id	sid	tid	token	lemma	pos
fcis-30136	1	1	frontiers	frontier	NOUN
fcis-30136	1	2	in	in	ADP
fcis-30136	1	3	computing	computing	NOUN
fcis-30136	1	4	and	and	CCONJ
fcis-30136	1	5	intelligent	intelligent	ADJ
fcis-30136	1	6	systems	system	NOUN
fcis-30136	1	7	issn	issn	VERB
fcis-30136	1	8	:	:	PUNCT
fcis-30136	1	9	2832	2832	NUM
fcis-30136	1	10	-	-	SYM
fcis-30136	1	11	6024	6024	NUM
fcis-30136	1	12	|	|	NOUN
fcis-30136	1	13	vol	vol	NOUN
fcis-30136	1	14	.	.	PROPN
fcis-30136	2	1	11	11	NUM
fcis-30136	2	2	,	,	PUNCT
fcis-30136	2	3	no	no	INTJ
fcis-30136	2	4	.	.	NOUN
fcis-30136	2	5	3	3	NUM
fcis-30136	2	6	,	,	PUNCT
fcis-30136	2	7	2025	2025	NUM
fcis-30136	2	8	85	85	NUM
fcis-30136	2	9	improvement	improvement	NOUN
fcis-30136	2	10	and	and	CCONJ
fcis-30136	2	11	implementation	implementation	NOUN
fcis-30136	2	12	of	of	ADP
fcis-30136	2	13	a	a	DET
fcis-30136	2	14	potato	potato	NOUN
fcis-30136	2	15	recognition	recognition	NOUN
fcis-30136	2	16	algorithm	algorithm	NOUN
fcis-30136	2	17	based	base	VERB
fcis-30136	2	18	on	on	ADP
fcis-30136	2	19	yolov8	yolov8	PROPN
fcis-30136	2	20	dongxuan	dongxuan	PROPN
fcis-30136	2	21	huang	huang	PROPN
fcis-30136	2	22	,	,	PUNCT
fcis-30136	2	23	mingge	mingge	NOUN
fcis-30136	2	24	sun	sun	PROPN
fcis-30136	2	25	*	*	PROPN
fcis-30136	2	26	,	,	PUNCT
fcis-30136	2	27	sen	sen	PROPN
fcis-30136	2	28	ye	ye	PROPN
fcis-30136	2	29	,	,	PUNCT
fcis-30136	2	30	jiaxuan	jiaxuan	ADJ
fcis-30136	2	31	chai	chai	NOUN
fcis-30136	2	32	school	school	NOUN
fcis-30136	2	33	of	of	ADP
fcis-30136	2	34	information	information	NOUN
fcis-30136	2	35	and	and	CCONJ
fcis-30136	2	36	control	control	PROPN
fcis-30136	2	37	engineering	engineering	PROPN
fcis-30136	2	38	,	,	PUNCT
fcis-30136	2	39	jilin	jilin	PROPN
fcis-30136	2	40	institute	institute	PROPN
fcis-30136	2	41	of	of	ADP
fcis-30136	2	42	chemical	chemical	PROPN
fcis-30136	2	43	technology	technology	PROPN
fcis-30136	2	44	,	,	PUNCT
fcis-30136	2	45	jilin	jilin	PROPN
fcis-30136	2	46	,	,	PUNCT
fcis-30136	2	47	jilin	jilin	PROPN
fcis-30136	2	48	132022	132022	NUM
fcis-30136	2	49	,	,	PUNCT
fcis-30136	2	50	china	china	PROPN
fcis-30136	2	51	*	*	PUNCT
fcis-30136	2	52	corresponding	correspond	VERB
fcis-30136	2	53	author	author	NOUN
fcis-30136	2	54	:	:	PUNCT
fcis-30136	2	55	mingge	mingge	NOUN
fcis-30136	2	56	sun	sun	PROPN
fcis-30136	2	57	abstract	abstract	NOUN
fcis-30136	2	58	:	:	PUNCT
fcis-30136	2	59	to	to	PART
fcis-30136	2	60	address	address	VERB
fcis-30136	2	61	issues	issue	NOUN
fcis-30136	2	62	such	such	ADJ
fcis-30136	2	63	as	as	ADP
fcis-30136	2	64	poor	poor	ADJ
fcis-30136	2	65	recognition	recognition	NOUN
fcis-30136	2	66	accuracy	accuracy	NOUN
fcis-30136	2	67	and	and	CCONJ
fcis-30136	2	68	low	low	ADJ
fcis-30136	2	69	efficiency	efficiency	NOUN
fcis-30136	2	70	caused	cause	VERB
fcis-30136	2	71	by	by	ADP
fcis-30136	2	72	the	the	DET
fcis-30136	2	73	large	large	ADJ
fcis-30136	2	74	number	number	NOUN
fcis-30136	2	75	of	of	ADP
fcis-30136	2	76	potato	potato	NOUN
fcis-30136	2	77	targets	target	NOUN
fcis-30136	2	78	that	that	PRON
fcis-30136	2	79	need	need	VERB
fcis-30136	2	80	to	to	PART
fcis-30136	2	81	be	be	AUX
fcis-30136	2	82	processed	process	VERB
fcis-30136	2	83	at	at	ADP
fcis-30136	2	84	once	once	ADV
fcis-30136	2	85	during	during	ADP
fcis-30136	2	86	potato	potato	NOUN
fcis-30136	2	87	identification	identification	NOUN
fcis-30136	3	1	,	,	PUNCT
fcis-30136	3	2	this	this	DET
fcis-30136	3	3	paper	paper	NOUN
fcis-30136	3	4	proposes	propose	VERB
fcis-30136	3	5	an	an	DET
fcis-30136	3	6	improved	improved	ADJ
fcis-30136	3	7	yolov8	yolov8	NOUN
fcis-30136	3	8	-	-	PUNCT
fcis-30136	3	9	based	base	VERB
fcis-30136	3	10	potato	potato	NOUN
fcis-30136	3	11	recognition	recognition	NOUN
fcis-30136	3	12	algorithm	algorithm	NOUN
fcis-30136	3	13	.	.	PUNCT
fcis-30136	4	1	the	the	DET
fcis-30136	4	2	improvements	improvement	NOUN
fcis-30136	4	3	include	include	VERB
fcis-30136	4	4	:	:	PUNCT
fcis-30136	4	5	introducing	introduce	VERB
fcis-30136	4	6	dynamicconv	dynamicconv	NOUN
fcis-30136	4	7	in	in	ADP
fcis-30136	4	8	the	the	DET
fcis-30136	4	9	backbone	backbone	NOUN
fcis-30136	4	10	network	network	NOUN
fcis-30136	4	11	to	to	PART
fcis-30136	4	12	enhance	enhance	VERB
fcis-30136	4	13	model	model	NOUN
fcis-30136	4	14	representation	representation	NOUN
fcis-30136	4	15	capability	capability	NOUN
fcis-30136	4	16	;	;	PUNCT
fcis-30136	4	17	replacing	replace	VERB
fcis-30136	4	18	the	the	DET
fcis-30136	4	19	c2f	c2f	NOUN
fcis-30136	4	20	module	module	NOUN
fcis-30136	4	21	in	in	ADP
fcis-30136	4	22	the	the	DET
fcis-30136	4	23	backbone	backbone	NOUN
fcis-30136	4	24	network	network	NOUN
fcis-30136	4	25	with	with	ADP
fcis-30136	4	26	c2f_saconv	c2f_saconv	NOUN
fcis-30136	4	27	to	to	PART
fcis-30136	4	28	strengthen	strengthen	VERB
fcis-30136	4	29	feature	feature	NOUN
fcis-30136	4	30	extraction	extraction	NOUN
fcis-30136	4	31	in	in	ADP
fcis-30136	4	32	object	object	NOUN
fcis-30136	4	33	detection	detection	NOUN
fcis-30136	4	34	and	and	CCONJ
fcis-30136	4	35	segmentation	segmentation	NOUN
fcis-30136	4	36	tasks	task	NOUN
fcis-30136	4	37	;	;	PUNCT
fcis-30136	4	38	incorporating	incorporate	VERB
fcis-30136	4	39	slim	slim	ADJ
fcis-30136	4	40	-	-	PUNCT
fcis-30136	4	41	neck	neck	NOUN
fcis-30136	4	42	in	in	ADP
fcis-30136	4	43	the	the	DET
fcis-30136	4	44	neck	neck	NOUN
fcis-30136	4	45	layer	layer	NOUN
fcis-30136	4	46	to	to	PART
fcis-30136	4	47	reduce	reduce	VERB
fcis-30136	4	48	model	model	NOUN
fcis-30136	4	49	parameters	parameter	NOUN
fcis-30136	4	50	and	and	CCONJ
fcis-30136	4	51	floating	float	VERB
fcis-30136	4	52	-	-	PUNCT
fcis-30136	4	53	point	point	NOUN
fcis-30136	4	54	operations	operation	NOUN
fcis-30136	4	55	;	;	PUNCT
fcis-30136	4	56	and	and	CCONJ
fcis-30136	4	57	adding	add	VERB
fcis-30136	4	58	an	an	DET
fcis-30136	4	59	ema	ema	PROPN
fcis-30136	4	60	attention	attention	NOUN
fcis-30136	4	61	mechanism	mechanism	NOUN
fcis-30136	4	62	to	to	PART
fcis-30136	4	63	improve	improve	VERB
fcis-30136	4	64	the	the	DET
fcis-30136	4	65	model	model	NOUN
fcis-30136	4	66	's	's	PART
fcis-30136	4	67	focus	focus	NOUN
fcis-30136	4	68	on	on	ADP
fcis-30136	4	69	different	different	ADJ
fcis-30136	4	70	potatoes	potato	NOUN
fcis-30136	4	71	.	.	PUNCT
fcis-30136	5	1	the	the	DET
fcis-30136	5	2	improved	improved	ADJ
fcis-30136	5	3	yolov8	yolov8	NOUN
fcis-30136	5	4	algorithm	algorithm	PROPN
fcis-30136	5	5	achieves	achieve	VERB
fcis-30136	5	6	an	an	DET
fcis-30136	5	7	average	average	ADJ
fcis-30136	5	8	recognition	recognition	NOUN
fcis-30136	5	9	accuracy	accuracy	NOUN
fcis-30136	5	10	of	of	ADP
fcis-30136	5	11	89.7	89.7	NUM
fcis-30136	5	12	%	%	NOUN
fcis-30136	5	13	,	,	PUNCT
fcis-30136	5	14	reduces	reduce	VERB
fcis-30136	5	15	computational	computational	ADJ
fcis-30136	5	16	load	load	NOUN
fcis-30136	5	17	by	by	ADP
fcis-30136	5	18	27.2	27.2	NUM
fcis-30136	5	19	%	%	NOUN
fcis-30136	5	20	,	,	PUNCT
fcis-30136	5	21	and	and	CCONJ
fcis-30136	5	22	facilitates	facilitate	VERB
fcis-30136	5	23	deployment	deployment	NOUN
fcis-30136	5	24	on	on	ADP
fcis-30136	5	25	resource	resource	NOUN
fcis-30136	5	26	-	-	PUNCT
fcis-30136	5	27	constrained	constrain	VERB
fcis-30136	5	28	embedded	embed	VERB
fcis-30136	5	29	devices	device	NOUN
fcis-30136	5	30	.	.	PUNCT
fcis-30136	6	1	keywords	keyword	NOUN
fcis-30136	6	2	:	:	PUNCT
fcis-30136	6	3	deep	deep	ADJ
fcis-30136	6	4	learning	learning	NOUN
fcis-30136	6	5	;	;	PUNCT
fcis-30136	6	6	yolov8	yolov8	NOUN
fcis-30136	6	7	;	;	PUNCT
fcis-30136	6	8	attention	attention	NOUN
fcis-30136	6	9	mechanism	mechanism	NOUN
fcis-30136	6	10	;	;	PUNCT
fcis-30136	6	11	potato	potato	NOUN
fcis-30136	6	12	recognition	recognition	NOUN
fcis-30136	6	13	.	.	PUNCT
fcis-30136	7	1	1	1	X
fcis-30136	7	2	.	.	X
fcis-30136	7	3	introduction	introduction	NOUN
fcis-30136	7	4	potatoes	potato	NOUN
fcis-30136	7	5	have	have	AUX
fcis-30136	7	6	now	now	ADV
fcis-30136	7	7	become	become	VERB
fcis-30136	7	8	china	china	PROPN
fcis-30136	7	9	's	's	PART
fcis-30136	7	10	fourth	fourth	ADV
fcis-30136	7	11	largest	large	ADJ
fcis-30136	7	12	staple	staple	ADJ
fcis-30136	7	13	crop	crop	NOUN
fcis-30136	7	14	after	after	ADP
fcis-30136	7	15	corn	corn	NOUN
fcis-30136	7	16	,	,	PUNCT
fcis-30136	7	17	rice	rice	NOUN
fcis-30136	7	18	,	,	PUNCT
fcis-30136	7	19	and	and	CCONJ
fcis-30136	7	20	wheat	wheat	NOUN
fcis-30136	7	21	,	,	PUNCT
fcis-30136	7	22	serving	serve	VERB
fcis-30136	7	23	as	as	ADP
fcis-30136	7	24	an	an	DET
fcis-30136	7	25	important	important	ADJ
fcis-30136	7	26	food	food	NOUN
fcis-30136	7	27	resource	resource	NOUN
fcis-30136	7	28	.	.	PUNCT
fcis-30136	8	1	china	china	PROPN
fcis-30136	8	2	boasts	boast	VERB
fcis-30136	8	3	vast	vast	ADJ
fcis-30136	8	4	potato	potato	NOUN
fcis-30136	8	5	planting	planting	NOUN
fcis-30136	8	6	areas	area	NOUN
fcis-30136	8	7	,	,	PUNCT
fcis-30136	8	8	particularly	particularly	ADV
fcis-30136	8	9	in	in	ADP
fcis-30136	8	10	northeast	northeast	ADJ
fcis-30136	8	11	china	china	PROPN
fcis-30136	8	12	,	,	PUNCT
fcis-30136	8	13	north	north	NOUN
fcis-30136	8	14	china	china	PROPN
fcis-30136	8	15	,	,	PUNCT
fcis-30136	8	16	and	and	CCONJ
fcis-30136	8	17	southwest	southwest	ADJ
fcis-30136	8	18	china	china	PROPN
fcis-30136	9	1	[	[	X
fcis-30136	9	2	1	1	NUM
fcis-30136	9	3	-	-	SYM
fcis-30136	9	4	2	2	NUM
fcis-30136	9	5	]	]	PUNCT
fcis-30136	9	6	.	.	PUNCT
fcis-30136	10	1	potatoes	potato	NOUN
fcis-30136	10	2	serve	serve	VERB
fcis-30136	10	3	multiple	multiple	ADJ
fcis-30136	10	4	purposes	purpose	NOUN
fcis-30136	10	5	:	:	PUNCT
fcis-30136	10	6	they	they	PRON
fcis-30136	10	7	can	can	AUX
fcis-30136	10	8	be	be	AUX
fcis-30136	10	9	cooked	cook	VERB
fcis-30136	10	10	as	as	ADP
fcis-30136	10	11	vegetables	vegetable	NOUN
fcis-30136	10	12	,	,	PUNCT
fcis-30136	10	13	consumed	consume	VERB
fcis-30136	10	14	as	as	ADP
fcis-30136	10	15	staple	staple	ADJ
fcis-30136	10	16	food	food	NOUN
fcis-30136	10	17	,	,	PUNCT
fcis-30136	10	18	or	or	CCONJ
fcis-30136	10	19	used	use	VERB
fcis-30136	10	20	as	as	ADP
fcis-30136	10	21	industrial	industrial	ADJ
fcis-30136	10	22	raw	raw	ADJ
fcis-30136	10	23	materials	material	NOUN
fcis-30136	10	24	for	for	ADP
fcis-30136	10	25	manufacturing	manufacturing	NOUN
fcis-30136	10	26	products	product	NOUN
fcis-30136	10	27	,	,	PUNCT
fcis-30136	10	28	making	make	VERB
fcis-30136	10	29	them	they	PRON
fcis-30136	10	30	a	a	DET
fcis-30136	10	31	versatile	versatile	ADJ
fcis-30136	10	32	grain	grain	NOUN
fcis-30136	10	33	resource	resource	NOUN
fcis-30136	10	34	[	[	X
fcis-30136	10	35	3	3	NUM
fcis-30136	10	36	]	]	PUNCT
fcis-30136	10	37	.	.	PUNCT
fcis-30136	11	1	currently	currently	ADV
fcis-30136	11	2	,	,	PUNCT
fcis-30136	11	3	potato	potato	NOUN
fcis-30136	11	4	classification	classification	NOUN
fcis-30136	11	5	primarily	primarily	ADV
fcis-30136	11	6	relies	rely	VERB
fcis-30136	11	7	on	on	ADP
fcis-30136	11	8	manual	manual	ADJ
fcis-30136	11	9	sorting	sorting	NOUN
fcis-30136	11	10	,	,	PUNCT
fcis-30136	11	11	which	which	PRON
fcis-30136	11	12	suffers	suffer	VERB
fcis-30136	11	13	from	from	ADP
fcis-30136	11	14	low	low	ADJ
fcis-30136	11	15	efficiency	efficiency	NOUN
fcis-30136	11	16	,	,	PUNCT
fcis-30136	11	17	high	high	ADJ
fcis-30136	11	18	labor	labor	NOUN
fcis-30136	11	19	intensity	intensity	NOUN
fcis-30136	11	20	,	,	PUNCT
fcis-30136	11	21	and	and	CCONJ
fcis-30136	11	22	significant	significant	ADJ
fcis-30136	11	23	variability	variability	NOUN
fcis-30136	11	24	between	between	ADP
fcis-30136	11	25	different	different	ADJ
fcis-30136	11	26	workers	worker	NOUN
fcis-30136	11	27	.	.	PUNCT
fcis-30136	12	1	traditional	traditional	ADJ
fcis-30136	12	2	manual	manual	ADJ
fcis-30136	12	3	sorting	sorting	NOUN
fcis-30136	12	4	can	can	AUX
fcis-30136	12	5	no	no	ADV
fcis-30136	12	6	longer	long	ADV
fcis-30136	12	7	meet	meet	VERB
fcis-30136	12	8	the	the	DET
fcis-30136	12	9	processing	processing	NOUN
fcis-30136	12	10	requirements	requirement	NOUN
fcis-30136	12	11	for	for	ADP
fcis-30136	12	12	subsequent	subsequent	ADJ
fcis-30136	12	13	potato	potato	NOUN
fcis-30136	12	14	production	production	NOUN
fcis-30136	12	15	.	.	PUNCT
fcis-30136	13	1	therefore	therefore	ADV
fcis-30136	13	2	,	,	PUNCT
fcis-30136	13	3	research	research	NOUN
fcis-30136	13	4	on	on	ADP
fcis-30136	13	5	automatic	automatic	ADJ
fcis-30136	13	6	potato	potato	NOUN
fcis-30136	13	7	sorting	sort	VERB
fcis-30136	13	8	robots	robot	NOUN
fcis-30136	13	9	holds	hold	VERB
fcis-30136	13	10	significant	significant	ADJ
fcis-30136	13	11	practical	practical	ADJ
fcis-30136	13	12	importance	importance	NOUN
fcis-30136	13	13	.	.	PUNCT
fcis-30136	14	1	in	in	ADP
fcis-30136	14	2	potato	potato	NOUN
fcis-30136	14	3	classification	classification	NOUN
fcis-30136	14	4	,	,	PUNCT
fcis-30136	14	5	operations	operation	NOUN
fcis-30136	14	6	are	be	AUX
fcis-30136	14	7	achieved	achieve	VERB
fcis-30136	14	8	by	by	ADP
fcis-30136	14	9	extracting	extract	VERB
fcis-30136	14	10	feature	feature	NOUN
fcis-30136	14	11	information	information	NOUN
fcis-30136	14	12	such	such	ADJ
fcis-30136	14	13	as	as	ADP
fcis-30136	14	14	size	size	NOUN
fcis-30136	14	15	,	,	PUNCT
fcis-30136	14	16	color	color	NOUN
fcis-30136	14	17	,	,	PUNCT
fcis-30136	14	18	and	and	CCONJ
fcis-30136	14	19	texture	texture	NOUN
fcis-30136	14	20	of	of	ADP
fcis-30136	14	21	different	different	ADJ
fcis-30136	14	22	potato	potato	NOUN
fcis-30136	14	23	varieties	variety	NOUN
fcis-30136	14	24	[	[	X
fcis-30136	14	25	4	4	NUM
fcis-30136	14	26	-	-	SYM
fcis-30136	14	27	5	5	NUM
fcis-30136	14	28	]	]	PUNCT
fcis-30136	14	29	.	.	PUNCT
fcis-30136	15	1	with	with	ADP
fcis-30136	15	2	advancements	advancement	NOUN
fcis-30136	15	3	in	in	ADP
fcis-30136	15	4	science	science	NOUN
fcis-30136	15	5	and	and	CCONJ
fcis-30136	15	6	technology	technology	NOUN
fcis-30136	15	7	,	,	PUNCT
fcis-30136	15	8	numerous	numerous	ADJ
fcis-30136	15	9	machine	machine	NOUN
fcis-30136	15	10	vision	vision	NOUN
fcis-30136	15	11	-	-	PUNCT
fcis-30136	15	12	based	base	VERB
fcis-30136	15	13	classification	classification	NOUN
fcis-30136	15	14	methods	method	NOUN
fcis-30136	15	15	for	for	ADP
fcis-30136	15	16	various	various	ADJ
fcis-30136	15	17	items	item	NOUN
fcis-30136	15	18	have	have	AUX
fcis-30136	15	19	emerged	emerge	VERB
fcis-30136	15	20	.	.	PUNCT
fcis-30136	16	1	alkateb	alkateb	NOUN
fcis-30136	16	2	g	g	PROPN
fcis-30136	16	3	et	et	PROPN
fcis-30136	16	4	al	al	PROPN
fcis-30136	16	5	.	.	PROPN
fcis-30136	16	6	enhanced	enhance	VERB
fcis-30136	16	7	potato	potato	NOUN
fcis-30136	16	8	disease	disease	NOUN
fcis-30136	16	9	management	management	NOUN
fcis-30136	16	10	using	use	VERB
fcis-30136	16	11	generative	generative	ADJ
fcis-30136	16	12	models	model	NOUN
fcis-30136	16	13	and	and	CCONJ
fcis-30136	16	14	convolutional	convolutional	ADJ
fcis-30136	16	15	neural	neural	ADJ
fcis-30136	16	16	networks	network	NOUN
fcis-30136	16	17	(	(	PUNCT
fcis-30136	16	18	cnns	cnns	PROPN
fcis-30136	16	19	)	)	PUNCT
fcis-30136	16	20	.	.	PUNCT
fcis-30136	17	1	traditional	traditional	ADJ
fcis-30136	17	2	disease	disease	NOUN
fcis-30136	17	3	identification	identification	NOUN
fcis-30136	17	4	methods	method	NOUN
fcis-30136	17	5	achieved	achieve	VERB
fcis-30136	17	6	over	over	ADP
fcis-30136	17	7	95	95	NUM
fcis-30136	17	8	%	%	NOUN
fcis-30136	17	9	success	success	NOUN
fcis-30136	17	10	in	in	ADP
fcis-30136	17	11	fully	fully	ADV
fcis-30136	17	12	accurate	accurate	ADJ
fcis-30136	17	13	detection	detection	NOUN
fcis-30136	17	14	and	and	CCONJ
fcis-30136	17	15	85	85	NUM
fcis-30136	17	16	%	%	NOUN
fcis-30136	17	17	success	success	NOUN
fcis-30136	17	18	in	in	ADP
fcis-30136	17	19	early	early	ADJ
fcis-30136	17	20	-	-	PUNCT
fcis-30136	17	21	stage	stage	NOUN
fcis-30136	17	22	detection	detection	NOUN
fcis-30136	17	23	[	[	X
fcis-30136	17	24	6	6	NUM
fcis-30136	17	25	]	]	PUNCT
fcis-30136	17	26	.	.	PUNCT
fcis-30136	18	1	liu	liu	PROPN
fcis-30136	18	2	yijun	yijun	PROPN
fcis-30136	18	3	et	et	PROPN
fcis-30136	18	4	al	al	PROPN
fcis-30136	18	5	.	.	PROPN
fcis-30136	18	6	developed	develop	VERB
fcis-30136	18	7	a	a	DET
fcis-30136	18	8	potato	potato	NOUN
fcis-30136	18	9	sprouting	sprouting	NOUN
fcis-30136	18	10	and	and	CCONJ
fcis-30136	18	11	surface	surface	NOUN
fcis-30136	18	12	damage	damage	NOUN
fcis-30136	18	13	detection	detection	NOUN
fcis-30136	18	14	model	model	NOUN
fcis-30136	18	15	based	base	VERB
fcis-30136	18	16	on	on	ADP
fcis-30136	18	17	the	the	DET
