id	sid	tid	token	lemma	pos
ajst-29358	1	1	academic	academic	ADJ
ajst-29358	1	2	journal	journal	NOUN
ajst-29358	1	3	of	of	ADP
ajst-29358	1	4	science	science	NOUN
ajst-29358	1	5	and	and	CCONJ
ajst-29358	1	6	technology	technology	NOUN
ajst-29358	1	7	issn	issn	NOUN
ajst-29358	1	8	:	:	PUNCT
ajst-29358	1	9	2771	2771	NUM
ajst-29358	1	10	-	-	SYM
ajst-29358	1	11	3032	3032	NUM
ajst-29358	1	12	|	|	NOUN
ajst-29358	1	13	vol	vol	NOUN
ajst-29358	1	14	.	.	PUNCT
ajst-29358	2	1	14	14	NUM
ajst-29358	2	2	,	,	PUNCT
ajst-29358	2	3	no	no	INTJ
ajst-29358	2	4	.	.	NOUN
ajst-29358	2	5	1	1	NUM
ajst-29358	2	6	,	,	PUNCT
ajst-29358	2	7	2025	2025	NUM
ajst-29358	2	8	150	150	NUM
ajst-29358	2	9	image	image	NOUN
ajst-29358	2	10	recognition	recognition	NOUN
ajst-29358	2	11	method	method	NOUN
ajst-29358	2	12	for	for	ADP
ajst-29358	2	13	corn	corn	NOUN
ajst-29358	2	14	northern	northern	ADJ
ajst-29358	2	15	leaf	leaf	NOUN
ajst-29358	2	16	blight	blight	NOUN
ajst-29358	2	17	of	of	ADP
ajst-29358	2	18	unmanned	unmanned	ADJ
ajst-29358	2	19	aerial	aerial	ADJ
ajst-29358	2	20	vehicles	vehicle	NOUN
ajst-29358	2	21	based	base	VERB
ajst-29358	2	22	on	on	ADP
ajst-29358	2	23	deep	deep	ADJ
ajst-29358	2	24	learning	learning	NOUN
ajst-29358	2	25	xianyuan	xianyuan	PROPN
ajst-29358	2	26	liu	liu	PROPN
ajst-29358	2	27	,	,	PUNCT
ajst-29358	2	28	wei	wei	PROPN
ajst-29358	2	29	su	su	PROPN
ajst-29358	2	30	jilin	jilin	PROPN
ajst-29358	2	31	institute	institute	PROPN
ajst-29358	2	32	of	of	ADP
ajst-29358	2	33	chemical	chemical	PROPN
ajst-29358	2	34	technology	technology	PROPN
ajst-29358	2	35	,	,	PUNCT
ajst-29358	2	36	jilin	jilin	PROPN
ajst-29358	2	37	132000	132000	NUM
ajst-29358	2	38	,	,	PUNCT
ajst-29358	2	39	china	china	PROPN
ajst-29358	2	40	abstract	abstract	NOUN
ajst-29358	2	41	:	:	PUNCT
ajst-29358	2	42	aiming	aim	VERB
ajst-29358	2	43	at	at	ADP
ajst-29358	2	44	the	the	DET
ajst-29358	2	45	problems	problem	NOUN
ajst-29358	2	46	of	of	ADP
ajst-29358	2	47	high	high	ADJ
ajst-29358	2	48	complexity	complexity	NOUN
ajst-29358	2	49	,	,	PUNCT
ajst-29358	2	50	false	false	ADJ
ajst-29358	2	51	detection	detection	NOUN
ajst-29358	2	52	and	and	CCONJ
ajst-29358	2	53	missing	miss	VERB
ajst-29358	2	54	detection	detection	NOUN
ajst-29358	2	55	of	of	ADP
ajst-29358	2	56	maize	maize	NOUN
ajst-29358	2	57	big	big	ADJ
ajst-29358	2	58	spot	spot	NOUN
ajst-29358	2	59	based	base	VERB
ajst-29358	2	60	on	on	ADP
ajst-29358	2	61	uav	uav	PROPN
ajst-29358	2	62	image	image	NOUN
ajst-29358	2	63	,	,	PUNCT
ajst-29358	2	64	an	an	DET
ajst-29358	2	65	improved	improved	ADJ
ajst-29358	2	66	algorithm	algorithm	NOUN
ajst-29358	2	67	for	for	ADP
ajst-29358	2	68	maize	maize	NOUN
ajst-29358	2	69	big	big	ADJ
ajst-29358	2	70	spot	spot	NOUN
ajst-29358	2	71	detection	detection	NOUN
ajst-29358	2	72	was	be	AUX
ajst-29358	2	73	proposed	propose	VERB
ajst-29358	2	74	.	.	PUNCT
ajst-29358	3	1	the	the	DET
ajst-29358	3	2	algorithm	algorithm	NOUN
ajst-29358	3	3	is	be	AUX
ajst-29358	3	4	based	base	VERB
ajst-29358	3	5	on	on	ADP
ajst-29358	3	6	ema	ema	PROPN
ajst-29358	3	7	and	and	CCONJ
ajst-29358	3	8	the	the	DET
ajst-29358	3	9	improved	improved	ADJ
ajst-29358	3	10	yolov8	yolov8	NOUN
ajst-29358	3	11	-	-	PUNCT
ajst-29358	3	12	bifpn	bifpn	PROPN
ajst-29358	3	13	feature	feature	NOUN
ajst-29358	3	14	pyramid	pyramid	NOUN
ajst-29358	3	15	network	network	NOUN
ajst-29358	3	16	.	.	PUNCT
ajst-29358	4	1	by	by	ADP
ajst-29358	4	2	adding	add	VERB
ajst-29358	4	3	the	the	DET
ajst-29358	4	4	efficient	efficient	ADJ
ajst-29358	4	5	multi	multi	ADJ
ajst-29358	4	6	-	-	ADJ
ajst-29358	4	7	scale	scale	ADJ
ajst-29358	4	8	ema	ema	PROPN
ajst-29358	4	9	attention	attention	NOUN
ajst-29358	4	10	mechanism	mechanism	NOUN
ajst-29358	4	11	module	module	NOUN
ajst-29358	4	12	,	,	PUNCT
ajst-29358	4	13	the	the	DET
ajst-29358	4	14	capturing	capture	VERB
ajst-29358	4	15	ability	ability	NOUN
ajst-29358	4	16	of	of	ADP
ajst-29358	4	17	detail	detail	NOUN
ajst-29358	4	18	information	information	NOUN
ajst-29358	4	19	is	be	AUX
ajst-29358	4	20	improved	improve	VERB
ajst-29358	4	21	,	,	PUNCT
ajst-29358	4	22	thus	thus	ADV
ajst-29358	4	23	enhancing	enhance	VERB
ajst-29358	4	24	the	the	DET
ajst-29358	4	25	feature	feature	NOUN
ajst-29358	4	26	extraction	extraction	NOUN
ajst-29358	4	27	ability	ability	NOUN
ajst-29358	4	28	of	of	ADP
ajst-29358	4	29	the	the	DET
ajst-29358	4	30	model	model	NOUN
ajst-29358	4	31	.	.	PUNCT
ajst-29358	5	1	the	the	DET
ajst-29358	5	2	yolov8	yolov8	NOUN
ajst-29358	5	3	structure	structure	NOUN
ajst-29358	5	4	is	be	AUX
ajst-29358	5	5	improved	improve	VERB
ajst-29358	5	6	into	into	ADP
ajst-29358	5	7	yolov8	yolov8	NOUN
ajst-29358	5	8	-	-	PUNCT
ajst-29358	5	9	bifpn	bifpn	PROPN
ajst-29358	5	10	feature	feature	NOUN
ajst-29358	5	11	pyramid	pyramid	NOUN
ajst-29358	5	12	network	network	NOUN
ajst-29358	5	13	by	by	ADP
ajst-29358	5	14	integrating	integrate	VERB
ajst-29358	5	15	bifpn	bifpn	PROPN
ajst-29358	5	16	structure	structure	NOUN
ajst-29358	5	17	,	,	PUNCT
ajst-29358	5	18	which	which	PRON
ajst-29358	5	19	can	can	AUX
ajst-29358	5	20	extract	extract	VERB
ajst-29358	5	21	context	context	NOUN
ajst-29358	5	22	information	information	NOUN
ajst-29358	5	23	more	more	ADV
ajst-29358	5	24	efficiently	efficiently	ADV
ajst-29358	5	25	.	.	PUNCT
ajst-29358	6	1	by	by	ADP
ajst-29358	6	2	introducing	introduce	VERB
ajst-29358	6	3	wlou	wlou	DET
ajst-29358	6	4	loss	loss	NOUN
ajst-29358	6	5	function	function	NOUN
ajst-29358	6	6	,	,	PUNCT
ajst-29358	6	7	low	low	ADJ
ajst-29358	6	8	quality	quality	NOUN
ajst-29358	6	9	samples	sample	NOUN
ajst-29358	6	10	in	in	ADP
ajst-29358	6	11	training	training	NOUN
ajst-29358	6	12	data	datum	NOUN
ajst-29358	6	13	are	be	AUX
ajst-29358	6	14	filtered	filter	VERB
ajst-29358	6	15	effectively	effectively	ADV
ajst-29358	6	16	,	,	PUNCT
ajst-29358	6	17	and	and	CCONJ
ajst-29358	6	18	the	the	DET
ajst-29358	6	19	generalization	generalization	NOUN
ajst-29358	6	20	ability	ability	NOUN
ajst-29358	6	21	of	of	ADP
ajst-29358	6	22	the	the	DET
ajst-29358	6	23	model	model	NOUN
ajst-29358	6	24	is	be	AUX
ajst-29358	6	25	improved	improve	VERB
ajst-29358	6	26	.	.	PUNCT
ajst-29358	7	1	in	in	ADP
ajst-29358	7	2	this	this	DET
ajst-29358	7	3	paper	paper	NOUN
ajst-29358	7	4	,	,	PUNCT
ajst-29358	7	5	the	the	DET
ajst-29358	7	6	accuracy	accuracy	NOUN
ajst-29358	7	7	p	p	NOUN
ajst-29358	7	8	was	be	AUX
ajst-29358	7	9	increased	increase	VERB
ajst-29358	7	10	by	by	ADP
ajst-29358	7	11	2.7	2.7	NUM
ajst-29358	7	12	%	%	NOUN
ajst-29358	7	13	,	,	PUNCT
ajst-29358	7	14	the	the	DET
ajst-29358	7	15	recall	recall	NOUN
ajst-29358	7	16	rate	rate	NOUN
ajst-29358	7	17	r	r	NOUN
ajst-29358	7	18	was	be	AUX
ajst-29358	7	19	increased	increase	VERB
ajst-29358	7	20	by	by	ADP
ajst-29358	7	21	4.7	4.7	NUM
ajst-29358	7	22	%	%	NOUN
ajst-29358	7	23	,	,	PUNCT
ajst-29358	7	24	the	the	DET
ajst-29358	7	25	average	average	ADJ
ajst-29358	7	26	accuracy	accuracy	NOUN
ajst-29358	7	27	map50	map50	PROPN
ajst-29358	7	28	was	be	AUX
ajst-29358	7	29	increased	increase	VERB
ajst-29358	7	30	by	by	ADP
ajst-29358	7	31	3.1	3.1	NUM
ajst-29358	7	32	%	%	NOUN
ajst-29358	7	33	,	,	PUNCT
ajst-29358	7	34	and	and	CCONJ
ajst-29358	7	35	the	the	DET
ajst-29358	7	36	model	model	NOUN
ajst-29358	7	37	size	size	NOUN
ajst-29358	7	38	was	be	AUX
ajst-29358	7	39	reduced	reduce	VERB
ajst-29358	7	40	by	by	ADP
ajst-29358	7	41	2.78	2.78	NUM
ajst-29358	7	42	m.	m.	NOUN
ajst-29358	7	43	compared	compare	VERB
ajst-29358	7	44	with	with	ADP
ajst-29358	7	45	other	other	ADJ
ajst-29358	7	46	yolo	yolo	ADJ
ajst-29358	7	47	algorithms	algorithm	NOUN
ajst-29358	7	48	,	,	PUNCT
ajst-29358	7	49	the	the	DET
ajst-29358	7	50	proposed	propose	VERB
ajst-29358	7	51	method	method	NOUN
ajst-29358	7	52	has	have	VERB
ajst-29358	7	53	significant	significant	ADJ
ajst-29358	7	54	advantages	advantage	NOUN
ajst-29358	7	55	in	in	ADP
ajst-29358	7	56	the	the	DET
ajst-29358	7	57	detection	detection	NOUN
ajst-29358	7	58	of	of	ADP
ajst-29358	7	59	corn	corn	NOUN
ajst-29358	7	60	big	big	ADJ
ajst-29358	7	61	spot	spot	NOUN
ajst-29358	7	62	disease	disease	NOUN
ajst-29358	7	63	.	.	PUNCT
ajst-29358	8	1	keywords	keyword	NOUN
ajst-29358	8	2	:	:	PUNCT
ajst-29358	8	3	ema	ema	NOUN
ajst-29358	8	4	attention	attention	NOUN
ajst-29358	8	5	mechanism	mechanism	NOUN
ajst-29358	8	6	;	;	PUNCT
ajst-29358	8	7	uav	uav	PROPN
ajst-29358	8	8	image	image	NOUN
ajst-29358	8	9	;	;	PUNCT
ajst-29358	8	10	yolov8	yolov8	PROPN
ajst-29358	8	11	-	-	PUNCT
ajst-29358	8	12	bifpn	bifpn	PROPN
ajst-29358	8	13	;	;	PUNCT
ajst-29358	8	14	corn	corn	NOUN
ajst-29358	8	15	large	large	ADJ
ajst-29358	8	16	spot	spot	NOUN
ajst-29358	8	17	disease	disease	NOUN
ajst-29358	8	18	.	.	PUNCT
ajst-29358	9	1	1	1	X
ajst-29358	9	2	.	.	X
ajst-29358	9	3	introduction	introduction	NOUN
ajst-29358	9	4	in	in	ADP
ajst-29358	9	5	china	china	PROPN
ajst-29358	9	6	's	's	PART
ajst-29358	9	7	agricultural	agricultural	ADJ
ajst-29358	9	8	structure	structure	NOUN
ajst-29358	9	9	,	,	PUNCT
ajst-29358	9	10	corn	corn	NOUN
ajst-29358	9	11	is	be	AUX
ajst-29358	9	12	the	the	DET
ajst-29358	9	13	largest	large	ADJ
ajst-29358	9	14	crop	crop	NOUN
ajst-29358	9	15	,	,	PUNCT
ajst-29358	9	16	accounting	account	VERB
ajst-29358	9	17	for	for	ADP
ajst-29358	9	18	nearly	nearly	ADV
ajst-29358	9	19	40	40	NUM
ajst-29358	9	20	%	%	NOUN
ajst-29358	9	21	of	of	ADP
ajst-29358	9	22	the	the	DET
ajst-29358	9	23	total	total	ADJ
ajst-29358	9	24	grain	grain	NOUN
ajst-29358	9	25	output	output	NOUN
ajst-29358	9	26	,	,	PUNCT
ajst-29358	9	27	with	with	ADP
ajst-29358	9	28	a	a	DET
ajst-29358	9	29	planting	planting	NOUN
ajst-29358	9	30	area	area	NOUN
ajst-29358	9	31	of	of	ADP
ajst-29358	9	32	650	650	NUM
ajst-29358	9	33	million	million	NUM
ajst-29358	9	34	mu	mu	NOUN
ajst-29358	9	35	.	.	PUNCT
ajst-29358	10	1	it	it	PRON
ajst-29358	10	2	is	be	AUX
ajst-29358	10	3	the	the	DET
ajst-29358	10	4	foundation	foundation	NOUN
ajst-29358	10	5	of	of	ADP
ajst-29358	10	6	grain	grain	NOUN
ajst-29358	10	7	production	production	NOUN
ajst-29358	10	8	,	,	PUNCT
ajst-29358	10	9	the	the	DET
ajst-29358	10	10	link	link	NOUN
ajst-29358	10	11	between	between	ADP
ajst-29358	10	12	crop	crop	NOUN
ajst-29358	10	13	cultivation	cultivation	NOUN
ajst-29358	10	14	and	and	CCONJ
ajst-29358	10	15	breeding	breeding	NOUN
ajst-29358	10	16	,	,	PUNCT
ajst-29358	10	17	more	more	ADJ
ajst-29358	10	18	than	than	ADP
ajst-29358	10	19	60	60	NUM
ajst-29358	10	20	%	%	NOUN
ajst-29358	10	21	of	of	ADP
ajst-29358	10	22	its	its	PRON
ajst-29358	10	23	output	output	NOUN
ajst-29358	10	24	is	be	AUX
ajst-29358	10	25	used	use	VERB
ajst-29358	10	26	for	for	ADP
ajst-29358	10	27	the	the	DET
ajst-29358	10	28	production	production	NOUN
ajst-29358	10	29	of	of	ADP
ajst-29358	10	30	feed	feed	NOUN
ajst-29358	10	31	,	,	PUNCT
ajst-29358	10	32	and	and	CCONJ
ajst-29358	10	33	it	it	PRON
ajst-29358	10	34	plays	play	VERB
ajst-29358	10	35	an	an	DET
ajst-29358	10	36	important	important	ADJ
ajst-29358	10	37	role	role	NOUN
ajst-29358	10	38	in	in	ADP
ajst-29358	10	39	the	the	DET
ajst-29358	10	40	economy	economy	NOUN
ajst-29358	10	41	,	,	PUNCT
ajst-29358	10	42	sustainable	sustainable	ADJ
ajst-29358	10	43	agriculture	agriculture	NOUN
ajst-29358	10	44	,	,	PUNCT
ajst-29358	10	45	and	and	CCONJ
ajst-29358	10	46	the	the	DET
ajst-29358	10	47	development	development	NOUN
ajst-29358	10	48	of	of	ADP
ajst-29358	10	49	rural	rural	ADJ
ajst-29358	10	50	livelihoods	livelihood	NOUN
ajst-29358	10	51	,	,	PUNCT
ajst-29358	10	52	and	and	CCONJ
ajst-29358	10	53	is	be	AUX
ajst-29358	10	54	also	also	ADV
ajst-29358	10	55	key	key	ADJ
ajst-29358	10	56	to	to	ADP
ajst-29358	10	57	ensuring	ensure	VERB
ajst-29358	10	58	food	food	NOUN
ajst-29358	10	59	security	security	NOUN
ajst-29358	10	60	and	and	CCONJ
ajst-29358	10	61	promoting	promote	VERB
ajst-29358	10	62	sustainable	sustainable	ADJ
ajst-29358	10	63	social	social	ADJ
ajst-29358	10	64	and	and	CCONJ
ajst-29358	10	65	economic	economic	ADJ
ajst-29358	10	66	development	development	NOUN
ajst-29358	10	67	.	.	PUNCT
ajst-29358	11	1	however	however	ADV
ajst-29358	11	2	,	,	PUNCT
ajst-29358	11	3	in	in	ADP
ajst-29358	11	4	recent	recent	ADJ
ajst-29358	11	5	years	year	NOUN
ajst-29358	11	6	,	,	PUNCT
ajst-29358	11	7	the	the	DET
ajst-29358	11	8	frequent	frequent	ADJ
ajst-29358	11	9	occurrence	occurrence	NOUN
ajst-29358	11	10	of	of	ADP
ajst-29358	11	11	corn	corn	NOUN
ajst-29358	11	12	diseases	disease	NOUN
ajst-29358	11	13	,	,	PUNCT
ajst-29358	11	14	especially	especially	ADV
ajst-29358	11	15	the	the	DET
ajst-29358	11	16	large	large	ADJ
ajst-29358	11	17	spot	spot	NOUN
ajst-29358	11	18	disease	disease	NOUN
ajst-29358	11	19	,	,	PUNCT
ajst-29358	11	20	poses	pose	VERB
ajst-29358	11	21	a	a	DET
ajst-29358	11	22	major	major	ADJ
ajst-29358	11	23	threat	threat	NOUN
ajst-29358	11	24	to	to	ADP
ajst-29358	11	25	corn	corn	NOUN
ajst-29358	11	26	leaves	leave	NOUN
ajst-29358	11	27	,	,	PUNCT
ajst-29358	11	28	affecting	affect	VERB
ajst-29358	11	29	photosynthesis	photosynthesis	NOUN
ajst-29358	11	30	and	and	CCONJ
ajst-29358	11	31	thus	thus	ADV
ajst-29358	11	32	threatening	threaten	VERB
ajst-29358	11	33	the	the	DET
ajst-29358	11	34	yield	yield	NOUN
ajst-29358	11	35	and	and	CCONJ
ajst-29358	11	36	quality	quality	NOUN
ajst-29358	11	37	of	of	ADP
ajst-29358	11	38	corn	corn	NOUN
ajst-29358	11	39	.	.	PUNCT
ajst-29358	12	1	due	due	ADP
ajst-29358	12	2	to	to	ADP
ajst-29358	12	3	the	the	DET
ajst-29358	12	4	inefficiency	inefficiency	NOUN
ajst-29358	12	5	,	,	PUNCT
ajst-29358	12	6	high	high	ADJ
ajst-29358	12	7	cost	cost	NOUN
ajst-29358	12	8	,	,	PUNCT
ajst-29358	12	9	subjectivity	subjectivity	NOUN
ajst-29358	12	10	,	,	PUNCT
ajst-29358	12	11	and	and	CCONJ
ajst-29358	12	12	lack	lack	NOUN
ajst-29358	12	13	of	of	ADP
ajst-29358	12	14	timeliness	timeliness	NOUN
ajst-29358	12	15	of	of	ADP
ajst-29358	12	16	traditional	traditional	ADJ
ajst-29358	12	17	manual	manual	ADJ
ajst-29358	12	18	detection	detection	NOUN
ajst-29358	12	19	methods	method	NOUN
ajst-29358	12	20	,	,	PUNCT
ajst-29358	12	21	machine	machine	NOUN
ajst-29358	12	22	vision	vision	NOUN
ajst-29358	12	23	technology	technology	NOUN
ajst-29358	12	24	is	be	AUX
ajst-29358	12	25	introduced	introduce	VERB
ajst-29358	12	26	to	to	PART
ajst-29358	12	27	achieve	achieve	VERB
ajst-29358	12	28	rapid	rapid	ADJ
ajst-29358	12	29	and	and	CCONJ
ajst-29358	12	30	effective	effective	ADJ
ajst-29358	12	31	identification	identification	NOUN
ajst-29358	12	32	of	of	ADP
ajst-29358	12	33	diseases	disease	NOUN
ajst-29358	12	34	,	,	PUNCT
ajst-29358	12	35	thereby	thereby	ADV
ajst-29358	12	36	formulating	formulate	VERB
ajst-29358	12	37	control	control	NOUN
ajst-29358	12	38	measures	measure	NOUN
ajst-29358	12	39	to	to	PART
ajst-29358	12	40	reduce	reduce	VERB
ajst-29358	12	41	the	the	DET
ajst-29358	12	42	negative	negative	ADJ
ajst-29358	12	43	impact	impact	NOUN
ajst-29358	12	44	of	of	ADP
ajst-29358	12	45	diseases	disease	NOUN
ajst-29358	12	46	,	,	PUNCT
ajst-29358	12	47	ensuring	ensure	VERB
ajst-29358	12	48	food	food	NOUN
ajst-29358	12	49	security	security	NOUN
ajst-29358	12	50	,	,	PUNCT
ajst-29358	12	51	and	and	CCONJ
ajst-29358	12	52	promoting	promote	VERB
ajst-29358	12	53	sustainable	sustainable	ADJ
ajst-29358	12	54	agricultural	agricultural	ADJ
ajst-29358	12	55	development	development	NOUN
ajst-29358	12	56	.	.	PUNCT
ajst-29358	13	1	with	with	ADP
ajst-29358	13	2	the	the	DET
ajst-29358	13	3	development	development	NOUN
ajst-29358	13	4	and	and	CCONJ
ajst-29358	13	5	cost	cost	NOUN
ajst-29358	13	6	reduction	reduction	NOUN
ajst-29358	13	7	of	of	ADP
ajst-29358	13	8	uav	uav	PROPN
ajst-29358	13	9	technology	technology	PROPN
ajst-29358	13	10	,	,	PUNCT
ajst-29358	13	11	its	its	PRON
ajst-29358	13	12	flexibility	flexibility	NOUN
ajst-29358	13	13	and	and	CCONJ
ajst-29358	13	14	accessibility	accessibility	NOUN
ajst-29358	13	15	have	have	AUX
ajst-29358	13	16	been	be	AUX
ajst-29358	13	17	widely	widely	ADV
ajst-29358	13	18	applied	apply	VERB
ajst-29358	13	19	in	in	ADP
ajst-29358	13	20	various	various	ADJ
ajst-29358	13	21	fields	field	NOUN
ajst-29358	13	22	,	,	PUNCT
ajst-29358	13	23	and	and	CCONJ
ajst-29358	13	24	its	its	PRON
ajst-29358	13	25	aerial	aerial	ADJ
ajst-29358	13	26	photography	photography	NOUN
ajst-29358	13	27	images	image	NOUN
ajst-29358	13	28	play	play	VERB
ajst-29358	13	29	an	an	DET
ajst-29358	13	30	important	important	ADJ
ajst-29358	13	31	role	role	NOUN
ajst-29358	13	32	in	in	ADP
ajst-29358	13	33	target	target	NOUN
ajst-29358	13	34	detection	detection	NOUN
ajst-29358	13	35	.	.	PUNCT
ajst-29358	14	1	combined	combine	VERB
ajst-29358	14	2	with	with	ADP
ajst-29358	14	3	machine	machine	NOUN
ajst-29358	14	4	vision	vision	NOUN
ajst-29358	14	5	technology	technology	NOUN
ajst-29358	14	6	,	,	PUNCT
ajst-29358	14	7	it	it	PRON
ajst-29358	14	8	can	can	AUX
ajst-29358	14	9	achieve	achieve	VERB
ajst-29358	14	10	rapid	rapid	ADJ
ajst-29358	14	11	and	and	CCONJ
ajst-29358	14	12	effective	effective	ADJ
ajst-29358	14	13	disease	disease	NOUN
ajst-29358	14	14	detection	detection	NOUN
ajst-29358	14	15	,	,	PUNCT
ajst-29358	14	16	providing	provide	VERB
ajst-29358	14	17	innovative	innovative	ADJ
ajst-29358	14	18	scientific	scientific	ADJ
ajst-29358	14	19	and	and	CCONJ
ajst-29358	14	20	technological	technological	ADJ
ajst-29358	14	21	means	mean	NOUN
ajst-29358	14	22	for	for	ADP
ajst-29358	14	23	agricultural	agricultural	ADJ
ajst-29358	14	24	production	production	NOUN
ajst-29358	14	25	.	.	PUNCT
ajst-29358	15	1	at	at	ADP
ajst-29358	15	2	present	present	ADJ
ajst-29358	15	3	,	,	PUNCT
ajst-29358	15	4	yolov8	yolov8	PROPN
ajst-29358	15	5	,	,	PUNCT
ajst-29358	15	6	as	as	ADP
ajst-29358	15	7	a	a	DET
ajst-29358	15	8	more	more	ADV
ajst-29358	15	9	advanced	advanced	ADJ
ajst-29358	15	10	target	target	NOUN
ajst-29358	15	11	detection	detection	NOUN
ajst-29358	15	12	model	model	NOUN
ajst-29358	15	13	,	,	PUNCT
ajst-29358	15	14	has	have	AUX
ajst-29358	15	15	shown	show	VERB
ajst-29358	15	16	great	great	ADJ
ajst-29358	15	17	potential	potential	NOUN
ajst-29358	15	18	in	in	ADP
ajst-29358	15	19	the	the	DET
ajst-29358	15	20	agricultural	agricultural	ADJ
ajst-29358	15	21	field	field	NOUN
ajst-29358	15	22	,	,	PUNCT
ajst-29358	15	23	especially	especially	ADV
ajst-29358	15	24	in	in	ADP
ajst-29358	15	25	crop	crop	NOUN
ajst-29358	15	26	disease	disease	NOUN
ajst-29358	15	27	detection	detection	NOUN
ajst-29358	15	28	.	.	PUNCT
ajst-29358	16	1	lu	lu	VERB
ajst-29358	16	2	et	et	PROPN
ajst-29358	16	3	al	al	PROPN
ajst-29358	16	4	.	.	PROPN
ajst-29358	16	5	proposed	propose	VERB
ajst-29358	16	6	the	the	DET
ajst-29358	16	7	yolov8_rice	yolov8_rice	PROPN
ajst-29358	16	8	model	model	NOUN
ajst-29358	16	9	,	,	PUNCT
ajst-29358	16	10	which	which	PRON
ajst-29358	16	11	significantly	significantly	ADV
ajst-29358	16	12	improved	improve	VERB
ajst-29358	16	13	the	the	DET
ajst-29358	16	14	recognition	recognition	NOUN
ajst-29358	16	15	ability	ability	NOUN
ajst-29358	16	16	of	of	ADP
ajst-29358	16	17	small	small	ADJ
ajst-29358	16	18	-	-	PUNCT
ajst-29358	16	19	sized	sized	ADJ
ajst-29358	16	20	and	and	CCONJ
ajst-29358	16	21	irregularly	irregularly	ADV
ajst-29358	16	22	shaped	shape	VERB
ajst-29358	16	23	disease	disease	NOUN
ajst-29358	16	24	types	type	NOUN
ajst-29358	16	25	by	by	ADP
ajst-29358	16	26	optimizing	optimize	VERB
ajst-29358	16	27	the	the	DET
ajst-29358	16	28	attention	attention	NOUN
ajst-29358	16	29	mechanism	mechanism	NOUN
ajst-29358	16	30	and	and	CCONJ
ajst-29358	16	31	introducing	introduce	VERB
ajst-29358	16	32	deformable	deformable	ADJ
ajst-29358	16	33	convolution	convolution	NOUN
ajst-29358	16	34	.	.	PUNCT
ajst-29358	17	1	in	in	ADP
ajst-29358	17	2	addition	addition	NOUN
ajst-29358	17	3	,	,	PUNCT
ajst-29358	17	4	he	he	PRON
ajst-29358	17	5	et	et	PROPN
ajst-29358	17	6	al	al	PROPN
ajst-29358	17	7	.	.	PROPN
ajst-29358	17	8	proposed	propose	VERB
ajst-29358	17	9	the	the	DET
ajst-29358	17	10	eds	ed	NOUN
ajst-29358	17	11	-	-	PUNCT
ajst-29358	17	12	yolov8	yolov8	NOUN
ajst-29358	17	13	algorithm	algorithm	NOUN
ajst-29358	17	14	,	,	PUNCT
ajst-29358	17	15	which	which	PRON
ajst-29358	17	16	significantly	significantly	ADV
ajst-29358	17	17	improved	improve	VERB
ajst-29358	17	18	the	the	DET
ajst-29358	17	19	accuracy	accuracy	NOUN
ajst-29358	17	20	of	of	ADP
ajst-29358	17	21	dense	dense	ADJ
ajst-29358	17	22	weed	weed	NOUN
ajst-29358	17	23	detection	detection	NOUN
ajst-29358	17	24	.	.	PUNCT
ajst-29358	18	1	sun	sun	PROPN
ajst-29358	18	2	et	et	PROPN
ajst-29358	18	3	al	al	PROPN
ajst-29358	18	4	.	.	PROPN
ajst-29358	18	5	proposed	propose	VERB
ajst-29358	18	6	a	a	DET
ajst-29358	18	7	lightweight	lightweight	ADJ
ajst-29358	18	8	network	network	NOUN
ajst-29358	18	9	to	to	PART
ajst-29358	18	10	improve	improve	VERB
ajst-29358	18	11	the	the	DET
ajst-29358	18	12	yolov8	yolov8	PROPN
ajst-29358	18	13	model	model	NOUN
ajst-29358	18	14	,	,	PUNCT
ajst-29358	18	15	effectively	effectively	ADV
ajst-29358	18	16	reducing	reduce	VERB
ajst-29358	18	17	the	the	DET
ajst-29358	18	18	model	model	NOUN
ajst-29358	18	19	's	's	PART
ajst-29358	18	20	parameter	parameter	NOUN
ajst-29358	18	21	volume	volume	NOUN
ajst-29358	18	22	and	and	CCONJ
ajst-29358	18	23	computational	computational	ADJ
ajst-29358	18	24	complexity	complexity	NOUN
ajst-29358	18	25	while	while	SCONJ
ajst-29358	18	26	maintaining	maintain	VERB
ajst-29358	18	27	accuracy	accuracy	NOUN
ajst-29358	18	28	in	in	ADP
ajst-29358	18	29	the	the	DET
ajst-29358	18	30	detection	detection	NOUN
ajst-29358	18	31	of	of	ADP
ajst-29358	18	32	tobacco	tobacco	NOUN
ajst-29358	18	33	pests	pest	NOUN
ajst-29358	18	34	in	in	ADP
ajst-29358	18	35	complex	complex	ADJ
ajst-29358	18	36	environments	environment	NOUN
ajst-29358	18	37	.	.	PUNCT
ajst-29358	19	1	deng	deng	PROPN
ajst-29358	19	2	pengfei	pengfei	PROPN
ajst-29358	19	3	et	et	PROPN
ajst-29358	19	4	al	al	PROPN
ajst-29358	19	5	.	.	PROPN
ajst-29358	19	6	used	use	VERB
ajst-29358	19	7	the	the	DET
ajst-29358	19	8	fusion	fusion	NOUN
ajst-29358	19	9	structure	structure	NOUN
ajst-29358	19	10	of	of	ADP
ajst-29358	19	11	inception	inception	NOUN
ajst-29358	19	12	and	and	CCONJ
ajst-29358	19	13	vgg	vgg	NOUN
ajst-29358	19	14	as	as	ADP
ajst-29358	19	15	the	the	DET
ajst-29358	19	16	feature	feature	NOUN
ajst-29358	19	17	extraction	extraction	NOUN
ajst-29358	19	18	model	model	NOUN
ajst-29358	19	19	,	,	PUNCT
ajst-29358	19	20	and	and	CCONJ
ajst-29358	19	21	compressed	compress	VERB
ajst-29358	19	22	the	the	DET
ajst-29358	19	23	model	model	NOUN
ajst-29358	19	24	through	through	ADP
ajst-29358	19	25	transfer	transfer	NOUN
ajst-29358	19	26	learning	learning	NOUN
ajst-29358	19	27	,	,	PUNCT
ajst-29358	19	28	channel	channel	NOUN
ajst-29358	19	29	pruning	pruning	NOUN
ajst-29358	19	30	,	,	PUNCT
ajst-29358	19	31	and	and	CCONJ
ajst-29358	19	32	knowledge	knowledge	NOUN
ajst-29358	19	33	distillation	distillation	NOUN
ajst-29358	19	34	,	,	PUNCT
ajst-29358	19	35	achieving	achieve	VERB
ajst-29358	19	36	rapid	rapid	ADJ
ajst-29358	19	37	identification	identification	NOUN
ajst-29358	19	38	of	of	ADP
ajst-29358	19	39	corn	corn	NOUN
ajst-29358	19	40	diseases	disease	NOUN
ajst-29358	19	41	.	.	PUNCT
ajst-29358	20	1	ye	ye	PRON
ajst-29358	20	2	et	et	PROPN
ajst-29358	20	3	al	al	PROPN
ajst-29358	20	4	.	.	PROPN
