id	sid	tid	token	lemma	pos
cana-2643	1	1	communications	communication	NOUN
cana-2643	1	2	on	on	ADP
cana-2643	1	3	applied	apply	VERB
cana-2643	1	4	nonlinear	nonlinear	ADJ
cana-2643	1	5	analysis	analysis	NOUN
cana-2643	1	6	issn	issn	NOUN
cana-2643	1	7	:	:	PUNCT
cana-2643	1	8	1074	1074	NUM
cana-2643	1	9	-	-	PUNCT
cana-2643	1	10	133x	133x	NUM
cana-2643	1	11	vol	vol	NOUN
cana-2643	1	12	32	32	NUM
cana-2643	1	13	no	no	NOUN
cana-2643	1	14	.	.	PUNCT
cana-2643	2	1	3s	3s	NUM
cana-2643	2	2	(	(	PUNCT
cana-2643	2	3	2025	2025	NUM
cana-2643	2	4	)	)	PUNCT
cana-2643	2	5	279	279	NUM
cana-2643	2	6	https://internationalpubls.com	https://internationalpubls.com	NUM
cana-2643	2	7	reinforced	reinforce	VERB
cana-2643	2	8	multi	multi	ADJ
cana-2643	2	9	-	-	ADJ
cana-2643	2	10	path	path	ADJ
cana-2643	2	11	graph	graph	NOUN
cana-2643	2	12	attention	attention	NOUN
cana-2643	2	13	network	network	NOUN
cana-2643	2	14	(	(	PUNCT
cana-2643	2	15	mpga	mpga	NOUN
cana-2643	2	16	-	-	PUNCT
cana-2643	2	17	rl	rl	NOUN
cana-2643	2	18	-	-	PUNCT
cana-2643	2	19	net	net	NOUN
cana-2643	2	20	)	)	PUNCT
cana-2643	2	21	for	for	ADP
cana-2643	2	22	adaptive	adaptive	ADJ
cana-2643	2	23	robotic	robotic	ADJ
cana-2643	2	24	vision	vision	NOUN
cana-2643	2	25	with	with	ADP
cana-2643	2	26	task	task	NOUN
cana-2643	2	27	-	-	PUNCT
cana-2643	2	28	specific	specific	ADJ
cana-2643	2	29	region	region	NOUN
cana-2643	2	30	prioritization	prioritization	NOUN
cana-2643	2	31	mr.b.rama	mr.b.rama	PROPN
cana-2643	2	32	subbaiah1	subbaiah1	NOUN
cana-2643	2	33	,	,	PUNCT
cana-2643	2	34	dr	dr	PROPN
cana-2643	2	35	p	p	PROPN
cana-2643	2	36	penchala	penchala	PROPN
cana-2643	2	37	prasad	prasad	PROPN
cana-2643	2	38	2	2	NUM
cana-2643	2	39	,	,	PUNCT
cana-2643	2	40	mr.g.vikram	mr.g.vikram	PROPN
cana-2643	2	41	chandra3	chandra3	NOUN
cana-2643	2	42	,	,	PUNCT
cana-2643	2	43	mr.m.suleman	mr.m.suleman	PROPN
cana-2643	2	44	basha4	basha4	NOUN
cana-2643	2	45	,	,	PUNCT
cana-2643	2	46	dr	dr	PROPN
cana-2643	2	47	p	p	PROPN
cana-2643	2	48	kiran	kiran	PROPN
cana-2643	2	49	rao5	rao5	PROPN
cana-2643	2	50	,	,	PUNCT
cana-2643	2	51	dr.r.senthil	dr.r.senthil	PROPN
cana-2643	2	52	ganesh	ganesh	NOUN
cana-2643	2	53	6	6	NUM
cana-2643	2	54	1assistant	1assistant	NUM
cana-2643	2	55	professor	professor	NOUN
cana-2643	2	56	,	,	PUNCT
cana-2643	2	57	department	department	NOUN
cana-2643	2	58	of	of	ADP
cana-2643	2	59	computer	computer	NOUN
cana-2643	2	60	science	science	PROPN
cana-2643	2	61	&	&	CCONJ
cana-2643	2	62	engineering	engineering	PROPN
cana-2643	2	63	,	,	PUNCT
cana-2643	2	64	rajeev	rajeev	PROPN
cana-2643	2	65	gandhi	gandhi	PROPN
cana-2643	2	66	memorial	memorial	PROPN
cana-2643	2	67	college	college	PROPN
cana-2643	2	68	of	of	ADP
cana-2643	2	69	engineering	engineering	PROPN
cana-2643	2	70	&	&	CCONJ
cana-2643	2	71	technology	technology	PROPN
cana-2643	2	72	,	,	PUNCT
cana-2643	2	73	nandyal	nandyal	PROPN
cana-2643	2	74	,	,	PUNCT
cana-2643	2	75	andhrapradesh	andhrapradesh	PROPN
cana-2643	2	76	,	,	PUNCT
cana-2643	2	77	india	india	PROPN
cana-2643	2	78	,	,	PUNCT
cana-2643	2	79	subhashrgmcet@gmail.com	subhashrgmcet@gmail.com	X
cana-2643	3	1	2associate	2associate	NUM
cana-2643	3	2	professor	professor	NOUN
cana-2643	3	3	,	,	PUNCT
cana-2643	3	4	department	department	NOUN
cana-2643	3	5	of	of	ADP
cana-2643	3	6	computer	computer	NOUN
cana-2643	3	7	science	science	NOUN
cana-2643	3	8	and	and	CCONJ
cana-2643	3	9	engineering	engineering	NOUN
cana-2643	3	10	(	(	PUNCT
cana-2643	3	11	data	data	NOUN
cana-2643	3	12	science	science	NOUN
cana-2643	3	13	)	)	PUNCT
cana-2643	3	14	,	,	PUNCT
cana-2643	3	15	rajeev	rajeev	PROPN
cana-2643	3	16	gandhi	gandhi	PROPN
cana-2643	3	17	memorial	memorial	PROPN
cana-2643	3	18	college	college	PROPN
cana-2643	3	19	of	of	ADP
cana-2643	3	20	engineering	engineering	NOUN
cana-2643	3	21	and	and	CCONJ
cana-2643	3	22	technology	technology	NOUN
cana-2643	3	23	,	,	PUNCT
cana-2643	3	24	nandyal	nandyal	PROPN
cana-2643	3	25	,	,	PUNCT
cana-2643	3	26	andhrapradesh	andhrapradesh	PROPN
cana-2643	3	27	,	,	PUNCT
cana-2643	3	28	india	india	PROPN
cana-2643	3	29	,	,	PUNCT
cana-2643	3	30	p.prasad71@gmail.com	p.prasad71@gmail.com	PROPN
cana-2643	3	31	3assistant	3assistant	PROPN
cana-2643	3	32	professor	professor	NOUN
cana-2643	3	33	,	,	PUNCT
cana-2643	3	34	department	department	NOUN
cana-2643	3	35	of	of	ADP
cana-2643	3	36	cse	cse	PROPN
cana-2643	3	37	(	(	PUNCT
cana-2643	3	38	data	data	NOUN
cana-2643	3	39	science	science	NOUN
cana-2643	3	40	)	)	PUNCT
cana-2643	3	41	,	,	PUNCT
cana-2643	3	42	rajeev	rajeev	PROPN
cana-2643	3	43	gandhi	gandhi	PROPN
cana-2643	3	44	memorial	memorial	PROPN
cana-2643	3	45	college	college	PROPN
cana-2643	3	46	of	of	ADP
cana-2643	3	47	engineering	engineering	PROPN
cana-2643	3	48	&	&	CCONJ
cana-2643	3	49	technology	technology	PROPN
cana-2643	3	50	,	,	PUNCT
cana-2643	3	51	nandyal	nandyal	PROPN
cana-2643	3	52	,	,	PUNCT
cana-2643	3	53	andhrapradesh	andhrapradesh	PROPN
cana-2643	3	54	,	,	PUNCT
cana-2643	3	55	india	india	PROPN
cana-2643	3	56	,	,	PUNCT
cana-2643	3	57	vikram.sam@gmail.com	vikram.sam@gmail.com	X
cana-2643	3	58	4associate	4associate	NUM
cana-2643	3	59	professor	professor	NOUN
cana-2643	3	60	,	,	PUNCT
cana-2643	3	61	department	department	NOUN
cana-2643	3	62	of	of	ADP
cana-2643	3	63	cse	cse	PROPN
cana-2643	3	64	(	(	PUNCT
cana-2643	3	65	data	data	NOUN
cana-2643	3	66	science	science	NOUN
cana-2643	3	67	)	)	PUNCT
cana-2643	3	68	,	,	PUNCT
cana-2643	3	69	rajeev	rajeev	PROPN
cana-2643	3	70	gandhi	gandhi	PROPN
cana-2643	3	71	memorial	memorial	PROPN
cana-2643	3	72	college	college	PROPN
cana-2643	3	73	of	of	ADP
cana-2643	3	74	engineering	engineering	PROPN
cana-2643	3	75	&	&	CCONJ
cana-2643	3	76	technology	technology	PROPN
cana-2643	3	77	,	,	PUNCT
cana-2643	3	78	nandyal	nandyal	PROPN
cana-2643	3	79	,	,	PUNCT
cana-2643	3	80	andhrapradesh	andhrapradesh	PROPN
cana-2643	3	81	,	,	PUNCT
cana-2643	3	82	india	india	PROPN
cana-2643	3	83	,	,	PUNCT
cana-2643	3	84	suleman.ndl@gmail.com	suleman.ndl@gmail.com	VERB
cana-2643	3	85	5assistant	5assistant	NUM
cana-2643	3	86	professor	professor	NOUN
cana-2643	3	87	,	,	PUNCT
cana-2643	3	88	department	department	NOUN
cana-2643	3	89	of	of	ADP
cana-2643	3	90	computer	computer	NOUN
cana-2643	3	91	science	science	NOUN
cana-2643	3	92	and	and	CCONJ
cana-2643	3	93	engineering	engineering	NOUN
cana-2643	3	94	(	(	PUNCT
cana-2643	3	95	data	data	NOUN
cana-2643	3	96	science	science	NOUN
cana-2643	3	97	)	)	PUNCT
cana-2643	3	98	,	,	PUNCT
cana-2643	3	99	rajeev	rajeev	PROPN
cana-2643	3	100	gandhi	gandhi	PROPN
cana-2643	3	101	memorial	memorial	PROPN
cana-2643	3	102	college	college	PROPN
cana-2643	3	103	of	of	ADP
cana-2643	3	104	engineering	engineering	NOUN
cana-2643	3	105	and	and	CCONJ
cana-2643	3	106	technology	technology	NOUN
cana-2643	3	107	,	,	PUNCT
cana-2643	3	108	nandyal	nandyal	PROPN
cana-2643	3	109	,	,	PUNCT
cana-2643	3	110	andhrapradesh	andhrapradesh	PROPN
cana-2643	3	111	,	,	PUNCT
cana-2643	3	112	india	india	PROPN
cana-2643	3	113	,	,	PUNCT
cana-2643	3	114	kiranraocse@gmail.com	kiranraocse@gmail.com	X
cana-2643	4	1	6associate	6associate	NUM
cana-2643	4	2	professor	professor	NOUN
cana-2643	4	3	,	,	PUNCT
cana-2643	4	4	department	department	PROPN
cana-2643	4	5	of	of	ADP
cana-2643	4	6	ece	ece	PROPN
cana-2643	4	7	,	,	PUNCT
cana-2643	4	8	sri	sri	PROPN
cana-2643	4	9	krishna	krishna	PROPN
cana-2643	4	10	college	college	PROPN
cana-2643	4	11	of	of	ADP
cana-2643	4	12	engineering	engineering	NOUN
cana-2643	4	13	and	and	CCONJ
cana-2643	4	14	technology	technology	NOUN
cana-2643	4	15	,	,	PUNCT
cana-2643	4	16	coimbatore	coimbatore	PROPN
cana-2643	4	17	drsenthilganesh@gmail.com	drsenthilganesh@gmail.com	PROPN
cana-2643	4	18	article	article	NOUN
cana-2643	4	19	history	history	NOUN
cana-2643	4	20	:	:	PUNCT
cana-2643	4	21	received	receive	VERB
cana-2643	4	22	:	:	PUNCT
cana-2643	4	23	21	21	NUM
cana-2643	4	24	-	-	SYM
cana-2643	4	25	09	09	NUM
cana-2643	4	26	-	-	PUNCT
cana-2643	4	27	2024	2024	NUM
cana-2643	4	28	revised	revise	VERB
cana-2643	4	29	:	:	PUNCT
cana-2643	4	30	30	30	NUM
cana-2643	4	31	-	-	SYM
cana-2643	4	32	10	10	NUM
cana-2643	4	33	-	-	PUNCT
cana-2643	4	34	2024	2024	NUM
cana-2643	4	35	accepted	accept	VERB
cana-2643	4	36	:	:	PUNCT
cana-2643	4	37	16	16	NUM
cana-2643	4	38	-	-	SYM
cana-2643	4	39	11	11	NUM
cana-2643	4	40	-	-	PUNCT
cana-2643	4	41	2024	2024	NUM
cana-2643	4	42	abstract	abstract	NOUN
cana-2643	4	43	:	:	PUNCT
cana-2643	4	44	robotic	robotic	ADJ
cana-2643	4	45	vision	vision	NOUN
cana-2643	4	46	plays	play	VERB
cana-2643	4	47	a	a	DET
cana-2643	4	48	crucial	crucial	ADJ
cana-2643	4	49	role	role	NOUN
cana-2643	4	50	in	in	ADP
cana-2643	4	51	enabling	enable	VERB
cana-2643	4	52	autonomous	autonomous	ADJ
cana-2643	4	53	systems	system	NOUN
cana-2643	4	54	to	to	AUX
cana-2643	4	55	perceive	perceive	VERB
cana-2643	4	56	,	,	PUNCT
cana-2643	4	57	understand	understand	VERB
cana-2643	4	58	,	,	PUNCT
cana-2643	4	59	and	and	CCONJ
cana-2643	4	60	interact	interact	VERB
cana-2643	4	61	with	with	ADP
cana-2643	4	62	complex	complex	ADJ
cana-2643	4	63	environments	environment	NOUN
cana-2643	4	64	.	.	PUNCT
cana-2643	5	1	however	however	ADV
cana-2643	5	2	,	,	PUNCT
cana-2643	5	3	accurately	accurately	ADV
cana-2643	5	4	segmenting	segment	VERB
cana-2643	5	5	and	and	CCONJ
cana-2643	5	6	prioritizing	prioritize	VERB
cana-2643	5	7	visual	visual	ADJ
cana-2643	5	8	regions	region	NOUN
cana-2643	5	9	of	of	ADP
cana-2643	5	10	interest	interest	NOUN
cana-2643	5	11	in	in	ADP
cana-2643	5	12	dynamic	dynamic	ADJ
cana-2643	5	13	scenes	scene	NOUN
cana-2643	5	14	is	be	AUX
cana-2643	5	15	challenging	challenge	VERB
cana-2643	5	16	due	due	ADJ
cana-2643	5	17	to	to	ADP
cana-2643	5	18	variations	variation	NOUN
cana-2643	5	19	in	in	ADP
cana-2643	5	20	object	object	NOUN
cana-2643	5	21	shapes	shape	NOUN
cana-2643	5	22	,	,	PUNCT
cana-2643	5	23	sizes	size	NOUN
cana-2643	5	24	,	,	PUNCT
cana-2643	5	25	and	and	CCONJ
cana-2643	5	26	spatial	spatial	ADJ
cana-2643	5	27	relationships	relationship	NOUN
cana-2643	5	28	.	.	PUNCT
cana-2643	6	1	traditional	traditional	ADJ
cana-2643	6	2	methods	method	NOUN
cana-2643	6	3	,	,	PUNCT
cana-2643	6	4	such	such	ADJ
cana-2643	6	5	as	as	ADP
cana-2643	6	6	convolutional	convolutional	ADJ
cana-2643	6	7	neural	neural	ADJ
cana-2643	6	8	networks	network	NOUN
cana-2643	6	9	(	(	PUNCT
cana-2643	6	10	cnns	cnns	PROPN
cana-2643	6	11	)	)	PUNCT
cana-2643	6	12	and	and	CCONJ
cana-2643	6	13	fully	fully	ADV
cana-2643	6	14	convolutional	convolutional	ADJ
cana-2643	6	15	networks	network	NOUN
cana-2643	6	16	(	(	PUNCT
cana-2643	6	17	fcns	fcns	NOUN
cana-2643	6	18	)	)	PUNCT
cana-2643	6	19	,	,	PUNCT
cana-2643	6	20	have	have	AUX
cana-2643	6	21	shown	show	VERB
cana-2643	6	22	promise	promise	NOUN
cana-2643	6	23	but	but	CCONJ
cana-2643	6	24	often	often	ADV
cana-2643	6	25	struggle	struggle	VERB
cana-2643	6	26	to	to	PART
cana-2643	6	27	integrate	integrate	VERB
cana-2643	6	28	both	both	DET
cana-2643	6	29	global	global	ADJ
cana-2643	6	30	context	context	NOUN
cana-2643	6	31	and	and	CCONJ
cana-2643	6	32	fine	fine	ADV
cana-2643	6	33	-	-	PUNCT
cana-2643	6	34	grained	grain	VERB
cana-2643	6	35	details	detail	NOUN
cana-2643	6	36	required	require	VERB
cana-2643	6	37	for	for	ADP
cana-2643	6	38	tasks	task	NOUN
cana-2643	6	39	like	like	ADP
cana-2643	6	40	object	object	NOUN
cana-2643	6	41	recognition	recognition	NOUN
cana-2643	6	42	and	and	CCONJ
cana-2643	6	43	navigation	navigation	NOUN
cana-2643	6	44	.	.	PUNCT
cana-2643	7	1	to	to	PART
cana-2643	7	2	address	address	VERB
cana-2643	7	3	these	these	DET
cana-2643	7	4	challenges	challenge	NOUN
cana-2643	7	5	,	,	PUNCT
cana-2643	7	6	we	we	PRON
cana-2643	7	7	present	present	VERB
cana-2643	7	8	a	a	DET
cana-2643	7	9	multi	multi	ADJ
cana-2643	7	10	-	-	ADJ
cana-2643	7	11	path	path	ADJ
cana-2643	7	12	graph	graph	NOUN
cana-2643	7	13	attention	attention	NOUN
cana-2643	7	14	network	network	NOUN
cana-2643	7	15	with	with	ADP
cana-2643	7	16	reinforcement	reinforcement	NOUN
cana-2643	7	17	learning	learning	NOUN
cana-2643	7	18	(	(	PUNCT
cana-2643	7	19	mpgarl	mpgarl	NOUN
cana-2643	7	20	-	-	PUNCT
cana-2643	7	21	net	net	NOUN
cana-2643	7	22	)	)	PUNCT
cana-2643	7	23	adapted	adapt	VERB
cana-2643	7	24	for	for	ADP
cana-2643	7	25	robotic	robotic	ADJ
cana-2643	7	26	vision	vision	NOUN
cana-2643	7	27	.	.	PUNCT
cana-2643	8	1	this	this	DET
cana-2643	8	2	framework	framework	NOUN
cana-2643	8	3	leverages	leverage	VERB
cana-2643	8	4	a	a	DET
cana-2643	8	5	multipath	multipath	NOUN
cana-2643	8	6	feature	feature	NOUN
cana-2643	8	7	extraction	extraction	NOUN
cana-2643	8	8	(	(	PUNCT
cana-2643	8	9	mpfe	mpfe	NOUN
cana-2643	8	10	)	)	PUNCT
cana-2643	8	11	module	module	NOUN
cana-2643	8	12	to	to	PART
cana-2643	8	13	capture	capture	VERB
cana-2643	8	14	multi	multi	ADJ
cana-2643	8	15	-	-	ADJ
cana-2643	8	16	scale	scale	ADJ
cana-2643	8	17	features	feature	NOUN
cana-2643	8	18	at	at	ADP
cana-2643	8	19	low	low	ADJ
cana-2643	8	20	,	,	PUNCT
cana-2643	8	21	medium	medium	ADJ
cana-2643	8	22	,	,	PUNCT
cana-2643	8	23	and	and	CCONJ
cana-2643	8	24	high	high	ADJ
cana-2643	8	25	resolutions	resolution	NOUN
cana-2643	8	26	,	,	PUNCT
cana-2643	8	27	fusing	fuse	VERB
cana-2643	8	28	them	they	PRON
cana-2643	8	29	using	use	VERB
cana-2643	8	30	an	an	DET
cana-2643	8	31	adaptive	adaptive	ADJ
cana-2643	8	32	attention	attention	NOUN
cana-2643	8	33	mechanism	mechanism	NOUN
cana-2643	8	34	that	that	PRON
cana-2643	8	35	assigns	assign	VERB
cana-2643	8	36	task	task	NOUN
cana-2643	8	37	-	-	PUNCT
cana-2643	8	38	specific	specific	ADJ
cana-2643	8	39	weights	weight	NOUN
cana-2643	8	40	to	to	ADP
cana-2643	8	41	each	each	DET
cana-2643	8	42	resolution	resolution	NOUN
cana-2643	8	43	level	level	NOUN
cana-2643	8	44	.	.	PUNCT
cana-2643	9	1	super	super	ADV
cana-2643	9	2	pixelbased	pixelbase	VERB
cana-2643	9	3	segmentation	segmentation	NOUN
cana-2643	9	4	is	be	AUX
cana-2643	9	5	then	then	ADV
cana-2643	9	6	applied	apply	VERB
cana-2643	9	7	to	to	ADP
cana-2643	9	8	the	the	DET
cana-2643	9	9	fused	fuse	VERB
cana-2643	9	10	feature	feature	NOUN
cana-2643	9	11	map	map	NOUN
cana-2643	9	12	,	,	PUNCT
cana-2643	9	13	representing	represent	VERB
cana-2643	9	14	regions	region	NOUN
cana-2643	9	15	as	as	ADP
cana-2643	9	16	graph	graph	NOUN
cana-2643	9	17	nodes	node	NOUN
cana-2643	9	18	,	,	PUNCT
cana-2643	9	19	with	with	ADP
cana-2643	9	20	graph	graph	NOUN
cana-2643	9	21	convolutional	convolutional	ADJ
cana-2643	9	22	networks	network	NOUN
cana-2643	9	23	(	(	PUNCT
cana-2643	9	24	gcns	gcns	PROPN
cana-2643	9	25	)	)	PUNCT
cana-2643	9	26	employed	employ	VERB
cana-2643	9	27	to	to	PART
cana-2643	9	28	model	model	VERB
cana-2643	9	29	spatial	spatial	ADJ
cana-2643	9	30	relationships	relationship	NOUN
cana-2643	9	31	between	between	ADP
cana-2643	9	32	regions	region	NOUN
cana-2643	9	33	.	.	PUNCT
cana-2643	10	1	reinforcement	reinforcement	NOUN
cana-2643	10	2	learning	learning	NOUN
cana-2643	10	3	(	(	PUNCT
cana-2643	10	4	rl	rl	NOUN
cana-2643	10	5	)	)	PUNCT
cana-2643	10	6	is	be	AUX
cana-2643	10	7	further	far	ADV
cana-2643	10	8	incorporated	incorporate	VERB
cana-2643	10	9	to	to	PART
cana-2643	10	10	dynamically	dynamically	ADV
cana-2643	10	11	adjust	adjust	VERB
cana-2643	10	12	attention	attention	NOUN
cana-2643	10	13	,	,	PUNCT
cana-2643	10	14	allowing	allow	VERB
cana-2643	10	15	the	the	DET
cana-2643	10	16	model	model	NOUN
cana-2643	10	17	to	to	PART
cana-2643	10	18	focus	focus	VERB
cana-2643	10	19	on	on	ADP
cana-2643	10	20	critical	critical	ADJ
cana-2643	10	21	areas	area	NOUN
cana-2643	10	22	,	,	PUNCT
cana-2643	10	23	such	such	ADJ
cana-2643	10	24	as	as	ADP
cana-2643	10	25	target	target	NOUN
cana-2643	10	26	objects	object	NOUN
cana-2643	10	27	or	or	CCONJ
cana-2643	10	28	pathways	pathway	NOUN
cana-2643	10	29	,	,	PUNCT
cana-2643	10	30	in	in	ADP
cana-2643	10	31	real	real	ADJ
cana-2643	10	32	-	-	PUNCT
cana-2643	10	33	time	time	NOUN
cana-2643	10	34	.	.	PUNCT
cana-2643	11	1	performance	performance	NOUN
cana-2643	11	2	evaluations	evaluation	NOUN
cana-2643	11	3	demonstrate	demonstrate	VERB
cana-2643	11	4	that	that	SCONJ
cana-2643	11	5	mpga	mpga	NOUN
cana-2643	11	6	-	-	PUNCT
cana-2643	11	7	rl	rl	NOUN
cana-2643	11	8	-	-	PUNCT
cana-2643	11	9	net	net	NOUN
cana-2643	11	10	enhances	enhance	NOUN
cana-2643	11	11	accuracy	accuracy	NOUN
cana-2643	11	12	in	in	ADP
cana-2643	11	13	robotic	robotic	ADJ
cana-2643	11	14	vision	vision	NOUN
cana-2643	11	15	tasks	task	NOUN
cana-2643	11	16	,	,	PUNCT
cana-2643	11	17	particularly	particularly	ADV
cana-2643	11	18	in	in	ADP
cana-2643	11	19	cluttered	cluttered	ADJ
cana-2643	11	20	or	or	CCONJ
cana-2643	11	21	dynamically	dynamically	ADV
cana-2643	11	22	changing	change	VERB
cana-2643	11	23	environments	environment	NOUN
cana-2643	11	24	,	,	PUNCT
cana-2643	11	25	achieving	achieve	VERB
cana-2643	11	26	higher	high	ADJ
cana-2643	11	27	precision	precision	NOUN
cana-2643	11	28	in	in	ADP
cana-2643	11	29	object	object	NOUN
cana-2643	11	30	detection	detection	NOUN
cana-2643	11	31	and	and	CCONJ
cana-2643	11	32	adaptive	adaptive	ADJ
cana-2643	11	33	focus	focus	NOUN
cana-2643	11	34	on	on	ADP
cana-2643	11	35	critical	critical	ADJ
cana-2643	11	36	regions	region	NOUN
cana-2643	11	37	compared	compare	VERB
cana-2643	11	38	to	to	ADP
cana-2643	11	39	conventional	conventional	ADJ
cana-2643	11	40	methods	method	NOUN
cana-2643	11	41	keywords	keyword	NOUN
cana-2643	11	42	:	:	PUNCT
cana-2643	11	43	robotic	robotic	ADJ
cana-2643	11	44	vision	vision	NOUN
cana-2643	11	45	·	·	PUNCT
cana-2643	11	46	computer	computer	NOUN
cana-2643	11	47	vision	vision	NOUN
cana-2643	11	48	·	·	PUNCT
cana-2643	11	49	graph	graph	VERB
cana-2643	11	50	convolutional	convolutional	ADJ
cana-2643	11	51	network	network	NOUN
cana-2643	11	52	(	(	PUNCT
cana-2643	11	53	gcn	gcn	NOUN
cana-2643	11	54	)	)	PUNCT
cana-2643	11	55	·	·	PUNCT
cana-2643	11	56	reinforcement	reinforcement	NOUN
cana-2643	11	57	learning	learning	NOUN
cana-2643	11	58	·	·	PUNCT
cana-2643	11	59	multi	multi	ADJ
cana-2643	11	60	-	-	ADJ
cana-2643	11	61	scale	scale	ADJ
cana-2643	11	62	feature	feature	NOUN
cana-2643	11	63	extraction	extraction	NOUN
cana-2643	11	64	·	·	PUNCT
cana-2643	11	65	multipath	multipath	NOUN
cana-2643	11	66	·	·	PUNCT
cana-2643	11	67	attention	attention	NOUN
cana-2643	11	68	mechanism	mechanism	NOUN
cana-2643	11	69	.	.	PUNCT
cana-2643	12	1	communications	communication	NOUN
cana-2643	12	2	on	on	ADP
cana-2643	12	3	applied	apply	VERB
cana-2643	12	4	nonlinear	nonlinear	ADJ
cana-2643	12	5	analysis	analysis	NOUN
cana-2643	12	6	issn	issn	NOUN
cana-2643	12	7	:	:	PUNCT
cana-2643	12	8	1074	1074	NUM
cana-2643	12	9	-	-	PUNCT
cana-2643	12	10	133x	133x	NUM
cana-2643	12	11	vol	vol	NOUN
cana-2643	12	12	32	32	NUM
cana-2643	12	13	no	no	NOUN
cana-2643	12	14	.	.	PUNCT
cana-2643	13	1	3s	3s	NUM
cana-2643	13	2	(	(	PUNCT
cana-2643	13	3	2025	2025	NUM
cana-2643	13	4	)	)	PUNCT
cana-2643	13	5	280	280	NUM
cana-2643	13	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	13	7	1	1	NUM
cana-2643	13	8	.	.	PUNCT
cana-2643	13	9	introduction	introduction	NOUN
cana-2643	13	10	robotic	robotic	ADJ
cana-2643	13	11	vision	vision	NOUN
cana-2643	13	12	systems	system	NOUN
cana-2643	13	13	are	be	AUX
cana-2643	13	14	critical	critical	ADJ
cana-2643	13	15	for	for	ADP
cana-2643	13	16	empowering	empower	VERB
cana-2643	13	17	autonomous	autonomous	ADJ
cana-2643	13	18	robots	robot	NOUN
cana-2643	13	19	with	with	ADP
cana-2643	13	20	the	the	DET
cana-2643	13	21	ability	ability	NOUN
cana-2643	13	22	to	to	PART
cana-2643	13	23	interpret	interpret	VERB
cana-2643	13	24	,	,	PUNCT
cana-2643	13	25	navigate	navigate	VERB
cana-2643	13	26	,	,	PUNCT
cana-2643	13	27	and	and	CCONJ
cana-2643	13	28	interact	interact	VERB
cana-2643	13	29	with	with	ADP
cana-2643	13	30	their	their	PRON
cana-2643	13	31	surroundings	surrounding	NOUN
cana-2643	13	32	.	.	PUNCT
cana-2643	14	1	applications	application	NOUN
cana-2643	14	2	in	in	ADP
cana-2643	14	3	areas	area	NOUN
cana-2643	14	4	such	such	ADJ
cana-2643	14	5	as	as	ADP
cana-2643	14	6	industrial	industrial	ADJ
cana-2643	14	7	automation	automation	NOUN
cana-2643	14	8	,	,	PUNCT
cana-2643	14	9	environmental	environmental	ADJ
cana-2643	14	10	monitoring	monitoring	NOUN
cana-2643	14	11	,	,	PUNCT
cana-2643	14	12	and	and	CCONJ
cana-2643	14	13	healthcare	healthcare	NOUN
cana-2643	14	14	demand	demand	VERB
cana-2643	14	15	a	a	DET
cana-2643	14	16	high	high	ADJ
cana-2643	14	17	level	level	NOUN
cana-2643	14	18	of	of	ADP
cana-2643	14	19	precision	precision	NOUN
cana-2643	14	20	and	and	CCONJ
cana-2643	14	21	adaptability	adaptability	NOUN
cana-2643	14	22	from	from	ADP
cana-2643	14	23	robotic	robotic	ADJ
cana-2643	14	24	vision	vision	NOUN
cana-2643	14	25	models	model	NOUN
cana-2643	14	26	to	to	PART
cana-2643	14	27	ensure	ensure	VERB
cana-2643	14	28	that	that	SCONJ
cana-2643	14	29	robots	robot	NOUN
cana-2643	14	30	can	can	AUX
cana-2643	14	31	make	make	VERB
cana-2643	14	32	informed	inform	VERB
cana-2643	14	33	,	,	PUNCT
cana-2643	14	34	real	real	ADJ
cana-2643	14	35	-	-	PUNCT
cana-2643	14	36	time	time	NOUN
cana-2643	14	37	decisions	decision	NOUN
cana-2643	14	38	based	base	VERB
cana-2643	14	39	on	on	ADP
cana-2643	14	40	visual	visual	ADJ
cana-2643	14	41	data	datum	NOUN
cana-2643	15	1	[	[	X
cana-2643	15	2	1	1	NUM
cana-2643	15	3	]	]	PUNCT
cana-2643	15	4	.	.	PUNCT
cana-2643	16	1	achieving	achieve	VERB
cana-2643	16	2	such	such	ADJ
cana-2643	16	3	accuracy	accuracy	NOUN
cana-2643	16	4	in	in	ADP
cana-2643	16	5	robotic	robotic	ADJ
cana-2643	16	6	perception	perception	NOUN
cana-2643	16	7	tasks	task	NOUN
cana-2643	16	8	requires	require	VERB
cana-2643	16	9	models	model	NOUN
cana-2643	16	10	that	that	PRON
cana-2643	16	11	can	can	AUX
cana-2643	16	12	dynamically	dynamically	ADV
cana-2643	16	13	prioritize	prioritize	VERB
cana-2643	16	14	and	and	CCONJ
cana-2643	16	15	focus	focus	VERB
cana-2643	16	16	on	on	ADP
cana-2643	16	17	relevant	relevant	ADJ
cana-2643	16	18	regions	region	NOUN
cana-2643	16	19	in	in	ADP
cana-2643	16	20	complex	complex	ADJ
cana-2643	16	21	scenes	scene	NOUN
cana-2643	16	22	.	.	PUNCT
cana-2643	17	1	however	however	ADV
cana-2643	17	2	,	,	PUNCT
cana-2643	17	3	this	this	DET
cana-2643	17	4	level	level	NOUN
cana-2643	17	5	of	of	ADP
cana-2643	17	6	adaptability	adaptability	NOUN
cana-2643	17	7	remains	remain	VERB
cana-2643	17	8	a	a	DET
cana-2643	17	9	challenge	challenge	NOUN
cana-2643	17	10	for	for	ADP
cana-2643	17	11	many	many	ADJ
cana-2643	17	12	existing	exist	VERB
cana-2643	17	13	approaches	approach	NOUN
cana-2643	17	14	,	,	PUNCT
cana-2643	17	15	as	as	SCONJ
cana-2643	17	16	robotic	robotic	ADJ
cana-2643	17	17	vision	vision	NOUN
cana-2643	17	18	tasks	task	NOUN
cana-2643	17	19	often	often	ADV
cana-2643	17	20	involve	involve	VERB
cana-2643	17	21	significant	significant	ADJ
cana-2643	17	22	variation	variation	NOUN
cana-2643	17	23	in	in	ADP
cana-2643	17	24	object	object	NOUN
cana-2643	17	25	shapes	shape	NOUN
cana-2643	17	26	,	,	PUNCT
cana-2643	17	27	sizes	size	NOUN
cana-2643	17	28	,	,	PUNCT
cana-2643	17	29	and	and	CCONJ
cana-2643	17	30	spatial	spatial	ADJ
cana-2643	17	31	relationships	relationship	NOUN
cana-2643	17	32	,	,	PUNCT
cana-2643	17	33	particularly	particularly	ADV
cana-2643	17	34	in	in	ADP
cana-2643	17	35	unstructured	unstructured	ADJ
cana-2643	17	36	environments	environment	NOUN
cana-2643	17	37	[	[	X
cana-2643	17	38	2	2	NUM
cana-2643	17	39	]	]	PUNCT
cana-2643	17	40	.	.	PUNCT
cana-2643	18	1	traditional	traditional	ADJ
cana-2643	18	2	computer	computer	NOUN
cana-2643	18	3	vision	vision	NOUN
cana-2643	18	4	systems	system	NOUN
cana-2643	18	5	have	have	AUX
cana-2643	18	6	relied	rely	VERB
cana-2643	18	7	heavily	heavily	ADV
cana-2643	18	8	on	on	ADP
cana-2643	18	9	convolutional	convolutional	ADJ
cana-2643	18	10	neural	neural	ADJ
cana-2643	18	11	networks	network	NOUN
cana-2643	18	12	(	(	PUNCT
cana-2643	18	13	cnns)[3	cnns)[3	X
cana-2643	18	14	]	]	PUNCT
cana-2643	18	15	as	as	SCONJ
cana-2643	18	16	shown	show	VERB
cana-2643	18	17	in	in	ADP
cana-2643	18	18	figure	figure	NOUN
cana-2643	18	19	1	1	NUM
cana-2643	18	20	,	,	PUNCT
cana-2643	18	21	which	which	PRON
cana-2643	18	22	are	be	AUX
cana-2643	18	23	highly	highly	ADV
cana-2643	18	24	effective	effective	ADJ
cana-2643	18	25	at	at	ADP
cana-2643	18	26	extracting	extract	VERB
cana-2643	18	27	visual	visual	ADJ
cana-2643	18	28	features	feature	NOUN
cana-2643	18	29	across	across	ADP
cana-2643	18	30	various	various	ADJ
cana-2643	18	31	image	image	NOUN
cana-2643	18	32	domains	domain	NOUN
cana-2643	18	33	.	.	PUNCT
cana-2643	19	1	however	however	ADV
cana-2643	19	2	,	,	PUNCT
cana-2643	19	3	cnns	cnn	NOUN
cana-2643	19	4	generally	generally	ADV
cana-2643	19	5	lack	lack	VERB
cana-2643	19	6	the	the	DET
cana-2643	19	7	flexibility	flexibility	NOUN
cana-2643	19	8	to	to	PART
cana-2643	19	9	adapt	adapt	VERB
cana-2643	19	10	focus	focus	NOUN
cana-2643	19	11	dynamically	dynamically	ADV
cana-2643	19	12	in	in	ADP
cana-2643	19	13	response	response	NOUN
cana-2643	19	14	to	to	ADP
cana-2643	19	15	task	task	NOUN
cana-2643	19	16	requirements	requirement	NOUN
cana-2643	19	17	or	or	CCONJ
cana-2643	19	18	environmental	environmental	ADJ
cana-2643	19	19	changes	change	NOUN
cana-2643	19	20	,	,	PUNCT
cana-2643	19	21	which	which	PRON
cana-2643	19	22	limits	limit	VERB
cana-2643	19	23	their	their	PRON
cana-2643	19	24	applicability	applicability	NOUN
cana-2643	19	25	in	in	ADP
cana-2643	19	26	complex	complex	ADJ
cana-2643	19	27	robotic	robotic	ADJ
cana-2643	19	28	tasks	task	NOUN
cana-2643	19	29	where	where	SCONJ
cana-2643	19	30	scene	scene	NOUN
cana-2643	19	31	context	context	NOUN
cana-2643	19	32	and	and	CCONJ
cana-2643	19	33	focus	focus	VERB
cana-2643	19	34	need	need	VERB
cana-2643	19	35	to	to	PART
cana-2643	19	36	be	be	AUX
cana-2643	19	37	continually	continually	ADV
cana-2643	19	38	reassessed	reassess	VERB
cana-2643	19	39	.	.	PUNCT
cana-2643	20	1	in	in	ADP
cana-2643	20	2	the	the	DET
cana-2643	20	3	context	context	NOUN
cana-2643	20	4	of	of	ADP
cana-2643	20	5	segmentation	segmentation	NOUN
cana-2643	20	6	,	,	PUNCT
cana-2643	20	7	u	u	NOUN
cana-2643	20	8	-	-	NOUN
cana-2643	20	9	net[4	net[4	PRON
cana-2643	20	10	]	]	PUNCT
cana-2643	20	11	and	and	CCONJ
cana-2643	20	12	segnet[5	segnet[5	PROPN
cana-2643	20	13	]	]	PUNCT
cana-2643	20	14	architectures	architecture	NOUN
cana-2643	20	15	have	have	AUX
cana-2643	20	16	shown	show	VERB
cana-2643	20	17	effectiveness	effectiveness	NOUN
cana-2643	20	18	in	in	ADP
cana-2643	20	19	balancing	balance	VERB
cana-2643	20	20	fine	fine	ADJ
cana-2643	20	21	detail	detail	NOUN
cana-2643	20	22	with	with	ADP
cana-2643	20	23	broader	broad	ADJ
cana-2643	20	24	spatial	spatial	ADJ
cana-2643	20	25	context	context	NOUN
cana-2643	20	26	,	,	PUNCT
cana-2643	20	27	making	make	VERB
cana-2643	20	28	them	they	PRON
cana-2643	20	29	popular	popular	ADJ
cana-2643	20	30	choices	choice	NOUN
cana-2643	20	31	in	in	ADP
cana-2643	20	32	both	both	CCONJ
cana-2643	20	33	general	general	ADJ
cana-2643	20	34	and	and	CCONJ
cana-2643	20	35	medical	medical	ADJ
cana-2643	20	36	image	image	NOUN
cana-2643	20	37	segmentation	segmentation	NOUN
cana-2643	20	38	.	.	PUNCT
cana-2643	21	1	figure	figure	NOUN
cana-2643	21	2	1	1	NUM
cana-2643	21	3	.	.	PUNCT
cana-2643	22	1	applying	apply	VERB
cana-2643	22	2	neural	neural	ADJ
cana-2643	22	3	networks	network	NOUN
cana-2643	22	4	to	to	ADP
cana-2643	22	5	robotic	robotic	ADJ
cana-2643	22	6	vision	vision	NOUN
cana-2643	22	7	and	and	CCONJ
cana-2643	22	8	guidance	guidance	NOUN
cana-2643	22	9	u	u	NOUN
cana-2643	22	10	-	-	NOUN
cana-2643	22	11	net[4	net[4	NUM
cana-2643	22	12	]	]	PUNCT
cana-2643	22	13	and	and	CCONJ
cana-2643	22	14	segnet[5	segnet[5	PROPN
cana-2643	22	15	]	]	PUNCT
cana-2643	22	16	architectures	architecture	NOUN
cana-2643	22	17	have	have	AUX
cana-2643	22	18	shown	show	VERB
cana-2643	22	19	effectiveness	effectiveness	NOUN
cana-2643	22	20	in	in	ADP
cana-2643	22	21	balancing	balance	VERB
cana-2643	22	22	fine	fine	ADJ
cana-2643	22	23	detail	detail	NOUN
cana-2643	22	24	with	with	ADP
cana-2643	22	25	broader	broad	ADJ
cana-2643	22	26	spatial	spatial	ADJ
cana-2643	22	27	context	context	NOUN
cana-2643	22	28	,	,	PUNCT
cana-2643	22	29	making	make	VERB
cana-2643	22	30	them	they	PRON
cana-2643	22	31	popular	popular	ADJ
cana-2643	22	32	choices	choice	NOUN
cana-2643	22	33	in	in	ADP
cana-2643	22	34	both	both	CCONJ
cana-2643	22	35	general	general	ADJ
cana-2643	22	36	and	and	CCONJ
cana-2643	22	37	medical	medical	ADJ
cana-2643	22	38	image	image	NOUN
cana-2643	22	39	segmentation	segmentation	NOUN
cana-2643	22	40	.	.	PUNCT
cana-2643	23	1	yet	yet	ADV
cana-2643	23	2	,	,	PUNCT
cana-2643	23	3	these	these	DET
cana-2643	23	4	models	model	NOUN
cana-2643	23	5	often	often	ADV
cana-2643	23	6	struggle	struggle	VERB
cana-2643	23	7	in	in	ADP
cana-2643	23	8	scenarios	scenario	NOUN
cana-2643	23	9	requiring	require	VERB
cana-2643	23	10	high	high	ADJ
cana-2643	23	11	precision	precision	NOUN
cana-2643	23	12	,	,	PUNCT
cana-2643	23	13	such	such	ADJ
cana-2643	23	14	as	as	ADP
cana-2643	23	15	segmenting	segment	VERB
cana-2643	23	16	small	small	ADJ
cana-2643	23	17	,	,	PUNCT
cana-2643	23	18	irregularly	irregularly	ADV
cana-2643	23	19	shaped	shape	VERB
cana-2643	23	20	objects	object	NOUN
cana-2643	23	21	or	or	CCONJ
cana-2643	23	22	recognizing	recognize	VERB
cana-2643	23	23	boundaries	boundary	NOUN
cana-2643	23	24	in	in	ADP
cana-2643	23	25	cluttered	cluttered	ADJ
cana-2643	23	26	or	or	CCONJ
cana-2643	23	27	dynamic	dynamic	ADJ
cana-2643	23	28	environments	environment	NOUN
cana-2643	23	29	.	.	PUNCT
cana-2643	24	1	such	such	ADJ
cana-2643	24	2	limitations	limitation	NOUN
cana-2643	24	3	highlight	highlight	VERB
cana-2643	24	4	the	the	DET
cana-2643	24	5	need	need	NOUN
cana-2643	24	6	for	for	ADP
cana-2643	24	7	models	model	NOUN
cana-2643	24	8	that	that	PRON
cana-2643	24	9	can	can	AUX
cana-2643	24	10	dynamically	dynamically	ADV
cana-2643	24	11	adapt	adapt	VERB
cana-2643	24	12	to	to	PART
cana-2643	24	13	prioritize	prioritize	VERB
cana-2643	24	14	region	region	NOUN
cana-2643	24	15	-	-	PUNCT
cana-2643	24	16	specific	specific	ADJ
cana-2643	24	17	focus	focus	NOUN
cana-2643	24	18	in	in	ADP
cana-2643	24	19	real	real	ADJ
cana-2643	24	20	time[6	time[6	NOUN
cana-2643	24	21	]	]	PUNCT
cana-2643	24	22	.	.	PUNCT
cana-2643	25	1	to	to	PART
cana-2643	25	2	address	address	VERB
cana-2643	25	3	some	some	PRON
cana-2643	25	4	of	of	ADP
cana-2643	25	5	these	these	DET
cana-2643	25	6	challenges	challenge	NOUN
cana-2643	25	7	,	,	PUNCT
cana-2643	25	8	recent	recent	ADJ
cana-2643	25	9	research	research	NOUN
cana-2643	25	10	has	have	AUX
cana-2643	25	11	explored	explore	VERB
cana-2643	25	12	graphbased	graphbased	ADJ
cana-2643	25	13	models	model	NOUN
cana-2643	25	14	,	,	PUNCT
cana-2643	25	15	including	include	VERB
cana-2643	25	16	graph	graph	NOUN
cana-2643	25	17	neural	neural	ADJ
cana-2643	25	18	networks	network	NOUN
cana-2643	25	19	(	(	PUNCT
cana-2643	25	20	gnns	gnns	NOUN
cana-2643	25	21	)	)	PUNCT
cana-2643	25	22	and	and	CCONJ
cana-2643	25	23	graph	graph	VERB
cana-2643	25	24	convolutional	convolutional	ADJ
cana-2643	25	25	networks	network	NOUN
cana-2643	25	26	(	(	PUNCT
cana-2643	25	27	gcns)[7	gcns)[7	NOUN
cana-2643	25	28	]	]	X
cana-2643	25	29	,	,	PUNCT
cana-2643	25	30	for	for	ADP
cana-2643	25	31	representing	represent	VERB
cana-2643	25	32	spatial	spatial	ADJ
cana-2643	25	33	relationships	relationship	NOUN
cana-2643	25	34	within	within	ADP
cana-2643	25	35	scenes	scene	NOUN
cana-2643	25	36	.	.	PUNCT
cana-2643	26	1	by	by	ADP
