id	sid	tid	token	lemma	pos
ajst-20799	1	1	academic	academic	ADJ
ajst-20799	1	2	journal	journal	NOUN
ajst-20799	1	3	of	of	ADP
ajst-20799	1	4	science	science	NOUN
ajst-20799	1	5	and	and	CCONJ
ajst-20799	1	6	technology	technology	NOUN
ajst-20799	1	7	issn	issn	NOUN
ajst-20799	1	8	:	:	PUNCT
ajst-20799	1	9	2771	2771	NUM
ajst-20799	1	10	-	-	SYM
ajst-20799	1	11	3032	3032	NUM
ajst-20799	1	12	|	|	NOUN
ajst-20799	1	13	vol	vol	NOUN
ajst-20799	1	14	.	.	PROPN
ajst-20799	2	1	10	10	NUM
ajst-20799	2	2	,	,	PUNCT
ajst-20799	2	3	no	no	INTJ
ajst-20799	2	4	.	.	NOUN
ajst-20799	2	5	3	3	NUM
ajst-20799	2	6	,	,	PUNCT
ajst-20799	2	7	2024	2024	NUM
ajst-20799	2	8	125	125	NUM
ajst-20799	2	9	vehicle	vehicle	NOUN
ajst-20799	2	10	detection	detection	NOUN
ajst-20799	2	11	and	and	CCONJ
ajst-20799	2	12	tracking	tracking	NOUN
ajst-20799	2	13	techniques	technique	NOUN
ajst-20799	2	14	based	base	VERB
ajst-20799	2	15	on	on	ADP
ajst-20799	2	16	deep	deep	ADJ
ajst-20799	2	17	learning	learning	NOUN
ajst-20799	2	18	in	in	ADP
ajst-20799	2	19	road	road	NOUN
ajst-20799	2	20	traffic	traffic	NOUN
ajst-20799	2	21	surveillance	surveillance	NOUN
ajst-20799	2	22	xiaolong	xiaolong	PROPN
ajst-20799	3	1	liu1	liu1	PROPN
ajst-20799	3	2	,	,	PUNCT
ajst-20799	3	3	a	a	PRON
ajst-20799	3	4	,	,	PUNCT
ajst-20799	3	5	nelson	nelson	PROPN
ajst-20799	3	6	c.	c.	PROPN
ajst-20799	3	7	rodelas2	rodelas2	PROPN
ajst-20799	3	8	,	,	PUNCT
ajst-20799	3	9	*	*	PUNCT
ajst-20799	3	10	1graduate	1graduate	NUM
ajst-20799	3	11	school	school	NOUN
ajst-20799	3	12	,	,	PUNCT
ajst-20799	3	13	university	university	NOUN
ajst-20799	3	14	of	of	ADP
ajst-20799	3	15	the	the	DET
ajst-20799	3	16	east	east	PROPN
ajst-20799	3	17	,	,	PUNCT
ajst-20799	3	18	manila	manila	PROPN
ajst-20799	3	19	,	,	PUNCT
ajst-20799	3	20	1008	1008	NUM
ajst-20799	3	21	,	,	PUNCT
ajst-20799	3	22	metro	metro	PROPN
ajst-20799	3	23	manila	manila	PROPN
ajst-20799	3	24	,	,	PUNCT
ajst-20799	3	25	philippines	philippine	NOUN
ajst-20799	3	26	2graduate	2graduate	NUM
ajst-20799	3	27	school	school	NOUN
ajst-20799	3	28	,	,	PUNCT
ajst-20799	3	29	university	university	NOUN
ajst-20799	3	30	of	of	ADP
ajst-20799	3	31	the	the	DET
ajst-20799	3	32	east	east	PROPN
ajst-20799	3	33	,	,	PUNCT
ajst-20799	3	34	manila	manila	PROPN
ajst-20799	3	35	,	,	PUNCT
ajst-20799	3	36	1008	1008	NUM
ajst-20799	3	37	,	,	PUNCT
ajst-20799	3	38	metro	metro	PROPN
ajst-20799	3	39	manila	manila	PROPN
ajst-20799	3	40	,	,	PUNCT
ajst-20799	3	41	philippines	philippines	PROPN
ajst-20799	3	42	aliu.xiaolong@ue.edu.ph	aliu.xiaolong@ue.edu.ph	PROPN
ajst-20799	3	43	,	,	PUNCT
ajst-20799	3	44	*	*	PUNCT
ajst-20799	3	45	corresponding	correspond	VERB
ajst-20799	3	46	author	author	NOUN
ajst-20799	3	47	:	:	PUNCT
ajst-20799	3	48	nelson.rodelas@ue.edu.ph	nelson.rodelas@ue.edu.ph	ADJ
ajst-20799	3	49	abstract	abstract	NOUN
ajst-20799	3	50	:	:	PUNCT
ajst-20799	3	51	this	this	DET
ajst-20799	3	52	article	article	NOUN
ajst-20799	3	53	discusses	discuss	VERB
ajst-20799	3	54	the	the	DET
ajst-20799	3	55	application	application	NOUN
ajst-20799	3	56	of	of	ADP
ajst-20799	3	57	deep	deep	ADJ
ajst-20799	3	58	learning	learning	NOUN
ajst-20799	3	59	in	in	ADP
ajst-20799	3	60	vehicle	vehicle	NOUN
ajst-20799	3	61	detection	detection	NOUN
ajst-20799	3	62	and	and	CCONJ
ajst-20799	3	63	tracking	tracking	NOUN
ajst-20799	3	64	technology	technology	NOUN
ajst-20799	3	65	,	,	PUNCT
ajst-20799	3	66	elaborating	elaborate	VERB
ajst-20799	3	67	on	on	ADP
ajst-20799	3	68	the	the	DET
ajst-20799	3	69	basic	basic	ADJ
ajst-20799	3	70	concepts	concept	NOUN
ajst-20799	3	71	of	of	ADP
ajst-20799	3	72	deep	deep	ADJ
ajst-20799	3	73	learning	learning	NOUN
ajst-20799	3	74	and	and	CCONJ
ajst-20799	3	75	its	its	PRON
ajst-20799	3	76	advantages	advantage	NOUN
ajst-20799	3	77	in	in	ADP
ajst-20799	3	78	vehicle	vehicle	NOUN
ajst-20799	3	79	target	target	NOUN
ajst-20799	3	80	detection	detection	NOUN
ajst-20799	3	81	.	.	PUNCT
ajst-20799	4	1	deep	deep	ADJ
ajst-20799	4	2	learning	learning	NOUN
ajst-20799	4	3	models	model	NOUN
ajst-20799	4	4	such	such	ADJ
ajst-20799	4	5	as	as	ADP
ajst-20799	4	6	convolutional	convolutional	ADJ
ajst-20799	4	7	neural	neural	ADJ
ajst-20799	4	8	networks	network	NOUN
ajst-20799	4	9	(	(	PUNCT
ajst-20799	4	10	cnns	cnns	PROPN
ajst-20799	4	11	)	)	PUNCT
ajst-20799	4	12	overcome	overcome	VERB
ajst-20799	4	13	the	the	DET
ajst-20799	4	14	reliance	reliance	NOUN
ajst-20799	4	15	on	on	ADP
ajst-20799	4	16	manual	manual	ADJ
ajst-20799	4	17	feature	feature	NOUN
ajst-20799	4	18	engineering	engineering	NOUN
ajst-20799	4	19	by	by	ADP
ajst-20799	4	20	automatically	automatically	ADV
ajst-20799	4	21	learning	learn	VERB
ajst-20799	4	22	image	image	NOUN
ajst-20799	4	23	features	feature	NOUN
ajst-20799	4	24	.	.	PUNCT
ajst-20799	5	1	the	the	DET
ajst-20799	5	2	article	article	NOUN
ajst-20799	5	3	focuses	focus	VERB
ajst-20799	5	4	on	on	ADP
ajst-20799	5	5	two	two	NUM
ajst-20799	5	6	deep	deep	ADJ
ajst-20799	5	7	learning	learning	NOUN
ajst-20799	5	8	detection	detection	NOUN
ajst-20799	5	9	frameworks	framework	NOUN
ajst-20799	5	10	,	,	PUNCT
ajst-20799	5	11	faster	fast	ADV
ajst-20799	5	12	r	r	NOUN
ajst-20799	5	13	-	-	PUNCT
ajst-20799	5	14	cnn	cnn	PROPN
ajst-20799	5	15	and	and	CCONJ
ajst-20799	5	16	yolo	yolo	PROPN
ajst-20799	5	17	.	.	PUNCT
ajst-20799	6	1	the	the	DET
ajst-20799	6	2	former	former	ADJ
ajst-20799	6	3	combines	combine	VERB
ajst-20799	6	4	region	region	NOUN
ajst-20799	6	5	proposal	proposal	NOUN
ajst-20799	6	6	networks	network	NOUN
ajst-20799	6	7	with	with	ADP
ajst-20799	6	8	region	region	NOUN
ajst-20799	6	9	classification	classification	NOUN
ajst-20799	6	10	networks	network	NOUN
ajst-20799	6	11	to	to	PART
ajst-20799	6	12	achieve	achieve	VERB
ajst-20799	6	13	end	end	NOUN
ajst-20799	6	14	-	-	PUNCT
ajst-20799	6	15	to	to	ADP
ajst-20799	6	16	-	-	PUNCT
ajst-20799	6	17	end	end	NOUN
ajst-20799	6	18	optimization	optimization	NOUN
ajst-20799	6	19	,	,	PUNCT
ajst-20799	6	20	while	while	SCONJ
ajst-20799	6	21	the	the	DET
ajst-20799	6	22	latter	latter	NOUN
ajst-20799	6	23	transforms	transform	VERB
ajst-20799	6	24	the	the	DET
ajst-20799	6	25	detection	detection	NOUN
ajst-20799	6	26	task	task	NOUN
ajst-20799	6	27	into	into	ADP
ajst-20799	6	28	a	a	DET
ajst-20799	6	29	regression	regression	NOUN
ajst-20799	6	30	problem	problem	NOUN
ajst-20799	6	31	,	,	PUNCT
ajst-20799	6	32	enabling	enable	VERB
ajst-20799	6	33	real	real	ADJ
ajst-20799	6	34	-	-	PUNCT
ajst-20799	6	35	time	time	NOUN
ajst-20799	6	36	detection	detection	NOUN
ajst-20799	6	37	in	in	ADP
ajst-20799	6	38	a	a	DET
ajst-20799	6	39	single	single	ADJ
ajst-20799	6	40	forward	forward	ADV
ajst-20799	6	41	pass	pass	NOUN
ajst-20799	6	42	.	.	PUNCT
ajst-20799	7	1	regarding	regard	VERB
ajst-20799	7	2	vehicle	vehicle	NOUN
ajst-20799	7	3	tracking	tracking	NOUN
ajst-20799	7	4	,	,	PUNCT
ajst-20799	7	5	the	the	DET
ajst-20799	7	6	article	article	NOUN
ajst-20799	7	7	explores	explore	VERB
ajst-20799	7	8	the	the	DET
ajst-20799	7	9	challenges	challenge	NOUN
ajst-20799	7	10	of	of	ADP
ajst-20799	7	11	multi	multi	ADJ
ajst-20799	7	12	-	-	ADJ
ajst-20799	7	13	object	object	ADJ
ajst-20799	7	14	tracking	tracking	NOUN
ajst-20799	7	15	such	such	ADJ
ajst-20799	7	16	as	as	ADP
ajst-20799	7	17	occlusion	occlusion	NOUN
ajst-20799	7	18	,	,	PUNCT
ajst-20799	7	19	cross	cross	NOUN
ajst-20799	7	20	-	-	NOUN
ajst-20799	7	21	movement	movement	NOUN
ajst-20799	7	22	,	,	PUNCT
ajst-20799	7	23	and	and	CCONJ
ajst-20799	7	24	the	the	DET
ajst-20799	7	25	tracking	tracking	NOUN
ajst-20799	7	26	requirements	requirement	NOUN
ajst-20799	7	27	of	of	ADP
ajst-20799	7	28	different	different	ADJ
ajst-20799	7	29	vehicle	vehicle	NOUN
ajst-20799	7	30	types	type	NOUN
ajst-20799	7	31	.	.	PUNCT
ajst-20799	8	1	deep	deep	ADJ
ajst-20799	8	2	learning	learning	NOUN
ajst-20799	8	3	applications	application	NOUN
ajst-20799	8	4	in	in	ADP
ajst-20799	8	5	this	this	DET
ajst-20799	8	6	field	field	NOUN
ajst-20799	8	7	,	,	PUNCT
ajst-20799	8	8	such	such	ADJ
ajst-20799	8	9	as	as	ADP
ajst-20799	8	10	the	the	DET
ajst-20799	8	11	deepsort	deepsort	NOUN
ajst-20799	8	12	and	and	CCONJ
ajst-20799	8	13	tracktor	tracktor	NOUN
ajst-20799	8	14	algorithms	algorithm	NOUN
ajst-20799	8	15	,	,	PUNCT
ajst-20799	8	16	combine	combine	VERB
ajst-20799	8	17	cnns	cnns	PROPN
ajst-20799	8	18	,	,	PUNCT
ajst-20799	8	19	rnns	rnn	NOUN
ajst-20799	8	20	,	,	PUNCT
ajst-20799	8	21	and	and	CCONJ
ajst-20799	8	22	traditional	traditional	ADJ
ajst-20799	8	23	tracking	tracking	NOUN
ajst-20799	8	24	methods	method	NOUN
ajst-20799	8	25	to	to	PART
ajst-20799	8	26	achieve	achieve	VERB
ajst-20799	8	27	feature	feature	NOUN
ajst-20799	8	28	learning	learning	NOUN
ajst-20799	8	29	,	,	PUNCT
ajst-20799	8	30	historical	historical	ADJ
ajst-20799	8	31	state	state	NOUN
ajst-20799	8	32	modeling	modeling	NOUN
ajst-20799	8	33	,	,	PUNCT
ajst-20799	8	34	and	and	CCONJ
ajst-20799	8	35	probabilistic	probabilistic	ADJ
ajst-20799	8	36	reasoning	reasoning	NOUN
ajst-20799	8	37	.	.	PUNCT
ajst-20799	9	1	performance	performance	NOUN
ajst-20799	9	2	evaluation	evaluation	NOUN
ajst-20799	9	3	is	be	AUX
ajst-20799	9	4	discussed	discuss	VERB
ajst-20799	9	5	in	in	ADP
ajst-20799	9	6	terms	term	NOUN
ajst-20799	9	7	of	of	ADP
ajst-20799	9	8	metrics	metric	NOUN
ajst-20799	9	9	like	like	ADP
ajst-20799	9	10	iou	iou	NOUN
ajst-20799	9	11	,	,	PUNCT
ajst-20799	9	12	precision	precision	NOUN
ajst-20799	9	13	,	,	PUNCT
ajst-20799	9	14	recall	recall	NOUN
ajst-20799	9	15	,	,	PUNCT
ajst-20799	9	16	and	and	CCONJ
ajst-20799	9	17	f1	f1	PROPN
ajst-20799	9	18	score	score	NOUN
ajst-20799	9	19	,	,	PUNCT
ajst-20799	9	20	comparing	compare	VERB
ajst-20799	9	21	and	and	CCONJ
ajst-20799	9	22	analyzing	analyze	VERB
ajst-20799	9	23	the	the	DET
ajst-20799	9	24	performance	performance	NOUN
ajst-20799	9	25	of	of	ADP
ajst-20799	9	26	different	different	ADJ
ajst-20799	9	27	algorithms	algorithm	NOUN
ajst-20799	9	28	in	in	ADP
ajst-20799	9	29	vehicle	vehicle	NOUN
ajst-20799	9	30	detection	detection	NOUN
ajst-20799	9	31	and	and	CCONJ
ajst-20799	9	32	tracking	tracking	NOUN
ajst-20799	9	33	tasks	task	NOUN
ajst-20799	9	34	.	.	PUNCT
ajst-20799	10	1	lastly	lastly	ADV
ajst-20799	10	2	,	,	PUNCT
ajst-20799	10	3	the	the	DET
ajst-20799	10	4	article	article	NOUN
ajst-20799	10	5	discusses	discuss	VERB
ajst-20799	10	6	the	the	DET
ajst-20799	10	7	balance	balance	NOUN
ajst-20799	10	8	between	between	ADP
ajst-20799	10	9	real	real	ADJ
ajst-20799	10	10	-	-	PUNCT
ajst-20799	10	11	time	time	NOUN
ajst-20799	10	12	and	and	CCONJ
ajst-20799	10	13	accuracy	accuracy	NOUN
ajst-20799	10	14	in	in	ADP
ajst-20799	10	15	deep	deep	ADJ
ajst-20799	10	16	learning	learning	NOUN
ajst-20799	10	17	-	-	PUNCT
ajst-20799	10	18	based	base	VERB
ajst-20799	10	19	vehicle	vehicle	NOUN
ajst-20799	10	20	detection	detection	NOUN
ajst-20799	10	21	and	and	CCONJ
ajst-20799	10	22	tracking	tracking	NOUN
ajst-20799	10	23	technology	technology	NOUN
ajst-20799	10	24	in	in	ADP
ajst-20799	10	25	road	road	NOUN
ajst-20799	10	26	traffic	traffic	NOUN
ajst-20799	10	27	monitoring	monitoring	NOUN
ajst-20799	10	28	,	,	PUNCT
ajst-20799	10	29	as	as	ADV
ajst-20799	10	30	well	well	ADV
ajst-20799	10	31	as	as	ADP
ajst-20799	10	32	its	its	PRON
ajst-20799	10	33	significant	significant	ADJ
ajst-20799	10	34	role	role	NOUN
ajst-20799	10	35	in	in	ADP
ajst-20799	10	36	traffic	traffic	NOUN
ajst-20799	10	37	accident	accident	NOUN
ajst-20799	10	38	warning	warning	NOUN
ajst-20799	10	39	and	and	CCONJ
ajst-20799	10	40	management	management	NOUN
ajst-20799	10	41	.	.	PUNCT
ajst-20799	11	1	keywords	keyword	NOUN
ajst-20799	11	2	:	:	PUNCT
ajst-20799	11	3	deep	deep	ADJ
ajst-20799	11	4	learning	learning	NOUN
ajst-20799	11	5	;	;	PUNCT
ajst-20799	11	6	tracktor	tracktor	NOUN
ajst-20799	11	7	algorithm	algorithm	NOUN
ajst-20799	11	8	;	;	PUNCT
ajst-20799	11	9	vehicle	vehicle	NOUN
ajst-20799	11	10	detection	detection	NOUN
ajst-20799	11	11	.	.	PUNCT
ajst-20799	12	1	1	1	X
ajst-20799	12	2	.	.	X
ajst-20799	12	3	introduction	introduction	NOUN
ajst-20799	12	4	with	with	ADP
ajst-20799	12	5	the	the	DET
ajst-20799	12	6	rapid	rapid	ADJ
ajst-20799	12	7	growth	growth	NOUN
ajst-20799	12	8	of	of	ADP
ajst-20799	12	9	urban	urban	ADJ
ajst-20799	12	10	traffic	traffic	NOUN
ajst-20799	12	11	,	,	PUNCT
ajst-20799	12	12	traditional	traditional	ADJ
ajst-20799	12	13	vehicle	vehicle	NOUN
ajst-20799	12	14	detection	detection	NOUN
ajst-20799	12	15	and	and	CCONJ
ajst-20799	12	16	tracking	tracking	NOUN
ajst-20799	12	17	technology	technology	NOUN
ajst-20799	12	18	face	face	VERB
ajst-20799	12	19	challenges	challenge	NOUN
ajst-20799	12	20	in	in	ADP
ajst-20799	12	21	dealing	deal	VERB
ajst-20799	12	22	with	with	ADP
ajst-20799	12	23	large	large	ADJ
ajst-20799	12	24	-	-	PUNCT
ajst-20799	12	25	scale	scale	NOUN
ajst-20799	12	26	data	datum	NOUN
ajst-20799	12	27	and	and	CCONJ
ajst-20799	12	28	complex	complex	ADJ
ajst-20799	12	29	scenarios	scenario	NOUN
ajst-20799	12	30	.	.	PUNCT
ajst-20799	13	1	based	base	VERB
ajst-20799	13	2	on	on	ADP
ajst-20799	13	3	the	the	DET
ajst-20799	13	4	advantages	advantage	NOUN
ajst-20799	13	5	of	of	ADP
ajst-20799	13	6	deep	deep	ADJ
ajst-20799	13	7	learning	learning	NOUN
ajst-20799	13	8	in	in	ADP
ajst-20799	13	9	feature	feature	NOUN
ajst-20799	13	10	learning	learning	NOUN
ajst-20799	13	11	and	and	CCONJ
ajst-20799	13	12	pattern	pattern	NOUN
ajst-20799	13	13	recognition	recognition	NOUN
ajst-20799	13	14	,	,	PUNCT
ajst-20799	13	15	deep	deep	ADJ
ajst-20799	13	16	learning	learning	NOUN
ajst-20799	13	17	-	-	PUNCT
ajst-20799	13	18	based	base	VERB
ajst-20799	13	19	vehicle	vehicle	NOUN
ajst-20799	13	20	detection	detection	NOUN
ajst-20799	13	21	and	and	CCONJ
ajst-20799	13	22	tracking	tracking	NOUN
ajst-20799	13	23	technology	technology	NOUN
ajst-20799	13	24	,	,	PUNCT
ajst-20799	13	25	along	along	ADP
ajst-20799	13	26	with	with	ADP
ajst-20799	13	27	trajectory	trajectory	NOUN
ajst-20799	13	28	prediction	prediction	NOUN
ajst-20799	13	29	methods	method	NOUN
ajst-20799	13	30	based	base	VERB
ajst-20799	13	31	on	on	ADP
ajst-20799	13	32	deep	deep	ADJ
ajst-20799	13	33	learning	learning	NOUN
ajst-20799	13	34	,	,	PUNCT
ajst-20799	13	35	have	have	AUX
ajst-20799	13	36	shown	show	VERB
ajst-20799	13	37	outstanding	outstanding	ADJ
ajst-20799	13	38	performance	performance	NOUN
ajst-20799	13	39	in	in	ADP
ajst-20799	13	40	long	long	ADJ
ajst-20799	13	41	-	-	PUNCT
ajst-20799	13	42	term	term	NOUN
ajst-20799	13	43	,	,	PUNCT
ajst-20799	13	44	multimodal	multimodal	NOUN
ajst-20799	13	45	motion	motion	NOUN
ajst-20799	13	46	,	,	PUNCT
ajst-20799	13	47	and	and	CCONJ
ajst-20799	13	48	vehicle	vehicle	NOUN
ajst-20799	13	49	-	-	PUNCT
ajst-20799	13	50	road	road	NOUN
ajst-20799	13	51	interaction	interaction	NOUN
ajst-20799	13	52	scenarios	scenario	NOUN
ajst-20799	13	53	.	.	PUNCT
ajst-20799	14	1	[	[	X
ajst-20799	14	2	1	1	X
ajst-20799	14	3	]	]	PUNCT
ajst-20799	14	4	this	this	PRON
ajst-20799	14	5	provides	provide	VERB
ajst-20799	14	6	new	new	ADJ
ajst-20799	14	7	solutions	solution	NOUN
ajst-20799	14	8	for	for	ADP
ajst-20799	14	9	vehicle	vehicle	NOUN
ajst-20799	14	10	detection	detection	NOUN
ajst-20799	14	11	and	and	CCONJ
ajst-20799	14	12	tracking	tracking	NOUN
ajst-20799	14	13	,	,	PUNCT
ajst-20799	14	14	and	and	CCONJ
ajst-20799	14	15	this	this	DET
ajst-20799	14	16	paper	paper	NOUN
ajst-20799	14	17	aims	aim	VERB
ajst-20799	14	18	to	to	PART
ajst-20799	14	19	explore	explore	VERB
ajst-20799	14	20	the	the	DET
ajst-20799	14	21	principles	principle	NOUN
ajst-20799	14	22	,	,	PUNCT
ajst-20799	14	23	methods	method	NOUN
ajst-20799	14	24	,	,	PUNCT
ajst-20799	14	25	and	and	CCONJ
ajst-20799	14	26	application	application	NOUN
ajst-20799	14	27	prospects	prospect	NOUN
ajst-20799	14	28	of	of	ADP
ajst-20799	14	29	this	this	DET
ajst-20799	14	30	technology	technology	NOUN
ajst-20799	14	31	.	.	PUNCT
ajst-20799	15	1	2	2	X
ajst-20799	15	2	.	.	X
ajst-20799	15	3	application	application	NOUN
ajst-20799	15	4	of	of	ADP
ajst-20799	15	5	deep	deep	ADJ
ajst-20799	15	6	learning	learning	NOUN
ajst-20799	15	7	in	in	ADP
ajst-20799	15	8	vehicle	vehicle	NOUN
ajst-20799	15	9	detection	detection	NOUN
ajst-20799	15	10	2.1	2.1	NUM
ajst-20799	15	11	.	.	PUNCT
ajst-20799	15	12	basic	basic	ADJ
ajst-20799	15	13	concepts	concept	NOUN
ajst-20799	15	14	of	of	ADP
ajst-20799	15	15	deep	deep	ADJ
ajst-20799	15	16	learning	learning	NOUN
ajst-20799	15	17	in	in	ADP
ajst-20799	15	18	deep	deep	ADJ
ajst-20799	15	19	learning	learning	NOUN
ajst-20799	15	20	,	,	PUNCT
ajst-20799	15	21	neural	neural	ADJ
ajst-20799	15	22	networks	network	NOUN
ajst-20799	15	23	are	be	AUX
ajst-20799	15	24	computational	computational	ADJ
ajst-20799	15	25	models	model	NOUN
ajst-20799	15	26	that	that	PRON
ajst-20799	15	27	mimic	mimic	VERB
ajst-20799	15	28	the	the	DET
ajst-20799	15	29	structure	structure	NOUN
ajst-20799	15	30	and	and	CCONJ
ajst-20799	15	31	function	function	NOUN
ajst-20799	15	32	of	of	ADP
ajst-20799	15	33	the	the	DET
ajst-20799	15	34	biological	biological	ADJ
ajst-20799	15	35	nervous	nervous	ADJ
ajst-20799	15	36	system	system	NOUN
ajst-20799	15	37	.	.	PUNCT
ajst-20799	16	1	they	they	PRON
ajst-20799	16	2	consist	consist	VERB
ajst-20799	16	3	of	of	ADP
ajst-20799	16	4	a	a	DET
ajst-20799	16	5	large	large	ADJ
ajst-20799	16	6	number	number	NOUN
ajst-20799	16	7	of	of	ADP
ajst-20799	16	8	artificial	artificial	ADJ
ajst-20799	16	9	neurons	neuron	NOUN
ajst-20799	16	10	interconnected	interconnect	VERB
ajst-20799	16	11	,	,	PUNCT
ajst-20799	16	12	forming	form	VERB
ajst-20799	16	13	a	a	DET
ajst-20799	16	14	network	network	NOUN
ajst-20799	16	15	structure	structure	NOUN
ajst-20799	16	16	where	where	SCONJ
ajst-20799	16	17	information	information	NOUN
ajst-20799	16	18	is	be	AUX
ajst-20799	16	19	passed	pass	VERB
ajst-20799	16	20	layer	layer	NOUN
ajst-20799	16	21	by	by	ADP
ajst-20799	16	22	layer	layer	NOUN
ajst-20799	16	23	to	to	PART
ajst-20799	16	24	achieve	achieve	VERB
ajst-20799	16	25	learning	learning	NOUN
ajst-20799	16	26	and	and	CCONJ
ajst-20799	16	27	decision	decision	NOUN
ajst-20799	16	28	-	-	PUNCT
ajst-20799	16	29	making	making	NOUN
ajst-20799	16	30	.	.	PUNCT
ajst-20799	17	1	the	the	DET
ajst-20799	17	2	basic	basic	ADJ
ajst-20799	17	3	unit	unit	NOUN
ajst-20799	17	4	of	of	ADP
ajst-20799	17	5	a	a	DET
ajst-20799	17	6	neural	neural	ADJ
ajst-20799	17	7	network	network	NOUN
ajst-20799	17	8	simulates	simulate	VERB
ajst-20799	17	9	some	some	DET
ajst-20799	17	10	functions	function	NOUN
ajst-20799	17	11	of	of	ADP
ajst-20799	17	12	biological	biological	ADJ
ajst-20799	17	13	neurons	neuron	NOUN
ajst-20799	17	14	;	;	PUNCT
ajst-20799	17	15	each	each	DET
ajst-20799	17	16	neuron	neuron	NOUN
ajst-20799	17	17	receives	receive	VERB
ajst-20799	17	18	a	a	DET
ajst-20799	17	19	set	set	NOUN
ajst-20799	17	20	of	of	ADP
ajst-20799	17	21	input	input	NOUN
ajst-20799	17	22	signals	signal	NOUN
ajst-20799	17	23	(	(	PUNCT
ajst-20799	17	24	from	from	ADP
ajst-20799	17	25	the	the	DET
ajst-20799	17	26	previous	previous	ADJ
ajst-20799	17	27	layer	layer	NOUN
ajst-20799	17	28	's	's	PART
ajst-20799	17	29	neurons	neuron	NOUN
ajst-20799	17	30	'	'	PART
ajst-20799	17	31	outputs	output	NOUN
ajst-20799	17	32	or	or	CCONJ
ajst-20799	17	33	directly	directly	ADV
ajst-20799	17	34	from	from	ADP
ajst-20799	17	35	input	input	NOUN
ajst-20799	17	36	data	datum	NOUN
ajst-20799	17	37	)	)	PUNCT
ajst-20799	17	38	,	,	PUNCT
ajst-20799	17	39	processes	process	VERB
ajst-20799	17	40	them	they	PRON
ajst-20799	17	41	through	through	ADP
ajst-20799	17	42	weighted	weight	VERB
ajst-20799	17	43	summation	summation	NOUN
ajst-20799	17	44	followed	follow	VERB
ajst-20799	17	45	by	by	ADP
ajst-20799	17	46	a	a	DET
ajst-20799	17	47	nonlinear	nonlinear	ADJ
ajst-20799	17	48	activation	activation	NOUN
ajst-20799	17	49	function	function	NOUN
ajst-20799	17	50	(	(	PUNCT
ajst-20799	17	51	such	such	ADJ
ajst-20799	17	52	as	as	ADP
ajst-20799	17	53	relu	relu	NOUN
ajst-20799	17	54	,	,	PUNCT
ajst-20799	17	55	sigmoid	sigmoid	NOUN
ajst-20799	17	56	,	,	PUNCT
ajst-20799	17	57	tanh	tanh	NOUN
ajst-20799	17	58	,	,	PUNCT
ajst-20799	17	59	etc	etc	X
ajst-20799	17	60	.	.	X
ajst-20799	17	61	)	)	PUNCT
ajst-20799	17	62	to	to	PART
ajst-20799	17	63	produce	produce	VERB
ajst-20799	17	64	a	a	DET
ajst-20799	17	65	single	single	ADJ
ajst-20799	17	66	output	output	NOUN
ajst-20799	17	67	signal	signal	NOUN
ajst-20799	17	68	.	.	PUNCT
ajst-20799	18	1	in	in	ADP
ajst-20799	18	2	cnns	cnn	NOUN
ajst-20799	18	3	for	for	ADP
ajst-20799	18	4	vehicle	vehicle	NOUN
ajst-20799	18	5	detection	detection	NOUN
ajst-20799	18	6	,	,	PUNCT
ajst-20799	18	7	a	a	DET
ajst-20799	18	8	neuron	neuron	NOUN
ajst-20799	18	9	may	may	AUX
ajst-20799	18	10	correspond	correspond	VERB
ajst-20799	18	11	to	to	ADP
ajst-20799	18	12	features	feature	NOUN
ajst-20799	18	13	of	of	ADP
ajst-20799	18	14	a	a	DET
ajst-20799	18	15	small	small	ADJ
ajst-20799	18	16	pixel	pixel	NOUN
ajst-20799	18	17	area	area	NOUN