ajst-29358	20	5	proposed	propose	VERB
ajst-29358	20	6	an	an	DET
ajst-29358	20	7	improved	improved	ADJ
ajst-29358	20	8	yolov8	yolov8	NOUN
ajst-29358	20	9	lightweight	lightweight	ADJ
ajst-29358	20	10	pest	pest	NOUN
ajst-29358	20	11	image	image	NOUN
ajst-29358	20	12	recognition	recognition	NOUN
ajst-29358	20	13	model	model	NOUN
ajst-29358	20	14	,	,	PUNCT
ajst-29358	20	15	which	which	PRON
ajst-29358	20	16	significantly	significantly	ADV
ajst-29358	20	17	improved	improve	VERB
ajst-29358	20	18	the	the	DET
ajst-29358	20	19	performance	performance	NOUN
ajst-29358	20	20	of	of	ADP
ajst-29358	20	21	small	small	ADJ
ajst-29358	20	22	target	target	NOUN
ajst-29358	20	23	detection	detection	NOUN
ajst-29358	20	24	by	by	ADP
ajst-29358	20	25	introducing	introduce	VERB
ajst-29358	20	26	biformer	biformer	NOUN
ajst-29358	20	27	attention	attention	NOUN
ajst-29358	20	28	mechanism	mechanism	NOUN
ajst-29358	20	29	and	and	CCONJ
ajst-29358	20	30	gelan	gelan	NOUN
ajst-29358	20	31	network	network	NOUN
ajst-29358	20	32	.	.	PUNCT
ajst-29358	21	1	liu	liu	PROPN
ajst-29358	21	2	et	et	PROPN
ajst-29358	21	3	al	al	PROPN
ajst-29358	21	4	.	.	PROPN
ajst-29358	21	5	proposed	propose	VERB
ajst-29358	21	6	a	a	DET
ajst-29358	21	7	feature	feature	NOUN
ajst-29358	21	8	-	-	PUNCT
ajst-29358	21	9	enhanced	enhance	VERB
ajst-29358	21	10	detection	detection	NOUN
ajst-29358	21	11	algorithm	algorithm	NOUN
ajst-29358	21	12	based	base	VERB
ajst-29358	21	13	on	on	ADP
ajst-29358	21	14	ssd	ssd	PROPN
ajst-29358	21	15	,	,	PUNCT
ajst-29358	21	16	integrating	integrate	VERB
ajst-29358	21	17	the	the	DET
ajst-29358	21	18	residual	residual	ADJ
ajst-29358	21	19	structure	structure	NOUN
ajst-29358	21	20	of	of	ADP
ajst-29358	21	21	the	the	DET
ajst-29358	21	22	resnet50	resnet50	NOUN
ajst-29358	21	23	network	network	NOUN
ajst-29358	21	24	into	into	ADP
ajst-29358	21	25	the	the	DET
ajst-29358	21	26	backbone	backbone	NOUN
ajst-29358	21	27	network	network	NOUN
ajst-29358	21	28	,	,	PUNCT
ajst-29358	21	29	improving	improve	VERB
ajst-29358	21	30	the	the	DET
ajst-29358	21	31	detection	detection	NOUN
ajst-29358	21	32	ability	ability	NOUN
ajst-29358	21	33	for	for	ADP
ajst-29358	21	34	small	small	ADJ
ajst-29358	21	35	targets	target	NOUN
ajst-29358	21	36	.	.	PUNCT
ajst-29358	22	1	the	the	DET
ajst-29358	22	2	yolov8	yolov8	PROPN
ajst-29358	22	3	model	model	NOUN
ajst-29358	22	4	has	have	AUX
ajst-29358	22	5	also	also	ADV
ajst-29358	22	6	been	be	AUX
ajst-29358	22	7	applied	apply	VERB
ajst-29358	22	8	to	to	ADP
ajst-29358	22	9	agricultural	agricultural	ADJ
ajst-29358	22	10	resource	resource	NOUN
ajst-29358	22	11	monitoring	monitoring	NOUN
ajst-29358	22	12	and	and	CCONJ
ajst-29358	22	13	yield	yield	NOUN
ajst-29358	22	14	estimation	estimation	NOUN
ajst-29358	22	15	.	.	PUNCT
ajst-29358	23	1	yu	yu	PROPN
ajst-29358	23	2	et	et	PROPN
ajst-29358	23	3	al	al	PROPN
ajst-29358	23	4	.	.	PROPN
ajst-29358	23	5	proposed	propose	VERB
ajst-29358	23	6	a	a	DET
ajst-29358	23	7	model	model	NOUN
ajst-29358	23	8	based	base	VERB
ajst-29358	23	9	on	on	ADP
ajst-29358	23	10	improved	improved	ADJ
ajst-29358	23	11	yolov8	yolov8	NOUN
ajst-29358	23	12	,	,	PUNCT
ajst-29358	23	13	using	use	VERB
ajst-29358	23	14	pconv	pconv	ADJ
ajst-29358	23	15	technology	technology	NOUN
ajst-29358	23	16	to	to	PART
ajst-29358	23	17	improve	improve	VERB
ajst-29358	23	18	model	model	NOUN
ajst-29358	23	19	detection	detection	NOUN
ajst-29358	23	20	accuracy	accuracy	NOUN
ajst-29358	23	21	in	in	ADP
ajst-29358	23	22	complex	complex	ADJ
ajst-29358	23	23	situations	situation	NOUN
ajst-29358	23	24	.	.	PUNCT
ajst-29358	24	1	with	with	ADP
ajst-29358	24	2	the	the	DET
ajst-29358	24	3	development	development	NOUN
ajst-29358	24	4	of	of	ADP
ajst-29358	24	5	agricultural	agricultural	ADJ
ajst-29358	24	6	automation	automation	NOUN
ajst-29358	24	7	and	and	CCONJ
ajst-29358	24	8	intelligence	intelligence	NOUN
ajst-29358	24	9	,	,	PUNCT
ajst-29358	24	10	the	the	DET
ajst-29358	24	11	application	application	NOUN
ajst-29358	24	12	of	of	ADP
ajst-29358	24	13	yolov8	yolov8	NOUN
ajst-29358	24	14	algorithm	algorithm	PROPN
ajst-29358	24	15	models	model	NOUN
ajst-29358	24	16	in	in	ADP
ajst-29358	24	17	agricultural	agricultural	ADJ
ajst-29358	24	18	robots	robot	NOUN
ajst-29358	24	19	and	and	CCONJ
ajst-29358	24	20	intelligent	intelligent	ADJ
ajst-29358	24	21	agricultural	agricultural	ADJ
ajst-29358	24	22	equipment	equipment	NOUN
ajst-29358	24	23	is	be	AUX
ajst-29358	24	24	also	also	ADV
ajst-29358	24	25	becoming	become	VERB
ajst-29358	24	26	more	more	ADV
ajst-29358	24	27	and	and	CCONJ
ajst-29358	24	28	more	more	ADV
ajst-29358	24	29	widespread	widespread	ADJ
ajst-29358	24	30	.	.	PUNCT
ajst-29358	25	1	guo	guo	PROPN
ajst-29358	25	2	et	et	PROPN
ajst-29358	25	3	al	al	PROPN
ajst-29358	25	4	.	.	PROPN
ajst-29358	25	5	used	use	VERB
ajst-29358	25	6	the	the	DET
ajst-29358	25	7	improved	improved	ADJ
ajst-29358	25	8	yolov8	yolov8	NOUN
ajst-29358	25	9	algorithm	algorithm	PROPN
ajst-29358	25	10	model	model	NOUN
ajst-29358	25	11	to	to	PART
ajst-29358	25	12	achieve	achieve	VERB
ajst-29358	25	13	early	early	ADJ
ajst-29358	25	14	detection	detection	NOUN
ajst-29358	25	15	of	of	ADP
ajst-29358	25	16	harmful	harmful	ADJ
ajst-29358	25	17	plants	plant	NOUN
ajst-29358	25	18	by	by	ADP
ajst-29358	25	19	combining	combine	VERB
ajst-29358	25	20	multi	multi	ADJ
ajst-29358	25	21	-	-	ADJ
ajst-29358	25	22	source	source	ADJ
ajst-29358	25	23	image	image	NOUN
ajst-29358	25	24	data	datum	NOUN
ajst-29358	25	25	.	.	PUNCT
ajst-29358	26	1	zhu	zhu	PROPN
ajst-29358	26	2	qi	qi	PROPN
ajst-29358	26	3	et	et	PROPN
ajst-29358	26	4	al	al	PROPN
ajst-29358	26	5	.	.	PROPN
ajst-29358	26	6	proposed	propose	VERB
ajst-29358	26	7	a	a	DET
ajst-29358	26	8	dscs	dscs	ADJ
ajst-29358	26	9	-	-	PUNCT
ajst-29358	26	10	yolo	yolo	NOUN
ajst-29358	26	11	-	-	PUNCT
ajst-29358	26	12	based	base	VERB
ajst-29358	26	13	apple	apple	NOUN
ajst-29358	26	14	surface	surface	NOUN
ajst-29358	26	15	defect	defect	NOUN
ajst-29358	26	16	detection	detection	NOUN
ajst-29358	26	17	method	method	NOUN
ajst-29358	26	18	,	,	PUNCT
ajst-29358	26	19	designed	design	VERB
ajst-29358	26	20	a	a	DET
ajst-29358	26	21	dscs	dscs	ADJ
ajst-29358	26	22	shallowdeep	shallowdeep	NOUN
ajst-29358	26	23	feature	feature	NOUN
ajst-29358	26	24	selection	selection	NOUN
ajst-29358	26	25	module	module	NOUN
ajst-29358	26	26	based	base	VERB
ajst-29358	26	27	on	on	ADP
ajst-29358	26	28	the	the	DET
ajst-29358	26	29	dense	dense	ADJ
ajst-29358	26	30	module	module	NOUN
ajst-29358	26	31	and	and	CCONJ
ajst-29358	26	32	se	se	NOUN
ajst-29358	26	33	module	module	NOUN
ajst-29358	26	34	,	,	PUNCT
ajst-29358	26	35	improved	improve	VERB
ajst-29358	26	36	the	the	DET
ajst-29358	26	37	original	original	ADJ
ajst-29358	26	38	decoupling	decouple	VERB
ajst-29358	26	39	head	head	NOUN
ajst-29358	26	40	with	with	ADP
ajst-29358	26	41	elu	elu	PROPN
ajst-29358	26	42	activation	activation	NOUN
ajst-29358	26	43	function	function	NOUN
ajst-29358	26	44	,	,	PUNCT
ajst-29358	26	45	and	and	CCONJ
ajst-29358	26	46	used	use	VERB
ajst-29358	26	47	the	the	DET
ajst-29358	26	48	wise	wise	ADJ
ajst-29358	26	49	-	-	PUNCT
ajst-29358	26	50	iou	iou	NOUN
ajst-29358	26	51	loss	loss	NOUN
ajst-29358	26	52	function	function	NOUN
ajst-29358	26	53	for	for	ADP
ajst-29358	26	54	nonlinear	nonlinear	ADJ
ajst-29358	26	55	gain	gain	NOUN
ajst-29358	26	56	,	,	PUNCT
ajst-29358	26	57	achieving	achieve	VERB
ajst-29358	26	58	dynamic	dynamic	ADJ
ajst-29358	26	59	focused	focused	ADJ
ajst-29358	26	60	learning	learning	NOUN
ajst-29358	26	61	of	of	ADP
ajst-29358	26	62	the	the	DET
ajst-29358	26	63	network	network	NOUN
ajst-29358	26	64	.	.	PUNCT
ajst-29358	27	1	liu	liu	PROPN
ajst-29358	27	2	et	et	PROPN
ajst-29358	27	3	al	al	PROPN
ajst-29358	27	4	.	.	PROPN
ajst-29358	27	5	proposed	propose	VERB
ajst-29358	27	6	a	a	DET
ajst-29358	27	7	lightweight	lightweight	ADJ
ajst-29358	27	8	apple	apple	NOUN
ajst-29358	27	9	detection	detection	NOUN
ajst-29358	27	10	algorithm	algorithm	NOUN
ajst-29358	27	11	faster	fast	ADV
ajst-29358	27	12	-	-	PUNCT
ajst-29358	27	13	yolo	yolo	PROPN
ajst-29358	27	14	-	-	PUNCT
ajst-29358	27	15	ap	ap	PROPN
ajst-29358	27	16	based	base	VERB
ajst-29358	27	17	on	on	ADP
ajst-29358	27	18	improved	improved	ADJ
ajst-29358	27	19	yolov8	yolov8	NOUN
ajst-29358	27	20	,	,	PUNCT
ajst-29358	27	21	which	which	PRON
ajst-29358	27	22	is	be	AUX
ajst-29358	27	23	specifically	specifically	ADV
ajst-29358	27	24	designed	design	VERB
ajst-29358	27	25	for	for	ADP
ajst-29358	27	26	edge	edge	NOUN
ajst-29358	27	27	computing	computing	NOUN
ajst-29358	27	28	devices	device	NOUN
ajst-29358	27	29	to	to	PART
ajst-29358	27	30	achieve	achieve	VERB
ajst-29358	27	31	real	real	ADJ
ajst-29358	27	32	-	-	PUNCT
ajst-29358	27	33	time	time	NOUN
ajst-29358	27	34	object	object	NOUN
ajst-29358	27	35	detection	detection	NOUN
ajst-29358	27	36	.	.	PUNCT
ajst-29358	28	1	wang	wang	PROPN
ajst-29358	28	2	et	et	PROPN
ajst-29358	28	3	al	al	PROPN
ajst-29358	28	4	.	.	PROPN
ajst-29358	28	5	proposed	propose	VERB
ajst-29358	28	6	a	a	DET
ajst-29358	28	7	yolov8	yolov8	NOUN
ajst-29358	28	8	+	+	NOUN
ajst-29358	28	9	model	model	NOUN
ajst-29358	28	10	for	for	ADP
ajst-29358	28	11	strawberry	strawberry	NOUN
ajst-29358	28	12	maturity	maturity	NOUN
ajst-29358	28	13	detection	detection	NOUN
ajst-29358	28	14	and	and	CCONJ
ajst-29358	28	15	classification	classification	NOUN
ajst-29358	28	16	,	,	PUNCT
ajst-29358	28	17	providing	provide	VERB
ajst-29358	28	18	technical	technical	ADJ
ajst-29358	28	19	support	support	NOUN
ajst-29358	28	20	for	for	ADP
ajst-29358	28	21	intelligent	intelligent	ADJ
ajst-29358	28	22	harvesting	harvesting	NOUN
ajst-29358	28	23	systems	system	NOUN
ajst-29358	28	24	.	.	PUNCT
ajst-29358	29	1	zhang	zhang	PROPN
ajst-29358	29	2	et	et	PROPN
ajst-29358	29	3	al	al	PROPN
ajst-29358	29	4	.	.	PROPN
ajst-29358	29	5	improved	improve	VERB
ajst-29358	29	6	the	the	DET
ajst-29358	29	7	yolov8	yolov8	NOUN
ajst-29358	29	8	151	151	NUM
ajst-29358	29	9	algorithm	algorithm	NOUN
ajst-29358	29	10	by	by	ADP
ajst-29358	29	11	introducing	introduce	VERB
ajst-29358	29	12	mish	mish	ADJ
ajst-29358	29	13	activation	activation	NOUN
ajst-29358	29	14	function	function	NOUN
ajst-29358	29	15	and	and	CCONJ
ajst-29358	29	16	denseblock	denseblock	NOUN
ajst-29358	29	17	structure	structure	NOUN
ajst-29358	29	18	,	,	PUNCT
ajst-29358	29	19	improving	improve	VERB
ajst-29358	29	20	the	the	DET
ajst-29358	29	21	accuracy	accuracy	NOUN
ajst-29358	29	22	and	and	CCONJ
ajst-29358	29	23	robustness	robustness	NOUN
ajst-29358	29	24	of	of	ADP
ajst-29358	29	25	agricultural	agricultural	ADJ
ajst-29358	29	26	pest	pest	NOUN
ajst-29358	29	27	and	and	CCONJ
ajst-29358	29	28	disease	disease	NOUN
ajst-29358	29	29	detection	detection	NOUN
ajst-29358	29	30	.	.	PUNCT
ajst-29358	30	1	zhu	zhu	PROPN
ajst-29358	30	2	et	et	PROPN
ajst-29358	30	3	al	al	PROPN
ajst-29358	30	4	.	.	PROPN
ajst-29358	30	5	proposed	propose	VERB
ajst-29358	30	6	the	the	DET
ajst-29358	30	7	poly	poly	ADJ
ajst-29358	30	8	-	-	PUNCT
ajst-29358	30	9	yolov8	yolov8	NOUN
ajst-29358	30	10	model	model	NOUN
ajst-29358	30	11	and	and	CCONJ
ajst-29358	30	12	designed	design	VERB
ajst-29358	30	13	a	a	DET
ajst-29358	30	14	new	new	ADJ
ajst-29358	30	15	loss	loss	NOUN
ajst-29358	30	16	calculation	calculation	NOUN
ajst-29358	30	17	algorithm	algorithm	NOUN
ajst-29358	30	18	,	,	PUNCT
ajst-29358	30	19	improving	improve	VERB
ajst-29358	30	20	the	the	DET
ajst-29358	30	21	model	model	NOUN
ajst-29358	30	22	's	's	PART
ajst-29358	30	23	detection	detection	NOUN
ajst-29358	30	24	ability	ability	NOUN
ajst-29358	30	25	for	for	ADP
ajst-29358	30	26	polygonal	polygonal	ADJ
ajst-29358	30	27	targets	target	NOUN
ajst-29358	30	28	.	.	PUNCT
ajst-29358	31	1	the	the	DET
ajst-29358	31	2	aforementioned	aforementioned	ADJ
ajst-29358	31	3	methods	method	NOUN
ajst-29358	31	4	have	have	VERB
ajst-29358	31	5	some	some	DET
ajst-29358	31	6	issues	issue	NOUN
ajst-29358	31	7	with	with	ADP
ajst-29358	31	8	realtime	realtime	NOUN
ajst-29358	31	9	and	and	CCONJ
ajst-29358	31	10	accuracy	accuracy	NOUN
ajst-29358	31	11	in	in	ADP
ajst-29358	31	12	detection	detection	NOUN
ajst-29358	31	13	speed	speed	NOUN
ajst-29358	31	14	and	and	CCONJ
ajst-29358	31	15	model	model	NOUN
ajst-29358	31	16	precision	precision	NOUN
ajst-29358	31	17	.	.	PUNCT
ajst-29358	32	1	in	in	ADP
ajst-29358	32	2	response	response	NOUN
ajst-29358	32	3	to	to	ADP
ajst-29358	32	4	these	these	DET
ajst-29358	32	5	issues	issue	NOUN
ajst-29358	32	6	,	,	PUNCT
ajst-29358	32	7	this	this	DET
ajst-29358	32	8	paper	paper	NOUN
ajst-29358	32	9	improves	improve	VERB
ajst-29358	32	10	the	the	DET
ajst-29358	32	11	yolov8n	yolov8n	PROPN
ajst-29358	32	12	as	as	ADP
ajst-29358	32	13	the	the	DET
ajst-29358	32	14	benchmark	benchmark	ADJ
ajst-29358	32	15	algorithm	algorithm	NOUN
ajst-29358	32	16	to	to	PART
ajst-29358	32	17	achieve	achieve	VERB
ajst-29358	32	18	uav	uav	PROPN
ajst-29358	32	19	image	image	NOUN
ajst-29358	32	20	corn	corn	NOUN
ajst-29358	32	21	large	large	ADJ
ajst-29358	32	22	spot	spot	NOUN
ajst-29358	32	23	disease	disease	NOUN
ajst-29358	32	24	detection	detection	NOUN
ajst-29358	32	25	,	,	PUNCT
ajst-29358	32	26	and	and	CCONJ
ajst-29358	32	27	the	the	DET
ajst-29358	32	28	main	main	ADJ
ajst-29358	32	29	improvements	improvement	NOUN
ajst-29358	32	30	and	and	CCONJ
ajst-29358	32	31	innovations	innovation	NOUN
ajst-29358	32	32	to	to	ADP
ajst-29358	32	33	the	the	DET
ajst-29358	32	34	yolov8	yolov8	NOUN
ajst-29358	32	35	algorithm	algorithm	PROPN
ajst-29358	32	36	are	be	AUX
ajst-29358	32	37	as	as	SCONJ
ajst-29358	32	38	follows	follow	VERB
ajst-29358	32	39	:	:	PUNCT
ajst-29358	32	40	first	first	ADV
ajst-29358	32	41	,	,	PUNCT
ajst-29358	32	42	a	a	DET
ajst-29358	32	43	new	new	ADJ
ajst-29358	32	44	attention	attention	NOUN
ajst-29358	32	45	mechanism	mechanism	NOUN
ajst-29358	32	46	module	module	NOUN
ajst-29358	32	47	is	be	AUX
ajst-29358	32	48	introduced	introduce	VERB
ajst-29358	32	49	to	to	PART
ajst-29358	32	50	improve	improve	VERB
ajst-29358	32	51	recognition	recognition	NOUN
ajst-29358	32	52	accuracy	accuracy	NOUN
ajst-29358	32	53	,	,	PUNCT
ajst-29358	32	54	then	then	ADV
ajst-29358	32	55	the	the	DET
ajst-29358	32	56	new	new	ADJ
ajst-29358	32	57	feature	feature	NOUN
ajst-29358	32	58	pyramid	pyramid	NOUN
ajst-29358	32	59	network	network	NOUN
ajst-29358	32	60	structure	structure	NOUN
ajst-29358	32	61	adopts	adopt	VERB
ajst-29358	32	62	cross	cross	ADJ
ajst-29358	32	63	-	-	ADJ
ajst-29358	32	64	scale	scale	ADJ
ajst-29358	32	65	connection	connection	NOUN
ajst-29358	32	66	methods	method	NOUN
ajst-29358	32	67	and	and	CCONJ
ajst-29358	32	68	weighted	weight	VERB
ajst-29358	32	69	feature	feature	NOUN
ajst-29358	32	70	methods	method	NOUN
ajst-29358	32	71	to	to	PART
ajst-29358	32	72	effectively	effectively	ADV
ajst-29358	32	73	integrate	integrate	VERB
ajst-29358	32	74	multi	multi	ADJ
ajst-29358	32	75	-	-	ADJ
ajst-29358	32	76	scale	scale	ADJ
ajst-29358	32	77	features	feature	NOUN
ajst-29358	32	78	,	,	PUNCT
ajst-29358	32	79	fully	fully	ADV
ajst-29358	32	80	solving	solve	VERB
ajst-29358	32	81	the	the	DET
ajst-29358	32	82	problem	problem	NOUN
ajst-29358	32	83	of	of	ADP
ajst-29358	32	84	large	large	ADJ
ajst-29358	32	85	scale	scale	NOUN
ajst-29358	32	86	differences	difference	NOUN
ajst-29358	32	87	and	and	CCONJ
ajst-29358	32	88	feature	feature	NOUN
ajst-29358	32	89	loss	loss	NOUN
ajst-29358	32	90	,	,	PUNCT
ajst-29358	32	91	in	in	ADP
ajst-29358	32	92	addition	addition	NOUN
ajst-29358	32	93	,	,	PUNCT
ajst-29358	32	94	the	the	DET
ajst-29358	32	95	introduction	introduction	NOUN
ajst-29358	32	96	of	of	ADP
ajst-29358	32	97	a	a	DET
ajst-29358	32	98	new	new	ADJ
ajst-29358	32	99	loss	loss	NOUN
ajst-29358	32	100	function	function	NOUN
ajst-29358	32	101	effectively	effectively	ADV
ajst-29358	32	102	reduces	reduce	VERB
ajst-29358	32	103	the	the	DET
ajst-29358	32	104	negative	negative	ADJ
ajst-29358	32	105	impact	impact	NOUN
ajst-29358	32	106	of	of	ADP
ajst-29358	32	107	lowquality	lowquality	NOUN
ajst-29358	32	108	samples	sample	NOUN
ajst-29358	32	109	,	,	PUNCT
ajst-29358	32	110	further	far	ADV
ajst-29358	32	111	enhancing	enhance	VERB
ajst-29358	32	112	the	the	DET
ajst-29358	32	113	model	model	NOUN
ajst-29358	32	114	's	's	PART
ajst-29358	32	115	generalization	generalization	NOUN
ajst-29358	32	116	performance	performance	NOUN
ajst-29358	32	117	.	.	PUNCT
ajst-29358	33	1	2	2	X
ajst-29358	33	2	.	.	X
ajst-29358	33	3	improved	improve	VERB
ajst-29358	33	4	yolov8	yolov8	NOUN
ajst-29358	33	5	algorithm	algorithm	PROPN
ajst-29358	33	6	the	the	DET
ajst-29358	33	7	yolov8	yolov8	NOUN
ajst-29358	33	8	algorithm	algorithm	PROPN
ajst-29358	33	9	is	be	AUX
ajst-29358	33	10	known	know	VERB
ajst-29358	33	11	for	for	ADP
ajst-29358	33	12	its	its	PRON
ajst-29358	33	13	fast	fast	ADJ
ajst-29358	33	14	speed	speed	NOUN
ajst-29358	33	15	and	and	CCONJ
ajst-29358	33	16	high	high	ADJ
ajst-29358	33	17	accuracy	accuracy	NOUN
ajst-29358	33	18	,	,	PUNCT
ajst-29358	33	19	and	and	CCONJ
ajst-29358	33	20	has	have	AUX
ajst-29358	33	21	been	be	AUX
ajst-29358	33	22	widely	widely	ADV
ajst-29358	33	23	applied	apply	VERB
ajst-29358	33	24	.	.	PUNCT
ajst-29358	34	1	the	the	DET
ajst-29358	34	2	innovation	innovation	NOUN
ajst-29358	34	3	of	of	ADP
ajst-29358	34	4	the	the	DET
ajst-29358	34	5	yolov8	yolov8	NOUN
ajst-29358	34	6	model	model	PROPN
ajst-29358	34	7	structure	structure	NOUN
ajst-29358	34	8	first	first	ADV
ajst-29358	34	9	uses	use	VERB
ajst-29358	34	10	a	a	DET
ajst-29358	34	11	deep	deep	ADJ
ajst-29358	34	12	convolutional	convolutional	ADJ
ajst-29358	34	13	neural	neural	ADJ
ajst-29358	34	14	network	network	NOUN
ajst-29358	34	15	as	as	ADP
ajst-29358	34	16	its	its	PRON
ajst-29358	34	17	feature	feature	NOUN
ajst-29358	34	18	extractor	extractor	NOUN
ajst-29358	34	19	,	,	PUNCT
ajst-29358	34	20	which	which	PRON
ajst-29358	34	21	greatly	greatly	ADV
ajst-29358	34	22	helps	help	VERB
ajst-29358	34	23	in	in	ADP
ajst-29358	34	24	capturing	capture	VERB
ajst-29358	34	25	richer	rich	ADJ
ajst-29358	34	26	image	image	NOUN
ajst-29358	34	27	features	feature	NOUN
ajst-29358	34	28	.	.	PUNCT
ajst-29358	35	1	at	at	ADP
ajst-29358	35	2	the	the	DET
ajst-29358	35	3	same	same	ADJ
ajst-29358	35	4	time	time	NOUN
ajst-29358	35	5	,	,	PUNCT
ajst-29358	35	6	multi	multi	ADJ
ajst-29358	35	7	-	-	ADJ
ajst-29358	35	8	scale	scale	ADJ
ajst-29358	35	9	feature	feature	NOUN
ajst-29358	35	10	fusion	fusion	NOUN
ajst-29358	35	11	technology	technology	NOUN
ajst-29358	35	12	is	be	AUX
ajst-29358	35	13	introduced	introduce	VERB
ajst-29358	35	14	into	into	ADP
ajst-29358	35	15	the	the	DET
ajst-29358	35	16	algorithm	algorithm	NOUN
ajst-29358	35	17	,	,	PUNCT
ajst-29358	35	18	making	make	VERB
ajst-29358	35	19	it	it	PRON
ajst-29358	35	20	more	more	ADV
ajst-29358	35	21	effective	effective	ADJ
ajst-29358	35	22	in	in	ADP
ajst-29358	35	23	handling	handle	VERB
ajst-29358	35	24	targets	target	NOUN
ajst-29358	35	25	of	of	ADP
ajst-29358	35	26	different	different	ADJ
ajst-29358	35	27	sizes	size	NOUN
ajst-29358	35	28	and	and	CCONJ
ajst-29358	35	29	improving	improve	VERB
ajst-29358	35	30	the	the	DET
ajst-29358	35	31	algorithm	algorithm	NOUN
ajst-29358	35	32	's	's	PART
ajst-29358	35	33	ability	ability	NOUN
ajst-29358	35	34	to	to	PART
ajst-29358	35	35	detect	detect	VERB
ajst-29358	35	36	small	small	ADJ
ajst-29358	35	37	targets	target	NOUN
ajst-29358	35	38	.	.	PUNCT
ajst-29358	36	1	yolov8	yolov8	NOUN
ajst-29358	36	2	also	also	ADV
ajst-29358	36	3	optimizes	optimize	VERB
ajst-29358	36	4	the	the	DET
ajst-29358	36	5	loss	loss	NOUN
ajst-29358	36	6	function	function	NOUN
ajst-29358	36	7	by	by	ADP
ajst-29358	36	8	introducing	introduce	VERB
ajst-29358	36	9	new	new	ADJ
ajst-29358	36	10	regularization	regularization	NOUN
ajst-29358	36	11	terms	term	NOUN
ajst-29358	36	12	and	and	CCONJ
ajst-29358	36	13	adjusting	adjust	VERB
ajst-29358	36	14	the	the	DET
ajst-29358	36	15	predicted	predict	VERB
ajst-29358	36	16	bounding	bounding	NOUN
ajst-29358	36	17	box	box	NOUN
ajst-29358	36	18	,	,	PUNCT
ajst-29358	36	19	further	far	ADV
ajst-29358	36	20	improving	improve	VERB
ajst-29358	36	21	the	the	DET
ajst-29358	36	22	accuracy	accuracy	NOUN
ajst-29358	36	23	of	of	ADP
ajst-29358	36	24	detection	detection	NOUN
ajst-29358	36	25	.	.	PUNCT
ajst-29358	37	1	yolov8	yolov8	NOUN
ajst-29358	37	2	has	have	AUX
ajst-29358	37	3	achieved	achieve	VERB
ajst-29358	37	4	simultaneous	simultaneous	ADJ
ajst-29358	37	5	improvement	improvement	NOUN
ajst-29358	37	6	in	in	ADP
ajst-29358	37	7	recognition	recognition	NOUN
ajst-29358	37	8	speed	speed	NOUN
ajst-29358	37	9	and	and	CCONJ
ajst-29358	37	10	detection	detection	NOUN
ajst-29358	37	11	accuracy	accuracy	NOUN
ajst-29358	37	12	through	through	ADP
ajst-29358	37	13	these	these	DET
ajst-29358	37	14	improvements	improvement	NOUN
ajst-29358	37	15	.	.	PUNCT
ajst-29358	38	1	the	the	DET
ajst-29358	38	2	optimization	optimization	NOUN
ajst-29358	38	3	of	of	ADP
ajst-29358	38	4	yolov8	yolov8	NOUN
ajst-29358	38	5	has	have	AUX
ajst-29358	38	6	achieved	achieve	VERB
ajst-29358	38	7	good	good	ADJ
ajst-29358	38	8	results	result	NOUN
ajst-29358	38	9	in	in	ADP
ajst-29358	38	10	target	target	NOUN
ajst-29358	38	11	detection	detection	NOUN
ajst-29358	38	12	tasks	task	NOUN
ajst-29358	38	13	.	.	PUNCT
ajst-29358	39	1	2.1	2.1	NUM
ajst-29358	39	2	.	.	PUNCT
ajst-29358	40	1	maize	maize	VERB
ajst-29358	40	2	large	large	ADJ
ajst-29358	40	3	spot	spot	NOUN
ajst-29358	40	4	disease	disease	NOUN
ajst-29358	40	5	recognition	recognition	NOUN
ajst-29358	40	6	algorithm	algorithm	NOUN
ajst-29358	40	7	model	model	NOUN
ajst-29358	40	8	to	to	PART
ajst-29358	40	9	improve	improve	VERB
ajst-29358	40	10	the	the	DET
ajst-29358	40	11	detection	detection	NOUN
ajst-29358	40	12	accuracy	accuracy	NOUN
ajst-29358	40	13	of	of	ADP
ajst-29358	40	14	maize	maize	NOUN
ajst-29358	40	15	large	large	ADJ
ajst-29358	40	16	spot	spot	NOUN
ajst-29358	40	17	disease	disease	NOUN
ajst-29358	40	18	,	,	PUNCT
ajst-29358	40	19	this	this	DET
ajst-29358	40	20	paper	paper	NOUN
ajst-29358	40	21	proposes	propose	VERB
ajst-29358	40	22	an	an	DET
ajst-29358	40	23	improved	improved	ADJ
ajst-29358	40	24	optimization	optimization	NOUN
ajst-29358	40	25	plan	plan	NOUN
ajst-29358	40	26	based	base	VERB
ajst-29358	40	27	on	on	ADP
ajst-29358	40	28	the	the	DET
ajst-29358	40	29	yolov8	yolov8	NOUN
ajst-29358	40	30	network	network	NOUN
ajst-29358	40	31	architecture	architecture	NOUN
ajst-29358	40	32	.	.	PUNCT
ajst-29358	41	1	the	the	DET
ajst-29358	41	2	model	model	NOUN
ajst-29358	41	3	is	be	AUX
ajst-29358	41	4	based	base	VERB
ajst-29358	41	5	on	on	ADP
ajst-29358	41	6	the	the	DET
ajst-29358	41	7	yolov8	yolov8	NOUN
ajst-29358	41	8	base	base	NOUN
ajst-29358	41	9	algorithm	algorithm	NOUN
ajst-29358	41	10	and	and	CCONJ
ajst-29358	41	11	is	be	AUX
ajst-29358	41	12	improved	improve	VERB
ajst-29358	41	13	and	and	CCONJ
ajst-29358	41	14	optimized	optimize	VERB
ajst-29358	41	15	,	,	PUNCT
ajst-29358	41	16	as	as	SCONJ
ajst-29358	41	17	shown	show	VERB
ajst-29358	41	18	in	in	ADP
ajst-29358	41	19	figure	figure	NOUN
ajst-29358	41	20	1	1	NUM
ajst-29358	41	21	.	.	PUNCT
ajst-29358	42	1	figure	figure	NOUN