cana-2643	26	2	treating	treat	VERB
cana-2643	26	3	regions	region	NOUN
cana-2643	26	4	of	of	ADP
cana-2643	26	5	interest	interest	NOUN
cana-2643	26	6	as	as	ADP
cana-2643	26	7	nodes	node	NOUN
cana-2643	26	8	within	within	ADP
cana-2643	26	9	a	a	DET
cana-2643	26	10	graph	graph	NOUN
cana-2643	26	11	,	,	PUNCT
cana-2643	26	12	gcns	gcns	PROPN
cana-2643	26	13	can	can	AUX
cana-2643	26	14	aggregate	aggregate	VERB
cana-2643	26	15	spatial	spatial	ADJ
cana-2643	26	16	information	information	NOUN
cana-2643	26	17	across	across	ADP
cana-2643	26	18	different	different	ADJ
cana-2643	26	19	parts	part	NOUN
cana-2643	26	20	of	of	ADP
cana-2643	26	21	an	an	DET
cana-2643	26	22	image	image	NOUN
cana-2643	26	23	,	,	PUNCT
cana-2643	26	24	providing	provide	VERB
cana-2643	26	25	an	an	DET
cana-2643	26	26	advantage	advantage	NOUN
cana-2643	26	27	in	in	ADP
cana-2643	26	28	capturing	capture	VERB
cana-2643	26	29	complex	complex	ADJ
cana-2643	26	30	relationships	relationship	NOUN
cana-2643	26	31	between	between	ADP
cana-2643	26	32	objects	object	NOUN
cana-2643	26	33	.	.	PUNCT
cana-2643	27	1	this	this	DET
cana-2643	27	2	approach	approach	NOUN
cana-2643	27	3	has	have	AUX
cana-2643	27	4	shown	show	VERB
cana-2643	27	5	particular	particular	ADJ
cana-2643	27	6	promise	promise	NOUN
cana-2643	27	7	in	in	ADP
cana-2643	27	8	tasks	task	NOUN
cana-2643	27	9	like	like	ADP
cana-2643	27	10	medical	medical	ADJ
cana-2643	27	11	imaging	imaging	NOUN
cana-2643	27	12	,	,	PUNCT
cana-2643	27	13	where	where	SCONJ
cana-2643	27	14	intricate	intricate	ADJ
cana-2643	27	15	boundary	boundary	ADJ
cana-2643	27	16	detection	detection	NOUN
cana-2643	27	17	is	be	AUX
cana-2643	27	18	crucial	crucial	ADJ
cana-2643	27	19	.	.	PUNCT
cana-2643	28	1	although	although	SCONJ
cana-2643	28	2	gcns	gcns	PROPN
cana-2643	28	3	improve	improve	VERB
cana-2643	28	4	spatial	spatial	ADJ
cana-2643	28	5	representation	representation	NOUN
cana-2643	28	6	and	and	CCONJ
cana-2643	28	7	structure	structure	NOUN
cana-2643	28	8	within	within	ADP
cana-2643	28	9	scenes	scene	NOUN
cana-2643	28	10	,	,	PUNCT
cana-2643	28	11	they	they	PRON
cana-2643	28	12	often	often	ADV
cana-2643	28	13	employ	employ	VERB
cana-2643	28	14	static	static	ADJ
cana-2643	28	15	structures	structure	NOUN
cana-2643	28	16	and	and	CCONJ
cana-2643	28	17	attention	attention	NOUN
cana-2643	28	18	mechanisms	mechanism	NOUN
cana-2643	28	19	that	that	PRON
cana-2643	28	20	lack	lack	VERB
cana-2643	28	21	the	the	DET
cana-2643	28	22	dynamic	dynamic	ADJ
cana-2643	28	23	prioritization	prioritization	NOUN
cana-2643	28	24	capabilities	capability	NOUN
cana-2643	28	25	required	require	VERB
cana-2643	28	26	in	in	ADP
cana-2643	28	27	real	real	ADJ
cana-2643	28	28	-	-	PUNCT
cana-2643	28	29	time	time	NOUN
cana-2643	28	30	robotic	robotic	ADJ
cana-2643	28	31	vision[8	vision[8	PROPN
cana-2643	28	32	]	]	PUNCT
cana-2643	28	33	.	.	PUNCT
cana-2643	29	1	communications	communication	NOUN
cana-2643	29	2	on	on	ADP
cana-2643	29	3	applied	apply	VERB
cana-2643	29	4	nonlinear	nonlinear	ADJ
cana-2643	29	5	analysis	analysis	NOUN
cana-2643	29	6	issn	issn	NOUN
cana-2643	29	7	:	:	PUNCT
cana-2643	29	8	1074	1074	NUM
cana-2643	29	9	-	-	PUNCT
cana-2643	29	10	133x	133x	NUM
cana-2643	29	11	vol	vol	NOUN
cana-2643	29	12	32	32	NUM
cana-2643	29	13	no	no	NOUN
cana-2643	29	14	.	.	PUNCT
cana-2643	30	1	3s	3s	NUM
cana-2643	30	2	(	(	PUNCT
cana-2643	30	3	2025	2025	NUM
cana-2643	30	4	)	)	PUNCT
cana-2643	30	5	281	281	NUM
cana-2643	30	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	30	7	moreover	moreover	ADV
cana-2643	30	8	,	,	PUNCT
cana-2643	30	9	fixed	fix	VERB
cana-2643	30	10	attention	attention	NOUN
cana-2643	30	11	weights	weight	NOUN
cana-2643	30	12	in	in	ADP
cana-2643	30	13	these	these	DET
cana-2643	30	14	models	model	NOUN
cana-2643	30	15	do	do	AUX
cana-2643	30	16	not	not	PART
cana-2643	30	17	adjust	adjust	VERB
cana-2643	30	18	according	accord	VERB
cana-2643	30	19	to	to	ADP
cana-2643	30	20	scene	scene	NOUN
cana-2643	30	21	demands	demand	NOUN
cana-2643	30	22	,	,	PUNCT
cana-2643	30	23	a	a	DET
cana-2643	30	24	limitation	limitation	NOUN
cana-2643	30	25	that	that	PRON
cana-2643	30	26	constrains	constrain	VERB
cana-2643	30	27	their	their	PRON
cana-2643	30	28	effectiveness	effectiveness	NOUN
cana-2643	30	29	in	in	ADP
cana-2643	30	30	dynamically	dynamically	ADV
cana-2643	30	31	changing	change	VERB
cana-2643	30	32	environments	environment	NOUN
cana-2643	30	33	.	.	PUNCT
cana-2643	31	1	motivation	motivation	NOUN
cana-2643	31	2	for	for	ADP
cana-2643	31	3	the	the	DET
cana-2643	31	4	proposed	propose	VERB
cana-2643	31	5	work	work	NOUN
cana-2643	31	6	:	:	PUNCT
cana-2643	31	7	despite	despite	SCONJ
cana-2643	31	8	significant	significant	ADJ
cana-2643	31	9	progress	progress	NOUN
cana-2643	31	10	,	,	PUNCT
cana-2643	31	11	current	current	ADJ
cana-2643	31	12	approaches	approach	NOUN
cana-2643	31	13	in	in	ADP
cana-2643	31	14	robotic	robotic	ADJ
cana-2643	31	15	vision	vision	NOUN
cana-2643	31	16	face	face	VERB
cana-2643	31	17	limitations	limitation	NOUN
cana-2643	31	18	in	in	ADP
cana-2643	31	19	dynamically	dynamically	ADV
cana-2643	31	20	integrating	integrate	VERB
cana-2643	31	21	global	global	ADJ
cana-2643	31	22	and	and	CCONJ
cana-2643	31	23	local	local	ADJ
cana-2643	31	24	context	context	NOUN
cana-2643	31	25	for	for	ADP
cana-2643	31	26	region	region	NOUN
cana-2643	31	27	-	-	PUNCT
cana-2643	31	28	specific	specific	ADJ
cana-2643	31	29	prioritization	prioritization	NOUN
cana-2643	31	30	.	.	PUNCT
cana-2643	32	1	while	while	SCONJ
cana-2643	32	2	cnn	cnn	PROPN
cana-2643	32	3	-	-	PUNCT
cana-2643	32	4	based	base	VERB
cana-2643	32	5	architectures	architecture	NOUN
cana-2643	32	6	offer	offer	VERB
cana-2643	32	7	robust	robust	ADJ
cana-2643	32	8	feature	feature	NOUN
cana-2643	32	9	extraction	extraction	NOUN
cana-2643	32	10	capabilities	capability	NOUN
cana-2643	32	11	,	,	PUNCT
cana-2643	32	12	their	their	PRON
cana-2643	32	13	static	static	ADJ
cana-2643	32	14	attention	attention	NOUN
cana-2643	32	15	mechanisms	mechanism	NOUN
cana-2643	32	16	limit	limit	VERB
cana-2643	32	17	adaptability	adaptability	NOUN
cana-2643	32	18	to	to	ADP
cana-2643	32	19	changing	change	VERB
cana-2643	32	20	task	task	NOUN
cana-2643	32	21	requirements	requirement	NOUN
cana-2643	32	22	.	.	PUNCT
cana-2643	33	1	in	in	ADP
cana-2643	33	2	contrast	contrast	NOUN
cana-2643	33	3	,	,	PUNCT
cana-2643	33	4	graph	graph	NOUN
cana-2643	33	5	-	-	PUNCT
cana-2643	33	6	based	base	VERB
cana-2643	33	7	approaches	approach	NOUN
cana-2643	33	8	provide	provide	VERB
cana-2643	33	9	a	a	DET
cana-2643	33	10	more	more	ADV
cana-2643	33	11	structured	structured	ADJ
cana-2643	33	12	way	way	NOUN
cana-2643	33	13	of	of	ADP
cana-2643	33	14	modeling	model	VERB
cana-2643	33	15	spatial	spatial	ADJ
cana-2643	33	16	relationships	relationship	NOUN
cana-2643	33	17	but	but	CCONJ
cana-2643	33	18	are	be	AUX
cana-2643	33	19	often	often	ADV
cana-2643	33	20	restricted	restrict	VERB
cana-2643	33	21	by	by	ADP
cana-2643	33	22	fixed	fix	VERB
cana-2643	33	23	graph	graph	NOUN
cana-2643	33	24	structures	structure	NOUN
cana-2643	33	25	and	and	CCONJ
cana-2643	33	26	attention	attention	NOUN
cana-2643	33	27	that	that	PRON
cana-2643	33	28	does	do	AUX
cana-2643	33	29	not	not	PART
cana-2643	33	30	adapt	adapt	VERB
cana-2643	33	31	to	to	ADP
cana-2643	33	32	real	real	ADJ
cana-2643	33	33	-	-	PUNCT
cana-2643	33	34	time	time	NOUN
cana-2643	33	35	feedback	feedback	NOUN
cana-2643	33	36	.	.	PUNCT
cana-2643	34	1	these	these	DET
cana-2643	34	2	limitations	limitation	NOUN
cana-2643	34	3	become	become	VERB
cana-2643	34	4	particularly	particularly	ADV
cana-2643	34	5	evident	evident	ADJ
cana-2643	34	6	in	in	ADP
cana-2643	34	7	real	real	ADJ
cana-2643	34	8	-	-	PUNCT
cana-2643	34	9	time	time	NOUN
cana-2643	34	10	applications	application	NOUN
cana-2643	34	11	requiring	require	VERB
cana-2643	34	12	adaptive	adaptive	ADJ
cana-2643	34	13	focus	focus	NOUN
cana-2643	34	14	,	,	PUNCT
cana-2643	34	15	such	such	ADJ
cana-2643	34	16	as	as	ADP
cana-2643	34	17	a	a	DET
cana-2643	34	18	robot	robot	NOUN
cana-2643	34	19	navigating	navigate	VERB
cana-2643	34	20	through	through	ADP
cana-2643	34	21	cluttered	cluttered	ADJ
cana-2643	34	22	environments	environment	NOUN
cana-2643	34	23	or	or	CCONJ
cana-2643	34	24	selecting	select	VERB
cana-2643	34	25	optimal	optimal	ADJ
cana-2643	34	26	grasp	grasp	NOUN
cana-2643	34	27	points	point	NOUN
cana-2643	34	28	on	on	ADP
cana-2643	34	29	objects	object	NOUN
cana-2643	34	30	with	with	ADP
cana-2643	34	31	irregular	irregular	ADJ
cana-2643	34	32	shapes	shape	NOUN
cana-2643	34	33	.	.	PUNCT
cana-2643	35	1	furthermore	furthermore	ADV
cana-2643	35	2	,	,	PUNCT
cana-2643	35	3	many	many	ADJ
cana-2643	35	4	current	current	ADJ
cana-2643	35	5	methods	method	NOUN
cana-2643	35	6	do	do	AUX
cana-2643	35	7	not	not	PART
cana-2643	35	8	fully	fully	ADV
cana-2643	35	9	leverage	leverage	VERB
cana-2643	35	10	multi	multi	ADJ
cana-2643	35	11	-	-	ADJ
cana-2643	35	12	scale	scale	ADJ
cana-2643	35	13	feature	feature	NOUN
cana-2643	35	14	integration	integration	NOUN
cana-2643	35	15	,	,	PUNCT
cana-2643	35	16	often	often	ADV
cana-2643	35	17	resulting	result	VERB
cana-2643	35	18	in	in	ADP
cana-2643	35	19	incomplete	incomplete	ADJ
cana-2643	35	20	segmentation	segmentation	NOUN
cana-2643	35	21	and	and	CCONJ
cana-2643	35	22	impacting	impact	VERB
cana-2643	35	23	task	task	NOUN
cana-2643	35	24	accuracy	accuracy	NOUN
cana-2643	35	25	,	,	PUNCT
cana-2643	35	26	especially	especially	ADV
cana-2643	35	27	in	in	ADP
cana-2643	35	28	scenes	scene	NOUN
cana-2643	35	29	with	with	ADP
cana-2643	35	30	intricate	intricate	ADJ
cana-2643	35	31	spatial	spatial	ADJ
cana-2643	35	32	dependencies	dependency	NOUN
cana-2643	35	33	.	.	PUNCT
cana-2643	36	1	to	to	PART
cana-2643	36	2	overcome	overcome	VERB
cana-2643	36	3	these	these	DET
cana-2643	36	4	challenges	challenge	NOUN
cana-2643	36	5	,	,	PUNCT
cana-2643	36	6	we	we	PRON
cana-2643	36	7	designed	design	VERB
cana-2643	36	8	a	a	DET
cana-2643	36	9	framework	framework	NOUN
cana-2643	36	10	,	,	PUNCT
cana-2643	36	11	with	with	ADP
cana-2643	36	12	multi	multi	ADJ
cana-2643	36	13	-	-	ADJ
cana-2643	36	14	path	path	ADJ
cana-2643	36	15	graph	graph	NOUN
cana-2643	36	16	attention	attention	NOUN
cana-2643	36	17	network	network	NOUN
cana-2643	36	18	with	with	ADP
cana-2643	36	19	reinforcement	reinforcement	NOUN
cana-2643	36	20	learning[9	learning[9	PROPN
cana-2643	36	21	]	]	X
cana-2643	36	22	(	(	PUNCT
cana-2643	36	23	mpga	mpga	NOUN
cana-2643	36	24	-	-	PUNCT
cana-2643	36	25	rl	rl	NOUN
cana-2643	36	26	-	-	PUNCT
cana-2643	36	27	net	net	NOUN
cana-2643	36	28	)	)	PUNCT
cana-2643	36	29	,	,	PUNCT
cana-2643	36	30	for	for	ADP
cana-2643	36	31	robotic	robotic	ADJ
cana-2643	36	32	vision	vision	NOUN
cana-2643	36	33	.	.	PUNCT
cana-2643	37	1	mpga	mpga	NOUN
cana-2643	37	2	-	-	PUNCT
cana-2643	37	3	rl	rl	NOUN
cana-2643	37	4	-	-	PUNCT
cana-2643	37	5	net	net	NOUN
cana-2643	37	6	integrates	integrate	NOUN
cana-2643	37	7	a	a	DET
cana-2643	37	8	multi	multi	ADJ
cana-2643	37	9	-	-	ADJ
cana-2643	37	10	scale	scale	ADJ
cana-2643	37	11	feature	feature	NOUN
cana-2643	37	12	extraction	extraction	NOUN
cana-2643	37	13	module	module	NOUN
cana-2643	37	14	with	with	ADP
cana-2643	37	15	graph	graph	NOUN
cana-2643	37	16	-	-	PUNCT
cana-2643	37	17	based	base	VERB
cana-2643	37	18	spatial	spatial	ADJ
cana-2643	37	19	modeling	modeling	NOUN
cana-2643	37	20	and	and	CCONJ
cana-2643	37	21	a	a	DET
cana-2643	37	22	reinforcement	reinforcement	NOUN
cana-2643	37	23	learning	learning	NOUN
cana-2643	37	24	-	-	PUNCT
cana-2643	37	25	driven	drive	VERB
cana-2643	37	26	attention	attention	NOUN
cana-2643	37	27	mechanism	mechanism	NOUN
cana-2643	37	28	that	that	PRON
cana-2643	37	29	enables	enable	VERB
cana-2643	37	30	real	real	ADJ
cana-2643	37	31	-	-	PUNCT
cana-2643	37	32	time	time	NOUN
cana-2643	37	33	prioritization	prioritization	NOUN
cana-2643	37	34	of	of	ADP
cana-2643	37	35	task	task	NOUN
cana-2643	37	36	-	-	PUNCT
cana-2643	37	37	relevant	relevant	ADJ
cana-2643	37	38	regions	region	NOUN
cana-2643	37	39	.	.	PUNCT
cana-2643	38	1	by	by	ADP
cana-2643	38	2	incorporating	incorporate	VERB
cana-2643	38	3	feedback	feedback	NOUN
cana-2643	38	4	-	-	PUNCT
cana-2643	38	5	based	base	VERB
cana-2643	38	6	reinforcement	reinforcement	NOUN
cana-2643	38	7	learning	learning	NOUN
cana-2643	38	8	,	,	PUNCT
cana-2643	38	9	mpga	mpga	NOUN
cana-2643	38	10	-	-	PUNCT
cana-2643	38	11	rl	rl	NOUN
cana-2643	38	12	-	-	PUNCT
cana-2643	38	13	net	net	NOUN
cana-2643	38	14	dynamically	dynamically	ADV
cana-2643	38	15	adjusts	adjust	VERB
cana-2643	38	16	its	its	PRON
cana-2643	38	17	attention	attention	NOUN
cana-2643	38	18	focus	focus	NOUN
cana-2643	38	19	,	,	PUNCT
cana-2643	38	20	concentrating	concentrate	VERB
cana-2643	38	21	on	on	ADP
cana-2643	38	22	critical	critical	ADJ
cana-2643	38	23	regions	region	NOUN
cana-2643	38	24	such	such	ADJ
cana-2643	38	25	as	as	ADP
cana-2643	38	26	navigable	navigable	ADJ
cana-2643	38	27	paths	path	NOUN
cana-2643	38	28	or	or	CCONJ
cana-2643	38	29	target	target	NOUN
cana-2643	38	30	objects	object	NOUN
cana-2643	38	31	,	,	PUNCT
cana-2643	38	32	which	which	PRON
cana-2643	38	33	enhances	enhance	VERB
cana-2643	38	34	adaptability	adaptability	NOUN
cana-2643	38	35	in	in	ADP
cana-2643	38	36	robotic	robotic	ADJ
cana-2643	38	37	vision	vision	NOUN
cana-2643	38	38	tasks	task	NOUN
cana-2643	38	39	.	.	PUNCT
cana-2643	39	1	objectives	objective	NOUN
cana-2643	39	2	of	of	ADP
cana-2643	39	3	the	the	DET
cana-2643	39	4	proposed	propose	VERB
cana-2643	39	5	work	work	NOUN
cana-2643	39	6	:	:	PUNCT
cana-2643	39	7	the	the	DET
cana-2643	39	8	primary	primary	ADJ
cana-2643	39	9	objective	objective	NOUN
cana-2643	39	10	of	of	ADP
cana-2643	39	11	this	this	DET
cana-2643	39	12	study	study	NOUN
cana-2643	39	13	is	be	AUX
cana-2643	39	14	to	to	PART
cana-2643	39	15	develop	develop	VERB
cana-2643	39	16	an	an	DET
cana-2643	39	17	adaptive	adaptive	ADJ
cana-2643	39	18	robotic	robotic	ADJ
cana-2643	39	19	vision	vision	NOUN
cana-2643	39	20	framework	framework	NOUN
cana-2643	39	21	capable	capable	ADJ
cana-2643	39	22	of	of	ADP
cana-2643	39	23	integrating	integrate	VERB
cana-2643	39	24	both	both	CCONJ
cana-2643	39	25	local	local	ADJ
cana-2643	39	26	and	and	CCONJ
cana-2643	39	27	global	global	ADJ
cana-2643	39	28	features	feature	NOUN
cana-2643	39	29	while	while	SCONJ
cana-2643	39	30	dynamically	dynamically	ADV
cana-2643	39	31	adjusting	adjust	VERB
cana-2643	39	32	focus	focus	NOUN
cana-2643	39	33	on	on	ADP
cana-2643	39	34	task	task	NOUN
cana-2643	39	35	-	-	PUNCT
cana-2643	39	36	specific	specific	ADJ
cana-2643	39	37	regions	region	NOUN
cana-2643	39	38	.	.	PUNCT
cana-2643	40	1	the	the	DET
cana-2643	40	2	specific	specific	ADJ
cana-2643	40	3	objectives	objective	NOUN
cana-2643	40	4	include	include	VERB
cana-2643	40	5	:	:	PUNCT
cana-2643	40	6	multi	multi	ADJ
cana-2643	40	7	-	-	ADJ
cana-2643	40	8	scale	scale	ADJ
cana-2643	40	9	feature	feature	NOUN
cana-2643	40	10	extraction	extraction	NOUN
cana-2643	40	11	:	:	PUNCT
cana-2643	40	12	to	to	PART
cana-2643	40	13	capture	capture	VERB
cana-2643	40	14	features	feature	NOUN
cana-2643	40	15	across	across	ADP
cana-2643	40	16	low	low	ADJ
cana-2643	40	17	,	,	PUNCT
cana-2643	40	18	medium	medium	ADJ
cana-2643	40	19	,	,	PUNCT
cana-2643	40	20	and	and	CCONJ
cana-2643	40	21	high	high	ADJ
cana-2643	40	22	resolutions	resolution	NOUN
cana-2643	40	23	using	use	VERB
cana-2643	40	24	a	a	DET
cana-2643	40	25	multi	multi	ADJ
cana-2643	40	26	-	-	ADJ
cana-2643	40	27	path	path	ADJ
cana-2643	40	28	feature	feature	NOUN
cana-2643	40	29	extraction	extraction	NOUN
cana-2643	40	30	(	(	PUNCT
cana-2643	40	31	mpfe	mpfe	NOUN
cana-2643	40	32	)	)	PUNCT
cana-2643	40	33	module	module	NOUN
cana-2643	40	34	,	,	PUNCT
cana-2643	40	35	which	which	PRON
cana-2643	40	36	enables	enable	VERB
cana-2643	40	37	the	the	DET
cana-2643	40	38	model	model	NOUN
cana-2643	40	39	to	to	PART
cana-2643	40	40	balance	balance	VERB
cana-2643	40	41	broad	broad	ADJ
cana-2643	40	42	contextual	contextual	ADJ
cana-2643	40	43	information	information	NOUN
cana-2643	40	44	with	with	ADP
cana-2643	40	45	finegrained	finegraine	VERB
cana-2643	40	46	details	detail	NOUN
cana-2643	40	47	essential	essential	ADJ
cana-2643	40	48	for	for	ADP
cana-2643	40	49	accurate	accurate	ADJ
cana-2643	40	50	segmentation	segmentation	NOUN
cana-2643	40	51	and	and	CCONJ
cana-2643	40	52	prioritization	prioritization	NOUN
cana-2643	40	53	.	.	PUNCT
cana-2643	41	1	graph	graph	NOUN
cana-2643	41	2	-	-	PUNCT
cana-2643	41	3	based	base	VERB
cana-2643	41	4	spatial	spatial	ADJ
cana-2643	41	5	relationship	relationship	NOUN
cana-2643	41	6	modeling	modeling	NOUN
cana-2643	41	7	:	:	PUNCT
cana-2643	41	8	to	to	PART
cana-2643	41	9	construct	construct	VERB
cana-2643	41	10	a	a	DET
cana-2643	41	11	graph	graph	NOUN
cana-2643	41	12	-	-	PUNCT
cana-2643	41	13	based	base	VERB
cana-2643	41	14	representation	representation	NOUN
cana-2643	41	15	using	use	VERB
cana-2643	41	16	super	super	ADJ
cana-2643	41	17	pixel	pixel	PROPN
cana-2643	41	18	segmentation	segmentation	NOUN
cana-2643	41	19	,	,	PUNCT
cana-2643	41	20	allowing	allow	VERB
cana-2643	41	21	the	the	DET
cana-2643	41	22	model	model	NOUN
cana-2643	41	23	to	to	PART
cana-2643	41	24	capture	capture	VERB
cana-2643	41	25	complex	complex	ADJ
cana-2643	41	26	spatial	spatial	ADJ
cana-2643	41	27	relationships	relationship	NOUN
cana-2643	41	28	among	among	ADP
cana-2643	41	29	regions	region	NOUN
cana-2643	41	30	for	for	ADP
cana-2643	41	31	improved	improved	ADJ
cana-2643	41	32	context	context	NOUN
cana-2643	41	33	in	in	ADP
cana-2643	41	34	robotic	robotic	ADJ
cana-2643	41	35	vision	vision	NOUN
cana-2643	41	36	tasks	task	NOUN
cana-2643	41	37	.	.	PUNCT
cana-2643	42	1	dynamic	dynamic	ADJ
cana-2643	42	2	attention	attention	NOUN
cana-2643	42	3	mechanism	mechanism	NOUN
cana-2643	42	4	via	via	ADP
cana-2643	42	5	reinforcement	reinforcement	NOUN
cana-2643	42	6	learning	learning	NOUN
cana-2643	42	7	:	:	PUNCT
cana-2643	42	8	to	to	PART
cana-2643	42	9	introduce	introduce	VERB
cana-2643	42	10	a	a	DET
cana-2643	42	11	reinforcement	reinforcement	NOUN
cana-2643	42	12	learning	learning	NOUN
cana-2643	42	13	component	component	NOUN
cana-2643	42	14	that	that	PRON
cana-2643	42	15	dynamically	dynamically	ADV
cana-2643	42	16	adjusts	adjust	VERB
cana-2643	42	17	attention	attention	NOUN
cana-2643	42	18	weights	weight	NOUN
cana-2643	42	19	,	,	PUNCT
cana-2643	42	20	enabling	enable	VERB
cana-2643	42	21	the	the	DET
cana-2643	42	22	model	model	NOUN
cana-2643	42	23	to	to	PART
cana-2643	42	24	emphasize	emphasize	VERB
cana-2643	42	25	critical	critical	ADJ
cana-2643	42	26	areas	area	NOUN
cana-2643	42	27	based	base	VERB
cana-2643	42	28	on	on	ADP
cana-2643	42	29	feedback	feedback	NOUN
cana-2643	42	30	in	in	ADP
cana-2643	42	31	real	real	ADJ
cana-2643	42	32	time[10	time[10	NOUN
cana-2643	42	33	]	]	PUNCT
cana-2643	42	34	.	.	PUNCT
cana-2643	43	1	precision	precision	NOUN
cana-2643	43	2	in	in	ADP
cana-2643	43	3	region	region	NOUN
cana-2643	43	4	-	-	PUNCT
cana-2643	43	5	specific	specific	ADJ
cana-2643	43	6	prioritization	prioritization	NOUN
cana-2643	43	7	:	:	PUNCT
cana-2643	43	8	to	to	PART
cana-2643	43	9	refine	refine	VERB
cana-2643	43	10	model	model	NOUN
cana-2643	43	11	focus	focus	NOUN
cana-2643	43	12	on	on	ADP
cana-2643	43	13	taskcritical	taskcritical	ADJ
cana-2643	43	14	regions	region	NOUN
cana-2643	43	15	,	,	PUNCT
cana-2643	43	16	especially	especially	ADV
cana-2643	43	17	in	in	ADP
cana-2643	43	18	complex	complex	ADJ
cana-2643	43	19	or	or	CCONJ
cana-2643	43	20	cluttered	cluttered	ADJ
cana-2643	43	21	environments	environment	NOUN
cana-2643	43	22	,	,	PUNCT
cana-2643	43	23	ensuring	ensure	VERB
cana-2643	43	24	that	that	SCONJ
cana-2643	43	25	region	region	NOUN
cana-2643	43	26	-	-	PUNCT
cana-2643	43	27	specific	specific	ADJ
cana-2643	43	28	prioritization	prioritization	NOUN
cana-2643	43	29	is	be	AUX
cana-2643	43	30	optimized	optimize	VERB
cana-2643	43	31	for	for	ADP
cana-2643	43	32	successful	successful	ADJ
cana-2643	43	33	task	task	NOUN
cana-2643	43	34	execution	execution	NOUN
cana-2643	43	35	2	2	NUM
cana-2643	43	36	.	.	PUNCT
cana-2643	43	37	related	relate	VERB
cana-2643	43	38	works	work	VERB
cana-2643	43	39	recent	recent	ADJ
cana-2643	43	40	advancements	advancement	NOUN
cana-2643	43	41	in	in	ADP
cana-2643	43	42	deep	deep	ADJ
cana-2643	43	43	learning	learning	NOUN
cana-2643	43	44	and	and	CCONJ
cana-2643	43	45	graph	graph	NOUN
cana-2643	43	46	-	-	PUNCT
cana-2643	43	47	based	base	VERB
cana-2643	43	48	neural	neural	ADJ
cana-2643	43	49	networks	network	NOUN
cana-2643	43	50	have	have	AUX
cana-2643	43	51	enabled	enable	VERB
cana-2643	43	52	significant	significant	ADJ
cana-2643	43	53	improvements	improvement	NOUN
cana-2643	43	54	in	in	ADP
cana-2643	43	55	robotic	robotic	ADJ
cana-2643	43	56	vision	vision	NOUN
cana-2643	43	57	,	,	PUNCT
cana-2643	43	58	yet	yet	CCONJ
cana-2643	43	59	challenges	challenge	NOUN
cana-2643	43	60	in	in	ADP
cana-2643	43	61	adaptive	adaptive	ADJ
cana-2643	43	62	region	region	NOUN
cana-2643	43	63	prioritization	prioritization	NOUN
cana-2643	43	64	and	and	CCONJ
cana-2643	43	65	multi	multi	ADJ
cana-2643	43	66	-	-	ADJ
cana-2643	43	67	scale	scale	ADJ
cana-2643	43	68	communications	communication	NOUN
cana-2643	43	69	on	on	ADP
cana-2643	43	70	applied	apply	VERB
cana-2643	43	71	nonlinear	nonlinear	ADJ
cana-2643	43	72	analysis	analysis	NOUN
cana-2643	43	73	issn	issn	NOUN
cana-2643	43	74	:	:	PUNCT
cana-2643	43	75	1074	1074	NUM
cana-2643	43	76	-	-	PUNCT
cana-2643	43	77	133x	133x	NUM
cana-2643	43	78	vol	vol	NOUN
cana-2643	43	79	32	32	NUM
cana-2643	44	1	no	no	NOUN
cana-2643	44	2	.	.	PUNCT
cana-2643	45	1	3s	3s	NUM
cana-2643	45	2	(	(	PUNCT
cana-2643	45	3	2025	2025	NUM
cana-2643	45	4	)	)	PUNCT
cana-2643	45	5	282	282	NUM
cana-2643	45	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	45	7	feature	feature	NOUN
cana-2643	45	8	integration	integration	NOUN
cana-2643	45	9	persist	persist	NOUN
cana-2643	45	10	.	.	PUNCT
cana-2643	46	1	this	this	DET
cana-2643	46	2	literature	literature	NOUN
cana-2643	46	3	survey	survey	NOUN
cana-2643	46	4	reviews	review	VERB
cana-2643	46	5	existing	exist	VERB
cana-2643	46	6	work	work	NOUN
cana-2643	46	7	in	in	ADP
cana-2643	46	8	cnn	cnn	PROPN
cana-2643	46	9	-	-	PUNCT
cana-2643	46	10	based	base	VERB
cana-2643	46	11	architectures	architecture	NOUN
cana-2643	46	12	,	,	PUNCT
cana-2643	46	13	attention	attention	NOUN
cana-2643	46	14	mechanisms	mechanism	NOUN
cana-2643	46	15	,	,	PUNCT
cana-2643	46	16	and	and	CCONJ
cana-2643	46	17	graph	graph	NOUN
cana-2643	46	18	-	-	PUNCT
cana-2643	46	19	based	base	VERB
cana-2643	46	20	approaches	approach	NOUN
cana-2643	46	21	,	,	PUNCT
cana-2643	46	22	highlighting	highlight	VERB
cana-2643	46	23	their	their	PRON
cana-2643	46	24	contributions	contribution	NOUN
cana-2643	46	25	and	and	CCONJ
cana-2643	46	26	limitations	limitation	NOUN
cana-2643	46	27	concerning	concern	VERB
cana-2643	46	28	robotic	robotic	ADJ
cana-2643	46	29	vision	vision	NOUN
cana-2643	46	30	tasks	task	NOUN
cana-2643	46	31	.	.	PUNCT
cana-2643	47	1	convolutional	convolutional	ADJ
cana-2643	47	2	neural	neural	ADJ
cana-2643	47	3	networks	network	NOUN
cana-2643	47	4	in	in	ADP
cana-2643	47	5	robotic	robotic	ADJ
cana-2643	47	6	vision	vision	NOUN
cana-2643	47	7	:	:	PUNCT
cana-2643	47	8	convolutional	convolutional	ADJ
cana-2643	47	9	neural	neural	ADJ
cana-2643	47	10	networks	network	NOUN
cana-2643	47	11	(	(	PUNCT
cana-2643	47	12	cnns	cnns	PROPN
cana-2643	47	13	)	)	PUNCT
cana-2643	47	14	have	have	AUX
cana-2643	47	15	been	be	AUX
cana-2643	47	16	foundational	foundational	ADJ
cana-2643	47	17	in	in	ADP
cana-2643	47	18	advancing	advance	VERB
cana-2643	47	19	robotic	robotic	ADJ
cana-2643	47	20	vision	vision	NOUN
cana-2643	47	21	due	due	ADP
cana-2643	47	22	to	to	ADP
cana-2643	47	23	their	their	PRON
cana-2643	47	24	powerful	powerful	ADJ
cana-2643	47	25	feature	feature	NOUN
cana-2643	47	26	extraction	extraction	NOUN
cana-2643	47	27	capabilities	capability	NOUN
cana-2643	47	28	across	across	ADP
cana-2643	47	29	various	various	ADJ
cana-2643	47	30	visual	visual	ADJ
cana-2643	47	31	domains	domain	NOUN
cana-2643	47	32	.	.	PUNCT
cana-2643	48	1	krizhevsky	krizhevsky	PROPN
cana-2643	48	2	et	et	PROPN
cana-2643	48	3	al	al	PROPN
cana-2643	48	4	.	.	PROPN
cana-2643	48	5	’s	’s	PART
cana-2643	48	6	work	work	NOUN
cana-2643	48	7	on	on	ADP
cana-2643	48	8	alexnet	alexnet	NOUN
cana-2643	48	9	demonstrated	demonstrate	VERB
cana-2643	48	10	the	the	DET
cana-2643	48	11	potential	potential	NOUN
cana-2643	48	12	of	of	ADP
cana-2643	48	13	deep	deep	ADJ
cana-2643	48	14	cnns	cnn	NOUN
cana-2643	48	15	for	for	ADP
cana-2643	48	16	high	high	ADJ
cana-2643	48	17	-	-	PUNCT
cana-2643	48	18	accuracy	accuracy	NOUN
cana-2643	48	19	object	object	NOUN
cana-2643	48	20	classification	classification	NOUN
cana-2643	48	21	,	,	PUNCT
cana-2643	48	22	setting	set	VERB
cana-2643	48	23	a	a	DET
cana-2643	48	24	precedent	precedent	NOUN
cana-2643	48	25	for	for	ADP
cana-2643	48	26	cnn	cnn	PROPN
cana-2643	48	27	application	application	NOUN
cana-2643	48	28	in	in	ADP
cana-2643	48	29	robotics	robotic	NOUN
cana-2643	48	30	.	.	PUNCT
cana-2643	49	1	since	since	SCONJ
cana-2643	49	2	then	then	ADV
cana-2643	49	3	,	,	PUNCT
cana-2643	49	4	numerous	numerous	ADJ
cana-2643	49	5	cnn	cnn	PROPN
cana-2643	49	6	-	-	PUNCT
cana-2643	49	7	based	base	VERB
cana-2643	49	8	architectures	architecture	NOUN
cana-2643	49	9	,	,	PUNCT
cana-2643	49	10	including	include	VERB
cana-2643	49	11	resnet	resnet	NOUN
cana-2643	49	12	and	and	CCONJ
cana-2643	49	13	vggnet	vggnet	NOUN
cana-2643	49	14	,	,	PUNCT
cana-2643	49	15	have	have	AUX
cana-2643	49	16	been	be	AUX
cana-2643	49	17	applied	apply	VERB
cana-2643	49	18	to	to	ADP
cana-2643	49	19	robotic	robotic	ADJ
cana-2643	49	20	vision	vision	NOUN
cana-2643	49	21	tasks	task	NOUN
cana-2643	49	22	.	.	PUNCT
cana-2643	50	1	these	these	DET
cana-2643	50	2	models	model	NOUN
cana-2643	50	3	effectively	effectively	ADV
cana-2643	50	4	capture	capture	VERB
cana-2643	50	5	hierarchical	hierarchical	ADJ
cana-2643	50	6	image	image	NOUN
cana-2643	50	7	features	feature	NOUN
cana-2643	50	8	,	,	PUNCT
cana-2643	50	9	but	but	CCONJ
cana-2643	50	10	they	they	PRON
cana-2643	50	11	rely	rely	VERB
cana-2643	50	12	on	on	ADP
cana-2643	50	13	static	static	ADJ
cana-2643	50	14	feature	feature	NOUN
cana-2643	50	15	extraction	extraction	NOUN
cana-2643	50	16	,	,	PUNCT
cana-2643	50	17	which	which	PRON
cana-2643	50	18	limits	limit	VERB
cana-2643	50	19	their	their	PRON
cana-2643	50	20	adaptability	adaptability	NOUN
cana-2643	50	21	to	to	ADP
cana-2643	50	22	task	task	NOUN
cana-2643	50	23	-	-	PUNCT
cana-2643	50	24	specific	specific	ADJ
cana-2643	50	25	focus	focus	NOUN
cana-2643	50	26	areas	area	NOUN
cana-2643	50	27	in	in	ADP
cana-2643	50	28	dynamic	dynamic	ADJ
cana-2643	50	29	scenes[11	scenes[11	NOUN
cana-2643	50	30	]	]	PUNCT
cana-2643	50	31	.	.	PUNCT
cana-2643	51	1	in	in	ADP
cana-2643	51	2	segmentation	segmentation	NOUN
cana-2643	51	3	tasks	task	NOUN
cana-2643	51	4	,	,	PUNCT
cana-2643	51	5	u	u	NOUN
cana-2643	51	6	-	-	NOUN
cana-2643	51	7	net	net	ADJ
cana-2643	51	8	and	and	CCONJ
cana-2643	51	9	segnet	segnet	NOUN
cana-2643	51	10	have	have	AUX
cana-2643	51	11	shown	show	VERB
cana-2643	51	12	impressive	impressive	ADJ
cana-2643	51	13	results	result	NOUN
cana-2643	51	14	in	in	ADP
cana-2643	51	15	preserving	preserve	VERB
cana-2643	51	16	spatial	spatial	ADJ
cana-2643	51	17	context	context	NOUN
cana-2643	51	18	through	through	ADP
cana-2643	51	19	encoder	encoder	NOUN
cana-2643	51	20	-	-	PUNCT
cana-2643	51	21	decoder	decoder	NOUN
cana-2643	51	22	architectures	architecture	NOUN
cana-2643	51	23	with	with	ADP
cana-2643	51	24	skip	skip	ADJ
cana-2643	51	25	connections	connection	NOUN
cana-2643	51	26	.	.	PUNCT
cana-2643	52	1	u	u	NOUN
cana-2643	52	2	-	-	NOUN
cana-2643	52	3	net	net	ADJ
cana-2643	52	4	,	,	PUNCT
cana-2643	52	5	for	for	ADP
cana-2643	52	6	example	example	NOUN
cana-2643	52	7	,	,	PUNCT
cana-2643	52	8	has	have	AUX
cana-2643	52	9	become	become	VERB
cana-2643	52	10	a	a	DET
cana-2643	52	11	preferred	preferred	ADJ
cana-2643	52	12	model	model	NOUN
cana-2643	52	13	in	in	ADP
cana-2643	52	14	medical	medical	ADJ
cana-2643	52	15	imaging	imaging	NOUN
cana-2643	52	16	due	due	ADP
cana-2643	52	17	to	to	ADP
cana-2643	52	18	its	its	PRON
cana-2643	52	19	ability	ability	NOUN
cana-2643	52	20	to	to	PART
cana-2643	52	21	capture	capture	VERB
cana-2643	52	22	fine	fine	ADV
cana-2643	52	23	-	-	PUNCT
cana-2643	52	24	grained	grain	VERB
cana-2643	52	25	details	detail	NOUN
cana-2643	52	26	while	while	SCONJ
cana-2643	52	27	retaining	retain	VERB
cana-2643	52	28	global	global	ADJ
cana-2643	52	29	context	context	NOUN
cana-2643	52	30	,	,	PUNCT
cana-2643	52	31	which	which	PRON
cana-2643	52	32	is	be	AUX
cana-2643	52	33	essential	essential	ADJ
cana-2643	52	34	for	for	ADP
cana-2643	52	35	boundary	boundary	ADJ
cana-2643	52	36	-	-	PUNCT
cana-2643	52	37	sensitive	sensitive	ADJ
cana-2643	52	38	tasks	task	NOUN
cana-2643	52	39	like	like	ADP
cana-2643	52	40	organ	organ	NOUN
cana-2643	52	41	segmentation	segmentation	NOUN
cana-2643	52	42	.	.	PUNCT
cana-2643	53	1	however	however	ADV
cana-2643	53	2	,	,	PUNCT
cana-2643	53	3	these	these	DET
cana-2643	53	4	models	model	NOUN
cana-2643	53	5	struggle	struggle	VERB
cana-2643	53	6	with	with	ADP
cana-2643	53	7	complex	complex	ADJ
cana-2643	53	8	environments	environment	NOUN
cana-2643	53	9	where	where	SCONJ
cana-2643	53	10	precise	precise	ADJ
cana-2643	53	11	,	,	PUNCT
cana-2643	53	12	real	real	ADJ
cana-2643	53	13	-	-	PUNCT
cana-2643	53	14	time	time	NOUN
cana-2643	53	15	focus	focus	NOUN
cana-2643	53	16	on	on	ADP
cana-2643	53	17	specific	specific	ADJ
cana-2643	53	18	regions	region	NOUN
cana-2643	53	19	is	be	AUX
cana-2643	53	20	critical	critical	ADJ
cana-2643	53	21	,	,	PUNCT
cana-2643	53	22	as	as	SCONJ
cana-2643	53	23	they	they	PRON
cana-2643	53	24	lack	lack	VERB
cana-2643	53	25	a	a	DET
cana-2643	53	26	mechanism	mechanism	NOUN
cana-2643	53	27	to	to	PART
cana-2643	53	28	adjust	adjust	VERB
cana-2643	53	29	focus	focus	VERB
cana-2643	53	30	dynamically	dynamically	ADV
cana-2643	53	31	.	.	PUNCT
cana-2643	54	1	according	accord	VERB
cana-2643	54	2	to	to	ADP
cana-2643	54	3	studies	study	NOUN
cana-2643	54	4	comparing	compare	VERB
cana-2643	54	5	u	u	NOUN
cana-2643	54	6	-	-	NOUN
cana-2643	54	7	net	net	ADJ
cana-2643	54	8	and	and	CCONJ
cana-2643	54	9	fcn	fcn	NOUN
cana-2643	54	10	models	model	NOUN
cana-2643	54	11	in	in	ADP
cana-2643	54	12	cluttered	cluttered	ADJ
cana-2643	54	13	robotic	robotic	ADJ
cana-2643	54	14	vision	vision	NOUN
cana-2643	54	15	environments	environment	NOUN
cana-2643	54	16	,	,	PUNCT
cana-2643	54	17	u	u	NOUN
cana-2643	54	18	-	-	NOUN
cana-2643	54	19	net	net	ADJ
cana-2643	54	20	’s	’s	PART
cana-2643	54	21	accuracy	accuracy	NOUN
cana-2643	54	22	declines	decline	VERB
cana-2643	54	23	by	by	ADP
cana-2643	54	24	up	up	ADP
cana-2643	54	25	to	to	PART
cana-2643	54	26	15	15	NUM
cana-2643	54	27	%	%	NOUN
cana-2643	54	28	in	in	ADP
cana-2643	54	29	handling	handle	VERB
cana-2643	54	30	irregular	irregular	ADJ
cana-2643	54	31	object	object	NOUN
cana-2643	54	32	boundaries	boundary	NOUN
cana-2643	54	33	.	.	PUNCT
cana-2643	55	1	this	this	DET
cana-2643	55	2	limitation	limitation	NOUN
cana-2643	55	3	necessitates	necessitate	VERB
cana-2643	55	4	further	further	ADJ
cana-2643	55	5	exploration	exploration	NOUN
cana-2643	55	6	into	into	ADP
cana-2643	55	7	adaptive	adaptive	ADJ
