costs	cost	NOUN
easat-8918	48	31	,	,	PUNCT
easat-8918	48	32	deeper	deep	ADJ
easat-8918	48	33	network	network	NOUN
easat-8918	48	34	architectures	architecture	NOUN
easat-8918	48	35	with	with	ADP
easat-8918	48	36	smaller	small	ADJ
easat-8918	48	37	filters	filter	NOUN
easat-8918	48	38	were	be	AUX
easat-8918	48	39	preferred	prefer	VERB
easat-8918	48	40	.	.	PUNCT
easat-8918	49	1	specifically	specifically	ADV
easat-8918	49	2	,	,	PUNCT
easat-8918	49	3	vgg	vgg	PROPN
easat-8918	50	1	[	[	X
easat-8918	50	2	19	19	NUM
easat-8918	50	3	]	]	PUNCT
easat-8918	50	4	stacked	stack	VERB
easat-8918	50	5	more	more	ADV
easat-8918	50	6	convolutional	convolutional	ADJ
easat-8918	50	7	layers	layer	NOUN
easat-8918	50	8	with	with	ADP
easat-8918	50	9	small	small	ADJ
easat-8918	50	10	kernel	kernel	NOUN
easat-8918	50	11	sizes	size	NOUN
easat-8918	50	12	,	,	PUNCT
easat-8918	50	13	as	as	SCONJ
easat-8918	50	14	illustrated	illustrate	VERB
easat-8918	50	15	in	in	ADP
easat-8918	50	16	figure	figure	NOUN
easat-8918	50	17	2	2	NUM
easat-8918	50	18	.	.	PUNCT
easat-8918	50	19	figure	figure	NOUN
easat-8918	50	20	2	2	NUM
easat-8918	50	21	.	.	PUNCT
easat-8918	51	1	the	the	DET
easat-8918	51	2	architecture	architecture	NOUN
easat-8918	51	3	of	of	ADP
easat-8918	51	4	the	the	DET
easat-8918	51	5	vgg	vgg	NOUN
easat-8918	51	6	.	.	PUNCT
easat-8918	52	1	as	as	SCONJ
easat-8918	52	2	accuracy	accuracy	NOUN
easat-8918	52	3	improved	improve	VERB
easat-8918	52	4	,	,	PUNCT
easat-8918	52	5	research	research	NOUN
easat-8918	52	6	shifted	shift	VERB
easat-8918	52	7	toward	toward	ADP
easat-8918	52	8	increasing	increase	VERB
easat-8918	52	9	the	the	DET
easat-8918	52	10	width	width	NOUN
easat-8918	52	11	of	of	ADP
easat-8918	52	12	networks	network	NOUN
easat-8918	52	13	.	.	PUNCT
easat-8918	53	1	googlenet	googlenet	PROPN
easat-8918	53	2	szegedy	szegedy	PROPN
easat-8918	53	3	,	,	PUNCT
easat-8918	53	4	et	et	PROPN
easat-8918	53	5	al	al	PROPN
easat-8918	53	6	.	.	PUNCT
easat-8918	54	1	[	[	X
easat-8918	54	2	20	20	NUM
easat-8918	54	3	]	]	PUNCT
easat-8918	54	4	expanded	expand	VERB
easat-8918	54	5	the	the	DET
easat-8918	54	6	width	width	NOUN
easat-8918	54	7	of	of	ADP
easat-8918	54	8	cnn	cnn	PROPN
easat-8918	54	9	architectures	architecture	NOUN
easat-8918	54	10	to	to	PART
easat-8918	54	11	enhance	enhance	VERB
easat-8918	54	12	the	the	DET
easat-8918	54	13	performance	performance	NOUN
easat-8918	54	14	of	of	ADP
easat-8918	54	15	imageprocessing	imageprocesse	VERB
easat-8918	54	16	applications	application	NOUN
easat-8918	54	17	.	.	PUNCT
easat-8918	55	1	additionally	additionally	ADV
easat-8918	55	2	,	,	PUNCT
easat-8918	55	3	it	it	PRON
easat-8918	55	4	reduced	reduce	VERB
easat-8918	55	5	the	the	DET
easat-8918	55	6	number	number	NOUN
easat-8918	55	7	of	of	ADP
easat-8918	55	8	parameters	parameter	NOUN
easat-8918	55	9	and	and	CCONJ
easat-8918	55	10	computational	computational	ADJ
easat-8918	55	11	costs	cost	NOUN
easat-8918	55	12	by	by	ADP
easat-8918	55	13	replacing	replace	VERB
easat-8918	55	14	large	large	ADJ
easat-8918	55	15	convolutional	convolutional	ADJ
easat-8918	55	16	kernels	kernel	NOUN
easat-8918	55	17	with	with	ADP
easat-8918	55	18	two	two	NUM
easat-8918	55	19	smaller	small	ADJ
easat-8918	55	20	ones	one	NOUN
easat-8918	55	21	.	.	PUNCT
easat-8918	56	1	the	the	DET
easat-8918	56	2	fundamental	fundamental	ADJ
easat-8918	56	3	convolutional	convolutional	ADJ
easat-8918	56	4	block	block	NOUN
easat-8918	56	5	in	in	ADP
easat-8918	56	6	googlenet	googlenet	NOUN
easat-8918	56	7	is	be	AUX
easat-8918	56	8	called	call	VERB
easat-8918	56	9	the	the	DET
easat-8918	56	10	inception	inception	ADJ
easat-8918	56	11	block	block	NOUN
easat-8918	56	12	(	(	PUNCT
easat-8918	56	13	see	see	VERB
easat-8918	56	14	figure	figure	NOUN
easat-8918	56	15	3	3	NUM
easat-8918	56	16	)	)	PUNCT
easat-8918	56	17	,	,	PUNCT
easat-8918	56	18	and	and	CCONJ
easat-8918	56	19	the	the	DET
easat-8918	56	20	googlenet	googlenet	NOUN
easat-8918	56	21	architecture	architecture	NOUN
easat-8918	56	22	is	be	AUX
easat-8918	56	23	shown	show	VERB
easat-8918	56	24	in	in	ADP
easat-8918	56	25	figure	figure	NOUN
easat-8918	56	26	4	4	NUM
easat-8918	56	27	.	.	PUNCT
easat-8918	56	28	figure	figure	VERB
easat-8918	56	29	3	3	NUM
easat-8918	56	30	.	.	PUNCT
easat-8918	57	1	the	the	DET
easat-8918	57	2	inception	inception	ADJ
easat-8918	57	3	block	block	NOUN
easat-8918	57	4	.	.	PUNCT
easat-8918	58	1	1348	1348	NUM
easat-8918	58	2	edelweiss	edelweiss	PROPN
easat-8918	58	3	applied	apply	VERB
easat-8918	58	4	science	science	NOUN
easat-8918	58	5	and	and	CCONJ
easat-8918	58	6	technology	technology	NOUN
easat-8918	58	7	issn	issn	PROPN
easat-8918	58	8	:	:	PUNCT
easat-8918	58	9	2576	2576	NUM
easat-8918	58	10	-	-	SYM
easat-8918	58	11	8484	8484	NUM
easat-8918	58	12	vol	vol	NOUN
easat-8918	58	13	.	.	PROPN
easat-8918	59	1	9	9	NUM
easat-8918	59	2	,	,	PUNCT
easat-8918	59	3	no	no	INTJ
easat-8918	59	4	.	.	NOUN
easat-8918	59	5	7	7	NUM
easat-8918	59	6	:	:	SYM
easat-8918	59	7	1345	1345	NUM
easat-8918	59	8	-	-	SYM
easat-8918	59	9	1359	1359	NUM
easat-8918	59	10	,	,	PUNCT
easat-8918	59	11	2025	2025	NUM
easat-8918	59	12	doi	doi	NOUN
easat-8918	59	13	:	:	PUNCT
easat-8918	59	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	59	15	©	©	PROPN
easat-8918	59	16	2025	2025	NUM
easat-8918	59	17	by	by	ADP
easat-8918	59	18	the	the	DET
easat-8918	59	19	author	author	NOUN
easat-8918	59	20	;	;	PUNCT
easat-8918	59	21	licensee	licensee	PROPN
easat-8918	59	22	learning	learn	VERB
easat-8918	59	23	gate	gate	NOUN
easat-8918	59	24	figure	figure	NOUN
easat-8918	59	25	4	4	NUM
easat-8918	59	26	.	.	PUNCT
easat-8918	59	27	googlenet	googlenet	NOUN
easat-8918	59	28	architecture	architecture	NOUN
easat-8918	59	29	.	.	PUNCT
easat-8918	60	1	although	although	SCONJ
easat-8918	60	2	vgg	vgg	NOUN
easat-8918	60	3	and	and	CCONJ
easat-8918	60	4	googlenet	googlenet	NOUN
easat-8918	60	5	methods	method	NOUN
easat-8918	60	6	are	be	AUX
easat-8918	60	7	effective	effective	ADJ
easat-8918	60	8	for	for	ADP
easat-8918	60	9	image	image	NOUN
easat-8918	60	10	applications	application	NOUN
easat-8918	60	11	,	,	PUNCT
easat-8918	60	12	they	they	PRON
easat-8918	60	13	come	come	VERB
easat-8918	60	14	with	with	ADP
easat-8918	60	15	two	two	NUM
easat-8918	60	16	main	main	ADJ
easat-8918	60	17	disadvantages	disadvantage	NOUN
easat-8918	60	18	:	:	PUNCT
easat-8918	60	19	if	if	SCONJ
easat-8918	60	20	the	the	DET
easat-8918	60	21	network	network	NOUN
easat-8918	60	22	is	be	AUX
easat-8918	60	23	too	too	ADV
easat-8918	60	24	deep	deep	ADJ
easat-8918	60	25	,	,	PUNCT
easat-8918	60	26	it	it	PRON
easat-8918	60	27	may	may	AUX
easat-8918	60	28	suffer	suffer	VERB
easat-8918	60	29	from	from	ADP
easat-8918	60	30	vanishing	vanish	VERB
easat-8918	60	31	or	or	CCONJ
easat-8918	60	32	exploding	explode	VERB
easat-8918	60	33	gradients	gradient	NOUN
easat-8918	60	34	;	;	PUNCT
easat-8918	60	35	if	if	SCONJ
easat-8918	60	36	the	the	DET
easat-8918	60	37	network	network	NOUN
easat-8918	60	38	is	be	AUX
easat-8918	60	39	too	too	ADV
easat-8918	60	40	wide	wide	ADJ
easat-8918	60	41	,	,	PUNCT
easat-8918	60	42	it	it	PRON
easat-8918	60	43	may	may	AUX
easat-8918	60	44	be	be	AUX
easat-8918	60	45	prone	prone	ADJ
easat-8918	60	46	to	to	ADP
easat-8918	60	47	overfitting	overfitte	VERB
easat-8918	60	48	.	.	PUNCT
easat-8918	61	1	to	to	PART
easat-8918	61	2	overcome	overcome	VERB
easat-8918	61	3	these	these	DET
easat-8918	61	4	issues	issue	NOUN
easat-8918	61	5	,	,	PUNCT
easat-8918	61	6	resnet	resnet	VERB
easat-8918	61	7	he	he	PRON
easat-8918	61	8	,	,	PUNCT
easat-8918	61	9	et	et	PROPN
easat-8918	61	10	al	al	PROPN
easat-8918	61	11	.	.	PUNCT
easat-8918	62	1	[	[	X
easat-8918	62	2	21	21	NUM
easat-8918	62	3	]	]	PUNCT
easat-8918	62	4	was	be	AUX
easat-8918	62	5	proposed	propose	VERB
easat-8918	62	6	in	in	ADP
easat-8918	62	7	2016	2016	NUM
easat-8918	62	8	.	.	PUNCT
easat-8918	63	1	to	to	PART
easat-8918	63	2	improve	improve	VERB
easat-8918	63	3	image	image	NOUN
easat-8918	63	4	recognition	recognition	NOUN
easat-8918	63	5	performance	performance	NOUN
easat-8918	63	6	,	,	PUNCT
easat-8918	63	7	residual	residual	ADJ
easat-8918	63	8	learning	learning	NOUN
easat-8918	63	9	was	be	AUX
easat-8918	63	10	introduced	introduce	VERB
easat-8918	63	11	into	into	ADP
easat-8918	63	12	each	each	DET
easat-8918	63	13	block	block	NOUN
easat-8918	63	14	of	of	ADP
easat-8918	63	15	the	the	DET
easat-8918	63	16	resnet	resnet	NOUN
easat-8918	63	17	architecture	architecture	NOUN
easat-8918	63	18	.	.	PUNCT
easat-8918	64	1	figure	figure	NOUN
easat-8918	64	2	5	5	NUM
easat-8918	64	3	shows	show	VERB
easat-8918	64	4	the	the	DET
easat-8918	64	5	structure	structure	NOUN
easat-8918	64	6	of	of	ADP
easat-8918	64	7	resnet-18	resnet-18	PROPN
easat-8918	64	8	,	,	PUNCT
easat-8918	64	9	while	while	SCONJ
easat-8918	64	10	the	the	DET
easat-8918	64	11	basic	basic	ADJ
easat-8918	64	12	building	building	NOUN
easat-8918	64	13	block	block	NOUN
easat-8918	64	14	of	of	ADP
easat-8918	64	15	resnet	resnet	NOUN
easat-8918	64	16	is	be	AUX
easat-8918	64	17	illustrated	illustrate	VERB
easat-8918	64	18	in	in	ADP
easat-8918	64	19	figure	figure	NOUN
easat-8918	64	20	6	6	NUM
easat-8918	64	21	.	.	PUNCT
easat-8918	64	22	figure	figure	NOUN
easat-8918	64	23	5	5	NUM
easat-8918	64	24	.	.	PUNCT
easat-8918	64	25	resnet-18	resnet-18	NOUN
easat-8918	64	26	architecture	architecture	NOUN
easat-8918	64	27	.	.	PUNCT
easat-8918	65	1	figure	figure	NOUN
easat-8918	65	2	6	6	NUM
easat-8918	65	3	.	.	PUNCT
easat-8918	65	4	resnet	resnet	ADJ
easat-8918	65	5	block	block	NOUN
easat-8918	65	6	.	.	PUNCT
easat-8918	66	1	traditional	traditional	ADJ
easat-8918	66	2	machine	machine	NOUN
easat-8918	66	3	learning	learn	VERB
easat-8918	66	4	algorithms	algorithm	NOUN
easat-8918	66	5	have	have	AUX
easat-8918	66	6	long	long	ADV
easat-8918	66	7	been	be	AUX
easat-8918	66	8	employed	employ	VERB
easat-8918	66	9	in	in	ADP
easat-8918	66	10	diverse	diverse	ADJ
easat-8918	66	11	image	image	NOUN
easat-8918	66	12	processing	processing	NOUN
easat-8918	66	13	applications	application	NOUN
easat-8918	66	14	[	[	X
easat-8918	66	15	22	22	NUM
easat-8918	66	16	-	-	SYM
easat-8918	66	17	27	27	NUM
easat-8918	66	18	]	]	PUNCT
easat-8918	66	19	.	.	PUNCT
easat-8918	67	1	for	for	ADP
easat-8918	67	2	nearly	nearly	ADV
easat-8918	67	3	a	a	PRON
easat-8918	67	4	decade	decade	NOUN
easat-8918	67	5	,	,	PUNCT
easat-8918	67	6	deep	deep	ADJ
easat-8918	67	7	networks	network	NOUN
easat-8918	67	8	have	have	AUX
easat-8918	67	9	seen	see	VERB
easat-8918	67	10	widespread	widespread	ADJ
easat-8918	67	11	use	use	NOUN
easat-8918	67	12	in	in	ADP
easat-8918	67	13	real	real	ADJ
easat-8918	67	14	-	-	PUNCT
easat-8918	67	15	world	world	NOUN
easat-8918	67	16	image	image	NOUN
easat-8918	67	17	1349	1349	NUM
easat-8918	67	18	edelweiss	edelweiss	PROPN
easat-8918	67	19	applied	apply	VERB
easat-8918	67	20	science	science	NOUN
easat-8918	67	21	and	and	CCONJ
easat-8918	67	22	technology	technology	NOUN
easat-8918	67	23	issn	issn	PROPN
easat-8918	67	24	:	:	PUNCT
easat-8918	67	25	2576	2576	NUM
easat-8918	67	26	-	-	SYM
easat-8918	67	27	8484	8484	NUM
easat-8918	67	28	vol	vol	NOUN
easat-8918	67	29	.	.	PROPN
easat-8918	68	1	9	9	NUM
easat-8918	68	2	,	,	PUNCT
easat-8918	68	3	no	no	INTJ
easat-8918	68	4	.	.	NOUN
easat-8918	68	5	7	7	NUM
easat-8918	68	6	:	:	SYM
easat-8918	68	7	1345	1345	NUM
easat-8918	68	8	-	-	SYM
easat-8918	68	9	1359	1359	NUM
easat-8918	68	10	,	,	PUNCT
easat-8918	68	11	2025	2025	NUM
easat-8918	68	12	doi	doi	NOUN
easat-8918	68	13	:	:	PUNCT
easat-8918	68	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	68	15	©	©	PROPN
easat-8918	68	16	2025	2025	NUM
easat-8918	68	17	by	by	ADP
easat-8918	68	18	the	the	DET
easat-8918	68	19	author	author	NOUN
easat-8918	68	20	;	;	PUNCT
easat-8918	68	21	licensee	licensee	PROPN
easat-8918	68	22	learning	learning	PROPN
easat-8918	68	23	gate	gate	PROPN
easat-8918	68	24	applications	application	NOUN
easat-8918	68	25	,	,	PUNCT
easat-8918	68	26	including	include	VERB
easat-8918	68	27	face	face	NOUN
easat-8918	68	28	recognition	recognition	NOUN
easat-8918	68	29	[	[	X
easat-8918	68	30	28	28	NUM
easat-8918	68	31	-	-	SYM
easat-8918	68	32	32	32	NUM
easat-8918	68	33	]	]	PUNCT
easat-8918	68	34	and	and	CCONJ
easat-8918	68	35	medical	medical	ADJ
easat-8918	68	36	diagnosis	diagnosis	NOUN
easat-8918	68	37	[	[	X
easat-8918	68	38	33	33	NUM
easat-8918	68	39	-	-	SYM
easat-8918	68	40	37	37	NUM
easat-8918	68	41	]	]	PUNCT
easat-8918	68	42	.	.	PUNCT
easat-8918	69	1	however	however	ADV
easat-8918	69	2	,	,	PUNCT
easat-8918	69	3	in	in	ADP
easat-8918	69	4	many	many	ADJ
easat-8918	69	5	applications	application	NOUN
easat-8918	69	6	,	,	PUNCT
easat-8918	69	7	such	such	ADJ
easat-8918	69	8	as	as	ADP
easat-8918	69	9	real	real	ADJ
easat-8918	69	10	noisy	noisy	ADJ
easat-8918	69	11	image	image	NOUN
easat-8918	69	12	scenarios	scenario	NOUN
easat-8918	69	13	,	,	PUNCT
easat-8918	69	14	the	the	DET
easat-8918	69	15	captured	capture	VERB
easat-8918	69	16	images	image	NOUN
easat-8918	69	17	are	be	AUX
easat-8918	69	18	often	often	ADV
easat-8918	69	19	insufficient	insufficient	ADJ
easat-8918	69	20	in	in	ADP
easat-8918	69	21	quality	quality	NOUN
easat-8918	69	22	,	,	PUNCT
easat-8918	69	23	and	and	CCONJ
easat-8918	69	24	deep	deep	ADJ
easat-8918	69	25	cnns	cnn	NOUN
easat-8918	69	26	tend	tend	VERB
easat-8918	69	27	to	to	PART
easat-8918	69	28	show	show	VERB
easat-8918	69	29	limited	limited	ADJ
easat-8918	69	30	performance	performance	NOUN
easat-8918	69	31	.	.	PUNCT
easat-8918	70	1	to	to	PART
easat-8918	70	2	address	address	VERB
easat-8918	70	3	this	this	PRON
easat-8918	70	4	,	,	PUNCT
easat-8918	70	5	generative	generative	ADJ
easat-8918	70	6	adversarial	adversarial	ADJ
easat-8918	70	7	networks	network	NOUN
easat-8918	70	8	(	(	PUNCT
easat-8918	70	9	gans	gans	PROPN
easat-8918	70	10	)	)	PUNCT
easat-8918	70	11	radford	radford	PROPN
easat-8918	70	12	,	,	PUNCT
easat-8918	70	13	et	et	PROPN
easat-8918	70	14	al	al	PROPN
easat-8918	70	15	.	.	PUNCT
easat-8918	71	1	[	[	X
easat-8918	71	2	38	38	NUM
easat-8918	71	3	]	]	PUNCT
easat-8918	71	4	were	be	AUX
easat-8918	71	5	developed	develop	VERB
easat-8918	71	6	.	.	PUNCT
easat-8918	72	1	gans	gan	NOUN
easat-8918	72	2	consist	consist	VERB
easat-8918	72	3	of	of	ADP
easat-8918	72	4	two	two	NUM
easat-8918	72	5	networks	network	NOUN
easat-8918	72	6	:	:	PUNCT
easat-8918	72	7	a	a	DET
easat-8918	72	8	generator	generator	NOUN
easat-8918	72	9	and	and	CCONJ
easat-8918	72	10	a	a	DET
easat-8918	72	11	discriminator	discriminator	NOUN
easat-8918	72	12	.	.	PUNCT
easat-8918	73	1	the	the	DET
easat-8918	73	2	generator	generator	NOUN
easat-8918	73	3	is	be	AUX
easat-8918	73	4	used	use	VERB
easat-8918	73	5	to	to	PART
easat-8918	73	6	produce	produce	VERB
easat-8918	73	7	samples	sample	NOUN
easat-8918	73	8	based	base	VERB
easat-8918	73	9	on	on	ADP
easat-8918	73	10	input	input	NOUN
easat-8918	73	11	data	datum	NOUN
easat-8918	73	12	,	,	PUNCT
easat-8918	73	13	while	while	SCONJ
easat-8918	73	14	the	the	DET
easat-8918	73	15	discriminator	discriminator	NOUN
easat-8918	73	16	evaluates	evaluate	VERB
easat-8918	73	17	the	the	DET
easat-8918	73	18	authenticity	authenticity	NOUN
easat-8918	73	19	of	of	ADP
easat-8918	73	20	both	both	CCONJ
easat-8918	73	21	the	the	DET
easat-8918	73	22	input	input	NOUN
easat-8918	73	23	samples	sample	NOUN
easat-8918	73	24	and	and	CCONJ
easat-8918	73	25	the	the	DET
easat-8918	73	26	generated	generate	VERB
easat-8918	73	27	ones	one	NOUN
easat-8918	73	28	.	.	PUNCT
easat-8918	74	1	these	these	DET
easat-8918	74	2	two	two	NUM
easat-8918	74	3	networks	network	NOUN
easat-8918	74	4	work	work	VERB
easat-8918	74	5	in	in	ADP
easat-8918	74	6	opposition	opposition	NOUN
easat-8918	74	7	:	:	PUNCT
easat-8918	74	8	if	if	SCONJ
easat-8918	74	9	the	the	DET
easat-8918	74	10	discriminator	discriminator	NOUN
easat-8918	74	11	can	can	AUX
easat-8918	74	12	accurately	accurately	ADV
easat-8918	74	13	distinguish	distinguish	VERB
easat-8918	74	14	real	real	ADJ
easat-8918	74	15	samples	sample	NOUN
easat-8918	74	16	and	and	CCONJ
easat-8918	74	17	the	the	DET
easat-8918	74	18	generator	generator	NOUN
easat-8918	74	19	can	can	AUX
easat-8918	74	20	produce	produce	VERB
easat-8918	74	21	convincing	convince	VERB
easat-8918	74	22	fake	fake	ADJ
easat-8918	74	23	ones	one	NOUN
easat-8918	74	24	,	,	PUNCT
easat-8918	74	25	the	the	DET
easat-8918	74	26	model	model	NOUN
easat-8918	74	27	is	be	AUX
easat-8918	74	28	considered	consider	VERB
easat-8918	74	29	successfully	successfully	ADV
easat-8918	74	30	trained	train	VERB
easat-8918	74	31	.	.	PUNCT
easat-8918	75	1	the	the	DET
easat-8918	75	2	architecture	architecture	NOUN
easat-8918	75	3	of	of	ADP
easat-8918	75	4	a	a	DET
easat-8918	75	5	gan	gan	PROPN
easat-8918	75	6	is	be	AUX
easat-8918	75	7	illustrated	illustrate	VERB
easat-8918	75	8	in	in	ADP
easat-8918	75	9	figure	figure	NOUN
easat-8918	75	10	7	7	NUM
easat-8918	75	11	.	.	PUNCT
easat-8918	76	1	due	due	ADP
easat-8918	76	2	to	to	ADP
easat-8918	76	3	their	their	PRON
easat-8918	76	4	ability	ability	NOUN
easat-8918	76	5	to	to	PART
easat-8918	76	6	generate	generate	VERB
easat-8918	76	7	complementary	complementary	ADJ
easat-8918	76	8	training	training	NOUN
easat-8918	76	9	examples	example	NOUN
easat-8918	76	10	,	,	PUNCT
easat-8918	76	11	gans	gan	NOUN
easat-8918	76	12	are	be	AUX
easat-8918	76	13	highly	highly	ADV
easat-8918	76	14	effective	effective	ADJ
easat-8918	76	15	in	in	ADP
easat-8918	76	16	working	work	VERB
easat-8918	76	17	with	with	ADP
easat-8918	76	18	small	small	ADJ
easat-8918	76	19	datasets	dataset	NOUN
easat-8918	76	20	,	,	PUNCT
easat-8918	76	21	especially	especially	ADV
easat-8918	76	22	in	in	ADP
easat-8918	76	23	tasks	task	NOUN
easat-8918	76	24	like	like	ADP
easat-8918	76	25	face	face	NOUN
easat-8918	76	26	recognition	recognition	NOUN
easat-8918	77	1	[	[	X
easat-8918	77	2	39	39	NUM
easat-8918	77	3	]	]	PUNCT
easat-8918	77	4	and	and	CCONJ
easat-8918	77	5	denoising	denoise	VERB
easat-8918	77	6	complex	complex	ADJ
easat-8918	77	7	noisy	noisy	ADJ
easat-8918	77	8	images	image	NOUN
easat-8918	78	1	[	[	X
easat-8918	78	2	40	40	NUM
easat-8918	78	3	]	]	PUNCT
easat-8918	78	4	.	.	PUNCT
easat-8918	79	1	figure	figure	NOUN
easat-8918	79	2	7	7	NUM
easat-8918	79	3	.	.	PUNCT
easat-8918	79	4	gan	gan	ADJ
easat-8918	79	5	architecture	architecture	NOUN
easat-8918	79	6	.	.	PUNCT
easat-8918	80	1	2.2	2.2	NUM
easat-8918	80	2	.	.	PUNCT
easat-8918	81	1	deep	deep	ADJ
easat-8918	81	2	learning	learning	NOUN
easat-8918	81	3	for	for	ADP
easat-8918	81	4	image	image	NOUN
easat-8918	81	5	denoising	denoise	VERB
easat-8918	81	6	with	with	ADP
easat-8918	81	7	the	the	DET
easat-8918	81	8	introduction	introduction	NOUN
easat-8918	81	9	of	of	ADP
easat-8918	81	10	the	the	DET
easat-8918	81	11	networks	network	NOUN
easat-8918	81	12	mentioned	mention	VERB
easat-8918	81	13	above	above	ADV
easat-8918	81	14	,	,	PUNCT
easat-8918	81	15	deep	deep	ADJ
easat-8918	81	16	learning	learning	NOUN
easat-8918	81	17	techniques	technique	NOUN
easat-8918	81	18	have	have	AUX
easat-8918	81	19	attracted	attract	VERB
easat-8918	81	20	considerable	considerable	ADJ
easat-8918	81	21	attention	attention	NOUN
easat-8918	81	22	in	in	ADP
easat-8918	81	23	the	the	DET
easat-8918	81	24	field	field	NOUN
easat-8918	81	25	of	of	ADP
easat-8918	81	26	image	image	NOUN
easat-8918	81	27	denoising	denoising	NOUN
easat-8918	81	28	.	.	PUNCT
easat-8918	82	1	researchers	researcher	NOUN
easat-8918	82	2	have	have	AUX
easat-8918	82	3	explored	explore	VERB
easat-8918	82	4	deep	deep	ADJ
easat-8918	82	5	neural	neural	ADJ
easat-8918	82	6	networks	network	NOUN
easat-8918	82	7	to	to	PART
easat-8918	82	8	address	address	VERB
easat-8918	82	9	the	the	DET
easat-8918	82	10	problem	problem	NOUN
easat-8918	82	11	of	of	ADP
easat-8918	82	12	noise	noise	NOUN
easat-8918	82	13	removal	removal	NOUN
easat-8918	82	14	.	.	PUNCT
easat-8918	83	1	however	however	ADV
easat-8918	83	2	,	,	PUNCT
easat-8918	83	3	there	there	PRON
easat-8918	83	4	are	be	VERB
easat-8918	83	5	significant	significant	ADJ
easat-8918	83	6	differences	difference	NOUN
easat-8918	83	7	among	among	ADP
easat-8918	83	8	various	various	ADJ
easat-8918	83	9	deep	deep	ADJ
easat-8918	83	10	learning	learning	NOUN
easat-8918	83	11	approaches	approach	NOUN
easat-8918	83	12	to	to	ADP
easat-8918	83	13	image	image	NOUN
easat-8918	83	14	denoising	denoise	VERB
easat-8918	83	15	.	.	PUNCT
easat-8918	84	1	specifically	specifically	ADV
easat-8918	84	2	,	,	PUNCT
easat-8918	84	3	discriminative	discriminative	NOUN
easat-8918	84	4	learning	learning	NOUN
easat-8918	84	5	based	base	VERB
easat-8918	84	6	on	on	ADP
easat-8918	84	7	deep	deep	ADJ
easat-8918	84	8	learning	learning	NOUN
easat-8918	84	9	has	have	AUX
easat-8918	84	10	proven	prove	VERB
easat-8918	84	11	effective	effective	ADJ
easat-8918	84	12	in	in	ADP
easat-8918	84	13	tackling	tackle	VERB
easat-8918	84	14	the	the	DET
easat-8918	84	15	problem	problem	NOUN
easat-8918	84	16	of	of	ADP
easat-8918	84	17	gaussian	gaussian	ADJ
easat-8918	84	18	noise	noise	NOUN
easat-8918	84	19	.	.	PUNCT
easat-8918	85	1	in	in	ADP
easat-8918	85	2	jain	jain	PROPN
easat-8918	85	3	and	and	CCONJ
easat-8918	85	4	seung	seung	PROPN
easat-8918	85	5	[	[	X
easat-8918	85	6	41	41	NUM
easat-8918	85	7	]	]	PUNCT
easat-8918	85	8	proposed	propose	VERB
easat-8918	85	9	using	use	VERB
easat-8918	85	10	convolutional	convolutional	ADJ
easat-8918	85	11	neural	neural	ADJ
easat-8918	85	12	networks	network	NOUN
easat-8918	85	13	(	(	PUNCT
easat-8918	85	14	cnns	cnns	PROPN
easat-8918	85	15	)	)	PUNCT
easat-8918	85	16	for	for	ADP
easat-8918	85	17	image	image	NOUN
easat-8918	85	18	denoising	denoising	NOUN
easat-8918	85	19	,	,	PUNCT
easat-8918	85	20	arguing	argue	VERB
easat-8918	85	21	that	that	SCONJ
easat-8918	85	22	cnns	cnn	NOUN
easat-8918	85	23	can	can	AUX
easat-8918	85	24	provide	provide	VERB
easat-8918	85	25	representations	representation	NOUN
easat-8918	85	26	like	like	ADP
easat-8918	85	27	,	,	PUNCT
easat-8918	85	28	or	or	CCONJ
easat-8918	85	29	even	even	ADV
easat-8918	85	30	better	well	ADJ
easat-8918	85	31	than	than	ADP
easat-8918	85	32	,	,	PUNCT
easat-8918	85	33	those	those	PRON
easat-8918	85	34	of	of	ADP
easat-8918	85	35	markov	markov	NOUN
easat-8918	85	36	random	random	ADJ
easat-8918	85	37	field	field	NOUN
easat-8918	85	38	(	(	PUNCT
easat-8918	85	39	mrf	mrf	NOUN
easat-8918	85	40	)	)	PUNCT
easat-8918	85	41	models	model	NOUN
easat-8918	86	1	[	[	X
easat-8918	86	2	42	42	NUM
easat-8918	86	3	]	]	PUNCT
easat-8918	86	4	.	.	PUNCT
easat-8918	87	1	in	in	ADP
easat-8918	87	2	burger	burger	NOUN
easat-8918	87	3	,	,	PUNCT
easat-8918	87	4	et	et	PROPN
easat-8918	87	5	al	al	PROPN
easat-8918	87	6	.	.	PUNCT
easat-8918	88	1	[	[	X
easat-8918	88	2	43	43	NUM
easat-8918	88	3	]	]	PUNCT
easat-8918	88	4	.	.	PUNCT
easat-8918	89	1	multi	multi	ADJ
easat-8918	89	2	-	-	ADJ
easat-8918	89	3	layer	layer	ADJ
easat-8918	89	4	perceptrons	perceptron	NOUN
easat-8918	89	5	(	(	PUNCT
easat-8918	89	6	mlps	mlp	NOUN
easat-8918	89	7	)	)	PUNCT
easat-8918	89	8	were	be	AUX
easat-8918	89	9	successfully	successfully	ADV
easat-8918	89	10	applied	apply	VERB
easat-8918	89	11	to	to	ADP
easat-8918	89	12	image	image	NOUN
easat-8918	89	13	denoising	denoise	VERB
easat-8918	89	14	.	.	PUNCT
easat-8918	90	1	in	in	ADP
easat-8918	90	2	xie	xie	PROPN
easat-8918	90	3	,	,	PUNCT
easat-8918	90	4	et	et	PROPN
easat-8918	90	5	al	al	PROPN
easat-8918	90	6	.	.	PUNCT
easat-8918	91	1	[	[	X
easat-8918	91	2	44	44	NUM
easat-8918	91	3	]	]	PUNCT
easat-8918	91	4	stacked	stack	VERB
easat-8918	91	5	sparse	sparse	ADJ
easat-8918	91	6	denoising	denoising	NOUN
easat-8918	91	7	autoencoders	autoencoder	NOUN
easat-8918	91	8	were	be	AUX
easat-8918	91	9	adopted	adopt	VERB
easat-8918	91	10	for	for	ADP
easat-8918	91	11	gaussian	gaussian	ADJ
easat-8918	91	12	noise	noise	NOUN
easat-8918	91	13	removal	removal	NOUN
easat-8918	91	14	and	and	CCONJ
easat-8918	91	15	achieved	achieve	VERB
easat-8918	91	16	results	result	NOUN
easat-8918	91	17	comparable	comparable	ADJ
easat-8918	91	18	to	to	ADP
easat-8918	91	19	k	k	X
easat-8918	91	20	-	-	PUNCT
easat-8918	91	21	svd	svd	PROPN
easat-8918	91	22	[	[	X
easat-8918	91	23	45	45	NUM
easat-8918	91	24	]	]	PUNCT
easat-8918	91	25	.	.	PUNCT
easat-8918	92	1	in	in	ADP
easat-8918	92	2	chen	chen	PROPN
easat-8918	92	3	and	and	CCONJ
easat-8918	92	4	pock	pock	NOUN
easat-8918	92	5	[	[	X
easat-8918	92	6	46	46	NUM
easat-8918	92	7	]	]	PUNCT
easat-8918	92	8	a	a	DET
easat-8918	92	9	trainable	trainable	ADJ
easat-8918	92	10	nonlinear	nonlinear	ADJ
easat-8918	92	11	reaction	reaction	NOUN
easat-8918	92	12	diffusion	diffusion	NOUN
easat-8918	92	13	(	(	PUNCT
easat-8918	92	14	tnrd	tnrd	NOUN
easat-8918	92	15	)	)	PUNCT
easat-8918	92	16	model	model	NOUN
easat-8918	92	17	was	be	AUX
easat-8918	92	18	proposed	propose	VERB
easat-8918	92	19	,	,	PUNCT
easat-8918	92	20	which	which	PRON
easat-8918	92	21	can	can	AUX
easat-8918	92	22	be	be	AUX
easat-8918	92	23	expressed	express	VERB
easat-8918	92	24	as	as	ADP
easat-8918	92	25	a	a	DET
easat-8918	92	26	feedforward	feedforward	ADJ
easat-8918	92	27	deep	deep	ADJ
easat-8918	92	28	network	network	NOUN
easat-8918	92	29	by	by	ADP
easat-8918	92	30	unrolling	unroll	VERB
easat-8918	92	31	a	a	DET
easat-8918	92	32	fixed	fix	VERB
easat-8918	92	33	number	number	NOUN
easat-8918	92	34	of	of	ADP
easat-8918	92	35	gradient	gradient	ADJ
easat-8918	92	36	descent	descent	NOUN
easat-8918	92	37	inference	inference	NOUN
easat-8918	92	38	steps	step	NOUN
easat-8918	92	39	.	.	PUNCT
easat-8918	93	1	among	among	ADP
easat-8918	93	2	these	these	DET
easat-8918	93	3	deep	deep	ADJ
easat-8918	93	4	neural	neural	ADJ
easat-8918	93	5	network	network	NOUN
easat-8918	93	6	-	-	PUNCT
easat-8918	93	7	based	base	VERB
easat-8918	93	8	approaches	approach	NOUN
easat-8918	93	9	,	,	PUNCT
easat-8918	93	10	mlp	mlp	PROPN
easat-8918	93	11	and	and	CCONJ
easat-8918	93	12	tnrd	tnrd	NOUN
easat-8918	93	13	have	have	AUX
easat-8918	93	14	shown	show	VERB
easat-8918	93	15	promising	promising	ADJ
easat-8918	93	16	performance	performance	NOUN
easat-8918	93	17	and	and	CCONJ
easat-8918	93	18	can	can	AUX