ajst-29358	42	2	1	1	NUM
ajst-29358	42	3	.	.	PUNCT
ajst-29358	42	4	improved	improve	VERB
ajst-29358	42	5	and	and	CCONJ
ajst-29358	42	6	optimized	optimize	VERB
ajst-29358	42	7	yolov8	yolov8	NOUN
ajst-29358	42	8	network	network	NOUN
ajst-29358	42	9	structure	structure	NOUN
ajst-29358	43	1	the	the	DET
ajst-29358	43	2	specific	specific	ADJ
ajst-29358	43	3	improvements	improvement	NOUN
ajst-29358	43	4	of	of	ADP
ajst-29358	43	5	the	the	DET
ajst-29358	43	6	improved	improved	ADJ
ajst-29358	43	7	model	model	NOUN
ajst-29358	43	8	compared	compare	VERB
ajst-29358	43	9	to	to	ADP
ajst-29358	43	10	the	the	DET
ajst-29358	43	11	pre	pre	ADJ
ajst-29358	43	12	-	-	ADJ
ajst-29358	43	13	improved	improved	ADJ
ajst-29358	43	14	model	model	NOUN
ajst-29358	43	15	are	be	AUX
ajst-29358	43	16	as	as	SCONJ
ajst-29358	43	17	follows	follow	VERB
ajst-29358	43	18	.	.	PUNCT
ajst-29358	44	1	backbone	backbone	NOUN
ajst-29358	44	2	network	network	NOUN
ajst-29358	44	3	:	:	PUNCT
ajst-29358	44	4	three	three	NUM
ajst-29358	44	5	multi	multi	ADJ
ajst-29358	44	6	-	-	ADJ
ajst-29358	44	7	scale	scale	ADJ
ajst-29358	44	8	attention	attention	NOUN
ajst-29358	44	9	ema	ema	PROPN
ajst-29358	44	10	modules	module	NOUN
ajst-29358	44	11	are	be	AUX
ajst-29358	44	12	introduced	introduce	VERB
ajst-29358	44	13	to	to	PART
ajst-29358	44	14	capture	capture	VERB
ajst-29358	44	15	detail	detail	NOUN
ajst-29358	44	16	information	information	NOUN
ajst-29358	44	17	and	and	CCONJ
ajst-29358	44	18	improve	improve	VERB
ajst-29358	44	19	the	the	DET
ajst-29358	44	20	model	model	NOUN
ajst-29358	44	21	's	's	PART
ajst-29358	44	22	feature	feature	NOUN
ajst-29358	44	23	extraction	extraction	NOUN
ajst-29358	44	24	capability	capability	NOUN
ajst-29358	44	25	,	,	PUNCT
ajst-29358	44	26	while	while	SCONJ
ajst-29358	44	27	retaining	retain	VERB
ajst-29358	44	28	the	the	DET
ajst-29358	44	29	original	original	ADJ
ajst-29358	44	30	sppf	sppf	ADJ
ajst-29358	44	31	module	module	NOUN
ajst-29358	44	32	in	in	ADP
ajst-29358	44	33	the	the	DET
ajst-29358	44	34	model	model	NOUN
ajst-29358	44	35	,	,	PUNCT
ajst-29358	44	36	which	which	PRON
ajst-29358	44	37	can	can	AUX
ajst-29358	44	38	solve	solve	VERB
ajst-29358	44	39	the	the	DET
ajst-29358	44	40	problem	problem	NOUN
ajst-29358	44	41	of	of	ADP
ajst-29358	44	42	repeated	repeat	VERB
ajst-29358	44	43	feature	feature	NOUN
ajst-29358	44	44	extraction	extraction	NOUN
ajst-29358	44	45	of	of	ADP
ajst-29358	44	46	images	image	NOUN
ajst-29358	44	47	by	by	ADP
ajst-29358	44	48	convolutional	convolutional	ADJ
ajst-29358	44	49	neural	neural	ADJ
ajst-29358	44	50	networks	network	NOUN
ajst-29358	44	51	,	,	PUNCT
ajst-29358	44	52	speed	speed	VERB
ajst-29358	44	53	up	up	ADP
ajst-29358	44	54	the	the	DET
ajst-29358	44	55	generation	generation	NOUN
ajst-29358	44	56	of	of	ADP
ajst-29358	44	57	candidate	candidate	NOUN
ajst-29358	44	58	boxes	box	NOUN
ajst-29358	44	59	,	,	PUNCT
ajst-29358	44	60	and	and	CCONJ
ajst-29358	44	61	save	save	VERB
ajst-29358	44	62	computational	computational	ADJ
ajst-29358	44	63	costs	cost	NOUN
ajst-29358	44	64	.	.	PUNCT
ajst-29358	45	1	neck	neck	NOUN
ajst-29358	45	2	network	network	NOUN
ajst-29358	45	3	:	:	PUNCT
ajst-29358	45	4	a	a	DET
ajst-29358	45	5	lightweight	lightweight	NOUN
ajst-29358	45	6	yolov8	yolov8	NOUN
ajst-29358	45	7	-	-	PUNCT
ajst-29358	45	8	bifpn	bifpn	PROPN
ajst-29358	45	9	feature	feature	NOUN
ajst-29358	45	10	pyramid	pyramid	NOUN
ajst-29358	45	11	network	network	NOUN
ajst-29358	45	12	structure	structure	NOUN
ajst-29358	45	13	is	be	AUX
ajst-29358	45	14	proposed	propose	VERB
ajst-29358	45	15	by	by	ADP
ajst-29358	45	16	fusing	fuse	VERB
ajst-29358	45	17	bifpn	bifpn	PROPN
ajst-29358	45	18	and	and	CCONJ
ajst-29358	45	19	the	the	DET
ajst-29358	45	20	original	original	ADJ
ajst-29358	45	21	yolov8	yolov8	NOUN
ajst-29358	45	22	structure	structure	NOUN
ajst-29358	45	23	to	to	PART
ajst-29358	45	24	enhance	enhance	VERB
ajst-29358	45	25	the	the	DET
ajst-29358	45	26	model	model	NOUN
ajst-29358	45	27	's	's	PART
ajst-29358	45	28	multiscale	multiscale	ADJ
ajst-29358	45	29	feature	feature	NOUN
ajst-29358	45	30	fusion	fusion	NOUN
ajst-29358	45	31	capability	capability	NOUN
ajst-29358	45	32	and	and	CCONJ
ajst-29358	45	33	improve	improve	VERB
ajst-29358	45	34	the	the	DET
ajst-29358	45	35	network	network	NOUN
ajst-29358	45	36	's	's	PART
ajst-29358	45	37	detection	detection	NOUN
ajst-29358	45	38	accuracy	accuracy	NOUN
ajst-29358	45	39	for	for	ADP
ajst-29358	45	40	small	small	ADJ
ajst-29358	45	41	targets	target	NOUN
ajst-29358	45	42	.	.	PUNCT
ajst-29358	46	1	152	152	NUM
ajst-29358	46	2	head	head	NOUN
ajst-29358	46	3	network	network	NOUN
ajst-29358	46	4	:	:	PUNCT
ajst-29358	46	5	the	the	DET
ajst-29358	46	6	original	original	ADJ
ajst-29358	46	7	network	network	NOUN
ajst-29358	46	8	loss	loss	NOUN
ajst-29358	46	9	function	function	NOUN
ajst-29358	46	10	is	be	AUX
ajst-29358	46	11	optimized	optimize	VERB
ajst-29358	46	12	with	with	ADP
ajst-29358	46	13	wiou	wiou	NOUN
ajst-29358	46	14	loss	loss	NOUN
ajst-29358	46	15	,	,	PUNCT
ajst-29358	46	16	introducing	introduce	VERB
ajst-29358	46	17	a	a	DET
ajst-29358	46	18	dynamic	dynamic	ADJ
ajst-29358	46	19	nonmonotonic	nonmonotonic	ADJ
ajst-29358	46	20	focusing	focus	VERB
ajst-29358	46	21	mechanism	mechanism	NOUN
ajst-29358	46	22	to	to	PART
ajst-29358	46	23	improve	improve	VERB
ajst-29358	46	24	the	the	DET
ajst-29358	46	25	model	model	NOUN
ajst-29358	46	26	's	's	PART
ajst-29358	46	27	generalization	generalization	NOUN
ajst-29358	46	28	capability	capability	NOUN
ajst-29358	46	29	.	.	PUNCT
ajst-29358	47	1	in	in	ADP
ajst-29358	47	2	summary	summary	NOUN
ajst-29358	47	3	,	,	PUNCT
ajst-29358	47	4	this	this	DET
ajst-29358	47	5	design	design	NOUN
ajst-29358	47	6	method	method	NOUN
ajst-29358	47	7	improves	improve	VERB
ajst-29358	47	8	the	the	DET
ajst-29358	47	9	existing	exist	VERB
ajst-29358	47	10	algorithm	algorithm	NOUN
ajst-29358	47	11	,	,	PUNCT
ajst-29358	47	12	achieving	achieve	VERB
ajst-29358	47	13	high	high	ADJ
ajst-29358	47	14	detection	detection	NOUN
ajst-29358	47	15	speed	speed	NOUN
ajst-29358	47	16	and	and	CCONJ
ajst-29358	47	17	model	model	NOUN
ajst-29358	47	18	lightweight	lightweight	NOUN
ajst-29358	47	19	while	while	SCONJ
ajst-29358	47	20	improving	improve	VERB
ajst-29358	47	21	detection	detection	NOUN
ajst-29358	47	22	accuracy	accuracy	NOUN
ajst-29358	47	23	,	,	PUNCT
ajst-29358	47	24	thereby	thereby	ADV
ajst-29358	47	25	enhancing	enhance	VERB
ajst-29358	47	26	the	the	DET
ajst-29358	47	27	extraction	extraction	NOUN
ajst-29358	47	28	capability	capability	NOUN
ajst-29358	47	29	of	of	ADP
ajst-29358	47	30	maize	maize	NOUN
ajst-29358	47	31	large	large	ADJ
ajst-29358	47	32	spot	spot	NOUN
ajst-29358	47	33	disease	disease	NOUN
ajst-29358	47	34	features	feature	VERB
ajst-29358	47	35	and	and	CCONJ
ajst-29358	47	36	enhancing	enhance	VERB
ajst-29358	47	37	the	the	DET
ajst-29358	47	38	stability	stability	NOUN
ajst-29358	47	39	of	of	ADP
ajst-29358	47	40	the	the	DET
ajst-29358	47	41	detection	detection	NOUN
ajst-29358	47	42	process	process	NOUN
ajst-29358	47	43	.	.	PUNCT
ajst-29358	48	1	2.2	2.2	NUM
ajst-29358	48	2	.	.	PUNCT
ajst-29358	48	3	backbone	backbone	NOUN
ajst-29358	48	4	network	network	NOUN
ajst-29358	48	5	improvement	improvement	NOUN
ajst-29358	48	6	.	.	PUNCT
ajst-29358	49	1	by	by	ADP
ajst-29358	49	2	introducing	introduce	VERB
ajst-29358	49	3	the	the	DET
ajst-29358	49	4	ema	ema	PROPN
ajst-29358	49	5	(	(	PUNCT
ajst-29358	49	6	efficient	efficient	ADJ
ajst-29358	49	7	multi	multi	ADJ
ajst-29358	49	8	-	-	ADJ
ajst-29358	49	9	scale	scale	ADJ
ajst-29358	49	10	attention	attention	NOUN
ajst-29358	49	11	)	)	PUNCT
ajst-29358	49	12	attention	attention	NOUN
ajst-29358	49	13	module	module	NOUN
ajst-29358	49	14	,	,	PUNCT
ajst-29358	49	15	the	the	DET
ajst-29358	49	16	ability	ability	NOUN
ajst-29358	49	17	of	of	ADP
ajst-29358	49	18	the	the	DET
ajst-29358	49	19	attention	attention	NOUN
ajst-29358	49	20	mechanism	mechanism	NOUN
ajst-29358	49	21	to	to	PART
ajst-29358	49	22	obtain	obtain	VERB
ajst-29358	49	23	local	local	ADJ
ajst-29358	49	24	important	important	ADJ
ajst-29358	49	25	information	information	NOUN
ajst-29358	49	26	is	be	AUX
ajst-29358	49	27	improved	improve	VERB
ajst-29358	49	28	.	.	PUNCT
ajst-29358	50	1	ema	ema	PROPN
ajst-29358	50	2	,	,	PUNCT
ajst-29358	50	3	as	as	ADP
ajst-29358	50	4	a	a	DET
ajst-29358	50	5	multi	multi	ADJ
ajst-29358	50	6	-	-	ADJ
ajst-29358	50	7	scale	scale	ADJ
ajst-29358	50	8	efficient	efficient	ADJ
ajst-29358	50	9	attention	attention	NOUN
ajst-29358	50	10	mechanism	mechanism	NOUN
ajst-29358	50	11	,	,	PUNCT
ajst-29358	50	12	reshapes	reshape	VERB
ajst-29358	50	13	the	the	DET
ajst-29358	50	14	channel	channel	NOUN
ajst-29358	50	15	dimension	dimension	NOUN
ajst-29358	50	16	while	while	SCONJ
ajst-29358	50	17	maintaining	maintain	VERB
ajst-29358	50	18	the	the	DET
ajst-29358	50	19	original	original	ADJ
ajst-29358	50	20	channel	channel	NOUN
ajst-29358	50	21	dimension	dimension	NOUN
ajst-29358	50	22	,	,	PUNCT
ajst-29358	50	23	improving	improve	VERB
ajst-29358	50	24	computational	computational	ADJ
ajst-29358	50	25	ability	ability	NOUN
ajst-29358	50	26	while	while	SCONJ
ajst-29358	50	27	minimizing	minimize	VERB
ajst-29358	50	28	the	the	DET
ajst-29358	50	29	loss	loss	NOUN
ajst-29358	50	30	of	of	ADP
ajst-29358	50	31	channel	channel	NOUN
ajst-29358	50	32	information	information	NOUN
ajst-29358	50	33	.	.	PUNCT
ajst-29358	51	1	ema	ema	PROPN
ajst-29358	51	2	is	be	AUX
ajst-29358	51	3	mainly	mainly	ADV
ajst-29358	51	4	responsible	responsible	ADJ
ajst-29358	51	5	for	for	ADP
ajst-29358	51	6	global	global	ADJ
ajst-29358	51	7	information	information	NOUN
ajst-29358	51	8	encoding	encoding	NOUN
ajst-29358	51	9	,	,	PUNCT
ajst-29358	51	10	can	can	AUX
ajst-29358	51	11	manage	manage	VERB
ajst-29358	51	12	the	the	DET
ajst-29358	51	13	channel	channel	NOUN
ajst-29358	51	14	weights	weight	NOUN
ajst-29358	51	15	of	of	ADP
ajst-29358	51	16	multiple	multiple	ADJ
ajst-29358	51	17	parallel	parallel	ADJ
ajst-29358	51	18	subnetworks	subnetwork	NOUN
ajst-29358	51	19	,	,	PUNCT
ajst-29358	51	20	and	and	CCONJ
ajst-29358	51	21	promotes	promote	VERB
ajst-29358	51	22	cross	cross	ADJ
ajst-29358	51	23	-	-	ADJ
ajst-29358	51	24	dimensional	dimensional	ADJ
ajst-29358	51	25	information	information	NOUN
ajst-29358	51	26	integration	integration	NOUN
ajst-29358	51	27	between	between	ADP
ajst-29358	51	28	these	these	DET
ajst-29358	51	29	subnetworks	subnetwork	NOUN
ajst-29358	51	30	,	,	PUNCT
ajst-29358	51	31	thereby	thereby	ADV
ajst-29358	51	32	producing	produce	VERB
ajst-29358	51	33	output	output	NOUN
ajst-29358	51	34	features	feature	NOUN
ajst-29358	51	35	.	.	PUNCT
ajst-29358	52	1	its	its	PRON
ajst-29358	52	2	structure	structure	NOUN
ajst-29358	52	3	is	be	AUX
ajst-29358	52	4	shown	show	VERB
ajst-29358	52	5	in	in	ADP
ajst-29358	52	6	figure	figure	NOUN
ajst-29358	52	7	2	2	NUM
ajst-29358	52	8	.	.	PUNCT
ajst-29358	52	9	when	when	SCONJ
ajst-29358	52	10	the	the	DET
ajst-29358	52	11	input	input	NOUN
ajst-29358	52	12	features	feature	VERB
ajst-29358	52	13	c	c	PROPN
ajst-29358	52	14	h	h	NOUN
ajst-29358	53	1	wr	wr	NOUN
ajst-29358	53	2			VERB
ajst-29358	53	3	x	x	PROPN
ajst-29358	53	4	are	be	AUX
ajst-29358	53	5	divided	divide	VERB
ajst-29358	53	6	into	into	ADP
ajst-29358	53	7	n	n	PRON
ajst-29358	53	8	sub	sub	NOUN
ajst-29358	53	9	-	-	ADJ
ajst-29358	53	10	features	feature	NOUN
ajst-29358	53	11			ADJ
ajst-29358	53	12			ADP
ajst-29358	53	13	c//c	c//c	PROPN
ajst-29358	53	14	h	h	NOUN
ajst-29358	53	15	0	0	NUM
ajst-29358	53	16	1	1	NUM
ajst-29358	53	17	,	,	PUNCT
ajst-29358	53	18	,	,	PUNCT
ajst-29358	53	19	,	,	PUNCT
ajst-29358	53	20	r	r	NOUN
ajst-29358	53	21	w	w	VERB
ajst-29358	53	22	i	i	PRON
ajst-29358	54	1	n	n	VERB
ajst-29358	54	2	i	i	PRON
ajst-29358	54	3			VERB
ajst-29358	54	4			VERB
ajst-29358	54	5			ADJ
ajst-29358	54	6	x	x	ADV
ajst-29358	54	7	x	x	X
ajst-29358	54	8	x	x	PUNCT
ajst-29358	54	9	x	x	X
ajst-29358	54	10	x	x	PUNCT
ajst-29358	54	11	according	accord	VERB
ajst-29358	54	12	to	to	ADP
ajst-29358	54	13	the	the	DET
ajst-29358	54	14	ema	ema	PROPN
ajst-29358	54	15	channel	channel	NOUN
ajst-29358	54	16	number	number	NOUN
ajst-29358	54	17	,	,	PUNCT
ajst-29358	54	18	let	let	VERB
ajst-29358	54	19	n<<c	n<<c	NOUN
ajst-29358	54	20	,	,	PUNCT
ajst-29358	54	21	and	and	CCONJ
ajst-29358	54	22	assume	assume	VERB
ajst-29358	54	23	that	that	SCONJ
ajst-29358	54	24	the	the	DET
ajst-29358	54	25	learned	learn	VERB
ajst-29358	54	26	weights	weight	NOUN
ajst-29358	54	27	can	can	AUX
ajst-29358	54	28	enhance	enhance	VERB
ajst-29358	54	29	the	the	DET
ajst-29358	54	30	feature	feature	NOUN
ajst-29358	54	31	representation	representation	NOUN
ajst-29358	54	32	of	of	ADP
ajst-29358	54	33	key	key	ADJ
ajst-29358	54	34	areas	area	NOUN
ajst-29358	54	35	in	in	ADP
ajst-29358	54	36	all	all	DET
ajst-29358	54	37	sub	sub	NOUN
ajst-29358	54	38	-	-	NOUN
ajst-29358	54	39	features	feature	NOUN
ajst-29358	54	40	.	.	PUNCT
ajst-29358	55	1	ema	ema	PROPN
ajst-29358	55	2	uses	use	VERB
ajst-29358	55	3	a	a	DET
ajst-29358	55	4	3x3	3x3	NUM
ajst-29358	55	5	branch	branch	NOUN
ajst-29358	55	6	and	and	CCONJ
ajst-29358	55	7	two	two	NUM
ajst-29358	55	8	1x1	1x1	NUM
ajst-29358	55	9	branches	branch	NOUN
ajst-29358	55	10	as	as	ADP
ajst-29358	55	11	parallel	parallel	ADJ
ajst-29358	55	12	paths	path	NOUN
ajst-29358	55	13	to	to	PART
ajst-29358	55	14	obtain	obtain	VERB
ajst-29358	55	15	feature	feature	NOUN
ajst-29358	55	16	weights	weight	NOUN
ajst-29358	55	17	,	,	PUNCT
ajst-29358	55	18	which	which	PRON
ajst-29358	55	19	can	can	AUX
ajst-29358	55	20	obtain	obtain	VERB
ajst-29358	55	21	multi	multi	ADJ
ajst-29358	55	22	-	-	ADJ
ajst-29358	55	23	scale	scale	ADJ
ajst-29358	55	24	features	feature	NOUN
ajst-29358	55	25	while	while	SCONJ
ajst-29358	55	26	avoiding	avoid	VERB
ajst-29358	55	27	dimensionality	dimensionality	NOUN
ajst-29358	55	28	reduction	reduction	NOUN
ajst-29358	55	29	.	.	PUNCT
ajst-29358	56	1	this	this	DET
ajst-29358	56	2	study	study	NOUN
ajst-29358	56	3	replaces	replace	VERB
ajst-29358	56	4	the	the	DET
ajst-29358	56	5	three	three	NUM
ajst-29358	56	6	c2f	c2f	NOUN
ajst-29358	56	7	in	in	ADP
ajst-29358	56	8	the	the	DET
ajst-29358	56	9	backbone	backbone	NOUN
ajst-29358	56	10	network	network	NOUN
ajst-29358	56	11	with	with	ADP
ajst-29358	56	12	ema	ema	PROPN
ajst-29358	56	13	.	.	PUNCT
ajst-29358	57	1	h	h	NOUN
ajst-29358	57	2	h	h	NOUN
ajst-29358	58	1	h	h	NOUN
ajst-29358	58	2	h	h	NOUN
ajst-29358	58	3	h	h	NOUN
ajst-29358	59	1	c	c	PROPN
ajst-29358	59	2	w	w	ADP
ajst-29358	59	3	∙	∙	PROPN
ajst-29358	59	4	 	 	SPACE
ajst-29358	59	5	∙	∙	PROPN
ajst-29358	59	6	 	 	SPACE
ajst-29358	59	7	∙	∙	PROPN
ajst-29358	59	8	 	 	SPACE
ajst-29358	59	9	∙	∙	PROPN
ajst-29358	59	10	n	n	CCONJ
ajst-29358	59	11	c//n	c//n	ADV
ajst-29358	59	12	c//n	c//n	INTJ
ajst-29358	59	13	c//n	c//n	ADV
ajst-29358	59	14	w	w	PROPN
ajst-29358	59	15	w	w	PROPN
ajst-29358	59	16	w	w	PROPN
ajst-29358	59	17	x	x	X
ajst-29358	59	18	c//n	c//n	PROPN
ajst-29358	59	19	w	w	PROPN
ajst-29358	59	20	n	n	PRON
ajst-29358	59	21	n*batch	n*batch	NOUN
ajst-29358	59	22	 	 	SPACE
ajst-29358	59	23	size	size	NOUN
ajst-29358	59	24	figure	figure	NOUN
ajst-29358	59	25	2	2	NUM
ajst-29358	59	26	.	.	PUNCT
ajst-29358	59	27	ema	ema	PROPN
ajst-29358	59	28	structure	structure	PROPN
ajst-29358	59	29	diagram	diagram	PROPN
ajst-29358	59	30	"	"	PUNCT
ajst-29358	59	31	n	n	CCONJ
ajst-29358	59	32	"	"	PUNCT
ajst-29358	59	33	represents	represent	VERB
ajst-29358	59	34	grouping	group	VERB
ajst-29358	59	35	,	,	PUNCT
ajst-29358	59	36	"	"	PUNCT
ajst-29358	59	37	x	x	SYM
ajst-29358	59	38	avg	avg	NOUN
ajst-29358	59	39	pool	pool	NOUN
ajst-29358	59	40	"	"	PUNCT
ajst-29358	59	41	represents	represent	VERB
ajst-29358	59	42	one	one	NUM
ajst-29358	59	43	-	-	PUNCT
ajst-29358	59	44	dimensional	dimensional	ADJ
ajst-29358	59	45	horizontal	horizontal	ADJ
ajst-29358	59	46	global	global	ADJ
ajst-29358	59	47	pool	pool	NOUN
ajst-29358	59	48	,	,	PUNCT
ajst-29358	59	49	"	"	PUNCT
ajst-29358	59	50	y	y	PROPN
ajst-29358	59	51	avg	avg	PROPN
ajst-29358	59	52	pool	pool	NOUN
ajst-29358	59	53	"	"	PUNCT
ajst-29358	59	54	represents	represent	VERB
ajst-29358	59	55	one	one	NUM
ajst-29358	59	56	-	-	PUNCT
ajst-29358	59	57	dimensional	dimensional	ADJ
ajst-29358	59	58	vertical	vertical	ADJ
ajst-29358	59	59	global	global	ADJ
ajst-29358	59	60	pool	pool	NOUN
ajst-29358	59	61	2.3	2.3	NUM
ajst-29358	59	62	.	.	PUNCT
ajst-29358	60	1	neck	neck	NOUN
ajst-29358	60	2	improvement	improvement	NOUN
ajst-29358	60	3	to	to	PART
ajst-29358	60	4	improve	improve	VERB
ajst-29358	60	5	the	the	DET
ajst-29358	60	6	detection	detection	NOUN
ajst-29358	60	7	accuracy	accuracy	NOUN
ajst-29358	60	8	of	of	ADP
ajst-29358	60	9	small	small	ADJ
ajst-29358	60	10	targets	target	NOUN
ajst-29358	60	11	such	such	ADJ
ajst-29358	60	12	as	as	ADP
ajst-29358	60	13	large	large	ADJ
ajst-29358	60	14	spot	spot	NOUN
ajst-29358	60	15	disease	disease	NOUN
ajst-29358	60	16	lesions	lesion	NOUN
ajst-29358	60	17	,	,	PUNCT
ajst-29358	60	18	yolov8	yolov8	PROPN
ajst-29358	60	19	has	have	AUX
ajst-29358	60	20	successfully	successfully	ADV
ajst-29358	60	21	integrated	integrate	VERB
ajst-29358	60	22	the	the	DET
ajst-29358	60	23	advantages	advantage	NOUN
ajst-29358	60	24	of	of	ADP
ajst-29358	60	25	pan	pan	NOUN
ajst-29358	60	26	(	(	PUNCT
ajst-29358	60	27	path	path	NOUN
ajst-29358	60	28	aggregation	aggregation	NOUN
ajst-29358	60	29	network	network	NOUN
ajst-29358	60	30	)	)	PUNCT
ajst-29358	60	31	and	and	CCONJ
ajst-29358	60	32	fpn	fpn	VERB
ajst-29358	60	33	(	(	PUNCT
ajst-29358	60	34	feature	feature	NOUN
ajst-29358	60	35	pyramid	pyramid	NOUN
ajst-29358	60	36	network	network	NOUN
ajst-29358	60	37	)	)	PUNCT
ajst-29358	60	38	structures	structure	NOUN
ajst-29358	60	39	,	,	PUNCT
ajst-29358	60	40	and	and	CCONJ
ajst-29358	60	41	this	this	DET
ajst-29358	60	42	paper	paper	NOUN
ajst-29358	60	43	further	far	ADV
ajst-29358	60	44	introduces	introduce	VERB
ajst-29358	60	45	the	the	DET
ajst-29358	60	46	concept	concept	NOUN
ajst-29358	60	47	of	of	ADP
ajst-29358	60	48	bifpn	bifpn	PROPN
ajst-29358	60	49	(	(	PUNCT
ajst-29358	60	50	bidirectional	bidirectional	ADJ
ajst-29358	60	51	feature	feature	NOUN
ajst-29358	60	52	pyramid	pyramid	NOUN
ajst-29358	60	53	network	network	NOUN
ajst-29358	60	54	)	)	PUNCT
ajst-29358	60	55	,	,	PUNCT
ajst-29358	60	56	proposing	propose	VERB
ajst-29358	60	57	an	an	DET
ajst-29358	60	58	innovative	innovative	ADJ
ajst-29358	60	59	yolov8	yolov8	NOUN
ajst-29358	60	60	-	-	PUNCT
ajst-29358	60	61	bifpn	bifpn	PROPN
ajst-29358	60	62	feature	feature	NOUN
ajst-29358	60	63	fusion	fusion	NOUN
ajst-29358	60	64	network	network	NOUN
ajst-29358	60	65	through	through	ADP
ajst-29358	60	66	new	new	ADJ
ajst-29358	60	67	crossscale	crossscale	ADJ
ajst-29358	60	68	connection	connection	NOUN
ajst-29358	60	69	methods	method	NOUN
ajst-29358	60	70	and	and	CCONJ
ajst-29358	60	71	feature	feature	NOUN
ajst-29358	60	72	fusion	fusion	NOUN
ajst-29358	60	73	methods	method	NOUN
ajst-29358	60	74	,	,	PUNCT
ajst-29358	60	75	the	the	DET
ajst-29358	60	76	detailed	detailed	ADJ
ajst-29358	60	77	network	network	NOUN
ajst-29358	60	78	structure	structure	NOUN
ajst-29358	60	79	is	be	AUX
ajst-29358	60	80	shown	show	VERB
ajst-29358	60	81	in	in	ADP
ajst-29358	60	82	figure	figure	NOUN
ajst-29358	60	83	3	3	NUM
ajst-29358	60	84	.	.	SYM
ajst-29358	60	85	153	153	NUM
ajst-29358	60	86	p3	p3	PROPN
ajst-29358	60	87	p4	p4	ADJ
ajst-29358	60	88	p5	p5	ADJ
ajst-29358	60	89	p6	p6	ADJ
ajst-29358	60	90	p7	p7	NOUN
ajst-29358	60	91	(	(	PUNCT
ajst-29358	60	92	a)bifpn	a)bifpn	PROPN
ajst-29358	60	93	(	(	PUNCT
ajst-29358	60	94	b)yolov8	b)yolov8	NOUN
ajst-29358	60	95	-	-	PUNCT
ajst-29358	60	96	bifpn	bifpn	PROPN
ajst-29358	60	97	figure	figure	NOUN
ajst-29358	60	98	3	3	NUM
ajst-29358	60	99	.	.	PUNCT
ajst-29358	61	1	bifpn	bifpn	PROPN
ajst-29358	61	2	,	,	PUNCT
ajst-29358	61	3	yolov8	yolov8	PROPN
ajst-29358	61	4	-	-	PUNCT
ajst-29358	61	5	bifpn	bifpn	PROPN
ajst-29358	61	6	feature	feature	NOUN
ajst-29358	61	7	network	network	NOUN
ajst-29358	61	8	structure	structure	NOUN
ajst-29358	61	9	the	the	DET
ajst-29358	61	10	construction	construction	NOUN
ajst-29358	61	11	method	method	NOUN
ajst-29358	61	12	of	of	ADP
ajst-29358	61	13	the	the	DET
ajst-29358	61	14	fpn	fpn	ADJ
ajst-29358	61	15	feature	feature	NOUN
ajst-29358	61	16	pyramid	pyramid	NOUN
ajst-29358	61	17	is	be	AUX
ajst-29358	61	18	top	top	ADJ
ajst-29358	61	19	-	-	PUNCT
ajst-29358	61	20	down	down	NOUN
ajst-29358	61	21	,	,	PUNCT
ajst-29358	61	22	which	which	PRON
ajst-29358	61	23	can	can	AUX
ajst-29358	61	24	capture	capture	VERB
ajst-29358	61	25	detailed	detailed	ADJ
ajst-29358	61	26	information	information	NOUN
ajst-29358	61	27	of	of	ADP
ajst-29358	61	28	images	image	NOUN
ajst-29358	61	29	at	at	ADP
ajst-29358	61	30	different	different	ADJ
ajst-29358	61	31	scales	scale	NOUN
ajst-29358	61	32	,	,	PUNCT
ajst-29358	61	33	thereby	thereby	ADV
ajst-29358	61	34	enhancing	enhance	VERB
ajst-29358	61	35	the	the	DET
ajst-29358	61	36	performance	performance	NOUN
ajst-29358	61	37	and	and	CCONJ
ajst-29358	61	38	semantic	semantic	ADJ
ajst-29358	61	39	information	information	NOUN
ajst-29358	61	40	robustness	robustness	NOUN
ajst-29358	61	41	of	of	ADP
ajst-29358	61	42	the	the	DET
ajst-29358	61	43	algorithm	algorithm	NOUN
ajst-29358	61	44	.	.	PUNCT
ajst-29358	62	1	nevertheless	nevertheless	ADV
ajst-29358	62	2	,	,	PUNCT
ajst-29358	62	3	the	the	DET
ajst-29358	62	4	issue	issue	NOUN
ajst-29358	62	5	of	of	ADP
ajst-29358	62	6	semantic	semantic	ADJ
ajst-29358	62	7	information	information	NOUN
ajst-29358	62	8	differences	difference	NOUN
ajst-29358	62	9	at	at	ADP
ajst-29358	62	10	different	different	ADJ
ajst-29358	62	11	levels	level	NOUN
ajst-29358	62	12	in	in	ADP
ajst-29358	62	13	fpn	fpn	PROPN
ajst-29358	62	14	still	still	ADV
ajst-29358	62	15	exists	exist	VERB
ajst-29358	62	16	in	in	ADP
ajst-29358	62	17	yolov8	yolov8	PROPN
ajst-29358	62	18	.	.	PUNCT
ajst-29358	63	1	the	the	DET
ajst-29358	63	2	construction	construction	NOUN
ajst-29358	63	3	method	method	NOUN
ajst-29358	63	4	of	of	ADP
ajst-29358	63	5	pan	pan	NOUN
ajst-29358	63	6	adds	add	VERB
ajst-29358	63	7	a	a	DET
ajst-29358	63	8	bottom	bottom	ADJ
ajst-29358	63	9	-	-	PUNCT
ajst-29358	63	10	up	up	ADP
ajst-29358	63	11	feature	feature	NOUN
ajst-29358	63	12	transmission	transmission	NOUN
ajst-29358	63	13	method	method	NOUN
ajst-29358	63	14	to	to	PART
ajst-29358	63	15	fpn	fpn	VERB
ajst-29358	63	16	,	,	PUNCT
ajst-29358	63	17	effectively	effectively	ADV
ajst-29358	63	18	enhancing	enhance	VERB
ajst-29358	63	19	the	the	DET
ajst-29358	63	20	transmission	transmission	NOUN
ajst-29358	63	21	of	of	ADP
ajst-29358	63	22	low	low	ADJ
ajst-29358	63	23	-	-	PUNCT
ajst-29358	63	24	level	level	NOUN
ajst-29358	63	25	positioning	positioning	NOUN
ajst-29358	63	26	features	feature	NOUN
ajst-29358	63	27	to	to	ADP
ajst-29358	63	28	the	the	DET
ajst-29358	63	29	high	high	ADJ
ajst-29358	63	30	level	level	NOUN
ajst-29358	63	31	,	,	PUNCT
ajst-29358	63	32	and	and	CCONJ
ajst-29358	63	33	constructing	construct	VERB
ajst-29358	63	34	a	a	DET
ajst-29358	63	35	fused	fuse	VERB
ajst-29358	63	36	feature	feature	NOUN