cana-2643	55	8	attention	attention	NOUN
cana-2643	55	9	mechanisms	mechanism	NOUN
cana-2643	55	10	that	that	PRON
cana-2643	55	11	can	can	AUX
cana-2643	55	12	prioritize	prioritize	VERB
cana-2643	55	13	region	region	NOUN
cana-2643	55	14	-	-	PUNCT
cana-2643	55	15	specific	specific	ADJ
cana-2643	55	16	focus	focus	NOUN
cana-2643	55	17	based	base	VERB
cana-2643	55	18	on	on	ADP
cana-2643	55	19	task	task	NOUN
cana-2643	55	20	demands	demand	NOUN
cana-2643	55	21	.	.	PUNCT
cana-2643	56	1	attention	attention	NOUN
cana-2643	56	2	mechanisms	mechanism	NOUN
cana-2643	56	3	for	for	ADP
cana-2643	56	4	adaptive	adaptive	ADJ
cana-2643	56	5	region	region	NOUN
cana-2643	56	6	prioritization	prioritization	NOUN
cana-2643	56	7	:	:	PUNCT
cana-2643	56	8	attention	attention	NOUN
cana-2643	56	9	mechanisms	mechanism	NOUN
cana-2643	56	10	have	have	AUX
cana-2643	56	11	been	be	AUX
cana-2643	56	12	integrated	integrate	VERB
cana-2643	56	13	into	into	ADP
cana-2643	56	14	neural	neural	ADJ
cana-2643	56	15	networks	network	NOUN
cana-2643	56	16	to	to	PART
cana-2643	56	17	address	address	VERB
cana-2643	56	18	the	the	DET
cana-2643	56	19	challenge	challenge	NOUN
cana-2643	56	20	of	of	ADP
cana-2643	56	21	identifying	identify	VERB
cana-2643	56	22	and	and	CCONJ
cana-2643	56	23	emphasizing	emphasize	VERB
cana-2643	56	24	important	important	ADJ
cana-2643	56	25	regions	region	NOUN
cana-2643	56	26	within	within	ADP
cana-2643	56	27	an	an	DET
cana-2643	56	28	image	image	NOUN
cana-2643	56	29	.	.	PUNCT
cana-2643	57	1	the	the	DET
cana-2643	57	2	"	"	PUNCT
cana-2643	57	3	attention	attention	NOUN
cana-2643	57	4	is	be	AUX
cana-2643	57	5	all	all	PRON
cana-2643	57	6	you	you	PRON
cana-2643	57	7	need	need	VERB
cana-2643	57	8	"	"	PUNCT
cana-2643	57	9	transformer	transformer	NOUN
cana-2643	57	10	model	model	NOUN
cana-2643	57	11	by	by	ADP
cana-2643	57	12	vaswani	vaswani	NOUN
cana-2643	57	13	et	et	NOUN
cana-2643	57	14	al.[12	al.[12	PROPN
cana-2643	57	15	]	]	PUNCT
cana-2643	57	16	was	be	AUX
cana-2643	57	17	a	a	DET
cana-2643	57	18	landmark	landmark	NOUN
cana-2643	57	19	in	in	ADP
cana-2643	57	20	this	this	DET
cana-2643	57	21	regard	regard	NOUN
cana-2643	57	22	,	,	PUNCT
cana-2643	57	23	demonstrating	demonstrate	VERB
cana-2643	57	24	how	how	SCONJ
cana-2643	57	25	attention	attention	NOUN
cana-2643	57	26	layers	layer	NOUN
cana-2643	57	27	can	can	AUX
cana-2643	57	28	focus	focus	VERB
cana-2643	57	29	computational	computational	ADJ
cana-2643	57	30	resources	resource	NOUN
cana-2643	57	31	on	on	ADP
cana-2643	57	32	critical	critical	ADJ
cana-2643	57	33	parts	part	NOUN
cana-2643	57	34	of	of	ADP
cana-2643	57	35	the	the	DET
cana-2643	57	36	input	input	NOUN
cana-2643	57	37	sequence	sequence	NOUN
cana-2643	57	38	.	.	PUNCT
cana-2643	58	1	this	this	DET
cana-2643	58	2	concept	concept	NOUN
cana-2643	58	3	has	have	AUX
cana-2643	58	4	since	since	ADV
cana-2643	58	5	been	be	AUX
cana-2643	58	6	applied	apply	VERB
cana-2643	58	7	to	to	ADP
cana-2643	58	8	vision	vision	NOUN
cana-2643	58	9	tasks	task	NOUN
cana-2643	58	10	in	in	ADP
cana-2643	58	11	models	model	NOUN
cana-2643	58	12	like	like	ADP
cana-2643	58	13	vision	vision	NOUN
cana-2643	58	14	transformers	transformer	NOUN
cana-2643	58	15	(	(	PUNCT
cana-2643	58	16	vits)[13	vits)[13	PROPN
cana-2643	58	17	]	]	PUNCT
cana-2643	58	18	,	,	PUNCT
cana-2643	58	19	where	where	SCONJ
cana-2643	58	20	attention	attention	NOUN
cana-2643	58	21	weights	weight	NOUN
cana-2643	58	22	are	be	AUX
cana-2643	58	23	computed	compute	VERB
cana-2643	58	24	to	to	PART
cana-2643	58	25	prioritize	prioritize	VERB
cana-2643	58	26	features	feature	NOUN
cana-2643	58	27	at	at	ADP
cana-2643	58	28	various	various	ADJ
cana-2643	58	29	levels	level	NOUN
cana-2643	58	30	.	.	PUNCT
cana-2643	59	1	although	although	SCONJ
cana-2643	59	2	transformers	transformer	NOUN
cana-2643	59	3	have	have	AUX
cana-2643	59	4	proven	prove	VERB
cana-2643	59	5	successful	successful	ADJ
cana-2643	59	6	in	in	ADP
cana-2643	59	7	enhancing	enhance	VERB
cana-2643	59	8	object	object	NOUN
cana-2643	59	9	detection	detection	NOUN
cana-2643	59	10	and	and	CCONJ
cana-2643	59	11	segmentation	segmentation	NOUN
cana-2643	59	12	accuracy	accuracy	NOUN
cana-2643	59	13	,	,	PUNCT
cana-2643	59	14	their	their	PRON
cana-2643	59	15	static	static	ADJ
cana-2643	59	16	attention	attention	NOUN
cana-2643	59	17	often	often	ADV
cana-2643	59	18	leads	lead	VERB
cana-2643	59	19	to	to	ADP
cana-2643	59	20	inefficiencies	inefficiency	NOUN
cana-2643	59	21	in	in	ADP
cana-2643	59	22	real	real	ADJ
cana-2643	59	23	-	-	PUNCT
cana-2643	59	24	time	time	NOUN
cana-2643	59	25	tasks	task	NOUN
cana-2643	59	26	where	where	SCONJ
cana-2643	59	27	attention	attention	NOUN
cana-2643	59	28	must	must	AUX
cana-2643	59	29	shift	shift	VERB
cana-2643	59	30	dynamically	dynamically	ADV
cana-2643	59	31	.	.	PUNCT
cana-2643	60	1	for	for	ADP
cana-2643	60	2	robotic	robotic	ADJ
cana-2643	60	3	vision	vision	NOUN
cana-2643	60	4	,	,	PUNCT
cana-2643	60	5	static	static	ADJ
cana-2643	60	6	attention	attention	NOUN
cana-2643	60	7	mechanisms	mechanism	NOUN
cana-2643	60	8	can	can	AUX
cana-2643	60	9	result	result	VERB
cana-2643	60	10	in	in	ADP
cana-2643	60	11	suboptimal	suboptimal	ADJ
cana-2643	60	12	focus	focus	NOUN
cana-2643	60	13	in	in	ADP
cana-2643	60	14	scenes	scene	NOUN
cana-2643	60	15	where	where	SCONJ
cana-2643	60	16	critical	critical	ADJ
cana-2643	60	17	regions	region	NOUN
cana-2643	60	18	change	change	VERB
cana-2643	60	19	with	with	ADP
cana-2643	60	20	context	context	NOUN
cana-2643	60	21	.	.	PUNCT
cana-2643	61	1	recent	recent	ADJ
cana-2643	61	2	studies	study	NOUN
cana-2643	61	3	have	have	AUX
cana-2643	61	4	applied	apply	VERB
cana-2643	61	5	multi	multi	ADJ
cana-2643	61	6	-	-	ADJ
cana-2643	61	7	scale	scale	ADJ
cana-2643	61	8	attention	attention	NOUN
cana-2643	61	9	networks	network	NOUN
cana-2643	61	10	for	for	ADP
cana-2643	61	11	object	object	NOUN
cana-2643	61	12	segmentation	segmentation	NOUN
cana-2643	61	13	,	,	PUNCT
cana-2643	61	14	which	which	PRON
cana-2643	61	15	adjust	adjust	VERB
cana-2643	61	16	attention	attention	NOUN
cana-2643	61	17	across	across	ADP
cana-2643	61	18	different	different	ADJ
cana-2643	61	19	feature	feature	NOUN
cana-2643	61	20	scales	scale	NOUN
cana-2643	61	21	.	.	PUNCT
cana-2643	62	1	such	such	ADJ
cana-2643	62	2	networks	network	NOUN
cana-2643	62	3	have	have	AUX
cana-2643	62	4	shown	show	VERB
cana-2643	62	5	a	a	DET
cana-2643	62	6	10–12	10–12	NUM
cana-2643	62	7	%	%	NOUN
cana-2643	62	8	increase	increase	NOUN
cana-2643	62	9	in	in	ADP
cana-2643	62	10	segmentation	segmentation	NOUN
cana-2643	62	11	accuracy	accuracy	NOUN
cana-2643	62	12	when	when	SCONJ
cana-2643	62	13	applied	apply	VERB
cana-2643	62	14	to	to	ADP
cana-2643	62	15	cluttered	cluttered	ADJ
cana-2643	62	16	environments	environment	NOUN
cana-2643	62	17	,	,	PUNCT
cana-2643	62	18	suggesting	suggest	VERB
cana-2643	62	19	the	the	DET
cana-2643	62	20	benefits	benefit	NOUN
cana-2643	62	21	of	of	ADP
cana-2643	62	22	multi	multi	ADJ
cana-2643	62	23	-	-	ADJ
cana-2643	62	24	scale	scale	ADJ
cana-2643	62	25	focus	focus	NOUN
cana-2643	62	26	.	.	PUNCT
cana-2643	63	1	however	however	ADV
cana-2643	63	2	,	,	PUNCT
cana-2643	63	3	the	the	DET
cana-2643	63	4	lack	lack	NOUN
cana-2643	63	5	of	of	ADP
cana-2643	63	6	a	a	DET
cana-2643	63	7	feedback	feedback	NOUN
cana-2643	63	8	-	-	PUNCT
cana-2643	63	9	driven	drive	VERB
cana-2643	63	10	mechanism	mechanism	NOUN
cana-2643	63	11	means	mean	VERB
cana-2643	63	12	these	these	DET
cana-2643	63	13	models	model	NOUN
cana-2643	63	14	can	can	AUX
cana-2643	63	15	not	not	PART
cana-2643	63	16	dynamically	dynamically	ADV
cana-2643	63	17	adapt	adapt	VERB
cana-2643	63	18	attention	attention	NOUN
cana-2643	63	19	based	base	VERB
cana-2643	63	20	on	on	ADP
cana-2643	63	21	evolving	evolve	VERB
cana-2643	63	22	scene	scene	NOUN
cana-2643	63	23	complexity	complexity	NOUN
cana-2643	63	24	.	.	PUNCT
cana-2643	64	1	this	this	DET
cana-2643	64	2	shortfall	shortfall	NOUN
cana-2643	64	3	underscores	underscore	VERB
cana-2643	64	4	the	the	DET
cana-2643	64	5	need	need	NOUN
cana-2643	64	6	for	for	ADP
cana-2643	64	7	reinforcement	reinforcement	NOUN
cana-2643	64	8	learning	learning	NOUN
cana-2643	64	9	-	-	PUNCT
cana-2643	64	10	based	base	VERB
cana-2643	64	11	attention	attention	NOUN
cana-2643	64	12	,	,	PUNCT
cana-2643	64	13	which	which	PRON
cana-2643	64	14	can	can	AUX
cana-2643	64	15	adapt	adapt	VERB
cana-2643	64	16	focus	focus	NOUN
cana-2643	64	17	based	base	VERB
cana-2643	64	18	on	on	ADP
cana-2643	64	19	task	task	NOUN
cana-2643	64	20	feedback	feedback	NOUN
cana-2643	64	21	,	,	PUNCT
cana-2643	64	22	thereby	thereby	ADV
cana-2643	64	23	improving	improve	VERB
cana-2643	64	24	performance	performance	NOUN
cana-2643	64	25	in	in	ADP
cana-2643	64	26	environments	environment	NOUN
cana-2643	64	27	where	where	SCONJ
cana-2643	64	28	target	target	NOUN
cana-2643	64	29	areas	area	NOUN
cana-2643	64	30	vary	vary	VERB
cana-2643	64	31	in	in	ADP
cana-2643	64	32	complexity	complexity	NOUN
cana-2643	64	33	and	and	CCONJ
cana-2643	64	34	importance	importance	NOUN
cana-2643	64	35	.	.	PUNCT
cana-2643	65	1	graph	graph	NOUN
cana-2643	65	2	-	-	PUNCT
cana-2643	65	3	based	base	VERB
cana-2643	65	4	models	model	NOUN
cana-2643	65	5	for	for	ADP
cana-2643	65	6	spatial	spatial	ADJ
cana-2643	65	7	relationship	relationship	NOUN
cana-2643	65	8	modeling	modeling	NOUN
cana-2643	65	9	:	:	PUNCT
cana-2643	65	10	graphbased	graphbased	ADJ
cana-2643	65	11	approaches	approach	NOUN
cana-2643	65	12	,	,	PUNCT
cana-2643	65	13	particularly	particularly	ADV
cana-2643	65	14	graph	graph	VERB
cana-2643	65	15	neural	neural	ADJ
cana-2643	65	16	networks	network	NOUN
cana-2643	65	17	(	(	PUNCT
cana-2643	65	18	gnns	gnns	NOUN
cana-2643	65	19	)	)	PUNCT
cana-2643	65	20	and	and	CCONJ
cana-2643	65	21	graph	graph	VERB
cana-2643	65	22	convolutional	convolutional	ADJ
cana-2643	65	23	networks	network	NOUN
cana-2643	65	24	(	(	PUNCT
cana-2643	65	25	gcns	gcns	PROPN
cana-2643	65	26	)	)	PUNCT
cana-2643	65	27	,	,	PUNCT
cana-2643	65	28	provide	provide	VERB
cana-2643	65	29	a	a	DET
cana-2643	65	30	structured	structured	ADJ
cana-2643	65	31	communications	communication	NOUN
cana-2643	65	32	on	on	ADP
cana-2643	65	33	applied	apply	VERB
cana-2643	65	34	nonlinear	nonlinear	ADJ
cana-2643	65	35	analysis	analysis	NOUN
cana-2643	65	36	issn	issn	NOUN
cana-2643	65	37	:	:	PUNCT
cana-2643	65	38	1074	1074	NUM
cana-2643	65	39	-	-	PUNCT
cana-2643	65	40	133x	133x	NUM
cana-2643	65	41	vol	vol	NOUN
cana-2643	65	42	32	32	NUM
cana-2643	65	43	no	no	NOUN
cana-2643	65	44	.	.	PUNCT
cana-2643	66	1	3s	3s	NUM
cana-2643	66	2	(	(	PUNCT
cana-2643	66	3	2025	2025	NUM
cana-2643	66	4	)	)	PUNCT
cana-2643	66	5	283	283	NUM
cana-2643	66	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	66	7	method	method	NOUN
cana-2643	66	8	for	for	ADP
cana-2643	66	9	modeling	model	VERB
cana-2643	66	10	spatial	spatial	ADJ
cana-2643	66	11	relationships	relationship	NOUN
cana-2643	66	12	between	between	ADP
cana-2643	66	13	objects	object	NOUN
cana-2643	66	14	in	in	ADP
cana-2643	66	15	an	an	DET
cana-2643	66	16	image	image	NOUN
cana-2643	66	17	.	.	PUNCT
cana-2643	67	1	kipf	kipf	PROPN
cana-2643	67	2	and	and	CCONJ
cana-2643	67	3	welling	well	VERB
cana-2643	67	4	’s	’s	PART
cana-2643	67	5	gcn	gcn	PROPN
cana-2643	67	6	model[14	model[14	PROPN
cana-2643	67	7	]	]	X
cana-2643	67	8	laid	lay	VERB
cana-2643	67	9	the	the	DET
cana-2643	67	10	groundwork	groundwork	NOUN
cana-2643	67	11	for	for	ADP
cana-2643	67	12	spatial	spatial	ADJ
cana-2643	67	13	graph	graph	NOUN
cana-2643	67	14	-	-	PUNCT
cana-2643	67	15	based	base	VERB
cana-2643	67	16	modeling	modeling	NOUN
cana-2643	67	17	by	by	ADP
cana-2643	67	18	introducing	introduce	VERB
cana-2643	67	19	convolutional	convolutional	ADJ
cana-2643	67	20	operations	operation	NOUN
cana-2643	67	21	on	on	ADP
cana-2643	67	22	graphs	graph	NOUN
cana-2643	67	23	,	,	PUNCT
cana-2643	67	24	enabling	enable	VERB
cana-2643	67	25	node	node	NOUN
cana-2643	67	26	-	-	PUNCT
cana-2643	67	27	level	level	NOUN
cana-2643	67	28	feature	feature	NOUN
cana-2643	67	29	aggregation	aggregation	NOUN
cana-2643	67	30	.	.	PUNCT
cana-2643	68	1	in	in	ADP
cana-2643	68	2	robotic	robotic	ADJ
cana-2643	68	3	vision	vision	NOUN
cana-2643	68	4	,	,	PUNCT
cana-2643	68	5	gcns	gcns	PROPN
cana-2643	68	6	have	have	AUX
cana-2643	68	7	shown	show	VERB
cana-2643	68	8	promise	promise	NOUN
cana-2643	68	9	for	for	ADP
cana-2643	68	10	tasks	task	NOUN
cana-2643	68	11	requiring	require	VERB
cana-2643	68	12	an	an	DET
cana-2643	68	13	understanding	understanding	NOUN
cana-2643	68	14	of	of	ADP
cana-2643	68	15	spatial	spatial	ADJ
cana-2643	68	16	dependencies	dependency	NOUN
cana-2643	68	17	between	between	ADP
cana-2643	68	18	objects	object	NOUN
cana-2643	68	19	,	,	PUNCT
cana-2643	68	20	such	such	ADJ
cana-2643	68	21	as	as	ADP
cana-2643	68	22	in	in	ADP
cana-2643	68	23	scene	scene	NOUN
cana-2643	68	24	graph	graph	NOUN
cana-2643	68	25	generation	generation	NOUN
cana-2643	68	26	for	for	ADP
cana-2643	68	27	object	object	NOUN
cana-2643	68	28	manipulation	manipulation	NOUN
cana-2643	68	29	.	.	PUNCT
cana-2643	69	1	by	by	ADP
cana-2643	69	2	representing	represent	VERB
cana-2643	69	3	regions	region	NOUN
cana-2643	69	4	as	as	ADP
cana-2643	69	5	graph	graph	NOUN
cana-2643	69	6	nodes	node	NOUN
cana-2643	69	7	,	,	PUNCT
cana-2643	69	8	gcns	gcns	PROPN
cana-2643	69	9	can	can	AUX
cana-2643	69	10	model	model	VERB
cana-2643	69	11	relationships	relationship	NOUN
cana-2643	69	12	between	between	ADP
cana-2643	69	13	regions	region	NOUN
cana-2643	69	14	,	,	PUNCT
cana-2643	69	15	making	make	VERB
cana-2643	69	16	them	they	PRON
cana-2643	69	17	particularly	particularly	ADV
cana-2643	69	18	useful	useful	ADJ
cana-2643	69	19	in	in	ADP
cana-2643	69	20	applications	application	NOUN
cana-2643	69	21	like	like	ADP
cana-2643	69	22	navigation	navigation	NOUN
cana-2643	69	23	or	or	CCONJ
cana-2643	69	24	object	object	VERB
cana-2643	69	25	interaction	interaction	NOUN
cana-2643	69	26	,	,	PUNCT
cana-2643	69	27	where	where	SCONJ
cana-2643	69	28	understanding	understand	VERB
cana-2643	69	29	spatial	spatial	ADJ
cana-2643	69	30	structure	structure	NOUN
cana-2643	69	31	is	be	AUX
cana-2643	69	32	critical	critical	ADJ
cana-2643	69	33	.	.	PUNCT
cana-2643	70	1	in	in	ADP
cana-2643	70	2	medical	medical	ADJ
cana-2643	70	3	imaging	imaging	NOUN
cana-2643	70	4	,	,	PUNCT
cana-2643	70	5	gcns	gcns	PROPN
cana-2643	70	6	have	have	AUX
cana-2643	70	7	achieved	achieve	VERB
cana-2643	70	8	a	a	DET
cana-2643	70	9	7–9	7–9	NUM
cana-2643	70	10	%	%	NOUN
cana-2643	70	11	improvement	improvement	NOUN
cana-2643	70	12	in	in	ADP
cana-2643	70	13	boundary	boundary	ADJ
cana-2643	70	14	detection	detection	NOUN
cana-2643	70	15	tasks	task	NOUN
cana-2643	70	16	over	over	ADP
cana-2643	70	17	cnn	cnn	PROPN
cana-2643	70	18	-	-	PUNCT
cana-2643	70	19	based	base	VERB
cana-2643	70	20	models	model	NOUN
cana-2643	70	21	by	by	ADP
cana-2643	70	22	better	well	ADV
cana-2643	70	23	capturing	capture	VERB
cana-2643	70	24	spatial	spatial	ADJ
cana-2643	70	25	relationships	relationship	NOUN
cana-2643	70	26	between	between	ADP
cana-2643	70	27	neighboring	neighboring	NOUN
cana-2643	70	28	regions	region	NOUN
cana-2643	70	29	.	.	PUNCT
cana-2643	71	1	however	however	ADV
cana-2643	71	2	,	,	PUNCT
cana-2643	71	3	conventional	conventional	ADJ
cana-2643	71	4	gcns	gcns	PROPN
cana-2643	71	5	still	still	ADV
cana-2643	71	6	face	face	VERB
cana-2643	71	7	limitations	limitation	NOUN
cana-2643	71	8	in	in	ADP
cana-2643	71	9	dynamically	dynamically	ADV
cana-2643	71	10	adjusting	adjust	VERB
cana-2643	71	11	the	the	DET
cana-2643	71	12	graph	graph	NOUN
cana-2643	71	13	structure	structure	NOUN
cana-2643	71	14	based	base	VERB
cana-2643	71	15	on	on	ADP
cana-2643	71	16	realtime	realtime	ADJ
cana-2643	71	17	feedback	feedback	NOUN
cana-2643	71	18	,	,	PUNCT
cana-2643	71	19	which	which	PRON
cana-2643	71	20	restricts	restrict	VERB
cana-2643	71	21	their	their	PRON
cana-2643	71	22	adaptability	adaptability	NOUN
cana-2643	71	23	to	to	ADP
cana-2643	71	24	dynamic	dynamic	ADJ
cana-2643	71	25	robotic	robotic	ADJ
cana-2643	71	26	vision	vision	NOUN
cana-2643	71	27	tasks	task	NOUN
cana-2643	71	28	.	.	PUNCT
cana-2643	72	1	to	to	PART
cana-2643	72	2	address	address	VERB
cana-2643	72	3	this	this	PRON
cana-2643	72	4	,	,	PUNCT
cana-2643	72	5	nguyen	nguyen	NOUN
cana-2643	72	6	et	et	PROPN
cana-2643	72	7	al	al	PROPN
cana-2643	72	8	.	.	PROPN
cana-2643	72	9	proposed	propose	VERB
cana-2643	72	10	a	a	DET
cana-2643	72	11	multi	multi	ADJ
cana-2643	72	12	-	-	ADJ
cana-2643	72	13	scale	scale	ADJ
cana-2643	72	14	gcn	gcn	NOUN
cana-2643	72	15	with	with	ADP
cana-2643	72	16	an	an	DET
cana-2643	72	17	attention	attention	NOUN
cana-2643	72	18	mechanism	mechanism	NOUN
cana-2643	72	19	for	for	ADP
cana-2643	72	20	semantic	semantic	ADJ
cana-2643	72	21	segmentation	segmentation	NOUN
cana-2643	72	22	,	,	PUNCT
cana-2643	72	23	which	which	PRON
cana-2643	72	24	improved	improve	VERB
cana-2643	72	25	iou	iou	NOUN
cana-2643	72	26	scores	score	NOUN
cana-2643	72	27	by	by	ADP
cana-2643	72	28	approximately	approximately	ADV
cana-2643	72	29	8	8	NUM
cana-2643	72	30	%	%	NOUN
cana-2643	72	31	over	over	ADP
cana-2643	72	32	standard	standard	PROPN
cana-2643	72	33	gcns	gcns	PROPN
cana-2643	72	34	in	in	ADP
cana-2643	72	35	complex	complex	ADJ
cana-2643	72	36	visual	visual	ADJ
cana-2643	72	37	environments	environment	NOUN
cana-2643	72	38	.	.	PUNCT
cana-2643	73	1	while	while	SCONJ
cana-2643	73	2	this	this	DET
cana-2643	73	3	model	model	NOUN
cana-2643	73	4	improved	improve	VERB
cana-2643	73	5	segmentation	segmentation	NOUN
cana-2643	73	6	by	by	ADP
cana-2643	73	7	integrating	integrate	VERB
cana-2643	73	8	multi	multi	ADJ
cana-2643	73	9	-	-	ADJ
cana-2643	73	10	scale	scale	ADJ
cana-2643	73	11	information	information	NOUN
cana-2643	73	12	,	,	PUNCT
cana-2643	73	13	it	it	PRON
cana-2643	73	14	lacked	lack	VERB
cana-2643	73	15	a	a	DET
cana-2643	73	16	reinforcement	reinforcement	NOUN
cana-2643	73	17	learning	learning	NOUN
cana-2643	73	18	mechanism	mechanism	NOUN
cana-2643	73	19	that	that	PRON
cana-2643	73	20	could	could	AUX
cana-2643	73	21	adapt	adapt	VERB
cana-2643	73	22	attention	attention	NOUN
cana-2643	73	23	based	base	VERB
cana-2643	73	24	on	on	ADP
cana-2643	73	25	task	task	NOUN
cana-2643	73	26	-	-	PUNCT
cana-2643	73	27	specific	specific	ADJ
cana-2643	73	28	performance	performance	NOUN
cana-2643	73	29	,	,	PUNCT
cana-2643	73	30	highlighting	highlight	VERB
cana-2643	73	31	the	the	DET
cana-2643	73	32	gap	gap	NOUN
cana-2643	73	33	for	for	ADP
cana-2643	73	34	reinforcement	reinforcement	NOUN
cana-2643	73	35	-	-	PUNCT
cana-2643	73	36	driven	drive	VERB
cana-2643	73	37	graph	graph	NOUN
cana-2643	73	38	attention	attention	NOUN
cana-2643	73	39	models[15	models[15	NOUN
cana-2643	73	40	]	]	PUNCT
cana-2643	73	41	.	.	PUNCT
cana-2643	74	1	reinforcement	reinforcement	NOUN
cana-2643	74	2	learning	learning	NOUN
cana-2643	74	3	in	in	ADP
cana-2643	74	4	robotic	robotic	ADJ
cana-2643	74	5	vision	vision	NOUN
cana-2643	74	6	:	:	PUNCT
cana-2643	74	7	reinforcement	reinforcement	NOUN
cana-2643	74	8	learning	learning	NOUN
cana-2643	74	9	(	(	PUNCT
cana-2643	74	10	rl	rl	NOUN
cana-2643	74	11	)	)	PUNCT
cana-2643	74	12	has	have	AUX
cana-2643	74	13	gained	gain	VERB
cana-2643	74	14	attention	attention	NOUN
cana-2643	74	15	as	as	ADP
cana-2643	74	16	a	a	DET
cana-2643	74	17	method	method	NOUN
cana-2643	74	18	for	for	ADP
cana-2643	74	19	enabling	enable	VERB
cana-2643	74	20	robots	robot	NOUN
cana-2643	74	21	to	to	PART
cana-2643	74	22	learn	learn	VERB
cana-2643	74	23	from	from	ADP
cana-2643	74	24	interaction	interaction	NOUN
cana-2643	74	25	with	with	ADP
cana-2643	74	26	their	their	PRON
cana-2643	74	27	environment	environment	NOUN
cana-2643	74	28	,	,	PUNCT
cana-2643	74	29	making	make	VERB
cana-2643	74	30	it	it	PRON
cana-2643	74	31	well	well	ADV
cana-2643	74	32	-	-	PUNCT
cana-2643	74	33	suited	suit	VERB
cana-2643	74	34	for	for	ADP
cana-2643	74	35	tasks	task	NOUN
cana-2643	74	36	requiring	require	VERB
cana-2643	74	37	adaptive	adaptive	ADJ
cana-2643	74	38	decision	decision	NOUN
cana-2643	74	39	-	-	PUNCT
cana-2643	74	40	making	making	NOUN
cana-2643	74	41	.	.	PUNCT
cana-2643	75	1	mnih	mnih	PROPN
cana-2643	75	2	et	et	PROPN
cana-2643	75	3	al	al	PROPN
cana-2643	75	4	.	.	PROPN
cana-2643	75	5	introduced	introduce	VERB
cana-2643	75	6	deep	deep	ADJ
cana-2643	75	7	q	q	NOUN
cana-2643	75	8	-	-	PUNCT
cana-2643	75	9	networks	network	NOUN
cana-2643	75	10	(	(	PUNCT
cana-2643	75	11	dqn	dqn	PROPN
cana-2643	75	12	)	)	PUNCT
cana-2643	75	13	,	,	PUNCT
cana-2643	75	14	which	which	PRON
cana-2643	75	15	combined	combine	VERB
cana-2643	75	16	deep	deep	ADJ
cana-2643	75	17	learning	learning	NOUN
cana-2643	75	18	with	with	ADP
cana-2643	75	19	rl	rl	PRON
cana-2643	75	20	to	to	PART
cana-2643	75	21	achieve	achieve	VERB
cana-2643	75	22	human	human	ADJ
cana-2643	75	23	-	-	PUNCT
cana-2643	75	24	level	level	NOUN
cana-2643	75	25	performance	performance	NOUN
cana-2643	75	26	in	in	ADP
cana-2643	75	27	video	video	NOUN
cana-2643	75	28	game	game	NOUN
cana-2643	75	29	environments	environment	NOUN
cana-2643	75	30	.	.	PUNCT
cana-2643	76	1	this	this	DET
cana-2643	76	2	breakthrough	breakthrough	NOUN
cana-2643	76	3	established	establish	VERB
cana-2643	76	4	rl	rl	PRON
cana-2643	76	5	as	as	ADP
cana-2643	76	6	a	a	DET
cana-2643	76	7	viable	viable	ADJ
cana-2643	76	8	approach	approach	NOUN
cana-2643	76	9	for	for	ADP
cana-2643	76	10	decision	decision	NOUN
cana-2643	76	11	-	-	PUNCT
cana-2643	76	12	based	base	VERB
cana-2643	76	13	adaptation	adaptation	NOUN
cana-2643	76	14	in	in	ADP
cana-2643	76	15	robotics	robotic	NOUN
cana-2643	76	16	.	.	PUNCT
cana-2643	77	1	in	in	ADP
cana-2643	77	2	the	the	DET
cana-2643	77	3	context	context	NOUN
cana-2643	77	4	of	of	ADP
cana-2643	77	5	robotic	robotic	ADJ
cana-2643	77	6	vision	vision	NOUN
cana-2643	77	7	,	,	PUNCT
cana-2643	77	8	rl	rl	PROPN
cana-2643	77	9	has	have	AUX
cana-2643	77	10	been	be	AUX
cana-2643	77	11	applied	apply	VERB
cana-2643	77	12	to	to	PART
cana-2643	77	13	optimize	optimize	VERB
cana-2643	77	14	task	task	NOUN
cana-2643	77	15	-	-	PUNCT
cana-2643	77	16	specific	specific	ADJ
cana-2643	77	17	attention	attention	NOUN
cana-2643	77	18	,	,	PUNCT
cana-2643	77	19	allowing	allow	VERB
cana-2643	77	20	robots	robot	NOUN
cana-2643	77	21	to	to	PART
cana-2643	77	22	dynamically	dynamically	ADV
cana-2643	77	23	adjust	adjust	VERB
cana-2643	77	24	their	their	PRON
cana-2643	77	25	focus	focus	NOUN
cana-2643	77	26	based	base	VERB
cana-2643	77	27	on	on	ADP
cana-2643	77	28	real	real	ADJ
cana-2643	77	29	-	-	PUNCT
cana-2643	77	30	time	time	NOUN
cana-2643	77	31	feedback	feedback	NOUN
cana-2643	77	32	.	.	PUNCT
cana-2643	78	1	for	for	ADP
cana-2643	78	2	instance	instance	NOUN
cana-2643	78	3	,	,	PUNCT
cana-2643	78	4	rl	rl	NOUN
cana-2643	78	5	-	-	PUNCT
cana-2643	78	6	based	base	VERB
cana-2643	78	7	attention	attention	NOUN
cana-2643	78	8	mechanisms	mechanism	NOUN
cana-2643	78	9	have	have	AUX
cana-2643	78	10	shown	show	VERB
cana-2643	78	11	success	success	NOUN
cana-2643	78	12	in	in	ADP
cana-2643	78	13	robotic	robotic	ADJ
cana-2643	78	14	navigation	navigation	NOUN
cana-2643	78	15	tasks	task	NOUN
cana-2643	78	16	by	by	ADP
cana-2643	78	17	improving	improve	VERB
cana-2643	78	18	the	the	DET
cana-2643	78	19	focus	focus	NOUN
cana-2643	78	20	on	on	ADP
cana-2643	78	21	relevant	relevant	ADJ
cana-2643	78	22	obstacles	obstacle	NOUN
cana-2643	78	23	and	and	CCONJ
cana-2643	78	24	pathways	pathway	NOUN
cana-2643	78	25	,	,	PUNCT
cana-2643	78	26	with	with	ADP
cana-2643	78	27	studies	study	NOUN
cana-2643	78	28	reporting	report	VERB
cana-2643	78	29	up	up	ADV
cana-2643	78	30	to	to	PART
cana-2643	78	31	20	20	NUM
cana-2643	78	32	%	%	NOUN
cana-2643	78	33	increases	increase	NOUN
cana-2643	78	34	in	in	ADP
cana-2643	78	35	navigation	navigation	NOUN
cana-2643	78	36	accuracy	accuracy	NOUN
cana-2643	78	37	over	over	ADP
cana-2643	78	38	static	static	ADJ
cana-2643	78	39	attention	attention	NOUN
cana-2643	78	40	models	model	NOUN
cana-2643	78	41	.	.	PUNCT
cana-2643	79	1	integrating	integrate	VERB
cana-2643	79	2	rl	rl	NOUN
cana-2643	79	3	with	with	ADP
cana-2643	79	4	attention	attention	NOUN
cana-2643	79	5	and	and	CCONJ
cana-2643	79	6	gcns	gcns	PROPN
cana-2643	79	7	could	could	AUX
cana-2643	79	8	potentially	potentially	ADV
cana-2643	79	9	offer	offer	VERB
cana-2643	79	10	substantial	substantial	ADJ
cana-2643	79	11	improvements	improvement	NOUN
cana-2643	79	12	in	in	ADP
cana-2643	79	13	real	real	ADJ
cana-2643	79	14	-	-	PUNCT
cana-2643	79	15	time	time	NOUN
cana-2643	79	16	robotic	robotic	ADJ
cana-2643	79	17	vision	vision	NOUN
cana-2643	79	18	,	,	PUNCT
cana-2643	79	19	enabling	enable	VERB
cana-2643	79	20	models	model	NOUN
cana-2643	79	21	to	to	PART
cana-2643	79	22	prioritize	prioritize	VERB
cana-2643	79	23	regions	region	NOUN
cana-2643	79	24	based	base	VERB
cana-2643	79	25	on	on	ADP
cana-2643	79	26	continuous	continuous	ADJ
cana-2643	79	27	feedback	feedback	NOUN
cana-2643	79	28	.	.	PUNCT
cana-2643	80	1	this	this	DET
cana-2643	80	2	approach	approach	NOUN
cana-2643	80	3	is	be	AUX
cana-2643	80	4	promising	promise	VERB
cana-2643	80	5	for	for	ADP
cana-2643	80	6	tasks	task	NOUN
cana-2643	80	7	such	such	ADJ
cana-2643	80	8	as	as	ADP
cana-2643	80	9	object	object	NOUN
cana-2643	80	10	manipulation	manipulation	NOUN
cana-2643	80	11	in	in	ADP
cana-2643	80	12	cluttered	cluttered	ADJ
cana-2643	80	13	environments	environment	NOUN
cana-2643	80	14	,	,	PUNCT
cana-2643	80	15	where	where	SCONJ
cana-2643	80	16	adaptive	adaptive	ADJ
cana-2643	80	17	focus	focus	NOUN
cana-2643	80	18	is	be	AUX
cana-2643	80	19	critical	critical	ADJ
cana-2643	80	20	for	for	ADP
cana-2643	80	21	identifying	identify	VERB
cana-2643	80	22	optimal	optimal	ADJ
cana-2643	80	23	grasp	grasp	NOUN
cana-2643	80	24	points	point	NOUN
cana-2643	80	25	or	or	CCONJ
cana-2643	80	26	avoiding	avoid	VERB
cana-2643	80	27	obstacles	obstacle	NOUN
cana-2643	80	28	.	.	PUNCT
cana-2643	81	1	statistical	statistical	ADJ
cana-2643	81	2	analysis	analysis	NOUN
cana-2643	81	3	of	of	ADP
cana-2643	81	4	robotic	robotic	ADJ
cana-2643	81	5	vision	vision	NOUN
cana-2643	81	6	models	model	NOUN
cana-2643	81	7	:	:	PUNCT
cana-2643	81	8	a	a	DET
cana-2643	81	9	comparative	comparative	ADJ
cana-2643	81	10	analysis	analysis	NOUN
cana-2643	81	11	of	of	ADP
cana-2643	81	12	cnns	cnn	NOUN
cana-2643	81	13	,	,	PUNCT
cana-2643	81	14	attention	attention	NOUN
cana-2643	81	15	-	-	PUNCT
cana-2643	81	16	based	base	VERB
cana-2643	81	17	models	model	NOUN
cana-2643	81	18	,	,	PUNCT
cana-2643	81	19	and	and	CCONJ
cana-2643	81	20	gcns	gcns	PROPN
cana-2643	81	21	highlights	highlight	VERB
cana-2643	81	22	their	their	PRON
cana-2643	81	23	respective	respective	ADJ
cana-2643	81	24	strengths	strength	NOUN
cana-2643	81	25	and	and	CCONJ
cana-2643	81	26	limitations	limitation	NOUN
cana-2643	81	27	.	.	PUNCT
cana-2643	82	1	cnns	cnns	PROPN
cana-2643	82	2	have	have	AUX
cana-2643	82	3	achieved	achieve	VERB
cana-2643	82	4	object	object	NOUN
cana-2643	82	5	detection	detection	NOUN
cana-2643	82	6	accuracies	accuracy	NOUN
cana-2643	82	7	upwards	upwards	ADV
cana-2643	82	8	of	of	ADP
cana-2643	82	9	90	90	NUM
cana-2643	82	10	%	%	NOUN
cana-2643	82	11	in	in	ADP
cana-2643	82	12	structured	structured	ADJ
cana-2643	82	13	environments	environment	NOUN
cana-2643	82	14	,	,	PUNCT
cana-2643	82	15	but	but	CCONJ
cana-2643	82	16	performance	performance	NOUN
cana-2643	82	17	drops	drop	VERB
cana-2643	82	18	in	in	ADP
cana-2643	82	19	complex	complex	ADJ
cana-2643	82	20	,	,	PUNCT
cana-2643	82	21	unstructured	unstructured	ADJ
cana-2643	82	22	scenes	scene	NOUN
cana-2643	82	23	due	due	ADJ
cana-2643	82	24	to	to	ADP
cana-2643	82	25	limited	limited	ADJ
cana-2643	82	26	adaptability	adaptability	NOUN
cana-2643	82	27	.	.	PUNCT
cana-2643	83	1	gcns	gcns	PROPN
cana-2643	83	2	,	,	PUNCT
cana-2643	83	3	on	on	ADP
cana-2643	83	4	the	the	DET
cana-2643	83	5	other	other	ADJ
cana-2643	83	6	hand	hand	NOUN
cana-2643	83	7	,	,	PUNCT
cana-2643	83	8	have	have	AUX
cana-2643	83	9	demonstrated	demonstrate	VERB
cana-2643	83	10	superior	superior	ADJ
cana-2643	83	11	spatial	spatial	ADJ
cana-2643	83	12	modeling	modeling	NOUN
cana-2643	83	13	capabilities	capability	NOUN
cana-2643	83	14	,	,	PUNCT
cana-2643	83	15	achieving	achieve	VERB
cana-2643	83	16	iou	iou	NOUN
cana-2643	83	17	improvements	improvement	NOUN
cana-2643	83	18	of	of	ADP
cana-2643	83	19	7–10	7–10	NUM
cana-2643	83	20	%	%	NOUN
cana-2643	83	21	in	in	ADP
cana-2643	83	22	cluttered	cluttered	ADJ
cana-2643	83	23	scenes	scene	NOUN
cana-2643	83	24	compared	compare	VERB
cana-2643	83	25	to	to	ADP
cana-2643	83	26	cnns	cnn	NOUN
cana-2643	83	27	.	.	PUNCT
cana-2643	84	1	multi	multi	ADJ
cana-2643	84	2	-	-	ADJ
cana-2643	84	3	scale	scale	ADJ
cana-2643	84	4	attention	attention	NOUN
cana-2643	84	5	models	model	NOUN
cana-2643	84	6	offer	offer	VERB
cana-2643	84	7	advantages	advantage	NOUN
cana-2643	84	8	in	in	ADP
cana-2643	84	9	dynamic	dynamic	ADJ
cana-2643	84	10	environments	environment	NOUN
cana-2643	84	11	,	,	PUNCT
cana-2643	84	12	with	with	ADP
cana-2643	84	13	studies	study	NOUN
cana-2643	84	14	reporting	report	VERB
cana-2643	84	15	a	a	DET
cana-2643	84	16	12	12	NUM
cana-2643	84	17	%	%	NOUN
cana-2643	84	18	increase	increase	NOUN
cana-2643	84	19	in	in	ADP
cana-2643	84	20	accuracy	accuracy	NOUN
cana-2643	84	21	when	when	SCONJ
cana-2643	84	22	attention	attention	NOUN
cana-2643	84	23	is	be	AUX
cana-2643	84	24	applied	apply	VERB
cana-2643	84	25	across	across	ADP
cana-2643	84	26	multiple	multiple	ADJ
cana-2643	84	27	scales	scale	NOUN
cana-2643	84	28	.	.	PUNCT
cana-2643	85	1	however	however	ADV
cana-2643	85	2	,	,	PUNCT
cana-2643	85	3	models	model	NOUN
cana-2643	85	4	that	that	PRON
cana-2643	85	5	combine	combine	VERB
cana-2643	85	6	multi	multi	ADJ
cana-2643	85	7	-	-	ADJ
cana-2643	85	8	scale	scale	ADJ
cana-2643	85	9	feature	feature	NOUN
cana-2643	85	10	extraction	extraction	NOUN
cana-2643	85	11	with	with	ADP
cana-2643	85	12	dynamic	dynamic	ADJ
cana-2643	85	13	reinforcement	reinforcement	NOUN
cana-2643	85	14	learn	learn	VERB
cana-2643	85	15	ing[16	ing[16	PROPN
cana-2643	85	16	]	]	PUNCT
cana-2643	85	17	are	be	AUX
cana-2643	85	18	expected	expect	VERB
cana-2643	85	19	to	to	PART
cana-2643	85	20	surpass	surpass	VERB
cana-2643	85	21	these	these	DET
cana-2643	85	22	baselines	baseline	NOUN
cana-2643	85	23	,	,	PUNCT
cana-2643	85	24	as	as	SCONJ
cana-2643	85	25	they	they	PRON
cana-2643	85	26	can	can	AUX
cana-2643	85	27	continuously	continuously	ADV
cana-2643	85	28	adjust	adjust	VERB
cana-2643	85	29	focus	focus	NOUN
cana-2643	85	30	based	base	VERB
cana-2643	85	31	on	on	ADP
cana-2643	85	32	feedback	feedback	NOUN
cana-2643	85	33	,	,	PUNCT
cana-2643	85	34	improving	improve	VERB
cana-2643	85	35	performance	performance	NOUN
cana-2643	85	36	across	across	ADP
cana-2643	85	37	various	various	ADJ
cana-2643	85	38	robotic	robotic	ADJ
cana-2643	85	39	vision	vision	NOUN
cana-2643	85	40	tasks	task	NOUN
cana-2643	85	41	.	.	PUNCT
cana-2643	86	1	communications	communication	NOUN
cana-2643	86	2	on	on	ADP
cana-2643	86	3	applied	apply	VERB
cana-2643	86	4	nonlinear	nonlinear	ADJ
cana-2643	86	5	analysis	analysis	NOUN
cana-2643	86	6	issn	issn	NOUN
cana-2643	86	7	:	:	PUNCT
cana-2643	86	8	1074	1074	NUM
cana-2643	86	9	-	-	PUNCT
cana-2643	86	10	133x	133x	NUM
cana-2643	86	11	vol	vol	NOUN
cana-2643	86	12	32	32	NUM
cana-2643	86	13	no	no	NOUN
cana-2643	86	14	.	.	PUNCT
cana-2643	87	1	3s	3s	NUM
