I
A
E
S
I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
r
t
if
ic
ia
l
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
V
ol
.
14
, N
o.
4
,
A
ugus
t
2025
, pp.
3153
~
3159
I
S
S
N
:
2252
-
8938
,
D
O
I
:
10.11591/
ij
a
i.
v
14
.i
4
.pp
3153
-
3159
3153
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
ai
.
ia
e
s
c
or
e
.c
om
U
r
b
an
i
n
c
i
d
e
n
t
d
e
t
e
c
t
i
on
b
ase
d
on
h
y
b
r
i
d
c
on
vol
u
t
i
o
n
al
n
e
u
r
al
n
e
t
w
or
k
s an
d
b
i
d
i
r
e
c
t
i
on
al
l
on
g sh
or
t
-
t
e
r
m
m
e
m
or
y
M
e
r
ye
m
A
you
, Jaou
ad
B
ou
m
h
id
i
L
3I
A
L
a
bor
a
t
or
y, F
a
c
ul
t
y of
S
c
i
e
nc
e
s
D
ha
r
E
l
M
e
hr
a
z
, S
i
di
M
oha
m
m
e
d
B
e
n A
bde
l
l
a
h
U
ni
ve
r
s
i
t
y, F
e
z
, M
or
oc
c
o
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
A
ug 8, 2024
R
e
vi
s
e
d
J
un 13, 2025
A
c
c
e
pt
e
d
J
ul
10, 2025
Rea
l
-
t
im
e
in
cid
en
t
de
te
cti
on
i
s
a
m
ajo
r
c
hal
le
nge
in
u
rba
n
ro
ad
s.
T
hi
s
p
ape
r
pro
pos
es
an
inn
ov
ati
ve
hyb
ri
d
me
th
od
for
in
cid
en
t
de
te
ct
ion
,
c
om
bin
ing
con
vol
ut
ion
al
n
eu
ra
l
ne
two
rk
s
(C
NN)
and
bid
ir
ec
tio
na
l
-
l
on
g
sh
o
rt
-
ter
m
mem
ory
(Bi
-
L
STM
).
C
NN
e
xtr
ac
ts
c
omp
le
x
sp
ati
al
f
ea
tur
es
f
ro
m
ra
w
dat
a,
whi
le
Bi
-
LS
TM
s
a
re
us
ed
fo
r
i
nc
ide
nt
de
te
cti
on
by
c
apt
ur
ing
l
on
g
-
t
erm
tem
por
al
de
pen
de
nci
es
pr
ese
nt
in
dat
a.
Th
e
p
rop
os
ed
a
lg
or
ith
m
i
s
e
v
alu
ate
d
usi
ng
s
im
ula
te
d
da
ta
fro
m
t
he
o
pe
n
-
s
ou
rce
so
ftw
ar
e
si
mu
lat
io
n
of
urb
an
mob
ili
t
y
(
SU
MO
)
.
Thi
s
c
omb
in
ati
on
im
pro
ve
s
in
ci
de
nt
d
et
ect
io
n'
s
a
c
cur
acy
and
r
ob
ust
ne
ss
b
y
ex
plo
it
ing
spa
ti
al
a
nd
t
emp
or
al
i
nfo
rm
at
ion
.
Ex
pe
ri
men
tal
res
ult
s
s
how th
at
o
ur hyb
ri
d
ap
pr
oac
h
o
utp
er
for
ms the
su
ppo
rt vec
to
r
m
ach
ine
(SV
M),
ran
do
m
for
es
t
(R
F),
a
nd
B
i
-
LST
M
al
gor
it
hms
,
wi
th
a
s
ub
sta
nti
al
dec
rea
se
i
n f
al
se
po
sit
iv
es
an
d t
he
sp
ee
d o
f
det
ec
ti
ng
ur
gen
t
sit
ua
tio
ns
.
K
e
y
w
o
r
d
s
:
B
id
ir
e
c
ti
ona
l
lo
ng s
hor
t
-
te
r
m
m
e
m
or
y
C
onvolut
io
na
l
ne
ur
a
l
ne
twor
ks
I
nc
id
e
nt
de
te
c
ti
on
M
a
c
hi
ne
l
e
a
r
ni
ng
U
r
ba
n r
oa
ds
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
M
e
r
ye
m
A
you
L
3I
A
L
a
bor
a
to
r
y,
F
a
c
ul
ty
of
S
c
ie
nc
e
s
D
ha
r
E
l
M
e
hr
a
z
,
S
id
i
M
oha
m
m
e
d B
e
n A
bde
ll
a
h U
ni
ve
r
s
it
y
F
e
z
3000, M
or
oc
c
o
E
m
a
il
:
m
e
r
ye
m
e
.a
you@
us
m
ba
.a
c
.m
a
1.
I
N
T
R
O
D
U
C
T
I
O
N
T
r
a
f
f
i
c
c
on
ge
s
t
io
n
c
a
u
s
e
s
m
a
ny
s
oc
io
-
e
c
on
om
i
c
a
nd
e
n
vi
r
onm
e
nt
a
l
is
s
ue
s
by
e
xt
e
n
de
d
tr
a
v
e
l
ti
m
e
,
hi
gh
e
r
f
u
e
l
c
on
s
um
p
ti
o
n,
a
nd
m
or
e
p
ol
l
ut
i
on
.
T
he
s
e
pr
o
bl
e
m
s
b
e
c
om
e
e
v
e
n
m
or
e
a
c
ut
e
du
e
to
e
v
e
r
-
i
nc
r
e
a
s
in
g
tr
a
f
f
i
c
a
nd
a
n
in
a
d
e
q
ua
te
r
oa
dw
a
y
s
y
s
t
e
m
. C
o
nv
e
n
ti
o
na
l
w
a
y
s
of
m
a
n
a
gi
ng
tr
a
f
f
i
c
,
s
u
c
h
a
s
bui
ld
i
ng
m
o
r
e
r
o
a
d
s
,
a
r
e
b
ot
h
e
x
pe
n
s
iv
e
a
nd
di
f
f
ic
ul
t
t
o
im
p
le
m
e
nt
in
ur
ba
n
c
e
nt
e
r
s
w
it
h
de
ns
e
p
opu
la
ti
o
ns
.
A
c
c
or
d
in
g
ly
,
i
nc
r
e
a
s
in
g
th
e
pr
e
c
i
s
io
n
a
n
d
e
f
f
e
c
t
iv
e
n
e
s
s
i
n
t
r
a
f
f
i
c
in
c
id
e
n
t
de
t
e
c
ti
on
i
s
a
n
i
m
por
t
a
nt
r
e
s
e
a
r
c
h
t
o
pi
c
.
D
ur
in
g
th
e
la
s
t
two
d
e
c
a
de
s
,
di
f
f
e
r
e
nt
m
e
th
ods
h
a
ve
be
e
n
s
ug
ge
s
te
d
to
d
e
te
c
t
s
u
c
h
e
v
e
nt
s
[
1]
f
r
om
c
la
s
s
ic
s
e
ns
or
-
ba
s
e
d
s
y
s
te
m
s
to
a
dva
nc
e
d
m
a
c
hi
ne
le
a
r
ni
ng
te
c
hni
que
s
[
2]
.
E
a
r
ly
te
c
hni
que
s
w
e
r
e
ba
s
e
d
on
lo
op
de
te
c
to
r
s
[
3]
,
[
4]
a
nd
s
ta
ti
s
ti
c
a
l
a
lg
or
it
hm
s
[
5]
,
[
6]
w
hi
c
h,
e
ve
n
if
e
f
f
e
c
ti
ve
,
a
r
e
c
ha
r
a
c
te
r
iz
e
d
by
hi
gh
im
pl
e
m
e
nt
a
ti
on
c
os
ts
a
nd
e
nvi
r
onm
e
nt
a
l
s
us
c
e
pt
ib
il
it
y.
H
ig
he
r
-
le
ve
l
m
ode
ls
,
s
uc
h
a
s
m
a
c
hi
ne
le
a
r
ni
ng
a
lg
or
it
hm
s
s
tu
dyi
ng
tr
a
f
f
ic
-
f
lo
w
pa
tt
e
r
ns
,
ha
ve
be
e
n
s
tu
di
e
d
a
s
w
e
ll
,
onl
y
to
e
nc
ount
e
r
di
f
f
ic
ul
ti
e
s
in
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
a
nd
to
gr
ow
pr
obl
e
m
a
ti
c
in
t
e
r
m
s
of
c
om
put
a
ti
ona
l
r
e
qui
r
e
m
e
nt
s
.
O
w
in
g
to
th
e
pr
ogr
e
s
s
in
de
e
p
le
a
r
ni
ng,
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
k
(
C
N
N
)
a
nd
lo
ng
s
hor
t
-
te
r
m
m
e
m
or
y
(
L
S
T
M
)
m
ode
ls
ha
ve
s
how
n
pr
om
is
in
g
pe
r
f
or
m
a
nc
e
f
or
th
e
in
c
id
e
nt
de
te
c
ti
on
ta
s
k.
C
N
N
s
a
r
e
a
ls
o
pa
r
ti
c
ul
a
r
ly
e
f
f
e
c
ti
ve
in
r
e
pr
e
s
e
nt
in
g
th
e
s
pa
ti
a
l
f
e
a
tu
r
e
s
of
tr
a
f
f
ic
da
ta
[
7]
–
[
9
]
,
w
hi
c
h
a
r
e
a
ls
o
a
bl
e
to
m
ode
l
r
oa
d
-
le
ve
l
va
r
ia
ti
ons
a
nd
c
onge
s
ti
on
pa
tt
e
r
ns
.
L
S
T
M
s
a
r
e
w
e
ll
known
f
or
m
ode
li
ng
te
m
por
a
l
d
e
pe
n
de
nc
ie
s
,
w
hi
c
h
a
ll
ow
th
e
s
ys
te
m
to
a
n
a
ly
z
e
th
e
s
e
que
nt
ia
l
tr
a
f
f
ic
f
lu
c
tu
a
ti
on
th
r
ough
ti
m
e
.