ajst-29358	63	37	pyramid	pyramid	NOUN
ajst-29358	63	38	.	.	PUNCT
ajst-29358	64	1	yolov8	yolov8	PROPN
ajst-29358	64	2	further	far	ADV
ajst-29358	64	3	optimizes	optimize	VERB
ajst-29358	64	4	on	on	ADP
ajst-29358	64	5	the	the	DET
ajst-29358	64	6	basis	basis	NOUN
ajst-29358	64	7	of	of	ADP
ajst-29358	64	8	pan	pan	NOUN
ajst-29358	64	9	,	,	PUNCT
ajst-29358	64	10	by	by	ADP
ajst-29358	64	11	removing	remove	VERB
ajst-29358	64	12	nodes	node	NOUN
ajst-29358	64	13	that	that	PRON
ajst-29358	64	14	do	do	AUX
ajst-29358	64	15	not	not	PART
ajst-29358	64	16	participate	participate	VERB
ajst-29358	64	17	in	in	ADP
ajst-29358	64	18	feature	feature	NOUN
ajst-29358	64	19	fusion	fusion	NOUN
ajst-29358	64	20	,	,	PUNCT
ajst-29358	64	21	both	both	PRON
ajst-29358	64	22	simplifying	simplify	VERB
ajst-29358	64	23	the	the	DET
ajst-29358	64	24	network	network	NOUN
ajst-29358	64	25	and	and	CCONJ
ajst-29358	64	26	improving	improve	VERB
ajst-29358	64	27	efficiency	efficiency	NOUN
ajst-29358	64	28	.	.	PUNCT
ajst-29358	65	1	however	however	ADV
ajst-29358	65	2	,	,	PUNCT
ajst-29358	65	3	these	these	DET
ajst-29358	65	4	improvements	improvement	NOUN
ajst-29358	65	5	still	still	ADV
ajst-29358	65	6	have	have	VERB
ajst-29358	65	7	certain	certain	ADJ
ajst-29358	65	8	limitations	limitation	NOUN
ajst-29358	65	9	in	in	ADP
ajst-29358	65	10	small	small	ADJ
ajst-29358	65	11	target	target	NOUN
ajst-29358	65	12	detection	detection	NOUN
ajst-29358	65	13	.	.	PUNCT
ajst-29358	66	1	small	small	ADJ
ajst-29358	66	2	targets	target	NOUN
ajst-29358	66	3	are	be	AUX
ajst-29358	66	4	easily	easily	ADV
ajst-29358	66	5	disturbed	disturb	VERB
ajst-29358	66	6	by	by	ADP
ajst-29358	66	7	large	large	ADJ
ajst-29358	66	8	targets	target	NOUN
ajst-29358	66	9	during	during	ADP
ajst-29358	66	10	feature	feature	NOUN
ajst-29358	66	11	extraction	extraction	NOUN
ajst-29358	66	12	,	,	PUNCT
ajst-29358	66	13	and	and	CCONJ
ajst-29358	66	14	inconspicuous	inconspicuous	ADJ
ajst-29358	66	15	feature	feature	NOUN
ajst-29358	66	16	information	information	NOUN
ajst-29358	66	17	is	be	AUX
ajst-29358	66	18	also	also	ADV
ajst-29358	66	19	easily	easily	ADV
ajst-29358	66	20	ignored	ignore	VERB
ajst-29358	66	21	by	by	ADP
ajst-29358	66	22	the	the	DET
ajst-29358	66	23	network	network	NOUN
ajst-29358	66	24	,	,	PUNCT
ajst-29358	66	25	leading	lead	VERB
ajst-29358	66	26	to	to	ADP
ajst-29358	66	27	the	the	DET
ajst-29358	66	28	gradual	gradual	ADJ
ajst-29358	66	29	loss	loss	NOUN
ajst-29358	66	30	of	of	ADP
ajst-29358	66	31	key	key	ADJ
ajst-29358	66	32	information	information	NOUN
ajst-29358	66	33	of	of	ADP
ajst-29358	66	34	small	small	ADJ
ajst-29358	66	35	targets	target	NOUN
ajst-29358	66	36	,	,	PUNCT
ajst-29358	66	37	thereby	thereby	ADV
ajst-29358	66	38	affecting	affect	VERB
ajst-29358	66	39	the	the	DET
ajst-29358	66	40	accuracy	accuracy	NOUN
ajst-29358	66	41	of	of	ADP
ajst-29358	66	42	detection	detection	NOUN
ajst-29358	66	43	.	.	PUNCT
ajst-29358	67	1	the	the	DET
ajst-29358	67	2	bifpn	bifpn	PROPN
ajst-29358	67	3	facilitates	facilitate	VERB
ajst-29358	67	4	the	the	DET
ajst-29358	67	5	connection	connection	NOUN
ajst-29358	67	6	of	of	ADP
ajst-29358	67	7	nodes	node	NOUN
ajst-29358	67	8	at	at	ADP
ajst-29358	67	9	the	the	DET
ajst-29358	67	10	input	input	NOUN
ajst-29358	67	11	and	and	CCONJ
ajst-29358	67	12	output	output	NOUN
ajst-29358	67	13	ends	end	VERB
ajst-29358	67	14	within	within	ADP
ajst-29358	67	15	the	the	DET
ajst-29358	67	16	same	same	ADJ
ajst-29358	67	17	level	level	NOUN
ajst-29358	67	18	via	via	ADP
ajst-29358	67	19	an	an	DET
ajst-29358	67	20	edge	edge	NOUN
ajst-29358	67	21	,	,	PUNCT
ajst-29358	67	22	thereby	thereby	ADV
ajst-29358	67	23	eliminating	eliminate	VERB
ajst-29358	67	24	nodes	node	NOUN
ajst-29358	67	25	incapable	incapable	ADJ
ajst-29358	67	26	of	of	ADP
ajst-29358	67	27	feature	feature	NOUN
ajst-29358	67	28	integration	integration	NOUN
ajst-29358	67	29	.	.	PUNCT
ajst-29358	68	1	moreover	moreover	ADV
ajst-29358	68	2	,	,	PUNCT
ajst-29358	68	3	it	it	PRON
ajst-29358	68	4	incorporates	incorporate	VERB
ajst-29358	68	5	weighted	weight	VERB
ajst-29358	68	6	feature	feature	NOUN
ajst-29358	68	7	fusion	fusion	NOUN
ajst-29358	68	8	techniques	technique	NOUN
ajst-29358	68	9	,	,	PUNCT
ajst-29358	68	10	which	which	PRON
ajst-29358	68	11	supplant	supplant	VERB
ajst-29358	68	12	the	the	DET
ajst-29358	68	13	conventional	conventional	ADJ
ajst-29358	68	14	feature	feature	NOUN
ajst-29358	68	15	concatenation	concatenation	NOUN
ajst-29358	68	16	approach	approach	NOUN
ajst-29358	68	17	,	,	PUNCT
ajst-29358	68	18	thus	thus	ADV
ajst-29358	68	19	more	more	ADV
ajst-29358	68	20	effectively	effectively	ADV
ajst-29358	68	21	preserving	preserve	VERB
ajst-29358	68	22	the	the	DET
ajst-29358	68	23	hierarchical	hierarchical	ADJ
ajst-29358	68	24	structure	structure	NOUN
ajst-29358	68	25	of	of	ADP
ajst-29358	68	26	features	feature	NOUN
ajst-29358	68	27	.	.	PUNCT
ajst-29358	69	1	in	in	ADP
ajst-29358	69	2	this	this	DET
ajst-29358	69	3	study	study	NOUN
ajst-29358	69	4	,	,	PUNCT
ajst-29358	69	5	an	an	DET
ajst-29358	69	6	enhancement	enhancement	NOUN
ajst-29358	69	7	to	to	ADP
ajst-29358	69	8	the	the	DET
ajst-29358	69	9	bifpn	bifpn	PROPN
ajst-29358	69	10	is	be	AUX
ajst-29358	69	11	proposed	propose	VERB
ajst-29358	69	12	,	,	PUNCT
ajst-29358	69	13	introducing	introduce	VERB
ajst-29358	69	14	a	a	DET
ajst-29358	69	15	novel	novel	ADJ
ajst-29358	69	16	cross	cross	ADJ
ajst-29358	69	17	-	-	ADJ
ajst-29358	69	18	layer	layer	ADJ
ajst-29358	69	19	connection	connection	NOUN
ajst-29358	69	20	methodology	methodology	NOUN
ajst-29358	69	21	.	.	PUNCT
ajst-29358	70	1	specifically	specifically	ADV
ajst-29358	70	2	,	,	PUNCT
ajst-29358	70	3	for	for	ADP
ajst-29358	70	4	the	the	DET
ajst-29358	70	5	second	second	ADJ
ajst-29358	70	6	and	and	CCONJ
ajst-29358	70	7	third	third	ADJ
ajst-29358	70	8	layers	layer	NOUN
ajst-29358	70	9	of	of	ADP
ajst-29358	70	10	the	the	DET
ajst-29358	70	11	network	network	NOUN
ajst-29358	70	12	,	,	PUNCT
ajst-29358	70	13	3×3	3×3	NUM
ajst-29358	70	14	and	and	CCONJ
ajst-29358	70	15	1×1	1×1	ADJ
ajst-29358	70	16	convolutional	convolutional	ADJ
ajst-29358	70	17	operations	operation	NOUN
ajst-29358	70	18	are	be	AUX
ajst-29358	70	19	respectively	respectively	ADV
ajst-29358	70	20	integrated	integrate	VERB
ajst-29358	70	21	to	to	PART
ajst-29358	70	22	adjust	adjust	VERB
ajst-29358	70	23	the	the	DET
ajst-29358	70	24	dimensions	dimension	NOUN
ajst-29358	70	25	of	of	ADP
ajst-29358	70	26	the	the	DET
ajst-29358	70	27	feature	feature	NOUN
ajst-29358	70	28	maps	map	NOUN
ajst-29358	70	29	to	to	ADP
ajst-29358	70	30	80×80	80×80	NUM
ajst-29358	70	31	,	,	PUNCT
ajst-29358	70	32	and	and	CCONJ
ajst-29358	70	33	subsequently	subsequently	ADV
ajst-29358	70	34	merge	merge	VERB
ajst-29358	70	35	these	these	PRON
ajst-29358	70	36	with	with	ADP
ajst-29358	70	37	the	the	DET
ajst-29358	70	38	outputs	output	NOUN
ajst-29358	70	39	of	of	ADP
ajst-29358	70	40	the	the	DET
ajst-29358	70	41	corresponding	corresponding	ADJ
ajst-29358	70	42	layers	layer	NOUN
ajst-29358	70	43	.	.	PUNCT
ajst-29358	71	1	additionally	additionally	ADV
ajst-29358	71	2	,	,	PUNCT
ajst-29358	71	3	a	a	DET
ajst-29358	71	4	1×1	1×1	ADJ
ajst-29358	71	5	convolutional	convolutional	ADJ
ajst-29358	71	6	layer	layer	NOUN
ajst-29358	71	7	is	be	AUX
ajst-29358	71	8	inserted	insert	VERB
ajst-29358	71	9	between	between	ADP
ajst-29358	71	10	the	the	DET
ajst-29358	71	11	input	input	NOUN
ajst-29358	71	12	and	and	CCONJ
ajst-29358	71	13	intermediate	intermediate	ADJ
ajst-29358	71	14	nodes	node	NOUN
ajst-29358	71	15	,	,	PUNCT
ajst-29358	71	16	and	and	CCONJ
ajst-29358	71	17	a	a	DET
ajst-29358	71	18	weighted	weight	VERB
ajst-29358	71	19	fusion	fusion	NOUN
ajst-29358	71	20	strategy	strategy	NOUN
ajst-29358	71	21	is	be	AUX
ajst-29358	71	22	employed	employ	VERB
ajst-29358	71	23	,	,	PUNCT
ajst-29358	71	24	which	which	PRON
ajst-29358	71	25	not	not	PART
ajst-29358	71	26	only	only	ADV
ajst-29358	71	27	reduces	reduce	VERB
ajst-29358	71	28	the	the	DET
ajst-29358	71	29	number	number	NOUN
ajst-29358	71	30	of	of	ADP
ajst-29358	71	31	model	model	NOUN
ajst-29358	71	32	parameters	parameter	NOUN
ajst-29358	71	33	but	but	CCONJ
ajst-29358	71	34	also	also	ADV
ajst-29358	71	35	more	more	ADV
ajst-29358	71	36	effectively	effectively	ADV
ajst-29358	71	37	retains	retain	VERB
ajst-29358	71	38	a	a	DET
ajst-29358	71	39	greater	great	ADJ
ajst-29358	71	40	amount	amount	NOUN
ajst-29358	71	41	of	of	ADP
ajst-29358	71	42	original	original	ADJ
ajst-29358	71	43	feature	feature	NOUN
ajst-29358	71	44	information	information	NOUN
ajst-29358	71	45	.	.	PUNCT
ajst-29358	72	1	2.4	2.4	NUM
ajst-29358	72	2	.	.	PUNCT
ajst-29358	72	3	loss	loss	NOUN
ajst-29358	72	4	function	function	VERB
ajst-29358	72	5	the	the	DET
ajst-29358	72	6	wiou	wiou	NOUN
ajst-29358	72	7	loss	loss	NOUN
ajst-29358	72	8	function	function	NOUN
ajst-29358	72	9	is	be	AUX
ajst-29358	72	10	used	use	VERB
ajst-29358	72	11	to	to	PART
ajst-29358	72	12	replace	replace	VERB
ajst-29358	72	13	the	the	DET
ajst-29358	72	14	original	original	ADJ
ajst-29358	72	15	loss	loss	NOUN
ajst-29358	72	16	function	function	NOUN
ajst-29358	72	17	to	to	PART
ajst-29358	72	18	improve	improve	VERB
ajst-29358	72	19	the	the	DET
ajst-29358	72	20	prediction	prediction	NOUN
ajst-29358	72	21	accuracy	accuracy	NOUN
ajst-29358	72	22	of	of	ADP
ajst-29358	72	23	anchor	anchor	NOUN
ajst-29358	72	24	frames	frame	NOUN
ajst-29358	72	25	.	.	PUNCT
ajst-29358	73	1	the	the	DET
ajst-29358	73	2	wiou	wiou	NOUN
ajst-29358	73	3	v3	v3	PROPN
ajst-29358	73	4	loss	loss	NOUN
ajst-29358	73	5	function	function	NOUN
ajst-29358	73	6	dynamically	dynamically	ADV
ajst-29358	73	7	allocates	allocate	VERB
ajst-29358	73	8	gradient	gradient	ADJ
ajst-29358	73	9	gains	gain	NOUN
ajst-29358	73	10	to	to	PART
ajst-29358	73	11	suppress	suppress	VERB
ajst-29358	73	12	harmful	harmful	ADJ
ajst-29358	73	13	gradients	gradient	NOUN
ajst-29358	73	14	generated	generate	VERB
ajst-29358	73	15	by	by	ADP
ajst-29358	73	16	low	low	ADJ
ajst-29358	73	17	-	-	PUNCT
ajst-29358	73	18	quality	quality	NOUN
ajst-29358	73	19	images	image	NOUN
ajst-29358	73	20	,	,	PUNCT
ajst-29358	73	21	overcoming	overcome	VERB
ajst-29358	73	22	the	the	DET
ajst-29358	73	23	insufficient	insufficient	ADJ
ajst-29358	73	24	localization	localization	NOUN
ajst-29358	73	25	performance	performance	NOUN
ajst-29358	73	26	of	of	ADP
ajst-29358	73	27	the	the	DET
ajst-29358	73	28	yolov8n	yolov8n	PROPN
ajst-29358	73	29	network	network	NOUN
ajst-29358	73	30	,	,	PUNCT
ajst-29358	73	31	allowing	allow	VERB
ajst-29358	73	32	the	the	DET
ajst-29358	73	33	network	network	NOUN
ajst-29358	73	34	to	to	PART
ajst-29358	73	35	more	more	ADV
ajst-29358	73	36	accurately	accurately	ADV
ajst-29358	73	37	locate	locate	ADJ
ajst-29358	73	38	lesions	lesion	NOUN
ajst-29358	73	39	in	in	ADP
ajst-29358	73	40	joint	joint	ADJ
ajst-29358	73	41	misalignment	misalignment	NOUN
ajst-29358	73	42	detection	detection	NOUN
ajst-29358	73	43	,	,	PUNCT
ajst-29358	73	44	thereby	thereby	ADV
ajst-29358	73	45	reducing	reduce	VERB
ajst-29358	73	46	errors	error	NOUN
ajst-29358	73	47	and	and	CCONJ
ajst-29358	73	48	improving	improve	VERB
ajst-29358	73	49	the	the	DET
ajst-29358	73	50	model	model	NOUN
ajst-29358	73	51	's	's	PART
ajst-29358	73	52	recognition	recognition	NOUN
ajst-29358	73	53	accuracy	accuracy	NOUN
ajst-29358	73	54	and	and	CCONJ
ajst-29358	73	55	generalization	generalization	NOUN
ajst-29358	73	56	capability	capability	NOUN
ajst-29358	73	57	.	.	PUNCT
ajst-29358	74	1	1wiouv	1wiouv	NUM
ajst-29358	74	2	wiou	wiou	NOUN
ajst-29358	74	3	ioul	ioul	NOUN
ajst-29358	74	4	r	r	NOUN
ajst-29358	74	5	l	l	NOUN
ajst-29358	74	6	(	(	PUNCT
ajst-29358	74	7	1	1	NUM
ajst-29358	74	8	)	)	SYM
ajst-29358	74	9	2	2	NUM
ajst-29358	74	10	2	2	NUM
ajst-29358	74	11	2	2	NUM
ajst-29358	74	12	2	2	NUM
ajst-29358	74	13	*	*	PUNCT
ajst-29358	74	14	(	(	PUNCT
ajst-29358	74	15	)	)	PUNCT
ajst-29358	74	16	(	(	PUNCT
ajst-29358	74	17	)	)	PUNCT
ajst-29358	74	18	exp	exp	NOUN
ajst-29358	74	19	(	(	PUNCT
ajst-29358	74	20	)	)	PUNCT
ajst-29358	74	21	(	(	PUNCT
ajst-29358	74	22	)	)	PUNCT
ajst-29358	74	23	gt	gt	PROPN
ajst-29358	74	24	gt	gt	PROPN
ajst-29358	74	25	wiou	wiou	NOUN
ajst-29358	74	26	g	g	PROPN
ajst-29358	74	27	g	g	NOUN
ajst-29358	74	28	x	x	PUNCT
ajst-29358	74	29	x	x	VERB
ajst-29358	74	30	y	y	NOUN
ajst-29358	74	31	y	y	NOUN
ajst-29358	74	32	r	r	PROPN
ajst-29358	74	33	w	w	PROPN
ajst-29358	74	34	h	h	PROPN
ajst-29358	74	35			PROPN
ajst-29358	74	36			VERB
ajst-29358	74	37			PROPN
ajst-29358	74	38			PRON
ajst-29358	74	39			PUNCT
ajst-29358	74	40	(	(	PUNCT
ajst-29358	74	41	2	2	X
ajst-29358	74	42	)	)	PUNCT
ajst-29358	74	43	wiou	wiou	NOUN
ajst-29358	74	44	loss	loss	NOUN
ajst-29358	74	45	function	function	NOUN
ajst-29358	74	46	is	be	AUX
ajst-29358	74	47	a	a	DET
ajst-29358	74	48	loss	loss	NOUN
ajst-29358	74	49	function	function	NOUN
ajst-29358	74	50	used	use	VERB
ajst-29358	74	51	for	for	ADP
ajst-29358	74	52	bounding	bound	VERB
ajst-29358	74	53	box	box	NOUN
ajst-29358	74	54	regression	regression	NOUN
ajst-29358	74	55	in	in	ADP
ajst-29358	74	56	target	target	NOUN
ajst-29358	74	57	detection	detection	NOUN
ajst-29358	74	58	,	,	PUNCT
ajst-29358	74	59	which	which	PRON
ajst-29358	74	60	optimizes	optimize	VERB
ajst-29358	74	61	the	the	DET
ajst-29358	74	62	quality	quality	NOUN
ajst-29358	74	63	assessment	assessment	NOUN
ajst-29358	74	64	of	of	ADP
ajst-29358	74	65	anchor	anchor	NOUN
ajst-29358	74	66	frames	frame	NOUN
ajst-29358	74	67	during	during	ADP
ajst-29358	74	68	training	training	NOUN
ajst-29358	74	69	by	by	ADP
ajst-29358	74	70	dynamically	dynamically	ADV
ajst-29358	74	71	allocating	allocate	VERB
ajst-29358	74	72	gradient	gradient	NOUN
ajst-29358	74	73	gains	gain	NOUN
ajst-29358	74	74	.	.	PUNCT
ajst-29358	75	1	compared	compare	VERB
ajst-29358	75	2	with	with	ADP
ajst-29358	75	3	the	the	DET
ajst-29358	75	4	ciou	ciou	NOUN
ajst-29358	75	5	loss	loss	NOUN
ajst-29358	75	6	function	function	NOUN
ajst-29358	75	7	,	,	PUNCT
ajst-29358	75	8	the	the	DET
ajst-29358	75	9	wiou	wiou	NOUN
ajst-29358	75	10	v3	v3	PROPN
ajst-29358	75	11	loss	loss	NOUN
ajst-29358	75	12	function	function	NOUN
ajst-29358	75	13	introduces	introduce	VERB
ajst-29358	75	14	a	a	DET
ajst-29358	75	15	dynamic	dynamic	ADJ
ajst-29358	75	16	non	non	ADJ
ajst-29358	75	17	-	-	ADJ
ajst-29358	75	18	monotonic	monotonic	ADJ
ajst-29358	75	19	mechanism	mechanism	NOUN
ajst-29358	75	20	,	,	PUNCT
ajst-29358	75	21	which	which	PRON
ajst-29358	75	22	uses	use	VERB
ajst-29358	75	23	"	"	PUNCT
ajst-29358	75	24	outlier	outlier	ADJ
ajst-29358	75	25	degree	degree	NOUN
ajst-29358	75	26	"	"	PUNCT
ajst-29358	75	27	instead	instead	ADV
ajst-29358	75	28	of	of	ADP
ajst-29358	75	29	iou	iou	VERB
ajst-29358	75	30	for	for	ADP
ajst-29358	75	31	anchor	anchor	NOUN
ajst-29358	75	32	frame	frame	NOUN
ajst-29358	75	33	quality	quality	NOUN
ajst-29358	75	34	assessment	assessment	NOUN
ajst-29358	75	35	and	and	CCONJ
ajst-29358	75	36	provides	provide	VERB
ajst-29358	75	37	a	a	DET
ajst-29358	75	38	wise	wise	ADJ
ajst-29358	75	39	gradient	gradient	NOUN
ajst-29358	75	40	gain	gain	NOUN
ajst-29358	75	41	allocation	allocation	NOUN
ajst-29358	75	42	strategy	strategy	NOUN
ajst-29358	75	43	.	.	PUNCT
ajst-29358	76	1	this	this	DET
ajst-29358	76	2	strategy	strategy	NOUN
ajst-29358	76	3	reduces	reduce	VERB
ajst-29358	76	4	the	the	DET
ajst-29358	76	5	competitiveness	competitiveness	NOUN
ajst-29358	76	6	of	of	ADP
ajst-29358	76	7	high	high	ADJ
ajst-29358	76	8	-	-	PUNCT
ajst-29358	76	9	quality	quality	NOUN
ajst-29358	76	10	anchor	anchor	NOUN
ajst-29358	76	11	frames	frame	NOUN
ajst-29358	76	12	while	while	SCONJ
ajst-29358	76	13	also	also	ADV
ajst-29358	76	14	reducing	reduce	VERB
ajst-29358	76	15	the	the	DET
ajst-29358	76	16	harmful	harmful	ADJ
ajst-29358	76	17	gradients	gradient	NOUN
ajst-29358	76	18	generated	generate	VERB
ajst-29358	76	19	by	by	ADP
ajst-29358	76	20	lowquality	lowquality	NOUN
ajst-29358	76	21	examples	example	NOUN
ajst-29358	76	22	,	,	PUNCT
ajst-29358	76	23	allowing	allow	VERB
ajst-29358	76	24	wiou	wiou	NOUN
ajst-29358	76	25	to	to	PART
ajst-29358	76	26	focus	focus	VERB
ajst-29358	76	27	more	more	ADV
ajst-29358	76	28	on	on	ADP
ajst-29358	76	29	ordinary	ordinary	ADJ
ajst-29358	76	30	quality	quality	NOUN
ajst-29358	76	31	anchor	anchor	NOUN
ajst-29358	76	32	frames	frame	NOUN
ajst-29358	76	33	and	and	CCONJ
ajst-29358	76	34	improve	improve	VERB
ajst-29358	76	35	the	the	DET
ajst-29358	76	36	overall	overall	ADJ
ajst-29358	76	37	performance	performance	NOUN
ajst-29358	76	38	of	of	ADP
ajst-29358	76	39	the	the	DET
ajst-29358	76	40	detector	detector	NOUN
ajst-29358	76	41	.	.	PUNCT
ajst-29358	77	1	3	3	X
ajst-29358	77	2	.	.	X
ajst-29358	77	3	experimental	experimental	ADJ
ajst-29358	77	4	data	datum	NOUN
ajst-29358	77	5	and	and	CCONJ
ajst-29358	77	6	evaluation	evaluation	NOUN
ajst-29358	77	7	metrics	metric	NOUN
ajst-29358	77	8	3.1	3.1	NUM
ajst-29358	77	9	.	.	PUNCT
ajst-29358	77	10	dataset	dataset	VERB
ajst-29358	77	11	the	the	DET
ajst-29358	77	12	corn	corn	NOUN
ajst-29358	77	13	large	large	ADJ
ajst-29358	77	14	spot	spot	NOUN
ajst-29358	77	15	disease	disease	NOUN
ajst-29358	77	16	image	image	NOUN
ajst-29358	77	17	acquisition	acquisition	NOUN
ajst-29358	77	18	platform	platform	NOUN
ajst-29358	77	19	is	be	AUX
ajst-29358	77	20	composed	compose	VERB
ajst-29358	77	21	of	of	ADP
ajst-29358	77	22	a	a	DET
ajst-29358	77	23	uav	uav	PROPN
ajst-29358	77	24	-	-	PUNCT
ajst-29358	77	25	mounted	mount	VERB
ajst-29358	77	26	camera	camera	NOUN
ajst-29358	77	27	.	.	PUNCT
ajst-29358	78	1	table	table	NOUN
ajst-29358	78	2	1	1	NUM
ajst-29358	78	3	shows	show	VERB
ajst-29358	78	4	the	the	DET
ajst-29358	78	5	corresponding	correspond	VERB
ajst-29358	78	6	equipment	equipment	NOUN
ajst-29358	78	7	models	model	NOUN
ajst-29358	78	8	and	and	CCONJ
ajst-29358	78	9	parameter	parameter	NOUN
ajst-29358	78	10	configurations	configuration	NOUN
ajst-29358	78	11	.	.	PUNCT
ajst-29358	79	1	table	table	NOUN
ajst-29358	79	2	1	1	NUM
ajst-29358	79	3	.	.	PUNCT
ajst-29358	80	1	uav	uav	PROPN
ajst-29358	80	2	platform	platform	PROPN
ajst-29358	80	3	parameters	parameter	NOUN
ajst-29358	80	4	numble	numble	ADJ
ajst-29358	80	5	name	name	NOUN
ajst-29358	80	6	scheme	scheme	NOUN
ajst-29358	80	7	2	2	NUM
ajst-29358	80	8	1	1	NUM
ajst-29358	80	9	uav	uav	PROPN
ajst-29358	80	10	dji	dji	PROPN
ajst-29358	80	11	mavic	mavic	PROPN
ajst-29358	80	12	air	air	NOUN
ajst-29358	81	1	2s	2s	NUM
ajst-29358	81	2	2	2	NUM
ajst-29358	81	3	image	image	NOUN
ajst-29358	81	4	resolution	resolution	NOUN
ajst-29358	81	5	6000×4000	6000×4000	NUM
ajst-29358	81	6	3	3	NUM
ajst-29358	81	7	maximum	maximum	ADJ
ajst-29358	81	8	flight	flight	NOUN
ajst-29358	81	9	time	time	NOUN
ajst-29358	81	10	34	34	NUM
ajst-29358	81	11	minutes	minute	NOUN
ajst-29358	81	12	4	4	NUM
ajst-29358	81	13	camera	camera	NOUN
ajst-29358	81	14	sensor	sensor	NOUN
ajst-29358	81	15	1/2	1/2	NUM
ajst-29358	81	16	inch	inch	NOUN
ajst-29358	81	17	cmos	cmos	NOUN
ajst-29358	81	18	,	,	PUNCT
ajst-29358	81	19	48mp	48mp	ADJ
ajst-29358	81	20	154	154	NUM
ajst-29358	81	21	to	to	PART
ajst-29358	81	22	construct	construct	VERB
ajst-29358	81	23	the	the	DET
ajst-29358	81	24	corn	corn	NOUN
ajst-29358	81	25	large	large	ADJ
ajst-29358	81	26	spot	spot	NOUN
ajst-29358	81	27	disease	disease	NOUN
ajst-29358	81	28	disease	disease	NOUN
ajst-29358	81	29	dataset	dataset	VERB
ajst-29358	81	30	,	,	PUNCT
ajst-29358	81	31	this	this	DET
ajst-29358	81	32	study	study	NOUN
ajst-29358	81	33	used	use	VERB
ajst-29358	81	34	the	the	DET
ajst-29358	81	35	labelimg	labelimg	ADJ
ajst-29358	81	36	tool	tool	NOUN
ajst-29358	81	37	to	to	PART
ajst-29358	81	38	accurately	accurately	ADV
ajst-29358	81	39	label	label	VERB
ajst-29358	81	40	4000	4000	NUM
ajst-29358	81	41	corn	corn	NOUN
ajst-29358	81	42	leaf	leaf	NOUN
ajst-29358	81	43	disease	disease	NOUN
ajst-29358	81	44	images	image	NOUN
ajst-29358	81	45	.	.	PUNCT
ajst-29358	82	1	to	to	PART
ajst-29358	82	2	improve	improve	VERB
ajst-29358	82	3	the	the	DET
ajst-29358	82	4	generalization	generalization	NOUN
ajst-29358	82	5	capability	capability	NOUN
ajst-29358	82	6	of	of	ADP
ajst-29358	82	7	the	the	DET
ajst-29358	82	8	model	model	NOUN
ajst-29358	82	9	during	during	ADP
ajst-29358	82	10	evaluation	evaluation	NOUN
ajst-29358	82	11	,	,	PUNCT
ajst-29358	82	12	the	the	DET
ajst-29358	82	13	dataset	dataset	NOUN
ajst-29358	82	14	was	be	AUX
ajst-29358	82	15	divided	divide	VERB
ajst-29358	82	16	in	in	ADP
ajst-29358	82	17	a	a	DET
ajst-29358	82	18	ratio	ratio	NOUN
ajst-29358	82	19	of	of	ADP
ajst-29358	82	20	8:1:1	8:1:1	NUM
ajst-29358	82	21	to	to	PART
ajst-29358	82	22	form	form	VERB
ajst-29358	82	23	the	the	DET
ajst-29358	82	24	training	training	NOUN
ajst-29358	82	25	set	set	NOUN
ajst-29358	82	26	,	,	PUNCT
ajst-29358	82	27	test	test	NOUN
ajst-29358	82	28	set	set	NOUN
ajst-29358	82	29	,	,	PUNCT
ajst-29358	82	30	and	and	CCONJ
ajst-29358	82	31	validation	validation	NOUN
ajst-29358	82	32	set	set	NOUN
ajst-29358	82	33	.	.	PUNCT
ajst-29358	83	1	a	a	DET
ajst-29358	83	2	total	total	NOUN
ajst-29358	83	3	of	of	ADP
ajst-29358	83	4	3200	3200	NUM
ajst-29358	83	5	,	,	PUNCT
ajst-29358	83	6	400	400	NUM
ajst-29358	83	7	,	,	PUNCT
ajst-29358	83	8	and	and	CCONJ
ajst-29358	83	9	400	400	NUM
ajst-29358	83	10	images	image	NOUN
ajst-29358	83	11	were	be	AUX
ajst-29358	83	12	obtained	obtain	VERB
ajst-29358	83	13	,	,	PUNCT
ajst-29358	83	14	meeting	meet	VERB
ajst-29358	83	15	the	the	DET
ajst-29358	83	16	needs	need	NOUN
ajst-29358	83	17	for	for	ADP
ajst-29358	83	18	comprehensive	comprehensive	ADJ
ajst-29358	83	19	training	training	NOUN
ajst-29358	83	20	and	and	CCONJ
ajst-29358	83	21	model	model	NOUN
ajst-29358	83	22	evaluation	evaluation	NOUN
ajst-29358	83	23	.	.	PUNCT
ajst-29358	84	1	this	this	DET
ajst-29358	84	2	annotation	annotation	NOUN
ajst-29358	84	3	and	and	CCONJ
ajst-29358	84	4	division	division	NOUN
ajst-29358	84	5	process	process	NOUN
ajst-29358	84	6	aims	aim	VERB
ajst-29358	84	7	to	to	PART
ajst-29358	84	8	provide	provide	VERB
ajst-29358	84	9	the	the	DET
ajst-29358	84	10	model	model	NOUN
ajst-29358	84	11	with	with	ADP
ajst-29358	84	12	accurate	accurate	ADJ
ajst-29358	84	13	enough	enough	ADJ
ajst-29358	84	14	training	training	NOUN
ajst-29358	84	15	data	datum	NOUN
ajst-29358	84	16	to	to	PART
ajst-29358	84	17	improve	improve	VERB
ajst-29358	84	18	its	its	PRON
ajst-29358	84	19	performance	performance	NOUN
ajst-29358	84	20	in	in	ADP
ajst-29358	84	21	practical	practical	ADJ
ajst-29358	84	22	applications	application	NOUN
ajst-29358	84	23	.	.	PUNCT
ajst-29358	85	1	figure	figure	NOUN
ajst-29358	85	2	4	4	NUM
ajst-29358	85	3	.	.	PUNCT
ajst-29358	86	1	dataset	dataset	VERB
ajst-29358	86	2	annotation	annotation	NOUN
ajst-29358	86	3	file	file	NOUN
ajst-29358	86	4	statistics	statistic	NOUN
ajst-29358	86	5	3.2	3.2	NUM
ajst-29358	86	6	.	.	PUNCT
ajst-29358	87	1	experimental	experimental	ADJ
ajst-29358	87	2	environment	environment	NOUN
ajst-29358	87	3	and	and	CCONJ
ajst-29358	87	4	parameter	parameter	NOUN
ajst-29358	87	5	configuration	configuration	NOUN
ajst-29358	87	6	table	table	NOUN
ajst-29358	87	7	2	2	NUM
ajst-29358	87	8	.	.	PUNCT