ajst-20799	18	18	in	in	ADP
ajst-20799	18	19	an	an	DET
ajst-20799	18	20	image	image	NOUN
ajst-20799	18	21	.	.	PUNCT
ajst-20799	19	1	typically	typically	ADV
ajst-20799	19	2	,	,	PUNCT
ajst-20799	19	3	several	several	ADJ
ajst-20799	19	4	convolutional	convolutional	ADJ
ajst-20799	19	5	layers	layer	NOUN
ajst-20799	19	6	are	be	AUX
ajst-20799	19	7	used	use	VERB
ajst-20799	19	8	to	to	PART
ajst-20799	19	9	extract	extract	VERB
ajst-20799	19	10	image	image	NOUN
ajst-20799	19	11	features	feature	NOUN
ajst-20799	19	12	,	,	PUNCT
ajst-20799	19	13	followed	follow	VERB
ajst-20799	19	14	by	by	ADP
ajst-20799	19	15	pooling	pool	VERB
ajst-20799	19	16	layers	layer	NOUN
ajst-20799	19	17	to	to	PART
ajst-20799	19	18	reduce	reduce	VERB
ajst-20799	19	19	dimensionality	dimensionality	NOUN
ajst-20799	19	20	,	,	PUNCT
ajst-20799	19	21	and	and	CCONJ
ajst-20799	19	22	finally	finally	ADV
ajst-20799	19	23	connected	connect	VERB
ajst-20799	19	24	to	to	ADP
ajst-20799	19	25	fully	fully	ADV
ajst-20799	19	26	connected	connected	ADJ
ajst-20799	19	27	layers	layer	NOUN
ajst-20799	19	28	or	or	CCONJ
ajst-20799	19	29	specific	specific	ADJ
ajst-20799	19	30	detection	detection	NOUN
ajst-20799	19	31	heads	head	NOUN
ajst-20799	19	32	(	(	PUNCT
ajst-20799	19	33	like	like	ADP
ajst-20799	19	34	yolo	yolo	PROPN
ajst-20799	19	35	's	's	PART
ajst-20799	19	36	fully	fully	ADV
ajst-20799	19	37	convolutional	convolutional	ADJ
ajst-20799	19	38	structure	structure	NOUN
ajst-20799	19	39	)	)	PUNCT
ajst-20799	19	40	for	for	ADP
ajst-20799	19	41	predicting	predict	VERB
ajst-20799	19	42	vehicle	vehicle	NOUN
ajst-20799	19	43	categories	category	NOUN
ajst-20799	19	44	and	and	CCONJ
ajst-20799	19	45	positions	position	NOUN
ajst-20799	19	46	.	.	PUNCT
ajst-20799	20	1	typical	typical	ADJ
ajst-20799	20	2	neural	neural	ADJ
ajst-20799	20	3	network	network	NOUN
ajst-20799	20	4	structures	structure	NOUN
ajst-20799	20	5	include	include	VERB
ajst-20799	20	6	convolutional	convolutional	ADJ
ajst-20799	20	7	neural	neural	ADJ
ajst-20799	20	8	networks	network	NOUN
ajst-20799	20	9	(	(	PUNCT
ajst-20799	20	10	cnns	cnns	PROPN
ajst-20799	20	11	)	)	PUNCT
ajst-20799	20	12	,	,	PUNCT
ajst-20799	20	13	recurrent	recurrent	ADJ
ajst-20799	20	14	neural	neural	ADJ
ajst-20799	20	15	networks	network	NOUN
ajst-20799	20	16	(	(	PUNCT
ajst-20799	20	17	rnns	rnns	PROPN
ajst-20799	20	18	)	)	PUNCT
ajst-20799	20	19	,	,	PUNCT
ajst-20799	20	20	and	and	CCONJ
ajst-20799	20	21	their	their	PRON
ajst-20799	20	22	variants	variant	NOUN
ajst-20799	20	23	or	or	CCONJ
ajst-20799	20	24	combinations	combination	NOUN
ajst-20799	20	25	.	.	PUNCT
ajst-20799	21	1	the	the	DET
ajst-20799	21	2	input	input	NOUN
ajst-20799	21	3	layer	layer	NOUN
ajst-20799	21	4	receives	receive	VERB
ajst-20799	21	5	raw	raw	ADJ
ajst-20799	21	6	data	datum	NOUN
ajst-20799	21	7	and	and	CCONJ
ajst-20799	21	8	transforms	transform	VERB
ajst-20799	21	9	it	it	PRON
ajst-20799	21	10	into	into	ADP
ajst-20799	21	11	a	a	DET
ajst-20799	21	12	format	format	NOUN
ajst-20799	21	13	understandable	understandable	ADJ
ajst-20799	21	14	by	by	ADP
ajst-20799	21	15	the	the	DET
ajst-20799	21	16	network	network	NOUN
ajst-20799	21	17	for	for	ADP
ajst-20799	21	18	subsequent	subsequent	ADJ
ajst-20799	21	19	processing	processing	NOUN
ajst-20799	21	20	.	.	PUNCT
ajst-20799	22	1	for	for	ADP
ajst-20799	22	2	cnns	cnn	NOUN
ajst-20799	22	3	,	,	PUNCT
ajst-20799	22	4	the	the	DET
ajst-20799	22	5	input	input	NOUN
ajst-20799	22	6	layer	layer	NOUN
ajst-20799	22	7	usually	usually	ADV
ajst-20799	22	8	takes	take	VERB
ajst-20799	22	9	image	image	NOUN
ajst-20799	22	10	data	datum	NOUN
ajst-20799	22	11	,	,	PUNCT
ajst-20799	22	12	such	such	ADJ
ajst-20799	22	13	as	as	ADP
ajst-20799	22	14	an	an	DET
ajst-20799	22	15	rgb	rgb	PROPN
ajst-20799	22	16	color	color	NOUN
ajst-20799	22	17	image	image	NOUN
ajst-20799	22	18	,	,	PUNCT
ajst-20799	22	19	representing	represent	VERB
ajst-20799	22	20	it	it	PRON
ajst-20799	22	21	as	as	ADP
ajst-20799	22	22	a	a	DET
ajst-20799	22	23	three	three	NUM
ajst-20799	22	24	-	-	PUNCT
ajst-20799	22	25	dimensional	dimensional	ADJ
ajst-20799	22	26	tensor	tensor	NOUN
ajst-20799	22	27	(	(	PUNCT
ajst-20799	22	28	height	height	NOUN
ajst-20799	22	29	,	,	PUNCT
ajst-20799	22	30	width	width	ADJ
ajst-20799	22	31	,	,	PUNCT
ajst-20799	22	32	number	number	NOUN
ajst-20799	22	33	of	of	ADP
ajst-20799	22	34	channels	channel	NOUN
ajst-20799	22	35	)	)	PUNCT
ajst-20799	22	36	,	,	PUNCT
ajst-20799	22	37	where	where	SCONJ
ajst-20799	22	38	each	each	DET
ajst-20799	22	39	pixel	pixel	PROPN
ajst-20799	22	40	's	's	PART
ajst-20799	22	41	value	value	NOUN
ajst-20799	22	42	corresponds	correspond	VERB
ajst-20799	22	43	to	to	ADP
ajst-20799	22	44	its	its	PRON
ajst-20799	22	45	red	red	ADJ
ajst-20799	22	46	,	,	PUNCT
ajst-20799	22	47	green	green	ADJ
ajst-20799	22	48	,	,	PUNCT
ajst-20799	22	49	blue	blue	ADJ
ajst-20799	22	50	channel	channel	NOUN
ajst-20799	22	51	intensity	intensity	NOUN
ajst-20799	22	52	.	.	PUNCT
ajst-20799	23	1	in	in	ADP
ajst-20799	23	2	rnns	rnns	PROPN
ajst-20799	23	3	,	,	PUNCT
ajst-20799	23	4	the	the	DET
ajst-20799	23	5	hidden	hide	VERB
ajst-20799	23	6	layer	layer	NOUN
ajst-20799	23	7	contains	contain	VERB
ajst-20799	23	8	recurrent	recurrent	ADJ
ajst-20799	23	9	units	unit	NOUN
ajst-20799	23	10	(	(	PUNCT
ajst-20799	23	11	like	like	ADP
ajst-20799	23	12	lstm	lstm	NOUN
ajst-20799	23	13	or	or	CCONJ
ajst-20799	23	14	gru	gru	NOUN
ajst-20799	23	15	)	)	PUNCT
ajst-20799	23	16	,	,	PUNCT
ajst-20799	23	17	which	which	PRON
ajst-20799	23	18	extract	extract	VERB
ajst-20799	23	19	features	feature	NOUN
ajst-20799	23	20	from	from	ADP
ajst-20799	23	21	data	datum	NOUN
ajst-20799	23	22	through	through	ADP
ajst-20799	23	23	multiple	multiple	ADJ
ajst-20799	23	24	computations	computation	NOUN
ajst-20799	23	25	and	and	CCONJ
ajst-20799	23	26	learning	learning	NOUN
ajst-20799	23	27	,	,	PUNCT
ajst-20799	23	28	with	with	ADP
ajst-20799	23	29	the	the	DET
ajst-20799	23	30	output	output	NOUN
ajst-20799	23	31	layer	layer	NOUN
ajst-20799	23	32	performing	perform	VERB
ajst-20799	23	33	classification	classification	NOUN
ajst-20799	23	34	or	or	CCONJ
ajst-20799	23	35	regression	regression	NOUN
ajst-20799	23	36	predictions	prediction	NOUN
ajst-20799	23	37	,	,	PUNCT
ajst-20799	23	38	capturing	capture	VERB
ajst-20799	23	39	long	long	ADJ
ajst-20799	23	40	-	-	PUNCT
ajst-20799	23	41	term	term	NOUN
ajst-20799	23	42	dependencies	dependency	NOUN
ajst-20799	23	43	in	in	ADP
ajst-20799	23	44	time	time	NOUN
ajst-20799	23	45	series	series	PROPN
ajst-20799	23	46	data	data	PROPN
ajst-20799	23	47	.	.	PUNCT
ajst-20799	24	1	the	the	DET
ajst-20799	24	2	working	work	VERB
ajst-20799	24	3	principle	principle	NOUN
ajst-20799	24	4	of	of	ADP
ajst-20799	24	5	neural	neural	ADJ
ajst-20799	24	6	networks	network	NOUN
ajst-20799	24	7	is	be	AUX
ajst-20799	24	8	based	base	VERB
ajst-20799	24	9	on	on	ADP
ajst-20799	24	10	the	the	DET
ajst-20799	24	11	backpropagation	backpropagation	NOUN
ajst-20799	24	12	algorithm	algorithm	NOUN
ajst-20799	24	13	,	,	PUNCT
ajst-20799	24	14	a	a	DET
ajst-20799	24	15	gradient	gradient	NOUN
ajst-20799	24	16	-	-	PUNCT
ajst-20799	24	17	based	base	VERB
ajst-20799	24	18	optimization	optimization	NOUN
ajst-20799	24	19	method	method	NOUN
ajst-20799	24	20	used	use	VERB
ajst-20799	24	21	to	to	PART
ajst-20799	24	22	compute	compute	VERB
ajst-20799	24	23	gradients	gradient	NOUN
ajst-20799	24	24	of	of	ADP
ajst-20799	24	25	all	all	DET
ajst-20799	24	26	learnable	learnable	ADJ
ajst-20799	24	27	parameters	parameter	NOUN
ajst-20799	24	28	in	in	ADP
ajst-20799	24	29	the	the	DET
ajst-20799	24	30	neural	neural	ADJ
ajst-20799	24	31	network	network	NOUN
ajst-20799	24	32	,	,	PUNCT
ajst-20799	24	33	indicating	indicate	VERB
ajst-20799	24	34	how	how	SCONJ
ajst-20799	24	35	small	small	ADJ
ajst-20799	24	36	changes	change	NOUN
ajst-20799	24	37	in	in	ADP
ajst-20799	24	38	parameters	parameter	NOUN
ajst-20799	24	39	affect	affect	VERB
ajst-20799	24	40	the	the	DET
ajst-20799	24	41	loss	loss	NOUN
ajst-20799	24	42	function	function	NOUN
ajst-20799	24	43	.	.	PUNCT
ajst-20799	25	1	through	through	ADP
ajst-20799	25	2	training	training	NOUN
ajst-20799	25	3	data	datum	NOUN
ajst-20799	25	4	,	,	PUNCT
ajst-20799	25	5	network	network	NOUN
ajst-20799	25	6	parameters	parameter	NOUN
ajst-20799	25	7	are	be	AUX
ajst-20799	25	8	adjusted	adjust	VERB
ajst-20799	25	9	continuously	continuously	ADV
ajst-20799	25	10	to	to	PART
ajst-20799	25	11	minimize	minimize	VERB
ajst-20799	25	12	the	the	DET
ajst-20799	25	13	error	error	NOUN
ajst-20799	25	14	between	between	ADP
ajst-20799	25	15	network	network	NOUN
ajst-20799	25	16	output	output	NOUN
ajst-20799	25	17	and	and	CCONJ
ajst-20799	25	18	actual	actual	ADJ
ajst-20799	25	19	results	result	NOUN
ajst-20799	25	20	.	.	PUNCT
ajst-20799	26	1	backpropagation	backpropagation	NOUN
ajst-20799	26	2	decomposes	decompose	VERB
ajst-20799	26	3	the	the	DET
ajst-20799	26	4	partial	partial	ADJ
ajst-20799	26	5	derivatives	derivative	NOUN
ajst-20799	26	6	(	(	PUNCT
ajst-20799	26	7	gradients	gradient	NOUN
ajst-20799	26	8	)	)	PUNCT
ajst-20799	26	9	of	of	ADP
ajst-20799	26	10	the	the	DET
ajst-20799	26	11	loss	loss	NOUN
ajst-20799	26	12	function	function	NOUN
ajst-20799	26	13	for	for	ADP
ajst-20799	26	14	each	each	DET
ajst-20799	26	15	parameter	parameter	NOUN
ajst-20799	26	16	using	use	VERB
ajst-20799	26	17	the	the	DET
ajst-20799	26	18	chain	chain	NOUN
ajst-20799	26	19	rule	rule	NOUN
ajst-20799	26	20	,	,	PUNCT
ajst-20799	26	21	comprising	comprising	NOUN
ajst-20799	26	22	gradients	gradient	NOUN
ajst-20799	26	23	of	of	ADP
ajst-20799	26	24	activation	activation	NOUN
ajst-20799	26	25	functions	function	NOUN
ajst-20799	26	26	,	,	PUNCT
ajst-20799	26	27	weights	weight	NOUN
ajst-20799	26	28	,	,	PUNCT
ajst-20799	26	29	biases	bias	NOUN
ajst-20799	26	30	,	,	PUNCT
ajst-20799	26	31	etc	etc	X
ajst-20799	26	32	.	.	X
ajst-20799	26	33	traditional	traditional	ADJ
ajst-20799	26	34	machine	machine	NOUN
ajst-20799	26	35	learning	learn	VERB
ajst-20799	26	36	methods	method	NOUN
ajst-20799	26	37	usually	usually	ADV
ajst-20799	26	38	require	require	VERB
ajst-20799	26	39	manual	manual	ADJ
ajst-20799	26	40	feature	feature	NOUN
ajst-20799	26	41	design	design	NOUN
ajst-20799	26	42	,	,	PUNCT
ajst-20799	26	43	where	where	SCONJ
ajst-20799	26	44	"	"	PUNCT
ajst-20799	26	45	features	feature	NOUN
ajst-20799	26	46	"	"	PUNCT
ajst-20799	26	47	refer	refer	VERB
ajst-20799	26	48	to	to	ADP
ajst-20799	26	49	key	key	ADJ
ajst-20799	26	50	information	information	NOUN
ajst-20799	26	51	extracted	extract	VERB
ajst-20799	26	52	from	from	ADP
ajst-20799	26	53	raw	raw	ADJ
ajst-20799	26	54	data	datum	NOUN
ajst-20799	26	55	,	,	PUNCT
ajst-20799	26	56	effectively	effectively	ADV
ajst-20799	26	57	reflecting	reflect	VERB
ajst-20799	26	58	the	the	DET
ajst-20799	26	59	core	core	NOUN
ajst-20799	26	60	attributes	attribute	NOUN
ajst-20799	26	61	or	or	CCONJ
ajst-20799	26	62	patterns	pattern	NOUN
ajst-20799	26	63	of	of	ADP
ajst-20799	26	64	data	datum	NOUN
ajst-20799	26	65	,	,	PUNCT
ajst-20799	26	66	facilitating	facilitate	VERB
ajst-20799	26	67	subsequent	subsequent	ADJ
ajst-20799	26	68	model	model	NOUN
ajst-20799	26	69	training	training	NOUN
ajst-20799	26	70	and	and	CCONJ
ajst-20799	26	71	prediction	prediction	NOUN
ajst-20799	26	72	.	.	PUNCT
ajst-20799	27	1	in	in	ADP
ajst-20799	27	2	contrast	contrast	NOUN
ajst-20799	27	3	,	,	PUNCT
ajst-20799	27	4	deep	deep	ADJ
ajst-20799	27	5	learning	learning	NOUN
ajst-20799	27	6	can	can	AUX
ajst-20799	27	7	automatically	automatically	ADV
ajst-20799	27	8	learn	learn	VERB
ajst-20799	27	9	feature	feature	NOUN
ajst-20799	27	10	representations	representation	NOUN
ajst-20799	27	11	of	of	ADP
ajst-20799	27	12	data	datum	NOUN
ajst-20799	27	13	through	through	ADP
ajst-20799	27	14	networks	network	NOUN
ajst-20799	27	15	.	.	PUNCT
ajst-20799	28	1	this	this	DET
ajst-20799	28	2	feature	feature	NOUN
ajst-20799	28	3	learning	learning	NOUN
ajst-20799	28	4	process	process	NOUN
ajst-20799	28	5	is	be	AUX
ajst-20799	28	6	done	do	VERB
ajst-20799	28	7	layer	layer	NOUN
ajst-20799	28	8	by	by	ADP
ajst-20799	28	9	layer	layer	NOUN
ajst-20799	28	10	,	,	PUNCT
ajst-20799	28	11	126	126	NUM
ajst-20799	28	12	significantly	significantly	ADV
ajst-20799	28	13	reducing	reduce	VERB
ajst-20799	28	14	reliance	reliance	NOUN
ajst-20799	28	15	on	on	ADP
ajst-20799	28	16	manual	manual	ADJ
ajst-20799	28	17	feature	feature	NOUN
ajst-20799	28	18	engineering	engineering	NOUN
ajst-20799	28	19	compared	compare	VERB
ajst-20799	28	20	to	to	ADP
ajst-20799	28	21	traditional	traditional	ADJ
ajst-20799	28	22	machine	machine	NOUN
ajst-20799	28	23	learning	learning	NOUN
ajst-20799	28	24	,	,	PUNCT
ajst-20799	28	25	with	with	ADP
ajst-20799	28	26	each	each	DET
ajst-20799	28	27	layer	layer	NOUN
ajst-20799	28	28	gradually	gradually	ADV
ajst-20799	28	29	increasing	increase	VERB
ajst-20799	28	30	the	the	DET
ajst-20799	28	31	data	datum	NOUN
ajst-20799	28	32	's	's	PART
ajst-20799	28	33	abstraction	abstraction	ADJ
ajst-20799	28	34	level	level	NOUN
ajst-20799	28	35	.	.	PUNCT
ajst-20799	29	1	by	by	ADP
ajst-20799	29	2	constructing	construct	VERB
ajst-20799	29	3	deep	deep	ADJ
ajst-20799	29	4	learning	learning	NOUN
ajst-20799	29	5	models	model	NOUN
ajst-20799	29	6	tailored	tailor	VERB
ajst-20799	29	7	for	for	ADP
ajst-20799	29	8	vehicle	vehicle	NOUN
ajst-20799	29	9	detection	detection	NOUN
ajst-20799	29	10	tasks	task	NOUN
ajst-20799	29	11	,	,	PUNCT
ajst-20799	29	12	specifically	specifically	ADV
ajst-20799	29	13	addressing	address	VERB
ajst-20799	29	14	the	the	DET
ajst-20799	29	15	recognition	recognition	NOUN
ajst-20799	29	16	and	and	CCONJ
ajst-20799	29	17	tracking	tracking	NOUN
ajst-20799	29	18	needs	need	NOUN
ajst-20799	29	19	of	of	ADP
ajst-20799	29	20	vehicles	vehicle	NOUN
ajst-20799	29	21	in	in	ADP
ajst-20799	29	22	various	various	ADJ
ajst-20799	29	23	environmental	environmental	ADJ
ajst-20799	29	24	conditions	condition	NOUN
ajst-20799	29	25	,	,	PUNCT
ajst-20799	29	26	customizing	customizing	NOUN
ajst-20799	29	27	designs	design	NOUN
ajst-20799	29	28	for	for	ADP
ajst-20799	29	29	challenges	challenge	NOUN
ajst-20799	29	30	like	like	ADP
ajst-20799	29	31	vehicle	vehicle	NOUN
ajst-20799	29	32	morphology	morphology	NOUN
ajst-20799	29	33	,	,	PUNCT
ajst-20799	29	34	background	background	NOUN
ajst-20799	29	35	complexity	complexity	NOUN
ajst-20799	29	36	,	,	PUNCT
ajst-20799	29	37	lighting	lighting	NOUN
ajst-20799	29	38	changes	change	NOUN
ajst-20799	29	39	,	,	PUNCT
ajst-20799	29	40	perspective	perspective	NOUN
ajst-20799	29	41	shifts	shift	NOUN
ajst-20799	29	42	,	,	PUNCT
ajst-20799	29	43	etc	etc	X
ajst-20799	29	44	.	.	X
ajst-20799	29	45	,	,	PUNCT
ajst-20799	29	46	using	use	VERB
ajst-20799	29	47	convolutional	convolutional	ADJ
ajst-20799	29	48	kernels	kernel	NOUN
ajst-20799	29	49	for	for	ADP
ajst-20799	29	50	sliding	slide	VERB
ajst-20799	29	51	window	window	NOUN
ajst-20799	29	52	scanning	scanning	NOUN
ajst-20799	29	53	of	of	ADP
ajst-20799	29	54	images	image	NOUN
ajst-20799	29	55	to	to	PART
ajst-20799	29	56	extract	extract	VERB
ajst-20799	29	57	local	local	ADJ
ajst-20799	29	58	features	feature	NOUN
ajst-20799	29	59	such	such	ADJ
ajst-20799	29	60	as	as	ADP
ajst-20799	29	61	vehicle	vehicle	NOUN
ajst-20799	29	62	edges	edge	NOUN
ajst-20799	29	63	,	,	PUNCT
ajst-20799	29	64	shapes	shape	NOUN
ajst-20799	29	65	,	,	PUNCT
ajst-20799	29	66	colors	color	NOUN
ajst-20799	29	67	,	,	PUNCT
ajst-20799	29	68	textures	texture	NOUN
ajst-20799	29	69	,	,	PUNCT
ajst-20799	29	70	etc	etc	X
ajst-20799	29	71	.	.	X
ajst-20799	29	72	,	,	PUNCT
ajst-20799	29	73	can	can	AUX
ajst-20799	29	74	achieve	achieve	VERB
ajst-20799	29	75	automatic	automatic	ADJ
ajst-20799	29	76	vehicle	vehicle	NOUN
ajst-20799	29	77	recognition	recognition	NOUN
ajst-20799	29	78	and	and	CCONJ
ajst-20799	29	79	tracking	tracking	NOUN
ajst-20799	29	80	.	.	PUNCT
ajst-20799	30	1	based	base	VERB
ajst-20799	30	2	on	on	ADP
ajst-20799	30	3	convolutional	convolutional	ADJ
ajst-20799	30	4	neural	neural	ADJ
ajst-20799	30	5	networks	network	NOUN
ajst-20799	30	6	(	(	PUNCT
ajst-20799	30	7	cnns	cnns	PROPN
ajst-20799	30	8	)	)	PUNCT
ajst-20799	30	9	,	,	PUNCT
ajst-20799	30	10	vehicle	vehicle	NOUN
ajst-20799	30	11	detection	detection	NOUN
ajst-20799	30	12	models	model	NOUN
ajst-20799	30	13	utilizing	utilize	VERB
ajst-20799	30	14	convolutional	convolutional	ADJ
ajst-20799	30	15	kernels	kernel	NOUN
ajst-20799	30	16	for	for	ADP
ajst-20799	30	17	sliding	slide	VERB
ajst-20799	30	18	window	window	NOUN
ajst-20799	30	19	scanning	scanning	NOUN
ajst-20799	30	20	of	of	ADP
ajst-20799	30	21	images	image	NOUN
ajst-20799	30	22	to	to	PART
ajst-20799	30	23	extract	extract	VERB
ajst-20799	30	24	local	local	ADJ
ajst-20799	30	25	features	feature	NOUN
ajst-20799	30	26	like	like	ADP
ajst-20799	30	27	vehicle	vehicle	NOUN
ajst-20799	30	28	edges	edge	NOUN
ajst-20799	30	29	,	,	PUNCT
ajst-20799	30	30	shapes	shape	NOUN
ajst-20799	30	31	,	,	PUNCT
ajst-20799	30	32	colors	color	NOUN
ajst-20799	30	33	,	,	PUNCT
ajst-20799	30	34	textures	texture	NOUN
ajst-20799	30	35	,	,	PUNCT
ajst-20799	30	36	etc	etc	X
ajst-20799	30	37	.	.	X
ajst-20799	30	38	,	,	PUNCT
ajst-20799	30	39	can	can	AUX
ajst-20799	30	40	effectively	effectively	ADV
ajst-20799	30	41	extract	extract	VERB
ajst-20799	30	42	vehicle	vehicle	NOUN
ajst-20799	30	43	features	feature	NOUN
ajst-20799	30	44	from	from	ADP
ajst-20799	30	45	images	image	NOUN
ajst-20799	30	46	and	and	CCONJ
ajst-20799	30	47	accurately	accurately	ADV
ajst-20799	30	48	recognize	recognize	VERB
ajst-20799	30	49	and	and	CCONJ
ajst-20799	30	50	locate	locate	VERB
ajst-20799	30	51	them	they	PRON
ajst-20799	30	52	.	.	PUNCT
ajst-20799	31	1	2.2	2.2	NUM
ajst-20799	31	2	.	.	PUNCT
ajst-20799	32	1	deep	deep	ADJ
ajst-20799	32	2	learning	learning	NOUN
ajst-20799	32	3	applications	application	NOUN
ajst-20799	32	4	in	in	ADP
ajst-20799	32	5	vehicle	vehicle	NOUN
ajst-20799	32	6	object	object	NOUN
ajst-20799	32	7	detection	detection	NOUN
ajst-20799	32	8	traditional	traditional	ADJ
ajst-20799	32	9	methods	method	NOUN
ajst-20799	32	10	of	of	ADP
ajst-20799	32	11	object	object	NOUN
ajst-20799	32	12	detection	detection	NOUN
ajst-20799	32	13	mainly	mainly	ADV
ajst-20799	32	14	involve	involve	VERB
ajst-20799	32	15	techniques	technique	NOUN
ajst-20799	32	16	based	base	VERB
ajst-20799	32	17	on	on	ADP
ajst-20799	32	18	feature	feature	NOUN
ajst-20799	32	19	engineering	engineering	NOUN
ajst-20799	32	20	and	and	CCONJ
ajst-20799	32	21	machine	machine	NOUN
ajst-20799	32	22	learning	learning	NOUN
ajst-20799	32	23	,	,	PUNCT
ajst-20799	32	24	such	such	ADJ
ajst-20799	32	25	as	as	ADP
ajst-20799	32	26	haar	haar	PROPN
ajst-20799	32	27	features	feature	NOUN
ajst-20799	32	28	,	,	PUNCT
ajst-20799	32	29	histogram	histogram	NOUN
ajst-20799	32	30	of	of	ADP
ajst-20799	32	31	oriented	orient	VERB
ajst-20799	32	32	gradients	gradient	NOUN
ajst-20799	32	33	(	(	PUNCT
ajst-20799	32	34	hog	hog	NOUN
ajst-20799	32	35	)	)	PUNCT
ajst-20799	32	36	features	feature	NOUN
ajst-20799	32	37	,	,	PUNCT
ajst-20799	32	38	and	and	CCONJ
ajst-20799	32	39	support	support	VERB
ajst-20799	32	40	vector	vector	NOUN
ajst-20799	32	41	machines	machine	NOUN
ajst-20799	32	42	(	(	PUNCT
ajst-20799	32	43	svm	svm	PROPN
ajst-20799	32	44	)	)	PUNCT
ajst-20799	32	45	.	.	PUNCT
ajst-20799	33	1	haar	haar	PROPN
ajst-20799	33	2	features	feature	NOUN
ajst-20799	33	3	,	,	PUNCT
ajst-20799	33	4	proposed	propose	VERB
ajst-20799	33	5	by	by	ADP
ajst-20799	33	6	paul	paul	PROPN
ajst-20799	33	7	viola	viola	PROPN
ajst-20799	33	8	and	and	CCONJ
ajst-20799	33	9	michael	michael	PROPN
ajst-20799	33	10	jones	jones	PROPN
ajst-20799	33	11	,	,	PUNCT
ajst-20799	33	12	are	be	AUX
ajst-20799	33	13	simple	simple	ADJ
ajst-20799	33	14	and	and	CCONJ
ajst-20799	33	15	efficient	efficient	ADJ
ajst-20799	33	16	visual	visual	ADJ
ajst-20799	33	17	features	feature	NOUN
ajst-20799	33	18	that	that	SCONJ
ajst-20799	33	19	essentially	essentially	ADV
ajst-20799	33	20	compute	compute	VERB
ajst-20799	33	21	contrast	contrast	NOUN
ajst-20799	33	22	on	on	ADP
ajst-20799	33	23	image	image	NOUN
ajst-20799	33	24	subregions	subregion	NOUN
ajst-20799	33	25	,	,	PUNCT
ajst-20799	33	26	typically	typically	ADV
ajst-20799	33	27	manifested	manifest	VERB
ajst-20799	33	28	as	as	ADP
ajst-20799	33	29	rectangular	rectangular	ADJ
ajst-20799	33	30	structures	structure	NOUN
ajst-20799	33	31	including	include	VERB
ajst-20799	33	32	single	single	ADJ
ajst-20799	33	33	rectangles	rectangle	NOUN
ajst-20799	33	34	,	,	PUNCT
ajst-20799	33	35	differential	differential	ADJ
ajst-20799	33	36	adjacent	adjacent	ADJ
ajst-20799	33	37	rectangles	rectangle	NOUN
ajst-20799	33	38	,	,	PUNCT
ajst-20799	33	39	and	and	CCONJ
ajst-20799	33	40	more	more	ADV
ajst-20799	33	41	complex	complex	ADJ
ajst-20799	33	42	multirectangle	multirectangle	ADJ