O
ur
w
or
k
c
ont
r
ib
ut
e
s
to
th
e
li
te
r
a
tu
r
e
by
de
m
ons
tr
a
ti
ng
th
a
t
e
xpl
oi
ti
ng
th
e
c
om
pl
e
m
e
nt
a
r
y
na
tu
r
e
of
C
N
N
a
nd
L
S
T
M
a
nd
f
us
in
g
th
e
ir
out
put
s
,
te
r
m
e
d
a
hyb
r
id
m
ode
l,
c
oul
d s
ig
ni
f
ic
a
nt
ly
i
m
pr
ove
t
he
a
c
c
ur
a
c
y a
nd r
obus
tn
e
s
s
of
t
r
a
f
f
ic
i
nc
id
e
nt
de
te
c
ti
on.
T
hi
s
w
or
k
is
a
f
ur
th
e
r
s
te
p
in
th
e
a
r
e
a
of
de
e
p
le
a
r
ni
ng
f
or
t
r
a
f
f
ic
a
na
ly
s
is
.
F
or
e
xa
m
pl
e
,
A
hm
a
dz
a
de
h
e
t
al
.
[
10]
us
e
d
a
1
D
-
C
N
N
to
e
xt
r
a
c
t
hi
gh
-
s
p
e
e
d
t
r
a
in
f
a
ul
t
s
ig
na
ls
a
nd s
how
e
d
th
e
pot
e
nt
ia
l
f
or
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
4
,
A
ugus
t
2025
:
3153
-
3159
3154
c
a
pt
ur
in
g
us
e
f
ul
s
pa
ti
a
l
f
e
a
tu
r
e
s
.
A
c
c
or
di
ng
to
L
i
e
t
al
.
[
11]
a
n
hybr
id
m
e
th
od
c
om
bi
ni
ng
C
N
N
a
nd
L
S
T
M
is
pr
opos
e
d
to
pr
e
di
c
t
tr
a
in
a
r
r
iv
a
l
de
la
ys
.
O
ur
w
or
k
e
xt
e
nds
th
e
s
ta
te
-
of
-
th
e
-
a
r
t
f
in
di
ngs
by
in
te
gr
a
ti
ng
C
N
N
s
f
or
s
pa
ti
a
l
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
a
nd
L
S
T
M
s
f
or
m
ode
li
ng
te
m
por
a
l
s
e
que
nc
e
s
,
pr
ovi
di
ng
a
c
om
pr
e
he
ns
iv
e
f
r
a
m
e
w
or
k f
or
de
te
c
ti
ng i
nc
id
e
nt
s
i
n r
e
a
l
ti
m
e
T
he
r
e
m
a
in
de
r
of
th
is
pa
pe
r
is
s
tr
uc
tu
r
e
d
a
s
f
ol
lo
w
s
.
S
e
c
ti
on
2
out
li
ne
s
th
e
da
ta
a
c
qui
s
it
io
n
pr
oc
e
s
s
a
nd
th
e
m
e
th
odol
ogy
a
dopt
e
d
f
or
de
te
c
ti
ng
in
c
id
e
nt
s
on
ur
ba
n
r
oa
d
ne
twor
ks
.
S
e
c
ti
on
3
pr
ovi
de
s
a
de
ta
il
e
d
de
s
c
r
ip
ti
on
of
th
e
pr
opos
e
d
a
lg
or
it
hm
.
S
e
c
ti
on
4
pr
e
s
e
n
ts
th
e
e
xpe
r
im
e
nt
a
l
r
e
s
ul
ts
a
lo
ng
w
it
h
a
c
om
pr
e
he
ns
iv
e
a
na
ly
s
is
. F
in
a
ll
y,
s
e
c
ti
on 5 c
onc
lu
de
s
t
h
e
pa
pe
r
.
2.
I
N
C
I
D
E
N
T
D
E
T
E
C
T
I
O
N
M
E
T
H
O
D
O
L
O
G
Y
2.1.
A
u
t
om
at
ic
i
n
c
id
e
n
t
d
e
t
e
c
t
io
n
s
ys
t
e
m
s
A
n
a
ut
om
a
ti
c
in
c
id
e
nt
de
te
c
ti
on
(
A
I
D
)
s
ys
t
e
m
is
e
s
s
e
nt
ia
l
f
or
th
e
c
ont
r
ol
of
tr
a
f
f
ic
f
lo
w
a
nd
opt
im
iz
in
g
c
a
pa
c
it
y
in
tr
a
ns
por
ta
ti
on
ne
twor
ks
.
T
he
s
e
s
ys
t
e
m
s
de
pl
oy
s
om
e
s
ta
te
-
of
-
th
e
-
a
r
t
te
c
hnol
ogi
e
s
to
s
e
ns
e
th
e
tr
a
f
f
ic
c
ondi
ti
ons
,
s
uc
h
a
s
a
c
c
id
e
nt
s
,
c
ong
e
s
ti
on,
a
m
ong
ot
he
r
s
,
in
ti
m
e
a
nd
w
it
h
hi
gh
a
c
c
ur
a
c
y.
T
hr
ough
th
e
us
e
of
da
ta
f
r
om
m
ul
ti
p
le
s
our
c
e
s
,
in
c
lu
di
ng
c
a
m
e
r
a
s
a
nd
s
e
n
s
or
s
.
T
he
r
e
a
r
e
di
f
f
e
r
e
nt
ty
pe
s
of
A
I
D
s
ys
te
m
s
a
va
il
a
bl
e
,
s
u
c
h
a
s
lo
op
f
in
de
r
s
.
T
he
s
e
s
ys
te
m
s
us
e
in
-
r
oa
d
e
m
be
dde
d
s
e
ns
or
s
to
r
e
c
ogni
z
e
c
ha
nge
s
[
12]
.
S
uc
h
s
ys
te
m
s
a
r
e
known
a
s
c
a
m
e
r
a
-
ba
s
e
d
be
c
a
u
s
e
tr
a
f
f
ic
is
r
e
c
or
de
d
on
f
il
m
by
m
ount
in
g
th
e
c
a
m
e
r
a
s
ove
r
th
e
r
oa
d.
A
lg
or
it
hm
s
[
13]
–
[
15]
a
r
e
us
e
d
to
pe
r
f
or
m
r
e
a
l
-
ti
m
e
a
na
ly
s
is
of
th
e
vi
de
o
to
id
e
nt
if
y
c
ol
li
s
io
ns
or
c
a
r
s
in
w
r
ong
di
r
e
c
ti
on.
A
I
D
s
y
s
te
m
s
u
s
e
c
om
pl
e
x
a
lg
or
it
hm
s
,
s
uc
h
a
s
m
a
c
hi
ne
le
a
r
ni
ng,
to
id
e
nt
if
y
a
nom
a
li
e
s
a
nd
is
s
u
e
a
le
r
ts
th
a
t
f
a
c
il
it
a
te
a
qui
c
k
r
e
s
pons
e
.
T
hi
s
f
a
s
t
di
s
c
ove
r
y
a
nd
r
e
s
pon
s
e
pr
oc
e
dur
e
gr
e
a
tl
y
m
it
ig
a
te
s
th
e
e
f
f
e
c
t
of
s
uc
h
in
c
id
e
nt
s
,
f
or
a
s
a
f
e
r
,
le
s
s
de
la
ye
d
f
lo
w
of
tr
a
f
f
ic
.
W
it
h
th
e
de
ve
lo
pm
e
nt
of
te
c
hnol
ogy,
d
e
e
p
le
a
r
ni
ng
m
ode
l
s
,
c
oul
d
m
a
ke
th
e
s
y
s
te
m
m
or
e
pr
e
c
i
s
e
a
nd
de
pe
nd
a
bl
e
,
w
hi
c
h
c
a
n
be
one
of
th
e
a
s
pe
c
ts
f
or
c
ons
tr
uc
ti
ng
a
s
m
a
r
te
r
a
n
d
m
or
e
r
obus
t
tr
a
ns
por
ta
ti
on
tr
a
f
f
ic
de
te
c
ti
on
pr
ovi
de
s
th
e
s
ys
te
m
tr
a
f
f
ic
tr
a
f
f
ic
-
r
e
la
te
d
in
f
or
m
a
ti
on
f
or
id
e
nt
if
yi
ng
th
e
oc
c
ur
r
e
nc
e
of
a
tr
a
f
f
ic
e
ve
nt
.
O
th
e
r
da
ta
, s
uc
h a
s
s
pe
e
d, volum
e
, a
nd oc
c
up
a
nc
y
[
16]
, a
r
e
of
te
n pr
ovi
de
d a
t
a
l
ow
e
r
r
a
te
.
2.2.
D
at
a c
ol
le
c
t
io
n
T
he
de
te
c
ti
on
a
lg
or
it
hm
f
unda
m
e
nt
a
ll
y
r
e
li
e
s
on
a
na
ly
z
in
g
c
ha
nge
s
in
tr
a
f
f
ic
da
ta
.
V
a
r
io
us
m
e
tr
ic
s
,
s
uc
h
a
s
v
e
lo
c
it
y,
oc
c
upa
nc
y
r
a
te
,
a
nd
tr
a
f
f
ic
f
lo
w
,
a
r
e
us
e
d
to
de
pi
c
t
tr
a
f
f
ic
c
ondi
ti
ons
.
I
n
th
is
s
tu
dy,
tr
a
f
f
ic
a
nd
in
c
id
e
nt
da
ta
w
e
r
e
ge
ne
r
a
te
d
us
in
g
s
im
ul
a
ti
on
of
ur
ba
n
m
obi
li
ty
(
S
U
M
O
)
.
I
nduc
ti
ve
lo
op
de
te
c
to
r
s
c
a
pt
ur
e
d
tr
a
f
f
ic
dyna
m
ic
s
a
t
a
30
-
s
e
c
ond
r
e
s
ol
ut
io
n,
m
e
a
s
ur
i
ng
s
pe
e
d,
vol
um
e
,
a
nd
oc
c
up
a
nc
y.