easat-8918	93	19	approach	approach	VERB
easat-8918	93	20	the	the	DET
easat-8918	93	21	denoising	denoise	VERB
easat-8918	93	22	quality	quality	NOUN
easat-8918	93	23	of	of	ADP
easat-8918	93	24	bm3d	bm3d	PROPN
easat-8918	93	25	[	[	X
easat-8918	93	26	47	47	NUM
easat-8918	93	27	]	]	PUNCT
easat-8918	93	28	.	.	PUNCT
easat-8918	94	1	in	in	ADP
easat-8918	94	2	chen	chen	PROPN
easat-8918	94	3	,	,	PUNCT
easat-8918	94	4	et	et	PROPN
easat-8918	94	5	al	al	PROPN
easat-8918	94	6	.	.	PUNCT
easat-8918	95	1	[	[	X
easat-8918	95	2	48	48	NUM
easat-8918	95	3	]	]	PUNCT
easat-8918	95	4	researchers	researcher	NOUN
easat-8918	95	5	proposed	propose	VERB
easat-8918	95	6	nafnet	nafnet	PROPN
easat-8918	95	7	,	,	PUNCT
easat-8918	95	8	a	a	DET
easat-8918	95	9	parameter	parameter	NOUN
easat-8918	95	10	-	-	PUNCT
easat-8918	95	11	efficient	efficient	ADJ
easat-8918	95	12	network	network	NOUN
easat-8918	95	13	with	with	ADP
easat-8918	95	14	nonlinear	nonlinear	ADJ
easat-8918	95	15	activation	activation	NOUN
easat-8918	95	16	-	-	PUNCT
easat-8918	95	17	free	free	ADJ
easat-8918	95	18	layers	layer	NOUN
easat-8918	95	19	,	,	PUNCT
easat-8918	95	20	which	which	PRON
easat-8918	95	21	outperformed	outperform	VERB
easat-8918	95	22	state	state	NOUN
easat-8918	95	23	-	-	PUNCT
easat-8918	95	24	of	of	ADP
easat-8918	95	25	-	-	PUNCT
easat-8918	95	26	the	the	DET
easat-8918	95	27	-	-	PUNCT
easat-8918	95	28	art	art	NOUN
easat-8918	95	29	(	(	PUNCT
easat-8918	95	30	sota	sota	NOUN
easat-8918	95	31	)	)	PUNCT
easat-8918	95	32	methods	method	NOUN
easat-8918	95	33	and	and	CCONJ
easat-8918	95	34	demonstrated	demonstrate	VERB
easat-8918	95	35	computational	computational	ADJ
easat-8918	95	36	efficiency	efficiency	NOUN
easat-8918	95	37	.	.	PUNCT
easat-8918	96	1	chen	chen	PROPN
easat-8918	96	2	,	,	PUNCT
easat-8918	96	3	et	et	PROPN
easat-8918	96	4	al	al	PROPN
easat-8918	96	5	.	.	PUNCT
easat-8918	97	1	[	[	X
easat-8918	97	2	49	49	NUM
easat-8918	97	3	]	]	PUNCT
easat-8918	97	4	proposed	propose	VERB
easat-8918	97	5	a	a	DET
easat-8918	97	6	simple	simple	ADJ
easat-8918	97	7	gan	gan	NOUN
easat-8918	97	8	that	that	PRON
easat-8918	97	9	takes	take	VERB
easat-8918	97	10	gaussian	gaussian	ADJ
easat-8918	97	11	noise	noise	NOUN
easat-8918	97	12	as	as	ADP
easat-8918	97	13	input	input	NOUN
easat-8918	97	14	to	to	PART
easat-8918	97	15	generate	generate	VERB
easat-8918	97	16	noisy	noisy	ADJ
easat-8918	97	17	patches	patch	NOUN
easat-8918	97	18	.	.	PUNCT
easat-8918	98	1	however	however	ADV
easat-8918	98	2	,	,	PUNCT
easat-8918	98	3	as	as	ADP
easat-8918	98	4	with	with	ADP
easat-8918	98	5	most	most	ADJ
easat-8918	98	6	conventional	conventional	ADJ
easat-8918	98	7	methods	method	NOUN
easat-8918	98	8	,	,	PUNCT
easat-8918	98	9	this	this	DET
easat-8918	98	10	gan	gan	PROPN
easat-8918	98	11	operates	operate	VERB
easat-8918	98	12	at	at	ADP
easat-8918	98	13	the	the	DET
easat-8918	98	14	image	image	NOUN
easat-8918	98	15	level	level	NOUN
easat-8918	98	16	,	,	PUNCT
easat-8918	98	17	treating	treat	VERB
easat-8918	98	18	images	image	NOUN
easat-8918	98	19	as	as	ADP
easat-8918	98	20	samples	sample	NOUN
easat-8918	98	21	and	and	CCONJ
easat-8918	98	22	attempting	attempt	VERB
easat-8918	98	23	to	to	PART
easat-8918	98	24	approximate	approximate	VERB
easat-8918	98	25	the	the	DET
easat-8918	98	26	probability	probability	NOUN
easat-8918	98	27	distribution	distribution	NOUN
easat-8918	98	28	of	of	ADP
easat-8918	98	29	real	real	ADJ
easat-8918	98	30	-	-	PUNCT
easat-8918	98	31	world	world	NOUN
easat-8918	98	32	noisy	noisy	ADJ
easat-8918	98	33	images	image	NOUN
easat-8918	98	34	.	.	PUNCT
easat-8918	99	1	to	to	PART
easat-8918	99	2	overcome	overcome	VERB
easat-8918	99	3	the	the	DET
easat-8918	99	4	limitations	limitation	NOUN
easat-8918	99	5	of	of	ADP
easat-8918	99	6	this	this	DET
easat-8918	99	7	approach	approach	NOUN
easat-8918	99	8	,	,	PUNCT
easat-8918	99	9	a	a	DET
easat-8918	99	10	new	new	ADJ
easat-8918	99	11	pixel	pixel	ADJ
easat-8918	99	12	-	-	PUNCT
easat-8918	99	13	level	level	NOUN
easat-8918	99	14	noise	noise	NOUN
easat-8918	99	15	-	-	PUNCT
easat-8918	99	16	aware	aware	ADJ
easat-8918	99	17	generative	generative	ADJ
easat-8918	99	18	adversarial	adversarial	ADJ
easat-8918	99	19	network	network	NOUN
easat-8918	99	20	(	(	PUNCT
easat-8918	99	21	pngan	pngan	PROPN
easat-8918	99	22	)	)	PUNCT
easat-8918	99	23	was	be	AUX
easat-8918	99	24	introduced	introduce	VERB
easat-8918	99	25	in	in	ADP
easat-8918	99	26	[	[	X
easat-8918	99	27	50	50	NUM
easat-8918	99	28	]	]	PUNCT
easat-8918	99	29	.	.	PUNCT
easat-8918	100	1	this	this	DET
easat-8918	100	2	novel	novel	ADJ
easat-8918	100	3	method	method	NOUN
easat-8918	100	4	performs	perform	VERB
easat-8918	100	5	alignment	alignment	NOUN
easat-8918	100	6	in	in	ADP
easat-8918	100	7	both	both	CCONJ
easat-8918	100	8	the	the	DET
easat-8918	100	9	image	image	NOUN
easat-8918	100	10	space	space	NOUN
easat-8918	100	11	and	and	CCONJ
easat-8918	100	12	noise	noise	NOUN
easat-8918	100	13	space	space	NOUN
easat-8918	100	14	simultaneously	simultaneously	ADV
easat-8918	100	15	during	during	ADP
easat-8918	100	16	training	training	NOUN
easat-8918	100	17	,	,	PUNCT
easat-8918	100	18	leading	lead	VERB
easat-8918	100	19	to	to	ADP
easat-8918	100	20	more	more	ADV
easat-8918	100	21	accurate	accurate	ADJ
easat-8918	100	22	noise	noise	NOUN
easat-8918	100	23	modeling	modeling	NOUN
easat-8918	100	24	.	.	PUNCT
easat-8918	101	1	1350	1350	NUM
easat-8918	101	2	edelweiss	edelweiss	PROPN
easat-8918	101	3	applied	apply	VERB
easat-8918	101	4	science	science	NOUN
easat-8918	101	5	and	and	CCONJ
easat-8918	101	6	technology	technology	NOUN
easat-8918	101	7	issn	issn	PROPN
easat-8918	101	8	:	:	PUNCT
easat-8918	101	9	2576	2576	NUM
easat-8918	101	10	-	-	SYM
easat-8918	101	11	8484	8484	NUM
easat-8918	101	12	vol	vol	NOUN
easat-8918	101	13	.	.	PROPN
easat-8918	102	1	9	9	NUM
easat-8918	102	2	,	,	PUNCT
easat-8918	102	3	no	no	INTJ
easat-8918	102	4	.	.	NOUN
easat-8918	102	5	7	7	NUM
easat-8918	102	6	:	:	SYM
easat-8918	102	7	1345	1345	NUM
easat-8918	102	8	-	-	SYM
easat-8918	102	9	1359	1359	NUM
easat-8918	102	10	,	,	PUNCT
easat-8918	102	11	2025	2025	NUM
easat-8918	102	12	doi	doi	NOUN
easat-8918	102	13	:	:	PUNCT
easat-8918	102	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	102	15	©	©	PROPN
easat-8918	102	16	2025	2025	NUM
easat-8918	102	17	by	by	ADP
easat-8918	102	18	the	the	DET
easat-8918	102	19	author	author	NOUN
easat-8918	102	20	;	;	PUNCT
easat-8918	102	21	licensee	licensee	PROPN
easat-8918	102	22	learning	learn	VERB
easat-8918	102	23	gate	gate	PROPN
easat-8918	102	24	in	in	ADP
easat-8918	102	25	zhang	zhang	PROPN
easat-8918	102	26	,	,	PUNCT
easat-8918	102	27	et	et	PROPN
easat-8918	102	28	al	al	PROPN
easat-8918	102	29	.	.	PUNCT
easat-8918	103	1	[	[	X
easat-8918	103	2	51	51	NUM
easat-8918	103	3	]	]	PUNCT
easat-8918	103	4	the	the	DET
easat-8918	103	5	structure	structure	NOUN
easat-8918	103	6	of	of	ADP
easat-8918	103	7	feedforward	feedforward	NOUN
easat-8918	103	8	denoising	denoise	VERB
easat-8918	103	9	convolutional	convolutional	ADJ
easat-8918	103	10	neural	neural	ADJ
easat-8918	103	11	networks	network	NOUN
easat-8918	103	12	(	(	PUNCT
easat-8918	103	13	dncnns	dncnn	NOUN
easat-8918	103	14	)	)	PUNCT
easat-8918	103	15	was	be	AUX
easat-8918	103	16	explored	explore	VERB
easat-8918	103	17	for	for	ADP
easat-8918	103	18	image	image	NOUN
easat-8918	103	19	noise	noise	NOUN
easat-8918	103	20	removal	removal	NOUN
easat-8918	103	21	.	.	PUNCT
easat-8918	104	1	to	to	PART
easat-8918	104	2	enhance	enhance	VERB
easat-8918	104	3	denoising	denoising	NOUN
easat-8918	104	4	performance	performance	NOUN
easat-8918	104	5	and	and	CCONJ
easat-8918	104	6	accelerate	accelerate	VERB
easat-8918	104	7	the	the	DET
easat-8918	104	8	training	training	NOUN
easat-8918	104	9	process	process	NOUN
easat-8918	104	10	,	,	PUNCT
easat-8918	104	11	residual	residual	ADJ
easat-8918	104	12	learning	learning	NOUN
easat-8918	104	13	and	and	CCONJ
easat-8918	104	14	batch	batch	NOUN
easat-8918	104	15	normalization	normalization	NOUN
easat-8918	104	16	were	be	AUX
easat-8918	104	17	employed	employ	VERB
easat-8918	104	18	.	.	PUNCT
easat-8918	105	1	the	the	DET
easat-8918	105	2	proposed	propose	VERB
easat-8918	105	3	network	network	NOUN
easat-8918	105	4	was	be	AUX
easat-8918	105	5	successfully	successfully	ADV
easat-8918	105	6	applied	apply	VERB
easat-8918	105	7	to	to	ADP
easat-8918	105	8	several	several	ADJ
easat-8918	105	9	common	common	ADJ
easat-8918	105	10	image	image	NOUN
easat-8918	105	11	degradation	degradation	NOUN
easat-8918	105	12	tasks	task	NOUN
easat-8918	105	13	,	,	PUNCT
easat-8918	105	14	including	include	VERB
easat-8918	105	15	gaussian	gaussian	ADJ
easat-8918	105	16	denoising	denoising	NOUN
easat-8918	105	17	,	,	PUNCT
easat-8918	105	18	singleimage	singleimage	NOUN
easat-8918	105	19	super	super	NOUN
easat-8918	105	20	-	-	NOUN
easat-8918	105	21	resolution	resolution	NOUN
easat-8918	105	22	,	,	PUNCT
easat-8918	105	23	and	and	CCONJ
easat-8918	105	24	jpeg	jpeg	NOUN
easat-8918	105	25	artifact	artifact	NOUN
easat-8918	105	26	removal	removal	NOUN
easat-8918	105	27	.	.	PUNCT
easat-8918	106	1	researchers	researcher	NOUN
easat-8918	106	2	in	in	ADP
easat-8918	106	3	zamir	zamir	PROPN
easat-8918	106	4	,	,	PUNCT
easat-8918	106	5	et	et	PROPN
easat-8918	106	6	al	al	PROPN
easat-8918	106	7	.	.	PUNCT
easat-8918	107	1	[	[	X
easat-8918	107	2	52	52	NUM
easat-8918	107	3	]	]	PUNCT
easat-8918	107	4	developed	develop	VERB
easat-8918	107	5	an	an	DET
easat-8918	107	6	efficient	efficient	ADJ
easat-8918	107	7	transformer	transformer	NOUN
easat-8918	107	8	-	-	PUNCT
easat-8918	107	9	based	base	VERB
easat-8918	107	10	model	model	NOUN
easat-8918	107	11	for	for	ADP
easat-8918	107	12	image	image	NOUN
easat-8918	107	13	denoising	denoising	NOUN
easat-8918	107	14	and	and	CCONJ
easat-8918	107	15	restoration	restoration	NOUN
easat-8918	107	16	,	,	PUNCT
easat-8918	107	17	capable	capable	ADJ
easat-8918	107	18	of	of	ADP
easat-8918	107	19	handling	handle	VERB
easat-8918	107	20	high	high	ADJ
easat-8918	107	21	-	-	PUNCT
easat-8918	107	22	resolution	resolution	NOUN
easat-8918	107	23	images	image	NOUN
easat-8918	107	24	.	.	PUNCT
easat-8918	108	1	to	to	PART
easat-8918	108	2	reduce	reduce	VERB
easat-8918	108	3	computational	computational	ADJ
easat-8918	108	4	demands	demand	NOUN
easat-8918	108	5	,	,	PUNCT
easat-8918	108	6	they	they	PRON
easat-8918	108	7	introduced	introduce	VERB
easat-8918	108	8	key	key	ADJ
easat-8918	108	9	architectural	architectural	ADJ
easat-8918	108	10	elements	element	NOUN
easat-8918	108	11	such	such	ADJ
easat-8918	108	12	as	as	ADP
easat-8918	108	13	a	a	DET
easat-8918	108	14	multi	multi	ADJ
easat-8918	108	15	-	-	ADJ
easat-8918	108	16	input	input	ADJ
easat-8918	108	17	self	self	NOUN
easat-8918	108	18	-	-	PUNCT
easat-8918	108	19	attention	attention	NOUN
easat-8918	108	20	(	(	PUNCT
easat-8918	108	21	sa	sa	NOUN
easat-8918	108	22	)	)	PUNCT
easat-8918	108	23	layer	layer	NOUN
easat-8918	108	24	and	and	CCONJ
easat-8918	108	25	a	a	DET
easat-8918	108	26	multi	multi	ADJ
easat-8918	108	27	-	-	ADJ
easat-8918	108	28	scale	scale	ADJ
easat-8918	108	29	hierarchical	hierarchical	ADJ
easat-8918	108	30	module	module	NOUN
easat-8918	108	31	with	with	ADP
easat-8918	108	32	reduced	reduced	ADJ
easat-8918	108	33	computational	computational	ADJ
easat-8918	108	34	complexity	complexity	NOUN
easat-8918	108	35	.	.	PUNCT
easat-8918	109	1	3	3	X
easat-8918	109	2	.	.	X
easat-8918	109	3	restormer	restormer	NOUN
easat-8918	109	4	and	and	CCONJ
easat-8918	109	5	residual	residual	ADJ
easat-8918	109	6	learning	learning	NOUN
easat-8918	109	7	of	of	ADP
easat-8918	109	8	dncnn	dncnn	PROPN
easat-8918	109	9	networks	network	NOUN
easat-8918	109	10	this	this	DET
easat-8918	109	11	section	section	NOUN
easat-8918	109	12	focuses	focus	VERB
easat-8918	109	13	on	on	ADP
easat-8918	109	14	two	two	NUM
easat-8918	109	15	recently	recently	ADV
easat-8918	109	16	developed	develop	VERB
easat-8918	109	17	image	image	NOUN
easat-8918	109	18	denoising	denoise	VERB
easat-8918	109	19	algorithms	algorithm	NOUN
easat-8918	109	20	:	:	PUNCT
easat-8918	109	21	the	the	DET
easat-8918	109	22	restoration	restoration	NOUN
easat-8918	109	23	transformer	transformer	NOUN
easat-8918	109	24	(	(	PUNCT
easat-8918	109	25	restormer	restormer	ADJ
easat-8918	109	26	)	)	PUNCT
easat-8918	109	27	network	network	NOUN
easat-8918	109	28	and	and	CCONJ
easat-8918	109	29	the	the	DET
easat-8918	109	30	residual	residual	ADJ
easat-8918	109	31	learning	learning	NOUN
easat-8918	109	32	of	of	ADP
easat-8918	109	33	the	the	DET
easat-8918	109	34	deep	deep	ADJ
easat-8918	109	35	cnn	cnn	PROPN
easat-8918	109	36	(	(	PUNCT
easat-8918	109	37	dncnn	dncnn	NOUN
easat-8918	109	38	)	)	PUNCT
easat-8918	109	39	network	network	NOUN
easat-8918	109	40	.	.	PUNCT
easat-8918	110	1	convolutional	convolutional	ADJ
easat-8918	110	2	neural	neural	ADJ
easat-8918	110	3	networks	network	NOUN
easat-8918	110	4	(	(	PUNCT
easat-8918	110	5	cnns	cnns	PROPN
easat-8918	110	6	)	)	PUNCT
easat-8918	110	7	,	,	PUNCT
easat-8918	110	8	which	which	PRON
easat-8918	110	9	have	have	AUX
easat-8918	110	10	demonstrated	demonstrate	VERB
easat-8918	110	11	strong	strong	ADJ
easat-8918	110	12	performance	performance	NOUN
easat-8918	110	13	in	in	ADP
easat-8918	110	14	image	image	NOUN
easat-8918	110	15	processing	processing	NOUN
easat-8918	110	16	applications	application	NOUN
easat-8918	110	17	,	,	PUNCT
easat-8918	110	18	will	will	AUX
easat-8918	110	19	be	be	AUX
easat-8918	110	20	utilized	utilize	VERB
easat-8918	110	21	in	in	ADP
easat-8918	110	22	the	the	DET
easat-8918	110	23	studied	study	VERB
easat-8918	110	24	algorithms	algorithm	NOUN
easat-8918	110	25	due	due	ADP
easat-8918	110	26	to	to	ADP
easat-8918	110	27	their	their	PRON
easat-8918	110	28	accessibility	accessibility	NOUN
easat-8918	110	29	to	to	ADP
easat-8918	110	30	large	large	ADJ
easat-8918	110	31	-	-	PUNCT
easat-8918	110	32	scale	scale	NOUN
easat-8918	110	33	datasets	dataset	NOUN
easat-8918	110	34	.	.	PUNCT
easat-8918	111	1	3.1	3.1	NUM
easat-8918	111	2	.	.	PUNCT
easat-8918	111	3	residual	residual	ADJ
easat-8918	111	4	learning	learning	NOUN
easat-8918	111	5	of	of	ADP
easat-8918	111	6	the	the	DET
easat-8918	111	7	denoising	denoise	VERB
easat-8918	111	8	cnn	cnn	PROPN
easat-8918	111	9	(	(	PUNCT
easat-8918	111	10	dncnn	dncnn	PROPN
easat-8918	111	11	)	)	PUNCT
easat-8918	111	12	network	network	NOUN
easat-8918	111	13	discriminative	discriminative	NOUN
easat-8918	111	14	model	model	NOUN
easat-8918	111	15	learning	learn	VERB
easat-8918	111	16	for	for	ADP
easat-8918	111	17	image	image	NOUN
easat-8918	111	18	denoising	denoising	NOUN
easat-8918	111	19	has	have	AUX
easat-8918	111	20	gained	gain	VERB
easat-8918	111	21	significant	significant	ADJ
easat-8918	111	22	attention	attention	NOUN
easat-8918	111	23	recently	recently	ADV
easat-8918	111	24	due	due	ADP
easat-8918	111	25	to	to	ADP
easat-8918	111	26	its	its	PRON
easat-8918	111	27	superior	superior	ADJ
easat-8918	111	28	denoising	denoising	NOUN
easat-8918	111	29	performance	performance	NOUN
easat-8918	111	30	.	.	PUNCT
easat-8918	112	1	in	in	ADP
easat-8918	112	2	zhang	zhang	PROPN
easat-8918	112	3	,	,	PUNCT
easat-8918	112	4	et	et	PROPN
easat-8918	112	5	al	al	PROPN
easat-8918	112	6	.	.	PUNCT
easat-8918	113	1	[	[	X
easat-8918	113	2	51	51	NUM
easat-8918	113	3	]	]	PUNCT
easat-8918	113	4	researchers	researcher	NOUN
easat-8918	113	5	explored	explore	VERB
easat-8918	113	6	the	the	DET
easat-8918	113	7	construction	construction	NOUN
easat-8918	113	8	of	of	ADP
easat-8918	113	9	feedforward	feedforward	NOUN
easat-8918	113	10	denoising	denoise	VERB
easat-8918	113	11	convolutional	convolutional	ADJ
easat-8918	113	12	neural	neural	ADJ
easat-8918	113	13	networks	network	NOUN
easat-8918	113	14	(	(	PUNCT
easat-8918	113	15	dncnns	dncnn	NOUN
easat-8918	113	16	)	)	PUNCT
easat-8918	113	17	by	by	ADP
easat-8918	113	18	incorporating	incorporate	VERB
easat-8918	113	19	advancements	advancement	NOUN
easat-8918	113	20	in	in	ADP
easat-8918	113	21	deep	deep	ADJ
easat-8918	113	22	architectures	architecture	NOUN
easat-8918	113	23	,	,	PUNCT
easat-8918	113	24	learning	learn	VERB
easat-8918	113	25	algorithms	algorithm	NOUN
easat-8918	113	26	,	,	PUNCT
easat-8918	113	27	and	and	CCONJ
easat-8918	113	28	regularization	regularization	NOUN
easat-8918	113	29	techniques	technique	NOUN
easat-8918	113	30	for	for	ADP
easat-8918	113	31	image	image	NOUN
easat-8918	113	32	denoising	denoise	VERB
easat-8918	113	33	.	.	PUNCT
easat-8918	114	1	specifically	specifically	ADV
easat-8918	114	2	,	,	PUNCT
easat-8918	114	3	residual	residual	ADJ
easat-8918	114	4	learning	learning	NOUN
easat-8918	114	5	and	and	CCONJ
easat-8918	114	6	batch	batch	NOUN
easat-8918	114	7	normalization	normalization	NOUN
easat-8918	114	8	are	be	AUX
easat-8918	114	9	employed	employ	VERB
easat-8918	114	10	to	to	AUX
easat-8918	114	11	both	both	PRON
easat-8918	114	12	accelerate	accelerate	VERB
easat-8918	114	13	the	the	DET
easat-8918	114	14	training	training	NOUN
easat-8918	114	15	process	process	NOUN
easat-8918	114	16	and	and	CCONJ
easat-8918	114	17	enhance	enhance	VERB
easat-8918	114	18	denoising	denoising	NOUN
easat-8918	114	19	performance	performance	NOUN
easat-8918	114	20	.	.	PUNCT
easat-8918	115	1	unlike	unlike	ADP
easat-8918	115	2	conventional	conventional	ADJ
easat-8918	115	3	discriminative	discriminative	NOUN
easat-8918	115	4	denoising	denoising	NOUN
easat-8918	115	5	models	model	NOUN
easat-8918	115	6	that	that	PRON
easat-8918	115	7	are	be	AUX
easat-8918	115	8	trained	train	VERB
easat-8918	115	9	for	for	ADP
easat-8918	115	10	a	a	DET
easat-8918	115	11	specific	specific	ADJ
easat-8918	115	12	noise	noise	NOUN
easat-8918	115	13	level	level	NOUN
easat-8918	115	14	(	(	PUNCT
easat-8918	115	15	e.g.	e.g.	ADV
easat-8918	115	16	,	,	PUNCT
easat-8918	115	17	additive	additive	ADJ
easat-8918	115	18	white	white	ADJ
easat-8918	115	19	gaussian	gaussian	PROPN
easat-8918	115	20	noise	noise	NOUN
easat-8918	115	21	awgn	awgn	NOUN
easat-8918	115	22	)	)	PUNCT
easat-8918	115	23	,	,	PUNCT
easat-8918	115	24	the	the	DET
easat-8918	115	25	dncnn	dncnn	PROPN
easat-8918	115	26	model	model	NOUN
easat-8918	115	27	can	can	AUX
easat-8918	115	28	remove	remove	VERB
easat-8918	115	29	gaussian	gaussian	ADJ
easat-8918	115	30	noise	noise	NOUN
easat-8918	115	31	with	with	ADP
easat-8918	115	32	unknown	unknown	ADJ
easat-8918	115	33	noise	noise	NOUN
easat-8918	115	34	levels	level	NOUN
easat-8918	115	35	(	(	PUNCT
easat-8918	115	36	i.e.	i.e.	X
easat-8918	115	37	,	,	PUNCT
easat-8918	115	38	blind	blind	ADJ
easat-8918	115	39	gaussian	gaussian	ADJ
easat-8918	115	40	denoising	denoising	NOUN
easat-8918	115	41	)	)	PUNCT
easat-8918	115	42	.	.	PUNCT
easat-8918	116	1	using	use	VERB
easat-8918	116	2	the	the	DET
easat-8918	116	3	residual	residual	ADJ
easat-8918	116	4	learning	learning	NOUN
easat-8918	116	5	strategy	strategy	NOUN
easat-8918	116	6	,	,	PUNCT
easat-8918	116	7	dncnn	dncnn	VERB
easat-8918	116	8	indirectly	indirectly	ADV
easat-8918	116	9	estimates	estimate	VERB
easat-8918	116	10	the	the	DET
easat-8918	116	11	clean	clean	ADJ
easat-8918	116	12	image	image	NOUN
easat-8918	116	13	embedded	embed	VERB
easat-8918	116	14	in	in	ADP
easat-8918	116	15	the	the	DET
easat-8918	116	16	hidden	hidden	ADJ
easat-8918	116	17	layers	layer	NOUN
easat-8918	116	18	.	.	PUNCT
easat-8918	117	1	the	the	DET
easat-8918	117	2	dncnn	dncnn	ADJ
easat-8918	117	3	network	network	NOUN
easat-8918	117	4	architecture	architecture	NOUN
easat-8918	117	5	is	be	AUX
easat-8918	117	6	shown	show	VERB
easat-8918	117	7	in	in	ADP
easat-8918	117	8	figure	figure	NOUN
easat-8918	117	9	8	8	NUM
easat-8918	117	10	.	.	PUNCT
easat-8918	118	1	3.1.1	3.1.1	NUM
easat-8918	118	2	.	.	PUNCT
easat-8918	119	1	residual	residual	ADJ
easat-8918	119	2	learning	learn	VERB
easat-8918	119	3	residual	residual	ADJ
easat-8918	119	4	learning	learning	NOUN
easat-8918	119	5	in	in	ADP
easat-8918	119	6	cnns	cnns	PROPN
easat-8918	119	7	was	be	AUX
easat-8918	119	8	initially	initially	ADV
easat-8918	119	9	proposed	propose	VERB
easat-8918	119	10	to	to	PART
easat-8918	119	11	address	address	VERB
easat-8918	119	12	the	the	DET
easat-8918	119	13	degradation	degradation	NOUN
easat-8918	119	14	of	of	ADP
easat-8918	119	15	performance	performance	NOUN
easat-8918	119	16	in	in	ADP
easat-8918	119	17	very	very	ADV
easat-8918	119	18	deep	deep	ADJ
easat-8918	119	19	networks	network	NOUN
easat-8918	119	20	.	.	PUNCT
easat-8918	120	1	as	as	ADP
easat-8918	120	2	network	network	NOUN
easat-8918	120	3	depth	depth	NOUN
easat-8918	120	4	increases	increase	NOUN
easat-8918	120	5	,	,	PUNCT
easat-8918	120	6	training	training	NOUN
easat-8918	120	7	accuracy	accuracy	NOUN
easat-8918	120	8	can	can	AUX
easat-8918	120	9	begin	begin	VERB
easat-8918	120	10	to	to	PART
easat-8918	120	11	degrade	degrade	VERB
easat-8918	120	12	.	.	PUNCT
easat-8918	121	1	with	with	ADP
easat-8918	121	2	residual	residual	ADJ
easat-8918	121	3	learning	learning	NOUN
easat-8918	121	4	,	,	PUNCT
easat-8918	121	5	deep	deep	ADJ
easat-8918	121	6	cnns	cnn	NOUN
easat-8918	121	7	can	can	AUX
easat-8918	121	8	be	be	AUX
easat-8918	121	9	trained	train	VERB
easat-8918	121	10	more	more	ADV
easat-8918	121	11	effectively	effectively	ADV
easat-8918	121	12	,	,	PUNCT
easat-8918	121	13	leading	lead	VERB
easat-8918	121	14	to	to	ADP
easat-8918	121	15	improvements	improvement	NOUN
easat-8918	121	16	in	in	ADP
easat-8918	121	17	tasks	task	NOUN
easat-8918	121	18	such	such	ADJ
easat-8918	121	19	as	as	ADP
easat-8918	121	20	image	image	NOUN
easat-8918	121	21	classification	classification	NOUN
easat-8918	121	22	and	and	CCONJ
easat-8918	121	23	object	object	NOUN
easat-8918	121	24	detection	detection	NOUN
easat-8918	122	1	[	[	X
easat-8918	122	2	53	53	NUM
easat-8918	122	3	]	]	PUNCT
easat-8918	122	4	.	.	PUNCT
easat-8918	123	1	the	the	DET
easat-8918	123	2	proposed	propose	VERB
easat-8918	123	3	dncnn	dncnn	PROPN
easat-8918	123	4	model	model	NOUN
easat-8918	123	5	adopts	adopt	VERB
easat-8918	123	6	the	the	DET
easat-8918	123	7	residual	residual	ADJ
easat-8918	123	8	learning	learning	NOUN
easat-8918	123	9	principle	principle	NOUN
easat-8918	123	10	by	by	ADP
easat-8918	123	11	using	use	VERB
easat-8918	123	12	a	a	DET
easat-8918	123	13	single	single	ADJ
easat-8918	123	14	residual	residual	ADJ
easat-8918	123	15	unit	unit	NOUN
easat-8918	123	16	to	to	PART
easat-8918	123	17	predict	predict	VERB
easat-8918	123	18	the	the	DET
easat-8918	123	19	residual	residual	ADJ
easat-8918	123	20	image	image	NOUN
easat-8918	123	21	(	(	PUNCT
easat-8918	123	22	i.e.	i.e.	X
easat-8918	123	23	,	,	PUNCT
easat-8918	123	24	the	the	DET
easat-8918	123	25	noise	noise	NOUN
easat-8918	123	26	)	)	PUNCT
easat-8918	123	27	,	,	PUNCT
easat-8918	123	28	rather	rather	ADV
easat-8918	123	29	than	than	ADP
easat-8918	123	30	directly	directly	ADV
easat-8918	123	31	predicting	predict	VERB
easat-8918	123	32	the	the	DET
easat-8918	123	33	clean	clean	ADJ
easat-8918	123	34	image	image	NOUN
easat-8918	123	35	.	.	PUNCT
easat-8918	124	1	3.1.2	3.1.2	X
easat-8918	124	2	.	.	NOUN
easat-8918	124	3	batch	batch	NOUN
easat-8918	124	4	normalization	normalization	NOUN
easat-8918	124	5	despite	despite	SCONJ
easat-8918	124	6	the	the	DET
easat-8918	124	7	simplicity	simplicity	NOUN
easat-8918	124	8	and	and	CCONJ
easat-8918	124	9	effectiveness	effectiveness	NOUN
easat-8918	124	10	of	of	ADP
easat-8918	124	11	mini	mini	ADJ
easat-8918	124	12	-	-	NOUN
easat-8918	124	13	batch	batch	ADJ
easat-8918	124	14	stochastic	stochastic	ADJ
easat-8918	124	15	gradient	gradient	ADJ
easat-8918	124	16	descent	descent	NOUN
easat-8918	124	17	(	(	PUNCT
easat-8918	124	18	sgd	sgd	PROPN
easat-8918	124	19	)	)	PUNCT
easat-8918	124	20	,	,	PUNCT
easat-8918	124	21	the	the	DET
easat-8918	124	22	performance	performance	NOUN
easat-8918	124	23	of	of	ADP
easat-8918	124	24	trained	train	VERB
easat-8918	124	25	cnn	cnn	PROPN
easat-8918	124	26	models	model	NOUN
easat-8918	124	27	can	can	AUX
easat-8918	124	28	degrade	degrade	VERB
easat-8918	124	29	due	due	ADJ
easat-8918	124	30	to	to	ADP
easat-8918	124	31	internal	internal	ADJ
easat-8918	124	32	covariate	covariate	ADJ
easat-8918	124	33	shift	shift	NOUN
easat-8918	124	34	,	,	PUNCT
easat-8918	124	35	that	that	ADV
easat-8918	124	36	is	is	ADV
easat-8918	124	37	,	,	PUNCT
easat-8918	124	38	changes	change	NOUN
easat-8918	124	39	in	in	ADP
easat-8918	124	40	the	the	DET
easat-8918	124	41	distribution	distribution	NOUN
easat-8918	124	42	of	of	ADP
easat-8918	124	43	nonlinear	nonlinear	ADJ
easat-8918	124	44	activations	activation	NOUN
easat-8918	124	45	during	during	ADP
easat-8918	124	46	training	training	NOUN
easat-8918	124	47	.	.	PUNCT
easat-8918	125	1	batch	batch	NOUN
easat-8918	125	2	normalization	normalization	NOUN
easat-8918	125	3	was	be	AUX
easat-8918	125	4	proposed	propose	VERB
easat-8918	125	5	to	to	PART
easat-8918	125	6	reduce	reduce	VERB
easat-8918	125	7	this	this	DET
easat-8918	125	8	effect	effect	NOUN
easat-8918	125	9	by	by	ADP
easat-8918	125	10	adding	add	VERB
easat-8918	125	11	a	a	DET
easat-8918	125	12	normalization	normalization	NOUN
easat-8918	125	13	step	step	NOUN
easat-8918	125	14	before	before	ADP
easat-8918	125	15	each	each	DET
easat-8918	125	16	nonlinearity	nonlinearity	NOUN
easat-8918	125	17	,	,	PUNCT
easat-8918	125	18	along	along	ADP
easat-8918	125	19	with	with	ADP
easat-8918	125	20	a	a	DET
easat-8918	125	21	scale	scale	NOUN
easat-8918	125	22	and	and	CCONJ
easat-8918	125	23	shift	shift	NOUN
easat-8918	125	24	operation	operation	NOUN
easat-8918	125	25	.	.	PUNCT
easat-8918	126	1	batch	batch	NOUN
easat-8918	126	2	normalization	normalization	NOUN
easat-8918	126	3	leads	lead	VERB
easat-8918	126	4	to	to	ADP
easat-8918	126	5	faster	fast	ADJ
easat-8918	126	6	training	training	NOUN
easat-8918	126	7	,	,	PUNCT
easat-8918	126	8	better	well	ADJ
easat-8918	126	9	performance	performance	NOUN
easat-8918	126	10	,	,	PUNCT
easat-8918	126	11	and	and	CCONJ
easat-8918	126	12	reduced	reduce	VERB
easat-8918	126	13	sensitivity	sensitivity	NOUN
easat-8918	126	14	to	to	ADP
easat-8918	126	15	weight	weight	NOUN