fcis-30136	18	18	faster	fast	ADJ
fcis-30136	18	19	-	-	PUNCT
fcis-30136	18	20	rcnn	rcnn	NOUN
fcis-30136	18	21	neural	neural	ADJ
fcis-30136	18	22	network	network	NOUN
fcis-30136	18	23	.	.	PUNCT
fcis-30136	19	1	they	they	PRON
fcis-30136	19	2	improved	improve	VERB
fcis-30136	19	3	the	the	DET
fcis-30136	19	4	model	model	NOUN
fcis-30136	19	5	using	use	VERB
fcis-30136	19	6	a	a	DET
fcis-30136	19	7	feature	feature	NOUN
fcis-30136	19	8	extraction	extraction	NOUN
fcis-30136	19	9	network	network	NOUN
fcis-30136	19	10	combining	combine	VERB
fcis-30136	19	11	resnet50	resnet50	NOUN
fcis-30136	19	12	residual	residual	ADJ
fcis-30136	19	13	networks	network	NOUN
fcis-30136	19	14	and	and	CCONJ
fcis-30136	19	15	multi	multi	ADJ
fcis-30136	19	16	-	-	ADJ
fcis-30136	19	17	scale	scale	ADJ
fcis-30136	19	18	fusion	fusion	NOUN
fcis-30136	19	19	of	of	ADP
fcis-30136	19	20	pyramid	pyramid	NOUN
fcis-30136	19	21	pooling	pool	VERB
fcis-30136	19	22	networks	network	NOUN
fcis-30136	19	23	,	,	PUNCT
fcis-30136	19	24	enabling	enable	VERB
fcis-30136	19	25	effective	effective	ADJ
fcis-30136	19	26	detection	detection	NOUN
fcis-30136	19	27	and	and	CCONJ
fcis-30136	19	28	evaluation	evaluation	NOUN
fcis-30136	19	29	of	of	ADP
fcis-30136	19	30	sprout	sprout	NOUN
fcis-30136	19	31	size	size	NOUN
fcis-30136	19	32	and	and	CCONJ
fcis-30136	19	33	surface	surface	NOUN
fcis-30136	19	34	damage	damage	NOUN
fcis-30136	19	35	severity	severity	NOUN
fcis-30136	19	36	[	[	X
fcis-30136	19	37	7	7	NUM
fcis-30136	19	38	]	]	PUNCT
fcis-30136	19	39	.	.	PUNCT
fcis-30136	20	1	akther	akther	PROPN
fcis-30136	20	2	j	j	PROPN
fcis-30136	20	3	et	et	PROPN
fcis-30136	20	4	al	al	PROPN
fcis-30136	20	5	.	.	PROPN
fcis-30136	20	6	created	create	VERB
fcis-30136	20	7	a	a	DET
fcis-30136	20	8	cnn	cnn	PROPN
fcis-30136	20	9	model	model	NOUN
fcis-30136	20	10	for	for	ADP
fcis-30136	20	11	real	real	ADJ
fcis-30136	20	12	-	-	PUNCT
fcis-30136	20	13	time	time	NOUN
fcis-30136	20	14	detection	detection	NOUN
fcis-30136	20	15	of	of	ADP
fcis-30136	20	16	early	early	ADJ
fcis-30136	20	17	and	and	CCONJ
fcis-30136	20	18	late	late	ADJ
fcis-30136	20	19	blight	blight	NOUN
fcis-30136	20	20	in	in	ADP
fcis-30136	20	21	potatoes	potato	NOUN
fcis-30136	20	22	.	.	PUNCT
fcis-30136	21	1	the	the	DET
fcis-30136	21	2	method	method	NOUN
fcis-30136	21	3	was	be	AUX
fcis-30136	21	4	evaluated	evaluate	VERB
fcis-30136	21	5	using	use	VERB
fcis-30136	21	6	classification	classification	NOUN
fcis-30136	21	7	optimizers	optimizer	NOUN
fcis-30136	21	8	,	,	PUNCT
fcis-30136	21	9	metrics	metric	NOUN
fcis-30136	21	10	,	,	PUNCT
fcis-30136	21	11	and	and	CCONJ
fcis-30136	21	12	loss	loss	NOUN
fcis-30136	21	13	functions	function	NOUN
fcis-30136	21	14	,	,	PUNCT
fcis-30136	21	15	with	with	ADP
fcis-30136	21	16	further	further	ADJ
fcis-30136	21	17	optimization	optimization	NOUN
fcis-30136	21	18	through	through	ADP
fcis-30136	21	19	layer	layer	NOUN
fcis-30136	21	20	-	-	PUNCT
fcis-30136	21	21	wise	wise	ADJ
fcis-30136	21	22	tensorboard	tensorboard	ADJ
fcis-30136	21	23	analysis	analysis	NOUN
fcis-30136	21	24	.	.	PUNCT
fcis-30136	22	1	the	the	DET
fcis-30136	22	2	final	final	ADJ
fcis-30136	22	3	model	model	NOUN
fcis-30136	22	4	achieved	achieve	VERB
fcis-30136	22	5	96.09	96.09	NUM
fcis-30136	22	6	%	%	NOUN
fcis-30136	22	7	accuracy	accuracy	NOUN
fcis-30136	22	8	on	on	ADP
fcis-30136	22	9	the	the	DET
fcis-30136	22	10	investigated	investigate	VERB
fcis-30136	22	11	dataset	dataset	NOUN
fcis-30136	22	12	[	[	X
fcis-30136	22	13	8	8	NUM
fcis-30136	22	14	]	]	PUNCT
fcis-30136	22	15	.	.	PUNCT
fcis-30136	23	1	zhang	zhang	PROPN
fcis-30136	23	2	w	w	PROPN
fcis-30136	23	3	et	et	PROPN
fcis-30136	23	4	al	al	PROPN
fcis-30136	23	5	.	.	PROPN
fcis-30136	23	6	proposed	propose	VERB
fcis-30136	23	7	a	a	DET
fcis-30136	23	8	seed	seed	NOUN
fcis-30136	23	9	potato	potato	NOUN
fcis-30136	23	10	bud	bud	NOUN
fcis-30136	23	11	recognition	recognition	NOUN
fcis-30136	23	12	method	method	NOUN
fcis-30136	23	13	based	base	VERB
fcis-30136	23	14	on	on	ADP
fcis-30136	23	15	an	an	DET
fcis-30136	23	16	improved	improved	ADJ
fcis-30136	23	17	yolov3	yolov3	NOUN
fcis-30136	23	18	-	-	PUNCT
fcis-30136	23	19	tiny	tiny	ADJ
fcis-30136	23	20	.	.	PUNCT
fcis-30136	24	1	by	by	ADP
fcis-30136	24	2	introducing	introduce	VERB
fcis-30136	24	3	the	the	DET
fcis-30136	24	4	ciou	ciou	NOUN
fcis-30136	24	5	bounding	bounding	NOUN
fcis-30136	24	6	box	box	NOUN
fcis-30136	24	7	regression	regression	NOUN
fcis-30136	24	8	loss	loss	NOUN
fcis-30136	24	9	function	function	NOUN
fcis-30136	24	10	to	to	PART
fcis-30136	24	11	enhance	enhance	VERB
fcis-30136	24	12	regression	regression	NOUN
fcis-30136	24	13	performance	performance	NOUN
fcis-30136	24	14	,	,	PUNCT
fcis-30136	24	15	the	the	DET
fcis-30136	24	16	model	model	NOUN
fcis-30136	24	17	achieved	achieve	VERB
fcis-30136	24	18	precision	precision	NOUN
fcis-30136	24	19	,	,	PUNCT
fcis-30136	24	20	recall	recall	NOUN
fcis-30136	24	21	,	,	PUNCT
fcis-30136	24	22	average	average	ADJ
fcis-30136	24	23	precision	precision	NOUN
fcis-30136	24	24	,	,	PUNCT
fcis-30136	24	25	and	and	CCONJ
fcis-30136	24	26	f1	f1	NOUN
fcis-30136	24	27	scores	score	NOUN
fcis-30136	24	28	of	of	ADP
fcis-30136	24	29	88.33	88.33	NUM
fcis-30136	24	30	%	%	NOUN
fcis-30136	24	31	,	,	PUNCT
fcis-30136	24	32	85.97	85.97	NUM
fcis-30136	24	33	%	%	NOUN
fcis-30136	24	34	,	,	PUNCT
fcis-30136	24	35	91.18	91.18	NUM
fcis-30136	24	36	%	%	NOUN
fcis-30136	24	37	,	,	PUNCT
fcis-30136	24	38	and	and	CCONJ
fcis-30136	24	39	87.13	87.13	NUM
fcis-30136	24	40	%	%	NOUN
fcis-30136	24	41	,	,	PUNCT
fcis-30136	24	42	respectively	respectively	ADV
fcis-30136	24	43	,	,	PUNCT
fcis-30136	24	44	for	for	ADP
fcis-30136	24	45	bud	bud	ADJ
fcis-30136	24	46	recognition	recognition	NOUN
fcis-30136	25	1	[	[	X
fcis-30136	25	2	9	9	NUM
fcis-30136	25	3	]	]	PUNCT
fcis-30136	25	4	.	.	PUNCT
fcis-30136	26	1	this	this	DET
fcis-30136	26	2	study	study	NOUN
fcis-30136	26	3	proposes	propose	VERB
fcis-30136	26	4	an	an	DET
fcis-30136	26	5	improved	improved	ADJ
fcis-30136	26	6	yolov8	yolov8	NOUN
fcis-30136	26	7	algorithm	algorithm	NOUN
fcis-30136	26	8	to	to	PART
fcis-30136	26	9	further	far	ADV
fcis-30136	26	10	enhance	enhance	VERB
fcis-30136	26	11	the	the	DET
fcis-30136	26	12	recognition	recognition	NOUN
fcis-30136	26	13	accuracy	accuracy	NOUN
fcis-30136	26	14	for	for	ADP
fcis-30136	26	15	different	different	ADJ
fcis-30136	26	16	potato	potato	NOUN
fcis-30136	26	17	varieties	variety	NOUN
fcis-30136	26	18	.	.	PUNCT
fcis-30136	27	1	the	the	DET
fcis-30136	27	2	algorithm	algorithm	NOUN
fcis-30136	27	3	primarily	primarily	ADV
fcis-30136	27	4	optimizes	optimize	VERB
fcis-30136	27	5	the	the	DET
fcis-30136	27	6	backbone	backbone	NOUN
fcis-30136	27	7	and	and	CCONJ
fcis-30136	27	8	neck	neck	NOUN
fcis-30136	27	9	modules	module	NOUN
fcis-30136	27	10	:	:	PUNCT
fcis-30136	27	11	dynamicconv	dynamicconv	NOUN
fcis-30136	27	12	is	be	AUX
fcis-30136	27	13	introduced	introduce	VERB
fcis-30136	27	14	into	into	ADP
fcis-30136	27	15	the	the	DET
fcis-30136	27	16	backbone	backbone	NOUN
fcis-30136	27	17	network	network	NOUN
fcis-30136	27	18	,	,	PUNCT
fcis-30136	27	19	replacing	replace	VERB
fcis-30136	27	20	traditional	traditional	ADJ
fcis-30136	27	21	static	static	ADJ
fcis-30136	27	22	convolution	convolution	NOUN
fcis-30136	27	23	with	with	ADP
fcis-30136	27	24	dynamic	dynamic	ADJ
fcis-30136	27	25	convolution	convolution	NOUN
fcis-30136	27	26	to	to	PART
fcis-30136	27	27	strengthen	strengthen	VERB
fcis-30136	27	28	the	the	DET
fcis-30136	27	29	model	model	NOUN
fcis-30136	27	30	's	's	PART
fcis-30136	27	31	expressive	expressive	ADJ
fcis-30136	27	32	capabilities	capability	NOUN
fcis-30136	27	33	;	;	PUNCT
fcis-30136	27	34	slim	slim	ADJ
fcis-30136	27	35	-	-	PUNCT
fcis-30136	27	36	neck	neck	NOUN
fcis-30136	27	37	is	be	AUX
fcis-30136	27	38	incorporated	incorporate	VERB
fcis-30136	27	39	into	into	ADP
fcis-30136	27	40	the	the	DET
fcis-30136	27	41	neck	neck	NOUN
fcis-30136	27	42	layer	layer	NOUN
fcis-30136	27	43	to	to	PART
fcis-30136	27	44	reduce	reduce	VERB
fcis-30136	27	45	model	model	NOUN
fcis-30136	27	46	parameters	parameter	NOUN
fcis-30136	27	47	and	and	CCONJ
fcis-30136	27	48	computational	computational	ADJ
fcis-30136	27	49	complexity	complexity	NOUN
fcis-30136	27	50	,	,	PUNCT
fcis-30136	27	51	broadening	broaden	VERB
fcis-30136	27	52	its	its	PRON
fcis-30136	27	53	applicability	applicability	NOUN
fcis-30136	27	54	;	;	PUNCT
fcis-30136	27	55	and	and	CCONJ
fcis-30136	27	56	the	the	DET
fcis-30136	27	57	ema	ema	PROPN
fcis-30136	27	58	attention	attention	NOUN
fcis-30136	27	59	mechanism	mechanism	NOUN
fcis-30136	27	60	is	be	AUX
fcis-30136	27	61	integrated	integrate	VERB
fcis-30136	27	62	to	to	PART
fcis-30136	27	63	improve	improve	VERB
fcis-30136	27	64	feature	feature	NOUN
fcis-30136	27	65	extraction	extraction	NOUN
fcis-30136	27	66	for	for	ADP
fcis-30136	27	67	diverse	diverse	ADJ
fcis-30136	27	68	characteristics	characteristic	NOUN
fcis-30136	27	69	.	.	PUNCT
fcis-30136	28	1	through	through	ADP
fcis-30136	28	2	these	these	DET
fcis-30136	28	3	enhancements	enhancement	NOUN
fcis-30136	28	4	,	,	PUNCT
fcis-30136	28	5	the	the	DET
fcis-30136	28	6	algorithm	algorithm	NOUN
fcis-30136	28	7	achieves	achieve	VERB
fcis-30136	28	8	reduced	reduce	VERB
fcis-30136	28	9	computational	computational	ADJ
fcis-30136	28	10	demands	demand	NOUN
fcis-30136	28	11	while	while	SCONJ
fcis-30136	28	12	significantly	significantly	ADV
fcis-30136	28	13	improving	improve	VERB
fcis-30136	28	14	potato	potato	NOUN
fcis-30136	28	15	recognition	recognition	NOUN
fcis-30136	28	16	accuracy	accuracy	NOUN
fcis-30136	28	17	.	.	PUNCT
fcis-30136	29	1	2	2	X
fcis-30136	29	2	.	.	X
fcis-30136	29	3	introduction	introduction	NOUN
fcis-30136	29	4	to	to	ADP
fcis-30136	29	5	yolov8	yolov8	PROPN
fcis-30136	29	6	yolov8	yolov8	PROPN
fcis-30136	29	7	was	be	AUX
fcis-30136	29	8	introduced	introduce	VERB
fcis-30136	29	9	by	by	ADP
fcis-30136	29	10	ultralytics	ultralytic	NOUN
fcis-30136	29	11	in	in	ADP
fcis-30136	29	12	early	early	ADJ
fcis-30136	29	13	2023	2023	NUM
fcis-30136	29	14	and	and	CCONJ
fcis-30136	29	15	can	can	AUX
fcis-30136	29	16	perform	perform	VERB
fcis-30136	29	17	object	object	NOUN
fcis-30136	29	18	detection	detection	NOUN
fcis-30136	29	19	,	,	PUNCT
fcis-30136	29	20	segmentation	segmentation	NOUN
fcis-30136	29	21	,	,	PUNCT
fcis-30136	29	22	and	and	CCONJ
fcis-30136	29	23	classification	classification	NOUN
fcis-30136	29	24	tasks	task	NOUN
fcis-30136	29	25	by	by	ADP
fcis-30136	29	26	selecting	select	VERB
fcis-30136	29	27	appropriate	appropriate	ADJ
fcis-30136	29	28	weight	weight	NOUN
fcis-30136	29	29	files	file	NOUN
fcis-30136	29	30	for	for	ADP
fcis-30136	29	31	different	different	ADJ
fcis-30136	29	32	scenarios	scenario	NOUN
fcis-30136	29	33	.	.	PUNCT
fcis-30136	30	1	the	the	DET
fcis-30136	30	2	network	network	NOUN
fcis-30136	30	3	architecture	architecture	NOUN
fcis-30136	30	4	of	of	ADP
fcis-30136	30	5	yolov8	yolov8	NOUN
fcis-30136	30	6	,	,	PUNCT
fcis-30136	30	7	as	as	SCONJ
fcis-30136	30	8	shown	show	VERB
fcis-30136	30	9	in	in	ADP
fcis-30136	30	10	figure	figure	NOUN
fcis-30136	30	11	1	1	NUM
fcis-30136	30	12	,	,	PUNCT
fcis-30136	30	13	consists	consist	VERB
fcis-30136	30	14	of	of	ADP
fcis-30136	30	15	five	five	NUM
fcis-30136	30	16	main	main	ADJ
fcis-30136	30	17	components	component	NOUN
fcis-30136	30	18	:	:	PUNCT
fcis-30136	30	19	input	input	NOUN
fcis-30136	30	20	,	,	PUNCT
fcis-30136	30	21	backbone	backbone	NOUN
fcis-30136	30	22	,	,	PUNCT
fcis-30136	30	23	neck	neck	NOUN
fcis-30136	30	24	(	(	PUNCT
fcis-30136	30	25	feature	feature	NOUN
fcis-30136	30	26	fusion	fusion	NOUN
fcis-30136	30	27	network	network	NOUN
fcis-30136	30	28	)	)	PUNCT
fcis-30136	30	29	,	,	PUNCT
fcis-30136	30	30	head	head	NOUN
fcis-30136	30	31	(	(	PUNCT
fcis-30136	30	32	detection	detection	NOUN
fcis-30136	30	33	head	head	NOUN
fcis-30136	30	34	)	)	PUNCT
fcis-30136	30	35	,	,	PUNCT
fcis-30136	30	36	and	and	CCONJ
fcis-30136	30	37	output	output	NOUN
fcis-30136	30	38	.	.	PUNCT
fcis-30136	31	1	the	the	DET
fcis-30136	31	2	backbone	backbone	NOUN
fcis-30136	31	3	of	of	ADP
fcis-30136	31	4	yolov8	yolov8	PROPN
fcis-30136	31	5	adopts	adopt	VERB
fcis-30136	31	6	the	the	DET
fcis-30136	31	7	elan	elan	PROPN
fcis-30136	31	8	concept	concept	NOUN
fcis-30136	31	9	from	from	ADP
fcis-30136	31	10	yolov7	yolov7	NOUN
fcis-30136	32	1	[	[	X
fcis-30136	32	2	10	10	NUM
fcis-30136	32	3	]	]	PUNCT
fcis-30136	32	4	,	,	PUNCT
fcis-30136	32	5	serving	serve	VERB
fcis-30136	32	6	to	to	PART
fcis-30136	32	7	provide	provide	VERB
fcis-30136	32	8	computational	computational	ADJ
fcis-30136	32	9	power	power	NOUN
fcis-30136	32	10	and	and	CCONJ
fcis-30136	32	11	feature	feature	NOUN
fcis-30136	32	12	extraction	extraction	NOUN
fcis-30136	32	13	.	.	PUNCT
fcis-30136	33	1	it	it	PRON
fcis-30136	33	2	comprises	comprise	VERB
fcis-30136	33	3	cbs	cbs	PROPN
fcis-30136	33	4	,	,	PUNCT
fcis-30136	33	5	c2f	c2f	PROPN
fcis-30136	33	6	,	,	PUNCT
fcis-30136	33	7	and	and	CCONJ
fcis-30136	33	8	sppf	sppf	ADJ
fcis-30136	33	9	modules	module	NOUN
fcis-30136	33	10	.	.	PUNCT
fcis-30136	34	1	the	the	DET
fcis-30136	34	2	c2f	c2f	NOUN
fcis-30136	34	3	module	module	NOUN
fcis-30136	34	4	enhances	enhance	VERB
fcis-30136	34	5	gradient	gradient	ADJ
fcis-30136	34	6	flow	flow	NOUN
fcis-30136	34	7	diversity	diversity	NOUN
fcis-30136	34	8	through	through	ADP
fcis-30136	34	9	residual	residual	ADJ
fcis-30136	34	10	connections	connection	NOUN
fcis-30136	34	11	,	,	PUNCT
fcis-30136	34	12	optimizing	optimize	VERB
fcis-30136	34	13	network	network	NOUN
fcis-30136	34	14	training	training	NOUN
fcis-30136	34	15	.	.	PUNCT
fcis-30136	35	1	for	for	ADP
fcis-30136	35	2	the	the	DET
fcis-30136	35	3	loss	loss	NOUN
fcis-30136	35	4	function	function	NOUN
fcis-30136	35	5	,	,	PUNCT
fcis-30136	35	6	yolov8	yolov8	NOUN
fcis-30136	35	7	replaces	replace	VERB
fcis-30136	35	8	traditional	traditional	ADJ
fcis-30136	35	9	iou	iou	NOUN
fcis-30136	35	10	matching	matching	NOUN
fcis-30136	35	11	or	or	CCONJ
fcis-30136	35	12	unilateral	unilateral	ADJ
fcis-30136	35	13	ratio	ratio	NOUN
fcis-30136	35	14	allocation	allocation	NOUN
fcis-30136	35	15	with	with	ADP
fcis-30136	35	16	the	the	DET
fcis-30136	35	17	taskalignedassigner	taskalignedassigner	NOUN
fcis-30136	35	18	positive	positive	ADJ
fcis-30136	35	19	sample	sample	NOUN
fcis-30136	35	20	allocation	allocation	NOUN
fcis-30136	35	21	strategy	strategy	NOUN
fcis-30136	35	22	—	—	PUNCT
fcis-30136	35	23	a	a	DET
fcis-30136	35	24	dynamic	dynamic	ADJ
fcis-30136	35	25	approach	approach	NOUN
fcis-30136	35	26	that	that	PRON
fcis-30136	35	27	selects	select	VERB
fcis-30136	35	28	positive	positive	ADJ
fcis-30136	35	29	samples	sample	NOUN
fcis-30136	35	30	based	base	VERB
fcis-30136	35	31	on	on	ADP
fcis-30136	35	32	weighted	weight	VERB
fcis-30136	35	33	scores	score	NOUN
fcis-30136	35	34	combining	combine	VERB
fcis-30136	35	35	classification	classification	NOUN
fcis-30136	35	36	and	and	CCONJ
fcis-30136	35	37	regression	regression	NOUN
fcis-30136	35	38	metrics	metric	NOUN
fcis-30136	35	39	.	.	PUNCT
fcis-30136	36	1	86	86	NUM
fcis-30136	36	2	figure	figure	NOUN
fcis-30136	36	3	1	1	NUM
fcis-30136	36	4	.	.	PUNCT
fcis-30136	37	1	yolov8	yolov8	NOUN
fcis-30136	37	2	network	network	NOUN
fcis-30136	37	3	structure	structure	NOUN
fcis-30136	37	4	3	3	NUM
fcis-30136	37	5	.	.	PUNCT
fcis-30136	37	6	improved	improve	VERB
fcis-30136	37	7	yolov8	yolov8	NOUN
fcis-30136	37	8	3.1	3.1	NUM
fcis-30136	37	9	.	.	PUNCT
fcis-30136	38	1	dynamicconv	dynamicconv	ADJ
fcis-30136	38	2	traditional	traditional	ADJ
fcis-30136	38	3	3×3	3×3	NUM
fcis-30136	38	4	convolution	convolution	NOUN
fcis-30136	38	5	kernels	kernel	NOUN
fcis-30136	38	6	have	have	AUX
fcis-30136	38	7	fixed	fix	VERB
fcis-30136	38	8	shapes	shape	NOUN
fcis-30136	38	9	and	and	CCONJ
fcis-30136	38	10	sizes	size	NOUN
fcis-30136	38	11	,	,	PUNCT
fcis-30136	38	12	which	which	PRON
fcis-30136	38	13	suffer	suffer	VERB
fcis-30136	38	14	from	from	ADP
fcis-30136	38	15	drawbacks	drawback	NOUN
fcis-30136	38	16	such	such	ADJ
fcis-30136	38	17	as	as	ADP
fcis-30136	38	18	large	large	ADJ
fcis-30136	38	19	parameter	parameter	NOUN
fcis-30136	38	20	counts	count	NOUN
fcis-30136	38	21	,	,	PUNCT
fcis-30136	38	22	poor	poor	ADJ
fcis-30136	38	23	scale	scale	NOUN