ajst-20799	33	43	combinations	combination	NOUN
ajst-20799	33	44	.	.	PUNCT
ajst-20799	34	1	haar	haar	PROPN
ajst-20799	34	2	features	feature	NOUN
ajst-20799	34	3	are	be	AUX
ajst-20799	34	4	often	often	ADV
ajst-20799	34	5	combined	combine	VERB
ajst-20799	34	6	with	with	ADP
ajst-20799	34	7	cascade	cascade	NOUN
ajst-20799	34	8	classifiers	classifier	NOUN
ajst-20799	34	9	and	and	CCONJ
ajst-20799	34	10	the	the	DET
ajst-20799	34	11	adaboost	adaboost	ADJ
ajst-20799	34	12	algorithm	algorithm	NOUN
ajst-20799	34	13	.	.	PUNCT
ajst-20799	35	1	cascade	cascade	NOUN
ajst-20799	35	2	classifiers	classifier	NOUN
ajst-20799	35	3	are	be	AUX
ajst-20799	35	4	composed	compose	VERB
ajst-20799	35	5	of	of	ADP
ajst-20799	35	6	multiple	multiple	ADJ
ajst-20799	35	7	weak	weak	ADJ
ajst-20799	35	8	classifiers	classifier	NOUN
ajst-20799	35	9	connected	connect	VERB
ajst-20799	35	10	in	in	ADP
ajst-20799	35	11	series	series	NOUN
ajst-20799	35	12	,	,	PUNCT
ajst-20799	35	13	with	with	ADP
ajst-20799	35	14	each	each	DET
ajst-20799	35	15	classifier	classifier	NOUN
ajst-20799	35	16	responsible	responsible	ADJ
ajst-20799	35	17	for	for	ADP
ajst-20799	35	18	screening	screen	VERB
ajst-20799	35	19	out	out	ADP
ajst-20799	35	20	a	a	DET
ajst-20799	35	21	portion	portion	NOUN
ajst-20799	35	22	of	of	ADP
ajst-20799	35	23	non	non	ADJ
ajst-20799	35	24	-	-	ADJ
ajst-20799	35	25	target	target	ADJ
ajst-20799	35	26	regions	region	NOUN
ajst-20799	35	27	,	,	PUNCT
ajst-20799	35	28	thereby	thereby	ADV
ajst-20799	35	29	reducing	reduce	VERB
ajst-20799	35	30	the	the	DET
ajst-20799	35	31	computational	computational	ADJ
ajst-20799	35	32	burden	burden	NOUN
ajst-20799	35	33	on	on	ADP
ajst-20799	35	34	subsequent	subsequent	ADJ
ajst-20799	35	35	classifiers	classifier	NOUN
ajst-20799	35	36	.	.	PUNCT
ajst-20799	36	1	histogram	histogram	NOUN
ajst-20799	36	2	of	of	ADP
ajst-20799	36	3	oriented	orient	VERB
ajst-20799	36	4	gradients	gradient	NOUN
ajst-20799	36	5	(	(	PUNCT
ajst-20799	36	6	hog	hog	NOUN
ajst-20799	36	7	)	)	PUNCT
ajst-20799	36	8	features	feature	NOUN
ajst-20799	36	9	,	,	PUNCT
ajst-20799	36	10	introduced	introduce	VERB
ajst-20799	36	11	by	by	ADP
ajst-20799	36	12	navneet	navneet	PROPN
ajst-20799	36	13	dalal	dalal	PROPN
ajst-20799	36	14	and	and	CCONJ
ajst-20799	36	15	bill	bill	PROPN
ajst-20799	36	16	triggs	triggs	PROPN
ajst-20799	36	17	,	,	PUNCT
ajst-20799	36	18	are	be	AUX
ajst-20799	36	19	feature	feature	NOUN
ajst-20799	36	20	descriptors	descriptor	NOUN
ajst-20799	36	21	based	base	VERB
ajst-20799	36	22	on	on	ADP
ajst-20799	36	23	local	local	ADJ
ajst-20799	36	24	gradient	gradient	NOUN
ajst-20799	36	25	direction	direction	NOUN
ajst-20799	36	26	histograms	histogram	NOUN
ajst-20799	36	27	.	.	PUNCT
ajst-20799	37	1	hog	hog	NOUN
ajst-20799	37	2	features	feature	NOUN
ajst-20799	37	3	have	have	AUX
ajst-20799	37	4	shown	show	VERB
ajst-20799	37	5	excellent	excellent	ADJ
ajst-20799	37	6	performance	performance	NOUN
ajst-20799	37	7	in	in	ADP
ajst-20799	37	8	tasks	task	NOUN
ajst-20799	37	9	like	like	ADP
ajst-20799	37	10	pedestrian	pedestrian	NOUN
ajst-20799	37	11	detection	detection	NOUN
ajst-20799	37	12	and	and	CCONJ
ajst-20799	37	13	vehicle	vehicle	NOUN
ajst-20799	37	14	detection	detection	NOUN
ajst-20799	37	15	,	,	PUNCT
ajst-20799	37	16	particularly	particularly	ADV
ajst-20799	37	17	becoming	become	VERB
ajst-20799	37	18	one	one	NUM
ajst-20799	37	19	of	of	ADP
ajst-20799	37	20	the	the	DET
ajst-20799	37	21	mainstream	mainstream	NOUN
ajst-20799	37	22	features	feature	NOUN
ajst-20799	37	23	in	in	ADP
ajst-20799	37	24	early	early	ADJ
ajst-20799	37	25	computer	computer	NOUN
ajst-20799	37	26	vision	vision	NOUN
ajst-20799	37	27	competitions	competition	NOUN
ajst-20799	37	28	like	like	ADP
ajst-20799	37	29	the	the	DET
ajst-20799	37	30	pascal	pascal	PROPN
ajst-20799	37	31	voc	voc	NOUN
ajst-20799	37	32	challenge	challenge	NOUN
ajst-20799	37	33	.	.	PUNCT
ajst-20799	38	1	while	while	SCONJ
ajst-20799	38	2	these	these	DET
ajst-20799	38	3	methods	method	NOUN
ajst-20799	38	4	demonstrate	demonstrate	VERB
ajst-20799	38	5	effectiveness	effectiveness	NOUN
ajst-20799	38	6	in	in	ADP
ajst-20799	38	7	specific	specific	ADJ
ajst-20799	38	8	scenarios	scenario	NOUN
ajst-20799	38	9	,	,	PUNCT
ajst-20799	38	10	their	their	PRON
ajst-20799	38	11	ability	ability	NOUN
ajst-20799	38	12	to	to	PART
ajst-20799	38	13	detect	detect	VERB
ajst-20799	38	14	objects	object	NOUN
ajst-20799	38	15	in	in	ADP
ajst-20799	38	16	complex	complex	ADJ
ajst-20799	38	17	environments	environment	NOUN
ajst-20799	38	18	and	and	CCONJ
ajst-20799	38	19	small	small	ADJ
ajst-20799	38	20	targets	target	NOUN
ajst-20799	38	21	is	be	AUX
ajst-20799	38	22	challenged	challenge	VERB
ajst-20799	38	23	as	as	SCONJ
ajst-20799	38	24	the	the	DET
ajst-20799	38	25	field	field	NOUN
ajst-20799	38	26	of	of	ADP
ajst-20799	38	27	computer	computer	NOUN
ajst-20799	38	28	vision	vision	NOUN
ajst-20799	38	29	evolves	evolve	VERB
ajst-20799	38	30	.	.	PUNCT
ajst-20799	39	1	in	in	ADP
ajst-20799	39	2	contrast	contrast	NOUN
ajst-20799	39	3	,	,	PUNCT
ajst-20799	39	4	deep	deep	ADJ
ajst-20799	39	5	learning	learning	NOUN
ajst-20799	39	6	-	-	PUNCT
ajst-20799	39	7	based	base	VERB
ajst-20799	39	8	object	object	NOUN
ajst-20799	39	9	detection	detection	NOUN
ajst-20799	39	10	methods	method	NOUN
ajst-20799	39	11	are	be	AUX
ajst-20799	39	12	better	well	ADV
ajst-20799	39	13	equipped	equip	VERB
ajst-20799	39	14	to	to	PART
ajst-20799	39	15	address	address	VERB
ajst-20799	39	16	these	these	DET
ajst-20799	39	17	challenges	challenge	NOUN
ajst-20799	39	18	.	.	PUNCT
ajst-20799	40	1	among	among	ADP
ajst-20799	40	2	them	they	PRON
ajst-20799	40	3	,	,	PUNCT
ajst-20799	40	4	faster	fast	ADV
ajst-20799	40	5	r	r	X
ajst-20799	40	6	-	-	PUNCT
ajst-20799	40	7	cnn	cnn	PROPN
ajst-20799	40	8	(	(	PUNCT
ajst-20799	40	9	faster	fast	ADJ
ajst-20799	40	10	region	region	NOUN
ajst-20799	40	11	-	-	PUNCT
ajst-20799	40	12	based	base	VERB
ajst-20799	40	13	convolutional	convolutional	ADJ
ajst-20799	40	14	neural	neural	ADJ
ajst-20799	40	15	network	network	NOUN
ajst-20799	40	16	)	)	PUNCT
ajst-20799	40	17	stands	stand	VERB
ajst-20799	40	18	as	as	ADP
ajst-20799	40	19	a	a	DET
ajst-20799	40	20	classic	classic	ADJ
ajst-20799	40	21	object	object	NOUN
ajst-20799	40	22	detection	detection	NOUN
ajst-20799	40	23	framework	framework	NOUN
ajst-20799	40	24	proposed	propose	VERB
ajst-20799	40	25	by	by	ADP
ajst-20799	40	26	ross	ross	PROPN
ajst-20799	40	27	girshick	girshick	PROPN
ajst-20799	40	28	and	and	CCONJ
ajst-20799	40	29	others	other	NOUN
ajst-20799	40	30	.	.	PUNCT
ajst-20799	41	1	it	it	PRON
ajst-20799	41	2	combines	combine	VERB
ajst-20799	41	3	a	a	DET
ajst-20799	41	4	region	region	NOUN
ajst-20799	41	5	proposal	proposal	NOUN
ajst-20799	41	6	network	network	NOUN
ajst-20799	41	7	(	(	PUNCT
ajst-20799	41	8	rpn	rpn	PROPN
ajst-20799	41	9	)	)	PUNCT
ajst-20799	41	10	and	and	CCONJ
ajst-20799	41	11	a	a	DET
ajst-20799	41	12	region	region	NOUN
ajst-20799	41	13	-	-	PUNCT
ajst-20799	41	14	based	base	VERB
ajst-20799	41	15	convolutional	convolutional	ADJ
ajst-20799	41	16	neural	neural	ADJ
ajst-20799	41	17	network	network	NOUN
ajst-20799	41	18	(	(	PUNCT
ajst-20799	41	19	rcnn	rcnn	PROPN
ajst-20799	41	20	)	)	PUNCT
ajst-20799	41	21	to	to	PART
ajst-20799	41	22	optimize	optimize	VERB
ajst-20799	41	23	the	the	DET
ajst-20799	41	24	entire	entire	ADJ
ajst-20799	41	25	model	model	NOUN
ajst-20799	41	26	from	from	ADP
ajst-20799	41	27	input	input	NOUN
ajst-20799	41	28	image	image	NOUN
ajst-20799	41	29	to	to	ADP
ajst-20799	41	30	final	final	ADJ
ajst-20799	41	31	detection	detection	NOUN
ajst-20799	41	32	output	output	NOUN
ajst-20799	41	33	as	as	ADP
ajst-20799	41	34	a	a	DET
ajst-20799	41	35	whole	whole	NOUN
ajst-20799	41	36	,	,	PUNCT
ajst-20799	41	37	reducing	reduce	VERB
ajst-20799	41	38	error	error	NOUN
ajst-20799	41	39	accumulation	accumulation	NOUN
ajst-20799	41	40	between	between	ADP
ajst-20799	41	41	feature	feature	NOUN
ajst-20799	41	42	extraction	extraction	NOUN
ajst-20799	41	43	and	and	CCONJ
ajst-20799	41	44	classification	classification	NOUN
ajst-20799	41	45	decisions	decision	NOUN
ajst-20799	41	46	seen	see	VERB
ajst-20799	41	47	in	in	ADP
ajst-20799	41	48	traditional	traditional	ADJ
ajst-20799	41	49	methods	method	NOUN
ajst-20799	41	50	,	,	PUNCT
ajst-20799	41	51	achieving	achieve	VERB
ajst-20799	41	52	end	end	NOUN
ajst-20799	41	53	-	-	PUNCT
ajst-20799	41	54	to	to	ADP
ajst-20799	41	55	-	-	PUNCT
ajst-20799	41	56	end	end	NOUN
ajst-20799	41	57	object	object	NOUN
ajst-20799	41	58	detection	detection	NOUN
ajst-20799	41	59	.	.	PUNCT
ajst-20799	42	1	the	the	DET
ajst-20799	42	2	rpn	rpn	PROPN
ajst-20799	42	3	generates	generate	VERB
ajst-20799	42	4	candidate	candidate	NOUN
ajst-20799	42	5	regions	region	NOUN
ajst-20799	42	6	,	,	PUNCT
ajst-20799	42	7	while	while	SCONJ
ajst-20799	42	8	the	the	DET
ajst-20799	42	9	rcnn	rcnn	NOUN
ajst-20799	42	10	,	,	PUNCT
ajst-20799	42	11	as	as	ADP
ajst-20799	42	12	the	the	DET
ajst-20799	42	13	first	first	ADJ
ajst-20799	42	14	stage	stage	NOUN
ajst-20799	42	15	of	of	ADP
ajst-20799	42	16	faster	fast	ADJ
ajst-20799	42	17	r	r	NOUN
ajst-20799	42	18	-	-	PUNCT
ajst-20799	42	19	cnn	cnn	PROPN
ajst-20799	42	20	,	,	PUNCT
ajst-20799	42	21	is	be	AUX
ajst-20799	42	22	responsible	responsible	ADJ
ajst-20799	42	23	for	for	ADP
ajst-20799	42	24	classifying	classify	VERB
ajst-20799	42	25	and	and	CCONJ
ajst-20799	42	26	locating	locate	VERB
ajst-20799	42	27	these	these	DET
ajst-20799	42	28	candidate	candidate	NOUN
ajst-20799	42	29	regions	region	NOUN
ajst-20799	42	30	.	.	PUNCT
ajst-20799	43	1	its	its	PRON
ajst-20799	43	2	core	core	NOUN
ajst-20799	43	3	task	task	NOUN
ajst-20799	43	4	is	be	AUX
ajst-20799	43	5	to	to	PART
ajst-20799	43	6	automatically	automatically	ADV
ajst-20799	43	7	generate	generate	VERB
ajst-20799	43	8	candidate	candidate	NOUN
ajst-20799	43	9	regions	region	NOUN
ajst-20799	43	10	(	(	PUNCT
ajst-20799	43	11	regions	region	NOUN
ajst-20799	43	12	of	of	ADP
ajst-20799	43	13	interests	interest	NOUN
ajst-20799	43	14	,	,	PUNCT
ajst-20799	43	15	rois	rois	PROPN
ajst-20799	43	16	)	)	PUNCT
ajst-20799	43	17	that	that	PRON
ajst-20799	43	18	may	may	AUX
ajst-20799	43	19	contain	contain	VERB
ajst-20799	43	20	target	target	NOUN
ajst-20799	43	21	objects	object	NOUN
ajst-20799	43	22	,	,	PUNCT
ajst-20799	43	23	reducing	reduce	VERB
ajst-20799	43	24	the	the	DET
ajst-20799	43	25	computational	computational	ADJ
ajst-20799	43	26	burden	burden	NOUN
ajst-20799	43	27	of	of	ADP
ajst-20799	43	28	subsequent	subsequent	ADJ
ajst-20799	43	29	processing	processing	NOUN
ajst-20799	43	30	and	and	CCONJ
ajst-20799	43	31	thus	thus	ADV
ajst-20799	43	32	balancing	balance	VERB
ajst-20799	43	33	the	the	DET
ajst-20799	43	34	accuracy	accuracy	NOUN
ajst-20799	43	35	and	and	CCONJ
ajst-20799	43	36	efficiency	efficiency	NOUN
ajst-20799	43	37	of	of	ADP
ajst-20799	43	38	object	object	NOUN
ajst-20799	43	39	detection	detection	NOUN
ajst-20799	43	40	.	.	PUNCT
ajst-20799	44	1	aditionally	aditionally	PROPN
ajst-20799	44	2	,	,	PUNCT
ajst-20799	44	3	yolo	yolo	PROPN
ajst-20799	44	4	(	(	PUNCT
ajst-20799	44	5	you	you	PRON
ajst-20799	44	6	only	only	ADV
ajst-20799	44	7	look	look	VERB
ajst-20799	44	8	once	once	ADV
ajst-20799	44	9	)	)	PUNCT
ajst-20799	44	10	is	be	AUX
ajst-20799	44	11	another	another	DET
ajst-20799	44	12	popular	popular	ADJ
ajst-20799	44	13	object	object	NOUN
ajst-20799	44	14	detection	detection	NOUN
ajst-20799	44	15	algorithm	algorithm	NOUN
ajst-20799	44	16	that	that	PRON
ajst-20799	44	17	innovatively	innovatively	ADV
ajst-20799	44	18	treats	treat	VERB
ajst-20799	44	19	the	the	DET
ajst-20799	44	20	entire	entire	ADJ
ajst-20799	44	21	object	object	NOUN
ajst-20799	44	22	detection	detection	NOUN
ajst-20799	44	23	task	task	NOUN
ajst-20799	44	24	as	as	ADP
ajst-20799	44	25	a	a	DET
ajst-20799	44	26	single	single	ADJ
ajst-20799	44	27	regression	regression	NOUN
ajst-20799	44	28	problem	problem	NOUN
ajst-20799	44	29	，	，	PUNCT
ajst-20799	44	30	yolov4	yolov4	PROPN
ajst-20799	44	31	is	be	AUX
ajst-20799	44	32	a	a	DET
ajst-20799	44	33	widely	widely	ADV
ajst-20799	44	34	used	use	VERB
ajst-20799	44	35	object	object	NOUN
ajst-20799	44	36	detection	detection	NOUN
ajst-20799	44	37	technology	technology	NOUN
ajst-20799	44	38	that	that	PRON
ajst-20799	44	39	boasts	boast	VERB
ajst-20799	44	40	high	high	ADJ
ajst-20799	44	41	accuracy	accuracy	NOUN
ajst-20799	44	42	and	and	CCONJ
ajst-20799	44	43	fast	fast	ADJ
ajst-20799	44	44	inference	inference	NOUN
ajst-20799	44	45	speed[2	speed[2	PROPN
ajst-20799	44	46	]	]	PUNCT
ajst-20799	44	47	,	,	PUNCT
ajst-20799	44	48	unlike	unlike	ADP
ajst-20799	44	49	traditional	traditional	ADJ
ajst-20799	44	50	two	two	NUM
ajst-20799	44	51	-	-	PUNCT
ajst-20799	44	52	stage	stage	NOUN
ajst-20799	44	53	or	or	CCONJ
ajst-20799	44	54	multi	multi	ADJ
ajst-20799	44	55	-	-	ADJ
ajst-20799	44	56	stage	stage	NOUN
ajst-20799	44	57	methods	method	NOUN
ajst-20799	44	58	.	.	PUNCT
ajst-20799	45	1	its	its	PRON
ajst-20799	45	2	core	core	ADJ
ajst-20799	45	3	idea	idea	NOUN
ajst-20799	45	4	is	be	AUX
ajst-20799	45	5	to	to	PART
ajst-20799	45	6	transform	transform	VERB
ajst-20799	45	7	the	the	DET
ajst-20799	45	8	object	object	NOUN
ajst-20799	45	9	detection	detection	NOUN
ajst-20799	45	10	problem	problem	NOUN
ajst-20799	45	11	into	into	ADP
ajst-20799	45	12	a	a	DET
ajst-20799	45	13	regression	regression	NOUN
ajst-20799	45	14	problem	problem	NOUN
ajst-20799	45	15	,	,	PUNCT
ajst-20799	45	16	requiring	require	VERB
ajst-20799	45	17	only	only	ADV
ajst-20799	45	18	one	one	NUM
ajst-20799	45	19	forward	forward	ADV
ajst-20799	45	20	pass	pass	VERB
ajst-20799	45	21	on	on	ADP
ajst-20799	45	22	the	the	DET
ajst-20799	45	23	input	input	NOUN
ajst-20799	45	24	image	image	NOUN
ajst-20799	45	25	to	to	PART
ajst-20799	45	26	directly	directly	ADV
ajst-20799	45	27	predict	predict	VERB
ajst-20799	45	28	the	the	DET
ajst-20799	45	29	class	class	NOUN
ajst-20799	45	30	and	and	CCONJ
ajst-20799	45	31	bounding	bound	VERB
ajst-20799	45	32	box	box	NOUN
ajst-20799	45	33	of	of	ADP
ajst-20799	45	34	the	the	DET
ajst-20799	45	35	target	target	NOUN
ajst-20799	45	36	at	at	ADP
ajst-20799	45	37	the	the	DET
ajst-20799	45	38	image	image	NOUN
ajst-20799	45	39	level	level	NOUN
ajst-20799	45	40	,	,	PUNCT
ajst-20799	45	41	achieving	achieve	VERB
ajst-20799	45	42	fast	fast	ADJ
ajst-20799	45	43	,	,	PUNCT
ajst-20799	45	44	real	real	ADJ
ajst-20799	45	45	-	-	PUNCT
ajst-20799	45	46	time	time	NOUN
ajst-20799	45	47	object	object	NOUN
ajst-20799	45	48	detection	detection	NOUN
ajst-20799	45	49	.	.	PUNCT
ajst-20799	46	1	the	the	DET
ajst-20799	46	2	yolo	yolo	ADJ
ajst-20799	46	3	algorithm	algorithm	NOUN
ajst-20799	46	4	segments	segment	VERB
ajst-20799	46	5	the	the	DET
ajst-20799	46	6	image	image	NOUN
ajst-20799	46	7	into	into	ADP
ajst-20799	46	8	grids	grid	NOUN
ajst-20799	46	9	,	,	PUNCT
ajst-20799	46	10	with	with	ADP
ajst-20799	46	11	each	each	DET
ajst-20799	46	12	grid	grid	NOUN
ajst-20799	46	13	cell	cell	NOUN
ajst-20799	46	14	responsible	responsible	ADJ
ajst-20799	46	15	for	for	ADP
ajst-20799	46	16	predicting	predict	VERB
ajst-20799	46	17	whether	whether	SCONJ
ajst-20799	46	18	there	there	PRON
ajst-20799	46	19	is	be	VERB
ajst-20799	46	20	an	an	DET
ajst-20799	46	21	object	object	NOUN
ajst-20799	46	22	in	in	ADP
ajst-20799	46	23	its	its	PRON
ajst-20799	46	24	coverage	coverage	NOUN
ajst-20799	46	25	area	area	NOUN
ajst-20799	46	26	,	,	PUNCT
ajst-20799	46	27	including	include	VERB
ajst-20799	46	28	predicting	predict	VERB
ajst-20799	46	29	the	the	DET
ajst-20799	46	30	presence	presence	NOUN
ajst-20799	46	31	of	of	ADP
ajst-20799	46	32	objects	object	NOUN
ajst-20799	46	33	,	,	PUNCT
ajst-20799	46	34	their	their	PRON
ajst-20799	46	35	positions	position	NOUN
ajst-20799	46	36	,	,	PUNCT
ajst-20799	46	37	and	and	CCONJ
ajst-20799	46	38	categories	category	NOUN
ajst-20799	46	39	,	,	PUNCT
ajst-20799	46	40	as	as	ADV
ajst-20799	46	41	well	well	ADV
ajst-20799	46	42	as	as	ADP
ajst-20799	46	43	multiple	multiple	ADJ
ajst-20799	46	44	bounding	bounding	NOUN
ajst-20799	46	45	boxes	box	NOUN
ajst-20799	46	46	(	(	PUNCT
ajst-20799	46	47	3	3	NUM
ajst-20799	46	48	in	in	ADP
ajst-20799	46	49	yolov3	yolov3	PROPN
ajst-20799	46	50	)	)	PUNCT
ajst-20799	46	51	containing	contain	VERB
ajst-20799	46	52	5	5	NUM
ajst-20799	46	53	coordinate	coordinate	NOUN
ajst-20799	46	54	values	value	NOUN
ajst-20799	46	55	(	(	PUNCT
ajst-20799	46	56	center	center	NOUN
ajst-20799	46	57	coordinates	coordinate	NOUN
ajst-20799	46	58	,	,	PUNCT
ajst-20799	46	59	width	width	ADJ
ajst-20799	46	60	,	,	PUNCT
ajst-20799	46	61	height	height	NOUN
ajst-20799	46	62	,	,	PUNCT
ajst-20799	46	63	and	and	CCONJ
ajst-20799	46	64	confidence	confidence	NOUN
ajst-20799	46	65	)	)	PUNCT
ajst-20799	46	66	and	and	CCONJ
ajst-20799	46	67	multiple	multiple	ADJ
ajst-20799	46	68	class	class	NOUN
ajst-20799	46	69	probabilities	probability	NOUN
ajst-20799	46	70	,	,	PUNCT
ajst-20799	46	71	thus	thus	ADV
ajst-20799	46	72	enabling	enable	VERB
ajst-20799	46	73	real	real	ADJ
ajst-20799	46	74	-	-	PUNCT
ajst-20799	46	75	time	time	NOUN
ajst-20799	46	76	object	object	NOUN
ajst-20799	46	77	detection	detection	NOUN
ajst-20799	46	78	capabilities	capability	NOUN
ajst-20799	46	79	.	.	PUNCT
ajst-20799	47	1	3	3	X
ajst-20799	47	2	.	.	PUNCT
ajst-20799	47	3	based	base	VERB
ajst-20799	47	4	on	on	ADP
ajst-20799	47	5	deep	deep	ADJ
ajst-20799	47	6	learning	learning	NOUN
ajst-20799	47	7	techniques	technique	NOUN
ajst-20799	47	8	for	for	ADP
ajst-20799	47	9	vehicle	vehicle	NOUN
ajst-20799	47	10	tracking	track	VERB
ajst-20799	47	11	3.1	3.1	NUM
ajst-20799	47	12	.	.	PUNCT
ajst-20799	48	1	overview	overview	NOUN
ajst-20799	48	2	of	of	ADP
ajst-20799	48	3	vehicle	vehicle	NOUN
ajst-20799	48	4	tracking	tracking	NOUN
ajst-20799	48	5	problem	problem	NOUN
ajst-20799	48	6	vehicle	vehicle	NOUN
ajst-20799	48	7	re	re	NOUN
ajst-20799	48	8	-	-	NOUN
ajst-20799	48	9	identification	identification	NOUN
ajst-20799	48	10	technology	technology	NOUN
ajst-20799	48	11	falls	fall	VERB
ajst-20799	48	12	within	within	ADP
ajst-20799	48	13	the	the	DET
ajst-20799	48	14	realm	realm	NOUN
ajst-20799	48	15	of	of	ADP
ajst-20799	48	16	urban	urban	ADJ
ajst-20799	48	17	intelligent	intelligent	ADJ
ajst-20799	48	18	transportation	transportation	NOUN
ajst-20799	48	19	and	and	CCONJ
ajst-20799	48	20	has	have	AUX
ajst-20799	48	21	garnered	garner	VERB
ajst-20799	48	22	widespread	widespread	ADJ
ajst-20799	48	23	attention	attention	NOUN
ajst-20799	48	24	due	due	ADP
ajst-20799	48	25	to	to	ADP
ajst-20799	48	26	its	its	PRON
ajst-20799	48	27	ability	ability	NOUN
ajst-20799	48	28	to	to	PART
ajst-20799	48	29	identify	identify	VERB
ajst-20799	48	30	vehicles	vehicle	NOUN
ajst-20799	48	31	based	base	VERB
ajst-20799	48	32	solely	solely	ADV
ajst-20799	48	33	on	on	ADP
ajst-20799	48	34	their	their	PRON
ajst-20799	48	35	appearance.[3	appearance.[3	NUM
ajst-20799	48	36	]	]	X
ajst-20799	48	37	multiple	multiple	ADJ
ajst-20799	48	38	-	-	PUNCT
ajst-20799	48	39	object	object	NOUN
ajst-20799	48	40	tracking	tracking	NOUN
ajst-20799	48	41	algorithms	algorithm	NOUN
ajst-20799	48	42	face	face	VERB
ajst-20799	48	43	challenges	challenge	NOUN
ajst-20799	48	44	in	in	ADP
ajst-20799	48	45	complex	complex	ADJ
ajst-20799	48	46	traffic	traffic	NOUN
ajst-20799	48	47	scenarios	scenario	NOUN
ajst-20799	48	48	where	where	SCONJ
ajst-20799	48	49	vehicles	vehicle	NOUN
ajst-20799	48	50	frequently	frequently	ADV
ajst-20799	48	51	experience	experience	VERB
ajst-20799	48	52	occlusion	occlusion	NOUN
ajst-20799	48	53	and	and	CCONJ
ajst-20799	48	54	intersecting	intersect	VERB
ajst-20799	48	55	motion	motion	NOUN
ajst-20799	48	56	trajectories	trajectory	NOUN
ajst-20799	48	57	.	.	PUNCT
ajst-20799	49	1	for	for	ADP
ajst-20799	49	2	instance	instance	NOUN
ajst-20799	49	3	,	,	PUNCT
ajst-20799	49	4	when	when	SCONJ
ajst-20799	49	5	a	a	DET
ajst-20799	49	6	car	car	NOUN
ajst-20799	49	7	is	be	AUX
ajst-20799	49	8	driving	drive	VERB
ajst-20799	49	9	in	in	ADP
ajst-20799	49	10	front	front	NOUN
ajst-20799	49	11	of	of	ADP
ajst-20799	49	12	a	a	DET
ajst-20799	49	13	bus	bus	NOUN
ajst-20799	49	14	,	,	PUNCT
ajst-20799	49	15	the	the	DET
ajst-20799	49	16	car	car	NOUN
ajst-20799	49	17	may	may	AUX
ajst-20799	49	18	be	be	AUX
ajst-20799	49	19	partially	partially	ADV
ajst-20799	49	20	or	or	CCONJ
ajst-20799	49	21	fully	fully	ADV
ajst-20799	49	22	occluded	occlude	VERB
ajst-20799	49	23	by	by	ADP
ajst-20799	49	24	the	the	DET
ajst-20799	49	25	bus	bus	NOUN
ajst-20799	49	26	,	,	PUNCT
ajst-20799	49	27	causing	cause	VERB
ajst-20799	49	28	its	its	PRON
ajst-20799	49	29	visual	visual	ADJ
ajst-20799	49	30	features	feature	NOUN
ajst-20799	49	31	such	such	ADJ
ajst-20799	49	32	as	as	ADP
ajst-20799	49	33	color	color	NOUN
ajst-20799	49	34	and	and	CCONJ
ajst-20799	49	35	shape	shape	NOUN
ajst-20799	49	36	to	to	PART