cana-2643	87	2	(	(	PUNCT
cana-2643	87	3	2025	2025	NUM
cana-2643	87	4	)	)	PUNCT
cana-2643	87	5	284	284	NUM
cana-2643	87	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	87	7	table	table	NOUN
cana-2643	87	8	1	1	NUM
cana-2643	87	9	:	:	PUNCT
cana-2643	87	10	summary	summary	NOUN
cana-2643	87	11	of	of	ADP
cana-2643	87	12	robotic	robotic	ADJ
cana-2643	87	13	vision	vision	NOUN
cana-2643	87	14	models	model	NOUN
cana-2643	87	15	:	:	PUNCT
cana-2643	87	16	advantages	advantage	NOUN
cana-2643	87	17	and	and	CCONJ
cana-2643	87	18	limitations	limitation	NOUN
cana-2643	87	19	model	model	NOUN
cana-2643	87	20	accuracy	accuracy	NOUN
cana-2643	87	21	advantages	advantage	NOUN
cana-2643	87	22	limitations	limitation	VERB
cana-2643	87	23	cnn	cnn	PROPN
cana-2643	87	24	(	(	PUNCT
cana-2643	87	25	base	base	NOUN
cana-2643	87	26	line	line	NOUN
cana-2643	87	27	)	)	PUNCT
cana-2643	88	1	[	[	X
cana-2643	88	2	17	17	NUM
cana-2643	88	3	]	]	SYM
cana-2643	88	4	85	85	NUM
cana-2643	88	5	%	%	NOUN
cana-2643	88	6	(	(	PUNCT
cana-2643	88	7	object	object	NOUN
cana-2643	88	8	detection	detection	NOUN
cana-2643	88	9	)	)	PUNCT
cana-2643	88	10	strong	strong	ADJ
cana-2643	88	11	feature	feature	NOUN
cana-2643	88	12	extraction	extraction	NOUN
cana-2643	88	13	,	,	PUNCT
cana-2643	88	14	simple	simple	ADJ
cana-2643	88	15	architecture	architecture	NOUN
cana-2643	88	16	,	,	PUNCT
cana-2643	88	17	and	and	CCONJ
cana-2643	88	18	effective	effective	ADJ
cana-2643	88	19	for	for	ADP
cana-2643	88	20	general	general	ADJ
cana-2643	88	21	image	image	NOUN
cana-2643	88	22	classification	classification	NOUN
cana-2643	88	23	.	.	PUNCT
cana-2643	89	1	limited	limited	ADJ
cana-2643	89	2	daptability	daptability	NOUN
cana-2643	89	3	;	;	PUNCT
cana-2643	89	4	struggles	struggle	VERB
cana-2643	89	5	in	in	ADP
cana-2643	89	6	complex	complex	ADJ
cana-2643	89	7	scenes	scene	NOUN
cana-2643	89	8	with	with	ADP
cana-2643	89	9	dynamic	dynamic	ADJ
cana-2643	89	10	objects	object	NOUN
cana-2643	89	11	and	and	CCONJ
cana-2643	89	12	lacks	lack	VERB
cana-2643	89	13	multiscale	multiscale	ADJ
cana-2643	89	14	focus	focus	NOUN
cana-2643	89	15	.	.	PUNCT
cana-2643	90	1	alex	alex	PROPN
cana-2643	90	2	net	net	PROPN
cana-2643	91	1	[	[	X
cana-2643	91	2	18	18	NUM
cana-2643	91	3	]	]	SYM
cana-2643	91	4	82	82	NUM
cana-2643	91	5	%	%	NOUN
cana-2643	91	6	(	(	PUNCT
cana-2643	91	7	image	image	NOUN
cana-2643	91	8	net	net	ADJ
cana-2643	91	9	classification	classification	NOUN
cana-2643	91	10	)	)	PUNCT
cana-2643	91	11	demonstrated	demonstrate	VERB
cana-2643	91	12	potential	potential	NOUN
cana-2643	91	13	of	of	ADP
cana-2643	91	14	deep	deep	ADJ
cana-2643	91	15	learning	learning	NOUN
cana-2643	91	16	for	for	ADP
cana-2643	91	17	large	large	ADJ
cana-2643	91	18	-	-	PUNCT
cana-2643	91	19	scale	scale	NOUN
cana-2643	91	20	classification	classification	NOUN
cana-2643	91	21	;	;	PUNCT
cana-2643	91	22	efficient	efficient	ADJ
cana-2643	91	23	for	for	ADP
cana-2643	91	24	early	early	ADJ
cana-2643	91	25	deep	deep	ADJ
cana-2643	91	26	learning	learning	NOUN
cana-2643	91	27	.	.	PUNCT
cana-2643	92	1	high	high	ADJ
cana-2643	92	2	computational	computational	ADJ
cana-2643	92	3	requirements	requirement	NOUN
cana-2643	92	4	;	;	PUNCT
cana-2643	92	5	lacks	lack	VERB
cana-2643	92	6	ability	ability	NOUN
cana-2643	92	7	to	to	PART
cana-2643	92	8	handle	handle	VERB
cana-2643	92	9	intricate	intricate	ADJ
cana-2643	92	10	spatial	spatial	ADJ
cana-2643	92	11	dependencies	dependency	NOUN
cana-2643	92	12	.	.	PUNCT
cana-2643	93	1	vgg	vgg	ADJ
cana-2643	93	2	net	net	NOUN
cana-2643	93	3	[	[	X
cana-2643	93	4	19	19	NUM
cana-2643	93	5	]	]	SYM
cana-2643	93	6	86	86	NUM
cana-2643	93	7	%	%	NOUN
cana-2643	93	8	(	(	PUNCT
cana-2643	93	9	image	image	NOUN
cana-2643	93	10	net	net	ADJ
cana-2643	93	11	classification	classification	NOUN
cana-2643	93	12	)	)	PUNCT
cana-2643	93	13	simplified	simplify	VERB
cana-2643	93	14	design	design	NOUN
cana-2643	93	15	with	with	ADP
cana-2643	93	16	improved	improved	ADJ
cana-2643	93	17	depth	depth	NOUN
cana-2643	93	18	for	for	ADP
cana-2643	93	19	more	more	ADV
cana-2643	93	20	accurate	accurate	ADJ
cana-2643	93	21	image	image	NOUN
cana-2643	93	22	classification	classification	NOUN
cana-2643	93	23	.	.	PUNCT
cana-2643	94	1	computationally	computationally	ADV
cana-2643	94	2	expensive	expensive	ADJ
cana-2643	94	3	due	due	ADJ
cana-2643	94	4	to	to	ADP
cana-2643	94	5	high	high	ADJ
cana-2643	94	6	parameter	parameter	NOUN
cana-2643	94	7	count	count	NOUN
cana-2643	94	8	;	;	PUNCT
cana-2643	94	9	memory	memory	NOUN
cana-2643	94	10	-	-	PUNCT
cana-2643	94	11	intensive	intensive	ADJ
cana-2643	94	12	and	and	CCONJ
cana-2643	94	13	slower	slow	ADJ
cana-2643	94	14	.	.	PUNCT
cana-2643	95	1	res	re	NOUN
cana-2643	95	2	net	net	NOUN
cana-2643	96	1	[	[	X
cana-2643	96	2	20	20	NUM
cana-2643	96	3	]	]	SYM
cana-2643	96	4	88	88	NUM
cana-2643	96	5	%	%	NOUN
cana-2643	96	6	(	(	PUNCT
cana-2643	96	7	image	image	NOUN
cana-2643	96	8	net	net	ADJ
cana-2643	96	9	classification	classification	NOUN
cana-2643	96	10	)	)	PUNCT
cana-2643	96	11	introduced	introduce	VERB
cana-2643	96	12	residual	residual	ADJ
cana-2643	96	13	connections	connection	NOUN
cana-2643	96	14	,	,	PUNCT
cana-2643	96	15	making	make	VERB
cana-2643	96	16	it	it	PRON
cana-2643	96	17	possible	possible	ADJ
cana-2643	96	18	to	to	PART
cana-2643	96	19	train	train	VERB
cana-2643	96	20	very	very	ADV
cana-2643	96	21	deep	deep	ADJ
cana-2643	96	22	networks	network	NOUN
cana-2643	96	23	effectively	effectively	ADV
cana-2643	96	24	.	.	PUNCT
cana-2643	97	1	high	high	ADJ
cana-2643	97	2	memory	memory	NOUN
cana-2643	97	3	and	and	CCONJ
cana-2643	97	4	processing	processing	NOUN
cana-2643	97	5	demands	demand	NOUN
cana-2643	97	6	;	;	PUNCT
cana-2643	97	7	limited	limited	ADJ
cana-2643	97	8	adaptability	adaptability	NOUN
cana-2643	97	9	in	in	ADP
cana-2643	97	10	real	real	ADJ
cana-2643	97	11	-	-	PUNCT
cana-2643	97	12	time	time	NOUN
cana-2643	97	13	and	and	CCONJ
cana-2643	97	14	dynamic	dynamic	ADJ
cana-2643	97	15	scene	scene	NOUN
cana-2643	97	16	contexts	contexts	NOUN
cana-2643	97	17	.	.	PUNCT
cana-2643	98	1	u	u	NOUN
cana-2643	98	2	-	-	NOUN
cana-2643	98	3	net	net	ADJ
cana-2643	98	4	[	[	X
cana-2643	98	5	21	21	NUM
cana-2643	98	6	]	]	SYM
cana-2643	98	7	80	80	NUM
cana-2643	98	8	%	%	NOUN
cana-2643	98	9	(	(	PUNCT
cana-2643	98	10	segmentation	segmentation	NOUN
cana-2643	98	11	iou	iou	NOUN
cana-2643	98	12	)	)	PUNCT
cana-2643	98	13	effective	effective	ADJ
cana-2643	98	14	encoder	encoder	NOUN
cana-2643	98	15	decoder	decoder	NOUN
cana-2643	98	16	structure	structure	NOUN
cana-2643	98	17	,	,	PUNCT
cana-2643	98	18	widely	widely	ADV
cana-2643	98	19	used	use	VERB
cana-2643	98	20	in	in	ADP
cana-2643	98	21	biomedical	biomedical	ADJ
cana-2643	98	22	segmentation	segmentation	NOUN
cana-2643	98	23	,	,	PUNCT
cana-2643	98	24	captures	capture	VERB
cana-2643	98	25	local	local	ADJ
cana-2643	98	26	and	and	CCONJ
cana-2643	98	27	global	global	ADJ
cana-2643	98	28	info	info	NOUN
cana-2643	98	29	.	.	PUNCT
cana-2643	99	1	less	less	ADV
cana-2643	99	2	effective	effective	ADJ
cana-2643	99	3	for	for	ADP
cana-2643	99	4	small	small	ADJ
cana-2643	99	5	and	and	CCONJ
cana-2643	99	6	irregular	irregular	ADJ
cana-2643	99	7	objects	object	NOUN
cana-2643	99	8	in	in	ADP
cana-2643	99	9	complex	complex	ADJ
cana-2643	99	10	or	or	CCONJ
cana-2643	99	11	cluttered	cluttered	ADJ
cana-2643	99	12	environments	environment	NOUN
cana-2643	99	13	.	.	PUNCT
cana-2643	100	1	segnet	segnet	PROPN
cana-2643	101	1	[	[	X
cana-2643	101	2	22	22	NUM
cana-2643	101	3	]	]	SYM
cana-2643	101	4	78	78	NUM
cana-2643	101	5	%	%	NOUN
cana-2643	101	6	(	(	PUNCT
cana-2643	101	7	segmentation	segmentation	NOUN
cana-2643	101	8	iou	iou	NOUN
cana-2643	101	9	)	)	PUNCT
cana-2643	101	10	maintains	maintain	VERB
cana-2643	101	11	spatial	spatial	ADJ
cana-2643	101	12	resolution	resolution	NOUN
cana-2643	101	13	,	,	PUNCT
cana-2643	101	14	suitable	suitable	ADJ
cana-2643	101	15	for	for	ADP
cana-2643	101	16	segmentation	segmentation	NOUN
cana-2643	101	17	tasks	task	NOUN
cana-2643	101	18	where	where	SCONJ
cana-2643	101	19	boundary	boundary	ADJ
cana-2643	101	20	preservation	preservation	NOUN
cana-2643	101	21	is	be	AUX
cana-2643	101	22	key	key	ADJ
cana-2643	101	23	.	.	PUNCT
cana-2643	102	1	limited	limited	ADJ
cana-2643	102	2	adaptability	adaptability	NOUN
cana-2643	102	3	;	;	PUNCT
cana-2643	102	4	struggles	struggle	NOUN
cana-2643	102	5	in	in	ADP
cana-2643	102	6	low	low	ADJ
cana-2643	102	7	contrast	contrast	NOUN
cana-2643	102	8	images	image	NOUN
cana-2643	102	9	and	and	CCONJ
cana-2643	102	10	lacks	lack	VERB
cana-2643	102	11	dynamic	dynamic	ADJ
cana-2643	102	12	region	region	NOUN
cana-2643	102	13	prioritization	prioritization	NOUN
cana-2643	102	14	.	.	PUNCT
cana-2643	103	1	fully	fully	ADV
cana-2643	103	2	convolutional	convolutional	ADJ
cana-2643	103	3	net	net	ADJ
cana-2643	103	4	work	work	NOUN
cana-2643	103	5	(	(	PUNCT
cana-2643	103	6	fcn	fcn	NOUN
cana-2643	103	7	)	)	PUNCT
cana-2643	104	1	[	[	X
cana-2643	104	2	23	23	NUM
cana-2643	104	3	]	]	SYM
cana-2643	104	4	77	77	NUM
cana-2643	104	5	%	%	NOUN
cana-2643	104	6	(	(	PUNCT
cana-2643	104	7	segmentation	segmentation	NOUN
cana-2643	104	8	iou	iou	NOUN
cana-2643	104	9	)	)	PUNCT
cana-2643	104	10	pioneered	pioneer	VERB
cana-2643	104	11	end	end	NOUN
cana-2643	104	12	-	-	PUNCT
cana-2643	104	13	to	to	ADP
cana-2643	104	14	-	-	PUNCT
cana-2643	104	15	end	end	NOUN
cana-2643	104	16	segmentation	segmentation	NOUN
cana-2643	104	17	;	;	PUNCT
cana-2643	104	18	enables	enable	VERB
cana-2643	104	19	dense	dense	ADJ
cana-2643	104	20	predictions	prediction	NOUN
cana-2643	104	21	for	for	ADP
cana-2643	104	22	entire	entire	ADJ
cana-2643	104	23	image	image	NOUN
cana-2643	104	24	.	.	PUNCT
cana-2643	105	1	less	less	ADV
cana-2643	105	2	effective	effective	ADJ
cana-2643	105	3	at	at	ADP
cana-2643	105	4	capturing	capture	VERB
cana-2643	105	5	complex	complex	ADJ
cana-2643	105	6	spatial	spatial	ADJ
cana-2643	105	7	dependencies	dependency	NOUN
cana-2643	105	8	and	and	CCONJ
cana-2643	105	9	prioritizing	prioritize	VERB
cana-2643	105	10	important	important	ADJ
cana-2643	105	11	regions	region	NOUN
cana-2643	105	12	.	.	PUNCT
cana-2643	106	1	graph	graph	VERB
cana-2643	106	2	convolutional	convolutional	ADJ
cana-2643	106	3	network	network	NOUN
cana-2643	106	4	(	(	PUNCT
cana-2643	106	5	gcn	gcn	NOUN
cana-2643	106	6	)	)	PUNCT
cana-2643	107	1	[	[	X
cana-2643	107	2	24	24	NUM
cana-2643	107	3	]	]	SYM
cana-2643	107	4	82	82	NUM
cana-2643	107	5	%	%	NOUN
cana-2643	107	6	(	(	PUNCT
cana-2643	107	7	segmentation	segmentation	NOUN
cana-2643	107	8	iou	iou	NOUN
cana-2643	107	9	)	)	PUNCT
cana-2643	107	10	models	model	NOUN
cana-2643	107	11	spatial	spatial	ADJ
cana-2643	107	12	relationships	relationship	NOUN
cana-2643	107	13	using	use	VERB
cana-2643	107	14	graph	graph	NOUN
cana-2643	107	15	structures	structure	NOUN
cana-2643	107	16	;	;	PUNCT
cana-2643	107	17	static	static	ADJ
cana-2643	107	18	structure	structure	NOUN
cana-2643	107	19	;	;	PUNCT
cana-2643	107	20	lacks	lack	VERB
cana-2643	107	21	ability	ability	NOUN
cana-2643	107	22	to	to	PART
cana-2643	107	23	dynamically	dynamically	ADV
cana-2643	107	24	adjust	adjust	VERB
cana-2643	107	25	attention	attention	NOUN
cana-2643	107	26	based	base	VERB
cana-2643	107	27	on	on	ADP
cana-2643	107	28	communications	communication	NOUN
cana-2643	107	29	on	on	ADP
cana-2643	107	30	applied	apply	VERB
cana-2643	107	31	nonlinear	nonlinear	ADJ
cana-2643	107	32	analysis	analysis	NOUN
cana-2643	107	33	issn	issn	NOUN
cana-2643	107	34	:	:	PUNCT
cana-2643	107	35	1074	1074	NUM
cana-2643	107	36	-	-	PUNCT
cana-2643	107	37	133x	133x	NUM
cana-2643	107	38	vol	vol	NOUN
cana-2643	107	39	32	32	NUM
cana-2643	107	40	no	no	NOUN
cana-2643	107	41	.	.	PUNCT
cana-2643	108	1	3s	3s	NUM
cana-2643	108	2	(	(	PUNCT
cana-2643	108	3	2025	2025	NUM
cana-2643	108	4	)	)	PUNCT
cana-2643	108	5	285	285	NUM
cana-2643	108	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	108	7	suitable	suitable	ADJ
cana-2643	108	8	for	for	ADP
cana-2643	108	9	applications	application	NOUN
cana-2643	108	10	requiring	require	VERB
cana-2643	108	11	object	object	NOUN
cana-2643	108	12	proximity	proximity	NOUN
cana-2643	108	13	.	.	PUNCT
cana-2643	109	1	task	task	NOUN
cana-2643	109	2	-	-	PUNCT
cana-2643	109	3	specific	specific	ADJ
cana-2643	109	4	focus	focus	NOUN
cana-2643	109	5	needs	need	NOUN
cana-2643	109	6	.	.	PUNCT
cana-2643	110	1	vision	vision	PROPN
cana-2643	110	2	transformer	transformer	PROPN
cana-2643	110	3	(	(	PUNCT
cana-2643	110	4	vit	vit	NOUN
cana-2643	110	5	)	)	PUNCT
cana-2643	111	1	[	[	X
cana-2643	111	2	25	25	NUM
cana-2643	111	3	]	]	SYM
cana-2643	111	4	85	85	NUM
cana-2643	111	5	%	%	NOUN
cana-2643	111	6	(	(	PUNCT
cana-2643	111	7	object	object	VERB
cana-2643	111	8	detec	detec	ADJ
cana-2643	111	9	tion	tion	NOUN
cana-2643	111	10	)	)	PUNCT
cana-2643	111	11	uses	use	VERB
cana-2643	111	12	self	self	NOUN
cana-2643	111	13	-	-	PUNCT
cana-2643	111	14	attention	attention	NOUN
cana-2643	111	15	for	for	ADP
cana-2643	111	16	capturing	capture	VERB
cana-2643	111	17	complex	complex	ADJ
cana-2643	111	18	dependencies	dependency	NOUN
cana-2643	111	19	;	;	PUNCT
cana-2643	111	20	performs	perform	VERB
cana-2643	111	21	well	well	ADV
cana-2643	111	22	on	on	ADP
cana-2643	111	23	large	large	ADJ
cana-2643	111	24	datasets	dataset	NOUN
cana-2643	111	25	.	.	PUNCT
cana-2643	112	1	high	high	ADJ
cana-2643	112	2	computational	computational	ADJ
cana-2643	112	3	load	load	NOUN
cana-2643	112	4	;	;	PUNCT
cana-2643	112	5	requires	require	VERB
cana-2643	112	6	extensive	extensive	ADJ
cana-2643	112	7	training	training	NOUN
cana-2643	112	8	data	datum	NOUN
cana-2643	112	9	and	and	CCONJ
cana-2643	112	10	struggles	struggle	NOUN
cana-2643	112	11	with	with	ADP
cana-2643	112	12	real	real	ADJ
cana-2643	112	13	-	-	PUNCT
cana-2643	112	14	time	time	NOUN
cana-2643	112	15	adjustments	adjustment	NOUN
cana-2643	112	16	.	.	PUNCT
cana-2643	113	1	multi	multi	ADJ
cana-2643	113	2	-	-	ADJ
cana-2643	113	3	scale	scale	ADJ
cana-2643	113	4	attention	attention	NOUN
cana-2643	113	5	network	network	NOUN
cana-2643	113	6	[	[	X
cana-2643	113	7	26	26	NUM
cana-2643	113	8	]	]	SYM
cana-2643	113	9	88	88	NUM
cana-2643	113	10	%	%	NOUN
cana-2643	113	11	(	(	PUNCT
cana-2643	113	12	segmentation	segmentation	NOUN
cana-2643	113	13	iou	iou	NOUN
cana-2643	113	14	)	)	PUNCT
cana-2643	113	15	adapts	adapt	VERB
cana-2643	113	16	attention	attention	NOUN
cana-2643	113	17	at	at	ADP
cana-2643	113	18	multiple	multiple	ADJ
cana-2643	113	19	scales	scale	NOUN
cana-2643	113	20	,	,	PUNCT
cana-2643	113	21	improving	improve	VERB
cana-2643	113	22	focus	focus	NOUN
cana-2643	113	23	on	on	ADP
cana-2643	113	24	key	key	ADJ
cana-2643	113	25	regions	region	NOUN
cana-2643	113	26	and	and	CCONJ
cana-2643	113	27	enhancing	enhance	VERB
cana-2643	113	28	segmentation	segmentation	NOUN
cana-2643	113	29	accuracy	accuracy	NOUN
cana-2643	113	30	.	.	PUNCT
cana-2643	114	1	fixed	fix	VERB
cana-2643	114	2	attention	attention	NOUN
cana-2643	114	3	;	;	PUNCT
cana-2643	114	4	lacks	lack	VERB
cana-2643	114	5	reinforcement	reinforcement	NOUN
cana-2643	114	6	learning	learning	NOUN
cana-2643	114	7	,	,	PUNCT
cana-2643	114	8	making	make	VERB
cana-2643	114	9	it	it	PRON
cana-2643	114	10	less	less	ADV
cana-2643	114	11	adaptive	adaptive	ADJ
cana-2643	114	12	to	to	ADP
cana-2643	114	13	dynamic	dynamic	ADJ
cana-2643	114	14	scenes	scene	NOUN
cana-2643	114	15	and	and	CCONJ
cana-2643	114	16	tasks	task	NOUN
cana-2643	114	17	.	.	PUNCT
cana-2643	115	1	multi	multi	ADJ
cana-2643	115	2	-	-	ADJ
cana-2643	115	3	scale	scale	ADJ
cana-2643	115	4	gcn	gcn	NOUN
cana-2643	115	5	[	[	X
cana-2643	115	6	27	27	NUM
cana-2643	115	7	]	]	SYM
cana-2643	115	8	88	88	NUM
cana-2643	115	9	%	%	NOUN
cana-2643	115	10	(	(	PUNCT
cana-2643	115	11	segmentation	segmentation	NOUN
cana-2643	115	12	iou	iou	NOUN
cana-2643	115	13	)	)	PUNCT
cana-2643	115	14	combines	combine	VERB
cana-2643	115	15	multi	multi	ADJ
cana-2643	115	16	-	-	ADJ
cana-2643	115	17	scale	scale	ADJ
cana-2643	115	18	graph	graph	NOUN
cana-2643	115	19	and	and	CCONJ
cana-2643	115	20	attention	attention	NOUN
cana-2643	115	21	for	for	ADP
cana-2643	115	22	improved	improve	VERB
cana-2643	115	23	spatial	spatial	ADJ
cana-2643	115	24	relationships	relationship	NOUN
cana-2643	115	25	and	and	CCONJ
cana-2643	115	26	seg	seg	NOUN
cana-2643	115	27	mentation	mentation	NOUN
cana-2643	115	28	in	in	ADP
cana-2643	115	29	cluttered	cluttered	ADJ
cana-2643	115	30	scenes	scene	NOUN
cana-2643	115	31	.	.	PUNCT
cana-2643	116	1	limited	limited	ADJ
cana-2643	116	2	real	real	ADJ
cana-2643	116	3	-	-	PUNCT
cana-2643	116	4	time	time	NOUN
cana-2643	116	5	adaptability	adaptability	NOUN
cana-2643	116	6	;	;	PUNCT
cana-2643	116	7	fixed	fix	VERB
cana-2643	116	8	attention	attention	NOUN
cana-2643	116	9	weights	weight	NOUN
cana-2643	116	10	can	can	AUX
cana-2643	116	11	lead	lead	VERB
cana-2643	116	12	to	to	ADP
cana-2643	116	13	inefficiencies	inefficiency	NOUN
cana-2643	116	14	when	when	SCONJ
cana-2643	116	15	scenes	scene	NOUN
cana-2643	116	16	change	change	VERB
cana-2643	116	17	.	.	PUNCT
cana-2643	117	1	squeezeandexcitation	squeezeandexcitation	NOUN
cana-2643	117	2	network	network	NOUN
cana-2643	117	3	[	[	X
cana-2643	117	4	28	28	NUM
cana-2643	117	5	]	]	SYM
cana-2643	117	6	87	87	NUM
cana-2643	117	7	%	%	NOUN
cana-2643	117	8	(	(	PUNCT
cana-2643	117	9	image	image	NOUN
cana-2643	117	10	classification	classification	NOUN
cana-2643	117	11	)	)	PUNCT
cana-2643	117	12	efficient	efficient	ADJ
cana-2643	117	13	channel	channel	NOUN
cana-2643	117	14	attention	attention	NOUN
cana-2643	117	15	boosts	boost	VERB
cana-2643	117	16	classification	classification	NOUN
cana-2643	117	17	accuracy	accuracy	NOUN
cana-2643	117	18	by	by	ADP
cana-2643	117	19	focusing	focus	VERB
cana-2643	117	20	on	on	ADP
cana-2643	117	21	key	key	ADJ
cana-2643	117	22	features	feature	NOUN
cana-2643	117	23	.	.	PUNCT
cana-2643	118	1	primarily	primarily	ADV
cana-2643	118	2	beneficial	beneficial	ADJ
cana-2643	118	3	for	for	ADP
cana-2643	118	4	classification	classification	NOUN
cana-2643	118	5	;	;	PUNCT
cana-2643	118	6	lacks	lack	VERB
cana-2643	118	7	explicit	explicit	ADJ
cana-2643	118	8	mechanisms	mechanism	NOUN
cana-2643	118	9	for	for	ADP
cana-2643	118	10	handling	handle	VERB
cana-2643	118	11	spatial	spatial	ADJ
cana-2643	118	12	dependencies	dependency	NOUN
cana-2643	118	13	in	in	ADP
cana-2643	118	14	segmentation	segmentation	NOUN
cana-2643	118	15	.	.	PUNCT
cana-2643	119	1	geometric	geometric	ADJ
cana-2643	119	2	deep	deep	ADJ
cana-2643	119	3	learning	learning	NOUN
cana-2643	119	4	[	[	X
cana-2643	119	5	29	29	NUM
cana-2643	119	6	]	]	SYM
cana-2643	119	7	n	n	CCONJ
cana-2643	119	8	/	/	SYM
cana-2643	119	9	a	a	DET
cana-2643	119	10	(	(	PUNCT
cana-2643	119	11	conceptual	conceptual	ADJ
cana-2643	119	12	)	)	PUNCT
cana-2643	119	13	extends	extend	VERB
cana-2643	119	14	deep	deep	ADJ
cana-2643	119	15	learning	learning	NOUN
cana-2643	119	16	to	to	AUX
cana-2643	119	17	non	non	ADJ
cana-2643	119	18	-	-	ADJ
cana-2643	119	19	euclidean	euclidean	ADJ
cana-2643	119	20	do	do	VERB
cana-2643	119	21	mains	main	NOUN
cana-2643	119	22	;	;	PUNCT
cana-2643	119	23	useful	useful	ADJ
cana-2643	119	24	for	for	ADP
cana-2643	119	25	analyzing	analyze	VERB
cana-2643	119	26	geometric	geometric	ADJ
cana-2643	119	27	structures	structure	NOUN
cana-2643	119	28	in	in	ADP
cana-2643	119	29	data	datum	NOUN
cana-2643	119	30	.	.	PUNCT
cana-2643	120	1	high	high	ADJ
cana-2643	120	2	computational	computational	ADJ
cana-2643	120	3	complexity	complexity	NOUN
cana-2643	120	4	;	;	PUNCT
cana-2643	120	5	not	not	PART
cana-2643	120	6	yet	yet	ADV
cana-2643	120	7	widely	widely	ADV
cana-2643	120	8	applied	apply	VERB
cana-2643	120	9	to	to	PART
cana-2643	120	10	realtime	realtime	VERB
cana-2643	120	11	applications	application	NOUN
cana-2643	120	12	in	in	ADP
cana-2643	120	13	robotic	robotic	ADJ
cana-2643	120	14	vision	vision	NOUN
cana-2643	120	15	.	.	PUNCT
cana-2643	121	1	multi	multi	ADJ
cana-2643	121	2	-	-	ADJ
cana-2643	121	3	scale	scale	ADJ
cana-2643	121	4	graph	graph	NOUN
cana-2643	121	5	convolutional	convolutional	ADJ
cana-2643	121	6	network	network	NOUN
cana-2643	121	7	with	with	ADP
cana-2643	121	8	attention	attention	NOUN
cana-2643	121	9	[	[	X
cana-2643	121	10	29	29	NUM
cana-2643	121	11	]	]	SYM
cana-2643	121	12	88	88	NUM
cana-2643	121	13	%	%	NOUN
cana-2643	121	14	(	(	PUNCT
cana-2643	121	15	segmentation	segmentation	NOUN
cana-2643	121	16	iou	iou	NOUN
cana-2643	121	17	)	)	PUNCT
cana-2643	121	18	integrates	integrate	VERB
cana-2643	121	19	multi	multi	ADJ
cana-2643	121	20	-	-	ADJ
cana-2643	121	21	scale	scale	ADJ
cana-2643	121	22	attention	attention	NOUN
cana-2643	121	23	with	with	ADP
cana-2643	121	24	gcn	gcn	NOUN
cana-2643	121	25	,	,	PUNCT
cana-2643	121	26	enhancing	enhance	VERB
cana-2643	121	27	performance	performance	NOUN
cana-2643	121	28	in	in	ADP
cana-2643	121	29	complex	complex	ADJ
cana-2643	121	30	segmentation	segmentation	NOUN
cana-2643	121	31	tasks	task	NOUN
cana-2643	121	32	.	.	PUNCT
cana-2643	122	1	lacks	lack	VERB
cana-2643	122	2	reinforcement	reinforcement	NOUN
cana-2643	122	3	learning	learning	NOUN
cana-2643	122	4	component	component	NOUN
cana-2643	122	5	,	,	PUNCT
cana-2643	122	6	reducing	reduce	VERB
cana-2643	122	7	its	its	PRON
cana-2643	122	8	ability	ability	NOUN
cana-2643	122	9	to	to	PART
cana-2643	122	10	adapt	adapt	VERB
cana-2643	122	11	focus	focus	NOUN
cana-2643	122	12	dynamically	dynamically	ADV
cana-2643	122	13	.	.	PUNCT
cana-2643	123	1	deep	deep	ADJ
cana-2643	123	2	q	q	NOUN
cana-2643	123	3	networks	network	NOUN
cana-2643	123	4	(	(	PUNCT
cana-2643	123	5	dqn	dqn	PROPN
cana-2643	123	6	)	)	PUNCT
cana-2643	124	1	[	[	X
cana-2643	124	2	30	30	NUM
cana-2643	124	3	]	]	SYM
cana-2643	124	4	n	n	CCONJ
cana-2643	124	5	/	/	SYM
cana-2643	124	6	a	a	PRON
cana-2643	124	7	(	(	PUNCT
cana-2643	124	8	reinforce	reinforce	VERB
cana-2643	124	9	ment	ment	ADJ
cana-2643	124	10	learning	learning	NOUN
cana-2643	124	11	)	)	PUNCT
cana-2643	124	12	integrates	integrate	VERB
cana-2643	125	1	deep	deep	ADJ
cana-2643	125	2	learn	learn	NOUN
cana-2643	125	3	ing	ing	ADJ
cana-2643	125	4	with	with	ADP
cana-2643	125	5	reinforcement	reinforcement	NOUN
cana-2643	125	6	learning	learning	NOUN
cana-2643	125	7	,	,	PUNCT
cana-2643	125	8	achieving	achieve	VERB
cana-2643	125	9	adaptive	adaptive	ADJ
cana-2643	125	10	decision	decision	NOUN
cana-2643	125	11	making	making	NOUN
cana-2643	125	12	.	.	PUNCT
cana-2643	126	1	primarily	primarily	ADV
cana-2643	126	2	designed	design	VERB
cana-2643	126	3	for	for	ADP
cana-2643	126	4	control	control	NOUN
cana-2643	126	5	tasks	task	NOUN
cana-2643	126	6	;	;	PUNCT
cana-2643	126	7	lacks	lack	VERB
cana-2643	126	8	built	build	VERB
cana-2643	126	9	-	-	PUNCT
cana-2643	126	10	in	in	ADP
cana-2643	126	11	capabilities	capability	NOUN
cana-2643	126	12	for	for	ADP
cana-2643	126	13	handling	handle	VERB
cana-2643	126	14	spatial	spatial	ADJ
cana-2643	126	15	data	datum	NOUN
cana-2643	126	16	or	or	CCONJ
cana-2643	126	17	segmentation	segmentation	NOUN
cana-2643	126	18	.	.	PUNCT
cana-2643	127	1	in	in	ADP
cana-2643	127	2	table	table	NOUN
cana-2643	127	3	1	1	NUM
cana-2643	127	4	,	,	PUNCT
cana-2643	127	5	a	a	DET
cana-2643	127	6	comprehensive	comprehensive	ADJ
cana-2643	127	7	comparison	comparison	NOUN
cana-2643	127	8	of	of	ADP
cana-2643	127	9	various	various	ADJ
cana-2643	127	10	robotic	robotic	ADJ
cana-2643	127	11	vision	vision	NOUN
cana-2643	127	12	models	model	NOUN
cana-2643	127	13	is	be	AUX
cana-2643	127	14	presented	present	VERB
cana-2643	127	15	.	.	PUNCT
cana-2643	128	1	this	this	DET
cana-2643	128	2	table	table	NOUN
cana-2643	128	3	highlights	highlight	NOUN
cana-2643	128	4	each	each	DET
cana-2643	128	5	model	model	NOUN
cana-2643	128	6	’s	’s	PART
cana-2643	128	7	performance	performance	NOUN
cana-2643	128	8	,	,	PUNCT
cana-2643	128	9	advantages	advantage	NOUN
cana-2643	128	10	,	,	PUNCT
cana-2643	128	11	and	and	CCONJ
cana-2643	128	12	limitations	limitation	NOUN
cana-2643	128	13	,	,	PUNCT
cana-2643	128	14	as	as	SCONJ
cana-2643	128	15	evidenced	evidence	VERB
cana-2643	128	16	in	in	ADP
cana-2643	128	17	recent	recent	ADJ
cana-2643	128	18	literature	literature	NOUN
cana-2643	128	19	.	.	PUNCT
cana-2643	129	1	communications	communication	NOUN
cana-2643	129	2	on	on	ADP
cana-2643	129	3	applied	apply	VERB
cana-2643	129	4	nonlinear	nonlinear	ADJ
cana-2643	129	5	analysis	analysis	NOUN
cana-2643	129	6	issn	issn	NOUN
cana-2643	129	7	:	:	PUNCT
cana-2643	129	8	1074	1074	NUM
cana-2643	129	9	-	-	PUNCT
cana-2643	129	10	133x	133x	NUM
cana-2643	129	11	vol	vol	NOUN
cana-2643	129	12	32	32	NUM
cana-2643	129	13	no	no	NOUN
cana-2643	129	14	.	.	PUNCT
cana-2643	130	1	3s	3s	NUM
cana-2643	130	2	(	(	PUNCT
cana-2643	130	3	2025	2025	NUM
cana-2643	130	4	)	)	PUNCT
cana-2643	130	5	286	286	NUM
cana-2643	130	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	130	7	×	×	NOUN
cana-2643	130	8	×	×	NOUN
cana-2643	130	9	×	×	NOUN
cana-2643	130	10	×	×	NOUN
cana-2643	130	11	the	the	DET
cana-2643	130	12	models	model	NOUN
cana-2643	130	13	vary	vary	VERB
cana-2643	130	14	widely	widely	ADV
cana-2643	130	15	in	in	ADP
cana-2643	130	16	their	their	PRON
cana-2643	130	17	focus	focus	NOUN
cana-2643	130	18	,	,	PUNCT
cana-2643	130	19	from	from	ADP
cana-2643	130	20	foundational	foundational	ADJ
cana-2643	130	21	cnns	cnn	NOUN
cana-2643	130	22	to	to	ADP
cana-2643	130	23	advanced	advanced	ADJ
cana-2643	130	24	architectures	architecture	NOUN
cana-2643	130	25	like	like	ADP
cana-2643	130	26	vision	vision	NOUN
cana-2643	130	27	transformers	transformer	NOUN
cana-2643	130	28	(	(	PUNCT
cana-2643	130	29	vits	vit	NOUN
cana-2643	130	30	)	)	PUNCT
cana-2643	130	31	and	and	CCONJ
cana-2643	130	32	graph	graph	VERB
cana-2643	130	33	convolutional	convolutional	ADJ
cana-2643	130	34	networks	network	NOUN
cana-2643	130	35	(	(	PUNCT
cana-2643	130	36	gcns	gcns	PROPN
cana-2643	130	37	)	)	PUNCT
cana-2643	130	38	,	,	PUNCT
cana-2643	130	39	each	each	PRON
cana-2643	130	40	contributing	contribute	VERB
cana-2643	130	41	unique	unique	ADJ
cana-2643	130	42	strengths	strength	NOUN
cana-2643	130	43	to	to	PART
cana-2643	130	44	robotic	robotic	ADJ
cana-2643	130	45	vision	vision	NOUN
cana-2643	130	46	applications	application	NOUN
cana-2643	130	47	.	.	PUNCT
cana-2643	131	1	3	3	X
cana-2643	131	2	.	.	X
cana-2643	131	3	methodology	methodology	NOUN
cana-2643	131	4	:	:	PUNCT
cana-2643	131	5	mpga	mpga	NOUN
cana-2643	131	6	-	-	PUNCT
cana-2643	131	7	rl	rl	NOUN
cana-2643	131	8	-	-	PUNCT
cana-2643	131	9	net	net	NOUN
cana-2643	131	10	for	for	ADP
cana-2643	131	11	adaptive	adaptive	ADJ
cana-2643	131	12	robotic	robotic	ADJ
cana-2643	131	13	vision	vision	NOUN
cana-2643	131	14	this	this	DET
cana-2643	131	15	section	section	NOUN
cana-2643	131	16	outlines	outline	VERB
cana-2643	131	17	the	the	DET
cana-2643	131	18	proposed	propose	VERB
cana-2643	131	19	multi	multi	ADJ
cana-2643	131	20	-	-	ADJ
cana-2643	131	21	path	path	ADJ
cana-2643	131	22	graph	graph	NOUN
cana-2643	131	23	attention	attention	NOUN
cana-2643	131	24	network	network	NOUN
cana-2643	131	25	with	with	ADP
cana-2643	131	26	reinforcement	reinforcement	NOUN
cana-2643	131	27	learning	learning	NOUN
cana-2643	131	28	(	(	PUNCT
cana-2643	131	29	mpga	mpga	NOUN
cana-2643	131	30	-	-	PUNCT
cana-2643	131	31	rl	rl	NOUN
cana-2643	131	32	-	-	PUNCT
cana-2643	131	33	net	net	NOUN
cana-2643	131	34	)	)	PUNCT
cana-2643	131	35	model	model	NOUN
cana-2643	131	36	,	,	PUNCT
cana-2643	131	37	specifically	specifically	ADV
cana-2643	131	38	designed	design	VERB
cana-2643	131	39	for	for	ADP
cana-2643	131	40	robotic	robotic	ADJ
cana-2643	131	41	vision	vision	NOUN
cana-2643	131	42	tasks	task	NOUN
cana-2643	131	43	.	.	PUNCT
cana-2643	132	1	the	the	DET
cana-2643	132	2	model	model	NOUN
cana-2643	132	3	integrates	integrate	VERB
cana-2643	132	4	multi	multi	ADJ
cana-2643	132	5	-	-	ADJ
cana-2643	132	6	path	path	ADJ
cana-2643	132	7	feature	feature	NOUN
cana-2643	132	8	extraction	extraction	NOUN
cana-2643	132	9	,	,	PUNCT
cana-2643	132	10	graph	graph	NOUN
cana-2643	132	11	-	-	PUNCT
cana-2643	132	12	based	base	VERB
cana-2643	132	13	feature	feature	NOUN
cana-2643	132	14	aggregation	aggregation	NOUN
cana-2643	132	15	using	use	VERB
cana-2643	132	16	superpixel	superpixel	ADJ
cana-2643	132	17	segmentation	segmentation	NOUN
cana-2643	132	18	,	,	PUNCT
cana-2643	132	19	and	and	CCONJ
cana-2643	132	20	reinforcement	reinforcement	NOUN
cana-2643	132	21	learning	learning	NOUN
cana-2643	132	22	(	(	PUNCT
cana-2643	132	23	rl	rl	NOUN
cana-2643	132	24	)	)	PUNCT
cana-2643	132	25	for	for	ADP
cana-2643	132	26	dynamic	dynamic	ADJ
cana-2643	132	27	region	region	NOUN
cana-2643	132	28	prioritization	prioritization	NOUN
cana-2643	132	29	,	,	PUNCT
cana-2643	132	30	focusing	focus	VERB
cana-2643	132	31	on	on	ADP
cana-2643	132	32	critical	critical	ADJ
cana-2643	132	33	regions	region	NOUN
cana-2643	132	34	within	within	ADP
cana-2643	132	35	complex	complex	ADJ
cana-2643	132	36	and	and	CCONJ
cana-2643	132	37	dynamic	dynamic	ADJ
cana-2643	132	38	scenes	scene	NOUN
cana-2643	132	39	,	,	PUNCT
cana-2643	132	40	such	such	ADJ
cana-2643	132	41	as	as	ADP
cana-2643	132	42	obstacles	obstacle	NOUN
cana-2643	132	43	and	and	CCONJ
cana-2643	132	44	pathways	pathway	NOUN
cana-2643	132	45	.	.	PUNCT
cana-2643	133	1	(	(	PUNCT
cana-2643	133	2	a)multi	a)multi	NOUN
cana-2643	133	3	-	-	ADJ
cana-2643	133	4	path	path	ADJ
cana-2643	133	5	feature	feature	NOUN
cana-2643	133	6	extraction	extraction	NOUN
cana-2643	133	7	(	(	PUNCT
cana-2643	133	8	mpfe	mpfe	PROPN
cana-2643	133	9	)	)	PUNCT
cana-2643	133	10	the	the	DET
cana-2643	133	11	multi	multi	ADJ
cana-2643	133	12	-	-	ADJ
cana-2643	133	13	path	path	ADJ
cana-2643	133	14	feature	feature	NOUN
cana-2643	133	15	extraction	extraction	NOUN
cana-2643	133	16	(	(	PUNCT
cana-2643	133	17	mpfe	mpfe	NOUN
cana-2643	133	18	)	)	PUNCT
cana-2643	133	19	module	module	NOUN
cana-2643	133	20	processes	process	NOUN
cana-2643	133	21	input	input	NOUN
cana-2643	133	22	scenes	scene	NOUN
cana-2643	133	23	through	through	ADP
cana-2643	133	24	three	three	NUM
cana-2643	133	25	parallel	parallel	ADJ
cana-2643	133	26	paths	path	NOUN
cana-2643	133	27	to	to	PART
cana-2643	133	28	capture	capture	VERB
cana-2643	133	29	features	feature	NOUN
cana-2643	133	30	at	at	ADP
cana-2643	133	31	various	various	ADJ
cana-2643	133	32	resolutions	resolution	NOUN
cana-2643	133	33	.	.	PUNCT
cana-2643	134	1	this	this	PRON
cana-2643	134	2	enables	enable	VERB
cana-2643	134	3	a	a	DET
cana-2643	134	4	comprehensive	comprehensive	ADJ
cana-2643	134	5	multi	multi	ADJ
cana-2643	134	6	-	-	ADJ
cana-2643	134	7	scale	scale	ADJ
cana-2643	134	8	feature	feature	NOUN
cana-2643	134	9	extraction	extraction	NOUN
cana-2643	134	10	mechanism	mechanism	NOUN
cana-2643	134	11	essential	essential	ADJ
cana-2643	134	12	for	for	ADP
cana-2643	134	13	detecting	detect	VERB
cana-2643	134	14	global	global	ADJ
cana-2643	134	15	structures	structure	NOUN
cana-2643	134	16	,	,	PUNCT
cana-2643	134	17	mid	mid	ADJ
cana-2643	134	18	-	-	ADJ
cana-2643	134	19	level	level	ADJ
cana-2643	134	20	regional	regional	ADJ
cana-2643	134	21	features	feature	NOUN
cana-2643	134	22	,	,	PUNCT
cana-2643	134	23	and	and	CCONJ
cana-2643	134	24	fine	fine	ADJ
cana-2643	134	25	details	detail	NOUN
cana-2643	134	26	,	,	PUNCT
cana-2643	134	27	such	such	ADJ
cana-2643	134	28	as	as	ADP
cana-2643	134	29	edges	edge	NOUN
cana-2643	134	30	and	and	CCONJ
cana-2643	134	31	small	small	ADJ
cana-2643	134	32	objects	object	NOUN
