I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
39
,
No
.
1
,
J
u
l
y
2
0
2
5
,
p
p
.
656
~
6
6
8
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
39
.i
1
.
pp
656
-
6
6
8
656
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Bo
o
sting
real
-
tim
e vehicle d
etec
tio
n in ur
ba
n t
ra
f
fic
using
a
no
v
el multi
-
a
ug
m
entatio
n
I
m
a
m
Ahma
d Ash
a
ri
1,
2
,
Wa
hy
ul Am
ien Sy
a
f
ei
3
,
Adi Wi
bo
wo
4
1
D
o
c
t
o
r
a
l
P
r
o
g
r
a
m
o
f
I
n
f
o
r
ma
t
i
o
n
S
y
s
t
e
ms,
U
n
i
v
e
r
si
t
a
s D
i
p
o
n
e
g
o
r
o
,
S
e
mar
a
n
g
,
I
n
d
o
n
e
s
i
a
2
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
a
s
H
a
r
a
p
a
n
B
a
n
g
sa,
P
u
r
w
o
k
e
r
t
o
,
I
n
d
o
n
e
si
a
3
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
s
i
t
a
s
D
i
p
o
n
e
g
o
r
o
,
S
e
mara
n
g
,
I
n
d
o
n
e
s
i
a
4
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
a
n
d
M
a
t
h
e
m
a
t
i
c
s,
U
n
i
v
e
r
s
i
t
a
s Di
p
o
n
e
g
o
r
o
,
S
e
mara
n
g
,
I
n
d
o
n
e
s
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
3
,
2
0
2
4
R
ev
is
ed
Mar
12
,
2
0
2
5
Acc
ep
ted
Mar
26
,
2
0
2
5
R
e
a
l
-
t
im
e
v
e
h
i
c
le
o
b
j
e
c
t
d
e
t
e
c
t
i
o
n
i
n
u
r
b
a
n
t
r
a
f
f
i
c
i
s
c
r
u
c
i
a
l
f
o
r
m
o
d
e
r
n
t
r
a
f
f
i
c
m
a
n
a
g
e
m
e
n
t
s
y
s
t
e
m
s
.
T
h
i
s
s
t
u
d
y
f
o
c
u
s
e
s
o
n
i
m
p
r
o
v
i
n
g
t
h
e
a
c
c
u
r
a
c
y
o
f
v
e
h
i
c
l
e
i
d
e
n
t
i
f
i
c
a
t
i
o
n
a
n
d
c
la
s
s
i
f
ic
a
t
i
o
n
i
n
h
e
a
v
y
t
r
a
f
f
ic
d
u
r
i
n
g
p
e
a
k
h
o
u
r
s
,
w
i
t
h
p
a
r
t
i
c
u
l
a
r
e
m
p
h
a
s
i
s
o
n
c
h
a
l
l
e
n
g
e
s
s
u
c
h
a
s
s
m
a
l
l
o
b
j
e
c
t
s
iz
e
s
a
n
d
i
n
t
e
r
f
e
re
n
c
e
f
r
o
m
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
s
.
T
h
e
u
s
e
o
f
m
u
l
t
i
-
l
a
b
e
l
im
a
g
e
s
e
n
a
b
l
e
s
t
h
e
s
im
u
l
t
a
n
e
o
u
s
d
e
t
e
c
t
i
o
n
o
f
v
a
r
i
o
u
s
v
e
h
i
c
l
e
t
y
p
e
s
w
i
t
h
i
n
a
s
i
n
g
l
e
f
r
a
m
e
,
p
r
o
v
id
i
n
g
m
o
r
e
d
e
t
a
i
l
e
d
i
n
f
o
r
m
a
t
i
o
n
a
b
o
u
t
t
r
a
f
f
i
c
c
o
n
d
i
t
i
o
n
s
.
Yo
u
o
n
l
y
l
o
o
k
o
n
c
e
(
Y
O
L
O
)
w
a
s
c
h
o
s
e
n
f
o
r
i
t
s
c
a
p
a
b
i
l
i
t
y
t
o
p
e
rf
o
r
m
r
e
a
l
-
t
i
m
e
o
b
j
e
c
t
d
e
te
c
t
i
o
n
w
i
t
h
h
i
g
h
a
c
c
u
ra
c
y
.
M
u
l
t
i
-
a
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s
w
e
r
e
a
p
p
l
i
e
d
t
o
e
n
r
i
c
h
t
h
e
t
r
a
i
n
i
n
g
d
a
t
a
,
m
a
k
i
n
g
t
h
e
m
o
d
e
l
m
o
r
e
r
o
b
u
s
t
t
o
v
a
r
y
i
n
g
l
i
g
h
t
i
n
g
c
o
n
d
i
t
i
o
n
s
,
v
i
e
w
p
o
i
n
t
s
,
o
b
j
e
c
t
o
c
c
l
u
s
i
o
n
s
,
a
n
d
i
s
s
u
e
s
r
e
l
a
t
e
d
t
o
s
m
a
l
l
o
b
j
e
c
t
s
.
YO
LO
v
8
n
a
n
d
Y
O
L
O
v
9
t
w
e
re
se
l
e
c
t
e
d
f
o
r
t
h
e
i
r
s
p
e
e
d
a
n
d
e
f
f
ic
i
e
n
c
y
.
M
o
d
e
l
s
w
i
t
h
o
u
t
a
u
g
m
e
n
t
a
t
i
o
n
,
1
0
s
i
n
g
l
e
-
a
u
g
m
e
n
ta
t
i
o
n
t
e
c
h
n
i
q
u
e
s
,
a
n
d
5
m
u
l
t
i
-
a
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s
w
e
r
e
t
e
s
t
e
d
.
T
h
e
r
e
s
u
l
t
s
s
h
o
w
t
h
a
t
Y
O
L
O
v
8
n
w
i
t
h
m
u
l
t
i
-
a
u
g
m
e
n
t
a
t
i
o
n
(
sc
a
l
i
n
g
,
z
o
o
m
i
n
,
b
r
i
g
h
t
n
e
s
s
a
d
j
u
s
t
m
e
n
t
,
c
o
l
o
r
j
i
t
te
r
,
a
n
d
n
o
i
s
e
i
n
j
e
c
t
i
o
n
)
a
c
h
i
e
v
e
d
t
h
e
h
i
g
h
e
s
t
m
A
P
5
0
-
9
5
s
c
o
r
e
o
f
0
.
5
3
6
,
s
u
r
p
a
s
s
i
n
g
Y
O
L
O
v
8
n
w
i
t
h
s
i
n
g
l
e
-
a
u
g
m
e
n
t
a
t
io
n
B
l
u
r
,
w
h
i
c
h
h
a
d
a
n
m
A
P
5
0
-
9
5
o
f
0
.
4
6
5
,
a
s
w
e
l
l
a
s
Y
O
LO
v
8
n
w
i
t
h
o
u
t
a
u
g
m
e
n
t
a
t
i
o
n
,
w
h
i
c
h
s
c
o
r
e
d
0
.
3
9
0
.
M
u
l
t
i
-
a
u
g
m
e
n
t
a
t
i
o
n
p
r
o
v
e
d
t
o
s
i
g
n
i
f
i
c
a
n
t
l
y
e
n
h
a
n
c
e
Y
O
L
O
’
s
p
e
r
f
o
r
m
a
n
c
e
.
K
ey
w
o
r
d
s
:
Mu
lti
-
au
g
m
en
tatio
n
Mu
lti
-
lab
el
Ur
b
an
tr
af
f
ic
Veh
icle
o
b
ject
d
etec
tio
n
YOL
O
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
I
m
am
Ah
m
ad
Ash
ar
i
Do
cto
r
al
Pro
g
r
a
m
o
f
I
n
f
o
r
m
at
io
n
Sy
s
tem
s
,
Un
iv
er
s
itas
Dip
o
n
eg
o
r
o
Sem
ar
an
g
5
0
2
4
1
,
I
n
d
o
n
esia
E
m
ail: im
am
ah
m
ad
ash
ar
i@
s
tu
d
en
ts
.
u
n
d
ip
.
ac
.
id
; im
am
a
h
m
a
d
ash
ar
i@
u
h
b
.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
A
s
a
d
e
v
e
l
o
p
i
n
g
c
o
u
n
t
r
y
,
I
n
d
o
n
e
s
i
a
f
a
c
es
s
i
g
n
i
f
i
c
a
n
t
c
h
al
l
en
g
e
s
r
e
l
a
t
e
d
t
o
h
i
g
h
p
o
p
u
l
a
t
i
o
n
g
r
o
w
t
h
,
i
n
a
d
e
q
u
a
t
e
i
n
f
r
as
t
r
u
c
t
u
r
e
,
a
n
d
i
n
c
r
e
a
s
i
n
g
t
r
a
f
f
i
c
c
o
n
g
e
s
ti
o
n
[
1
]
.