V
e
lo
c
it
y
r
e
pr
e
s
e
nt
s
th
e
a
ve
r
a
ge
s
pe
e
d
of
ve
hi
c
le
s
w
it
hi
n
e
a
c
h
30
-
s
e
c
o
nd
in
te
r
va
l,
vol
um
e
in
di
c
a
te
s
th
e
num
be
r
o
f
ve
hi
c
le
s
pa
s
s
in
g
th
r
ough
e
a
c
h
l
a
ne
,
a
nd
oc
c
upa
nc
y
r
e
f
le
c
ts
th
e
pr
opor
ti
on
of
ti
m
e
th
e
de
te
c
to
r
w
a
s
oc
c
upi
e
d
by ve
hi
c
le
s
dur
in
g t
he
de
te
c
ti
on w
in
dow
.
T
r
a
f
f
i
c
in
c
id
e
n
t
s
c
a
n
ta
ke
v
a
r
i
ou
s
f
or
m
s
,
i
nc
lu
di
n
g
a
c
c
i
de
nt
s
,
r
o
a
d
w
or
k
s
,
a
d
ve
r
s
e
w
e
a
th
e
r
c
o
ndi
ti
o
n
s
,
or
h
e
a
vy
c
ong
e
s
ti
on
,
a
ll
of
w
hi
c
h
d
is
r
up
t
th
e
nor
m
a
l
f
l
ow
of
ve
hi
c
l
e
s
.
T
o
il
lu
s
tr
a
t
e
t
hi
s
ph
e
n
om
e
no
n,
w
e
s
im
ul
a
t
e
d
a
n
i
nc
i
d
e
nt
b
y
in
te
nt
i
on
a
ll
y
s
to
ppi
ng
a
c
a
r
a
t
a
pr
e
de
f
i
ne
d
lo
c
a
ti
o
n
f
or
a
f
ix
e
d
pe
r
io
d
.
T
hi
s
c
on
tr
o
ll
e
d
s
c
e
na
r
io
a
ll
ow
s
u
s
t
o a
na
ly
z
e
t
h
e
i
m
p
a
c
t
of
s
uc
h di
s
r
upt
io
n
s
on t
r
a
f
f
i
c
f
lo
w
a
nd
e
v
a
l
ua
te
t
h
e
e
f
f
e
c
ti
ve
ne
s
s
of
our
de
t
e
c
ti
on
m
od
e
l.
F
ig
ur
e
1
pr
ovi
de
s
a
v
i
s
u
a
l
r
e
pr
e
s
e
nt
a
ti
on
of
t
hi
s
s
im
ul
a
t
e
d
i
nc
id
e
nt
.
F
i
gur
e
2
f
ur
th
e
r
i
ll
u
s
tr
a
te
s
how
t
h
e
in
c
id
e
nt
a
f
f
e
c
ts
t
r
a
f
f
i
c
m
e
tr
i
c
s
a
s
d
e
t
e
c
te
d b
y
up
s
tr
e
a
m
a
nd
do
w
n
s
tr
e
a
m
l
oo
p d
e
t
e
c
to
r
s
.
I
n
th
e
e
ve
nt
of
a
tr
a
f
f
ic
in
c
id
e
nt
,
di
s
ti
nc
ti
ve
pa
tt
e
r
ns
e
m
e
r
ge
in
bot
h
ups
tr
e
a
m
a
nd
dow
n
s
tr
e
a
m
lo
op
de
te
c
to
r
s
.
E
xpe
c
te
d
c
ha
nge
s
in
c
lu
de
a
r
e
duc
ti
on
in
up
s
tr
e
a
m
s
pe
e
d
a
nd
vol
um
e
,
a
lo
ng
w
it
h
a
de
c
r
e
a
s
e
in
dow
ns
tr
e
a
m
vol
um
e
.
C
onc
ur
r
e
nt
ly
,
th
e
r
e
is
a
n
in
c
r
e
a
s
e
in
ups
tr
e
a
m
oc
c
upa
nc
y
a
nd
va
r
ia
ti
ons
in
dow
n
s
tr
e
a
m
s
pe
e
d a
nd oc
c
upa
nc
y. I
n
(
1)
de
s
c
r
ib
e
s
how
V
i
s
c
om
put
e
d.
=
∑
=
1
(
1)
W
he
r
e
N
is
th
e
num
be
r
of
ve
hi
c
le
s
a
t
a
de
te
c
ti
on
lo
c
a
ti
on
ove
r
a
gi
ve
n
ti
m
e
pe
r
io
d,
a
nd
v
i
is
th
e
i
th
ve
lo
c
it
y.
O
c
c
upa
nc
y r
a
te
i
s
c
om
put
e
d a
s
out
li
ne
d i
n (
2)
.
=
∑
=
1
(
2)
W
he
r
e
L
is
th
e
le
ngt
h
of
th
e
obs
e
r
ve
d
r
oa
d,
a
nd
L
i
is
th
e
le
ngt
h
of
th
e
i
th
ve
hi
c
le
.
T
he
num
be
r
of
c
a
r
s
th
a
t
pa
s
s
th
r
ough
a
de
te
c
ti
on
poi
nt
in
a
pr
e
de
te
r
m
in
e
d
a
m
ount
o
f
ti
m
e
is
r
e
f
e
r
r
e
d
to
a
s
tr
a
f
f
ic
f
lo
w
(
T
)
.
T
he
m
a
th
e
m
a
ti
c
a
l
f
or
m
ul
a
ti
on of
t
r
a
f
f
ic
f
lo
w
i
s
gi
ve
n i
n (
3)
.
=
∑
=
0
(
3)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
U
r
ban inc
id
e
nt
de
te
c
ti
on ba
s
e
d on hy
br
id
c
onv
ol
ut
io
nal
ne
ur
al
ne
tw
or
k
s
…
(
M
e
r
y
e
m
A
y
ou)
3155
W
he
r
e
is
th
e
ti
m
e
in
te
r
va
l
a
nd
N
i
is
th
e
num
be
r
of
ve
hi
c
le
s
s
e
e
n
a
t
a
de
te
c
ti
on
lo
c
a
ti
on
w
it
hi
n
a
1
-
s
e
c
ond
in
te
r
va
l.
F
ig
ur
e
1.
T
r
a
f
f
ic
i
nc
id
e
nt
pa
tt
e
r
ns
F
ig
ur
e
2
. I
ll
us
tr
a
ti
on of
a
r
oa
d i
nc
id
e
nt
3.
D
E
S
I
G
N
O
F
I
N
C
I
D
E
N
T
D
E
T
E
C
T
I
O
N
M
O
D
E
L
B
A
S
E
D
O
N
A
H
Y
B
R
I
D
BI
-
L
S
T
M
A
N
D
1
D
C
N
N
3.1.
C
on
vol
u
t
io
n
al
n
e
u
r
al
n
e
t
w
or
k
C
N
N
s
[
17]
–
[
19]
ty
pi
c
a
ll
y c
ons
is
t
of
a
f
ul
ly
c
onne
c
te
d l
a
ye
r
, po
ol
in
g l
a
ye
r
s
, r
e
la
te
d w
e
ig
ht
s
, a
nd one
or
m
or
e
c
onvolut
io
na
l
la
ye
r
s
.
T
o
e
xt
r
a
c
t
f
e
a
tu
r
e
s
,
th
e
c
onvolut
io
na
l
la
ye
r
m
a
ke
s
us
e
of
lo
c
a
l
c
or
r
e
la
ti
ons
in
th
e
da
ta
.
E
xc
e
pt
th
e
c
onvolut
io
n
pr
oc
e
dur
e
be
in
g
li
m
it
e
d
to
o
ne
di
m
e
ns
io
n,
th
e
1D
C
N
N
is
c
om
pa
r
a
bl
e
to
two
-
di
m
e
ns
io
na
l
C
N
N
s
.
A
s
a
r
e
s
ul
t,
it
ha
s
a
s
ha
ll
ow
a
r
c
hi
te
c
tu
r
e
th
a
t
c
a
n
be
tr
a
in
e
d
on
e
m
be
dde
d
de
ve
lo
pm
e
nt
boa
r
ds
or
e
ve
n
or
di
na
r
y
C
P
U
s
[
20]
.
F
or
c
la
s
s
if
ic
a
ti
on
a
ppl
ic
a
ti
ons
,
th
e
c
onvolut
io
n
m
e
th
od
e
xt
r
a
c
ts
us
e
f
ul
hi
e
r
a
r
c
hi
c
a
l
f
e
a
tu
r
e
s
f
r
om
a
da
ta
s
e
t.
3.2.
B
id
ir
e
c
t
io
n
al
-
lo
n
g s
h
or
t
-
t
e
r
m
m
e
m
o
r
y
T
he
bi
di
r
e
c
ti
ona
l
-
lo
ng
s
hor
t
-
te
r
m
m
e
m
or
y
(
Bi
-
L
S
T
M
)
,
w
hi
c
h
e
xt
r
a
c
ts
s
ig
ni
f
ic
a
nt
c
h
a
r
a
c
te
r
is
ti
c
s
[
21]
,
[
22]
,
w
a
s
a
ppl
ie
d
in
th
is
r
e
s
e
a
r
c
h.
A
n
in
put
ga
te
,
out
put
ga
te
,
f
or
ge
t
ga
te
,
a
nd
m
e
m
or
y
uni
t
(
c
e
ll
)
a
r
e
c
om
pone
nt
s
of
t
he
L
S
T
M
s
tr
uc
tu
r
e
[
23]
. T
he
a
c
ti
va
ti
on of
t
hi
s
ga
te
i
s
f
ound us
in
g (
4)
[
24]
.
=
(
[
,
ℎ
−
1
,
−
1
]
+
)
(
4)
T
he
in
put
s
e
que
nc
e
is
de
not
e
d
by
;
ℎ
−
1
r
e
pr
e
s
e
nt
s
th
e
pr
e
vi
ous
bl
oc
k
out
put
;
−
1
de
not
e
s
th
e
pr
e
vi
ous
L
S
T
M
bl
oc
k
m
e
m
or
y;
de
not
e
s
th
e
bi
a
s
v
e
c
to
r
;
r
e
pr
e
s
e
nt
s
th
e
in
di
vi
dua
l
w
e
ig
ht
ve
c
to
r
s
;
a
nd
de
not
e
s
th
e
s
ig
m
oi
d
f
unc
ti
on.