fcis-30136	38	24	invariance	invariance	NOUN
fcis-30136	38	25	,	,	PUNCT
fcis-30136	38	26	and	and	CCONJ
fcis-30136	38	27	limited	limited	ADJ
fcis-30136	38	28	local	local	ADJ
fcis-30136	38	29	receptive	receptive	ADJ
fcis-30136	38	30	fields	field	NOUN
fcis-30136	38	31	during	during	ADP
fcis-30136	38	32	convolution	convolution	NOUN
fcis-30136	38	33	operations	operation	NOUN
fcis-30136	38	34	[	[	X
fcis-30136	38	35	11	11	NUM
fcis-30136	38	36	]	]	PUNCT
fcis-30136	38	37	.	.	PUNCT
fcis-30136	39	1	to	to	PART
fcis-30136	39	2	address	address	VERB
fcis-30136	39	3	these	these	DET
fcis-30136	39	4	limitations	limitation	NOUN
fcis-30136	39	5	,	,	PUNCT
fcis-30136	39	6	the	the	DET
fcis-30136	39	7	dynamicconv	dynamicconv	ADJ
fcis-30136	39	8	module	module	NOUN
fcis-30136	39	9	is	be	AUX
fcis-30136	39	10	constructed	construct	VERB
fcis-30136	39	11	by	by	ADP
fcis-30136	39	12	integrating	integrate	VERB
fcis-30136	39	13	dynamic	dynamic	ADJ
fcis-30136	39	14	convolution	convolution	NOUN
fcis-30136	39	15	(	(	PUNCT
fcis-30136	39	16	dynamicconv	dynamicconv	NOUN
fcis-30136	39	17	)	)	PUNCT
fcis-30136	39	18	into	into	ADP
fcis-30136	39	19	the	the	DET
fcis-30136	39	20	backbone	backbone	NOUN
fcis-30136	39	21	network	network	NOUN
fcis-30136	39	22	.	.	PUNCT
fcis-30136	40	1	dynamicconv	dynamicconv	ADJ
fcis-30136	40	2	dynamically	dynamically	ADJ
fcis-30136	40	3	selects	select	NOUN
fcis-30136	40	4	or	or	CCONJ
fcis-30136	40	5	combines	combine	VERB
fcis-30136	40	6	different	different	ADJ
fcis-30136	40	7	convolution	convolution	NOUN
fcis-30136	40	8	kernels	kernel	NOUN
fcis-30136	40	9	based	base	VERB
fcis-30136	40	10	on	on	ADP
fcis-30136	40	11	input	input	NOUN
fcis-30136	40	12	samples	sample	NOUN
fcis-30136	40	13	,	,	PUNCT
fcis-30136	40	14	enabling	enable	VERB
fcis-30136	40	15	the	the	DET
fcis-30136	40	16	network	network	NOUN
fcis-30136	40	17	structure	structure	NOUN
fcis-30136	40	18	to	to	PART
fcis-30136	40	19	adaptively	adaptively	ADV
fcis-30136	40	20	adjust	adjust	VERB
fcis-30136	40	21	its	its	PRON
fcis-30136	40	22	parameters	parameter	NOUN
fcis-30136	40	23	according	accord	VERB
fcis-30136	40	24	to	to	ADP
fcis-30136	40	25	varying	vary	VERB
fcis-30136	40	26	input	input	NOUN
fcis-30136	40	27	data	datum	NOUN
fcis-30136	40	28	.	.	PUNCT
fcis-30136	41	1	by	by	ADP
fcis-30136	41	2	employing	employ	VERB
fcis-30136	41	3	multiple	multiple	ADJ
fcis-30136	41	4	convolution	convolution	NOUN
fcis-30136	41	5	kernels	kernel	NOUN
fcis-30136	41	6	within	within	ADP
fcis-30136	41	7	a	a	DET
fcis-30136	41	8	single	single	ADJ
fcis-30136	41	9	convolutional	convolutional	ADJ
fcis-30136	41	10	layer	layer	NOUN
fcis-30136	41	11	and	and	CCONJ
fcis-30136	41	12	integrating	integrate	VERB
fcis-30136	41	13	attention	attention	NOUN
fcis-30136	41	14	mechanisms	mechanism	NOUN
fcis-30136	41	15	to	to	PART
fcis-30136	41	16	fuse	fuse	VERB
fcis-30136	41	17	information	information	NOUN
fcis-30136	41	18	from	from	ADP
fcis-30136	41	19	different	different	ADJ
fcis-30136	41	20	kernels	kernel	NOUN
fcis-30136	41	21	,	,	PUNCT
fcis-30136	41	22	dynamicconv	dynamicconv	NOUN
fcis-30136	41	23	enriches	enrich	VERB
fcis-30136	41	24	feature	feature	NOUN
fcis-30136	41	25	extraction	extraction	NOUN
fcis-30136	41	26	without	without	ADP
fcis-30136	41	27	significantly	significantly	ADV
fcis-30136	41	28	increasing	increase	VERB
fcis-30136	41	29	computational	computational	ADJ
fcis-30136	41	30	complexity	complexity	NOUN
fcis-30136	41	31	.	.	PUNCT
fcis-30136	42	1	as	as	SCONJ
fcis-30136	42	2	shown	show	VERB
fcis-30136	42	3	in	in	ADP
fcis-30136	42	4	figure	figure	NOUN
fcis-30136	42	5	2	2	NUM
fcis-30136	42	6	,	,	PUNCT
fcis-30136	42	7	dynamicconv	dynamicconv	ADJ
fcis-30136	42	8	sets	set	NOUN
fcis-30136	42	9	k	k	PROPN
fcis-30136	42	10	convolution	convolution	NOUN
fcis-30136	42	11	kernels	kernel	NOUN
fcis-30136	42	12	with	with	ADP
fcis-30136	42	13	the	the	DET
fcis-30136	42	14	same	same	ADJ
fcis-30136	42	15	scale	scale	NOUN
fcis-30136	42	16	and	and	CCONJ
fcis-30136	42	17	input	input	NOUN
fcis-30136	42	18	data	datum	NOUN
fcis-30136	42	19	channel	channel	NOUN
fcis-30136	42	20	dimensions	dimension	NOUN
fcis-30136	42	21	at	at	ADP
fcis-30136	42	22	specific	specific	ADJ
fcis-30136	42	23	layers	layer	NOUN
fcis-30136	42	24	.	.	PUNCT
fcis-30136	43	1	these	these	DET
fcis-30136	43	2	kernels	kernel	NOUN
fcis-30136	43	3	are	be	AUX
fcis-30136	43	4	fused	fuse	VERB
fcis-30136	43	5	through	through	ADP
fcis-30136	43	6	their	their	PRON
fcis-30136	43	7	respective	respective	ADJ
fcis-30136	43	8	attention	attention	NOUN
fcis-30136	43	9	weights	weight	NOUN
fcis-30136	43	10	π_k	π_k	PART
fcis-30136	43	11	to	to	PART
fcis-30136	43	12	obtain	obtain	VERB
fcis-30136	43	13	the	the	DET
fcis-30136	43	14	convolution	convolution	NOUN
fcis-30136	43	15	kernel	kernel	NOUN
fcis-30136	43	16	parameters	parameter	NOUN
fcis-30136	43	17	for	for	ADP
fcis-30136	43	18	that	that	DET
fcis-30136	43	19	layer	layer	NOUN
fcis-30136	43	20	.	.	PUNCT
fcis-30136	44	1	the	the	DET
fcis-30136	44	2	π_k	π_k	PUNCT
fcis-30136	44	3	is	be	AUX
fcis-30136	44	4	obtained	obtain	VERB
fcis-30136	44	5	through	through	ADP
fcis-30136	44	6	the	the	DET
fcis-30136	44	7	attention	attention	NOUN
fcis-30136	44	8	mechanism	mechanism	NOUN
fcis-30136	44	9	based	base	VERB
fcis-30136	44	10	on	on	ADP
fcis-30136	44	11	the	the	DET
fcis-30136	44	12	input	input	NOUN
fcis-30136	44	13	data	datum	NOUN
fcis-30136	44	14	x.	x.	NOUN
fcis-30136	44	15	first	first	ADV
fcis-30136	44	16	,	,	PUNCT
fcis-30136	44	17	global	global	ADJ
fcis-30136	44	18	average	average	ADJ
fcis-30136	44	19	pooling	pooling	NOUN
fcis-30136	44	20	(	(	PUNCT
fcis-30136	44	21	avg_pool	avg_pool	NOUN
fcis-30136	44	22	)	)	PUNCT
fcis-30136	44	23	is	be	AUX
fcis-30136	44	24	performed	perform	VERB
fcis-30136	44	25	to	to	PART
fcis-30136	44	26	obtain	obtain	VERB
fcis-30136	44	27	global	global	ADJ
fcis-30136	44	28	segmentation	segmentation	NOUN
fcis-30136	44	29	features	feature	NOUN
fcis-30136	44	30	.	.	PUNCT
fcis-30136	45	1	then	then	ADV
fcis-30136	45	2	,	,	PUNCT
fcis-30136	45	3	two	two	NUM
fcis-30136	45	4	fc	fc	PROPN
fcis-30136	45	5	layers	layer	NOUN
fcis-30136	45	6	map	map	VERB
fcis-30136	45	7	these	these	DET
fcis-30136	45	8	features	feature	NOUN
fcis-30136	45	9	to	to	ADP
fcis-30136	45	10	k	k	ADJ
fcis-30136	45	11	-	-	ADJ
fcis-30136	45	12	dimensional	dimensional	ADJ
fcis-30136	45	13	space	space	NOUN
fcis-30136	45	14	,	,	PUNCT
fcis-30136	45	15	followed	follow	VERB
fcis-30136	45	16	by	by	ADP
fcis-30136	45	17	softmax	softmax	NOUN
fcis-30136	45	18	normalization	normalization	NOUN
fcis-30136	45	19	to	to	PART
fcis-30136	45	20	allocate	allocate	VERB
fcis-30136	45	21	the	the	DET
fcis-30136	45	22	obtained	obtain	VERB
fcis-30136	45	23	k	k	PROPN
fcis-30136	45	24	attention	attention	NOUN
fcis-30136	45	25	weights	weight	NOUN
fcis-30136	45	26	to	to	ADP
fcis-30136	45	27	the	the	DET
fcis-30136	45	28	k	k	PROPN
fcis-30136	45	29	convolution	convolution	NOUN
fcis-30136	45	30	kernels	kernel	NOUN
fcis-30136	45	31	of	of	ADP
fcis-30136	45	32	this	this	DET
fcis-30136	45	33	layer	layer	NOUN
fcis-30136	45	34	.	.	PUNCT
fcis-30136	46	1	by	by	ADP
fcis-30136	46	2	multiplying	multiply	VERB
fcis-30136	46	3	the	the	DET
fcis-30136	46	4	obtained	obtain	VERB
fcis-30136	46	5	weight	weight	NOUN
fcis-30136	46	6	coefficients	coefficient	NOUN
fcis-30136	46	7	with	with	ADP
fcis-30136	46	8	their	their	PRON
fcis-30136	46	9	corresponding	correspond	VERB
fcis-30136	46	10	biases	bias	NOUN
fcis-30136	46	11	,	,	PUNCT
fcis-30136	46	12	the	the	DET
fcis-30136	46	13	data	data	NOUN
fcis-30136	46	14	information	information	NOUN
fcis-30136	46	15	at	at	ADP
fcis-30136	46	16	that	that	DET
fcis-30136	46	17	convolution	convolution	NOUN
fcis-30136	46	18	position	position	NOUN
fcis-30136	46	19	can	can	AUX
fcis-30136	46	20	be	be	AUX
fcis-30136	46	21	obtained	obtain	VERB
fcis-30136	46	22	.	.	PUNCT
fcis-30136	47	1	the	the	DET
fcis-30136	47	2	summation	summation	NOUN
fcis-30136	47	3	of	of	ADP
fcis-30136	47	4	all	all	DET
fcis-30136	47	5	convolution	convolution	NOUN
fcis-30136	47	6	weights	weight	NOUN
fcis-30136	47	7	is	be	AUX
fcis-30136	47	8	then	then	ADV
fcis-30136	47	9	computed	compute	VERB
fcis-30136	47	10	,	,	PUNCT
fcis-30136	47	11	and	and	CCONJ
fcis-30136	47	12	the	the	DET
fcis-30136	47	13	matrix	matrix	NOUN
fcis-30136	47	14	multiplication	multiplication	NOUN
fcis-30136	47	15	of	of	ADP
fcis-30136	47	16	this	this	DET
fcis-30136	47	17	summation	summation	NOUN
fcis-30136	47	18	result	result	NOUN
fcis-30136	47	19	with	with	ADP
fcis-30136	47	20	x	x	PUNCT
fcis-30136	47	21	completes	complete	VERB
fcis-30136	47	22	the	the	DET
fcis-30136	47	23	dynamicconv	dynamicconv	ADJ
fcis-30136	47	24	operation	operation	NOUN
fcis-30136	47	25	.	.	PUNCT
fcis-30136	48	1	the	the	DET
fcis-30136	48	2	dynamic	dynamic	ADJ
fcis-30136	48	3	perception	perception	NOUN
fcis-30136	48	4	model	model	NOUN
fcis-30136	48	5	of	of	ADP
fcis-30136	48	6	dynamicconv	dynamicconv	NOUN
fcis-30136	48	7	is	be	AUX
fcis-30136	48	8	shown	show	VERB
fcis-30136	48	9	in	in	ADP
fcis-30136	48	10	equation	equation	NOUN
fcis-30136	48	11	(	(	PUNCT
fcis-30136	48	12	1	1	NUM
fcis-30136	48	13	)	)	PUNCT
fcis-30136	48	14	.	.	PUNCT
fcis-30136	49	1	，	，	PUNCT
fcis-30136	49	2	1	1	NUM
fcis-30136	49	3	the	the	DET
fcis-30136	49	4	weights	weight	NOUN
fcis-30136	49	5	and	and	CCONJ
fcis-30136	49	6	biases	bias	NOUN
fcis-30136	49	7	required	require	VERB
fcis-30136	49	8	by	by	ADP
fcis-30136	49	9	the	the	DET
fcis-30136	49	10	dynamic	dynamic	ADJ
fcis-30136	49	11	perception	perception	NOUN
fcis-30136	49	12	model	model	NOUN
fcis-30136	49	13	can	can	AUX
fcis-30136	49	14	be	be	AUX
fcis-30136	49	15	derived	derive	VERB
fcis-30136	49	16	from	from	ADP
fcis-30136	49	17	equations	equation	NOUN
fcis-30136	49	18	(	(	PUNCT
fcis-30136	49	19	2	2	NUM
fcis-30136	49	20	)	)	PUNCT
fcis-30136	49	21	and	and	CCONJ
fcis-30136	49	22	(	(	PUNCT
fcis-30136	49	23	3	3	NUM
fcis-30136	49	24	)	)	PUNCT
fcis-30136	49	25	.	.	PUNCT
fcis-30136	49	26	  	  	SPACE
fcis-30136	50	1	，	，	PUNCT
fcis-30136	50	2	2	2	NUM
fcis-30136	50	3	  	  	SPACE
fcis-30136	50	4	，	，	PROPN
fcis-30136	50	5	3	3	NUM
fcis-30136	50	6	.	.	PUNCT
fcis-30136	50	7	.	.	PUNCT
fcis-30136	51	1	0	0	NUM
fcis-30136	52	1	1	1	NUM
fcis-30136	52	2	,	,	PUNCT
fcis-30136	52	3	  	  	SPACE
fcis-30136	52	4	1	1	NUM
fcis-30136	52	5	，	，	PROPN
fcis-30136	52	6	4	4	NUM
fcis-30136	52	7	equation	equation	NOUN
fcis-30136	52	8	(	(	PUNCT
fcis-30136	52	9	4	4	NUM
fcis-30136	52	10	)	)	PUNCT
fcis-30136	52	11	represents	represent	VERB
fcis-30136	52	12	the	the	DET
fcis-30136	52	13	constraint	constraint	NOUN
fcis-30136	52	14	on	on	ADP
fcis-30136	52	15	the	the	DET
fcis-30136	52	16	weight	weight	NOUN
fcis-30136	52	17	coefficients	coefficient	NOUN
fcis-30136	52	18	.	.	PUNCT
fcis-30136	53	1	the	the	DET
fcis-30136	53	2	weight	weight	NOUN
fcis-30136	53	3	coefficients	coefficient	NOUN
fcis-30136	53	4	are	be	AUX
fcis-30136	53	5	not	not	PART
fcis-30136	53	6	fixed	fix	VERB
fcis-30136	53	7	and	and	CCONJ
fcis-30136	53	8	can	can	AUX
fcis-30136	53	9	be	be	AUX
fcis-30136	53	10	adaptively	adaptively	ADV
fcis-30136	53	11	selected	select	VERB
fcis-30136	53	12	based	base	VERB
fcis-30136	53	13	on	on	ADP
fcis-30136	53	14	the	the	DET
fcis-30136	53	15	input	input	NOUN
fcis-30136	53	16	data	datum	NOUN
fcis-30136	53	17	,	,	PUNCT
fcis-30136	53	18	thereby	thereby	ADV
fcis-30136	53	19	enabling	enable	VERB
fcis-30136	53	20	the	the	DET
fcis-30136	53	21	model	model	NOUN
fcis-30136	53	22	to	to	PART
fcis-30136	53	23	possess	possess	VERB
fcis-30136	53	24	stronger	strong	ADJ
fcis-30136	53	25	feature	feature	NOUN
fcis-30136	53	26	representation	representation	NOUN
fcis-30136	53	27	capabilities	capability	NOUN
fcis-30136	53	28	.	.	PUNCT
fcis-30136	54	1	figure	figure	NOUN
fcis-30136	54	2	2	2	NUM
fcis-30136	54	3	.	.	PUNCT
fcis-30136	54	4	dynamicconv	dynamicconv	ADJ
fcis-30136	54	5	structure	structure	NOUN
fcis-30136	54	6	87	87	NUM
fcis-30136	54	7	3.2	3.2	NUM
fcis-30136	54	8	.	.	PUNCT
fcis-30136	55	1	sac	sac	NOUN
fcis-30136	55	2	in	in	ADP
fcis-30136	55	3	current	current	ADJ
fcis-30136	55	4	object	object	NOUN
fcis-30136	55	5	detection	detection	NOUN
fcis-30136	55	6	research	research	NOUN
fcis-30136	55	7	,	,	PUNCT
fcis-30136	55	8	it	it	PRON
fcis-30136	55	9	has	have	AUX
fcis-30136	55	10	been	be	AUX
fcis-30136	55	11	demonstrated	demonstrate	VERB
fcis-30136	55	12	that	that	SCONJ
fcis-30136	55	13	mechanisms	mechanism	NOUN
fcis-30136	55	14	employing	employ	VERB
fcis-30136	55	15	dual	dual	ADJ
fcis-30136	55	16	observation	observation	NOUN
fcis-30136	55	17	and	and	CCONJ
fcis-30136	55	18	reasoning	reasoning	NOUN
fcis-30136	55	19	can	can	AUX
fcis-30136	55	20	exhibit	exhibit	VERB
fcis-30136	55	21	superior	superior	ADJ
fcis-30136	55	22	performance	performance	NOUN
fcis-30136	55	23	.	.	PUNCT
fcis-30136	56	1	sac	sac	NOUN
fcis-30136	56	2	(	(	PUNCT
fcis-30136	56	3	switchable	switchable	ADJ
fcis-30136	56	4	atrous	atrous	ADJ
fcis-30136	56	5	convolution	convolution	NOUN
fcis-30136	56	6	)	)	PUNCT
fcis-30136	56	7	integrates	integrate	VERB
fcis-30136	56	8	this	this	DET
fcis-30136	56	9	dual	dual	ADJ
fcis-30136	56	10	observation	observation	NOUN
fcis-30136	56	11	and	and	CCONJ
fcis-30136	56	12	reasoning	reasoning	NOUN
fcis-30136	56	13	mechanism	mechanism	NOUN
fcis-30136	56	14	into	into	ADP
fcis-30136	56	15	the	the	DET
fcis-30136	56	16	backbone	backbone	NOUN
fcis-30136	56	17	network	network	NOUN
fcis-30136	56	18	of	of	ADP
fcis-30136	56	19	object	object	NOUN
fcis-30136	56	20	detection	detection	NOUN
fcis-30136	56	21	systems	system	NOUN
fcis-30136	56	22	,	,	PUNCT
fcis-30136	56	23	deploying	deploy	VERB
fcis-30136	56	24	this	this	DET
fcis-30136	56	25	mechanism	mechanism	NOUN
fcis-30136	56	26	at	at	ADP
fcis-30136	56	27	both	both	CCONJ
fcis-30136	56	28	macro	macro	ADJ
fcis-30136	56	29	and	and	CCONJ
fcis-30136	56	30	micro	micro	ADJ
fcis-30136	56	31	levels	level	NOUN
fcis-30136	56	32	.	.	PUNCT
fcis-30136	57	1	this	this	DET
fcis-30136	57	2	approach	approach	NOUN
fcis-30136	57	3	significantly	significantly	ADV
fcis-30136	57	4	enhances	enhance	VERB
fcis-30136	57	5	detector	detector	NOUN
fcis-30136	57	6	performance	performance	NOUN
fcis-30136	57	7	without	without	ADP
fcis-30136	57	8	substantially	substantially	ADV
fcis-30136	57	9	increasing	increase	VERB
fcis-30136	57	10	model	model	NOUN
fcis-30136	57	11	parameters	parameter	NOUN
fcis-30136	57	12	or	or	CCONJ
fcis-30136	57	13	weight	weight	NOUN
fcis-30136	57	14	size	size	NOUN
fcis-30136	57	15	,	,	PUNCT
fcis-30136	57	16	thereby	thereby	ADV
fcis-30136	57	17	maintaining	maintain	VERB
fcis-30136	57	18	inference	inference	NOUN
fcis-30136	57	19	speed	speed	NOUN
fcis-30136	57	20	.	.	PUNCT
fcis-30136	58	1	the	the	DET
fcis-30136	58	2	core	core	NOUN
fcis-30136	58	3	concept	concept	NOUN
fcis-30136	58	4	of	of	ADP
fcis-30136	58	5	sac	sac	NOUN
fcis-30136	58	6	lies	lie	VERB
fcis-30136	58	7	in	in	ADP
fcis-30136	58	8	applying	apply	VERB
fcis-30136	58	9	convolution	convolution	NOUN
fcis-30136	58	10	operations	operation	NOUN
fcis-30136	58	11	with	with	ADP
fcis-30136	58	12	different	different	ADJ
fcis-30136	58	13	dilation	dilation	NOUN
fcis-30136	58	14	rates	rate	NOUN
fcis-30136	58	15	to	to	ADP
fcis-30136	58	16	the	the	DET
fcis-30136	58	17	same	same	ADJ
fcis-30136	58	18	input	input	NOUN
fcis-30136	58	19	features	feature	NOUN
fcis-30136	58	20	.	.	PUNCT
fcis-30136	59	1	dilated	dilated	ADJ
fcis-30136	59	2	convolution	convolution	NOUN
fcis-30136	59	3	expands	expand	VERB
fcis-30136	59	4	the	the	DET
fcis-30136	59	5	receptive	receptive	ADJ
fcis-30136	59	6	field	field	NOUN
fcis-30136	59	7	by	by	ADP
fcis-30136	59	8	introducing	introduce	VERB
fcis-30136	59	9	additional	additional	ADJ
fcis-30136	59	10	spaces	space	NOUN
fcis-30136	59	11	(	(	PUNCT
fcis-30136	59	12	dilation	dilation	NOUN
fcis-30136	59	13	)	)	PUNCT
fcis-30136	59	14	in	in	ADP
fcis-30136	59	15	the	the	DET
fcis-30136	59	16	convolution	convolution	NOUN
fcis-30136	59	17	kernel	kernel	NOUN
fcis-30136	59	18	without	without	ADP
fcis-30136	59	19	increasing	increase	VERB
fcis-30136	59	20	parameters	parameter	NOUN
fcis-30136	59	21	or	or	CCONJ
fcis-30136	59	22	computational	computational	ADJ
fcis-30136	59	23	load	load	NOUN
fcis-30136	59	24	.	.	PUNCT
fcis-30136	60	1	sac	sac	NOUN