A
c
c
o
r
d
i
n
g
t
o
t
h
e
Di
r
e
c
t
o
r
at
e
G
e
n
e
r
al
o
f
L
a
n
d
T
r
a
n
s
p
o
r
t
a
t
i
o
n
o
f
t
h
e
M
i
n
is
t
r
y
o
f
T
r
a
n
s
p
o
r
t
a
t
i
o
n
o
f
t
h
e
R
e
p
u
b
l
i
c
o
f
I
n
d
o
n
e
s
i
a
,
t
r
a
f
f
i
c
ac
c
id
e
n
t
s
i
n
2
0
1
9
w
e
r
e
p
r
i
m
a
r
i
l
y
c
a
u
s
e
d
b
y
m
o
t
o
r
c
y
c
l
e
s
,
w
h
i
c
h
a
c
c
o
u
n
t
e
d
f
o
r
m
o
r
e
t
h
a
n
7
0
%
o
f
t
o
t
a
l
t
r
a
f
f
i
c
a
c
c
i
d
e
n
t
s
[
2
]
.
R
e
s
e
a
r
c
h
c
o
n
d
u
c
t
e
d
i
n
I
n
d
o
n
e
s
i
a
,
s
p
e
cif
i
c
a
l
l
y
o
n
t
h
e
T
r
a
n
s
y
o
g
i
C
i
b
u
b
u
r
r
o
a
d
,
s
h
o
w
s
t
h
a
t
v
e
h
i
c
le
v
o
l
u
m
e
i
n
c
r
e
a
s
es
s
i
g
n
i
f
i
c
a
n
t
l
y
d
u
r
i
n
g
p
e
a
k
h
o
u
r
s
[
3
]
.
E
s
t
i
m
a
ti
n
g
t
r
a
f
f
i
c
d
e
n
s
i
t
y
i
s
c
r
u
c
i
al
f
o
r
i
m
p
r
o
v
i
n
g
t
r
a
n
s
p
o
r
t
a
t
i
o
n
s
y
s
t
e
m
s
as
i
t
h
e
l
p
s
d
e
s
i
g
n
m
o
r
e
e
f
f
i
c
i
e
n
t
t
r
a
f
f
i
c
m
a
n
a
g
e
m
e
n
t
s
t
r
a
t
e
g
i
es
.
V
e
h
ic
l
e
r
e
c
o
g
n
it
i
o
n
a
n
d
c
o
u
n
t
i
n
g
a
r
e
t
h
e
tw
o
p
r
i
m
a
r
y
s
t
e
p
s
i
n
es
t
i
m
at
i
n
g
t
r
af
f
i
c
d
e
n
s
i
t
y
[
4
]
.
T
h
e
m
a
i
n
c
h
a
ll
e
n
g
e
i
n
t
h
is
t
e
c
h
n
o
l
o
g
y
is
a
c
h
ie
v
i
n
g
f
a
s
t
,
a
cc
u
r
a
te
d
e
t
e
c
t
i
o
n
i
n
c
o
m
p
l
e
x
e
n
v
i
r
o
n
m
e
n
t
s
.
S
t
at
e
-
of
-
th
e
-
a
r
t
(
S
o
T
A)
m
o
d
e
l
s
,
t
r
a
i
n
e
d
o
n
d
i
v
e
r
s
e
d
a
t
a
s
et
s
s
u
c
h
as
MS
C
OC
O
,
a
r
e
r
o
b
u
s
t
b
u
t
o
f
t
e
n
s
tr
u
g
g
l
e
w
i
t
h
m
o
n
o
t
o
n
o
u
s
b
ac
k
g
r
o
u
n
d
s
i
n
s
u
r
v
e
i
ll
a
n
c
e
o
r
r
o
ad
v
i
d
e
o
s
,
le
a
d
i
n
g
to
r
e
d
u
c
e
d
p
e
r
f
o
r
m
a
n
c
e
[
5
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
B
o
o
s
tin
g
r
ea
l
-
time
ve
h
icle
d
et
ec
tio
n
in
u
r
b
a
n
tr
a
ffic u
s
in
g
a
n
o
ve
l m
u
lti
-
a
u
g
men
ta
tio
n
(
I
m
a
m
A
h
ma
d
A
s
h
a
r
i
)
657
T
h
e
i
n
a
b
i
l
it
y
t
o
o
b
ta
i
n
i
m
a
g
es
t
h
at
m
e
e
t
d
e
s
i
r
e
d
c
r
it
e
r
i
a
i
s
p
r
o
b
l
e
m
a
t
i
c
,
r
e
s
u
lt
i
n
g
i
n
d
ata
s
e
t
b
i
a
s
,
o
v
e
r
f
i
t
t
i
n
g
,
a
n
d
i
n
a
c
c
u
r
a
t
e
o
u
t
c
o
m
e
s
[
6
]
.
W
h
i
l
e
d
e
e
p
l
e
ar
n
i
n
g
a
l
g
o
r
i
t
h
m
s
h
a
v
e
d
e
m
o
n
s
t
r
a
t
e
d
o
u
t
s
t
a
n
d
i
n
g
p
e
r
f
o
r
m
a
n
c
e
i
n
v
a
r
i
o
u
s
c
o
m
p
u
t
e
r
v
i
s
i
o
n
t
as
k
s
,
li
m
i
t
e
d
l
a
b
e
l
e
d
d
a
t
a
c
a
n
l
e
a
d
t
o
o
v
e
r
f
i
tt
i
n
g
,
h
i
n
d
e
r
i
n
g
t
h
e
n
e
t
w
o
r
k
’
s
g
e
n
e
r
a
li
z
a
ti
o
n
t
o
u
n
s
e
e
n
d
a
t
a
[
7
]
,
[
8
]
.
Da
t
a
a
u
g
m
en
t
a
t
i
o
n
is
o
n
e
o
f
t
h
e
m
o
s
t
wi
d
el
y
u
s
e
d
t
o
o
ls
i
n
d
e
e
p
l
e
a
r
n
i
n
g
,
u
n
d
e
r
p
i
n
n
i
n
g
m
a
n
y
r
e
c
e
n
t
a
d
v
a
n
c
e
s
i
n
f
ie
l
d
s
s
u
c
h
as
cl
a
s
s
i
f
ic
a
t
i
o
n
,
g
e
n
e
r
a
t
i
v
e
m
o
d
e
ls
,
a
n
d
r
e
p
r
e
s
e
n
t
a
ti
o
n
l
e
a
r
n
i
n
g
[
9
]
.
A
u
g
m
e
n
t
a
t
i
o
n
m
o
d
u
l
e
s
s
i
g
n
i
f
i
ca
n
t
l
y
r
e
d
u
c
e
t
h
e
n
e
e
d
f
o
r
m
a
n
u
a
l
l
a
b
e
l
i
n
g
a
n
d
f
a
c
i
l
it
a
t
e
t
h
e
a
d
o
p
t
i
o
n
o
f
r
e
al
-
w
o
r
l
d
a
p
p
l
i
ca
t
i
o
n
s
[
1
0
]
.
Ad
d
i
t
i
o
n
a
l
l
y
,
a
u
g
m
e
n
ta
t
i
o
n
t
e
ch
n
i
q
u
e
s
a
r
e
h
i
g
h
l
y
e
f
f
e
c
t
i
v
e
i
n
a
d
d
r
e
s
s
i
n
g
d
a
t
a
s
ca
r
c
i
t
y
c
h
a
l
l
e
n
g
es
[
1
1
]
.
D
a
t
a
a
u
g
m
e
n
t
a
t
i
o
n
i
s
a
p
p
l
i
e
d
t
o
i
n
c
r
e
a
s
e
t
h
e
n
u
m
b
e
r
o
f
i
m
a
g
e
s
i
n
t
h
e
d
a
ta
s
et
,
a
s
t
h
e
d
i
s
t
r
i
b
u
ti
o
n
o
f
i
m
a
g
e
s
a
c
r
o
s
s
c
l
a
s
s
es
i
s
o
f
t
e
n
u
n
e
v
e
n
[
1
1
]
.
I
m
a
g
e
a
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s
a
r
t
i
f
i
c
i
al
l
y
e
x
p
a
n
d
t
h
e
d
a
t
a
s
e
t
,
e
n
a
b
l
i
n
g
s
y
s
t
em
s
t
o
l
e
a
r
n
h
o
w
i
m
a
g
e
s
a
p
p
e
a
r
f
r
o
m
d
i
f
f
e
r
e
n
t
p
e
r
s
p
e
c
t
i
v
es
,
s
u
c
h
as
w
h
e
n
v
iew
e
d
f
r
o
m
v
a
r
i
o
u
s
a
n
g
l
e
s
o
r
b
l
u
r
r
e
d
d
u
e
t
o
a
d
v
e
r
s
e
w
e
a
t
h
e
r
c
o
n
d
i
t
i
o
n
s
[
1
2
]
.
O
n
e
o
f
t
h
e
a
u
g
m
e
n
t
a
t
i
o
n
m
e
t
h
o
d
s
u
s
e
d
i
n
i
n
t
e
l
l
i
g
e
n
t
t
r
a
n
s
p
o
r
t
a
t
i
o
n
c
a
s
e
s
t
u
d
i
es
i
n
c
l
u
d
e
s
h
o
r
i
z
o
n
t
a
l
f
l
i
p
p
i
n
g
,
r
o
t
a
t
i
o
n
,
a
n
d
G
a
u
s
s
i
an
n
o
i
s
e
.
I
n
t
h
is
r
e
s
ea
r
c
h
,
d
a
t
a
a
u
g
m
e
n
t
a
t
i
o
n
m
e
t
h
o
d
s
a
r
e
a
p
p
l
i
e
d
t
o
e
x
p
a
n
d
t
h
e
d
a
t
a
s
et
a
n
d
a
d
d
r
e
s
s
c
l
as
s
i
m
b
a
l
a
n
c
e
i
s
s
u
es
[
4
]
.