easat-8918	126	16	initialization	initialization	NOUN
easat-8918	126	17	[	[	X
easat-8918	126	18	21	21	NUM
easat-8918	126	19	]	]	PUNCT
easat-8918	126	20	.	.	PUNCT
easat-8918	127	1	1351	1351	NUM
easat-8918	127	2	edelweiss	edelweiss	PROPN
easat-8918	127	3	applied	apply	VERB
easat-8918	127	4	science	science	NOUN
easat-8918	127	5	and	and	CCONJ
easat-8918	127	6	technology	technology	NOUN
easat-8918	127	7	issn	issn	PROPN
easat-8918	127	8	:	:	PUNCT
easat-8918	127	9	2576	2576	NUM
easat-8918	127	10	-	-	SYM
easat-8918	127	11	8484	8484	NUM
easat-8918	127	12	vol	vol	NOUN
easat-8918	127	13	.	.	PROPN
easat-8918	128	1	9	9	NUM
easat-8918	128	2	,	,	PUNCT
easat-8918	128	3	no	no	INTJ
easat-8918	128	4	.	.	NOUN
easat-8918	128	5	7	7	NUM
easat-8918	128	6	:	:	SYM
easat-8918	128	7	1345	1345	NUM
easat-8918	128	8	-	-	SYM
easat-8918	128	9	1359	1359	NUM
easat-8918	128	10	,	,	PUNCT
easat-8918	128	11	2025	2025	NUM
easat-8918	128	12	doi	doi	NOUN
easat-8918	128	13	:	:	PUNCT
easat-8918	128	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	128	15	©	©	PROPN
easat-8918	128	16	2025	2025	NUM
easat-8918	128	17	by	by	ADP
easat-8918	128	18	the	the	DET
easat-8918	128	19	author	author	NOUN
easat-8918	128	20	;	;	PUNCT
easat-8918	128	21	licensee	licensee	PROPN
easat-8918	128	22	learning	learn	VERB
easat-8918	128	23	gate	gate	PROPN
easat-8918	128	24	figure	figure	NOUN
easat-8918	128	25	8	8	NUM
easat-8918	128	26	.	.	PUNCT
easat-8918	129	1	dncnn	dncnn	PROPN
easat-8918	129	2	network	network	NOUN
easat-8918	129	3	architecture	architecture	NOUN
easat-8918	129	4	.	.	PUNCT
easat-8918	130	1	3.2	3.2	NUM
easat-8918	130	2	.	.	PUNCT
easat-8918	130	3	restoration	restoration	NOUN
easat-8918	130	4	transformer	transformer	NOUN
easat-8918	130	5	(	(	PUNCT
easat-8918	130	6	restormer	restormer	ADJ
easat-8918	130	7	)	)	PUNCT
easat-8918	130	8	network	network	NOUN
easat-8918	130	9	image	image	NOUN
easat-8918	130	10	restoration	restoration	NOUN
easat-8918	130	11	is	be	AUX
easat-8918	130	12	the	the	DET
easat-8918	130	13	task	task	NOUN
easat-8918	130	14	of	of	ADP
easat-8918	130	15	reconstructing	reconstruct	VERB
easat-8918	130	16	a	a	DET
easat-8918	130	17	high	high	ADJ
easat-8918	130	18	-	-	PUNCT
easat-8918	130	19	quality	quality	NOUN
easat-8918	130	20	image	image	NOUN
easat-8918	130	21	by	by	ADP
easat-8918	130	22	removing	remove	VERB
easat-8918	130	23	degradations	degradation	NOUN
easat-8918	130	24	(	(	PUNCT
easat-8918	130	25	e.g.	e.g.	ADV
easat-8918	130	26	,	,	PUNCT
easat-8918	130	27	noise	noise	NOUN
easat-8918	130	28	,	,	PUNCT
easat-8918	130	29	blur	blur	NOUN
easat-8918	130	30	,	,	PUNCT
easat-8918	130	31	rain	rain	NOUN
easat-8918	130	32	streaks	streak	NOUN
easat-8918	130	33	)	)	PUNCT
easat-8918	130	34	from	from	ADP
easat-8918	130	35	a	a	DET
easat-8918	130	36	corrupted	corrupted	ADJ
easat-8918	130	37	input	input	NOUN
easat-8918	130	38	.	.	PUNCT
easat-8918	131	1	convolutional	convolutional	ADJ
easat-8918	131	2	neural	neural	ADJ
easat-8918	131	3	networks	network	NOUN
easat-8918	131	4	(	(	PUNCT
easat-8918	131	5	cnns	cnns	PROPN
easat-8918	131	6	)	)	PUNCT
easat-8918	131	7	have	have	AUX
easat-8918	131	8	become	become	VERB
easat-8918	131	9	the	the	DET
easat-8918	131	10	preferred	prefer	VERB
easat-8918	131	11	choice	choice	NOUN
easat-8918	131	12	over	over	ADP
easat-8918	131	13	traditional	traditional	ADJ
easat-8918	131	14	restoration	restoration	NOUN
easat-8918	131	15	approaches	approach	NOUN
easat-8918	131	16	due	due	ADJ
easat-8918	131	17	to	to	ADP
easat-8918	131	18	their	their	PRON
easat-8918	131	19	strong	strong	ADJ
easat-8918	131	20	performance	performance	NOUN
easat-8918	131	21	in	in	ADP
easat-8918	131	22	learning	learn	VERB
easat-8918	131	23	generalizable	generalizable	ADJ
easat-8918	131	24	priors	prior	NOUN
easat-8918	131	25	from	from	ADP
easat-8918	131	26	large	large	ADJ
easat-8918	131	27	-	-	PUNCT
easat-8918	131	28	scale	scale	NOUN
easat-8918	131	29	data	datum	NOUN
easat-8918	131	30	[	[	X
easat-8918	131	31	52	52	NUM
easat-8918	131	32	]	]	PUNCT
easat-8918	131	33	.	.	PUNCT
easat-8918	132	1	the	the	DET
easat-8918	132	2	core	core	NOUN
easat-8918	132	3	operation	operation	NOUN
easat-8918	132	4	in	in	ADP
easat-8918	132	5	cnns	cnns	PROPN
easat-8918	132	6	is	be	AUX
easat-8918	132	7	convolution	convolution	NOUN
easat-8918	132	8	,	,	PUNCT
easat-8918	132	9	which	which	PRON
easat-8918	132	10	offers	offer	VERB
easat-8918	132	11	local	local	ADJ
easat-8918	132	12	connectivity	connectivity	NOUN
easat-8918	132	13	and	and	CCONJ
easat-8918	132	14	translation	translation	NOUN
easat-8918	132	15	equivariance	equivariance	NOUN
easat-8918	132	16	.	.	PUNCT
easat-8918	133	1	while	while	SCONJ
easat-8918	133	2	these	these	DET
easat-8918	133	3	properties	property	NOUN
easat-8918	133	4	provide	provide	VERB
easat-8918	133	5	efficiency	efficiency	NOUN
easat-8918	133	6	and	and	CCONJ
easat-8918	133	7	generalization	generalization	NOUN
easat-8918	133	8	,	,	PUNCT
easat-8918	133	9	they	they	PRON
easat-8918	133	10	also	also	ADV
easat-8918	133	11	introduce	introduce	VERB
easat-8918	133	12	two	two	NUM
easat-8918	133	13	key	key	ADJ
easat-8918	133	14	limitations	limitation	NOUN
easat-8918	133	15	:	:	PUNCT
easat-8918	133	16	to	to	PART
easat-8918	133	17	address	address	VERB
easat-8918	133	18	these	these	DET
easat-8918	133	19	issues	issue	NOUN
easat-8918	133	20	,	,	PUNCT
easat-8918	133	21	the	the	DET
easat-8918	133	22	self	self	NOUN
easat-8918	133	23	-	-	PUNCT
easat-8918	133	24	attention	attention	NOUN
easat-8918	133	25	(	(	PUNCT
easat-8918	133	26	sa	sa	NOUN
easat-8918	133	27	)	)	PUNCT
easat-8918	133	28	mechanism	mechanism	NOUN
easat-8918	133	29	was	be	AUX
easat-8918	133	30	introduced	introduce	VERB
easat-8918	133	31	,	,	PUNCT
easat-8918	133	32	which	which	PRON
easat-8918	133	33	computes	compute	VERB
easat-8918	133	34	a	a	DET
easat-8918	133	35	pixel	pixel	NOUN
easat-8918	133	36	’s	’s	PART
easat-8918	133	37	response	response	NOUN
easat-8918	133	38	based	base	VERB
easat-8918	133	39	on	on	ADP
easat-8918	133	40	a	a	DET
easat-8918	133	41	weighted	weight	VERB
easat-8918	133	42	sum	sum	NOUN
easat-8918	133	43	of	of	ADP
easat-8918	133	44	all	all	DET
easat-8918	133	45	positions	position	NOUN
easat-8918	133	46	in	in	ADP
easat-8918	133	47	the	the	DET
easat-8918	133	48	input	input	NOUN
easat-8918	133	49	[	[	X
easat-8918	133	50	54	54	NUM
easat-8918	133	51	,	,	PUNCT
easat-8918	133	52	55	55	NUM
easat-8918	133	53	]	]	PUNCT
easat-8918	133	54	.	.	PUNCT
easat-8918	134	1	sa	sa	PROPN
easat-8918	134	2	has	have	AUX
easat-8918	134	3	become	become	VERB
easat-8918	134	4	a	a	DET
easat-8918	134	5	fundamental	fundamental	ADJ
easat-8918	134	6	component	component	NOUN
easat-8918	134	7	in	in	ADP
easat-8918	134	8	transformer	transformer	NOUN
easat-8918	134	9	models	model	NOUN
easat-8918	134	10	,	,	PUNCT
easat-8918	134	11	which	which	PRON
easat-8918	134	12	are	be	AUX
easat-8918	134	13	optimized	optimize	VERB
easat-8918	134	14	for	for	ADP
easat-8918	134	15	parallelization	parallelization	NOUN
easat-8918	134	16	and	and	CCONJ
easat-8918	134	17	effective	effective	ADJ
easat-8918	134	18	representation	representation	NOUN
easat-8918	134	19	learning	learn	VERB
easat-8918	134	20	[	[	X
easat-8918	134	21	56	56	NUM
easat-8918	134	22	]	]	PUNCT
easat-8918	134	23	.	.	PUNCT
easat-8918	135	1	transformers	transformer	NOUN
easat-8918	135	2	have	have	AUX
easat-8918	135	3	demonstrated	demonstrate	VERB
easat-8918	135	4	state	state	NOUN
easat-8918	135	5	-	-	PUNCT
easat-8918	135	6	of	of	ADP
easat-8918	135	7	-	-	PUNCT
easat-8918	135	8	the	the	DET
easat-8918	135	9	-	-	PUNCT
easat-8918	135	10	art	art	NOUN
easat-8918	135	11	performance	performance	NOUN
easat-8918	135	12	in	in	ADP
easat-8918	135	13	natural	natural	ADJ
easat-8918	135	14	language	language	NOUN
easat-8918	135	15	processing	processing	NOUN
easat-8918	135	16	[	[	X
easat-8918	135	17	57	57	NUM
easat-8918	135	18	]	]	PUNCT
easat-8918	135	19	and	and	CCONJ
easat-8918	135	20	high	high	ADJ
easat-8918	135	21	-	-	PUNCT
easat-8918	135	22	level	level	NOUN
easat-8918	135	23	computer	computer	NOUN
easat-8918	135	24	vision	vision	NOUN
easat-8918	135	25	tasks	task	NOUN
easat-8918	135	26	[	[	X
easat-8918	135	27	58	58	NUM
easat-8918	135	28	,	,	PUNCT
easat-8918	135	29	59	59	NUM
easat-8918	135	30	]	]	PUNCT
easat-8918	135	31	.	.	PUNCT
easat-8918	136	1	although	although	SCONJ
easat-8918	136	2	sa	sa	PROPN
easat-8918	136	3	is	be	AUX
easat-8918	136	4	highly	highly	ADV
easat-8918	136	5	effective	effective	ADJ
easat-8918	136	6	at	at	ADP
easat-8918	136	7	capturing	capture	VERB
easat-8918	136	8	long	long	ADJ
easat-8918	136	9	-	-	PUNCT
easat-8918	136	10	range	range	NOUN
easat-8918	136	11	pixel	pixel	NOUN
easat-8918	136	12	interactions	interaction	NOUN
easat-8918	136	13	,	,	PUNCT
easat-8918	136	14	its	its	PRON
easat-8918	136	15	computational	computational	ADJ
easat-8918	136	16	complexity	complexity	NOUN
easat-8918	136	17	increases	increase	VERB
easat-8918	136	18	quadratically	quadratically	ADV
easat-8918	136	19	with	with	ADP
easat-8918	136	20	spatial	spatial	ADJ
easat-8918	136	21	resolution	resolution	NOUN
easat-8918	136	22	,	,	PUNCT
easat-8918	136	23	making	make	VERB
easat-8918	136	24	it	it	PRON
easat-8918	136	25	impractical	impractical	ADJ
easat-8918	136	26	for	for	ADP
easat-8918	136	27	high	high	ADJ
easat-8918	136	28	-	-	PUNCT
easat-8918	136	29	resolution	resolution	NOUN
easat-8918	136	30	image	image	NOUN
easat-8918	136	31	processing	processing	NOUN
easat-8918	136	32	,	,	PUNCT
easat-8918	136	33	common	common	ADJ
easat-8918	136	34	in	in	ADP
easat-8918	136	35	restoration	restoration	NOUN
easat-8918	136	36	tasks	task	NOUN
easat-8918	136	37	.	.	PUNCT
easat-8918	137	1	research	research	NOUN
easat-8918	137	2	on	on	ADP
easat-8918	137	3	adapting	adapt	VERB
easat-8918	137	4	transformers	transformer	NOUN
easat-8918	137	5	to	to	ADP
easat-8918	137	6	image	image	NOUN
easat-8918	137	7	restoration	restoration	NOUN
easat-8918	137	8	is	be	AUX
easat-8918	137	9	still	still	ADV
easat-8918	137	10	limited	limit	VERB
easat-8918	137	11	[	[	X
easat-8918	137	12	60	60	NUM
easat-8918	137	13	,	,	PUNCT
easat-8918	137	14	61	61	NUM
easat-8918	137	15	]	]	PUNCT
easat-8918	137	16	.	.	PUNCT
easat-8918	138	1	to	to	PART
easat-8918	138	2	reduce	reduce	VERB
easat-8918	138	3	computational	computational	ADJ
easat-8918	138	4	burden	burden	NOUN
easat-8918	138	5	,	,	PUNCT
easat-8918	138	6	some	some	DET
easat-8918	138	7	methods	method	NOUN
easat-8918	138	8	apply	apply	VERB
easat-8918	138	9	sa	sa	NOUN
easat-8918	138	10	in	in	ADP
easat-8918	138	11	small	small	ADJ
easat-8918	138	12	8×8	8×8	NUM
easat-8918	138	13	spatial	spatial	ADJ
easat-8918	138	14	windows	window	NOUN
easat-8918	138	15	or	or	CCONJ
easat-8918	138	16	divide	divide	VERB
easat-8918	138	17	the	the	DET
easat-8918	138	18	input	input	NOUN
easat-8918	138	19	image	image	NOUN
easat-8918	138	20	into	into	ADP
easat-8918	138	21	non	non	ADJ
easat-8918	138	22	-	-	ADJ
easat-8918	138	23	overlapping	overlapping	ADJ
easat-8918	138	24	48×48	48×48	NUM
easat-8918	138	25	patches	patch	NOUN
easat-8918	138	26	to	to	PART
easat-8918	138	27	perform	perform	VERB
easat-8918	138	28	attention	attention	NOUN
easat-8918	138	29	independently	independently	ADV
easat-8918	138	30	within	within	ADP
easat-8918	138	31	each	each	DET
easat-8918	138	32	patch	patch	NOUN
easat-8918	138	33	.	.	PUNCT
easat-8918	139	1	however	however	ADV
easat-8918	139	2	,	,	PUNCT
easat-8918	139	3	limiting	limit	VERB
easat-8918	139	4	the	the	DET
easat-8918	139	5	spatial	spatial	ADJ
easat-8918	139	6	scope	scope	NOUN
easat-8918	139	7	of	of	ADP
easat-8918	139	8	sa	sa	PROPN
easat-8918	139	9	conflicts	conflict	NOUN
easat-8918	139	10	with	with	ADP
easat-8918	139	11	the	the	DET
easat-8918	139	12	goal	goal	NOUN
easat-8918	139	13	of	of	ADP
easat-8918	139	14	capturing	capture	VERB
easat-8918	139	15	true	true	ADJ
easat-8918	139	16	long	long	ADJ
easat-8918	139	17	-	-	PUNCT
easat-8918	139	18	range	range	NOUN
easat-8918	139	19	pixel	pixel	NOUN
easat-8918	139	20	relationships	relationship	NOUN
easat-8918	139	21	,	,	PUNCT
easat-8918	139	22	especially	especially	ADV
easat-8918	139	23	in	in	ADP
easat-8918	139	24	high	high	ADJ
easat-8918	139	25	-	-	PUNCT
easat-8918	139	26	resolution	resolution	NOUN
easat-8918	139	27	images	image	NOUN
easat-8918	139	28	.	.	PUNCT
easat-8918	140	1	in	in	ADP
easat-8918	140	2	zamir	zamir	PROPN
easat-8918	140	3	,	,	PUNCT
easat-8918	140	4	et	et	PROPN
easat-8918	140	5	al	al	PROPN
easat-8918	140	6	.	.	PUNCT
easat-8918	141	1	[	[	X
easat-8918	141	2	52	52	NUM
easat-8918	141	3	]	]	PUNCT
easat-8918	141	4	researchers	researcher	NOUN
easat-8918	141	5	proposed	propose	VERB
easat-8918	141	6	an	an	DET
easat-8918	141	7	efficient	efficient	ADJ
easat-8918	141	8	transformer	transformer	NOUN
easat-8918	141	9	model	model	NOUN
easat-8918	141	10	capable	capable	ADJ
easat-8918	141	11	of	of	ADP
easat-8918	141	12	handling	handle	VERB
easat-8918	141	13	highresolution	highresolution	NOUN
easat-8918	141	14	images	image	NOUN
easat-8918	141	15	for	for	ADP
easat-8918	141	16	restoration	restoration	NOUN
easat-8918	141	17	tasks	task	NOUN
easat-8918	141	18	.	.	PUNCT
easat-8918	142	1	to	to	PART
easat-8918	142	2	manage	manage	VERB
easat-8918	142	3	computational	computational	ADJ
easat-8918	142	4	demands	demand	NOUN
easat-8918	142	5	,	,	PUNCT
easat-8918	142	6	they	they	PRON
easat-8918	142	7	introduced	introduce	VERB
easat-8918	142	8	a	a	DET
easat-8918	142	9	multihead	multihead	NOUN
easat-8918	142	10	self	self	NOUN
easat-8918	142	11	-	-	PUNCT
easat-8918	142	12	attention	attention	NOUN
easat-8918	142	13	layer	layer	NOUN
easat-8918	142	14	and	and	CCONJ
easat-8918	142	15	a	a	DET
easat-8918	142	16	multi	multi	ADJ
easat-8918	142	17	-	-	ADJ
easat-8918	142	18	scale	scale	ADJ
easat-8918	142	19	hierarchical	hierarchical	ADJ
easat-8918	142	20	module	module	NOUN
easat-8918	142	21	,	,	PUNCT
easat-8918	142	22	which	which	PRON
easat-8918	142	23	requires	require	VERB
easat-8918	142	24	fewer	few	ADJ
easat-8918	142	25	resources	resource	NOUN
easat-8918	142	26	than	than	ADP
easat-8918	142	27	a	a	DET
easat-8918	142	28	single	single	ADJ
easat-8918	142	29	-	-	PUNCT
easat-8918	142	30	scale	scale	NOUN
easat-8918	142	31	network	network	NOUN
easat-8918	142	32	[	[	X
easat-8918	142	33	58	58	NUM
easat-8918	142	34	]	]	PUNCT
easat-8918	142	35	.	.	PUNCT
easat-8918	143	1	a	a	DET
easat-8918	143	2	progressive	progressive	ADJ
easat-8918	143	3	training	training	NOUN
easat-8918	143	4	strategy	strategy	NOUN
easat-8918	143	5	was	be	AUX
easat-8918	143	6	adopted	adopt	VERB
easat-8918	143	7	to	to	PART
easat-8918	143	8	help	help	VERB
easat-8918	143	9	the	the	DET
easat-8918	143	10	model	model	NOUN
easat-8918	143	11	learn	learn	VERB
easat-8918	143	12	image	image	NOUN
easat-8918	143	13	statistics	statistic	NOUN
easat-8918	143	14	from	from	ADP
easat-8918	143	15	large	large	ADJ
easat-8918	143	16	datasets	dataset	NOUN
easat-8918	143	17	,	,	PUNCT
easat-8918	143	18	enhancing	enhance	VERB
easat-8918	143	19	contextual	contextual	ADJ
easat-8918	143	20	understanding	understanding	NOUN
easat-8918	143	21	and	and	CCONJ
easat-8918	143	22	improving	improve	VERB
easat-8918	143	23	quality	quality	NOUN
easat-8918	143	24	during	during	ADP
easat-8918	143	25	inference	inference	NOUN
easat-8918	143	26	[	[	X
easat-8918	143	27	52	52	NUM
easat-8918	143	28	]	]	PUNCT
easat-8918	143	29	.	.	PUNCT
easat-8918	144	1	the	the	DET
easat-8918	144	2	restormer	restormer	ADJ
easat-8918	144	3	architecture	architecture	NOUN
easat-8918	144	4	for	for	ADP
easat-8918	144	5	high	high	ADJ
easat-8918	144	6	-	-	PUNCT
easat-8918	144	7	resolution	resolution	NOUN
easat-8918	144	8	image	image	NOUN
easat-8918	144	9	restoration	restoration	NOUN
easat-8918	144	10	is	be	AUX
easat-8918	144	11	shown	show	VERB
easat-8918	144	12	in	in	ADP
easat-8918	144	13	figure	figure	NOUN
easat-8918	144	14	9	9	NUM
easat-8918	144	15	.	.	PUNCT
easat-8918	145	1	it	it	PRON
easat-8918	145	2	consists	consist	VERB
easat-8918	145	3	of	of	ADP
easat-8918	145	4	a	a	DET
easat-8918	145	5	multi	multi	ADJ
easat-8918	145	6	-	-	ADJ
easat-8918	145	7	scale	scale	ADJ
easat-8918	145	8	hierarchical	hierarchical	ADJ
easat-8918	145	9	design	design	NOUN
easat-8918	145	10	with	with	ADP
easat-8918	145	11	efficient	efficient	ADJ
easat-8918	145	12	transformer	transformer	NOUN
easat-8918	145	13	blocks	block	NOUN
easat-8918	145	14	.	.	PUNCT
easat-8918	146	1	the	the	DET
easat-8918	146	2	core	core	NOUN
easat-8918	146	3	modules	module	NOUN
easat-8918	146	4	of	of	ADP
easat-8918	146	5	a	a	DET
easat-8918	146	6	transformer	transformer	NOUN
easat-8918	146	7	block	block	NOUN
easat-8918	146	8	include	include	VERB
easat-8918	146	9	:	:	PUNCT
easat-8918	146	10	●	●	PUNCT
easat-8918	146	11	multi	multi	ADJ
easat-8918	146	12	-	-	ADJ
easat-8918	146	13	dconv	dconv	ADJ
easat-8918	146	14	transpose	transpose	NOUN
easat-8918	146	15	attention	attention	NOUN
easat-8918	146	16	(	(	PUNCT
easat-8918	146	17	mdta	mdta	PROPN
easat-8918	146	18	):	):	PUNCT
easat-8918	146	19	mdta	mdta	PROPN
easat-8918	146	20	enables	enable	VERB
easat-8918	146	21	spatially	spatially	ADV
easat-8918	146	22	enriched	enrich	VERB
easat-8918	146	23	channel	channel	NOUN
easat-8918	146	24	-	-	PUNCT
easat-8918	146	25	wise	wise	ADJ
easat-8918	146	26	query	query	NOUN
easat-8918	146	27	-	-	PUNCT
easat-8918	146	28	key	key	NOUN
easat-8918	146	29	interactions	interaction	NOUN
easat-8918	146	30	,	,	PUNCT
easat-8918	146	31	rather	rather	ADV
easat-8918	146	32	than	than	ADP
easat-8918	146	33	operating	operate	VERB
easat-8918	146	34	across	across	ADP
easat-8918	146	35	spatial	spatial	ADJ
easat-8918	146	36	dimensions	dimension	NOUN
easat-8918	146	37	.	.	PUNCT
easat-8918	147	1	●	●	PUNCT
easat-8918	147	2	gated	gate	VERB
easat-8918	147	3	-	-	PUNCT
easat-8918	147	4	dconv	dconv	NOUN
easat-8918	147	5	feed	feed	NOUN
easat-8918	147	6	-	-	PUNCT
easat-8918	147	7	forward	forward	NOUN
easat-8918	147	8	network	network	NOUN
easat-8918	147	9	(	(	PUNCT
easat-8918	147	10	gdfn	gdfn	NOUN
easat-8918	147	11	):	):	PUNCT
easat-8918	147	12	gdfn	gdfn	NOUN
easat-8918	147	13	performs	perform	VERB
easat-8918	147	14	controlled	control	VERB
easat-8918	147	15	feature	feature	NOUN
easat-8918	147	16	transformation	transformation	NOUN
easat-8918	147	17	,	,	PUNCT
easat-8918	147	18	allowing	allow	VERB
easat-8918	147	19	more	more	ADV
easat-8918	147	20	effective	effective	ADJ
easat-8918	147	21	propagation	propagation	NOUN
easat-8918	147	22	of	of	ADP
easat-8918	147	23	useful	useful	ADJ
easat-8918	147	24	information	information	NOUN
easat-8918	147	25	.	.	PUNCT
easat-8918	148	1	1352	1352	NUM
easat-8918	148	2	edelweiss	edelweiss	PROPN
easat-8918	148	3	applied	apply	VERB
easat-8918	148	4	science	science	NOUN
easat-8918	148	5	and	and	CCONJ
easat-8918	148	6	technology	technology	NOUN
easat-8918	148	7	issn	issn	PROPN
easat-8918	148	8	:	:	PUNCT
easat-8918	148	9	2576	2576	NUM
easat-8918	148	10	-	-	SYM
easat-8918	148	11	8484	8484	NUM
easat-8918	148	12	vol	vol	NOUN
easat-8918	148	13	.	.	PROPN
easat-8918	149	1	9	9	NUM
easat-8918	149	2	,	,	PUNCT
easat-8918	149	3	no	no	INTJ
easat-8918	149	4	.	.	NOUN
easat-8918	149	5	7	7	NUM
easat-8918	149	6	:	:	SYM
easat-8918	149	7	1345	1345	NUM
easat-8918	149	8	-	-	SYM
easat-8918	149	9	1359	1359	NUM
easat-8918	149	10	,	,	PUNCT
easat-8918	149	11	2025	2025	NUM
easat-8918	149	12	doi	doi	NOUN
easat-8918	149	13	:	:	PUNCT
easat-8918	149	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	149	15	©	©	PROPN
easat-8918	149	16	2025	2025	NUM
easat-8918	149	17	by	by	ADP
easat-8918	149	18	the	the	DET
easat-8918	149	19	author	author	NOUN
easat-8918	149	20	;	;	PUNCT
easat-8918	149	21	licensee	licensee	PROPN
easat-8918	149	22	learning	learn	VERB
easat-8918	149	23	gate	gate	NOUN
easat-8918	149	24	figure	figure	NOUN
easat-8918	149	25	9	9	NUM
easat-8918	149	26	.	.	PUNCT
easat-8918	149	27	restoration	restoration	NOUN
easat-8918	149	28	transformer	transformer	NOUN
easat-8918	149	29	(	(	PUNCT
easat-8918	149	30	restormer	restormer	ADJ
easat-8918	149	31	)	)	PUNCT
easat-8918	149	32	network	network	NOUN
easat-8918	149	33	architecture	architecture	NOUN
easat-8918	149	34	zamir	zamir	PROPN
easat-8918	149	35	,	,	PUNCT
easat-8918	149	36	et	et	PROPN
easat-8918	149	37	al	al	PROPN
easat-8918	149	38	.	.	PUNCT
easat-8918	150	1	[	[	X
easat-8918	150	2	52	52	NUM
easat-8918	150	3	]	]	PUNCT
easat-8918	150	4	.	.	PUNCT
easat-8918	151	1	4	4	X
easat-8918	151	2	.	.	X
easat-8918	151	3	experimental	experimental	ADJ
easat-8918	151	4	results	result	NOUN
easat-8918	151	5	and	and	CCONJ
easat-8918	151	6	discussions	discussion	NOUN
easat-8918	151	7	since	since	SCONJ
easat-8918	151	8	noise	noise	NOUN
easat-8918	151	9	reduction	reduction	NOUN
easat-8918	151	10	in	in	ADP
easat-8918	151	11	images	image	NOUN
easat-8918	151	12	is	be	AUX
easat-8918	151	13	an	an	DET
easat-8918	151	14	indispensable	indispensable	ADJ
easat-8918	151	15	step	step	NOUN
easat-8918	151	16	in	in	ADP
easat-8918	151	17	many	many	ADJ
easat-8918	151	18	practical	practical	ADJ
easat-8918	151	19	applications	application	NOUN
easat-8918	151	20	,	,	PUNCT
easat-8918	151	21	it	it	PRON
easat-8918	151	22	remains	remain	VERB
easat-8918	151	23	a	a	DET
easat-8918	151	24	classic	classic	ADJ
easat-8918	151	25	yet	yet	CCONJ
easat-8918	151	26	active	active	ADJ
easat-8918	151	27	topic	topic	NOUN
easat-8918	151	28	in	in	ADP
easat-8918	151	29	low	low	ADJ
easat-8918	151	30	-	-	PUNCT
easat-8918	151	31	level	level	NOUN
easat-8918	151	32	vision	vision	NOUN
easat-8918	151	33	.	.	PUNCT
easat-8918	152	1	the	the	DET
easat-8918	152	2	goal	goal	NOUN
easat-8918	152	3	of	of	ADP
easat-8918	152	4	image	image	NOUN
easat-8918	152	5	denoising	denoising	NOUN
easat-8918	152	6	is	be	AUX
easat-8918	152	7	to	to	PART
easat-8918	152	8	obtain	obtain	VERB
easat-8918	152	9	a	a	DET
easat-8918	152	10	clean	clean	ADJ
easat-8918	152	11	image	image	NOUN
easat-8918	152	12	x	x	VERB
easat-8918	152	13	from	from	ADP
easat-8918	152	14	a	a	DET
easat-8918	152	15	noisy	noisy	ADJ
easat-8918	152	16	signal	signal	NOUN
easat-8918	152	17	y	y	PROPN
easat-8918	152	18	that	that	PRON
easat-8918	152	19	follows	follow	VERB
easat-8918	152	20	the	the	DET
easat-8918	152	21	image	image	NOUN
easat-8918	152	22	degradation	degradation	NOUN
easat-8918	152	23	model	model	NOUN
easat-8918	152	24	y	y	PROPN
easat-8918	152	25	=	=	PUNCT
easat-8918	152	26	x	x	PROPN
easat-8918	153	1	+	+	CCONJ
easat-8918	153	2	v.	v.	ADP
easat-8918	153	3	a	a	DET
easat-8918	153	4	common	common	ADJ
easat-8918	153	5	assumption	assumption	NOUN
easat-8918	153	6	is	be	AUX
easat-8918	153	7	that	that	SCONJ
easat-8918	153	8	v	v	NOUN
easat-8918	153	9	is	be	AUX
easat-8918	153	10	additive	additive	ADJ
easat-8918	153	11	white	white	ADJ
easat-8918	153	12	gaussian	gaussian	PROPN
easat-8918	153	13	noise	noise	NOUN
easat-8918	153	14	(	(	PUNCT
easat-8918	153	15	awgn	awgn	NOUN
easat-8918	153	16	)	)	PUNCT
easat-8918	153	17	with	with	ADP
easat-8918	153	18	standard	standard	ADJ
easat-8918	153	19	deviation	deviation	NOUN
easat-8918	153	20	σ	σ	PROPN
easat-8918	153	21	.	.	PROPN
easat-8918	153	22	from	from	ADP
easat-8918	153	23	a	a	DET
easat-8918	153	24	bayesian	bayesian	NOUN
easat-8918	153	25	perspective	perspective	NOUN
easat-8918	153	26	,	,	PUNCT
easat-8918	153	27	when	when	SCONJ
easat-8918	153	28	the	the	DET
easat-8918	153	29	probability	probability	NOUN
easat-8918	153	30	distribution	distribution	NOUN
easat-8918	153	31	is	be	AUX
easat-8918	153	32	known	know	VERB
easat-8918	153	33	,	,	PUNCT
easat-8918	153	34	modeling	model	VERB
easat-8918	153	35	the	the	DET
easat-8918	153	36	prior	prior	NOUN
easat-8918	153	37	of	of	ADP
easat-8918	153	38	the	the	DET
easat-8918	153	39	image	image	NOUN
easat-8918	153	40	plays	play	VERB
easat-8918	153	41	a	a	DET
easat-8918	153	42	central	central	ADJ
easat-8918	153	43	role	role	NOUN
easat-8918	153	44	in	in	ADP
easat-8918	153	45	noise	noise	NOUN
easat-8918	153	46	removal	removal	NOUN
easat-8918	153	47	.	.	PUNCT
easat-8918	154	1	in	in	ADP
easat-8918	154	2	this	this	DET
easat-8918	154	3	section	section	NOUN
easat-8918	154	4	,	,	PUNCT
easat-8918	154	5	comparisons	comparison	NOUN
easat-8918	154	6	are	be	AUX
easat-8918	154	7	made	make	VERB
easat-8918	154	8	between	between	ADP
easat-8918	154	9	two	two	NUM
easat-8918	154	10	recently	recently	ADV
easat-8918	154	11	introduced	introduce	VERB
easat-8918	154	12	algorithms	algorithm	NOUN
easat-8918	154	13	for	for	ADP
easat-8918	154	14	image	image	NOUN
easat-8918	154	15	denoising	denoising	NOUN
easat-8918	154	16	,	,	PUNCT
easat-8918	154	17	as	as	SCONJ
easat-8918	154	18	discussed	discuss	VERB
easat-8918	154	19	in	in	ADP
easat-8918	154	20	the	the	DET
easat-8918	154	21	previous	previous	ADJ
easat-8918	154	22	section	section	NOUN
easat-8918	154	23	.	.	PUNCT
easat-8918	155	1	different	different	ADJ
easat-8918	155	2	images	image	NOUN
easat-8918	155	3	from	from	ADP
easat-8918	155	4	the	the	DET
easat-8918	155	5	bsd68	bsd68	PROPN
easat-8918	155	6	dataset	dataset	NOUN
easat-8918	155	7	were	be	AUX
easat-8918	155	8	used	use	VERB
easat-8918	155	9	for	for	ADP
easat-8918	155	10	the	the	DET
easat-8918	155	11	comparisons	comparison	NOUN
easat-8918	155	12	.	.	PUNCT
easat-8918	156	1	for	for	ADP
easat-8918	156	2	each	each	DET
easat-8918	156	3	resulting	result	VERB
easat-8918	156	4	image	image	NOUN
easat-8918	156	5	,	,	PUNCT
easat-8918	156	6	the	the	DET
easat-8918	156	7	peak	peak	NOUN
easat-8918	156	8	signal	signal	NOUN
easat-8918	156	9	-	-	PUNCT
easat-8918	156	10	to	to	ADP
easat-8918	156	11	-	-	PUNCT
easat-8918	156	12	noise	noise	NOUN
easat-8918	156	13	ratio	ratio	NOUN
easat-8918	156	14	(	(	PUNCT
easat-8918	156	15	psnr	psnr	NOUN
easat-8918	156	16	)	)	PUNCT
easat-8918	156	17	was	be	AUX
easat-8918	156	18	calculated	calculate	VERB
easat-8918	156	19	.	.	PUNCT
easat-8918	157	1	psnr	psnr	NOUN
easat-8918	157	2	calculates	calculate	VERB
easat-8918	157	3	the	the	DET
easat-8918	157	4	peak	peak	NOUN
easat-8918	157	5	signal	signal	NOUN
easat-8918	157	6	-	-	PUNCT
easat-8918	157	7	to	to	ADP
easat-8918	157	8	-	-	PUNCT
easat-8918	157	9	noise	noise	NOUN