fcis-30136	60	2	leverages	leverage	NOUN
fcis-30136	60	3	this	this	DET
fcis-30136	60	4	characteristic	characteristic	NOUN
fcis-30136	60	5	to	to	PART
fcis-30136	60	6	capture	capture	VERB
fcis-30136	60	7	multi	multi	ADJ
fcis-30136	60	8	-	-	ADJ
fcis-30136	60	9	scale	scale	ADJ
fcis-30136	60	10	features	feature	NOUN
fcis-30136	60	11	.	.	PUNCT
fcis-30136	61	1	by	by	ADP
fcis-30136	61	2	utilizing	utilize	VERB
fcis-30136	61	3	identical	identical	ADJ
fcis-30136	61	4	weights	weight	NOUN
fcis-30136	61	5	for	for	ADP
fcis-30136	61	6	convolutions	convolution	NOUN
fcis-30136	61	7	with	with	ADP
fcis-30136	61	8	different	different	ADJ
fcis-30136	61	9	dilation	dilation	NOUN
fcis-30136	61	10	rates	rate	NOUN
fcis-30136	61	11	,	,	PUNCT
fcis-30136	61	12	sac	sac	NOUN
fcis-30136	61	13	can	can	AUX
fcis-30136	61	14	convert	convert	VERB
fcis-30136	61	15	traditional	traditional	ADJ
fcis-30136	61	16	convolution	convolution	NOUN
fcis-30136	61	17	layers	layer	NOUN
fcis-30136	61	18	into	into	ADP
fcis-30136	61	19	sac	sac	NOUN
fcis-30136	61	20	layers	layer	NOUN
fcis-30136	61	21	,	,	PUNCT
fcis-30136	61	22	enabling	enable	VERB
fcis-30136	61	23	soft	soft	ADJ
fcis-30136	61	24	-	-	PUNCT
fcis-30136	61	25	switching	switching	NOUN
fcis-30136	61	26	between	between	ADP
fcis-30136	61	27	different	different	ADJ
fcis-30136	61	28	dilation	dilation	NOUN
fcis-30136	61	29	rates	rate	NOUN
fcis-30136	61	30	during	during	ADP
fcis-30136	61	31	convolution	convolution	NOUN
fcis-30136	61	32	operations	operation	NOUN
fcis-30136	61	33	.	.	PUNCT
fcis-30136	62	1	although	although	SCONJ
fcis-30136	62	2	sac	sac	NOUN
fcis-30136	62	3	switches	switch	NOUN
fcis-30136	62	4	between	between	ADP
fcis-30136	62	5	different	different	ADJ
fcis-30136	62	6	dilation	dilation	NOUN
fcis-30136	62	7	rates	rate	NOUN
fcis-30136	62	8	,	,	PUNCT
fcis-30136	62	9	all	all	DET
fcis-30136	62	10	operations	operation	NOUN
fcis-30136	62	11	share	share	VERB
fcis-30136	62	12	the	the	DET
fcis-30136	62	13	same	same	ADJ
fcis-30136	62	14	weights	weight	NOUN
fcis-30136	62	15	with	with	ADP
fcis-30136	62	16	only	only	ADV
fcis-30136	62	17	one	one	NUM
fcis-30136	62	18	trainable	trainable	ADJ
fcis-30136	62	19	variation	variation	NOUN
fcis-30136	62	20	.	.	PUNCT
fcis-30136	63	1	this	this	DET
fcis-30136	63	2	design	design	NOUN
fcis-30136	63	3	reduces	reduce	VERB
fcis-30136	63	4	model	model	NOUN
fcis-30136	63	5	complexity	complexity	NOUN
fcis-30136	63	6	while	while	SCONJ
fcis-30136	63	7	preserving	preserve	VERB
fcis-30136	63	8	flexibility	flexibility	NOUN
fcis-30136	63	9	.	.	PUNCT
fcis-30136	64	1	sac	sac	NOUN
fcis-30136	64	2	also	also	ADV
fcis-30136	64	3	incorporates	incorporate	VERB
fcis-30136	64	4	two	two	NUM
fcis-30136	64	5	global	global	ADJ
fcis-30136	64	6	context	context	NOUN
fcis-30136	64	7	modules	module	NOUN
fcis-30136	64	8	that	that	PRON
fcis-30136	64	9	augment	augment	VERB
fcis-30136	64	10	input	input	NOUN
fcis-30136	64	11	features	feature	NOUN
fcis-30136	64	12	with	with	ADP
fcis-30136	64	13	image	image	NOUN
fcis-30136	64	14	-	-	PUNCT
fcis-30136	64	15	level	level	NOUN
fcis-30136	64	16	information	information	NOUN
fcis-30136	64	17	.	.	PUNCT
fcis-30136	65	1	this	this	DET
fcis-30136	65	2	operation	operation	NOUN
fcis-30136	65	3	helps	help	VERB
fcis-30136	65	4	the	the	DET
fcis-30136	65	5	network	network	NOUN
fcis-30136	65	6	better	well	ADV
fcis-30136	65	7	understand	understand	VERB
fcis-30136	65	8	and	and	CCONJ
fcis-30136	65	9	process	process	VERB
fcis-30136	65	10	the	the	DET
fcis-30136	65	11	holistic	holistic	ADJ
fcis-30136	65	12	content	content	NOUN
fcis-30136	65	13	of	of	ADP
fcis-30136	65	14	images	image	NOUN
fcis-30136	65	15	,	,	PUNCT
fcis-30136	65	16	thereby	thereby	ADV
fcis-30136	65	17	improving	improve	VERB
fcis-30136	65	18	the	the	DET
fcis-30136	65	19	quality	quality	NOUN
fcis-30136	65	20	and	and	CCONJ
fcis-30136	65	21	accuracy	accuracy	NOUN
fcis-30136	65	22	of	of	ADP
fcis-30136	65	23	feature	feature	NOUN
fcis-30136	65	24	extraction	extraction	NOUN
fcis-30136	65	25	.	.	PUNCT
fcis-30136	66	1	figure	figure	NOUN
fcis-30136	66	2	3	3	NUM
fcis-30136	66	3	illustrates	illustrate	VERB
fcis-30136	66	4	the	the	DET
fcis-30136	66	5	specific	specific	ADJ
fcis-30136	66	6	implementation	implementation	NOUN
fcis-30136	66	7	of	of	ADP
fcis-30136	66	8	sac	sac	NOUN
fcis-30136	66	9	.	.	PUNCT
fcis-30136	66	10	figure	figure	NOUN
fcis-30136	66	11	3	3	NUM
fcis-30136	66	12	.	.	PUNCT
fcis-30136	66	13	detailed	detailed	ADJ
fcis-30136	66	14	architecture	architecture	NOUN
fcis-30136	66	15	of	of	ADP
fcis-30136	66	16	sac	sac	NOUN
fcis-30136	66	17	transformation	transformation	NOUN
fcis-30136	66	18	3.3	3.3	NUM
fcis-30136	66	19	.	.	PUNCT
fcis-30136	67	1	feature	feature	NOUN
fcis-30136	67	2	fusion	fusion	NOUN
fcis-30136	67	3	network	network	NOUN
fcis-30136	67	4	figure	figure	NOUN
fcis-30136	67	5	4	4	NUM
fcis-30136	67	6	.	.	PUNCT
fcis-30136	67	7	vov	vov	PROPN
fcis-30136	67	8	-	-	PUNCT
fcis-30136	67	9	gscsp	gscsp	PROPN
fcis-30136	67	10	module	module	NOUN
fcis-30136	67	11	the	the	DET
fcis-30136	67	12	neck	neck	NOUN
fcis-30136	67	13	is	be	AUX
fcis-30136	67	14	the	the	DET
fcis-30136	67	15	component	component	NOUN
fcis-30136	67	16	that	that	PRON
fcis-30136	67	17	connects	connect	VERB
fcis-30136	67	18	the	the	DET
fcis-30136	67	19	backbone	backbone	NOUN
fcis-30136	67	20	and	and	CCONJ
fcis-30136	67	21	head	head	NOUN
fcis-30136	67	22	,	,	PUNCT
fcis-30136	67	23	responsible	responsible	ADJ
fcis-30136	67	24	for	for	ADP
fcis-30136	67	25	processing	processing	NOUN
fcis-30136	67	26	and	and	CCONJ
fcis-30136	67	27	fusing	fuse	VERB
fcis-30136	67	28	feature	feature	NOUN
fcis-30136	67	29	maps	map	NOUN
fcis-30136	67	30	of	of	ADP
fcis-30136	67	31	different	different	ADJ
fcis-30136	67	32	scales	scale	NOUN
fcis-30136	67	33	.	.	PUNCT
fcis-30136	68	1	in	in	ADP
fcis-30136	68	2	potato	potato	NOUN
fcis-30136	68	3	recognition	recognition	NOUN
fcis-30136	68	4	scenarios	scenario	NOUN
fcis-30136	68	5	with	with	ADP
fcis-30136	68	6	dense	dense	ADJ
fcis-30136	68	7	occlusions	occlusion	NOUN
fcis-30136	68	8	,	,	PUNCT
fcis-30136	68	9	the	the	DET
fcis-30136	68	10	yolov8	yolov8	NOUN
fcis-30136	68	11	algorithm	algorithm	PROPN
fcis-30136	68	12	may	may	AUX
fcis-30136	68	13	lead	lead	VERB
fcis-30136	68	14	to	to	ADP
fcis-30136	68	15	missed	miss	VERB
fcis-30136	68	16	detections	detection	NOUN
fcis-30136	68	17	.	.	PUNCT
fcis-30136	69	1	therefore	therefore	ADV
fcis-30136	69	2	,	,	PUNCT
fcis-30136	69	3	the	the	DET
fcis-30136	69	4	slim	slim	ADJ
fcis-30136	69	5	-	-	PUNCT
fcis-30136	69	6	neck	neck	NOUN
fcis-30136	69	7	structure	structure	NOUN
fcis-30136	69	8	is	be	AUX
fcis-30136	69	9	introduced	introduce	VERB
fcis-30136	69	10	to	to	PART
fcis-30136	69	11	improve	improve	VERB
fcis-30136	69	12	the	the	DET
fcis-30136	69	13	neck	neck	NOUN
fcis-30136	69	14	network	network	NOUN
fcis-30136	69	15	of	of	ADP
fcis-30136	69	16	yolov8	yolov8	PROPN
fcis-30136	69	17	.	.	PUNCT
fcis-30136	70	1	slim	slim	ADJ
fcis-30136	70	2	-	-	PUNCT
fcis-30136	70	3	neck	neck	NOUN
fcis-30136	70	4	is	be	AUX
fcis-30136	70	5	constructed	construct	VERB
fcis-30136	70	6	using	use	VERB
fcis-30136	70	7	modern	modern	ADJ
fcis-30136	70	8	lightweight	lightweight	ADJ
fcis-30136	70	9	depthwise	depthwise	NOUN
fcis-30136	70	10	separable	separable	ADJ
fcis-30136	70	11	convolution	convolution	NOUN
fcis-30136	70	12	(	(	PUNCT
fcis-30136	70	13	gsconv	gsconv	NOUN
fcis-30136	70	14	)	)	PUNCT
fcis-30136	70	15	,	,	PUNCT
fcis-30136	70	16	gs	gs	PROPN
fcis-30136	70	17	bottlenecks	bottleneck	NOUN
fcis-30136	70	18	,	,	PUNCT
fcis-30136	70	19	and	and	CCONJ
fcis-30136	70	20	a	a	DET
fcis-30136	70	21	cross	cross	ADJ
fcis-30136	70	22	-	-	ADJ
fcis-30136	70	23	stage	stage	ADJ
fcis-30136	70	24	partial	partial	ADJ
fcis-30136	70	25	network	network	NOUN
fcis-30136	70	26	module	module	NOUN
fcis-30136	70	27	(	(	PUNCT
fcis-30136	70	28	vov	vov	PROPN
fcis-30136	70	29	-	-	PUNCT
fcis-30136	70	30	gscsp	gscsp	PROPN
fcis-30136	70	31	)	)	PUNCT
fcis-30136	70	32	.	.	PUNCT
fcis-30136	71	1	traditional	traditional	ADJ
fcis-30136	71	2	convolutional	convolutional	ADJ
fcis-30136	71	3	neural	neural	ADJ
fcis-30136	71	4	networks	network	NOUN
fcis-30136	71	5	convert	convert	VERB
fcis-30136	71	6	spatial	spatial	ADJ
fcis-30136	71	7	information	information	NOUN
fcis-30136	71	8	into	into	ADP
fcis-30136	71	9	channel	channel	NOUN
fcis-30136	71	10	information	information	NOUN
fcis-30136	71	11	,	,	PUNCT
fcis-30136	71	12	but	but	CCONJ
fcis-30136	71	13	this	this	DET
fcis-30136	71	14	process	process	NOUN
fcis-30136	71	15	inevitably	inevitably	ADV
fcis-30136	71	16	leads	lead	VERB
fcis-30136	71	17	to	to	ADP
fcis-30136	71	18	partial	partial	ADJ
fcis-30136	71	19	loss	loss	NOUN
fcis-30136	71	20	of	of	ADP
fcis-30136	71	21	semantic	semantic	ADJ
fcis-30136	71	22	information	information	NOUN
fcis-30136	71	23	during	during	ADP
fcis-30136	71	24	each	each	DET
fcis-30136	71	25	spatial	spatial	ADJ
fcis-30136	71	26	88	88	NUM
fcis-30136	71	27	compression	compression	NOUN
fcis-30136	71	28	and	and	CCONJ
fcis-30136	71	29	channel	channel	NOUN
fcis-30136	71	30	expansion	expansion	NOUN
fcis-30136	71	31	of	of	ADP
fcis-30136	71	32	feature	feature	NOUN
fcis-30136	71	33	maps	map	NOUN
fcis-30136	71	34	.	.	PUNCT
fcis-30136	72	1	gsconv	gsconv	NOUN
fcis-30136	72	2	independently	independently	ADV
fcis-30136	72	3	performs	perform	VERB
fcis-30136	72	4	convolution	convolution	NOUN
fcis-30136	72	5	on	on	ADP
fcis-30136	72	6	each	each	DET
fcis-30136	72	7	channel	channel	NOUN
fcis-30136	72	8	of	of	ADP
fcis-30136	72	9	the	the	DET
fcis-30136	72	10	input	input	NOUN
fcis-30136	72	11	feature	feature	NOUN
fcis-30136	72	12	map	map	NOUN
fcis-30136	72	13	,	,	PUNCT
fcis-30136	72	14	concatenates	concatenate	VERB
fcis-30136	72	15	the	the	DET
fcis-30136	72	16	convolution	convolution	NOUN
fcis-30136	72	17	results	result	NOUN
fcis-30136	72	18	,	,	PUNCT
fcis-30136	72	19	and	and	CCONJ
fcis-30136	72	20	then	then	ADV
fcis-30136	72	21	applies	apply	VERB
fcis-30136	72	22	a	a	DET
fcis-30136	72	23	shuffle	shuffle	ADJ
fcis-30136	72	24	operation	operation	NOUN
fcis-30136	72	25	to	to	PART
fcis-30136	72	26	rearrange	rearrange	VERB
fcis-30136	72	27	feature	feature	NOUN
fcis-30136	72	28	channels	channel	NOUN
fcis-30136	72	29	,	,	PUNCT
fcis-30136	72	30	thereby	thereby	ADV
fcis-30136	72	31	enhancing	enhance	VERB
fcis-30136	72	32	information	information	NOUN
fcis-30136	72	33	flow	flow	NOUN
fcis-30136	72	34	between	between	ADP
fcis-30136	72	35	features	feature	NOUN
fcis-30136	72	36	.	.	PUNCT
fcis-30136	73	1	the	the	DET
fcis-30136	73	2	gs	gs	PROPN
fcis-30136	73	3	bottleneck	bottleneck	NOUN
fcis-30136	73	4	built	build	VERB
fcis-30136	73	5	upon	upon	SCONJ
fcis-30136	73	6	gsconv	gsconv	NOUN
fcis-30136	73	7	improves	improve	VERB
fcis-30136	73	8	nonlinear	nonlinear	ADJ
fcis-30136	73	9	feature	feature	NOUN
fcis-30136	73	10	expression	expression	NOUN
fcis-30136	73	11	and	and	CCONJ
fcis-30136	73	12	information	information	NOUN
fcis-30136	73	13	reuse	reuse	NOUN
fcis-30136	73	14	.	.	PUNCT
fcis-30136	74	1	the	the	DET
fcis-30136	74	2	vov	vov	PROPN
fcis-30136	74	3	-	-	PUNCT
fcis-30136	74	4	gscsp	gscsp	PROPN
fcis-30136	74	5	module	module	NOUN
fcis-30136	74	6	,	,	PUNCT
fcis-30136	74	7	designed	design	VERB
fcis-30136	74	8	with	with	ADP
fcis-30136	74	9	a	a	DET
fcis-30136	74	10	one	one	NUM
fcis-30136	74	11	-	-	PUNCT
fcis-30136	74	12	time	time	NOUN
fcis-30136	74	13	aggregation	aggregation	NOUN
fcis-30136	74	14	method	method	NOUN
fcis-30136	74	15	,	,	PUNCT
fcis-30136	74	16	enables	enable	VERB
fcis-30136	74	17	effective	effective	ADJ
fcis-30136	74	18	information	information	NOUN
fcis-30136	74	19	fusion	fusion	NOUN
fcis-30136	74	20	between	between	ADP
fcis-30136	74	21	feature	feature	NOUN
fcis-30136	74	22	maps	map	NOUN
fcis-30136	74	23	at	at	ADP
fcis-30136	74	24	different	different	ADJ
fcis-30136	74	25	stages	stage	NOUN
fcis-30136	74	26	.	.	PUNCT
fcis-30136	75	1	this	this	DET
fcis-30136	75	2	approach	approach	NOUN
fcis-30136	75	3	reduces	reduce	VERB
fcis-30136	75	4	computational	computational	ADJ
fcis-30136	75	5	complexity	complexity	NOUN
fcis-30136	75	6	while	while	SCONJ
fcis-30136	75	7	maintaining	maintain	VERB
fcis-30136	75	8	sufficient	sufficient	ADJ
fcis-30136	75	9	accuracy	accuracy	NOUN
fcis-30136	75	10	,	,	PUNCT
fcis-30136	75	11	which	which	PRON
fcis-30136	75	12	is	be	AUX
fcis-30136	75	13	particularly	particularly	ADV
fcis-30136	75	14	beneficial	beneficial	ADJ
fcis-30136	75	15	for	for	ADP
fcis-30136	75	16	resource	resource	NOUN
fcis-30136	75	17	-	-	PUNCT
fcis-30136	75	18	constrained	constrain	VERB
fcis-30136	75	19	environments	environment	NOUN
fcis-30136	75	20	.	.	PUNCT
fcis-30136	76	1	these	these	DET
fcis-30136	76	2	modular	modular	ADJ
fcis-30136	76	3	designs	design	NOUN
fcis-30136	76	4	embody	embody	VERB
fcis-30136	76	5	the	the	DET
fcis-30136	76	6	slim	slim	ADJ
fcis-30136	76	7	-	-	PUNCT
fcis-30136	76	8	neck	neck	NOUN
fcis-30136	76	9	philosophy	philosophy	NOUN
fcis-30136	76	10	,	,	PUNCT
fcis-30136	76	11	aiming	aim	VERB
fcis-30136	76	12	to	to	PART
fcis-30136	76	13	reduce	reduce	VERB
fcis-30136	76	14	computational	computational	ADJ
fcis-30136	76	15	complexity	complexity	NOUN
fcis-30136	76	16	and	and	CCONJ
fcis-30136	76	17	inference	inference	NOUN
fcis-30136	76	18	time	time	NOUN
fcis-30136	76	19	without	without	ADP
fcis-30136	76	20	compromising	compromise	VERB
fcis-30136	76	21	accuracy	accuracy	NOUN
fcis-30136	76	22	.	.	PUNCT
fcis-30136	77	1	through	through	ADP
fcis-30136	77	2	such	such	ADJ
fcis-30136	77	3	modular	modular	ADJ
fcis-30136	77	4	design	design	NOUN
fcis-30136	77	5	,	,	PUNCT
fcis-30136	77	6	flexible	flexible	ADJ
fcis-30136	77	7	network	network	NOUN
fcis-30136	77	8	architectures	architecture	NOUN
fcis-30136	77	9	can	can	AUX
fcis-30136	77	10	be	be	AUX
fcis-30136	77	11	constructed	construct	VERB
fcis-30136	77	12	to	to	PART
fcis-30136	77	13	suit	suit	VERB
fcis-30136	77	14	specific	specific	ADJ
fcis-30136	77	15	task	task	NOUN
fcis-30136	77	16	requirements	requirement	NOUN
fcis-30136	77	17	.	.	PUNCT
fcis-30136	78	1	the	the	DET
fcis-30136	78	2	structure	structure	NOUN
fcis-30136	78	3	of	of	ADP
fcis-30136	78	4	the	the	DET
fcis-30136	78	5	vov	vov	PROPN
fcis-30136	78	6	-	-	PUNCT
fcis-30136	78	7	gscsp	gscsp	PROPN
fcis-30136	78	8	framework	framework	NOUN
fcis-30136	78	9	is	be	AUX
fcis-30136	78	10	illustrated	illustrate	VERB
fcis-30136	78	11	in	in	ADP
fcis-30136	78	12	figure	figure	NOUN
fcis-30136	78	13	4	4	NUM
fcis-30136	78	14	.	.	ADP
fcis-30136	78	15	3.4	3.4	NUM
fcis-30136	78	16	.	.	PUNCT
fcis-30136	79	1	attention	attention	NOUN
fcis-30136	79	2	module	module	NOUN
fcis-30136	79	3	the	the	DET
fcis-30136	79	4	efficient	efficient	ADJ
fcis-30136	79	5	multi	multi	ADJ
fcis-30136	79	6	-	-	ADJ
fcis-30136	79	7	scale	scale	ADJ
fcis-30136	79	8	attention	attention	NOUN
fcis-30136	79	9	(	(	PUNCT
fcis-30136	79	10	ema	ema	PROPN
fcis-30136	79	11	)	)	PUNCT
fcis-30136	79	12	module	module	NOUN
fcis-30136	79	13	is	be	AUX
fcis-30136	79	14	a	a	DET
fcis-30136	79	15	novel	novel	ADJ
fcis-30136	79	16	attention	attention	NOUN
fcis-30136	79	17	mechanism	mechanism	NOUN
fcis-30136	79	18	that	that	PRON
fcis-30136	79	19	reshapes	reshape	VERB
fcis-30136	79	20	partial	partial	ADJ
fcis-30136	79	21	channel	channel	NOUN
fcis-30136	79	22	and	and	CCONJ
fcis-30136	79	23	batch	batch	NOUN
fcis-30136	79	24	dimensions	dimension	NOUN
fcis-30136	79	25	while	while	SCONJ
fcis-30136	79	26	grouping	group	VERB
fcis-30136	79	27	channel	channel	NOUN
fcis-30136	79	28	dimensions	dimension	NOUN
fcis-30136	79	29	into	into	ADP
fcis-30136	79	30	multiple	multiple	ADJ
fcis-30136	79	31	sub	sub	NOUN
fcis-30136	79	32	-	-	NOUN
fcis-30136	79	33	features	feature	NOUN
fcis-30136	79	34	.	.	PUNCT
fcis-30136	80	1	this	this	DET
fcis-30136	80	2	approach	approach	NOUN
fcis-30136	80	3	preserves	preserve	VERB
fcis-30136	80	4	critical	critical	ADJ
fcis-30136	80	5	channel	channel	NOUN
fcis-30136	80	6	information	information	NOUN
fcis-30136	80	7	,	,	PUNCT
fcis-30136	80	8	reduces	reduce	VERB
fcis-30136	80	9	computational	computational	ADJ
fcis-30136	80	10	overhead	overhead	NOUN
fcis-30136	80	11	,	,	PUNCT
fcis-30136	80	12	and	and	CCONJ
fcis-30136	80	13	enhances	enhance	VERB
fcis-30136	80	14	the	the	DET