F
u
r
t
h
e
r
m
o
r
e
,
i
n
v
e
h
i
c
l
e
s
p
e
e
d
d
e
t
e
c
ti
o
n
r
e
s
e
a
r
c
h
,
t
e
c
h
n
i
q
u
es
s
u
c
h
a
s
M
o
s
a
i
c
,
r
a
n
d
o
m
p
e
r
s
p
e
c
t
i
v
e
,
Mi
x
u
p
,
H
SV
a
d
j
u
s
t
m
e
n
t
s
,
v
e
r
ti
c
a
l f
l
i
p
(
F
l
i
p
u
d
)
,
a
n
d
h
o
r
i
z
o
n
t
a
l
f
li
p
(
F
li
p
l
r
)
a
r
e
e
m
p
l
o
y
e
d
f
o
r
d
et
e
c
t
i
n
g
t
h
r
e
e
t
y
p
es
o
f
v
e
h
i
cl
es
:
ca
r
s
,
b
u
s
es
,
a
n
d
t
r
u
c
k
s
[
1
2
]
.
O
t
h
e
r
s
t
u
d
ies
e
m
p
l
o
y
b
r
i
g
h
t
n
e
s
s
a
n
d
c
o
n
t
r
a
s
t
a
d
j
u
s
t
m
e
n
t
t
e
c
h
n
iq
u
e
s
f
o
r
r
e
c
o
g
n
i
z
i
n
g
b
l
a
c
k
s
m
o
k
e
f
r
o
m
v
e
h
i
c
l
e
s
i
n
r
o
a
d
t
r
a
f
f
ic
m
o
n
i
t
o
r
i
n
g
v
i
d
e
o
s
[
1
3
]
.
R
e
s
e
a
r
c
h
r
e
l
at
e
d
t
o
r
o
ad
a
r
e
a
s
i
n
v
o
l
v
e
s
d
i
v
i
d
i
n
g
t
h
e
r
o
a
d
i
n
t
o
s
e
c
ti
o
n
s
a
n
d
a
u
g
m
e
n
t
i
n
g
t
h
e
m
u
s
i
n
g
m
e
t
h
o
d
s
s
u
c
h
as
h
o
r
i
z
o
n
t
a
l
f
l
i
p
p
i
n
g
,
c
o
l
o
r
j
i
tt
e
r
,
A
u
t
o
A
u
g
,
a
n
d
M
i
x
u
p
[
1
4
]
.
A
d
d
i
t
i
o
n
a
l
l
y
,
i
n
f
l
o
o
d
d
e
p
t
h
d
e
t
e
c
t
i
o
n
s
t
u
d
ie
s
,
a
u
g
m
e
n
t
a
ti
o
n
t
e
c
h
n
i
q
u
e
s
s
u
c
h
a
s
M
o
s
ai
c
,
la
b
e
l
s
m
o
o
t
h
i
n
g
,
a
n
d
c
o
s
i
n
e
a
n
n
e
a
l
i
n
g
a
r
e
a
p
p
l
i
e
d
i
n
d
e
p
e
n
d
e
n
t
l
y
,
w
i
t
h
M
o
s
ai
c
p
r
o
v
i
n
g
m
o
s
t
e
f
f
e
c
ti
v
e
f
o
r
t
h
is
c
as
e
[
1
5
]
.
I
n
a
s
t
u
d
y
a
i
m
e
d
at
i
n
cr
e
a
s
i
n
g
t
r
ai
n
i
n
g
s
et
d
i
v
e
r
s
i
t
y
a
n
d
p
r
e
v
e
n
t
i
n
g
o
v
er
f
i
t
t
i
n
g
,
m
u
lt
i
-
l
a
b
el
d
at
a
a
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s
s
u
c
h
as
r
a
n
d
o
m
c
r
o
p
p
i
n
g
,
f
l
i
p
p
i
n
g
,
r
o
t
a
t
i
o
n
,
a
n
d
c
o
l
o
r
j
i
tt
e
r
i
n
g
w
e
r
e
a
p
p
l
i
e
d
t
o
i
m
p
r
o
v
e
t
h
e
ac
c
u
r
a
c
y
a
n
d
e
f
f
i
c
i
e
n
c
y
o
f
i
m
a
g
e
a
n
n
o
t
a
ti
o
n
p
r
o
c
e
s
s
es
[
1
6
]
.
Y
O
L
O
is
a
S
o
T
A
,
r
e
a
l
-
ti
m
e
o
b
j
e
c
t
d
e
t
e
c
ti
o
n
a
l
g
o
r
i
t
h
m
t
h
a
t
h
a
s
g
a
i
n
e
d
s
i
g
n
i
f
i
c
a
n
t
a
t
t
e
n
ti
o
n
i
n
t
h
e
c
o
m
p
u
t
e
r
v
i
s
i
o
n
c
o
m
m
u
n
i
t
y
.
Y
O
L
O
i
s
a
s
i
n
g
l
e
-
s
t
a
g
e
d
e
t
e
c
to
r
,
m
e
a
n
i
n
g
i
t
d
e
t
e
ct
s
al
l
o
b
j
ec
t
s
i
n
a
n
i
m
a
g
e
w
it
h
a
s
i
n
g
l
e
f
o
r
w
a
r
d
p
a
s
s
t
h
r
o
u
g
h
t
h
e
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
(
C
N
N
)
[
1
7
]
.
I
t
s
r
e
a
l
-
t
i
m
e
d
e
t
e
c
t
i
o
n
c
a
p
a
b
i
l
it
y
m
a
k
e
s
i
t
i
d
e
a
l
f
o
r
a
p
p
l
i
c
at
i
o
n
s
r
e
q
u
i
r
i
n
g
h
i
g
h
-
s
p
e
e
d
p
r
o
c
e
s
s
i
n
g
.
I
n
o
n
e
s
t
u
d
y
,
Y
O
L
O
v
4
p
r
o
v
e
d
e
f
f
e
c
t
i
v
e
f
o
r
r
e
a
l
-
t
i
m
e
t
r
a
f
f
i
c
s
i
g
n
d
e
t
e
c
t
i
o
n
i
n
a
u
t
o
n
o
m
o
u
s
v
e
h
i
c
l
es
[
1
8
]
.
A
d
d
i
t
i
o
n
a
l
l
y
,
Y
O
L
O
h
as
b
e
e
n
s
u
c
c
es
s
f
u
l
l
y
u
s
e
d
t
o
d
e
t
e
c
t
a
n
d
c
l
as
s
i
f
y
h
u
m
a
n
h
a
n
d
a
c
t
i
o
n
s
i
n
e
g
o
c
e
n
t
r
i
c
v
i
d
e
o
s
[
1
9
]
.
A
n
o
t
h
e
r
s
t
u
d
y
s
h
o
w
e
d
t
h
a
t
t
h
e
Y
O
L
O
m
o
d
e
l
s
i
g
n
i
f
i
c
a
n
t
l
y
i
m
p
r
o
v
e
d
B
al
l
T
r
a
c
k
i
n
g
,
G
o
a
l
A
li
g
n
m
e
n
t
,
a
n
d
R
o
b
o
t
A
v
o
i
d
a
n
c
e
t
as
k
s
i
n
h
u
m
a
n
o
i
d
r
o
b
o
t
s
o
c
c
e
r
[
2
0
]
.
A
n
e
n
h
a
n
c
e
d
v
e
r
s
i
o
n
o
f
Y
O
L
O
v
4
w
i
t
h
S
e
m
i
DSC
o
n
v
a
n
d
t
h
e
F
I
o
U
l
o
s
s
f
u
n
c
t
i
o
n
d
e
m
o
n
s
t
r
a
t
e
d
s
i
g
n
i
f
i
c
a
n
t
p
e
r
f
o
r
m
a
n
c
e
i
n
u
n
d
e
r
w
a
t
e
r
t
a
r
g
e
t
d
e
t
e
c
ti
o
n
[
2
1
]
.
I
n
m
e
d
i
c
a
l
im
a
g
i
n
g
,
Y
O
L
O
a
c
h
i
e
v
e
d
f
as
t
er
c
o
m
p
u
t
a
t
i
o
n
t
i
m
e
s
c
o
m
p
a
r
e
d
t
o
S
S
D
n
et
w
o
r
k
s
f
o
r
o
v
a
r
i
a
n
t
u
m
o
r
c
l
a
s
s
i
f
i
c
at
i
o
n
i
n
u
l
t
r
a
s
o
u
n
d
i
m
a
g
es
[
2
2
]
.
A
s
y
s
t
e
m
a
ti
c
l
it
e
r
a
t
u
r
e
r
e
v
i
e
w
(
S
L
R
)
o
f
s
t
u
d
i
es
o
n
t
r
a
f
f
i
c
s
i
g
n
d
e
t
e
ct
i
o
n
a
n
d
r
e
c
o
g
n
i
t
i
o
n
u
s
i
n
g
Y
O
L
O
,
p
u
b
l
i
s
h
e
d
b
e
t
w
e
e
n
2
0
1
6
a
n
d
2
0
2
2
,
h
i
g
h
l
i
g
h
t
e
d
a
d
v
a
n
c
e
m
e
n
t
s
i
n
d
e
t
e
c
ti
o
n
,
c
l
a
s
s
i
f
i
ca
t
i
o
n
,
a
n
d
p
r
o
c
e
s
s
i
n
g
s
p
e
e
d
f
o
r
t
r
a
f
f
i
c
s
i
g
n
r
e
c
o
g
n
i
t
i
o
n
s
y
s
te
m
s
a
c
r
o
s
s
v
a
r
i
o
u
s
YO
L
O
v
e
r
s
i
o
n
s
.