ajst-20799	49	37	become	become	VERB
ajst-20799	49	38	blurry	blurry	ADJ
ajst-20799	49	39	or	or	CCONJ
ajst-20799	49	40	even	even	ADV
ajst-20799	49	41	invisible	invisible	ADJ
ajst-20799	49	42	from	from	ADP
ajst-20799	49	43	the	the	DET
ajst-20799	49	44	camera	camera	NOUN
ajst-20799	49	45	's	's	PART
ajst-20799	49	46	perspective	perspective	NOUN
ajst-20799	49	47	.	.	PUNCT
ajst-20799	50	1	this	this	DET
ajst-20799	50	2	variation	variation	NOUN
ajst-20799	50	3	in	in	ADP
ajst-20799	50	4	target	target	NOUN
ajst-20799	50	5	appearance	appearance	NOUN
ajst-20799	50	6	increases	increase	VERB
ajst-20799	50	7	the	the	DET
ajst-20799	50	8	difficulty	difficulty	NOUN
ajst-20799	50	9	of	of	ADP
ajst-20799	50	10	tracking	tracking	NOUN
ajst-20799	50	11	.	.	PUNCT
ajst-20799	51	1	different	different	ADJ
ajst-20799	51	2	types	type	NOUN
ajst-20799	51	3	of	of	ADP
ajst-20799	51	4	vehicles	vehicle	NOUN
ajst-20799	51	5	(	(	PUNCT
ajst-20799	51	6	e.g.	e.g.	ADV
ajst-20799	51	7	,	,	PUNCT
ajst-20799	51	8	cars	car	NOUN
ajst-20799	51	9	,	,	PUNCT
ajst-20799	51	10	trucks	truck	NOUN
ajst-20799	51	11	,	,	PUNCT
ajst-20799	51	12	bicycles	bicycle	NOUN
ajst-20799	51	13	)	)	PUNCT
ajst-20799	51	14	vary	vary	VERB
ajst-20799	51	15	significantly	significantly	ADV
ajst-20799	51	16	in	in	ADP
ajst-20799	51	17	size	size	NOUN
ajst-20799	51	18	and	and	CCONJ
ajst-20799	51	19	motion	motion	NOUN
ajst-20799	51	20	characteristics	characteristic	NOUN
ajst-20799	51	21	.	.	PUNCT
ajst-20799	52	1	for	for	ADP
ajst-20799	52	2	example	example	NOUN
ajst-20799	52	3	,	,	PUNCT
ajst-20799	52	4	small	small	ADJ
ajst-20799	52	5	cars	car	NOUN
ajst-20799	52	6	,	,	PUNCT
ajst-20799	52	7	large	large	ADJ
ajst-20799	52	8	buses	bus	NOUN
ajst-20799	52	9	,	,	PUNCT
ajst-20799	52	10	trucks	truck	NOUN
ajst-20799	52	11	,	,	PUNCT
ajst-20799	52	12	motorcycles	motorcycle	NOUN
ajst-20799	52	13	,	,	PUNCT
ajst-20799	52	14	and	and	CCONJ
ajst-20799	52	15	bicycles	bicycle	NOUN
ajst-20799	52	16	differ	differ	VERB
ajst-20799	52	17	in	in	ADP
ajst-20799	52	18	size	size	NOUN
ajst-20799	52	19	,	,	PUNCT
ajst-20799	52	20	shape	shape	NOUN
ajst-20799	52	21	,	,	PUNCT
ajst-20799	52	22	speed	speed	NOUN
ajst-20799	52	23	,	,	PUNCT
ajst-20799	52	24	acceleration	acceleration	NOUN
ajst-20799	52	25	,	,	PUNCT
ajst-20799	52	26	etc	etc	X
ajst-20799	52	27	.	.	X
ajst-20799	52	28	tracking	track	VERB
ajst-20799	52	29	algorithms	algorithm	NOUN
ajst-20799	52	30	need	need	VERB
ajst-20799	52	31	to	to	PART
ajst-20799	52	32	adapt	adapt	VERB
ajst-20799	52	33	well	well	ADV
ajst-20799	52	34	and	and	CCONJ
ajst-20799	52	35	generalize	generalize	VERB
ajst-20799	52	36	effectively	effectively	ADV
ajst-20799	52	37	to	to	PART
ajst-20799	52	38	handle	handle	VERB
ajst-20799	52	39	different	different	ADJ
ajst-20799	52	40	types	type	NOUN
ajst-20799	52	41	of	of	ADP
ajst-20799	52	42	vehicles	vehicle	NOUN
ajst-20799	52	43	,	,	PUNCT
ajst-20799	52	44	regardless	regardless	ADV
ajst-20799	52	45	of	of	ADP
ajst-20799	52	46	their	their	PRON
ajst-20799	52	47	size	size	NOUN
ajst-20799	52	48	,	,	PUNCT
ajst-20799	52	49	changing	change	VERB
ajst-20799	52	50	shapes	shape	NOUN
ajst-20799	52	51	,	,	PUNCT
ajst-20799	52	52	or	or	CCONJ
ajst-20799	52	53	complex	complex	ADJ
ajst-20799	52	54	motion	motion	NOUN
ajst-20799	52	55	patterns	pattern	NOUN
ajst-20799	52	56	.	.	PUNCT
ajst-20799	53	1	real	real	ADJ
ajst-20799	53	2	-	-	PUNCT
ajst-20799	53	3	time	time	NOUN
ajst-20799	53	4	processing	processing	NOUN
ajst-20799	53	5	is	be	AUX
ajst-20799	53	6	an	an	DET
ajst-20799	53	7	indispensable	indispensable	ADJ
ajst-20799	53	8	feature	feature	NOUN
ajst-20799	53	9	of	of	ADP
ajst-20799	53	10	multiple	multiple	ADJ
ajst-20799	53	11	-	-	PUNCT
ajst-20799	53	12	object	object	NOUN
ajst-20799	53	13	tracking	tracking	NOUN
ajst-20799	53	14	algorithms	algorithm	NOUN
ajst-20799	53	15	in	in	ADP
ajst-20799	53	16	practical	practical	ADJ
ajst-20799	53	17	applications	application	NOUN
ajst-20799	53	18	,	,	PUNCT
ajst-20799	53	19	yet	yet	CCONJ
ajst-20799	53	20	it	it	PRON
ajst-20799	53	21	poses	pose	VERB
ajst-20799	53	22	a	a	DET
ajst-20799	53	23	significant	significant	ADJ
ajst-20799	53	24	challenge	challenge	NOUN
ajst-20799	53	25	.	.	PUNCT
ajst-20799	54	1	these	these	DET
ajst-20799	54	2	algorithms	algorithm	NOUN
ajst-20799	54	3	must	must	AUX
ajst-20799	54	4	process	process	VERB
ajst-20799	54	5	large	large	ADJ
ajst-20799	54	6	-	-	PUNCT
ajst-20799	54	7	scale	scale	NOUN
ajst-20799	54	8	data	datum	NOUN
ajst-20799	54	9	and	and	CCONJ
ajst-20799	54	10	achieve	achieve	VERB
ajst-20799	54	11	real	real	ADJ
ajst-20799	54	12	-	-	PUNCT
ajst-20799	54	13	time	time	NOUN
ajst-20799	54	14	tracking	tracking	NOUN
ajst-20799	54	15	within	within	ADP
ajst-20799	54	16	a	a	DET
ajst-20799	54	17	short	short	ADJ
ajst-20799	54	18	timeframe	timeframe	NOUN
ajst-20799	54	19	,	,	PUNCT
ajst-20799	54	20	updating	update	VERB
ajst-20799	54	21	target	target	NOUN
ajst-20799	54	22	state	state	NOUN
ajst-20799	54	23	information	information	NOUN
ajst-20799	54	24	promptly	promptly	ADV
ajst-20799	54	25	to	to	PART
ajst-20799	54	26	make	make	VERB
ajst-20799	54	27	timely	timely	ADJ
ajst-20799	54	28	decisions	decision	NOUN
ajst-20799	54	29	and	and	CCONJ
ajst-20799	54	30	avoid	avoid	VERB
ajst-20799	54	31	safety	safety	NOUN
ajst-20799	54	32	risks	risk	NOUN
ajst-20799	54	33	or	or	CCONJ
ajst-20799	54	34	monitoring	monitor	VERB
ajst-20799	54	35	failures	failure	NOUN
ajst-20799	54	36	caused	cause	VERB
ajst-20799	54	37	by	by	ADP
ajst-20799	54	38	processing	processing	NOUN
ajst-20799	54	39	delays	delay	NOUN
ajst-20799	54	40	.	.	PUNCT
ajst-20799	55	1	hence	hence	ADV
ajst-20799	55	2	,	,	PUNCT
ajst-20799	55	3	multiple	multiple	ADJ
ajst-20799	55	4	-	-	PUNCT
ajst-20799	55	5	object	object	NOUN
ajst-20799	55	6	tracking	tracking	NOUN
ajst-20799	55	7	algorithms	algorithm	NOUN
ajst-20799	55	8	need	need	VERB
ajst-20799	55	9	to	to	PART
ajst-20799	55	10	swiftly	swiftly	ADV
ajst-20799	55	11	perform	perform	VERB
ajst-20799	55	12	a	a	DET
ajst-20799	55	13	series	series	NOUN
ajst-20799	55	14	of	of	ADP
ajst-20799	55	15	operations	operation	NOUN
ajst-20799	55	16	including	include	VERB
ajst-20799	55	17	object	object	NOUN
ajst-20799	55	18	detection	detection	NOUN
ajst-20799	55	19	,	,	PUNCT
ajst-20799	55	20	association	association	NOUN
ajst-20799	55	21	,	,	PUNCT
ajst-20799	55	22	and	and	CCONJ
ajst-20799	55	23	state	state	NOUN
ajst-20799	55	24	updating	updating	NOUN
ajst-20799	55	25	on	on	ADP
ajst-20799	55	26	each	each	DET
ajst-20799	55	27	frame	frame	NOUN
ajst-20799	55	28	of	of	ADP
ajst-20799	55	29	the	the	DET
ajst-20799	55	30	image	image	NOUN
ajst-20799	55	31	.	.	PUNCT
ajst-20799	56	1	they	they	PRON
ajst-20799	56	2	must	must	AUX
ajst-20799	56	3	output	output	VERB
ajst-20799	56	4	the	the	DET
ajst-20799	56	5	latest	late	ADJ
ajst-20799	56	6	positions	position	NOUN
ajst-20799	56	7	,	,	PUNCT
ajst-20799	56	8	speeds	speed	NOUN
ajst-20799	56	9	,	,	PUNCT
ajst-20799	56	10	orientations	orientation	NOUN
ajst-20799	56	11	,	,	PUNCT
ajst-20799	56	12	etc	etc	X
ajst-20799	56	13	.	.	X
ajst-20799	56	14	,	,	PUNCT
ajst-20799	56	15	of	of	ADP
ajst-20799	56	16	all	all	DET
ajst-20799	56	17	targets	target	NOUN
ajst-20799	56	18	in	in	ADP
ajst-20799	56	19	milliseconds	millisecond	NOUN
ajst-20799	56	20	,	,	PUNCT
ajst-20799	56	21	enabling	enable	VERB
ajst-20799	56	22	decision	decision	NOUN
ajst-20799	56	23	-	-	PUNCT
ajst-20799	56	24	making	make	VERB
ajst-20799	56	25	modules	module	NOUN
ajst-20799	56	26	to	to	PART
ajst-20799	56	27	plan	plan	VERB
ajst-20799	56	28	driving	drive	VERB
ajst-20799	56	29	paths	path	NOUN
ajst-20799	56	30	and	and	CCONJ
ajst-20799	56	31	avoid	avoid	VERB
ajst-20799	56	32	collision	collision	NOUN
ajst-20799	56	33	risks	risk	NOUN
ajst-20799	56	34	.	.	PUNCT
ajst-20799	57	1	deep	deep	ADJ
ajst-20799	57	2	learning	learning	NOUN
ajst-20799	57	3	-	-	PUNCT
ajst-20799	57	4	based	base	VERB
ajst-20799	57	5	vehicle	vehicle	NOUN
ajst-20799	57	6	tracking	tracking	NOUN
ajst-20799	57	7	technology	technology	NOUN
ajst-20799	57	8	utilizes	utilize	VERB
ajst-20799	57	9	the	the	DET
ajst-20799	57	10	powerful	powerful	ADJ
ajst-20799	57	11	representation	representation	NOUN
ajst-20799	57	12	capabilities	capability	NOUN
ajst-20799	57	13	of	of	ADP
ajst-20799	57	14	convolutional	convolutional	ADJ
ajst-20799	57	15	neural	neural	ADJ
ajst-20799	57	16	networks	network	NOUN
ajst-20799	57	17	(	(	PUNCT
ajst-20799	57	18	cnns	cnns	PROPN
ajst-20799	57	19	)	)	PUNCT
ajst-20799	57	20	in	in	ADP
ajst-20799	57	21	deep	deep	ADJ
ajst-20799	57	22	learning	learning	NOUN
ajst-20799	57	23	.	.	PUNCT
ajst-20799	58	1	it	it	PRON
ajst-20799	58	2	designs	design	VERB
ajst-20799	58	3	specific	specific	ADJ
ajst-20799	58	4	feature	feature	NOUN
ajst-20799	58	5	extraction	extraction	NOUN
ajst-20799	58	6	network	network	NOUN
ajst-20799	58	7	structures	structure	NOUN
ajst-20799	58	8	to	to	PART
ajst-20799	58	9	address	address	VERB
ajst-20799	58	10	the	the	DET
ajst-20799	58	11	challenges	challenge	NOUN
ajst-20799	58	12	and	and	CCONJ
ajst-20799	58	13	requirements	requirement	NOUN
ajst-20799	58	14	of	of	ADP
ajst-20799	58	15	multiple	multiple	ADJ
ajst-20799	58	16	-	-	PUNCT
ajst-20799	58	17	object	object	NOUN
ajst-20799	58	18	tracking	tracking	NOUN
ajst-20799	58	19	algorithms	algorithm	NOUN
ajst-20799	58	20	.	.	PUNCT
ajst-20799	59	1	by	by	ADP
ajst-20799	59	2	designing	design	VERB
ajst-20799	59	3	feature	feature	NOUN
ajst-20799	59	4	extraction	extraction	NOUN
ajst-20799	59	5	networks	network	NOUN
ajst-20799	59	6	suitable	suitable	ADJ
ajst-20799	59	7	for	for	ADP
ajst-20799	59	8	complex	complex	ADJ
ajst-20799	59	9	scenes	scene	NOUN
ajst-20799	59	10	,	,	PUNCT
ajst-20799	59	11	such	such	ADJ
ajst-20799	59	12	as	as	ADP
ajst-20799	59	13	cnn	cnn	PROPN
ajst-20799	59	14	-	-	PUNCT
ajst-20799	59	15	based	base	VERB
ajst-20799	59	16	feature	feature	NOUN
ajst-20799	59	17	extraction	extraction	NOUN
ajst-20799	59	18	modules	module	NOUN
ajst-20799	59	19	,	,	PUNCT
ajst-20799	59	20	it	it	PRON
ajst-20799	59	21	can	can	AUX
ajst-20799	59	22	effectively	effectively	ADV
ajst-20799	59	23	capture	capture	VERB
ajst-20799	59	24	spatial	spatial	ADJ
ajst-20799	59	25	and	and	CCONJ
ajst-20799	59	26	temporal	temporal	ADJ
ajst-20799	59	27	information	information	NOUN
ajst-20799	59	28	of	of	ADP
ajst-20799	59	29	targets	target	NOUN
ajst-20799	59	30	,	,	PUNCT
ajst-20799	59	31	automatically	automatically	ADV
ajst-20799	59	32	extracting	extract	VERB
ajst-20799	59	33	hierarchical	hierarchical	ADJ
ajst-20799	59	34	and	and	CCONJ
ajst-20799	59	35	abstract	abstract	ADJ
ajst-20799	59	36	features	feature	NOUN
ajst-20799	59	37	from	from	ADP
ajst-20799	59	38	input	input	NOUN
ajst-20799	59	39	images	image	NOUN
ajst-20799	59	40	,	,	PUNCT
ajst-20799	59	41	covering	cover	VERB
ajst-20799	59	42	spatial	spatial	ADJ
ajst-20799	59	43	127	127	NUM
ajst-20799	59	44	layouts	layout	NOUN
ajst-20799	59	45	,	,	PUNCT
ajst-20799	59	46	textures	texture	NOUN
ajst-20799	59	47	,	,	PUNCT
ajst-20799	59	48	colors	color	NOUN
ajst-20799	59	49	,	,	PUNCT
ajst-20799	59	50	etc	etc	X
ajst-20799	59	51	.	.	X
ajst-20799	59	52	,	,	PUNCT
ajst-20799	59	53	of	of	ADP
ajst-20799	59	54	targets	target	NOUN
ajst-20799	59	55	as	as	ADV
ajst-20799	59	56	well	well	ADV
ajst-20799	59	57	as	as	ADP
ajst-20799	59	58	their	their	PRON
ajst-20799	59	59	dynamic	dynamic	ADJ
ajst-20799	59	60	features	feature	NOUN
ajst-20799	59	61	evolving	evolve	VERB
ajst-20799	59	62	over	over	ADP
ajst-20799	59	63	time	time	NOUN
ajst-20799	59	64	.	.	PUNCT
ajst-20799	60	1	furthermore	furthermore	ADV
ajst-20799	60	2	,	,	PUNCT
ajst-20799	60	3	it	it	PRON
ajst-20799	60	4	adopts	adopt	VERB
ajst-20799	60	5	target	target	NOUN
ajst-20799	60	6	matching	matching	NOUN
ajst-20799	60	7	and	and	CCONJ
ajst-20799	60	8	trajectory	trajectory	NOUN
ajst-20799	60	9	prediction	prediction	NOUN
ajst-20799	60	10	techniques	technique	NOUN
ajst-20799	60	11	,	,	PUNCT
ajst-20799	60	12	combining	combine	VERB
ajst-20799	60	13	tools	tool	NOUN
ajst-20799	60	14	like	like	ADP
ajst-20799	60	15	recurrent	recurrent	ADJ
ajst-20799	60	16	neural	neural	ADJ
ajst-20799	60	17	networks	network	NOUN
ajst-20799	60	18	(	(	PUNCT
ajst-20799	60	19	rnns	rnns	PROPN
ajst-20799	60	20	)	)	PUNCT
ajst-20799	60	21	or	or	CCONJ
ajst-20799	60	22	long	long	ADJ
ajst-20799	60	23	short	short	ADJ
ajst-20799	60	24	-	-	PUNCT
ajst-20799	60	25	term	term	NOUN
ajst-20799	60	26	memory	memory	NOUN
ajst-20799	60	27	networks	network	NOUN
ajst-20799	60	28	(	(	PUNCT
ajst-20799	60	29	lstms	lstms	ADJ
ajst-20799	60	30	)	)	PUNCT
ajst-20799	60	31	for	for	ADP
ajst-20799	60	32	sequence	sequence	NOUN
ajst-20799	60	33	modeling	modeling	NOUN
ajst-20799	60	34	.	.	PUNCT
ajst-20799	61	1	these	these	DET
ajst-20799	61	2	networks	network	NOUN
ajst-20799	61	3	can	can	AUX
ajst-20799	61	4	capture	capture	VERB
ajst-20799	61	5	the	the	DET
ajst-20799	61	6	temporal	temporal	ADJ
ajst-20799	61	7	dependencies	dependency	NOUN
ajst-20799	61	8	of	of	ADP
ajst-20799	61	9	target	target	NOUN
ajst-20799	61	10	motion	motion	NOUN
ajst-20799	61	11	,	,	PUNCT
ajst-20799	61	12	predict	predict	VERB
ajst-20799	61	13	future	future	ADJ
ajst-20799	61	14	positions	position	NOUN
ajst-20799	61	15	,	,	PUNCT
ajst-20799	61	16	speeds	speed	NOUN
ajst-20799	61	17	,	,	PUNCT
ajst-20799	61	18	directions	direction	NOUN
ajst-20799	61	19	,	,	PUNCT
ajst-20799	61	20	etc	etc	X
ajst-20799	61	21	.	.	X
ajst-20799	61	22	,	,	PUNCT
ajst-20799	61	23	based	base	VERB
ajst-20799	61	24	on	on	ADP
ajst-20799	61	25	historical	historical	ADJ
ajst-20799	61	26	state	state	NOUN
ajst-20799	61	27	information	information	NOUN
ajst-20799	61	28	,	,	PUNCT
ajst-20799	61	29	enabling	enable	VERB
ajst-20799	61	30	continuous	continuous	ADJ
ajst-20799	61	31	tracking	tracking	NOUN
ajst-20799	61	32	and	and	CCONJ
ajst-20799	61	33	prediction	prediction	NOUN
ajst-20799	61	34	of	of	ADP
ajst-20799	61	35	target	target	NOUN
ajst-20799	61	36	motion	motion	NOUN
ajst-20799	61	37	trajectories	trajectory	NOUN
ajst-20799	61	38	,	,	PUNCT
ajst-20799	61	39	thereby	thereby	ADV
ajst-20799	61	40	enhancing	enhance	VERB
ajst-20799	61	41	tracking	track	VERB
ajst-20799	61	42	stability	stability	NOUN
ajst-20799	61	43	and	and	CCONJ
ajst-20799	61	44	robustness	robustness	NOUN
ajst-20799	61	45	.	.	PUNCT
ajst-20799	62	1	for	for	ADP
ajst-20799	62	2	instance	instance	NOUN
ajst-20799	62	3	,	,	PUNCT
ajst-20799	62	4	in	in	ADP
ajst-20799	62	5	a	a	DET
ajst-20799	62	6	highway	highway	NOUN
ajst-20799	62	7	monitoring	monitoring	NOUN
ajst-20799	62	8	scenario	scenario	NOUN
ajst-20799	62	9	,	,	PUNCT
ajst-20799	62	10	a	a	DET
ajst-20799	62	11	car	car	NOUN
ajst-20799	62	12	gradually	gradually	ADV
ajst-20799	62	13	accelerates	accelerate	VERB
ajst-20799	62	14	and	and	CCONJ
ajst-20799	62	15	changes	change	NOUN
ajst-20799	62	16	lanes	lane	NOUN
ajst-20799	62	17	over	over	ADP
ajst-20799	62	18	several	several	ADJ
ajst-20799	62	19	frames	frame	NOUN
ajst-20799	62	20	of	of	ADP
ajst-20799	62	21	images	image	NOUN
ajst-20799	62	22	.	.	PUNCT
ajst-20799	63	1	an	an	DET
ajst-20799	63	2	lstm	lstm	NOUN
ajst-20799	63	3	-	-	PUNCT
ajst-20799	63	4	based	base	VERB
ajst-20799	63	5	trajectory	trajectory	NOUN
ajst-20799	63	6	prediction	prediction	NOUN
ajst-20799	63	7	model	model	NOUN
ajst-20799	63	8	can	can	AUX
ajst-20799	63	9	learn	learn	VERB
ajst-20799	63	10	the	the	DET
ajst-20799	63	11	speed	speed	NOUN
ajst-20799	63	12	trend	trend	NOUN
ajst-20799	63	13	of	of	ADP
ajst-20799	63	14	the	the	DET
ajst-20799	63	15	car	car	NOUN
ajst-20799	63	16	from	from	ADP
ajst-20799	63	17	past	past	ADJ
ajst-20799	63	18	frames	frame	NOUN
ajst-20799	63	19	and	and	CCONJ
ajst-20799	63	20	the	the	DET
ajst-20799	63	21	lane	lane	NOUN
ajst-20799	63	22	-	-	PUNCT
ajst-20799	63	23	changing	change	VERB
ajst-20799	63	24	pattern	pattern	NOUN
ajst-20799	63	25	.	.	PUNCT
ajst-20799	64	1	even	even	ADV
ajst-20799	64	2	if	if	SCONJ
ajst-20799	64	3	other	other	ADJ
ajst-20799	64	4	vehicles	vehicle	NOUN
ajst-20799	64	5	temporarily	temporarily	ADV
ajst-20799	64	6	obstruct	obstruct	VERB
ajst-20799	64	7	the	the	DET
ajst-20799	64	8	line	line	NOUN
ajst-20799	64	9	of	of	ADP
ajst-20799	64	10	sight	sight	NOUN
ajst-20799	64	11	,	,	PUNCT
ajst-20799	64	12	the	the	DET
ajst-20799	64	13	prediction	prediction	NOUN
ajst-20799	64	14	model	model	NOUN
ajst-20799	64	15	can	can	AUX
ajst-20799	64	16	maintain	maintain	VERB
ajst-20799	64	17	effective	effective	ADJ
ajst-20799	64	18	tracking	tracking	NOUN
ajst-20799	64	19	of	of	ADP
ajst-20799	64	20	the	the	DET
ajst-20799	64	21	target	target	NOUN
ajst-20799	64	22	car	car	NOUN
ajst-20799	64	23	based	base	VERB
ajst-20799	64	24	on	on	ADP
ajst-20799	64	25	the	the	DET
ajst-20799	64	26	established	establish	VERB
ajst-20799	64	27	motion	motion	NOUN
ajst-20799	64	28	model	model	NOUN
ajst-20799	64	29	.	.	PUNCT
ajst-20799	65	1	combining	combine	VERB
ajst-20799	65	2	traditional	traditional	ADJ
ajst-20799	65	3	target	target	NOUN
ajst-20799	65	4	association	association	NOUN
ajst-20799	65	5	and	and	CCONJ
ajst-20799	65	6	state	state	NOUN
ajst-20799	65	7	estimation	estimation	NOUN
ajst-20799	65	8	algorithms	algorithm	NOUN
ajst-20799	65	9	such	such	ADJ
ajst-20799	65	10	as	as	ADP
ajst-20799	65	11	kalman	kalman	NOUN
ajst-20799	65	12	filters	filter	NOUN
ajst-20799	65	13	or	or	CCONJ
ajst-20799	65	14	graph	graph	NOUN
ajst-20799	65	15	-	-	PUNCT
ajst-20799	65	16	based	base	VERB
ajst-20799	65	17	tracking	tracking	NOUN
ajst-20799	65	18	methods	method	NOUN
ajst-20799	65	19	with	with	ADP
ajst-20799	65	20	deep	deep	ADJ
ajst-20799	65	21	learning	learning	NOUN
ajst-20799	65	22	frameworks	framework	NOUN
ajst-20799	65	23	for	for	ADP
ajst-20799	65	24	probabilistic	probabilistic	ADJ
ajst-20799	65	25	reasoning	reasoning	NOUN
ajst-20799	65	26	and	and	CCONJ
ajst-20799	65	27	optimal	optimal	ADJ
ajst-20799	65	28	state	state	NOUN
ajst-20799	65	29	estimation	estimation	NOUN
ajst-20799	65	30	can	can	AUX
ajst-20799	65	31	effectively	effectively	ADV
ajst-20799	65	32	address	address	VERB
ajst-20799	65	33	issues	issue	NOUN
ajst-20799	65	34	like	like	ADP
ajst-20799	65	35	target	target	NOUN
ajst-20799	65	36	occlusion	occlusion	NOUN
ajst-20799	65	37	and	and	CCONJ
ajst-20799	65	38	intersecting	intersect	VERB
ajst-20799	65	39	motion	motion	NOUN
ajst-20799	65	40	,	,	PUNCT
ajst-20799	65	41	improving	improve	VERB
ajst-20799	65	42	the	the	DET
ajst-20799	65	43	accuracy	accuracy	NOUN
ajst-20799	65	44	and	and	CCONJ
ajst-20799	65	45	efficiency	efficiency	NOUN
ajst-20799	65	46	of	of	ADP
ajst-20799	65	47	multiple	multiple	ADJ
ajst-20799	65	48	-	-	PUNCT
ajst-20799	65	49	object	object	NOUN
ajst-20799	65	50	tracking	tracking	NOUN
ajst-20799	65	51	algorithms	algorithm	NOUN
ajst-20799	65	52	.	.	PUNCT
ajst-20799	66	1	for	for	ADP
ajst-20799	66	2	example	example	NOUN
ajst-20799	66	3	,	,	PUNCT
ajst-20799	66	4	when	when	SCONJ
ajst-20799	66	5	two	two	NUM
ajst-20799	66	6	cars	car	NOUN
ajst-20799	66	7	traveling	travel	VERB
ajst-20799	66	8	in	in	ADP
ajst-20799	66	9	the	the	DET
ajst-20799	66	10	same	same	ADJ
ajst-20799	66	11	direction	direction	NOUN
ajst-20799	66	12	at	at	ADP
ajst-20799	66	13	a	a	DET
ajst-20799	66	14	crossroad	crossroad	NOUN
ajst-20799	66	15	slow	slow	NOUN
ajst-20799	66	16	down	down	ADP
ajst-20799	66	17	as	as	SCONJ
ajst-20799	66	18	they	they	PRON
ajst-20799	66	19	approach	approach	VERB
ajst-20799	66	20	,	,	PUNCT
ajst-20799	66	21	then	then	ADV
ajst-20799	66	22	almost	almost	ADV
ajst-20799	66	23	simultaneously	simultaneously	ADV
ajst-20799	66	24	enter	enter	VERB
ajst-20799	66	25	different	different	ADJ
ajst-20799	66	26	turning	turn	VERB
ajst-20799	66	27	lanes	lane	NOUN
ajst-20799	66	28	causing	cause	VERB
ajst-20799	66	29	severe	severe	ADJ
ajst-20799	66	30	visual	visual	ADJ
ajst-20799	66	31	intersection	intersection	NOUN
ajst-20799	66	32	,	,	PUNCT
ajst-20799	66	33	the	the	DET
ajst-20799	66	34	multipleobject	multipleobject	NOUN
ajst-20799	66	35	tracking	tracking	NOUN
ajst-20799	66	36	system	system	NOUN
ajst-20799	66	37	will	will	AUX
ajst-20799	66	38	construct	construct	VERB
ajst-20799	66	39	a	a	DET
ajst-20799	66	40	graph	graph	NOUN
ajst-20799	66	41	model	model	NOUN
ajst-20799	66	42	representing	represent	VERB
ajst-20799	66	43	potential	potential	ADJ
ajst-20799	66	44	correlations	correlation	NOUN
ajst-20799	66	45	between	between	ADP
ajst-20799	66	46	targets	target	NOUN
ajst-20799	66	47	.	.	PUNCT
ajst-20799	67	1	it	it	PRON
ajst-20799	67	2	combines	combine	VERB
ajst-20799	67	3	each	each	DET
ajst-20799	67	4	target	target	NOUN
ajst-20799	67	5	's	's	PART
ajst-20799	67	6	observation	observation	NOUN
ajst-20799	67	7	information	information	NOUN
ajst-20799	67	8	(	(	PUNCT
ajst-20799	67	9	e.g.	e.g.	ADV
ajst-20799	67	10	,	,	PUNCT
ajst-20799	67	11	detection	detection	NOUN
ajst-20799	67	12	box	box	NOUN
ajst-20799	67	13	position	position	NOUN
ajst-20799	67	14	,	,	PUNCT
ajst-20799	67	15	size	size	NOUN