cana-2643	134	33	.	.	PUNCT
cana-2643	135	1	path	path	NOUN
cana-2643	135	2	1	1	NUM
cana-2643	135	3	:	:	PUNCT
cana-2643	135	4	low	low	ADJ
cana-2643	135	5	-	-	PUNCT
cana-2643	135	6	resolution	resolution	NOUN
cana-2643	135	7	global	global	ADJ
cana-2643	135	8	features	feature	NOUN
cana-2643	135	9	path	path	NOUN
cana-2643	135	10	1	1	NUM
cana-2643	135	11	captures	capture	NOUN
cana-2643	135	12	broad	broad	ADJ
cana-2643	135	13	,	,	PUNCT
cana-2643	135	14	global	global	ADJ
cana-2643	135	15	features	feature	NOUN
cana-2643	135	16	by	by	ADP
cana-2643	135	17	applying	apply	VERB
cana-2643	135	18	convolution	convolution	NOUN
cana-2643	135	19	and	and	CCONJ
cana-2643	135	20	pooling	pool	VERB
cana-2643	135	21	operations	operation	NOUN
cana-2643	135	22	that	that	PRON
cana-2643	135	23	significantly	significantly	ADV
cana-2643	135	24	reduce	reduce	VERB
cana-2643	135	25	spatial	spatial	ADJ
cana-2643	135	26	resolution	resolution	NOUN
cana-2643	135	27	.	.	PUNCT
cana-2643	136	1	for	for	ADP
cana-2643	136	2	an	an	DET
cana-2643	136	3	input	input	NOUN
cana-2643	136	4	scene	scene	NOUN
cana-2643	136	5	of	of	ADP
cana-2643	136	6	size	size	NOUN
cana-2643	136	7	512	512	NUM
cana-2643	136	8	512	512	NUM
cana-2643	136	9	,	,	PUNCT
cana-2643	136	10	this	this	DET
cana-2643	136	11	path	path	NOUN
cana-2643	136	12	outputs	output	VERB
cana-2643	136	13	a	a	DET
cana-2643	136	14	feature	feature	NOUN
cana-2643	136	15	map	map	NOUN
cana-2643	136	16	of	of	ADP
cana-2643	136	17	size	size	NOUN
cana-2643	136	18	64	64	NUM
cana-2643	136	19	64	64	NUM
cana-2643	136	20	,	,	PUNCT
cana-2643	136	21	allowing	allow	VERB
cana-2643	136	22	the	the	DET
cana-2643	136	23	model	model	NOUN
cana-2643	136	24	to	to	PART
cana-2643	136	25	capture	capture	VERB
cana-2643	136	26	large	large	ADJ
cana-2643	136	27	-	-	PUNCT
cana-2643	136	28	scale	scale	NOUN
cana-2643	136	29	structures	structure	NOUN
cana-2643	136	30	and	and	CCONJ
cana-2643	136	31	the	the	DET
cana-2643	136	32	overall	overall	ADJ
cana-2643	136	33	context	context	NOUN
cana-2643	136	34	of	of	ADP
cana-2643	136	35	the	the	DET
cana-2643	136	36	scene	scene	NOUN
cana-2643	136	37	.	.	PUNCT
cana-2643	137	1	path	path	NOUN
cana-2643	137	2	2	2	NUM
cana-2643	137	3	:	:	PUNCT
cana-2643	137	4	medium	medium	ADJ
cana-2643	137	5	-	-	PUNCT
cana-2643	137	6	resolution	resolution	NOUN
cana-2643	137	7	features	feature	VERB
cana-2643	137	8	path	path	NOUN
cana-2643	137	9	2	2	NUM
cana-2643	137	10	balances	balance	NOUN
cana-2643	137	11	global	global	ADJ
cana-2643	137	12	context	context	NOUN
cana-2643	137	13	with	with	ADP
cana-2643	137	14	local	local	ADJ
cana-2643	137	15	details	detail	NOUN
cana-2643	137	16	by	by	ADP
cana-2643	137	17	retaining	retain	VERB
cana-2643	137	18	more	more	ADJ
cana-2643	137	19	spatial	spatial	ADJ
cana-2643	137	20	resolution	resolution	NOUN
cana-2643	137	21	than	than	ADP
cana-2643	137	22	path	path	NOUN
cana-2643	137	23	1	1	NUM
cana-2643	137	24	.	.	PUNCT
cana-2643	138	1	with	with	ADP
cana-2643	138	2	fewer	few	ADJ
cana-2643	138	3	pooling	pool	VERB
cana-2643	138	4	operations	operation	NOUN
cana-2643	138	5	,	,	PUNCT
cana-2643	138	6	this	this	DET
cana-2643	138	7	path	path	NOUN
cana-2643	138	8	outputs	output	VERB
cana-2643	138	9	a	a	DET
cana-2643	138	10	feature	feature	NOUN
cana-2643	138	11	map	map	NOUN
cana-2643	138	12	of	of	ADP
cana-2643	138	13	size	size	NOUN
cana-2643	138	14	128	128	NUM
cana-2643	138	15	128	128	NUM
cana-2643	138	16	,	,	PUNCT
cana-2643	138	17	optimized	optimize	VERB
cana-2643	138	18	for	for	ADP
cana-2643	138	19	capturing	capture	VERB
cana-2643	138	20	mid	mid	ADJ
cana-2643	138	21	-	-	ADJ
cana-2643	138	22	level	level	ADJ
cana-2643	138	23	structures	structure	NOUN
cana-2643	138	24	within	within	ADP
cana-2643	138	25	the	the	DET
cana-2643	138	26	scene	scene	NOUN
cana-2643	138	27	,	,	PUNCT
cana-2643	138	28	while	while	SCONJ
cana-2643	138	29	preserving	preserve	VERB
cana-2643	138	30	some	some	DET
cana-2643	138	31	global	global	ADJ
cana-2643	138	32	context	context	NOUN
cana-2643	138	33	.	.	PUNCT
cana-2643	139	1	path	path	NOUN
cana-2643	139	2	3	3	NUM
cana-2643	139	3	:	:	PUNCT
cana-2643	139	4	high	high	ADJ
cana-2643	139	5	-	-	PUNCT
cana-2643	139	6	resolution	resolution	NOUN
cana-2643	139	7	local	local	ADJ
cana-2643	139	8	features	feature	NOUN
cana-2643	139	9	path	path	NOUN
cana-2643	139	10	3	3	NUM
cana-2643	139	11	preserves	preserve	VERB
cana-2643	139	12	high	high	ADJ
cana-2643	139	13	-	-	PUNCT
cana-2643	139	14	resolution	resolution	NOUN
cana-2643	139	15	details	detail	NOUN
cana-2643	139	16	by	by	ADP
cana-2643	139	17	applying	apply	VERB
cana-2643	139	18	minimal	minimal	ADJ
cana-2643	139	19	down	down	ADP
cana-2643	139	20	sampling	sample	VERB
cana-2643	139	21	,	,	PUNCT
cana-2643	139	22	producing	produce	VERB
cana-2643	139	23	a	a	DET
cana-2643	139	24	feature	feature	NOUN
cana-2643	139	25	map	map	NOUN
cana-2643	139	26	with	with	ADP
cana-2643	139	27	a	a	DET
cana-2643	139	28	resolution	resolution	NOUN
cana-2643	139	29	of	of	ADP
cana-2643	139	30	256	256	NUM
cana-2643	139	31	256	256	NUM
cana-2643	139	32	.	.	PUNCT
cana-2643	140	1	this	this	DET
cana-2643	140	2	path	path	NOUN
cana-2643	140	3	focuses	focus	VERB
cana-2643	140	4	on	on	ADP
cana-2643	140	5	finegrained	finegrained	ADJ
cana-2643	140	6	features	feature	NOUN
cana-2643	140	7	,	,	PUNCT
cana-2643	140	8	such	such	ADJ
cana-2643	140	9	as	as	ADP
cana-2643	140	10	object	object	NOUN
cana-2643	140	11	edges	edge	NOUN
cana-2643	140	12	,	,	PUNCT
cana-2643	140	13	small	small	ADJ
cana-2643	140	14	obstacles	obstacle	NOUN
cana-2643	140	15	,	,	PUNCT
cana-2643	140	16	or	or	CCONJ
cana-2643	140	17	detailed	detailed	ADJ
cana-2643	140	18	scene	scene	NOUN
cana-2643	140	19	boundaries	boundary	NOUN
cana-2643	140	20	,	,	PUNCT
cana-2643	140	21	which	which	PRON
cana-2643	140	22	are	be	AUX
cana-2643	140	23	critical	critical	ADJ
cana-2643	140	24	for	for	ADP
cana-2643	140	25	accurate	accurate	ADJ
cana-2643	140	26	object	object	NOUN
cana-2643	140	27	recognition	recognition	NOUN
cana-2643	140	28	and	and	CCONJ
cana-2643	140	29	path	path	NOUN
cana-2643	140	30	planning	planning	NOUN
cana-2643	140	31	.	.	PUNCT
cana-2643	141	1	(	(	PUNCT
cana-2643	141	2	b)feature	b)feature	NOUN
cana-2643	141	3	attention	attention	NOUN
cana-2643	141	4	fusion	fusion	NOUN
cana-2643	141	5	to	to	PART
cana-2643	141	6	combine	combine	VERB
cana-2643	141	7	the	the	DET
cana-2643	141	8	outputs	output	NOUN
cana-2643	141	9	from	from	ADP
cana-2643	141	10	the	the	DET
cana-2643	141	11	three	three	NUM
cana-2643	141	12	paths	path	NOUN
cana-2643	141	13	,	,	PUNCT
cana-2643	141	14	feature	feature	NOUN
cana-2643	141	15	attention	attention	NOUN
cana-2643	141	16	fusion	fusion	NOUN
cana-2643	141	17	adaptively	adaptively	ADV
cana-2643	141	18	weights	weight	VERB
cana-2643	141	19	the	the	DET
cana-2643	141	20	importance	importance	NOUN
cana-2643	141	21	of	of	ADP
cana-2643	141	22	each	each	DET
cana-2643	141	23	path	path	NOUN
cana-2643	141	24	.	.	PUNCT
cana-2643	142	1	this	this	PRON
cana-2643	142	2	allows	allow	VERB
cana-2643	142	3	the	the	DET
cana-2643	142	4	most	most	ADV
cana-2643	142	5	relevant	relevant	ADJ
cana-2643	142	6	features	feature	NOUN
cana-2643	142	7	for	for	ADP
cana-2643	142	8	a	a	DET
cana-2643	142	9	given	give	VERB
cana-2643	142	10	scene	scene	NOUN
cana-2643	142	11	to	to	PART
cana-2643	142	12	be	be	AUX
cana-2643	142	13	prioritized	prioritize	VERB
cana-2643	142	14	.	.	PUNCT
cana-2643	143	1	where	where	SCONJ
cana-2643	143	2	wp	wp	PROPN
cana-2643	143	3	represents	represent	VERB
cana-2643	143	4	trainable	trainable	ADJ
cana-2643	143	5	weights	weight	NOUN
cana-2643	143	6	applied	apply	VERB
cana-2643	143	7	to	to	ADP
cana-2643	143	8	the	the	DET
cana-2643	143	9	feature	feature	NOUN
cana-2643	143	10	map	map	NOUN
cana-2643	143	11	fp	fp	INTJ
cana-2643	143	12	from	from	ADP
cana-2643	143	13	path	path	NOUN
cana-2643	143	14	p	p	X
cana-2643	143	15	,	,	PUNCT
cana-2643	143	16	and	and	CCONJ
cana-2643	143	17	the	the	DET
cana-2643	143	18	softmax	softmax	NOUN
cana-2643	143	19	function	function	NOUN
cana-2643	143	20	ensures	ensure	VERB
cana-2643	143	21	that	that	SCONJ
cana-2643	143	22	the	the	DET
cana-2643	143	23	attention	attention	NOUN
cana-2643	143	24	scores	score	NOUN
cana-2643	143	25	sum	sum	VERB
cana-2643	143	26	to	to	ADP
cana-2643	143	27	1	1	NUM
cana-2643	143	28	.	.	PUNCT
cana-2643	144	1	the	the	DET
cana-2643	144	2	final	final	ADJ
cana-2643	144	3	fused	fuse	VERB
cana-2643	144	4	feature	feature	NOUN
cana-2643	144	5	map	map	NOUN
cana-2643	144	6	ffused	ffuse	VERB
cana-2643	144	7	is	be	AUX
cana-2643	144	8	computed	compute	VERB
cana-2643	144	9	as	as	ADP
cana-2643	144	10	a	a	DET
cana-2643	144	11	weighted	weighted	ADJ
cana-2643	144	12	sum	sum	NOUN
cana-2643	144	13	of	of	ADP
cana-2643	144	14	the	the	DET
cana-2643	144	15	feature	feature	NOUN
cana-2643	144	16	maps	map	NOUN
cana-2643	144	17	from	from	ADP
cana-2643	144	18	all	all	DET
cana-2643	144	19	paths	path	NOUN
cana-2643	144	20	:	:	PUNCT
cana-2643	144	21	ffused	ffuse	VERB
cana-2643	144	22	=	=	PUNCT
cana-2643	145	1	α1f1	α1f1	X
cana-2643	146	1	+	+	CCONJ
cana-2643	146	2	α2f2	α2f2	PUNCT
cana-2643	146	3	+	+	CCONJ
cana-2643	146	4	α3f3	α3f3	X
cana-2643	146	5	this	this	DET
cana-2643	146	6	attention	attention	NOUN
cana-2643	146	7	mechanism	mechanism	NOUN
cana-2643	146	8	enables	enable	VERB
cana-2643	146	9	the	the	DET
cana-2643	146	10	model	model	NOUN
cana-2643	146	11	to	to	PART
cana-2643	146	12	dynamically	dynamically	ADV
cana-2643	146	13	focus	focus	VERB
cana-2643	146	14	on	on	ADP
cana-2643	146	15	different	different	ADJ
cana-2643	146	16	feature	feature	NOUN
cana-2643	146	17	scales	scale	NOUN
cana-2643	146	18	depending	depend	VERB
cana-2643	146	19	on	on	ADP
cana-2643	146	20	the	the	DET
cana-2643	146	21	input	input	NOUN
cana-2643	146	22	scene	scene	NOUN
cana-2643	146	23	’s	’s	PART
cana-2643	146	24	specific	specific	ADJ
cana-2643	146	25	characteristics	characteristic	NOUN
cana-2643	146	26	,	,	PUNCT
cana-2643	146	27	prioritizing	prioritize	VERB
cana-2643	146	28	broader	broad	ADJ
cana-2643	146	29	structures	structure	NOUN
cana-2643	146	30	when	when	SCONJ
cana-2643	146	31	necessary	necessary	ADJ
cana-2643	146	32	or	or	CCONJ
cana-2643	146	33	honing	hone	VERB
cana-2643	146	34	in	in	ADP
cana-2643	146	35	on	on	ADP
cana-2643	146	36	finer	fine	ADJ
cana-2643	146	37	details	detail	NOUN
cana-2643	146	38	when	when	SCONJ
cana-2643	146	39	required	require	VERB
cana-2643	146	40	.	.	PUNCT
cana-2643	147	1	communications	communication	NOUN
cana-2643	147	2	on	on	ADP
cana-2643	147	3	applied	apply	VERB
cana-2643	147	4	nonlinear	nonlinear	ADJ
cana-2643	147	5	analysis	analysis	NOUN
cana-2643	147	6	issn	issn	NOUN
cana-2643	147	7	:	:	PUNCT
cana-2643	147	8	1074	1074	NUM
cana-2643	147	9	-	-	PUNCT
cana-2643	147	10	133x	133x	NUM
cana-2643	147	11	vol	vol	NOUN
cana-2643	147	12	32	32	NUM
cana-2643	147	13	no	no	NOUN
cana-2643	147	14	.	.	PUNCT
cana-2643	148	1	3s	3s	NUM
cana-2643	148	2	(	(	PUNCT
cana-2643	148	3	2025	2025	NUM
cana-2643	148	4	)	)	PUNCT
cana-2643	148	5	287	287	NUM
cana-2643	148	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	148	7	×	×	NOUN
cana-2643	148	8	×	×	NOUN
cana-2643	148	9	∈	∈	PROPN
cana-2643	148	10	∈	∈	PROPN
cana-2643	148	11	∈	∈	PROPN
cana-2643	148	12	∈	∈	PROPN
cana-2643	148	13	∈	∈	PROPN
cana-2643	148	14	∈	∈	PROPN
cana-2643	148	15	once	once	ADV
cana-2643	148	16	multi	multi	ADJ
cana-2643	148	17	-	-	ADJ
cana-2643	148	18	scale	scale	ADJ
cana-2643	148	19	features	feature	NOUN
cana-2643	148	20	are	be	AUX
cana-2643	148	21	extracted	extract	VERB
cana-2643	148	22	from	from	ADP
cana-2643	148	23	the	the	DET
cana-2643	148	24	mpfe	mpfe	PROPN
cana-2643	148	25	module	module	NOUN
cana-2643	148	26	and	and	CCONJ
cana-2643	148	27	fused	fuse	VERB
cana-2643	148	28	through	through	ADP
cana-2643	148	29	attention	attention	NOUN
cana-2643	148	30	,	,	PUNCT
cana-2643	148	31	the	the	DET
cana-2643	148	32	next	next	ADJ
cana-2643	148	33	step	step	NOUN
cana-2643	148	34	is	be	AUX
cana-2643	148	35	to	to	PART
cana-2643	148	36	model	model	VERB
cana-2643	148	37	spatial	spatial	ADJ
cana-2643	148	38	relationships	relationship	NOUN
cana-2643	148	39	between	between	ADP
cana-2643	148	40	different	different	ADJ
cana-2643	148	41	regions	region	NOUN
cana-2643	148	42	using	use	VERB
cana-2643	148	43	a	a	DET
cana-2643	148	44	graph	graph	NOUN
cana-2643	148	45	convolutional	convolutional	ADJ
cana-2643	148	46	network	network	NOUN
cana-2643	148	47	(	(	PUNCT
cana-2643	148	48	gcn	gcn	NOUN
cana-2643	148	49	)	)	PUNCT
cana-2643	148	50	.	.	PUNCT
cana-2643	149	1	transition	transition	NOUN
cana-2643	149	2	from	from	ADP
cana-2643	149	3	feature	feature	NOUN
cana-2643	149	4	maps	map	NOUN
cana-2643	149	5	to	to	PART
cana-2643	149	6	graph	graph	VERB
cana-2643	149	7	representation	representation	NOUN
cana-2643	149	8	the	the	DET
cana-2643	149	9	fused	fuse	VERB
cana-2643	149	10	feature	feature	NOUN
cana-2643	149	11	map	map	NOUN
cana-2643	149	12	is	be	AUX
cana-2643	149	13	divided	divide	VERB
cana-2643	149	14	into	into	ADP
cana-2643	149	15	regions	region	NOUN
cana-2643	149	16	using	use	VERB
cana-2643	149	17	slic	slic	NOUN
cana-2643	149	18	(	(	PUNCT
cana-2643	149	19	simple	simple	ADJ
cana-2643	149	20	linear	linear	ADJ
cana-2643	149	21	iterative	iterative	NOUN
cana-2643	149	22	clustering	clustering	NOUN
cana-2643	149	23	)	)	PUNCT
cana-2643	149	24	,	,	PUNCT
cana-2643	149	25	grouping	group	VERB
cana-2643	149	26	spatially	spatially	ADV
cana-2643	149	27	connected	connect	VERB
cana-2643	149	28	pixels	pixel	NOUN
cana-2643	149	29	with	with	ADP
cana-2643	149	30	similar	similar	ADJ
cana-2643	149	31	features	feature	NOUN
cana-2643	149	32	into	into	ADP
cana-2643	149	33	superpixels	superpixel	NOUN
cana-2643	149	34	.	.	PUNCT
cana-2643	150	1	each	each	DET
cana-2643	150	2	superpixel	superpixel	NOUN
cana-2643	150	3	forms	form	VERB
cana-2643	150	4	a	a	DET
cana-2643	150	5	region	region	NOUN
cana-2643	150	6	in	in	ADP
cana-2643	150	7	the	the	DET
cana-2643	150	8	scene	scene	NOUN
cana-2643	150	9	and	and	CCONJ
cana-2643	150	10	is	be	AUX
cana-2643	150	11	represented	represent	VERB
cana-2643	150	12	as	as	ADP
cana-2643	150	13	a	a	DET
cana-2643	150	14	node	node	NOUN
cana-2643	150	15	in	in	ADP
cana-2643	150	16	the	the	DET
cana-2643	150	17	graph	graph	NOUN
cana-2643	150	18	.	.	PUNCT
cana-2643	151	1	given	give	VERB
cana-2643	151	2	a	a	DET
cana-2643	151	3	feature	feature	NOUN
cana-2643	151	4	map	map	NOUN
cana-2643	151	5	ffused	ffuse	VERB
cana-2643	151	6	of	of	ADP
cana-2643	151	7	size	size	NOUN
cana-2643	151	8	h	h	PROPN
cana-2643	151	9	w	w	PROPN
cana-2643	151	10	c	c	PROPN
cana-2643	151	11	,	,	PUNCT
cana-2643	151	12	slic	slic	NOUN
cana-2643	151	13	is	be	AUX
cana-2643	151	14	applied	apply	VERB
cana-2643	151	15	to	to	PART
cana-2643	151	16	segment	segment	VERB
cana-2643	151	17	the	the	DET
cana-2643	151	18	feature	feature	NOUN
cana-2643	151	19	map	map	NOUN
cana-2643	151	20	into	into	ADP
cana-2643	151	21	nregions	nregion	NOUN
cana-2643	151	22	super	super	ADJ
cana-2643	151	23	pixels	pixel	NOUN
cana-2643	151	24	,	,	PUNCT
cana-2643	151	25	where	where	SCONJ
cana-2643	151	26	nregions	nregion	NOUN
cana-2643	151	27	is	be	AUX
cana-2643	151	28	the	the	DET
cana-2643	151	29	desired	desire	VERB
cana-2643	151	30	number	number	NOUN
cana-2643	151	31	of	of	ADP
cana-2643	151	32	regions	region	NOUN
cana-2643	151	33	:	:	PUNCT
cana-2643	151	34	super	super	ADJ
cana-2643	151	35	pixels	pixel	NOUN
cana-2643	151	36	=	=	SYM
cana-2643	151	37	slic(ffused	slic(ffuse	VERB
cana-2643	151	38	,	,	PUNCT
cana-2643	151	39	nregions	nregion	NOUN
cana-2643	151	40	)	)	PUNCT
cana-2643	151	41	each	each	DET
cana-2643	151	42	super	super	PROPN
cana-2643	151	43	pixel	pixel	PROPN
cana-2643	151	44	corresponds	correspond	VERB
cana-2643	151	45	to	to	ADP
cana-2643	151	46	a	a	DET
cana-2643	151	47	set	set	NOUN
cana-2643	151	48	of	of	ADP
cana-2643	151	49	spatially	spatially	ADV
cana-2643	151	50	connected	connect	VERB
cana-2643	151	51	pixels	pixel	NOUN
cana-2643	151	52	and	and	CCONJ
cana-2643	151	53	becomes	become	VERB
cana-2643	151	54	a	a	DET
cana-2643	151	55	node	node	PROPN
cana-2643	151	56	vi	vi	NOUN
cana-2643	151	57	in	in	ADP
cana-2643	151	58	the	the	DET
cana-2643	151	59	graph	graph	NOUN
cana-2643	151	60	.	.	PUNCT
cana-2643	152	1	the	the	DET
cana-2643	152	2	feature	feature	NOUN
cana-2643	152	3	vector	vector	NOUN
cana-2643	152	4	for	for	ADP
cana-2643	152	5	each	each	DET
cana-2643	152	6	node	node	NOUN
cana-2643	152	7	is	be	AUX
cana-2643	152	8	obtained	obtain	VERB
cana-2643	152	9	by	by	ADP
cana-2643	152	10	averaging	average	VERB
cana-2643	152	11	feature	feature	NOUN
cana-2643	152	12	values	value	NOUN
cana-2643	152	13	within	within	ADP
cana-2643	152	14	the	the	DET
cana-2643	152	15	corresponding	corresponding	ADJ
cana-2643	152	16	super	super	ADJ
cana-2643	152	17	pixel	pixel	NOUN
cana-2643	152	18	,	,	PUNCT
cana-2643	152	19	transforming	transform	VERB
cana-2643	152	20	the	the	DET
cana-2643	152	21	feature	feature	NOUN
cana-2643	152	22	map	map	NOUN
cana-2643	152	23	into	into	ADP
cana-2643	152	24	a	a	DET
cana-2643	152	25	graph	graph	NOUN
cana-2643	152	26	structure	structure	NOUN
cana-2643	152	27	where	where	SCONJ
cana-2643	152	28	each	each	DET
cana-2643	152	29	node	node	NOUN
cana-2643	152	30	represents	represent	VERB
cana-2643	152	31	a	a	DET
cana-2643	152	32	region	region	NOUN
cana-2643	152	33	of	of	ADP
cana-2643	152	34	the	the	DET
cana-2643	152	35	scene	scene	NOUN
cana-2643	152	36	.	.	PUNCT
cana-2643	153	1	graph	graph	VERB
cana-2643	153	2	construction	construction	NOUN
cana-2643	153	3	after	after	ADP
cana-2643	153	4	identifying	identify	VERB
cana-2643	153	5	the	the	DET
cana-2643	153	6	super	super	ADJ
cana-2643	153	7	pixels	pixel	NOUN
cana-2643	153	8	,	,	PUNCT
cana-2643	153	9	a	a	DET
cana-2643	153	10	graph	graph	NOUN
cana-2643	153	11	g	g	NOUN
cana-2643	153	12	=	=	SYM
cana-2643	153	13	(	(	PUNCT
cana-2643	153	14	v	v	NOUN
cana-2643	153	15	,	,	PUNCT
cana-2643	153	16	e	e	NOUN
cana-2643	153	17	)	)	PUNCT
cana-2643	153	18	is	be	AUX
cana-2643	153	19	constructed	construct	VERB
cana-2643	153	20	.	.	PUNCT
cana-2643	154	1	each	each	DET
cana-2643	154	2	node	node	PROPN
cana-2643	154	3	vi	vi	PROPN
cana-2643	154	4	v	v	NOUN
cana-2643	154	5	represents	represent	VERB
cana-2643	154	6	a	a	DET
cana-2643	154	7	super	super	ADJ
cana-2643	154	8	pixel	pixel	NOUN
cana-2643	154	9	,	,	PUNCT
cana-2643	154	10	and	and	CCONJ
cana-2643	154	11	edges	edge	VERB
cana-2643	154	12	eij	eij	PROPN
cana-2643	154	13	e	e	PROPN
cana-2643	154	14	are	be	AUX
cana-2643	154	15	created	create	VERB
cana-2643	154	16	between	between	ADP
cana-2643	154	17	nodes	node	NOUN
cana-2643	154	18	to	to	PART
cana-2643	154	19	model	model	VERB
cana-2643	154	20	spatial	spatial	ADJ
cana-2643	154	21	relationships	relationship	NOUN
cana-2643	154	22	between	between	ADP
cana-2643	154	23	regions	region	NOUN
cana-2643	154	24	.	.	PUNCT
cana-2643	155	1	nodes	node	NOUN
cana-2643	155	2	are	be	AUX
cana-2643	155	3	connected	connect	VERB
cana-2643	155	4	if	if	SCONJ
cana-2643	155	5	their	their	PRON
cana-2643	155	6	super	super	ADJ
cana-2643	155	7	pixels	pixel	NOUN
cana-2643	155	8	are	be	AUX
cana-2643	155	9	adjacent	adjacent	ADJ
cana-2643	155	10	or	or	CCONJ
cana-2643	155	11	share	share	VERB
cana-2643	155	12	similar	similar	ADJ
cana-2643	155	13	feature	feature	NOUN
cana-2643	155	14	characteristics	characteristic	NOUN
cana-2643	155	15	.	.	PUNCT
cana-2643	156	1	the	the	DET
cana-2643	156	2	adjacency	adjacency	PROPN
cana-2643	156	3	matrix	matrix	NOUN
cana-2643	156	4	a	a	DET
cana-2643	156	5	rn×n	rn×n	NOUN
cana-2643	156	6	,	,	PUNCT
cana-2643	156	7	where	where	SCONJ
cana-2643	156	8	n	n	PRON
cana-2643	156	9	is	be	AUX
cana-2643	156	10	the	the	DET
cana-2643	156	11	number	number	NOUN
cana-2643	156	12	of	of	ADP
cana-2643	156	13	superpixels	superpixel	NOUN
cana-2643	156	14	,	,	PUNCT
cana-2643	156	15	defines	define	VERB
cana-2643	156	16	the	the	DET
cana-2643	156	17	connectivity	connectivity	NOUN
cana-2643	156	18	between	between	ADP
cana-2643	156	19	nodes	node	NOUN
cana-2643	156	20	.	.	PUNCT
cana-2643	157	1	the	the	DET
cana-2643	157	2	feature	feature	NOUN
cana-2643	157	3	matrix	matrix	NOUN
cana-2643	157	4	x	x	PUNCT
cana-2643	157	5	rn×c	rn×c	NOUN
cana-2643	157	6	,	,	PUNCT
cana-2643	157	7	where	where	SCONJ
cana-2643	157	8	c	c	PROPN
cana-2643	157	9	is	be	AUX
cana-2643	157	10	the	the	DET
cana-2643	157	11	feature	feature	NOUN
cana-2643	157	12	dimension	dimension	NOUN
cana-2643	157	13	,	,	PUNCT
cana-2643	157	14	stores	store	VERB
cana-2643	157	15	the	the	DET
cana-2643	157	16	feature	feature	NOUN
cana-2643	157	17	vectors	vector	NOUN
cana-2643	157	18	for	for	ADP
cana-2643	157	19	each	each	PRON
cana-2643	157	20	node.gcn	node.gcn	PUNCT
cana-2643	157	21	for	for	ADP
cana-2643	157	22	feature	feature	NOUN
cana-2643	157	23	aggregation	aggregation	NOUN
cana-2643	157	24	the	the	DET
cana-2643	157	25	graph	graph	NOUN
cana-2643	157	26	convolutional	convolutional	ADJ
cana-2643	157	27	network	network	NOUN
cana-2643	157	28	(	(	PUNCT
cana-2643	157	29	gcn	gcn	NOUN
cana-2643	157	30	)	)	PUNCT
cana-2643	157	31	aggregates	aggregate	VERB
cana-2643	157	32	spatial	spatial	ADJ
cana-2643	157	33	relationships	relationship	NOUN
cana-2643	157	34	between	between	ADP
cana-2643	157	35	nodes	node	NOUN
cana-2643	157	36	in	in	ADP
cana-2643	157	37	the	the	DET
cana-2643	157	38	graph	graph	NOUN
cana-2643	157	39	by	by	ADP
cana-2643	157	40	propagating	propagate	VERB
cana-2643	157	41	information	information	NOUN
cana-2643	157	42	from	from	ADP
cana-2643	157	43	neighboring	neighboring	NOUN
cana-2643	157	44	nodes	node	NOUN
cana-2643	157	45	.	.	PUNCT
cana-2643	158	1	each	each	DET
cana-2643	158	2	node	node	NOUN
cana-2643	158	3	’s	’s	PART
cana-2643	158	4	features	feature	NOUN
cana-2643	158	5	are	be	AUX
cana-2643	158	6	updated	update	VERB
cana-2643	158	7	by	by	ADP
cana-2643	158	8	combining	combine	VERB
cana-2643	158	9	its	its	PRON
cana-2643	158	10	own	own	ADJ
cana-2643	158	11	features	feature	NOUN
cana-2643	158	12	with	with	ADP
cana-2643	158	13	those	those	PRON
cana-2643	158	14	from	from	ADP
cana-2643	158	15	its	its	PRON
cana-2643	158	16	neighbors	neighbor	NOUN
cana-2643	158	17	:	:	PUNCT
cana-2643	158	18	x′	x′	PROPN
cana-2643	158	19	=	=	SYM
cana-2643	158	20	σ	σ	PROPN
cana-2643	158	21	(	(	PUNCT
cana-2643	158	22	axwgcn	axwgcn	PROPN
cana-2643	158	23	)	)	PUNCT
cana-2643	158	24	where	where	SCONJ
cana-2643	158	25	:	:	PUNCT
cana-2643	158	26	x	x	X
cana-2643	158	27	∈	∈	NOUN
cana-2643	158	28	rn×c	rn×c	NOUN
cana-2643	158	29	is	be	AUX
cana-2643	158	30	the	the	DET
cana-2643	158	31	feature	feature	NOUN
cana-2643	158	32	matrix	matrix	NOUN
cana-2643	158	33	containing	contain	VERB
cana-2643	158	34	node	node	ADJ
cana-2643	158	35	features	feature	NOUN
cana-2643	158	36	.	.	PUNCT
cana-2643	159	1	a	a	DET
cana-2643	159	2	rn×n	rn×n	PROPN
cana-2643	159	3	is	be	AUX
cana-2643	159	4	the	the	DET
cana-2643	159	5	adjacency	adjacency	PROPN
cana-2643	159	6	matrix	matrix	NOUN
cana-2643	159	7	.	.	PUNCT
cana-2643	160	1	wgc	wgc	PROPN
cana-2643	160	2	rc×c′	rc×c′	PROPN
cana-2643	160	3	is	be	AUX
cana-2643	160	4	the	the	DET
cana-2643	160	5	trainable	trainable	ADJ
cana-2643	160	6	weight	weight	NOUN
cana-2643	160	7	matrix	matrix	NOUN
cana-2643	160	8	.	.	PUNCT
cana-2643	161	1	c′	c′	VERB
cana-2643	161	2	is	be	AUX
cana-2643	161	3	the	the	DET
cana-2643	161	4	new	new	ADJ
cana-2643	161	5	feature	feature	NOUN
cana-2643	161	6	dimension	dimension	NOUN
cana-2643	161	7	after	after	ADP
cana-2643	161	8	graph	graph	NOUN
cana-2643	161	9	convolution	convolution	NOUN
cana-2643	161	10	.	.	PUNCT
cana-2643	162	1	σ	σ	NOUN
cana-2643	162	2	is	be	AUX
cana-2643	162	3	a	a	DET
cana-2643	162	4	non	non	ADJ
cana-2643	162	5	-	-	ADJ
cana-2643	162	6	linear	linear	ADJ
cana-2643	162	7	activation	activation	NOUN
cana-2643	162	8	function	function	NOUN
cana-2643	162	9	(	(	PUNCT
cana-2643	162	10	e.g.	e.g.	ADV
cana-2643	162	11	,	,	PUNCT
cana-2643	162	12	relu	relu	NOUN
cana-2643	162	13	)	)	PUNCT
cana-2643	162	14	.	.	PUNCT
cana-2643	163	1	attention	attention	NOUN
cana-2643	163	2	mechanism	mechanism	NOUN
cana-2643	163	3	with	with	ADP
cana-2643	163	4	reinforcement	reinforcement	NOUN
cana-2643	163	5	learning	learning	NOUN
cana-2643	163	6	(	(	PUNCT
cana-2643	163	7	rl	rl	NOUN
cana-2643	163	8	-	-	PUNCT
cana-2643	163	9	attention	attention	NOUN
cana-2643	163	10	)	)	PUNCT
cana-2643	163	11	after	after	ADP
cana-2643	163	12	feature	feature	NOUN
cana-2643	163	13	aggregation	aggregation	NOUN
cana-2643	163	14	through	through	ADP
cana-2643	163	15	the	the	DET
cana-2643	163	16	gcn	gcn	NOUN
cana-2643	163	17	,	,	PUNCT
cana-2643	163	18	an	an	DET
cana-2643	163	19	attention	attention	NOUN
cana-2643	163	20	mechanism	mechanism	NOUN
cana-2643	163	21	is	be	AUX
cana-2643	163	22	applied	apply	VERB
cana-2643	163	23	to	to	PART
cana-2643	163	24	emphasize	emphasize	VERB
cana-2643	163	25	critical	critical	ADJ
cana-2643	163	26	regions	region	NOUN
cana-2643	163	27	,	,	PUNCT
cana-2643	163	28	such	such	ADJ
cana-2643	163	29	as	as	ADP
cana-2643	163	30	obstacles	obstacle	NOUN
cana-2643	163	31	or	or	CCONJ
cana-2643	163	32	key	key	ADJ
cana-2643	163	33	navigational	navigational	ADJ
cana-2643	163	34	paths	path	NOUN
cana-2643	163	35	.	.	PUNCT
cana-2643	164	1	the	the	DET
cana-2643	164	2	attention	attention	NOUN
cana-2643	164	3	mechanism	mechanism	NOUN
cana-2643	164	4	assigns	assign	VERB
cana-2643	164	5	dynamic	dynamic	ADJ
cana-2643	164	6	weights	weight	NOUN
cana-2643	164	7	βi	βi	PRON
cana-2643	164	8	to	to	ADP
cana-2643	164	9	the	the	DET
cana-2643	164	10	feature	feature	NOUN
cana-2643	164	11	vectors	vector	NOUN
cana-2643	164	12	of	of	ADP
cana-2643	164	13	each	each	DET
cana-2643	164	14	region	region	NOUN
cana-2643	164	15	i	i	PRON
cana-2643	164	16	,	,	PUNCT
cana-2643	164	17	learned	learn	VERB
cana-2643	164	18	through	through	ADP
cana-2643	164	19	reinforcement	reinforcement	NOUN
cana-2643	164	20	learning	learning	NOUN
cana-2643	164	21	(	(	PUNCT
cana-2643	164	22	rl	rl	NOUN
cana-2643	164	23	)	)	PUNCT
cana-2643	164	24	,	,	PUNCT
cana-2643	164	25	which	which	PRON
cana-2643	164	26	optimizes	optimize	VERB
cana-2643	164	27	the	the	DET
cana-2643	164	28	attention	attention	NOUN
cana-2643	164	29	focus	focus	NOUN
cana-2643	164	30	based	base	VERB
cana-2643	164	31	on	on	ADP
cana-2643	164	32	real	real	ADJ
cana-2643	164	33	-	-	PUNCT
cana-2643	164	34	time	time	NOUN
cana-2643	164	35	performance	performance	NOUN
cana-2643	164	36	.	.	PUNCT
cana-2643	165	1	communications	communication	NOUN
cana-2643	165	2	on	on	ADP
cana-2643	165	3	applied	apply	VERB
cana-2643	165	4	nonlinear	nonlinear	ADJ
cana-2643	165	5	analysis	analysis	NOUN
cana-2643	165	6	issn	issn	NOUN
cana-2643	165	7	:	:	PUNCT
cana-2643	165	8	1074	1074	NUM
cana-2643	165	9	-	-	PUNCT
cana-2643	165	10	133x	133x	NUM
cana-2643	165	11	vol	vol	NOUN
cana-2643	165	12	32	32	NUM
cana-2643	165	13	no	no	NOUN
cana-2643	165	14	.	.	PUNCT
cana-2643	166	1	3s	3s	NUM
cana-2643	166	2	(	(	PUNCT
cana-2643	166	3	2025	2025	NUM
cana-2643	166	4	)	)	PUNCT
cana-2643	166	5	288	288	NUM
cana-2643	166	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	166	7	let	let	VERB
cana-2643	166	8	x′	x′	PROPN
cana-2643	166	9	represent	represent	VERB
cana-2643	166	10	the	the	DET
cana-2643	166	11	updated	update	VERB
cana-2643	166	12	feature	feature	NOUN
cana-2643	166	13	matrix	matrix	NOUN
cana-2643	166	14	from	from	ADP
cana-2643	166	15	the	the	DET
cana-2643	166	16	gcn	gcn	NOUN
cana-2643	166	17	.	.	PUNCT
cana-2643	167	1	the	the	DET
cana-2643	167	2	attention	attention	NOUN
cana-2643	167	3	mechanism	mechanism	NOUN
cana-2643	167	4	computes	compute	VERB
cana-2643	167	5	attention	attention	NOUN
cana-2643	167	6	weights	weight	NOUN
cana-2643	167	7	βi	βi	PRON
cana-2643	167	8	for	for	ADP
cana-2643	167	9	each	each	DET
cana-2643	167	10	region	region	NOUN
cana-2643	167	11	i	i	PRON
cana-2643	167	12	:	:	PUNCT
cana-2643	167	13	x̂i	x̂i	PROPN
cana-2643	167	14	=	=	SYM
cana-2643	167	15	βi	βi	PROPN
cana-2643	167	16	·	·	PUNCT
cana-2643	167	17	xi	xi	X
cana-2643	167	18	′	′	NUM
cana-2643	167	19	the	the	DET
cana-2643	167	20	overall	overall	ADJ
cana-2643	167	21	mpga	mpga	NOUN
cana-2643	167	22	-	-	PUNCT
cana-2643	167	23	rl	rl	NOUN
cana-2643	167	24	-	-	PUNCT
cana-2643	167	25	net	net	ADJ
cana-2643	167	26	framework	framework	NOUN
cana-2643	167	27	integrates	integrate	VERB
cana-2643	167	28	the	the	DET
cana-2643	167	29	mpfe	mpfe	PROPN
cana-2643	167	30	module	module	NOUN
cana-2643	167	31	,	,	PUNCT
cana-2643	167	32	slicbased	slicbase	VERB
cana-2643	167	33	graph	graph	NOUN
cana-2643	167	34	construction	construction	NOUN
cana-2643	167	35	,	,	PUNCT
cana-2643	167	36	gcn	gcn	NOUN
cana-2643	167	37	-	-	PUNCT
cana-2643	167	38	based	base	VERB
cana-2643	167	39	feature	feature	NOUN
cana-2643	167	40	aggregation	aggregation	NOUN
cana-2643	167	41	,	,	PUNCT
cana-2643	167	42	and	and	CCONJ
cana-2643	167	43	rl	rl	NOUN
cana-2643	167	44	-	-	PUNCT
cana-2643	167	45	based	base	VERB
cana-2643	167	46	attention	attention	NOUN
cana-2643	167	47	.	.	PUNCT
cana-2643	168	1	this	this	DET
cana-2643	168	2	flow	flow	NOUN
cana-2643	168	3	ensures	ensure	VERB
cana-2643	168	4	that	that	SCONJ
cana-2643	168	5	the	the	DET
cana-2643	168	6	model	model	NOUN
cana-2643	168	7	captures	capture	VERB
cana-2643	168	8	both	both	DET
cana-2643	168	9	local	local	ADJ
cana-2643	168	10	and	and	CCONJ
cana-2643	168	11	global	global	ADJ
cana-2643	168	12	dependencies	dependency	NOUN
cana-2643	168	13	within	within	ADP
cana-2643	168	14	the	the	DET
cana-2643	168	15	scene	scene	NOUN
cana-2643	168	16	,	,	PUNCT
cana-2643	168	17	dynamically	dynamically	ADV
cana-2643	168	18	emphasizes	emphasize	VERB
cana-2643	168	19	important	important	ADJ
cana-2643	168	20	regions	region	NOUN
cana-2643	168	21	,	,	PUNCT
cana-2643	168	22	and	and	CCONJ
cana-2643	168	23	enhances	enhance	VERB
cana-2643	168	24	the	the	DET
cana-2643	168	25	robot	robot	NOUN
cana-2643	168	26	’s	’s	PART
cana-2643	168	27	ability	ability	NOUN
cana-2643	168	28	to	to	PART
cana-2643	168	29	interpret	interpret	VERB
cana-2643	168	30	,	,	PUNCT
cana-2643	168	31	navigate	navigate	VERB
cana-2643	168	32	,	,	PUNCT
cana-2643	168	33	and	and	CCONJ
cana-2643	168	34	interact	interact	VERB
cana-2643	168	35	effectively	effectively	ADV
cana-2643	168	36	in	in	ADP
cana-2643	168	37	complex	complex	ADJ
cana-2643	168	38	environments	environment	NOUN
cana-2643	168	39	.	.	PUNCT
cana-2643	169	1	4	4	X
cana-2643	169	2	.	.	NOUN
cana-2643	169	3	results	result	NOUN
cana-2643	169	4	and	and	CCONJ
cana-2643	169	5	discussion	discussion	VERB
cana-2643	169	6	the	the	DET
cana-2643	169	7	multi	multi	ADJ
cana-2643	169	8	-	-	ADJ
cana-2643	169	9	path	path	ADJ
cana-2643	169	10	feature	feature	NOUN
cana-2643	169	11	extraction	extraction	NOUN
cana-2643	169	12	(	(	PUNCT
cana-2643	169	13	mpfe	mpfe	NOUN
cana-2643	169	14	)	)	PUNCT
cana-2643	169	15	module	module	NOUN
cana-2643	169	16	in	in	ADP
cana-2643	169	17	the	the	DET
cana-2643	169	18	mpga	mpga	NOUN
cana-2643	169	19	-	-	PUNCT
cana-2643	169	20	rl	rl	NOUN
cana-2643	169	21	-	-	PUNCT
cana-2643	169	22	net	net	NOUN
cana-2643	169	23	framework	framework	NOUN
cana-2643	169	24	applies	apply	VERB
cana-2643	169	25	a	a	DET
cana-2643	169	26	multi	multi	ADJ
cana-2643	169	27	-	-	ADJ
cana-2643	169	28	resolution	resolution	ADJ
cana-2643	169	29	approach	approach	NOUN
cana-2643	169	30	to	to	PART
cana-2643	169	31	ensure	ensure	VERB
cana-2643	169	32	comprehensive	comprehensive	ADJ
cana-2643	169	33	feature	feature	NOUN