easat-8918	157	10	ratio	ratio	NOUN
easat-8918	157	11	between	between	ADP
easat-8918	157	12	two	two	NUM
easat-8918	157	13	images	image	NOUN
easat-8918	157	14	in	in	ADP
easat-8918	157	15	decibels	decibel	NOUN
easat-8918	157	16	.	.	PUNCT
easat-8918	158	1	this	this	DET
easat-8918	158	2	ratio	ratio	NOUN
easat-8918	158	3	is	be	AUX
easat-8918	158	4	used	use	VERB
easat-8918	158	5	as	as	ADP
easat-8918	158	6	a	a	DET
easat-8918	158	7	quality	quality	NOUN
easat-8918	158	8	metric	metric	NOUN
easat-8918	158	9	between	between	ADP
easat-8918	158	10	the	the	DET
easat-8918	158	11	original	original	ADJ
easat-8918	158	12	image	image	NOUN
easat-8918	158	13	and	and	CCONJ
easat-8918	158	14	the	the	DET
easat-8918	158	15	compressed	compressed	ADJ
easat-8918	158	16	or	or	CCONJ
easat-8918	158	17	denoised	denoise	VERB
easat-8918	158	18	version	version	NOUN
easat-8918	158	19	.	.	PUNCT
easat-8918	159	1	the	the	PRON
easat-8918	159	2	higher	high	ADJ
easat-8918	159	3	the	the	DET
easat-8918	159	4	psnr	psnr	NOUN
easat-8918	159	5	value	value	NOUN
easat-8918	159	6	,	,	PUNCT
easat-8918	159	7	the	the	PRON
easat-8918	159	8	better	well	ADJ
easat-8918	159	9	the	the	DET
easat-8918	159	10	quality	quality	NOUN
easat-8918	159	11	of	of	ADP
easat-8918	159	12	the	the	DET
easat-8918	159	13	denoised	denoise	VERB
easat-8918	159	14	image	image	NOUN
easat-8918	159	15	.	.	PUNCT
easat-8918	160	1	to	to	PART
easat-8918	160	2	compute	compute	VERB
easat-8918	160	3	psnr	psnr	NOUN
easat-8918	160	4	,	,	PUNCT
easat-8918	160	5	the	the	DET
easat-8918	160	6	mean	mean	NOUN
easat-8918	160	7	squared	square	VERB
easat-8918	160	8	error	error	NOUN
easat-8918	160	9	(	(	PUNCT
easat-8918	160	10	mse	mse	NOUN
easat-8918	160	11	)	)	PUNCT
easat-8918	160	12	is	be	AUX
easat-8918	160	13	first	first	ADV
easat-8918	160	14	calculated	calculate	VERB
easat-8918	160	15	using	use	VERB
easat-8918	160	16	equation	equation	NOUN
easat-8918	160	17	(	(	PUNCT
easat-8918	160	18	1	1	NUM
easat-8918	160	19	):	):	PUNCT
easat-8918	160	20	𝑀𝑆𝐸	𝑀𝑆𝐸	PROPN
easat-8918	160	21	=	=	SYM
easat-8918	160	22	∑𝑀,𝑁	∑𝑀,𝑁	PROPN
easat-8918	161	1	[	[	X
easat-8918	161	2	𝐼1(𝑚,𝑛)−𝐼2(𝑚,𝑛)]2	𝐼1(𝑚,𝑛)−𝐼2(𝑚,𝑛)]2	ADJ
easat-8918	161	3	𝑀×𝑁	𝑀×𝑁	NOUN
easat-8918	161	4	(	(	PUNCT
easat-8918	161	5	1	1	NUM
easat-8918	161	6	)	)	PUNCT
easat-8918	161	7	in	in	ADP
easat-8918	161	8	the	the	DET
easat-8918	161	9	above	above	ADJ
easat-8918	161	10	equation	equation	NOUN
easat-8918	161	11	,	,	PUNCT
easat-8918	161	12	m	m	PROPN
easat-8918	161	13	and	and	CCONJ
easat-8918	161	14	n	n	PRON
easat-8918	161	15	represent	represent	VERB
easat-8918	161	16	the	the	DET
easat-8918	161	17	number	number	NOUN
easat-8918	161	18	of	of	ADP
easat-8918	161	19	rows	row	NOUN
easat-8918	161	20	and	and	CCONJ
easat-8918	161	21	columns	column	NOUN
easat-8918	161	22	in	in	ADP
easat-8918	161	23	the	the	DET
easat-8918	161	24	image	image	NOUN
easat-8918	161	25	matrix	matrix	NOUN
easat-8918	161	26	.	.	PUNCT
easat-8918	162	1	then	then	ADV
easat-8918	162	2	,	,	PUNCT
easat-8918	162	3	using	use	VERB
easat-8918	162	4	equation	equation	NOUN
easat-8918	162	5	(	(	PUNCT
easat-8918	162	6	2	2	NUM
easat-8918	162	7	)	)	PUNCT
easat-8918	162	8	,	,	PUNCT
easat-8918	162	9	the	the	DET
easat-8918	162	10	psnr	psnr	NOUN
easat-8918	162	11	is	be	AUX
easat-8918	162	12	calculated	calculate	VERB
easat-8918	162	13	as	as	SCONJ
easat-8918	162	14	follows	follow	VERB
easat-8918	162	15	:	:	PUNCT
easat-8918	162	16	𝑃𝑆𝑁𝑅	𝑃𝑆𝑁𝑅	NOUN
easat-8918	162	17	=	=	PUNCT
easat-8918	162	18	10	10	NUM
easat-8918	162	19	𝑙𝑜𝑔10	𝑙𝑜𝑔10	X
easat-8918	162	20	(	(	PUNCT
easat-8918	162	21	2552	2552	NUM
easat-8918	162	22	𝑀𝑆𝐸	𝑀𝑆𝐸	NOUN
easat-8918	162	23	)	)	PUNCT
easat-8918	162	24	(	(	PUNCT
easat-8918	162	25	2	2	X
easat-8918	162	26	)	)	PUNCT
easat-8918	162	27	the	the	DET
easat-8918	162	28	psnr	psnr	NOUN
easat-8918	162	29	results	result	VERB
easat-8918	162	30	for	for	ADP
easat-8918	162	31	an	an	DET
easat-8918	162	32	image	image	NOUN
easat-8918	162	33	from	from	ADP
easat-8918	162	34	the	the	DET
easat-8918	162	35	bsd68	bsd68	PROPN
easat-8918	162	36	dataset	dataset	VERB
easat-8918	162	37	with	with	ADP
easat-8918	162	38	gaussian	gaussian	ADJ
easat-8918	162	39	noise	noise	NOUN
easat-8918	162	40	added	add	VERB
easat-8918	162	41	at	at	ADP
easat-8918	162	42	a	a	DET
easat-8918	162	43	level	level	NOUN
easat-8918	162	44	of	of	ADP
easat-8918	162	45	50	50	NUM
easat-8918	162	46	were	be	AUX
easat-8918	162	47	observed	observe	VERB
easat-8918	162	48	and	and	CCONJ
easat-8918	162	49	compared	compare	VERB
easat-8918	162	50	using	use	VERB
easat-8918	162	51	different	different	ADJ
easat-8918	162	52	denoising	denoising	NOUN
easat-8918	162	53	algorithms	algorithm	NOUN
easat-8918	162	54	,	,	PUNCT
easat-8918	162	55	as	as	SCONJ
easat-8918	162	56	shown	show	VERB
easat-8918	162	57	in	in	ADP
easat-8918	162	58	figure	figure	NOUN
easat-8918	162	59	10	10	NUM
easat-8918	162	60	.	.	PUNCT
easat-8918	163	1	among	among	ADP
easat-8918	163	2	the	the	DET
easat-8918	163	3	five	five	NUM
easat-8918	163	4	algorithms	algorithm	NOUN
easat-8918	163	5	compared	compare	VERB
easat-8918	163	6	,	,	PUNCT
easat-8918	163	7	bm3d	bm3d	PROPN
easat-8918	163	8	had	have	VERB
easat-8918	163	9	the	the	DET
easat-8918	163	10	lowest	low	ADJ
easat-8918	163	11	performance	performance	NOUN
easat-8918	163	12	with	with	ADP
easat-8918	163	13	a	a	DET
easat-8918	163	14	psnr	psnr	NOUN
easat-8918	163	15	of	of	ADP
easat-8918	163	16	26.21	26.21	NUM
easat-8918	163	17	db	db	NOUN
easat-8918	163	18	,	,	PUNCT
easat-8918	163	19	while	while	SCONJ
easat-8918	163	20	restormer	restormer	PROPN
easat-8918	163	21	achieved	achieve	VERB
easat-8918	163	22	the	the	DET
easat-8918	163	23	highest	high	ADJ
easat-8918	163	24	psnr	psnr	NOUN
easat-8918	163	25	value	value	NOUN
easat-8918	163	26	with	with	ADP
easat-8918	163	27	27.849	27.849	NUM
easat-8918	163	28	db	db	PROPN
easat-8918	163	29	.	.	PROPN
easat-8918	163	30	1353	1353	NUM
easat-8918	163	31	edelweiss	edelweiss	PROPN
easat-8918	163	32	applied	apply	VERB
easat-8918	163	33	science	science	NOUN
easat-8918	163	34	and	and	CCONJ
easat-8918	163	35	technology	technology	NOUN
easat-8918	163	36	issn	issn	PROPN
easat-8918	163	37	:	:	PUNCT
easat-8918	163	38	2576	2576	NUM
easat-8918	163	39	-	-	SYM
easat-8918	163	40	8484	8484	NUM
easat-8918	163	41	vol	vol	NOUN
easat-8918	163	42	.	.	PROPN
easat-8918	164	1	9	9	NUM
easat-8918	164	2	,	,	PUNCT
easat-8918	164	3	no	no	INTJ
easat-8918	164	4	.	.	NOUN
easat-8918	164	5	7	7	NUM
easat-8918	164	6	:	:	SYM
easat-8918	164	7	1345	1345	NUM
easat-8918	164	8	-	-	SYM
easat-8918	164	9	1359	1359	NUM
easat-8918	164	10	,	,	PUNCT
easat-8918	164	11	2025	2025	NUM
easat-8918	164	12	doi	doi	NOUN
easat-8918	164	13	:	:	PUNCT
easat-8918	164	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	164	15	©	©	PROPN
easat-8918	164	16	2025	2025	NUM
easat-8918	164	17	by	by	ADP
easat-8918	164	18	the	the	DET
easat-8918	164	19	author	author	NOUN
easat-8918	164	20	;	;	PUNCT
easat-8918	164	21	licensee	licensee	PROPN
easat-8918	164	22	learning	learn	VERB
easat-8918	164	23	gate	gate	NOUN
easat-8918	164	24	original	original	ADJ
easat-8918	164	25	image	image	NOUN
easat-8918	164	26	noisy	noisy	ADJ
easat-8918	164	27	image	image	NOUN
easat-8918	164	28	/	/	SYM
easat-8918	164	29	14.76db	14.76db	NUM
easat-8918	164	30	denoised	denoise	VERB
easat-8918	164	31	bm3d	bm3d	PROPN
easat-8918	164	32	/	/	SYM
easat-8918	164	33	26.21db	26.21db	NUM
easat-8918	164	34	denoised	denoise	VERB
easat-8918	164	35	tnrd	tnrd	NOUN
easat-8918	164	36	/	/	SYM
easat-8918	164	37	26.59db	26.59db	NUM
easat-8918	164	38	denoised	denoise	VERB
easat-8918	164	39	dncnn	dncnn	ADJ
easat-8918	164	40	/	/	SYM
easat-8918	164	41	26.90db	26.90db	NUM
easat-8918	164	42	denoised	denoise	VERB
easat-8918	164	43	restormer	restormer	NOUN
easat-8918	164	44	/	/	PUNCT
easat-8918	164	45	27.849	27.849	NUM
easat-8918	164	46	db	db	PROPN
easat-8918	164	47	figure	figure	NOUN
easat-8918	164	48	10	10	NUM
easat-8918	164	49	.	.	PUNCT
easat-8918	165	1	comparison	comparison	NOUN
easat-8918	165	2	of	of	ADP
easat-8918	165	3	denoising	denoising	NOUN
easat-8918	165	4	results	result	NOUN
easat-8918	165	5	for	for	ADP
easat-8918	165	6	image	image	NOUN
easat-8918	165	7	0010	0010	NUM
easat-8918	165	8	from	from	ADP
easat-8918	165	9	the	the	DET
easat-8918	165	10	bsd68	bsd68	PROPN
easat-8918	165	11	dataset	dataset	NOUN
easat-8918	165	12	at	at	ADP
easat-8918	165	13	gaussian	gaussian	ADJ
easat-8918	165	14	noise	noise	NOUN
easat-8918	165	15	level	level	NOUN
easat-8918	165	16	50	50	NUM
easat-8918	165	17	.	.	PUNCT
easat-8918	166	1	another	another	DET
easat-8918	166	2	comparison	comparison	NOUN
easat-8918	166	3	was	be	AUX
easat-8918	166	4	carried	carry	VERB
easat-8918	166	5	out	out	ADP
easat-8918	166	6	on	on	ADP
easat-8918	166	7	the	the	DET
easat-8918	166	8	images	image	NOUN
easat-8918	166	9	“	"	PUNCT
easat-8918	166	10	cameraman	cameraman	ADJ
easat-8918	166	11	,	,	PUNCT
easat-8918	166	12	”	"	PUNCT
easat-8918	166	13	“	"	PUNCT
easat-8918	166	14	bsd68	bsd68	PROPN
easat-8918	166	15	-	-	PUNCT
easat-8918	166	16	10	10	NUM
easat-8918	166	17	,	,	PUNCT
easat-8918	166	18	”	"	PUNCT
easat-8918	166	19	and	and	CCONJ
easat-8918	166	20	“	"	PUNCT
easat-8918	166	21	bsd68	bsd68	PROPN
easat-8918	166	22	-	-	PUNCT
easat-8918	166	23	09	09	NUM
easat-8918	166	24	”	"	PUNCT
easat-8918	166	25	with	with	ADP
easat-8918	166	26	different	different	ADJ
easat-8918	166	27	levels	level	NOUN
easat-8918	166	28	of	of	ADP
easat-8918	166	29	gaussian	gaussian	ADJ
easat-8918	166	30	noise	noise	NOUN
easat-8918	166	31	(	(	PUNCT
easat-8918	166	32	10	10	NUM
easat-8918	166	33	,	,	PUNCT
easat-8918	166	34	15	15	NUM
easat-8918	166	35	,	,	PUNCT
easat-8918	166	36	25	25	NUM
easat-8918	166	37	,	,	PUNCT
easat-8918	166	38	and	and	CCONJ
easat-8918	166	39	50	50	NUM
easat-8918	166	40	)	)	PUNCT
easat-8918	166	41	using	use	VERB
easat-8918	166	42	the	the	DET
easat-8918	166	43	restormer	restormer	NOUN
easat-8918	166	44	denoising	denoising	NOUN
easat-8918	166	45	algorithm	algorithm	NOUN
easat-8918	166	46	.	.	PUNCT
easat-8918	167	1	it	it	PRON
easat-8918	167	2	is	be	AUX
easat-8918	167	3	evident	evident	ADJ
easat-8918	167	4	that	that	SCONJ
easat-8918	167	5	the	the	DET
easat-8918	167	6	psnr	psnr	NOUN
easat-8918	167	7	value	value	NOUN
easat-8918	167	8	increases	increase	VERB
easat-8918	167	9	inversely	inversely	ADV
easat-8918	167	10	with	with	ADP
easat-8918	167	11	the	the	DET
easat-8918	167	12	noise	noise	NOUN
easat-8918	167	13	level	level	NOUN
easat-8918	167	14	.	.	PUNCT
easat-8918	168	1	the	the	PRON
easat-8918	168	2	higher	high	ADJ
easat-8918	168	3	the	the	DET
easat-8918	168	4	gaussian	gaussian	ADJ
easat-8918	168	5	noise	noise	NOUN
easat-8918	168	6	,	,	PUNCT
easat-8918	168	7	the	the	PRON
easat-8918	168	8	lower	low	ADJ
easat-8918	168	9	the	the	DET
easat-8918	168	10	psnr	psnr	NOUN
easat-8918	168	11	value	value	NOUN
easat-8918	168	12	,	,	PUNCT
easat-8918	168	13	and	and	CCONJ
easat-8918	168	14	vice	vice	ADV
easat-8918	168	15	versa	versa	ADV
easat-8918	168	16	.	.	PUNCT
easat-8918	169	1	the	the	DET
easat-8918	169	2	psnr	psnr	NOUN
easat-8918	169	3	results	result	VERB
easat-8918	169	4	for	for	ADP
easat-8918	169	5	the	the	DET
easat-8918	169	6	denoised	denoise	VERB
easat-8918	169	7	images	image	NOUN
easat-8918	169	8	are	be	AUX
easat-8918	169	9	presented	present	VERB
easat-8918	169	10	in	in	ADP
easat-8918	169	11	figures	figure	NOUN
easat-8918	169	12	11	11	NUM
easat-8918	169	13	,	,	PUNCT
easat-8918	169	14	12	12	NUM
easat-8918	169	15	,	,	PUNCT
easat-8918	169	16	and	and	CCONJ
easat-8918	169	17	13	13	NUM
easat-8918	169	18	.	.	PUNCT
easat-8918	170	1	for	for	ADP
easat-8918	170	2	a	a	DET
easat-8918	170	3	better	well	ADJ
easat-8918	170	4	comparison	comparison	NOUN
easat-8918	170	5	,	,	PUNCT
easat-8918	170	6	a	a	DET
easat-8918	170	7	summary	summary	NOUN
easat-8918	170	8	of	of	ADP
easat-8918	170	9	the	the	DET
easat-8918	170	10	denoising	denoising	NOUN
easat-8918	170	11	results	result	NOUN
easat-8918	170	12	is	be	AUX
easat-8918	170	13	listed	list	VERB
easat-8918	170	14	in	in	ADP
easat-8918	170	15	table	table	NOUN
easat-8918	170	16	1	1	NUM
easat-8918	170	17	.	.	SYM
easat-8918	170	18	1354	1354	NUM
easat-8918	170	19	edelweiss	edelweiss	PROPN
easat-8918	170	20	applied	apply	VERB
easat-8918	170	21	science	science	NOUN
easat-8918	170	22	and	and	CCONJ
easat-8918	170	23	technology	technology	NOUN
easat-8918	170	24	issn	issn	PROPN
easat-8918	170	25	:	:	PUNCT
easat-8918	170	26	2576	2576	NUM
easat-8918	170	27	-	-	SYM
easat-8918	170	28	8484	8484	NUM
easat-8918	170	29	vol	vol	NOUN
easat-8918	170	30	.	.	PROPN
easat-8918	171	1	9	9	NUM
easat-8918	171	2	,	,	PUNCT
easat-8918	171	3	no	no	INTJ
easat-8918	171	4	.	.	NOUN
easat-8918	171	5	7	7	NUM
easat-8918	171	6	:	:	SYM
easat-8918	171	7	1345	1345	NUM
easat-8918	171	8	-	-	SYM
easat-8918	171	9	1359	1359	NUM
easat-8918	171	10	,	,	PUNCT
easat-8918	171	11	2025	2025	NUM
easat-8918	171	12	doi	doi	NOUN
easat-8918	171	13	:	:	PUNCT
easat-8918	171	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	171	15	©	©	PROPN
easat-8918	171	16	2025	2025	NUM
easat-8918	171	17	by	by	ADP
easat-8918	171	18	the	the	DET
easat-8918	171	19	author	author	NOUN
easat-8918	171	20	;	;	PUNCT
easat-8918	171	21	licensee	licensee	PROPN
easat-8918	171	22	learning	learn	VERB
easat-8918	171	23	gate	gate	NOUN
easat-8918	171	24	σ	σ	PROPN
easat-8918	171	25	=	=	PROPN
easat-8918	171	26	10	10	NUM
easat-8918	171	27	σ	σ	NOUN
easat-8918	171	28	=	=	SYM
easat-8918	171	29	15	15	NUM
easat-8918	171	30	σ	σ	NOUN
easat-8918	171	31	=	=	SYM
easat-8918	171	32	25	25	NUM
easat-8918	171	33	σ	σ	NOUN
easat-8918	171	34	=	=	SYM
easat-8918	171	35	50	50	NUM
easat-8918	171	36	psnr	psnr	NOUN
easat-8918	171	37	30.50	30.50	NUM
easat-8918	171	38	db	db	PROPN
easat-8918	171	39	psnr	psnr	NOUN
easat-8918	171	40	29.30	29.30	NUM
easat-8918	171	41	db	db	PROPN
easat-8918	171	42	psnr	psnr	NOUN
easat-8918	171	43	27.97	27.97	NUM
easat-8918	171	44	db	db	PROPN
easat-8918	171	45	psnr	psnr	NOUN
easat-8918	171	46	27.96	27.96	NUM
easat-8918	171	47	db	db	PROPN
easat-8918	171	48	figure	figure	NOUN
easat-8918	171	49	11	11	NUM
easat-8918	171	50	.	.	PUNCT
easat-8918	172	1	gaussian	gaussian	ADJ
easat-8918	172	2	image	image	NOUN
easat-8918	172	3	denoising	denoise	VERB
easat-8918	172	4	on	on	ADP
easat-8918	172	5	the	the	DET
easat-8918	172	6	cameraman	cameraman	ADJ
easat-8918	172	7	image	image	NOUN
easat-8918	172	8	using	use	VERB
easat-8918	172	9	the	the	DET
easat-8918	172	10	restormer	restormer	ADJ
easat-8918	172	11	algorithm	algorithm	NOUN
easat-8918	172	12	at	at	ADP
easat-8918	172	13	noise	noise	NOUN
easat-8918	172	14	levels	level	NOUN
easat-8918	172	15	10	10	NUM
easat-8918	172	16	,	,	PUNCT
easat-8918	172	17	15	15	NUM
easat-8918	172	18	,	,	PUNCT
easat-8918	172	19	25	25	NUM
easat-8918	172	20	,	,	PUNCT
easat-8918	172	21	and	and	CCONJ
easat-8918	172	22	50	50	NUM
easat-8918	172	23	.	.	PUNCT
easat-8918	173	1	σ	σ	NOUN
easat-8918	173	2	=	=	SYM
easat-8918	173	3	10	10	NUM
easat-8918	173	4	σ	σ	NOUN
easat-8918	173	5	=	=	SYM
easat-8918	173	6	15	15	NUM
easat-8918	173	7	σ	σ	NOUN
easat-8918	173	8	=	=	SYM
easat-8918	173	9	25	25	NUM
easat-8918	173	10	σ	σ	NOUN
easat-8918	173	11	=	=	NOUN
easat-8918	173	12	50	50	NUM
easat-8918	173	13	1355	1355	NUM
easat-8918	173	14	edelweiss	edelweiss	PROPN
easat-8918	173	15	applied	apply	VERB
easat-8918	173	16	science	science	NOUN
easat-8918	173	17	and	and	CCONJ
easat-8918	173	18	technology	technology	NOUN
easat-8918	173	19	issn	issn	PROPN
easat-8918	173	20	:	:	PUNCT
easat-8918	173	21	2576	2576	NUM
easat-8918	173	22	-	-	SYM
easat-8918	173	23	8484	8484	NUM
easat-8918	173	24	vol	vol	NOUN
easat-8918	173	25	.	.	PROPN
easat-8918	173	26	9	9	NUM
easat-8918	173	27	,	,	PUNCT
easat-8918	173	28	no	no	INTJ
easat-8918	173	29	.	.	NOUN
easat-8918	173	30	7	7	NUM
easat-8918	173	31	:	:	SYM
easat-8918	173	32	1345	1345	NUM
easat-8918	173	33	-	-	SYM
easat-8918	173	34	1359	1359	NUM
easat-8918	173	35	,	,	PUNCT
easat-8918	173	36	2025	2025	NUM
easat-8918	173	37	doi	doi	NOUN
easat-8918	173	38	:	:	PUNCT
easat-8918	173	39	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	173	40	©	©	PROPN
easat-8918	173	41	2025	2025	NUM
easat-8918	173	42	by	by	ADP
easat-8918	173	43	the	the	DET
easat-8918	173	44	author	author	NOUN
easat-8918	173	45	;	;	PUNCT
easat-8918	173	46	licensee	licensee	PROPN
easat-8918	173	47	learning	learn	VERB
easat-8918	173	48	gate	gate	PROPN
easat-8918	173	49	psnr	psnr	PROPN
easat-8918	173	50	30.65	30.65	NUM
easat-8918	173	51	db	db	PROPN
easat-8918	173	52	psnr	psnr	NOUN
easat-8918	173	53	29.4	29.4	NUM
easat-8918	173	54	db	db	PROPN
easat-8918	173	55	psnr	psnr	NOUN
easat-8918	173	56	28.5	28.5	NUM
easat-8918	173	57	db	db	PROPN
easat-8918	173	58	psnr	psnr	NOUN
easat-8918	173	59	27.91	27.91	NUM
easat-8918	173	60	db	db	PROPN
easat-8918	173	61	figure	figure	NOUN
easat-8918	173	62	12	12	NUM
easat-8918	173	63	.	.	PUNCT
easat-8918	174	1	gaussian	gaussian	ADJ
easat-8918	174	2	image	image	NOUN
easat-8918	174	3	denoising	denoise	VERB
easat-8918	174	4	on	on	ADP
easat-8918	174	5	image	image	NOUN
easat-8918	174	6	0010	0010	NUM
easat-8918	174	7	from	from	ADP
easat-8918	174	8	the	the	DET
easat-8918	174	9	bsd68	bsd68	PROPN
easat-8918	174	10	dataset	dataset	NOUN
easat-8918	174	11	using	use	VERB
easat-8918	174	12	the	the	DET
easat-8918	174	13	restormer	restormer	ADJ
easat-8918	174	14	algorithm	algorithm	NOUN
easat-8918	174	15	at	at	ADP
easat-8918	174	16	noise	noise	NOUN
easat-8918	174	17	levels	level	NOUN
easat-8918	174	18	10	10	NUM
easat-8918	174	19	,	,	PUNCT
easat-8918	174	20	15	15	NUM
easat-8918	174	21	,	,	PUNCT
easat-8918	174	22	25	25	NUM
easat-8918	174	23	,	,	PUNCT
easat-8918	174	24	and	and	CCONJ
easat-8918	174	25	50	50	NUM
easat-8918	174	26	.	.	PUNCT
easat-8918	175	1	σ	σ	NOUN
easat-8918	175	2	=	=	SYM
easat-8918	175	3	10	10	NUM
easat-8918	175	4	σ	σ	NOUN
easat-8918	175	5	=	=	SYM
easat-8918	175	6	15	15	NUM
easat-8918	175	7	σ	σ	NOUN
easat-8918	175	8	=	=	SYM
easat-8918	175	9	25	25	NUM
easat-8918	175	10	σ	σ	NOUN
easat-8918	175	11	=	=	SYM
easat-8918	175	12	50	50	NUM
easat-8918	175	13	psnr	psnr	NOUN
easat-8918	175	14	31.51	31.51	NUM
easat-8918	175	15	db	db	PROPN
easat-8918	175	16	psnr	psnr	NOUN
easat-8918	175	17	30.17	30.17	NUM
easat-8918	175	18	db	db	PROPN
easat-8918	175	19	psnr	psnr	NOUN
easat-8918	175	20	29.05	29.05	NUM
easat-8918	175	21	db	db	PROPN
easat-8918	175	22	psnr	psnr	NOUN
easat-8918	175	23	27.93	27.93	NUM
easat-8918	175	24	db	db	PROPN
easat-8918	175	25	figure	figure	NOUN
easat-8918	175	26	13	13	NUM
easat-8918	175	27	.	.	PUNCT
easat-8918	176	1	gaussian	gaussian	ADJ
easat-8918	176	2	image	image	NOUN
easat-8918	176	3	denoising	denoise	VERB
easat-8918	176	4	on	on	ADP
easat-8918	176	5	image	image	NOUN
easat-8918	176	6	0009	0009	NUM
easat-8918	176	7	from	from	ADP
easat-8918	176	8	the	the	DET
easat-8918	176	9	bsd68	bsd68	PROPN
easat-8918	176	10	dataset	dataset	NOUN
easat-8918	176	11	using	use	VERB
easat-8918	176	12	the	the	DET
easat-8918	176	13	restormer	restormer	ADJ
easat-8918	176	14	algorithm	algorithm	NOUN
easat-8918	176	15	at	at	ADP
easat-8918	176	16	noise	noise	NOUN
easat-8918	176	17	levels	level	NOUN
easat-8918	176	18	10	10	NUM
easat-8918	176	19	,	,	PUNCT
easat-8918	176	20	15	15	NUM
easat-8918	176	21	,	,	PUNCT
easat-8918	176	22	25	25	NUM
easat-8918	176	23	,	,	PUNCT
easat-8918	176	24	and	and	CCONJ
easat-8918	176	25	50	50	NUM
easat-8918	176	26	.	.	X
easat-8918	177	1	1356	1356	NUM
easat-8918	177	2	edelweiss	edelweiss	PROPN
easat-8918	177	3	applied	apply	VERB
easat-8918	177	4	science	science	NOUN
easat-8918	177	5	and	and	CCONJ
easat-8918	177	6	technology	technology	NOUN
easat-8918	177	7	issn	issn	PROPN
easat-8918	177	8	:	:	PUNCT
easat-8918	177	9	2576	2576	NUM
easat-8918	177	10	-	-	SYM
easat-8918	177	11	8484	8484	NUM
easat-8918	177	12	vol	vol	NOUN
easat-8918	177	13	.	.	PROPN
easat-8918	178	1	9	9	NUM
easat-8918	178	2	,	,	PUNCT
easat-8918	178	3	no	no	INTJ
easat-8918	178	4	.	.	NOUN
easat-8918	178	5	7	7	NUM
easat-8918	178	6	:	:	SYM
easat-8918	178	7	1345	1345	NUM
easat-8918	178	8	-	-	SYM
easat-8918	178	9	1359	1359	NUM
easat-8918	178	10	,	,	PUNCT
easat-8918	178	11	2025	2025	NUM
easat-8918	178	12	doi	doi	NOUN
easat-8918	178	13	:	:	PUNCT
easat-8918	178	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	178	15	©	©	PROPN
easat-8918	178	16	2025	2025	NUM
easat-8918	178	17	by	by	ADP
easat-8918	178	18	the	the	DET
easat-8918	178	19	author	author	NOUN
easat-8918	178	20	;	;	PUNCT
easat-8918	178	21	licensee	licensee	PROPN
easat-8918	178	22	learning	learn	VERB
easat-8918	178	23	gate	gate	NOUN
easat-8918	178	24	table	table	NOUN
easat-8918	178	25	1	1	NUM
easat-8918	178	26	.	.	PUNCT
easat-8918	179	1	psnr	psnr	NOUN
easat-8918	179	2	(	(	PUNCT
easat-8918	179	3	db	db	PROPN
easat-8918	179	4	)	)	PUNCT
easat-8918	179	5	results	result	NOUN
easat-8918	179	6	obtained	obtain	VERB
easat-8918	179	7	using	use	VERB
easat-8918	179	8	the	the	DET
easat-8918	179	9	restormer	restormer	ADJ
easat-8918	179	10	algorithm	algorithm	NOUN
easat-8918	179	11	at	at	ADP
easat-8918	179	12	noise	noise	NOUN
easat-8918	179	13	levels	level	NOUN
easat-8918	179	14	10	10	NUM
easat-8918	179	15	,	,	PUNCT
easat-8918	179	16	15	15	NUM
easat-8918	179	17	,	,	PUNCT
easat-8918	179	18	25	25	NUM
easat-8918	179	19	,	,	PUNCT
easat-8918	179	20	&	&	CCONJ
easat-8918	179	21	50	50	NUM
easat-8918	179	22	.	.	PUNCT
easat-8918	179	23	noise	noise	NOUN
easat-8918	179	24	level	level	NOUN
easat-8918	179	25	cameraman	cameraman	NOUN
easat-8918	179	26	bsd68	bsd68	PROPN
easat-8918	179	27	-	-	PUNCT
easat-8918	179	28	09	09	NUM
easat-8918	179	29	bsd68	bsd68	PROPN
easat-8918	179	30	-	-	PUNCT
easat-8918	179	31	10	10	NUM
easat-8918	179	32	10	10	NUM
easat-8918	179	33	30.50	30.50	NUM
easat-8918	179	34	31.51	31.51	NUM
easat-8918	179	35	30.65	30.65	NUM
easat-8918	179	36	15	15	NUM
easat-8918	179	37	29.30	29.30	NUM
easat-8918	179	38	30.17	30.17	NUM
easat-8918	179	39	29.4	29.4	NUM
easat-8918	179	40	25	25	NUM
easat-8918	179	41	27.97	27.97	NUM
easat-8918	179	42	29.05	29.05	NUM
easat-8918	179	43	28.5	28.5	NUM
easat-8918	179	44	50	50	NUM
easat-8918	179	45	27.96	27.96	NUM
easat-8918	179	46	27.93	27.93	NUM
easat-8918	179	47	27.91	27.91	NUM
easat-8918	179	48	denoising	denoising	NOUN
easat-8918	179	49	was	be	AUX
easat-8918	179	50	performed	perform	VERB
easat-8918	179	51	on	on	ADP
easat-8918	179	52	the	the	DET
easat-8918	179	53	bsd68	bsd68	PROPN
easat-8918	179	54	dataset	dataset	NOUN
easat-8918	179	55	(	(	PUNCT
easat-8918	179	56	consisting	consist	VERB
easat-8918	179	57	of	of	ADP
easat-8918	179	58	68	68	NUM
easat-8918	179	59	grayscale	grayscale	NOUN
easat-8918	179	60	images	image	NOUN
easat-8918	179	61	)	)	PUNCT
easat-8918	179	62	with	with	ADP
easat-8918	179	63	noise	noise	NOUN
easat-8918	179	64	levels	level	NOUN
easat-8918	179	65	of	of	ADP
easat-8918	179	66	15	15	NUM
easat-8918	179	67	,	,	PUNCT
easat-8918	179	68	25	25	NUM
easat-8918	179	69	,	,	PUNCT
easat-8918	179	70	and	and	CCONJ
easat-8918	179	71	50	50	NUM
easat-8918	179	72	using	use	VERB
easat-8918	179	73	different	different	ADJ
easat-8918	179	74	methods	method	NOUN
easat-8918	179	75	.	.	PUNCT
easat-8918	180	1	the	the	DET
easat-8918	180	2	average	average	ADJ
easat-8918	180	3	psnr	psnr	NOUN
easat-8918	180	4	(	(	PUNCT
easat-8918	180	5	db	db	NOUN
easat-8918	180	6	)	)	PUNCT
easat-8918	180	7	results	result	NOUN
easat-8918	180	8	were	be	AUX
easat-8918	180	9	calculated	calculate	VERB
easat-8918	180	10	and	and	CCONJ
easat-8918	180	11	are	be	AUX
easat-8918	180	12	presented	present	VERB
easat-8918	180	13	in	in	ADP
easat-8918	180	14	table	table	NOUN
easat-8918	180	15	2	2	NUM
easat-8918	180	16	.	.	PUNCT
easat-8918	181	1	the	the	DET
easat-8918	181	2	best	good	ADJ
easat-8918	181	3	results	result	NOUN
easat-8918	181	4	for	for	ADP
easat-8918	181	5	each	each	DET
easat-8918	181	6	noise	noise	NOUN
easat-8918	181	7	level	level	NOUN
easat-8918	181	8	are	be	AUX
easat-8918	181	9	highlighted	highlight	VERB
easat-8918	181	10	in	in	ADP
easat-8918	181	11	bold	bold	ADJ
easat-8918	181	12	.	.	PUNCT
easat-8918	182	1	the	the	DET
easat-8918	182	2	comparison	comparison	NOUN
easat-8918	182	3	shows	show	VERB
easat-8918	182	4	that	that	SCONJ
easat-8918	182	5	the	the	DET
easat-8918	182	6	dncnn	dncnn	NOUN
easat-8918	182	7	and	and	CCONJ
easat-8918	182	8	restormer	restormer	ADJ
easat-8918	182	9	algorithms	algorithm	NOUN
easat-8918	182	10	produce	produce	VERB
easat-8918	182	11	results	result	NOUN
easat-8918	182	12	competitive	competitive	ADJ
easat-8918	182	13	with	with	ADP
easat-8918	182	14	the	the	DET
easat-8918	182	15	bm3d	bm3d	PROPN
easat-8918	182	16	and	and	CCONJ
easat-8918	182	17	tnrd	tnrd	NOUN
easat-8918	182	18	algorithms	algorithm	NOUN
easat-8918	182	19	from	from	ADP
easat-8918	182	20	the	the	DET
easat-8918	182	21	literature	literature	NOUN
easat-8918	182	22	.	.	PUNCT
easat-8918	183	1	while	while	SCONJ
easat-8918	183	2	restormer	restormer	NOUN
easat-8918	183	3	outperforms	outperform	NOUN
easat-8918	183	4	dncnn	dncnn	VERB
easat-8918	183	5	at	at	ADP
easat-8918	183	6	higher	high	ADJ
easat-8918	183	7	noise	noise	NOUN
easat-8918	183	8	levels	level	NOUN
easat-8918	183	9	(	(	PUNCT
easat-8918	183	10	σ	σ	X
easat-8918	183	11	=	=	SYM
easat-8918	183	12	50	50	NUM
easat-8918	183	13	)	)	PUNCT
easat-8918	183	14	,	,	PUNCT
easat-8918	183	15	dncnn	dncnn	PROPN
easat-8918	183	16	yields	yield	VERB
easat-8918	183	17	better	well	ADJ
easat-8918	183	18	performance	performance	NOUN
easat-8918	183	19	at	at	ADP
easat-8918	183	20	lower	low	ADJ
easat-8918	183	21	noise	noise	NOUN
easat-8918	183	22	levels	level	NOUN
easat-8918	183	23	(	(	PUNCT
easat-8918	183	24	σ	σ	NOUN
easat-8918	183	25	=	=	SYM
easat-8918	183	26	25	25	NUM
easat-8918	183	27	and	and	CCONJ
easat-8918	183	28	σ	σ	NUM
easat-8918	183	29	=	=	SYM
easat-8918	183	30	15	15	NUM
easat-8918	183	31	)	)	PUNCT
easat-8918	183	32	.	.	PUNCT
easat-8918	184	1	table	table	NOUN
easat-8918	184	2	2	2	NUM
easat-8918	184	3	.	.	PUNCT
easat-8918	185	1	psnr	psnr	NOUN
easat-8918	185	2	(	(	PUNCT
easat-8918	185	3	db	db	PROPN
easat-8918	185	4	)	)	PUNCT
easat-8918	185	5	results	result	NOUN