ajst-20799	67	16	)	)	PUNCT
ajst-20799	67	17	with	with	ADP
ajst-20799	67	18	the	the	DET
ajst-20799	67	19	state	state	NOUN
ajst-20799	67	20	prediction	prediction	NOUN
ajst-20799	67	21	provided	provide	VERB
ajst-20799	67	22	by	by	ADP
ajst-20799	67	23	kalman	kalman	NOUN
ajst-20799	67	24	filters	filter	NOUN
ajst-20799	67	25	to	to	PART
ajst-20799	67	26	determine	determine	VERB
ajst-20799	67	27	the	the	DET
ajst-20799	67	28	true	true	ADJ
ajst-20799	67	29	identity	identity	NOUN
ajst-20799	67	30	of	of	ADP
ajst-20799	67	31	each	each	DET
ajst-20799	67	32	detection	detection	NOUN
ajst-20799	67	33	box	box	NOUN
ajst-20799	67	34	in	in	ADP
ajst-20799	67	35	the	the	DET
ajst-20799	67	36	current	current	ADJ
ajst-20799	67	37	frame	frame	NOUN
ajst-20799	67	38	via	via	ADP
ajst-20799	67	39	maximum	maximum	ADJ
ajst-20799	67	40	a	a	DET
ajst-20799	67	41	posteriori	posteriori	NOUN
ajst-20799	67	42	(	(	PUNCT
ajst-20799	67	43	map	map	NOUN
ajst-20799	67	44	)	)	PUNCT
ajst-20799	67	45	inference	inference	NOUN
ajst-20799	67	46	,	,	PUNCT
ajst-20799	67	47	as	as	ADV
ajst-20799	67	48	well	well	ADV
ajst-20799	67	49	as	as	ADP
ajst-20799	67	50	their	their	PRON
ajst-20799	67	51	correct	correct	ADJ
ajst-20799	67	52	trajectories	trajectory	NOUN
ajst-20799	67	53	after	after	ADP
ajst-20799	67	54	the	the	DET
ajst-20799	67	55	intersecting	intersecting	ADJ
ajst-20799	67	56	motion	motion	NOUN
ajst-20799	67	57	.	.	PUNCT
ajst-20799	68	1	3.2	3.2	NUM
ajst-20799	68	2	.	.	PUNCT
ajst-20799	68	3	application	application	NOUN
ajst-20799	68	4	of	of	ADP
ajst-20799	68	5	deep	deep	ADJ
ajst-20799	68	6	learning	learning	NOUN
ajst-20799	68	7	in	in	ADP
ajst-20799	68	8	vehicle	vehicle	NOUN
ajst-20799	68	9	tracking	tracking	NOUN
ajst-20799	68	10	in	in	ADP
ajst-20799	68	11	the	the	DET
ajst-20799	68	12	combination	combination	NOUN
ajst-20799	68	13	of	of	ADP
ajst-20799	68	14	deep	deep	ADJ
ajst-20799	68	15	learning	learning	NOUN
ajst-20799	68	16	and	and	CCONJ
ajst-20799	68	17	tracking	tracking	NOUN
ajst-20799	68	18	algorithms	algorithm	NOUN
ajst-20799	68	19	,	,	PUNCT
ajst-20799	68	20	deepsort	deepsort	NOUN
ajst-20799	68	21	is	be	AUX
ajst-20799	68	22	a	a	DET
ajst-20799	68	23	deep	deep	ADJ
ajst-20799	68	24	learning	learning	NOUN
ajst-20799	68	25	-	-	PUNCT
ajst-20799	68	26	based	base	VERB
ajst-20799	68	27	online	online	ADJ
ajst-20799	68	28	multiobject	multiobject	NOUN
ajst-20799	68	29	tracking	tracking	NOUN
ajst-20799	68	30	method	method	NOUN
ajst-20799	68	31	.	.	PUNCT
ajst-20799	69	1	its	its	PRON
ajst-20799	69	2	core	core	ADJ
ajst-20799	69	3	idea	idea	NOUN
ajst-20799	69	4	is	be	AUX
ajst-20799	69	5	to	to	PART
ajst-20799	69	6	integrate	integrate	VERB
ajst-20799	69	7	the	the	DET
ajst-20799	69	8	representation	representation	NOUN
ajst-20799	69	9	power	power	NOUN
ajst-20799	69	10	of	of	ADP
ajst-20799	69	11	deep	deep	ADJ
ajst-20799	69	12	learning	learning	NOUN
ajst-20799	69	13	with	with	ADP
ajst-20799	69	14	traditional	traditional	ADJ
ajst-20799	69	15	techniques	technique	NOUN
ajst-20799	69	16	like	like	ADP
ajst-20799	69	17	kalman	kalman	NOUN
ajst-20799	69	18	filtering	filtering	NOUN
ajst-20799	69	19	and	and	CCONJ
ajst-20799	69	20	the	the	DET
ajst-20799	69	21	hungarian	hungarian	ADJ
ajst-20799	69	22	algorithm	algorithm	NOUN
ajst-20799	69	23	to	to	PART
ajst-20799	69	24	achieve	achieve	VERB
ajst-20799	69	25	robust	robust	ADJ
ajst-20799	69	26	target	target	NOUN
ajst-20799	69	27	tracking	tracking	NOUN
ajst-20799	69	28	.	.	PUNCT
ajst-20799	70	1	deepsort	deepsort	NOUN
ajst-20799	70	2	combines	combine	VERB
ajst-20799	70	3	the	the	DET
ajst-20799	70	4	characteristics	characteristic	NOUN
ajst-20799	70	5	of	of	ADP
ajst-20799	70	6	convolutional	convolutional	ADJ
ajst-20799	70	7	neural	neural	ADJ
ajst-20799	70	8	networks	network	NOUN
ajst-20799	70	9	(	(	PUNCT
ajst-20799	70	10	cnns	cnns	PROPN
ajst-20799	70	11	)	)	PUNCT
ajst-20799	70	12	and	and	CCONJ
ajst-20799	70	13	recurrent	recurrent	ADJ
ajst-20799	70	14	neural	neural	ADJ
ajst-20799	70	15	networks	network	NOUN
ajst-20799	70	16	(	(	PUNCT
ajst-20799	70	17	rnns	rnns	PROPN
ajst-20799	70	18	)	)	PUNCT
ajst-20799	70	19	,	,	PUNCT
ajst-20799	70	20	enabling	enable	VERB
ajst-20799	70	21	feature	feature	NOUN
ajst-20799	70	22	extraction	extraction	NOUN
ajst-20799	70	23	and	and	CCONJ
ajst-20799	70	24	historical	historical	ADJ
ajst-20799	70	25	state	state	NOUN
ajst-20799	70	26	modeling	modeling	NOUN
ajst-20799	70	27	of	of	ADP
ajst-20799	70	28	targets	target	NOUN
ajst-20799	70	29	.	.	PUNCT
ajst-20799	71	1	specifically	specifically	ADV
ajst-20799	71	2	,	,	PUNCT
ajst-20799	71	3	it	it	PRON
ajst-20799	71	4	utilizes	utilize	VERB
ajst-20799	71	5	pre	pre	ADJ
ajst-20799	71	6	-	-	ADJ
ajst-20799	71	7	trained	train	VERB
ajst-20799	71	8	cnns	cnn	NOUN
ajst-20799	71	9	(	(	PUNCT
ajst-20799	71	10	e.g.	e.g.	ADV
ajst-20799	71	11	,	,	PUNCT
ajst-20799	71	12	resnet	resnet	NOUN
ajst-20799	71	13	)	)	PUNCT
ajst-20799	71	14	to	to	PART
ajst-20799	71	15	extract	extract	VERB
ajst-20799	71	16	appearance	appearance	NOUN
ajst-20799	71	17	features	feature	NOUN
ajst-20799	71	18	of	of	ADP
ajst-20799	71	19	targets	target	NOUN
ajst-20799	71	20	(	(	PUNCT
ajst-20799	71	21	strong	strong	ADJ
ajst-20799	71	22	invariance	invariance	NOUN
ajst-20799	71	23	to	to	ADP
ajst-20799	71	24	lighting	lighting	NOUN
ajst-20799	71	25	,	,	PUNCT
ajst-20799	71	26	pose	pose	VERB
ajst-20799	71	27	,	,	PUNCT
ajst-20799	71	28	partial	partial	ADJ
ajst-20799	71	29	occlusion	occlusion	NOUN
ajst-20799	71	30	,	,	PUNCT
ajst-20799	71	31	etc	etc	X
ajst-20799	71	32	.	.	X
ajst-20799	71	33	)	)	PUNCT
ajst-20799	71	34	.	.	PUNCT
ajst-20799	72	1	in	in	ADP
ajst-20799	72	2	each	each	DET
ajst-20799	72	3	frame	frame	NOUN
ajst-20799	72	4	,	,	PUNCT
ajst-20799	72	5	deepsort	deepsort	NOUN
ajst-20799	72	6	first	first	ADV
ajst-20799	72	7	uses	use	VERB
ajst-20799	72	8	the	the	DET
ajst-20799	72	9	cnn	cnn	PROPN
ajst-20799	72	10	network	network	NOUN
ajst-20799	72	11	to	to	PART
ajst-20799	72	12	extract	extract	VERB
ajst-20799	72	13	features	feature	NOUN
ajst-20799	72	14	of	of	ADP
ajst-20799	72	15	targets	target	NOUN
ajst-20799	72	16	,	,	PUNCT
ajst-20799	72	17	then	then	ADV
ajst-20799	72	18	models	model	NOUN
ajst-20799	72	19	and	and	CCONJ
ajst-20799	72	20	predicts	predict	VERB
ajst-20799	72	21	the	the	DET
ajst-20799	72	22	motion	motion	NOUN
ajst-20799	72	23	trajectory	trajectory	NOUN
ajst-20799	72	24	of	of	ADP
ajst-20799	72	25	targets	target	NOUN
ajst-20799	72	26	using	use	VERB
ajst-20799	72	27	the	the	DET
ajst-20799	72	28	rnn	rnn	NOUN
ajst-20799	72	29	network	network	NOUN
ajst-20799	72	30	.	.	PUNCT
ajst-20799	73	1	it	it	PRON
ajst-20799	73	2	uses	use	VERB
ajst-20799	73	3	a	a	DET
ajst-20799	73	4	metric	metric	ADJ
ajst-20799	73	5	learning	learn	VERB
ajst-20799	73	6	network	network	NOUN
ajst-20799	73	7	to	to	PART
ajst-20799	73	8	calculate	calculate	VERB
ajst-20799	73	9	the	the	DET
ajst-20799	73	10	similarity	similarity	NOUN
ajst-20799	73	11	between	between	ADP
ajst-20799	73	12	newly	newly	ADV
ajst-20799	73	13	detected	detect	VERB
ajst-20799	73	14	targets	target	NOUN
ajst-20799	73	15	and	and	CCONJ
ajst-20799	73	16	existing	exist	VERB
ajst-20799	73	17	trajectories	trajectory	NOUN
ajst-20799	73	18	,	,	PUNCT
ajst-20799	73	19	combined	combine	VERB
ajst-20799	73	20	with	with	ADP
ajst-20799	73	21	the	the	DET
ajst-20799	73	22	hungarian	hungarian	ADJ
ajst-20799	73	23	algorithm	algorithm	NOUN
ajst-20799	73	24	for	for	ADP
ajst-20799	73	25	data	data	PROPN
ajst-20799	73	26	association	association	NOUN
ajst-20799	73	27	,	,	PUNCT
ajst-20799	73	28	thereby	thereby	ADV
ajst-20799	73	29	achieving	achieve	VERB
ajst-20799	73	30	accurate	accurate	ADJ
ajst-20799	73	31	tracking	tracking	NOUN
ajst-20799	73	32	and	and	CCONJ
ajst-20799	73	33	state	state	NOUN
ajst-20799	73	34	prediction	prediction	NOUN
ajst-20799	73	35	of	of	ADP
ajst-20799	73	36	multiple	multiple	ADJ
ajst-20799	73	37	targets	target	NOUN
ajst-20799	73	38	.	.	PUNCT
ajst-20799	74	1	deepsort	deepsort	NOUN
ajst-20799	74	2	maintains	maintain	VERB
ajst-20799	74	3	a	a	DET
ajst-20799	74	4	historical	historical	ADJ
ajst-20799	74	5	feature	feature	NOUN
ajst-20799	74	6	queue	queue	NOUN
ajst-20799	74	7	for	for	ADP
ajst-20799	74	8	each	each	DET
ajst-20799	74	9	target	target	NOUN
ajst-20799	74	10	,	,	PUNCT
ajst-20799	74	11	so	so	CCONJ
ajst-20799	74	12	when	when	SCONJ
ajst-20799	74	13	a	a	DET
ajst-20799	74	14	target	target	NOUN
ajst-20799	74	15	briefly	briefly	NOUN
ajst-20799	74	16	disappears	disappear	VERB
ajst-20799	74	17	and	and	CCONJ
ajst-20799	74	18	reappears	reappear	VERB
ajst-20799	74	19	,	,	PUNCT
ajst-20799	74	20	it	it	PRON
ajst-20799	74	21	can	can	AUX
ajst-20799	74	22	compare	compare	VERB
ajst-20799	74	23	the	the	DET
ajst-20799	74	24	current	current	ADJ
ajst-20799	74	25	detected	detect	VERB
ajst-20799	74	26	target	target	NOUN
ajst-20799	74	27	features	feature	VERB
ajst-20799	74	28	with	with	ADP
ajst-20799	74	29	the	the	DET
ajst-20799	74	30	historical	historical	ADJ
ajst-20799	74	31	feature	feature	NOUN
ajst-20799	74	32	queue	queue	NOUN
ajst-20799	74	33	,	,	PUNCT
ajst-20799	74	34	leveraging	leverage	VERB
ajst-20799	74	35	static	static	ADJ
ajst-20799	74	36	appearance	appearance	NOUN
ajst-20799	74	37	features	feature	NOUN
ajst-20799	74	38	extracted	extract	VERB
ajst-20799	74	39	by	by	ADP
ajst-20799	74	40	deep	deep	ADJ
ajst-20799	74	41	learning	learning	NOUN
ajst-20799	74	42	instead	instead	ADV
ajst-20799	74	43	of	of	ADP
ajst-20799	74	44	dynamic	dynamic	ADJ
ajst-20799	74	45	sequence	sequence	NOUN
ajst-20799	74	46	modeling	model	VERB
ajst-20799	74	47	through	through	ADP
ajst-20799	74	48	rnns	rnn	NOUN
ajst-20799	74	49	.	.	PUNCT
ajst-20799	75	1	apart	apart	ADV
ajst-20799	75	2	from	from	ADP
ajst-20799	75	3	deepsort	deepsort	NOUN
ajst-20799	75	4	,	,	PUNCT
ajst-20799	75	5	there	there	PRON
ajst-20799	75	6	are	be	VERB
ajst-20799	75	7	other	other	ADJ
ajst-20799	75	8	deep	deep	ADJ
ajst-20799	75	9	learning	learning	NOUN
ajst-20799	75	10	-	-	PUNCT
ajst-20799	75	11	based	base	VERB
ajst-20799	75	12	multi	multi	ADJ
ajst-20799	75	13	-	-	ADJ
ajst-20799	75	14	object	object	ADJ
ajst-20799	75	15	tracking	tracking	NOUN
ajst-20799	75	16	methods	method	NOUN
ajst-20799	75	17	like	like	ADP
ajst-20799	75	18	tracktor	tracktor	NOUN
ajst-20799	75	19	.	.	PUNCT
ajst-20799	76	1	tracktor	tracktor	NOUN
ajst-20799	76	2	is	be	AUX
ajst-20799	76	3	an	an	DET
ajst-20799	76	4	innovative	innovative	ADJ
ajst-20799	76	5	multi	multi	ADJ
ajst-20799	76	6	-	-	ADJ
ajst-20799	76	7	object	object	ADJ
ajst-20799	76	8	tracking	tracking	NOUN
ajst-20799	76	9	algorithm	algorithm	NOUN
ajst-20799	76	10	that	that	PRON
ajst-20799	76	11	combines	combine	VERB
ajst-20799	76	12	the	the	DET
ajst-20799	76	13	feature	feature	NOUN
ajst-20799	76	14	learning	learn	VERB
ajst-20799	76	15	ability	ability	NOUN
ajst-20799	76	16	of	of	ADP
ajst-20799	76	17	deep	deep	ADJ
ajst-20799	76	18	learning	learning	NOUN
ajst-20799	76	19	with	with	ADP
ajst-20799	76	20	traditional	traditional	ADJ
ajst-20799	76	21	tracking	tracking	NOUN
ajst-20799	76	22	optimization	optimization	NOUN
ajst-20799	76	23	strategies	strategy	NOUN
ajst-20799	76	24	.	.	PUNCT
ajst-20799	77	1	its	its	PRON
ajst-20799	77	2	effectiveness	effectiveness	NOUN
ajst-20799	77	3	depends	depend	VERB
ajst-20799	77	4	on	on	ADP
ajst-20799	77	5	whether	whether	SCONJ
ajst-20799	77	6	the	the	DET
ajst-20799	77	7	model	model	NOUN
ajst-20799	77	8	can	can	AUX
ajst-20799	77	9	extract	extract	VERB
ajst-20799	77	10	task	task	NOUN
ajst-20799	77	11	-	-	PUNCT
ajst-20799	77	12	relevant	relevant	ADJ
ajst-20799	77	13	and	and	CCONJ
ajst-20799	77	14	discriminative	discriminative	NOUN
ajst-20799	77	15	information	information	NOUN
ajst-20799	77	16	from	from	ADP
ajst-20799	77	17	raw	raw	ADJ
ajst-20799	77	18	data	datum	NOUN
ajst-20799	77	19	(	(	PUNCT
ajst-20799	77	20	such	such	ADJ
ajst-20799	77	21	as	as	ADP
ajst-20799	77	22	images	image	NOUN
ajst-20799	77	23	or	or	CCONJ
ajst-20799	77	24	video	video	NOUN
ajst-20799	77	25	frames	frame	NOUN
ajst-20799	77	26	)	)	PUNCT
ajst-20799	77	27	.	.	PUNCT
ajst-20799	78	1	the	the	DET
ajst-20799	78	2	tracktor	tracktor	NOUN
ajst-20799	78	3	algorithm	algorithm	NOUN
ajst-20799	78	4	achieves	achieve	VERB
ajst-20799	78	5	stable	stable	ADJ
ajst-20799	78	6	tracking	tracking	NOUN
ajst-20799	78	7	and	and	CCONJ
ajst-20799	78	8	state	state	NOUN
ajst-20799	78	9	estimation	estimation	NOUN
ajst-20799	78	10	of	of	ADP
ajst-20799	78	11	targets	target	NOUN
ajst-20799	78	12	through	through	ADP
ajst-20799	78	13	effective	effective	ADJ
ajst-20799	78	14	feature	feature	NOUN
ajst-20799	78	15	representation	representation	NOUN
ajst-20799	78	16	and	and	CCONJ
ajst-20799	78	17	target	target	NOUN
ajst-20799	78	18	matching	matching	NOUN
ajst-20799	78	19	strategies	strategy	NOUN
ajst-20799	78	20	.	.	PUNCT
ajst-20799	79	1	by	by	ADP
ajst-20799	79	2	carefully	carefully	ADV
ajst-20799	79	3	designing	design	VERB
ajst-20799	79	4	network	network	NOUN
ajst-20799	79	5	structures	structure	NOUN
ajst-20799	79	6	and	and	CCONJ
ajst-20799	79	7	training	training	NOUN
ajst-20799	79	8	strategies	strategy	NOUN
ajst-20799	79	9	,	,	PUNCT
ajst-20799	79	10	tracktor	tracktor	NOUN
ajst-20799	79	11	ensures	ensure	VERB
ajst-20799	79	12	that	that	PRON
ajst-20799	79	13	learned	learn	VERB
ajst-20799	79	14	features	feature	NOUN
ajst-20799	79	15	have	have	VERB
ajst-20799	79	16	good	good	ADJ
ajst-20799	79	17	invariance	invariance	NOUN
ajst-20799	79	18	to	to	ADP
ajst-20799	79	19	factors	factor	NOUN
ajst-20799	79	20	like	like	ADP
ajst-20799	79	21	lighting	lighting	NOUN
ajst-20799	79	22	changes	change	NOUN
ajst-20799	79	23	,	,	PUNCT
ajst-20799	79	24	viewpoint	viewpoint	NOUN
ajst-20799	79	25	variations	variation	NOUN
ajst-20799	79	26	,	,	PUNCT
ajst-20799	79	27	and	and	CCONJ
ajst-20799	79	28	partial	partial	ADJ
ajst-20799	79	29	occlusions	occlusion	NOUN
ajst-20799	79	30	,	,	PUNCT
ajst-20799	79	31	thereby	thereby	ADV
ajst-20799	79	32	improving	improve	VERB
ajst-20799	79	33	tracking	tracking	NOUN
ajst-20799	79	34	robustness	robustness	NOUN
ajst-20799	79	35	.	.	PUNCT
ajst-20799	80	1	it	it	PRON
ajst-20799	80	2	combines	combine	VERB
ajst-20799	80	3	the	the	DET
ajst-20799	80	4	feature	feature	NOUN
ajst-20799	80	5	learning	learn	VERB
ajst-20799	80	6	ability	ability	NOUN
ajst-20799	80	7	of	of	ADP
ajst-20799	80	8	deep	deep	ADJ
ajst-20799	80	9	learning	learning	NOUN
ajst-20799	80	10	with	with	ADP
ajst-20799	80	11	tracking	track	VERB
ajst-20799	80	12	algorithm	algorithm	NOUN
ajst-20799	80	13	optimization	optimization	NOUN
ajst-20799	80	14	strategies	strategy	NOUN
ajst-20799	80	15	,	,	PUNCT
ajst-20799	80	16	making	make	VERB
ajst-20799	80	17	it	it	PRON
ajst-20799	80	18	suitable	suitable	ADJ
ajst-20799	80	19	for	for	ADP
ajst-20799	80	20	addressing	address	VERB
ajst-20799	80	21	target	target	NOUN
ajst-20799	80	22	tracking	tracking	NOUN
ajst-20799	80	23	challenges	challenge	NOUN
ajst-20799	80	24	in	in	ADP
ajst-20799	80	25	complex	complex	ADJ
ajst-20799	80	26	scenes	scene	NOUN
ajst-20799	80	27	,	,	PUNCT
ajst-20799	80	28	although	although	SCONJ
ajst-20799	80	29	there	there	PRON
ajst-20799	80	30	may	may	AUX
ajst-20799	80	31	be	be	AUX
ajst-20799	80	32	differences	difference	NOUN
ajst-20799	80	33	in	in	ADP
ajst-20799	80	34	specific	specific	ADJ
ajst-20799	80	35	implementations	implementation	NOUN
ajst-20799	80	36	and	and	CCONJ
ajst-20799	80	37	technical	technical	ADJ
ajst-20799	80	38	paths	path	NOUN
ajst-20799	80	39	.	.	PUNCT
ajst-20799	81	1	4	4	X
ajst-20799	81	2	.	.	X
ajst-20799	81	3	performance	performance	NOUN
ajst-20799	81	4	evaluation	evaluation	NOUN
ajst-20799	81	5	and	and	CCONJ
ajst-20799	81	6	comparison	comparison	NOUN
ajst-20799	81	7	of	of	ADP
ajst-20799	81	8	vehicle	vehicle	NOUN
ajst-20799	81	9	detection	detection	NOUN
ajst-20799	81	10	and	and	CCONJ
ajst-20799	81	11	tracking	tracking	NOUN
ajst-20799	81	12	techniques	technique	NOUN
ajst-20799	81	13	4.1	4.1	NUM
ajst-20799	81	14	.	.	PUNCT
ajst-20799	82	1	detection	detection	NOUN
ajst-20799	82	2	accuracy	accuracy	NOUN
ajst-20799	82	3	evaluation	evaluation	NOUN
ajst-20799	82	4	metrics	metric	NOUN
ajst-20799	82	5	the	the	DET
ajst-20799	82	6	intersection	intersection	NOUN
ajst-20799	82	7	over	over	ADP
ajst-20799	82	8	union	union	NOUN
ajst-20799	82	9	(	(	PUNCT
ajst-20799	82	10	iou	iou	NOUN
ajst-20799	82	11	)	)	PUNCT
ajst-20799	82	12	metric	metric	NOUN
ajst-20799	82	13	is	be	AUX
ajst-20799	82	14	widely	widely	ADV
ajst-20799	82	15	used	use	VERB
ajst-20799	82	16	in	in	ADP
ajst-20799	82	17	the	the	DET
ajst-20799	82	18	computer	computer	NOUN
ajst-20799	82	19	vision	vision	NOUN
ajst-20799	82	20	field	field	NOUN
ajst-20799	82	21	to	to	PART
ajst-20799	82	22	measure	measure	VERB
ajst-20799	82	23	the	the	DET
ajst-20799	82	24	degree	degree	NOUN
ajst-20799	82	25	of	of	ADP
ajst-20799	82	26	overlap	overlap	NOUN
ajst-20799	82	27	between	between	ADP
ajst-20799	82	28	the	the	DET
ajst-20799	82	29	output	output	NOUN
ajst-20799	82	30	of	of	ADP
ajst-20799	82	31	object	object	NOUN
ajst-20799	82	32	detection	detection	NOUN
ajst-20799	82	33	algorithms	algorithm	NOUN
ajst-20799	82	34	and	and	CCONJ
ajst-20799	82	35	the	the	DET
ajst-20799	82	36	ground	ground	NOUN
ajst-20799	82	37	truth	truth	NOUN
ajst-20799	82	38	bounding	bounding	NOUN
ajst-20799	82	39	boxes	box	NOUN
ajst-20799	82	40	.	.	PUNCT
ajst-20799	83	1	it	it	PRON
ajst-20799	83	2	is	be	AUX
ajst-20799	83	3	a	a	DET
ajst-20799	83	4	key	key	ADJ
ajst-20799	83	5	metric	metric	NOUN
ajst-20799	83	6	for	for	ADP
ajst-20799	83	7	assessing	assess	VERB
ajst-20799	83	8	model	model	NOUN
ajst-20799	83	9	prediction	prediction	NOUN
ajst-20799	83	10	accuracy	accuracy	NOUN
ajst-20799	83	11	,	,	PUNCT
ajst-20799	83	12	localization	localization	NOUN
ajst-20799	83	13	accuracy	accuracy	NOUN
ajst-20799	83	14	,	,	PUNCT
ajst-20799	83	15	and	and	CCONJ
ajst-20799	83	16	overall	overall	ADJ
ajst-20799	83	17	object	object	NOUN
ajst-20799	83	18	detection	detection	NOUN
ajst-20799	83	19	performance	performance	NOUN
ajst-20799	83	20	.	.	PUNCT
ajst-20799	84	1	specifically	specifically	ADV
ajst-20799	84	2	,	,	PUNCT
ajst-20799	84	3	iou	iou	PROPN
ajst-20799	84	4	is	be	AUX
ajst-20799	84	5	calculated	calculate	VERB
ajst-20799	84	6	by	by	ADP
ajst-20799	84	7	dividing	divide	VERB
ajst-20799	84	8	the	the	DET
ajst-20799	84	9	intersection	intersection	NOUN
ajst-20799	84	10	area	area	NOUN
ajst-20799	84	11	of	of	ADP
ajst-20799	84	12	the	the	DET
ajst-20799	84	13	ground	ground	NOUN
ajst-20799	84	14	truth	truth	NOUN
ajst-20799	84	15	bounding	bound	VERB
ajst-20799	84	16	box	box	NOUN
ajst-20799	84	17	and	and	CCONJ
ajst-20799	84	18	the	the	DET
ajst-20799	84	19	detected	detect	VERB
ajst-20799	84	20	bounding	bounding	NOUN
ajst-20799	84	21	box	box	NOUN
ajst-20799	84	22	by	by	ADP
ajst-20799	84	23	their	their	PRON
ajst-20799	84	24	union	union	NOUN
ajst-20799	84	25	area	area	NOUN
ajst-20799	84	26	,	,	PUNCT
ajst-20799	84	27	given	give	VERB
ajst-20799	84	28	as	as	ADP
ajst-20799	84	29	iou	iou	NOUN
ajst-20799	84	30	=	=	SYM
ajst-20799	84	31	(	(	PUNCT
ajst-20799	84	32	intersection	intersection	NOUN
ajst-20799	84	33	area	area	NOUN
ajst-20799	84	34	)	)	PUNCT
ajst-20799	84	35	/	/	PUNCT
ajst-20799	84	36	(	(	PUNCT
ajst-20799	84	37	union	union	NOUN
ajst-20799	84	38	area	area	NOUN
ajst-20799	84	39	)	)	PUNCT
ajst-20799	84	40	.	.	PUNCT
ajst-20799	85	1	the	the	DET
ajst-20799	85	2	iou	iou	PROPN
ajst-20799	85	3	value	value	NOUN
ajst-20799	85	4	ranges	range	VERB
ajst-20799	85	5	from	from	ADP
ajst-20799	85	6	0	0	NUM
ajst-20799	85	7	to	to	ADP
ajst-20799	85	8	1	1	NUM
ajst-20799	85	9	,	,	PUNCT
ajst-20799	85	10	where	where	SCONJ
ajst-20799	85	11	a	a	DET
ajst-20799	85	12	value	value	NOUN
ajst-20799	85	13	closer	close	ADV
ajst-20799	85	14	to	to	ADP
ajst-20799	85	15	1	1	NUM
ajst-20799	85	16	indicates	indicate	VERB
ajst-20799	85	17	a	a	DET
ajst-20799	85	18	higher	high	ADJ
ajst-20799	85	19	degree	degree	NOUN
ajst-20799	85	20	of	of	ADP
ajst-20799	85	21	match	match	NOUN
ajst-20799	85	22	between	between	ADP
ajst-20799	85	23	the	the	DET
ajst-20799	85	24	detection	detection	NOUN
ajst-20799	85	25	result	result	NOUN
ajst-20799	85	26	and	and	CCONJ
ajst-20799	85	27	the	the	DET
ajst-20799	85	28	ground	ground	NOUN
ajst-20799	85	29	truth	truth	NOUN
ajst-20799	85	30	.	.	PUNCT
ajst-20799	86	1	as	as	SCONJ
ajst-20799	86	2	the	the	DET
ajst-20799	86	3	iou	iou	NOUN
ajst-20799	86	4	value	value	NOUN
ajst-20799	86	5	decreases	decrease	VERB
ajst-20799	86	6	,	,	PUNCT
ajst-20799	86	7	the	the	DET
ajst-20799	86	8	overlap	overlap	NOUN
ajst-20799	86	9	between	between	ADP
ajst-20799	86	10	the	the	DET
ajst-20799	86	11	predicted	predict	VERB
ajst-20799	86	12	bounding	bounding	NOUN
ajst-20799	86	13	box	box	NOUN
ajst-20799	86	14	and	and	CCONJ
ajst-20799	86	15	the	the	DET
ajst-20799	86	16	ground	ground	NOUN
ajst-20799	86	17	truth	truth	NOUN