T
h
e
r
es
u
l
ts
s
h
o
w
t
h
a
t
YO
L
O
i
s
c
a
p
a
b
l
e
o
f
d
e
t
e
ct
in
g
a
n
d
r
e
c
o
g
n
i
z
i
n
g
t
r
a
f
f
i
c
s
i
g
n
s
w
i
t
h
h
i
g
h
a
c
c
u
r
a
cy
a
n
d
f
a
s
t
p
r
o
c
e
s
s
i
n
g
t
i
m
es
[
23]
.
F
u
r
t
h
e
r
m
o
r
e
,
r
e
s
e
a
r
c
h
u
s
i
n
g
t
h
e
l
a
t
e
s
t
YO
L
O
m
o
d
e
l
,
Y
O
L
O
v
9
,
h
as
s
h
o
w
n
th
a
t
i
t
c
a
n
h
a
n
d
l
e
m
u
lt
i
p
l
e
a
d
v
e
r
s
e
c
o
n
d
i
t
i
o
n
s
s
i
m
u
l
t
a
n
e
o
u
s
l
y
i
n
o
b
j
e
c
t
d
e
t
e
ct
i
o
n
t
a
s
k
s
[
2
4
]
.
M
a
n
y
s
t
u
d
ie
s
h
a
v
e
d
e
m
o
n
s
tr
a
t
e
d
t
h
a
t
d
a
t
a
a
u
g
m
e
n
t
a
ti
o
n
c
a
n
s
i
g
n
i
f
i
c
a
n
tl
y
i
m
p
r
o
v
e
Y
O
L
O
’
s
p
e
r
f
o
r
m
a
n
c
e
i
n
o
b
j
e
c
t
d
e
t
e
ct
i
o
n
a
n
d
c
l
a
s
s
i
f
i
c
at
i
o
n
.
F
o
r
i
n
s
t
a
n
c
e
,
r
e
s
e
a
r
c
h
u
s
i
n
g
Y
OL
O
v
5
em
p
l
o
y
e
d
t
r
a
d
i
t
i
o
n
a
l
d
a
t
a
a
u
g
m
e
n
t
a
ti
o
n
m
e
t
h
o
d
s
s
u
c
h
a
s
n
o
is
e
a
d
d
i
ti
o
n
,
c
r
o
p
p
i
n
g
,
f
l
i
p
p
i
n
g
,
r
o
t
a
t
i
o
n
,
b
r
i
g
h
tn
e
s
s
,
a
n
d
c
o
n
t
r
a
s
t
a
d
j
u
s
t
m
e
n
ts
t
o
e
n
r
i
c
h
t
h
e
d
a
tas
e
t
[
2
5
]
.
T
h
es
e
t
e
c
h
n
i
q
u
e
s
h
a
v
e
b
e
e
n
a
p
p
l
i
e
d
t
o
YO
L
O
v
5
,
as
w
el
l
a
s
Y
O
L
O
v
6
,
Y
O
L
O
v
7
,
a
n
d
Y
O
L
O
v
8
,
t
o
e
n
h
a
n
c
e
v
e
h
i
c
l
e
d
e
t
e
c
t
i
o
n
a
n
d
c
l
a
s
s
i
f
i
c
at
i
o
n
i
n
s
u
r
v
e
i
l
la
n
c
e
f
o
o
t
a
g
e
[
2
6
]
.
S
p
e
c
i
f
i
ca
l
l
y
,
i
n
Y
O
L
O
v
7
,
r
e
s
ea
r
c
h
e
r
s
e
x
p
a
n
d
e
d
t
h
e
D
A
W
N
d
a
t
a
s
et
b
y
a
p
p
l
y
i
n
g
a
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s
s
u
c
h
a
s
b
l
u
r
,
s
a
t
u
r
a
t
i
o
n
,
b
r
i
g
h
t
n
e
s
s
,
d
a
r
k
n
e
s
s
,
n
o
i
s
e
,
e
x
p
o
s
u
r
e
,
h
u
e
,
a
n
d
g
r
a
y
s
c
a
l
e
[
2
7
]
.
A
d
d
i
t
i
o
n
a
l
l
y
,
f
o
r
p
o
t
h
o
l
e
d
e
t
e
c
t
i
o
n
u
s
i
n
g
Y
O
L
O
v
8
,
t
e
c
h
n
i
q
u
e
s
s
u
c
h
a
s
r
o
t
at
i
o
n
,
s
ca
l
i
n
g
,
a
n
d
f
l
i
p
p
i
n
g
w
e
r
e
a
p
p
l
i
e
d
t
o
c
r
e
a
t
e
m
o
r
e
i
m
a
g
e
v
a
r
i
a
t
i
o
n
s
,
t
h
e
r
e
b
y
i
m
p
r
o
v
i
n
g
d
e
t
e
c
ti
o
n
p
e
r
f
o
r
m
a
n
c
e
[
2
8
]
.
I
n
v
e
h
i
c
l
e
d
et
e
c
ti
o
n
c
a
s
e
s
tu
d
i
e
s
,
a
u
g
m
e
n
t
at
i
o
n
m
e
t
h
o
d
s
s
u
c
h
as
r
o
t
at
i
o
n
,
w
i
d
th
a
n
d
h
e
i
g
h
t
s
h
i
f
ts
,
b
r
i
g
h
t
n
es
s
a
d
j
u
s
t
m
e
n
ts
,
z
o
o
m
i
n
g
,
a
n
d
h
o
r
i
z
o
n
t
a
l
f
li
p
p
i
n
g
w
e
r
e
u
s
e
d
t
o
e
n
h
a
n
c
e
t
h
e
ac
c
u
r
a
c
y
o
f
Y
O
L
O
v
2
a
n
d
Y
O
L
O
v
4
[
2
9
]
.
D
u
r
i
n
g
t
r
a
f
f
i
c
v
e
h
ic
l
e
d
at
a
s
et
a
u
g
m
e
n
t
a
ti
o
n
,
m
u
l
t
i
-
l
a
b
e
l
c
h
a
l
l
e
n
g
es
a
r
is
e
,
an
d
t
r
a
i
n
i
n
g
C
NN
m
o
d
e
ls
o
n
s
m
a
l
l
d
a
t
as
e
ts
is
a
ls
o
c
h
a
ll
e
n
g
in
g
.
T
h
e
r
e
f
o
r
e
,
m
o
d
e
l
e
n
h
a
n
c
e
m
e
n
t
t
h
r
o
u
g
h
m
u
l
t
i
-
a
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
es
is
n
e
c
e
s
s
a
r
y
t
o
f
u
r
t
h
e
r
i
m
p
r
o
v
e
p
e
r
f
o
r
m
a
n
c
e
[
1
6
]
.
T
h
e
c
o
n
t
r
i
b
u
t
i
o
n
o
f
t
h
i
s
r
es
e
a
r
c
h
i
s
t
h
e
d
e
v
e
l
o
p
m
e
n
t
o
f
t
h
e
Y
O
L
O
m
o
d
e
l
u
s
i
n
g
m
u
l
t
i
-
a
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s
t
o
e
n
h
a
n
c
e
r
e
a
l
-
ti
m
e
d
e
t
e
c
t
i
o
n
a
c
c
u
r
a
c
y
o
f
v
e
h
i
c
l
e
t
y
p
e
s
a
n
d
c
o
u
n
t
s
f
r
o
m
C
C
T
V
f
o
o
t
a
g
e
.
T
h
e
m
o
d
e
l
i
s
e
x
p
e
c
te
d
t
o
i
m
p
r
o
v
e
a
c
c
u
r
a
cy
u
n
d
e
r
d
i
v
e
r
s
e
a
n
d
c
o
m
p
l
e
x
u
r
b
a
n
t
r
a
f
f
i
c
c
o
n
d
i
t
i
o
n
s
.
T
h
e
d
ev
e
l
o
p
m
e
n
t
p
r
i
m
a
r
i
l
y
f
o
c
u
s
e
s
o
n
a
d
d
r
e
s
s
i
n
g
c
h
al
l
e
n
g
e
s
r
e
l
at
e
d
t
o
t
r
a
f
f
i
c
d
e
n
s
it
y
d
u
r
in
g
p
e
a
k
h
o
u
r
s
,
s
m
a
l
l
o
b
j
ec
t
d
e
te
c
t
i
o
n
,
a
n
d
v
a
r
y
i
n
g
l
i
g
h
t
i
n
g
c
o
n
d
i
t
i
o
n
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
1
,
J
u
ly
20
25
:
656
-
6
6
8
658
2.
M
E
T
H
O
D
2
.
1
.