easat-8918	185	6	obtained	obtain	VERB
easat-8918	185	7	using	use	VERB
easat-8918	185	8	different	different	ADJ
easat-8918	185	9	methods	method	NOUN
easat-8918	185	10	on	on	ADP
easat-8918	185	11	the	the	DET
easat-8918	185	12	bsd68	bsd68	PROPN
easat-8918	185	13	dataset	dataset	NOUN
easat-8918	185	14	at	at	ADP
easat-8918	185	15	noise	noise	NOUN
easat-8918	185	16	levels	level	NOUN
easat-8918	185	17	15	15	NUM
easat-8918	185	18	,	,	PUNCT
easat-8918	185	19	25	25	NUM
easat-8918	185	20	,	,	PUNCT
easat-8918	185	21	&	&	CCONJ
easat-8918	185	22	50	50	NUM
easat-8918	185	23	.	.	PUNCT
easat-8918	186	1	noise	noise	NOUN
easat-8918	186	2	level	level	NOUN
easat-8918	186	3	bm3d	bm3d	PROPN
easat-8918	186	4	tnrsd	tnrsd	NOUN
easat-8918	186	5	dncnn	dncnn	VERB
easat-8918	186	6	restormer	restormer	PROPN
easat-8918	186	7	15	15	NUM
easat-8918	187	1	31.08	31.08	NUM
easat-8918	187	2	31.42	31.42	NUM
easat-8918	187	3	31.46	31.46	NUM
easat-8918	187	4	30.71	30.71	NUM
easat-8918	187	5	25	25	NUM
easat-8918	187	6	28.57	28.57	NUM
easat-8918	187	7	28.92	28.92	NUM
easat-8918	187	8	29.02	29.02	NUM
easat-8918	187	9	28.82	28.82	NUM
easat-8918	187	10	50	50	NUM
easat-8918	187	11	25.62	25.62	NUM
easat-8918	187	12	25.97	25.97	NUM
easat-8918	187	13	26.10	26.10	NUM
easat-8918	187	14	27.95	27.95	NUM
easat-8918	187	15	5	5	NUM
easat-8918	187	16	.	.	PUNCT
easat-8918	187	17	conclusions	conclusion	NOUN
easat-8918	187	18	this	this	DET
easat-8918	187	19	study	study	NOUN
easat-8918	187	20	thoroughly	thoroughly	ADV
easat-8918	187	21	investigated	investigate	VERB
easat-8918	187	22	image	image	NOUN
easat-8918	187	23	denoising	denoising	NOUN
easat-8918	187	24	performance	performance	NOUN
easat-8918	187	25	using	use	VERB
easat-8918	187	26	two	two	NUM
easat-8918	187	27	cutting	cutting	NOUN
easat-8918	187	28	-	-	PUNCT
easat-8918	187	29	edge	edge	NOUN
easat-8918	187	30	deep	deep	ADJ
easat-8918	187	31	learning	learning	NOUN
easat-8918	187	32	networks	network	NOUN
easat-8918	187	33	:	:	PUNCT
easat-8918	187	34	the	the	DET
easat-8918	187	35	restoration	restoration	NOUN
easat-8918	187	36	transformer	transformer	NOUN
easat-8918	187	37	(	(	PUNCT
easat-8918	187	38	restormer	restormer	PROPN
easat-8918	187	39	)	)	PUNCT
easat-8918	187	40	and	and	CCONJ
easat-8918	187	41	the	the	DET
easat-8918	187	42	deep	deep	ADJ
easat-8918	187	43	cnn	cnn	PROPN
easat-8918	187	44	(	(	PUNCT
easat-8918	187	45	dncnn	dncnn	PROPN
easat-8918	187	46	)	)	PUNCT
easat-8918	187	47	with	with	ADP
easat-8918	187	48	residual	residual	ADJ
easat-8918	187	49	learning	learning	NOUN
easat-8918	187	50	.	.	PUNCT
easat-8918	188	1	the	the	DET
easat-8918	188	2	denoising	denoising	NOUN
easat-8918	188	3	process	process	NOUN
easat-8918	188	4	was	be	AUX
easat-8918	188	5	rigorously	rigorously	ADV
easat-8918	188	6	applied	apply	VERB
easat-8918	188	7	to	to	ADP
easat-8918	188	8	images	image	NOUN
easat-8918	188	9	from	from	ADP
easat-8918	188	10	the	the	DET
easat-8918	188	11	bsd68	bsd68	PROPN
easat-8918	188	12	dataset	dataset	NOUN
easat-8918	188	13	and	and	CCONJ
easat-8918	188	14	the	the	DET
easat-8918	188	15	classic	classic	ADJ
easat-8918	188	16	cameraman	cameraman	ADJ
easat-8918	188	17	image	image	NOUN
easat-8918	188	18	,	,	PUNCT
easat-8918	188	19	both	both	PRON
easat-8918	188	20	corrupted	corrupt	VERB
easat-8918	188	21	with	with	ADP
easat-8918	188	22	various	various	ADJ
easat-8918	188	23	levels	level	NOUN
easat-8918	188	24	of	of	ADP
easat-8918	188	25	gaussian	gaussian	ADJ
easat-8918	188	26	noise	noise	NOUN
easat-8918	188	27	.	.	PUNCT
easat-8918	189	1	to	to	PART
easat-8918	189	2	quantitatively	quantitatively	ADV
easat-8918	189	3	assess	assess	VERB
easat-8918	189	4	denoising	denoising	NOUN
easat-8918	189	5	effectiveness	effectiveness	NOUN
easat-8918	189	6	,	,	PUNCT
easat-8918	189	7	peak	peak	NOUN
easat-8918	189	8	signal	signal	NOUN
easat-8918	189	9	-	-	PUNCT
easat-8918	189	10	to	to	ADP
easat-8918	189	11	-	-	PUNCT
easat-8918	189	12	noise	noise	NOUN
easat-8918	189	13	ratio	ratio	NOUN
easat-8918	189	14	(	(	PUNCT
easat-8918	189	15	psnr	psnr	NOUN
easat-8918	189	16	)	)	PUNCT
easat-8918	189	17	values	value	NOUN
easat-8918	189	18	were	be	AUX
easat-8918	189	19	calculated	calculate	VERB
easat-8918	189	20	.	.	PUNCT
easat-8918	190	1	these	these	DET
easat-8918	190	2	values	value	NOUN
easat-8918	190	3	serve	serve	VERB
easat-8918	190	4	as	as	ADP
easat-8918	190	5	a	a	DET
easat-8918	190	6	crucial	crucial	ADJ
easat-8918	190	7	metric	metric	NOUN
easat-8918	190	8	for	for	ADP
easat-8918	190	9	measuring	measure	VERB
easat-8918	190	10	the	the	DET
easat-8918	190	11	fidelity	fidelity	NOUN
easat-8918	190	12	between	between	ADP
easat-8918	190	13	the	the	DET
easat-8918	190	14	original	original	ADJ
easat-8918	190	15	and	and	CCONJ
easat-8918	190	16	denoised	denoise	VERB
easat-8918	190	17	images	image	NOUN
easat-8918	190	18	.	.	PUNCT
easat-8918	191	1	the	the	DET
easat-8918	191	2	denoising	denoise	VERB
easat-8918	191	3	capabilities	capability	NOUN
easat-8918	191	4	of	of	ADP
easat-8918	191	5	dncnn	dncnn	NOUN
easat-8918	191	6	and	and	CCONJ
easat-8918	191	7	restormer	restormer	PROPN
easat-8918	191	8	were	be	AUX
easat-8918	191	9	directly	directly	ADV
easat-8918	191	10	compared	compare	VERB
easat-8918	191	11	against	against	ADP
easat-8918	191	12	two	two	NUM
easat-8918	191	13	established	establish	VERB
easat-8918	191	14	algorithms	algorithm	NOUN
easat-8918	191	15	in	in	ADP
easat-8918	191	16	image	image	NOUN
easat-8918	191	17	denoising	denoise	VERB
easat-8918	191	18	literature	literature	NOUN
easat-8918	191	19	:	:	PUNCT
easat-8918	191	20	bm3d	bm3d	NOUN
easat-8918	191	21	and	and	CCONJ
easat-8918	191	22	tnrd	tnrd	NOUN
easat-8918	191	23	.	.	PUNCT
easat-8918	192	1	our	our	PRON
easat-8918	192	2	findings	finding	NOUN
easat-8918	192	3	revealed	reveal	VERB
easat-8918	192	4	that	that	SCONJ
easat-8918	192	5	dncnn	dncnn	PROPN
easat-8918	192	6	and	and	CCONJ
easat-8918	192	7	restormer	restormer	PROPN
easat-8918	192	8	consistently	consistently	ADV
easat-8918	192	9	achieved	achieve	VERB
easat-8918	192	10	the	the	DET
easat-8918	192	11	highest	high	ADJ
easat-8918	192	12	psnr	psnr	NOUN
easat-8918	192	13	results	result	NOUN
easat-8918	192	14	for	for	ADP
easat-8918	192	15	images	image	NOUN
easat-8918	192	16	subjected	subject	VERB
easat-8918	192	17	to	to	ADP
easat-8918	192	18	gaussian	gaussian	ADJ
easat-8918	192	19	noise	noise	NOUN
easat-8918	192	20	levels	level	NOUN
easat-8918	192	21	of	of	ADP
easat-8918	192	22	15	15	NUM
easat-8918	192	23	,	,	PUNCT
easat-8918	192	24	25	25	NUM
easat-8918	192	25	,	,	PUNCT
easat-8918	192	26	and	and	CCONJ
easat-8918	192	27	50	50	NUM
easat-8918	192	28	.	.	PUNCT
easat-8918	193	1	these	these	DET
easat-8918	193	2	results	result	NOUN
easat-8918	193	3	underscore	underscore	VERB
easat-8918	193	4	the	the	DET
easat-8918	193	5	significant	significant	ADJ
easat-8918	193	6	potential	potential	NOUN
easat-8918	193	7	of	of	ADP
easat-8918	193	8	both	both	DET
easat-8918	193	9	deep	deep	ADJ
easat-8918	193	10	learning	learning	NOUN
easat-8918	193	11	architectures	architecture	NOUN
easat-8918	193	12	,	,	PUNCT
easat-8918	193	13	particularly	particularly	ADV
easat-8918	193	14	highlighting	highlight	VERB
easat-8918	193	15	the	the	DET
easat-8918	193	16	superior	superior	ADJ
easat-8918	193	17	performance	performance	NOUN
easat-8918	193	18	of	of	ADP
easat-8918	193	19	transformer	transformer	ADJ
easat-8918	193	20	architectures	architecture	NOUN
easat-8918	193	21	like	like	ADP
easat-8918	193	22	restormer	restormer	NOUN
easat-8918	193	23	in	in	ADP
easat-8918	193	24	image	image	NOUN
easat-8918	193	25	restoration	restoration	NOUN
easat-8918	193	26	tasks	task	NOUN
easat-8918	193	27	.	.	PUNCT
easat-8918	194	1	their	their	PRON
easat-8918	194	2	capabilities	capability	NOUN
easat-8918	194	3	were	be	AUX
easat-8918	194	4	shown	show	VERB
easat-8918	194	5	to	to	PART
easat-8918	194	6	surpass	surpass	VERB
easat-8918	194	7	those	those	PRON
easat-8918	194	8	of	of	ADP
easat-8918	194	9	traditional	traditional	ADJ
easat-8918	194	10	methodologies	methodology	NOUN
easat-8918	194	11	in	in	ADP
easat-8918	194	12	terms	term	NOUN
easat-8918	194	13	of	of	ADP
easat-8918	194	14	denoising	denoise	VERB
easat-8918	194	15	quality	quality	NOUN
easat-8918	194	16	.	.	PUNCT
easat-8918	195	1	transparency	transparency	NOUN
easat-8918	195	2	:	:	PUNCT
easat-8918	195	3	the	the	DET
easat-8918	195	4	author	author	NOUN
easat-8918	195	5	confirms	confirm	VERB
easat-8918	195	6	that	that	SCONJ
easat-8918	195	7	the	the	DET
easat-8918	195	8	manuscript	manuscript	NOUN
easat-8918	195	9	is	be	AUX
easat-8918	195	10	an	an	DET
easat-8918	195	11	honest	honest	ADJ
easat-8918	195	12	,	,	PUNCT
easat-8918	195	13	accurate	accurate	ADJ
easat-8918	195	14	,	,	PUNCT
easat-8918	195	15	and	and	CCONJ
easat-8918	195	16	transparent	transparent	ADJ
easat-8918	195	17	account	account	NOUN
easat-8918	195	18	of	of	ADP
easat-8918	195	19	the	the	DET
easat-8918	195	20	study	study	NOUN
easat-8918	195	21	;	;	PUNCT
easat-8918	195	22	that	that	SCONJ
easat-8918	195	23	no	no	DET
easat-8918	195	24	vital	vital	ADJ
easat-8918	195	25	features	feature	NOUN
easat-8918	195	26	of	of	ADP
easat-8918	195	27	the	the	DET
easat-8918	195	28	study	study	NOUN
easat-8918	195	29	have	have	AUX
easat-8918	195	30	been	be	AUX
easat-8918	195	31	omitted	omit	VERB
easat-8918	195	32	;	;	PUNCT
easat-8918	195	33	and	and	CCONJ
easat-8918	195	34	that	that	SCONJ
easat-8918	195	35	any	any	DET
easat-8918	195	36	discrepancies	discrepancy	NOUN
easat-8918	195	37	from	from	ADP
easat-8918	195	38	the	the	DET
easat-8918	195	39	study	study	NOUN
easat-8918	195	40	as	as	SCONJ
easat-8918	195	41	planned	plan	VERB
easat-8918	195	42	have	have	AUX
easat-8918	195	43	been	be	AUX
easat-8918	195	44	explained	explain	VERB
easat-8918	195	45	.	.	PUNCT
easat-8918	196	1	this	this	DET
easat-8918	196	2	study	study	NOUN
easat-8918	196	3	followed	follow	VERB
easat-8918	196	4	all	all	DET
easat-8918	196	5	ethical	ethical	ADJ
easat-8918	196	6	practices	practice	NOUN
easat-8918	196	7	during	during	ADP
easat-8918	196	8	writing	writing	NOUN
easat-8918	196	9	.	.	PUNCT
easat-8918	197	1	copyright	copyright	NOUN
easat-8918	197	2	:	:	PUNCT
easat-8918	197	3	©	©	PROPN
easat-8918	197	4	2025	2025	NUM
easat-8918	197	5	by	by	ADP
easat-8918	197	6	the	the	DET
easat-8918	197	7	author	author	NOUN
easat-8918	197	8	.	.	PUNCT
easat-8918	198	1	this	this	DET
easat-8918	198	2	open	open	ADJ
easat-8918	198	3	-	-	PUNCT
easat-8918	198	4	access	access	NOUN
easat-8918	198	5	article	article	NOUN
easat-8918	198	6	is	be	AUX
easat-8918	198	7	distributed	distribute	VERB
easat-8918	198	8	under	under	ADP
easat-8918	198	9	the	the	DET
easat-8918	198	10	terms	term	NOUN
easat-8918	198	11	and	and	CCONJ
easat-8918	198	12	conditions	condition	NOUN
easat-8918	198	13	of	of	ADP
easat-8918	198	14	the	the	DET
easat-8918	198	15	creative	creative	ADJ
easat-8918	198	16	commons	common	NOUN
easat-8918	198	17	attribution	attribution	NOUN
easat-8918	198	18	(	(	PUNCT
easat-8918	198	19	cc	cc	NOUN
easat-8918	198	20	by	by	ADP
easat-8918	198	21	)	)	PUNCT
easat-8918	198	22	license	license	NOUN
easat-8918	198	23	(	(	PUNCT
easat-8918	198	24	https://creativecommons.org/licenses/by/4.0/	https://creativecommons.org/licenses/by/4.0/	PROPN
easat-8918	198	25	)	)	PUNCT
easat-8918	198	26	.	.	PUNCT
easat-8918	199	1	references	reference	NOUN
easat-8918	199	2	[	[	X
easat-8918	199	3	1	1	NUM
easat-8918	199	4	]	]	PUNCT
easat-8918	199	5	l.	l.	PROPN
easat-8918	199	6	fan	fan	PROPN
easat-8918	199	7	,	,	PUNCT
easat-8918	199	8	f.	f.	PROPN
easat-8918	199	9	zhang	zhang	PROPN
easat-8918	199	10	,	,	PUNCT
easat-8918	199	11	h.	h.	PROPN
easat-8918	199	12	fan	fan	PROPN
easat-8918	199	13	,	,	PUNCT
easat-8918	199	14	and	and	CCONJ
easat-8918	199	15	c.	c.	PROPN
easat-8918	199	16	zhang	zhang	PROPN
easat-8918	199	17	,	,	PUNCT
easat-8918	199	18	"	"	PUNCT
easat-8918	199	19	brief	brief	ADJ
easat-8918	199	20	review	review	NOUN
easat-8918	199	21	of	of	ADP
easat-8918	199	22	image	image	NOUN
easat-8918	199	23	denoising	denoising	NOUN
easat-8918	199	24	techniques	technique	NOUN
easat-8918	199	25	,	,	PUNCT
easat-8918	199	26	"	"	PUNCT
easat-8918	199	27	visual	visual	ADJ
easat-8918	199	28	computing	computing	NOUN
easat-8918	199	29	for	for	ADP
easat-8918	199	30	industry	industry	NOUN
easat-8918	199	31	,	,	PUNCT
easat-8918	199	32	biomedicine	biomedicine	NOUN
easat-8918	199	33	and	and	CCONJ
easat-8918	199	34	art	art	NOUN
easat-8918	199	35	,	,	PUNCT
easat-8918	199	36	vol	vol	NOUN
easat-8918	199	37	.	.	PROPN
easat-8918	199	38	2	2	NUM
easat-8918	199	39	,	,	PUNCT
easat-8918	199	40	no	no	INTJ
easat-8918	199	41	.	.	NOUN
easat-8918	199	42	1	1	NUM
easat-8918	199	43	,	,	PUNCT
easat-8918	199	44	p.	p.	NOUN
easat-8918	199	45	7	7	NUM
easat-8918	199	46	,	,	PUNCT
easat-8918	199	47	2019	2019	NUM
easat-8918	199	48	.	.	PUNCT
easat-8918	200	1	https://doi.org/10.1186/s42492-019-0016-7	https://doi.org/10.1186/s42492-019-0016-7	PROPN
easat-8918	201	1	[	[	X
easat-8918	201	2	2	2	NUM
easat-8918	201	3	]	]	X
easat-8918	201	4	g.	g.	PROPN
easat-8918	201	5	eleyan	eleyan	PROPN
easat-8918	201	6	and	and	CCONJ
easat-8918	201	7	m.	m.	PROPN
easat-8918	201	8	s.	s.	PROPN
easat-8918	201	9	salman	salman	PROPN
easat-8918	201	10	,	,	PUNCT
easat-8918	201	11	"	"	PUNCT
easat-8918	201	12	convergence	convergence	NOUN
easat-8918	201	13	analysis	analysis	NOUN
easat-8918	201	14	of	of	ADP
easat-8918	201	15	the	the	DET
easat-8918	201	16	mixed	mixed	ADJ
easat-8918	201	17	-	-	PUNCT
easat-8918	201	18	norm	norm	NOUN
easat-8918	201	19	lms	lm	NOUN
easat-8918	201	20	and	and	CCONJ
easat-8918	201	21	two	two	NUM
easat-8918	201	22	versions	version	NOUN
easat-8918	201	23	for	for	ADP
easat-8918	201	24	sparse	sparse	ADJ
easat-8918	201	25	system	system	NOUN
easat-8918	201	26	identification	identification	NOUN
easat-8918	201	27	,	,	PUNCT
easat-8918	201	28	"	"	PUNCT
easat-8918	201	29	signal	signal	NOUN
easat-8918	201	30	,	,	PUNCT
easat-8918	201	31	image	image	NOUN
easat-8918	201	32	and	and	CCONJ
easat-8918	201	33	video	video	NOUN
easat-8918	201	34	processing	processing	NOUN
easat-8918	201	35	,	,	PUNCT
easat-8918	201	36	vol	vol	NOUN
easat-8918	201	37	.	.	PROPN
easat-8918	201	38	14	14	NUM
easat-8918	201	39	,	,	PUNCT
easat-8918	201	40	no	no	INTJ
easat-8918	201	41	.	.	NOUN
easat-8918	201	42	5	5	NUM
easat-8918	201	43	,	,	PUNCT
easat-8918	201	44	pp	pp	ADJ
easat-8918	201	45	.	.	PUNCT
easat-8918	202	1	965	965	NUM
easat-8918	202	2	-	-	SYM
easat-8918	202	3	970	970	NUM
easat-8918	202	4	,	,	PUNCT
easat-8918	202	5	2020	2020	NUM
easat-8918	202	6	.	.	PUNCT
easat-8918	203	1	https://doi.org/10.1007/s11760019-01628-9	https://doi.org/10.1007/s11760019-01628-9	PROPN
easat-8918	203	2	https://creativecommons.org/licenses/by/4.0/	https://creativecommons.org/licenses/by/4.0/	PROPN
easat-8918	203	3	https://doi.org/10.1186/s42492-019-0016-7	https://doi.org/10.1186/s42492-019-0016-7	PROPN
easat-8918	203	4	https://doi.org/10.1007/s11760-019-01628-9	https://doi.org/10.1007/s11760-019-01628-9	NUM
easat-8918	203	5	https://doi.org/10.1007/s11760-019-01628-9	https://doi.org/10.1007/s11760-019-01628-9	NUM
easat-8918	203	6	1357	1357	NUM
easat-8918	203	7	edelweiss	edelweiss	PROPN
easat-8918	203	8	applied	apply	VERB
easat-8918	203	9	science	science	NOUN
easat-8918	203	10	and	and	CCONJ
easat-8918	203	11	technology	technology	NOUN
easat-8918	203	12	issn	issn	PROPN
easat-8918	203	13	:	:	PUNCT
easat-8918	203	14	2576	2576	NUM
easat-8918	203	15	-	-	SYM
easat-8918	203	16	8484	8484	NUM
easat-8918	203	17	vol	vol	NOUN
easat-8918	203	18	.	.	PROPN
easat-8918	204	1	9	9	NUM
easat-8918	204	2	,	,	PUNCT
easat-8918	204	3	no	no	INTJ
easat-8918	204	4	.	.	NOUN
easat-8918	204	5	7	7	NUM
easat-8918	204	6	:	:	SYM
easat-8918	204	7	1345	1345	NUM
easat-8918	204	8	-	-	SYM
easat-8918	204	9	1359	1359	NUM
easat-8918	204	10	,	,	PUNCT
easat-8918	204	11	2025	2025	NUM
easat-8918	204	12	doi	doi	NOUN
easat-8918	204	13	:	:	PUNCT
easat-8918	204	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	204	15	©	©	PROPN
easat-8918	204	16	2025	2025	NUM
easat-8918	204	17	by	by	ADP
easat-8918	204	18	the	the	DET
easat-8918	204	19	author	author	NOUN
easat-8918	204	20	;	;	PUNCT
easat-8918	204	21	licensee	licensee	PROPN
easat-8918	204	22	learning	learning	NOUN
easat-8918	204	23	gate	gate	NOUN
easat-8918	205	1	[	[	X
easat-8918	205	2	3	3	NUM
easat-8918	205	3	]	]	X
easat-8918	205	4	g.	g.	PROPN
easat-8918	205	5	eleyan	eleyan	PROPN
easat-8918	205	6	,	,	PUNCT
easat-8918	205	7	m.	m.	PROPN
easat-8918	205	8	s.	s.	PROPN
easat-8918	205	9	salman	salman	PROPN
easat-8918	205	10	,	,	PUNCT
easat-8918	205	11	and	and	CCONJ
easat-8918	205	12	c.	c.	PROPN
easat-8918	205	13	turan	turan	PROPN
easat-8918	205	14	,	,	PUNCT
easat-8918	205	15	"	"	PUNCT
easat-8918	205	16	two	two	NUM
easat-8918	205	17	-	-	PUNCT
easat-8918	205	18	dimensional	dimensional	ADJ
easat-8918	205	19	sparse	sparse	ADJ
easat-8918	205	20	lms	lm	NOUN
easat-8918	205	21	for	for	ADP
easat-8918	205	22	image	image	NOUN
easat-8918	205	23	denoising	denoising	NOUN
easat-8918	205	24	,	,	PUNCT
easat-8918	205	25	"	"	PUNCT
easat-8918	205	26	in	in	ADP
easat-8918	205	27	twelfth	twelfth	ADJ
easat-8918	205	28	international	international	ADJ
easat-8918	205	29	conference	conference	NOUN
easat-8918	205	30	on	on	ADP
easat-8918	205	31	electronics	electronic	NOUN
easat-8918	205	32	,	,	PUNCT
easat-8918	205	33	computer	computer	NOUN
easat-8918	205	34	and	and	CCONJ
easat-8918	205	35	computation	computation	NOUN
easat-8918	205	36	,	,	PUNCT
easat-8918	205	37	2015	2015	NUM
easat-8918	205	38	.	.	PUNCT
easat-8918	206	1	https://doi.org/10.1109/icecco.2015.7416909	https://doi.org/10.1109/icecco.2015.7416909	PROPN
easat-8918	207	1	[	[	X
easat-8918	207	2	4	4	NUM
easat-8918	207	3	]	]	PUNCT
easat-8918	207	4	m.	m.	NOUN
easat-8918	207	5	s.	s.	PROPN
easat-8918	207	6	salman	salman	PROPN
easat-8918	207	7	and	and	CCONJ
easat-8918	207	8	a.	a.	NOUN
easat-8918	207	9	eleyan	eleyan	PROPN
easat-8918	207	10	,	,	PUNCT
easat-8918	207	11	"	"	PUNCT
easat-8918	207	12	an	an	DET
easat-8918	207	13	efficient	efficient	ADJ
easat-8918	207	14	2	2	NUM
easat-8918	207	15	-	-	PUNCT
easat-8918	207	16	d	d	NOUN
easat-8918	207	17	recursive	recursive	ADJ
easat-8918	207	18	inverse	inverse	NOUN
easat-8918	207	19	algorithm	algorithm	NOUN
easat-8918	207	20	for	for	ADP
easat-8918	207	21	image	image	NOUN
easat-8918	207	22	de	de	X
easat-8918	207	23	-	-	ADJ
easat-8918	207	24	noising	noising	ADJ
easat-8918	207	25	,	,	PUNCT
easat-8918	207	26	"	"	PUNCT
easat-8918	207	27	in	in	ADP
easat-8918	207	28	paper	paper	NOUN
easat-8918	207	29	presented	present	VERB
easat-8918	207	30	at	at	ADP
easat-8918	207	31	the	the	DET
easat-8918	207	32	image	image	NOUN
easat-8918	207	33	processing	processing	NOUN
easat-8918	207	34	&	&	CCONJ
easat-8918	207	35	communications	communication	NOUN
easat-8918	207	36	challenges	challenge	NOUN
easat-8918	207	37	conference	conference	NOUN
easat-8918	207	38	,	,	PUNCT
easat-8918	207	39	2013	2013	NUM
easat-8918	207	40	.	.	PUNCT
easat-8918	208	1	[	[	X
easat-8918	208	2	5	5	X
easat-8918	208	3	]	]	X
easat-8918	208	4	g.	g.	PROPN
easat-8918	208	5	eleyan	eleyan	PROPN
easat-8918	208	6	and	and	CCONJ
easat-8918	208	7	m.	m.	PROPN
easat-8918	208	8	salman	salman	PROPN
easat-8918	208	9	,	,	PUNCT
easat-8918	208	10	"	"	PUNCT
easat-8918	208	11	image	image	NOUN
easat-8918	208	12	denoising	denoise	VERB
easat-8918	208	13	with	with	ADP
easat-8918	208	14	two	two	NUM
easat-8918	208	15	-	-	PUNCT
easat-8918	208	16	dimensional	dimensional	ADJ
easat-8918	208	17	zero	zero	NUM
easat-8918	208	18	attracting	attract	VERB
easat-8918	208	19	lms	lm	NOUN
easat-8918	208	20	algorithm	algorithm	NOUN
easat-8918	208	21	,	,	PUNCT
easat-8918	208	22	"	"	PUNCT
easat-8918	208	23	journal	journal	NOUN
easat-8918	208	24	of	of	ADP
easat-8918	208	25	engineering	engineering	NOUN
easat-8918	208	26	sciences	science	NOUN
easat-8918	208	27	,	,	PUNCT
easat-8918	208	28	vol	vol	NOUN
easat-8918	208	29	.	.	PROPN
easat-8918	208	30	25	25	NUM
easat-8918	208	31	,	,	PUNCT
easat-8918	208	32	no	no	INTJ
easat-8918	208	33	.	.	NOUN
easat-8918	208	34	5	5	NUM
easat-8918	208	35	,	,	PUNCT
easat-8918	208	36	pp	pp	ADJ
easat-8918	208	37	.	.	PUNCT
easat-8918	209	1	539	539	NUM
easat-8918	209	2	-	-	SYM
easat-8918	209	3	545	545	NUM
easat-8918	209	4	,	,	PUNCT
easat-8918	209	5	2019	2019	NUM
easat-8918	209	6	.	.	PUNCT
easat-8918	210	1	https://doi.org/10.5505/pajes.2018.06982	https://doi.org/10.5505/pajes.2018.06982	PROPN
easat-8918	211	1	[	[	X
easat-8918	211	2	6	6	NUM
easat-8918	211	3	]	]	PUNCT
easat-8918	211	4	v.	v.	ADP
easat-8918	211	5	atamoradov	atamoradov	PROPN
easat-8918	211	6	,	,	PUNCT
easat-8918	211	7	a.	a.	NOUN
easat-8918	211	8	eleyan	eleyan	NOUN
easat-8918	211	9	,	,	PUNCT
easat-8918	211	10	and	and	CCONJ
easat-8918	211	11	b.	b.	PROPN
easat-8918	211	12	kralik	kralik	PROPN
easat-8918	211	13	,	,	PUNCT
easat-8918	211	14	"	"	PUNCT
easat-8918	211	15	performance	performance	NOUN
easat-8918	211	16	evaluation	evaluation	NOUN
easat-8918	211	17	for	for	ADP
easat-8918	211	18	face	face	NOUN
easat-8918	211	19	recognition	recognition	NOUN
easat-8918	211	20	using	use	VERB
easat-8918	211	21	wavelet	wavelet	NOUN
easat-8918	211	22	-	-	PUNCT
easat-8918	211	23	based	base	VERB
easat-8918	211	24	image	image	NOUN
easat-8918	211	25	denoising	denoising	NOUN
easat-8918	211	26	,	,	PUNCT
easat-8918	211	27	"	"	PUNCT
easat-8918	211	28	presented	present	VERB
easat-8918	211	29	at	at	ADP
easat-8918	211	30	the	the	DET
easat-8918	211	31	international	international	ADJ
easat-8918	211	32	conference	conference	NOUN
easat-8918	211	33	on	on	ADP
easat-8918	211	34	technological	technological	ADJ
easat-8918	211	35	advances	advance	NOUN
easat-8918	211	36	in	in	ADP
easat-8918	211	37	electrical	electrical	ADJ
easat-8918	211	38	,	,	PUNCT
easat-8918	211	39	electronics	electronic	NOUN
easat-8918	211	40	and	and	CCONJ
easat-8918	211	41	computer	computer	NOUN
easat-8918	211	42	engineering	engineering	NOUN
easat-8918	211	43	,	,	PUNCT
easat-8918	211	44	2013	2013	NUM
easat-8918	211	45	.	.	PUNCT
easat-8918	212	1	[	[	X
easat-8918	212	2	7	7	NUM
easat-8918	212	3	]	]	PUNCT
easat-8918	212	4	a.	a.	NOUN
easat-8918	212	5	m.	m.	NOUN
easat-8918	212	6	ashir	ashir	NOUN
easat-8918	212	7	and	and	CCONJ
easat-8918	212	8	a.	a.	NOUN
easat-8918	212	9	eleyan	eleyan	NOUN
easat-8918	212	10	,	,	PUNCT
easat-8918	212	11	"	"	PUNCT
easat-8918	212	12	a	a	DET
easat-8918	212	13	multi	multi	ADJ
easat-8918	212	14	-	-	ADJ
easat-8918	212	15	resolution	resolution	ADJ
easat-8918	212	16	approach	approach	NOUN
easat-8918	212	17	for	for	ADP
easat-8918	212	18	edge	edge	NOUN
easat-8918	212	19	detection	detection	NOUN
easat-8918	212	20	using	use	VERB
easat-8918	212	21	ant	ant	ADJ
easat-8918	212	22	colony	colony	NOUN
easat-8918	212	23	optimization	optimization	NOUN
easat-8918	212	24	,	,	PUNCT
easat-8918	212	25	"	"	PUNCT
easat-8918	212	26	in	in	ADP
easat-8918	212	27	23rd	23rd	ADJ
easat-8918	212	28	signal	signal	NOUN
easat-8918	212	29	processing	processing	NOUN
easat-8918	212	30	and	and	CCONJ
easat-8918	212	31	communications	communication	NOUN
easat-8918	212	32	applications	application	NOUN
easat-8918	212	33	conference	conference	NOUN
easat-8918	212	34	,	,	PUNCT
easat-8918	212	35	2015	2015	NUM
easat-8918	212	36	.	.	PUNCT
easat-8918	213	1	[	[	X
easat-8918	213	2	8	8	NUM
easat-8918	213	3	]	]	X
easat-8918	213	4	a.	a.	NOUN
easat-8918	213	5	muhammad	muhammad	PROPN
easat-8918	213	6	,	,	PUNCT
easat-8918	213	7	i.	i.	PROPN
easat-8918	213	8	bala	bala	PROPN
easat-8918	213	9	,	,	PUNCT
easat-8918	213	10	m.	m.	PROPN
easat-8918	213	11	s.	s.	PROPN
easat-8918	213	12	salman	salman	PROPN
easat-8918	213	13	,	,	PUNCT
easat-8918	213	14	and	and	CCONJ
easat-8918	213	15	a.	a.	NOUN
easat-8918	213	16	eleyan	eleyan	NOUN
easat-8918	213	17	,	,	PUNCT
easat-8918	213	18	"	"	PUNCT
easat-8918	213	19	dwt	dwt	NOUN
easat-8918	213	20	subbands	subband	NOUN
easat-8918	213	21	fusion	fusion	NOUN
easat-8918	213	22	using	use	VERB
easat-8918	213	23	ant	ant	ADJ
easat-8918	213	24	colony	colony	NOUN
easat-8918	213	25	optimization	optimization	NOUN
easat-8918	213	26	for	for	ADP
easat-8918	213	27	edge	edge	NOUN
easat-8918	213	28	detection	detection	NOUN
easat-8918	213	29	,	,	PUNCT
easat-8918	213	30	"	"	PUNCT
easat-8918	213	31	in	in	ADP
easat-8918	213	32	22nd	22nd	NOUN
easat-8918	213	33	signal	signal	NOUN
easat-8918	213	34	processing	processing	NOUN
easat-8918	213	35	and	and	CCONJ
easat-8918	213	36	communications	communication	NOUN
easat-8918	213	37	applications	application	NOUN
easat-8918	213	38	conference	conference	NOUN
easat-8918	213	39	,	,	PUNCT
easat-8918	213	40	2014	2014	NUM
easat-8918	213	41	.	.	PUNCT
easat-8918	214	1	[	[	X
easat-8918	214	2	9	9	NUM
easat-8918	214	3	]	]	PUNCT
easat-8918	214	4	j.	j.	PROPN
easat-8918	214	5	zhang	zhang	PROPN
easat-8918	214	6	and	and	CCONJ
easat-8918	214	7	b.	b.	PROPN
easat-8918	214	8	ghanem	ghanem	PROPN
easat-8918	214	9	,	,	PUNCT
easat-8918	214	10	"	"	PUNCT
easat-8918	214	11	ista	ista	NOUN
easat-8918	214	12	-	-	PUNCT
easat-8918	214	13	net	net	NOUN
easat-8918	214	14	:	:	PUNCT
easat-8918	214	15	interpretable	interpretable	ADJ
easat-8918	214	16	optimization	optimization	NOUN
easat-8918	214	17	-	-	PUNCT
easat-8918	214	18	inspired	inspire	VERB
easat-8918	214	19	deep	deep	ADJ
easat-8918	214	20	network	network	NOUN
easat-8918	214	21	for	for	ADP
easat-8918	214	22	image	image	NOUN
easat-8918	214	23	compressive	compressive	ADJ
easat-8918	214	24	sensing	sensing	NOUN
easat-8918	214	25	,	,	PUNCT
easat-8918	214	26	"	"	PUNCT
easat-8918	214	27	in	in	ADP
easat-8918	214	28	proceedings	proceeding	NOUN
easat-8918	214	29	of	of	ADP
easat-8918	214	30	the	the	DET
easat-8918	214	31	ieee	ieee	NOUN
easat-8918	214	32	conference	conference	NOUN
easat-8918	214	33	on	on	ADP
easat-8918	214	34	computer	computer	NOUN
easat-8918	214	35	vision	vision	NOUN
easat-8918	214	36	and	and	CCONJ
easat-8918	214	37	pattern	pattern	NOUN
easat-8918	214	38	recognition	recognition	NOUN
easat-8918	214	39	(	(	PUNCT
easat-8918	214	40	cvpr	cvpr	NOUN
easat-8918	214	41	)	)	PUNCT