ajst-20799	86	18	bounding	bound	VERB
ajst-20799	86	19	box	box	NOUN
ajst-20799	86	20	decreases	decrease	VERB
ajst-20799	86	21	,	,	PUNCT
ajst-20799	86	22	leading	lead	VERB
ajst-20799	86	23	to	to	ADP
ajst-20799	86	24	reduced	reduce	VERB
ajst-20799	86	25	localization	localization	NOUN
ajst-20799	86	26	accuracy	accuracy	NOUN
ajst-20799	86	27	.	.	PUNCT
ajst-20799	87	1	predictions	prediction	NOUN
ajst-20799	87	2	with	with	ADP
ajst-20799	87	3	iou	iou	PROPN
ajst-20799	87	4	below	below	ADP
ajst-20799	87	5	a	a	DET
ajst-20799	87	6	certain	certain	ADJ
ajst-20799	87	7	threshold	threshold	NOUN
ajst-20799	87	8	(	(	PUNCT
ajst-20799	87	9	e.g.	e.g.	ADV
ajst-20799	87	10	,	,	PUNCT
ajst-20799	87	11	0.5	0.5	NUM
ajst-20799	87	12	)	)	PUNCT
ajst-20799	87	13	may	may	AUX
ajst-20799	87	14	be	be	AUX
ajst-20799	87	15	considered	consider	VERB
ajst-20799	87	16	as	as	ADP
ajst-20799	87	17	false	false	ADJ
ajst-20799	87	18	positives	positive	NOUN
ajst-20799	87	19	or	or	CCONJ
ajst-20799	87	20	false	false	ADJ
ajst-20799	87	21	negatives	negative	NOUN
ajst-20799	87	22	.	.	PUNCT
ajst-20799	88	1	in	in	ADP
ajst-20799	88	2	addition	addition	NOUN
ajst-20799	88	3	to	to	ADP
ajst-20799	88	4	the	the	DET
ajst-20799	88	5	iou	iou	NOUN
ajst-20799	88	6	metric	metric	NOUN
ajst-20799	88	7	,	,	PUNCT
ajst-20799	88	8	there	there	PRON
ajst-20799	88	9	are	be	VERB
ajst-20799	88	10	other	other	ADJ
ajst-20799	88	11	commonly	commonly	ADV
ajst-20799	88	12	used	use	VERB
ajst-20799	88	13	detection	detection	NOUN
ajst-20799	88	14	accuracy	accuracy	NOUN
ajst-20799	88	15	evaluation	evaluation	NOUN
ajst-20799	88	16	metrics	metric	NOUN
ajst-20799	88	17	such	such	ADJ
ajst-20799	88	18	as	as	ADP
ajst-20799	88	19	precision	precision	NOUN
ajst-20799	88	20	(	(	PUNCT
ajst-20799	88	21	measuring	measure	VERB
ajst-20799	88	22	the	the	DET
ajst-20799	88	23	proportion	proportion	NOUN
ajst-20799	88	24	of	of	ADP
ajst-20799	88	25	true	true	ADJ
ajst-20799	88	26	positives	positive	NOUN
ajst-20799	88	27	among	among	ADP
ajst-20799	88	28	all	all	DET
ajst-20799	88	29	samples	sample	NOUN
ajst-20799	88	30	classified	classify	VERB
ajst-20799	88	31	as	as	ADP
ajst-20799	88	32	positive	positive	ADJ
ajst-20799	88	33	by	by	ADP
ajst-20799	88	34	the	the	DET
ajst-20799	88	35	model	model	NOUN
ajst-20799	88	36	)	)	PUNCT
ajst-20799	88	37	,	,	PUNCT
ajst-20799	88	38	recall	recall	INTJ
ajst-20799	88	39	(	(	PUNCT
ajst-20799	88	40	measuring	measure	VERB
ajst-20799	88	41	the	the	DET
ajst-20799	88	42	proportion	proportion	NOUN
ajst-20799	88	43	of	of	ADP
ajst-20799	88	44	actual	actual	ADJ
ajst-20799	88	45	positive	positive	ADJ
ajst-20799	88	46	samples	sample	NOUN
ajst-20799	88	47	that	that	PRON
ajst-20799	88	48	are	be	AUX
ajst-20799	88	49	successfully	successfully	ADV
ajst-20799	88	50	detected	detect	VERB
ajst-20799	88	51	by	by	ADP
ajst-20799	88	52	the	the	DET
ajst-20799	88	53	model	model	NOUN
ajst-20799	88	54	)	)	PUNCT
ajst-20799	88	55	,	,	PUNCT
ajst-20799	88	56	and	and	CCONJ
ajst-20799	88	57	f1	f1	NOUN
ajst-20799	88	58	score	score	NOUN
ajst-20799	88	59	(	(	PUNCT
ajst-20799	88	60	the	the	DET
ajst-20799	88	61	harmonic	harmonic	ADJ
ajst-20799	88	62	mean	mean	NOUN
ajst-20799	88	63	of	of	ADP
ajst-20799	88	64	precision	precision	NOUN
ajst-20799	88	65	and	and	CCONJ
ajst-20799	88	66	recall	recall	NOUN
ajst-20799	88	67	,	,	PUNCT
ajst-20799	88	68	providing	provide	VERB
ajst-20799	88	69	a	a	DET
ajst-20799	88	70	single	single	ADJ
ajst-20799	88	71	value	value	NOUN
ajst-20799	88	72	to	to	PART
ajst-20799	88	73	comprehensively	comprehensively	ADV
ajst-20799	88	74	reflect	reflect	VERB
ajst-20799	88	75	the	the	DET
ajst-20799	88	76	model	model	NOUN
ajst-20799	88	77	's	's	PART
ajst-20799	88	78	performance	performance	NOUN
ajst-20799	88	79	in	in	ADP
ajst-20799	88	80	terms	term	NOUN
ajst-20799	88	81	of	of	ADP
ajst-20799	88	82	both	both	DET
ajst-20799	88	83	precision	precision	NOUN
ajst-20799	88	84	and	and	CCONJ
ajst-20799	88	85	recall	recall	NOUN
ajst-20799	88	86	)	)	PUNCT
ajst-20799	88	87	.	.	PUNCT
ajst-20799	89	1	precision	precision	NOUN
ajst-20799	89	2	focuses	focus	VERB
ajst-20799	89	3	on	on	ADP
ajst-20799	89	4	the	the	DET
ajst-20799	89	5	reliability	reliability	NOUN
ajst-20799	89	6	of	of	ADP
ajst-20799	89	7	the	the	DET
ajst-20799	89	8	model	model	NOUN
ajst-20799	89	9	's	's	PART
ajst-20799	89	10	predictions	prediction	NOUN
ajst-20799	89	11	,	,	PUNCT
ajst-20799	89	12	indicating	indicate	VERB
ajst-20799	89	13	the	the	DET
ajst-20799	89	14	proportion	proportion	NOUN
ajst-20799	89	15	of	of	ADP
ajst-20799	89	16	correctly	correctly	ADV
ajst-20799	89	17	identified	identify	VERB
ajst-20799	89	18	positive	positive	ADJ
ajst-20799	89	19	samples	sample	NOUN
ajst-20799	89	20	among	among	ADP
ajst-20799	89	21	the	the	DET
ajst-20799	89	22	detected	detect	VERB
ajst-20799	89	23	positives	positive	NOUN
ajst-20799	89	24	.	.	PUNCT
ajst-20799	90	1	recall	recall	PROPN
ajst-20799	90	2	emphasizes	emphasize	VERB
ajst-20799	90	3	the	the	DET
ajst-20799	90	4	comprehensiveness	comprehensiveness	NOUN
ajst-20799	90	5	of	of	ADP
ajst-20799	90	6	the	the	DET
ajst-20799	90	7	model	model	NOUN
ajst-20799	90	8	in	in	ADP
ajst-20799	90	9	finding	find	VERB
ajst-20799	90	10	targets	target	NOUN
ajst-20799	90	11	,	,	PUNCT
ajst-20799	90	12	representing	represent	VERB
ajst-20799	90	13	the	the	DET
ajst-20799	90	14	proportion	proportion	NOUN
ajst-20799	90	15	of	of	ADP
ajst-20799	90	16	all	all	DET
ajst-20799	90	17	true	true	ADJ
ajst-20799	90	18	positive	positive	ADJ
ajst-20799	90	19	samples	sample	NOUN
ajst-20799	90	20	that	that	PRON
ajst-20799	90	21	are	be	AUX
ajst-20799	90	22	detected	detect	VERB
ajst-20799	90	23	.	.	PUNCT
ajst-20799	91	1	f1	f1	NOUN
ajst-20799	91	2	score	score	NOUN
ajst-20799	91	3	balances	balance	NOUN
ajst-20799	91	4	precision	precision	NOUN
ajst-20799	91	5	and	and	CCONJ
ajst-20799	91	6	recall	recall	NOUN
ajst-20799	91	7	,	,	PUNCT
ajst-20799	91	8	providing	provide	VERB
ajst-20799	91	9	a	a	DET
ajst-20799	91	10	comprehensive	comprehensive	ADJ
ajst-20799	91	11	evaluation	evaluation	NOUN
ajst-20799	91	12	metric	metric	NOUN
ajst-20799	91	13	that	that	PRON
ajst-20799	91	14	considers	consider	VERB
ajst-20799	91	15	both	both	CCONJ
ajst-20799	91	16	the	the	DET
ajst-20799	91	17	accuracy	accuracy	NOUN
ajst-20799	91	18	and	and	CCONJ
ajst-20799	91	19	completeness	completeness	NOUN
ajst-20799	91	20	of	of	ADP
ajst-20799	91	21	the	the	DET
ajst-20799	91	22	algorithm	algorithm	NOUN
ajst-20799	91	23	.	.	PUNCT
ajst-20799	92	1	in	in	ADP
ajst-20799	92	2	practical	practical	ADJ
ajst-20799	92	3	applications	application	NOUN
ajst-20799	92	4	such	such	ADJ
ajst-20799	92	5	as	as	ADP
ajst-20799	92	6	vehicle	vehicle	NOUN
ajst-20799	92	7	detection	detection	NOUN
ajst-20799	92	8	and	and	CCONJ
ajst-20799	92	9	tracking	tracking	NOUN
ajst-20799	92	10	technologies	technology	NOUN
ajst-20799	92	11	,	,	PUNCT
ajst-20799	92	12	the	the	DET
ajst-20799	92	13	iou	iou	NOUN
ajst-20799	92	14	metric	metric	NOUN
ajst-20799	92	15	is	be	AUX
ajst-20799	92	16	commonly	commonly	ADV
ajst-20799	92	17	used	use	VERB
ajst-20799	92	18	to	to	PART
ajst-20799	92	19	evaluate	evaluate	VERB
ajst-20799	92	20	the	the	DET
ajst-20799	92	21	accuracy	accuracy	NOUN
ajst-20799	92	22	and	and	CCONJ
ajst-20799	92	23	robustness	robustness	NOUN
ajst-20799	92	24	of	of	ADP
ajst-20799	92	25	object	object	NOUN
ajst-20799	92	26	detection	detection	NOUN
ajst-20799	92	27	algorithms	algorithm	NOUN
ajst-20799	92	28	.	.	PUNCT
ajst-20799	93	1	it	it	PRON
ajst-20799	93	2	quantifies	quantify	VERB
ajst-20799	93	3	the	the	DET
ajst-20799	93	4	overlap	overlap	NOUN
ajst-20799	93	5	between	between	ADP
ajst-20799	93	6	the	the	DET
ajst-20799	93	7	model	model	NOUN
ajst-20799	93	8	's	's	PART
ajst-20799	93	9	predicted	predict	VERB
ajst-20799	93	10	bounding	bounding	NOUN
ajst-20799	93	11	boxes	box	NOUN
ajst-20799	93	12	and	and	CCONJ
ajst-20799	93	13	the	the	DET
ajst-20799	93	14	actual	actual	ADJ
ajst-20799	93	15	ground	ground	NOUN
ajst-20799	93	16	truth	truth	NOUN
ajst-20799	93	17	bounding	bounding	NOUN
ajst-20799	93	18	boxes	box	NOUN
ajst-20799	93	19	,	,	PUNCT
ajst-20799	93	20	providing	provide	VERB
ajst-20799	93	21	an	an	DET
ajst-20799	93	22	objective	objective	NOUN
ajst-20799	93	23	and	and	CCONJ
ajst-20799	93	24	consistent	consistent	ADJ
ajst-20799	93	25	128	128	NUM
ajst-20799	93	26	standard	standard	NOUN
ajst-20799	93	27	for	for	ADP
ajst-20799	93	28	assessing	assess	VERB
ajst-20799	93	29	the	the	DET
ajst-20799	93	30	performance	performance	NOUN
ajst-20799	93	31	of	of	ADP
ajst-20799	93	32	different	different	ADJ
ajst-20799	93	33	object	object	NOUN
ajst-20799	93	34	detection	detection	NOUN
ajst-20799	93	35	algorithms	algorithm	NOUN
ajst-20799	93	36	.	.	PUNCT
ajst-20799	94	1	by	by	ADP
ajst-20799	94	2	calculating	calculate	VERB
ajst-20799	94	3	the	the	DET
ajst-20799	94	4	iou	iou	NOUN
ajst-20799	94	5	metric	metric	NOUN
ajst-20799	94	6	,	,	PUNCT
ajst-20799	94	7	the	the	DET
ajst-20799	94	8	degree	degree	NOUN
ajst-20799	94	9	of	of	ADP
ajst-20799	94	10	match	match	NOUN
ajst-20799	94	11	between	between	ADP
ajst-20799	94	12	the	the	DET
ajst-20799	94	13	detection	detection	NOUN
ajst-20799	94	14	results	result	NOUN
ajst-20799	94	15	and	and	CCONJ
ajst-20799	94	16	the	the	DET
ajst-20799	94	17	ground	ground	NOUN
ajst-20799	94	18	truth	truth	NOUN
ajst-20799	94	19	can	can	AUX
ajst-20799	94	20	be	be	AUX
ajst-20799	94	21	determined	determine	VERB
ajst-20799	94	22	,	,	PUNCT
ajst-20799	94	23	enabling	enable	VERB
ajst-20799	94	24	performance	performance	NOUN
ajst-20799	94	25	evaluation	evaluation	NOUN
ajst-20799	94	26	and	and	CCONJ
ajst-20799	94	27	comparison	comparison	NOUN
ajst-20799	94	28	of	of	ADP
ajst-20799	94	29	algorithms	algorithm	NOUN
ajst-20799	94	30	.	.	PUNCT
ajst-20799	95	1	4.2	4.2	NUM
ajst-20799	95	2	.	.	PUNCT
ajst-20799	96	1	performance	performance	NOUN
ajst-20799	96	2	comparison	comparison	NOUN
ajst-20799	96	3	analysis	analysis	NOUN
ajst-20799	96	4	of	of	ADP
ajst-20799	96	5	different	different	ADJ
ajst-20799	96	6	algorithms	algorithm	NOUN
ajst-20799	96	7	for	for	ADP
ajst-20799	96	8	vehicle	vehicle	NOUN
ajst-20799	96	9	detection	detection	NOUN
ajst-20799	96	10	tasks	task	NOUN
ajst-20799	96	11	,	,	PUNCT
ajst-20799	96	12	commonly	commonly	ADV
ajst-20799	96	13	used	use	VERB
ajst-20799	96	14	deep	deep	ADJ
ajst-20799	96	15	learning	learning	NOUN
ajst-20799	96	16	methods	method	NOUN
ajst-20799	96	17	include	include	VERB
ajst-20799	96	18	faster	fast	ADJ
ajst-20799	96	19	r	r	NOUN
ajst-20799	96	20	-	-	PUNCT
ajst-20799	96	21	cnn	cnn	PROPN
ajst-20799	96	22	,	,	PUNCT
ajst-20799	96	23	yolo	yolo	PROPN
ajst-20799	96	24	(	(	PUNCT
ajst-20799	96	25	you	you	PRON
ajst-20799	96	26	only	only	ADV
ajst-20799	96	27	look	look	VERB
ajst-20799	96	28	once	once	ADV
ajst-20799	96	29	)	)	PUNCT
ajst-20799	96	30	,	,	PUNCT
ajst-20799	96	31	and	and	CCONJ
ajst-20799	96	32	ssd	ssd	NOUN
ajst-20799	96	33	(	(	PUNCT
ajst-20799	96	34	single	single	ADJ
ajst-20799	96	35	shot	shot	NOUN
ajst-20799	96	36	multibox	multibox	PROPN
ajst-20799	96	37	detector).[4	detector).[4	PROPN
ajst-20799	96	38	]	]	PUNCT
ajst-20799	96	39	faster	fast	ADV
ajst-20799	96	40	r	r	NOUN
ajst-20799	96	41	-	-	PUNCT
ajst-20799	96	42	cnn	cnn	PROPN
ajst-20799	96	43	excels	excel	NOUN
ajst-20799	96	44	in	in	ADP
ajst-20799	96	45	accuracy	accuracy	NOUN
ajst-20799	96	46	,	,	PUNCT
ajst-20799	96	47	especially	especially	ADV
ajst-20799	96	48	for	for	ADP
ajst-20799	96	49	detecting	detect	VERB
ajst-20799	96	50	small	small	ADJ
ajst-20799	96	51	objects	object	NOUN
ajst-20799	96	52	and	and	CCONJ
ajst-20799	96	53	in	in	ADP
ajst-20799	96	54	complex	complex	ADJ
ajst-20799	96	55	scenarios	scenario	NOUN
ajst-20799	96	56	.	.	PUNCT
ajst-20799	97	1	its	its	PRON
ajst-20799	97	2	two	two	NUM
ajst-20799	97	3	-	-	PUNCT
ajst-20799	97	4	stage	stage	NOUN
ajst-20799	97	5	design	design	NOUN
ajst-20799	97	6	allows	allow	VERB
ajst-20799	97	7	for	for	ADP
ajst-20799	97	8	multiple	multiple	ADJ
ajst-20799	97	9	iterations	iteration	NOUN
ajst-20799	97	10	of	of	ADP
ajst-20799	97	11	optimization	optimization	NOUN
ajst-20799	97	12	within	within	ADP
ajst-20799	97	13	the	the	DET
ajst-20799	97	14	network	network	NOUN
ajst-20799	97	15	,	,	PUNCT
ajst-20799	97	16	thus	thus	ADV
ajst-20799	97	17	improving	improve	VERB
ajst-20799	97	18	localization	localization	NOUN
ajst-20799	97	19	accuracy	accuracy	NOUN
ajst-20799	97	20	.	.	PUNCT
ajst-20799	98	1	yolo	yolo	PROPN
ajst-20799	98	2	boasts	boast	VERB
ajst-20799	98	3	excellent	excellent	ADJ
ajst-20799	98	4	real	real	ADJ
ajst-20799	98	5	-	-	PUNCT
ajst-20799	98	6	time	time	NOUN
ajst-20799	98	7	performance	performance	NOUN
ajst-20799	98	8	and	and	CCONJ
ajst-20799	98	9	ease	ease	NOUN
ajst-20799	98	10	of	of	ADP
ajst-20799	98	11	deployment	deployment	NOUN
ajst-20799	98	12	,	,	PUNCT
ajst-20799	98	13	making	make	VERB
ajst-20799	98	14	it	it	PRON
ajst-20799	98	15	suitable	suitable	ADJ
ajst-20799	98	16	for	for	ADP
ajst-20799	98	17	large	large	ADJ
ajst-20799	98	18	-	-	PUNCT
ajst-20799	98	19	scale	scale	NOUN
ajst-20799	98	20	object	object	NOUN
ajst-20799	98	21	detection	detection	NOUN
ajst-20799	98	22	tasks	task	NOUN
ajst-20799	98	23	like	like	ADP
ajst-20799	98	24	vehicle	vehicle	NOUN
ajst-20799	98	25	detection	detection	NOUN
ajst-20799	98	26	on	on	ADP
ajst-20799	98	27	highways	highway	NOUN
ajst-20799	98	28	,	,	PUNCT
ajst-20799	98	29	where	where	SCONJ
ajst-20799	98	30	it	it	PRON
ajst-20799	98	31	often	often	ADV
ajst-20799	98	32	exhibits	exhibit	VERB
ajst-20799	98	33	good	good	ADJ
ajst-20799	98	34	performance	performance	NOUN
ajst-20799	98	35	.	.	PUNCT
ajst-20799	99	1	ssd	ssd	NOUN
ajst-20799	99	2	balances	balance	NOUN
ajst-20799	99	3	accuracy	accuracy	NOUN
ajst-20799	99	4	and	and	CCONJ
ajst-20799	99	5	speed	speed	NOUN
ajst-20799	99	6	,	,	PUNCT
ajst-20799	99	7	with	with	ADP
ajst-20799	99	8	its	its	PRON
ajst-20799	99	9	multi	multi	ADJ
ajst-20799	99	10	-	-	ADJ
ajst-20799	99	11	scale	scale	ADJ
ajst-20799	99	12	detection	detection	NOUN
ajst-20799	99	13	mechanism	mechanism	NOUN
ajst-20799	99	14	making	make	VERB
ajst-20799	99	15	it	it	PRON
ajst-20799	99	16	adaptable	adaptable	ADJ
ajst-20799	99	17	to	to	ADP
ajst-20799	99	18	different	different	ADJ
ajst-20799	99	19	sizes	size	NOUN
ajst-20799	99	20	of	of	ADP
ajst-20799	99	21	vehicles	vehicle	NOUN
ajst-20799	99	22	,	,	PUNCT
ajst-20799	99	23	especially	especially	ADV
ajst-20799	99	24	addressing	address	VERB
ajst-20799	99	25	common	common	ADJ
ajst-20799	99	26	distance	distance	NOUN
ajst-20799	99	27	and	and	CCONJ
ajst-20799	99	28	perspective	perspective	NOUN
ajst-20799	99	29	variations	variation	NOUN
ajst-20799	99	30	encountered	encounter	VERB
ajst-20799	99	31	in	in	ADP
ajst-20799	99	32	vehicle	vehicle	NOUN
ajst-20799	99	33	detection	detection	NOUN
ajst-20799	99	34	tasks	task	NOUN
ajst-20799	99	35	.	.	PUNCT
ajst-20799	100	1	for	for	ADP
ajst-20799	100	2	vehicle	vehicle	NOUN
ajst-20799	100	3	tracking	tracking	NOUN
ajst-20799	100	4	tasks	task	NOUN
ajst-20799	100	5	,	,	PUNCT
ajst-20799	100	6	commonly	commonly	ADV
ajst-20799	100	7	used	use	VERB
ajst-20799	100	8	deep	deep	ADJ
ajst-20799	100	9	learning	learning	NOUN
ajst-20799	100	10	methods	method	NOUN
ajst-20799	100	11	include	include	VERB
ajst-20799	100	12	deepsort	deepsort	NOUN
ajst-20799	100	13	(	(	PUNCT
ajst-20799	100	14	deep	deep	ADJ
ajst-20799	100	15	simple	simple	ADJ
ajst-20799	100	16	online	online	ADJ
ajst-20799	100	17	and	and	CCONJ
ajst-20799	100	18	realtime	realtime	ADJ
ajst-20799	100	19	tracking	tracking	NOUN
ajst-20799	100	20	)	)	PUNCT
ajst-20799	100	21	and	and	CCONJ
ajst-20799	100	22	tracktor	tracktor	NOUN
ajst-20799	100	23	.	.	PUNCT
ajst-20799	101	1	for	for	ADP
ajst-20799	101	2	instance	instance	NOUN
ajst-20799	101	3	,	,	PUNCT
ajst-20799	101	4	pre	pre	ADJ
ajst-20799	101	5	-	-	ADJ
ajst-20799	101	6	trained	train	VERB
ajst-20799	101	7	models	model	NOUN
ajst-20799	101	8	like	like	ADP
ajst-20799	101	9	yolo	yolo	NOUN
ajst-20799	101	10	or	or	CCONJ
ajst-20799	101	11	ssd	ssd	NOUN
ajst-20799	101	12	can	can	AUX
ajst-20799	101	13	be	be	AUX
ajst-20799	101	14	used	use	VERB
ajst-20799	101	15	for	for	ADP
ajst-20799	101	16	initial	initial	ADJ
ajst-20799	101	17	vehicle	vehicle	NOUN
ajst-20799	101	18	detection	detection	NOUN
ajst-20799	101	19	in	in	ADP
ajst-20799	101	20	images	image	NOUN
ajst-20799	101	21	,	,	PUNCT
ajst-20799	101	22	followed	follow	VERB
ajst-20799	101	23	by	by	ADP
ajst-20799	101	24	feeding	feed	VERB
ajst-20799	101	25	the	the	DET
ajst-20799	101	26	detected	detect	VERB
ajst-20799	101	27	vehicle	vehicle	NOUN
ajst-20799	101	28	regions	region	NOUN
ajst-20799	101	29	into	into	ADP
ajst-20799	101	30	specialized	specialized	PROPN
ajst-20799	101	31	cnn	cnn	PROPN
ajst-20799	101	32	models	model	NOUN
ajst-20799	101	33	such	such	ADJ
ajst-20799	101	34	as	as	ADP
ajst-20799	101	35	resnet	resnet	NOUN
ajst-20799	101	36	or	or	CCONJ
ajst-20799	101	37	mobilenet	mobilenet	NOUN
ajst-20799	101	38	for	for	ADP
ajst-20799	101	39	feature	feature	NOUN
ajst-20799	101	40	extraction	extraction	NOUN
ajst-20799	101	41	.	.	PUNCT
ajst-20799	102	1	the	the	DET
ajst-20799	102	2	deepsort	deepsort	NOUN
ajst-20799	102	3	algorithm	algorithm	NOUN
ajst-20799	102	4	combines	combine	VERB
ajst-20799	102	5	convolutional	convolutional	ADJ
ajst-20799	102	6	neural	neural	ADJ
ajst-20799	102	7	networks	network	NOUN
ajst-20799	102	8	(	(	PUNCT
ajst-20799	102	9	cnns	cnns	PROPN
ajst-20799	102	10	)	)	PUNCT
ajst-20799	102	11	and	and	CCONJ
ajst-20799	102	12	recurrent	recurrent	ADJ
ajst-20799	102	13	neural	neural	ADJ
ajst-20799	102	14	networks	network	NOUN
ajst-20799	102	15	(	(	PUNCT
ajst-20799	102	16	rnns	rnns	PROPN
ajst-20799	102	17	)	)	PUNCT
ajst-20799	102	18	,	,	PUNCT
ajst-20799	102	19	where	where	SCONJ
ajst-20799	102	20	rnns	rnn	NOUN
ajst-20799	102	21	can	can	AUX
ajst-20799	102	22	memorize	memorize	VERB
ajst-20799	102	23	feature	feature	NOUN
ajst-20799	102	24	information	information	NOUN
ajst-20799	102	25	of	of	ADP
ajst-20799	102	26	the	the	DET
ajst-20799	102	27	same	same	ADJ
ajst-20799	102	28	vehicle	vehicle	NOUN
ajst-20799	102	29	across	across	ADP
ajst-20799	102	30	past	past	PROPN
ajst-20799	102	31	frames	frame	NOUN
ajst-20799	102	32	,	,	PUNCT
ajst-20799	102	33	forming	form	VERB
ajst-20799	102	34	a	a	DET
ajst-20799	102	35	"	"	PUNCT
ajst-20799	102	36	trajectory	trajectory	NOUN
ajst-20799	102	37	embedding	embed	VERB
ajst-20799	102	38	"	"	PUNCT
ajst-20799	102	39	to	to	PART
ajst-20799	102	40	achieve	achieve	VERB
ajst-20799	102	41	accurate	accurate	ADJ
ajst-20799	102	42	tracking	tracking	NOUN
ajst-20799	102	43	through	through	ADP
ajst-20799	102	44	feature	feature	NOUN
ajst-20799	102	45	extraction	extraction	NOUN
ajst-20799	102	46	and	and	CCONJ
ajst-20799	102	47	historical	historical	ADJ
ajst-20799	102	48	state	state	NOUN
ajst-20799	102	49	modeling	modeling	NOUN
ajst-20799	102	50	.	.	PUNCT
ajst-20799	103	1	the	the	DET
ajst-20799	103	2	tracktor	tracktor	NOUN
ajst-20799	103	3	algorithm	algorithm	NOUN
ajst-20799	103	4	typically	typically	ADV
ajst-20799	103	5	extends	extend	VERB
ajst-20799	103	6	existing	exist	VERB
ajst-20799	103	7	detectors	detector	NOUN
ajst-20799	103	8	(	(	PUNCT
ajst-20799	103	9	e.g.	e.g.	ADV
ajst-20799	103	10	,	,	PUNCT
ajst-20799	103	11	faster	fast	ADJ
ajst-20799	103	12	r	r	NOUN
ajst-20799	103	13	-	-	PUNCT
ajst-20799	103	14	cnn	cnn	PROPN
ajst-20799	103	15	,	,	PUNCT
ajst-20799	103	16	mask	mask	VERB
ajst-20799	103	17	r	r	NOUN
ajst-20799	103	18	-	-	PUNCT
ajst-20799	103	19	cnn	cnn	NOUN
ajst-20799	103	20	)	)	PUNCT
ajst-20799	103	21	by	by	ADP
ajst-20799	103	22	designing	design	VERB
ajst-20799	103	23	effective	effective	ADJ
ajst-20799	103	24	feature	feature	NOUN
ajst-20799	103	25	representations	representation	NOUN
ajst-20799	103	26	and	and	CCONJ
ajst-20799	103	27	target	target	NOUN
ajst-20799	103	28	matching	matching	NOUN
ajst-20799	103	29	strategies	strategy	NOUN
ajst-20799	103	30	to	to	PART
ajst-20799	103	31	improve	improve	VERB
ajst-20799	103	32	tracking	tracking	NOUN
ajst-20799	103	33	accuracy	accuracy	NOUN
ajst-20799	103	34	and	and	CCONJ
ajst-20799	103	35	robustness	robustness	NOUN
ajst-20799	103	36	.	.	PUNCT
ajst-20799	104	1	when	when	SCONJ
ajst-20799	104	2	conducting	conduct	VERB
ajst-20799	104	3	a	a	DET
ajst-20799	104	4	comparative	comparative	ADJ
ajst-20799	104	5	analysis	analysis	NOUN
ajst-20799	104	6	of	of	ADP
ajst-20799	104	7	different	different	ADJ
ajst-20799	104	8	deep	deep	ADJ
ajst-20799	104	9	learning	learning	NOUN
ajst-20799	104	10	methods	method	NOUN
ajst-20799	104	11	for	for	ADP
ajst-20799	104	12	vehicle	vehicle	NOUN
ajst-20799	104	13	detection	detection	NOUN
ajst-20799	104	14	and	and	CCONJ
ajst-20799	104	15	tracking	tracking	NOUN
ajst-20799	104	16	tasks	task	NOUN
ajst-20799	104	17	,	,	PUNCT
ajst-20799	104	18	it	it	PRON
ajst-20799	104	19	's	be	AUX