M
et
ho
do
lo
g
y
W
e
p
r
esen
t
a
wo
r
k
f
lo
w
f
o
r
r
e
al
-
tim
e
m
u
lti
-
lab
el
v
e
h
icle
d
et
ec
tio
n
in
tr
a
f
f
ic
u
s
in
g
th
e
YOL
O
m
o
d
el
en
h
an
ce
d
with
m
u
lti
-
a
u
g
m
en
tatio
n
tech
n
iq
u
es.
T
h
e
d
iag
r
a
m
b
elo
w
illu
s
tr
ates
ea
ch
s
ta
g
e,
f
r
o
m
r
aw
d
ata
ac
q
u
is
itio
n
to
m
o
d
el
ev
alu
atio
n
,
with
an
em
p
h
asis
o
n
th
e
u
s
e
o
f
m
u
lti
-
au
g
m
en
tatio
n
tech
n
iq
u
es
to
im
p
r
o
v
e
th
e
YOL
O
m
o
d
el
’
s
p
er
f
o
r
m
a
n
ce
.
T
h
e
wo
r
k
f
lo
w
d
ia
g
r
am
is
s
h
o
wn
in
Fig
u
r
e
1
.
Fig
u
r
e
1
illu
s
tr
ates
th
e
wo
r
k
f
l
o
w
f
o
r
v
eh
icle
d
etec
tio
n
i
n
tr
af
f
ic
u
s
in
g
th
e
YOL
O
m
o
d
el
en
h
an
ce
d
with
m
u
lti
-
au
g
m
en
tatio
n
tech
n
iq
u
es.
T
h
e
p
r
o
ce
s
s
b
eg
in
s
with
v
id
eo
ca
p
tu
r
e
f
r
o
m
C
C
T
V
ca
m
er
as
r
ec
o
r
d
in
g
v
eh
icle
tr
af
f
ic
f
o
r
a
n
aly
s
is
.
T
h
e
v
id
eo
is
th
en
p
r
o
ce
s
s
ed
in
th
e
d
ata
p
r
ep
r
o
ce
s
s
in
g
s
tag
e
to
p
r
e
p
ar
e
it
f
o
r
an
n
o
tatio
n
.
I
m
a
g
es
ar
e
ac
q
u
ir
ed
b
y
s
p
litt
in
g
th
e
v
id
eo
r
ec
o
r
d
in
g
s
i
n
to
f
r
am
es
at
5
-
m
in
u
te
in
ter
v
als
d
u
r
in
g
d
ata
p
r
ep
r
o
ce
s
s
in
g
.
E
ac
h
v
e
h
i
cle
in
th
e
im
ag
es
i
s
th
en
an
n
o
tated
b
y
ty
p
e,
in
clu
d
i
n
g
d
etaile
d
in
f
o
r
m
atio
n
s
u
ch
as
lo
ca
tio
n
an
d
b
o
u
n
d
in
g
b
o
x
co
o
r
d
i
n
ates.
T
h
is
en
s
u
r
es
th
e
d
ata
is
r
ea
d
y
f
o
r
f
u
r
th
e
r
an
aly
s
is
o
r
m
o
d
el
tr
ain
in
g
.
Af
ter
an
n
o
tatio
n
,
th
e
d
ata
is
s
p
lit in
to
8
0
% f
o
r
tr
ain
i
n
g
an
d
2
0
% f
o
r
v
alid
atio
n
,
e
n
s
u
r
in
g
th
e
m
o
d
el
is
p
r
o
p
er
l
y
tr
ai
n
ed
a
n
d
ev
alu
a
te
d
o
n
u
n
s
ee
n
d
ata
[
3
0
]
-
[
3
2
]
.
N
ex
t,
v
a
r
io
u
s
au
g
m
en
tatio
n
tec
h
n
iq
u
es
a
r
e
ap
p
lied
to
th
e
d
ataset
to
i
n
cr
ea
s
e
d
ata
v
ar
iab
ilit
y
an
d
im
p
r
o
v
e
m
o
d
el
p
e
r
f
o
r
m
an
ce
.
T
h
e
ap
p
lied
au
g
m
en
tatio
n
tech
n
iq
u
es
in
clu
d
e
b
lu
r
,
b
r
ig
h
tn
ess
ad
ju
s
tm
en
t,
co
n
tr
ast
ad
ju
s
tm
en
t,
co
lo
r
jitt
er
,
cr
o
p
p
in
g
,
f
lip
p
in
g
,
n
o
is
e
i
n
j
e
c
t
i
o
n
,
r
o
t
at
i
o
n
,
s
ca
l
i
n
g
,
a
n
d
z
o
o
m
i
n
.
T
h
e
i
m
p
l
e
m
e
n
t
e
d
a
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
es
a
r
e
il
l
u
s
t
r
a
t
e
d
i
n
Fi
g
u
r
e
2
.
Fig
u
r
e
1
.
Mo
d
el
wo
r
k
f
lo
w
Fig
u
r
e
2
.
I
m
p
lem
en
te
d
au
g
m
e
n
tatio
n
tech
n
iq
u
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
B
o
o
s
tin
g
r
ea
l
-
time
ve
h
icle
d
et
ec
tio
n
in
u
r
b
a
n
tr
a
ffic u
s
in
g
a
n
o
ve
l m
u
lti
-
a
u
g
men
ta
tio
n
(
I
m
a
m
A
h
ma
d
A
s
h
a
r
i
)
659
Fig
u
r
e
2
s
h
o
ws
th
at
ea
c
h
au
g
m
en
tatio
n
tech
n
i
q
u
e
tr
i
p
les
th
e
d
ataset
s
ize,
ex
p
an
d
in
g
5
7
6
f
r
am
es
in
to
1
,
7
2
8
a
u
g
m
e
n
ted
f
r
am
es
to
s
im
u
late
r
ea
l
-
wo
r
ld
co
n
d
itio
n
s
lik
e
lig
h
tin
g
ch
an
g
es,
co
lo
r
v
a
r
iatio
n
s
,
an
d
v
is
u
al
d
is
tu
r
b
an
ce
s
,
en
h
a
n
cin
g
t
h
e
m
o
d
el
’
s
ac
cu
r
ac
y
an
d
r
o
b
u
s
tn
ess
.
T
h
e
n
ex
t
s
tag
e
in
v
o
lv
es
test
in
g
YOL
Ov
8
n
a
n
d
YOL
Ov
9
t
with
th
ese
au
g
m
e
n
tatio
n
s
to
ev
alu
ate
v
e
h
icle
d
etec
tio
n
in
tr
af
f
ic
v
id
eo
s
.
T
h
e
p
r
o
ce
s
s
in
clu
d
es
tr
ain
in
g
e
v
alu
atio
n
,
test
in
g
o
n
v
ar
ied
r
o
a
d
c
o
n
d
itio
n
s
,
an
d
f
in
al
ass
ess
m
en
t
u
s
in
g
test
in
g
m
etr
ics
to
m
ea
s
u
r
e
ac
cu
r
ac
y
a
n
d
id
e
n
tify
ar
ea
s
f
o
r
im
p
r
o
v
em
e
n
t.
T
h
is
wo
r
k
f
l
o
w,
f
r
o
m
d
ata
ac
q
u
is
itio
n
to
m
o
d
el
ev
al
u
atio
n
,
em
p
h
asiz
es a
u
g
m
en
tatio
n
’
s
r
o
le
in
en
h
an
ci
n
g
YOL
O
’
s
p
er
f
o
r
m
an
ce
,
as illu
s
tr
ated
in
Fig
u
r
e
3
.
Fig
u
r
e
3
s
h
o
ws
th
e
d
esig
n
f
o
r
test
in
g
th
e
f
lip
p
i
n
g
au
g
m
en
ta
tio
n
tech
n
iq
u
e
with
YOL
O.
T
h
e
p
r
o
ce
s
s
b
eg
in
s
with
tr
af
f
ic
im
ag
e
tr
ai
n
in
g
s
am
p
les
u
n
d
er
g
o
in
g
h
o
r
i
zo
n
tal
f
lip
(
_
=
1
.
0
−
_
)
an
d
v
er
tical
f
lip
(
_
=
1
.
0
−
_
)
,
p
r
o
d
u
cin
g
tr
an
s
f
o
r
m
ed
im
ag
es.
T
h
e
a
u
g
m
en
te
d
im
ag
es
a
r
e
f
e
d
in
to
th
e
YOL
O
f
ea
tu
r
e
ex
tr
ac
to
r
to
ca
p
tu
r
e
ess
en
tial
f
ea
tu
r
es.
T
h
e
YOL
O
o
u
tp
u
t
is
p
r
o
c
ess
ed
b
y
th
r
ee
m
ai
n
h
ea
d
s
:
th
e
class
if
icatio
n
h
ea
d
f
o
r
o
b
ject
class
if
icatio
n
,
th
e
d
etec
tio
n
h
ea
d
f
o
r
o
b
ject
d
ete
ctio
n
an
d
m
ar
k
in
g
,
an
d
th
e
ap
p
ea
r
an
ce
em
b
ed
d
i
n
g
h
ea
d
f
o
r
o
b
ject
em
b
ed
d
in
g
.
T
h
is
tech
n
iq
u
e
aim
s
to
en
h
an
ce
YOL
O
m
o
d
el
p
er
f
o
r
m
an
ce
b
y
r
ec
o
g
n
izin
g
o
b
jects f
r
o
m
v
a
r
io
u
s
o
r
ie
n
tatio
n
s
.
Fig
u
r
e
3
.