easat-8918	214	42	,	,	PUNCT
easat-8918	214	43	2018	2018	NUM
easat-8918	214	44	.	.	PUNCT
easat-8918	215	1	[	[	X
easat-8918	215	2	10	10	NUM
easat-8918	215	3	]	]	PUNCT
easat-8918	215	4	m.	m.	NOUN
easat-8918	215	5	alkandari	alkandari	PROPN
easat-8918	215	6	,	,	PUNCT
easat-8918	215	7	f.	f.	PROPN
easat-8918	215	8	almanaye	almanaye	PROPN
easat-8918	215	9	,	,	PUNCT
easat-8918	215	10	m.	m.	NOUN
easat-8918	215	11	alsoori	alsoori	PROPN
easat-8918	215	12	,	,	PUNCT
easat-8918	215	13	h.	h.	PROPN
easat-8918	215	14	farfoura	farfoura	NOUN
easat-8918	215	15	,	,	PUNCT
easat-8918	215	16	a.	a.	NOUN
easat-8918	215	17	alshammari	alshammari	NOUN
easat-8918	215	18	,	,	PUNCT
easat-8918	215	19	and	and	CCONJ
easat-8918	215	20	a.	a.	NOUN
easat-8918	215	21	eleyan	eleyan	NOUN
easat-8918	215	22	,	,	PUNCT
easat-8918	215	23	"	"	PUNCT
easat-8918	215	24	kuwaiti	kuwaiti	PROPN
easat-8918	215	25	-	-	PUNCT
easat-8918	215	26	arabic	arabic	ADJ
easat-8918	215	27	sentiment	sentiment	NOUN
easat-8918	215	28	analysis	analysis	NOUN
easat-8918	215	29	:	:	PUNCT
easat-8918	215	30	a	a	DET
easat-8918	215	31	comparative	comparative	ADJ
easat-8918	215	32	study	study	NOUN
easat-8918	215	33	,	,	PUNCT
easat-8918	215	34	"	"	PUNCT
easat-8918	215	35	in	in	ADP
easat-8918	215	36	6th	6th	ADJ
easat-8918	215	37	international	international	ADJ
easat-8918	215	38	conference	conference	NOUN
easat-8918	215	39	on	on	ADP
easat-8918	215	40	bio	bio	NOUN
easat-8918	215	41	-	-	NOUN
easat-8918	215	42	engineering	engineering	NOUN
easat-8918	215	43	for	for	ADP
easat-8918	215	44	smart	smart	ADJ
easat-8918	215	45	technologies	technology	NOUN
easat-8918	215	46	,	,	PUNCT
easat-8918	215	47	2025	2025	NUM
easat-8918	215	48	.	.	PUNCT
easat-8918	216	1	[	[	X
easat-8918	216	2	11	11	NUM
easat-8918	216	3	]	]	X
easat-8918	216	4	b.	b.	PROPN
easat-8918	216	5	alwazzan	alwazzan	PROPN
easat-8918	216	6	et	et	PROPN
easat-8918	216	7	al	al	PROPN
easat-8918	216	8	.	.	PROPN
easat-8918	216	9	,	,	PUNCT
easat-8918	216	10	"	"	PUNCT
easat-8918	216	11	brain	brain	NOUN
easat-8918	216	12	activity	activity	NOUN
easat-8918	216	13	monitoring	monitoring	NOUN
easat-8918	216	14	using	use	VERB
easat-8918	216	15	machine	machine	NOUN
easat-8918	216	16	learning	learning	NOUN
easat-8918	216	17	and	and	CCONJ
easat-8918	216	18	dry	dry	ADJ
easat-8918	216	19	-	-	PUNCT
easat-8918	216	20	eeg	eeg	NOUN
easat-8918	216	21	signals	signal	NOUN
easat-8918	216	22	,	,	PUNCT
easat-8918	216	23	"	"	PUNCT
easat-8918	216	24	in	in	ADP
easat-8918	216	25	second	second	ADJ
easat-8918	216	26	jordanian	jordanian	ADJ
easat-8918	216	27	international	international	ADJ
easat-8918	216	28	biomedical	biomedical	ADJ
easat-8918	216	29	engineering	engineering	NOUN
easat-8918	216	30	conference	conference	NOUN
easat-8918	216	31	,	,	PUNCT
easat-8918	216	32	2024	2024	NUM
easat-8918	216	33	.	.	PUNCT
easat-8918	217	1	[	[	X
easat-8918	217	2	12	12	NUM
easat-8918	217	3	]	]	PUNCT
easat-8918	217	4	m.	m.	NOUN
easat-8918	217	5	alkandari	alkandari	PROPN
easat-8918	217	6	,	,	PUNCT
easat-8918	217	7	f.	f.	PROPN
easat-8918	217	8	almanaye	almanaye	PROPN
easat-8918	217	9	,	,	PUNCT
easat-8918	217	10	s.	s.	PROPN
easat-8918	217	11	alhashan	alhashan	PROPN
easat-8918	217	12	,	,	PUNCT
easat-8918	217	13	d.	d.	PROPN
easat-8918	217	14	alenezi	alenezi	PROPN
easat-8918	217	15	,	,	PUNCT
easat-8918	217	16	and	and	CCONJ
easat-8918	217	17	a.	a.	NOUN
easat-8918	217	18	eleyan	eleyan	NOUN
easat-8918	217	19	,	,	PUNCT
easat-8918	217	20	"	"	PUNCT
easat-8918	217	21	deep	deep	ADJ
easat-8918	217	22	learning	learning	NOUN
easat-8918	217	23	-	-	PUNCT
easat-8918	217	24	based	base	VERB
easat-8918	217	25	eeg	eeg	NOUN
easat-8918	217	26	signals	signal	NOUN
easat-8918	217	27	interpretation	interpretation	NOUN
easat-8918	217	28	for	for	ADP
easat-8918	217	29	motor	motor	NOUN
easat-8918	217	30	imagery	imagery	NOUN
easat-8918	217	31	,	,	PUNCT
easat-8918	217	32	"	"	PUNCT
easat-8918	217	33	in	in	ADP
easat-8918	217	34	international	international	ADJ
easat-8918	217	35	conference	conference	NOUN
easat-8918	217	36	on	on	ADP
easat-8918	217	37	bio	bio	NOUN
easat-8918	217	38	-	-	NOUN
easat-8918	217	39	engineering	engineering	NOUN
easat-8918	217	40	for	for	ADP
easat-8918	217	41	smart	smart	ADJ
easat-8918	217	42	technologies	technology	NOUN
easat-8918	217	43	,	,	PUNCT
easat-8918	217	44	2025	2025	NUM
easat-8918	217	45	.	.	PUNCT
easat-8918	218	1	[	[	X
easat-8918	218	2	13	13	NUM
easat-8918	218	3	]	]	PUNCT
easat-8918	218	4	a.	a.	NOUN
easat-8918	218	5	eleyan	eleyan	PROPN
easat-8918	218	6	,	,	PUNCT
easat-8918	218	7	f.	f.	PROPN
easat-8918	218	8	bayram	bayram	PROPN
easat-8918	218	9	,	,	PUNCT
easat-8918	218	10	and	and	CCONJ
easat-8918	218	11	g.	g.	PROPN
easat-8918	218	12	eleyan	eleyan	PROPN
easat-8918	218	13	,	,	PUNCT
easat-8918	218	14	"	"	PUNCT
easat-8918	218	15	spectrogram	spectrogram	NOUN
easat-8918	218	16	based	base	VERB
easat-8918	218	17	arrhythmia	arrhythmia	NOUN
easat-8918	218	18	classification	classification	NOUN
easat-8918	218	19	using	use	VERB
easat-8918	218	20	three	three	NUM
easat-8918	218	21	-	-	PUNCT
easat-8918	218	22	channel	channel	NOUN
easat-8918	218	23	deep	deep	ADJ
easat-8918	218	24	learning	learning	NOUN
easat-8918	218	25	model	model	NOUN
easat-8918	218	26	with	with	ADP
easat-8918	218	27	feature	feature	NOUN
easat-8918	218	28	fusion	fusion	NOUN
easat-8918	218	29	,	,	PUNCT
easat-8918	218	30	"	"	PUNCT
easat-8918	218	31	applied	apply	VERB
easat-8918	218	32	sciences	science	NOUN
easat-8918	218	33	,	,	PUNCT
easat-8918	218	34	vol	vol	NOUN
easat-8918	218	35	.	.	PROPN
easat-8918	218	36	14	14	NUM
easat-8918	218	37	,	,	PUNCT
easat-8918	218	38	no	no	INTJ
easat-8918	218	39	.	.	NOUN
easat-8918	218	40	21	21	NUM
easat-8918	218	41	,	,	PUNCT
easat-8918	218	42	p.	p.	NOUN
easat-8918	218	43	9936	9936	NUM
easat-8918	218	44	,	,	PUNCT
easat-8918	218	45	2024	2024	NUM
easat-8918	218	46	.	.	PUNCT
easat-8918	219	1	https://doi.org/10.3390/app14219936	https://doi.org/10.3390/app14219936	PROPN
easat-8918	220	1	[	[	X
easat-8918	220	2	14	14	NUM
easat-8918	220	3	]	]	PUNCT
easat-8918	220	4	z.	z.	PROPN
easat-8918	220	5	lu	lu	PROPN
easat-8918	220	6	et	et	PROPN
easat-8918	220	7	al	al	PROPN
easat-8918	220	8	.	.	PROPN
easat-8918	220	9	,	,	PUNCT
easat-8918	220	10	"	"	PUNCT
easat-8918	220	11	fast	fast	ADJ
easat-8918	220	12	single	single	ADJ
easat-8918	220	13	image	image	NOUN
easat-8918	220	14	super	super	NOUN
easat-8918	220	15	-	-	NOUN
easat-8918	220	16	resolution	resolution	NOUN
easat-8918	220	17	via	via	ADP
easat-8918	220	18	dilated	dilate	VERB
easat-8918	220	19	residual	residual	ADJ
easat-8918	220	20	networks	network	NOUN
easat-8918	220	21	,	,	PUNCT
easat-8918	220	22	"	"	PUNCT
easat-8918	220	23	institute	institute	NOUN
easat-8918	220	24	of	of	ADP
easat-8918	220	25	electrical	electrical	ADJ
easat-8918	220	26	and	and	CCONJ
easat-8918	220	27	electronics	electronic	NOUN
easat-8918	220	28	engineers	engineer	NOUN
easat-8918	220	29	.	.	PUNCT
easat-8918	221	1	access	access	NOUN
easat-8918	221	2	,	,	PUNCT
easat-8918	221	3	vol	vol	NOUN
easat-8918	221	4	.	.	PROPN
easat-8918	221	5	7	7	NUM
easat-8918	221	6	,	,	PUNCT
easat-8918	221	7	pp	pp	ADJ
easat-8918	221	8	.	.	PUNCT
easat-8918	222	1	109729	109729	NUM
easat-8918	222	2	-	-	SYM
easat-8918	222	3	109738	109738	NUM
easat-8918	222	4	,	,	PUNCT
easat-8918	222	5	2018	2018	NUM
easat-8918	222	6	.	.	PUNCT
easat-8918	223	1	[	[	X
easat-8918	223	2	15	15	NUM
easat-8918	223	3	]	]	X
easat-8918	223	4	y.	y.	NOUN
easat-8918	223	5	lecun	lecun	PROPN
easat-8918	223	6	,	,	PUNCT
easat-8918	223	7	l.	l.	PROPN
easat-8918	223	8	bottou	bottou	PROPN
easat-8918	223	9	,	,	PUNCT
easat-8918	223	10	y.	y.	PROPN
easat-8918	223	11	bengio	bengio	PROPN
easat-8918	223	12	,	,	PUNCT
easat-8918	223	13	and	and	CCONJ
easat-8918	223	14	p.	p.	PROPN
easat-8918	223	15	haffner	haffner	NOUN
easat-8918	223	16	,	,	PUNCT
easat-8918	223	17	"	"	PUNCT
easat-8918	223	18	gradient	gradient	NOUN
easat-8918	223	19	-	-	PUNCT
easat-8918	223	20	based	base	VERB
easat-8918	223	21	learning	learning	NOUN
easat-8918	223	22	applied	apply	VERB
easat-8918	223	23	to	to	ADP
easat-8918	223	24	document	document	NOUN
easat-8918	223	25	recognition	recognition	NOUN
easat-8918	223	26	,	,	PUNCT
easat-8918	223	27	"	"	PUNCT
easat-8918	223	28	proceedings	proceeding	NOUN
easat-8918	223	29	of	of	ADP
easat-8918	223	30	the	the	DET
easat-8918	223	31	institute	institute	NOUN
easat-8918	223	32	of	of	ADP
easat-8918	223	33	electrical	electrical	ADJ
easat-8918	223	34	and	and	CCONJ
easat-8918	223	35	electronics	electronic	NOUN
easat-8918	223	36	engineers	engineer	NOUN
easat-8918	223	37	,	,	PUNCT
easat-8918	223	38	vol	vol	NOUN
easat-8918	223	39	.	.	PROPN
easat-8918	223	40	86	86	NUM
easat-8918	223	41	,	,	PUNCT
easat-8918	223	42	no	no	INTJ
easat-8918	223	43	.	.	NOUN
easat-8918	223	44	11	11	NUM
easat-8918	223	45	,	,	PUNCT
easat-8918	223	46	pp	pp	ADJ
easat-8918	223	47	.	.	PUNCT
easat-8918	223	48	2278	2278	NUM
easat-8918	223	49	-	-	SYM
easat-8918	223	50	2324	2324	NUM
easat-8918	223	51	,	,	PUNCT
easat-8918	223	52	2002	2002	NUM
easat-8918	223	53	.	.	PUNCT
easat-8918	224	1	[	[	X
easat-8918	224	2	16	16	NUM
easat-8918	224	3	]	]	PUNCT
easat-8918	224	4	a.	a.	NOUN
easat-8918	224	5	krizhevsky	krizhevsky	PROPN
easat-8918	224	6	,	,	PUNCT
easat-8918	224	7	i.	i.	PROPN
easat-8918	224	8	sutskever	sutskever	PROPN
easat-8918	224	9	,	,	PUNCT
easat-8918	224	10	and	and	CCONJ
easat-8918	224	11	g.	g.	PROPN
easat-8918	224	12	e.	e.	PROPN
easat-8918	224	13	hinton	hinton	PROPN
easat-8918	224	14	,	,	PUNCT
easat-8918	224	15	"	"	PUNCT
easat-8918	224	16	imagenet	imagenet	ADJ
easat-8918	224	17	classification	classification	NOUN
easat-8918	224	18	with	with	ADP
easat-8918	224	19	deep	deep	ADJ
easat-8918	224	20	convolutional	convolutional	ADJ
easat-8918	224	21	neural	neural	ADJ
easat-8918	224	22	networks	network	NOUN
easat-8918	224	23	,	,	PUNCT
easat-8918	224	24	"	"	PUNCT
easat-8918	224	25	advances	advance	NOUN
easat-8918	224	26	in	in	ADP
easat-8918	224	27	neural	neural	ADJ
easat-8918	224	28	information	information	NOUN
easat-8918	224	29	processing	processing	NOUN
easat-8918	224	30	systems	system	NOUN
easat-8918	224	31	,	,	PUNCT
easat-8918	224	32	vol	vol	NOUN
easat-8918	224	33	.	.	PROPN
easat-8918	224	34	25	25	NUM
easat-8918	224	35	,	,	PUNCT
easat-8918	224	36	2012	2012	NUM
easat-8918	224	37	.	.	PUNCT
easat-8918	225	1	[	[	X
easat-8918	225	2	17	17	NUM
easat-8918	225	3	]	]	PUNCT
easat-8918	225	4	a.	a.	NOUN
easat-8918	225	5	daalah	daalah	PROPN
easat-8918	225	6	,	,	PUNCT
easat-8918	225	7	s.	s.	PROPN
easat-8918	225	8	aldousari	aldousari	PROPN
easat-8918	225	9	,	,	PUNCT
easat-8918	225	10	s.	s.	PROPN
easat-8918	225	11	alrashed	alrashe	VERB
easat-8918	225	12	,	,	PUNCT
easat-8918	225	13	d.	d.	PROPN
easat-8918	225	14	alenezi	alenezi	PROPN
easat-8918	225	15	,	,	PUNCT
easat-8918	225	16	a.	a.	NOUN
easat-8918	225	17	alhajri	alhajri	PROPN
easat-8918	225	18	,	,	PUNCT
easat-8918	225	19	and	and	CCONJ
easat-8918	225	20	a.	a.	NOUN
easat-8918	225	21	eleyan	eleyan	NOUN
easat-8918	225	22	,	,	PUNCT
easat-8918	225	23	"	"	PUNCT
easat-8918	225	24	plantleaf	plantleaf	ADJ
easat-8918	225	25	disease	disease	NOUN
easat-8918	225	26	classification	classification	NOUN
easat-8918	225	27	using	use	VERB
easat-8918	225	28	fine	fine	ADV
easat-8918	225	29	-	-	PUNCT
easat-8918	225	30	tuned	tune	VERB
easat-8918	225	31	pre	pre	ADJ
easat-8918	225	32	-	-	ADJ
easat-8918	225	33	trained	train	VERB
easat-8918	225	34	cnn	cnn	PROPN
easat-8918	225	35	models	model	NOUN
easat-8918	225	36	,	,	PUNCT
easat-8918	225	37	"	"	PUNCT
easat-8918	225	38	in	in	ADP
easat-8918	225	39	6th	6th	ADJ
easat-8918	225	40	international	international	ADJ
easat-8918	225	41	conference	conference	NOUN
easat-8918	225	42	on	on	ADP
easat-8918	225	43	bio	bio	NOUN
easat-8918	225	44	-	-	NOUN
easat-8918	225	45	engineering	engineering	NOUN
easat-8918	225	46	for	for	ADP
easat-8918	225	47	smart	smart	ADJ
easat-8918	225	48	technologies	technology	NOUN
easat-8918	225	49	,	,	PUNCT
easat-8918	225	50	2025	2025	NUM
easat-8918	225	51	.	.	PUNCT
easat-8918	226	1	[	[	X
easat-8918	226	2	18	18	NUM
easat-8918	226	3	]	]	X
easat-8918	226	4	l.	l.	PROPN
easat-8918	226	5	bottou	bottou	PROPN
easat-8918	226	6	,	,	PUNCT
easat-8918	226	7	"	"	PUNCT
easat-8918	226	8	large	large	ADJ
easat-8918	226	9	-	-	PUNCT
easat-8918	226	10	scale	scale	NOUN
easat-8918	226	11	machine	machine	NOUN
easat-8918	226	12	learning	learn	VERB
easat-8918	226	13	with	with	ADP
easat-8918	226	14	stochastic	stochastic	ADJ
easat-8918	226	15	gradient	gradient	ADJ
easat-8918	226	16	descent	descent	NOUN
easat-8918	226	17	,	,	PUNCT
easat-8918	226	18	"	"	PUNCT
easat-8918	226	19	in	in	ADP
easat-8918	226	20	19th	19th	ADJ
easat-8918	226	21	international	international	ADJ
easat-8918	226	22	conference	conference	NOUN
easat-8918	226	23	on	on	ADP
easat-8918	226	24	computational	computational	ADJ
easat-8918	226	25	statistics	statistic	NOUN
easat-8918	226	26	,	,	PUNCT
easat-8918	226	27	2010	2010	NUM
easat-8918	226	28	.	.	PUNCT
easat-8918	227	1	[	[	X
easat-8918	227	2	19	19	NUM
easat-8918	227	3	]	]	PUNCT
easat-8918	227	4	k.	k.	NOUN
easat-8918	227	5	simonyan	simonyan	PROPN
easat-8918	227	6	and	and	CCONJ
easat-8918	227	7	a.	a.	NOUN
easat-8918	227	8	zisserman	zisserman	PROPN
easat-8918	227	9	,	,	PUNCT
easat-8918	227	10	"	"	PUNCT
easat-8918	227	11	very	very	ADV
easat-8918	227	12	deep	deep	ADJ
easat-8918	227	13	convolutional	convolutional	ADJ
easat-8918	227	14	networks	network	NOUN
easat-8918	227	15	for	for	ADP
easat-8918	227	16	large	large	ADJ
easat-8918	227	17	-	-	PUNCT
easat-8918	227	18	scale	scale	NOUN
easat-8918	227	19	image	image	NOUN
easat-8918	227	20	recognition	recognition	NOUN
easat-8918	227	21	,	,	PUNCT
easat-8918	227	22	"	"	PUNCT
easat-8918	227	23	arxiv	arxiv	PROPN
easat-8918	227	24	preprint	preprint	NOUN
easat-8918	227	25	arxiv:1409.1556	arxiv:1409.1556	NOUN
easat-8918	227	26	,	,	PUNCT
easat-8918	227	27	2014	2014	NUM
easat-8918	227	28	.	.	PUNCT
easat-8918	228	1	[	[	X
easat-8918	228	2	20	20	NUM
easat-8918	228	3	]	]	X
easat-8918	228	4	c.	c.	NOUN
easat-8918	228	5	szegedy	szegedy	VERB
easat-8918	228	6	et	et	PROPN
easat-8918	228	7	al	al	PROPN
easat-8918	228	8	.	.	PROPN
easat-8918	228	9	,	,	PUNCT
easat-8918	228	10	"	"	PUNCT
easat-8918	228	11	going	go	VERB
easat-8918	228	12	deeper	deeply	ADV
easat-8918	228	13	with	with	ADP
easat-8918	228	14	convolutions	convolution	NOUN
easat-8918	228	15	,	,	PUNCT
easat-8918	228	16	"	"	PUNCT
easat-8918	228	17	in	in	ADP
easat-8918	228	18	proceedings	proceeding	NOUN
easat-8918	228	19	of	of	ADP
easat-8918	228	20	the	the	DET
easat-8918	228	21	nstitute	nstitute	NOUN
easat-8918	228	22	of	of	ADP
easat-8918	228	23	electrical	electrical	ADJ
easat-8918	228	24	and	and	CCONJ
easat-8918	228	25	electronics	electronic	NOUN
easat-8918	228	26	engineers.history	engineers.history	PROPN
easat-8918	228	27	conference	conference	NOUN
easat-8918	228	28	on	on	ADP
easat-8918	228	29	computer	computer	NOUN
easat-8918	228	30	vision	vision	NOUN
easat-8918	228	31	and	and	CCONJ
easat-8918	228	32	pattern	pattern	NOUN
easat-8918	228	33	recognition	recognition	NOUN
easat-8918	228	34	,	,	PUNCT
easat-8918	228	35	2015	2015	NUM
easat-8918	228	36	.	.	PUNCT
easat-8918	229	1	[	[	X
easat-8918	229	2	21	21	NUM
easat-8918	229	3	]	]	PUNCT
easat-8918	229	4	k.	k.	NOUN
easat-8918	230	1	he	he	PROPN
easat-8918	230	2	,	,	PUNCT
easat-8918	230	3	x.	x.	PROPN
easat-8918	230	4	zhang	zhang	PROPN
easat-8918	230	5	,	,	PUNCT
easat-8918	230	6	s.	s.	PROPN
easat-8918	230	7	ren	ren	PROPN
easat-8918	230	8	,	,	PUNCT
easat-8918	230	9	and	and	CCONJ
easat-8918	230	10	j.	j.	PROPN
easat-8918	230	11	sun	sun	PROPN
easat-8918	230	12	,	,	PUNCT
easat-8918	230	13	"	"	PUNCT
easat-8918	230	14	deep	deep	ADJ
easat-8918	230	15	residual	residual	ADJ
easat-8918	230	16	learning	learning	NOUN
easat-8918	230	17	for	for	ADP
easat-8918	230	18	image	image	NOUN
easat-8918	230	19	recognition	recognition	NOUN
easat-8918	230	20	,	,	PUNCT
easat-8918	230	21	"	"	PUNCT
easat-8918	230	22	in	in	ADP
easat-8918	230	23	proceedings	proceeding	NOUN
easat-8918	230	24	of	of	ADP
easat-8918	230	25	the	the	DET
easat-8918	230	26	ieee	ieee	NOUN
easat-8918	230	27	conference	conference	NOUN
easat-8918	230	28	on	on	ADP
easat-8918	230	29	computer	computer	NOUN
easat-8918	230	30	vision	vision	NOUN
easat-8918	230	31	and	and	CCONJ
easat-8918	230	32	pattern	pattern	NOUN
easat-8918	230	33	recognition	recognition	NOUN
easat-8918	230	34	2016	2016	NUM
easat-8918	230	35	.	.	PUNCT
easat-8918	231	1	[	[	X
easat-8918	231	2	22	22	NUM
easat-8918	231	3	]	]	X
easat-8918	231	4	g.	g.	PROPN
easat-8918	231	5	sehu	sehu	PROPN
easat-8918	231	6	,	,	PUNCT
easat-8918	231	7	a.	a.	PROPN
easat-8918	231	8	b.	b.	PROPN
easat-8918	231	9	ashir	ashir	PROPN
easat-8918	231	10	,	,	PUNCT
easat-8918	231	11	and	and	CCONJ
easat-8918	231	12	a.	a.	NOUN
easat-8918	231	13	eleyan	eleyan	NOUN
easat-8918	231	14	,	,	PUNCT
easat-8918	231	15	"	"	PUNCT
easat-8918	231	16	character	character	NOUN
easat-8918	231	17	recognition	recognition	NOUN
easat-8918	231	18	using	use	VERB
easat-8918	231	19	correlation	correlation	NOUN
easat-8918	231	20	&	&	CCONJ
easat-8918	231	21	hamming	hamming	PROPN
easat-8918	231	22	distance	distance	NOUN
easat-8918	231	23	,	,	PUNCT
easat-8918	231	24	"	"	PUNCT
easat-8918	231	25	in	in	ADP
easat-8918	231	26	signal	signal	ADJ
easat-8918	231	27	processing	processing	NOUN
easat-8918	231	28	and	and	CCONJ
easat-8918	231	29	communications	communication	NOUN
easat-8918	231	30	applications	application	NOUN
easat-8918	231	31	conference	conference	NOUN
easat-8918	231	32	,	,	PUNCT
easat-8918	231	33	2015	2015	NUM
easat-8918	231	34	.	.	PUNCT
easat-8918	232	1	[	[	X
easat-8918	232	2	23	23	NUM
easat-8918	232	3	]	]	PUNCT
easat-8918	232	4	j.	j.	PROPN
easat-8918	232	5	larour	larour	PROPN
easat-8918	232	6	,	,	PUNCT
easat-8918	232	7	l.	l.	PROPN
easat-8918	232	8	e.	e.	PROPN
easat-8918	232	9	aranchuk	aranchuk	PROPN
easat-8918	232	10	,	,	PUNCT
easat-8918	232	11	y.	y.	PROPN
easat-8918	232	12	danisman	danisman	PROPN
easat-8918	232	13	,	,	PUNCT
easat-8918	232	14	a.	a.	NOUN
easat-8918	232	15	eleyan	eleyan	NOUN
easat-8918	232	16	,	,	PUNCT
easat-8918	232	17	and	and	CCONJ
easat-8918	232	18	m.	m.	PROPN
easat-8918	232	19	f.	f.	PROPN
easat-8918	232	20	yilmaz	yilmaz	PROPN
easat-8918	232	21	,	,	PUNCT
easat-8918	232	22	"	"	PUNCT
easat-8918	232	23	modeling	modeling	NOUN
easat-8918	232	24	of	of	ADP
easat-8918	232	25	the	the	DET
easat-8918	232	26	l	l	ADJ
easat-8918	232	27	-	-	ADJ
easat-8918	232	28	shell	shell	ADJ
easat-8918	232	29	copper	copper	NOUN
easat-8918	232	30	x	x	ADJ
easat-8918	232	31	-	-	ADJ
easat-8918	232	32	pinch	pinch	ADJ
easat-8918	232	33	plasma	plasma	NOUN
easat-8918	232	34	produced	produce	VERB
easat-8918	232	35	by	by	ADP
easat-8918	232	36	the	the	DET
easat-8918	232	37	compact	compact	ADJ
easat-8918	232	38	generator	generator	NOUN
easat-8918	232	39	of	of	ADP
easat-8918	232	40	ecole	ecole	PROPN
easat-8918	232	41	polytechnique	polytechnique	NOUN
easat-8918	232	42	using	use	VERB
easat-8918	232	43	pattern	pattern	NOUN
easat-8918	232	44	recognition	recognition	NOUN
easat-8918	232	45	,	,	PUNCT
easat-8918	232	46	"	"	PUNCT
easat-8918	232	47	physics	physics	NOUN
easat-8918	232	48	of	of	ADP
easat-8918	232	49	plasmas	plasma	NOUN
easat-8918	232	50	,	,	PUNCT
easat-8918	232	51	vol	vol	NOUN
easat-8918	232	52	.	.	PROPN
easat-8918	232	53	23	23	NUM
easat-8918	232	54	,	,	PUNCT
easat-8918	232	55	no	no	INTJ
easat-8918	232	56	.	.	NOUN
easat-8918	232	57	3	3	NUM
easat-8918	232	58	,	,	PUNCT
easat-8918	232	59	2016	2016	NUM
easat-8918	232	60	.	.	PUNCT
easat-8918	233	1	[	[	X
easat-8918	233	2	24	24	NUM
easat-8918	233	3	]	]	PUNCT
easat-8918	233	4	f.	f.	PROPN
easat-8918	233	5	m.	m.	PROPN
easat-8918	233	6	yilmaz	yilmaz	PROPN
easat-8918	233	7	,	,	PUNCT
easat-8918	233	8	a.	a.	PROPN
easat-8918	233	9	eleyan	eleyan	PROPN
easat-8918	233	10	,	,	PUNCT
easat-8918	233	11	l.	l.	PROPN
easat-8918	233	12	e.	e.	PROPN
easat-8918	233	13	aranchuk	aranchuk	PROPN
easat-8918	233	14	,	,	PUNCT
easat-8918	233	15	and	and	CCONJ
easat-8918	233	16	j.	j.	PROPN
easat-8918	233	17	larour	larour	PROPN
easat-8918	233	18	,	,	PUNCT
easat-8918	233	19	"	"	PUNCT
easat-8918	233	20	spectroscopic	spectroscopic	ADJ
easat-8918	233	21	analysis	analysis	NOUN
easat-8918	233	22	of	of	ADP
easat-8918	233	23	x	x	ADJ
easat-8918	233	24	-	-	ADJ
easat-8918	233	25	pinch	pinch	ADJ
easat-8918	233	26	plasma	plasma	NOUN
easat-8918	233	27	produced	produce	VERB
easat-8918	233	28	on	on	ADP
easat-8918	233	29	the	the	DET
easat-8918	233	30	compact	compact	ADJ
easat-8918	233	31	lc	lc	PROPN
easat-8918	233	32	-	-	PUNCT
easat-8918	233	33	generator	generator	NOUN
easat-8918	233	34	of	of	ADP
easat-8918	233	35	ecole	ecole	PROPN
easat-8918	233	36	polytechnique	polytechnique	NOUN
easat-8918	233	37	using	use	VERB
easat-8918	233	38	artificial	artificial	ADJ
easat-8918	233	39	neural	neural	ADJ
easat-8918	233	40	networks	network	NOUN
easat-8918	233	41	,	,	PUNCT
easat-8918	233	42	"	"	PUNCT
easat-8918	233	43	high	high	ADJ
easat-8918	233	44	energy	energy	NOUN
easat-8918	233	45	density	density	NOUN
easat-8918	233	46	physics	physics	PROPN
easat-8918	233	47	,	,	PUNCT
easat-8918	233	48	vol	vol	NOUN
easat-8918	233	49	.	.	PROPN
easat-8918	233	50	12	12	NUM
easat-8918	233	51	,	,	PUNCT
easat-8918	233	52	pp	pp	ADJ
easat-8918	233	53	.	.	PUNCT
easat-8918	234	1	1	1	NUM
easat-8918	234	2	-	-	SYM
easat-8918	234	3	4	4	NUM
easat-8918	234	4	,	,	PUNCT
easat-8918	234	5	2014	2014	NUM
easat-8918	234	6	.	.	PUNCT
easat-8918	235	1	https://doi.org/10.1016/j.hedp.2014.04.001	https://doi.org/10.1016/j.hedp.2014.04.001	PROPN
easat-8918	235	2	[	[	X
easat-8918	235	3	25	25	NUM
easat-8918	235	4	]	]	PUNCT
easat-8918	235	5	g.	g.	PROPN
easat-8918	235	6	hu	hu	PROPN
easat-8918	235	7	et	et	PROPN
easat-8918	235	8	al	al	PROPN
easat-8918	235	9	.	.	PROPN
easat-8918	235	10	,	,	PUNCT
easat-8918	235	11	"	"	PUNCT
easat-8918	235	12	when	when	SCONJ
easat-8918	235	13	face	face	NOUN
easat-8918	235	14	recognition	recognition	NOUN
easat-8918	235	15	meets	meet	VERB
easat-8918	235	16	with	with	ADP
easat-8918	235	17	deep	deep	ADJ
easat-8918	235	18	learning	learning	NOUN
easat-8918	235	19	:	:	PUNCT
easat-8918	235	20	an	an	DET
easat-8918	235	21	evaluation	evaluation	NOUN
easat-8918	235	22	of	of	ADP
easat-8918	235	23	convolutional	convolutional	ADJ
easat-8918	235	24	neural	neural	ADJ
easat-8918	235	25	networks	network	NOUN
easat-8918	235	26	for	for	ADP
easat-8918	235	27	face	face	NOUN
easat-8918	235	28	recognition	recognition	NOUN
easat-8918	235	29	,	,	PUNCT
easat-8918	235	30	"	"	PUNCT
easat-8918	235	31	in	in	ADP
easat-8918	235	32	proceedings	proceeding	NOUN
easat-8918	235	33	of	of	ADP
easat-8918	235	34	the	the	DET
easat-8918	235	35	ieee	ieee	NOUN
easat-8918	235	36	international	international	PROPN
easat-8918	235	37	conference	conference	NOUN
easat-8918	235	38	on	on	ADP
easat-8918	235	39	computer	computer	NOUN
easat-8918	235	40	vision	vision	NOUN
easat-8918	235	41	workshops	workshop	NOUN
easat-8918	235	42	,	,	PUNCT
easat-8918	235	43	2015	2015	NUM
easat-8918	235	44	.	.	PUNCT
easat-8918	236	1	[	[	X
easat-8918	236	2	26	26	NUM
easat-8918	236	3	]	]	X
easat-8918	236	4	i.	i.	NOUN
easat-8918	236	5	mesecan	mesecan	PROPN
easat-8918	236	6	,	,	PUNCT
easat-8918	236	7	a.	a.	NOUN
easat-8918	236	8	eleyan	eleyan	NOUN
easat-8918	236	9	,	,	PUNCT
easat-8918	236	10	and	and	CCONJ
easat-8918	236	11	b.	b.	PROPN
easat-8918	236	12	karlık	karlık	NOUN
easat-8918	236	13	,	,	PUNCT
easat-8918	236	14	"	"	PUNCT
easat-8918	236	15	sift	sift	NOUN
easat-8918	236	16	-	-	PUNCT
easat-8918	236	17	based	base	VERB
easat-8918	236	18	iris	iris	NOUN
easat-8918	236	19	recognition	recognition	NOUN
easat-8918	236	20	using	use	VERB
easat-8918	236	21	sub	sub	NOUN
easat-8918	236	22	-	-	NOUN
easat-8918	236	23	segments	segment	NOUN
easat-8918	236	24	,	,	PUNCT
easat-8918	236	25	the	the	DET
easat-8918	236	26	international	international	ADJ
easat-8918	236	27	conference	conference	NOUN
easat-8918	236	28	on	on	ADP
easat-8918	236	29	technological	technological	ADJ
easat-8918	236	30	advances	advance	NOUN
easat-8918	236	31	in	in	ADP
easat-8918	236	32	electrical	electrical	ADJ
easat-8918	236	33	,	,	PUNCT
easat-8918	236	34	"	"	PUNCT
easat-8918	236	35	in	in	ADP
easat-8918	236	36	electronics	electronics	NOUN
easat-8918	236	37	and	and	CCONJ
easat-8918	236	38	computer	computer	NOUN
easat-8918	236	39	engineering	engineering	NOUN
easat-8918	236	40	2013	2013	NUM
easat-8918	236	41	.	.	PUNCT
easat-8918	237	1	[	[	X
easat-8918	237	2	27	27	NUM
easat-8918	237	3	]	]	PUNCT
easat-8918	237	4	a.	a.	NOUN
easat-8918	237	5	m.	m.	NOUN
easat-8918	237	6	ashir	ashir	NOUN
easat-8918	237	7	and	and	CCONJ
easat-8918	237	8	a.	a.	NOUN
easat-8918	237	9	eleyan	eleyan	NOUN
easat-8918	237	10	,	,	PUNCT
easat-8918	237	11	"	"	PUNCT
easat-8918	237	12	compressive	compressive	ADJ
easat-8918	237	13	sensing	sensing	NOUN
easat-8918	237	14	based	base	VERB
easat-8918	237	15	facial	facial	ADJ
easat-8918	237	16	expression	expression	NOUN
easat-8918	237	17	recognition	recognition	NOUN
easat-8918	237	18	,	,	PUNCT
easat-8918	237	19	"	"	PUNCT
easat-8918	237	20	in	in	ADP
easat-8918	237	21	24th	24th	ADJ
easat-8918	237	22	signal	signal	NOUN
easat-8918	237	23	processing	processing	NOUN
easat-8918	237	24	and	and	CCONJ
easat-8918	237	25	communication	communication	NOUN
easat-8918	237	26	application	application	NOUN
easat-8918	237	27	conference	conference	NOUN
easat-8918	237	28	.	.	PUNCT
easat-8918	237	29	,	,	PUNCT
easat-8918	237	30	2016	2016	NUM
easat-8918	237	31	.	.	PUNCT
easat-8918	238	1	[	[	X
easat-8918	238	2	28	28	NUM
easat-8918	238	3	]	]	X
easat-8918	238	4	a.	a.	NOUN
easat-8918	238	5	eleyan	eleyan	PROPN
easat-8918	238	6	,	,	PUNCT
easat-8918	238	7	k.	k.	PROPN
easat-8918	238	8	kose	kose	PROPN
easat-8918	238	9	,	,	PUNCT
easat-8918	238	10	and	and	CCONJ
easat-8918	238	11	a.	a.	PROPN
easat-8918	238	12	e.	e.	PROPN
easat-8918	238	13	cetin	cetin	PROPN
easat-8918	238	14	,	,	PUNCT
easat-8918	238	15	"	"	PUNCT
easat-8918	238	16	image	image	NOUN
easat-8918	238	17	feature	feature	NOUN