ajst-29358	88	1	yolov8	yolov8	NOUN
ajst-29358	88	2	model	model	PROPN
ajst-29358	88	3	training	training	NOUN
ajst-29358	88	4	environment	environment	NOUN
ajst-29358	88	5	configuration	configuration	NOUN
ajst-29358	88	6	parameter	parameter	PROPN
ajst-29358	88	7	cpu	cpu	PROPN
ajst-29358	88	8	intel(r	intel(r	PROPN
ajst-29358	88	9	)	)	PUNCT
ajst-29358	88	10	core(tm	core(tm	NOUN
ajst-29358	88	11	)	)	PUNCT
ajst-29358	88	12	i7	i7	NOUN
ajst-29358	88	13	-	-	PUNCT
ajst-29358	88	14	12700	12700	NUM
ajst-29358	88	15	memory	memory	NOUN
ajst-29358	88	16	32	32	NUM
ajst-29358	88	17	g	g	NOUN
ajst-29358	88	18	gpu	gpu	NOUN
ajst-29358	88	19	video	video	NOUN
ajst-29358	88	20	memory	memory	NOUN
ajst-29358	88	21	training	training	NOUN
ajst-29358	88	22	environment	environment	NOUN
ajst-29358	88	23	operating	operating	NOUN
ajst-29358	88	24	system	system	NOUN
ajst-29358	88	25	development	development	NOUN
ajst-29358	88	26	environment	environment	PROPN
ajst-29358	88	27	nvidia	nvidia	PROPN
ajst-29358	88	28	geforce	geforce	PROPN
ajst-29358	88	29	rtx4070	rtx4070	VERB
ajst-29358	88	30	12	12	NUM
ajst-29358	88	31	g	g	NOUN
ajst-29358	88	32	cuda	cuda	NOUN
ajst-29358	88	33	12.4	12.4	NUM
ajst-29358	88	34	windows10	windows10	NOUN
ajst-29358	88	35	(	(	PUNCT
ajst-29358	88	36	64	64	NUM
ajst-29358	88	37	-	-	PUNCT
ajst-29358	88	38	bit	bit	NOUN
ajst-29358	88	39	)	)	PUNCT
ajst-29358	88	40	python3.11	python3.11	VERB
ajst-29358	88	41	pytorch2.3	pytorch2.3	VERB
ajst-29358	88	42	parameter	parameter	NOUN
ajst-29358	88	43	settings	setting	NOUN
ajst-29358	88	44	:	:	PUNCT
ajst-29358	88	45	the	the	DET
ajst-29358	88	46	training	training	NOUN
ajst-29358	88	47	period	period	NOUN
ajst-29358	88	48	is	be	AUX
ajst-29358	88	49	set	set	VERB
ajst-29358	88	50	to	to	ADP
ajst-29358	88	51	150	150	NUM
ajst-29358	88	52	epochs	epoch	NOUN
ajst-29358	88	53	,	,	PUNCT
ajst-29358	88	54	under	under	ADP
ajst-29358	88	55	which	which	PRON
ajst-29358	88	56	the	the	DET
ajst-29358	88	57	model	model	NOUN
ajst-29358	88	58	can	can	AUX
ajst-29358	88	59	reach	reach	VERB
ajst-29358	88	60	a	a	DET
ajst-29358	88	61	converged	converged	ADJ
ajst-29358	88	62	state	state	NOUN
ajst-29358	88	63	;	;	PUNCT
ajst-29358	88	64	the	the	DET
ajst-29358	88	65	learning	learning	NOUN
ajst-29358	88	66	rate	rate	NOUN
ajst-29358	88	67	is	be	AUX
ajst-29358	88	68	set	set	VERB
ajst-29358	88	69	to	to	ADP
ajst-29358	88	70	0.01	0.01	NUM
ajst-29358	88	71	,	,	PUNCT
ajst-29358	88	72	and	and	CCONJ
ajst-29358	88	73	the	the	DET
ajst-29358	88	74	number	number	NOUN
ajst-29358	88	75	of	of	ADP
ajst-29358	88	76	images	image	NOUN
ajst-29358	88	77	processed	process	VERB
ajst-29358	88	78	per	per	ADP
ajst-29358	88	79	batch	batch	NOUN
ajst-29358	88	80	is	be	AUX
ajst-29358	88	81	16	16	NUM
ajst-29358	88	82	,	,	PUNCT
ajst-29358	88	83	ensuring	ensure	VERB
ajst-29358	88	84	that	that	SCONJ
ajst-29358	88	85	the	the	DET
ajst-29358	88	86	gpu	gpu	NOUN
ajst-29358	88	87	can	can	AUX
ajst-29358	88	88	operate	operate	VERB
ajst-29358	88	89	at	at	ADP
ajst-29358	88	90	full	full	ADJ
ajst-29358	88	91	power	power	NOUN
ajst-29358	88	92	;	;	PUNCT
ajst-29358	88	93	the	the	DET
ajst-29358	88	94	input	input	NOUN
ajst-29358	88	95	image	image	NOUN
ajst-29358	88	96	size	size	NOUN
ajst-29358	88	97	is	be	AUX
ajst-29358	88	98	640×640	640×640	NUM
ajst-29358	88	99	pixels	pixel	NOUN
ajst-29358	88	100	,	,	PUNCT
ajst-29358	88	101	and	and	CCONJ
ajst-29358	88	102	the	the	DET
ajst-29358	88	103	images	image	NOUN
ajst-29358	88	104	are	be	AUX
ajst-29358	88	105	scaled	scale	VERB
ajst-29358	88	106	to	to	ADP
ajst-29358	88	107	a	a	DET
ajst-29358	88	108	uniform	uniform	ADJ
ajst-29358	88	109	size	size	NOUN
ajst-29358	88	110	to	to	PART
ajst-29358	88	111	improve	improve	VERB
ajst-29358	88	112	processing	processing	NOUN
ajst-29358	88	113	speed	speed	NOUN
ajst-29358	88	114	.	.	PUNCT
ajst-29358	89	1	3.3	3.3	NUM
ajst-29358	89	2	.	.	PUNCT
ajst-29358	89	3	evaluation	evaluation	NOUN
ajst-29358	89	4	metrics	metric	NOUN
ajst-29358	89	5	in	in	ADP
ajst-29358	89	6	this	this	DET
ajst-29358	89	7	paper	paper	NOUN
ajst-29358	89	8	,	,	PUNCT
ajst-29358	89	9	the	the	DET
ajst-29358	89	10	following	follow	VERB
ajst-29358	89	11	evaluation	evaluation	NOUN
ajst-29358	89	12	metrics	metric	NOUN
ajst-29358	89	13	are	be	AUX
ajst-29358	89	14	mainly	mainly	ADV
ajst-29358	89	15	used	use	VERB
ajst-29358	89	16	to	to	PART
ajst-29358	89	17	measure	measure	VERB
ajst-29358	89	18	the	the	DET
ajst-29358	89	19	performance	performance	NOUN
ajst-29358	89	20	of	of	ADP
ajst-29358	89	21	the	the	DET
ajst-29358	89	22	model	model	NOUN
ajst-29358	89	23	:	:	PUNCT
ajst-29358	89	24	precision	precision	NOUN
ajst-29358	89	25	(	(	PUNCT
ajst-29358	89	26	p	p	NOUN
ajst-29358	89	27	)	)	PUNCT
ajst-29358	89	28	,	,	PUNCT
ajst-29358	89	29	recall	recall	NOUN
ajst-29358	89	30	rate	rate	NOUN
ajst-29358	89	31	(	(	PUNCT
ajst-29358	89	32	r	r	NOUN
ajst-29358	89	33	)	)	PUNCT
ajst-29358	89	34	,	,	PUNCT
ajst-29358	89	35	average	average	ADJ
ajst-29358	89	36	precision	precision	NOUN
ajst-29358	89	37	(	(	PUNCT
ajst-29358	89	38	map	map	NOUN
ajst-29358	89	39	)	)	PUNCT
ajst-29358	89	40	,	,	PUNCT
ajst-29358	89	41	f1	f1	NOUN
ajst-29358	89	42	score	score	NOUN
ajst-29358	89	43	,	,	PUNCT
ajst-29358	89	44	model	model	NOUN
ajst-29358	89	45	parameter	parameter	NOUN
ajst-29358	89	46	quantity	quantity	NOUN
ajst-29358	89	47	(	(	PUNCT
ajst-29358	89	48	params	param	NOUN
ajst-29358	89	49	)	)	PUNCT
ajst-29358	89	50	,	,	PUNCT
ajst-29358	89	51	total	total	ADJ
ajst-29358	89	52	floating	float	VERB
ajst-29358	89	53	-	-	PUNCT
ajst-29358	89	54	point	point	NOUN
ajst-29358	89	55	operations	operation	NOUN
ajst-29358	89	56	(	(	PUNCT
ajst-29358	89	57	flops	flop	NOUN
ajst-29358	89	58	)	)	PUNCT
ajst-29358	89	59	.	.	PUNCT
ajst-29358	90	1	these	these	DET
ajst-29358	90	2	metrics	metric	NOUN
ajst-29358	90	3	together	together	ADV
ajst-29358	90	4	form	form	VERB
ajst-29358	90	5	a	a	DET
ajst-29358	90	6	comprehensive	comprehensive	ADJ
ajst-29358	90	7	evaluation	evaluation	NOUN
ajst-29358	90	8	system	system	NOUN
ajst-29358	90	9	,	,	PUNCT
ajst-29358	90	10	and	and	CCONJ
ajst-29358	90	11	the	the	DET
ajst-29358	90	12	related	relate	VERB
ajst-29358	90	13	formulas	formula	NOUN
ajst-29358	90	14	are	be	AUX
ajst-29358	90	15	as	as	SCONJ
ajst-29358	90	16	follows	follow	VERB
ajst-29358	90	17	:	:	PUNCT
ajst-29358	90	18	tp	tp	PART
ajst-29358	90	19	p=	p=	PROPN
ajst-29358	90	20	tp+fp	tp+fp	PROPN
ajst-29358	90	21	(	(	PUNCT
ajst-29358	90	22	3	3	NUM
ajst-29358	90	23	)	)	PUNCT
ajst-29358	90	24	tp	tp	X
ajst-29358	90	25	r=	r=	ADJ
ajst-29358	90	26	tp+fn	tp+fn	NOUN
ajst-29358	90	27	(	(	PUNCT
ajst-29358	90	28	4	4	NUM
ajst-29358	90	29	)	)	SYM
ajst-29358	90	30	1	1	NUM
ajst-29358	90	31	0	0	NUM
ajst-29358	90	32	ap=	ap=	NOUN
ajst-29358	90	33	(	(	PUNCT
ajst-29358	90	34	)	)	PUNCT
ajst-29358	90	35	p	p	NOUN
ajst-29358	90	36	r	r	NOUN
ajst-29358	90	37	dr	dr	PROPN
ajst-29358	90	38	(	(	PUNCT
ajst-29358	90	39	5	5	NUM
ajst-29358	90	40	)	)	PUNCT
ajst-29358	90	41	0map=	0map=	PUNCT
ajst-29358	91	1	n	n	CCONJ
ajst-29358	91	2	i	i	PRON
ajst-29358	92	1	i	i	PRON
ajst-29358	92	2	ap	ap	VERB
ajst-29358	93	1	n	n	ADV
ajst-29358	93	2			PROPN
ajst-29358	93	3			X
ajst-29358	93	4	(	(	PUNCT
ajst-29358	93	5	6	6	NUM
ajst-29358	93	6	)	)	PUNCT
ajst-29358	93	7	155	155	NUM
ajst-29358	93	8	2pr	2pr	NUM
ajst-29358	93	9	f1=	f1=	PROPN
ajst-29358	93	10	p+r	p+r	NUM
ajst-29358	93	11	(	(	PUNCT
ajst-29358	93	12	7	7	NUM
ajst-29358	93	13	)	)	PUNCT
ajst-29358	93	14	precision	precision	NOUN
ajst-29358	93	15	and	and	CCONJ
ajst-29358	93	16	recall	recall	NOUN
ajst-29358	93	17	rate	rate	NOUN
ajst-29358	93	18	can	can	AUX
ajst-29358	93	19	be	be	AUX
ajst-29358	93	20	measured	measure	VERB
ajst-29358	93	21	using	use	VERB
ajst-29358	93	22	true	true	ADJ
ajst-29358	93	23	positives	positive	NOUN
ajst-29358	93	24	(	(	PUNCT
ajst-29358	93	25	tp	tp	NOUN
ajst-29358	93	26	)	)	PUNCT
ajst-29358	93	27	,	,	PUNCT
ajst-29358	93	28	true	true	ADJ
ajst-29358	93	29	negatives	negative	NOUN
ajst-29358	93	30	(	(	PUNCT
ajst-29358	93	31	tn	tn	NOUN
ajst-29358	93	32	)	)	PUNCT
ajst-29358	93	33	,	,	PUNCT
ajst-29358	93	34	false	false	ADJ
ajst-29358	93	35	positives	positive	NOUN
ajst-29358	93	36	(	(	PUNCT
ajst-29358	93	37	fp	fp	NOUN
ajst-29358	93	38	)	)	PUNCT
ajst-29358	93	39	,	,	PUNCT
ajst-29358	93	40	and	and	CCONJ
ajst-29358	93	41	false	false	ADJ
ajst-29358	93	42	negatives	negative	NOUN
ajst-29358	93	43	(	(	PUNCT
ajst-29358	93	44	fn	fn	NOUN
ajst-29358	93	45	)	)	PUNCT
ajst-29358	93	46	.	.	PUNCT
ajst-29358	94	1	ap	ap	PROPN
ajst-29358	94	2	represents	represent	VERB
ajst-29358	94	3	the	the	DET
ajst-29358	94	4	average	average	ADJ
ajst-29358	94	5	precision	precision	NOUN
ajst-29358	94	6	on	on	ADP
ajst-29358	94	7	each	each	DET
ajst-29358	94	8	category	category	NOUN
ajst-29358	94	9	,	,	PUNCT
ajst-29358	94	10	which	which	PRON
ajst-29358	94	11	can	can	AUX
ajst-29358	94	12	measure	measure	VERB
ajst-29358	94	13	the	the	DET
ajst-29358	94	14	model	model	NOUN
ajst-29358	94	15	's	's	PART
ajst-29358	94	16	performance	performance	NOUN
ajst-29358	94	17	on	on	ADP
ajst-29358	94	18	different	different	ADJ
ajst-29358	94	19	categories	category	NOUN
ajst-29358	94	20	,	,	PUNCT
ajst-29358	94	21	and	and	CCONJ
ajst-29358	94	22	map	map	NOUN
ajst-29358	94	23	is	be	AUX
ajst-29358	94	24	the	the	DET
ajst-29358	94	25	average	average	ADJ
ajst-29358	94	26	value	value	NOUN
ajst-29358	94	27	of	of	ADP
ajst-29358	94	28	all	all	DET
ajst-29358	94	29	aps	ap	NOUN
ajst-29358	94	30	,	,	PUNCT
ajst-29358	94	31	which	which	PRON
ajst-29358	94	32	can	can	AUX
ajst-29358	94	33	measure	measure	VERB
ajst-29358	94	34	the	the	DET
ajst-29358	94	35	model	model	NOUN
ajst-29358	94	36	's	's	PART
ajst-29358	94	37	performance	performance	NOUN
ajst-29358	94	38	on	on	ADP
ajst-29358	94	39	all	all	DET
ajst-29358	94	40	categories	category	NOUN
ajst-29358	94	41	.	.	PUNCT
ajst-29358	95	1	the	the	DET
ajst-29358	95	2	definition	definition	NOUN
ajst-29358	95	3	of	of	ADP
ajst-29358	95	4	map	map	NOUN
ajst-29358	95	5	is	be	AUX
ajst-29358	95	6	shown	show	VERB
ajst-29358	95	7	in	in	ADP
ajst-29358	95	8	formula	formula	NOUN
ajst-29358	95	9	(	(	PUNCT
ajst-29358	95	10	6	6	NUM
ajst-29358	95	11	)	)	PUNCT
ajst-29358	95	12	.	.	PUNCT
ajst-29358	96	1	map50	map50	PROPN
ajst-29358	96	2	is	be	AUX
ajst-29358	96	3	the	the	DET
ajst-29358	96	4	total	total	ADJ
ajst-29358	96	5	category	category	NOUN
ajst-29358	96	6	average	average	NOUN
ajst-29358	96	7	precision	precision	NOUN
ajst-29358	96	8	when	when	SCONJ
ajst-29358	96	9	the	the	DET
ajst-29358	96	10	iou	iou	NOUN
ajst-29358	96	11	threshold	threshold	NOUN
ajst-29358	96	12	is	be	AUX
ajst-29358	96	13	set	set	VERB
ajst-29358	96	14	to	to	ADP
ajst-29358	96	15	0.5	0.5	NUM
ajst-29358	96	16	.	.	PUNCT
ajst-29358	97	1	4	4	NUM
ajst-29358	97	2	.	.	X
ajst-29358	97	3	experimental	experimental	ADJ
ajst-29358	97	4	results	result	NOUN
ajst-29358	97	5	and	and	CCONJ
ajst-29358	97	6	analysis	analysis	NOUN
ajst-29358	97	7	4.1	4.1	NUM
ajst-29358	97	8	.	.	PUNCT
ajst-29358	98	1	ablation	ablation	NOUN
ajst-29358	98	2	study	study	NOUN
ajst-29358	98	3	to	to	PART
ajst-29358	98	4	ensure	ensure	VERB
ajst-29358	98	5	the	the	DET
ajst-29358	98	6	rationality	rationality	NOUN
ajst-29358	98	7	of	of	ADP
ajst-29358	98	8	the	the	DET
ajst-29358	98	9	ablation	ablation	NOUN
ajst-29358	98	10	study	study	NOUN
ajst-29358	98	11	comparison	comparison	NOUN
ajst-29358	98	12	,	,	PUNCT
ajst-29358	98	13	all	all	DET
ajst-29358	98	14	participating	participate	VERB
ajst-29358	98	15	models	model	NOUN
ajst-29358	98	16	are	be	AUX
ajst-29358	98	17	trained	train	VERB
ajst-29358	98	18	based	base	VERB
ajst-29358	98	19	on	on	ADP
ajst-29358	98	20	the	the	DET
ajst-29358	98	21	same	same	ADJ
ajst-29358	98	22	dataset	dataset	NOUN
ajst-29358	98	23	and	and	CCONJ
ajst-29358	98	24	parameter	parameter	NOUN
ajst-29358	98	25	environment	environment	NOUN
ajst-29358	98	26	.	.	PUNCT
ajst-29358	99	1	in	in	ADP
ajst-29358	99	2	this	this	DET
ajst-29358	99	3	study	study	NOUN
ajst-29358	99	4	,	,	PUNCT
ajst-29358	99	5	yolov8n	yolov8n	PROPN
ajst-29358	99	6	is	be	AUX
ajst-29358	99	7	used	use	VERB
ajst-29358	99	8	as	as	ADP
ajst-29358	99	9	the	the	DET
ajst-29358	99	10	benchmark	benchmark	ADJ
ajst-29358	99	11	algorithm	algorithm	NOUN
ajst-29358	99	12	to	to	PART
ajst-29358	99	13	implement	implement	VERB
ajst-29358	99	14	a	a	DET
ajst-29358	99	15	series	series	NOUN
ajst-29358	99	16	of	of	ADP
ajst-29358	99	17	improvement	improvement	NOUN
ajst-29358	99	18	measures	measure	NOUN
ajst-29358	99	19	,	,	PUNCT
ajst-29358	99	20	and	and	CCONJ
ajst-29358	99	21	the	the	DET
ajst-29358	99	22	corresponding	corresponding	ADJ
ajst-29358	99	23	ablation	ablation	NOUN
ajst-29358	99	24	experiments	experiment	NOUN
ajst-29358	99	25	are	be	AUX
ajst-29358	99	26	completed	complete	VERB
ajst-29358	99	27	.	.	PUNCT
ajst-29358	100	1	the	the	DET
ajst-29358	100	2	results	result	NOUN
ajst-29358	100	3	of	of	ADP
ajst-29358	100	4	the	the	DET
ajst-29358	100	5	experiments	experiment	NOUN
ajst-29358	100	6	are	be	AUX
ajst-29358	100	7	shown	show	VERB
ajst-29358	100	8	in	in	ADP
ajst-29358	100	9	table	table	NOUN
ajst-29358	100	10	1	1	NUM
ajst-29358	100	11	.	.	PUNCT
ajst-29358	100	12	table	table	NOUN
ajst-29358	100	13	3	3	NUM
ajst-29358	100	14	.	.	PUNCT
ajst-29358	100	15	ablation	ablation	NOUN
ajst-29358	100	16	study	study	NOUN
ajst-29358	100	17	results	result	VERB
ajst-29358	100	18	experiment	experiment	NOUN
ajst-29358	100	19	improvement	improvement	NOUN
ajst-29358	100	20	backbone	backbone	NOUN
ajst-29358	100	21	improvement	improvement	NOUN
ajst-29358	100	22	neck	neck	NOUN
ajst-29358	100	23	improvement	improvement	NOUN
ajst-29358	100	24	loss	loss	NOUN
ajst-29358	100	25	p/%	p/%	PROPN
ajst-29358	100	26	r/%	r/%	PROPN
ajst-29358	100	27	map	map	NOUN
ajst-29358	100	28	50/%	50/%	NUM
ajst-29358	100	29	params	param	NOUN
ajst-29358	100	30	/m	/m	PUNCT
ajst-29358	101	1	flop	flop	VERB
ajst-29358	101	2	s	s	PART
ajst-29358	101	3	/	/	SYM
ajst-29358	101	4	g	g	NOUN
ajst-29358	101	5	f1/%	f1/%	NOUN
ajst-29358	101	6	1	1	NUM
ajst-29358	101	7	×	×	NOUN
ajst-29358	101	8	×	×	NOUN
ajst-29358	101	9	×	×	NOUN
ajst-29358	101	10	80.1	80.1	NUM
ajst-29358	101	11	73.4	73.4	NUM
ajst-29358	101	12	77.2	77.2	NUM
ajst-29358	101	13	9.31	9.31	NUM
ajst-29358	101	14	20.4	20.4	NUM
ajst-29358	101	15	76.6	76.6	NUM
ajst-29358	101	16	2	2	NUM
ajst-29358	101	17	√	√	NUM
ajst-29358	101	18	×	×	NOUN
ajst-29358	101	19	×	×	NOUN
ajst-29358	101	20	81.1	81.1	NUM
ajst-29358	101	21	74.2	74.2	NUM
ajst-29358	101	22	78.1	78.1	NUM
ajst-29358	101	23	10.35	10.35	NUM
ajst-29358	101	24	21.5	21.5	NUM
ajst-29358	101	25	77.5	77.5	NUM
ajst-29358	101	26	3	3	NUM
ajst-29358	101	27	4	4	NUM
ajst-29358	101	28	5	5	NUM
ajst-29358	101	29	6	6	NUM
ajst-29358	101	30	7	7	NUM
ajst-29358	101	31	8	8	NUM
ajst-29358	101	32	×	×	NOUN
ajst-29358	101	33	×	×	NOUN
ajst-29358	101	34	√	√	NUM
ajst-29358	101	35	√	√	NUM
ajst-29358	101	36	×	×	NOUN
ajst-29358	101	37	√	√	NUM
ajst-29358	101	38	√	√	NUM
ajst-29358	101	39	×	×	NOUN
ajst-29358	101	40	√	√	NUM
ajst-29358	101	41	×	×	NOUN
ajst-29358	101	42	√	√	NUM
ajst-29358	101	43	√	√	NUM
ajst-29358	101	44	×	×	NOUN
ajst-29358	101	45	√	√	NUM
ajst-29358	101	46	×	×	NOUN
ajst-29358	101	47	√	√	NOUN
ajst-29358	101	48	√	√	NUM
ajst-29358	101	49	√	√	NUM
ajst-29358	101	50	80.1	80.1	NUM
ajst-29358	101	51	79.8	79.8	NUM
ajst-29358	101	52	81.4	81.4	NUM
ajst-29358	101	53	80.3	80.3	NUM
ajst-29358	101	54	81.7	81.7	NUM
ajst-29358	101	55	82.8	82.8	NUM
ajst-29358	101	56	75.4	75.4	NUM
ajst-29358	101	57	73.8	73.8	NUM
ajst-29358	101	58	74.7	74.7	NUM
ajst-29358	101	59	73.6	73.6	NUM
ajst-29358	101	60	76.9	76.9	NUM
ajst-29358	101	61	78.1	78.1	NUM
ajst-29358	101	62	77.7	77.7	NUM
ajst-29358	101	63	77.4	77.4	NUM
ajst-29358	101	64	78.8	78.8	NUM
ajst-29358	101	65	77.5	77.5	NUM
ajst-29358	101	66	79.2	79.2	NUM
ajst-29358	101	67	80.3	80.3	NUM
ajst-29358	101	68	7.51	7.51	NUM
ajst-29358	101	69	11.13	11.13	NUM
ajst-29358	101	70	6.72	6.72	NUM
ajst-29358	101	71	10.72	10.72	NUM
ajst-29358	101	72	7.48	7.48	NUM
ajst-29358	101	73	6.53	6.53	NUM
ajst-29358	101	74	16.7	16.7	NUM
ajst-29358	101	75	21.7	21.7	NUM
ajst-29358	101	76	16.3	16.3	NUM
ajst-29358	101	77	21.5	21.5	NUM
ajst-29358	101	78	16.7	16.7	NUM
ajst-29358	101	79	16.1	16.1	NUM
ajst-29358	101	80	77.8	77.8	NUM
ajst-29358	101	81	76.7	76.7	NUM
ajst-29358	101	82	77.9	77.9	NUM
ajst-29358	101	83	76.8	76.8	NUM
ajst-29358	101	84	79.2	79.2	NUM
ajst-29358	101	85	80.4	80.4	NUM
ajst-29358	101	86	in	in	ADP
ajst-29358	101	87	-	-	PUNCT
ajst-29358	101	88	depth	depth	NOUN
ajst-29358	101	89	analysis	analysis	NOUN
ajst-29358	101	90	of	of	ADP
ajst-29358	101	91	the	the	DET
ajst-29358	101	92	experimental	experimental	ADJ
ajst-29358	101	93	results	result	NOUN
ajst-29358	101	94	:	:	PUNCT
ajst-29358	101	95	experiment	experiment	NOUN
ajst-29358	101	96	2	2	NUM
ajst-29358	101	97	integrated	integrate	VERB
ajst-29358	101	98	the	the	DET
ajst-29358	101	99	ema	ema	PROPN
ajst-29358	101	100	attention	attention	NOUN
ajst-29358	101	101	mechanism	mechanism	NOUN
ajst-29358	101	102	into	into	ADP
ajst-29358	101	103	the	the	DET
ajst-29358	101	104	backbone	backbone	NOUN
ajst-29358	101	105	network	network	NOUN
ajst-29358	101	106	,	,	PUNCT
ajst-29358	101	107	and	and	CCONJ
ajst-29358	101	108	most	most	ADJ
ajst-29358	101	109	performance	performance	NOUN
ajst-29358	101	110	indicators	indicator	NOUN
ajst-29358	101	111	have	have	AUX
ajst-29358	101	112	improved	improve	VERB
ajst-29358	101	113	,	,	PUNCT
ajst-29358	101	114	indicating	indicate	VERB
ajst-29358	101	115	that	that	SCONJ
ajst-29358	101	116	the	the	DET
ajst-29358	101	117	optimization	optimization	NOUN
ajst-29358	101	118	of	of	ADP
ajst-29358	101	119	the	the	DET
ajst-29358	101	120	backbone	backbone	NOUN
ajst-29358	101	121	network	network	NOUN
ajst-29358	101	122	has	have	AUX
ajst-29358	101	123	enhanced	enhance	VERB
ajst-29358	101	124	the	the	DET
ajst-29358	101	125	model	model	NOUN
ajst-29358	101	126	's	's	PART
ajst-29358	101	127	feature	feature	NOUN
ajst-29358	101	128	extraction	extraction	NOUN
ajst-29358	101	129	capability	capability	NOUN
ajst-29358	101	130	and	and	CCONJ
ajst-29358	101	131	improved	improve	VERB
ajst-29358	101	132	detection	detection	NOUN
ajst-29358	101	133	accuracy	accuracy	NOUN
ajst-29358	101	134	.	.	PUNCT
ajst-29358	102	1	experiment	experiment	NOUN
ajst-29358	102	2	3	3	NUM
ajst-29358	102	3	integrated	integrate	VERB
ajst-29358	102	4	the	the	DET
ajst-29358	102	5	yolov8	yolov8	NOUN
ajst-29358	102	6	-	-	PUNCT
ajst-29358	102	7	bifpn	bifpn	PROPN
ajst-29358	102	8	structure	structure	NOUN
ajst-29358	102	9	into	into	ADP
ajst-29358	102	10	the	the	DET
ajst-29358	102	11	neck	neck	NOUN
ajst-29358	102	12	of	of	ADP
ajst-29358	102	13	yolov8n	yolov8n	NOUN
ajst-29358	102	14	,	,	PUNCT
ajst-29358	102	15	and	and	CCONJ
ajst-29358	102	16	compared	compare	VERB
ajst-29358	102	17	with	with	ADP
ajst-29358	102	18	the	the	DET
ajst-29358	102	19	original	original	ADJ
ajst-29358	102	20	algorithm	algorithm	NOUN
ajst-29358	102	21	,	,	PUNCT
ajst-29358	102	22	all	all	DET
ajst-29358	102	23	performance	performance	NOUN
ajst-29358	102	24	indicators	indicator	NOUN
ajst-29358	102	25	have	have	AUX
ajst-29358	102	26	significantly	significantly	ADV
ajst-29358	102	27	improved	improve	VERB
ajst-29358	102	28	except	except	SCONJ
ajst-29358	102	29	for	for	ADP
ajst-29358	102	30	precision	precision	NOUN
ajst-29358	102	31	,	,	PUNCT
ajst-29358	102	32	ensuring	ensure	VERB
ajst-29358	102	33	that	that	SCONJ
ajst-29358	102	34	the	the	DET
ajst-29358	102	35	detection	detection	NOUN
ajst-29358	102	36	accuracy	accuracy	NOUN
ajst-29358	102	37	is	be	AUX
ajst-29358	102	38	improved	improve	VERB
ajst-29358	102	39	while	while	SCONJ
ajst-29358	102	40	the	the	DET
ajst-29358	102	41	model	model	NOUN
ajst-29358	102	42	is	be	AUX
ajst-29358	102	43	simplified	simplify	VERB
ajst-29358	102	44	.	.	PUNCT
ajst-29358	103	1	considering	consider	VERB
ajst-29358	103	2	that	that	DET
ajst-29358	103	3	precision	precision	NOUN
ajst-29358	103	4	is	be	AUX
ajst-29358	103	5	not	not	PART
ajst-29358	103	6	a	a	DET
ajst-29358	103	7	core	core	NOUN
ajst-29358	103	8	indicator	indicator	NOUN
ajst-29358	103	9	of	of	ADP
ajst-29358	103	10	model	model	NOUN
ajst-29358	103	11	evaluation	evaluation	NOUN
ajst-29358	103	12	,	,	PUNCT
ajst-29358	103	13	it	it	PRON
ajst-29358	103	14	can	can	AUX
ajst-29358	103	15	be	be	AUX
ajst-29358	103	16	considered	consider	VERB
ajst-29358	103	17	that	that	SCONJ
ajst-29358	103	18	the	the	DET
ajst-29358	103	19	improvement	improvement	NOUN
ajst-29358	103	20	of	of	ADP
ajst-29358	103	21	the	the	DET
ajst-29358	103	22	network	network	NOUN
ajst-29358	103	23	neck	neck	NOUN
ajst-29358	103	24	is	be	AUX
ajst-29358	103	25	very	very	ADV
ajst-29358	103	26	successful	successful	ADJ
ajst-29358	103	27	.	.	PUNCT
ajst-29358	104	1	finally	finally	ADV
ajst-29358	104	2	,	,	PUNCT
ajst-29358	104	3	experiment	experiment	NOUN
ajst-29358	104	4	4	4	NUM
ajst-29358	104	5	used	use	VERB
ajst-29358	104	6	wiou	wiou	NOUN
ajst-29358	104	7	as	as	ADP
ajst-29358	104	8	a	a	DET
ajst-29358	104	9	new	new	ADJ
ajst-29358	104	10	method	method	NOUN
ajst-29358	104	11	for	for	ADP
ajst-29358	104	12	bounding	bound	VERB
ajst-29358	104	13	box	box	NOUN
ajst-29358	104	14	regression	regression	NOUN
ajst-29358	104	15	,	,	PUNCT
ajst-29358	104	16	and	and	CCONJ
ajst-29358	104	17	compared	compare	VERB
ajst-29358	104	18	with	with	ADP
ajst-29358	104	19	the	the	DET
ajst-29358	104	20	benchmark	benchmark	ADJ
ajst-29358	104	21	algorithm	algorithm	NOUN
ajst-29358	104	22	,	,	PUNCT
ajst-29358	104	23	it	it	PRON
ajst-29358	104	24	successfully	successfully	ADV
ajst-29358	104	25	improved	improve	VERB
ajst-29358	104	26	the	the	DET
ajst-29358	104	27	network	network	NOUN
ajst-29358	104	28	's	's	PART
ajst-29358	104	29	detection	detection	NOUN
ajst-29358	104	30	accuracy	accuracy	NOUN
ajst-29358	104	31	without	without	ADP
ajst-29358	104	32	increasing	increase	VERB
ajst-29358	104	33	the	the	DET
ajst-29358	104	34	model	model	NOUN
ajst-29358	104	35	's	's	PART
ajst-29358	104	36	complexity	complexity	NOUN
ajst-29358	104	37	.	.	PUNCT
ajst-29358	105	1	figure	figure	NOUN
ajst-29358	105	2	5	5	NUM
ajst-29358	105	3	.	.	PUNCT
ajst-29358	105	4	ablation	ablation	NOUN
ajst-29358	105	5	study	study	NOUN
ajst-29358	105	6	map50	map50	PROPN
ajst-29358	105	7	comparison	comparison	NOUN
ajst-29358	105	8	in	in	ADP
ajst-29358	105	9	the	the	DET
ajst-29358	105	10	yolov8n	yolov8n	NOUN
ajst-29358	105	11	and	and	CCONJ
ajst-29358	105	12	the	the	DET
ajst-29358	105	13	improved	improved	ADJ
ajst-29358	105	14	backbone	backbone	NOUN
ajst-29358	105	15	network	network	NOUN
ajst-29358	105	16	structures	structure	NOUN
ajst-29358	105	17	,	,	PUNCT
ajst-29358	105	18	after	after	ADP
ajst-29358	105	19	introducing	introduce	VERB
ajst-29358	105	20	the	the	DET
ajst-29358	105	21	wiou	wiou	NOUN
ajst-29358	105	22	loss	loss	NOUN
ajst-29358	105	23	function	function	NOUN
ajst-29358	105	24	,	,	PUNCT
ajst-29358	105	25	the	the	DET
ajst-29358	105	26	map50	map50	PROPN
ajst-29358	105	27	increased	increase	VERB
ajst-29358	105	28	by	by	ADP
ajst-29358	105	29	0.2	0.2	NUM
ajst-29358	105	30	%	%	NOUN
ajst-29358	105	31	and	and	CCONJ