Fli
p
p
in
g
au
g
m
en
tati
o
n
tech
n
i
q
u
e
m
o
d
el
d
esig
n
wit
h
YOL
O
2
.
2
.
Resea
rc
h
ins
t
rum
ent
T
h
e
d
ataset
was
co
llected
f
r
o
m
s
u
r
v
eillan
ce
ca
m
er
as
at
th
e
Fatm
awa
ti
tr
af
f
ic
lig
h
ts
in
Sem
ar
an
g
C
ity
b
etwe
en
Dec
em
b
er
1
9
,
2
0
2
3
,
a
n
d
Feb
r
u
ar
y
1
5
,
2
0
2
4
,
f
r
o
m
0
6
:
0
0
to
0
7
:0
0
W
I
B
.
Vid
eo
s
wer
e
r
ec
o
r
d
e
d
in
H.
2
6
4
f
o
r
m
at
at
1
2
8
0
×
9
6
0
r
eso
lu
tio
n
a
n
d
2
5
FP
S,
f
o
c
u
s
in
g
o
n
v
eh
icles
in
th
e
r
ig
h
t
lan
e.
Fra
m
es
wer
e
ex
tr
ac
ted
ev
er
y
5
m
in
u
tes
u
s
in
g
a
ca
lcu
lated
in
ter
v
al
o
f
7
5
0
0
f
r
am
es,
y
ield
in
g
7
2
0
im
ag
es
ca
teg
o
r
ized
in
to
m
o
to
r
cy
cles
(
3
1
,
4
8
1
)
,
ca
r
s
(
1
2
,
4
0
2
)
,
tr
u
ck
s
(
1
,
1
8
4
)
,
an
d
b
u
s
es
(
2
8
0
)
.
Fo
r
tr
ain
in
g
,
5
7
6
s
am
p
les
wer
e
u
s
ed
,
wh
ile
1
4
4
s
am
p
les
wer
e
allo
ca
ted
f
o
r
v
alid
atio
n
.
An
n
o
tati
o
n
was
co
n
d
u
cted
u
s
in
g
R
o
b
o
f
lo
w
[
3
3
]
.
Mo
d
el
an
aly
s
is
was p
er
f
o
r
m
ed
o
n
Go
o
g
l
e
C
o
lab
u
s
in
g
Py
th
o
n
3
.
1
0
.
1
2
an
d
Py
T
o
r
c
h
2
.
3
.
0
with
C
UDA
1
2
.
1
s
u
p
p
o
r
t.
2
.
3
.
Da
t
a
a
na
ly
s
is
t
ec
hn
iqu
e
s
T
h
e
d
ata
an
al
y
s
is
b
eg
an
with
co
llectin
g
f
r
am
es
f
r
o
m
C
C
T
V
tr
af
f
ic
v
id
e
o
s
,
wh
ich
wer
e
u
p
lo
ad
ed
t
o
R
o
b
o
f
lo
w
f
o
r
d
ataset
m
an
ag
e
m
en
t
an
d
an
n
o
tatio
n
[
3
4
]
.
Ve
h
icle
class
es
m
o
to
r
cy
cles,
ca
r
s
,
b
u
s
es,
an
d
tr
u
ck
s
wer
e
cr
ea
ted
to
ca
teg
o
r
ize
an
d
tr
ain
th
e
m
o
d
el
ef
f
ec
tiv
ely
.
E
ac
h
im
ag
e
was
an
n
o
tated
b
y
m
ar
k
in
g
v
eh
icles
ac
co
r
d
in
g
to
th
eir
class
,
en
s
u
r
in
g
ac
cu
r
ate
lab
elin
g
f
o
r
tr
ain
in
g
.
T
h
e
d
ataset
was
th
en
s
p
li
t
in
to
8
0
%
tr
ain
in
g
(
5
7
6
im
ag
es)
an
d
2
0
%
v
alid
atio
n
(
1
4
4
im
a
g
es).
Au
g
m
en
t
at
io
n
was
p
er
f
o
r
m
ed
u
s
in
g
Py
th
o
n
,
tr
ip
lin
g
th
e
d
ataset
b
y
ap
p
ly
in
g
two
au
g
m
en
tatio
n
tech
n
i
q
u
es
p
er
i
m
ag
e
wh
ile
k
ee
p
in
g
o
n
e
as
th
e
o
r
ig
in
al.
T
h
e
au
g
m
en
tatio
n
tech
n
iq
u
e
v
alu
e
s
ar
e
d
etailed
in
T
ab
le
1
.
E
ac
h
au
g
m
en
tatio
n
tech
n
iq
u
e
f
r
o
m
T
a
b
le
1
u
s
es
s
p
ec
if
i
c
v
alu
es
to
g
e
n
er
ate
im
ag
e
v
ar
iatio
n
s
.
T
h
e
b
l
u
r
tec
h
n
iq
u
e
u
s
es
k
e
r
n
el
s
izes
o
f
1
f
o
r
i
m
ag
e
2
a
n
d
2
f
o
r
i
m
a
g
e
3
.
Fo
r
b
r
ig
h
tn
ess
ad
ju
s
tm
en
t,
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
1
,
J
u
ly
20
25
:
656
-
6
6
8
660
b
r
ig
h
tn
ess
f
ac
to
r
s
ar
e
0
.
8
f
o
r
i
m
ag
e
2
an
d
1
.
2
f
o
r
i
m
a
g
e
3
.
C
o
n
tr
ast
ad
ju
s
tm
en
t
u
s
es
alp
h
a
v
alu
es
o
f
1
.
5
f
o
r
i
m
ag
e
2
a
n
d
2
.
0
f
o
r
i
m
a
g
e
3
.
I
n
co
lo
r
jitt
er
ad
j
u
s
tm
en
t,
a
c
o
m
b
in
atio
n
o
f
b
r
ig
h
tn
ess
,
co
n
tr
ast,
an
d
s
atu
r
atio
n
f
ac
to
r
s
is
r
an
d
o
m
ize
d
with
in
a
r
an
g
e
o
f
0
.
6
to
1
.
4
,
wh
ile
th
e
h
u
e
f
ac
to
r
is
r
an
d
o
m
ized
b
e
twee
n
-
0
.
1
a
n
d
0
.
1
.
I
m
ag
e
cr
o
p
p
in
g
is
d
o
n
e
with
t
wo
ty
p
es
o
f
cr
o
p
s
:
to
p
cr
o
p
f
r
o
m
(
0
,
0
)
to
(
wid
th
,
h
eig
h
t
*
0
.
5
)
o
n
i
m
ag
e
2
,
an
d
b
o
tto
m
c
r
o
p
f
r
o
m
(
0
,
h
ei
g
h
t
*
0
.
5
)
to
(
wid
th
,
h
eig
h
t)
o
n
i
m
ag
e
3
.
Fli
p
p
in
g
in
clu
d
es
b
o
th
h
o
r
izo
n
tal
a
n
d
v
er
tical
f
lip
s
,
with
th
e
h
o
r
izo
n
tal
f
lip
ce
n
ter
at
x
_
ce
n
ter
=
1
.
0
f
o
r
i
m
ag
e
2
an
d
-
x
_
ce
n
ter
f
o
r
v
er
tical
f
lip
o
n
i
m
ag
e
3
.
No
is
e
in
jectio
n
a
d
d
s
Gau
s
s
ian
n
o
is
e
with
v
alu
es
r
an
d
o
m
ized
b
etwe
en
0
a
n
d
0
.
1
f
o
r
b
o
th
i
m
ag
e
2
an
d
i
m
ag
e
3
.
R
o
tatio
n
is
ap
p
l
ied
at
9
0
d
eg
r
e
es
f
o
r
i
m
ag
e
2
an
d
2
7
0
d
eg
r
ee
s
f
o
r
i
m
ag
e
3
.
Scalin
g
ap
p
lies
a
f
ac
to
r
b
etwe
en
0
.
8
a
n
d
1
.
2
f
o
r
b
o
th
i
m
ag
e
2
an
d
i
m
a
g
e
3
,
w
h
ile
zo
o
m
-
in
u
s
es
a
f
ac
to
r
o
f
1
.
2
f
o
r
i
m
a
g
e
2
an
d
1
.
5
f
o
r
i
m
a
g
e
3
.
E
ac
h
v
alu
e
is
ca
r
ef
u
lly
d
esig
n
e
d
to
c
r
ea
t
e
s
ig
n
if
ican
t
im
ag
e
v
ar
iatio
n
s
,
im
p
r
o
v
in
g
m
o
d
el
p
er
f
o
r
m
an
ce
u
n
d
er
v
ar
io
u
s
v
eh
icle
r
ec
o
g
n
itio
n
c
o
n
d
itio
n
s
.
An
ex
a
m
p
le
o
f
th
e
o
u
t
p
u
t
f
r
o
m
ea
ch
d
ata
au
g
m
en
tatio
n
tech
n
iq
u
e
u
s
in
g
th
e
s
ec
o
n
d
v
al
u
e
ca
n
b
e
s
ee
n
i
n
Fig
u
r
e
4
.
T
ab
le
1
.