easat-8918	238	18	extraction	extraction	NOUN
easat-8918	238	19	using	use	VERB
easat-8918	238	20	compressive	compressive	ADJ
easat-8918	238	21	sensing	sensing	NOUN
easat-8918	238	22	,	,	PUNCT
easat-8918	238	23	image	image	NOUN
easat-8918	238	24	processing	processing	NOUN
easat-8918	238	25	"	"	PUNCT
easat-8918	238	26	in	in	ADP
easat-8918	238	27	communications	communication	NOUN
easat-8918	238	28	challenges	challenge	NOUN
easat-8918	238	29	conference	conference	NOUN
easat-8918	238	30	,	,	PUNCT
easat-8918	238	31	2013	2013	NUM
easat-8918	238	32	.	.	PUNCT
easat-8918	239	1	[	[	X
easat-8918	239	2	29	29	NUM
easat-8918	239	3	]	]	PUNCT
easat-8918	239	4	a.	a.	NOUN
easat-8918	239	5	eleyan	eleyan	PROPN
easat-8918	239	6	,	,	PUNCT
easat-8918	239	7	h.	h.	PROPN
easat-8918	239	8	demirel	demirel	PROPN
easat-8918	239	9	,	,	PUNCT
easat-8918	239	10	and	and	CCONJ
easat-8918	239	11	h.	h.	PROPN
easat-8918	239	12	ozkaramanli	ozkaramanli	PROPN
easat-8918	239	13	,	,	PUNCT
easat-8918	239	14	"	"	PUNCT
easat-8918	239	15	face	face	NOUN
easat-8918	239	16	recognition	recognition	NOUN
easat-8918	239	17	using	use	VERB
easat-8918	239	18	dual	dual	ADJ
easat-8918	239	19	-	-	PUNCT
easat-8918	239	20	tree	tree	NOUN
easat-8918	239	21	wavelet	wavelet	NOUN
easat-8918	239	22	transform	transform	NOUN
easat-8918	239	23	,	,	PUNCT
easat-8918	239	24	"	"	PUNCT
easat-8918	239	25	in	in	ADP
easat-8918	239	26	proceedings	proceeding	NOUN
easat-8918	239	27	of	of	ADP
easat-8918	239	28	the	the	DET
easat-8918	239	29	ieee	ieee	NOUN
easat-8918	239	30	international	international	ADJ
easat-8918	239	31	symposium	symposium	NOUN
easat-8918	239	32	on	on	ADP
easat-8918	239	33	signal	signal	NOUN
easat-8918	239	34	processing	processing	NOUN
easat-8918	239	35	and	and	CCONJ
easat-8918	239	36	information	information	NOUN
easat-8918	239	37	technology	technology	NOUN
easat-8918	239	38	,	,	PUNCT
easat-8918	239	39	2008	2008	NUM
easat-8918	239	40	.	.	PUNCT
easat-8918	240	1	[	[	X
easat-8918	240	2	30	30	NUM
easat-8918	240	3	]	]	PUNCT
easat-8918	240	4	a.	a.	NOUN
easat-8918	240	5	m.	m.	PROPN
easat-8918	240	6	ashir	ashir	PROPN
easat-8918	240	7	,	,	PUNCT
easat-8918	240	8	a.	a.	NOUN
easat-8918	240	9	eleyan	eleyan	NOUN
easat-8918	240	10	,	,	PUNCT
easat-8918	240	11	and	and	CCONJ
easat-8918	240	12	b.	b.	PROPN
easat-8918	240	13	akdemir	akdemir	PROPN
easat-8918	240	14	,	,	PUNCT
easat-8918	240	15	"	"	PUNCT
easat-8918	240	16	facial	facial	ADJ
easat-8918	240	17	expression	expression	NOUN
easat-8918	240	18	recognition	recognition	NOUN
easat-8918	240	19	with	with	ADP
easat-8918	240	20	dynamic	dynamic	ADJ
easat-8918	240	21	cascaded	cascade	VERB
easat-8918	240	22	classifier	classifier	NOUN
easat-8918	240	23	,	,	PUNCT
easat-8918	240	24	"	"	PUNCT
easat-8918	240	25	neural	neural	ADJ
easat-8918	240	26	computing	computing	NOUN
easat-8918	240	27	and	and	CCONJ
easat-8918	240	28	applications	application	NOUN
easat-8918	240	29	,	,	PUNCT
easat-8918	240	30	vol	vol	NOUN
easat-8918	240	31	.	.	PROPN
easat-8918	240	32	32	32	NUM
easat-8918	240	33	,	,	PUNCT
easat-8918	240	34	no	no	INTJ
easat-8918	240	35	.	.	NOUN
easat-8918	240	36	10	10	NUM
easat-8918	240	37	,	,	PUNCT
easat-8918	240	38	pp	pp	ADJ
easat-8918	240	39	.	.	PUNCT
easat-8918	241	1	6295	6295	NUM
easat-8918	241	2	-	-	SYM
easat-8918	241	3	6309	6309	NUM
easat-8918	241	4	,	,	PUNCT
easat-8918	241	5	2020	2020	NUM
easat-8918	241	6	.	.	PUNCT
easat-8918	242	1	https://doi.org/10.1007/s00521-019-04138-4	https://doi.org/10.1007/s00521-019-04138-4	PROPN
easat-8918	242	2	https://doi.org/10.1109/icecco.2015.7416909	https://doi.org/10.1109/icecco.2015.7416909	PROPN
easat-8918	242	3	https://doi.org/10.5505/pajes.2018.06982	https://doi.org/10.5505/pajes.2018.06982	PROPN
easat-8918	242	4	https://doi.org/10.3390/app14219936	https://doi.org/10.3390/app14219936	PROPN
easat-8918	242	5	https://doi.org/10.1016/j.hedp.2014.04.001	https://doi.org/10.1016/j.hedp.2014.04.001	PROPN
easat-8918	242	6	https://doi.org/10.1007/s00521-019-04138-4	https://doi.org/10.1007/s00521-019-04138-4	PROPN
easat-8918	242	7	1358	1358	NUM
easat-8918	242	8	edelweiss	edelweiss	PROPN
easat-8918	242	9	applied	apply	VERB
easat-8918	242	10	science	science	NOUN
easat-8918	242	11	and	and	CCONJ
easat-8918	242	12	technology	technology	NOUN
easat-8918	242	13	issn	issn	PROPN
easat-8918	242	14	:	:	PUNCT
easat-8918	242	15	2576	2576	NUM
easat-8918	242	16	-	-	SYM
easat-8918	242	17	8484	8484	NUM
easat-8918	242	18	vol	vol	NOUN
easat-8918	242	19	.	.	PROPN
easat-8918	243	1	9	9	NUM
easat-8918	243	2	,	,	PUNCT
easat-8918	243	3	no	no	INTJ
easat-8918	243	4	.	.	NOUN
easat-8918	243	5	7	7	NUM
easat-8918	243	6	:	:	SYM
easat-8918	243	7	1345	1345	NUM
easat-8918	243	8	-	-	SYM
easat-8918	243	9	1359	1359	NUM
easat-8918	243	10	,	,	PUNCT
easat-8918	243	11	2025	2025	NUM
easat-8918	243	12	doi	doi	NOUN
easat-8918	243	13	:	:	PUNCT
easat-8918	243	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	243	15	©	©	PROPN
easat-8918	243	16	2025	2025	NUM
easat-8918	243	17	by	by	ADP
easat-8918	243	18	the	the	DET
easat-8918	243	19	author	author	NOUN
easat-8918	243	20	;	;	PUNCT
easat-8918	243	21	licensee	licensee	PROPN
easat-8918	243	22	learning	learning	NOUN
easat-8918	243	23	gate	gate	NOUN
easat-8918	244	1	[	[	X
easat-8918	244	2	31	31	NUM
easat-8918	244	3	]	]	PUNCT
easat-8918	244	4	a.	a.	NOUN
easat-8918	244	5	eleyan	eleyan	NOUN
easat-8918	244	6	,	,	PUNCT
easat-8918	244	7	"	"	PUNCT
easat-8918	244	8	statistical	statistical	ADJ
easat-8918	244	9	local	local	ADJ
easat-8918	244	10	descriptors	descriptor	NOUN
easat-8918	244	11	for	for	ADP
easat-8918	244	12	face	face	NOUN
easat-8918	244	13	recognition	recognition	NOUN
easat-8918	244	14	:	:	PUNCT
easat-8918	244	15	a	a	DET
easat-8918	244	16	comprehensive	comprehensive	ADJ
easat-8918	244	17	study	study	NOUN
easat-8918	244	18	,	,	PUNCT
easat-8918	244	19	"	"	PUNCT
easat-8918	244	20	multimedia	multimedia	NOUN
easat-8918	244	21	tools	tool	NOUN
easat-8918	244	22	and	and	CCONJ
easat-8918	244	23	applications	application	NOUN
easat-8918	244	24	,	,	PUNCT
easat-8918	244	25	vol	vol	NOUN
easat-8918	244	26	.	.	PROPN
easat-8918	244	27	82	82	NUM
easat-8918	244	28	,	,	PUNCT
easat-8918	244	29	no	no	INTJ
easat-8918	244	30	.	.	NOUN
easat-8918	244	31	21	21	NUM
easat-8918	244	32	,	,	PUNCT
easat-8918	244	33	pp	pp	ADJ
easat-8918	244	34	.	.	PUNCT
easat-8918	244	35	32485	32485	NUM
easat-8918	244	36	-	-	SYM
easat-8918	244	37	32504	32504	NUM
easat-8918	244	38	,	,	PUNCT
easat-8918	244	39	2023	2023	NUM
easat-8918	244	40	.	.	PUNCT
easat-8918	244	41	https://doi.org/10.1007/s11042-023-14482-2	https://doi.org/10.1007/s11042-023-14482-2	PUNCT
easat-8918	245	1	[	[	X
easat-8918	245	2	32	32	NUM
easat-8918	245	3	]	]	PUNCT
easat-8918	245	4	m.	m.	PROPN
easat-8918	245	5	al	al	PROPN
easat-8918	245	6	-	-	PUNCT
easat-8918	245	7	azzeh	azzeh	PROPN
easat-8918	245	8	,	,	PUNCT
easat-8918	245	9	a.	a.	NOUN
easat-8918	245	10	eleyan	eleyan	PROPN
easat-8918	245	11	,	,	PUNCT
easat-8918	245	12	and	and	CCONJ
easat-8918	245	13	h.	h.	PROPN
easat-8918	245	14	demirel	demirel	PROPN
easat-8918	245	15	,	,	PUNCT
easat-8918	245	16	"	"	PUNCT
easat-8918	245	17	pca	pca	NOUN
easat-8918	245	18	-	-	PUNCT
easat-8918	245	19	based	base	VERB
easat-8918	245	20	face	face	NOUN
easat-8918	245	21	recognition	recognition	NOUN
easat-8918	245	22	from	from	ADP
easat-8918	245	23	video	video	NOUN
easat-8918	245	24	using	use	VERB
easat-8918	245	25	super	super	NOUN
easat-8918	245	26	-	-	NOUN
easat-8918	245	27	resolution	resolution	NOUN
easat-8918	245	28	,	,	PUNCT
easat-8918	245	29	"	"	PUNCT
easat-8918	245	30	in	in	ADP
easat-8918	245	31	23rd	23rd	ADJ
easat-8918	245	32	international	international	ADJ
easat-8918	245	33	symposium	symposium	NOUN
easat-8918	245	34	on	on	ADP
easat-8918	245	35	computer	computer	NOUN
easat-8918	245	36	and	and	CCONJ
easat-8918	245	37	information	information	NOUN
easat-8918	245	38	sciences	science	NOUN
easat-8918	245	39	,	,	PUNCT
easat-8918	245	40	2008	2008	NUM
easat-8918	245	41	.	.	PUNCT
easat-8918	246	1	[	[	X
easat-8918	246	2	33	33	NUM
easat-8918	246	3	]	]	PUNCT
easat-8918	246	4	a.	a.	NOUN
easat-8918	246	5	eleyan	eleyan	NOUN
easat-8918	246	6	,	,	PUNCT
easat-8918	246	7	"	"	PUNCT
easat-8918	246	8	breast	breast	NOUN
easat-8918	246	9	cancer	cancer	NOUN
easat-8918	246	10	classification	classification	NOUN
easat-8918	246	11	using	use	VERB
easat-8918	246	12	moments	moment	NOUN
easat-8918	246	13	,	,	PUNCT
easat-8918	246	14	"	"	PUNCT
easat-8918	246	15	in	in	ADP
easat-8918	246	16	20th	20th	ADJ
easat-8918	246	17	signal	signal	NOUN
easat-8918	246	18	processing	processing	NOUN
easat-8918	246	19	and	and	CCONJ
easat-8918	246	20	communications	communication	NOUN
easat-8918	246	21	applications	application	NOUN
easat-8918	246	22	conference	conference	NOUN
easat-8918	246	23	,	,	PUNCT
easat-8918	246	24	2012	2012	NUM
easat-8918	246	25	.	.	PUNCT
easat-8918	247	1	[	[	X
easat-8918	247	2	34	34	NUM
easat-8918	247	3	]	]	PUNCT
easat-8918	247	4	a.	a.	NOUN
easat-8918	247	5	eleyan	eleyan	PROPN
easat-8918	247	6	,	,	PUNCT
easat-8918	247	7	z.	z.	PROPN
easat-8918	247	8	al	al	PROPN
easat-8918	247	9	-	-	PUNCT
easat-8918	247	10	barakeh	barakeh	PROPN
easat-8918	247	11	,	,	PUNCT
easat-8918	247	12	r.	r.	PROPN
easat-8918	247	13	ghandour	ghandour	PROPN
easat-8918	247	14	,	,	PUNCT
easat-8918	247	15	b.	b.	PROPN
easat-8918	247	16	neji	neji	PROPN
easat-8918	247	17	,	,	PUNCT
easat-8918	247	18	and	and	CCONJ
easat-8918	247	19	a.	a.	NOUN
easat-8918	247	20	eleyan	eleyan	NOUN
easat-8918	247	21	,	,	PUNCT
easat-8918	247	22	"	"	PUNCT
easat-8918	247	23	brain	brain	NOUN
easat-8918	247	24	tumor	tumor	NOUN
easat-8918	247	25	detection	detection	NOUN
easat-8918	247	26	via	via	ADP
easat-8918	247	27	ensemble	ensemble	ADJ
easat-8918	247	28	cnn	cnn	PROPN
easat-8918	247	29	-	-	PUNCT
easat-8918	247	30	based	base	VERB
easat-8918	247	31	deep	deep	ADJ
easat-8918	247	32	learning	learning	NOUN
easat-8918	247	33	models	model	NOUN
easat-8918	247	34	,	,	PUNCT
easat-8918	247	35	"	"	PUNCT
easat-8918	247	36	in	in	ADP
easat-8918	247	37	6th	6th	ADJ
easat-8918	247	38	international	international	ADJ
easat-8918	247	39	conference	conference	NOUN
easat-8918	247	40	on	on	ADP
easat-8918	247	41	bio	bio	NOUN
easat-8918	247	42	-	-	NOUN
easat-8918	247	43	engineering	engineering	NOUN
easat-8918	247	44	for	for	ADP
easat-8918	247	45	smart	smart	ADJ
easat-8918	247	46	technologies	technology	NOUN
easat-8918	247	47	,	,	PUNCT
easat-8918	247	48	2025	2025	NUM
easat-8918	247	49	.	.	PUNCT
easat-8918	248	1	[	[	X
easat-8918	248	2	35	35	NUM
easat-8918	248	3	]	]	PUNCT
easat-8918	248	4	a.	a.	NOUN
easat-8918	248	5	eleyan	eleyan	NOUN
easat-8918	248	6	and	and	CCONJ
easat-8918	248	7	e.	e.	PROPN
easat-8918	248	8	alboghbaish	alboghbaish	PROPN
easat-8918	248	9	,	,	PUNCT
easat-8918	248	10	"	"	PUNCT
easat-8918	248	11	multi	multi	ADJ
easat-8918	248	12	-	-	ADJ
easat-8918	248	13	classifier	classifier	ADJ
easat-8918	248	14	deep	deep	ADJ
easat-8918	248	15	learning	learning	NOUN
easat-8918	248	16	based	base	VERB
easat-8918	248	17	system	system	NOUN
easat-8918	248	18	for	for	ADP
easat-8918	248	19	ecg	ecg	PROPN
easat-8918	248	20	classification	classification	NOUN
easat-8918	248	21	using	use	VERB
easat-8918	248	22	fourier	fourier	NOUN
easat-8918	248	23	transform	transform	NOUN
easat-8918	248	24	,	,	PUNCT
easat-8918	248	25	"	"	PUNCT
easat-8918	248	26	in	in	ADP
easat-8918	248	27	5th	5th	ADJ
easat-8918	248	28	international	international	ADJ
easat-8918	248	29	conference	conference	NOUN
easat-8918	248	30	on	on	ADP
easat-8918	248	31	bioengineering	bioengineere	VERB
easat-8918	248	32	for	for	ADP
easat-8918	248	33	smart	smart	ADJ
easat-8918	248	34	technologies	technology	NOUN
easat-8918	248	35	,	,	PUNCT
easat-8918	248	36	2023	2023	NUM
easat-8918	248	37	.	.	PUNCT
easat-8918	249	1	[	[	X
easat-8918	249	2	36	36	NUM
easat-8918	249	3	]	]	X
easat-8918	249	4	q.	q.	PROPN
easat-8918	249	5	li	li	PROPN
easat-8918	249	6	,	,	PUNCT
easat-8918	249	7	w.	w.	PROPN
easat-8918	249	8	cai	cai	PROPN
easat-8918	249	9	,	,	PUNCT
easat-8918	249	10	x.	x.	PROPN
easat-8918	249	11	wang	wang	PROPN
easat-8918	249	12	,	,	PUNCT
easat-8918	249	13	y.	y.	PROPN
easat-8918	249	14	zhou	zhou	PROPN
easat-8918	249	15	,	,	PUNCT
easat-8918	249	16	d.	d.	PROPN
easat-8918	249	17	d.	d.	PROPN
easat-8918	249	18	feng	feng	PROPN
easat-8918	249	19	,	,	PUNCT
easat-8918	249	20	and	and	CCONJ
easat-8918	249	21	m.	m.	PROPN
easat-8918	249	22	chen	chen	PROPN
easat-8918	249	23	,	,	PUNCT
easat-8918	249	24	"	"	PUNCT
easat-8918	249	25	medical	medical	ADJ
easat-8918	249	26	image	image	NOUN
easat-8918	249	27	classification	classification	NOUN
easat-8918	249	28	with	with	ADP
easat-8918	249	29	convolutional	convolutional	ADJ
easat-8918	249	30	neural	neural	ADJ
easat-8918	249	31	network	network	NOUN
easat-8918	249	32	,	,	PUNCT
easat-8918	249	33	"	"	PUNCT
easat-8918	249	34	in	in	ADP
easat-8918	249	35	13th	13th	ADJ
easat-8918	249	36	international	international	ADJ
easat-8918	249	37	conference	conference	NOUN
easat-8918	249	38	on	on	ADP
easat-8918	249	39	control	control	NOUN
easat-8918	249	40	automation	automation	NOUN
easat-8918	249	41	robotics	robotic	NOUN
easat-8918	249	42	&	&	CCONJ
easat-8918	249	43	vision	vision	PROPN
easat-8918	249	44	,	,	PUNCT
easat-8918	249	45	2014	2014	NUM
easat-8918	249	46	.	.	PUNCT
easat-8918	250	1	[	[	X
easat-8918	250	2	37	37	NUM
easat-8918	250	3	]	]	X
easat-8918	250	4	e.	e.	PROPN
easat-8918	250	5	alboghbaish	alboghbaish	PROPN
easat-8918	250	6	,	,	PUNCT
easat-8918	250	7	a.	a.	NOUN
easat-8918	250	8	eleyan	eleyan	NOUN
easat-8918	250	9	,	,	PUNCT
easat-8918	250	10	and	and	CCONJ
easat-8918	250	11	a.	a.	NOUN
easat-8918	250	12	eleyan	eleyan	NOUN
easat-8918	250	13	,	,	PUNCT
easat-8918	250	14	"	"	PUNCT
easat-8918	250	15	performance	performance	NOUN
easat-8918	250	16	comparison	comparison	NOUN
easat-8918	250	17	between	between	ADP
easat-8918	250	18	transform	transform	NOUN
easat-8918	250	19	-	-	PUNCT
easat-8918	250	20	based	base	VERB
easat-8918	250	21	deep	deep	ADJ
easat-8918	250	22	learning	learning	NOUN
easat-8918	250	23	approaches	approach	NOUN
easat-8918	250	24	for	for	ADP
easat-8918	250	25	ecg	ecg	PROPN
easat-8918	250	26	signal	signal	PROPN
easat-8918	250	27	classification	classification	NOUN
easat-8918	250	28	,	,	PUNCT
easat-8918	250	29	"	"	PUNCT
easat-8918	250	30	in	in	ADP
easat-8918	250	31	11th	11th	ADJ
easat-8918	250	32	international	international	ADJ
easat-8918	250	33	conference	conference	NOUN
easat-8918	250	34	on	on	ADP
easat-8918	250	35	electrical	electrical	ADJ
easat-8918	250	36	and	and	CCONJ
easat-8918	250	37	electronics	electronic	NOUN
easat-8918	250	38	engineering	engineering	NOUN
easat-8918	250	39	,	,	PUNCT
easat-8918	250	40	2024	2024	NUM
easat-8918	250	41	.	.	PUNCT
easat-8918	251	1	https://doi.org/10.1109/iceee62185.2024.10779262	https://doi.org/10.1109/iceee62185.2024.10779262	X
easat-8918	252	1	[	[	X
easat-8918	252	2	38	38	NUM
easat-8918	252	3	]	]	PUNCT
easat-8918	252	4	a.	a.	PROPN
easat-8918	252	5	radford	radford	PROPN
easat-8918	252	6	,	,	PUNCT
easat-8918	252	7	l.	l.	PROPN
easat-8918	252	8	metz	metz	PROPN
easat-8918	252	9	,	,	PUNCT
easat-8918	252	10	and	and	CCONJ
easat-8918	252	11	s.	s.	PROPN
easat-8918	252	12	chintala	chintala	PROPN
easat-8918	252	13	,	,	PUNCT
easat-8918	252	14	"	"	PUNCT
easat-8918	252	15	unsupervised	unsupervised	ADJ
easat-8918	252	16	representation	representation	NOUN
easat-8918	252	17	learning	learn	VERB
easat-8918	252	18	with	with	ADP
easat-8918	252	19	deep	deep	ADJ
easat-8918	252	20	convolutional	convolutional	ADJ
easat-8918	252	21	generative	generative	ADJ
easat-8918	252	22	adversarial	adversarial	ADJ
easat-8918	252	23	networks	network	NOUN
easat-8918	252	24	,	,	PUNCT
easat-8918	252	25	"	"	PUNCT
easat-8918	252	26	arxiv	arxiv	PROPN
easat-8918	252	27	preprint	preprint	NOUN
easat-8918	252	28	arxiv:1511.06434	arxiv:1511.06434	NOUN
easat-8918	252	29	,	,	PUNCT
easat-8918	252	30	2015	2015	NUM
easat-8918	252	31	.	.	PUNCT
easat-8918	253	1	[	[	X
easat-8918	253	2	39	39	NUM
easat-8918	253	3	]	]	PUNCT
easat-8918	253	4	l.	l.	PROPN
easat-8918	253	5	tran	tran	PROPN
easat-8918	253	6	,	,	PUNCT
easat-8918	253	7	x.	x.	PROPN
easat-8918	253	8	yin	yin	PROPN
easat-8918	253	9	,	,	PUNCT
easat-8918	253	10	and	and	CCONJ
easat-8918	253	11	x.	x.	NOUN
easat-8918	253	12	liu	liu	PROPN
easat-8918	253	13	,	,	PUNCT
easat-8918	253	14	"	"	PUNCT
easat-8918	253	15	disentangled	disentangle	VERB
easat-8918	253	16	representation	representation	NOUN
easat-8918	253	17	learning	learn	VERB
easat-8918	253	18	gan	gan	NOUN
easat-8918	253	19	for	for	ADP
easat-8918	253	20	pose	pose	NOUN
easat-8918	253	21	-	-	PUNCT
easat-8918	253	22	invariant	invariant	ADJ
easat-8918	253	23	face	face	NOUN
easat-8918	253	24	recognition	recognition	NOUN
easat-8918	253	25	,	,	PUNCT
easat-8918	253	26	"	"	PUNCT
easat-8918	253	27	in	in	ADP
easat-8918	253	28	proceedings	proceeding	NOUN
easat-8918	253	29	of	of	ADP
easat-8918	253	30	theinstitute	theinstitute	NOUN
easat-8918	253	31	of	of	ADP
easat-8918	253	32	electrical	electrical	ADJ
easat-8918	253	33	and	and	CCONJ
easat-8918	253	34	electronics	electronic	NOUN
easat-8918	253	35	engineers	engineer	NOUN
easat-8918	253	36	.	.	PUNCT
easat-8918	254	1	conference	conference	NOUN
easat-8918	254	2	on	on	ADP
easat-8918	254	3	computer	computer	NOUN
easat-8918	254	4	vision	vision	NOUN
easat-8918	254	5	and	and	CCONJ
easat-8918	254	6	pattern	pattern	NOUN
easat-8918	254	7	recognition	recognition	NOUN
easat-8918	254	8	,	,	PUNCT
easat-8918	254	9	2017	2017	NUM
easat-8918	254	10	.	.	PUNCT
easat-8918	255	1	[	[	X
easat-8918	255	2	40	40	NUM
easat-8918	255	3	]	]	PUNCT
easat-8918	255	4	m.	m.	NOUN
easat-8918	255	5	kas	kas	PROPN
easat-8918	255	6	,	,	PUNCT
easat-8918	255	7	a.	a.	NOUN
easat-8918	255	8	chahi	chahi	PROPN
easat-8918	255	9	,	,	PUNCT
easat-8918	255	10	i.	i.	PROPN
easat-8918	255	11	kajo	kajo	PROPN
easat-8918	255	12	,	,	PUNCT
easat-8918	255	13	and	and	CCONJ
easat-8918	255	14	y.	y.	PROPN
easat-8918	255	15	ruichek	ruichek	PROPN
easat-8918	255	16	,	,	PUNCT
easat-8918	255	17	"	"	PUNCT
easat-8918	255	18	eigengan	eigengan	VERB
easat-8918	255	19	:	:	PUNCT
easat-8918	255	20	an	an	DET
easat-8918	255	21	svd	svd	PROPN
easat-8918	255	22	subspace	subspace	NOUN
easat-8918	255	23	-	-	PUNCT
easat-8918	255	24	based	base	VERB
easat-8918	255	25	learning	learning	NOUN
easat-8918	255	26	for	for	ADP
easat-8918	255	27	image	image	NOUN
easat-8918	255	28	generation	generation	NOUN
easat-8918	255	29	using	use	VERB
easat-8918	255	30	conditional	conditional	ADJ
easat-8918	255	31	gan	gan	NOUN
easat-8918	255	32	,	,	PUNCT
easat-8918	255	33	"	"	PUNCT
easat-8918	255	34	knowledge	knowledge	NOUN
easat-8918	255	35	-	-	PUNCT
easat-8918	255	36	based	base	VERB
easat-8918	255	37	systems	system	NOUN
easat-8918	255	38	,	,	PUNCT
easat-8918	255	39	vol	vol	NOUN
easat-8918	255	40	.	.	PROPN
easat-8918	255	41	293	293	NUM
easat-8918	255	42	,	,	PUNCT
easat-8918	255	43	p.	p.	NOUN
easat-8918	255	44	111691	111691	NUM
easat-8918	255	45	,	,	PUNCT
easat-8918	255	46	2024	2024	NUM
easat-8918	255	47	.	.	PUNCT
easat-8918	256	1	https://doi.org/10.1016/j.knosys.2024.111691	https://doi.org/10.1016/j.knosys.2024.111691	PROPN
easat-8918	256	2	[	[	X
easat-8918	256	3	41	41	NUM
easat-8918	256	4	]	]	X
easat-8918	256	5	v.	v.	ADP
easat-8918	256	6	jain	jain	PROPN
easat-8918	256	7	and	and	CCONJ
easat-8918	256	8	s.	s.	PROPN
easat-8918	256	9	seung	seung	PROPN
easat-8918	256	10	,	,	PUNCT
easat-8918	256	11	"	"	PUNCT
easat-8918	256	12	natural	natural	ADJ
easat-8918	256	13	image	image	NOUN
easat-8918	256	14	denoising	denoise	VERB
easat-8918	256	15	with	with	ADP
easat-8918	256	16	convolutional	convolutional	ADJ
easat-8918	256	17	networks	network	NOUN
easat-8918	256	18	,	,	PUNCT
easat-8918	256	19	"	"	PUNCT
easat-8918	256	20	advances	advance	NOUN
easat-8918	256	21	in	in	ADP
easat-8918	256	22	neural	neural	ADJ
easat-8918	256	23	information	information	NOUN
easat-8918	256	24	processing	processing	NOUN
easat-8918	256	25	systems	system	NOUN
easat-8918	256	26	,	,	PUNCT
easat-8918	256	27	vol	vol	NOUN
easat-8918	256	28	.	.	PROPN
easat-8918	256	29	21	21	NUM
easat-8918	256	30	,	,	PUNCT
easat-8918	256	31	pp	pp	ADJ
easat-8918	256	32	.	.	PUNCT
easat-8918	257	1	769–776	769–776	NUM
easat-8918	257	2	,	,	PUNCT
easat-8918	257	3	2009	2009	NUM
easat-8918	257	4	.	.	PUNCT
easat-8918	258	1	[	[	X
easat-8918	258	2	42	42	NUM
easat-8918	258	3	]	]	PUNCT
easat-8918	258	4	x.	x.	PROPN
easat-8918	258	5	lan	lan	PROPN
easat-8918	258	6	,	,	PUNCT
easat-8918	258	7	s.	s.	PROPN
easat-8918	258	8	roth	roth	PROPN
easat-8918	258	9	,	,	PUNCT
easat-8918	258	10	d.	d.	PROPN
easat-8918	258	11	huttenlocher	huttenlocher	ADV
easat-8918	258	12	,	,	PUNCT
easat-8918	258	13	and	and	CCONJ
easat-8918	258	14	m.	m.	PROPN
easat-8918	258	15	j.	j.	PROPN
easat-8918	258	16	black	black	PROPN
easat-8918	258	17	,	,	PUNCT
easat-8918	258	18	"	"	PUNCT
easat-8918	258	19	efficient	efficient	ADJ
easat-8918	258	20	belief	belief	NOUN
easat-8918	258	21	propagation	propagation	NOUN
easat-8918	258	22	with	with	ADP
easat-8918	258	23	learned	learn	VERB
easat-8918	258	24	higher	high	ADJ
easat-8918	258	25	-	-	PUNCT
easat-8918	258	26	order	order	NOUN
easat-8918	258	27	markov	markov	NOUN
easat-8918	258	28	random	random	ADJ
easat-8918	258	29	fields	field	NOUN
easat-8918	258	30	,	,	PUNCT
easat-8918	258	31	"	"	PUNCT
easat-8918	258	32	in	in	ADP
easat-8918	258	33	proceedings	proceeding	NOUN
easat-8918	258	34	of	of	ADP
easat-8918	258	35	the	the	DET
easat-8918	258	36	european	european	PROPN
easat-8918	258	37	conference	conference	PROPN
easat-8918	258	38	on	on	ADP
easat-8918	258	39	computer	computer	NOUN
easat-8918	258	40	vision	vision	NOUN
easat-8918	258	41	,	,	PUNCT
easat-8918	258	42	2006	2006	NUM
easat-8918	258	43	.	.	PUNCT
easat-8918	259	1	[	[	X
easat-8918	259	2	43	43	NUM
easat-8918	259	3	]	]	X
easat-8918	259	4	h.	h.	PROPN
easat-8918	259	5	c.	c.	PROPN
easat-8918	259	6	burger	burger	PROPN
easat-8918	259	7	,	,	PUNCT
easat-8918	259	8	c.	c.	PROPN
easat-8918	259	9	j.	j.	PROPN
easat-8918	259	10	schuler	schuler	PROPN
easat-8918	259	11	,	,	PUNCT
easat-8918	259	12	and	and	CCONJ
easat-8918	259	13	s.	s.	PROPN
easat-8918	259	14	harmeling	harmeling	PROPN
easat-8918	259	15	,	,	PUNCT
easat-8918	259	16	"	"	PUNCT
easat-8918	259	17	image	image	NOUN
easat-8918	259	18	denoising	denoising	NOUN
easat-8918	259	19	:	:	PUNCT
easat-8918	259	20	can	can	AUX
easat-8918	259	21	plain	plain	VERB
easat-8918	259	22	neural	neural	ADJ
easat-8918	259	23	networks	network	NOUN
easat-8918	259	24	compete	compete	VERB
easat-8918	259	25	with	with	ADP
easat-8918	259	26	bm3d	bm3d	NOUN
easat-8918	259	27	?	?	PUNCT
easat-8918	259	28	,	,	PUNCT
easat-8918	259	29	"	"	PUNCT
easat-8918	259	30	in	in	ADP
easat-8918	259	31	conference	conference	NOUN
easat-8918	259	32	on	on	ADP
easat-8918	259	33	computer	computer	NOUN
easat-8918	259	34	vision	vision	NOUN
easat-8918	259	35	and	and	CCONJ
easat-8918	259	36	pattern	pattern	NOUN
easat-8918	259	37	recognition	recognition	NOUN
easat-8918	259	38	2012	2012	NUM
easat-8918	259	39	.	.	PUNCT
easat-8918	260	1	[	[	X
easat-8918	260	2	44	44	NUM
easat-8918	260	3	]	]	PUNCT
easat-8918	260	4	j.	j.	PROPN
easat-8918	260	5	xie	xie	PROPN
easat-8918	260	6	,	,	PUNCT
easat-8918	260	7	l.	l.	PROPN
easat-8918	260	8	xu	xu	PROPN
easat-8918	260	9	,	,	PUNCT
easat-8918	260	10	and	and	CCONJ
easat-8918	260	11	e.	e.	PROPN
easat-8918	260	12	chen	chen	PROPN
easat-8918	260	13	,	,	PUNCT
easat-8918	260	14	"	"	PUNCT
easat-8918	260	15	image	image	NOUN
easat-8918	260	16	denoising	denoising	NOUN
easat-8918	260	17	and	and	CCONJ
easat-8918	260	18	inpainting	inpainte	VERB
easat-8918	260	19	with	with	ADP
easat-8918	260	20	deep	deep	ADJ
easat-8918	260	21	neural	neural	ADJ
easat-8918	260	22	networks	network	NOUN
easat-8918	260	23	:	:	PUNCT
easat-8918	260	24	advances	advance	NOUN
easat-8918	260	25	in	in	ADP
easat-8918	260	26	neural	neural	ADJ
easat-8918	260	27	"	"	PUNCT
easat-8918	260	28	in	in	ADP
easat-8918	260	29	information	information	NOUN
easat-8918	260	30	processing	processing	NOUN
easat-8918	260	31	systems	system	NOUN
easat-8918	260	32	,	,	PUNCT
easat-8918	260	33	2012	2012	NUM
easat-8918	260	34	.	.	PUNCT
easat-8918	261	1	[	[	X
easat-8918	261	2	45	45	NUM
easat-8918	261	3	]	]	PUNCT
easat-8918	261	4	m.	m.	NOUN
easat-8918	261	5	elad	elad	PROPN
easat-8918	261	6	and	and	CCONJ
easat-8918	261	7	m.	m.	NOUN
easat-8918	261	8	aharon	aharon	PROPN
easat-8918	261	9	,	,	PUNCT
easat-8918	261	10	"	"	PUNCT
easat-8918	261	11	image	image	NOUN
easat-8918	261	12	denoising	denoise	VERB
easat-8918	261	13	via	via	ADP
easat-8918	261	14	sparse	sparse	ADJ
easat-8918	261	15	and	and	CCONJ
easat-8918	261	16	redundant	redundant	ADJ
easat-8918	261	17	representations	representation	NOUN
easat-8918	261	18	over	over	ADP
easat-8918	261	19	learned	learn	VERB
easat-8918	261	20	dictionaries	dictionary	NOUN
easat-8918	261	21	,	,	PUNCT
easat-8918	261	22	"	"	PUNCT
easat-8918	261	23	institute	institute	NOUN
easat-8918	261	24	of	of	ADP
easat-8918	261	25	electrical	electrical	ADJ
easat-8918	261	26	and	and	CCONJ
easat-8918	261	27	electronics	electronic	NOUN
easat-8918	261	28	engineers	engineer	NOUN
easat-8918	261	29	.	.	PUNCT
easat-8918	262	1	transactions	transaction	NOUN
easat-8918	262	2	on	on	ADP
easat-8918	262	3	image	image	NOUN
easat-8918	262	4	processing	processing	NOUN
easat-8918	262	5	,	,	PUNCT
easat-8918	262	6	vol	vol	NOUN
easat-8918	262	7	.	.	PROPN
easat-8918	262	8	15	15	NUM
easat-8918	262	9	,	,	PUNCT
easat-8918	262	10	no	no	INTJ
easat-8918	262	11	.	.	NOUN
easat-8918	262	12	12	12	NUM
easat-8918	262	13	,	,	PUNCT
easat-8918	262	14	pp	pp	ADJ
easat-8918	262	15	.	.	PUNCT
easat-8918	263	1	3736	3736	NUM
easat-8918	263	2	-	-	SYM
easat-8918	263	3	3745	3745	NUM
easat-8918	263	4	,	,	PUNCT
easat-8918	263	5	2006	2006	NUM
easat-8918	263	6	.	.	PUNCT
easat-8918	264	1	[	[	X
easat-8918	264	2	46	46	NUM
easat-8918	264	3	]	]	X
easat-8918	264	4	y.	y.	PROPN
easat-8918	264	5	chen	chen	PROPN
easat-8918	264	6	and	and	CCONJ
easat-8918	264	7	t.	t.	NOUN
easat-8918	264	8	pock	pock	NOUN