cana-2643	169	34	extraction	extraction	NOUN
cana-2643	169	35	from	from	ADP
cana-2643	169	36	the	the	DET
cana-2643	169	37	input	input	NOUN
cana-2643	169	38	scene	scene	NOUN
cana-2643	169	39	,	,	PUNCT
cana-2643	169	40	addressing	address	VERB
cana-2643	169	41	both	both	CCONJ
cana-2643	169	42	broad	broad	ADJ
cana-2643	169	43	contextual	contextual	ADJ
cana-2643	169	44	elements	element	NOUN
cana-2643	169	45	and	and	CCONJ
cana-2643	169	46	fine	fine	ADV
cana-2643	169	47	-	-	PUNCT
cana-2643	169	48	grained	grain	VERB
cana-2643	169	49	details	detail	NOUN
cana-2643	169	50	essential	essential	ADJ
cana-2643	169	51	for	for	ADP
cana-2643	169	52	adaptive	adaptive	ADJ
cana-2643	169	53	robotic	robotic	ADJ
cana-2643	169	54	vision	vision	NOUN
cana-2643	169	55	.	.	PUNCT
cana-2643	170	1	the	the	DET
cana-2643	170	2	process	process	NOUN
cana-2643	170	3	begins	begin	VERB
cana-2643	170	4	with	with	ADP
cana-2643	170	5	the	the	DET
cana-2643	170	6	low	low	ADJ
cana-2643	170	7	-	-	PUNCT
cana-2643	170	8	resolution	resolution	NOUN
cana-2643	170	9	path	path	NOUN
cana-2643	170	10	,	,	PUNCT
cana-2643	170	11	where	where	SCONJ
cana-2643	170	12	the	the	DET
cana-2643	170	13	input	input	NOUN
cana-2643	170	14	scene	scene	NOUN
cana-2643	170	15	is	be	AUX
cana-2643	170	16	significantly	significantly	ADV
cana-2643	170	17	down	down	ADV
cana-2643	170	18	sampled	sample	VERB
cana-2643	170	19	,	,	PUNCT
cana-2643	170	20	capturing	capture	VERB
cana-2643	170	21	only	only	ADV
cana-2643	170	22	broad	broad	ADJ
cana-2643	170	23	shapes	shape	NOUN
cana-2643	170	24	and	and	CCONJ
cana-2643	170	25	the	the	DET
cana-2643	170	26	overall	overall	ADJ
cana-2643	170	27	layout	layout	NOUN
cana-2643	170	28	of	of	ADP
cana-2643	170	29	objects	object	NOUN
cana-2643	170	30	.	.	PUNCT
cana-2643	171	1	this	this	DET
cana-2643	171	2	path	path	NOUN
cana-2643	171	3	enables	enable	VERB
cana-2643	171	4	the	the	DET
cana-2643	171	5	model	model	NOUN
cana-2643	171	6	to	to	PART
cana-2643	171	7	grasp	grasp	VERB
cana-2643	171	8	the	the	DET
cana-2643	171	9	general	general	ADJ
cana-2643	171	10	spatial	spatial	ADJ
cana-2643	171	11	structure	structure	NOUN
cana-2643	171	12	of	of	ADP
cana-2643	171	13	the	the	DET
cana-2643	171	14	scene	scene	NOUN
cana-2643	171	15	while	while	SCONJ
cana-2643	171	16	minimizing	minimize	VERB
cana-2643	171	17	computational	computational	ADJ
cana-2643	171	18	load	load	NOUN
cana-2643	171	19	by	by	ADP
cana-2643	171	20	discarding	discard	VERB
cana-2643	171	21	fine	fine	ADJ
cana-2643	171	22	details	detail	NOUN
cana-2643	171	23	.	.	PUNCT
cana-2643	172	1	the	the	DET
cana-2643	172	2	medium	medium	ADJ
cana-2643	172	3	-	-	PUNCT
cana-2643	172	4	resolution	resolution	NOUN
cana-2643	172	5	path	path	NOUN
cana-2643	172	6	offers	offer	VERB
cana-2643	172	7	a	a	DET
cana-2643	172	8	balanced	balanced	ADJ
cana-2643	172	9	perspective	perspective	NOUN
cana-2643	172	10	by	by	ADP
cana-2643	172	11	retaining	retain	VERB
cana-2643	172	12	intermediate	intermediate	ADJ
cana-2643	172	13	spatial	spatial	ADJ
cana-2643	172	14	resolution	resolution	NOUN
cana-2643	172	15	,	,	PUNCT
cana-2643	172	16	allowing	allow	VERB
cana-2643	172	17	for	for	ADP
cana-2643	172	18	the	the	DET
cana-2643	172	19	recognition	recognition	NOUN
cana-2643	172	20	of	of	ADP
cana-2643	172	21	object	object	NOUN
cana-2643	172	22	boundaries	boundary	NOUN
cana-2643	172	23	and	and	CCONJ
cana-2643	172	24	shapes	shape	NOUN
cana-2643	172	25	with	with	ADP
cana-2643	172	26	moderate	moderate	ADJ
cana-2643	172	27	detail	detail	NOUN
cana-2643	172	28	.	.	PUNCT
cana-2643	173	1	this	this	DET
cana-2643	173	2	path	path	NOUN
cana-2643	173	3	effectively	effectively	ADV
cana-2643	173	4	captures	capture	VERB
cana-2643	173	5	mid	mid	ADJ
cana-2643	173	6	-	-	ADJ
cana-2643	173	7	level	level	ADJ
cana-2643	173	8	features	feature	NOUN
cana-2643	173	9	,	,	PUNCT
cana-2643	173	10	which	which	PRON
cana-2643	173	11	are	be	AUX
cana-2643	173	12	vital	vital	ADJ
cana-2643	173	13	for	for	ADP
cana-2643	173	14	differentiating	differentiate	VERB
cana-2643	173	15	individual	individual	ADJ
cana-2643	173	16	objects	object	NOUN
cana-2643	173	17	—	—	PUNCT
cana-2643	173	18	such	such	ADJ
cana-2643	173	19	as	as	ADP
cana-2643	173	20	distinct	distinct	ADJ
cana-2643	173	21	trash	trash	NOUN
cana-2643	173	22	bins	bin	NOUN
cana-2643	173	23	in	in	ADP
cana-2643	173	24	a	a	DET
cana-2643	173	25	cluttered	cluttered	ADJ
cana-2643	173	26	environment	environment	NOUN
cana-2643	173	27	—	—	PUNCT
cana-2643	173	28	and	and	CCONJ
cana-2643	173	29	for	for	ADP
cana-2643	173	30	understanding	understand	VERB
cana-2643	173	31	their	their	PRON
cana-2643	173	32	spatial	spatial	ADJ
cana-2643	173	33	relationships	relationship	NOUN
cana-2643	173	34	.	.	PUNCT
cana-2643	174	1	the	the	DET
cana-2643	174	2	high	high	ADJ
cana-2643	174	3	-	-	PUNCT
cana-2643	174	4	resolution	resolution	NOUN
cana-2643	174	5	path	path	NOUN
cana-2643	174	6	,	,	PUNCT
cana-2643	174	7	on	on	ADP
cana-2643	174	8	the	the	DET
cana-2643	174	9	other	other	ADJ
cana-2643	174	10	hand	hand	NOUN
cana-2643	174	11	,	,	PUNCT
cana-2643	174	12	preserves	preserve	VERB
cana-2643	174	13	fine	fine	ADJ
cana-2643	174	14	details	detail	NOUN
cana-2643	174	15	by	by	ADP
cana-2643	174	16	keeping	keep	VERB
cana-2643	174	17	the	the	DET
cana-2643	174	18	scene	scene	NOUN
cana-2643	174	19	close	close	ADV
cana-2643	174	20	to	to	ADP
cana-2643	174	21	its	its	PRON
cana-2643	174	22	original	original	ADJ
cana-2643	174	23	resolution	resolution	NOUN
cana-2643	174	24	.	.	PUNCT
cana-2643	175	1	this	this	DET
cana-2643	175	2	path	path	NOUN
cana-2643	175	3	focuses	focus	VERB
cana-2643	175	4	on	on	ADP
cana-2643	175	5	small	small	ADJ
cana-2643	175	6	-	-	PUNCT
cana-2643	175	7	scale	scale	NOUN
cana-2643	175	8	features	feature	NOUN
cana-2643	175	9	like	like	ADP
cana-2643	175	10	edges	edge	NOUN
cana-2643	175	11	,	,	PUNCT
cana-2643	175	12	textures	texture	NOUN
cana-2643	175	13	,	,	PUNCT
cana-2643	175	14	and	and	CCONJ
cana-2643	175	15	labels	label	NOUN
cana-2643	175	16	,	,	PUNCT
cana-2643	175	17	which	which	PRON
cana-2643	175	18	are	be	AUX
cana-2643	175	19	crucial	crucial	ADJ
cana-2643	175	20	for	for	ADP
cana-2643	175	21	precise	precise	ADJ
cana-2643	175	22	object	object	NOUN
cana-2643	175	23	identification	identification	NOUN
cana-2643	175	24	and	and	CCONJ
cana-2643	175	25	understanding	understand	VERB
cana-2643	175	26	intricate	intricate	ADJ
cana-2643	175	27	surface	surface	NOUN
cana-2643	175	28	characteristics	characteristic	NOUN
cana-2643	175	29	.	.	PUNCT
cana-2643	176	1	by	by	ADP
cana-2643	176	2	capturing	capture	VERB
cana-2643	176	3	these	these	DET
cana-2643	176	4	highresolution	highresolution	NOUN
cana-2643	176	5	features	feature	NOUN
cana-2643	176	6	,	,	PUNCT
cana-2643	176	7	the	the	DET
cana-2643	176	8	model	model	NOUN
cana-2643	176	9	gains	gain	VERB
cana-2643	176	10	the	the	DET
cana-2643	176	11	ability	ability	NOUN
cana-2643	176	12	to	to	PART
cana-2643	176	13	recognize	recognize	VERB
cana-2643	176	14	subtle	subtle	ADJ
cana-2643	176	15	distinctions	distinction	NOUN
cana-2643	176	16	,	,	PUNCT
cana-2643	176	17	such	such	ADJ
cana-2643	176	18	as	as	ADP
cana-2643	176	19	identifying	identify	VERB
cana-2643	176	20	specific	specific	ADJ
cana-2643	176	21	text	text	NOUN
cana-2643	176	22	on	on	ADP
cana-2643	176	23	labels	label	NOUN
cana-2643	176	24	or	or	CCONJ
cana-2643	176	25	texture	texture	ADJ
cana-2643	176	26	variations	variation	NOUN
cana-2643	176	27	on	on	ADP
cana-2643	176	28	bin	bin	NOUN
cana-2643	176	29	surfaces	surface	NOUN
cana-2643	176	30	.	.	PUNCT
cana-2643	177	1	following	follow	VERB
cana-2643	177	2	feature	feature	NOUN
cana-2643	177	3	extraction	extraction	NOUN
cana-2643	177	4	at	at	ADP
cana-2643	177	5	each	each	DET
cana-2643	177	6	resolution	resolution	NOUN
cana-2643	177	7	,	,	PUNCT
cana-2643	177	8	mpfe	mpfe	PROPN
cana-2643	177	9	employs	employ	VERB
cana-2643	177	10	an	an	DET
cana-2643	177	11	attentionbased	attentionbase	VERB
cana-2643	177	12	fusion	fusion	NOUN
cana-2643	177	13	mechanism	mechanism	NOUN
cana-2643	177	14	to	to	PART
cana-2643	177	15	combine	combine	VERB
cana-2643	177	16	the	the	DET
cana-2643	177	17	outputs	output	NOUN
cana-2643	177	18	from	from	ADP
cana-2643	177	19	all	all	DET
cana-2643	177	20	three	three	NUM
cana-2643	177	21	paths	path	NOUN
cana-2643	177	22	.	.	PUNCT
cana-2643	178	1	this	this	DET
cana-2643	178	2	fusion	fusion	NOUN
cana-2643	178	3	process	process	NOUN
cana-2643	178	4	adaptively	adaptively	ADV
cana-2643	178	5	weighs	weigh	VERB
cana-2643	178	6	each	each	DET
cana-2643	178	7	path	path	NOUN
cana-2643	178	8	’s	’s	PART
cana-2643	178	9	contribution	contribution	NOUN
cana-2643	178	10	based	base	VERB
cana-2643	178	11	on	on	ADP
cana-2643	178	12	its	its	PRON
cana-2643	178	13	relevance	relevance	NOUN
cana-2643	178	14	to	to	ADP
cana-2643	178	15	the	the	DET
cana-2643	178	16	scene	scene	NOUN
cana-2643	178	17	’s	’s	PART
cana-2643	178	18	requirements	requirement	NOUN
cana-2643	178	19	,	,	PUNCT
cana-2643	178	20	resulting	result	VERB
cana-2643	178	21	in	in	ADP
cana-2643	178	22	a	a	DET
cana-2643	178	23	composite	composite	ADJ
cana-2643	178	24	feature	feature	NOUN
cana-2643	178	25	map	map	NOUN
cana-2643	178	26	that	that	PRON
cana-2643	178	27	integrates	integrate	VERB
cana-2643	178	28	both	both	CCONJ
cana-2643	178	29	global	global	ADJ
cana-2643	178	30	context	context	NOUN
cana-2643	178	31	and	and	CCONJ
cana-2643	178	32	specific	specific	ADJ
cana-2643	178	33	details	detail	NOUN
cana-2643	178	34	.	.	PUNCT
cana-2643	179	1	this	this	DET
cana-2643	179	2	multi	multi	ADJ
cana-2643	179	3	-	-	ADJ
cana-2643	179	4	scale	scale	ADJ
cana-2643	179	5	representation	representation	NOUN
cana-2643	179	6	,	,	PUNCT
cana-2643	179	7	produced	produce	VERB
cana-2643	179	8	by	by	ADP
cana-2643	179	9	mpfe	mpfe	PROPN
cana-2643	179	10	,	,	PUNCT
cana-2643	179	11	equips	equip	VERB
cana-2643	179	12	the	the	DET
cana-2643	179	13	robotic	robotic	ADJ
cana-2643	179	14	vision	vision	NOUN
cana-2643	179	15	system	system	NOUN
cana-2643	179	16	with	with	ADP
cana-2643	179	17	a	a	DET
cana-2643	179	18	rich	rich	ADJ
cana-2643	179	19	,	,	PUNCT
cana-2643	179	20	balanced	balanced	ADJ
cana-2643	179	21	view	view	NOUN
cana-2643	179	22	of	of	ADP
cana-2643	179	23	the	the	DET
cana-2643	179	24	scene	scene	NOUN
cana-2643	179	25	,	,	PUNCT
cana-2643	179	26	ready	ready	ADJ
cana-2643	179	27	for	for	ADP
cana-2643	179	28	further	further	ADJ
cana-2643	179	29	processing	processing	NOUN
cana-2643	179	30	in	in	ADP
cana-2643	179	31	subsequent	subsequent	ADJ
cana-2643	179	32	stages	stage	NOUN
cana-2643	179	33	like	like	ADP
cana-2643	179	34	segmentation	segmentation	NOUN
cana-2643	179	35	and	and	CCONJ
cana-2643	179	36	region	region	NOUN
cana-2643	179	37	prioritization	prioritization	NOUN
cana-2643	179	38	.	.	PUNCT
cana-2643	180	1	the	the	DET
cana-2643	180	2	mpfe	mpfe	PROPN
cana-2643	180	3	module	module	NOUN
cana-2643	180	4	’s	’s	PART
cana-2643	180	5	ability	ability	NOUN
cana-2643	180	6	to	to	PART
cana-2643	180	7	capture	capture	VERB
cana-2643	180	8	and	and	CCONJ
cana-2643	180	9	integrate	integrate	VERB
cana-2643	180	10	multi	multi	ADJ
cana-2643	180	11	-	-	ADJ
cana-2643	180	12	resolution	resolution	ADJ
cana-2643	180	13	features	feature	NOUN
cana-2643	180	14	enhances	enhance	VERB
cana-2643	180	15	mpga	mpga	NOUN
cana-2643	180	16	-	-	PUNCT
cana-2643	180	17	rl	rl	NOUN
cana-2643	180	18	-	-	PUNCT
cana-2643	180	19	net	net	NOUN
cana-2643	180	20	’s	’s	PART
cana-2643	180	21	adaptability	adaptability	NOUN
cana-2643	180	22	in	in	ADP
cana-2643	180	23	complex	complex	ADJ
cana-2643	180	24	environments	environment	NOUN
cana-2643	180	25	,	,	PUNCT
cana-2643	180	26	enabling	enable	VERB
cana-2643	180	27	precise	precise	ADJ
cana-2643	180	28	scene	scene	NOUN
cana-2643	180	29	interpretation	interpretation	NOUN
cana-2643	180	30	that	that	PRON
cana-2643	180	31	is	be	AUX
cana-2643	180	32	essential	essential	ADJ
cana-2643	180	33	for	for	ADP
cana-2643	180	34	real	real	ADJ
cana-2643	180	35	-	-	PUNCT
cana-2643	180	36	time	time	NOUN
cana-2643	180	37	robotic	robotic	ADJ
cana-2643	180	38	navigation	navigation	NOUN
cana-2643	180	39	and	and	CCONJ
cana-2643	180	40	interaction	interaction	NOUN
cana-2643	180	41	in	in	ADP
cana-2643	180	42	diverse	diverse	ADJ
cana-2643	180	43	settings	setting	NOUN
cana-2643	180	44	.	.	PUNCT
cana-2643	181	1	as	as	SCONJ
cana-2643	181	2	shown	show	VERB
cana-2643	181	3	in	in	ADP
cana-2643	181	4	table	table	NOUN
cana-2643	181	5	2	2	NUM
cana-2643	181	6	,	,	PUNCT
cana-2643	181	7	mpga	mpga	NOUN
cana-2643	181	8	-	-	PUNCT
cana-2643	181	9	rl	rl	NOUN
cana-2643	181	10	-	-	PUNCT
cana-2643	181	11	net	net	NOUN
cana-2643	181	12	demonstrates	demonstrate	VERB
cana-2643	181	13	communications	communication	NOUN
cana-2643	181	14	on	on	ADP
cana-2643	181	15	applied	apply	VERB
cana-2643	181	16	nonlinear	nonlinear	ADJ
cana-2643	181	17	analysis	analysis	NOUN
cana-2643	181	18	issn	issn	NOUN
cana-2643	181	19	:	:	PUNCT
cana-2643	181	20	1074	1074	NUM
cana-2643	181	21	-	-	PUNCT
cana-2643	181	22	133x	133x	NUM
cana-2643	181	23	vol	vol	NOUN
cana-2643	181	24	32	32	NUM
cana-2643	181	25	no	no	NOUN
cana-2643	181	26	.	.	PUNCT
cana-2643	182	1	3s	3s	NUM
cana-2643	182	2	(	(	PUNCT
cana-2643	182	3	2025	2025	NUM
cana-2643	182	4	)	)	PUNCT
cana-2643	182	5	289	289	NUM
cana-2643	182	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	182	7	fig	fig	NOUN
cana-2643	182	8	.	.	PUNCT
cana-2643	183	1	2	2	X
cana-2643	183	2	.	.	X
cana-2643	183	3	multi	multi	ADJ
cana-2643	183	4	-	-	ADJ
cana-2643	183	5	path	path	ADJ
cana-2643	183	6	feature	feature	NOUN
cana-2643	183	7	extraction	extraction	NOUN
cana-2643	183	8	(	(	PUNCT
cana-2643	183	9	mpfe	mpfe	NOUN
cana-2643	183	10	)	)	PUNCT
cana-2643	183	11	stages	stage	NOUN
cana-2643	183	12	for	for	ADP
cana-2643	183	13	adaptive	adaptive	ADJ
cana-2643	183	14	robotic	robotic	ADJ
cana-2643	183	15	vision	vision	NOUN
cana-2643	183	16	enhanced	enhance	VERB
cana-2643	183	17	performance	performance	NOUN
cana-2643	183	18	over	over	ADP
cana-2643	183	19	the	the	DET
cana-2643	183	20	u	u	NOUN
cana-2643	183	21	-	-	ADJ
cana-2643	183	22	net	net	ADJ
cana-2643	183	23	variant	variant	NOUN
cana-2643	183	24	in	in	ADP
cana-2643	183	25	terms	term	NOUN
cana-2643	183	26	of	of	ADP
cana-2643	183	27	accuracy	accuracy	NOUN
cana-2643	183	28	,	,	PUNCT
cana-2643	183	29	precision	precision	NOUN
cana-2643	183	30	,	,	PUNCT
cana-2643	183	31	recall	recall	NOUN
cana-2643	183	32	,	,	PUNCT
cana-2643	183	33	and	and	CCONJ
cana-2643	183	34	f1	f1	NOUN
cana-2643	183	35	-	-	PUNCT
cana-2643	183	36	score	score	NOUN
cana-2643	183	37	,	,	PUNCT
cana-2643	183	38	underscoring	underscore	VERB
cana-2643	183	39	its	its	PRON
cana-2643	183	40	ability	ability	NOUN
cana-2643	183	41	to	to	PART
cana-2643	183	42	capture	capture	VERB
cana-2643	183	43	complex	complex	ADJ
cana-2643	183	44	features	feature	NOUN
cana-2643	183	45	in	in	ADP
cana-2643	183	46	cluttered	cluttered	ADJ
cana-2643	183	47	scenes	scene	NOUN
cana-2643	183	48	while	while	SCONJ
cana-2643	183	49	remaining	remain	VERB
cana-2643	183	50	computationally	computationally	ADV
cana-2643	183	51	efficient	efficient	ADJ
cana-2643	183	52	.	.	PUNCT
cana-2643	184	1	this	this	PRON
cana-2643	184	2	is	be	AUX
cana-2643	184	3	attributed	attribute	VERB
cana-2643	184	4	to	to	ADP
cana-2643	184	5	its	its	PRON
cana-2643	184	6	multi	multi	ADJ
cana-2643	184	7	-	-	ADJ
cana-2643	184	8	resolution	resolution	ADJ
cana-2643	184	9	approach	approach	NOUN
cana-2643	184	10	,	,	PUNCT
cana-2643	184	11	where	where	SCONJ
cana-2643	184	12	each	each	DET
cana-2643	184	13	resolution	resolution	NOUN
cana-2643	184	14	path	path	NOUN
cana-2643	184	15	captures	capture	VERB
cana-2643	184	16	distinct	distinct	ADJ
cana-2643	184	17	levels	level	NOUN
cana-2643	184	18	of	of	ADP
cana-2643	184	19	detail	detail	NOUN
cana-2643	184	20	:	:	PUNCT
cana-2643	184	21	low	low	ADJ
cana-2643	184	22	resolution	resolution	NOUN
cana-2643	184	23	for	for	ADP
cana-2643	184	24	broad	broad	ADJ
cana-2643	184	25	spatial	spatial	ADJ
cana-2643	184	26	structure	structure	NOUN
cana-2643	184	27	,	,	PUNCT
cana-2643	184	28	medium	medium	NOUN
cana-2643	184	29	for	for	ADP
cana-2643	184	30	object	object	NOUN
cana-2643	184	31	boundaries	boundary	NOUN
cana-2643	184	32	.	.	PUNCT
cana-2643	185	1	and	and	CCONJ
cana-2643	185	2	high	high	ADJ
cana-2643	185	3	resolution	resolution	NOUN
cana-2643	185	4	for	for	ADP
cana-2643	185	5	fine	fine	ADJ
cana-2643	185	6	details	detail	NOUN
cana-2643	185	7	,	,	PUNCT
cana-2643	185	8	as	as	SCONJ
cana-2643	185	9	detailed	detailed	ADJ
cana-2643	185	10	in	in	ADP
cana-2643	185	11	table	table	NOUN
cana-2643	185	12	3	3	NUM
cana-2643	185	13	.	.	PUNCT
cana-2643	186	1	the	the	DET
cana-2643	186	2	attention	attention	NOUN
cana-2643	186	3	-	-	PUNCT
cana-2643	186	4	based	base	VERB
cana-2643	186	5	fusion	fusion	NOUN
cana-2643	186	6	further	far	ADV
cana-2643	186	7	combines	combine	VERB
cana-2643	186	8	these	these	DET
cana-2643	186	9	paths	path	NOUN
cana-2643	186	10	to	to	PART
cana-2643	186	11	create	create	VERB
cana-2643	186	12	a	a	DET
cana-2643	186	13	comprehensive	comprehensive	ADJ
cana-2643	186	14	feature	feature	NOUN
cana-2643	186	15	map	map	NOUN
cana-2643	186	16	,	,	PUNCT
cana-2643	186	17	dynamically	dynamically	ADV
cana-2643	186	18	weighting	weight	VERB
cana-2643	186	19	contributions	contribution	NOUN
cana-2643	186	20	from	from	ADP
cana-2643	186	21	each	each	DET
cana-2643	186	22	path	path	NOUN
cana-2643	186	23	based	base	VERB
cana-2643	186	24	on	on	ADP
cana-2643	186	25	scene	scene	NOUN
cana-2643	186	26	requirements	requirement	NOUN
cana-2643	186	27	.	.	PUNCT
cana-2643	187	1	this	this	DET
cana-2643	187	2	design	design	NOUN
cana-2643	187	3	reduces	reduce	VERB
cana-2643	187	4	processing	processing	NOUN
cana-2643	187	5	time	time	NOUN
cana-2643	187	6	and	and	CCONJ
cana-2643	187	7	memory	memory	NOUN
cana-2643	187	8	usage	usage	NOUN
cana-2643	187	9	,	,	PUNCT
cana-2643	187	10	making	make	VERB
cana-2643	187	11	mpga	mpga	NOUN
cana-2643	187	12	-	-	PUNCT
cana-2643	187	13	rl	rl	NOUN
cana-2643	187	14	-	-	PUNCT
cana-2643	187	15	net	net	NOUN
cana-2643	187	16	a	a	DET
cana-2643	187	17	suitable	suitable	ADJ
cana-2643	187	18	choice	choice	NOUN
cana-2643	187	19	for	for	ADP
cana-2643	187	20	adaptive	adaptive	ADJ
cana-2643	187	21	robotic	robotic	ADJ
cana-2643	187	22	vision	vision	NOUN
cana-2643	187	23	tasks	task	NOUN
cana-2643	187	24	in	in	ADP
cana-2643	187	25	diverse	diverse	ADJ
cana-2643	187	26	and	and	CCONJ
cana-2643	187	27	real	real	ADJ
cana-2643	187	28	-	-	PUNCT
cana-2643	187	29	time	time	NOUN
cana-2643	187	30	settings	setting	NOUN
cana-2643	187	31	.	.	PUNCT
cana-2643	188	1	table	table	NOUN
cana-2643	188	2	4	4	NUM
cana-2643	188	3	presents	present	VERB
cana-2643	188	4	a	a	DET
cana-2643	188	5	resolution	resolution	NOUN
cana-2643	188	6	-	-	PUNCT
cana-2643	188	7	wise	wise	ADJ
cana-2643	188	8	performance	performance	NOUN
cana-2643	188	9	evaluation	evaluation	NOUN
cana-2643	188	10	of	of	ADP
cana-2643	188	11	mpga	mpga	NOUN
cana-2643	188	12	-	-	PUNCT
cana-2643	188	13	rl	rl	NOUN
cana-2643	188	14	.	.	NOUN
cana-2643	188	15	table	table	NOUN
cana-2643	189	1	2	2	NUM
cana-2643	189	2	.	.	PUNCT
cana-2643	189	3	quantitative	quantitative	ADJ
cana-2643	189	4	performance	performance	NOUN
cana-2643	189	5	comparison	comparison	NOUN
cana-2643	189	6	between	between	ADP
cana-2643	189	7	mpga	mpga	NOUN
cana-2643	189	8	-	-	PUNCT
cana-2643	189	9	rl	rl	NOUN
cana-2643	189	10	-	-	PUNCT
cana-2643	189	11	net	net	ADJ
cana-2643	189	12	and	and	CCONJ
cana-2643	189	13	u	u	NOUN
cana-2643	189	14	-	-	ADJ
cana-2643	189	15	net	net	ADJ
cana-2643	189	16	variant	variant	NOUN
cana-2643	189	17	metric	metric	ADJ
cana-2643	189	18	mpga	mpga	NOUN
cana-2643	189	19	-	-	PUNCT
cana-2643	189	20	rl	rl	NOUN
cana-2643	189	21	-	-	PUNCT
cana-2643	189	22	net	net	NOUN
cana-2643	189	23	(	(	PUNCT
cana-2643	189	24	proposed	propose	VERB
cana-2643	189	25	)	)	PUNCT
cana-2643	189	26	u	u	NOUN
cana-2643	189	27	-	-	NOUN
cana-2643	189	28	netvariant	netvariant	ADJ
cana-2643	189	29	(	(	PUNCT
cana-2643	189	30	baseline	baseline	NOUN
cana-2643	189	31	)	)	PUNCT
cana-2643	189	32	improvement	improvement	NOUN
cana-2643	189	33	(	(	PUNCT
cana-2643	189	34	%	%	NOUN
cana-2643	189	35	)	)	PUNCT
cana-2643	189	36	accuracy	accuracy	NOUN
cana-2643	189	37	92.5	92.5	NUM
cana-2643	189	38	%	%	NOUN
cana-2643	189	39	86.7	86.7	NUM
cana-2643	189	40	%	%	NOUN
cana-2643	189	41	+6.7	+6.7	ADP
cana-2643	189	42	%	%	NOUN
cana-2643	189	43	precision	precision	NOUN
cana-2643	189	44	91.8	91.8	NUM
cana-2643	189	45	%	%	NOUN
cana-2643	189	46	85.2	85.2	NUM
cana-2643	189	47	%	%	NOUN
cana-2643	189	48	+7.8	+7.8	NOUN
cana-2643	189	49	%	%	NOUN
cana-2643	189	50	recall	recall	VERB
cana-2643	189	51	93.3	93.3	NUM
cana-2643	189	52	%	%	NOUN
cana-2643	189	53	87.0	87.0	NUM
cana-2643	189	54	%	%	NOUN
cana-2643	189	55	+7.2	+7.2	NUM
cana-2643	189	56	%	%	NOUN
cana-2643	189	57	f1	f1	NOUN
cana-2643	189	58	-	-	PUNCT
cana-2643	189	59	score	score	NOUN
cana-2643	189	60	92.5	92.5	NUM
cana-2643	189	61	%	%	NOUN
cana-2643	189	62	86.0	86.0	NUM
cana-2643	189	63	%	%	NOUN
cana-2643	189	64	+7.6	+7.6	NUM
cana-2643	189	65	%	%	NOUN
cana-2643	189	66	processing	processing	NOUN
cana-2643	189	67	time	time	NOUN
cana-2643	189	68	(	(	PUNCT
cana-2643	189	69	ms	ms	NOUN
cana-2643	189	70	)	)	PUNCT
cana-2643	189	71	150	150	NUM
cana-2643	189	72	ms	ms	NOUN
cana-2643	189	73	210	210	NUM
cana-2643	189	74	ms	ms	NOUN
cana-2643	189	75	-28.6	-28.6	PROPN
cana-2643	189	76	%	%	NOUN
cana-2643	189	77	memory	memory	NOUN
cana-2643	189	78	usage	usage	NOUN
cana-2643	189	79	(	(	PUNCT
cana-2643	189	80	mb	mb	NOUN
cana-2643	189	81	)	)	PUNCT
cana-2643	189	82	500	500	NUM
cana-2643	189	83	mb	mb	NOUN
cana-2643	189	84	650	650	NUM
cana-2643	189	85	mb	mb	NOUN
cana-2643	189	86	-23.1	-23.1	NUM
cana-2643	189	87	%	%	NOUN
cana-2643	189	88	communications	communication	NOUN
cana-2643	189	89	on	on	ADP
cana-2643	189	90	applied	apply	VERB
cana-2643	189	91	nonlinear	nonlinear	ADJ
cana-2643	189	92	analysis	analysis	NOUN
cana-2643	189	93	issn	issn	NOUN
cana-2643	189	94	:	:	PUNCT
cana-2643	189	95	1074	1074	NUM
cana-2643	189	96	-	-	PUNCT
cana-2643	189	97	133x	133x	NUM
cana-2643	189	98	vol	vol	NOUN
cana-2643	189	99	32	32	NUM
cana-2643	189	100	no	no	NOUN
cana-2643	189	101	.	.	PUNCT
cana-2643	190	1	3s	3s	NUM
cana-2643	190	2	(	(	PUNCT
cana-2643	190	3	2025	2025	NUM
cana-2643	190	4	)	)	PUNCT
cana-2643	190	5	290	290	NUM
cana-2643	190	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	190	7	table	table	NOUN
cana-2643	190	8	3	3	NUM
cana-2643	190	9	.	.	PUNCT
cana-2643	190	10	resolution	resolution	NOUN
cana-2643	190	11	-	-	PUNCT
cana-2643	190	12	wise	wise	ADJ
cana-2643	190	13	performance	performance	NOUN
cana-2643	190	14	analysis	analysis	NOUN
cana-2643	190	15	of	of	ADP
cana-2643	190	16	mpga	mpga	NOUN
cana-2643	190	17	-	-	PUNCT
cana-2643	190	18	rl	rl	NOUN
cana-2643	190	19	-	-	PUNCT
cana-2643	190	20	net	net	NOUN
cana-2643	190	21	resolution	resolution	NOUN
cana-2643	190	22	path	path	NOUN
cana-2643	190	23	precision	precision	NOUN
cana-2643	190	24	(	(	PUNCT
cana-2643	190	25	%	%	INTJ
cana-2643	190	26	)	)	PUNCT
cana-2643	190	27	recall	recall	NOUN
cana-2643	190	28	(	(	PUNCT
cana-2643	190	29	%	%	NOUN
cana-2643	190	30	)	)	PUNCT
cana-2643	190	31	f1	f1	NOUN
cana-2643	190	32	-	-	PUNCT
cana-2643	190	33	score	score	NOUN
cana-2643	190	34	(	(	PUNCT
cana-2643	190	35	%	%	INTJ
cana-2643	190	36	)	)	PUNCT
cana-2643	190	37	processin	processin	NOUN
cana-2643	190	38	g	g	PROPN
cana-2643	190	39	time	time	NOUN
cana-2643	190	40	(	(	PUNCT
cana-2643	190	41	ms	ms	NOUN
cana-2643	190	42	)	)	PUNCT
cana-2643	190	43	low	low	ADJ
cana-2643	190	44	resolution	resolution	NOUN
cana-2643	190	45	87.4	87.4	NUM
cana-2643	190	46	%	%	NOUN
cana-2643	190	47	88.2	88.2	NUM
cana-2643	190	48	%	%	NOUN
cana-2643	190	49	87.8	87.8	NUM
cana-2643	190	50	%	%	NOUN
cana-2643	190	51	40	40	NUM
cana-2643	190	52	ms	ms	NOUN
cana-2643	190	53	medium	medium	ADJ
cana-2643	190	54	resolution	resolution	NOUN
cana-2643	190	55	91.1	91.1	NUM
cana-2643	190	56	%	%	NOUN
cana-2643	190	57	90.6	90.6	NUM
cana-2643	190	58	%	%	NOUN
cana-2643	190	59	90.8	90.8	NUM
cana-2643	190	60	%	%	NOUN
cana-2643	190	61	55	55	NUM
cana-2643	190	62	ms	ms	NOUN
cana-2643	190	63	high	high	ADJ
cana-2643	190	64	resolution	resolution	NOUN
cana-2643	190	65	92.5	92.5	NUM
cana-2643	190	66	%	%	NOUN
cana-2643	190	67	93.0	93.0	NUM
cana-2643	190	68	%	%	NOUN
cana-2643	190	69	92.7	92.7	NUM
cana-2643	190	70	%	%	NOUN
cana-2643	190	71	55	55	NUM
cana-2643	190	72	ms	ms	ADJ
cana-2643	190	73	attention	attention	NOUN
cana-2643	190	74	-	-	PUNCT
cana-2643	190	75	based	base	VERB
cana-2643	190	76	fusion	fusion	NOUN
cana-2643	190	77	92.5	92.5	NUM
cana-2643	190	78	%	%	NOUN
cana-2643	190	79	93.3	93.3	NUM
cana-2643	190	80	%	%	NOUN
cana-2643	190	81	92.5	92.5	NUM
cana-2643	190	82	%	%	NOUN
cana-2643	190	83	150	150	NUM
cana-2643	190	84	ms	ms	NOUN
cana-2643	190	85	net	net	NOUN
cana-2643	190	86	,	,	PUNCT
cana-2643	190	87	illustrating	illustrate	VERB
cana-2643	190	88	how	how	SCONJ
cana-2643	190	89	each	each	DET
cana-2643	190	90	resolution	resolution	NOUN
cana-2643	190	91	path	path	NOUN
cana-2643	190	92	(	(	PUNCT
cana-2643	190	93	low	low	ADJ
cana-2643	190	94	,	,	PUNCT
cana-2643	190	95	medium	medium	ADJ
cana-2643	190	96	,	,	PUNCT
cana-2643	190	97	and	and	CCONJ
cana-2643	190	98	high	high	ADJ
cana-2643	190	99	)	)	PUNCT
cana-2643	190	100	contributes	contribute	VERB
cana-2643	190	101	uniquely	uniquely	ADV
cana-2643	190	102	to	to	ADP
cana-2643	190	103	segmentation	segmentation	NOUN
cana-2643	190	104	metrics	metric	NOUN
cana-2643	190	105	,	,	PUNCT
cana-2643	190	106	including	include	VERB
cana-2643	190	107	mean	mean	ADJ
cana-2643	190	108	iou	iou	PROPN
cana-2643	190	109	,	,	PUNCT
cana-2643	190	110	pixel	pixel	PROPN
cana-2643	190	111	accuracy	accuracy	NOUN
cana-2643	190	112	,	,	PUNCT
cana-2643	190	113	boundary	boundary	ADJ
cana-2643	190	114	f1	f1	NOUN
cana-2643	190	115	score	score	NOUN
cana-2643	190	116	,	,	PUNCT
cana-2643	190	117	inference	inference	NOUN
cana-2643	190	118	time	time	NOUN
cana-2643	190	119	,	,	PUNCT
cana-2643	190	120	and	and	CCONJ
cana-2643	190	121	memory	memory	NOUN
cana-2643	190	122	usage[31	usage[31	NOUN
cana-2643	190	123	]	]	X
cana-2643	190	124	.	.	PUNCT
cana-2643	191	1	each	each	DET
cana-2643	191	2	path	path	NOUN
cana-2643	191	3	provides	provide	VERB
cana-2643	191	4	different	different	ADJ
cana-2643	191	5	levels	level	NOUN
cana-2643	191	6	of	of	ADP
cana-2643	191	7	detail	detail	NOUN
cana-2643	191	8	and	and	CCONJ
cana-2643	191	9	spatial	spatial	ADJ
cana-2643	191	10	context	context	NOUN
cana-2643	191	11	,	,	PUNCT
cana-2643	191	12	with	with	ADP
cana-2643	191	13	low	low	ADJ
cana-2643	191	14	resolution	resolution	NOUN
cana-2643	191	15	focusing	focus	VERB
cana-2643	191	16	on	on	ADP
cana-2643	191	17	broad	broad	ADJ
cana-2643	191	18	shapes	shape	NOUN
cana-2643	191	19	,	,	PUNCT
cana-2643	191	20	medium	medium	ADJ
cana-2643	191	21	resolution	resolution	NOUN
cana-2643	191	22	enhancing	enhance	VERB
cana-2643	191	23	object	object	NOUN
cana-2643	191	24	boundaries	boundary	NOUN
cana-2643	191	25	,	,	PUNCT
cana-2643	191	26	and	and	CCONJ
cana-2643	191	27	high	high	ADJ
cana-2643	191	28	resolution	resolution	NOUN
cana-2643	191	29	capturing	capture	VERB
cana-2643	191	30	fine	fine	ADJ
cana-2643	191	31	details	detail	NOUN
cana-2643	191	32	.	.	PUNCT
cana-2643	192	1	this	this	DET
cana-2643	192	2	layered	layered	ADJ
cana-2643	192	3	approach	approach	NOUN
cana-2643	192	4	aligns	align	VERB
cana-2643	192	5	with	with	ADP
cana-2643	192	6	studies	study	NOUN
cana-2643	192	7	on	on	ADP
cana-2643	192	8	multi	multi	ADJ
cana-2643	192	9	-	-	ADJ
cana-2643	192	10	resolution	resolution	ADJ
cana-2643	192	11	feature	feature	NOUN
cana-2643	192	12	extraction	extraction	NOUN
cana-2643	192	13	for	for	ADP
cana-2643	192	14	scene	scene	NOUN
cana-2643	192	15	segmentation	segmentation	NOUN
cana-2643	192	16	,	,	PUNCT
cana-2643	192	17	where	where	SCONJ
cana-2643	192	18	different	different	ADJ
cana-2643	192	19	resolutions	resolution	NOUN
cana-2643	192	20	improve	improve	VERB
cana-2643	192	21	overall	overall	ADJ
cana-2643	192	22	accuracy	accuracy	NOUN
cana-2643	192	23	and	and	CCONJ
cana-2643	192	24	computational	computational	ADJ
cana-2643	192	25	efficiency	efficiency	NOUN
cana-2643	192	26	.	.	PUNCT
cana-2643	193	1	table	table	NOUN
cana-2643	193	2	5	5	NUM
cana-2643	193	3	shows	show	VERB
cana-2643	193	4	an	an	DET
cana-2643	193	5	ablation	ablation	NOUN
cana-2643	193	6	study	study	NOUN
cana-2643	193	7	of	of	ADP
cana-2643	193	8	mpga	mpga	NOUN
cana-2643	193	9	-	-	PUNCT
cana-2643	193	10	rl	rl	NOUN
cana-2643	193	11	-	-	PUNCT
cana-2643	193	12	net	net	NOUN
cana-2643	193	13	’s	’s	PART
cana-2643	193	14	components	component	NOUN
cana-2643	193	15	,	,	PUNCT
cana-2643	193	16	testing	test	VERB
cana-2643	193	17	the	the	DET
cana-2643	193	18	impact	impact	NOUN
cana-2643	193	19	of	of	ADP
cana-2643	193	20	each	each	DET
cana-2643	193	21	resolution	resolution	NOUN
cana-2643	193	22	path	path	NOUN
cana-2643	193	23	individually	individually	ADV
cana-2643	193	24	and	and	CCONJ
cana-2643	193	25	in	in	ADP
cana-2643	193	26	combination	combination	NOUN
cana-2643	193	27	.	.	PUNCT
cana-2643	194	1	the	the	DET
cana-2643	194	2	results	result	NOUN
cana-2643	194	3	demonstrate	demonstrate	VERB
cana-2643	194	4	how	how	SCONJ
cana-2643	194	5	adding	add	VERB
cana-2643	194	6	each	each	DET
cana-2643	194	7	path	path	NOUN
cana-2643	194	8	incrementally	incrementally	ADV
cana-2643	194	9	enhances	enhance	VERB
cana-2643	194	10	mean	mean	VERB
cana-2643	194	11	iou	iou	NOUN
cana-2643	194	12	and	and	CCONJ
cana-2643	194	13	table	table	NOUN
cana-2643	194	14	4	4	NUM
cana-2643	194	15	.	.	PUNCT
cana-2643	194	16	resolution	resolution	NOUN
cana-2643	194	17	-	-	PUNCT
cana-2643	194	18	wise	wise	ADJ
cana-2643	194	19	performance	performance	NOUN
cana-2643	194	20	evaluation	evaluation	NOUN
cana-2643	194	21	of	of	ADP
cana-2643	194	22	mpga	mpga	NOUN
cana-2643	194	23	-	-	PUNCT
cana-2643	194	24	rl	rl	NOUN
cana-2643	194	25	-	-	PUNCT
cana-2643	194	26	net	net	NOUN
cana-2643	194	27	resolutio	resolutio	NOUN
cana-2643	194	28	n	n	PRON
cana-2643	194	29	path	path	NOUN
cana-2643	195	1	mean	mean	VERB
cana-2643	195	2	i	i	PRON
cana-2643	195	3	ou	ou	VERB
cana-2643	195	4	(	(	PUNCT
cana-2643	195	5	%	%	INTJ
cana-2643	195	6	)	)	PUNCT
cana-2643	195	7	pixel	pixel	PROPN
cana-2643	195	8	accu	accu	PROPN
cana-2643	195	9	racy	racy	PROPN
cana-2643	195	10	(	(	PUNCT
cana-2643	195	11	%	%	INTJ
cana-2643	195	12	)	)	PUNCT
cana-2643	195	13	boundary	boundary	PROPN
cana-2643	195	14	f1	f1	PROPN
cana-2643	195	15	sco	sco	PROPN
cana-2643	195	16	re	re	X
cana-2643	195	17	(	(	PUNCT
cana-2643	195	18	%	%	INTJ
cana-2643	195	19	)	)	PUNCT
cana-2643	195	20	inference	inference	NOUN
cana-2643	195	21	time	time	NOUN
cana-2643	195	22	(	(	PUNCT
cana-2643	195	23	ms	ms	NOUN
cana-2643	195	24	)	)	PUNCT
cana-2643	195	25	memory	memory	NOUN
cana-2643	195	26	usage	usage	NOUN
cana-2643	195	27	(	(	PUNCT
cana-2643	195	28	mb	mb	NOUN
cana-2643	195	29	)	)	PUNCT
cana-2643	195	30	low	low	PROPN
cana-2643	195	31	resolu	resolu	PROPN
cana-2643	195	32	tion	tion	NOUN
cana-2643	195	33	80.2	80.2	NUM
cana-2643	195	34	%	%	NOUN
cana-2643	195	35	88.5	88.5	NUM
cana-2643	195	36	%	%	NOUN
cana-2643	195	37	84.3	84.3	NUM
cana-2643	195	38	%	%	NOUN
cana-2643	195	39	40	40	NUM
cana-2643	195	40	ms	ms	NOUN
cana-2643	195	41	150	150	NUM
cana-2643	195	42	mb	mb	NOUN
cana-2643	195	43	medium	medium	ADJ