Au
g
m
en
tatio
n
v
alu
es
No
A
u
g
V
a
l
u
e
A
u
g
m
e
n
t
a
t
i
o
n
f
a
c
t
o
r
(
i
m
a
g
e
)
R
e
f
e
r
e
n
c
e
s
1
2
3
1
B
l
u
r
K
e
r
n
e
l
s
i
z
e
-
1
2
[
3
5
]
2
B
r
i
g
h
t
n
e
ss A
d
j
u
st
m
e
n
t
B
r
i
g
h
t
n
e
ss
f
a
c
t
o
r
-
0
.
8
1
.
2
[
3
6
]
3
C
o
n
t
r
a
st
A
d
j
u
s
t
me
n
t
A
l
p
h
a
-
1
.
5
2
.
0
[
3
7
]
4
C
o
l
o
r
j
i
t
t
e
r
(
B
r
i
g
h
t
n
e
s,
c
o
n
t
r
a
st
,
sat
u
r
a
t
i
o
n
)
a
n
d
h
u
e
-
R
a
n
d
(
0
.
6
,
1
.
4
)
a
n
d
R
a
n
d
(
-
0
.
1
,
0
.
1
)
R
a
n
d
(
0
.
6
,
1
.
4
)
a
n
d
R
a
n
d
(
-
0
.
1
,
0
.
1
)
[
3
8
]
5
C
r
o
p
p
i
n
g
C
r
o
p
h
e
i
g
h
t
f
r
a
c
t
i
o
n
-
To
p
c
r
o
p
(
0
,
0
)
t
o
(
w
i
d
t
h
,
h
e
i
g
h
t
*
0
.
5
)
B
o
t
t
o
m
C
r
o
p
(
0
,
h
e
i
g
h
t
*
0
.
5
)
t
o
(
w
i
d
t
h
,
h
e
i
g
h
t
)
[
3
9
]
6
F
l
i
p
p
i
n
g
H
o
r
i
z
o
n
t
a
l
a
n
d
v
e
r
t
i
c
a
l
f
l
i
p
-
H
o
r
i
z
o
n
t
a
l
f
lip
x
_
c
e
n
t
e
r
=
1
.
0
-
x
_
c
e
n
t
e
r
V
e
r
t
i
c
a
l
f
l
i
p
y
_
c
e
n
t
e
r
=
1
.
0
-
y
_
c
e
n
t
e
r
[
4
0
]
7
N
o
i
se
i
n
j
e
c
t
i
o
n
G
a
u
ss
i
a
n
n
o
i
se
-
R
a
n
d
(
0
,
0
.
1
)
R
a
n
d
(
0
,
0
.
1
)
[
3
9
]
8
R
o
t
a
t
i
o
n
R
o
t
a
t
i
o
n
-
90
’
2
7
0
’
[
4
1
]
9
S
c
a
l
i
n
g
S
c
a
l
e
i
ma
g
e
-
R
a
n
d
(
0
.
8
,
1
.
2
)
R
a
n
d
(
0
.
8
,
1
.
2
)
[
4
2
]
10
Zo
o
m
i
n
Zo
o
m
i
n
-
1
.
2
1
.
5
[
4
3
]
Fig
u
r
e
4
.
E
x
am
p
le
o
f
d
ata
a
u
g
m
en
tatio
n
o
u
tp
u
ts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
B
o
o
s
tin
g
r
ea
l
-
time
ve
h
icle
d
et
ec
tio
n
in
u
r
b
a
n
tr
a
ffic u
s
in
g
a
n
o
ve
l m
u
lti
-
a
u
g
men
ta
tio
n
(
I
m
a
m
A
h
ma
d
A
s
h
a
r
i
)
661
2
.
4
.
M
ulti
-
a
ug
m
ent
a
t
io
n
Mu
lti
-
au
g
m
en
tatio
n
in
im
ag
e
d
ataset
p
r
o
ce
s
s
in
g
en
h
an
ce
s
m
o
d
el
ad
ap
ta
b
ilit
y
to
r
ea
l
-
wo
r
ld
co
n
d
itio
n
s
b
y
in
cr
ea
s
in
g
d
atas
et
v
ar
iab
ilit
y
with
o
u
t
ad
d
itio
n
al
d
ata
c
o
llectio
n
.
Ap
p
ly
in
g
m
u
ltip
le
tech
n
i
q
u
es
s
im
u
ltan
eo
u
s
ly
im
p
r
o
v
es
o
b
je
ct
r
ec
o
g
n
itio
n
u
n
d
er
d
if
f
er
e
n
t
lig
h
tin
g
,
s
izes,
o
r
ien
tatio
n
s
,
a
n
d
im
ag
e
q
u
alities
,
m
ak
in
g
th
e
m
o
d
el
m
o
r
e
r
o
b
u
s
t.
T
h
e
au
g
m
en
tatio
n
tech
n
iq
u
es
u
s
ed
i
n
th
is
s
tu
d
y
ad
d
r
ess
k
ey
o
b
ject
r
ec
o
g
n
itio
n
ch
allen
g
es a
n
d
ar
e
lis
ted
in
T
ab
le
2
.
T
ab
le
2
s
h
o
ws
th
at
t
h
e
f
ir
s
t
co
m
b
in
atio
n
f
o
c
u
s
es
o
n
s
m
al
l
o
b
jects
an
d
lig
h
tin
g
,
w
h
ile
th
e
s
ec
o
n
d
d
ea
ls
with
v
ar
iatio
n
s
in
an
g
le
an
d
lig
h
tin
g
.
T
h
e
th
ir
d
c
o
m
b
i
n
atio
n
tar
g
ets
s
ize
an
d
co
lo
r
v
ar
iatio
n
s
,
wh
ile
th
e
f
o
u
r
th
f
o
cu
s
es
o
n
d
etail
en
h
a
n
ce
m
en
t.
T
h
e
f
in
al
co
m
b
in
ati
o
n
in
co
r
p
o
r
ates
all
au
g
m
en
ta
tio
n
tech
n
iq
u
es
to
en
s
u
r
e
o
v
er
all
r
o
b
u
s
tn
ess
.
Ma
th
em
atica
lly
,
if
is
th
e
o
r
ig
in
a
l
d
ataset,
th
en
af
ter
ap
p
ly
i
n
g
t
h
e
au
g
m
e
n
tatio
n
co
m
b
in
atio
n
s
,
th
e
n
ew
d
atase
t
ca
n
b
e
r
e
p
r
esen
ted
as
Y
=
∑
A
i
(X)
n
i
-
1
,
wh
er
e
is
th
e
i
-
th
au
g
m
en
tatio
n
tech
n
iq
u
e
ap
p
lied
to
d
ataset
.
T
h
is
r
esu
lts
in
a
lar
g
e
r
a
n
d
m
o
r
e
v
ar
ie
d
d
ataset,
en
h
a
n
cin
g
th
e
m
o
d
el
’
s
ab
ilit
y
to
g
en
er
alize
.
T
ab
le
2
.
Mu
lti
au
g
m
en
tatio
n
tech
n
iq
u
e
No
A
u
g
m
e
n
t
a
t
i
o
n
t
e
c
h
n
i
q
u
e
c
o
m
b
i
n
a
t
i
o
n
F
o
c
u
s
R
e
f
e
r
e
n
c
e
1
S
c
a
l
i
n
g
+
C
r
o
p
p
i
n
g
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
st
me
n
t
+
N
o
i
s
e
i
n
j
e
c
t
i
o
n
+
B
l
u
r
S
mal
l
o
b
j
e
c
t
s
a
n
d
l
i
g
h
t
i
n
g
[
4
4
]
,
[
4
5
]
,
[
4
6
]
2
R
o
t
a
t
i
o
n
+
F
l
i
p
p
i
n
g
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
st
me
n
t
+
C
o
n
t
r
a
st
a
d
j
u
s
t
me
n
t
+
C
o
l
o
r
j
i
t
t
e
r
A
n
g
l
e
a
n
d
l
i
g
h
t
i
n
g
v
a
r
i
a
t
i
o
n
[
4
5
]
,
[
4
7
]
3
S
c
a
l
i
n
g
+
Z
o
o
m
i
n
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
s
t
me
n
t
+
C
o
l
o
r
j
i
t
t
e
r
+
N
o
i
se
i
n
j
e
c
t
i
o
n
S
i
z
e
a
n
d
c
o
l
o
r
v
a
r
i
a
t
i
o
n
[
4
5
]
,
[
4
8
]
4
C
r
o
p
p
i
n
g
+
Z
o
o
m
i
n
+
C
o
n
t
r
a
s
t
a
d
j
u
s
t
men
t
+
N
o
i
se
i
n
j
e
c
t
i
o
n
+
B
l
u
r
D
e
t
a
i
l
e
n
h
a
n
c
e
m
e
n
t
[
4
5
]
,
[
4
8
]
,
[
4
9
]
,
[
4
0
]
5
S
c
a
l
i
n
g
+
R
o
t
a
t
i
o
n
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
s
t
me
n
t
+
C
o
n
t
r
a
st
a
d
j
u
s
t
me
n
t
+
N
o
i
se
i
n
j
e
c
t
i
o
n
C
o
m
p
l
e
t
e
c
o
m
b
i
n
a
t
i
o
n
f
o
r
r
o
b
u
st
n
e
s
s
[
4
0
]
,
[
5
0
]
,
[
5
1
]
2
.
5
.