ajst-20799	104	20	essential	essential	ADJ
ajst-20799	104	21	to	to	PART
ajst-20799	104	22	consider	consider	VERB
ajst-20799	104	23	the	the	DET
ajst-20799	104	24	accuracy	accuracy	NOUN
ajst-20799	104	25	of	of	ADP
ajst-20799	104	26	detection	detection	NOUN
ajst-20799	104	27	and	and	CCONJ
ajst-20799	104	28	tracking	tracking	NOUN
ajst-20799	104	29	.	.	PUNCT
ajst-20799	105	1	this	this	PRON
ajst-20799	105	2	involves	involve	VERB
ajst-20799	105	3	evaluating	evaluate	VERB
ajst-20799	105	4	whether	whether	SCONJ
ajst-20799	105	5	the	the	DET
ajst-20799	105	6	algorithms	algorithm	NOUN
ajst-20799	105	7	can	can	AUX
ajst-20799	105	8	accurately	accurately	ADV
ajst-20799	105	9	detect	detect	VERB
ajst-20799	105	10	and	and	CCONJ
ajst-20799	105	11	track	track	NOUN
ajst-20799	105	12	vehicle	vehicle	NOUN
ajst-20799	105	13	targets	target	NOUN
ajst-20799	105	14	,	,	PUNCT
ajst-20799	105	15	with	with	ADP
ajst-20799	105	16	high	high	ADJ
ajst-20799	105	17	accuracy	accuracy	NOUN
ajst-20799	105	18	indicating	indicate	VERB
ajst-20799	105	19	precise	precise	ADJ
ajst-20799	105	20	localization	localization	NOUN
ajst-20799	105	21	of	of	ADP
ajst-20799	105	22	vehicle	vehicle	NOUN
ajst-20799	105	23	positions	position	NOUN
ajst-20799	105	24	,	,	PUNCT
ajst-20799	105	25	reduced	reduce	VERB
ajst-20799	105	26	false	false	ADJ
ajst-20799	105	27	positives	positive	NOUN
ajst-20799	105	28	(	(	PUNCT
ajst-20799	105	29	misidentifying	misidentify	VERB
ajst-20799	105	30	non	non	ADJ
ajst-20799	105	31	-	-	NOUN
ajst-20799	105	32	vehicles	vehicle	NOUN
ajst-20799	105	33	as	as	ADP
ajst-20799	105	34	vehicles	vehicle	NOUN
ajst-20799	105	35	)	)	PUNCT
ajst-20799	105	36	,	,	PUNCT
ajst-20799	105	37	and	and	CCONJ
ajst-20799	105	38	false	false	ADJ
ajst-20799	105	39	negatives	negative	NOUN
ajst-20799	105	40	(	(	PUNCT
ajst-20799	105	41	failing	fail	VERB
ajst-20799	105	42	to	to	PART
ajst-20799	105	43	detect	detect	VERB
ajst-20799	105	44	actual	actual	ADJ
ajst-20799	105	45	vehicles	vehicle	NOUN
ajst-20799	105	46	)	)	PUNCT
ajst-20799	105	47	,	,	PUNCT
ajst-20799	105	48	thus	thus	ADV
ajst-20799	105	49	avoiding	avoid	VERB
ajst-20799	105	50	misidentifications	misidentification	NOUN
ajst-20799	105	51	and	and	CCONJ
ajst-20799	105	52	missed	miss	VERB
ajst-20799	105	53	detections	detection	NOUN
ajst-20799	105	54	.	.	PUNCT
ajst-20799	106	1	secondly	secondly	ADV
ajst-20799	106	2	,	,	PUNCT
ajst-20799	106	3	real	real	ADJ
ajst-20799	106	4	-	-	PUNCT
ajst-20799	106	5	time	time	NOUN
ajst-20799	106	6	performance	performance	NOUN
ajst-20799	106	7	is	be	AUX
ajst-20799	106	8	crucial	crucial	ADJ
ajst-20799	106	9	for	for	ADP
ajst-20799	106	10	applications	application	NOUN
ajst-20799	106	11	like	like	ADP
ajst-20799	106	12	autonomous	autonomous	ADJ
ajst-20799	106	13	driving	driving	NOUN
ajst-20799	106	14	and	and	CCONJ
ajst-20799	106	15	video	video	NOUN
ajst-20799	106	16	surveillance	surveillance	NOUN
ajst-20799	106	17	,	,	PUNCT
ajst-20799	106	18	requiring	require	VERB
ajst-20799	106	19	algorithms	algorithm	NOUN
ajst-20799	106	20	to	to	PART
ajst-20799	106	21	complete	complete	VERB
ajst-20799	106	22	detection	detection	NOUN
ajst-20799	106	23	within	within	ADP
ajst-20799	106	24	milliseconds	millisecond	NOUN
ajst-20799	106	25	to	to	PART
ajst-20799	106	26	ensure	ensure	VERB
ajst-20799	106	27	real	real	ADJ
ajst-20799	106	28	-	-	PUNCT
ajst-20799	106	29	time	time	NOUN
ajst-20799	106	30	synchronization	synchronization	NOUN
ajst-20799	106	31	with	with	ADP
ajst-20799	106	32	video	video	NOUN
ajst-20799	106	33	streams	stream	NOUN
ajst-20799	106	34	,	,	PUNCT
ajst-20799	106	35	i.e.	i.e.	X
ajst-20799	106	36	,	,	PUNCT
ajst-20799	106	37	whether	whether	SCONJ
ajst-20799	106	38	the	the	DET
ajst-20799	106	39	algorithms	algorithm	NOUN
ajst-20799	106	40	can	can	AUX
ajst-20799	106	41	rapidly	rapidly	ADV
ajst-20799	106	42	and	and	CCONJ
ajst-20799	106	43	efficiently	efficiently	ADV
ajst-20799	106	44	perform	perform	VERB
ajst-20799	106	45	object	object	NOUN
ajst-20799	106	46	detection	detection	NOUN
ajst-20799	106	47	and	and	CCONJ
ajst-20799	106	48	tracking	tracking	NOUN
ajst-20799	106	49	in	in	ADP
ajst-20799	106	50	real	real	ADJ
ajst-20799	106	51	-	-	PUNCT
ajst-20799	106	52	time	time	NOUN
ajst-20799	106	53	video	video	NOUN
ajst-20799	106	54	streams	stream	NOUN
ajst-20799	106	55	.	.	PUNCT
ajst-20799	107	1	robustness	robustness	NOUN
ajst-20799	107	2	of	of	ADP
ajst-20799	107	3	the	the	DET
ajst-20799	107	4	algorithm	algorithm	NOUN
ajst-20799	107	5	should	should	AUX
ajst-20799	107	6	also	also	ADV
ajst-20799	107	7	be	be	AUX
ajst-20799	107	8	considered	consider	VERB
ajst-20799	107	9	,	,	PUNCT
ajst-20799	107	10	assessing	assess	VERB
ajst-20799	107	11	its	its	PRON
ajst-20799	107	12	performance	performance	NOUN
ajst-20799	107	13	across	across	ADP
ajst-20799	107	14	various	various	ADJ
ajst-20799	107	15	environmental	environmental	ADJ
ajst-20799	107	16	conditions	condition	NOUN
ajst-20799	107	17	such	such	ADJ
ajst-20799	107	18	as	as	ADP
ajst-20799	107	19	urban	urban	ADJ
ajst-20799	107	20	roads	road	NOUN
ajst-20799	107	21	,	,	PUNCT
ajst-20799	107	22	highways	highway	NOUN
ajst-20799	107	23	,	,	PUNCT
ajst-20799	107	24	rural	rural	ADJ
ajst-20799	107	25	roads	road	NOUN
ajst-20799	107	26	,	,	PUNCT
ajst-20799	107	27	tunnels	tunnel	NOUN
ajst-20799	107	28	,	,	PUNCT
ajst-20799	107	29	nighttime	nighttime	NOUN
ajst-20799	107	30	,	,	PUNCT
ajst-20799	107	31	rainy	rainy	ADJ
ajst-20799	107	32	or	or	CCONJ
ajst-20799	107	33	snowy	snowy	ADJ
ajst-20799	107	34	weather	weather	NOUN
ajst-20799	107	35	,	,	PUNCT
ajst-20799	107	36	etc	etc	X
ajst-20799	107	37	.	.	X
ajst-20799	107	38	,	,	PUNCT
ajst-20799	107	39	i.e.	i.e.	X
ajst-20799	107	40	,	,	PUNCT
ajst-20799	107	41	the	the	DET
ajst-20799	107	42	algorithm	algorithm	NOUN
ajst-20799	107	43	's	's	PART
ajst-20799	107	44	adaptability	adaptability	NOUN
ajst-20799	107	45	to	to	ADP
ajst-20799	107	46	different	different	ADJ
ajst-20799	107	47	scenes	scene	NOUN
ajst-20799	107	48	,	,	PUNCT
ajst-20799	107	49	lighting	lighting	NOUN
ajst-20799	107	50	conditions	condition	NOUN
ajst-20799	107	51	,	,	PUNCT
ajst-20799	107	52	and	and	CCONJ
ajst-20799	107	53	target	target	NOUN
ajst-20799	107	54	scales	scale	NOUN
ajst-20799	107	55	.	.	PUNCT
ajst-20799	108	1	5	5	X
ajst-20799	108	2	.	.	X
ajst-20799	108	3	application	application	NOUN
ajst-20799	108	4	in	in	ADP
ajst-20799	108	5	road	road	NOUN
ajst-20799	108	6	traffic	traffic	NOUN
ajst-20799	108	7	monitoring	monitor	VERB
ajst-20799	108	8	5.1	5.1	NUM
ajst-20799	108	9	.	.	PUNCT
ajst-20799	108	10	balancing	balance	VERB
ajst-20799	108	11	real	real	ADJ
ajst-20799	108	12	-	-	PUNCT
ajst-20799	108	13	time	time	NOUN
ajst-20799	108	14	and	and	CCONJ
ajst-20799	108	15	accuracy	accuracy	NOUN
ajst-20799	108	16	a	a	DET
ajst-20799	108	17	real	real	ADJ
ajst-20799	108	18	-	-	PUNCT
ajst-20799	108	19	time	time	NOUN
ajst-20799	108	20	monitoring	monitoring	NOUN
ajst-20799	108	21	system	system	NOUN
ajst-20799	108	22	refers	refer	VERB
ajst-20799	108	23	to	to	ADP
ajst-20799	108	24	the	the	DET
ajst-20799	108	25	use	use	NOUN
ajst-20799	108	26	of	of	ADP
ajst-20799	108	27	cameras	camera	NOUN
ajst-20799	108	28	,	,	PUNCT
ajst-20799	108	29	radars	radar	NOUN
ajst-20799	108	30	,	,	PUNCT
ajst-20799	108	31	sensors	sensor	NOUN
ajst-20799	108	32	,	,	PUNCT
ajst-20799	108	33	and	and	CCONJ
ajst-20799	108	34	other	other	ADJ
ajst-20799	108	35	devices	device	NOUN
ajst-20799	108	36	deployed	deploy	VERB
ajst-20799	108	37	at	at	ADP
ajst-20799	108	38	key	key	ADJ
ajst-20799	108	39	locations	location	NOUN
ajst-20799	108	40	such	such	ADJ
ajst-20799	108	41	as	as	ADP
ajst-20799	108	42	roads	road	NOUN
ajst-20799	108	43	,	,	PUNCT
ajst-20799	108	44	intersections	intersection	NOUN
ajst-20799	108	45	,	,	PUNCT
ajst-20799	108	46	tunnels	tunnel	NOUN
ajst-20799	108	47	,	,	PUNCT
ajst-20799	108	48	and	and	CCONJ
ajst-20799	108	49	bridges	bridge	NOUN
ajst-20799	108	50	.	.	PUNCT
ajst-20799	109	1	one	one	NUM
ajst-20799	109	2	of	of	ADP
ajst-20799	109	3	the	the	DET
ajst-20799	109	4	most	most	ADV
ajst-20799	109	5	fundamental	fundamental	ADJ
ajst-20799	109	6	requirements	requirement	NOUN
ajst-20799	109	7	for	for	ADP
ajst-20799	109	8	a	a	DET
ajst-20799	109	9	real	real	ADJ
ajst-20799	109	10	-	-	PUNCT
ajst-20799	109	11	time	time	NOUN
ajst-20799	109	12	monitoring	monitoring	NOUN
ajst-20799	109	13	system	system	NOUN
ajst-20799	109	14	is	be	AUX
ajst-20799	109	15	the	the	DET
ajst-20799	109	16	real	real	ADJ
ajst-20799	109	17	-	-	PUNCT
ajst-20799	109	18	time	time	NOUN
ajst-20799	109	19	detection	detection	NOUN
ajst-20799	109	20	and	and	CCONJ
ajst-20799	109	21	tracking	tracking	NOUN
ajst-20799	109	22	of	of	ADP
ajst-20799	109	23	vehicles.[5	vehicles.[5	NOUN
ajst-20799	109	24	]	]	PUNCT
ajst-20799	109	25	this	this	PRON
ajst-20799	109	26	means	mean	VERB
ajst-20799	109	27	that	that	SCONJ
ajst-20799	109	28	the	the	DET
ajst-20799	109	29	system	system	NOUN
ajst-20799	109	30	needs	need	VERB
ajst-20799	109	31	to	to	PART
ajst-20799	109	32	accurately	accurately	ADV
ajst-20799	109	33	identify	identify	VERB
ajst-20799	109	34	and	and	CCONJ
ajst-20799	109	35	track	track	VERB
ajst-20799	109	36	vehicles	vehicle	NOUN
ajst-20799	109	37	that	that	PRON
ajst-20799	109	38	appear	appear	VERB
ajst-20799	109	39	on	on	ADP
ajst-20799	109	40	the	the	DET
ajst-20799	109	41	road	road	NOUN
ajst-20799	109	42	within	within	ADP
ajst-20799	109	43	a	a	DET
ajst-20799	109	44	short	short	ADJ
ajst-20799	109	45	period	period	NOUN
ajst-20799	109	46	.	.	PUNCT
ajst-20799	110	1	it	it	PRON
ajst-20799	110	2	requires	require	VERB
ajst-20799	110	3	almost	almost	ADV
ajst-20799	110	4	real	real	ADJ
ajst-20799	110	5	-	-	PUNCT
ajst-20799	110	6	time	time	NOUN
ajst-20799	110	7	capture	capture	NOUN
ajst-20799	110	8	and	and	CCONJ
ajst-20799	110	9	analysis	analysis	NOUN
ajst-20799	110	10	of	of	ADP
ajst-20799	110	11	every	every	DET
ajst-20799	110	12	frame	frame	NOUN
ajst-20799	110	13	in	in	ADP
ajst-20799	110	14	the	the	DET
ajst-20799	110	15	video	video	NOUN
ajst-20799	110	16	stream	stream	NOUN
ajst-20799	110	17	to	to	PART
ajst-20799	110	18	promptly	promptly	ADV
ajst-20799	110	19	identify	identify	VERB
ajst-20799	110	20	and	and	CCONJ
ajst-20799	110	21	track	track	VERB
ajst-20799	110	22	newly	newly	ADV
ajst-20799	110	23	appeared	appear	VERB
ajst-20799	110	24	vehicles	vehicle	NOUN
ajst-20799	110	25	,	,	PUNCT
ajst-20799	110	26	update	update	VERB
ajst-20799	110	27	the	the	DET
ajst-20799	110	28	status	status	NOUN
ajst-20799	110	29	of	of	ADP
ajst-20799	110	30	existing	exist	VERB
ajst-20799	110	31	vehicles	vehicle	NOUN
ajst-20799	110	32	,	,	PUNCT
ajst-20799	110	33	and	and	CCONJ
ajst-20799	110	34	respond	respond	VERB
ajst-20799	110	35	quickly	quickly	ADV
ajst-20799	110	36	to	to	ADP
ajst-20799	110	37	abnormal	abnormal	ADJ
ajst-20799	110	38	situations	situation	NOUN
ajst-20799	110	39	.	.	PUNCT
ajst-20799	111	1	the	the	DET
ajst-20799	111	2	vehicle	vehicle	NOUN
ajst-20799	111	3	detection	detection	NOUN
ajst-20799	111	4	and	and	CCONJ
ajst-20799	111	5	tracking	tracking	NOUN
ajst-20799	111	6	in	in	ADP
ajst-20799	111	7	real	real	ADJ
ajst-20799	111	8	-	-	PUNCT
ajst-20799	111	9	time	time	NOUN
ajst-20799	111	10	monitoring	monitoring	NOUN
ajst-20799	111	11	systems	system	NOUN
ajst-20799	111	12	require	require	VERB
ajst-20799	111	13	high	high	ADJ
ajst-20799	111	14	real	real	ADJ
ajst-20799	111	15	-	-	PUNCT
ajst-20799	111	16	time	time	NOUN
ajst-20799	111	17	capabilities	capability	NOUN
ajst-20799	111	18	.	.	PUNCT
ajst-20799	112	1	therefore	therefore	ADV
ajst-20799	112	2	,	,	PUNCT
ajst-20799	112	3	detection	detection	NOUN
ajst-20799	112	4	models	model	NOUN
ajst-20799	112	5	that	that	PRON
ajst-20799	112	6	can	can	AUX
ajst-20799	112	7	accurately	accurately	ADV
ajst-20799	112	8	identify	identify	VERB
ajst-20799	112	9	vehicle	vehicle	NOUN
ajst-20799	112	10	boundaries	boundary	NOUN
ajst-20799	112	11	,	,	PUNCT
ajst-20799	112	12	types	type	NOUN
ajst-20799	112	13	,	,	PUNCT
ajst-20799	112	14	and	and	CCONJ
ajst-20799	112	15	even	even	ADV
ajst-20799	112	16	vehicle	vehicle	NOUN
ajst-20799	112	17	attributes	attribute	NOUN
ajst-20799	112	18	such	such	ADJ
ajst-20799	112	19	as	as	ADP
ajst-20799	112	20	license	license	NOUN
ajst-20799	112	21	plate	plate	NOUN
ajst-20799	112	22	numbers	number	NOUN
ajst-20799	112	23	and	and	CCONJ
ajst-20799	112	24	colors	color	NOUN
ajst-20799	112	25	should	should	AUX
ajst-20799	112	26	be	be	AUX
ajst-20799	112	27	selected	select	VERB
ajst-20799	112	28	or	or	CCONJ
ajst-20799	112	29	developed	develop	VERB
ajst-20799	112	30	.	.	PUNCT
ajst-20799	113	1	these	these	DET
ajst-20799	113	2	models	model	NOUN
ajst-20799	113	3	should	should	AUX
ajst-20799	113	4	be	be	AUX
ajst-20799	113	5	able	able	ADJ
ajst-20799	113	6	to	to	PART
ajst-20799	113	7	complete	complete	VERB
ajst-20799	113	8	target	target	NOUN
ajst-20799	113	9	detection	detection	NOUN
ajst-20799	113	10	and	and	CCONJ
ajst-20799	113	11	tracking	tracking	NOUN
ajst-20799	113	12	tasks	task	NOUN
ajst-20799	113	13	quickly	quickly	ADV
ajst-20799	113	14	and	and	CCONJ
ajst-20799	113	15	timely	timely	ADJ
ajst-20799	113	16	,	,	PUNCT
ajst-20799	113	17	reducing	reduce	VERB
ajst-20799	113	18	false	false	ADJ
ajst-20799	113	19	positives	positive	NOUN
ajst-20799	113	20	(	(	PUNCT
ajst-20799	113	21	misidentifying	misidentify	VERB
ajst-20799	113	22	non	non	ADJ
ajst-20799	113	23	-	-	NOUN
ajst-20799	113	24	vehicles	vehicle	NOUN
ajst-20799	113	25	as	as	ADP
ajst-20799	113	26	vehicles	vehicle	NOUN
ajst-20799	113	27	)	)	PUNCT
ajst-20799	113	28	and	and	CCONJ
ajst-20799	113	29	missed	miss	VERB
ajst-20799	113	30	detections	detection	NOUN
ajst-20799	113	31	(	(	PUNCT
ajst-20799	113	32	failing	fail	VERB
ajst-20799	113	33	to	to	PART
ajst-20799	113	34	detect	detect	VERB
ajst-20799	113	35	actual	actual	ADJ
ajst-20799	113	36	vehicles	vehicle	NOUN
ajst-20799	113	37	)	)	PUNCT
ajst-20799	113	38	.	.	PUNCT
ajst-20799	114	1	real	real	ADJ
ajst-20799	114	2	-	-	PUNCT
ajst-20799	114	3	time	time	NOUN
ajst-20799	114	4	monitoring	monitoring	NOUN
ajst-20799	114	5	systems	system	NOUN
ajst-20799	114	6	also	also	ADV
ajst-20799	114	7	need	need	VERB
ajst-20799	114	8	to	to	PART
ajst-20799	114	9	consider	consider	VERB
ajst-20799	114	10	accuracy	accuracy	NOUN
ajst-20799	114	11	issues	issue	NOUN
ajst-20799	114	12	.	.	PUNCT
ajst-20799	115	1	in	in	ADP
ajst-20799	115	2	continuous	continuous	ADJ
ajst-20799	115	3	video	video	NOUN
ajst-20799	115	4	streams	stream	NOUN
ajst-20799	115	5	,	,	PUNCT
ajst-20799	115	6	the	the	DET
ajst-20799	115	7	system	system	NOUN
ajst-20799	115	8	must	must	AUX
ajst-20799	115	9	ensure	ensure	VERB
ajst-20799	115	10	accurate	accurate	ADJ
ajst-20799	115	11	detection	detection	NOUN
ajst-20799	115	12	and	and	CCONJ
ajst-20799	115	13	tracking	tracking	NOUN
ajst-20799	115	14	results	result	NOUN
ajst-20799	115	15	for	for	ADP
ajst-20799	115	16	vehicles	vehicle	NOUN
ajst-20799	115	17	.	.	PUNCT
ajst-20799	116	1	even	even	ADV
ajst-20799	116	2	in	in	ADP
ajst-20799	116	3	complex	complex	ADJ
ajst-20799	116	4	situations	situation	NOUN
ajst-20799	116	5	such	such	ADJ
ajst-20799	116	6	as	as	ADP
ajst-20799	116	7	vehicle	vehicle	NOUN
ajst-20799	116	8	occlusion	occlusion	NOUN
ajst-20799	116	9	,	,	PUNCT
ajst-20799	116	10	deformation	deformation	NOUN
ajst-20799	116	11	,	,	PUNCT
ajst-20799	116	12	rapid	rapid	ADJ
ajst-20799	116	13	movement	movement	NOUN
ajst-20799	116	14	,	,	PUNCT
ajst-20799	116	15	or	or	CCONJ
ajst-20799	116	16	changes	change	NOUN
ajst-20799	116	17	in	in	ADP
ajst-20799	116	18	lighting	lighting	NOUN
ajst-20799	116	19	conditions	condition	NOUN
ajst-20799	116	20	,	,	PUNCT
ajst-20799	116	21	there	there	PRON
ajst-20799	116	22	should	should	AUX
ajst-20799	116	23	be	be	AUX
ajst-20799	116	24	no	no	DET
ajst-20799	116	25	i	i	NOUN
ajst-20799	116	26	d	d	NOUN
ajst-20799	116	27	switching	switch	VERB
ajst-20799	116	28	or	or	CCONJ
ajst-20799	116	29	tracking	tracking	NOUN
ajst-20799	116	30	loss	loss	NOUN
ajst-20799	116	31	,	,	PUNCT
ajst-20799	116	32	avoiding	avoid	VERB
ajst-20799	116	33	false	false	ADJ
ajst-20799	116	34	positives	positive	NOUN
ajst-20799	116	35	or	or	CCONJ
ajst-20799	116	36	missed	miss	VERB
ajst-20799	116	37	detections	detection	NOUN
ajst-20799	116	38	.	.	PUNCT
ajst-20799	117	1	5.2	5.2	NUM
ajst-20799	117	2	.	.	PUNCT
ajst-20799	117	3	traffic	traffic	NOUN
ajst-20799	117	4	accident	accident	NOUN
ajst-20799	117	5	warning	warning	NOUN
ajst-20799	117	6	and	and	CCONJ
ajst-20799	117	7	management	management	NOUN
ajst-20799	117	8	traffic	traffic	NOUN
ajst-20799	117	9	accident	accident	NOUN
ajst-20799	117	10	warning	warning	NOUN
ajst-20799	117	11	and	and	CCONJ
ajst-20799	117	12	management	management	NOUN
ajst-20799	117	13	systems	system	NOUN
ajst-20799	117	14	are	be	AUX
ajst-20799	117	15	crucial	crucial	ADJ
ajst-20799	117	16	components	component	NOUN
ajst-20799	117	17	of	of	ADP
ajst-20799	117	18	modern	modern	ADJ
ajst-20799	117	19	intelligent	intelligent	ADJ
ajst-20799	117	20	transportation	transportation	NOUN
ajst-20799	117	21	systems	system	NOUN
ajst-20799	117	22	.	.	PUNCT
ajst-20799	118	1	deep	deep	ADJ
ajst-20799	118	2	learning	learning	NOUN
ajst-20799	118	3	-	-	PUNCT
ajst-20799	118	4	based	base	VERB
ajst-20799	118	5	traffic	traffic	NOUN
ajst-20799	118	6	accident	accident	NOUN
ajst-20799	118	7	prediction	prediction	NOUN
ajst-20799	118	8	methods	method	NOUN
ajst-20799	118	9	involve	involve	VERB
ajst-20799	118	10	training	training	NOUN
ajst-20799	118	11	and	and	CCONJ
ajst-20799	118	12	learning	learn	VERB
ajst-20799	118	13	deep	deep	ADJ
ajst-20799	118	14	learning	learning	NOUN
ajst-20799	118	15	models	model	NOUN
ajst-20799	118	16	using	use	VERB
ajst-20799	118	17	road	road	NOUN
ajst-20799	118	18	traffic	traffic	NOUN
ajst-20799	118	19	data	datum	NOUN
ajst-20799	118	20	.	.	PUNCT
ajst-20799	119	1	detailed	detailed	ADJ
ajst-20799	119	2	information	information	NOUN
ajst-20799	119	3	such	such	ADJ
ajst-20799	119	4	as	as	ADP
ajst-20799	119	5	the	the	DET
ajst-20799	119	6	location	location	NOUN
ajst-20799	119	7	,	,	PUNCT
ajst-20799	119	8	time	time	NOUN
ajst-20799	119	9	,	,	PUNCT
ajst-20799	119	10	type	type	NOUN
ajst-20799	119	11	,	,	PUNCT
ajst-20799	119	12	and	and	CCONJ
ajst-20799	119	13	severity	severity	NOUN
ajst-20799	119	14	of	of	ADP
ajst-20799	119	15	traffic	traffic	NOUN
ajst-20799	119	16	accidents	accident	NOUN
ajst-20799	119	17	that	that	PRON
ajst-20799	119	18	occurred	occur	VERB
ajst-20799	119	19	in	in	ADP
ajst-20799	119	20	the	the	DET
ajst-20799	119	21	recent	recent	ADJ
ajst-20799	119	22	past	past	NOUN
ajst-20799	119	23	is	be	AUX
ajst-20799	119	24	provided	provide	VERB
ajst-20799	119	25	as	as	ADP
ajst-20799	119	26	learning	learn	VERB
ajst-20799	119	27	samples	sample	NOUN
ajst-20799	119	28	to	to	ADP
ajst-20799	119	29	the	the	DET
ajst-20799	119	30	model	model	NOUN
ajst-20799	119	31	.	.	PUNCT
ajst-20799	120	1	dynamic	dynamic	ADJ
ajst-20799	120	2	information	information	NOUN
ajst-20799	120	3	such	such	ADJ
ajst-20799	120	4	as	as	ADP
ajst-20799	120	5	vehicle	vehicle	NOUN
ajst-20799	120	6	position	position	NOUN
ajst-20799	120	7	,	,	PUNCT
ajst-20799	120	8	speed	speed	NOUN
ajst-20799	120	9	,	,	PUNCT
ajst-20799	120	10	acceleration	acceleration	NOUN
ajst-20799	120	11	,	,	PUNCT
ajst-20799	120	12	and	and	CCONJ
ajst-20799	120	13	direction	direction	NOUN
ajst-20799	120	14	is	be	AUX
ajst-20799	120	15	collected	collect	VERB
ajst-20799	120	16	through	through	ADP
ajst-20799	120	17	gps	gps	PROPN
ajst-20799	120	18	,	,	PUNCT
ajst-20799	120	19	onboard	onboard	ADP
ajst-20799	120	20	sensors	sensor	NOUN
ajst-20799	120	21	,	,	PUNCT
ajst-20799	120	22	and	and	CCONJ
ajst-20799	120	23	other	other	ADJ
ajst-20799	120	24	devices	device	NOUN
ajst-20799	120	25	to	to	PART
ajst-20799	120	26	analyze	analyze	VERB
ajst-20799	120	27	vehicle	vehicle	NOUN
ajst-20799	120	28	behavior	behavior	NOUN
ajst-20799	120	29	patterns	pattern	NOUN
ajst-20799	120	30	.	.	PUNCT
ajst-20799	121	1	this	this	DET
ajst-20799	121	2	analysis	analysis	NOUN
ajst-20799	121	3	can	can	AUX
ajst-20799	121	4	effectively	effectively	ADV
ajst-20799	121	5	identify	identify	VERB
ajst-20799	121	6	the	the	DET
ajst-20799	121	7	possibility	possibility	NOUN
ajst-20799	121	8	of	of	ADP
ajst-20799	121	9	traffic	traffic	NOUN
ajst-20799	121	10	accidents	accident	NOUN
ajst-20799	121	11	occurring	occur	VERB
ajst-20799	121	12	and	and	CCONJ
ajst-20799	121	13	simultaneously	simultaneously	ADV
ajst-20799	121	14	annotate	annotate	VERB
ajst-20799	121	15	accident	accident	NOUN
ajst-20799	121	16	information	information	NOUN
ajst-20799	121	17	in	in	ADP
ajst-20799	121	18	real	real	ADJ
ajst-20799	121	19	-	-	PUNCT
ajst-20799	121	20	time	time	NOUN
ajst-20799	121	21	on	on	ADP
ajst-20799	121	22	the	the	DET
ajst-20799	121	23	geographical	geographical	ADJ
ajst-20799	121	24	information	information	NOUN
ajst-20799	121	25	system	system	NOUN