fcis-30136	80	15	model	model	NOUN
fcis-30136	80	16	's	's	PART
fcis-30136	80	17	feature	feature	NOUN
fcis-30136	80	18	processing	processing	NOUN
fcis-30136	80	19	capability	capability	NOUN
fcis-30136	80	20	.	.	PUNCT
fcis-30136	81	1	by	by	ADP
fcis-30136	81	2	encoding	encode	VERB
fcis-30136	81	3	global	global	ADJ
fcis-30136	81	4	information	information	NOUN
fcis-30136	81	5	,	,	PUNCT
fcis-30136	81	6	the	the	DET
fcis-30136	81	7	ema	ema	PROPN
fcis-30136	81	8	module	module	NOUN
fcis-30136	81	9	recalibrates	recalibrate	VERB
fcis-30136	81	10	channel	channel	NOUN
fcis-30136	81	11	weights	weight	NOUN
fcis-30136	81	12	in	in	ADP
fcis-30136	81	13	each	each	DET
fcis-30136	81	14	parallel	parallel	ADJ
fcis-30136	81	15	branch	branch	NOUN
fcis-30136	81	16	and	and	CCONJ
fcis-30136	81	17	captures	capture	VERB
fcis-30136	81	18	pixellevel	pixellevel	ADJ
fcis-30136	81	19	relationships	relationship	NOUN
fcis-30136	81	20	through	through	ADP
fcis-30136	81	21	cross	cross	ADJ
fcis-30136	81	22	-	-	ADJ
fcis-30136	81	23	dimensional	dimensional	ADJ
fcis-30136	81	24	interactions	interaction	NOUN
fcis-30136	81	25	,	,	PUNCT
fcis-30136	81	26	thereby	thereby	ADV
fcis-30136	81	27	strengthening	strengthen	VERB
fcis-30136	81	28	feature	feature	NOUN
fcis-30136	81	29	representation	representation	NOUN
fcis-30136	81	30	.	.	PUNCT
fcis-30136	82	1	as	as	SCONJ
fcis-30136	82	2	shown	show	VERB
fcis-30136	82	3	in	in	ADP
fcis-30136	82	4	figure	figure	NOUN
fcis-30136	82	5	5	5	NUM
fcis-30136	82	6	,	,	PUNCT
fcis-30136	82	7	g	g	PROPN
fcis-30136	82	8	represents	represent	VERB
fcis-30136	82	9	the	the	DET
fcis-30136	82	10	number	number	NOUN
fcis-30136	82	11	of	of	ADP
fcis-30136	82	12	groups	group	NOUN
fcis-30136	82	13	into	into	ADP
fcis-30136	82	14	which	which	PRON
fcis-30136	82	15	input	input	NOUN
fcis-30136	82	16	channels	channel	NOUN
fcis-30136	82	17	are	be	AUX
fcis-30136	82	18	divided	divide	VERB
fcis-30136	82	19	,	,	PUNCT
fcis-30136	82	20	while	while	SCONJ
fcis-30136	82	21	x_avg_pool	x_avg_pool	PROPN
fcis-30136	82	22	and	and	CCONJ
fcis-30136	82	23	y_avg_pool	y_avg_pool	NOUN
fcis-30136	82	24	denote	denote	VERB
fcis-30136	82	25	1d	1d	NUM
fcis-30136	82	26	average	average	ADJ
fcis-30136	82	27	pooling	pooling	NOUN
fcis-30136	82	28	along	along	ADP
fcis-30136	82	29	horizontal	horizontal	ADJ
fcis-30136	82	30	and	and	CCONJ
fcis-30136	82	31	vertical	vertical	ADJ
fcis-30136	82	32	directions	direction	NOUN
fcis-30136	82	33	,	,	PUNCT
fcis-30136	82	34	respectively	respectively	ADV
fcis-30136	82	35	.	.	PUNCT
fcis-30136	83	1	the	the	DET
fcis-30136	83	2	ema	ema	PROPN
fcis-30136	83	3	partitions	partition	VERB
fcis-30136	83	4	multi	multi	ADJ
fcis-30136	83	5	-	-	ADJ
fcis-30136	83	6	scale	scale	ADJ
fcis-30136	83	7	spatial	spatial	ADJ
fcis-30136	83	8	information	information	NOUN
fcis-30136	83	9	into	into	ADP
fcis-30136	83	10	three	three	NUM
fcis-30136	83	11	parallel	parallel	ADJ
fcis-30136	83	12	pathways	pathway	NOUN
fcis-30136	83	13	:	:	PUNCT
fcis-30136	83	14	two	two	NUM
fcis-30136	83	15	pathways	pathway	NOUN
fcis-30136	83	16	in	in	ADP
fcis-30136	83	17	the	the	DET
fcis-30136	83	18	1×1	1×1	ADJ
fcis-30136	83	19	branch	branch	NOUN
fcis-30136	83	20	utilize	utilize	VERB
fcis-30136	83	21	1d	1d	NUM
fcis-30136	83	22	global	global	ADJ
fcis-30136	83	23	average	average	ADJ
fcis-30136	83	24	pooling	pool	VERB
fcis-30136	83	25	operations	operation	NOUN
fcis-30136	83	26	along	along	ADP
fcis-30136	83	27	horizontal	horizontal	ADJ
fcis-30136	83	28	and	and	CCONJ
fcis-30136	83	29	vertical	vertical	ADJ
fcis-30136	83	30	spatial	spatial	ADJ
fcis-30136	83	31	directions	direction	NOUN
fcis-30136	83	32	for	for	ADP
fcis-30136	83	33	channel	channel	NOUN
fcis-30136	83	34	encoding	encoding	NOUN
fcis-30136	83	35	,	,	PUNCT
fcis-30136	83	36	while	while	SCONJ
fcis-30136	83	37	the	the	DET
fcis-30136	83	38	third	third	ADJ
fcis-30136	83	39	pathway	pathway	NOUN
fcis-30136	83	40	in	in	ADP
fcis-30136	83	41	the	the	DET
fcis-30136	83	42	3×3	3×3	NUM
fcis-30136	83	43	branch	branch	NOUN
fcis-30136	83	44	employs	employ	NOUN
fcis-30136	83	45	stacked	stack	VERB
fcis-30136	83	46	3×3	3×3	NUM
fcis-30136	83	47	kernels	kernel	NOUN
fcis-30136	83	48	to	to	PART
fcis-30136	83	49	capture	capture	VERB
fcis-30136	83	50	multi	multi	ADJ
fcis-30136	83	51	-	-	ADJ
fcis-30136	83	52	scale	scale	ADJ
fcis-30136	83	53	feature	feature	NOUN
fcis-30136	83	54	representations	representation	NOUN
fcis-30136	83	55	.	.	PUNCT
fcis-30136	84	1	output	output	NOUN
fcis-30136	84	2	features	feature	NOUN
fcis-30136	84	3	from	from	ADP
fcis-30136	84	4	both	both	DET
fcis-30136	84	5	branches	branch	NOUN
fcis-30136	84	6	undergo	undergo	VERB
fcis-30136	84	7	sigmoid	sigmoid	NOUN
fcis-30136	84	8	activation	activation	NOUN
fcis-30136	84	9	and	and	CCONJ
fcis-30136	84	10	normalization	normalization	NOUN
fcis-30136	84	11	,	,	PUNCT
fcis-30136	84	12	then	then	ADV
fcis-30136	84	13	merge	merge	VERB
fcis-30136	84	14	through	through	ADP
fcis-30136	84	15	a	a	DET
fcis-30136	84	16	cross	cross	ADJ
fcis-30136	84	17	-	-	ADJ
fcis-30136	84	18	dimensional	dimensional	ADJ
fcis-30136	84	19	interaction	interaction	NOUN
fcis-30136	84	20	module	module	NOUN
fcis-30136	84	21	to	to	PART
fcis-30136	84	22	establish	establish	VERB
fcis-30136	84	23	pixel	pixel	ADJ
fcis-30136	84	24	-	-	PUNCT
fcis-30136	84	25	level	level	NOUN
fcis-30136	84	26	pairwise	pairwise	NOUN
fcis-30136	84	27	relationships	relationship	NOUN
fcis-30136	84	28	.	.	PUNCT
fcis-30136	85	1	after	after	ADP
fcis-30136	85	2	final	final	ADJ
fcis-30136	85	3	sigmoid	sigmoid	NOUN
fcis-30136	85	4	modulation	modulation	NOUN
fcis-30136	85	5	,	,	PUNCT
fcis-30136	85	6	the	the	DET
fcis-30136	85	7	enhanced	enhance	VERB
fcis-30136	85	8	/	/	SYM
fcis-30136	85	9	attenuated	attenuate	VERB
fcis-30136	85	10	feature	feature	NOUN
fcis-30136	85	11	maps	map	NOUN
fcis-30136	85	12	are	be	AUX
fcis-30136	85	13	combined	combine	VERB
fcis-30136	85	14	with	with	ADP
fcis-30136	85	15	original	original	ADJ
fcis-30136	85	16	inputs	input	NOUN
fcis-30136	85	17	to	to	PART
fcis-30136	85	18	produce	produce	VERB
fcis-30136	85	19	the	the	DET
fcis-30136	85	20	ultimate	ultimate	ADJ
fcis-30136	85	21	output	output	NOUN
fcis-30136	85	22	.	.	PUNCT
fcis-30136	86	1	figure	figure	NOUN
fcis-30136	86	2	5	5	NUM
fcis-30136	86	3	.	.	PUNCT
fcis-30136	86	4	ema	ema	PROPN
fcis-30136	86	5	architecture	architecture	PROPN
fcis-30136	86	6	diagram	diagram	NOUN
fcis-30136	86	7	4	4	NUM
fcis-30136	86	8	.	.	PUNCT
fcis-30136	87	1	experiments	experiment	NOUN
fcis-30136	87	2	and	and	CCONJ
fcis-30136	87	3	results	result	VERB
fcis-30136	87	4	analysis	analysis	NOUN
fcis-30136	87	5	4.1	4.1	NUM
fcis-30136	87	6	.	.	PUNCT
fcis-30136	88	1	experimental	experimental	ADJ
fcis-30136	88	2	environment	environment	NOUN
fcis-30136	88	3	figure	figure	NOUN
fcis-30136	88	4	6	6	NUM
fcis-30136	88	5	.	.	PUNCT
fcis-30136	89	1	image	image	NOUN
fcis-30136	89	2	examples	example	NOUN
fcis-30136	89	3	of	of	ADP
fcis-30136	89	4	five	five	NUM
fcis-30136	89	5	categories	category	NOUN
fcis-30136	89	6	89	89	NUM
fcis-30136	89	7	the	the	DET
fcis-30136	89	8	equipment	equipment	NOUN
fcis-30136	89	9	operating	operating	NOUN
fcis-30136	89	10	system	system	NOUN
fcis-30136	89	11	used	use	VERB
fcis-30136	89	12	in	in	ADP
fcis-30136	89	13	this	this	DET
fcis-30136	89	14	experiment	experiment	NOUN
fcis-30136	89	15	was	be	AUX
fcis-30136	89	16	windows	window	NOUN
fcis-30136	89	17	11	11	NUM
fcis-30136	89	18	.	.	PUNCT
fcis-30136	90	1	the	the	DET
fcis-30136	90	2	model	model	NOUN
fcis-30136	90	3	framework	framework	NOUN
fcis-30136	90	4	utilized	utilize	VERB
fcis-30136	90	5	the	the	DET
fcis-30136	90	6	pytorch	pytorch	NOUN
fcis-30136	90	7	2.2.1	2.2.1	NUM
fcis-30136	90	8	deep	deep	ADJ
fcis-30136	90	9	learning	learning	NOUN
fcis-30136	90	10	framework	framework	NOUN
fcis-30136	90	11	.	.	PUNCT
fcis-30136	91	1	the	the	DET
fcis-30136	91	2	central	central	ADJ
fcis-30136	91	3	processing	processing	NOUN
fcis-30136	91	4	unit	unit	NOUN
fcis-30136	91	5	(	(	PUNCT
fcis-30136	91	6	cpu	cpu	PROPN
fcis-30136	91	7	)	)	PUNCT
fcis-30136	91	8	was	be	AUX
fcis-30136	91	9	a	a	DET
fcis-30136	91	10	12th	12th	ADJ
fcis-30136	91	11	gen	gen	PROPN
fcis-30136	91	12	intel(r	intel(r	PROPN
fcis-30136	91	13	)	)	PUNCT
fcis-30136	91	14	core	core	NOUN
fcis-30136	91	15	(	(	PUNCT
fcis-30136	91	16	tm	tm	NOUN
fcis-30136	91	17	)	)	PUNCT
fcis-30136	91	18	i7	i7	NOUN
fcis-30136	91	19	-	-	PUNCT
fcis-30136	91	20	12700h	12700h	NUM
fcis-30136	91	21	,	,	PUNCT
fcis-30136	91	22	with	with	ADP
fcis-30136	91	23	an	an	DET
fcis-30136	91	24	nvidia	nvidia	PROPN
fcis-30136	91	25	geforce	geforce	NOUN
fcis-30136	91	26	rtx	rtx	PROPN
fcis-30136	91	27	3060	3060	NUM
fcis-30136	91	28	graphics	graphic	NOUN
fcis-30136	91	29	card	card	NOUN
fcis-30136	91	30	providing	provide	VERB
fcis-30136	91	31	6	6	NUM
fcis-30136	91	32	gb	gb	NOUN
fcis-30136	91	33	of	of	ADP
fcis-30136	91	34	video	video	NOUN
fcis-30136	91	35	memory	memory	NOUN
fcis-30136	91	36	.	.	PUNCT
fcis-30136	92	1	the	the	DET
fcis-30136	92	2	device	device	NOUN
fcis-30136	92	3	was	be	AUX
fcis-30136	92	4	equipped	equip	VERB
fcis-30136	92	5	with	with	ADP
fcis-30136	92	6	16	16	NUM
fcis-30136	92	7	gb	gb	NOUN
fcis-30136	92	8	of	of	ADP
fcis-30136	92	9	ram	ram	NOUN
fcis-30136	92	10	,	,	PUNCT
fcis-30136	92	11	running	run	VERB
fcis-30136	92	12	on	on	ADP
fcis-30136	92	13	program	program	NOUN
fcis-30136	92	14	compiler	compiler	NOUN
fcis-30136	92	15	version	version	NOUN
fcis-30136	92	16	2023	2023	NUM
fcis-30136	92	17	with	with	ADP
fcis-30136	92	18	python	python	PROPN
fcis-30136	92	19	3.11	3.11	NUM
fcis-30136	92	20	and	and	CCONJ
fcis-30136	92	21	cuda	cuda	NOUN
fcis-30136	92	22	version	version	PROPN
fcis-30136	92	23	12.1	12.1	NUM
fcis-30136	92	24	.	.	PUNCT
fcis-30136	93	1	4.2	4.2	NUM
fcis-30136	93	2	.	.	PUNCT
fcis-30136	93	3	dataset	dataset	VERB
fcis-30136	93	4	the	the	DET
fcis-30136	93	5	dataset	dataset	NOUN
fcis-30136	93	6	used	use	VERB
fcis-30136	93	7	in	in	ADP
fcis-30136	93	8	this	this	DET
fcis-30136	93	9	experiment	experiment	NOUN
fcis-30136	93	10	is	be	AUX
fcis-30136	93	11	potato	potato	NOUN
fcis-30136	93	12	-	-	PUNCT
fcis-30136	93	13	eruvaka	eruvaka	ADJ
fcis-30136	93	14	,	,	PUNCT
fcis-30136	93	15	with	with	ADP
fcis-30136	93	16	example	example	NOUN
fcis-30136	93	17	images	image	NOUN
fcis-30136	93	18	shown	show	VERB
fcis-30136	93	19	in	in	ADP
fcis-30136	93	20	figure	figure	NOUN
fcis-30136	93	21	6	6	NUM
fcis-30136	93	22	.	.	PUNCT
fcis-30136	94	1	this	this	DET
fcis-30136	94	2	dataset	dataset	NOUN
fcis-30136	94	3	contains	contain	VERB
fcis-30136	94	4	8,003	8,003	NUM
fcis-30136	94	5	images	image	NOUN
fcis-30136	94	6	of	of	ADP
fcis-30136	94	7	five	five	NUM
fcis-30136	94	8	potato	potato	NOUN
fcis-30136	94	9	categories	category	NOUN
fcis-30136	94	10	:	:	PUNCT
fcis-30136	94	11	normal	normal	ADJ
fcis-30136	94	12	potato	potato	NOUN
fcis-30136	94	13	,	,	PUNCT
fcis-30136	94	14	damaged	damage	VERB
fcis-30136	94	15	potato	potato	NOUN
fcis-30136	94	16	,	,	PUNCT
fcis-30136	94	17	defected	defect	VERB
fcis-30136	94	18	potato	potato	NOUN
fcis-30136	94	19	,	,	PUNCT
fcis-30136	94	20	diseased	diseased	ADJ
fcis-30136	94	21	potato	potato	NOUN
fcis-30136	94	22	,	,	PUNCT
fcis-30136	94	23	and	and	CCONJ
fcis-30136	94	24	sprouted	sprout	VERB
fcis-30136	94	25	potato	potato	NOUN
fcis-30136	94	26	.	.	PUNCT
fcis-30136	95	1	each	each	DET
fcis-30136	95	2	image	image	NOUN
fcis-30136	95	3	has	have	VERB
fcis-30136	95	4	a	a	DET
fcis-30136	95	5	resolution	resolution	NOUN
fcis-30136	95	6	of	of	ADP
fcis-30136	95	7	640×640	640×640	NUM
fcis-30136	95	8	pixels	pixel	NOUN
fcis-30136	95	9	.	.	PUNCT
fcis-30136	96	1	the	the	DET
fcis-30136	96	2	dataset	dataset	NOUN
fcis-30136	96	3	is	be	AUX
fcis-30136	96	4	divided	divide	VERB
fcis-30136	96	5	into	into	ADP
fcis-30136	96	6	training	training	NOUN
fcis-30136	96	7	,	,	PUNCT
fcis-30136	96	8	validation	validation	NOUN
fcis-30136	96	9	,	,	PUNCT
fcis-30136	96	10	and	and	CCONJ
fcis-30136	96	11	test	test	NOUN
fcis-30136	96	12	sets	set	NOUN
fcis-30136	96	13	with	with	ADP
fcis-30136	96	14	an	an	DET
fcis-30136	96	15	8:1:1	8:1:1	NUM
fcis-30136	96	16	ratio	ratio	NOUN
fcis-30136	96	17	of	of	ADP
fcis-30136	96	18	image	image	NOUN
fcis-30136	96	19	quantities	quantity	NOUN
fcis-30136	96	20	.	.	PUNCT
fcis-30136	97	1	4.3	4.3	NUM
fcis-30136	97	2	.	.	PUNCT
fcis-30136	97	3	evaluation	evaluation	NOUN
fcis-30136	97	4	metrics	metric	NOUN
fcis-30136	97	5	this	this	DET
fcis-30136	97	6	experiment	experiment	NOUN
fcis-30136	97	7	employs	employ	VERB
fcis-30136	97	8	precision	precision	NOUN
fcis-30136	97	9	(	(	PUNCT
fcis-30136	97	10	p	p	NOUN
fcis-30136	97	11	)	)	PUNCT
fcis-30136	97	12	,	,	PUNCT
fcis-30136	97	13	recall	recall	INTJ
fcis-30136	97	14	(	(	PUNCT
fcis-30136	97	15	r	r	NOUN
fcis-30136	97	16	)	)	PUNCT
fcis-30136	97	17	,	,	PUNCT
fcis-30136	97	18	mean	mean	VERB
fcis-30136	97	19	average	average	ADJ
fcis-30136	97	20	precision	precision	NOUN
fcis-30136	97	21	(	(	PUNCT
fcis-30136	97	22	map	map	NOUN
fcis-30136	97	23	)	)	PUNCT
fcis-30136	97	24	,	,	PUNCT
fcis-30136	97	25	computational	computational	ADJ
fcis-30136	97	26	complexity	complexity	NOUN
fcis-30136	97	27	(	(	PUNCT
fcis-30136	97	28	gflops	gflop	NOUN
fcis-30136	97	29	)	)	PUNCT
fcis-30136	97	30	,	,	PUNCT
fcis-30136	97	31	and	and	CCONJ
fcis-30136	97	32	frames	frame	NOUN
fcis-30136	97	33	per	per	ADP
fcis-30136	97	34	second	second	ADJ
fcis-30136	97	35	(	(	PUNCT
fcis-30136	97	36	fps	fps	PROPN
fcis-30136	97	37	)	)	PUNCT
fcis-30136	97	38	as	as	ADP
fcis-30136	97	39	evaluation	evaluation	NOUN
fcis-30136	97	40	metrics	metric	NOUN
fcis-30136	97	41	.	.	PUNCT
fcis-30136	98	1	precision	precision	NOUN
fcis-30136	98	2	represents	represent	VERB
fcis-30136	98	3	the	the	DET
fcis-30136	98	4	proportion	proportion	NOUN
fcis-30136	98	5	of	of	ADP
fcis-30136	98	6	correctly	correctly	ADV
fcis-30136	98	7	predicted	predict	VERB
fcis-30136	98	8	positive	positive	ADJ
fcis-30136	98	9	samples	sample	NOUN
fcis-30136	98	10	among	among	ADP
fcis-30136	98	11	all	all	ADV
fcis-30136	98	12	predicted	predict	VERB
fcis-30136	98	13	positive	positive	ADJ
fcis-30136	98	14	samples	sample	NOUN
fcis-30136	98	15	;	;	PUNCT
fcis-30136	98	16	recall	recall	NOUN
fcis-30136	98	17	denotes	denote	VERB
fcis-30136	98	18	the	the	DET
fcis-30136	98	19	proportion	proportion	NOUN
fcis-30136	98	20	of	of	ADP
fcis-30136	98	21	correctly	correctly	ADV
fcis-30136	98	22	predicted	predict	VERB
fcis-30136	98	23	positive	positive	ADJ
fcis-30136	98	24	samples	sample	NOUN
fcis-30136	98	25	among	among	ADP
fcis-30136	98	26	all	all	DET
fcis-30136	98	27	actual	actual	ADJ
fcis-30136	98	28	positive	positive	ADJ
fcis-30136	98	29	samples	sample	NOUN
fcis-30136	98	30	;	;	PUNCT
fcis-30136	98	31	mean	mean	ADJ
fcis-30136	98	32	average	average	ADJ
fcis-30136	98	33	precision	precision	NOUN
fcis-30136	98	34	(	(	PUNCT
fcis-30136	98	35	map	map	NOUN
fcis-30136	98	36	)	)	PUNCT
fcis-30136	98	37	is	be	AUX
fcis-30136	98	38	obtained	obtain	VERB
fcis-30136	98	39	by	by	ADP
fcis-30136	98	40	averaging	average	VERB
fcis-30136	98	41	the	the	DET
fcis-30136	98	42	results	result	NOUN
fcis-30136	98	43	of	of	ADP
fcis-30136	98	44	integrating	integrate	VERB
fcis-30136	98	45	the	the	DET
fcis-30136	98	46	precision	precision	NOUN
fcis-30136	98	47	-	-	PUNCT
fcis-30136	98	48	recall	recall	NOUN
fcis-30136	98	49	(	(	PUNCT
fcis-30136	98	50	p	p	NOUN
fcis-30136	98	51	-	-	PUNCT
fcis-30136	98	52	r	r	NOUN
fcis-30136	98	53	)	)	PUNCT
fcis-30136	98	54	curves	curve	NOUN
fcis-30136	98	55	for	for	ADP
fcis-30136	98	56	each	each	DET
fcis-30136	98	57	category	category	NOUN
fcis-30136	98	58	.	.	PUNCT
fcis-30136	99	1	the	the	DET
fcis-30136	99	2	equations	equation	NOUN
fcis-30136	99	3	for	for	ADP
fcis-30136	99	4	precision	precision	NOUN
fcis-30136	99	5	,	,	PUNCT
fcis-30136	99	6	recall	recall	NOUN
fcis-30136	99	7	,	,	PUNCT
fcis-30136	99	8	and	and	CCONJ
fcis-30136	99	9	mean	mean	ADJ
fcis-30136	99	10	average	average	ADJ
fcis-30136	99	11	precision	precision	NOUN
fcis-30136	99	12	are	be	AUX
fcis-30136	99	13	shown	show	VERB
fcis-30136	99	14	in	in	ADP
fcis-30136	99	15	formulas	formula	NOUN
fcis-30136	99	16	(	(	PUNCT
fcis-30136	99	17	5	5	NUM
fcis-30136	99	18	)	)	PUNCT
fcis-30136	99	19	to	to	ADP
fcis-30136	99	20	(	(	PUNCT
fcis-30136	99	21	7	7	NUM
fcis-30136	99	22	)	)	PUNCT