ajst-29358	105	32	0.3	0.3	NUM
ajst-29358	105	33	%	%	NOUN
ajst-29358	105	34	respectively	respectively	ADV
ajst-29358	105	35	,	,	PUNCT
ajst-29358	105	36	the	the	DET
ajst-29358	105	37	recall	recall	NOUN
ajst-29358	105	38	rate	rate	NOUN
ajst-29358	105	39	increased	increase	VERB
ajst-29358	105	40	by	by	ADP
ajst-29358	105	41	0.4	0.4	NUM
ajst-29358	105	42	%	%	NOUN
ajst-29358	105	43	and	and	CCONJ
ajst-29358	105	44	0.2	0.2	NUM
ajst-29358	105	45	%	%	NOUN
ajst-29358	105	46	respectively	respectively	ADV
ajst-29358	105	47	,	,	PUNCT
ajst-29358	105	48	but	but	CCONJ
ajst-29358	105	49	the	the	DET
ajst-29358	105	50	precision	precision	NOUN
ajst-29358	105	51	decreased	decrease	VERB
ajst-29358	105	52	,	,	PUNCT
ajst-29358	105	53	indicating	indicate	VERB
ajst-29358	105	54	that	that	SCONJ
ajst-29358	105	55	the	the	DET
ajst-29358	105	56	improvement	improvement	NOUN
ajst-29358	105	57	effect	effect	NOUN
ajst-29358	105	58	is	be	AUX
ajst-29358	105	59	not	not	PART
ajst-29358	105	60	significant	significant	ADJ
ajst-29358	105	61	.	.	PUNCT
ajst-29358	106	1	however	however	ADV
ajst-29358	106	2	,	,	PUNCT
ajst-29358	106	3	when	when	SCONJ
ajst-29358	106	4	wiou	wiou	NOUN
ajst-29358	106	5	was	be	AUX
ajst-29358	106	6	applied	apply	VERB
ajst-29358	106	7	to	to	ADP
ajst-29358	106	8	the	the	DET
ajst-29358	106	9	improved	improve	VERB
ajst-29358	106	10	neck	neck	NOUN
ajst-29358	106	11	network	network	NOUN
ajst-29358	106	12	structure	structure	NOUN
ajst-29358	106	13	,	,	PUNCT
ajst-29358	106	14	the	the	DET
ajst-29358	106	15	map50	map50	PROPN
ajst-29358	106	16	significantly	significantly	ADV
ajst-29358	106	17	increased	increase	VERB
ajst-29358	106	18	by	by	ADP
ajst-29358	106	19	2	2	NUM
ajst-29358	106	20	%	%	NOUN
ajst-29358	106	21	,	,	PUNCT
ajst-29358	106	22	the	the	DET
ajst-29358	106	23	recall	recall	NOUN
ajst-29358	106	24	rate	rate	NOUN
ajst-29358	106	25	increased	increase	VERB
ajst-29358	106	26	by	by	ADP
ajst-29358	106	27	3.5	3.5	NUM
ajst-29358	106	28	%	%	NOUN
ajst-29358	106	29	,	,	PUNCT
ajst-29358	106	30	and	and	CCONJ
ajst-29358	106	31	the	the	DET
ajst-29358	106	32	precision	precision	NOUN
ajst-29358	106	33	also	also	ADV
ajst-29358	106	34	improved	improve	VERB
ajst-29358	106	35	,	,	PUNCT
ajst-29358	106	36	reaching	reach	VERB
ajst-29358	106	37	1.6	1.6	NUM
ajst-29358	106	38	%	%	NOUN
ajst-29358	106	39	.	.	PUNCT
ajst-29358	107	1	these	these	DET
ajst-29358	107	2	comparative	comparative	ADJ
ajst-29358	107	3	results	result	NOUN
ajst-29358	107	4	indicate	indicate	VERB
ajst-29358	107	5	that	that	SCONJ
ajst-29358	107	6	the	the	DET
ajst-29358	107	7	wiou	wiou	NOUN
ajst-29358	107	8	loss	loss	NOUN
ajst-29358	107	9	function	function	NOUN
ajst-29358	107	10	performed	perform	VERB
ajst-29358	107	11	better	well	ADV
ajst-29358	107	12	in	in	ADP
ajst-29358	107	13	the	the	DET
ajst-29358	107	14	network	network	NOUN
ajst-29358	107	15	optimized	optimize	VERB
ajst-29358	107	16	with	with	ADP
ajst-29358	107	17	the	the	DET
ajst-29358	107	18	neck	neck	NOUN
ajst-29358	107	19	structure	structure	NOUN
ajst-29358	107	20	,	,	PUNCT
ajst-29358	107	21	implying	imply	VERB
ajst-29358	107	22	that	that	SCONJ
ajst-29358	107	23	the	the	DET
ajst-29358	107	24	wiou	wiou	NOUN
ajst-29358	107	25	loss	loss	NOUN
ajst-29358	107	26	function	function	NOUN
ajst-29358	107	27	is	be	AUX
ajst-29358	107	28	more	more	ADV
ajst-29358	107	29	suitable	suitable	ADJ
ajst-29358	107	30	for	for	ADP
ajst-29358	107	31	networks	network	NOUN
ajst-29358	107	32	that	that	PRON
ajst-29358	107	33	have	have	AUX
ajst-29358	107	34	undergone	undergo	VERB
ajst-29358	107	35	multi	multi	ADJ
ajst-29358	107	36	-	-	ADJ
ajst-29358	107	37	scale	scale	ADJ
ajst-29358	107	38	feature	feature	NOUN
ajst-29358	107	39	fusion	fusion	NOUN
ajst-29358	107	40	.	.	PUNCT
ajst-29358	108	1	experiment	experiment	NOUN
ajst-29358	108	2	no	no	NOUN
ajst-29358	108	3	.	.	NOUN
ajst-29358	108	4	8	8	NUM
ajst-29358	108	5	shows	show	VERB
ajst-29358	108	6	that	that	SCONJ
ajst-29358	108	7	after	after	ADP
ajst-29358	108	8	integrating	integrate	VERB
ajst-29358	108	9	all	all	DET
ajst-29358	108	10	the	the	DET
ajst-29358	108	11	improvements	improvement	NOUN
ajst-29358	108	12	proposed	propose	VERB
ajst-29358	108	13	in	in	ADP
ajst-29358	108	14	this	this	DET
ajst-29358	108	15	paper	paper	NOUN
ajst-29358	108	16	into	into	ADP
ajst-29358	108	17	yolov8n	yolov8n	PROPN
ajst-29358	108	18	,	,	PUNCT
ajst-29358	108	19	compared	compare	VERB
ajst-29358	108	20	with	with	ADP
ajst-29358	108	21	the	the	DET
ajst-29358	108	22	original	original	ADJ
ajst-29358	108	23	algorithm	algorithm	NOUN
ajst-29358	108	24	,	,	PUNCT
ajst-29358	108	25	the	the	DET
ajst-29358	108	26	method	method	NOUN
ajst-29358	108	27	proposed	propose	VERB
ajst-29358	108	28	in	in	ADP
ajst-29358	108	29	this	this	DET
ajst-29358	108	30	paper	paper	NOUN
ajst-29358	108	31	increased	increase	VERB
ajst-29358	108	32	the	the	DET
ajst-29358	108	33	map50	map50	NOUN
ajst-29358	108	34	by	by	ADP
ajst-29358	108	35	3.1	3.1	NUM
ajst-29358	108	36	%	%	NOUN
ajst-29358	108	37	,	,	PUNCT
ajst-29358	108	38	precision	precision	NOUN
ajst-29358	108	39	by	by	ADP
ajst-29358	108	40	2.7	2.7	NUM
ajst-29358	108	41	%	%	NOUN
ajst-29358	108	42	,	,	PUNCT
ajst-29358	108	43	recall	recall	NOUN
ajst-29358	108	44	rate	rate	NOUN
ajst-29358	108	45	by	by	ADP
ajst-29358	108	46	4.7	4.7	NUM
ajst-29358	108	47	%	%	NOUN
ajst-29358	108	48	,	,	PUNCT
ajst-29358	108	49	reduced	reduce	VERB
ajst-29358	108	50	the	the	DET
ajst-29358	108	51	parameter	parameter	NOUN
ajst-29358	108	52	volume	volume	NOUN
ajst-29358	108	53	by	by	ADP
ajst-29358	108	54	2.78	2.78	NUM
ajst-29358	108	55	million	million	NUM
ajst-29358	108	56	,	,	PUNCT
ajst-29358	108	57	reduced	reduce	VERB
ajst-29358	108	58	the	the	DET
ajst-29358	108	59	total	total	ADJ
ajst-29358	108	60	floating	float	VERB
ajst-29358	108	61	-	-	PUNCT
ajst-29358	108	62	point	point	NOUN
ajst-29358	108	63	operations	operation	NOUN
ajst-29358	108	64	by	by	ADP
ajst-29358	108	65	4.3	4.3	NUM
ajst-29358	108	66	g	g	NOUN
ajst-29358	108	67	,	,	PUNCT
ajst-29358	108	68	and	and	CCONJ
ajst-29358	108	69	increased	increase	VERB
ajst-29358	108	70	the	the	DET
ajst-29358	108	71	f1	f1	NOUN
ajst-29358	108	72	score	score	NOUN
ajst-29358	108	73	by	by	ADP
ajst-29358	108	74	3.8	3.8	NUM
ajst-29358	108	75	%	%	NOUN
ajst-29358	108	76	.	.	PUNCT
ajst-29358	109	1	these	these	DET
ajst-29358	109	2	improvements	improvement	NOUN
ajst-29358	109	3	make	make	VERB
ajst-29358	109	4	the	the	DET
ajst-29358	109	5	algorithm	algorithm	NOUN
ajst-29358	109	6	156	156	NUM
ajst-29358	109	7	superior	superior	ADJ
ajst-29358	109	8	to	to	ADP
ajst-29358	109	9	yolov8n	yolov8n	NOUN
ajst-29358	109	10	in	in	ADP
ajst-29358	109	11	all	all	DET
ajst-29358	109	12	performance	performance	NOUN
ajst-29358	109	13	indicators	indicator	NOUN
ajst-29358	109	14	.	.	PUNCT
ajst-29358	110	1	as	as	SCONJ
ajst-29358	110	2	shown	show	VERB
ajst-29358	110	3	in	in	ADP
ajst-29358	110	4	figure	figure	NOUN
ajst-29358	110	5	5	5	NUM
ajst-29358	110	6	,	,	PUNCT
ajst-29358	110	7	which	which	PRON
ajst-29358	110	8	compares	compare	VERB
ajst-29358	110	9	the	the	DET
ajst-29358	110	10	changes	change	NOUN
ajst-29358	110	11	in	in	ADP
ajst-29358	110	12	map50	map50	PROPN
ajst-29358	110	13	,	,	PUNCT
ajst-29358	110	14	it	it	PRON
ajst-29358	110	15	not	not	PART
ajst-29358	110	16	only	only	ADV
ajst-29358	110	17	shows	show	VERB
ajst-29358	110	18	the	the	DET
ajst-29358	110	19	improvement	improvement	NOUN
ajst-29358	110	20	in	in	ADP
ajst-29358	110	21	detection	detection	NOUN
ajst-29358	110	22	accuracy	accuracy	NOUN
ajst-29358	110	23	but	but	CCONJ
ajst-29358	110	24	also	also	ADV
ajst-29358	110	25	the	the	DET
ajst-29358	110	26	simplification	simplification	NOUN
ajst-29358	110	27	of	of	ADP
ajst-29358	110	28	the	the	DET
ajst-29358	110	29	model	model	NOUN
ajst-29358	110	30	,	,	PUNCT
ajst-29358	110	31	meeting	meet	VERB
ajst-29358	110	32	the	the	DET
ajst-29358	110	33	needs	need	NOUN
ajst-29358	110	34	for	for	ADP
ajst-29358	110	35	real	real	ADJ
ajst-29358	110	36	-	-	PUNCT
ajst-29358	110	37	time	time	NOUN
ajst-29358	110	38	and	and	CCONJ
ajst-29358	110	39	accuracy	accuracy	NOUN
ajst-29358	110	40	.	.	PUNCT
ajst-29358	111	1	4.2	4.2	NUM
ajst-29358	111	2	.	.	PUNCT
ajst-29358	112	1	ema	ema	PROPN
ajst-29358	112	2	quantity	quantity	PROPN
ajst-29358	112	3	comparison	comparison	NOUN
ajst-29358	112	4	experiment	experiment	NOUN
ajst-29358	112	5	there	there	PRON
ajst-29358	112	6	are	be	VERB
ajst-29358	112	7	4	4	NUM
ajst-29358	112	8	c2f	c2f	NOUN
ajst-29358	112	9	modules	module	NOUN
ajst-29358	112	10	in	in	ADP
ajst-29358	112	11	the	the	DET
ajst-29358	112	12	yolov8	yolov8	NOUN
ajst-29358	112	13	backbone	backbone	NOUN
ajst-29358	112	14	network	network	NOUN
ajst-29358	112	15	.	.	PUNCT
ajst-29358	113	1	adding	add	VERB
ajst-29358	113	2	different	different	ADJ
ajst-29358	113	3	quantities	quantity	NOUN
ajst-29358	113	4	of	of	ADP
ajst-29358	113	5	ema	ema	PROPN
ajst-29358	113	6	will	will	AUX
ajst-29358	113	7	have	have	VERB
ajst-29358	113	8	different	different	ADJ
ajst-29358	113	9	effects	effect	NOUN
ajst-29358	113	10	on	on	ADP
ajst-29358	113	11	the	the	DET
ajst-29358	113	12	model	model	NOUN
ajst-29358	113	13	.	.	PUNCT
ajst-29358	114	1	this	this	DET
ajst-29358	114	2	experiment	experiment	NOUN
ajst-29358	114	3	compares	compare	VERB
ajst-29358	114	4	the	the	DET
ajst-29358	114	5	model	model	NOUN
ajst-29358	114	6	by	by	ADP
ajst-29358	114	7	adding	add	VERB
ajst-29358	114	8	different	different	ADJ
ajst-29358	114	9	quantities	quantity	NOUN
ajst-29358	114	10	of	of	ADP
ajst-29358	114	11	ema	ema	PROPN
ajst-29358	114	12	modules	module	NOUN
ajst-29358	114	13	after	after	ADP
ajst-29358	114	14	the	the	DET
ajst-29358	114	15	c2f	c2f	NOUN
ajst-29358	114	16	modules	module	NOUN
ajst-29358	114	17	in	in	ADP
ajst-29358	114	18	the	the	DET
ajst-29358	114	19	yolov8	yolov8	NOUN
ajst-29358	114	20	backbone	backbone	PROPN
ajst-29358	114	21	network	network	NOUN
ajst-29358	114	22	.	.	PUNCT
ajst-29358	115	1	the	the	DET
ajst-29358	115	2	results	result	NOUN
ajst-29358	115	3	are	be	AUX
ajst-29358	115	4	shown	show	VERB
ajst-29358	115	5	in	in	ADP
ajst-29358	115	6	table	table	NOUN
ajst-29358	115	7	2	2	NUM
ajst-29358	115	8	.	.	PUNCT
ajst-29358	115	9	table	table	NOUN
ajst-29358	115	10	4	4	NUM
ajst-29358	115	11	.	.	PUNCT
ajst-29358	116	1	ema	ema	PROPN
ajst-29358	116	2	quantity	quantity	PROPN
ajst-29358	116	3	comparison	comparison	NOUN
ajst-29358	116	4	experiment	experiment	NOUN
ajst-29358	116	5	results	result	NOUN
ajst-29358	116	6	method	method	VERB
ajst-29358	116	7	map50	map50	ADJ
ajst-29358	116	8	%	%	NOUN
ajst-29358	116	9	params	param	NOUN
ajst-29358	116	10	/	/	SYM
ajst-29358	116	11	m	m	NOUN
ajst-29358	116	12	flops	flop	NOUN
ajst-29358	116	13	/	/	SYM
ajst-29358	116	14	g	g	PROPN
ajst-29358	116	15	yolov8n	yolov8n	PROPN
ajst-29358	116	16	77.2	77.2	NUM
ajst-29358	116	17	9.31	9.31	NUM
ajst-29358	116	18	20.4	20.4	NUM
ajst-29358	116	19	1ema	1ema	NUM
ajst-29358	116	20	77.4	77.4	NUM
ajst-29358	116	21	9.34	9.34	NUM
ajst-29358	116	22	20.9	20.9	NUM
ajst-29358	116	23	2ema	2ema	NUM
ajst-29358	116	24	3ema	3ema	NUM
ajst-29358	116	25	4ema	4ema	NUM
ajst-29358	116	26	77.6	77.6	NUM
ajst-29358	116	27	77.8	77.8	NUM
ajst-29358	116	28	77.8	77.8	NUM
ajst-29358	116	29	9.36	9.36	NUM
ajst-29358	116	30	9.38	9.38	NUM
ajst-29358	116	31	9.41	9.41	NUM
ajst-29358	116	32	21.2	21.2	NUM
ajst-29358	116	33	21.5	21.5	NUM
ajst-29358	116	34	21.9	21.9	NUM
ajst-29358	116	35	from	from	ADP
ajst-29358	116	36	the	the	DET
ajst-29358	116	37	experimental	experimental	ADJ
ajst-29358	116	38	data	datum	NOUN
ajst-29358	116	39	in	in	ADP
ajst-29358	116	40	table	table	NOUN
ajst-29358	116	41	2	2	NUM
ajst-29358	116	42	,	,	PUNCT
ajst-29358	116	43	it	it	PRON
ajst-29358	116	44	is	be	AUX
ajst-29358	116	45	known	know	VERB
ajst-29358	116	46	that	that	SCONJ
ajst-29358	116	47	adding	add	VERB
ajst-29358	116	48	the	the	DET
ajst-29358	116	49	ema	ema	PROPN
ajst-29358	116	50	attention	attention	NOUN
ajst-29358	116	51	mechanism	mechanism	NOUN
ajst-29358	116	52	can	can	AUX
ajst-29358	116	53	improve	improve	VERB
ajst-29358	116	54	the	the	DET
ajst-29358	116	55	model	model	NOUN
ajst-29358	116	56	's	's	PART
ajst-29358	116	57	feature	feature	NOUN
ajst-29358	116	58	extraction	extraction	NOUN
ajst-29358	116	59	capability	capability	NOUN
ajst-29358	116	60	and	and	CCONJ
ajst-29358	116	61	thereby	thereby	ADV
ajst-29358	116	62	improve	improve	VERB
ajst-29358	116	63	detection	detection	NOUN
ajst-29358	116	64	accuracy	accuracy	NOUN
ajst-29358	116	65	,	,	PUNCT
ajst-29358	116	66	but	but	CCONJ
ajst-29358	116	67	at	at	ADP
ajst-29358	116	68	the	the	DET
ajst-29358	116	69	same	same	ADJ
ajst-29358	116	70	time	time	NOUN
ajst-29358	116	71	,	,	PUNCT
ajst-29358	116	72	it	it	PRON
ajst-29358	116	73	causes	cause	VERB
ajst-29358	116	74	an	an	DET
ajst-29358	116	75	increase	increase	NOUN
ajst-29358	116	76	in	in	ADP
ajst-29358	116	77	the	the	DET
ajst-29358	116	78	number	number	NOUN
ajst-29358	116	79	of	of	ADP
ajst-29358	116	80	parameters	parameter	NOUN
ajst-29358	116	81	,	,	PUNCT
ajst-29358	116	82	which	which	PRON
ajst-29358	116	83	will	will	AUX
ajst-29358	116	84	slow	slow	VERB
ajst-29358	116	85	down	down	ADP
ajst-29358	116	86	the	the	DET
ajst-29358	116	87	model	model	NOUN
ajst-29358	116	88	's	's	PART
ajst-29358	116	89	computation	computation	NOUN
ajst-29358	116	90	speed	speed	NOUN
ajst-29358	116	91	.	.	PUNCT
ajst-29358	117	1	therefore	therefore	ADV
ajst-29358	117	2	,	,	PUNCT
ajst-29358	117	3	as	as	ADP
ajst-29358	117	4	the	the	DET
ajst-29358	117	5	number	number	NOUN
ajst-29358	117	6	of	of	ADP
ajst-29358	117	7	ema	ema	PROPN
ajst-29358	117	8	introductions	introduction	NOUN
ajst-29358	117	9	increases	increase	NOUN
ajst-29358	117	10	,	,	PUNCT
ajst-29358	117	11	a	a	DET
ajst-29358	117	12	balance	balance	NOUN
ajst-29358	117	13	point	point	NOUN
ajst-29358	117	14	needs	need	VERB
ajst-29358	117	15	to	to	PART
ajst-29358	117	16	be	be	AUX
ajst-29358	117	17	sought	seek	VERB
ajst-29358	117	18	between	between	ADP
ajst-29358	117	19	the	the	DET
ajst-29358	117	20	number	number	NOUN
ajst-29358	117	21	of	of	ADP
ajst-29358	117	22	parameters	parameter	NOUN
ajst-29358	117	23	and	and	CCONJ
ajst-29358	117	24	the	the	DET
ajst-29358	117	25	improvement	improvement	NOUN
ajst-29358	117	26	of	of	ADP
ajst-29358	117	27	accuracy	accuracy	NOUN
ajst-29358	117	28	.	.	PUNCT
ajst-29358	118	1	when	when	SCONJ
ajst-29358	118	2	adding	add	VERB
ajst-29358	118	3	up	up	ADP
ajst-29358	118	4	to	to	PART
ajst-29358	118	5	3	3	NUM
ajst-29358	118	6	ema	ema	NOUN
ajst-29358	118	7	mechanisms	mechanism	NOUN
ajst-29358	118	8	,	,	PUNCT
ajst-29358	118	9	accuracy	accuracy	NOUN
ajst-29358	118	10	is	be	AUX
ajst-29358	118	11	improved	improve	VERB
ajst-29358	118	12	and	and	CCONJ
ajst-29358	118	13	the	the	DET
ajst-29358	118	14	increase	increase	NOUN
ajst-29358	118	15	in	in	ADP
ajst-29358	118	16	the	the	DET
ajst-29358	118	17	number	number	NOUN
ajst-29358	118	18	of	of	ADP
ajst-29358	118	19	parameters	parameter	NOUN
ajst-29358	118	20	is	be	AUX
ajst-29358	118	21	within	within	ADP
ajst-29358	118	22	an	an	DET
ajst-29358	118	23	acceptable	acceptable	ADJ
ajst-29358	118	24	range	range	NOUN
ajst-29358	118	25	.	.	PUNCT
ajst-29358	119	1	however	however	ADV
ajst-29358	119	2	,	,	PUNCT
ajst-29358	119	3	after	after	ADP
ajst-29358	119	4	adding	add	VERB
ajst-29358	119	5	4	4	NUM
ajst-29358	119	6	ema	ema	PROPN
ajst-29358	119	7	mechanisms	mechanism	NOUN
ajst-29358	119	8	,	,	PUNCT
ajst-29358	119	9	the	the	DET
ajst-29358	119	10	accuracy	accuracy	NOUN
ajst-29358	119	11	did	do	AUX
ajst-29358	119	12	not	not	PART
ajst-29358	119	13	significantly	significantly	ADV
ajst-29358	119	14	improve	improve	VERB
ajst-29358	119	15	,	,	PUNCT
ajst-29358	119	16	and	and	CCONJ
ajst-29358	119	17	the	the	DET
ajst-29358	119	18	number	number	NOUN
ajst-29358	119	19	of	of	ADP
ajst-29358	119	20	parameters	parameter	NOUN
ajst-29358	119	21	continued	continue	VERB
ajst-29358	119	22	to	to	PART
ajst-29358	119	23	increase	increase	VERB
ajst-29358	119	24	.	.	PUNCT
ajst-29358	120	1	therefore	therefore	ADV
ajst-29358	120	2	,	,	PUNCT
ajst-29358	120	3	adding	add	VERB
ajst-29358	120	4	3	3	NUM
ajst-29358	120	5	ema	ema	PROPN
ajst-29358	120	6	mechanisms	mechanism	NOUN
ajst-29358	120	7	in	in	ADP
ajst-29358	120	8	the	the	DET
ajst-29358	120	9	backbone	backbone	NOUN
ajst-29358	120	10	network	network	NOUN
ajst-29358	120	11	can	can	AUX
ajst-29358	120	12	achieve	achieve	VERB
ajst-29358	120	13	an	an	DET
ajst-29358	120	14	ideal	ideal	ADJ
ajst-29358	120	15	balance	balance	NOUN
ajst-29358	120	16	between	between	ADP
ajst-29358	120	17	speed	speed	NOUN
ajst-29358	120	18	and	and	CCONJ
ajst-29358	120	19	accuracy	accuracy	NOUN
ajst-29358	120	20	improvement	improvement	NOUN
ajst-29358	120	21	.	.	PUNCT
ajst-29358	121	1	4.3	4.3	NUM
ajst-29358	121	2	.	.	PUNCT
ajst-29358	121	3	comparative	comparative	ADJ
ajst-29358	121	4	experiment	experiment	NOUN
ajst-29358	121	5	to	to	PART
ajst-29358	121	6	fully	fully	ADV
ajst-29358	121	7	evaluate	evaluate	VERB
ajst-29358	121	8	the	the	DET
ajst-29358	121	9	advantages	advantage	NOUN
ajst-29358	121	10	of	of	ADP
ajst-29358	121	11	the	the	DET
ajst-29358	121	12	improved	improved	ADJ
ajst-29358	121	13	yolov8	yolov8	NOUN
ajst-29358	121	14	algorithm	algorithm	PROPN
ajst-29358	121	15	proposed	propose	VERB
ajst-29358	121	16	in	in	ADP
ajst-29358	121	17	this	this	DET
ajst-29358	121	18	paper	paper	NOUN
ajst-29358	121	19	in	in	ADP
ajst-29358	121	20	the	the	DET
ajst-29358	121	21	detection	detection	NOUN
ajst-29358	121	22	of	of	ADP
ajst-29358	121	23	corn	corn	NOUN
ajst-29358	121	24	large	large	ADJ
ajst-29358	121	25	spot	spot	NOUN
ajst-29358	121	26	disease	disease	NOUN
ajst-29358	121	27	,	,	PUNCT
ajst-29358	121	28	this	this	DET
ajst-29358	121	29	paper	paper	NOUN
ajst-29358	121	30	compares	compare	VERB
ajst-29358	121	31	the	the	DET
ajst-29358	121	32	improved	improved	ADJ
ajst-29358	121	33	algorithm	algorithm	NOUN
ajst-29358	121	34	with	with	ADP
ajst-29358	121	35	the	the	DET
ajst-29358	121	36	yolov8n	yolov8n	PROPN
ajst-29358	121	37	benchmark	benchmark	NOUN
ajst-29358	121	38	algorithm	algorithm	NOUN
ajst-29358	121	39	and	and	CCONJ
ajst-29358	121	40	other	other	ADJ
ajst-29358	121	41	representative	representative	ADJ
ajst-29358	121	42	network	network	NOUN
ajst-29358	121	43	models	model	NOUN
ajst-29358	121	44	in	in	ADP
ajst-29358	121	45	the	the	DET
ajst-29358	121	46	field	field	NOUN
ajst-29358	121	47	,	,	PUNCT
ajst-29358	121	48	including	include	VERB
ajst-29358	121	49	fasterrcnn	fasterrcnn	PROPN
ajst-29358	121	50	,	,	PUNCT
ajst-29358	121	51	ssd	ssd	NOUN
ajst-29358	121	52	,	,	PUNCT
ajst-29358	121	53	retinanet	retinanet	NOUN
ajst-29358	121	54	,	,	PUNCT
ajst-29358	121	55	yolov5	yolov5	NOUN
ajst-29358	121	56	,	,	PUNCT
ajst-29358	121	57	yolov6	yolov6	PROPN
ajst-29358	121	58	,	,	PUNCT
ajst-29358	121	59	yolov7	yolov7	NOUN
ajst-29358	121	60	-	-	PUNCT
ajst-29358	121	61	tiny	tiny	ADJ
ajst-29358	121	62	,	,	PUNCT
ajst-29358	121	63	and	and	CCONJ
ajst-29358	121	64	yolov8	yolov8	PROPN
ajst-29358	121	65	,	,	PUNCT
ajst-29358	121	66	as	as	SCONJ
ajst-29358	121	67	shown	show	VERB
ajst-29358	121	68	in	in	ADP
ajst-29358	121	69	table	table	NOUN
ajst-29358	121	70	3	3	NUM
ajst-29358	121	71	.	.	PUNCT
ajst-29358	121	72	table	table	NOUN
ajst-29358	121	73	5	5	NUM
ajst-29358	121	74	.	.	PUNCT
ajst-29358	121	75	comparative	comparative	ADJ
ajst-29358	121	76	experiment	experiment	NOUN
ajst-29358	121	77	results	result	NOUN
ajst-29358	121	78	of	of	ADP
ajst-29358	121	79	different	different	ADJ
ajst-29358	121	80	models	model	NOUN
ajst-29358	121	81	model	model	VERB
ajst-29358	121	82	p/%	p/%	PROPN
ajst-29358	121	83	r/%	r/%	PROPN
ajst-29358	121	84	map50/%	map50/%	NOUN
ajst-29358	121	85	f1/%	f1/%	NOUN
ajst-29358	121	86	params	param	NOUN
ajst-29358	121	87	/	/	SYM
ajst-29358	121	88	m	m	NOUN
ajst-29358	121	89	flops	flop	NOUN
ajst-29358	121	90	/	/	SYM
ajst-29358	121	91	g	g	NOUN
ajst-29358	121	92	faster	fast	ADV
ajst-29358	121	93	r	r	NOUN
ajst-29358	121	94	-	-	PUNCT
ajst-29358	121	95	cnn	cnn	PROPN
ajst-29358	121	96	74.1	74.1	NUM
ajst-29358	121	97	78.5	78.5	NUM
ajst-29358	121	98	75.6	75.6	NUM
ajst-29358	121	99	76.2	76.2	NUM
ajst-29358	121	100	76.73	76.73	NUM
ajst-29358	121	101	95.7	95.7	NUM
ajst-29358	121	102	ssd	ssd	NOUN
ajst-29358	121	103	80.4	80.4	NUM
ajst-29358	121	104	77.4	77.4	NUM
ajst-29358	121	105	79.4	79.4	NUM
ajst-29358	121	106	78.9	78.9	NUM
ajst-29358	121	107	24.71	24.71	NUM
ajst-29358	121	108	57.2	57.2	NUM
ajst-29358	121	109	retinanet	retinanet	NOUN
ajst-29358	121	110	74.6	74.6	NUM
ajst-29358	121	111	73.1	73.1	NUM
ajst-29358	121	112	74.7	74.7	NUM
ajst-29358	121	113	73.8	73.8	NUM
ajst-29358	121	114	18.9	18.9	NUM
ajst-29358	121	115	60.4	60.4	NUM
ajst-29358	121	116	yolov5	yolov5	NOUN
ajst-29358	121	117	75.3	75.3	NUM
ajst-29358	121	118	76.4	76.4	NUM
ajst-29358	121	119	76.3	76.3	NUM
ajst-29358	121	120	75.8	75.8	NUM
ajst-29358	121	121	6.51	6.51	NUM
ajst-29358	121	122	15.8	15.8	NUM
ajst-29358	121	123	yolov6	yolov6	NOUN
ajst-29358	121	124	77.1	77.1	NUM
ajst-29358	121	125	73.7	73.7	NUM
ajst-29358	121	126	77.3	77.3	NUM
ajst-29358	121	127	75.4	75.4	NUM
ajst-29358	121	128	9.50	9.50	NUM
ajst-29358	121	129	17.5	17.5	NUM
ajst-29358	121	130	yolov7	yolov7	NOUN
ajst-29358	121	131	-	-	PUNCT
ajst-29358	121	132	tiny	tiny	ADJ
ajst-29358	121	133	78.5	78.5	NUM
ajst-29358	121	134	76.9	76.9	NUM
ajst-29358	121	135	77.7	77.7	NUM
ajst-29358	121	136	78.2	78.2	NUM
ajst-29358	121	137	6.14	6.14	NUM
ajst-29358	121	138	12.1	12.1	NUM
ajst-29358	121	139	yolov8	yolov8	NOUN
ajst-29358	121	140	80.1	80.1	NUM
ajst-29358	121	141	73.4	73.4	NUM
ajst-29358	121	142	77.2	77.2	NUM
ajst-29358	121	143	76.6	76.6	NUM
ajst-29358	121	144	9.31	9.31	NUM
ajst-29358	121	145	20.4	20.4	NUM
ajst-29358	121	146	improved	improve	VERB
ajst-29358	121	147	yolov8n	yolov8n	NOUN
ajst-29358	121	148	82.8	82.8	NUM
ajst-29358	121	149	78.1	78.1	NUM
ajst-29358	121	150	80.3	80.3	NUM
ajst-29358	121	151	80.4	80.4	NUM
ajst-29358	121	152	6.53	6.53	NUM
ajst-29358	121	153	16.1	16.1	NUM
ajst-29358	121	154	the	the	DET
ajst-29358	121	155	results	result	NOUN
ajst-29358	121	156	of	of	ADP
ajst-29358	121	157	the	the	DET
ajst-29358	121	158	comparative	comparative	ADJ
ajst-29358	121	159	experiment	experiment	NOUN
ajst-29358	121	160	show	show	VERB
ajst-29358	121	161	that	that	SCONJ
ajst-29358	121	162	the	the	DET
ajst-29358	121	163	improved	improved	ADJ
ajst-29358	121	164	algorithm	algorithm	NOUN
ajst-29358	121	165	proposed	propose	VERB
ajst-29358	121	166	in	in	ADP
ajst-29358	121	167	this	this	DET
ajst-29358	121	168	paper	paper	NOUN
ajst-29358	121	169	has	have	AUX
ajst-29358	121	170	achieved	achieve	VERB
ajst-29358	121	171	the	the	DET
ajst-29358	121	172	highest	high	ADJ
ajst-29358	121	173	level	level	NOUN
ajst-29358	121	174	in	in	ADP
ajst-29358	121	175	terms	term	NOUN
ajst-29358	121	176	of	of	ADP
ajst-29358	121	177	precision	precision	NOUN
ajst-29358	121	178	and	and	CCONJ
ajst-29358	121	179	map50	map50	PROPN
ajst-29358	121	180	values	value	NOUN
ajst-29358	121	181	.	.	PUNCT
ajst-29358	122	1	the	the	DET
ajst-29358	122	2	algorithm	algorithm	NOUN
ajst-29358	122	3	not	not	PART
ajst-29358	122	4	only	only	ADV