cana-2643	195	44	resolutio	resolutio	NOUN
cana-2643	195	45	n	n	CCONJ
cana-2643	195	46	84.5	84.5	NUM
cana-2643	195	47	%	%	NOUN
cana-2643	195	48	91.1	91.1	NUM
cana-2643	195	49	%	%	NOUN
cana-2643	195	50	87.6	87.6	NUM
cana-2643	195	51	%	%	NOUN
cana-2643	195	52	55	55	NUM
cana-2643	195	53	ms	ms	NOUN
cana-2643	195	54	200	200	NUM
cana-2643	195	55	mb	mb	ADP
cana-2643	195	56	high	high	ADJ
cana-2643	195	57	reso	reso	NOUN
cana-2643	195	58	lution	lution	PROPN
cana-2643	195	59	85.8	85.8	NUM
cana-2643	195	60	%	%	NOUN
cana-2643	195	61	92.4	92.4	NUM
cana-2643	195	62	%	%	NOUN
cana-2643	195	63	89.7	89.7	NUM
cana-2643	195	64	%	%	NOUN
cana-2643	195	65	65	65	NUM
cana-2643	195	66	ms	ms	NOUN
cana-2643	195	67	250	250	NUM
cana-2643	195	68	mb	mb	ADP
cana-2643	195	69	boundary	boundary	ADJ
cana-2643	195	70	f1	f1	PROPN
cana-2643	195	71	score	score	NOUN
cana-2643	195	72	,	,	PUNCT
cana-2643	195	73	with	with	ADP
cana-2643	195	74	the	the	DET
cana-2643	195	75	highest	high	ADJ
cana-2643	195	76	metrics	metric	NOUN
cana-2643	195	77	achieved	achieve	VERB
cana-2643	195	78	by	by	ADP
cana-2643	195	79	combining	combine	VERB
cana-2643	195	80	all	all	DET
cana-2643	195	81	three	three	NUM
cana-2643	195	82	paths	path	NOUN
cana-2643	195	83	and	and	CCONJ
cana-2643	195	84	applying	apply	VERB
cana-2643	195	85	attention	attention	NOUN
cana-2643	195	86	-	-	PUNCT
cana-2643	195	87	based	base	VERB
cana-2643	195	88	fusion	fusion	NOUN
cana-2643	195	89	.	.	PUNCT
cana-2643	196	1	the	the	DET
cana-2643	196	2	attention	attention	NOUN
cana-2643	196	3	-	-	PUNCT
cana-2643	196	4	based	base	VERB
cana-2643	196	5	fusion	fusion	NOUN
cana-2643	196	6	adaptively	adaptively	ADV
cana-2643	196	7	balances	balance	VERB
cana-2643	196	8	contributions	contribution	NOUN
cana-2643	196	9	,	,	PUNCT
cana-2643	196	10	resulting	result	VERB
cana-2643	196	11	in	in	ADP
cana-2643	196	12	optimal	optimal	ADJ
cana-2643	196	13	segmentation	segmentation	NOUN
cana-2643	196	14	precision	precision	NOUN
cana-2643	196	15	and	and	CCONJ
cana-2643	196	16	efficiency	efficiency	NOUN
cana-2643	196	17	,	,	PUNCT
cana-2643	196	18	consistent	consistent	ADJ
cana-2643	196	19	with	with	ADP
cana-2643	196	20	findings	finding	NOUN
cana-2643	196	21	in	in	ADP
cana-2643	196	22	adaptive	adaptive	ADJ
cana-2643	196	23	feature	feature	NOUN
cana-2643	196	24	fusion	fusion	NOUN
cana-2643	196	25	.	.	PUNCT
cana-2643	197	1	table	table	NOUN
cana-2643	197	2	6	6	NUM
cana-2643	197	3	provides	provide	VERB
cana-2643	197	4	insights	insight	NOUN
cana-2643	197	5	into	into	ADP
cana-2643	197	6	the	the	DET
cana-2643	197	7	adaptive	adaptive	ADJ
cana-2643	197	8	fusion	fusion	NOUN
cana-2643	197	9	weights	weight	NOUN
cana-2643	197	10	assigned	assign	VERB
cana-2643	197	11	to	to	ADP
cana-2643	197	12	each	each	DET
cana-2643	197	13	resolution	resolution	NOUN
cana-2643	197	14	path	path	NOUN
cana-2643	197	15	across	across	ADP
cana-2643	197	16	different	different	ADJ
cana-2643	197	17	scene	scene	NOUN
cana-2643	197	18	complexities	complexity	NOUN
cana-2643	197	19	(	(	PUNCT
cana-2643	197	20	simple	simple	ADJ
cana-2643	197	21	,	,	PUNCT
cana-2643	197	22	moderate	moderate	ADJ
cana-2643	197	23	,	,	PUNCT
cana-2643	197	24	and	and	CCONJ
cana-2643	197	25	complex	complex	ADJ
cana-2643	197	26	)	)	PUNCT
cana-2643	197	27	.	.	PUNCT
cana-2643	198	1	the	the	DET
cana-2643	198	2	adaptive	adaptive	ADJ
cana-2643	198	3	fusion	fusion	NOUN
cana-2643	198	4	mechanism	mechanism	NOUN
cana-2643	198	5	prioritizes	prioritize	VERB
cana-2643	198	6	high	high	ADJ
cana-2643	198	7	-	-	PUNCT
cana-2643	198	8	resolution	resolution	NOUN
cana-2643	198	9	features	feature	NOUN
cana-2643	198	10	in	in	ADP
cana-2643	198	11	complex	complex	ADJ
cana-2643	198	12	scenes	scene	NOUN
cana-2643	198	13	to	to	PART
cana-2643	198	14	capture	capture	VERB
cana-2643	198	15	intricate	intricate	ADJ
cana-2643	198	16	details	detail	NOUN
cana-2643	198	17	,	,	PUNCT
cana-2643	198	18	while	while	SCONJ
cana-2643	198	19	simple	simple	ADJ
cana-2643	198	20	scenes	scene	NOUN
cana-2643	198	21	rely	rely	VERB
cana-2643	198	22	more	more	ADV
cana-2643	198	23	on	on	ADP
cana-2643	198	24	lowresolution	lowresolution	NOUN
cana-2643	198	25	paths	path	NOUN
cana-2643	198	26	to	to	PART
cana-2643	198	27	minimize	minimize	VERB
cana-2643	198	28	computational	computational	ADJ
cana-2643	198	29	load	load	NOUN
cana-2643	198	30	.	.	PUNCT
cana-2643	199	1	this	this	DET
cana-2643	199	2	adaptive	adaptive	ADJ
cana-2643	199	3	weighting	weighting	NOUN
cana-2643	199	4	aligns	align	VERB
cana-2643	199	5	with	with	ADP
cana-2643	199	6	approaches	approach	NOUN
cana-2643	199	7	in	in	ADP
cana-2643	199	8	scene	scene	NOUN
cana-2643	199	9	-	-	PUNCT
cana-2643	199	10	adaptive	adaptive	ADJ
cana-2643	199	11	attention	attention	NOUN
cana-2643	199	12	mechanisms	mechanism	NOUN
cana-2643	199	13	and	and	CCONJ
cana-2643	199	14	highlights	highlight	NOUN
cana-2643	199	15	mpgarl	mpgarl	PROPN
cana-2643	199	16	-	-	PUNCT
cana-2643	199	17	net	net	NOUN
cana-2643	199	18	’s	’s	PART
cana-2643	199	19	efficiency	efficiency	NOUN
cana-2643	199	20	in	in	ADP
cana-2643	199	21	diverse	diverse	ADJ
cana-2643	199	22	environmental	environmental	ADJ
cana-2643	199	23	contexts	contexts	NOUN
cana-2643	199	24	.	.	PUNCT
cana-2643	200	1	figure	figure	VERB
cana-2643	200	2	4shows	4shows	NUM
cana-2643	200	3	a	a	DET
cana-2643	200	4	side	side	NOUN
cana-2643	200	5	-	-	PUNCT
cana-2643	200	6	by	by	ADP
cana-2643	200	7	communications	communication	NOUN
cana-2643	200	8	on	on	ADP
cana-2643	200	9	applied	apply	VERB
cana-2643	200	10	nonlinear	nonlinear	ADJ
cana-2643	200	11	analysis	analysis	NOUN
cana-2643	200	12	issn	issn	NOUN
cana-2643	200	13	:	:	PUNCT
cana-2643	200	14	1074	1074	NUM
cana-2643	200	15	-	-	PUNCT
cana-2643	200	16	133x	133x	NUM
cana-2643	200	17	vol	vol	NOUN
cana-2643	200	18	32	32	NUM
cana-2643	200	19	no	no	NOUN
cana-2643	200	20	.	.	PUNCT
cana-2643	201	1	3s	3s	NUM
cana-2643	201	2	(	(	PUNCT
cana-2643	201	3	2025	2025	NUM
cana-2643	201	4	)	)	PUNCT
cana-2643	201	5	291	291	NUM
cana-2643	201	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	201	7	table	table	NOUN
cana-2643	201	8	5	5	NUM
cana-2643	201	9	.	.	PUNCT
cana-2643	201	10	ablation	ablation	NOUN
cana-2643	201	11	study	study	NOUN
cana-2643	201	12	of	of	ADP
cana-2643	201	13	mpga	mpga	NOUN
cana-2643	201	14	-	-	PUNCT
cana-2643	201	15	rl	rl	NOUN
cana-2643	201	16	-	-	PUNCT
cana-2643	201	17	net	net	ADJ
cana-2643	201	18	components	component	NOUN
cana-2643	201	19	configuration	configuration	NOUN
cana-2643	201	20	mean	mean	VERB
cana-2643	201	21	i	i	PRON
cana-2643	201	22	ou	ou	VERB
cana-2643	201	23	(	(	PUNCT
cana-2643	201	24	%	%	INTJ
cana-2643	201	25	)	)	PUNCT
cana-2643	201	26	pixel	pixel	PROPN
cana-2643	201	27	accu	accu	PROPN
cana-2643	201	28	racy	racy	PROPN
cana-2643	201	29	(	(	PUNCT
cana-2643	201	30	%	%	INTJ
cana-2643	201	31	)	)	PUNCT
cana-2643	201	32	boundary	boundary	PROPN
cana-2643	201	33	f1	f1	PROPN
cana-2643	201	34	sco	sco	PROPN
cana-2643	201	35	re	re	X
cana-2643	201	36	(	(	PUNCT
cana-2643	201	37	%	%	INTJ
cana-2643	201	38	)	)	PUNCT
cana-2643	201	39	inference	inference	NOUN
cana-2643	201	40	time	time	NOUN
cana-2643	201	41	(	(	PUNCT
cana-2643	201	42	ms	ms	NOUN
cana-2643	201	43	)	)	PUNCT
cana-2643	201	44	low	low	ADJ
cana-2643	201	45	resolution	resolution	NOUN
cana-2643	201	46	only	only	ADV
cana-2643	201	47	80.2	80.2	NUM
cana-2643	201	48	%	%	NOUN
cana-2643	201	49	88.5	88.5	NUM
cana-2643	201	50	%	%	NOUN
cana-2643	201	51	84.3	84.3	NUM
cana-2643	201	52	%	%	NOUN
cana-2643	201	53	40	40	NUM
cana-2643	201	54	ms	ms	NOUN
cana-2643	201	55	medium	medium	ADJ
cana-2643	201	56	resolution	resolution	NOUN
cana-2643	201	57	only	only	ADV
cana-2643	201	58	84.5	84.5	NUM
cana-2643	201	59	%	%	NOUN
cana-2643	201	60	91.1	91.1	NUM
cana-2643	201	61	%	%	NOUN
cana-2643	201	62	87.6	87.6	NUM
cana-2643	201	63	%	%	NOUN
cana-2643	201	64	55	55	NUM
cana-2643	201	65	ms	ms	NOUN
cana-2643	201	66	high	high	ADJ
cana-2643	201	67	resolution	resolution	NOUN
cana-2643	201	68	only	only	ADV
cana-2643	201	69	85.8	85.8	NUM
cana-2643	201	70	%	%	NOUN
cana-2643	201	71	92.4	92.4	NUM
cana-2643	201	72	%	%	NOUN
cana-2643	201	73	89.7	89.7	NUM
cana-2643	201	74	%	%	NOUN
cana-2643	201	75	65	65	NUM
cana-2643	201	76	ms	ms	PROPN
cana-2643	201	77	low	low	ADJ
cana-2643	201	78	+	+	CCONJ
cana-2643	201	79	medium	medium	ADJ
cana-2643	201	80	resolution	resolution	NOUN
cana-2643	201	81	86.1	86.1	NUM
cana-2643	201	82	%	%	NOUN
cana-2643	201	83	92.8	92.8	NUM
cana-2643	201	84	%	%	NOUN
cana-2643	201	85	90.1	90.1	NUM
cana-2643	201	86	%	%	NOUN
cana-2643	201	87	95	95	NUM
cana-2643	201	88	ms	ms	NOUN
cana-2643	201	89	low	low	ADJ
cana-2643	201	90	+	+	CCONJ
cana-2643	201	91	medium	medium	ADJ
cana-2643	201	92	+	+	CCONJ
cana-2643	201	93	high	high	ADJ
cana-2643	201	94	resolution	resolution	NOUN
cana-2643	201	95	88.2	88.2	NUM
cana-2643	201	96	%	%	NOUN
cana-2643	201	97	93.5	93.5	NUM
cana-2643	201	98	%	%	NOUN
cana-2643	201	99	91.4	91.4	NUM
cana-2643	201	100	%	%	NOUN
cana-2643	201	101	135	135	NUM
cana-2643	201	102	ms	ms	NOUN
cana-2643	201	103	with	with	ADP
cana-2643	201	104	attention	attention	NOUN
cana-2643	201	105	based	base	VERB
cana-2643	201	106	fusion	fusion	NOUN
cana-2643	201	107	89.0	89.0	NUM
cana-2643	201	108	%	%	NOUN
cana-2643	201	109	94.2	94.2	NUM
cana-2643	201	110	%	%	NOUN
cana-2643	201	111	92.1	92.1	NUM
cana-2643	201	112	%	%	NOUN
cana-2643	201	113	150	150	NUM
cana-2643	201	114	ms	ms	ADJ
cana-2643	201	115	side	side	NOUN
cana-2643	201	116	performance	performance	NOUN
cana-2643	201	117	comparison	comparison	NOUN
cana-2643	201	118	across	across	ADP
cana-2643	201	119	key	key	ADJ
cana-2643	201	120	metrics	metric	NOUN
cana-2643	201	121	.	.	PUNCT
cana-2643	202	1	mpga	mpga	NOUN
cana-2643	202	2	-	-	PUNCT
cana-2643	202	3	rl	rl	NOUN
cana-2643	202	4	-	-	PUNCT
cana-2643	202	5	net	net	NOUN
cana-2643	202	6	outperforms	outperform	NOUN
cana-2643	202	7	fcn-8s	fcn-8s	ADJ
cana-2643	202	8	in	in	ADP
cana-2643	202	9	segmentation	segmentation	NOUN
cana-2643	202	10	accuracy	accuracy	NOUN
cana-2643	202	11	,	,	PUNCT
cana-2643	202	12	represented	represent	VERB
cana-2643	202	13	by	by	ADP
cana-2643	202	14	the	the	DET
cana-2643	202	15	mean	mean	ADJ
cana-2643	202	16	iou	iou	NOUN
cana-2643	202	17	,	,	PUNCT
cana-2643	202	18	pixel	pixel	PROPN
cana-2643	202	19	accuracy	accuracy	NOUN
cana-2643	202	20	,	,	PUNCT
cana-2643	202	21	and	and	CCONJ
cana-2643	202	22	boundary	boundary	ADJ
cana-2643	202	23	f1	f1	PROPN
cana-2643	202	24	score	score	NOUN
cana-2643	202	25	.	.	PUNCT
cana-2643	203	1	these	these	DET
cana-2643	203	2	gains	gain	NOUN
cana-2643	203	3	indicate	indicate	VERB
cana-2643	203	4	mpga	mpga	NOUN
cana-2643	203	5	-	-	PUNCT
cana-2643	203	6	rl	rl	NOUN
cana-2643	203	7	-	-	PUNCT
cana-2643	203	8	net	net	NOUN
cana-2643	203	9	’s	’s	PART
cana-2643	203	10	ability	ability	NOUN
cana-2643	203	11	to	to	PART
cana-2643	203	12	capture	capture	VERB
cana-2643	203	13	detailed	detailed	ADJ
cana-2643	203	14	object	object	NOUN
cana-2643	203	15	boundaries	boundary	NOUN
cana-2643	203	16	and	and	CCONJ
cana-2643	203	17	spatial	spatial	ADJ
cana-2643	203	18	context	context	NOUN
cana-2643	203	19	,	,	PUNCT
cana-2643	203	20	essential	essential	ADJ
cana-2643	203	21	for	for	ADP
cana-2643	203	22	high	high	ADJ
cana-2643	203	23	-	-	PUNCT
cana-2643	203	24	stakes	stake	NOUN
cana-2643	203	25	fig	fig	NOUN
cana-2643	203	26	.	.	PUNCT
cana-2643	204	1	3	3	X
cana-2643	204	2	.	.	X
cana-2643	204	3	attention	attention	NOUN
cana-2643	204	4	-	-	PUNCT
cana-2643	204	5	based	base	VERB
cana-2643	204	6	fusion	fusion	NOUN
cana-2643	204	7	weight	weight	NOUN
cana-2643	204	8	analysis	analysis	NOUN
cana-2643	204	9	across	across	ADP
cana-2643	204	10	scene	scene	NOUN
cana-2643	204	11	types	type	NOUN
cana-2643	204	12	communications	communication	NOUN
cana-2643	204	13	on	on	ADP
cana-2643	204	14	applied	apply	VERB
cana-2643	204	15	nonlinear	nonlinear	ADJ
cana-2643	204	16	analysis	analysis	NOUN
cana-2643	204	17	issn	issn	NOUN
cana-2643	204	18	:	:	PUNCT
cana-2643	204	19	1074	1074	NUM
cana-2643	204	20	-	-	PUNCT
cana-2643	204	21	133x	133x	NUM
cana-2643	204	22	vol	vol	NOUN
cana-2643	204	23	32	32	NUM
cana-2643	204	24	no	no	NOUN
cana-2643	204	25	.	.	PUNCT
cana-2643	205	1	3s	3s	NUM
cana-2643	205	2	(	(	PUNCT
cana-2643	205	3	2025	2025	NUM
cana-2643	205	4	)	)	PUNCT
cana-2643	205	5	292	292	NUM
cana-2643	205	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	205	7	robotic	robotic	ADJ
cana-2643	205	8	vision	vision	NOUN
cana-2643	205	9	applications	application	NOUN
cana-2643	205	10	.	.	PUNCT
cana-2643	206	1	additionally	additionally	ADV
cana-2643	206	2	,	,	PUNCT
cana-2643	206	3	mpga	mpga	NOUN
cana-2643	206	4	-	-	PUNCT
cana-2643	206	5	rl	rl	NOUN
cana-2643	206	6	-	-	PUNCT
cana-2643	206	7	net	net	NOUN
cana-2643	206	8	demonstrates	demonstrate	VERB
cana-2643	206	9	computational	computational	ADJ
cana-2643	206	10	efficiency	efficiency	NOUN
cana-2643	206	11	with	with	ADP
cana-2643	206	12	reduced	reduce	VERB
cana-2643	206	13	inference	inference	NOUN
cana-2643	206	14	time	time	NOUN
cana-2643	206	15	and	and	CCONJ
cana-2643	206	16	memory	memory	NOUN
cana-2643	206	17	usage	usage	NOUN
cana-2643	206	18	,	,	PUNCT
cana-2643	206	19	making	make	VERB
cana-2643	206	20	it	it	PRON
cana-2643	206	21	more	more	ADV
cana-2643	206	22	suitable	suitable	ADJ
cana-2643	206	23	for	for	ADP
cana-2643	206	24	real	real	ADJ
cana-2643	206	25	-	-	PUNCT
cana-2643	206	26	time	time	NOUN
cana-2643	206	27	processing	processing	NOUN
cana-2643	206	28	needs	need	NOUN
cana-2643	206	29	.	.	PUNCT
cana-2643	207	1	this	this	DET
cana-2643	207	2	efficiency	efficiency	NOUN
cana-2643	207	3	aligns	align	VERB
cana-2643	207	4	with	with	ADP
cana-2643	207	5	findings	finding	NOUN
cana-2643	207	6	in	in	ADP
cana-2643	207	7	multi	multi	ADJ
cana-2643	207	8	-	-	ADJ
cana-2643	207	9	resolution	resolution	ADJ
cana-2643	207	10	feature	feature	NOUN
cana-2643	207	11	extraction	extraction	NOUN
cana-2643	207	12	,	,	PUNCT
cana-2643	207	13	where	where	SCONJ
cana-2643	207	14	models	model	NOUN
cana-2643	207	15	balance	balance	NOUN
cana-2643	207	16	accuracy	accuracy	NOUN
cana-2643	207	17	and	and	CCONJ
cana-2643	207	18	resource	resource	NOUN
cana-2643	207	19	management	management	NOUN
cana-2643	207	20	by	by	ADP
cana-2643	207	21	combining	combine	VERB
cana-2643	207	22	low	low	ADJ
cana-2643	207	23	and	and	CCONJ
cana-2643	207	24	high	high	ADJ
cana-2643	207	25	-	-	PUNCT
cana-2643	207	26	level	level	NOUN
cana-2643	207	27	details	detail	NOUN
cana-2643	207	28	.	.	PUNCT
cana-2643	208	1	figure	figure	NOUN
cana-2643	208	2	3	3	NUM
cana-2643	208	3	provides	provide	VERB
cana-2643	208	4	an	an	DET
cana-2643	208	5	analysis	analysis	NOUN
cana-2643	208	6	of	of	ADP
cana-2643	208	7	the	the	DET
cana-2643	208	8	adaptive	adaptive	ADJ
cana-2643	208	9	attention	attention	NOUN
cana-2643	208	10	-	-	PUNCT
cana-2643	208	11	based	base	VERB
cana-2643	208	12	fusion	fusion	NOUN
cana-2643	208	13	weights	weight	NOUN
cana-2643	208	14	,	,	PUNCT
cana-2643	208	15	which	which	PRON
cana-2643	208	16	adjust	adjust	VERB
cana-2643	208	17	according	accord	VERB
cana-2643	208	18	to	to	ADP
cana-2643	208	19	the	the	DET
cana-2643	208	20	complexity	complexity	NOUN
cana-2643	208	21	of	of	ADP
cana-2643	208	22	the	the	DET
cana-2643	208	23	scene	scene	NOUN
cana-2643	208	24	.	.	PUNCT
cana-2643	209	1	in	in	ADP
cana-2643	209	2	simple	simple	ADJ
cana-2643	209	3	scenes	scene	NOUN
cana-2643	209	4	,	,	PUNCT
cana-2643	209	5	mpgarl	mpgarl	NOUN
cana-2643	209	6	-	-	PUNCT
cana-2643	209	7	net	net	NOUN
cana-2643	209	8	relies	rely	VERB
cana-2643	209	9	more	more	ADJ
cana-2643	209	10	on	on	ADP
cana-2643	209	11	the	the	DET
cana-2643	209	12	low	low	ADJ
cana-2643	209	13	-	-	PUNCT
cana-2643	209	14	resolution	resolution	NOUN
cana-2643	209	15	path	path	NOUN
cana-2643	209	16	(	(	PUNCT
cana-2643	209	17	weight	weight	NOUN
cana-2643	209	18	of	of	ADP
cana-2643	209	19	0.5	0.5	NUM
cana-2643	209	20	)	)	PUNCT
cana-2643	209	21	for	for	ADP
cana-2643	209	22	basic	basic	ADJ
cana-2643	209	23	spatial	spatial	ADJ
cana-2643	209	24	information	information	NOUN
cana-2643	209	25	,	,	PUNCT
cana-2643	209	26	minimizing	minimize	VERB
cana-2643	209	27	computational	computational	ADJ
cana-2643	209	28	demands	demand	NOUN
cana-2643	209	29	.	.	PUNCT
cana-2643	210	1	in	in	ADP
cana-2643	210	2	moderate	moderate	ADJ
cana-2643	210	3	scenes	scene	NOUN
cana-2643	210	4	,	,	PUNCT
cana-2643	210	5	the	the	DET
cana-2643	210	6	mediumresolution	mediumresolution	NOUN
cana-2643	210	7	path	path	NOUN
cana-2643	210	8	’s	’s	PART
cana-2643	210	9	weight	weight	NOUN
cana-2643	210	10	increases	increase	NOUN
cana-2643	210	11	(	(	PUNCT
cana-2643	210	12	0.4	0.4	NUM
cana-2643	210	13	)	)	PUNCT
cana-2643	210	14	,	,	PUNCT
cana-2643	210	15	offering	offer	VERB
cana-2643	210	16	additional	additional	ADJ
cana-2643	210	17	boundary	boundary	ADJ
cana-2643	210	18	information	information	NOUN
cana-2643	210	19	to	to	PART
cana-2643	210	20	enhance	enhance	VERB
cana-2643	210	21	segmentation	segmentation	NOUN
cana-2643	210	22	quality	quality	NOUN
cana-2643	210	23	.	.	PUNCT
cana-2643	211	1	for	for	ADP
cana-2643	211	2	complex	complex	ADJ
cana-2643	211	3	scenes	scene	NOUN
cana-2643	211	4	,	,	PUNCT
cana-2643	211	5	the	the	DET
cana-2643	211	6	high	high	ADJ
cana-2643	211	7	-	-	PUNCT
cana-2643	211	8	resolution	resolution	NOUN
cana-2643	211	9	path	path	NOUN
cana-2643	211	10	receives	receive	VERB
cana-2643	211	11	the	the	DET
cana-2643	211	12	highest	high	ADJ
cana-2643	211	13	weight	weight	NOUN
cana-2643	211	14	(	(	PUNCT
cana-2643	211	15	0.5	0.5	NUM
cana-2643	211	16	)	)	PUNCT
cana-2643	211	17	,	,	PUNCT
cana-2643	211	18	allowing	allow	VERB
cana-2643	211	19	the	the	DET
cana-2643	211	20	model	model	NOUN
cana-2643	211	21	to	to	PART
cana-2643	211	22	focus	focus	VERB
cana-2643	211	23	on	on	ADP
cana-2643	211	24	fine	fine	ADV
cana-2643	211	25	-	-	PUNCT
cana-2643	211	26	grained	grain	VERB
cana-2643	211	27	details	detail	NOUN
cana-2643	211	28	such	such	ADJ
cana-2643	211	29	as	as	ADP
cana-2643	211	30	textures	texture	NOUN
cana-2643	211	31	and	and	CCONJ
cana-2643	211	32	edges	edge	NOUN
cana-2643	211	33	.	.	PUNCT
cana-2643	212	1	this	this	DET
cana-2643	212	2	adaptive	adaptive	ADJ
cana-2643	212	3	weighting	weighting	NOUN
cana-2643	212	4	mechanism	mechanism	NOUN
cana-2643	212	5	aligns	align	VERB
cana-2643	212	6	with	with	ADP
cana-2643	212	7	recent	recent	ADJ
cana-2643	212	8	advancements	advancement	NOUN
cana-2643	212	9	in	in	ADP
cana-2643	212	10	scene	scene	NOUN
cana-2643	212	11	-	-	PUNCT
cana-2643	212	12	adaptive	adaptive	ADJ
cana-2643	212	13	attention	attention	NOUN
cana-2643	212	14	,	,	PUNCT
cana-2643	212	15	ensuring	ensure	VERB
cana-2643	212	16	that	that	SCONJ
cana-2643	212	17	mpga	mpga	NOUN
cana-2643	212	18	-	-	PUNCT
cana-2643	212	19	rl	rl	NOUN
cana-2643	212	20	-	-	PUNCT
cana-2643	212	21	net	net	NOUN
cana-2643	212	22	maintains	maintain	VERB
cana-2643	212	23	a	a	DET
cana-2643	212	24	balance	balance	NOUN
cana-2643	212	25	between	between	ADP
cana-2643	212	26	detail	detail	NOUN
cana-2643	212	27	and	and	CCONJ
cana-2643	212	28	efficiency	efficiency	NOUN
cana-2643	212	29	by	by	ADP
cana-2643	212	30	tailoring	tailor	VERB
cana-2643	212	31	feature	feature	NOUN
cana-2643	212	32	extraction	extraction	NOUN
cana-2643	212	33	based	base	VERB
cana-2643	212	34	on	on	ADP
cana-2643	212	35	scene	scene	NOUN
cana-2643	212	36	demands	demand	NOUN
cana-2643	212	37	.	.	PUNCT
cana-2643	213	1	together	together	ADV
cana-2643	213	2	,	,	PUNCT
cana-2643	213	3	figures	figure	NOUN
cana-2643	213	4	3	3	NUM
cana-2643	213	5	and	and	CCONJ
cana-2643	213	6	4	4	NUM
cana-2643	213	7	underscore	underscore	VERB
cana-2643	213	8	the	the	DET
cana-2643	213	9	adaptability	adaptability	NOUN
cana-2643	213	10	and	and	CCONJ
cana-2643	213	11	precision	precision	NOUN
cana-2643	213	12	of	of	ADP
cana-2643	213	13	mpga	mpga	NOUN
cana-2643	213	14	-	-	PUNCT
cana-2643	213	15	rl	rl	NOUN
cana-2643	213	16	-	-	PUNCT
cana-2643	213	17	net	net	NOUN
cana-2643	213	18	,	,	PUNCT
cana-2643	213	19	making	make	VERB
cana-2643	213	20	it	it	PRON
cana-2643	213	21	a	a	DET
cana-2643	213	22	robust	robust	ADJ
cana-2643	213	23	choice	choice	NOUN
cana-2643	213	24	for	for	ADP
cana-2643	213	25	dynamic	dynamic	ADJ
cana-2643	213	26	,	,	PUNCT
cana-2643	213	27	real	real	ADJ
cana-2643	213	28	-	-	PUNCT
cana-2643	213	29	time	time	NOUN
cana-2643	213	30	robotic	robotic	ADJ
cana-2643	213	31	vision	vision	NOUN
cana-2643	213	32	tasks	task	NOUN
cana-2643	213	33	across	across	ADP
cana-2643	213	34	a	a	DET
cana-2643	213	35	range	range	NOUN
cana-2643	213	36	of	of	ADP
cana-2643	213	37	environmental	environmental	ADJ
cana-2643	213	38	complexities	complexity	NOUN
cana-2643	213	39	fig	fig	NOUN
cana-2643	213	40	.	.	PUNCT
cana-2643	214	1	4	4	X
cana-2643	214	2	.	.	NOUN
cana-2643	214	3	comparative	comparative	ADJ
cana-2643	214	4	performance	performance	NOUN
cana-2643	214	5	analysis	analysis	NOUN
cana-2643	214	6	of	of	ADP
cana-2643	214	7	mpga	mpga	NOUN
cana-2643	214	8	-	-	PUNCT
cana-2643	214	9	rl	rl	NOUN
cana-2643	214	10	-	-	PUNCT
cana-2643	214	11	net	net	NOUN
cana-2643	214	12	vs	vs	ADP
cana-2643	214	13	fcn-8s	fcn-8s	ADJ
cana-2643	214	14	5	5	NUM
cana-2643	214	15	.	.	PUNCT
cana-2643	214	16	discussion	discussion	NOUN
cana-2643	214	17	and	and	CCONJ
cana-2643	214	18	conclusion	conclusion	NOUN
cana-2643	214	19	the	the	DET
cana-2643	214	20	mpga	mpga	NOUN
cana-2643	214	21	-	-	PUNCT
cana-2643	214	22	rl	rl	NOUN
cana-2643	214	23	-	-	PUNCT
cana-2643	214	24	net	net	NOUN
cana-2643	214	25	framework	framework	NOUN
cana-2643	214	26	,	,	PUNCT
cana-2643	214	27	with	with	ADP
cana-2643	214	28	its	its	PRON
cana-2643	214	29	multi	multi	ADJ
cana-2643	214	30	-	-	ADJ
cana-2643	214	31	path	path	ADJ
cana-2643	214	32	feature	feature	NOUN
cana-2643	214	33	extraction	extraction	NOUN
cana-2643	214	34	(	(	PUNCT
cana-2643	214	35	mpfe	mpfe	NOUN
cana-2643	214	36	)	)	PUNCT
cana-2643	214	37	module	module	NOUN
cana-2643	214	38	and	and	CCONJ
cana-2643	214	39	attention	attention	NOUN
cana-2643	214	40	-	-	PUNCT
cana-2643	214	41	based	base	VERB
cana-2643	214	42	fusion	fusion	NOUN
cana-2643	214	43	,	,	PUNCT
cana-2643	214	44	offers	offer	VERB
cana-2643	214	45	a	a	DET
cana-2643	214	46	comprehensive	comprehensive	ADJ
cana-2643	214	47	solution	solution	NOUN
cana-2643	214	48	for	for	ADP
cana-2643	214	49	adaptive	adaptive	ADJ
cana-2643	214	50	robotic	robotic	ADJ
cana-2643	214	51	vision	vision	NOUN
cana-2643	214	52	in	in	ADP
cana-2643	214	53	complex	complex	ADJ
cana-2643	214	54	and	and	CCONJ
cana-2643	214	55	dynamic	dynamic	ADJ
cana-2643	214	56	environments	environment	NOUN
cana-2643	214	57	.	.	PUNCT
cana-2643	215	1	the	the	DET
cana-2643	215	2	comparative	comparative	ADJ
cana-2643	215	3	results	result	NOUN
cana-2643	215	4	presented	present	VERB
cana-2643	215	5	in	in	ADP
cana-2643	215	6	figure	figure	NOUN
cana-2643	215	7	4	4	NUM
cana-2643	215	8	highlight	highlight	VERB
cana-2643	215	9	mpga	mpga	NOUN
cana-2643	215	10	-	-	PUNCT
cana-2643	215	11	rl	rl	NOUN
cana-2643	215	12	-	-	PUNCT
cana-2643	215	13	net	net	NOUN
cana-2643	215	14	’s	’s	PART
cana-2643	215	15	improvements	improvement	NOUN
cana-2643	215	16	in	in	ADP
cana-2643	215	17	mean	mean	ADJ
cana-2643	215	18	iou	iou	PROPN
cana-2643	215	19	,	,	PUNCT
cana-2643	215	20	pixel	pixel	PROPN
cana-2643	215	21	accuracy	accuracy	NOUN
cana-2643	215	22	,	,	PUNCT
cana-2643	215	23	and	and	CCONJ
cana-2643	215	24	boundary	boundary	ADJ
cana-2643	215	25	f1	f1	NOUN
cana-2643	215	26	score	score	NOUN
cana-2643	215	27	over	over	ADP
cana-2643	215	28	baseline	baseline	NOUN
cana-2643	215	29	models	model	NOUN
cana-2643	215	30	,	,	PUNCT
cana-2643	215	31	particularly	particularly	ADV
cana-2643	215	32	in	in	ADP
cana-2643	215	33	terms	term	NOUN
cana-2643	215	34	of	of	ADP
cana-2643	215	35	accurate	accurate	ADJ
cana-2643	215	36	boundary	boundary	ADJ
cana-2643	215	37	segmentation	segmentation	NOUN
cana-2643	215	38	and	and	CCONJ
cana-2643	215	39	reduced	reduce	VERB
cana-2643	215	40	computational	computational	ADJ
cana-2643	215	41	load	load	NOUN
cana-2643	215	42	.	.	PUNCT
cana-2643	216	1	the	the	DET
cana-2643	216	2	framework	framework	NOUN
cana-2643	216	3	’s	’s	PART
cana-2643	216	4	architecture	architecture	NOUN
cana-2643	216	5	leverages	leverage	VERB
cana-2643	216	6	a	a	DET
cana-2643	216	7	multi	multi	ADJ
cana-2643	216	8	-	-	ADJ
cana-2643	216	9	resolution	resolution	ADJ
cana-2643	216	10	approach	approach	NOUN
cana-2643	216	11	to	to	PART
cana-2643	216	12	enhance	enhance	VERB
cana-2643	216	13	detail	detail	NOUN
cana-2643	216	14	capture	capture	NOUN
cana-2643	216	15	across	across	ADP
cana-2643	216	16	low	low	ADJ
cana-2643	216	17	,	,	PUNCT
cana-2643	216	18	medium	medium	ADJ
cana-2643	216	19	,	,	PUNCT
cana-2643	216	20	and	and	CCONJ
cana-2643	216	21	high	high	ADJ
cana-2643	216	22	resolutions	resolution	NOUN
cana-2643	216	23	,	,	PUNCT
cana-2643	216	24	each	each	PRON
cana-2643	216	25	contributing	contribute	VERB
cana-2643	216	26	uniquely	uniquely	ADV
cana-2643	216	27	to	to	ADP
cana-2643	216	28	the	the	DET
cana-2643	216	29	final	final	ADJ
cana-2643	216	30	segmentation	segmentation	NOUN
cana-2643	216	31	output	output	NOUN
cana-2643	216	32	.	.	PUNCT
cana-2643	217	1	this	this	DET
cana-2643	217	2	setup	setup	NOUN
cana-2643	217	3	enables	enable	VERB
cana-2643	217	4	mpga	mpga	NOUN
cana-2643	217	5	-	-	PUNCT
cana-2643	217	6	rl	rl	NOUN
cana-2643	217	7	-	-	PUNCT
cana-2643	217	8	net	net	NOUN
cana-2643	217	9	to	to	PART
cana-2643	217	10	maintain	maintain	VERB
cana-2643	217	11	efficiency	efficiency	NOUN
cana-2643	217	12	while	while	SCONJ
cana-2643	217	13	ensuring	ensure	VERB
cana-2643	217	14	high	high	ADJ
cana-2643	217	15	precision	precision	NOUN
cana-2643	217	16	,	,	PUNCT
cana-2643	217	17	especially	especially	ADV
cana-2643	217	18	in	in	ADP
cana-2643	217	19	scenes	scene	NOUN
cana-2643	217	20	with	with	ADP
cana-2643	217	21	varying	vary	VERB
cana-2643	217	22	complexities	complexity	NOUN
cana-2643	217	23	.	.	PUNCT
cana-2643	218	1	figure	figure	VERB
cana-2643	218	2	4	4	NUM
cana-2643	218	3	showcases	showcase	VERB
cana-2643	218	4	the	the	DET
cana-2643	218	5	adaptive	adaptive	ADJ
cana-2643	218	6	nature	nature	NOUN
cana-2643	218	7	of	of	ADP
cana-2643	218	8	the	the	DET
cana-2643	218	9	attention	attention	NOUN
cana-2643	218	10	-	-	PUNCT
cana-2643	218	11	based	base	VERB
cana-2643	218	12	fusion	fusion	NOUN
cana-2643	218	13	mechanism	mechanism	NOUN
cana-2643	218	14	,	,	PUNCT
cana-2643	218	15	which	which	PRON
cana-2643	218	16	dynamically	dynamically	ADV
cana-2643	218	17	assigns	assign	VERB
cana-2643	218	18	weight	weight	NOUN
cana-2643	218	19	to	to	ADP
cana-2643	218	20	each	each	DET
cana-2643	218	21	resolution	resolution	NOUN
cana-2643	218	22	path	path	NOUN
cana-2643	218	23	based	base	VERB
cana-2643	218	24	on	on	ADP
cana-2643	218	25	the	the	DET
cana-2643	218	26	scene	scene	NOUN
cana-2643	218	27	’s	’s	PART
cana-2643	218	28	communications	communication	NOUN
cana-2643	218	29	on	on	ADP
cana-2643	218	30	applied	apply	VERB
cana-2643	218	31	nonlinear	nonlinear	ADJ
cana-2643	218	32	analysis	analysis	NOUN
cana-2643	218	33	issn	issn	NOUN
cana-2643	218	34	:	:	PUNCT
cana-2643	218	35	1074	1074	NUM
cana-2643	218	36	-	-	PUNCT
cana-2643	218	37	133x	133x	NUM
cana-2643	218	38	vol	vol	NOUN
cana-2643	218	39	32	32	NUM
cana-2643	218	40	no	no	NOUN
cana-2643	218	41	.	.	PUNCT
cana-2643	219	1	3s	3s	NUM
cana-2643	219	2	(	(	PUNCT
cana-2643	219	3	2025	2025	NUM
cana-2643	219	4	)	)	PUNCT
cana-2643	219	5	293	293	NUM
cana-2643	219	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	219	7	specific	specific	ADJ
cana-2643	219	8	requirements	requirement	NOUN
cana-2643	219	9	.	.	PUNCT
cana-2643	220	1	for	for	ADP
cana-2643	220	2	simple	simple	ADJ
cana-2643	220	3	scenes	scene	NOUN
cana-2643	220	4	,	,	PUNCT
cana-2643	220	5	the	the	DET
cana-2643	220	6	low	low	ADJ
cana-2643	220	7	-	-	PUNCT
cana-2643	220	8	resolution	resolution	NOUN
cana-2643	220	9	path	path	NOUN
cana-2643	220	10	is	be	AUX
cana-2643	220	11	prioritized	prioritize	VERB
cana-2643	220	12	,	,	PUNCT
cana-2643	220	13	ensuring	ensure	VERB
cana-2643	220	14	computational	computational	ADJ
cana-2643	220	15	efficiency	efficiency	NOUN
cana-2643	220	16	.	.	PUNCT
cana-2643	221	1	in	in	ADP
cana-2643	221	2	contrast	contrast	NOUN
cana-2643	221	3	,	,	PUNCT
cana-2643	221	4	complex	complex	ADJ
cana-2643	221	5	scenes	scene	NOUN
cana-2643	221	6	trigger	trigger	VERB
cana-2643	221	7	a	a	DET
cana-2643	221	8	higher	high	ADJ
cana-2643	221	9	weight	weight	NOUN
cana-2643	221	10	allocation	allocation	NOUN
cana-2643	221	11	for	for	ADP
cana-2643	221	12	the	the	DET
cana-2643	221	13	highresolution	highresolution	NOUN
cana-2643	221	14	path	path	NOUN
cana-2643	221	15	,	,	PUNCT
cana-2643	221	16	enabling	enable	VERB
cana-2643	221	17	fine	fine	ADV
cana-2643	221	18	-	-	PUNCT
cana-2643	221	19	grained	grain	VERB
cana-2643	221	20	feature	feature	NOUN
cana-2643	221	21	capture	capture	NOUN
cana-2643	221	22	essential	essential	ADJ
cana-2643	221	23	for	for	ADP
cana-2643	221	24	tasks	task	NOUN
cana-2643	221	25	like	like	ADP
cana-2643	221	26	object	object	NOUN
cana-2643	221	27	recognition	recognition	NOUN
cana-2643	221	28	in	in	ADP
cana-2643	221	29	cluttered	cluttered	ADJ
cana-2643	221	30	environments	environment	NOUN
cana-2643	221	31	.	.	PUNCT
cana-2643	222	1	this	this	DET
cana-2643	222	2	adaptive	adaptive	ADJ
cana-2643	222	3	fusion	fusion	NOUN
cana-2643	222	4	aligns	align	VERB
cana-2643	222	5	with	with	ADP
cana-2643	222	6	recent	recent	ADJ
cana-2643	222	7	advancements	advancement	NOUN
cana-2643	222	8	in	in	ADP
cana-2643	222	9	scene	scene	NOUN
cana-2643	222	10	-	-	PUNCT
cana-2643	222	11	adaptive	adaptive	ADJ
cana-2643	222	12	attention	attention	NOUN
cana-2643	222	13	,	,	PUNCT
cana-2643	222	14	reinforcing	reinforce	VERB
cana-2643	222	15	mpga	mpga	NOUN
cana-2643	222	16	-	-	PUNCT
cana-2643	222	17	rl	rl	NOUN
cana-2643	222	18	-	-	PUNCT
cana-2643	222	19	net	net	NOUN
cana-2643	222	20	’s	’s	PART
cana-2643	222	21	suitability	suitability	NOUN
cana-2643	222	22	for	for	ADP
cana-2643	222	23	applications	application	NOUN
cana-2643	222	24	that	that	PRON
cana-2643	222	25	require	require	VERB
cana-2643	222	26	a	a	DET
cana-2643	222	27	flexible	flexible	ADJ
cana-2643	222	28	,	,	PUNCT
cana-2643	222	29	contextaware	contextaware	NOUN
cana-2643	222	30	approach	approach	NOUN
cana-2643	222	31	.	.	PUNCT
cana-2643	223	1	in	in	ADP
cana-2643	223	2	conclusion	conclusion	NOUN
cana-2643	223	3	,	,	PUNCT
cana-2643	223	4	mpga	mpga	NOUN
cana-2643	223	5	-	-	PUNCT
cana-2643	223	6	rl	rl	NOUN
cana-2643	223	7	-	-	PUNCT
cana-2643	223	8	net	net	NOUN
cana-2643	223	9	significantly	significantly	ADV
cana-2643	223	10	advances	advance	VERB
cana-2643	223	11	the	the	DET
cana-2643	223	12	field	field	NOUN
cana-2643	223	13	of	of	ADP
cana-2643	223	14	robotic	robotic	ADJ
cana-2643	223	15	vision	vision	NOUN
cana-2643	223	16	by	by	ADP
cana-2643	223	17	balancing	balance	VERB
cana-2643	223	18	multiresolution	multiresolution	NOUN
cana-2643	223	19	feature	feature	NOUN
cana-2643	223	20	extraction	extraction	NOUN
cana-2643	223	21	with	with	ADP
cana-2643	223	22	adaptive	adaptive	ADJ
cana-2643	223	23	attention	attention	NOUN
cana-2643	223	24	mechanisms	mechanism	NOUN
cana-2643	223	25	.	.	PUNCT
cana-2643	224	1	this	this	DET
cana-2643	224	2	balance	balance	NOUN
cana-2643	224	3	facilitates	facilitate	VERB
cana-2643	224	4	accurate	accurate	ADJ
cana-2643	224	5	,	,	PUNCT
cana-2643	224	6	real	real	ADJ
cana-2643	224	7	-	-	PUNCT
cana-2643	224	8	time	time	NOUN