fcis-30136	99	23	.	.	PUNCT
fcis-30136	100	1	5	5	NUM
fcis-30136	100	2	6	6	NUM
fcis-30136	100	3	∑	∑	PROPN
fcis-30136	100	4	d	d	SYM
fcis-30136	100	5	7	7	NUM
fcis-30136	100	6	tp	tp	NOUN
fcis-30136	100	7	represents	represent	VERB
fcis-30136	100	8	the	the	DET
fcis-30136	100	9	number	number	NOUN
fcis-30136	100	10	of	of	ADP
fcis-30136	100	11	correctly	correctly	ADV
fcis-30136	100	12	classified	classify	VERB
fcis-30136	100	13	positive	positive	ADJ
fcis-30136	100	14	samples	sample	NOUN
fcis-30136	100	15	,	,	PUNCT
fcis-30136	100	16	fp	fp	X
fcis-30136	100	17	represents	represent	VERB
fcis-30136	100	18	the	the	DET
fcis-30136	100	19	number	number	NOUN
fcis-30136	100	20	of	of	ADP
fcis-30136	100	21	correctly	correctly	ADV
fcis-30136	100	22	classified	classified	ADJ
fcis-30136	100	23	negative	negative	ADJ
fcis-30136	100	24	samples	sample	NOUN
fcis-30136	100	25	,	,	PUNCT
fcis-30136	100	26	and	and	CCONJ
fcis-30136	100	27	fn	fn	NOUN
fcis-30136	100	28	represents	represent	VERB
fcis-30136	100	29	the	the	DET
fcis-30136	100	30	number	number	NOUN
fcis-30136	100	31	of	of	ADP
fcis-30136	100	32	incorrectly	incorrectly	ADV
fcis-30136	100	33	classified	classify	VERB
fcis-30136	100	34	positive	positive	ADJ
fcis-30136	100	35	samples	sample	NOUN
fcis-30136	100	36	.	.	PUNCT
fcis-30136	101	1	4.4	4.4	NUM
fcis-30136	101	2	.	.	PUNCT
fcis-30136	101	3	ablation	ablation	NOUN
fcis-30136	101	4	experiment	experiment	NOUN
fcis-30136	101	5	as	as	SCONJ
fcis-30136	101	6	can	can	AUX
fcis-30136	101	7	be	be	AUX
fcis-30136	101	8	seen	see	VERB
fcis-30136	101	9	from	from	ADP
fcis-30136	101	10	table	table	NOUN
fcis-30136	101	11	1	1	NUM
fcis-30136	101	12	,	,	PUNCT
fcis-30136	101	13	the	the	DET
fcis-30136	101	14	map	map	NOUN
fcis-30136	101	15	value	value	NOUN
fcis-30136	101	16	of	of	ADP
fcis-30136	101	17	the	the	DET
fcis-30136	101	18	unimproved	unimproved	ADJ
fcis-30136	101	19	yolov8	yolov8	NOUN
fcis-30136	101	20	algorithm	algorithm	PROPN
fcis-30136	101	21	is	be	AUX
fcis-30136	101	22	86.4	86.4	NUM
fcis-30136	101	23	%	%	NOUN
fcis-30136	101	24	,	,	PUNCT
fcis-30136	101	25	while	while	SCONJ
fcis-30136	101	26	the	the	DET
fcis-30136	101	27	improved	improved	ADJ
fcis-30136	101	28	yolov8	yolov8	NOUN
fcis-30136	101	29	algorithm	algorithm	PROPN
fcis-30136	101	30	achieves	achieve	VERB
fcis-30136	101	31	a	a	DET
fcis-30136	101	32	map	map	NOUN
fcis-30136	101	33	of	of	ADP
fcis-30136	101	34	89.7	89.7	NUM
fcis-30136	101	35	%	%	NOUN
fcis-30136	101	36	,	,	PUNCT
fcis-30136	101	37	representing	represent	VERB
fcis-30136	101	38	an	an	DET
fcis-30136	101	39	increase	increase	NOUN
fcis-30136	101	40	of	of	ADP
fcis-30136	101	41	3.3	3.3	NUM
fcis-30136	101	42	percentage	percentage	NOUN
fcis-30136	101	43	points	point	NOUN
fcis-30136	101	44	.	.	PUNCT
fcis-30136	102	1	the	the	DET
fcis-30136	102	2	average	average	ADJ
fcis-30136	102	3	precision	precision	NOUN
fcis-30136	102	4	after	after	ADP
fcis-30136	102	5	implementing	implement	VERB
fcis-30136	102	6	different	different	ADJ
fcis-30136	102	7	module	module	NOUN
fcis-30136	102	8	improvements	improvement	NOUN
fcis-30136	102	9	reaches	reach	VERB
fcis-30136	102	10	89.1	89.1	NUM
fcis-30136	102	11	%	%	NOUN
fcis-30136	102	12	,	,	PUNCT
fcis-30136	102	13	88.8	88.8	NUM
fcis-30136	102	14	%	%	NOUN
fcis-30136	102	15	,	,	PUNCT
fcis-30136	102	16	and	and	CCONJ
fcis-30136	102	17	89.5	89.5	NUM
fcis-30136	102	18	%	%	NOUN
fcis-30136	102	19	respectively	respectively	ADV
fcis-30136	102	20	,	,	PUNCT
fcis-30136	102	21	which	which	PRON
fcis-30136	102	22	are	be	AUX
fcis-30136	102	23	2.7	2.7	NUM
fcis-30136	102	24	%	%	NOUN
fcis-30136	102	25	,	,	PUNCT
fcis-30136	102	26	2.4	2.4	NUM
fcis-30136	102	27	%	%	NOUN
fcis-30136	102	28	,	,	PUNCT
fcis-30136	102	29	and	and	CCONJ
fcis-30136	102	30	3.1	3.1	NUM
fcis-30136	102	31	%	%	NOUN
fcis-30136	102	32	higher	high	ADJ
fcis-30136	102	33	than	than	ADP
fcis-30136	102	34	the	the	DET
fcis-30136	102	35	original	original	ADJ
fcis-30136	102	36	yolov8	yolov8	NOUN
fcis-30136	102	37	algorithm	algorithm	NOUN
fcis-30136	102	38	.	.	PUNCT
fcis-30136	103	1	improvement	improvement	NOUN
fcis-30136	103	2	4	4	NUM
fcis-30136	103	3	demonstrates	demonstrate	VERB
fcis-30136	103	4	the	the	DET
fcis-30136	103	5	best	good	ADJ
fcis-30136	103	6	performance	performance	NOUN
fcis-30136	103	7	in	in	ADP
fcis-30136	103	8	detecting	detect	VERB
fcis-30136	103	9	potato	potato	NOUN
fcis-30136	103	10	,	,	PUNCT
fcis-30136	103	11	damaged	damage	VERB
fcis-30136	103	12	,	,	PUNCT
fcis-30136	103	13	diseased	diseased	ADJ
fcis-30136	103	14	,	,	PUNCT
fcis-30136	103	15	and	and	CCONJ
fcis-30136	103	16	sprouted	sprout	VERB
fcis-30136	103	17	categories	category	NOUN
fcis-30136	103	18	,	,	PUNCT
fcis-30136	103	19	proving	prove	VERB
fcis-30136	103	20	that	that	SCONJ
fcis-30136	103	21	our	our	PRON
fcis-30136	103	22	algorithm	algorithm	NOUN
fcis-30136	103	23	enhancements	enhancement	NOUN
fcis-30136	103	24	can	can	AUX
fcis-30136	103	25	effectively	effectively	ADV
fcis-30136	103	26	recognize	recognize	VERB
fcis-30136	103	27	different	different	ADJ
fcis-30136	103	28	potato	potato	NOUN
fcis-30136	103	29	types	type	NOUN
fcis-30136	103	30	.	.	PUNCT
fcis-30136	104	1	after	after	ADP
fcis-30136	104	2	implementing	implement	VERB
fcis-30136	104	3	dynamicconv	dynamicconv	ADJ
fcis-30136	104	4	and	and	CCONJ
fcis-30136	104	5	sac	sac	NOUN
fcis-30136	104	6	module	module	NOUN
fcis-30136	104	7	improvements	improvement	NOUN
fcis-30136	104	8	,	,	PUNCT
fcis-30136	104	9	the	the	DET
fcis-30136	104	10	model	model	NOUN
fcis-30136	104	11	's	's	PART
fcis-30136	104	12	average	average	ADJ
fcis-30136	104	13	precision	precision	NOUN
fcis-30136	104	14	shows	show	VERB
fcis-30136	104	15	consistent	consistent	ADJ
fcis-30136	104	16	improvement	improvement	NOUN
fcis-30136	104	17	.	.	PUNCT
fcis-30136	105	1	although	although	SCONJ
fcis-30136	105	2	improvement	improvement	NOUN
fcis-30136	105	3	2	2	NUM
fcis-30136	105	4	exhibits	exhibit	VERB
fcis-30136	105	5	a	a	DET
fcis-30136	105	6	slight	slight	ADJ
fcis-30136	105	7	decrease	decrease	NOUN
fcis-30136	105	8	in	in	ADP
fcis-30136	105	9	average	average	ADJ
fcis-30136	105	10	precision	precision	NOUN
fcis-30136	105	11	compared	compare	VERB
fcis-30136	105	12	to	to	ADP
fcis-30136	105	13	improvement	improvement	NOUN
fcis-30136	105	14	1	1	NUM
fcis-30136	105	15	,	,	PUNCT
fcis-30136	105	16	it	it	PRON
fcis-30136	105	17	successfully	successfully	ADV
fcis-30136	105	18	reduces	reduce	VERB
fcis-30136	105	19	model	model	NOUN
fcis-30136	105	20	size	size	NOUN
fcis-30136	105	21	,	,	PUNCT
fcis-30136	105	22	decreases	decrease	VERB
fcis-30136	105	23	computational	computational	ADJ
fcis-30136	105	24	requirements	requirement	NOUN
fcis-30136	105	25	,	,	PUNCT
fcis-30136	105	26	and	and	CCONJ
fcis-30136	105	27	enhances	enhance	VERB
fcis-30136	105	28	detection	detection	NOUN
fcis-30136	105	29	speed	speed	NOUN
fcis-30136	105	30	.	.	PUNCT
fcis-30136	106	1	improvement	improvement	NOUN
fcis-30136	106	2	3	3	NUM
fcis-30136	106	3	further	far	ADV
fcis-30136	106	4	elevates	elevate	VERB
fcis-30136	106	5	average	average	ADJ
fcis-30136	106	6	precision	precision	NOUN
fcis-30136	106	7	while	while	SCONJ
fcis-30136	106	8	continuing	continue	VERB
fcis-30136	106	9	to	to	PART
fcis-30136	106	10	reduce	reduce	VERB
fcis-30136	106	11	computational	computational	ADJ
fcis-30136	106	12	costs	cost	NOUN
fcis-30136	106	13	and	and	CCONJ
fcis-30136	106	14	model	model	NOUN
fcis-30136	106	15	size	size	NOUN
fcis-30136	106	16	.	.	PUNCT
fcis-30136	107	1	in	in	ADP
fcis-30136	107	2	table	table	NOUN
fcis-30136	107	3	1	1	NUM
fcis-30136	107	4	,	,	PUNCT
fcis-30136	107	5	potato	potato	NOUN
fcis-30136	107	6	,	,	PUNCT
fcis-30136	107	7	damaged	damage	VERB
fcis-30136	107	8	,	,	PUNCT
fcis-30136	107	9	defected	defect	VERB
fcis-30136	107	10	,	,	PUNCT
fcis-30136	107	11	diseased	diseased	ADJ
fcis-30136	107	12	,	,	PUNCT
fcis-30136	107	13	and	and	CCONJ
fcis-30136	107	14	sprouted	sprout	VERB
fcis-30136	107	15	represent	represent	VERB
fcis-30136	107	16	the	the	DET
fcis-30136	107	17	precision	precision	NOUN
fcis-30136	107	18	values	value	NOUN
fcis-30136	107	19	corresponding	correspond	VERB
fcis-30136	107	20	to	to	ADP
fcis-30136	107	21	each	each	DET
fcis-30136	107	22	respective	respective	ADJ
fcis-30136	107	23	category	category	NOUN
fcis-30136	107	24	.	.	PUNCT
fcis-30136	108	1	table	table	NOUN
fcis-30136	108	2	1	1	NUM
fcis-30136	108	3	.	.	PUNCT
fcis-30136	108	4	ablation	ablation	NOUN
fcis-30136	108	5	experiment	experiment	NOUN
fcis-30136	108	6	results	result	NOUN
fcis-30136	108	7	project	project	VERB
fcis-30136	108	8	dynamicconv	dynamicconv	ADJ
fcis-30136	108	9	sac	sac	NOUN
fcis-30136	108	10	vovgscsp	vovgscsp	NOUN
fcis-30136	108	11	ema	ema	PROPN
fcis-30136	108	12	ap/%	ap/%	PROPN
fcis-30136	108	13	potato	potato	NOUN
fcis-30136	109	1	damaged	damage	VERB
fcis-30136	109	2	yolov8	yolov8	NOUN
fcis-30136	109	3	×	×	NOUN
fcis-30136	109	4	×	×	NOUN
fcis-30136	109	5	×	×	NOUN
fcis-30136	109	6	×	×	NOUN
fcis-30136	109	7	90.0	90.0	NUM
fcis-30136	109	8	80.0	80.0	NUM
fcis-30136	109	9	project	project	NOUN
fcis-30136	109	10	1	1	NUM
fcis-30136	109	11	√	√	NUM
fcis-30136	110	1	×	×	NOUN
fcis-30136	111	1	×	×	NOUN
fcis-30136	111	2	×	×	NOUN
fcis-30136	111	3	94.1	94.1	NUM
fcis-30136	111	4	88.0	88.0	NUM
fcis-30136	111	5	project	project	NOUN
fcis-30136	111	6	2	2	NUM
fcis-30136	111	7	√	√	NUM
fcis-30136	111	8	√	√	NUM
fcis-30136	111	9	×	×	PROPN
fcis-30136	111	10	×	×	NOUN
fcis-30136	111	11	88.8	88.8	NUM
fcis-30136	111	12	88.3	88.3	NUM
fcis-30136	111	13	project	project	NOUN
fcis-30136	111	14	3	3	NUM
fcis-30136	111	15	√	√	NOUN
fcis-30136	111	16	√	√	NUM
fcis-30136	111	17	√	√	NUM
fcis-30136	111	18	×	×	NOUN
fcis-30136	111	19	94.3	94.3	NUM
fcis-30136	111	20	88.4	88.4	NUM
fcis-30136	111	21	project	project	NOUN
fcis-30136	111	22	4	4	NUM
fcis-30136	111	23	√	√	NOUN
fcis-30136	111	24	√	√	NUM
fcis-30136	111	25	√	√	NUM
fcis-30136	111	26	√	√	NUM
fcis-30136	111	27	94.8	94.8	NUM
fcis-30136	111	28	89.8	89.8	NUM
fcis-30136	111	29	ap/%	ap/%	PROPN
fcis-30136	111	30	p/%	p/%	PROPN
fcis-30136	111	31	r/%	r/%	PROPN
fcis-30136	111	32	map/%	map/%	PROPN
fcis-30136	111	33	gflops/109	gflops/109	PROPN
fcis-30136	111	34	defected	defect	VERB
fcis-30136	111	35	diseased	disease	VERB
fcis-30136	111	36	sprouted	sprout	VERB
fcis-30136	111	37	85.0	85.0	NUM
fcis-30136	111	38	87.6	87.6	NUM
fcis-30136	111	39	89.5	89.5	NUM
fcis-30136	111	40	80.3	80.3	NUM
fcis-30136	111	41	77.5	77.5	NUM
fcis-30136	111	42	86.4	86.4	NUM
fcis-30136	111	43	8.1	8.1	NUM
fcis-30136	111	44	82.7	82.7	NUM
fcis-30136	111	45	89.1	89.1	NUM
fcis-30136	111	46	91.8	91.8	NUM
fcis-30136	111	47	82.2	82.2	NUM
fcis-30136	111	48	82.9	82.9	NUM
fcis-30136	111	49	89.1	89.1	NUM
fcis-30136	111	50	7.1	7.1	NUM
fcis-30136	111	51	82.0	82.0	NUM
fcis-30136	111	52	88.3	88.3	NUM
fcis-30136	111	53	91.1	91.1	NUM
fcis-30136	111	54	82.2	82.2	NUM
fcis-30136	111	55	83.4	83.4	NUM
fcis-30136	111	56	88.8	88.8	NUM
fcis-30136	111	57	6.4	6.4	NUM
fcis-30136	111	58	84.6	84.6	NUM
fcis-30136	111	59	89.0	89.0	NUM
fcis-30136	111	60	91.2	91.2	NUM
fcis-30136	111	61	82.5	82.5	NUM
fcis-30136	111	62	89.5	89.5	NUM
fcis-30136	111	63	89.5	89.5	NUM
fcis-30136	111	64	5.8	5.8	NUM
fcis-30136	111	65	82.5	82.5	NUM
fcis-30136	111	66	89.5	89.5	NUM
fcis-30136	111	67	91.9	91.9	NUM
fcis-30136	111	68	82.2	82.2	NUM
fcis-30136	111	69	89.7	89.7	NUM
fcis-30136	111	70	89.7	89.7	NUM
fcis-30136	111	71	5.9	5.9	NUM
fcis-30136	111	72	4.5	4.5	NUM
fcis-30136	111	73	.	.	PUNCT
fcis-30136	112	1	controlled	control	VERB
fcis-30136	112	2	experiment	experiment	NOUN
fcis-30136	112	3	to	to	PART
fcis-30136	112	4	further	far	ADV
fcis-30136	112	5	validate	validate	VERB
fcis-30136	112	6	the	the	DET
fcis-30136	112	7	effectiveness	effectiveness	NOUN
fcis-30136	112	8	of	of	ADP
fcis-30136	112	9	the	the	DET
fcis-30136	112	10	improved	improved	ADJ
fcis-30136	112	11	algorithm	algorithm	NOUN
fcis-30136	112	12	in	in	ADP
fcis-30136	112	13	enhancing	enhance	VERB
fcis-30136	112	14	performance	performance	NOUN
fcis-30136	112	15	,	,	PUNCT
fcis-30136	112	16	we	we	PRON
fcis-30136	112	17	compared	compare	VERB
fcis-30136	112	18	it	it	PRON
fcis-30136	112	19	with	with	ADP
fcis-30136	112	20	ssd	ssd	NOUN
fcis-30136	112	21	[	[	X
fcis-30136	112	22	12	12	NUM
fcis-30136	112	23	]	]	PUNCT
fcis-30136	112	24	,	,	PUNCT
fcis-30136	112	25	faster	fast	ADJ
fcis-30136	112	26	-	-	PUNCT
fcis-30136	112	27	rcnn	rcnn	NOUN
fcis-30136	113	1	[	[	X
fcis-30136	113	2	13	13	NUM
fcis-30136	113	3	]	]	PUNCT
fcis-30136	113	4	,	,	PUNCT
fcis-30136	113	5	yolov5[14	yolov5[14	PROPN
fcis-30136	113	6	]	]	X
fcis-30136	113	7	,	,	PUNCT
fcis-30136	113	8	and	and	CCONJ
fcis-30136	113	9	yolov10	yolov10	NOUN
fcis-30136	114	1	[	[	X
fcis-30136	114	2	15	15	NUM
fcis-30136	114	3	]	]	PUNCT
fcis-30136	114	4	,	,	PUNCT
fcis-30136	114	5	with	with	ADP
fcis-30136	114	6	the	the	DET
fcis-30136	114	7	comparison	comparison	NOUN
fcis-30136	114	8	results	result	NOUN
fcis-30136	114	9	shown	show	VERB
fcis-30136	114	10	in	in	ADP
fcis-30136	114	11	table	table	NOUN
fcis-30136	114	12	2	2	NUM
fcis-30136	114	13	.	.	PUNCT
fcis-30136	115	1	the	the	DET
fcis-30136	115	2	comparative	comparative	ADJ
fcis-30136	115	3	experiments	experiment	NOUN
fcis-30136	115	4	demonstrate	demonstrate	VERB
fcis-30136	115	5	that	that	SCONJ
fcis-30136	115	6	the	the	DET
fcis-30136	115	7	improved	improved	ADJ
fcis-30136	115	8	algorithm	algorithm	NOUN
fcis-30136	115	9	outperforms	outperform	VERB
fcis-30136	115	10	others	other	NOUN
fcis-30136	115	11	in	in	ADP
fcis-30136	115	12	terms	term	NOUN
fcis-30136	115	13	of	of	ADP
fcis-30136	115	14	recognition	recognition	NOUN
fcis-30136	115	15	accuracy	accuracy	NOUN
fcis-30136	115	16	,	,	PUNCT
fcis-30136	115	17	mean	mean	ADJ
fcis-30136	115	18	average	average	ADJ
fcis-30136	115	19	precision	precision	NOUN
fcis-30136	115	20	(	(	PUNCT
fcis-30136	115	21	map	map	NOUN
fcis-30136	115	22	)	)	PUNCT
fcis-30136	115	23	,	,	PUNCT
fcis-30136	115	24	and	and	CCONJ
fcis-30136	115	25	computational	computational	ADJ
fcis-30136	115	26	complexity	complexity	NOUN
fcis-30136	115	27	across	across	ADP
fcis-30136	115	28	all	all	DET
fcis-30136	115	29	categories	category	NOUN
fcis-30136	115	30	.	.	PUNCT
fcis-30136	116	1	the	the	DET
fcis-30136	116	2	complexity	complexity	NOUN
fcis-30136	116	3	of	of	ADP
fcis-30136	116	4	the	the	DET
fcis-30136	116	5	faster	fast	ADJ
fcis-30136	116	6	-	-	PUNCT
fcis-30136	116	7	rcnn	rcnn	NOUN
fcis-30136	116	8	algorithm	algorithm	PROPN
fcis-30136	116	9	significantly	significantly	ADV
fcis-30136	116	10	exceeds	exceed	VERB
fcis-30136	116	11	that	that	PRON
fcis-30136	116	12	of	of	ADP
fcis-30136	116	13	other	other	ADJ
fcis-30136	116	14	methods	method	NOUN
fcis-30136	116	15	,	,	PUNCT
fcis-30136	116	16	while	while	SCONJ
fcis-30136	116	17	its	its	PRON
fcis-30136	116	18	map	map	NOUN
fcis-30136	116	19	and	and	CCONJ
fcis-30136	116	20	fps	fps	VERB
fcis-30136	116	21	show	show	VERB
fcis-30136	116	22	no	no	DET
fcis-30136	116	23	clear	clear	ADJ
fcis-30136	116	24	advantages	advantage	NOUN
fcis-30136	116	25	,	,	PUNCT
fcis-30136	116	26	indicating	indicate	VERB
fcis-30136	116	27	poor	poor	ADJ
fcis-30136	116	28	performance	performance	NOUN
fcis-30136	116	29	on	on	ADP
fcis-30136	116	30	our	our	PRON
fcis-30136	116	31	dataset	dataset	NOUN
fcis-30136	116	32	.	.	PUNCT
fcis-30136	117	1	compared	compare	VERB
fcis-30136	117	2	to	to	ADP
fcis-30136	117	3	the	the	DET
fcis-30136	117	4	ssd	ssd	NOUN
fcis-30136	117	5	algorithm	algorithm	NOUN
fcis-30136	117	6	,	,	PUNCT
fcis-30136	117	7	the	the	DET
fcis-30136	117	8	improved	improved	ADJ
fcis-30136	117	9	algorithm	algorithm	NOUN
fcis-30136	117	10	achieves	achieve	VERB
fcis-30136	117	11	a	a	DET
fcis-30136	117	12	16	16	NUM
fcis-30136	117	13	%	%	NOUN
fcis-30136	117	14	increase	increase	NOUN
fcis-30136	117	15	in	in	ADP
fcis-30136	117	16	map	map	NOUN
fcis-30136	117	17	and	and	CCONJ
fcis-30136	117	18	a	a	DET
fcis-30136	117	19	90.6	90.6	NUM
fcis-30136	117	20	%	%	NOUN
fcis-30136	117	21	reduction	reduction	NOUN
fcis-30136	117	22	in	in	ADP
fcis-30136	117	23	computational	computational	ADJ
fcis-30136	117	24	load	load	NOUN
fcis-30136	117	25	.	.	PUNCT
fcis-30136	118	1	although	although	SCONJ
fcis-30136	118	2	yolov5	yolov5	NOUN
fcis-30136	118	3	and	and	CCONJ
fcis-30136	118	4	yolov10	yolov10	NOUN