T
h
e
in
put
ga
te
ut
il
iz
e
s
a
b
a
s
ic
n
e
ur
a
l
ne
twor
k
w
it
h
a
ta
nh
a
c
ti
va
ti
on
f
unc
ti
on
a
nd
th
e
e
f
f
e
c
t
of
th
e
pr
e
vi
ous
m
e
m
or
y
bl
oc
k
to
c
r
e
a
te
ne
w
m
e
m
or
y.
T
he
ope
r
a
ti
ons
pe
r
f
or
m
e
d
f
or
th
e
s
e
c
a
lc
ul
a
ti
ons
by (
5)
a
nd (
6)
.
=
(
[
,
ℎ
−
1
,
−
1
]
+
)
(
5)
=
∙
−
1
+
∙
tan
ℎ
(
[
,
ℎ
−
1
,
−
1
]
)
+
(
6)
A
c
ons
c
io
u
s
ly
de
s
ig
ni
ng
a
nd
r
e
tr
ie
vi
ng
lo
ng
-
te
r
m
in
f
or
m
a
ti
on
c
a
n
a
ll
e
vi
a
te
th
e
pr
obl
e
m
of
lo
ng
-
te
r
m
de
pe
nde
nc
y t
ha
t
th
e
L
S
T
M
n
e
twor
ks
a
r
e
pr
one
t
o i
n a
c
tu
a
l
a
ppl
ic
a
ti
ons
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
4
,
A
ugus
t
2025
:
3153
-
3159
3156
3.3.
H
yb
r
id
CNN
-
Bi
-
L
S
T
M
m
od
e
l
f
o
r
i
n
c
id
e
n
t
d
e
t
e
c
t
io
n
F
or
s
im
pl
e
ope
r
a
ti
on
a
nd
f
a
s
te
r
c
om
put
a
ti
on
s
pe
e
d,
th
e
1D
C
N
N
is
e
m
pl
oye
d
to
e
xt
r
a
c
t
s
pa
ti
a
l
f
e
a
tu
r
e
s
f
r
om
th
e
r
a
w
tr
a
f
f
ic
da
ta
.
T
hi
s
a
r
c
hi
te
c
tu
r
e
is
pa
r
ti
c
u
la
r
ly
e
f
f
e
c
ti
ve
in
r
e
duc
in
g
th
e
c
om
put
a
ti
ona
l
c
om
pl
e
xi
ty
w
hi
le
r
e
ta
in
in
g
e
s
s
e
nt
ia
l
s
p
a
ti
a
l
pa
tt
e
r
ns
.
I
n
a
dd
it
io
n,
th
e
B
i
-
L
S
T
M
is
a
dopt
e
d
f
or
in
c
id
e
nt
de
te
c
ti
on
a
nd
f
or
c
a
pt
ur
in
g
lo
ng
-
te
r
m
te
m
por
a
l
de
pe
nde
nc
ie
s
in
ti
m
e
s
e
r
ie
s
da
ta
,
m
a
ki
ng
th
e
m
ode
l
m
or
e
a
c
c
ur
a
te
i
n r
e
c
ogni
z
in
g c
om
pl
e
x t
r
a
f
f
ic
dyna
m
ic
s
.
A
lg
or
it
hm
1 i
ll
us
tr
a
te
s
t
he
pr
opos
e
d 1D
C
N
N
-
Bi
-
L
S
T
M
.
A
lg
or
it
hm
1:
H
ybr
id
pr
opos
e
d m
e
th
od 1D C
N
N
-
Bi
-
L
S
T
M
S
te
p 1:
in
it
ia
li
z
e
t
r
a
in
in
g s
e
t
:
−
C
ons
id
e
r
a
t
r
a
in
in
g s
e
t
=
{
(
,
)
|
∈
ℝ
,
∈
ℝ
,
=
1
,
…
.
}
S
te
p 2:
da
ta
s
e
t
s
pl
it
ti
ng
S
te
p 3:
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
w
it
h 1 D C
N
N
:
−
A
ppl
y a
1D
C
N
N
l
a
ye
r
t
o e
xt
r
a
c
t
lo
c
a
l
f
e
a
tu
r
e
s
f
r
om
t
he
i
nput
da
ta
.
S
te
p 4:
te
m
por
a
l
f
e
a
tu
r
e
pr
oc
e
s
s
in
g
w
it
h
Bi
-
L
S
T
M
−
U
s
e
Bi
-
L
S
T
M
la
ye
r
s
to
e
nha
nc
e
th
e
e
xt
r
a
c
te
d
f
e
a
tu
r
e
s
by
c
a
pt
ur
in
g
te
m
por
a
l
de
pe
nde
nc
ie
s
a
nd
pa
tt
e
r
ns
.
S
te
p 5:
pr
e
ve
nt
ove
r
f
it
ti
ng
:
−
I
nc
or
por
a
te
dr
opout
la
ye
r
s
t
o pr
e
ve
nt
ove
r
f
it
ti
ng.
S
te
p 6:
N
or
m
a
li
z
a
ti
on:
−
U
s
e
a
ba
tc
h nor
m
a
li
z
a
ti
on
la
y
e
r
t
o s
ta
bi
li
z
e
a
nd
a
c
c
e
le
r
a
te
t
he
t
r
a
in
in
g pr
oc
e
s
s
.
S
te
p 7:
de
ns
e
l
a
ye
r
s
pr
oc
e
s
s
in
g
:
−
A
dd
f
ul
ly
c
onne
c
te
d
la
ye
r
s
t
o i
nt
e
gr
a
te
t
he
e
xt
r
a
c
te
d f
e
a
tu
r
e
s
a
n
d m
a
ke
de
te
c
ti
on.
S
te
p 8:
pe
r
f
or
m
a
nc
e
m
e
a
s
ur
e
m
e
nt
:
−
I
m
pl
e
m
e
nt
pe
r
f
or
m
a
nc
e
m
e
a
s
ur
e
s
t
o a
s
s
e
s
s
t
he
c
la
s
s
if
ic
a
ti
on a
c
c
ur
a
c
y a
nd othe
r
r
e
le
va
nt
m
e
tr
ic
4.
R
E
S
U
L
T
S
A
N
D
D
I
S
C
U
S
S
I
O
N
4.1.
P
e
r
f
or
m
an
c
e
e
val
u
at
io
n
c
r
it
e
r
ia
T
o e
va
lu
a
te
t
he
e
f
f
e
c
ti
ve
ne
s
s
of
t
he
c
a
te
gor
iz
a
ti
on mode
ls
, t
hr
e
e
m
e
tr
ic
s
w
e
r
e
us
e
d:
a
c
c
ur
a
c
y, r
e
c
a
ll
,
a
nd
pr
e
c
is
io
n.
A
s
de
m
ons
tr
a
te
d
by
(
7)
,
a
c
c
ur
a
c
y
m
e
a
s
ur
e
s
a
c
l
a
s
s
if
ie
r
'
s
ove
r
a
ll
c
or
r
e
c
tn
e
s
s
.
R
e
c
a
ll
qu
a
nt
if
ie
s
th
e
pe
r
c
e
nt
a
ge
of
pos
it
iv
e
r
e
c
or
ds
th
a
t
th
e
c
la
s
s
if
ie
r
a
c
c
ur
a
te
ly
pr
e
di
c
ts
.
A
s
a
r
e
s
ul
t,
r
e
c
a
ll
is
c
a
l
c
ul
a
te
d
us
in
g
(
8)
.
C
onve
r
s
e
ly
,
pr
e
c
is
io
n,
a
s
s
how
n
in
(
9)
,
is
th
e
pe
r
c
e
nt
a
ge
of
tr
ue
pos
it
iv
e
(
T
P
)
r
e
c
or
ds
a
m
ong
a
ll
pos
it
iv
e
ly
pr
e
di
c
te
d r
e
c
or
ds
.
=
(
+
)
+
+
+
(
7)
R
e
c
a
l
l
=
+
(
8)
=
+
(
9)
4.2
.
S
im
u
la
t
io
n
of
u
r
b
an
r
oad
m
ob
il
it
y
S
U
M
O
m
o
d
e
l
s
a
n
d
e
xa
m
in
e
s
s
e
v
e
r
a
l
a
s
pe
c
t
s
of
tr
a
n
s
por
ta
ti
o
n
a
nd
m
obi
li
t
y
in
ur
ba
n
s
e
t
ti
n
g
s
u
s
in
g
c
om
put
e
r
-
ba
s
e
d
s
im
ul
a
ti
on
m
e
th
od
ol
o
gi
e
s
.
T
o
r
e
s
e
a
r
c
h a
n
d
f
or
e
c
a
s
t
how
va
r
io
u
s
e
nt
it
i
e
s
w
i
ll
m
o
ve
a
n
d i
nt
e
r
a
c
t
w
it
h
in
c
i
ti
e
s
,
t
hi
s
m
e
t
ho
d
i
nvo
lv
e
s
b
ui
l
di
n
g
v
ir
t
ua
l
m
o
d
e
l
s
of
m
e
tr
op
ol
it
a
n
e
nvi
r
on
m
e
nt
s
,
c
om
pl
e
t
e
w
it
h
r
oa
d
ne
t
w
or
k
s
,
v
e
hi
c
l
e
s
,
a
nd
o
th
e
r
r
e
l
e
v
a
n
t
e
l
e
m
e
n
ts
[
25]
.
A
lo
ng
s
i
de
th
e
a
f
or
e
m
e
nt
i
on
e
d
f
u
nc
ti
o
na
li
t
ie
s
,
S
U
M
O
ut
il
iz
e
s
tr
a
f
f
ic
c
ont
r
ol
in
t
e
r
f
a
c
e
(
T
r
a
C
I
)
t
o
e
na
bl
e
s
e
a
m
l
e
s
s
r
un
ti
m
e
in
t
e
r
a
c
ti
o
n
w
it
h
e
xt
e
r
n
a
l
pr
ogr
a
m
s
[
26]
.
4.3.