ajst-29358	122	5	has	have	VERB
ajst-29358	122	6	excellent	excellent	ADJ
ajst-29358	122	7	performance	performance	NOUN
ajst-29358	122	8	but	but	CCONJ
ajst-29358	122	9	also	also	ADV
ajst-29358	122	10	the	the	DET
ajst-29358	122	11	fastest	fast	ADJ
ajst-29358	122	12	detection	detection	NOUN
ajst-29358	122	13	speed	speed	NOUN
ajst-29358	122	14	,	,	PUNCT
ajst-29358	122	15	and	and	CCONJ
ajst-29358	122	16	the	the	DET
ajst-29358	122	17	parameter	parameter	NOUN
ajst-29358	122	18	volume	volume	NOUN
ajst-29358	122	19	is	be	AUX
ajst-29358	122	20	not	not	PART
ajst-29358	122	21	much	much	ADV
ajst-29358	122	22	different	different	ADJ
ajst-29358	122	23	from	from	ADP
ajst-29358	122	24	other	other	ADJ
ajst-29358	122	25	models	model	NOUN
ajst-29358	122	26	.	.	PUNCT
ajst-29358	123	1	compared	compare	VERB
ajst-29358	123	2	with	with	ADP
ajst-29358	123	3	the	the	DET
ajst-29358	123	4	benchmark	benchmark	ADJ
ajst-29358	123	5	algorithm	algorithm	NOUN
ajst-29358	123	6	yolov8n	yolov8n	PROPN
ajst-29358	123	7	,	,	PUNCT
ajst-29358	123	8	the	the	DET
ajst-29358	123	9	improved	improved	ADJ
ajst-29358	123	10	algorithm	algorithm	NOUN
ajst-29358	123	11	shows	show	VERB
ajst-29358	123	12	significant	significant	ADJ
ajst-29358	123	13	advantages	advantage	NOUN
ajst-29358	123	14	in	in	ADP
ajst-29358	123	15	all	all	DET
ajst-29358	123	16	indicators	indicator	NOUN
ajst-29358	123	17	.	.	PUNCT
ajst-29358	124	1	considering	consider	VERB
ajst-29358	124	2	all	all	DET
ajst-29358	124	3	performance	performance	NOUN
ajst-29358	124	4	indicators	indicator	NOUN
ajst-29358	124	5	,	,	PUNCT
ajst-29358	124	6	the	the	DET
ajst-29358	124	7	comprehensive	comprehensive	ADJ
ajst-29358	124	8	performance	performance	NOUN
ajst-29358	124	9	of	of	ADP
ajst-29358	124	10	the	the	DET
ajst-29358	124	11	improved	improved	ADJ
ajst-29358	124	12	algorithm	algorithm	NOUN
ajst-29358	124	13	proposed	propose	VERB
ajst-29358	124	14	in	in	ADP
ajst-29358	124	15	this	this	DET
ajst-29358	124	16	paper	paper	NOUN
ajst-29358	124	17	is	be	AUX
ajst-29358	124	18	the	the	DET
ajst-29358	124	19	best	good	ADJ
ajst-29358	124	20	,	,	PUNCT
ajst-29358	124	21	with	with	ADP
ajst-29358	124	22	good	good	ADJ
ajst-29358	124	23	detection	detection	NOUN
ajst-29358	124	24	ability	ability	NOUN
ajst-29358	124	25	for	for	ADP
ajst-29358	124	26	corn	corn	NOUN
ajst-29358	124	27	large	large	ADJ
ajst-29358	124	28	spot	spot	NOUN
ajst-29358	124	29	disease	disease	NOUN
ajst-29358	124	30	in	in	ADP
ajst-29358	124	31	uav	uav	PROPN
ajst-29358	124	32	images	image	NOUN
ajst-29358	124	33	,	,	PUNCT
ajst-29358	124	34	and	and	CCONJ
ajst-29358	124	35	good	good	ADJ
ajst-29358	124	36	performance	performance	NOUN
ajst-29358	124	37	in	in	ADP
ajst-29358	124	38	both	both	PRON
ajst-29358	124	39	speed	speed	NOUN
ajst-29358	124	40	and	and	CCONJ
ajst-29358	124	41	accuracy	accuracy	NOUN
ajst-29358	124	42	.	.	PUNCT
ajst-29358	125	1	to	to	PART
ajst-29358	125	2	intuitively	intuitively	ADV
ajst-29358	125	3	show	show	VERB
ajst-29358	125	4	the	the	DET
ajst-29358	125	5	superiority	superiority	NOUN
ajst-29358	125	6	of	of	ADP
ajst-29358	125	7	this	this	DET
ajst-29358	125	8	algorithm	algorithm	NOUN
ajst-29358	125	9	,	,	PUNCT
ajst-29358	125	10	the	the	DET
ajst-29358	125	11	map50	map50	PROPN
ajst-29358	125	12	change	change	NOUN
ajst-29358	125	13	curve	curve	NOUN
ajst-29358	125	14	is	be	AUX
ajst-29358	125	15	compared	compare	VERB
ajst-29358	125	16	,	,	PUNCT
ajst-29358	125	17	and	and	CCONJ
ajst-29358	125	18	the	the	DET
ajst-29358	125	19	experimental	experimental	ADJ
ajst-29358	125	20	results	result	NOUN
ajst-29358	125	21	are	be	AUX
ajst-29358	125	22	shown	show	VERB
ajst-29358	125	23	in	in	ADP
ajst-29358	125	24	figure	figure	NOUN
ajst-29358	125	25	6	6	NUM
ajst-29358	125	26	.	.	PUNCT
ajst-29358	126	1	from	from	ADP
ajst-29358	126	2	figure	figure	NOUN
ajst-29358	126	3	6	6	NUM
ajst-29358	126	4	,	,	PUNCT
ajst-29358	126	5	it	it	PRON
ajst-29358	126	6	can	can	AUX
ajst-29358	126	7	be	be	AUX
ajst-29358	126	8	seen	see	VERB
ajst-29358	126	9	that	that	SCONJ
ajst-29358	126	10	all	all	DET
ajst-29358	126	11	algorithms	algorithm	NOUN
ajst-29358	126	12	can	can	AUX
ajst-29358	126	13	converge	converge	VERB
ajst-29358	126	14	through	through	ADP
ajst-29358	126	15	continuous	continuous	ADJ
ajst-29358	126	16	iteration	iteration	NOUN
ajst-29358	126	17	,	,	PUNCT
ajst-29358	126	18	but	but	CCONJ
ajst-29358	126	19	the	the	DET
ajst-29358	126	20	curve	curve	NOUN
ajst-29358	126	21	of	of	ADP
ajst-29358	126	22	the	the	DET
ajst-29358	126	23	improved	improved	ADJ
ajst-29358	126	24	yolov8n	yolov8n	NOUN
ajst-29358	126	25	is	be	AUX
ajst-29358	126	26	significantly	significantly	ADV
ajst-29358	126	27	higher	high	ADJ
ajst-29358	126	28	than	than	ADP
ajst-29358	126	29	other	other	ADJ
ajst-29358	126	30	algorithms	algorithm	NOUN
ajst-29358	126	31	,	,	PUNCT
ajst-29358	126	32	proving	prove	VERB
ajst-29358	126	33	that	that	SCONJ
ajst-29358	126	34	the	the	DET
ajst-29358	126	35	improved	improved	ADJ
ajst-29358	126	36	yolov8	yolov8	NOUN
ajst-29358	126	37	algorithm	algorithm	PROPN
ajst-29358	126	38	has	have	AUX
ajst-29358	126	39	effectively	effectively	ADV
ajst-29358	126	40	improved	improve	VERB
ajst-29358	126	41	detection	detection	NOUN
ajst-29358	126	42	accuracy	accuracy	NOUN
ajst-29358	126	43	.	.	PUNCT
ajst-29358	127	1	figure	figure	NOUN
ajst-29358	127	2	6	6	NUM
ajst-29358	127	3	.	.	PUNCT
ajst-29358	127	4	comparison	comparison	NOUN
ajst-29358	127	5	of	of	ADP
ajst-29358	127	6	map50	map50	PROPN
ajst-29358	127	7	for	for	ADP
ajst-29358	127	8	different	different	ADJ
ajst-29358	127	9	models	model	NOUN
ajst-29358	127	10	5	5	NUM
ajst-29358	127	11	.	.	X
ajst-29358	128	1	summary	summary	NOUN
ajst-29358	128	2	in	in	ADP
ajst-29358	128	3	this	this	DET
ajst-29358	128	4	study	study	NOUN
ajst-29358	128	5	,	,	PUNCT
ajst-29358	128	6	considering	consider	VERB
ajst-29358	128	7	the	the	DET
ajst-29358	128	8	high	high	ADJ
ajst-29358	128	9	complexity	complexity	NOUN
ajst-29358	128	10	and	and	CCONJ
ajst-29358	128	11	low	low	ADJ
ajst-29358	128	12	157	157	NUM
ajst-29358	128	13	distinguishability	distinguishability	NOUN
ajst-29358	128	14	of	of	ADP
ajst-29358	128	15	the	the	DET
ajst-29358	128	16	morphological	morphological	ADJ
ajst-29358	128	17	and	and	CCONJ
ajst-29358	128	18	color	color	NOUN
ajst-29358	128	19	features	feature	NOUN
ajst-29358	128	20	of	of	ADP
ajst-29358	128	21	corn	corn	NOUN
ajst-29358	128	22	large	large	ADJ
ajst-29358	128	23	spot	spot	NOUN
ajst-29358	128	24	disease	disease	NOUN
ajst-29358	128	25	lesions	lesion	NOUN
ajst-29358	128	26	,	,	PUNCT
ajst-29358	128	27	an	an	DET
ajst-29358	128	28	improved	improved	ADJ
ajst-29358	128	29	yolov8	yolov8	NOUN
ajst-29358	128	30	corn	corn	NOUN
ajst-29358	128	31	large	large	ADJ
ajst-29358	128	32	spot	spot	NOUN
ajst-29358	128	33	disease	disease	NOUN
ajst-29358	128	34	detection	detection	NOUN
ajst-29358	128	35	model	model	NOUN
ajst-29358	128	36	was	be	AUX
ajst-29358	128	37	proposed	propose	VERB
ajst-29358	128	38	,	,	PUNCT
ajst-29358	128	39	specifically	specifically	ADV
ajst-29358	128	40	for	for	ADP
ajst-29358	128	41	detecting	detect	VERB
ajst-29358	128	42	corn	corn	NOUN
ajst-29358	128	43	large	large	ADJ
ajst-29358	128	44	spot	spot	NOUN
ajst-29358	128	45	disease	disease	NOUN
ajst-29358	128	46	.	.	PUNCT
ajst-29358	129	1	the	the	DET
ajst-29358	129	2	model	model	NOUN
ajst-29358	129	3	's	's	PART
ajst-29358	129	4	training	training	NOUN
ajst-29358	129	5	was	be	AUX
ajst-29358	129	6	based	base	VERB
ajst-29358	129	7	on	on	ADP
ajst-29358	129	8	a	a	DET
ajst-29358	129	9	mixed	mixed	ADJ
ajst-29358	129	10	dataset	dataset	NOUN
ajst-29358	129	11	,	,	PUNCT
ajst-29358	129	12	which	which	PRON
ajst-29358	129	13	was	be	AUX
ajst-29358	129	14	composed	compose	VERB
ajst-29358	129	15	of	of	ADP
ajst-29358	129	16	some	some	DET
ajst-29358	129	17	public	public	ADJ
ajst-29358	129	18	dataset	dataset	NOUN
ajst-29358	129	19	images	image	NOUN
ajst-29358	129	20	and	and	CCONJ
ajst-29358	129	21	our	our	PRON
ajst-29358	129	22	own	own	ADJ
ajst-29358	129	23	collected	collect	VERB
ajst-29358	129	24	corn	corn	NOUN
ajst-29358	129	25	large	large	ADJ
ajst-29358	129	26	spot	spot	NOUN
ajst-29358	129	27	disease	disease	PROPN
ajst-29358	129	28	uav	uav	PROPN
ajst-29358	129	29	images	image	NOUN
ajst-29358	129	30	.	.	PUNCT
ajst-29358	130	1	the	the	DET
ajst-29358	130	2	following	follow	VERB
ajst-29358	130	3	are	be	AUX
ajst-29358	130	4	the	the	DET
ajst-29358	130	5	main	main	ADJ
ajst-29358	130	6	conclusions	conclusion	NOUN
ajst-29358	130	7	of	of	ADP
ajst-29358	130	8	this	this	DET
ajst-29358	130	9	study	study	NOUN
ajst-29358	130	10	:	:	PUNCT
ajst-29358	130	11	first	first	ADV
ajst-29358	130	12	,	,	PUNCT
ajst-29358	130	13	a	a	DET
ajst-29358	130	14	corn	corn	NOUN
ajst-29358	130	15	large	large	ADJ
ajst-29358	130	16	spot	spot	NOUN
ajst-29358	130	17	disease	disease	NOUN
ajst-29358	130	18	image	image	NOUN
ajst-29358	130	19	dataset	dataset	VERB
ajst-29358	130	20	under	under	ADP
ajst-29358	130	21	natural	natural	ADJ
ajst-29358	130	22	conditions	condition	NOUN
ajst-29358	130	23	was	be	AUX
ajst-29358	130	24	successfully	successfully	ADV
ajst-29358	130	25	constructed	construct	VERB
ajst-29358	130	26	,	,	PUNCT
ajst-29358	130	27	and	and	CCONJ
ajst-29358	130	28	its	its	PRON
ajst-29358	130	29	data	data	NOUN
ajst-29358	130	30	integrity	integrity	NOUN
ajst-29358	130	31	and	and	CCONJ
ajst-29358	130	32	reliability	reliability	NOUN
ajst-29358	130	33	were	be	AUX
ajst-29358	130	34	ensured	ensure	VERB
ajst-29358	130	35	,	,	PUNCT
ajst-29358	130	36	which	which	PRON
ajst-29358	130	37	is	be	AUX
ajst-29358	130	38	crucial	crucial	ADJ
ajst-29358	130	39	for	for	ADP
ajst-29358	130	40	model	model	NOUN
ajst-29358	130	41	training	training	NOUN
ajst-29358	130	42	and	and	CCONJ
ajst-29358	130	43	evaluation	evaluation	NOUN
ajst-29358	130	44	.	.	PUNCT
ajst-29358	131	1	second	second	ADJ
ajst-29358	131	2	,	,	PUNCT
ajst-29358	131	3	the	the	DET
ajst-29358	131	4	ema	ema	PROPN
ajst-29358	131	5	attention	attention	NOUN
ajst-29358	131	6	mechanism	mechanism	NOUN
ajst-29358	131	7	was	be	AUX
ajst-29358	131	8	added	add	VERB
ajst-29358	131	9	to	to	ADP
ajst-29358	131	10	the	the	DET
ajst-29358	131	11	backbone	backbone	NOUN
ajst-29358	131	12	network	network	NOUN
ajst-29358	131	13	,	,	PUNCT
ajst-29358	131	14	the	the	DET
ajst-29358	131	15	yolov8	yolov8	PROPN
ajst-29358	131	16	-	-	PUNCT
ajst-29358	131	17	bifpn	bifpn	PROPN
ajst-29358	131	18	was	be	AUX
ajst-29358	131	19	designed	design	VERB
ajst-29358	131	20	to	to	PART
ajst-29358	131	21	optimize	optimize	VERB
ajst-29358	131	22	the	the	DET
ajst-29358	131	23	model	model	NOUN
ajst-29358	131	24	's	's	PART
ajst-29358	131	25	neck	neck	NOUN
ajst-29358	131	26	structure	structure	NOUN
ajst-29358	131	27	,	,	PUNCT
ajst-29358	131	28	and	and	CCONJ
ajst-29358	131	29	a	a	DET
ajst-29358	131	30	new	new	ADJ
ajst-29358	131	31	wiou	wiou	NOUN
ajst-29358	131	32	loss	loss	NOUN
ajst-29358	131	33	function	function	NOUN
ajst-29358	131	34	was	be	AUX
ajst-29358	131	35	introduced	introduce	VERB
ajst-29358	131	36	to	to	PART
ajst-29358	131	37	improve	improve	VERB
ajst-29358	131	38	the	the	DET
ajst-29358	131	39	yolov8	yolov8	NOUN
ajst-29358	131	40	algorithm	algorithm	PROPN
ajst-29358	131	41	,	,	PUNCT
ajst-29358	131	42	making	make	VERB
ajst-29358	131	43	it	it	PRON
ajst-29358	131	44	specifically	specifically	ADV
ajst-29358	131	45	applicable	applicable	ADJ
ajst-29358	131	46	to	to	ADP
ajst-29358	131	47	the	the	DET
ajst-29358	131	48	detection	detection	NOUN
ajst-29358	131	49	of	of	ADP
ajst-29358	131	50	corn	corn	NOUN
ajst-29358	131	51	large	large	ADJ
ajst-29358	131	52	spot	spot	NOUN
ajst-29358	131	53	disease	disease	NOUN
ajst-29358	131	54	in	in	ADP
ajst-29358	131	55	uav	uav	PROPN
ajst-29358	131	56	images	image	NOUN
ajst-29358	131	57	.	.	PUNCT
ajst-29358	132	1	finally	finally	ADV
ajst-29358	132	2	,	,	PUNCT
ajst-29358	132	3	compared	compare	VERB
ajst-29358	132	4	with	with	ADP
ajst-29358	132	5	the	the	DET
ajst-29358	132	6	benchmark	benchmark	ADJ
ajst-29358	132	7	algorithm	algorithm	NOUN
ajst-29358	132	8	yolov8n	yolov8n	PROPN
ajst-29358	132	9	,	,	PUNCT
ajst-29358	132	10	the	the	DET
ajst-29358	132	11	proposed	propose	VERB
ajst-29358	132	12	algorithm	algorithm	NOUN
ajst-29358	132	13	shows	show	VERB
ajst-29358	132	14	superiority	superiority	NOUN
ajst-29358	132	15	in	in	ADP
ajst-29358	132	16	many	many	ADJ
ajst-29358	132	17	aspects	aspect	NOUN
ajst-29358	132	18	,	,	PUNCT
ajst-29358	132	19	in	in	ADP
ajst-29358	132	20	comparison	comparison	NOUN
ajst-29358	132	21	with	with	ADP
ajst-29358	132	22	other	other	ADJ
ajst-29358	132	23	advanced	advanced	ADJ
ajst-29358	132	24	methods	method	NOUN
ajst-29358	132	25	,	,	PUNCT
ajst-29358	132	26	it	it	PRON
ajst-29358	132	27	has	have	VERB
ajst-29358	132	28	the	the	DET
ajst-29358	132	29	best	good	ADJ
ajst-29358	132	30	comprehensive	comprehensive	ADJ
ajst-29358	132	31	performance	performance	NOUN
ajst-29358	132	32	,	,	PUNCT
ajst-29358	132	33	meeting	meet	VERB
ajst-29358	132	34	the	the	DET
ajst-29358	132	35	requirements	requirement	NOUN
ajst-29358	132	36	for	for	ADP
ajst-29358	132	37	real	real	ADJ
ajst-29358	132	38	-	-	PUNCT
ajst-29358	132	39	time	time	NOUN
ajst-29358	132	40	and	and	CCONJ
ajst-29358	132	41	accuracy	accuracy	NOUN
ajst-29358	132	42	.	.	PUNCT
ajst-29358	133	1	in	in	ADP
ajst-29358	133	2	future	future	ADJ
ajst-29358	133	3	research	research	NOUN
ajst-29358	133	4	,	,	PUNCT
ajst-29358	133	5	we	we	PRON
ajst-29358	133	6	plan	plan	VERB
ajst-29358	133	7	to	to	PART
ajst-29358	133	8	increase	increase	VERB
ajst-29358	133	9	the	the	DET
ajst-29358	133	10	types	type	NOUN
ajst-29358	133	11	of	of	ADP
ajst-29358	133	12	diseases	disease	NOUN
ajst-29358	133	13	and	and	CCONJ
ajst-29358	133	14	the	the	DET
ajst-29358	133	15	number	number	NOUN
ajst-29358	133	16	of	of	ADP
ajst-29358	133	17	samples	sample	NOUN
ajst-29358	133	18	to	to	PART
ajst-29358	133	19	improve	improve	VERB
ajst-29358	133	20	the	the	DET
ajst-29358	133	21	model	model	NOUN
ajst-29358	133	22	's	's	PART
ajst-29358	133	23	accuracy	accuracy	NOUN
ajst-29358	133	24	and	and	CCONJ
ajst-29358	133	25	recognition	recognition	NOUN
ajst-29358	133	26	ability	ability	NOUN
ajst-29358	133	27	.	.	PUNCT
ajst-29358	134	1	in	in	ADP
ajst-29358	134	2	addition	addition	NOUN
ajst-29358	134	3	,	,	PUNCT
ajst-29358	134	4	we	we	PRON
ajst-29358	134	5	will	will	AUX
ajst-29358	134	6	continue	continue	VERB
ajst-29358	134	7	to	to	PART
ajst-29358	134	8	explore	explore	VERB
ajst-29358	134	9	the	the	DET
ajst-29358	134	10	combination	combination	NOUN
ajst-29358	134	11	and	and	CCONJ
ajst-29358	134	12	optimization	optimization	NOUN
ajst-29358	134	13	improvement	improvement	NOUN
ajst-29358	134	14	of	of	ADP
ajst-29358	134	15	different	different	ADJ
ajst-29358	134	16	network	network	NOUN
ajst-29358	134	17	models	model	NOUN
ajst-29358	134	18	to	to	PART
ajst-29358	134	19	further	far	ADV
ajst-29358	134	20	improve	improve	VERB
ajst-29358	134	21	the	the	DET
ajst-29358	134	22	model	model	NOUN
ajst-29358	134	23	's	's	PART
ajst-29358	134	24	training	training	NOUN
ajst-29358	134	25	speed	speed	NOUN
ajst-29358	134	26	and	and	CCONJ
ajst-29358	134	27	accuracy	accuracy	NOUN
ajst-29358	134	28	.	.	PUNCT
ajst-29358	135	1	references	reference	NOUN
ajst-29358	135	2	[	[	X
ajst-29358	135	3	1	1	NUM
ajst-29358	135	4	]	]	X
ajst-29358	135	5	han	han	PROPN
ajst-29358	135	6	dongfeng	dongfeng	PROPN
ajst-29358	135	7	,	,	PUNCT
ajst-29358	135	8	li	li	PROPN
ajst-29358	135	9	feng	feng	PROPN
ajst-29358	135	10	,	,	PUNCT
ajst-29358	135	11	qin	qin	PROPN
ajst-29358	135	12	quan	quan	PROPN
ajst-29358	135	13	,	,	PUNCT
ajst-29358	135	14	et	et	PROPN
ajst-29358	135	15	al	al	PROPN
ajst-29358	135	16	.	.	PUNCT
ajst-29358	136	1	a	a	DET
ajst-29358	136	2	refined	refined	ADJ
ajst-29358	136	3	identification	identification	NOUN
ajst-29358	136	4	method	method	NOUN
ajst-29358	136	5	for	for	ADP
ajst-29358	136	6	summer	summer	NOUN
ajst-29358	136	7	corn	corn	NOUN
ajst-29358	136	8	planting	planting	NOUN
ajst-29358	136	9	area	area	NOUN
ajst-29358	136	10	based	base	VERB
ajst-29358	136	11	on	on	ADP
ajst-29358	136	12	gee	gee	NOUN
ajst-29358	136	13	and	and	CCONJ
ajst-29358	136	14	sentinel-1/2	sentinel-1/2	PROPN
ajst-29358	136	15	data[j	data[j	PROPN
ajst-29358	136	16	]	]	PUNCT
ajst-29358	136	17	.	.	PUNCT
ajst-29358	137	1	journal	journal	PROPN
ajst-29358	137	2	of	of	ADP
ajst-29358	137	3	marine	marine	ADJ
ajst-29358	137	4	meteorology	meteorology	NOUN
ajst-29358	137	5	,	,	PUNCT
ajst-29358	137	6	2024	2024	NUM
ajst-29358	137	7	,	,	PUNCT
ajst-29358	137	8	44(03	44(03	NOUN
ajst-29358	137	9	):	):	PUNCT
ajst-29358	137	10	122	122	NUM
ajst-29358	137	11	-	-	SYM
ajst-29358	137	12	132	132	NUM
ajst-29358	137	13	.	.	PUNCT
ajst-29358	138	1	[	[	X
ajst-29358	138	2	2	2	NUM
ajst-29358	138	3	]	]	PUNCT
ajst-29358	138	4	xu	xu	PROPN
ajst-29358	138	5	xingbin	xingbin	PROPN
ajst-29358	138	6	.	.	PUNCT
ajst-29358	139	1	characteristics	characteristic	NOUN
ajst-29358	139	2	of	of	ADP
ajst-29358	139	3	corn	corn	NOUN
ajst-29358	139	4	disease	disease	NOUN
ajst-29358	139	5	occurrence	occurrence	NOUN
ajst-29358	139	6	and	and	CCONJ
ajst-29358	139	7	control	control	NOUN
ajst-29358	139	8	techniques[j	techniques[j	PROPN
ajst-29358	139	9	]	]	PUNCT
ajst-29358	139	10	.	.	PUNCT
ajst-29358	140	1	modern	modern	ADJ
ajst-29358	140	2	agriculture	agriculture	NOUN
ajst-29358	140	3	,	,	PUNCT
ajst-29358	140	4	2024	2024	NUM
ajst-29358	140	5	,	,	PUNCT
ajst-29358	140	6	(	(	PUNCT
ajst-29358	140	7	08	08	NUM
ajst-29358	140	8	):	):	PUNCT
ajst-29358	140	9	47	47	NUM
ajst-29358	140	10	-	-	SYM
ajst-29358	140	11	49	49	NUM
ajst-29358	140	12	.	.	PUNCT
ajst-29358	141	1	[	[	X
ajst-29358	141	2	3	3	X
ajst-29358	141	3	]	]	X
ajst-29358	141	4	dai	dai	PROPN
ajst-29358	141	5	linhua	linhua	PROPN
ajst-29358	141	6	,	,	PUNCT
ajst-29358	141	7	li	li	PROPN
ajst-29358	141	8	yuansong	yuansong	PROPN
ajst-29358	141	9	,	,	PUNCT
ajst-29358	141	10	shi	shi	PROPN
ajst-29358	141	11	rui	rui	PROPN
ajst-29358	141	12	.	.	PROPN
ajst-29358	142	1	improved	improve	VERB
ajst-29358	142	2	yolov8n	yolov8n	NOUN
ajst-29358	142	3	algorithm	algorithm	NOUN
ajst-29358	142	4	for	for	ADP
ajst-29358	142	5	rice	rice	NOUN
ajst-29358	142	6	leaf	leaf	NOUN
ajst-29358	142	7	disease	disease	NOUN
ajst-29358	142	8	detection[j	detection[j	PROPN
ajst-29358	142	9	]	]	PUNCT
ajst-29358	142	10	.	.	PUNCT
ajst-29358	143	1	journal	journal	PROPN
ajst-29358	143	2	of	of	ADP
ajst-29358	143	3	hubei	hubei	PROPN
ajst-29358	143	4	university	university	PROPN
ajst-29358	143	5	for	for	ADP
ajst-29358	143	6	nationalities	nationality	NOUN
ajst-29358	143	7	(	(	PUNCT
ajst-29358	143	8	natural	natural	ADJ
ajst-29358	143	9	science	science	NOUN
ajst-29358	143	10	edition	edition	NOUN
ajst-29358	143	11	)	)	PUNCT
ajst-29358	143	12	,	,	PUNCT
ajst-29358	143	13	2024	2024	NUM
ajst-29358	143	14	,	,	PUNCT
ajst-29358	143	15	42(03	42(03	NOUN
ajst-29358	143	16	):	):	PUNCT
ajst-29358	143	17	382	382	NUM
ajst-29358	143	18	-	-	SYM
ajst-29358	143	19	388	388	NUM
ajst-29358	143	20	.	.	PUNCT
ajst-29358	144	1	[	[	X
ajst-29358	144	2	4	4	NUM
ajst-29358	144	3	]	]	X
ajst-29358	144	4	meng	meng	PROPN
ajst-29358	144	5	wei	wei	PROPN
ajst-29358	144	6	,	,	PUNCT
ajst-29358	144	7	wang	wang	PROPN
ajst-29358	144	8	jinghui	jinghui	PROPN
ajst-29358	144	9	.	.	PUNCT
ajst-29358	144	10	progress	progress	NOUN
ajst-29358	144	11	in	in	ADP
ajst-29358	144	12	the	the	DET
ajst-29358	144	13	application	application	NOUN
ajst-29358	144	14	of	of	ADP
ajst-29358	144	15	uav	uav	PROPN
ajst-29358	144	16	technology	technology	PROPN
ajst-29358	144	17	in	in	ADP
ajst-29358	144	18	forest	forest	NOUN
ajst-29358	144	19	pest	pest	NOUN
ajst-29358	144	20	and	and	CCONJ
ajst-29358	144	21	disease	disease	PROPN
ajst-29358	144	22	monitoring[j	monitoring[j	PROPN
ajst-29358	144	23	/	/	SYM
ajst-29358	144	24	ol	ol	PROPN
ajst-29358	144	25	]	]	PUNCT
ajst-29358	144	26	.	.	PUNCT
ajst-29358	145	1	forest	forest	NOUN
ajst-29358	145	2	science	science	NOUN
ajst-29358	145	3	and	and	CCONJ
ajst-29358	145	4	communication	communication	NOUN
ajst-29358	145	5	,	,	PUNCT
ajst-29358	145	6	1	1	NUM
ajst-29358	145	7	-	-	SYM
ajst-29358	145	8	6	6	NUM
ajst-29358	146	1	[	[	NOUN
ajst-29358	146	2	2024	2024	NUM
ajst-29358	146	3	-	-	SYM
ajst-29358	146	4	10	10	NUM
ajst-29358	146	5	-	-	SYM
ajst-29358	146	6	10	10	NUM
ajst-29358	146	7	]	]	PUNCT
ajst-29358	146	8	.	.	PUNCT
ajst-29358	147	1	[	[	X
ajst-29358	147	2	5	5	NUM
ajst-29358	147	3	]	]	X
ajst-29358	147	4	lu	lu	PROPN
ajst-29358	147	5	,	,	PUNCT
ajst-29358	147	6	y.	y.	PROPN
ajst-29358	147	7	,	,	PUNCT
ajst-29358	147	8	yu	yu	PROPN
ajst-29358	147	9	,	,	PUNCT
ajst-29358	147	10	j.	j.	PROPN
ajst-29358	147	11	h.	h.	PROPN
ajst-29358	147	12	,	,	PUNCT
ajst-29358	147	13	zhu	zhu	PROPN
ajst-29358	147	14	,	,	PUNCT
ajst-29358	147	15	x.	x.	PROPN
ajst-29358	147	16	f.	f.	PROPN
ajst-29358	147	17	,	,	PUNCT
ajst-29358	147	18	zhang	zhang	PROPN
ajst-29358	147	19	,	,	PUNCT
ajst-29358	147	20	b.	b.	PROPN
ajst-29358	147	21	f.	f.	PROPN
ajst-29358	147	22	,	,	PUNCT
ajst-29358	147	23	&	&	CCONJ
ajst-29358	147	24	sun	sun	PROPN
ajst-29358	147	25	,	,	PUNCT
ajst-29358	147	26	z.	z.	PROPN
ajst-29358	147	27	f.	f.	PROPN
ajst-29358	147	28	(	(	PUNCT
ajst-29358	147	29	2024	2024	NUM
ajst-29358	147	30	)	)	PUNCT
ajst-29358	147	31	.	.	PUNCT
ajst-29358	148	1	yolov8	yolov8	NOUN
ajst-29358	148	2	-	-	PUNCT
ajst-29358	148	3	rice	rice	PROPN
ajst-29358	148	4	:	:	PUNCT
ajst-29358	148	5	a	a	DET
ajst-29358	148	6	rice	rice	NOUN
ajst-29358	148	7	leaf	leaf	NOUN
ajst-29358	148	8	disease	disease	NOUN
ajst-29358	148	9	detection	detection	NOUN
ajst-29358	148	10	model	model	NOUN
ajst-29358	148	11	based	base	VERB
ajst-29358	148	12	on	on	ADP
ajst-29358	148	13	yolov8	yolov8	PROPN
ajst-29358	148	14	.	.	PUNCT
ajst-29358	149	1	paddy	paddy	NOUN
ajst-29358	149	2	and	and	CCONJ
ajst-29358	149	3	water	water	NOUN
ajst-29358	149	4	environment	environment	NOUN
ajst-29358	149	5	.	.	PUNCT
ajst-29358	150	1	[	[	X
ajst-29358	150	2	6	6	NUM
ajst-29358	150	3	]	]	PUNCT
ajst-29358	150	4	he	he	PRON
ajst-29358	150	5	,	,	PUNCT
ajst-29358	150	6	c.	c.	PROPN
ajst-29358	150	7	c.	c.	PROPN
ajst-29358	150	8	,	,	PUNCT
ajst-29358	150	9	wan	wan	PROPN
ajst-29358	150	10	,	,	PUNCT
ajst-29358	150	11	f.	f.	PROPN
ajst-29358	150	12	x.	x.	PROPN
ajst-29358	150	13	,	,	PUNCT
ajst-29358	150	14	ma	ma	PROPN
ajst-29358	150	15	,	,	PUNCT
ajst-29358	150	16	g.	g.	PROPN
ajst-29358	150	17	j.	j.	PROPN
ajst-29358	150	18	,	,	PUNCT
ajst-29358	150	19	mou	mou	PROPN
ajst-29358	150	20	,	,	PUNCT
ajst-29358	150	21	x.	x.	PROPN
ajst-29358	150	22	b.	b.	PROPN
ajst-29358	150	23	,	,	PUNCT
ajst-29358	150	24	zhang	zhang	PROPN
ajst-29358	150	25	,	,	PUNCT
ajst-29358	150	26	k.	k.	PROPN
ajst-29358	150	27	k.	k.	PROPN
ajst-29358	150	28	,	,	PUNCT
ajst-29358	150	29	wu	wu	PROPN
ajst-29358	150	30	,	,	PUNCT
ajst-29358	150	31	x.	x.	PROPN
ajst-29358	150	32	f.	f.	PROPN
ajst-29358	150	33	,	,	PUNCT
ajst-29358	150	34	&	&	CCONJ
ajst-29358	150	35	huang	huang	PROPN
ajst-29358	150	36	,	,	PUNCT
ajst-29358	150	37	x.	x.	PROPN
ajst-29358	150	38	p.	p.	NOUN
ajst-29358	150	39	(	(	PUNCT
ajst-29358	150	40	2024	2024	NUM
ajst-29358	150	41	)	)	PUNCT
ajst-29358	150	42	.	.	PUNCT