easat-8918	264	9	,	,	PUNCT
easat-8918	264	10	"	"	PUNCT
easat-8918	264	11	trainable	trainable	ADJ
easat-8918	264	12	nonlinear	nonlinear	ADJ
easat-8918	264	13	reaction	reaction	NOUN
easat-8918	264	14	diffusion	diffusion	NOUN
easat-8918	264	15	:	:	PUNCT
easat-8918	264	16	a	a	DET
easat-8918	264	17	flexible	flexible	ADJ
easat-8918	264	18	framework	framework	NOUN
easat-8918	264	19	for	for	ADP
easat-8918	264	20	fast	fast	ADJ
easat-8918	264	21	and	and	CCONJ
easat-8918	264	22	effective	effective	ADJ
easat-8918	264	23	image	image	NOUN
easat-8918	264	24	restoration	restoration	NOUN
easat-8918	264	25	,	,	PUNCT
easat-8918	264	26	"	"	PUNCT
easat-8918	264	27	institute	institute	NOUN
easat-8918	264	28	of	of	ADP
easat-8918	264	29	electrical	electrical	ADJ
easat-8918	264	30	and	and	CCONJ
easat-8918	264	31	electronics	electronic	NOUN
easat-8918	264	32	engineers	engineer	NOUN
easat-8918	264	33	.	.	PUNCT
easat-8918	265	1	transactions	transaction	NOUN
easat-8918	265	2	on	on	ADP
easat-8918	265	3	pattern	pattern	NOUN
easat-8918	265	4	analysis	analysis	NOUN
easat-8918	265	5	and	and	CCONJ
easat-8918	265	6	machine	machine	NOUN
easat-8918	265	7	intelligence	intelligence	NOUN
easat-8918	265	8	,	,	PUNCT
easat-8918	265	9	vol	vol	NOUN
easat-8918	265	10	.	.	PROPN
easat-8918	265	11	39	39	NUM
easat-8918	265	12	,	,	PUNCT
easat-8918	265	13	no	no	INTJ
easat-8918	265	14	.	.	NOUN
easat-8918	265	15	6	6	NUM
easat-8918	265	16	,	,	PUNCT
easat-8918	265	17	pp	pp	ADJ
easat-8918	265	18	.	.	PUNCT
easat-8918	266	1	1256	1256	NUM
easat-8918	266	2	-	-	SYM
easat-8918	266	3	1272	1272	NUM
easat-8918	266	4	,	,	PUNCT
easat-8918	266	5	2016	2016	NUM
easat-8918	266	6	.	.	PUNCT
easat-8918	267	1	[	[	X
easat-8918	267	2	47	47	NUM
easat-8918	267	3	]	]	PUNCT
easat-8918	267	4	k.	k.	PROPN
easat-8918	267	5	dabov	dabov	PROPN
easat-8918	267	6	,	,	PUNCT
easat-8918	267	7	a.	a.	PROPN
easat-8918	267	8	foi	foi	PROPN
easat-8918	267	9	,	,	PUNCT
easat-8918	267	10	v.	v.	CCONJ
easat-8918	267	11	katkovnik	katkovnik	X
easat-8918	267	12	,	,	PUNCT
easat-8918	267	13	and	and	CCONJ
easat-8918	267	14	k.	k.	PROPN
easat-8918	267	15	egiazarian	egiazarian	PROPN
easat-8918	267	16	,	,	PUNCT
easat-8918	267	17	"	"	PUNCT
easat-8918	267	18	image	image	NOUN
easat-8918	267	19	denoising	denoise	VERB
easat-8918	267	20	by	by	ADP
easat-8918	267	21	sparse	sparse	ADJ
easat-8918	267	22	3	3	NUM
easat-8918	267	23	-	-	PUNCT
easat-8918	267	24	d	d	NOUN
easat-8918	267	25	transform	transform	NOUN
easat-8918	267	26	-	-	PUNCT
easat-8918	267	27	domain	domain	NOUN
easat-8918	267	28	collaborative	collaborative	ADJ
easat-8918	267	29	filtering	filtering	NOUN
easat-8918	267	30	,	,	PUNCT
easat-8918	267	31	"	"	PUNCT
easat-8918	267	32	institute	institute	NOUN
easat-8918	267	33	of	of	ADP
easat-8918	267	34	electrical	electrical	ADJ
easat-8918	267	35	and	and	CCONJ
easat-8918	267	36	electronics	electronic	NOUN
easat-8918	267	37	engineers.transactions	engineers.transaction	NOUN
easat-8918	267	38	on	on	ADP
easat-8918	267	39	image	image	NOUN
easat-8918	267	40	processing	processing	NOUN
easat-8918	267	41	,	,	PUNCT
easat-8918	267	42	vol	vol	NOUN
easat-8918	267	43	.	.	PROPN
easat-8918	267	44	16	16	NUM
easat-8918	267	45	,	,	PUNCT
easat-8918	267	46	no	no	INTJ
easat-8918	267	47	.	.	NOUN
easat-8918	267	48	8	8	NUM
easat-8918	267	49	,	,	PUNCT
easat-8918	267	50	pp	pp	ADJ
easat-8918	267	51	.	.	PUNCT
easat-8918	267	52	2080	2080	NUM
easat-8918	267	53	-	-	SYM
easat-8918	267	54	2095	2095	NUM
easat-8918	267	55	,	,	PUNCT
easat-8918	267	56	2007	2007	NUM
easat-8918	267	57	.	.	PUNCT
easat-8918	268	1	[	[	X
easat-8918	268	2	48	48	NUM
easat-8918	268	3	]	]	PUNCT
easat-8918	268	4	l.	l.	PROPN
easat-8918	268	5	chen	chen	PROPN
easat-8918	268	6	,	,	PUNCT
easat-8918	268	7	x.	x.	PROPN
easat-8918	268	8	chu	chu	PROPN
easat-8918	268	9	,	,	PUNCT
easat-8918	268	10	x.	x.	PROPN
easat-8918	268	11	zhang	zhang	PROPN
easat-8918	268	12	,	,	PUNCT
easat-8918	268	13	and	and	CCONJ
easat-8918	268	14	j.	j.	PROPN
easat-8918	268	15	sun	sun	PROPN
easat-8918	268	16	,	,	PUNCT
easat-8918	268	17	"	"	PUNCT
easat-8918	268	18	simple	simple	ADJ
easat-8918	268	19	baselines	baseline	NOUN
easat-8918	268	20	for	for	ADP
easat-8918	268	21	image	image	NOUN
easat-8918	268	22	restoration	restoration	NOUN
easat-8918	268	23	,	,	PUNCT
easat-8918	268	24	"	"	PUNCT
easat-8918	268	25	in	in	ADP
easat-8918	268	26	proceedings	proceeding	NOUN
easat-8918	268	27	of	of	ADP
easat-8918	268	28	the	the	DET
easat-8918	268	29	european	european	PROPN
easat-8918	268	30	conference	conference	PROPN
easat-8918	268	31	on	on	ADP
easat-8918	268	32	computer	computer	NOUN
easat-8918	268	33	vision	vision	NOUN
easat-8918	268	34	,	,	PUNCT
easat-8918	268	35	2022	2022	NUM
easat-8918	268	36	.	.	PUNCT
easat-8918	269	1	[	[	X
easat-8918	269	2	49	49	NUM
easat-8918	269	3	]	]	PUNCT
easat-8918	269	4	j.	j.	PROPN
easat-8918	269	5	chen	chen	PROPN
easat-8918	269	6	,	,	PUNCT
easat-8918	269	7	j.	j.	PROPN
easat-8918	269	8	chen	chen	PROPN
easat-8918	269	9	,	,	PUNCT
easat-8918	269	10	h.	h.	PROPN
easat-8918	269	11	chao	chao	PROPN
easat-8918	269	12	,	,	PUNCT
easat-8918	269	13	and	and	CCONJ
easat-8918	269	14	m.	m.	PROPN
easat-8918	269	15	yang	yang	PROPN
easat-8918	269	16	,	,	PUNCT
easat-8918	269	17	"	"	PUNCT
easat-8918	269	18	image	image	NOUN
easat-8918	269	19	blind	blind	ADJ
easat-8918	269	20	denoising	denoising	NOUN
easat-8918	269	21	with	with	ADP
easat-8918	269	22	generative	generative	ADJ
easat-8918	269	23	adversarial	adversarial	ADJ
easat-8918	269	24	network	network	NOUN
easat-8918	269	25	based	base	VERB
easat-8918	269	26	noise	noise	NOUN
easat-8918	269	27	modeling	modeling	NOUN
easat-8918	269	28	,	,	PUNCT
easat-8918	269	29	"	"	PUNCT
easat-8918	269	30	in	in	ADP
easat-8918	269	31	proceedings	proceeding	NOUN
easat-8918	269	32	of	of	ADP
easat-8918	269	33	the	the	DET
easat-8918	269	34	ieee	ieee	NOUN
easat-8918	269	35	conference	conference	NOUN
easat-8918	269	36	on	on	ADP
easat-8918	269	37	computer	computer	NOUN
easat-8918	269	38	vision	vision	NOUN
easat-8918	269	39	and	and	CCONJ
easat-8918	269	40	pattern	pattern	NOUN
easat-8918	269	41	recognition	recognition	NOUN
easat-8918	269	42	,	,	PUNCT
easat-8918	269	43	2018	2018	NUM
easat-8918	269	44	.	.	PUNCT
easat-8918	270	1	[	[	X
easat-8918	270	2	50	50	NUM
easat-8918	270	3	]	]	X
easat-8918	270	4	y.	y.	PROPN
easat-8918	270	5	cai	cai	PROPN
easat-8918	270	6	,	,	PUNCT
easat-8918	270	7	x.	x.	PROPN
easat-8918	270	8	hu	hu	PROPN
easat-8918	270	9	,	,	PUNCT
easat-8918	270	10	h.	h.	PROPN
easat-8918	270	11	wang	wang	PROPN
easat-8918	270	12	,	,	PUNCT
easat-8918	270	13	y.	y.	PROPN
easat-8918	270	14	zhang	zhang	PROPN
easat-8918	270	15	,	,	PUNCT
easat-8918	270	16	h.	h.	PROPN
easat-8918	270	17	pfister	pfister	PROPN
easat-8918	270	18	,	,	PUNCT
easat-8918	270	19	and	and	CCONJ
easat-8918	270	20	d.	d.	PROPN
easat-8918	270	21	wei	wei	PROPN
easat-8918	270	22	,	,	PUNCT
easat-8918	270	23	"	"	PUNCT
easat-8918	270	24	learning	learn	VERB
easat-8918	270	25	to	to	PART
easat-8918	270	26	generate	generate	VERB
easat-8918	270	27	realistic	realistic	ADJ
easat-8918	270	28	noisy	noisy	ADJ
easat-8918	270	29	images	image	NOUN
easat-8918	270	30	via	via	ADP
easat-8918	270	31	pixellevel	pixellevel	NOUN
easat-8918	270	32	noise	noise	NOUN
easat-8918	270	33	-	-	PUNCT
easat-8918	270	34	aware	aware	ADJ
easat-8918	270	35	adversarial	adversarial	ADJ
easat-8918	270	36	training	training	NOUN
easat-8918	270	37	,	,	PUNCT
easat-8918	270	38	"	"	PUNCT
easat-8918	270	39	advances	advance	NOUN
easat-8918	270	40	in	in	ADP
easat-8918	270	41	neural	neural	ADJ
easat-8918	270	42	information	information	NOUN
easat-8918	270	43	processing	processing	NOUN
easat-8918	270	44	systems	system	NOUN
easat-8918	270	45	,	,	PUNCT
easat-8918	270	46	vol	vol	NOUN
easat-8918	270	47	.	.	PROPN
easat-8918	271	1	34	34	NUM
easat-8918	271	2	,	,	PUNCT
easat-8918	271	3	pp	pp	ADJ
easat-8918	271	4	.	.	PUNCT
easat-8918	272	1	3259	3259	NUM
easat-8918	272	2	-	-	SYM
easat-8918	272	3	3270	3270	NUM
easat-8918	272	4	,	,	PUNCT
easat-8918	272	5	2021	2021	NUM
easat-8918	272	6	.	.	PUNCT
easat-8918	273	1	[	[	X
easat-8918	273	2	51	51	NUM
easat-8918	273	3	]	]	PUNCT
easat-8918	273	4	k.	k.	PROPN
easat-8918	273	5	zhang	zhang	PROPN
easat-8918	273	6	,	,	PUNCT
easat-8918	273	7	w.	w.	PROPN
easat-8918	273	8	zuo	zuo	PROPN
easat-8918	273	9	,	,	PUNCT
easat-8918	273	10	y.	y.	PROPN
easat-8918	273	11	chen	chen	PROPN
easat-8918	273	12	,	,	PUNCT
easat-8918	273	13	d.	d.	PROPN
easat-8918	273	14	meng	meng	PROPN
easat-8918	273	15	,	,	PUNCT
easat-8918	273	16	and	and	CCONJ
easat-8918	273	17	l.	l.	PROPN
easat-8918	273	18	zhang	zhang	PROPN
easat-8918	273	19	,	,	PUNCT
easat-8918	273	20	"	"	PUNCT
easat-8918	273	21	beyond	beyond	ADP
easat-8918	273	22	a	a	DET
easat-8918	273	23	gaussian	gaussian	ADJ
easat-8918	273	24	denoiser	denoiser	NOUN
easat-8918	273	25	:	:	PUNCT
easat-8918	273	26	residual	residual	ADJ
easat-8918	273	27	learning	learning	NOUN
easat-8918	273	28	of	of	ADP
easat-8918	273	29	deep	deep	ADJ
easat-8918	273	30	cnn	cnn	PROPN
easat-8918	273	31	for	for	ADP
easat-8918	273	32	image	image	NOUN
easat-8918	273	33	denoising	denoising	NOUN
easat-8918	273	34	,	,	PUNCT
easat-8918	273	35	"	"	PUNCT
easat-8918	273	36	institute	institute	NOUN
easat-8918	273	37	of	of	ADP
easat-8918	273	38	electrical	electrical	ADJ
easat-8918	273	39	and	and	CCONJ
easat-8918	273	40	electronics	electronic	NOUN
easat-8918	273	41	engineers	engineer	NOUN
easat-8918	273	42	.	.	PUNCT
easat-8918	274	1	transactions	transaction	NOUN
easat-8918	274	2	on	on	ADP
easat-8918	274	3	image	image	NOUN
easat-8918	274	4	processing	processing	NOUN
easat-8918	274	5	,	,	PUNCT
easat-8918	274	6	vol	vol	NOUN
easat-8918	274	7	.	.	PROPN
easat-8918	274	8	26	26	NUM
easat-8918	274	9	,	,	PUNCT
easat-8918	274	10	no	no	INTJ
easat-8918	274	11	.	.	NOUN
easat-8918	274	12	7	7	NUM
easat-8918	274	13	,	,	PUNCT
easat-8918	274	14	pp	pp	ADJ
easat-8918	274	15	.	.	PUNCT
easat-8918	274	16	3142	3142	NUM
easat-8918	274	17	-	-	SYM
easat-8918	274	18	3155	3155	NUM
easat-8918	274	19	,	,	PUNCT
easat-8918	274	20	2017	2017	NUM
easat-8918	274	21	.	.	PUNCT
easat-8918	275	1	[	[	X
easat-8918	275	2	52	52	NUM
easat-8918	275	3	]	]	PUNCT
easat-8918	275	4	s.	s.	PROPN
easat-8918	275	5	w.	w.	PROPN
easat-8918	275	6	zamir	zamir	PROPN
easat-8918	275	7	,	,	PUNCT
easat-8918	275	8	a.	a.	PROPN
easat-8918	275	9	arora	arora	PROPN
easat-8918	275	10	,	,	PUNCT
easat-8918	275	11	s.	s.	PROPN
easat-8918	275	12	khan	khan	PROPN
easat-8918	275	13	,	,	PUNCT
easat-8918	275	14	m.	m.	NOUN
easat-8918	275	15	hayat	hayat	PROPN
easat-8918	275	16	,	,	PUNCT
easat-8918	275	17	f.	f.	PROPN
easat-8918	275	18	s.	s.	PROPN
easat-8918	275	19	khan	khan	PROPN
easat-8918	275	20	,	,	PUNCT
easat-8918	275	21	and	and	CCONJ
easat-8918	275	22	m.	m.	PROPN
easat-8918	275	23	yang	yang	PROPN
easat-8918	275	24	,	,	PUNCT
easat-8918	275	25	"	"	PUNCT
easat-8918	275	26	restormer	restormer	ADJ
easat-8918	275	27	:	:	PUNCT
easat-8918	275	28	efficient	efficient	ADJ
easat-8918	275	29	transformer	transformer	NOUN
easat-8918	275	30	for	for	ADP
easat-8918	275	31	highresolution	highresolution	NOUN
easat-8918	275	32	image	image	NOUN
easat-8918	275	33	restoration	restoration	NOUN
easat-8918	275	34	,	,	PUNCT
easat-8918	275	35	"	"	PUNCT
easat-8918	275	36	in	in	ADP
easat-8918	275	37	proceedings	proceeding	NOUN
easat-8918	275	38	of	of	ADP
easat-8918	275	39	the	the	DET
easat-8918	275	40	2022	2022	NUM
easat-8918	275	41	ieee	ieee	NOUN
easat-8918	275	42	/	/	SYM
easat-8918	275	43	cvf	cvf	NOUN
easat-8918	275	44	conference	conference	NOUN
easat-8918	275	45	on	on	ADP
easat-8918	275	46	computer	computer	NOUN
easat-8918	275	47	vision	vision	NOUN
easat-8918	275	48	and	and	CCONJ
easat-8918	275	49	pattern	pattern	NOUN
easat-8918	275	50	recognition	recognition	NOUN
easat-8918	275	51	(	(	PUNCT
easat-8918	275	52	cvpr	cvpr	NOUN
easat-8918	275	53	)	)	PUNCT
easat-8918	275	54	,	,	PUNCT
easat-8918	275	55	2022	2022	NUM
easat-8918	275	56	.	.	PUNCT
easat-8918	276	1	[	[	X
easat-8918	276	2	53	53	NUM
easat-8918	276	3	]	]	PUNCT
easat-8918	276	4	s.	s.	PROPN
easat-8918	276	5	ioffe	ioffe	PROPN
easat-8918	276	6	and	and	CCONJ
easat-8918	276	7	c.	c.	PROPN
easat-8918	276	8	szegedy	szegedy	PROPN
easat-8918	276	9	,	,	PUNCT
easat-8918	276	10	"	"	PUNCT
easat-8918	276	11	batch	batch	VERB
easat-8918	276	12	normalization	normalization	NOUN
easat-8918	276	13	:	:	PUNCT
easat-8918	276	14	accelerating	accelerate	VERB
easat-8918	276	15	deep	deep	ADJ
easat-8918	276	16	network	network	NOUN
easat-8918	276	17	training	training	NOUN
easat-8918	276	18	by	by	ADP
easat-8918	276	19	reducing	reduce	VERB
easat-8918	276	20	internal	internal	ADJ
easat-8918	276	21	covariate	covariate	ADJ
easat-8918	276	22	shift	shift	NOUN
easat-8918	276	23	,	,	PUNCT
easat-8918	276	24	"	"	PUNCT
easat-8918	276	25	in	in	ADP
easat-8918	276	26	proceedings	proceeding	NOUN
easat-8918	276	27	of	of	ADP
easat-8918	276	28	the	the	DET
easat-8918	276	29	international	international	ADJ
easat-8918	276	30	conference	conference	NOUN
easat-8918	276	31	on	on	ADP
easat-8918	276	32	machine	machine	NOUN
easat-8918	276	33	learning	learning	NOUN
easat-8918	276	34	,	,	PUNCT
easat-8918	276	35	2015	2015	NUM
easat-8918	276	36	.	.	PUNCT
easat-8918	277	1	[	[	X
easat-8918	277	2	54	54	NUM
easat-8918	277	3	]	]	PUNCT
easat-8918	277	4	a.	a.	NOUN
easat-8918	277	5	dosovitskiy	dosovitskiy	NOUN
easat-8918	277	6	et	et	PROPN
easat-8918	277	7	al	al	PROPN
easat-8918	277	8	.	.	PROPN
easat-8918	277	9	,	,	PUNCT
easat-8918	277	10	"	"	PUNCT
easat-8918	277	11	an	an	DET
easat-8918	277	12	image	image	NOUN
easat-8918	277	13	is	be	AUX
easat-8918	277	14	worth	worth	ADJ
easat-8918	277	15	16x16	16x16	NUM
easat-8918	277	16	words	word	NOUN
easat-8918	277	17	:	:	PUNCT
easat-8918	277	18	transformers	transformer	NOUN
easat-8918	277	19	for	for	ADP
easat-8918	277	20	image	image	NOUN
easat-8918	277	21	recognition	recognition	NOUN
easat-8918	277	22	at	at	ADP
easat-8918	277	23	scale	scale	NOUN
easat-8918	277	24	,	,	PUNCT
easat-8918	277	25	"	"	PUNCT
easat-8918	277	26	arxiv	arxiv	PROPN
easat-8918	277	27	preprint	preprint	NOUN
easat-8918	277	28	arxiv:2010.11929	arxiv:2010.11929	NOUN
easat-8918	277	29	,	,	PUNCT
easat-8918	277	30	2020	2020	NUM
easat-8918	277	31	.	.	PUNCT
easat-8918	278	1	[	[	X
easat-8918	278	2	55	55	NUM
easat-8918	278	3	]	]	PUNCT
easat-8918	278	4	h.	h.	PROPN
easat-8918	278	5	zhang	zhang	PROPN
easat-8918	278	6	,	,	PUNCT
easat-8918	278	7	i.	i.	PROPN
easat-8918	278	8	goodfellow	goodfellow	PROPN
easat-8918	278	9	,	,	PUNCT
easat-8918	278	10	d.	d.	PROPN
easat-8918	278	11	metaxas	metaxas	PROPN
easat-8918	278	12	,	,	PUNCT
easat-8918	278	13	and	and	CCONJ
easat-8918	278	14	a.	a.	NOUN
easat-8918	278	15	odena	odena	NOUN
easat-8918	278	16	,	,	PUNCT
easat-8918	278	17	"	"	PUNCT
easat-8918	278	18	self	self	NOUN
easat-8918	278	19	-	-	PUNCT
easat-8918	278	20	attention	attention	NOUN
easat-8918	278	21	generative	generative	ADJ
easat-8918	278	22	adversarial	adversarial	ADJ
easat-8918	278	23	networks	network	NOUN
easat-8918	278	24	,	,	PUNCT
easat-8918	278	25	"	"	PUNCT
easat-8918	278	26	in	in	ADP
easat-8918	278	27	proceedings	proceeding	NOUN
easat-8918	278	28	of	of	ADP
easat-8918	278	29	the	the	DET
easat-8918	278	30	international	international	ADJ
easat-8918	278	31	conference	conference	NOUN
easat-8918	278	32	on	on	ADP
easat-8918	278	33	machine	machine	NOUN
easat-8918	278	34	learning	learning	NOUN
easat-8918	278	35	,	,	PUNCT
easat-8918	278	36	2019	2019	NUM
easat-8918	278	37	.	.	PUNCT
easat-8918	279	1	[	[	X
easat-8918	279	2	56	56	NUM
easat-8918	279	3	]	]	X
easat-8918	279	4	s.	s.	PROPN
easat-8918	279	5	khan	khan	PROPN
easat-8918	279	6	,	,	PUNCT
easat-8918	279	7	m.	m.	NOUN
easat-8918	279	8	naseer	naseer	PROPN
easat-8918	279	9	,	,	PUNCT
easat-8918	279	10	m.	m.	NOUN
easat-8918	279	11	hayat	hayat	PROPN
easat-8918	279	12	,	,	PUNCT
easat-8918	279	13	s.	s.	PROPN
easat-8918	279	14	w.	w.	PROPN
easat-8918	279	15	zamir	zamir	PROPN
easat-8918	279	16	,	,	PUNCT
easat-8918	279	17	f.	f.	PROPN
easat-8918	279	18	s.	s.	PROPN
easat-8918	279	19	khan	khan	PROPN
easat-8918	279	20	,	,	PUNCT
easat-8918	279	21	and	and	CCONJ
easat-8918	279	22	m.	m.	NOUN
easat-8918	279	23	shah	shah	NOUN
easat-8918	279	24	,	,	PUNCT
easat-8918	279	25	"	"	PUNCT
easat-8918	279	26	transformers	transformer	NOUN
easat-8918	279	27	in	in	ADP
easat-8918	279	28	vision	vision	NOUN
easat-8918	279	29	:	:	PUNCT
easat-8918	279	30	a	a	DET
easat-8918	279	31	survey	survey	NOUN
easat-8918	279	32	,	,	PUNCT
easat-8918	279	33	"	"	PUNCT
easat-8918	279	34	acm	acm	PROPN
easat-8918	279	35	computing	computing	NOUN
easat-8918	279	36	surveys	survey	NOUN
easat-8918	279	37	vol	vol	NOUN
easat-8918	279	38	.	.	PROPN
easat-8918	279	39	54	54	NUM
easat-8918	279	40	,	,	PUNCT
easat-8918	279	41	no	no	NOUN
easat-8918	279	42	.	.	PUNCT
easat-8918	279	43	10s	10	NOUN
easat-8918	279	44	,	,	PUNCT
easat-8918	279	45	pp	pp	ADJ
easat-8918	279	46	.	.	PUNCT
easat-8918	280	1	1	1	NUM
easat-8918	280	2	-	-	SYM
easat-8918	280	3	41	41	NUM
easat-8918	280	4	,	,	PUNCT
easat-8918	280	5	2022	2022	NUM
easat-8918	280	6	.	.	PUNCT
easat-8918	281	1	https://doi.org/10.1007/s11042-023-14482-2	https://doi.org/10.1007/s11042-023-14482-2	X
easat-8918	282	1	https://doi.org/10.1109/iceee62185.2024.10779262	https://doi.org/10.1109/iceee62185.2024.10779262	PROPN
easat-8918	282	2	https://doi.org/10.1016/j.knosys.2024.111691	https://doi.org/10.1016/j.knosys.2024.111691	PROPN
easat-8918	282	3	1359	1359	NUM
easat-8918	282	4	edelweiss	edelweiss	PROPN
easat-8918	282	5	applied	apply	VERB
easat-8918	282	6	science	science	NOUN
easat-8918	282	7	and	and	CCONJ
easat-8918	282	8	technology	technology	NOUN
easat-8918	282	9	issn	issn	PROPN
easat-8918	282	10	:	:	PUNCT
easat-8918	282	11	2576	2576	NUM
easat-8918	282	12	-	-	SYM
easat-8918	282	13	8484	8484	NUM
easat-8918	282	14	vol	vol	NOUN
easat-8918	282	15	.	.	PROPN
easat-8918	283	1	9	9	NUM
easat-8918	283	2	,	,	PUNCT
easat-8918	283	3	no	no	INTJ
easat-8918	283	4	.	.	NOUN
easat-8918	283	5	7	7	NUM
easat-8918	283	6	:	:	SYM
easat-8918	283	7	1345	1345	NUM
easat-8918	283	8	-	-	SYM
easat-8918	283	9	1359	1359	NUM
easat-8918	283	10	,	,	PUNCT
easat-8918	283	11	2025	2025	NUM
easat-8918	283	12	doi	doi	NOUN
easat-8918	283	13	:	:	PUNCT
easat-8918	283	14	10.55214/25768484.v9i7.8918	10.55214/25768484.v9i7.8918	PROPN
easat-8918	283	15	©	©	PROPN
easat-8918	283	16	2025	2025	NUM
easat-8918	283	17	by	by	ADP
easat-8918	283	18	the	the	DET
easat-8918	283	19	author	author	NOUN
easat-8918	283	20	;	;	PUNCT
easat-8918	283	21	licensee	licensee	PROPN
easat-8918	283	22	learning	learning	NOUN
easat-8918	283	23	gate	gate	NOUN
easat-8918	284	1	[	[	X
easat-8918	284	2	57	57	NUM
easat-8918	284	3	]	]	PUNCT
easat-8918	284	4	w.	w.	PROPN
easat-8918	284	5	fedus	fedus	PROPN
easat-8918	284	6	,	,	PUNCT
easat-8918	284	7	b.	b.	PROPN
easat-8918	284	8	zoph	zoph	PROPN
easat-8918	284	9	,	,	PUNCT
easat-8918	284	10	and	and	CCONJ
easat-8918	284	11	n.	n.	PROPN
easat-8918	284	12	shazeer	shazeer	NOUN
easat-8918	284	13	,	,	PUNCT
easat-8918	284	14	"	"	PUNCT
easat-8918	284	15	switch	switch	NOUN
easat-8918	284	16	transformers	transformer	NOUN
easat-8918	284	17	:	:	PUNCT
easat-8918	284	18	scaling	scale	VERB
easat-8918	284	19	to	to	ADP
easat-8918	284	20	trillion	trillion	NUM
easat-8918	284	21	parameter	parameter	NOUN
easat-8918	284	22	models	model	NOUN
easat-8918	284	23	with	with	ADP
easat-8918	284	24	simple	simple	ADJ
easat-8918	284	25	and	and	CCONJ
easat-8918	284	26	efficient	efficient	ADJ
easat-8918	284	27	sparsity	sparsity	NOUN
easat-8918	284	28	,	,	PUNCT
easat-8918	284	29	"	"	PUNCT
easat-8918	284	30	journal	journal	NOUN
easat-8918	284	31	of	of	ADP
easat-8918	284	32	machine	machine	NOUN
easat-8918	284	33	learning	learn	VERB
easat-8918	284	34	research	research	NOUN
easat-8918	284	35	,	,	PUNCT
easat-8918	284	36	vol	vol	NOUN
easat-8918	284	37	.	.	PROPN
easat-8918	284	38	23	23	NUM
easat-8918	284	39	,	,	PUNCT
easat-8918	284	40	no	no	INTJ
easat-8918	284	41	.	.	NOUN
easat-8918	284	42	120	120	NUM
easat-8918	284	43	,	,	PUNCT
easat-8918	284	44	pp	pp	ADJ
easat-8918	284	45	.	.	PUNCT
easat-8918	285	1	1	1	NUM
easat-8918	285	2	-	-	SYM
easat-8918	285	3	39	39	NUM
easat-8918	285	4	,	,	PUNCT
easat-8918	285	5	2022	2022	NUM
easat-8918	285	6	.	.	PUNCT
easat-8918	286	1	https://arxiv.org/abs/2101.03961	https://arxiv.org/abs/2101.03961	PROPN
easat-8918	286	2	[	[	X
easat-8918	286	3	58	58	NUM
easat-8918	286	4	]	]	X
easat-8918	286	5	n.	n.	NOUN
easat-8918	286	6	carion	carion	PROPN
easat-8918	286	7	,	,	PUNCT
easat-8918	286	8	f.	f.	PROPN
easat-8918	286	9	massa	massa	PROPN
easat-8918	286	10	,	,	PUNCT
easat-8918	286	11	g.	g.	PROPN
easat-8918	286	12	synnaeve	synnaeve	PROPN
easat-8918	286	13	,	,	PUNCT
easat-8918	286	14	n.	n.	PROPN
easat-8918	286	15	usunier	usunier	PROPN
easat-8918	286	16	,	,	PUNCT
easat-8918	286	17	a.	a.	NOUN
easat-8918	286	18	kirillov	kirillov	PROPN
easat-8918	286	19	,	,	PUNCT
easat-8918	286	20	and	and	CCONJ
easat-8918	286	21	s.	s.	PROPN
easat-8918	286	22	zagoruyko	zagoruyko	PROPN
easat-8918	286	23	,	,	PUNCT
easat-8918	286	24	"	"	PUNCT
easat-8918	286	25	end	end	NOUN
easat-8918	286	26	-	-	PUNCT
easat-8918	286	27	to	to	ADP
easat-8918	286	28	-	-	PUNCT
easat-8918	286	29	end	end	NOUN
easat-8918	286	30	object	object	NOUN
easat-8918	286	31	detection	detection	NOUN
easat-8918	286	32	with	with	ADP
easat-8918	286	33	transformers	transformer	NOUN
easat-8918	286	34	,	,	PUNCT
easat-8918	286	35	"	"	PUNCT
easat-8918	286	36	in	in	ADP
easat-8918	286	37	european	european	ADJ
easat-8918	286	38	conference	conference	NOUN
easat-8918	286	39	on	on	ADP
easat-8918	286	40	computer	computer	NOUN
easat-8918	286	41	vision	vision	NOUN
easat-8918	286	42	,	,	PUNCT
easat-8918	286	43	2020	2020	NUM
easat-8918	286	44	.	.	PUNCT
easat-8918	287	1	[	[	X
easat-8918	287	2	59	59	NUM
easat-8918	287	3	]	]	X
easat-8918	287	4	h.	h.	PROPN
easat-8918	287	5	touvron	touvron	PROPN
easat-8918	287	6	,	,	PUNCT
easat-8918	287	7	m.	m.	NOUN
easat-8918	287	8	cord	cord	NOUN
easat-8918	287	9	,	,	PUNCT
easat-8918	287	10	m.	m.	NOUN
easat-8918	287	11	douze	douze	PROPN
easat-8918	287	12	,	,	PUNCT
easat-8918	287	13	f.	f.	PROPN
easat-8918	287	14	massa	massa	PROPN
easat-8918	287	15	,	,	PUNCT
easat-8918	287	16	a.	a.	NOUN
easat-8918	287	17	sablayrolles	sablayrolle	NOUN
easat-8918	287	18	,	,	PUNCT
easat-8918	287	19	and	and	CCONJ
easat-8918	287	20	h.	h.	PROPN
easat-8918	287	21	jégou	jégou	PROPN
easat-8918	287	22	,	,	PUNCT
easat-8918	287	23	"	"	PUNCT
easat-8918	287	24	training	train	VERB
easat-8918	287	25	data	data	NOUN
easat-8918	287	26	-	-	PUNCT
easat-8918	287	27	efficient	efficient	ADJ
easat-8918	287	28	image	image	NOUN
easat-8918	287	29	transformers	transformer	NOUN
easat-8918	287	30	&	&	CCONJ
easat-8918	287	31	distillation	distillation	NOUN
easat-8918	287	32	through	through	ADP
easat-8918	287	33	attention	attention	NOUN
easat-8918	287	34	,	,	PUNCT
easat-8918	287	35	"	"	PUNCT
easat-8918	287	36	in	in	ADP
easat-8918	287	37	international	international	ADJ
easat-8918	287	38	conference	conference	NOUN
easat-8918	287	39	on	on	ADP
easat-8918	287	40	machine	machine	NOUN
easat-8918	287	41	learning	learning	NOUN
easat-8918	287	42	,	,	PUNCT
easat-8918	287	43	2021	2021	NUM
easat-8918	287	44	.	.	PUNCT
easat-8918	288	1	[	[	X
easat-8918	288	2	60	60	NUM
easat-8918	288	3	]	]	X
easat-8918	288	4	h.	h.	PROPN
easat-8918	288	5	chen	chen	PROPN
easat-8918	288	6	et	et	PROPN
easat-8918	288	7	al	al	PROPN
easat-8918	288	8	.	.	PROPN
easat-8918	288	9	,	,	PUNCT
easat-8918	288	10	"	"	PUNCT
easat-8918	288	11	pre	pre	ADJ
easat-8918	288	12	-	-	ADJ
easat-8918	288	13	trained	train	VERB
easat-8918	288	14	image	image	NOUN
easat-8918	288	15	processing	processing	NOUN
easat-8918	288	16	transformer	transformer	NOUN
easat-8918	288	17	,	,	PUNCT
easat-8918	288	18	"	"	PUNCT
easat-8918	288	19	in	in	ADP
easat-8918	288	20	proceedings	proceeding	NOUN
easat-8918	288	21	of	of	ADP
easat-8918	288	22	the	the	DET
easat-8918	288	23	ieee	ieee	NOUN
easat-8918	288	24	/	/	SYM
easat-8918	288	25	cvf	cvf	NOUN
easat-8918	288	26	conference	conference	NOUN
easat-8918	288	27	on	on	ADP
easat-8918	288	28	computer	computer	NOUN
easat-8918	288	29	vision	vision	NOUN
easat-8918	288	30	and	and	CCONJ
easat-8918	288	31	pattern	pattern	NOUN
easat-8918	288	32	recognition	recognition	NOUN
easat-8918	288	33	,	,	PUNCT
easat-8918	288	34	2021	2021	NUM
easat-8918	288	35	.	.	PUNCT
easat-8918	289	1	[	[	X
easat-8918	289	2	61	61	NUM
easat-8918	289	3	]	]	PUNCT
easat-8918	289	4	j.	j.	PROPN
easat-8918	289	5	liang	liang	PROPN
easat-8918	289	6	,	,	PUNCT
easat-8918	289	7	j.	j.	PROPN
easat-8918	289	8	cao	cao	PROPN
easat-8918	289	9	,	,	PUNCT
easat-8918	289	10	g.	g.	PROPN
easat-8918	289	11	sun	sun	PROPN
easat-8918	289	12	,	,	PUNCT
easat-8918	289	13	k.	k.	PROPN
easat-8918	289	14	zhang	zhang	PROPN
easat-8918	289	15	,	,	PUNCT
easat-8918	289	16	l.	l.	PROPN
easat-8918	289	17	van	van	PROPN
easat-8918	289	18	gool	gool	PROPN
easat-8918	289	19	,	,	PUNCT
easat-8918	289	20	and	and	CCONJ
easat-8918	289	21	r.	r.	PROPN
easat-8918	289	22	timofte	timofte	PROPN
easat-8918	289	23	,	,	PUNCT
easat-8918	289	24	"	"	PUNCT
easat-8918	289	25	swinir	swinir	NOUN
easat-8918	289	26	:	:	PUNCT
easat-8918	289	27	image	image	NOUN
easat-8918	289	28	restoration	restoration	NOUN
easat-8918	289	29	using	use	VERB
easat-8918	289	30	swin	swin	PROPN
easat-8918	289	31	transformer	transformer	PROPN
easat-8918	289	32	,	,	PUNCT
easat-8918	289	33	"	"	PUNCT
easat-8918	289	34	in	in	ADP
easat-8918	289	35	proceedings	proceeding	NOUN
easat-8918	289	36	of	of	ADP
easat-8918	289	37	the	the	DET
easat-8918	289	38	ieee	ieee	NOUN
easat-8918	289	39	/	/	SYM
easat-8918	289	40	cvf	cvf	NOUN
easat-8918	289	41	international	international	ADJ
easat-8918	289	42	conference	conference	NOUN
easat-8918	289	43	on	on	ADP
easat-8918	289	44	computer	computer	NOUN
easat-8918	289	45	vision	vision	NOUN
easat-8918	289	46	,	,	PUNCT
easat-8918	289	47	2021	2021	NUM
easat-8918	289	48	.	.	PUNCT
easat-8918	289	49	https://arxiv.org/abs/2101.03961	https://arxiv.org/abs/2101.03961	NOUN