T
h
e
s
c
e
n
ar
io
u
s
e
d
i
n
t
h
e
s
im
u
la
t
io
n
R
e
s
e
a
r
c
h
by
B
ut
t
a
nd
S
ha
f
iq
ue
[
27]
,
ups
tr
e
a
m
a
nd
dow
n
s
tr
e
a
m
de
te
c
to
r
s
ta
ti
on
s
a
r
e
s
e
pa
r
a
te
d
by
tr
a
f
f
ic
li
ght
s
or
c
r
os
s
r
oa
ds
,
s
im
ul
a
ti
ng
a
tr
a
f
f
ic
e
nvi
r
onm
e
nt
.
T
hi
s
s
c
e
na
r
io
f
a
it
hf
ul
ly
de
pi
c
ts
a
n
ur
ba
n
tr
a
f
f
ic
s
ys
te
m
w
it
h a
hi
gh f
r
e
que
nc
y of
j
unc
ti
ons
. T
hi
s
s
e
tu
p i
s
pa
r
ti
c
u
la
r
ly
r
e
le
va
nt
f
or
a
s
ys
te
m
m
a
na
gi
ng i
nc
id
e
nt
s
w
it
h
two
s
ta
ti
ons
.
I
n
th
is
a
r
r
a
nge
m
e
nt
,
de
te
c
to
r
s
a
r
e
pl
a
c
e
d
im
m
e
di
a
te
ly
be
f
or
e
tr
a
f
f
ic
in
te
r
s
e
c
ti
on
s
.
A
two
-
s
ta
ti
on
s
e
tu
p
in
c
lu
de
s
a
n
ups
tr
e
a
m
d
e
te
c
to
r
s
ta
ti
on
lo
c
a
te
d
be
f
or
e
th
e
in
te
r
s
e
c
ti
on
li
nki
ng
two
s
e
gm
e
nt
s
,
a
nd
a
dow
ns
tr
e
a
m
de
te
c
to
r
s
ta
ti
on
s
it
ua
te
d
be
f
or
e
th
e
ne
xt
c
r
o
s
s
r
oa
d
a
t
th
e
e
nd
of
th
e
s
ubs
e
que
nt
s
e
gm
e
nt
.
T
r
a
f
f
ic
s
ig
na
ls
in
th
is
s
c
e
na
r
io
a
r
e
e
xpe
c
t
e
d
to
s
ig
ni
f
ic
a
nt
ly
im
pa
c
t
th
e
c
a
pt
ur
e
d
tr
a
f
f
ic
in
f
or
m
a
ti
on
f
r
om
th
e
s
e
de
te
c
to
r
s
,
w
hi
c
h i
nc
lu
de
s
c
ha
r
a
c
te
r
is
ti
c
s
l
ik
e
oc
c
upa
n
c
y, s
pe
e
d, a
nd t
r
a
f
f
ic
f
lo
w
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
U
r
ban inc
id
e
nt
de
te
c
ti
on ba
s
e
d on hy
br
id
c
onv
ol
ut
io
nal
ne
ur
al
ne
tw
or
k
s
…
(
M
e
r
y
e
m
A
y
ou)
3157
4.4.
R
e
s
u
lt
s
an
al
ys
is
T
he
in
put
s
c
on
s
is
t
of
tr
a
f
f
i
c
f
lo
w
,
oc
c
up
a
nc
y
r
a
t
e
,
a
nd
v
e
lo
c
it
y,
w
hi
c
h
a
r
e
ta
k
e
n
f
r
om
in
du
c
ti
ve
lo
op
de
te
c
to
r
s
lo
c
a
t
e
d ups
tr
e
a
m
a
n
d
d
ow
n
s
tr
e
a
m
. T
h
e
a
lg
or
it
hm
s
pr
o
duc
e
two
out
put
s
:
on
e
f
or
a
n
in
c
id
e
nt
a
nd
z
e
r
o
f
or
a
no
n
-
in
c
id
e
nt
.
T
hr
e
e
s
e
pa
r
a
te
i
nc
i
de
nt
lo
c
a
ti
on
s
, 60,
4
00, a
nd
600
m
e
t
e
r
s
f
r
om
r
e
f
e
r
e
nc
e
po
in
ts
,
h
a
ve
da
t
a
r
e
c
or
de
d
e
v
e
r
y
30
s
e
c
ond
s
.
T
he
d
a
ta
s
e
t
i
s
m
a
d
e
up
of
5
,
752
s
a
m
pl
e
s
th
a
t
w
e
r
e
g
a
th
e
r
e
d
ov
e
r
a
pe
r
io
d
of
72
hour
s
.
T
hi
r
ty
of
th
e
s
e
in
c
id
e
nt
c
a
s
e
s
a
r
e
c
ol
le
c
te
d
a
t
th
e
pr
e
c
is
e
in
c
i
de
nt
oc
c
ur
r
e
n
c
e
in
s
ta
nc
e
s
a
nd
i
s
ol
a
te
d
f
or
t
e
s
t
in
g.
A
f
ur
th
e
r
1
,
72
1 s
a
m
pl
e
s
a
r
e
s
e
t
a
s
id
e
f
or
t
e
s
ti
ng.
T
a
b
le
1 di
s
pl
a
y
s
t
h
e
obt
a
in
e
d r
e
s
ul
t
s
.
T
o
s
how
th
e
e
f
f
e
c
ti
ve
ne
s
s
of
our
m
ode
l,
w
e
c
om
pa
r
e
d
th
e
out
c
om
e
s
of
C
N
N
-
Bi
-
L
S
T
M
w
it
h
th
os
e
obt
a
in
e
d
f
r
om
th
e
s
uppor
t
ve
c
to
r
m
a
c
hi
ne
(
S
V
M
)
,
r
a
ndom
f
or
e
s
t
(
R
F
)
,
a
nd
B
i
-
L
S
T
M
a
lg
or
it
hm
s
.
T
he
pe
r
f
or
m
a
nc
e
c
om
pa
r
is
on
w
a
s
f
oc
us
e
d
on
tr
a
in
in
g
ti
m
e
,
a
c
c
ur
a
c
y,
a
nd
a
ppr
oxi
m
a
ti
on
a
bi
li
ty
us
in
g
th
e
r
oot
m
e
a
n
s
qua
r
e
e
r
r
or
(
R
M
S
E
)
c
r
it
e
r
io
n.
T
he
R
M
S
E
is
gi
ve
n
by
th
e
f
ol
lo
w
in
g
e
qua
ti
on,
w
he
r
e
s
ta
nds
f
or
th
e
e
s
ti
m
a
te
d
va
lu
e
,
or
th
e
e
s
ti
m
a
te
d
c
onc
e
nt
r
a
ti
on,
a
nd
f
or
th
e
obs
e
r
va
ti
on,
w
hi
c
h
is
ba
s
e
d
in
r
e
la
ti
on
to
a
c
c
or
di
ng t
o t
he
numbe
r
n of
obs
e
r
va
ti
ons
.
O
ur
pr
opos
e
d
m
e
th
od
w
a
s
e
va
lu
a
t
e
d
by
c
om
pa
r
in
g
th
e
C
N
N
-
Bi
-
L
S
T
M
m
ode
l’
s
pe
r
f
or
m
a
nc
e
w
it
h
th
a
t
of
S
V
M
,
R
F
a
nd
B
i
-
L
S
T
M
,
a
s
pr
e
s
e
nt
e
d
in
th
e
r
e
s
ul
ts
s
e
c
ti
on.
T
he
e
va
lu
a
ti
on
m
e
tr
ic
s
in
c
lu
de
d
R
M
S
E
,
tr
a
in
in
g
dur
a
ti
on,
a
c
c
ur
a
c
y
le
ve
l,
a
nd
e
s
ti
m
a
ti
on
s
iz
e
.
R
M
S
E
s
e
r
ve
d
a
s
th
e
pr
im
a
r
y
pe
r
f
or
m
a
nc
e
in
di
c
a
to
r
a
nd
w
a
s
c
om
put
e
d
us
in
g
th
e
f
ol
lo
w
in
g
(
10)
,
w
he
r
e
is
th
e
pr
e
di
c
te
d
va
lu
e
,
is
th
e
obs
e
r
ve
d
va
lu
e
,
a
nd
r
e
pr
e
s
e
nt
s
t
he
t
ot
a
l
num
be
r
of
obs
e
r
va
ti
ons
.
=
√
∑
(
−
)
2
=
1
(
10)
T
a
bl
e
2
s
how
th
a
t
C
N
N
-
Bi
-
L
S
T
M
obt
a
in
s
th
e
lo
w
e
r
R
M
S
E
th
a
n
B
i
-
L
S
T
M
,
w
hi
le
th
e
tr
a
in
in
g
ti
m
e
in
c
r
e
a
s
e
v
e
r
y
s
li
ght
ly
.
I
t
is
s
ig
ni
f
ic
a
nt
ly
be
tt
e
r
th
a
n
R
M
S
E
o
f
S
V
M
a
nd
R
F
,
w
hi
le
tr
a
in
in
g
s
p
e
e
d
is
ve
r
y
f
a
s
te
r
.
F
ig
ur
e
3
s
how
s
th
a
t
our
a
ppr
oa
c
h
C
N
N
-
Bi
-
L
S
T
M
,
de
m
ons
tr
a
te
d
in
c
r
e
a
s
e
d
a
c
c
ur
a
c
y,
in
di
c
a
ti
ng
it
s
s
upe
r
io
r
e
f
f
e
c
ti
ve
ne
s
s
i
n i
nc
id
e
nt
de
t
e
c
ti
on, by c
om
pa
r
in
g i
t
w
it
h B
i
-
L
S
T
M
, S
V
M
, a
nd R
F
a
lg
or
it
hm
s
.