cana-2643	224	9	scene	scene	NOUN
cana-2643	224	10	interpretation	interpretation	NOUN
cana-2643	224	11	in	in	ADP
cana-2643	224	12	diverse	diverse	ADJ
cana-2643	224	13	settings	setting	NOUN
cana-2643	224	14	,	,	PUNCT
cana-2643	224	15	from	from	ADP
cana-2643	224	16	simple	simple	ADJ
cana-2643	224	17	layouts	layout	NOUN
cana-2643	224	18	to	to	ADP
cana-2643	224	19	cluttered	cluttered	ADJ
cana-2643	224	20	,	,	PUNCT
cana-2643	224	21	dynamic	dynamic	ADJ
cana-2643	224	22	scenes	scene	NOUN
cana-2643	224	23	.	.	PUNCT
cana-2643	225	1	future	future	ADJ
cana-2643	225	2	research	research	NOUN
cana-2643	225	3	could	could	AUX
cana-2643	225	4	explore	explore	VERB
cana-2643	225	5	extending	extend	VERB
cana-2643	225	6	mpga	mpga	NOUN
cana-2643	225	7	-	-	PUNCT
cana-2643	225	8	rl	rl	NOUN
cana-2643	225	9	-	-	PUNCT
cana-2643	225	10	net	net	NOUN
cana-2643	225	11	to	to	ADP
cana-2643	225	12	other	other	ADJ
cana-2643	225	13	domains	domain	NOUN
cana-2643	225	14	,	,	PUNCT
cana-2643	225	15	such	such	ADJ
cana-2643	225	16	as	as	ADP
cana-2643	225	17	autonomous	autonomous	ADJ
cana-2643	225	18	driving	driving	NOUN
cana-2643	225	19	or	or	CCONJ
cana-2643	225	20	healthcare	healthcare	NOUN
cana-2643	225	21	,	,	PUNCT
cana-2643	225	22	where	where	SCONJ
cana-2643	225	23	real	real	ADJ
cana-2643	225	24	-	-	PUNCT
cana-2643	225	25	time	time	NOUN
cana-2643	225	26	,	,	PUNCT
cana-2643	225	27	adaptive	adaptive	ADJ
cana-2643	225	28	region	region	NOUN
cana-2643	225	29	prioritization	prioritization	NOUN
cana-2643	225	30	is	be	AUX
cana-2643	225	31	crucial	crucial	ADJ
cana-2643	225	32	.	.	PUNCT
cana-2643	226	1	additionally	additionally	ADV
cana-2643	226	2	,	,	PUNCT
cana-2643	226	3	integrating	integrate	VERB
cana-2643	226	4	domain	domain	NOUN
cana-2643	226	5	-	-	PUNCT
cana-2643	226	6	specific	specific	ADJ
cana-2643	226	7	reinforcement	reinforcement	NOUN
cana-2643	226	8	learning	learning	NOUN
cana-2643	226	9	could	could	AUX
cana-2643	226	10	further	far	ADV
cana-2643	226	11	enhance	enhance	VERB
cana-2643	226	12	the	the	DET
cana-2643	226	13	framework	framework	NOUN
cana-2643	226	14	’s	’s	PART
cana-2643	226	15	adaptability	adaptability	NOUN
cana-2643	226	16	,	,	PUNCT
cana-2643	226	17	tailoring	tailor	VERB
cana-2643	226	18	attention	attention	NOUN
cana-2643	226	19	weights	weight	NOUN
cana-2643	226	20	to	to	ADP
cana-2643	226	21	specific	specific	ADJ
cana-2643	226	22	tasks	task	NOUN
cana-2643	226	23	or	or	CCONJ
cana-2643	226	24	environments	environment	NOUN
cana-2643	226	25	references	reference	NOUN
cana-2643	226	26	[	[	X
cana-2643	226	27	1	1	NUM
cana-2643	226	28	]	]	X
cana-2643	226	29	wan	wan	PROPN
cana-2643	226	30	,	,	PUNCT
cana-2643	226	31	s.	s.	PROPN
cana-2643	226	32	,	,	PUNCT
cana-2643	226	33	&	&	CCONJ
cana-2643	226	34	goudos	goudo	NOUN
cana-2643	226	35	,	,	PUNCT
cana-2643	226	36	s.	s.	PROPN
cana-2643	226	37	(	(	PUNCT
cana-2643	226	38	2020	2020	NUM
cana-2643	226	39	)	)	PUNCT
cana-2643	226	40	.	.	PUNCT
cana-2643	227	1	faster	fast	ADV
cana-2643	227	2	r	r	X
cana-2643	227	3	-	-	PUNCT
cana-2643	227	4	cnn	cnn	PROPN
cana-2643	227	5	for	for	ADP
cana-2643	227	6	multi	multi	ADJ
cana-2643	227	7	-	-	ADJ
cana-2643	227	8	class	class	ADJ
cana-2643	227	9	fruit	fruit	NOUN
cana-2643	227	10	detection	detection	NOUN
cana-2643	227	11	using	use	VERB
cana-2643	227	12	a	a	DET
cana-2643	227	13	robotic	robotic	ADJ
cana-2643	227	14	vision	vision	NOUN
cana-2643	227	15	system	system	NOUN
cana-2643	227	16	.	.	PUNCT
cana-2643	228	1	in	in	ADP
cana-2643	228	2	computer	computer	NOUN
cana-2643	228	3	networks	network	NOUN
cana-2643	228	4	(	(	PUNCT
cana-2643	228	5	vol	vol	NOUN
cana-2643	228	6	.	.	NOUN
cana-2643	228	7	168	168	NUM
cana-2643	228	8	,	,	PUNCT
cana-2643	228	9	p.	p.	NOUN
cana-2643	228	10	107036	107036	NUM
cana-2643	228	11	)	)	PUNCT
cana-2643	228	12	.	.	PUNCT
cana-2643	229	1	elsevier	elsevier	PROPN
cana-2643	229	2	bv	bv	PROPN
cana-2643	229	3	.	.	PROPN
cana-2643	229	4	https://doi.org/10.1016/j.comnet.2019.107036	https://doi.org/10.1016/j.comnet.2019.107036	PROPN
cana-2643	229	5	[	[	X
cana-2643	229	6	2	2	NUM
cana-2643	229	7	]	]	PUNCT
cana-2643	229	8	m.	m.	NOUN
cana-2643	229	9	halstead	halstead	PROPN
cana-2643	229	10	,	,	PUNCT
cana-2643	229	11	c.	c.	PROPN
cana-2643	229	12	mccool	mccool	PROPN
cana-2643	229	13	,	,	PUNCT
cana-2643	229	14	s.	s.	PROPN
cana-2643	229	15	denman	denman	PROPN
cana-2643	229	16	,	,	PUNCT
cana-2643	229	17	t.	t.	PROPN
cana-2643	229	18	perez	perez	PROPN
cana-2643	229	19	and	and	CCONJ
cana-2643	229	20	c.	c.	PROPN
cana-2643	229	21	fookes	fooke	NOUN
cana-2643	229	22	,	,	PUNCT
cana-2643	229	23	"	"	PUNCT
cana-2643	229	24	fruit	fruit	NOUN
cana-2643	229	25	quantity	quantity	NOUN
cana-2643	229	26	and	and	CCONJ
cana-2643	229	27	ripeness	ripeness	NOUN
cana-2643	229	28	estimation	estimation	NOUN
cana-2643	229	29	using	use	VERB
cana-2643	229	30	a	a	DET
cana-2643	229	31	robotic	robotic	ADJ
cana-2643	229	32	vision	vision	NOUN
cana-2643	229	33	system	system	NOUN
cana-2643	229	34	,	,	PUNCT
cana-2643	229	35	"	"	PUNCT
cana-2643	229	36	in	in	ADP
cana-2643	229	37	ieee	ieee	NOUN
cana-2643	229	38	robotics	robotic	NOUN
cana-2643	229	39	and	and	CCONJ
cana-2643	229	40	automation	automation	NOUN
cana-2643	229	41	letters	letter	NOUN
cana-2643	229	42	,	,	PUNCT
cana-2643	229	43	vol	vol	NOUN
cana-2643	229	44	.	.	PROPN
cana-2643	230	1	3	3	NUM
cana-2643	230	2	,	,	PUNCT
cana-2643	230	3	no	no	INTJ
cana-2643	230	4	.	.	NOUN
cana-2643	230	5	4	4	NUM
cana-2643	230	6	,	,	PUNCT
cana-2643	230	7	pp	pp	ADJ
cana-2643	230	8	.	.	PUNCT
cana-2643	231	1	2995	2995	NUM
cana-2643	231	2	-	-	SYM
cana-2643	231	3	3002	3002	NUM
cana-2643	231	4	,	,	PUNCT
cana-2643	231	5	oct	oct	PROPN
cana-2643	231	6	.	.	PROPN
cana-2643	231	7	2018	2018	NUM
cana-2643	231	8	,	,	PUNCT
cana-2643	231	9	doi	doi	NOUN
cana-2643	231	10	:	:	PUNCT
cana-2643	231	11	10.1109	10.1109	NUM
cana-2643	231	12	/	/	SYM
cana-2643	231	13	lra.2018.2849514	lra.2018.2849514	NOUN
cana-2643	231	14	.	.	PUNCT
cana-2643	232	1	[	[	X
cana-2643	232	2	3	3	NUM
cana-2643	232	3	]	]	X
cana-2643	232	4	beysolow	beysolow	PROPN
cana-2643	232	5	ii	ii	PROPN
cana-2643	232	6	,	,	PUNCT
cana-2643	232	7	t.	t.	PROPN
cana-2643	232	8	(	(	PUNCT
cana-2643	232	9	2017	2017	NUM
cana-2643	232	10	)	)	PUNCT
cana-2643	232	11	.	.	PUNCT
cana-2643	233	1	convolutional	convolutional	ADJ
cana-2643	233	2	neural	neural	ADJ
cana-2643	233	3	networks	network	NOUN
cana-2643	233	4	(	(	PUNCT
cana-2643	233	5	cnns	cnns	PROPN
cana-2643	233	6	)	)	PUNCT
cana-2643	233	7	.	.	PUNCT
cana-2643	234	1	in	in	ADP
cana-2643	234	2	introduction	introduction	NOUN
cana-2643	234	3	to	to	ADP
cana-2643	234	4	deep	deep	ADJ
cana-2643	234	5	learning	learning	NOUN
cana-2643	234	6	using	use	VERB
cana-2643	234	7	r	r	NOUN
cana-2643	234	8	(	(	PUNCT
cana-2643	234	9	pp	pp	ADJ
cana-2643	234	10	.	.	PUNCT
cana-2643	235	1	101–112	101–112	NUM
cana-2643	235	2	)	)	PUNCT
cana-2643	235	3	.	.	PUNCT
cana-2643	236	1	apress	apress	PROPN
cana-2643	236	2	.	.	PUNCT
cana-2643	237	1	https://doi.org/10.1007/978-14842-2734-3_5	https://doi.org/10.1007/978-14842-2734-3_5	PROPN
cana-2643	238	1	[	[	X
cana-2643	238	2	4	4	NUM
cana-2643	238	3	]	]	X
cana-2643	238	4	v.	v.	CCONJ
cana-2643	238	5	badrinarayanan	badrinarayanan	PROPN
cana-2643	238	6	,	,	PUNCT
cana-2643	238	7	a.	a.	PROPN
cana-2643	238	8	kendall	kendall	PROPN
cana-2643	238	9	and	and	CCONJ
cana-2643	238	10	r.	r.	PROPN
cana-2643	238	11	cipolla	cipolla	PROPN
cana-2643	238	12	,	,	PUNCT
cana-2643	238	13	"	"	PUNCT
cana-2643	238	14	segnet	segnet	NOUN
cana-2643	238	15	:	:	PUNCT
cana-2643	238	16	a	a	DET
cana-2643	238	17	deep	deep	ADJ
cana-2643	238	18	convolutional	convolutional	ADJ
cana-2643	238	19	encoder	encoder	NOUN
cana-2643	238	20	-	-	PUNCT
cana-2643	238	21	decoder	decoder	NOUN
cana-2643	238	22	architecture	architecture	NOUN
cana-2643	238	23	for	for	ADP
cana-2643	238	24	image	image	NOUN
cana-2643	238	25	segmentation	segmentation	NOUN
cana-2643	238	26	,	,	PUNCT
cana-2643	238	27	"	"	PUNCT
cana-2643	238	28	in	in	ADP
cana-2643	238	29	ieee	ieee	NOUN
cana-2643	238	30	transactions	transaction	NOUN
cana-2643	238	31	on	on	ADP
cana-2643	238	32	pattern	pattern	NOUN
cana-2643	238	33	analysis	analysis	NOUN
cana-2643	238	34	and	and	CCONJ
cana-2643	238	35	machine	machine	NOUN
cana-2643	238	36	intelligence	intelligence	NOUN
cana-2643	238	37	,	,	PUNCT
cana-2643	238	38	vol	vol	NOUN
cana-2643	238	39	.	.	PROPN
cana-2643	238	40	39	39	NUM
cana-2643	238	41	,	,	PUNCT
cana-2643	238	42	no	no	INTJ
cana-2643	238	43	.	.	NOUN
cana-2643	238	44	12	12	NUM
cana-2643	238	45	,	,	PUNCT
cana-2643	238	46	pp	pp	ADJ
cana-2643	238	47	.	.	PUNCT
cana-2643	239	1	2481	2481	NUM
cana-2643	239	2	-	-	SYM
cana-2643	239	3	2495	2495	NUM
cana-2643	239	4	,	,	PUNCT
cana-2643	239	5	1	1	NUM
cana-2643	239	6	dec.2017	dec.2017	NUM
cana-2643	239	7	,	,	PUNCT
cana-2643	239	8	doi	doi	NOUN
cana-2643	239	9	:	:	PUNCT
cana-2643	239	10	10.1109	10.1109	NUM
cana-2643	239	11	/	/	SYM
cana-2643	239	12	tpami.2016.2644615	tpami.2016.2644615	NOUN
cana-2643	239	13	.	.	PUNCT
cana-2643	240	1	[	[	X
cana-2643	240	2	5	5	NUM
cana-2643	240	3	]	]	PUNCT
cana-2643	240	4	r.	r.	PROPN
cana-2643	240	5	hamaguchi	hamaguchi	PROPN
cana-2643	240	6	,	,	PUNCT
cana-2643	240	7	a.	a.	PROPN
cana-2643	240	8	fujita	fujita	PROPN
cana-2643	240	9	,	,	PUNCT
cana-2643	240	10	k.	k.	PROPN
cana-2643	240	11	nemoto	nemoto	PROPN
cana-2643	240	12	,	,	PUNCT
cana-2643	240	13	t.	t.	NOUN
cana-2643	240	14	imaizumi	imaizumi	PROPN
cana-2643	240	15	and	and	CCONJ
cana-2643	240	16	s.	s.	PROPN
cana-2643	240	17	hikosaka	hikosaka	PROPN
cana-2643	240	18	,	,	PUNCT
cana-2643	240	19	"	"	PUNCT
cana-2643	240	20	effective	effective	ADJ
cana-2643	240	21	use	use	NOUN
cana-2643	240	22	of	of	ADP
cana-2643	240	23	dilated	dilated	ADJ
cana-2643	240	24	convolutions	convolution	NOUN
cana-2643	240	25	for	for	ADP
cana-2643	240	26	segmenting	segment	VERB
cana-2643	240	27	small	small	ADJ
cana-2643	240	28	object	object	NOUN
cana-2643	240	29	instances	instance	NOUN
cana-2643	240	30	in	in	ADP
cana-2643	240	31	remote	remote	ADJ
cana-2643	240	32	sensing	sense	VERB
cana-2643	240	33	imagery	imagery	NOUN
cana-2643	240	34	,	,	PUNCT
cana-2643	240	35	"	"	PUNCT
cana-2643	240	36	2018	2018	NUM
cana-2643	240	37	ieee	ieee	NOUN
cana-2643	240	38	winter	winter	NOUN
cana-2643	240	39	conference	conference	NOUN
cana-2643	240	40	on	on	ADP
cana-2643	240	41	applications	application	NOUN
cana-2643	240	42	of	of	ADP
cana-2643	240	43	computer	computer	NOUN
cana-2643	240	44	vision	vision	NOUN
cana-2643	240	45	(	(	PUNCT
cana-2643	240	46	wacv	wacv	NOUN
cana-2643	240	47	)	)	PUNCT
cana-2643	240	48	,	,	PUNCT
cana-2643	240	49	lake	lake	PROPN
cana-2643	240	50	tahoe	tahoe	PROPN
cana-2643	240	51	,	,	PUNCT
cana-2643	240	52	nv	nv	PROPN
cana-2643	240	53	,	,	PUNCT
cana-2643	240	54	usa	usa	PROPN
cana-2643	240	55	,	,	PUNCT
cana-2643	240	56	2018	2018	NUM
cana-2643	240	57	,	,	PUNCT
cana-2643	240	58	pp	pp	ADJ
cana-2643	240	59	.	.	PUNCT
cana-2643	240	60	1442	1442	NUM
cana-2643	240	61	-	-	SYM
cana-2643	240	62	1450	1450	NUM
cana-2643	240	63	,	,	PUNCT
cana-2643	240	64	doi	doi	NOUN
cana-2643	240	65	:	:	PUNCT
cana-2643	240	66	10.1109	10.1109	NUM
cana-2643	240	67	/	/	SYM
cana-2643	240	68	wacv.2018.00162	wacv.2018.00162	NOUN
cana-2643	240	69	.	.	PUNCT
cana-2643	241	1	[	[	X
cana-2643	241	2	6	6	NUM
cana-2643	241	3	]	]	X
cana-2643	241	4	rao	rao	PROPN
cana-2643	241	5	pk	pk	PROPN
cana-2643	241	6	,	,	PUNCT
cana-2643	241	7	chatterjee	chatterjee	PROPN
cana-2643	241	8	s	s	PROPN
cana-2643	241	9	,	,	PUNCT
cana-2643	241	10	nagaraju	nagaraju	ADJ
cana-2643	241	11	k	k	PROPN
cana-2643	241	12	,	,	PUNCT
cana-2643	241	13	khan	khan	PROPN
cana-2643	241	14	sb	sb	PROPN
cana-2643	241	15	,	,	PUNCT
cana-2643	241	16	almusharraf	almusharraf	PROPN
cana-2643	241	17	a	a	PRON
cana-2643	241	18	,	,	PUNCT
cana-2643	241	19	alharbi	alharbi	PROPN
cana-2643	241	20	ai	ai	NOUN
cana-2643	241	21	.	.	NOUN
cana-2643	241	22	fusion	fusion	NOUN
cana-2643	241	23	of	of	ADP
cana-2643	241	24	graph	graph	NOUN
cana-2643	241	25	and	and	CCONJ
cana-2643	241	26	tabular	tabular	VERB
cana-2643	241	27	deep	deep	ADJ
cana-2643	241	28	learning	learning	NOUN
cana-2643	241	29	models	model	NOUN
cana-2643	241	30	for	for	ADP
cana-2643	241	31	predicting	predict	VERB
cana-2643	241	32	chronic	chronic	ADJ
cana-2643	241	33	kidney	kidney	NOUN
cana-2643	241	34	disease	disease	NOUN
cana-2643	241	35	.	.	PUNCT
cana-2643	242	1	diagnostics	diagnostic	NOUN
cana-2643	242	2	.	.	PUNCT
cana-2643	243	1	2023	2023	NUM
cana-2643	243	2	;	;	PUNCT
cana-2643	243	3	13(12):1981	13(12):1981	NUM
cana-2643	243	4	.	.	X
cana-2643	244	1	https://doi.org/10.3390/diagnostics13121981	https://doi.org/10.3390/diagnostics13121981	NOUN
cana-2643	244	2	[	[	X
cana-2643	244	3	7	7	X
cana-2643	244	4	]	]	X
cana-2643	244	5	rao	rao	PROPN
cana-2643	244	6	pk	pk	PROPN
cana-2643	244	7	,	,	PUNCT
cana-2643	244	8	chatterjee	chatterjee	PROPN
cana-2643	244	9	s	s	PROPN
cana-2643	244	10	,	,	PUNCT
cana-2643	244	11	janardhan	janardhan	PROPN
cana-2643	244	12	m	m	PROPN
cana-2643	244	13	,	,	PUNCT
cana-2643	244	14	nagaraju	nagaraju	PROPN
cana-2643	244	15	k	k	PROPN
cana-2643	244	16	,	,	PUNCT
cana-2643	244	17	khan	khan	PROPN
cana-2643	244	18	sb	sb	PROPN
cana-2643	244	19	,	,	PUNCT
cana-2643	244	20	almusharraf	almusharraf	PROPN
cana-2643	244	21	a	a	PRON
cana-2643	244	22	,	,	PUNCT
cana-2643	244	23	al	al	PROPN
cana-2643	244	24	-	-	PUNCT
cana-2643	244	25	harbe	harbe	NOUN
cana-2643	244	26	ai	ai	NOUN
cana-2643	244	27	.	.	PUNCT
cana-2643	245	1	optimizing	optimize	VERB
cana-2643	245	2	inference	inference	NOUN
cana-2643	245	3	distribution	distribution	NOUN
cana-2643	245	4	for	for	ADP
cana-2643	245	5	efficient	efficient	ADJ
cana-2643	245	6	kidney	kidney	NOUN
cana-2643	245	7	tumor	tumor	NOUN
cana-2643	245	8	segmentation	segmentation	NOUN
cana-2643	245	9	using	use	VERB
cana-2643	245	10	a	a	DET
cana-2643	245	11	unet	unet	NOUN
cana-2643	245	12	-	-	PUNCT
cana-2643	245	13	pwp	pwp	NOUN
cana-2643	245	14	deep	deep	ADJ
cana-2643	245	15	-	-	PUNCT
cana-2643	245	16	learning	learn	VERB
cana-2643	245	17	model	model	NOUN
cana-2643	245	18	with	with	ADP
cana-2643	245	19	xai	xai	PROPN
cana-2643	245	20	on	on	ADP
cana-2643	245	21	ct	ct	NUM
cana-2643	245	22	scan	scan	NOUN
cana-2643	245	23	images	image	NOUN
cana-2643	245	24	.	.	PUNCT
cana-2643	246	1	diagnostics	diagnostic	NOUN
cana-2643	246	2	.	.	PUNCT
cana-2643	247	1	2023	2023	NUM
cana-2643	247	2	;	;	PUNCT
cana-2643	247	3	13(20):3244	13(20):3244	X
cana-2643	247	4	.	.	PUNCT
cana-2643	248	1	https://doi.org/10.3390/diagnostics13203244	https://doi.org/10.3390/diagnostics13203244	PROPN
cana-2643	249	1	[	[	X
cana-2643	249	2	8	8	NUM
cana-2643	249	3	]	]	X
cana-2643	249	4	van	van	PROPN
cana-2643	249	5	hasselt	hasselt	PROPN
cana-2643	249	6	,	,	PUNCT
cana-2643	249	7	h.	h.	PROPN
cana-2643	249	8	,	,	PUNCT
cana-2643	249	9	guez	guez	PROPN
cana-2643	249	10	,	,	PUNCT
cana-2643	249	11	a.	a.	PROPN
cana-2643	249	12	,	,	PUNCT
cana-2643	249	13	&	&	CCONJ
cana-2643	249	14	silver	silver	PROPN
cana-2643	249	15	,	,	PUNCT
cana-2643	249	16	d.	d.	PROPN
cana-2643	249	17	(	(	PUNCT
cana-2643	249	18	2016	2016	NUM
cana-2643	249	19	)	)	PUNCT
cana-2643	249	20	.	.	PUNCT
cana-2643	250	1	deep	deep	ADJ
cana-2643	250	2	reinforcement	reinforcement	NOUN
cana-2643	250	3	learning	learning	NOUN
cana-2643	250	4	with	with	ADP
cana-2643	250	5	double	double	ADJ
cana-2643	250	6	q	q	NOUN
cana-2643	250	7	-	-	PUNCT
cana-2643	250	8	learning	learning	NOUN
cana-2643	250	9	.	.	PUNCT
cana-2643	251	1	in	in	ADP
cana-2643	251	2	proceedings	proceeding	NOUN
cana-2643	251	3	of	of	ADP
cana-2643	251	4	the	the	DET
cana-2643	251	5	aaai	aaai	PROPN
cana-2643	251	6	conference	conference	NOUN
cana-2643	251	7	on	on	ADP
cana-2643	251	8	artificial	artificial	ADJ
cana-2643	251	9	intelligence	intelligence	NOUN
cana-2643	251	10	(	(	PUNCT
cana-2643	251	11	vol	vol	NOUN
cana-2643	251	12	.	.	PROPN
cana-2643	251	13	30	30	NUM
cana-2643	251	14	,	,	PUNCT
cana-2643	251	15	issue	issue	NOUN
cana-2643	251	16	1	1	NUM
cana-2643	251	17	)	)	PUNCT
cana-2643	251	18	.	.	PUNCT
cana-2643	252	1	association	association	NOUN
cana-2643	252	2	for	for	ADP
cana-2643	252	3	the	the	DET
cana-2643	252	4	advancement	advancement	NOUN
cana-2643	252	5	of	of	ADP
cana-2643	252	6	artificial	artificial	ADJ
cana-2643	252	7	intelligence	intelligence	NOUN
cana-2643	252	8	(	(	PUNCT
cana-2643	252	9	aaai	aaai	PROPN
cana-2643	252	10	)	)	PUNCT
cana-2643	252	11	.	.	PUNCT
cana-2643	253	1	https://doi.org/10.1609/aaai.v30i1.10295	https://doi.org/10.1609/aaai.v30i1.10295	PRON
cana-2643	254	1	[	[	X
cana-2643	254	2	9	9	NUM
cana-2643	254	3	]	]	X
cana-2643	254	4	kaelbling	kaelbling	NOUN
cana-2643	254	5	,	,	PUNCT
cana-2643	254	6	l.	l.	PROPN
cana-2643	254	7	p.	p.	PROPN
cana-2643	254	8	,	,	PUNCT
cana-2643	254	9	littman	littman	PROPN
cana-2643	254	10	,	,	PUNCT
cana-2643	254	11	m.	m.	PROPN
cana-2643	254	12	l.	l.	PROPN
cana-2643	254	13	,	,	PUNCT
cana-2643	254	14	&	&	CCONJ
cana-2643	254	15	moore	moore	PROPN
cana-2643	254	16	,	,	PUNCT
cana-2643	254	17	a.	a.	PROPN
cana-2643	254	18	w.	w.	PROPN
cana-2643	254	19	(	(	PUNCT
cana-2643	254	20	1996	1996	NUM
cana-2643	254	21	)	)	PUNCT
cana-2643	254	22	.	.	PUNCT
cana-2643	255	1	reinforcement	reinforcement	NOUN
cana-2643	255	2	learning	learning	NOUN
cana-2643	255	3	:	:	PUNCT
cana-2643	255	4	a	a	DET
cana-2643	255	5	survey	survey	NOUN
cana-2643	255	6	.	.	PUNCT
cana-2643	256	1	journal	journal	NOUN
cana-2643	256	2	of	of	ADP
cana-2643	256	3	artificial	artificial	ADJ
cana-2643	256	4	intelligence	intelligence	NOUN
cana-2643	256	5	research	research	NOUN
cana-2643	256	6	,	,	PUNCT
cana-2643	256	7	4	4	NUM
cana-2643	256	8	,	,	PUNCT
cana-2643	256	9	237	237	NUM
cana-2643	256	10	-	-	SYM
cana-2643	256	11	285	285	NUM
cana-2643	256	12	.	.	PUNCT
cana-2643	257	1	[	[	X
cana-2643	257	2	10	10	NUM
cana-2643	257	3	]	]	X
cana-2643	257	4	feng	feng	PROPN
cana-2643	257	5	,	,	PUNCT
cana-2643	257	6	s.	s.	PROPN
cana-2643	257	7	,	,	PUNCT
cana-2643	257	8	zhao	zhao	PROPN
cana-2643	257	9	,	,	PUNCT
cana-2643	257	10	l.	l.	PROPN
cana-2643	257	11	,	,	PUNCT
cana-2643	257	12	shi	shi	PROPN
cana-2643	257	13	,	,	PUNCT
cana-2643	257	14	h.	h.	PROPN
cana-2643	257	15	,	,	PUNCT
cana-2643	257	16	wang	wang	PROPN
cana-2643	257	17	,	,	PUNCT
cana-2643	257	18	m.	m.	NOUN
cana-2643	257	19	,	,	PUNCT
cana-2643	257	20	shen	shen	PROPN
cana-2643	257	21	,	,	PUNCT
cana-2643	257	22	s.	s.	PROPN
cana-2643	257	23	,	,	PUNCT
cana-2643	257	24	&	&	CCONJ
cana-2643	257	25	wang	wang	PROPN
cana-2643	257	26	,	,	PUNCT
cana-2643	257	27	w.	w.	PROPN
cana-2643	257	28	(	(	PUNCT
cana-2643	257	29	2023	2023	NUM
cana-2643	257	30	)	)	PUNCT
cana-2643	257	31	.	.	PUNCT
cana-2643	258	1	onedimensional	onedimensional	ADJ
cana-2643	258	2	vggnet	vggnet	PROPN
cana-2643	258	3	for	for	ADP
cana-2643	258	4	highdimensional	highdimensional	ADJ
cana-2643	258	5	data	datum	NOUN
cana-2643	258	6	.	.	PUNCT
cana-2643	259	1	in	in	ADP
cana-2643	259	2	applied	apply	VERB
cana-2643	259	3	soft	soft	ADJ
cana-2643	259	4	computing	computing	NOUN
cana-2643	259	5	(	(	PUNCT
cana-2643	259	6	vol	vol	NOUN
cana-2643	259	7	.	.	PROPN
cana-2643	259	8	135	135	NUM
cana-2643	259	9	,	,	PUNCT
cana-2643	259	10	p.	p.	NOUN
cana-2643	259	11	110035	110035	NUM
cana-2643	259	12	)	)	PUNCT
cana-2643	259	13	.	.	PUNCT
cana-2643	260	1	elsevier	elsevier	PROPN
cana-2643	260	2	bv	bv	PROPN
cana-2643	260	3	.	.	PROPN
cana-2643	260	4	https://doi.org/10.1016/j.asoc.2023.110035	https://doi.org/10.1016/j.asoc.2023.110035	PROPN
cana-2643	260	5	[	[	X
cana-2643	260	6	11	11	NUM
cana-2643	260	7	]	]	SYM
cana-2643	260	8	vaswani	vaswani	NOUN
cana-2643	260	9	,	,	PUNCT
cana-2643	260	10	a.	a.	NOUN
cana-2643	260	11	(	(	PUNCT
cana-2643	260	12	2017	2017	NUM
cana-2643	260	13	)	)	PUNCT
cana-2643	260	14	.	.	PUNCT
cana-2643	261	1	attention	attention	NOUN
cana-2643	261	2	is	be	AUX
cana-2643	261	3	all	all	PRON
cana-2643	261	4	you	you	PRON
cana-2643	261	5	need	need	VERB
cana-2643	261	6	.	.	PUNCT
cana-2643	262	1	advances	advance	NOUN
cana-2643	262	2	in	in	ADP
cana-2643	262	3	neural	neural	ADJ
cana-2643	262	4	information	information	NOUN
cana-2643	262	5	processing	processing	NOUN
cana-2643	262	6	systems	system	NOUN
cana-2643	262	7	.	.	PUNCT
cana-2643	263	1	[	[	X
cana-2643	263	2	12	12	NUM
cana-2643	263	3	]	]	PUNCT
cana-2643	263	4	h.	h.	PROPN
cana-2643	263	5	zhang	zhang	PROPN
cana-2643	263	6	,	,	PUNCT
cana-2643	263	7	j.	j.	PROPN
cana-2643	263	8	duan	duan	PROPN
cana-2643	263	9	,	,	PUNCT
cana-2643	263	10	m.	m.	PROPN
cana-2643	263	11	xue	xue	PROPN
cana-2643	263	12	,	,	PUNCT
cana-2643	263	13	j.	j.	PROPN
cana-2643	263	14	song	song	PROPN
cana-2643	263	15	,	,	PUNCT
cana-2643	263	16	l.	l.	PROPN
cana-2643	263	17	sun	sun	PROPN
cana-2643	263	18	and	and	CCONJ
cana-2643	263	19	m.	m.	NOUN
cana-2643	263	20	song	song	NOUN
cana-2643	263	21	,	,	PUNCT
cana-2643	263	22	"	"	PUNCT
cana-2643	263	23	bootstrapping	bootstrappe	VERB
cana-2643	263	24	vits	vit	NOUN
cana-2643	263	25	:	:	PUNCT
cana-2643	263	26	towards	towards	ADP
cana-2643	263	27	liberating	liberate	VERB
cana-2643	263	28	vision	vision	NOUN
cana-2643	263	29	transformers	transformer	NOUN
cana-2643	263	30	from	from	ADP
cana-2643	263	31	pre	pre	ADJ
cana-2643	263	32	-	-	NOUN
cana-2643	263	33	training	training	NOUN
cana-2643	263	34	,	,	PUNCT
cana-2643	263	35	"	"	PUNCT
cana-2643	263	36	2022	2022	NUM
cana-2643	263	37	ieee	ieee	NOUN
cana-2643	263	38	/	/	SYM
cana-2643	263	39	cvf	cvf	NOUN
cana-2643	263	40	conference	conference	NOUN
cana-2643	263	41	on	on	ADP
cana-2643	263	42	computer	computer	NOUN
cana-2643	263	43	vision	vision	NOUN
cana-2643	263	44	and	and	CCONJ
cana-2643	263	45	pattern	pattern	NOUN
cana-2643	263	46	recognition	recognition	NOUN
cana-2643	263	47	(	(	PUNCT
cana-2643	263	48	cvpr	cvpr	NOUN
cana-2643	263	49	)	)	PUNCT
cana-2643	263	50	,	,	PUNCT
cana-2643	263	51	new	new	PROPN
cana-2643	263	52	orleans	orleans	PROPN
cana-2643	263	53	,	,	PUNCT
cana-2643	263	54	la	la	PROPN
cana-2643	263	55	,	,	PUNCT
cana-2643	263	56	usa	usa	PROPN
cana-2643	263	57	,	,	PUNCT
cana-2643	263	58	2022	2022	NUM
cana-2643	263	59	,	,	PUNCT
cana-2643	263	60	pp	pp	ADV
cana-2643	263	61	.	.	PUNCT
cana-2643	263	62	8934	8934	NUM
cana-2643	263	63	-	-	SYM
cana-2643	263	64	8943	8943	NUM
cana-2643	263	65	,	,	PUNCT
cana-2643	263	66	doi	doi	NOUN
cana-2643	263	67	:	:	PUNCT
cana-2643	263	68	10.1109	10.1109	NUM
cana-2643	263	69	/	/	SYM
cana-2643	263	70	cvpr52688.2022.00874	cvpr52688.2022.00874	NOUN
cana-2643	263	71	.	.	PUNCT
cana-2643	264	1	[	[	X
cana-2643	264	2	13	13	NUM
cana-2643	264	3	]	]	PUNCT
cana-2643	264	4	kipf	kipf	PROPN
cana-2643	264	5	,	,	PUNCT
cana-2643	264	6	t.	t.	PROPN
cana-2643	264	7	n.	n.	PROPN
cana-2643	264	8	,	,	PUNCT
cana-2643	264	9	&	&	CCONJ
cana-2643	264	10	welling	well	VERB
cana-2643	264	11	,	,	PUNCT
cana-2643	264	12	m.	m.	NOUN
cana-2643	264	13	(	(	PUNCT
cana-2643	264	14	2016	2016	NUM
cana-2643	264	15	)	)	PUNCT
cana-2643	264	16	.	.	PUNCT
cana-2643	265	1	variational	variational	ADJ
cana-2643	265	2	graph	graph	NOUN
cana-2643	265	3	auto	auto	NOUN
cana-2643	265	4	-	-	PUNCT
cana-2643	265	5	encoders	encoder	NOUN
cana-2643	265	6	.	.	PUNCT
cana-2643	266	1	arxiv	arxiv	PROPN
cana-2643	266	2	preprint	preprint	VERB
cana-2643	266	3	arxiv:1611.07308	arxiv:1611.07308	ADP
cana-2643	266	4	.	.	PUNCT
cana-2643	267	1	communications	communication	NOUN
cana-2643	267	2	on	on	ADP
cana-2643	267	3	applied	apply	VERB
cana-2643	267	4	nonlinear	nonlinear	ADJ
cana-2643	267	5	analysis	analysis	NOUN
cana-2643	267	6	issn	issn	NOUN
cana-2643	267	7	:	:	PUNCT
cana-2643	267	8	1074	1074	NUM
cana-2643	267	9	-	-	PUNCT
cana-2643	267	10	133x	133x	NUM
cana-2643	267	11	vol	vol	NOUN
cana-2643	267	12	32	32	NUM
cana-2643	267	13	no	no	NOUN
cana-2643	267	14	.	.	PUNCT
cana-2643	268	1	3s	3s	NUM
cana-2643	268	2	(	(	PUNCT
cana-2643	268	3	2025	2025	NUM
cana-2643	268	4	)	)	PUNCT
cana-2643	268	5	294	294	NUM
cana-2643	268	6	https://internationalpubls.com	https://internationalpubls.com	X
cana-2643	269	1	[	[	X
cana-2643	269	2	14	14	NUM
cana-2643	269	3	]	]	PUNCT
cana-2643	269	4	x.	x.	NOUN
cana-2643	269	5	cai	cai	PROPN
cana-2643	269	6	and	and	CCONJ
cana-2643	269	7	y.	y.	PROPN
cana-2643	269	8	chen	chen	PROPN
cana-2643	269	9	,	,	PUNCT
cana-2643	269	10	"	"	PUNCT
cana-2643	269	11	multipath	multipath	NOUN
cana-2643	269	12	routing	routing	NOUN
cana-2643	269	13	for	for	ADP
cana-2643	269	14	traffic	traffic	NOUN
cana-2643	269	15	engineering	engineering	NOUN
cana-2643	269	16	with	with	ADP
cana-2643	269	17	hypergraph	hypergraph	NOUN
cana-2643	269	18	attention	attention	NOUN
cana-2643	269	19	enhanced	enhance	VERB
cana-2643	269	20	multi	multi	ADJ
cana-2643	269	21	-	-	ADJ
cana-2643	269	22	agent	agent	ADJ
cana-2643	269	23	reinforcement	reinforcement	NOUN
cana-2643	269	24	learning	learning	NOUN
cana-2643	269	25	,	,	PUNCT
cana-2643	269	26	"	"	PUNCT
cana-2643	269	27	2022	2022	NUM
cana-2643	269	28	31st	31st	NOUN
cana-2643	269	29	wireless	wireless	NOUN
cana-2643	269	30	and	and	CCONJ
cana-2643	269	31	optical	optical	ADJ
cana-2643	269	32	communications	communication	NOUN
cana-2643	269	33	conference	conference	NOUN
cana-2643	269	34	(	(	PUNCT
cana-2643	269	35	wocc	wocc	PROPN
cana-2643	269	36	)	)	PUNCT
cana-2643	269	37	,	,	PUNCT
cana-2643	269	38	shenzhen	shenzhen	PROPN
cana-2643	269	39	,	,	PUNCT
cana-2643	269	40	china	china	PROPN
cana-2643	269	41	,	,	PUNCT
cana-2643	269	42	2022	2022	NUM
cana-2643	269	43	,	,	PUNCT
cana-2643	269	44	pp	pp	ADJ
cana-2643	269	45	.	.	PUNCT
cana-2643	270	1	103108	103108	NUM
cana-2643	270	2	,	,	PUNCT
cana-2643	270	3	doi	doi	NOUN
cana-2643	270	4	:	:	PUNCT
cana-2643	270	5	10.1109	10.1109	NUM
cana-2643	270	6	/	/	SYM
cana-2643	270	7	wocc55104.2022.9880574	wocc55104.2022.9880574	NOUN
cana-2643	270	8	.	.	PUNCT
cana-2643	271	1	[	[	X
cana-2643	271	2	15	15	NUM
cana-2643	271	3	]	]	X
cana-2643	271	4	joshua	joshua	PROPN
cana-2643	271	5	wong	wong	PROPN
cana-2643	271	6	,	,	PUNCT
cana-2643	271	7	emily	emily	PROPN
cana-2643	271	8	a.	a.	PROPN
cana-2643	271	9	nack	nack	PROPN
cana-2643	271	10	,	,	PUNCT
cana-2643	271	11	zachary	zachary	PROPN
cana-2643	271	12	steelman	steelman	PROPN
cana-2643	271	13	et	et	PROPN
cana-2643	271	14	al	al	PROPN
cana-2643	271	15	.	.	PROPN
cana-2643	271	16	,	,	PUNCT
cana-2643	271	17	"	"	PUNCT
cana-2643	271	18	a	a	DET
cana-2643	271	19	methodology	methodology	NOUN
cana-2643	271	20	for	for	ADP
cana-2643	271	21	representing	represent	VERB
cana-2643	271	22	and	and	CCONJ
cana-2643	271	23	assessing	assess	VERB
cana-2643	271	24	artificial	artificial	ADJ
cana-2643	271	25	intelligence	intelligence	NOUN
cana-2643	271	26	decision	decision	NOUN
cana-2643	271	27	aids	aid	NOUN
cana-2643	271	28	within	within	ADP
cana-2643	271	29	modeling	modeling	NOUN
cana-2643	271	30	and	and	CCONJ
cana-2643	271	31	simulation	simulation	NOUN
cana-2643	271	32	"	"	PUNCT
cana-2643	271	33	,	,	PUNCT
cana-2643	271	34	artificial	artificial	ADJ
cana-2643	271	35	intelligence	intelligence	NOUN
cana-2643	271	36	and	and	CCONJ
cana-2643	271	37	machine	machine	NOUN
cana-2643	271	38	learning	learning	NOUN
cana-2643	271	39	for	for	ADP
cana-2643	271	40	multi	multi	ADJ
cana-2643	271	41	-	-	ADJ
cana-2643	271	42	domain	domain	ADJ
cana-2643	271	43	operations	operation	NOUN
cana-2643	271	44	applications	application	NOUN
cana-2643	271	45	vi	vi	PROPN
cana-2643	271	46	,	,	PUNCT
cana-2643	271	47	21	21	NUM
cana-2643	271	48	,	,	PUNCT
cana-2643	271	49	(	(	PUNCT
cana-2643	271	50	2024	2024	NUM
cana-2643	271	51	)	)	PUNCT
cana-2643	271	52	;	;	PUNCT
cana-2643	271	53	doi:10.1117/12.3013180	doi:10.1117/12.3013180	PROPN
cana-2643	271	54	[	[	PUNCT
cana-2643	271	55	16	16	NUM
cana-2643	271	56	]	]	X
cana-2643	271	57	xiong	xiong	PROPN
cana-2643	271	58	,	,	PUNCT
cana-2643	271	59	f.	f.	PROPN
cana-2643	271	60	,	,	PUNCT
cana-2643	271	61	xiao	xiao	PROPN
cana-2643	271	62	,	,	PUNCT
cana-2643	271	63	y.	y.	PROPN
cana-2643	271	64	,	,	PUNCT
cana-2643	271	65	cao	cao	PROPN
cana-2643	271	66	,	,	PUNCT
cana-2643	271	67	z.	z.	PROPN
cana-2643	271	68	,	,	PUNCT
cana-2643	271	69	gong	gong	PROPN
cana-2643	271	70	,	,	PUNCT
cana-2643	271	71	k.	k.	PROPN
cana-2643	271	72	,	,	PUNCT
cana-2643	271	73	fang	fang	PROPN
cana-2643	271	74	,	,	PUNCT
cana-2643	271	75	z.	z.	PROPN
cana-2643	271	76	,	,	PUNCT
cana-2643	271	77	&	&	CCONJ
cana-2643	271	78	zhou	zhou	PROPN
cana-2643	271	79	,	,	PUNCT
cana-2643	271	80	j.	j.	PROPN
cana-2643	271	81	t.	t.	PROPN
cana-2643	271	82	(	(	PUNCT
cana-2643	271	83	2019	2019	NUM
cana-2643	271	84	,	,	PUNCT
cana-2643	271	85	may	may	AUX
cana-2643	271	86	)	)	PUNCT
cana-2643	271	87	.	.	PUNCT
cana-2643	272	1	good	good	ADJ
cana-2643	272	2	practices	practice	NOUN
cana-2643	272	3	on	on	ADP
cana-2643	272	4	building	build	VERB
cana-2643	272	5	effective	effective	ADJ
cana-2643	272	6	cnn	cnn	PROPN
cana-2643	272	7	baseline	baseline	PROPN
cana-2643	272	8	model	model	NOUN
cana-2643	272	9	for	for	ADP
cana-2643	272	10	person	person	NOUN
cana-2643	272	11	re	re	NOUN
cana-2643	272	12	-	-	NOUN
cana-2643	272	13	identification	identification	NOUN
cana-2643	272	14	.	.	PUNCT
cana-2643	273	1	in	in	ADP
cana-2643	273	2	tenth	tenth	ADJ
cana-2643	273	3	international	international	ADJ
cana-2643	273	4	conference	conference	NOUN
cana-2643	273	5	on	on	ADP
cana-2643	273	6	graphics	graphic	NOUN
cana-2643	273	7	and	and	CCONJ
cana-2643	273	8	image	image	NOUN
cana-2643	273	9	processing	processing	NOUN
cana-2643	273	10	(	(	PUNCT
cana-2643	273	11	icgip	icgip	PROPN
cana-2643	273	12	2018	2018	NUM
cana-2643	273	13	)	)	PUNCT
cana-2643	273	14	(	(	PUNCT
cana-2643	273	15	vol	vol	NOUN
cana-2643	273	16	.	.	NOUN
cana-2643	273	17	11069	11069	NUM
cana-2643	273	18	,	,	PUNCT
cana-2643	273	19	pp	pp	ADJ
cana-2643	273	20	.	.	PUNCT
cana-2643	274	1	142	142	NUM
cana-2643	274	2	-	-	SYM
cana-2643	274	3	152	152	NUM
cana-2643	274	4	)	)	PUNCT
cana-2643	274	5	.	.	PUNCT
cana-2643	275	1	spie	spie	NOUN
cana-2643	275	2	.	.	PUNCT
cana-2643	276	1	[	[	X
cana-2643	276	2	17	17	NUM
cana-2643	276	3	]	]	SYM
cana-2643	276	4	yuan	yuan	NOUN
cana-2643	276	5	,	,	PUNCT
cana-2643	276	6	z.	z.	PROPN
cana-2643	276	7	w.	w.	PROPN
cana-2643	276	8	,	,	PUNCT
cana-2643	276	9	&	&	CCONJ
cana-2643	276	10	zhang	zhang	PROPN
cana-2643	276	11	,	,	PUNCT
cana-2643	276	12	j.	j.	PROPN
cana-2643	276	13	(	(	PUNCT
cana-2643	276	14	2016	2016	NUM
cana-2643	276	15	,	,	PUNCT
cana-2643	276	16	august	august	PROPN
cana-2643	276	17	)	)	PUNCT
cana-2643	276	18	.	.	PUNCT
cana-2643	277	1	feature	feature	NOUN
cana-2643	277	2	extraction	extraction	NOUN
cana-2643	277	3	and	and	CCONJ
cana-2643	277	4	image	image	NOUN
cana-2643	277	5	retrieval	retrieval	NOUN
cana-2643	277	6	based	base	VERB
cana-2643	277	7	on	on	ADP
cana-2643	277	8	alexnet	alexnet	NOUN
cana-2643	277	9	.	.	PUNCT
cana-2643	278	1	in	in	ADP
cana-2643	278	2	eighth	eighth	ADJ
cana-2643	278	3	international	international	ADJ
cana-2643	278	4	conference	conference	NOUN
cana-2643	278	5	on	on	ADP
cana-2643	278	6	digital	digital	ADJ
cana-2643	278	7	image	image	NOUN
cana-2643	278	8	processing	processing	NOUN
cana-2643	278	9	(	(	PUNCT
cana-2643	278	10	icdip	icdip	NOUN
cana-2643	278	11	2016	2016	NUM
cana-2643	278	12	)	)	PUNCT
cana-2643	278	13	(	(	PUNCT
cana-2643	278	14	vol	vol	NOUN
cana-2643	278	15	.	.	NOUN
cana-2643	278	16	10033	10033	NUM
cana-2643	278	17	,	,	PUNCT
cana-2643	278	18	pp	pp	ADJ
cana-2643	278	19	.	.	PUNCT
cana-2643	278	20	65	65	NUM
cana-2643	278	21	-	-	SYM
cana-2643	278	22	69	69	NUM
cana-2643	278	23	)	)	PUNCT
cana-2643	278	24	.	.	PUNCT
cana-2643	279	1	spie	spie	NOUN