fcis-30136	118	5	,	,	PUNCT
fcis-30136	118	6	as	as	ADP
fcis-30136	118	7	members	member	NOUN
fcis-30136	118	8	of	of	ADP
fcis-30136	118	9	the	the	DET
fcis-30136	118	10	yolov8	yolov8	PROPN
fcis-30136	118	11	series	series	PROPN
fcis-30136	118	12	,	,	PUNCT
fcis-30136	118	13	exhibit	exhibit	VERB
fcis-30136	118	14	comparable	comparable	ADJ
fcis-30136	118	15	performance	performance	NOUN
fcis-30136	118	16	to	to	ADP
fcis-30136	118	17	the	the	DET
fcis-30136	118	18	improved	improved	ADJ
fcis-30136	118	19	algorithm	algorithm	NOUN
fcis-30136	118	20	in	in	ADP
fcis-30136	118	21	most	most	ADJ
fcis-30136	118	22	aspects	aspect	NOUN
fcis-30136	118	23	,	,	PUNCT
fcis-30136	118	24	the	the	DET
fcis-30136	118	25	proposed	propose	VERB
fcis-30136	118	26	method	method	NOUN
fcis-30136	118	27	still	still	ADV
fcis-30136	118	28	shows	show	VERB
fcis-30136	118	29	improvements	improvement	NOUN
fcis-30136	118	30	:	:	PUNCT
fcis-30136	118	31	map	map	VERB
fcis-30136	118	32	increases	increase	NOUN
fcis-30136	118	33	by	by	ADP
fcis-30136	118	34	1.9	1.9	NUM
fcis-30136	118	35	%	%	NOUN
fcis-30136	118	36	and	and	CCONJ
fcis-30136	118	37	2.3	2.3	NUM
fcis-30136	118	38	%	%	NOUN
fcis-30136	118	39	,	,	PUNCT
fcis-30136	118	40	respectively	respectively	ADV
fcis-30136	118	41	,	,	PUNCT
fcis-30136	118	42	while	while	SCONJ
fcis-30136	118	43	computational	computational	ADJ
fcis-30136	118	44	load	load	NOUN
fcis-30136	118	45	decreases	decrease	VERB
fcis-30136	118	46	by	by	ADP
fcis-30136	118	47	16.9	16.9	NUM
fcis-30136	118	48	%	%	NOUN
fcis-30136	118	49	and	and	CCONJ
fcis-30136	118	50	28.1	28.1	NUM
fcis-30136	118	51	%	%	NOUN
fcis-30136	118	52	.	.	PUNCT
fcis-30136	119	1	these	these	DET
fcis-30136	119	2	results	result	NOUN
fcis-30136	119	3	confirm	confirm	VERB
fcis-30136	119	4	that	that	SCONJ
fcis-30136	119	5	our	our	PRON
fcis-30136	119	6	improved	improved	ADJ
fcis-30136	119	7	algorithm	algorithm	NOUN
fcis-30136	119	8	demonstrates	demonstrate	VERB
fcis-30136	119	9	strong	strong	ADJ
fcis-30136	119	10	potato	potato	NOUN
fcis-30136	119	11	recognition	recognition	NOUN
fcis-30136	119	12	capabilities	capability	NOUN
fcis-30136	119	13	.	.	PUNCT
fcis-30136	120	1	despite	despite	SCONJ
fcis-30136	120	2	a	a	DET
fcis-30136	120	3	slight	slight	ADJ
fcis-30136	120	4	reduction	reduction	NOUN
fcis-30136	120	5	in	in	ADP
fcis-30136	120	6	fps	fps	PROPN
fcis-30136	120	7	,	,	PUNCT
fcis-30136	120	8	its	its	PRON
fcis-30136	120	9	detection	detection	NOUN
fcis-30136	120	10	accuracy	accuracy	NOUN
fcis-30136	120	11	and	and	CCONJ
fcis-30136	120	12	computational	computational	ADJ
fcis-30136	120	13	efficiency	efficiency	NOUN
fcis-30136	120	14	still	still	ADV
fcis-30136	120	15	meet	meet	VERB
fcis-30136	120	16	practical	practical	ADJ
fcis-30136	120	17	requirements	requirement	NOUN
fcis-30136	120	18	,	,	PUNCT
fcis-30136	120	19	making	make	VERB
fcis-30136	120	20	it	it	PRON
fcis-30136	120	21	more	more	ADV
fcis-30136	120	22	suitable	suitable	ADJ
fcis-30136	120	23	for	for	ADP
fcis-30136	120	24	deployment	deployment	NOUN
fcis-30136	120	25	on	on	ADP
fcis-30136	120	26	resource	resource	NOUN
fcis-30136	120	27	-	-	PUNCT
fcis-30136	120	28	constrained	constrain	VERB
fcis-30136	120	29	embedded	embed	VERB
fcis-30136	120	30	devices	device	NOUN
fcis-30136	120	31	.	.	PUNCT
fcis-30136	121	1	90	90	NUM
fcis-30136	121	2	table	table	NOUN
fcis-30136	121	3	2	2	NUM
fcis-30136	121	4	.	.	PUNCT
fcis-30136	121	5	comparative	comparative	ADJ
fcis-30136	121	6	experiment	experiment	NOUN
fcis-30136	121	7	results	result	NOUN
fcis-30136	121	8	project	project	NOUN
fcis-30136	121	9	ap/%	ap/%	PROPN
fcis-30136	121	10	map/%	map/%	PROPN
fcis-30136	121	11	gflops/109	gflops/109	PROPN
fcis-30136	121	12	fps	fps	PROPN
fcis-30136	121	13	/	/	SYM
fcis-30136	121	14	ms	ms	PROPN
fcis-30136	121	15	potato	potato	NOUN
fcis-30136	121	16	damaged	damage	VERB
fcis-30136	121	17	defected	defect	VERB
fcis-30136	121	18	diseased	diseased	ADJ
fcis-30136	121	19	sprouted	sprout	VERB
fcis-30136	121	20	ssd	ssd	NOUN
fcis-30136	121	21	79.9	79.9	NUM
fcis-30136	121	22	74.3	74.3	NUM
fcis-30136	121	23	72.6	72.6	NUM
fcis-30136	121	24	63.1	63.1	NUM
fcis-30136	121	25	78.4	78.4	NUM
fcis-30136	121	26	73.7	73.7	NUM
fcis-30136	121	27	62.7	62.7	NUM
fcis-30136	121	28	134	134	NUM
fcis-30136	121	29	faster	fast	ADJ
fcis-30136	121	30	-	-	PUNCT
fcis-30136	121	31	rcnn	rcnn	NOUN
fcis-30136	121	32	82.6	82.6	NUM
fcis-30136	121	33	75.3	75.3	NUM
fcis-30136	121	34	71.5	71.5	NUM
fcis-30136	121	35	61.2	61.2	NUM
fcis-30136	121	36	82.3	82.3	NUM
fcis-30136	121	37	74.6	74.6	NUM
fcis-30136	121	38	369.8	369.8	NUM
fcis-30136	121	39	72	72	NUM
fcis-30136	121	40	yolov5	yolov5	NOUN
fcis-30136	121	41	94.2	94.2	NUM
fcis-30136	121	42	88.0	88.0	NUM
fcis-30136	121	43	78.4	78.4	NUM
fcis-30136	121	44	88.8	88.8	NUM
fcis-30136	121	45	89.6	89.6	NUM
fcis-30136	121	46	87.8	87.8	NUM
fcis-30136	121	47	7.1	7.1	NUM
fcis-30136	121	48	158	158	NUM
fcis-30136	121	49	yolov10	yolov10	NOUN
fcis-30136	121	50	91.5	91.5	NUM
fcis-30136	121	51	88.1	88.1	NUM
fcis-30136	121	52	81.2	81.2	NUM
fcis-30136	121	53	88.7	88.7	NUM
fcis-30136	121	54	87.4	87.4	NUM
fcis-30136	121	55	87.4	87.4	NUM
fcis-30136	121	56	8.2	8.2	NUM
fcis-30136	121	57	119	119	NUM
fcis-30136	121	58	improved	improved	ADJ
fcis-30136	121	59	yolov8	yolov8	NOUN
fcis-30136	121	60	94.8	94.8	NUM
fcis-30136	121	61	89.8	89.8	NUM
fcis-30136	121	62	82.5	82.5	NUM
fcis-30136	121	63	89.5	89.5	NUM
fcis-30136	121	64	91.9	91.9	NUM
fcis-30136	121	65	89.7	89.7	NUM
fcis-30136	121	66	5.9	5.9	NUM
fcis-30136	121	67	61	61	NUM
fcis-30136	121	68	5	5	NUM
fcis-30136	121	69	.	.	PUNCT
fcis-30136	121	70	conclusion	conclusion	NOUN
fcis-30136	121	71	the	the	DET
fcis-30136	121	72	improved	improved	ADJ
fcis-30136	121	73	yolov8	yolov8	NOUN
fcis-30136	121	74	algorithm	algorithm	PROPN
fcis-30136	121	75	was	be	AUX
fcis-30136	121	76	subjected	subject	VERB
fcis-30136	121	77	to	to	ADP
fcis-30136	121	78	various	various	ADJ
fcis-30136	121	79	comparative	comparative	ADJ
fcis-30136	121	80	experiments	experiment	NOUN
fcis-30136	121	81	and	and	CCONJ
fcis-30136	121	82	ablation	ablation	NOUN
fcis-30136	121	83	experiments	experiment	NOUN
fcis-30136	121	84	.	.	PUNCT
fcis-30136	122	1	the	the	DET
fcis-30136	122	2	enhanced	enhanced	ADJ
fcis-30136	122	3	algorithm	algorithm	NOUN
fcis-30136	122	4	achieved	achieve	VERB
fcis-30136	122	5	a	a	DET
fcis-30136	122	6	detection	detection	NOUN
fcis-30136	122	7	accuracy	accuracy	NOUN
fcis-30136	122	8	of	of	ADP
fcis-30136	122	9	89.7	89.7	NUM
fcis-30136	122	10	%	%	NOUN
fcis-30136	122	11	,	,	PUNCT
fcis-30136	122	12	representing	represent	VERB
fcis-30136	122	13	a	a	DET
fcis-30136	122	14	3.3	3.3	NUM
fcis-30136	122	15	%	%	NOUN
fcis-30136	122	16	improvement	improvement	NOUN
fcis-30136	122	17	over	over	ADP
fcis-30136	122	18	the	the	DET
fcis-30136	122	19	original	original	ADJ
fcis-30136	122	20	algorithm	algorithm	NOUN
fcis-30136	122	21	while	while	SCONJ
fcis-30136	122	22	reducing	reduce	VERB
fcis-30136	122	23	instances	instance	NOUN
fcis-30136	122	24	of	of	ADP
fcis-30136	122	25	missed	miss	VERB
fcis-30136	122	26	and	and	CCONJ
fcis-30136	122	27	false	false	ADJ
fcis-30136	122	28	detections	detection	NOUN
fcis-30136	122	29	.	.	PUNCT
fcis-30136	123	1	to	to	PART
fcis-30136	123	2	validate	validate	VERB
fcis-30136	123	3	the	the	DET
fcis-30136	123	4	robustness	robustness	NOUN
fcis-30136	123	5	and	and	CCONJ
fcis-30136	123	6	generalizability	generalizability	NOUN
fcis-30136	123	7	of	of	ADP
fcis-30136	123	8	the	the	DET
fcis-30136	123	9	improved	improved	ADJ
fcis-30136	123	10	algorithm	algorithm	NOUN
fcis-30136	123	11	,	,	PUNCT
fcis-30136	123	12	multiple	multiple	ADJ
fcis-30136	123	13	experiments	experiment	NOUN
fcis-30136	123	14	were	be	AUX
fcis-30136	123	15	conducted	conduct	VERB
fcis-30136	123	16	on	on	ADP
fcis-30136	123	17	a	a	DET
fcis-30136	123	18	potato	potato	NOUN
fcis-30136	123	19	dataset	dataset	NOUN
fcis-30136	123	20	.	.	PUNCT
fcis-30136	124	1	the	the	DET
fcis-30136	124	2	modified	modify	VERB
fcis-30136	124	3	algorithm	algorithm	NOUN
fcis-30136	124	4	demonstrated	demonstrate	VERB
fcis-30136	124	5	enhanced	enhanced	ADJ
fcis-30136	124	6	detection	detection	NOUN
fcis-30136	124	7	sensitivity	sensitivity	NOUN
fcis-30136	124	8	and	and	CCONJ
fcis-30136	124	9	accuracy	accuracy	NOUN
fcis-30136	124	10	in	in	ADP
fcis-30136	124	11	potato	potato	NOUN
fcis-30136	124	12	classification	classification	NOUN
fcis-30136	124	13	tasks	task	NOUN
fcis-30136	124	14	,	,	PUNCT
fcis-30136	124	15	with	with	ADP
fcis-30136	124	16	a	a	DET
fcis-30136	124	17	27.2	27.2	NUM
fcis-30136	124	18	%	%	NOUN
fcis-30136	124	19	reduction	reduction	NOUN
fcis-30136	124	20	in	in	ADP
fcis-30136	124	21	computational	computational	ADJ
fcis-30136	124	22	load	load	NOUN
fcis-30136	124	23	compared	compare	VERB
fcis-30136	124	24	to	to	ADP
fcis-30136	124	25	the	the	DET
fcis-30136	124	26	original	original	ADJ
fcis-30136	124	27	algorithm	algorithm	NOUN
fcis-30136	124	28	,	,	PUNCT
fcis-30136	124	29	making	make	VERB
fcis-30136	124	30	it	it	PRON
fcis-30136	124	31	more	more	ADV
fcis-30136	124	32	suitable	suitable	ADJ
fcis-30136	124	33	for	for	ADP
fcis-30136	124	34	deployment	deployment	NOUN
fcis-30136	124	35	on	on	ADP
fcis-30136	124	36	resource	resource	NOUN
fcis-30136	124	37	-	-	PUNCT
fcis-30136	124	38	constrained	constrain	VERB
fcis-30136	124	39	embedded	embed	VERB
fcis-30136	124	40	devices	device	NOUN
fcis-30136	124	41	.	.	PUNCT
fcis-30136	125	1	however	however	ADV
fcis-30136	125	2	,	,	PUNCT
fcis-30136	125	3	the	the	DET
fcis-30136	125	4	improved	improved	ADJ
fcis-30136	125	5	algorithm	algorithm	NOUN
fcis-30136	125	6	still	still	ADV
fcis-30136	125	7	exhibits	exhibit	VERB
fcis-30136	125	8	limitations	limitation	NOUN
fcis-30136	125	9	in	in	ADP
fcis-30136	125	10	detection	detection	NOUN
fcis-30136	125	11	speed	speed	NOUN
fcis-30136	125	12	.	.	PUNCT
fcis-30136	126	1	future	future	ADJ
fcis-30136	126	2	research	research	NOUN
fcis-30136	126	3	will	will	AUX
fcis-30136	126	4	focus	focus	VERB
fcis-30136	126	5	on	on	ADP
fcis-30136	126	6	employing	employ	VERB
fcis-30136	126	7	more	more	ADV
fcis-30136	126	8	flexible	flexible	ADJ
fcis-30136	126	9	backbone	backbone	NOUN
fcis-30136	126	10	networks	network	NOUN
fcis-30136	126	11	and	and	CCONJ
fcis-30136	126	12	image	image	NOUN
fcis-30136	126	13	preprocessing	preprocesse	VERB
fcis-30136	126	14	techniques	technique	NOUN
fcis-30136	126	15	to	to	PART
fcis-30136	126	16	refine	refine	VERB
fcis-30136	126	17	the	the	DET
fcis-30136	126	18	algorithm	algorithm	NOUN
fcis-30136	126	19	.	.	PUNCT
fcis-30136	127	1	the	the	DET
fcis-30136	127	2	goal	goal	NOUN
fcis-30136	127	3	is	be	AUX
fcis-30136	127	4	to	to	PART
fcis-30136	127	5	optimize	optimize	VERB
fcis-30136	127	6	it	it	PRON
fcis-30136	127	7	into	into	ADP
fcis-30136	127	8	a	a	DET
fcis-30136	127	9	more	more	ADV
fcis-30136	127	10	lightweight	lightweight	ADJ
fcis-30136	127	11	and	and	CCONJ
fcis-30136	127	12	faster	fast	ADJ
fcis-30136	127	13	detection	detection	NOUN
fcis-30136	127	14	model	model	NOUN
fcis-30136	127	15	while	while	SCONJ
fcis-30136	127	16	maintaining	maintain	VERB
fcis-30136	127	17	the	the	DET
fcis-30136	127	18	current	current	ADJ
fcis-30136	127	19	level	level	NOUN
fcis-30136	127	20	of	of	ADP
fcis-30136	127	21	detection	detection	NOUN
fcis-30136	127	22	accuracy	accuracy	NOUN
fcis-30136	127	23	.	.	PUNCT
fcis-30136	128	1	references	reference	NOUN
fcis-30136	128	2	[	[	X
fcis-30136	128	3	1	1	NUM
fcis-30136	128	4	]	]	PUNCT
fcis-30136	128	5	tang	tang	X
fcis-30136	128	6	jianzhao	jianzhao	PROPN
fcis-30136	128	7	,	,	PUNCT
fcis-30136	128	8	wang	wang	PROPN
fcis-30136	128	9	jing	jing	PROPN
fcis-30136	128	10	,	,	PUNCT
fcis-30136	128	11	xiao	xiao	PROPN
fcis-30136	128	12	dengpan	dengpan	PROPN
fcis-30136	128	13	,	,	PUNCT
fcis-30136	128	14	et	et	PROPN
fcis-30136	128	15	al	al	PROPN
fcis-30136	128	16	.	.	PUNCT
fcis-30136	128	17	research	research	NOUN
fcis-30136	128	18	progress	progress	NOUN
fcis-30136	128	19	and	and	CCONJ
fcis-30136	128	20	development	development	NOUN
fcis-30136	128	21	prospects	prospect	NOUN
fcis-30136	128	22	of	of	ADP
fcis-30136	128	23	potato	potato	NOUN
fcis-30136	128	24	growth	growth	NOUN
fcis-30136	128	25	models	model	NOUN
fcis-30136	129	1	[	[	X
fcis-30136	129	2	j	j	X
fcis-30136	129	3	]	]	X
fcis-30136	129	4	.	.	PUNCT
fcis-30136	130	1	scientia	scientia	PROPN
fcis-30136	130	2	agricultura	agricultura	PROPN
fcis-30136	130	3	sinica,2021,54(05):921	sinica,2021,54(05):921	PROPN
fcis-30136	130	4	-	-	NUM
fcis-30136	130	5	932	932	NUM
fcis-30136	130	6	.	.	PUNCT
fcis-30136	131	1	[	[	X
fcis-30136	131	2	2	2	NUM
fcis-30136	131	3	]	]	X
fcis-30136	131	4	hao	hao	PROPN
fcis-30136	131	5	ruoshi	ruoshi	PROPN
fcis-30136	131	6	,	,	PUNCT
fcis-30136	131	7	lyu	lyu	PROPN
fcis-30136	131	8	jianfei	jianfei	PROPN
fcis-30136	131	9	,	,	PUNCT
fcis-30136	131	10	wang	wang	PROPN
fcis-30136	131	11	aoxue	aoxue	PROPN
fcis-30136	131	12	,	,	PUNCT
fcis-30136	131	13	et	et	PROPN
fcis-30136	131	14	al	al	PROPN
fcis-30136	131	15	.	.	PUNCT
fcis-30136	131	16	new	new	ADJ
fcis-30136	131	17	characteristics	characteristic	NOUN
fcis-30136	131	18	and	and	CCONJ
fcis-30136	131	19	development	development	NOUN
fcis-30136	131	20	trends	trend	NOUN
fcis-30136	131	21	of	of	ADP
fcis-30136	131	22	china	china	PROPN
fcis-30136	131	23	's	's	PART
fcis-30136	131	24	potato	potato	NOUN
fcis-30136	131	25	industry	industry	NOUN
fcis-30136	131	26	under	under	ADP
fcis-30136	131	27	the	the	DET
fcis-30136	131	28	background	background	NOUN
fcis-30136	131	29	of	of	ADP
fcis-30136	131	30	high	high	ADJ
fcis-30136	131	31	-	-	PUNCT
fcis-30136	131	32	quality	quality	NOUN
fcis-30136	131	33	development	development	NOUN
fcis-30136	132	1	[	[	X
fcis-30136	132	2	j	j	X
fcis-30136	132	3	]	]	X
fcis-30136	132	4	.	.	PUNCT
fcis-30136	133	1	agricultural	agricultural	ADJ
fcis-30136	133	2	outlook,2024,20(01):7	outlook,2024,20(01):7	PROPN
fcis-30136	133	3	-	-	PUNCT
fcis-30136	133	4	12	12	NUM
fcis-30136	133	5	.	.	PUNCT
fcis-30136	134	1	[	[	X
fcis-30136	134	2	3	3	NUM
fcis-30136	134	3	]	]	X
fcis-30136	134	4	li	li	PROPN
fcis-30136	134	5	zhonghui	zhonghui	PROPN
fcis-30136	134	6	,	,	PUNCT
fcis-30136	134	7	zhang	zhang	PROPN
fcis-30136	134	8	haocheng	haocheng	PROPN
fcis-30136	134	9	,	,	PUNCT
fcis-30136	134	10	wang	wang	PROPN
fcis-30136	134	11	xiuli	xiuli	PROPN
fcis-30136	134	12	.	.	PUNCT
fcis-30136	135	1	current	current	ADJ
fcis-30136	135	2	status	status	NOUN
fcis-30136	135	3	and	and	CCONJ
fcis-30136	135	4	development	development	NOUN
fcis-30136	135	5	prospects	prospect	NOUN
fcis-30136	135	6	of	of	ADP
fcis-30136	135	7	potato	potato	NOUN
fcis-30136	135	8	supply	supply	NOUN
fcis-30136	135	9	and	and	CCONJ
fcis-30136	135	10	demand	demand	NOUN
fcis-30136	135	11	in	in	ADP
fcis-30136	135	12	china	china	PROPN
fcis-30136	136	1	[	[	X
fcis-30136	136	2	j	j	X
fcis-30136	136	3	]	]	X
fcis-30136	136	4	.	.	PUNCT
fcis-30136	137	1	agricultural	agricultural	ADJ
fcis-30136	137	2	outlook,2022	outlook,2022	NOUN
fcis-30136	137	3	,	,	PUNCT
fcis-30136	137	4	18(06	18(06	NUM
fcis-30136	137	5	):	):	PUNCT
fcis-30136	137	6	79	79	NUM
fcis-30136	137	7	-	-	SYM
fcis-30136	137	8	85	85	NUM
fcis-30136	137	9	.	.	PUNCT
fcis-30136	138	1	[	[	X
fcis-30136	138	2	4	4	X
fcis-30136	138	3	]	]	X
fcis-30136	138	4	liang	liang	PROPN
fcis-30136	138	5	jurong	jurong	PROPN
fcis-30136	138	6	,	,	PUNCT
fcis-30136	138	7	huang	huang	PROPN
fcis-30136	138	8	biao	biao	PROPN
fcis-30136	138	9	,	,	PUNCT
fcis-30136	138	10	tian	tian	PROPN
fcis-30136	138	11	haijiang	haijiang	PROPN
fcis-30136	138	12	,	,	PUNCT
fcis-30136	138	13	et	et	PROPN
fcis-30136	138	14	al	al	PROPN
fcis-30136	138	15	.	.	PUNCT
fcis-30136	138	16	research	research	NOUN
fcis-30136	138	17	status	status	NOUN
fcis-30136	138	18	of	of	ADP
fcis-30136	138	19	automatic	automatic	ADJ
fcis-30136	138	20	potato	potato	NOUN
fcis-30136	138	21	grading	grade	VERB
fcis-30136	138	22	machines	machine	NOUN
fcis-30136	139	1	[	[	X
fcis-30136	139	2	j	j	X
fcis-30136	139	3	]	]	X
fcis-30136	139	4	.	.	PUNCT
fcis-30136	140	1	hans	hans	PROPN
fcis-30136	140	2	journal	journal	PROPN
fcis-30136	140	3	of	of	ADP
fcis-30136	140	4	agricultural	agricultural	ADJ
fcis-30136	140	5	sciences	science	NOUN
fcis-30136	140	6	,	,	PUNCT
fcis-30136	140	7	2021	2021	NUM