T
a
bl
e
1. R
e
s
ul
ts
obt
a
in
e
d f
or
i
nc
id
e
nt
de
te
c
ti
on us
in
g C
N
N
,
B
i
-
L
S
T
M
, C
N
N
-
BI
-
L
S
T
M
,
R
-
e
xt
r
e
m
e
l
e
a
r
ni
ng
m
a
c
hi
ne
(
R
-
ELM
)
A
l
gor
i
t
hm
I
nc
i
de
nt
dur
a
t
i
on
D
ur
a
t
i
on r
a
t
e
(
DR
)
(
%
)
M
e
a
n t
i
m
e
-
to
-
de
t
e
c
t
(
M
T
T
D
)
(
s
)
F
a
l
s
e
a
l
a
r
m
r
a
t
e
(
F
A
R
)
(%)
S
V
M
L
e
s
s
t
ha
n 4m
i
n
82.81
83.12
2.1
L
e
s
s
t
ha
n 6m
i
n
90.19
38.41
0.40
RF
L
e
s
s
t
ha
n 4m
i
n
83.78
82.60
2.1
L
e
s
s
t
ha
n 6m
i
n
91.87
38.32
0.43
Bi
-
L
S
T
M
L
e
s
s
t
ha
n 4m
i
n
85.82
81.66
2.2
L
e
s
s
t
ha
n 6m
i
n
90.87
39.95
0.42
C
N
N
-
Bi
-
L
S
T
M
L
e
s
s
t
ha
n 4m
i
n
87.65
80.98
2.2
L
e
s
s
t
ha
n 6m
i
n
93.78
37.94
0.46
T
a
bl
e
2. P
e
r
f
or
m
a
nc
e
m
e
a
s
ur
e
m
e
nt
b
a
s
e
d on
R
M
S
E
a
nd t
r
a
in
in
g t
im
e
P
e
r
f
or
m
a
nc
e
C
N
N
-
Bi
-
L
S
T
M
Bi
-
L
S
T
M
S
V
M
RF
R
M
S
E
0.0911
0.0998
0.7066
0.8129
T
r
a
i
ni
ng t
i
m
e
(
s
)
1.1433
2.0123
3.0923
4.5032
A
c
c
ur
a
c
y
0.95
0.92
0.90
0.91
F
ig
ur
e
3
.
A
c
c
ur
a
c
y c
om
pa
r
is
on of
C
N
N
-
Bi
-
L
S
T
M
,
Bi
-
L
S
T
M
,
R
F
,
a
nd S
V
M
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
4
,
A
ugus
t
2025
:
3153
-
3159
3158
5.
C
O
N
C
L
U
S
I
O
N
A
N
D
F
U
T
U
R
E
D
I
R
E
C
T
I
O
N
S
T
hi
s
s
tu
dy pr
opos
e
s
a
hybr
id
m
ode
l
c
om
bi
ni
ng 1D C
N
N
a
nd
Bi
-
L
S
T
M
ne
twor
ks
t
o e
nha
nc
e
A
I
D
on
ur
ba
n
r
oa
ds
.
T
he
1D
C
N
N
s
e
f
f
e
c
ti
ve
ly
e
xt
r
a
c
t
s
pa
ti
a
l
f
e
a
tu
r
e
s
f
r
om
r
a
w
tr
a
f
f
ic
da
ta
,
a
ll
ow
in
g
f
or
a
m
or
e
r
e
f
in
e
d
r
e
pr
e
s
e
nt
a
ti
on
of
lo
c
a
l
s
tr
uc
tu
r
e
s
.
I
n
pa
r
a
ll
e
l,
t
he
Bi
-
L
S
T
M
s
c
a
pt
ur
e
lo
ng
-
te
r
m
te
m
por
a
l
de
pe
nde
nc
ie
s
,
w
hi
c
h
a
r
e
e
s
s
e
nt
ia
l
in
tr
a
f
f
ic
f
lu
c
tu
a
ti
ons
th
a
t
m
a
y
in
di
c
a
te
a
n
in
c
id
e
nt
.
T
hi
s
hybr
id
C
N
N
-
Bi
-
L
S
T
M
a
r
c
hi
te
c
tu
r
e
le
ve
r
a
ge
s
th
e
s
tr
e
ngt
hs
of
bot
h
a
ppr
oa
c
he
s
,
c
om
bi
ni
ng
th
e
f
e
a
tu
r
e
e
xt
r
a
c
ti
on
pow
e
r
o
f
C
N
N
s
w
it
h
th
e
te
m
por
a
l
s
e
que
nc
e
pr
oc
e
s
s
in
g
c
a
pa
bi
li
ti
e
s
of
Bi
-
L
S
T
M
s
.
T
he
pe
r
f
or
m
a
nc
e
of
th
e
pr
opos
e
d
m
ode
l
w
a
s
c
om
pa
r
e
d
to
th
a
t
of
popula
r
a
lg
or
it
hm
s
,
in
c
lu
di
ng
S
V
M
,
R
F
,
a
nd
Bi
-
L
S
T
M
a
lo
ne
.
T
he
r
e
s
ul
ts
de
m
ons
tr
a
te
a
c
le
a
r
a
dva
nt
a
ge
of
th
e
C
N
N
-
Bi
-
L
S
T
M
m
ode
l,
w
it
h
im
pr
ove
d
pe
r
f
or
m
a
nc
e
a
nd
hi
gh
de
te
c
ti
on
a
c
c
ur
a
c
y,
hi
ghl
ig
ht
in
g
it
s
pot
e
nt
ia
l
f
or
r
e
a
l
-
ti
m
e
in
c
id
e
nt
de
te
c
t
io
n
in
c
om
pl
e
x
ur
ba
n
e
nvi
r
onm
e
nt
s
.
A
lt
hough
th
e
pr
opos
e
d
a
ppr
oa
c
h
e
nha
nc
e
s
th
e
de
te
c
ti
on
pe
r
f
or
m
a
nc
e
a
nd
s
how
s
pr
om
is
in
g
r
e
s
ul
ts
,
f
ut
ur
e
w
or
k
c
oul
d
c
ons
id
e
r
s
e
ve
r
a
l
ty
pe
s
of
unc
e
r
ta
in
ty
t
ha
t
c
a
n a
f
f
e
c
t
th
e
s
ys
te
m
.
F
U
N
D
I
N
G
I
N
F
O
R
M
A
T
I
O
N
T
hi
s
r
e
s
e
a
r
c
h
r
e
c
e
iv
e
d
no
s
pe
c
if
ic
gr
a
nt
f
r
om
a
ny
f
undi
ng
a
ge
nc
y
in
th
e
publ
ic
,
c
om
m
e
r
c
ia
l,
or
not
-
f
or
-
pr
o
f
it
s
e
c
to
r
s
.
A
U
T
H
O
R
C
O
N
T
R
I
B
U
T
I
O
N
S
S
T
A
T
E
M
E
N
T
T
hi
s
jo
ur
na
l
us
e
s
th
e
C
ont
r
ib
ut
or
R
ol
e
s
T
a
xonomy
(
C
R
e
d
iT
)
to
r
e
c
ogni
z
e
in
di
vi
dua
l
a
ut
hor
c
ont
r
ib
ut
io
ns
, r
e
duc
e
a
ut
hor
s
hi
p di
s
put
e
s
,
a
nd f
a
c
il
it
a
te
c
ol
la
bo
r
a
ti
on.
N
am
e
o
f
A
u
t
h
or
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
M
e
r
ye
m
A
you
✓
✓
✓
✓
✓
✓
✓
✓
✓
J
a
oua
d B
oum
hi
di
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
onc
e
pt
ua
li
z
a
ti
on
M
:
M
e
th
odol
ogy
So
:
So
f
twa
r
e
Va
:
Va
li
da
ti
on
Fo
:
Fo
r
m
a
l
a
na
ly
s
is
I
:
I
nve
s
ti
ga
ti
on
R
:
R
e
s
our
c
e
s
D
:
D
a
ta
C
ur
a
ti
on
O
:
W
r
it
in
g
-
O
r
ig
in
a
l
D
r
a
f
t
E
:
W
r
it
in
g
-
R
e
vi
e
w
&
E
di
ti
ng
Vi
:
Vi
s
ua
li
z
a
ti
on
Su
:
Su
pe
r
vi
s
io
n
P
:
P
r
oj
e
c
t
a
dm
in
is
tr
a
ti
on
Fu
:
Fu
ndi
ng a
c
qui
s
it
io
n
C
O
N
F
L
I
C
T
O
F
I
N
T
E
R
E
S
T
S
T
A
T
E
M
E
N
T
T
he
a
ut
hor
s
s
t
a
te
no c
onf
li
c
t
of
i
nt
e
r
e
s
t.
D
A
T
A
A
V
A
I
L
A
B
I
L
I
T
Y
T
he
da
ta
u
s
e
d i
n t
he
pa
pe
r
w
il
l
be
a
v
a
il
a
bl
e
upon r
e
que
s
t.
R
E
F
E
R
E
N
C
E
S
[
1]
M
.
A
you,
Y
.
T
r
a
r
di
,
L
.
C
ha
ki
r
,
a
nd
J
.
B
oum
hi
di
,
“
D
a
t
a
-
d
r
i
ve
n
t
r
a
f
f
i
c
i
nc
i
de
nt
de
t
e
c
t
i
on
i
n
ur
ba
n r
oa
ds
ba
s
e
d
on
m
a
c
hi
ne
l
e
a
r
ni
ng
a
l
gor
i
t
hm
s
,”
E
3S W
e
b of
C
onf
e
r
e
nc
e
s
, vol
. 469, 2023, doi
:
10.1051/
e
3
s
c
onf
/
202346900102.
[
2]
Q
.
Q
u,
Y
.
S
he
n,
M
.
Y
a
ng,
R
.
Z
ha
ng,
a
nd
H
.
Z
ha
ng,
“
E
xpr
e
s
s
w
a
y
t
r
a
f
f
i
c
i
nc
i
de
nt
de
t
e
c
t
i
on
us
i
ng
a
de
e
p
l
e
a
r
ni
ng
a
ppr
oa
c
h
ba
s
e
d
on s
pa
t
i
ot
e
m
por
a
l
f
e
a
t
ur
e
s
w
i
t
h m
ul
t
i
l
e
ve
l
f
us
i
on,”
J
ou
r
nal
of
T
r
an
s
por
t
at
i
on E
ngi
ne
e
r
i
ng, P
ar
t
A
:
Sy
s
t
e
m
s
,
vol
. 150, no. 6, 2024
,
doi
:
10.1061/
j
t
e
pbs
.t
e
e
ng
-
8001.