ajst-29358	151	1	analysis	analysis	NOUN
ajst-29358	151	2	of	of	ADP
ajst-29358	151	3	the	the	DET
ajst-29358	151	4	impact	impact	NOUN
ajst-29358	151	5	of	of	ADP
ajst-29358	151	6	different	different	ADJ
ajst-29358	151	7	improvement	improvement	NOUN
ajst-29358	151	8	methods	method	NOUN
ajst-29358	151	9	based	base	VERB
ajst-29358	151	10	on	on	ADP
ajst-29358	151	11	yolov8	yolov8	NOUN
ajst-29358	151	12	for	for	ADP
ajst-29358	151	13	weed	weed	NOUN
ajst-29358	151	14	detection	detection	NOUN
ajst-29358	151	15	.	.	PUNCT
ajst-29358	152	1	agriculture	agriculture	NOUN
ajst-29358	152	2	-	-	PUNCT
ajst-29358	152	3	basel	basel	PROPN
ajst-29358	152	4	.	.	PUNCT
ajst-29358	153	1	[	[	X
ajst-29358	153	2	7	7	X
ajst-29358	153	3	]	]	X
ajst-29358	153	4	sun	sun	NOUN
ajst-29358	153	5	,	,	PUNCT
ajst-29358	153	6	d.	d.	PROPN
ajst-29358	153	7	z.	z.	PROPN
ajst-29358	153	8	,	,	PUNCT
ajst-29358	153	9	zhang	zhang	PROPN
ajst-29358	153	10	,	,	PUNCT
ajst-29358	153	11	k.	k.	PROPN
ajst-29358	153	12	,	,	PUNCT
ajst-29358	153	13	zhong	zhong	PROPN
ajst-29358	153	14	,	,	PUNCT
ajst-29358	153	15	h.	h.	PROPN
ajst-29358	153	16	s.	s.	PROPN
ajst-29358	153	17	,	,	PUNCT
ajst-29358	153	18	xie	xie	PROPN
ajst-29358	153	19	,	,	PUNCT
ajst-29358	153	20	j.	j.	PROPN
ajst-29358	153	21	x.	x.	PROPN
ajst-29358	153	22	,	,	PUNCT
ajst-29358	153	23	xue	xue	PROPN
ajst-29358	153	24	,	,	PUNCT
ajst-29358	153	25	x.	x.	NOUN
ajst-29358	153	26	y.	y.	PROPN
ajst-29358	153	27	,	,	PUNCT
ajst-29358	153	28	yan	yan	PROPN
ajst-29358	153	29	,	,	PUNCT
ajst-29358	153	30	m.	m.	PROPN
ajst-29358	153	31	l.	l.	PROPN
ajst-29358	153	32	,	,	PUNCT
ajst-29358	153	33	wu	wu	PROPN
ajst-29358	153	34	,	,	PUNCT
ajst-29358	153	35	w.	w.	PROPN
ajst-29358	153	36	b.	b.	PROPN
ajst-29358	153	37	,	,	PUNCT
ajst-29358	153	38	&	&	CCONJ
ajst-29358	153	39	li	li	PROPN
ajst-29358	153	40	,	,	PUNCT
ajst-29358	153	41	j.	j.	PROPN
ajst-29358	153	42	h.	h.	PROPN
ajst-29358	153	43	(	(	PUNCT
ajst-29358	153	44	2024	2024	NUM
ajst-29358	153	45	)	)	PUNCT
ajst-29358	153	46	.	.	PUNCT
ajst-29358	154	1	efficient	efficient	ADJ
ajst-29358	154	2	tobacco	tobacco	NOUN
ajst-29358	154	3	pest	pest	NOUN
ajst-29358	154	4	detection	detection	NOUN
ajst-29358	154	5	in	in	ADP
ajst-29358	154	6	complex	complex	ADJ
ajst-29358	154	7	environments	environment	NOUN
ajst-29358	154	8	using	use	VERB
ajst-29358	154	9	an	an	DET
ajst-29358	154	10	enhanced	enhanced	ADJ
ajst-29358	154	11	yolov8	yolov8	NOUN
ajst-29358	154	12	model	model	NOUN
ajst-29358	154	13	.	.	PUNCT
ajst-29358	155	1	agriculture	agriculture	NOUN
ajst-29358	155	2	-	-	PUNCT
ajst-29358	155	3	basel	basel	PROPN
ajst-29358	155	4	.	.	PUNCT
ajst-29358	156	1	[	[	X
ajst-29358	156	2	8	8	NUM
ajst-29358	156	3	]	]	X
ajst-29358	156	4	deng	deng	PROPN
ajst-29358	156	5	pengfei	pengfei	PROPN
ajst-29358	156	6	,	,	PUNCT
ajst-29358	156	7	guan	guan	PROPN
ajst-29358	156	8	zheng	zheng	PROPN
ajst-29358	156	9	,	,	PUNCT
ajst-29358	156	10	wang	wang	PROPN
ajst-29358	156	11	yuyang	yuyang	PROPN
ajst-29358	156	12	,	,	PUNCT
ajst-29358	156	13	et	et	PROPN
ajst-29358	156	14	al	al	PROPN
ajst-29358	156	15	.	.	PUNCT
ajst-29358	157	1	a	a	DET
ajst-29358	157	2	method	method	NOUN
ajst-29358	157	3	for	for	ADP
ajst-29358	157	4	corn	corn	NOUN
ajst-29358	157	5	disease	disease	NOUN
ajst-29358	157	6	recognition	recognition	NOUN
ajst-29358	157	7	based	base	VERB
ajst-29358	157	8	on	on	ADP
ajst-29358	157	9	transfer	transfer	NOUN
ajst-29358	157	10	learning	learning	NOUN
ajst-29358	157	11	and	and	CCONJ
ajst-29358	157	12	model	model	NOUN
ajst-29358	157	13	compression[j	compression[j	PROPN
ajst-29358	157	14	]	]	PUNCT
ajst-29358	157	15	.	.	PUNCT
ajst-29358	158	1	computer	computer	NOUN
ajst-29358	158	2	science	science	NOUN
ajst-29358	158	3	,	,	PUNCT
ajst-29358	158	4	2022	2022	NUM
ajst-29358	158	5	,	,	PUNCT
ajst-29358	158	6	49(s2	49(s2	NUM
ajst-29358	158	7	):	):	PUNCT
ajst-29358	158	8	444	444	NUM
ajst-29358	158	9	-	-	SYM
ajst-29358	158	10	449	449	NUM
ajst-29358	158	11	.	.	PUNCT
ajst-29358	159	1	[	[	X
ajst-29358	159	2	9	9	NUM
ajst-29358	159	3	]	]	SYM
ajst-29358	159	4	ye	ye	PROPN
ajst-29358	159	5	,	,	PUNCT
ajst-29358	159	6	r.	r.	PROPN
ajst-29358	159	7	,	,	PUNCT
ajst-29358	159	8	gao	gao	PROPN
ajst-29358	159	9	,	,	PUNCT
ajst-29358	159	10	q.	q.	PROPN
ajst-29358	159	11	,	,	PUNCT
ajst-29358	159	12	qian	qian	PROPN
ajst-29358	159	13	,	,	PUNCT
ajst-29358	159	14	y.	y.	PROPN
ajst-29358	159	15	,	,	PUNCT
ajst-29358	159	16	sun	sun	PROPN
ajst-29358	159	17	,	,	PUNCT
ajst-29358	159	18	j.	j.	PROPN
ajst-29358	159	19	h.	h.	PROPN
ajst-29358	159	20	,	,	PUNCT
ajst-29358	159	21	&	&	CCONJ
ajst-29358	159	22	li	li	PROPN
ajst-29358	159	23	,	,	PUNCT
ajst-29358	159	24	t.	t.	PROPN
ajst-29358	159	25	(	(	PUNCT
ajst-29358	159	26	2024	2024	NUM
ajst-29358	159	27	)	)	PUNCT
ajst-29358	159	28	.	.	PUNCT
ajst-29358	160	1	improved	improve	VERB
ajst-29358	160	2	yolov8	yolov8	NOUN
ajst-29358	160	3	and	and	CCONJ
ajst-29358	160	4	sahi	sahi	NOUN
ajst-29358	160	5	model	model	NOUN
ajst-29358	160	6	for	for	ADP
ajst-29358	160	7	the	the	DET
ajst-29358	160	8	collaborative	collaborative	ADJ
ajst-29358	160	9	detection	detection	NOUN
ajst-29358	160	10	of	of	ADP
ajst-29358	160	11	small	small	ADJ
ajst-29358	160	12	targets	target	NOUN
ajst-29358	160	13	at	at	ADP
ajst-29358	160	14	the	the	DET
ajst-29358	160	15	micro	micro	ADJ
ajst-29358	160	16	scale	scale	NOUN
ajst-29358	160	17	:	:	PUNCT
ajst-29358	160	18	a	a	DET
ajst-29358	160	19	case	case	NOUN
ajst-29358	160	20	study	study	NOUN
ajst-29358	160	21	of	of	ADP
ajst-29358	160	22	pest	pest	NOUN
ajst-29358	160	23	detection	detection	NOUN
ajst-29358	160	24	in	in	ADP
ajst-29358	160	25	tea	tea	NOUN
ajst-29358	160	26	.	.	PUNCT
ajst-29358	161	1	agronomy	agronomy	NOUN
ajst-29358	161	2	-	-	PUNCT
ajst-29358	161	3	basel	basel	NOUN
ajst-29358	161	4	.	.	PUNCT
ajst-29358	162	1	[	[	X
ajst-29358	162	2	10	10	NUM
ajst-29358	162	3	]	]	X
ajst-29358	162	4	liu	liu	PROPN
ajst-29358	162	5	w	w	PROPN
ajst-29358	162	6	j	j	PROPN
ajst-29358	162	7	,	,	PUNCT
ajst-29358	162	8	qiang	qiang	PROPN
ajst-29358	162	9	j	j	PROPN
ajst-29358	162	10	,	,	PUNCT
ajst-29358	162	11	li	li	PROPN
ajst-29358	162	12	x	x	SYM
ajst-29358	162	13	x	x	PROPN
ajst-29358	162	14	,	,	PUNCT
ajst-29358	162	15	et	et	PROPN
ajst-29358	162	16	al	al	PROPN
ajst-29358	162	17	.	.	PUNCT
ajst-29358	162	18	uav	uav	PROPN
ajst-29358	162	19	image	image	PROPN
ajst-29358	162	20	small	small	ADJ
ajst-29358	162	21	object	object	NOUN
ajst-29358	162	22	detection	detection	NOUN
ajst-29358	162	23	based	base	VERB
ajst-29358	162	24	on	on	ADP
ajst-29358	162	25	composite	composite	ADJ
ajst-29358	162	26	backbone	backbone	NOUN
ajst-29358	162	27	network[j	network[j	NOUN
ajst-29358	162	28	]	]	PUNCT
ajst-29358	162	29	.	.	PUNCT
ajst-29358	163	1	mobile	mobile	ADJ
ajst-29358	163	2	information	information	NOUN
ajst-29358	163	3	systems	system	NOUN
ajst-29358	163	4	,	,	PUNCT
ajst-29358	163	5	2022(15):1	2022(15):1	NUM
ajst-29358	163	6	-	-	SYM
ajst-29358	163	7	11	11	NUM
ajst-29358	163	8	.	.	PUNCT
ajst-29358	164	1	[	[	X
ajst-29358	164	2	11	11	NUM
ajst-29358	164	3	]	]	X
ajst-29358	164	4	yu	yu	PROPN
ajst-29358	164	5	,	,	PUNCT
ajst-29358	164	6	x.	x.	PROPN
ajst-29358	164	7	,	,	PUNCT
ajst-29358	164	8	yin	yin	PROPN
ajst-29358	164	9	,	,	PUNCT
ajst-29358	164	10	d.	d.	PROPN
ajst-29358	164	11	m.	m.	PROPN
ajst-29358	164	12	,	,	PUNCT
ajst-29358	164	13	xu	xu	PROPN
ajst-29358	164	14	,	,	PUNCT
ajst-29358	164	15	h.	h.	PROPN
ajst-29358	164	16	g.	g.	PROPN
ajst-29358	164	17	,	,	PUNCT
ajst-29358	164	18	espinosa	espinosa	PROPN
ajst-29358	164	19	,	,	PUNCT
ajst-29358	164	20	f.	f.	PROPN
ajst-29358	164	21	p.	p.	PROPN
ajst-29358	164	22	,	,	PUNCT
ajst-29358	164	23	schmidhalter	schmidhalter	NOUN
ajst-29358	164	24	,	,	PUNCT
ajst-29358	164	25	u.	u.	PROPN
ajst-29358	164	26	,	,	PUNCT
ajst-29358	164	27	nie	nie	PROPN
ajst-29358	164	28	,	,	PUNCT
ajst-29358	164	29	c.	c.	PROPN
ajst-29358	164	30	w.	w.	PROPN
ajst-29358	164	31	,	,	PUNCT
ajst-29358	164	32	bai	bai	PROPN
ajst-29358	164	33	,	,	PUNCT
ajst-29358	164	34	y.	y.	PROPN
ajst-29358	164	35	,	,	PUNCT
ajst-29358	164	36	sankaran	sankaran	VERB
ajst-29358	164	37	,	,	PUNCT
ajst-29358	164	38	s.	s.	PROPN
ajst-29358	164	39	,	,	PUNCT
ajst-29358	164	40	ming	ming	PROPN
ajst-29358	164	41	,	,	PUNCT
ajst-29358	164	42	b.	b.	PROPN
ajst-29358	164	43	,	,	PUNCT
ajst-29358	164	44	cui	cui	PROPN
ajst-29358	164	45	,	,	PUNCT
ajst-29358	164	46	n.	n.	PROPN
ajst-29358	164	47	b.	b.	PROPN
ajst-29358	164	48	,	,	PUNCT
ajst-29358	164	49	wu	wu	PROPN
ajst-29358	164	50	,	,	PUNCT
ajst-29358	164	51	w.	w.	PROPN
ajst-29358	164	52	b.	b.	PROPN
ajst-29358	164	53	,	,	PUNCT
ajst-29358	164	54	&	&	CCONJ
ajst-29358	164	55	jin	jin	PROPN
ajst-29358	164	56	,	,	PUNCT
ajst-29358	164	57	x.	x.	PROPN
ajst-29358	164	58	l.	l.	PROPN
ajst-29358	164	59	(	(	PUNCT
ajst-29358	164	60	2024	2024	NUM
ajst-29358	164	61	)	)	PUNCT
ajst-29358	164	62	.	.	PUNCT
ajst-29358	165	1	maize	maize	NOUN
ajst-29358	165	2	tassel	tassel	NOUN
ajst-29358	165	3	number	number	NOUN
ajst-29358	165	4	and	and	CCONJ
ajst-29358	165	5	tasseling	tasseling	NOUN
ajst-29358	165	6	stage	stage	NOUN
ajst-29358	165	7	monitoring	monitoring	NOUN
ajst-29358	165	8	based	base	VERB
ajst-29358	165	9	on	on	ADP
ajst-29358	165	10	near	near	ADJ
ajst-29358	165	11	-	-	PUNCT
ajst-29358	165	12	ground	ground	NOUN
ajst-29358	165	13	and	and	CCONJ
ajst-29358	165	14	uav	uav	PROPN
ajst-29358	165	15	rgb	rgb	PROPN
ajst-29358	165	16	images	image	NOUN
ajst-29358	165	17	by	by	ADP
ajst-29358	165	18	improved	improved	ADJ
ajst-29358	165	19	yolov8	yolov8	NOUN
ajst-29358	165	20	.	.	PUNCT
ajst-29358	166	1	precision	precision	NOUN
ajst-29358	166	2	agriculture	agriculture	NOUN
ajst-29358	166	3	.	.	PUNCT
ajst-29358	167	1	[	[	X
ajst-29358	167	2	12	12	NUM
ajst-29358	167	3	]	]	X
ajst-29358	167	4	guo	guo	PROPN
ajst-29358	167	5	,	,	PUNCT
ajst-29358	167	6	b.	b.	PROPN
ajst-29358	167	7	l.	l.	PROPN
ajst-29358	167	8	,	,	PUNCT
ajst-29358	167	9	ling	ling	PROPN
ajst-29358	167	10	,	,	PUNCT
ajst-29358	167	11	s.	s.	PROPN
ajst-29358	167	12	k.	k.	PROPN
ajst-29358	167	13	,	,	PUNCT
ajst-29358	167	14	tan	tan	PROPN
ajst-29358	167	15	,	,	PUNCT
ajst-29358	167	16	h.	h.	PROPN
ajst-29358	167	17	y.	y.	PROPN
ajst-29358	167	18	,	,	PUNCT
ajst-29358	167	19	wang	wang	PROPN
ajst-29358	167	20	,	,	PUNCT
ajst-29358	167	21	s.	s.	PROPN
ajst-29358	167	22	,	,	PUNCT
ajst-29358	167	23	wu	wu	PROPN
ajst-29358	167	24	,	,	PUNCT
ajst-29358	167	25	c.	c.	PROPN
ajst-29358	167	26	l.	l.	PROPN
ajst-29358	167	27	,	,	PUNCT
ajst-29358	167	28	&	&	CCONJ
ajst-29358	167	29	yang	yang	PROPN
ajst-29358	167	30	,	,	PUNCT
ajst-29358	167	31	d.	d.	PROPN
ajst-29358	167	32	s.	s.	PROPN
ajst-29358	167	33	(	(	PUNCT
ajst-29358	167	34	2024	2024	NUM
ajst-29358	167	35	)	)	PUNCT
ajst-29358	167	36	.	.	PUNCT
ajst-29358	168	1	detection	detection	NOUN
ajst-29358	168	2	of	of	ADP
ajst-29358	168	3	the	the	DET
ajst-29358	168	4	grassland	grassland	NOUN
ajst-29358	168	5	weed	weed	VERB
ajst-29358	168	6	phlomoides	phlomoide	NOUN
ajst-29358	168	7	umbrosa	umbrosa	VERB
ajst-29358	168	8	using	use	VERB
ajst-29358	168	9	multi	multi	ADJ
ajst-29358	168	10	-	-	ADJ
ajst-29358	168	11	source	source	NOUN
ajst-29358	168	12	imagery	imagery	NOUN
ajst-29358	168	13	and	and	CCONJ
ajst-29358	168	14	an	an	DET
ajst-29358	168	15	improved	improved	ADJ
ajst-29358	168	16	yolov8	yolov8	NOUN
ajst-29358	168	17	network	network	NOUN
ajst-29358	168	18	.	.	PUNCT
ajst-29358	169	1	agronomy	agronomy	NOUN
ajst-29358	169	2	-	-	PUNCT
ajst-29358	169	3	basel	basel	NOUN
ajst-29358	169	4	.	.	PUNCT
ajst-29358	170	1	[	[	X
ajst-29358	170	2	13	13	NUM
ajst-29358	170	3	]	]	X
ajst-29358	170	4	zhu	zhu	PROPN
ajst-29358	170	5	qi	qi	PROPN
ajst-29358	170	6	,	,	PUNCT
ajst-29358	170	7	zhou	zhou	PROPN
ajst-29358	170	8	deqiang	deqiang	PROPN
ajst-29358	170	9	,	,	PUNCT
ajst-29358	170	10	sheng	sheng	PROPN
ajst-29358	170	11	weifeng	weifeng	PROPN
ajst-29358	170	12	,	,	PUNCT
ajst-29358	170	13	et	et	PROPN
ajst-29358	170	14	al	al	PROPN
ajst-29358	170	15	.	.	PUNCT
ajst-29358	171	1	a	a	DET
ajst-29358	171	2	dscs	dscs	ADV
ajst-29358	171	3	-	-	PUNCT
ajst-29358	171	4	yolobased	yolobase	VERB
ajst-29358	171	5	apple	apple	NOUN
ajst-29358	171	6	surface	surface	NOUN
ajst-29358	171	7	defect	defect	NOUN
ajst-29358	171	8	detection	detection	NOUN
ajst-29358	171	9	method	method	NOUN
ajst-29358	171	10	[	[	X
ajst-29358	171	11	j	j	X
ajst-29358	171	12	]	]	X
ajst-29358	171	13	.	.	PUNCT
ajst-29358	172	1	journal	journal	PROPN
ajst-29358	172	2	of	of	ADP
ajst-29358	172	3	nanjing	nanjing	PROPN
ajst-29358	172	4	agricultural	agricultural	PROPN
ajst-29358	172	5	university	university	PROPN
ajst-29358	172	6	,	,	PUNCT
ajst-29358	172	7	2024	2024	NUM
ajst-29358	172	8	,	,	PUNCT
ajst-29358	172	9	47(03	47(03	NUM
ajst-29358	172	10	):	):	PUNCT
ajst-29358	172	11	592	592	NUM
ajst-29358	172	12	-	-	SYM
ajst-29358	172	13	601	601	NUM
ajst-29358	172	14	.	.	PUNCT
ajst-29358	173	1	[	[	X
ajst-29358	173	2	14	14	NUM
ajst-29358	173	3	]	]	X
ajst-29358	173	4	liu	liu	PROPN
ajst-29358	173	5	,	,	PUNCT
ajst-29358	173	6	z.	z.	PROPN
ajst-29358	173	7	f.	f.	PROPN
ajst-29358	173	8	,	,	PUNCT
ajst-29358	173	9	abeyrathna	abeyrathna	PROPN
ajst-29358	173	10	,	,	PUNCT
ajst-29358	173	11	r.	r.	PROPN
ajst-29358	173	12	m.	m.	PROPN
ajst-29358	173	13	r.	r.	PROPN
ajst-29358	173	14	d.	d.	PROPN
ajst-29358	173	15	,	,	PUNCT
ajst-29358	173	16	sampurno	sampurno	PROPN
ajst-29358	173	17	,	,	PUNCT
ajst-29358	173	18	r.	r.	PROPN
ajst-29358	173	19	m.	m.	PROPN
ajst-29358	173	20	,	,	PUNCT
ajst-29358	173	21	nakaguchi	nakaguchi	PROPN
ajst-29358	173	22	,	,	PUNCT
ajst-29358	173	23	v.	v.	ADP
ajst-29358	173	24	m.	m.	NOUN
ajst-29358	173	25	,	,	PUNCT
ajst-29358	173	26	&	&	CCONJ
ajst-29358	173	27	ahamed	ahamed	PROPN
ajst-29358	173	28	,	,	PUNCT
ajst-29358	173	29	t.	t.	PROPN
ajst-29358	173	30	(	(	PUNCT
ajst-29358	173	31	2024	2024	NUM
ajst-29358	173	32	)	)	PUNCT
ajst-29358	173	33	.	.	PUNCT
ajst-29358	174	1	faster	fast	ADV
ajst-29358	174	2	-	-	PUNCT
ajst-29358	174	3	yolo	yolo	PROPN
ajst-29358	174	4	-	-	PUNCT
ajst-29358	174	5	ap	ap	PROPN
ajst-29358	174	6	:	:	PUNCT
ajst-29358	174	7	a	a	DET
ajst-29358	174	8	lightweight	lightweight	ADJ
ajst-29358	174	9	apple	apple	NOUN
ajst-29358	174	10	detection	detection	NOUN
ajst-29358	174	11	algorithm	algorithm	NOUN
ajst-29358	174	12	based	base	VERB
ajst-29358	174	13	on	on	ADP
ajst-29358	174	14	improved	improved	ADJ
ajst-29358	174	15	yolov8	yolov8	NOUN
ajst-29358	174	16	with	with	ADP
ajst-29358	174	17	a	a	DET
ajst-29358	174	18	new	new	ADJ
ajst-29358	174	19	efficient	efficient	ADJ
ajst-29358	174	20	pdwconv	pdwconv	NOUN
ajst-29358	174	21	in	in	ADP
ajst-29358	174	22	orchard	orchard	NOUN
ajst-29358	174	23	.	.	PUNCT
ajst-29358	175	1	computers	computer	NOUN
ajst-29358	175	2	and	and	CCONJ
ajst-29358	175	3	electronics	electronic	NOUN
ajst-29358	175	4	in	in	ADP
ajst-29358	175	5	agriculture	agriculture	NOUN
ajst-29358	175	6	.	.	PUNCT
ajst-29358	176	1	[	[	X
ajst-29358	176	2	15	15	NUM
ajst-29358	176	3	]	]	X
ajst-29358	176	4	wang	wang	PROPN
ajst-29358	176	5	,	,	PUNCT
ajst-29358	176	6	c.	c.	PROPN
ajst-29358	176	7	l.	l.	PROPN
ajst-29358	176	8	,	,	PUNCT
ajst-29358	176	9	wang	wang	PROPN
ajst-29358	176	10	,	,	PUNCT
ajst-29358	176	11	h.	h.	PROPN
ajst-29358	176	12	m.	m.	PROPN
ajst-29358	176	13	,	,	PUNCT
ajst-29358	176	14	han	han	PROPN
ajst-29358	176	15	,	,	PUNCT
ajst-29358	176	16	q.	q.	PROPN
ajst-29358	176	17	y.	y.	PROPN
ajst-29358	176	18	,	,	PUNCT
ajst-29358	176	19	zhang	zhang	PROPN
ajst-29358	176	20	,	,	PUNCT
ajst-29358	176	21	z.	z.	PROPN
ajst-29358	176	22	g.	g.	PROPN
ajst-29358	176	23	,	,	PUNCT
ajst-29358	176	24	kong	kong	PROPN
ajst-29358	176	25	,	,	PUNCT
ajst-29358	176	26	d.	d.	PROPN
ajst-29358	176	27	d.	d.	PROPN
ajst-29358	176	28	,	,	PUNCT
ajst-29358	176	29	&	&	CCONJ
ajst-29358	176	30	zou	zou	PROPN
ajst-29358	176	31	,	,	PUNCT
ajst-29358	176	32	x.	x.	PROPN
ajst-29358	176	33	j.	j.	PROPN
ajst-29358	176	34	(	(	PUNCT
ajst-29358	176	35	2024	2024	NUM
ajst-29358	176	36	)	)	PUNCT
ajst-29358	176	37	.	.	PUNCT
ajst-29358	177	1	strawberry	strawberry	NOUN
ajst-29358	177	2	detection	detection	NOUN
ajst-29358	177	3	and	and	CCONJ
ajst-29358	177	4	ripeness	ripeness	NOUN
ajst-29358	177	5	classification	classification	NOUN
ajst-29358	177	6	using	use	VERB
ajst-29358	177	7	yolov8	yolov8	NOUN
ajst-29358	177	8	+	+	NOUN
ajst-29358	177	9	model	model	NOUN
ajst-29358	177	10	and	and	CCONJ
ajst-29358	177	11	image	image	NOUN
ajst-29358	177	12	processing	processing	NOUN
ajst-29358	177	13	method	method	NOUN
ajst-29358	177	14	.	.	PUNCT
ajst-29358	178	1	agriculture	agriculture	NOUN
ajst-29358	178	2	-	-	PUNCT
ajst-29358	178	3	basel	basel	PROPN
ajst-29358	178	4	.	.	PUNCT
ajst-29358	179	1	[	[	X
ajst-29358	179	2	16	16	NUM
ajst-29358	179	3	]	]	X
ajst-29358	179	4	zhang	zhang	PROPN
ajst-29358	179	5	,	,	PUNCT
ajst-29358	179	6	l.	l.	PROPN
ajst-29358	179	7	j.	j.	PROPN
ajst-29358	179	8	,	,	PUNCT
ajst-29358	179	9	ding	ding	PROPN
ajst-29358	179	10	,	,	PUNCT
ajst-29358	179	11	g.	g.	PROPN
ajst-29358	179	12	c.	c.	PROPN
ajst-29358	179	13	,	,	PUNCT
ajst-29358	179	14	li	li	PROPN
ajst-29358	179	15	,	,	PUNCT
ajst-29358	179	16	c.	c.	PROPN
ajst-29358	179	17	r.	r.	PROPN
ajst-29358	179	18	,	,	PUNCT
ajst-29358	179	19	&	&	CCONJ
ajst-29358	179	20	li	li	PROPN
ajst-29358	179	21	,	,	PUNCT
ajst-29358	179	22	d.	d.	PROPN
ajst-29358	179	23	m.	m.	PROPN
ajst-29358	179	24	(	(	PUNCT
ajst-29358	179	25	2024	2024	NUM
ajst-29358	179	26	)	)	PUNCT
ajst-29358	179	27	.	.	PUNCT
ajst-29358	180	1	dcfyolov8	dcfyolov8	NOUN
ajst-29358	180	2	:	:	PUNCT
ajst-29358	180	3	an	an	DET
ajst-29358	180	4	improved	improved	ADJ
ajst-29358	180	5	algorithm	algorithm	NOUN
ajst-29358	180	6	for	for	ADP
ajst-29358	180	7	aggregating	aggregate	VERB
ajst-29358	180	8	low	low	ADJ
ajst-29358	180	9	-	-	PUNCT
ajst-29358	180	10	level	level	NOUN
ajst-29358	180	11	features	feature	NOUN
ajst-29358	180	12	to	to	PART
ajst-29358	180	13	detect	detect	VERB
ajst-29358	180	14	agricultural	agricultural	ADJ
ajst-29358	180	15	pests	pest	NOUN
ajst-29358	180	16	and	and	CCONJ
ajst-29358	180	17	diseases	disease	NOUN
ajst-29358	180	18	.	.	PUNCT
ajst-29358	181	1	agronomy	agronomy	NOUN
ajst-29358	181	2	-	-	PUNCT
ajst-29358	181	3	basel	basel	NOUN
ajst-29358	181	4	.	.	PUNCT
ajst-29358	182	1	[	[	X
ajst-29358	182	2	17	17	NUM
ajst-29358	182	3	]	]	X
ajst-29358	182	4	zhu	zhu	PROPN
ajst-29358	182	5	,	,	PUNCT
ajst-29358	182	6	r.	r.	PROPN
ajst-29358	182	7	x.	x.	PROPN
ajst-29358	182	8	,	,	PUNCT
ajst-29358	182	9	hao	hao	PROPN
ajst-29358	182	10	,	,	PUNCT
ajst-29358	182	11	f.	f.	PROPN
ajst-29358	182	12	q.	q.	PROPN
ajst-29358	182	13	,	,	PUNCT
ajst-29358	182	14	&	&	CCONJ
ajst-29358	182	15	ma	ma	PROPN
ajst-29358	182	16	,	,	PUNCT
ajst-29358	182	17	d.	d.	PROPN
ajst-29358	182	18	x.	x.	PROPN
ajst-29358	182	19	(	(	PUNCT
ajst-29358	182	20	2024	2024	NUM
ajst-29358	182	21	)	)	PUNCT
ajst-29358	182	22	.	.	PUNCT
ajst-29358	183	1	research	research	NOUN
ajst-29358	183	2	on	on	ADP
ajst-29358	183	3	polygon	polygon	NOUN
ajst-29358	183	4	pest	pest	NOUN
ajst-29358	183	5	-	-	PUNCT
ajst-29358	183	6	infected	infect	VERB
ajst-29358	183	7	leaf	leaf	NOUN
ajst-29358	183	8	region	region	NOUN
ajst-29358	183	9	detection	detection	NOUN
ajst-29358	183	10	based	base	VERB
ajst-29358	183	11	on	on	ADP
ajst-29358	183	12	yolov8	yolov8	PROPN
ajst-29358	183	13	.	.	PUNCT
ajst-29358	184	1	agriculture	agriculture	NOUN
ajst-29358	184	2	-	-	PUNCT
ajst-29358	184	3	basel	basel	PROPN
ajst-29358	184	4	.	.	PUNCT
ajst-29358	185	1	[	[	X
ajst-29358	185	2	18	18	NUM
ajst-29358	185	3	]	]	X
ajst-29358	185	4	ouyang	ouyang	PROPN
ajst-29358	185	5	d	d	PROPN
ajst-29358	185	6	l	l	PROPN
ajst-29358	185	7	,	,	PUNCT
ajst-29358	185	8	he	he	PRON
ajst-29358	185	9	s	s	VERB
ajst-29358	185	10	,	,	PUNCT
ajst-29358	185	11	zhang	zhang	PROPN
ajst-29358	185	12	g	g	PROPN
ajst-29358	185	13	z	z	PROPN
ajst-29358	185	14	,	,	PUNCT
ajst-29358	185	15	et	et	PROPN
ajst-29358	185	16	al	al	PROPN
ajst-29358	185	17	.	.	PROPN
ajst-29358	186	1	efficient	efficient	ADJ
ajst-29358	186	2	multi	multi	ADJ
ajst-29358	186	3	scale	scale	NOUN
ajst-29358	186	4	attention	attention	NOUN
ajst-29358	186	5	module	module	NOUN
ajst-29358	186	6	with	with	ADP
ajst-29358	186	7	cross	cross	ADJ
ajst-29358	186	8	-	-	ADJ
ajst-29358	186	9	spatial	spatial	ADJ
ajst-29358	186	10	learning[c	learning[c	PROPN
ajst-29358	186	11	]	]	PUNCT
ajst-29358	186	12	//	//	NUM
ajst-29358	186	13	proceedings	proceeding	NOUN
ajst-29358	186	14	of	of	ADP
ajst-29358	186	15	the	the	DET
ajst-29358	186	16	ieee	ieee	NOUN
ajst-29358	186	17	international	international	PROPN
ajst-29358	186	18	conference	conference	NOUN
ajst-29358	186	19	on	on	ADP
ajst-29358	186	20	acoustics	acoustics	PROPN
ajst-29358	186	21	vancouver	vancouver	NOUN
ajst-29358	186	22	:	:	PUNCT
ajst-29358	186	23	ieee,2023:12021	ieee,2023:12021	PROPN
ajst-29358	186	24	-	-	PROPN
ajst-29358	186	25	12031	12031	NUM
ajst-29358	186	26	.	.	PUNCT
ajst-29358	187	1	[	[	X
ajst-29358	187	2	19	19	NUM
ajst-29358	187	3	]	]	X
ajst-29358	187	4	tan	tan	PROPN
ajst-29358	187	5	m	m	PROPN
ajst-29358	187	6	,	,	PUNCT
ajst-29358	187	7	pang	pang	NOUN
ajst-29358	187	8	r	r	NOUN
ajst-29358	187	9	,	,	PUNCT
ajst-29358	187	10	le	le	X
ajst-29358	187	11	q	q	PROPN
ajst-29358	187	12	v.	v.	ADP
ajst-29358	187	13	efficientdet	efficientdet	PROPN
ajst-29358	187	14	:	:	PUNCT
ajst-29358	187	15	scalable	scalable	ADJ
ajst-29358	187	16	and	and	CCONJ
ajst-29358	187	17	efficient	efficient	ADJ
ajst-29358	187	18	object	object	NOUN
ajst-29358	187	19	detection[c]//proceedings	detection[c]//proceeding	NOUN
ajst-29358	187	20	of	of	ADP
ajst-29358	187	21	the	the	DET
ajst-29358	187	22	ieee	ieee	NOUN
ajst-29358	187	23	/	/	SYM
ajst-29358	187	24	cvf	cvf	NOUN
ajst-29358	187	25	conference	conference	NOUN
ajst-29358	187	26	on	on	ADP
ajst-29358	187	27	computer	computer	NOUN
ajst-29358	187	28	vision	vision	NOUN
ajst-29358	187	29	and	and	CCONJ
ajst-29358	187	30	pattern	pattern	NOUN
ajst-29358	187	31	recognition	recognition	NOUN
ajst-29358	187	32	.	.	PUNCT
ajst-29358	188	1	2020	2020	NUM
ajst-29358	188	2	:	:	PUNCT
ajst-29358	189	1	1078110790	1078110790	NUM
ajst-29358	189	2	[	[	X
ajst-29358	189	3	20	20	NUM
ajst-29358	189	4	]	]	X
ajst-29358	189	5	tong	tong	PROPN
ajst-29358	189	6	z	z	PROPN
ajst-29358	189	7	,	,	PUNCT
ajst-29358	189	8	chen	chen	PROPN
ajst-29358	189	9	y	y	PROPN
ajst-29358	189	10	,	,	PUNCT
ajst-29358	189	11	xu	xu	PROPN
ajst-29358	190	1	z	z	X
ajst-29358	190	2	,	,	PUNCT
ajst-29358	190	3	et	et	PROPN
ajst-29358	190	4	al	al	PROPN
ajst-29358	190	5	.	.	PUNCT
ajst-29358	191	1	wise	wise	ADJ
ajst-29358	191	2	-	-	PUNCT
ajst-29358	191	3	iou	iou	NOUN
ajst-29358	191	4	:	:	PUNCT
ajst-29358	191	5	bounding	bound	VERB
ajst-29358	191	6	box	box	NOUN
ajst-29358	191	7	regression	regression	NOUN
ajst-29358	191	8	loss	loss	NOUN
ajst-29358	191	9	with	with	ADP
ajst-29358	191	10	dynamic	dynamic	ADJ
ajst-29358	191	11	focusing	focus	VERB
ajst-29358	191	12	mechanism[j	mechanism[j	PROPN
ajst-29358	191	13	]	]	PUNCT
ajst-29358	191	14	.	.	PUNCT
ajst-29358	192	1	arxiv	arxiv	PROPN
ajst-29358	192	2	preprint	preprint	PROPN
ajst-29358	192	3	arxiv:2301.10051	arxiv:2301.10051	PROPN
ajst-29358	192	4	,	,	PUNCT
ajst-29358	192	5	2023	2023	NUM
ajst-29358	192	6	.	.	PUNCT