fcis-30136	140	8	,	,	PUNCT
fcis-30136	140	9	11	11	NUM
fcis-30136	140	10	:	:	SYM
fcis-30136	140	11	138	138	NUM
fcis-30136	140	12	.	.	PUNCT
fcis-30136	141	1	[	[	X
fcis-30136	141	2	5	5	X
fcis-30136	141	3	]	]	X
fcis-30136	141	4	li	li	PROPN
fcis-30136	141	5	qi	qi	PROPN
fcis-30136	141	6	,	,	PUNCT
fcis-30136	141	7	wang	wang	PROPN
fcis-30136	141	8	jun	jun	PROPN
fcis-30136	141	9	.	.	PROPN
fcis-30136	141	10	potato	potato	NOUN
fcis-30136	141	11	grading	grade	VERB
fcis-30136	141	12	method	method	NOUN
fcis-30136	141	13	based	base	VERB
fcis-30136	141	14	on	on	ADP
fcis-30136	141	15	fusion	fusion	NOUN
fcis-30136	141	16	of	of	ADP
fcis-30136	141	17	color	color	NOUN
fcis-30136	141	18	and	and	CCONJ
fcis-30136	141	19	texture	texture	ADJ
fcis-30136	141	20	features[j	features[j	NOUN
fcis-30136	141	21	]	]	PUNCT
fcis-30136	141	22	.	.	PUNCT
fcis-30136	142	1	science	science	NOUN
fcis-30136	142	2	technology	technology	NOUN
fcis-30136	142	3	and	and	CCONJ
fcis-30136	142	4	engineering	engineering	NOUN
fcis-30136	142	5	,	,	PUNCT
fcis-30136	142	6	2019	2019	NUM
fcis-30136	142	7	,	,	PUNCT
fcis-30136	142	8	19(25	19(25	NUM
fcis-30136	142	9	):	):	PUNCT
fcis-30136	142	10	273	273	NUM
fcis-30136	142	11	-	-	SYM
fcis-30136	142	12	279	279	NUM
fcis-30136	142	13	.	.	PUNCT
fcis-30136	143	1	[	[	X
fcis-30136	143	2	6	6	NUM
fcis-30136	143	3	]	]	X
fcis-30136	143	4	al	al	PROPN
fcis-30136	143	5	-	-	PUNCT
fcis-30136	143	6	kateb	kateb	PROPN
fcis-30136	143	7	g	g	NOUN
fcis-30136	143	8	,	,	PUNCT
fcis-30136	143	9	mijwil	mijwil	VERB
fcis-30136	143	10	m	m	NOUN
fcis-30136	143	11	m	m	NOUN
fcis-30136	143	12	,	,	PUNCT
fcis-30136	143	13	aljanabi	aljanabi	PROPN
fcis-30136	143	14	m	m	PROPN
fcis-30136	143	15	,	,	PUNCT
fcis-30136	143	16	et	et	PROPN
fcis-30136	143	17	al	al	PROPN
fcis-30136	143	18	.	.	PUNCT
fcis-30136	143	19	ai	ai	PROPN
fcis-30136	143	20	-	-	PUNCT
fcis-30136	143	21	potatoguard	potatoguard	NOUN
fcis-30136	143	22	:	:	PUNCT
fcis-30136	143	23	leveraging	leverage	VERB
fcis-30136	143	24	generative	generative	ADJ
fcis-30136	143	25	models	model	NOUN
fcis-30136	143	26	for	for	ADP
fcis-30136	143	27	early	early	ADJ
fcis-30136	143	28	detection	detection	NOUN
fcis-30136	143	29	of	of	ADP
fcis-30136	143	30	potato	potato	NOUN
fcis-30136	143	31	diseases	disease	NOUN
fcis-30136	143	32	[	[	X
fcis-30136	143	33	j	j	X
fcis-30136	143	34	]	]	X
fcis-30136	143	35	.	.	PUNCT
fcis-30136	144	1	potato	potato	NOUN
fcis-30136	144	2	research	research	NOUN
fcis-30136	144	3	,	,	PUNCT
fcis-30136	144	4	2024	2024	NUM
fcis-30136	144	5	:	:	PUNCT
fcis-30136	144	6	1	1	NUM
fcis-30136	144	7	-	-	SYM
fcis-30136	144	8	15	15	NUM
fcis-30136	144	9	.	.	PUNCT
fcis-30136	145	1	[	[	X
fcis-30136	145	2	7	7	X
fcis-30136	145	3	]	]	X
fcis-30136	145	4	liu	liu	PROPN
fcis-30136	145	5	yijun	yijun	PROPN
fcis-30136	145	6	.	.	PUNCT
fcis-30136	146	1	research	research	NOUN
fcis-30136	146	2	on	on	ADP
fcis-30136	146	3	potato	potato	NOUN
fcis-30136	146	4	external	external	ADJ
fcis-30136	146	5	quality	quality	NOUN
fcis-30136	146	6	grading	grading	NOUN
fcis-30136	146	7	method	method	NOUN
fcis-30136	146	8	based	base	VERB
fcis-30136	146	9	on	on	ADP
fcis-30136	146	10	machine	machine	NOUN
fcis-30136	146	11	vision	vision	NOUN
fcis-30136	147	1	[	[	X
fcis-30136	147	2	d	d	X
fcis-30136	147	3	]	]	X
fcis-30136	147	4	.	.	PUNCT
fcis-30136	148	1	chinese	chinese	PROPN
fcis-30136	148	2	academy	academy	PROPN
fcis-30136	148	3	of	of	ADP
fcis-30136	148	4	agricultural	agricultural	ADJ
fcis-30136	148	5	mechanization	mechanization	NOUN
fcis-30136	148	6	sciences	science	NOUN
fcis-30136	148	7	,	,	PUNCT
fcis-30136	148	8	doi	doi	NOUN
fcis-30136	148	9	:	:	PUNCT
fcis-30136	148	10	10.27629	10.27629	NUM
fcis-30136	148	11	/	/	SYM
fcis-30136	148	12	d.	d.	NOUN
fcis-30136	148	13	cnki	cnki	PROPN
fcis-30136	148	14	.	.	PUNCT
fcis-30136	148	15	gznjk	gznjk	NOUN
fcis-30136	148	16	.	.	PUNCT
fcis-30136	149	1	2023.000009	2023.000009	NUM
fcis-30136	149	2	.	.	PUNCT
fcis-30136	150	1	[	[	X
fcis-30136	150	2	8	8	NUM
fcis-30136	150	3	]	]	PUNCT
fcis-30136	150	4	akther	akther	PROPN
fcis-30136	150	5	j	j	PROPN
fcis-30136	150	6	,	,	PUNCT
fcis-30136	150	7	nayan	nayan	PROPN
fcis-30136	150	8	a	a	DET
fcis-30136	150	9	a	a	PROPN
fcis-30136	150	10	,	,	PUNCT
fcis-30136	150	11	harun	harun	PROPN
fcis-30136	150	12	-	-	PUNCT
fcis-30136	150	13	or	or	CCONJ
fcis-30136	150	14	-	-	PUNCT
fcis-30136	150	15	roshid	roshid	VERB
fcis-30136	150	16	m.	m.	NOUN
fcis-30136	150	17	potato	potato	NOUN
fcis-30136	150	18	leaves	leave	VERB
fcis-30136	150	19	blight	blight	NOUN
fcis-30136	150	20	disease	disease	NOUN
fcis-30136	150	21	recognition	recognition	NOUN
fcis-30136	150	22	and	and	CCONJ
fcis-30136	150	23	categorization	categorization	NOUN
fcis-30136	150	24	using	use	VERB
fcis-30136	150	25	deep	deep	ADJ
fcis-30136	150	26	learning[j	learning[j	NOUN
fcis-30136	150	27	]	]	PUNCT
fcis-30136	150	28	.	.	PUNCT
fcis-30136	151	1	engineering	engineering	NOUN
fcis-30136	151	2	journal	journal	NOUN
fcis-30136	151	3	,	,	PUNCT
fcis-30136	151	4	2023	2023	NUM
fcis-30136	151	5	,	,	PUNCT
fcis-30136	151	6	27(9	27(9	NUM
fcis-30136	151	7	):	):	PUNCT
fcis-30136	151	8	27	27	NUM
fcis-30136	151	9	-	-	SYM
fcis-30136	151	10	38	38	NUM
fcis-30136	151	11	.	.	PUNCT
fcis-30136	152	1	[	[	X
fcis-30136	152	2	9	9	NUM
fcis-30136	152	3	]	]	X
fcis-30136	152	4	zhang	zhang	PROPN
fcis-30136	152	5	w	w	PROPN
fcis-30136	152	6	,	,	PUNCT
fcis-30136	152	7	yuelin	yuelin	PROPN
fcis-30136	152	8	h	h	PROPN
fcis-30136	152	9	a	a	PRON
fcis-30136	152	10	n	n	NOUN
fcis-30136	152	11	,	,	PUNCT
fcis-30136	152	12	huang	huang	PROPN
fcis-30136	152	13	c	c	PROPN
fcis-30136	152	14	,	,	PUNCT
fcis-30136	152	15	et	et	PROPN
fcis-30136	152	16	al	al	PROPN
fcis-30136	152	17	.	.	PUNCT
fcis-30136	152	18	recognition	recognition	PROPN
fcis-30136	152	19	method	method	NOUN
fcis-30136	152	20	for	for	ADP
fcis-30136	152	21	seed	seed	NOUN
fcis-30136	152	22	potato	potato	NOUN
fcis-30136	152	23	buds	bud	NOUN
fcis-30136	152	24	based	base	VERB
fcis-30136	152	25	on	on	ADP
fcis-30136	152	26	improved	improved	ADJ
fcis-30136	152	27	yolov3tiny[j	yolov3tiny[j	PROPN
fcis-30136	152	28	]	]	PUNCT
fcis-30136	152	29	.	.	PUNCT
fcis-30136	153	1	inmateh	inmateh	NOUN
fcis-30136	153	2	-	-	PUNCT
fcis-30136	153	3	agricultural	agricultural	ADJ
fcis-30136	153	4	engineering	engineering	NOUN
fcis-30136	153	5	,	,	PUNCT
fcis-30136	153	6	2022	2022	NUM
fcis-30136	153	7	,	,	PUNCT
fcis-30136	153	8	67(2	67(2	NUM
fcis-30136	153	9	)	)	PUNCT
fcis-30136	153	10	.	.	PUNCT
fcis-30136	154	1	[	[	X
fcis-30136	154	2	10	10	NUM
fcis-30136	154	3	]	]	X
fcis-30136	154	4	wang	wang	PROPN
fcis-30136	154	5	c	c	PROPN
fcis-30136	154	6	y	y	PROPN
fcis-30136	154	7	,	,	PUNCT
fcis-30136	154	8	bochkovskiy	bochkovskiy	VERB
fcis-30136	154	9	a	a	PRON
fcis-30136	154	10	,	,	PUNCT
fcis-30136	154	11	liao	liao	PROPN
fcis-30136	154	12	h	h	PROPN
fcis-30136	154	13	y	y	PROPN
fcis-30136	154	14	m.	m.	NOUN
fcis-30136	154	15	yolov7	yolov7	NOUN
fcis-30136	154	16	:	:	PUNCT
fcis-30136	154	17	trainable	trainable	ADJ
fcis-30136	154	18	bag	bag	NOUN
fcis-30136	154	19	-	-	PUNCT
fcis-30136	154	20	of	of	ADP
fcis-30136	154	21	-	-	PUNCT
fcis-30136	154	22	freebies	freebie	NOUN
fcis-30136	154	23	sets	set	VERB
fcis-30136	154	24	new	new	ADJ
fcis-30136	154	25	state	state	NOUN
fcis-30136	154	26	-	-	PUNCT
fcis-30136	154	27	of	of	ADP
fcis-30136	154	28	-	-	PUNCT
fcis-30136	154	29	the	the	DET
fcis-30136	154	30	-	-	PUNCT
fcis-30136	154	31	art	art	NOUN
fcis-30136	154	32	for	for	ADP
fcis-30136	154	33	real	real	ADJ
fcis-30136	154	34	-	-	PUNCT
fcis-30136	154	35	time	time	NOUN
fcis-30136	154	36	object	object	NOUN
fcis-30136	154	37	detectors[c]//proceedings	detectors[c]//proceeding	NOUN
fcis-30136	154	38	of	of	ADP
fcis-30136	154	39	the	the	DET
fcis-30136	154	40	ieee	ieee	NOUN
fcis-30136	154	41	/	/	SYM
fcis-30136	154	42	cvf	cvf	NOUN
fcis-30136	154	43	conference	conference	NOUN
fcis-30136	154	44	on	on	ADP
fcis-30136	154	45	computer	computer	NOUN
fcis-30136	154	46	vision	vision	NOUN
fcis-30136	154	47	and	and	CCONJ
fcis-30136	154	48	pattern	pattern	NOUN
fcis-30136	154	49	recognition	recognition	NOUN
fcis-30136	154	50	.	.	PUNCT
fcis-30136	155	1	2023	2023	NUM
fcis-30136	155	2	:	:	PUNCT
fcis-30136	155	3	7464	7464	NUM
fcis-30136	155	4	-	-	SYM
fcis-30136	155	5	7475	7475	NUM
fcis-30136	155	6	.	.	PUNCT
fcis-30136	156	1	[	[	X
fcis-30136	156	2	11	11	NUM
fcis-30136	156	3	]	]	X
fcis-30136	156	4	c	c	PROPN
fcis-30136	156	5	g	g	PROPN
fcis-30136	156	6	s	s	PROPN
fcis-30136	156	7	,	,	PUNCT
fcis-30136	156	8	upadhyay	upadhyay	PROPN
fcis-30136	156	9	a	a	X
fcis-30136	156	10	,	,	PUNCT
fcis-30136	156	11	zhang	zhang	PROPN
fcis-30136	156	12	y	y	PROPN
fcis-30136	156	13	,	,	PUNCT
fcis-30136	156	14	et	et	PROPN
fcis-30136	156	15	al.field	al.field	NOUN
fcis-30136	156	16	-	-	PUNCT
fcis-30136	156	17	based	base	VERB
fcis-30136	156	18	multispecies	multispecie	NOUN
fcis-30136	156	19	weed	weed	VERB
fcis-30136	156	20	and	and	CCONJ
fcis-30136	156	21	crop	crop	NOUN
fcis-30136	156	22	detection	detection	NOUN
fcis-30136	156	23	using	use	VERB
fcis-30136	156	24	ground	ground	NOUN
fcis-30136	156	25	robots	robot	NOUN
fcis-30136	156	26	and	and	CCONJ
fcis-30136	156	27	advanced	advanced	ADJ
fcis-30136	156	28	yolo	yolo	ADJ
fcis-30136	156	29	models	model	NOUN
fcis-30136	156	30	:	:	PUNCT
fcis-30136	156	31	a	a	DET
fcis-30136	156	32	data	datum	NOUN
fcis-30136	156	33	and	and	CCONJ
fcis-30136	156	34	model	model	ADJ
fcis-30136	156	35	-	-	ADJ
fcis-30136	156	36	centric	centric	ADJ
fcis-30136	156	37	approach[j].smart	approach[j].smart	NOUN
fcis-30136	156	38	agricultural	agricultural	ADJ
fcis-30136	156	39	technology	technology	NOUN
fcis-30136	156	40	,	,	PUNCT
fcis-30136	156	41	2024	2024	NUM
fcis-30136	156	42	,	,	PUNCT
fcis-30136	156	43	9100538	9100538	NUM
fcis-30136	156	44	-	-	SYM
fcis-30136	156	45	100538	100538	NUM
fcis-30136	156	46	.	.	PUNCT
fcis-30136	157	1	[	[	X
fcis-30136	157	2	12	12	NUM
fcis-30136	157	3	]	]	X
fcis-30136	157	4	liu	liu	PROPN
fcis-30136	157	5	w	w	PROPN
fcis-30136	157	6	,	,	PUNCT
fcis-30136	157	7	anguelov	anguelov	NOUN
fcis-30136	157	8	d	d	NOUN
fcis-30136	157	9	,	,	PUNCT
fcis-30136	157	10	erhan	erhan	ADP
fcis-30136	157	11	d	d	PROPN
fcis-30136	157	12	,	,	PUNCT
fcis-30136	157	13	et	et	PROPN
fcis-30136	157	14	al	al	PROPN
fcis-30136	157	15	.	.	PROPN
fcis-30136	157	16	ssd	ssd	PROPN
fcis-30136	157	17	:	:	PUNCT
fcis-30136	157	18	single	single	ADJ
fcis-30136	157	19	shot	shot	PROPN
fcis-30136	157	20	multibox	multibox	PROPN
fcis-30136	157	21	detector[c]//computer	detector[c]//computer	PROPN
fcis-30136	157	22	vision	vision	NOUN
fcis-30136	157	23	–	–	PUNCT
fcis-30136	157	24	eccv	eccv	NOUN
fcis-30136	157	25	2016	2016	NUM
fcis-30136	157	26	:	:	PUNCT
fcis-30136	157	27	14th	14th	ADJ
fcis-30136	157	28	european	european	ADJ
fcis-30136	157	29	conference	conference	PROPN
fcis-30136	157	30	,	,	PUNCT
fcis-30136	157	31	amsterdam	amsterdam	PROPN
fcis-30136	157	32	,	,	PUNCT
fcis-30136	157	33	the	the	DET
fcis-30136	157	34	netherlands	netherlands	PROPN
fcis-30136	157	35	,	,	PUNCT
fcis-30136	157	36	october	october	PROPN
fcis-30136	157	37	11–14	11–14	NUM
fcis-30136	157	38	,	,	PUNCT
fcis-30136	157	39	2016	2016	NUM
fcis-30136	157	40	,	,	PUNCT
fcis-30136	157	41	proceedings	proceeding	NOUN
fcis-30136	157	42	,	,	PUNCT
fcis-30136	157	43	part	part	NOUN
fcis-30136	157	44	i	i	PRON
fcis-30136	157	45	14	14	NUM
fcis-30136	157	46	.	.	PUNCT
fcis-30136	158	1	springer	springer	NOUN
fcis-30136	158	2	international	international	ADJ
fcis-30136	158	3	publishing	publishing	NOUN
fcis-30136	158	4	,	,	PUNCT
fcis-30136	158	5	2016	2016	NUM
fcis-30136	158	6	:	:	PUNCT
fcis-30136	158	7	21	21	NUM
fcis-30136	158	8	-	-	SYM
fcis-30136	158	9	37	37	NUM
fcis-30136	158	10	.	.	PUNCT
fcis-30136	159	1	[	[	X
fcis-30136	159	2	13	13	NUM
fcis-30136	159	3	]	]	X
fcis-30136	159	4	shih	shih	PROPN
fcis-30136	159	5	k	k	PROPN
fcis-30136	159	6	h	h	PROPN
fcis-30136	159	7	,	,	PUNCT
fcis-30136	159	8	chiu	chiu	PROPN
fcis-30136	159	9	c	c	PROPN
fcis-30136	159	10	t	t	PROPN
fcis-30136	159	11	,	,	PUNCT
fcis-30136	159	12	lin	lin	PROPN
fcis-30136	159	13	j	j	PROPN
fcis-30136	159	14	a	a	PROPN
fcis-30136	159	15	,	,	PUNCT
fcis-30136	159	16	et	et	PROPN
fcis-30136	159	17	al	al	PROPN
fcis-30136	159	18	.	.	PUNCT
fcis-30136	159	19	real	real	ADJ
fcis-30136	159	20	-	-	PUNCT
fcis-30136	159	21	time	time	NOUN
fcis-30136	159	22	object	object	NOUN
fcis-30136	159	23	detection	detection	NOUN
fcis-30136	159	24	with	with	ADP
fcis-30136	159	25	reduced	reduced	ADJ
fcis-30136	159	26	region	region	NOUN
fcis-30136	159	27	proposal	proposal	NOUN
fcis-30136	159	28	network	network	NOUN
fcis-30136	159	29	via	via	ADP
fcis-30136	159	30	multi	multi	ADJ
fcis-30136	159	31	-	-	ADJ
fcis-30136	159	32	feature	feature	ADJ
fcis-30136	159	33	concatenation[j	concatenation[j	NOUN
fcis-30136	159	34	]	]	PUNCT
fcis-30136	159	35	.	.	PUNCT
fcis-30136	160	1	ieee	ieee	NOUN
fcis-30136	160	2	transactions	transaction	NOUN
fcis-30136	160	3	on	on	ADP
fcis-30136	160	4	neural	neural	ADJ
fcis-30136	160	5	networks	network	NOUN
fcis-30136	160	6	and	and	CCONJ
fcis-30136	160	7	learning	learning	NOUN
fcis-30136	160	8	systems	system	NOUN
fcis-30136	160	9	,	,	PUNCT
fcis-30136	160	10	2019	2019	NUM
fcis-30136	160	11	,	,	PUNCT
fcis-30136	160	12	31(6	31(6	NUM
fcis-30136	160	13	):	):	PUNCT
fcis-30136	160	14	2164	2164	NUM
fcis-30136	160	15	-	-	SYM
fcis-30136	160	16	2173	2173	NUM
fcis-30136	160	17	.	.	PUNCT
fcis-30136	161	1	[	[	X
fcis-30136	161	2	14	14	NUM
fcis-30136	161	3	]	]	X
fcis-30136	161	4	zhu	zhu	PROPN
fcis-30136	161	5	x	x	SYM
fcis-30136	161	6	,	,	PUNCT
fcis-30136	161	7	lyu	lyu	X
fcis-30136	161	8	s	s	X
fcis-30136	161	9	,	,	PUNCT
fcis-30136	161	10	wang	wang	PROPN
fcis-30136	161	11	x	x	PROPN
fcis-30136	161	12	,	,	PUNCT
fcis-30136	161	13	et	et	PROPN
fcis-30136	161	14	al	al	PROPN
fcis-30136	161	15	.	.	PUNCT
fcis-30136	162	1	tph	tph	PROPN
fcis-30136	162	2	-	-	PUNCT
fcis-30136	162	3	yolov5	yolov5	PROPN
fcis-30136	162	4	:	:	PUNCT
fcis-30136	162	5	improved	improved	ADJ
fcis-30136	162	6	yolov5	yolov5	NOUN
fcis-30136	162	7	based	base	VERB
fcis-30136	162	8	on	on	ADP
fcis-30136	162	9	transformer	transformer	NOUN
fcis-30136	162	10	prediction	prediction	NOUN
fcis-30136	162	11	head	head	NOUN
fcis-30136	162	12	for	for	ADP
fcis-30136	162	13	object	object	NOUN
fcis-30136	162	14	detection	detection	NOUN
fcis-30136	162	15	on	on	ADP
fcis-30136	162	16	drone	drone	NOUN
fcis-30136	162	17	-	-	PUNCT
fcis-30136	162	18	captured	capture	VERB
fcis-30136	162	19	scenarios[c]//	scenarios[c]//	PROPN
fcis-30136	162	20	proceedings	proceeding	NOUN
fcis-30136	162	21	of	of	ADP
fcis-30136	162	22	the	the	DET
fcis-30136	162	23	ieee	ieee	NOUN
fcis-30136	162	24	/	/	SYM
fcis-30136	162	25	cvf	cvf	NOUN
fcis-30136	162	26	international	international	ADJ
fcis-30136	162	27	conference	conference	NOUN
fcis-30136	162	28	on	on	ADP
fcis-30136	162	29	computer	computer	NOUN
fcis-30136	162	30	vision	vision	NOUN
fcis-30136	162	31	.	.	PUNCT
fcis-30136	163	1	2021	2021	NUM
fcis-30136	163	2	:	:	PUNCT
fcis-30136	163	3	2778	2778	NUM
fcis-30136	163	4	-	-	SYM
fcis-30136	163	5	2788	2788	NUM
fcis-30136	163	6	.	.	PUNCT
fcis-30136	164	1	[	[	X
fcis-30136	164	2	15	15	NUM
fcis-30136	164	3	]	]	X
fcis-30136	164	4	wang	wang	PROPN
fcis-30136	164	5	a	a	PROPN
fcis-30136	164	6	,	,	PUNCT
fcis-30136	164	7	chen	chen	PROPN
fcis-30136	164	8	h	h	PROPN
fcis-30136	164	9	,	,	PUNCT
fcis-30136	164	10	liu	liu	PROPN
fcis-30136	164	11	l	l	PROPN
fcis-30136	164	12	,	,	PUNCT
fcis-30136	164	13	et	et	PROPN
fcis-30136	164	14	al	al	PROPN
fcis-30136	164	15	.	.	PROPN
fcis-30136	164	16	yolov10	yolov10	PROPN
fcis-30136	164	17	:	:	PUNCT
fcis-30136	164	18	real	real	ADJ
fcis-30136	164	19	-	-	PUNCT
fcis-30136	164	20	time	time	NOUN
fcis-30136	164	21	end	end	NOUN
fcis-30136	164	22	-	-	PUNCT
fcis-30136	164	23	to	to	ADP
fcis-30136	164	24	-	-	PUNCT
fcis-30136	164	25	end	end	NOUN
fcis-30136	164	26	object	object	NOUN
fcis-30136	164	27	detection[j	detection[j	PROPN
fcis-30136	164	28	]	]	PUNCT
fcis-30136	164	29	.	.	PUNCT
fcis-30136	165	1	arxiv	arxiv	PROPN
fcis-30136	165	2	preprint	preprint	PROPN
fcis-30136	165	3	arxiv:2405	arxiv:2405	PROPN
fcis-30136	165	4	.	.	PROPN
fcis-30136	165	5	14458	14458	NUM
fcis-30136	165	6	,	,	PUNCT
fcis-30136	165	7	2024	2024	NUM
fcis-30136	165	8	.	.	PUNCT