[
3]
S
.
C
he
n,
W
.
W
a
ng,
a
nd
H
.
V.
Z
uyl
e
n,
“
C
on
s
t
r
uc
t
s
uppor
t
ve
c
t
or
m
a
c
hi
ne
e
n
s
e
m
bl
e
t
o
d
e
t
e
c
t
t
r
a
f
f
i
c
i
nc
i
de
nt
,”
E
x
pe
r
t
S
y
s
t
e
m
s
w
i
t
h
A
ppl
i
c
at
i
ons
, vol
. 36, no. 8, pp. 10976
–
10986, 2009, doi
:
10.1016/
j
.e
s
w
a
.2009.02.039.
[
4]
J
. L
a
c
he
r
r
e
, J
. L
. C
a
s
t
i
l
l
o
-
S
e
que
r
a
, a
nd D
.
M
a
ur
i
c
i
o, “
F
a
c
t
or
s
, pr
e
di
c
t
i
on, a
nd e
xpl
a
i
na
bi
l
i
t
y of
ve
hi
c
l
e
a
c
c
i
de
nt
r
i
s
k due
t
o dr
i
vi
ng
be
ha
vi
or
t
hr
ough
m
a
c
hi
ne
l
e
a
r
ni
ng:
a
s
ys
t
e
m
a
t
i
c
l
i
t
e
r
a
t
ur
e
r
e
vi
e
w
,
2013
–
2023,”
C
om
put
at
i
on
,
vol
.
12,
no.
7,
p.
131,
2024,
doi
:
10.3390/
c
om
put
a
t
i
on12070131.
[
5]
S
. S
. P
e
ddi
r
a
j
u, K
. H
a
r
a
p
a
na
ha
l
l
i
, E
. A
nde
r
t
,
a
nd A
.
S
hr
i
va
s
t
a
va
, “
I
nc
i
de
nt
N
e
t
:
t
r
a
f
f
i
c
i
nc
i
de
nt
de
t
e
c
t
i
on, l
oc
a
l
i
z
a
t
i
on a
nd
s
e
v
e
r
i
t
y
e
s
t
i
m
a
t
i
on
w
i
t
h
s
pa
r
s
e
s
e
ns
i
ng
,”
I
E
E
E
C
onf
e
r
e
nc
e
on
I
nt
e
l
l
i
ge
nt
T
r
ans
por
t
at
i
on
Sy
s
t
e
m
s
,
P
r
oc
e
e
di
ngs
,
I
T
SC
,
pp. 1470
–
1475, A
ug. 2024, doi
:
10.1109/
I
T
S
C
58415.2024.10920070.
[
6]
P
.
C
ha
kr
a
bor
t
y,
C
.
H
e
gde
,
a
nd
A
.
S
ha
r
m
a
,
“
D
a
t
a
-
dr
i
ve
n
pa
r
a
l
l
e
l
i
z
a
bl
e
t
r
a
f
f
i
c
i
nc
i
de
nt
de
t
e
c
t
i
on
us
i
ng
s
pa
t
i
o
-
t
e
m
por
a
l
l
y
de
noi
s
e
d
r
obus
t
t
hr
e
s
hol
ds
,”
T
r
ans
por
t
at
i
on
R
e
s
e
ar
c
h
P
ar
t
C
:
E
m
e
r
gi
ng
T
e
c
hnol
ogi
e
s
,
vol
.
105,
pp.
81
–
99,
2019,
doi
:
10.1016/
j
.t
r
c
.2019.05.034.
[
7]
N
.
F
a
qi
r
,
C
.
L
oqm
a
n,
a
nd
J
.
B
oum
hi
di
,
“
C
om
bi
ne
d
e
xt
r
e
m
e
l
e
a
r
ni
ng
m
a
c
hi
ne
a
nd
m
a
x
pr
e
s
s
ur
e
a
l
gor
i
t
hm
s
f
or
t
r
a
f
f
i
c
s
i
gna
l
c
ont
r
ol
,”
I
nt
e
l
l
i
ge
nt
Sy
s
t
e
m
s
w
i
t
h A
ppl
i
c
at
i
ons
, vol
. 19, 2023, doi
:
10.1016/
j
.i
s
w
a
.2023.200255.
[
8]
R
.
H
a
s
a
bi
,
R
.
K
ur
ni
a
w
a
n,
S
ugi
a
r
t
o,
R
.
N
.
T
.
W
a
hyuni
,
a
nd
E
.
N
u
r
m
a
w
a
t
i
,
“
U
r
ba
n
t
r
a
f
f
i
c
c
onge
s
t
i
on
a
nd
i
t
s
a
s
s
oc
i
a
t
i
on
w
i
t
h
ga
s
s
t
a
t
i
on
de
ns
i
t
y:
i
ns
i
ght
s
f
r
om
G
oogl
e
M
a
ps
d
a
t
a
,”
I
ndone
s
i
an
J
ou
r
nal
of
E
l
e
c
t
r
i
c
al
E
ngi
ne
e
r
i
ng
and
C
om
put
e
r
Sc
i
e
n
c
e
,
vol
.
35,
no. 3, pp. 1618
–
1626, 2024, doi
:
10.11591/
i
j
e
e
c
s
.v35.i
3.pp1618
-
1626.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
U
r
ban inc
id
e
nt
de
te
c
ti
on ba
s
e
d on hy
br
id
c
onv
ol
ut
io
nal
ne
ur
al
ne
tw
or
k
s
…
(
M
e
r
y
e
m
A
y
ou)
3159
[
9]
H
.
O
.
A
w
or
i
nde
,
A
.
E
.
A
de
ni
yi
,
S
.
A
de
ba
yo,
F
.
A
de
ni
j
i
,
a
nd
O
.
J
.
A
r
oba
,
“
D
e
ve
l
opm
e
nt
of
a
pr
i
o
r
i
t
i
z
e
d
t
r
a
f
f
i
c
l
i
ght
c
ont
r
o
l
s
ys
t
e
m
f
or
e
m
e
r
ge
nc
y
v
e
hi
c
l
e
s
,”
I
A
E
S
I
nt
e
r
nat
i
onal
J
our
nal
of
A
r
t
i
f
i
c
i
al
I
nt
e
l
l
i
ge
nc
e
,
vol
.
13,
no.
4,
pp.
4019
–
4028,
2024
,
doi
:
10.11591/
i
j
a
i
.v13.i
4.pp4019
-
4028.
[
10]
M
.
A
hm
a
dz
a
de
h,
S
.
M
.
Z
a
hr
a
i
,
a
nd
M
.
B
i
t
a
r
a
f
,
“
A
n
i
nt
e
gr
a
t
e
d
de
e
p
ne
ur
a
l
n
e
t
w
or
k
m
ode
l
c
om
bi
ni
ng
1D
C
N
N
a
nd
L
S
T
M
f
or
s
t
r
uc
t
ur
a
l
he
a
l
t
h
m
oni
t
or
i
ng
ut
i
l
i
z
i
ng
m
ul
t
i
s
e
ns
or
t
i
m
e
-
s
e
r
i
e
s
da
t
a
,”
St
r
uc
t
ur
a
l
H
e
al
t
h
M
oni
t
or
i
ng
,
vol
.
24,
no.
1,
pp.
447
–
465,
2025, doi
:
10.1177/
147592172412
39041.
[
11]
J
.
L
i
,
X
.
X
u,
M
.
Z
ha
o,
a
nd
R
.
S
hi
,
“
T
r
a
i
n
a
r
r
i
va
l
de
l
a
y
pr
e
di
c
t
i
on
ba
s
e
d
on
a
C
N
N
-
L
S
T
M
a
ppr
oa
c
h
,”
21
s
t
C
O
T
A
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
of
T
r
ans
po
r
t
at
i
on P
r
of
e
s
s
i
onal
s
,
2021,
pp.
555
-
563
,
doi
:
10.1061/
9
780784483565.054
.
[
12]
J
.
F
r
i
e
dm
a
n,
T
.
H
a
s
t
i
e
,
a
nd
R
.
T
i
bs
hi
r
a
ni
,
“
R
e
gul
a
r
i
z
a
t
i
on
pa
t
hs
f
or
ge
ne
r
a
l
i
z
e
d
l
i
ne
a
r
m
ode
l
s
vi
a
c
oor
di
na
t
e
de
s
c
e
nt
,”
J
ou
r
nal
of
St
at
i
s
t
i
c
al
Sof
t
w
ar
e
, vol
. 33, no. 1, pp. 1
–
22, 2010, doi
:
10.18637/
j
s
s
.v033.i
01.
[
13]
M
. J
. R
.
H
e
a
l
y,
C
.
R
. R
a
o, a
nd
S
. K
.
M
i
t
r
a
, “
G
e
n
e
r
a
l
i
z
e
d
i
nve
r
s
e
of
m
a
t
r
i
c
e
s
a
nd i
t
s
a
ppl
i
c
a
t
i
o
ns
,”
J
our
nal
of
t
he
R
oy
al
St
at
i
s
t
i
c
a
l
Soc
i
e
t
y
. Se
r
i
e
s
A
(
G
e
ne
r
al
)
, vol
. 135, no. 3, 1972, doi
:
10.2307/
2344631.
[
14]
S
. H
i
r
e
c
he
a
nd A
. D
e
nna
i
,
“
M
a
c
hi
ne
l
e
a
r
ni
ng t
e
c
hni
que
s
f
or
r
oa
d t
r
a
f
f
i
c
a
ut
om
a
t
i
c
i
nc
i
de
nt
de
t
e
c
t
i
on s
y
s
t
e
m
s
:
a
r
e
vi
e
w
,”
i
n
Sm
ar
t
E
ne
r
gy
E
m
pow
e
r
m
e
nt
i
n Sm
a
r
t
and R
e
s
i
l
i
e
nt
C
i
t
i
e
s
, 2020, pp. 60
–
69
, doi
:
10.1
007/
978
-
3
-
030
-
37207
-
1_7.