ajst-20799	121	26	(	(	PUNCT
ajst-20799	121	27	gis	gis	NOUN
ajst-20799	121	28	)	)	PUNCT
ajst-20799	121	29	of	of	ADP
ajst-20799	121	30	the	the	DET
ajst-20799	121	31	traffic	traffic	NOUN
ajst-20799	121	32	management	management	NOUN
ajst-20799	121	33	department	department	NOUN
ajst-20799	121	34	,	,	PUNCT
ajst-20799	121	35	visually	visually	ADV
ajst-20799	121	36	displaying	display	VERB
ajst-20799	121	37	accident	accident	NOUN
ajst-20799	121	38	locations	location	NOUN
ajst-20799	121	39	and	and	CCONJ
ajst-20799	121	40	surrounding	surround	VERB
ajst-20799	121	41	traffic	traffic	NOUN
ajst-20799	121	42	conditions	condition	NOUN
ajst-20799	121	43	.	.	PUNCT
ajst-20799	122	1	these	these	DET
ajst-20799	122	2	models	model	NOUN
ajst-20799	122	3	can	can	AUX
ajst-20799	122	4	utilize	utilize	VERB
ajst-20799	122	5	various	various	ADJ
ajst-20799	122	6	data	datum	NOUN
ajst-20799	122	7	sources	source	NOUN
ajst-20799	122	8	such	such	ADJ
ajst-20799	122	9	as	as	ADP
ajst-20799	122	10	historical	historical	ADJ
ajst-20799	122	11	traffic	traffic	NOUN
ajst-20799	122	12	data	datum	NOUN
ajst-20799	122	13	,	,	PUNCT
ajst-20799	122	14	road	road	NOUN
ajst-20799	122	15	conditions	condition	NOUN
ajst-20799	122	16	,	,	PUNCT
ajst-20799	122	17	and	and	CCONJ
ajst-20799	122	18	vehicle	vehicle	NOUN
ajst-20799	122	19	motion	motion	NOUN
ajst-20799	122	20	trajectories	trajectory	NOUN
ajst-20799	122	21	,	,	PUNCT
ajst-20799	122	22	undergo	undergo	VERB
ajst-20799	122	23	preprocessing	preprocessing	NOUN
ajst-20799	122	24	operations	operation	NOUN
ajst-20799	122	25	like	like	ADP
ajst-20799	122	26	cleaning	clean	VERB
ajst-20799	122	27	,	,	PUNCT
ajst-20799	122	28	formatting	formatting	NOUN
ajst-20799	122	29	,	,	PUNCT
ajst-20799	122	30	and	and	CCONJ
ajst-20799	122	31	spatiotemporal	spatiotemporal	ADJ
ajst-20799	122	32	alignment	alignment	NOUN
ajst-20799	122	33	to	to	PART
ajst-20799	122	34	form	form	VERB
ajst-20799	122	35	structured	structure	VERB
ajst-20799	122	36	,	,	PUNCT
ajst-20799	122	37	standardized	standardize	VERB
ajst-20799	122	38	datasets	dataset	NOUN
ajst-20799	122	39	for	for	ADP
ajst-20799	122	40	training	training	NOUN
ajst-20799	122	41	and	and	CCONJ
ajst-20799	122	42	prediction	prediction	NOUN
ajst-20799	122	43	by	by	ADP
ajst-20799	122	44	deep	deep	ADJ
ajst-20799	122	45	learning	learning	NOUN
ajst-20799	122	46	models	model	NOUN
ajst-20799	122	47	.	.	PUNCT
ajst-20799	123	1	by	by	ADP
ajst-20799	123	2	learning	learn	VERB
ajst-20799	123	3	the	the	DET
ajst-20799	123	4	patterns	pattern	NOUN
ajst-20799	123	5	and	and	CCONJ
ajst-20799	123	6	trends	trend	NOUN
ajst-20799	123	7	of	of	ADP
ajst-20799	123	8	traffic	traffic	NOUN
ajst-20799	123	9	accidents	accident	NOUN
ajst-20799	123	10	,	,	PUNCT
ajst-20799	123	11	these	these	DET
ajst-20799	123	12	models	model	NOUN
ajst-20799	123	13	can	can	AUX
ajst-20799	123	14	predict	predict	VERB
ajst-20799	123	15	potential	potential	ADJ
ajst-20799	123	16	traffic	traffic	NOUN
ajst-20799	123	17	accidents	accident	NOUN
ajst-20799	123	18	.	.	PUNCT
ajst-20799	124	1	predictions	prediction	NOUN
ajst-20799	124	2	can	can	AUX
ajst-20799	124	3	be	be	AUX
ajst-20799	124	4	global	global	ADJ
ajst-20799	124	5	(	(	PUNCT
ajst-20799	124	6	for	for	ADP
ajst-20799	124	7	an	an	DET
ajst-20799	124	8	entire	entire	ADJ
ajst-20799	124	9	city	city	NOUN
ajst-20799	124	10	or	or	CCONJ
ajst-20799	124	11	specific	specific	ADJ
ajst-20799	124	12	area	area	NOUN
ajst-20799	124	13	)	)	PUNCT
ajst-20799	124	14	or	or	CCONJ
ajst-20799	124	15	local	local	ADJ
ajst-20799	124	16	(	(	PUNCT
ajst-20799	124	17	for	for	ADP
ajst-20799	124	18	specific	specific	ADJ
ajst-20799	124	19	road	road	NOUN
ajst-20799	124	20	sections	section	NOUN
ajst-20799	124	21	or	or	CCONJ
ajst-20799	124	22	intersections	intersection	NOUN
ajst-20799	124	23	)	)	PUNCT
ajst-20799	124	24	.	.	PUNCT
ajst-20799	125	1	setting	set	VERB
ajst-20799	125	2	reasonable	reasonable	ADJ
ajst-20799	125	3	warning	warning	NOUN
ajst-20799	125	4	thresholds	threshold	NOUN
ajst-20799	125	5	,	,	PUNCT
ajst-20799	125	6	when	when	SCONJ
ajst-20799	125	7	the	the	DET
ajst-20799	125	8	model	model	NOUN
ajst-20799	125	9	predicts	predict	VERB
ajst-20799	125	10	a	a	DET
ajst-20799	125	11	risk	risk	NOUN
ajst-20799	125	12	exceeding	exceed	VERB
ajst-20799	125	13	the	the	DET
ajst-20799	125	14	threshold	threshold	NOUN
ajst-20799	125	15	,	,	PUNCT
ajst-20799	125	16	triggers	trigger	VERB
ajst-20799	125	17	a	a	DET
ajst-20799	125	18	traffic	traffic	NOUN
ajst-20799	125	19	accident	accident	NOUN
ajst-20799	125	20	warning	warning	NOUN
ajst-20799	125	21	.	.	PUNCT
ajst-20799	126	1	the	the	DET
ajst-20799	126	2	warning	warning	NOUN
ajst-20799	126	3	information	information	NOUN
ajst-20799	126	4	may	may	AUX
ajst-20799	126	5	include	include	VERB
ajst-20799	126	6	prediction	prediction	NOUN
ajst-20799	126	7	time	time	NOUN
ajst-20799	126	8	,	,	PUNCT
ajst-20799	126	9	location	location	NOUN
ajst-20799	126	10	,	,	PUNCT
ajst-20799	126	11	risk	risk	NOUN
ajst-20799	126	12	level	level	NOUN
ajst-20799	126	13	,	,	PUNCT
ajst-20799	126	14	potential	potential	ADJ
ajst-20799	126	15	impact	impact	NOUN
ajst-20799	126	16	range	range	NOUN
ajst-20799	126	17	,	,	PUNCT
ajst-20799	126	18	etc	etc	X
ajst-20799	126	19	.	.	X
ajst-20799	126	20	upon	upon	SCONJ
ajst-20799	126	21	receiving	receive	VERB
ajst-20799	126	22	the	the	DET
ajst-20799	126	23	warning	warning	NOUN
ajst-20799	126	24	,	,	PUNCT
ajst-20799	126	25	rescue	rescue	NOUN
ajst-20799	126	26	vehicles	vehicle	NOUN
ajst-20799	126	27	such	such	ADJ
ajst-20799	126	28	as	as	ADP
ajst-20799	126	29	ambulances	ambulance	NOUN
ajst-20799	126	30	129	129	NUM
ajst-20799	126	31	and	and	CCONJ
ajst-20799	126	32	fire	fire	NOUN
ajst-20799	126	33	trucks	truck	NOUN
ajst-20799	126	34	can	can	AUX
ajst-20799	126	35	quickly	quickly	ADV
ajst-20799	126	36	reach	reach	VERB
ajst-20799	126	37	the	the	DET
ajst-20799	126	38	scene	scene	NOUN
ajst-20799	126	39	based	base	VERB
ajst-20799	126	40	on	on	ADP
ajst-20799	126	41	the	the	DET
ajst-20799	126	42	system	system	NOUN
ajst-20799	126	43	's	's	PART
ajst-20799	126	44	optimal	optimal	ADJ
ajst-20799	126	45	route	route	NOUN
ajst-20799	126	46	,	,	PUNCT
ajst-20799	126	47	ensuring	ensure	VERB
ajst-20799	126	48	timely	timely	ADJ
ajst-20799	126	49	medical	medical	ADJ
ajst-20799	126	50	treatment	treatment	NOUN
ajst-20799	126	51	for	for	ADP
ajst-20799	126	52	the	the	DET
ajst-20799	126	53	injured	injure	VERB
ajst-20799	126	54	.	.	PUNCT
ajst-20799	127	1	the	the	DET
ajst-20799	127	2	traffic	traffic	NOUN
ajst-20799	127	3	management	management	PROPN
ajst-20799	127	4	department	department	NOUN
ajst-20799	127	5	should	should	AUX
ajst-20799	127	6	promptly	promptly	ADV
ajst-20799	127	7	initiate	initiate	VERB
ajst-20799	127	8	emergency	emergency	NOUN
ajst-20799	127	9	plans	plan	NOUN
ajst-20799	127	10	,	,	PUNCT
ajst-20799	127	11	mobilize	mobilize	VERB
ajst-20799	127	12	rescue	rescue	NOUN
ajst-20799	127	13	forces	force	NOUN
ajst-20799	127	14	,	,	PUNCT
ajst-20799	127	15	manage	manage	VERB
ajst-20799	127	16	traffic	traffic	NOUN
ajst-20799	127	17	flow	flow	NOUN
ajst-20799	127	18	,	,	PUNCT
ajst-20799	127	19	and	and	CCONJ
ajst-20799	127	20	provide	provide	VERB
ajst-20799	127	21	road	road	NOUN
ajst-20799	127	22	condition	condition	NOUN
ajst-20799	127	23	information	information	NOUN
ajst-20799	127	24	to	to	PART
ajst-20799	127	25	reduce	reduce	VERB
ajst-20799	127	26	accident	accident	NOUN
ajst-20799	127	27	impacts	impact	NOUN
ajst-20799	127	28	and	and	CCONJ
ajst-20799	127	29	casualties	casualty	NOUN
ajst-20799	127	30	,	,	PUNCT
ajst-20799	127	31	thereby	thereby	ADV
ajst-20799	127	32	optimizing	optimize	VERB
ajst-20799	127	33	traffic	traffic	NOUN
ajst-20799	127	34	signal	signal	NOUN
ajst-20799	127	35	timing	timing	NOUN
ajst-20799	127	36	,	,	PUNCT
ajst-20799	127	37	enhancing	enhance	VERB
ajst-20799	127	38	law	law	NOUN
ajst-20799	127	39	enforcement	enforcement	NOUN
ajst-20799	127	40	supervision	supervision	NOUN
ajst-20799	127	41	,	,	PUNCT
ajst-20799	127	42	improving	improve	VERB
ajst-20799	127	43	road	road	NOUN
ajst-20799	127	44	infrastructure	infrastructure	NOUN
ajst-20799	127	45	,	,	PUNCT
ajst-20799	127	46	conducting	conduct	VERB
ajst-20799	127	47	safety	safety	NOUN
ajst-20799	127	48	education	education	NOUN
ajst-20799	127	49	,	,	PUNCT
ajst-20799	127	50	etc	etc	X
ajst-20799	127	51	.	.	X
ajst-20799	127	52	,	,	PUNCT
ajst-20799	127	53	to	to	PART
ajst-20799	127	54	reduce	reduce	VERB
ajst-20799	127	55	the	the	DET
ajst-20799	127	56	probability	probability	NOUN
ajst-20799	127	57	of	of	ADP
ajst-20799	127	58	traffic	traffic	NOUN
ajst-20799	127	59	accidents	accident	NOUN
ajst-20799	127	60	at	at	ADP
ajst-20799	127	61	the	the	DET
ajst-20799	127	62	source	source	NOUN
ajst-20799	127	63	.	.	PUNCT
ajst-20799	128	1	6	6	NUM
ajst-20799	128	2	.	.	X
ajst-20799	128	3	conclusion	conclusion	NOUN
ajst-20799	128	4	deep	deep	ADJ
ajst-20799	128	5	learning	learning	NOUN
ajst-20799	128	6	has	have	AUX
ajst-20799	128	7	brought	bring	VERB
ajst-20799	128	8	revolutionary	revolutionary	ADJ
ajst-20799	128	9	progress	progress	NOUN
ajst-20799	128	10	to	to	PART
ajst-20799	128	11	vehicle	vehicle	NOUN
ajst-20799	128	12	detection	detection	NOUN
ajst-20799	128	13	and	and	CCONJ
ajst-20799	128	14	tracking	tracking	NOUN
ajst-20799	128	15	technology	technology	NOUN
ajst-20799	128	16	.	.	PUNCT
ajst-20799	129	1	through	through	ADP
ajst-20799	129	2	automatic	automatic	ADJ
ajst-20799	129	3	feature	feature	NOUN
ajst-20799	129	4	learning	learning	NOUN
ajst-20799	129	5	and	and	CCONJ
ajst-20799	129	6	advanced	advanced	ADJ
ajst-20799	129	7	representation	representation	NOUN
ajst-20799	129	8	capabilities	capability	NOUN
ajst-20799	129	9	,	,	PUNCT
ajst-20799	129	10	it	it	PRON
ajst-20799	129	11	can	can	AUX
ajst-20799	129	12	overcome	overcome	VERB
ajst-20799	129	13	the	the	DET
ajst-20799	129	14	limitations	limitation	NOUN
ajst-20799	129	15	of	of	ADP
ajst-20799	129	16	traditional	traditional	ADJ
ajst-20799	129	17	methods	method	NOUN
ajst-20799	129	18	in	in	ADP
ajst-20799	129	19	complex	complex	ADJ
ajst-20799	129	20	environments	environment	NOUN
ajst-20799	129	21	,	,	PUNCT
ajst-20799	129	22	significantly	significantly	ADV
ajst-20799	129	23	improving	improve	VERB
ajst-20799	129	24	detection	detection	NOUN
ajst-20799	129	25	accuracy	accuracy	NOUN
ajst-20799	129	26	and	and	CCONJ
ajst-20799	129	27	tracking	track	VERB
ajst-20799	129	28	stability	stability	NOUN
ajst-20799	129	29	.	.	PUNCT
ajst-20799	130	1	in	in	ADP
ajst-20799	130	2	practical	practical	ADJ
ajst-20799	130	3	applications	application	NOUN
ajst-20799	130	4	,	,	PUNCT
ajst-20799	130	5	deep	deep	ADJ
ajst-20799	130	6	learning	learning	NOUN
ajst-20799	130	7	models	model	NOUN
ajst-20799	130	8	such	such	ADJ
ajst-20799	130	9	as	as	ADP
ajst-20799	130	10	faster	fast	ADJ
ajst-20799	130	11	r	r	NOUN
ajst-20799	130	12	-	-	PUNCT
ajst-20799	130	13	cnn	cnn	PROPN
ajst-20799	130	14	,	,	PUNCT
ajst-20799	130	15	yolo	yolo	PROPN
ajst-20799	130	16	,	,	PUNCT
ajst-20799	130	17	deepsort	deepsort	NOUN
ajst-20799	130	18	,	,	PUNCT
ajst-20799	130	19	and	and	CCONJ
ajst-20799	130	20	tracktor	tracktor	NOUN
ajst-20799	130	21	provide	provide	VERB
ajst-20799	130	22	strong	strong	ADJ
ajst-20799	130	23	support	support	NOUN
ajst-20799	130	24	for	for	ADP
ajst-20799	130	25	road	road	NOUN
ajst-20799	130	26	traffic	traffic	NOUN
ajst-20799	130	27	monitoring	monitoring	NOUN
ajst-20799	130	28	systems	system	NOUN
ajst-20799	130	29	with	with	ADP
ajst-20799	130	30	their	their	PRON
ajst-20799	130	31	excellent	excellent	ADJ
ajst-20799	130	32	performance	performance	NOUN
ajst-20799	130	33	and	and	CCONJ
ajst-20799	130	34	real	real	ADJ
ajst-20799	130	35	-	-	PUNCT
ajst-20799	130	36	time	time	NOUN
ajst-20799	130	37	capabilities	capability	NOUN
ajst-20799	130	38	.	.	PUNCT
ajst-20799	131	1	these	these	DET
ajst-20799	131	2	technologies	technology	NOUN
ajst-20799	131	3	not	not	PART
ajst-20799	131	4	only	only	ADV
ajst-20799	131	5	enable	enable	VERB
ajst-20799	131	6	precise	precise	ADJ
ajst-20799	131	7	detection	detection	NOUN
ajst-20799	131	8	and	and	CCONJ
ajst-20799	131	9	tracking	tracking	NOUN
ajst-20799	131	10	of	of	ADP
ajst-20799	131	11	vehicles	vehicle	NOUN
ajst-20799	131	12	but	but	CCONJ
ajst-20799	131	13	also	also	ADV
ajst-20799	131	14	effectively	effectively	ADV
ajst-20799	131	15	predict	predict	VERB
ajst-20799	131	16	traffic	traffic	NOUN
ajst-20799	131	17	accidents	accident	NOUN
ajst-20799	131	18	,	,	PUNCT
ajst-20799	131	19	providing	provide	VERB
ajst-20799	131	20	decision	decision	NOUN
ajst-20799	131	21	support	support	NOUN
ajst-20799	131	22	for	for	ADP
ajst-20799	131	23	traffic	traffic	NOUN
ajst-20799	131	24	management	management	NOUN
ajst-20799	131	25	departments	department	NOUN
ajst-20799	131	26	.	.	PUNCT
ajst-20799	132	1	in	in	ADP
ajst-20799	132	2	the	the	DET
ajst-20799	132	3	future	future	NOUN
ajst-20799	132	4	,	,	PUNCT
ajst-20799	132	5	with	with	ADP
ajst-20799	132	6	the	the	DET
ajst-20799	132	7	continuous	continuous	ADJ
ajst-20799	132	8	development	development	NOUN
ajst-20799	132	9	and	and	CCONJ
ajst-20799	132	10	optimization	optimization	NOUN
ajst-20799	132	11	of	of	ADP
ajst-20799	132	12	deep	deep	ADJ
ajst-20799	132	13	learning	learning	NOUN
ajst-20799	132	14	technology	technology	NOUN
ajst-20799	132	15	,	,	PUNCT
ajst-20799	132	16	the	the	DET
ajst-20799	132	17	level	level	NOUN
ajst-20799	132	18	of	of	ADP
ajst-20799	132	19	intelligence	intelligence	NOUN
ajst-20799	132	20	in	in	ADP
ajst-20799	132	21	vehicle	vehicle	NOUN
ajst-20799	132	22	detection	detection	NOUN
ajst-20799	132	23	and	and	CCONJ
ajst-20799	132	24	tracking	tracking	NOUN
ajst-20799	132	25	is	be	AUX
ajst-20799	132	26	expected	expect	VERB
ajst-20799	132	27	to	to	PART
ajst-20799	132	28	further	far	ADV
ajst-20799	132	29	improve	improve	VERB
ajst-20799	132	30	,	,	PUNCT
ajst-20799	132	31	playing	play	VERB
ajst-20799	132	32	a	a	DET
ajst-20799	132	33	greater	great	ADJ
ajst-20799	132	34	role	role	NOUN
ajst-20799	132	35	in	in	ADP
ajst-20799	132	36	building	build	VERB
ajst-20799	132	37	a	a	DET
ajst-20799	132	38	safer	safe	ADJ
ajst-20799	132	39	and	and	CCONJ
ajst-20799	132	40	more	more	ADV
ajst-20799	132	41	efficient	efficient	ADJ
ajst-20799	132	42	traffic	traffic	NOUN
ajst-20799	132	43	environment	environment	NOUN
ajst-20799	132	44	.	.	PUNCT
ajst-20799	133	1	references	reference	NOUN
ajst-20799	133	2	[	[	X
ajst-20799	133	3	1	1	NUM
ajst-20799	133	4	]	]	SYM
ajst-20799	133	5	yang	yang	PROPN
ajst-20799	133	6	,	,	PUNCT
ajst-20799	133	7	r.	r.	PROPN
ajst-20799	133	8	,	,	PUNCT
ajst-20799	133	9	&	&	CCONJ
ajst-20799	133	10	zhang	zhang	PROPN
ajst-20799	133	11	,	,	PUNCT
ajst-20799	133	12	g.	g.	PROPN
ajst-20799	133	13	(	(	PUNCT
ajst-20799	133	14	2024	2024	NUM
ajst-20799	133	15	)	)	PUNCT
ajst-20799	133	16	.	.	PUNCT
ajst-20799	134	1	a	a	DET
ajst-20799	134	2	review	review	NOUN
ajst-20799	134	3	of	of	ADP
ajst-20799	134	4	intelligent	intelligent	ADJ
ajst-20799	134	5	vehicle	vehicle	NOUN
ajst-20799	134	6	trajectory	trajectory	NOUN
ajst-20799	134	7	prediction	prediction	NOUN
ajst-20799	134	8	based	base	VERB
ajst-20799	134	9	on	on	ADP
ajst-20799	134	10	deep	deep	ADJ
ajst-20799	134	11	learning	learning	NOUN
ajst-20799	134	12	.	.	PUNCT
ajst-20799	135	1	automotive	automotive	ADJ
ajst-20799	135	2	abstracts	abstract	NOUN
ajst-20799	135	3	,	,	PUNCT
ajst-20799	135	4	2024(02	2024(02	NUM
ajst-20799	135	5	)	)	PUNCT
ajst-20799	135	6	,	,	PUNCT
ajst-20799	135	7	1	1	NUM
ajst-20799	135	8	-	-	SYM
ajst-20799	135	9	9	9	NUM
ajst-20799	135	10	.	.	PUNCT
ajst-20799	135	11	doi	doi	NOUN
ajst-20799	135	12	:	:	PUNCT
ajst-20799	135	13	10.19822	10.19822	NUM
ajst-20799	135	14	/	/	SYM
ajst-20799	135	15	j.cnki.16716329.20230085	j.cnki.16716329.20230085	NOUN
ajst-20799	135	16	.	.	PUNCT
ajst-20799	136	1	[	[	X
ajst-20799	136	2	2	2	NUM
ajst-20799	136	3	]	]	X
ajst-20799	136	4	fang	fang	X
ajst-20799	136	5	,	,	PUNCT
ajst-20799	136	6	h.	h.	PROPN
ajst-20799	136	7	(	(	PUNCT
ajst-20799	136	8	2023	2023	NUM
ajst-20799	136	9	)	)	PUNCT
ajst-20799	136	10	.	.	PUNCT
ajst-20799	137	1	research	research	NOUN
ajst-20799	137	2	on	on	ADP
ajst-20799	137	3	key	key	ADJ
ajst-20799	137	4	technologies	technology	NOUN
ajst-20799	137	5	of	of	ADP
ajst-20799	137	6	vehicle	vehicle	NOUN
ajst-20799	137	7	identification	identification	NOUN
ajst-20799	137	8	system	system	NOUN
ajst-20799	137	9	based	base	VERB
ajst-20799	137	10	on	on	ADP
ajst-20799	137	11	deep	deep	ADJ
ajst-20799	137	12	learning	learning	NOUN
ajst-20799	137	13	(	(	PUNCT
ajst-20799	137	14	doctoral	doctoral	ADJ
ajst-20799	137	15	dissertation	dissertation	NOUN
ajst-20799	137	16	)	)	PUNCT
ajst-20799	137	17	.	.	PUNCT
ajst-20799	138	1	huazhong	huazhong	PROPN
ajst-20799	138	2	university	university	PROPN
ajst-20799	138	3	of	of	ADP
ajst-20799	138	4	science	science	NOUN
ajst-20799	138	5	and	and	CCONJ
ajst-20799	138	6	technology	technology	NOUN
ajst-20799	138	7	.	.	PUNCT
ajst-20799	139	1	doi	doi	NOUN
ajst-20799	139	2	:	:	PUNCT
ajst-20799	139	3	10.27157	10.27157	NUM
ajst-20799	139	4	/	/	SYM
ajst-20799	139	5	d.cnki.ghzku.2021.005408	d.cnki.ghzku.2021.005408	ADJ
ajst-20799	139	6	.	.	PUNCT
ajst-20799	140	1	[	[	X
ajst-20799	140	2	3	3	NUM
ajst-20799	140	3	]	]	SYM
ajst-20799	140	4	hu	hu	PROPN
ajst-20799	140	5	,	,	PUNCT
ajst-20799	140	6	z.	z.	PROPN
ajst-20799	140	7	(	(	PUNCT
ajst-20799	140	8	2023	2023	NUM
ajst-20799	140	9	)	)	PUNCT
ajst-20799	140	10	.	.	PUNCT
ajst-20799	141	1	research	research	NOUN
ajst-20799	141	2	on	on	ADP
ajst-20799	141	3	cross	cross	ADJ
ajst-20799	141	4	-	-	ADJ
ajst-20799	141	5	camera	camera	ADJ
ajst-20799	141	6	vehicle	vehicle	NOUN
ajst-20799	141	7	reidentification	reidentification	NOUN
ajst-20799	141	8	technology	technology	NOUN
ajst-20799	141	9	based	base	VERB
ajst-20799	141	10	on	on	ADP
ajst-20799	141	11	deep	deep	ADJ
ajst-20799	141	12	learning	learning	NOUN
ajst-20799	141	13	(	(	PUNCT
ajst-20799	141	14	doctoral	doctoral	ADJ
ajst-20799	141	15	dissertation	dissertation	NOUN
ajst-20799	141	16	)	)	PUNCT
ajst-20799	141	17	.	.	PUNCT
ajst-20799	142	1	guizhou	guizhou	PROPN
ajst-20799	142	2	university	university	PROPN
ajst-20799	142	3	.	.	PUNCT
ajst-20799	143	1	doi	doi	NOUN
ajst-20799	143	2	:	:	PUNCT
ajst-20799	143	3	10.27047	10.27047	NUM
ajst-20799	143	4	/	/	SYM
ajst-20799	143	5	d.cnki.ggudu.2023.000071	d.cnki.ggudu.2023.000071	NOUN
ajst-20799	143	6	.	.	PUNCT
ajst-20799	144	1	[	[	X
ajst-20799	144	2	4	4	NUM
ajst-20799	144	3	]	]	X
ajst-20799	144	4	cao	cao	PROPN
ajst-20799	144	5	,	,	PUNCT
ajst-20799	144	6	c.	c.	PROPN
ajst-20799	144	7	(	(	PUNCT
ajst-20799	144	8	2022	2022	NUM
ajst-20799	144	9	)	)	PUNCT
ajst-20799	144	10	.	.	PUNCT
ajst-20799	145	1	research	research	NOUN
ajst-20799	145	2	on	on	ADP
ajst-20799	145	3	vehicle	vehicle	NOUN
ajst-20799	145	4	object	object	NOUN
ajst-20799	145	5	detection	detection	NOUN
ajst-20799	145	6	based	base	VERB
ajst-20799	145	7	on	on	ADP
ajst-20799	145	8	deep	deep	ADJ
ajst-20799	145	9	learning	learning	NOUN
ajst-20799	145	10	(	(	PUNCT
ajst-20799	145	11	master	master	NOUN
ajst-20799	145	12	's	's	PART
ajst-20799	145	13	thesis	thesis	NOUN
ajst-20799	145	14	)	)	PUNCT
ajst-20799	145	15	.	.	PUNCT
ajst-20799	146	1	nanjing	nanjing	PROPN
ajst-20799	146	2	university	university	PROPN
ajst-20799	146	3	of	of	ADP
ajst-20799	146	4	information	information	NOUN
ajst-20799	146	5	science	science	PROPN
ajst-20799	146	6	&	&	CCONJ
ajst-20799	146	7	technology	technology	PROPN
ajst-20799	146	8	.	.	PUNCT
ajst-20799	147	1	doi	doi	NOUN
ajst-20799	147	2	:	:	PUNCT
ajst-20799	147	3	10.27248	10.27248	NUM
ajst-20799	147	4	/	/	SYM
ajst-20799	147	5	d.cnki.gnjqc.2021.000628	d.cnki.gnjqc.2021.000628	NOUN
ajst-20799	147	6	.	.	PUNCT
ajst-20799	148	1	[	[	X
ajst-20799	148	2	5	5	NUM
ajst-20799	148	3	]	]	X
ajst-20799	148	4	wang	wang	PROPN
ajst-20799	148	5	,	,	PUNCT
ajst-20799	148	6	y.	y.	PROPN
ajst-20799	148	7	(	(	PUNCT
ajst-20799	148	8	2022	2022	NUM
ajst-20799	148	9	)	)	PUNCT
ajst-20799	148	10	.	.	PUNCT
ajst-20799	149	1	design	design	NOUN
ajst-20799	149	2	of	of	ADP
ajst-20799	149	3	road	road	NOUN
ajst-20799	149	4	vehicle	vehicle	NOUN
ajst-20799	149	5	detection	detection	NOUN
ajst-20799	149	6	system	system	NOUN
ajst-20799	149	7	based	base	VERB
ajst-20799	149	8	on	on	ADP
ajst-20799	149	9	deep	deep	ADJ
ajst-20799	149	10	learning	learning	NOUN
ajst-20799	149	11	(	(	PUNCT
ajst-20799	149	12	master	master	NOUN
ajst-20799	149	13	's	's	PART
ajst-20799	149	14	thesis	thesis	NOUN
ajst-20799	149	15	)	)	PUNCT
ajst-20799	149	16	.	.	PUNCT
ajst-20799	150	1	wuhan	wuhan	PROPN
ajst-20799	150	2	university	university	PROPN
ajst-20799	150	3	.	.	PUNCT
ajst-20799	151	1	doi	doi	NOUN
ajst-20799	151	2	:	:	PUNCT
ajst-20799	151	3	10.27379	10.27379	NUM
ajst-20799	151	4	/	/	SYM
ajst-20799	151	5	d.cnki.gwhdu.2022.000356	d.cnki.gwhdu.2022.000356	NOUN
ajst-20799	151	6	.	.	PUNCT