[
15]
O
.
E
l
S
a
hl
y
a
nd
A
.
A
bde
l
f
a
t
a
h,
“
A
s
ys
t
e
m
a
t
i
c
r
e
vi
e
w
of
t
r
a
f
f
i
c
i
nc
i
de
nt
de
t
e
c
t
i
on
a
l
gor
i
t
hm
s
,”
Sus
t
ai
nabi
l
i
t
y
,
vol
.
14,
no.
22,
2022, doi
:
10.3390/
s
u142214859.
[
16]
Y
.
S
un,
T
.
M
a
l
l
i
c
k,
P
.
B
a
l
a
pr
a
ka
s
h,
a
nd
J
.
M
a
c
f
a
r
l
a
ne
,
“
A
da
t
a
-
c
e
nt
r
i
c
w
e
a
k
s
upe
r
vi
s
e
d
l
e
a
r
ni
ng
f
or
hi
ghw
a
y
t
r
a
f
f
i
c
i
nc
i
de
nt
de
t
e
c
t
i
on,”
A
c
c
i
de
nt
A
nal
y
s
i
s
and P
r
e
v
e
nt
i
on
, vol
. 176, 2022, doi
:
10.1016/
j
.a
a
p.2022.106779.
[
17]
J
.
S
c
hm
i
dhube
r
,
“
D
e
e
p
l
e
a
r
ni
ng
i
n
ne
ur
a
l
ne
t
w
or
ks
:
a
n
ove
r
vi
e
w
,”
N
e
ur
al
N
e
t
w
or
k
s
,
vol
.
61,
pp.
85
–
117,
2015
,
doi
:
10.1016/
j
.ne
une
t
.2014.09.003.
[
18]
G
. V
.
H
oudt
, C
.
M
os
qu
e
r
a
, a
nd G
.
N
á
pol
e
s
,
“
A
r
e
vi
e
w
on
t
he
l
ong s
hor
t
-
t
e
r
m
m
e
m
or
y m
ode
l
,”
A
r
t
i
f
i
c
i
al
I
nt
e
l
l
i
ge
nc
e
R
e
v
i
e
w
, vol
.
53, no. 8, pp. 5929
–
5955, 2020, doi
:
10.1007/
s
10462
-
020
-
09838
-
1.
[
19]
T
.
G
a
nokr
a
t
a
na
a
,
M
.
K
e
t
c
ha
m
,
a
nd
P
.
P
r
a
m
ke
a
w
,
“
A
dva
nc
e
m
e
nt
s
i
n
c
a
t
a
r
a
c
t
de
t
e
c
t
i
on:
t
he
s
ys
t
e
m
a
t
i
c
de
ve
l
opm
e
nt
of
L
e
N
e
t
-
c
onvol
ut
i
ona
l
ne
ur
a
l
ne
t
w
or
k m
ode
l
s
,”
J
our
nal
of
I
m
agi
ng
, vol
. 9, no. 10, 2023, doi
:
10.3390/
j
i
m
a
gi
ng9100197.
[
20]
M
.
M
.
R
.
K
.
M
a
m
un
a
nd
A
.
A
l
oua
ni
,
“
F
A
-
1D
-
C
N
N
i
m
pl
e
m
e
nt
a
t
i
on
t
o
i
m
pr
ove
di
a
gnos
i
s
of
he
a
r
t
di
s
e
a
s
e
r
i
s
k
l
e
ve
l
,
”
i
n
P
r
oc
e
e
di
ngs
of
t
he
6t
h
W
or
l
d
C
ong
r
e
s
s
on
E
l
e
c
t
r
i
c
al
E
ngi
ne
e
r
i
ng
and
C
om
put
e
r
Sy
s
t
e
m
s
and
Sc
i
e
nc
e
,
2020,
pp.
122.1
-
122.9
,
doi
:
10.11159/
i
c
be
s
20.122.
[
21]
X.
-
T
.
T
r
a
n,
T
. T
.
-
T
.
D
o,
a
nd
C
.
-
T
.
L
i
n
,
“
E
a
r
l
y
de
t
e
c
t
i
on
of
hu
m
a
n
de
c
i
s
i
on
-
m
a
ki
ng
i
n
c
onc
e
a
l
e
d
obj
e
c
t
vi
s
ua
l
s
e
a
r
c
hi
ng
t
a
s
k
s
:
a
n
EEG
-
B
i
L
S
T
M
s
t
udy
,”
A
nnual
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
of
t
he
I
E
E
E
E
ngi
ne
e
r
i
ng
i
n
M
e
di
c
i
ne
and
B
i
ol
ogy
Soc
i
e
t
y
(
E
M
B
C
)
,
2023
, pp.
1
–
4
,
doi
:
10.1109/
E
M
B
C
40787.2023.10340547
.
[
22]
M
.
W
a
qa
s
a
nd
U
.
W
.
H
um
phr
i
e
s
,
“
A
c
r
i
t
i
c
a
l
r
e
vi
e
w
o
f
R
N
N
a
nd
L
S
T
M
va
r
i
a
nt
s
i
n
hydr
ol
ogi
c
a
l
t
i
m
e
s
e
r
i
e
s
pr
e
di
c
t
i
ons
,”
M
e
t
hods
X
, vol
. 13, 2024, doi
:
10.1016/
j
.m
e
x.2024.102946.
[
23]
K
.
W
i
l
br
a
nd
e
t
al
.
,
“
P
r
e
di
c
t
i
ng
s
t
r
e
a
m
f
l
ow
w
i
t
h
L
S
T
M
ne
t
w
or
ks
us
i
ng
gl
oba
l
da
t
a
s
e
t
s
,”
F
r
ont
i
e
r
s
i
n
W
at
e
r
,
vol
.
5,
2023,
doi
:
10.3389/
f
r
w
a
.2023.1166124.
[
24]
O
.
Y
i
l
di
r
i
m
,
M
.
T
a
l
o,
E
.
J
.
C
i
a
c
c
i
o,
R
.
S
.
T
a
n,
a
nd
U
.
R
.
A
c
h
a
r
ya
,
“
A
c
c
ur
a
t
e
de
e
p
ne
ur
a
l
ne
t
w
or
k
m
ode
l
t
o
de
t
e
c
t
c
a
r
di
a
c
a
r
r
hyt
hm
i
a
on
m
or
e
t
ha
n
10,000
i
ndi
vi
dua
l
s
ubj
e
c
t
E
C
G
r
e
c
or
ds
,”
C
om
put
e
r
M
e
t
hods
and
P
r
ogr
am
s
i
n
B
i
om
e
di
c
i
ne
,
vol
.
197,
2020, doi
:
10.1016/
j
.c
m
pb.2020.105740.
[
25]
A
.
K
e
l
e
r
,
W
.
S
un,
a
nd
J
.
-
D
.
S
c
hm
öc
ke
r
,
“
G
e
n
e
r
a
t
i
ng
a
nd
c
a
l
i
br
a
t
i
ng
a
m
i
c
r
o
s
c
opi
c
t
r
a
f
f
i
c
f
l
ow
s
i
m
ul
a
t
i
on
ne
t
w
or
k
of
K
yot
o,”
SU
M
O
C
onf
e
r
e
nc
e
P
r
o
c
e
e
di
ngs
, vol
. 4, pp. 189
–
195, 2023, doi
:
10.52825/
s
c
p.
v4i
.226.
[
26]
A
.
W
e
ge
ne
r
,
M
.
P
i
ór
kow
s
ki
,
M
.
R
a
ya
,
H
.
H
e
l
l
br
üc
k,
S
.
F
i
s
c
he
r
,
a
nd
J
.
P
.
H
u
ba
ux,
“
T
r
a
C
I
:
a
n
i
nt
e
r
f
a
c
e
f
or
c
oupl
i
ng
r
oa
d
t
r
a
f
f
i
c
a
nd
ne
t
w
or
k
s
i
m
ul
a
t
or
s
,”
P
r
oc
e
e
di
ng
s
of
t
he
11t
h
C
om
m
uni
c
at
i
ons
an
d
N
e
t
w
o
r
k
i
ng
Si
m
ul
at
i
on
Sy
m
pos
i
um
,
C
N
S’
08
,
pp. 155
–
163, 2008, doi
:
10.1145/
1400713.1400740.
[
27]
M
.
S
.
B
ut
t
a
nd
M
.
A
.
S
ha
f
i
que
,
“
A
l
i
t
e
r
a
t
ur
e
r
e
vi
e
w
:
A
I
m
ode
l
s
f
or
r
oa
d
s
a
f
e
t
y
f
or
pr
e
di
c
t
i
on
of
c
r
a
s
h
f
r
e
que
nc
y
a
nd
s
e
ve
r
i
t
y
,”
D
i
s
c
ov
e
r
C
i
v
i
l
E
ngi
ne
e
r
i
ng
, vol
. 2, no. 1, 2025, doi
:
10.1007/
s
44290
-
025
-
00255
-
3.
B
I
O
G
R
A
P
H
I
E
S
O
F
A
U
T
H
O
R
S
Meryem
Ayou
is
obtained
a
master’s
degree
in
decisional
c
omputing
an
d
intelligent
vision
from
Sidi
Mohamed
Ben
Abdellah
University,
Mor
occo,
in
2019.
Now
she
is
a
Ph.D.
s
tudent
at
Sidi
Mohammed
Ben
Abdellah
University,
Morocco,
in
2019.
Her
research
interests
include
machine
learning
and
intellig
ent
transportati
on
systems.
She
can
be
contacted
at email
:
meryeme.a
you@
usmba.ac.ma
.
Jaouad Boumhidi
is a Profe
ssor of Compute
r Science
at the Fac
ult
y of Scienc
es
,
Fez,
Morocco.
He
received
a
Ph
.
D
.
in
Computer
Science
from
the
University
of
Sidi
Mohamed
ben
Abdellah
in
2005.
His
researc
h
interests
include
deep
learning
and
intelligent
transporta
tion systems.
He can be contacted at email:
jaouad.boumh
idi@
usmba.ac.
ma
.
Evaluation Warning : The document was created with Spire.PDF for Python.