YO
L
O
m
o
dels
YOL
O
is
wid
ely
u
s
ed
in
tr
an
s
p
o
r
tatio
n
s
y
s
tem
s
f
o
r
its
h
ig
h
d
etec
tio
n
ac
cu
r
ac
y
an
d
r
ea
l
-
tim
e
p
er
f
o
r
m
an
ce
[
5
2
]
.
YOL
Ov
8
i
n
tr
o
d
u
ce
s
f
iv
e
s
ca
led
v
er
s
io
n
s
-
YOL
Ov
8
n
,
YOL
Ov
8
s
,
YOL
Ov
8
m
,
YOL
Ov
8
l,
an
d
YOL
Ov
8
x
-
ca
ter
in
g
to
d
if
f
er
en
t
a
p
p
licatio
n
s
,
with
YOL
Ov
8
n
o
p
tim
ized
f
o
r
s
p
ee
d
,
ac
h
iev
in
g
8
0
.
4
m
s
o
n
a
C
PU
u
s
in
g
ONNX
an
d
0
.
9
9
m
s
o
n
an
A
1
0
0
GPU
u
s
in
g
T
en
s
o
r
R
T
.
T
r
ain
ed
o
n
th
e
C
OC
O
d
ataset,
wh
ich
in
clu
d
es 8
0
o
b
ject
class
es su
c
h
as c
ar
s
,
m
o
to
r
c
y
cles,
b
u
s
es,
an
d
tr
u
c
k
s
,
YOL
Ov
8
n
is
id
ea
l
f
o
r
r
ea
l
-
tim
e
tr
a
f
f
ic
m
o
n
ito
r
in
g
o
n
lo
w
-
p
o
we
r
d
ev
ices
[
5
3
]
,
[
5
4
]
.
Similar
ly
,
YOL
Ov
9
o
f
f
er
s
f
iv
e
v
ar
ian
ts
-
YO
L
Ov
9
t,
YOL
Ov
9
s
,
YOL
Ov
9
m
,
YOL
Ov
9
c,
an
d
YOL
Ov
9
e
-
d
esig
n
ed
f
o
r
v
ar
i
o
u
s
n
ee
d
s
,
with
YOL
Ov
9
t
b
alan
cin
g
s
p
ee
d
an
d
ac
cu
r
ac
y
,
f
ea
tu
r
in
g
2
M
p
ar
a
m
eter
s
,
7
.
7
B
FLOPs,
an
d
a
m
APv
al
5
0
-
9
5
s
co
r
e
o
f
3
8
.
3
%
[
5
5
]
.
T
h
is
s
tu
d
y
f
o
cu
s
es
o
n
YOL
Ov
8
n
an
d
YOL
Ov
9
t,
s
elec
tin
g
lig
h
tweig
h
t
m
o
d
els
o
p
tim
ized
f
o
r
r
ea
l
-
tim
e
u
s
e
wh
ile
m
ain
tain
in
g
s
u
f
f
icien
t
ac
cu
r
ac
y
,
with
ad
ju
s
ted
d
e
f
au
lt YO
L
O
p
ar
am
eter
s
d
etailed
in
T
a
b
l
e
3
.
T
h
e
d
ef
au
lt
b
atch
s
ize
in
YOL
O
was
r
ed
u
ce
d
f
r
o
m
6
4
to
8
to
m
in
im
ize
o
v
e
r
f
itti
n
g
,
p
ar
ti
cu
lar
ly
o
n
u
n
clea
r
o
r
o
v
er
ly
b
r
i
g
h
t
im
a
g
es,
wh
ile
th
e
p
atien
ce
s
ize
w
as
in
cr
ea
s
ed
f
r
o
m
5
to
1
0
to
p
r
ev
en
t
p
r
em
at
u
r
e
s
to
p
p
in
g
.
T
h
e
d
r
o
p
o
u
t
v
alu
e
was
ad
ju
s
ted
f
r
o
m
0
.
0
to
0
.
2
t
o
en
h
an
ce
r
o
b
u
s
tn
ess
,
an
d
th
e
im
ag
e
s
ize
wa
s
s
e
t
to
6
4
0
×6
4
0
p
i
x
els
to
b
ala
n
ce
s
m
all
o
b
ject
d
etec
tio
n
an
d
i
n
f
er
en
ce
s
p
ee
d
[
5
9
]
-
[
6
2
]
.
T
h
e
an
aly
s
is
p
r
o
ce
s
s
f
o
llo
ws
s
tr
u
ctu
r
ed
s
tag
es,
in
clu
d
in
g
p
r
ep
r
o
ce
s
s
in
g
,
lab
elin
g
,
an
n
o
tatio
n
,
d
ata
au
g
m
en
tatio
n
,
an
d
d
ataset
s
p
li
ttin
g
f
o
r
tr
ain
in
g
an
d
v
alid
atio
n
.
Mo
d
el
tr
ain
i
n
g
u
s
es
8
0
%
o
f
th
e
d
ataset
o
v
e
r
6
4
ep
o
ch
s
with
p
ar
am
eter
s
lis
ted
in
T
ab
le
2
,
wh
ile
v
alid
atio
n
with
th
e
r
em
ain
in
g
2
0
%
ass
ess
es
p
r
ec
is
io
n
,
r
ec
all,
an
d
m
AP.
Fin
ally
,
test
in
g
with
1
0
s
elec
ted
f
r
am
es
ev
alu
ates
th
e
b
est
m
o
d
el
’
s
r
eliab
ilit
y
,
en
s
u
r
i
n
g
r
o
b
u
s
t
r
ea
l
-
wo
r
ld
p
er
f
o
r
m
an
ce
.
T
ab
le
3
.
Ad
ju
s
ted
p
ar
am
eter
s
P
a
r
a
me
t
e
r
V
a
l
u
e
R
e
f
e
r
e
n
c
e
s
B
a
t
c
h
8
[
5
6
]
P
a
t
i
e
n
c
e
10
[
5
7
]
D
r
o
p
o
u
t
0
.
2
[
5
8
]
I
mag
e
s
i
z
e
6
4
0
[
5
9
]
2
.
6
.
P
er
f
o
r
m
a
nce
ev
a
lua
t
io
n m
ea
s
ures
E
v
alu
atin
g
a
class
if
icatio
n
m
o
d
el
r
e
q
u
ir
es
a
co
m
p
r
eh
en
s
i
v
e
an
aly
s
is
u
s
in
g
s
ev
er
al
k
e
y
m
etr
ics.
Pre
cisi
o
n
,
wh
ich
m
ea
s
u
r
es
th
e
ac
cu
r
ac
y
o
f
p
o
s
itiv
e
p
r
e
d
ic
tio
n
s
am
o
n
g
all
p
r
ed
icted
p
o
s
itiv
es,
an
d
r
ec
all,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
1
,
J
u
ly
20
25
:
656
-
6
6
8
662
wh
ich
ev
alu
ates
th
e
p
r
o
p
o
r
tio
n
o
f
co
r
r
ec
tly
id
e
n
tifie
d
p
o
s
itiv
es
o
u
t
o
f
all
ac
t
u
al
p
o
s
itiv
es,
ar
e
cr
itical
m
etr
ics.
Ad
d
itio
n
ally
,
th
e
m
ea
n
av
er
a
g
e
p
r
ec
is
io
n
(
m
AP)
m
et
r
ic
i
s
u
s
ed
to
e
v
alu
ate
d
etec
ted
b
o
u
n
d
in
g
b
o
x
es
b
y
co
m
p
ar
in
g
th
em
to
g
r
o
u
n
d
-
tr
u
th
b
o
x
es
an
d
ass
ig
n
in
g
a
co
r
r
e
s
p
o
n
d
in
g
s
co
r
e
[
5
9
]
.
T
h
e
e
q
u
a
tio
n
s
f
o
r
p
r
ec
is
io
n
,
r
ec
all
,
an
d
m
AP c
an
b
e
s
ee
n
i
n
(
1
)
-
(
3
)
.
=
+
(
1
)
=
+
(
2
)
mAP
=
1
∑
(
)
=
1
(
3
)
W
h
er
e
FP
r
ep
r
esen
ts
(
f
alse p
o
s
itiv
e
)
,
T
N
in
d
icate
s
(
tr
u
e
n
eg
ativ
e
)
,
T
P m
ea
n
s
(
tr
u
e
p
o
s
itiv
e
)
,
an
d
FN
in
d
icate
s
(
f
alse
n
eg
ativ
e
)
.
AP
is
(
av
er
a
g
e
p
r
ec
is
io
n
)
,
AP
j
in
d
icate
s
th
e
av
er
ag
e
p
r
ec
is
io
n
f
o
r
ca
te
g
o
r
y
i,
a
n
d
N
is
th
e
n
u
m
b
er
o
f
class
es.
T
h
is
s
tu
d
y
u
s
es
m
ea
n
p
r
ec
is
io
n
m
A
P5
0
an
d
m
AP5
0
-
9
5
as
t
h
e
p
r
im
ar
y
m
etr
ics
f
o
r
m
ea
s
u
r
in
g
d
etec
tio
n
ac
cu
r
ac
y
.
Flo
atin
g
-
p
o
in
t
o
p
er
atio
n
s
(
GFLO
Ps
)
,
th
e
n
u
m
b
er
o
f
p
a
r
am
eter
s
(
Par
am
s
)
,
an
d
d
etec
tio
n
f
r
am
es
p
e
r
s
ec
o
n
d
(
FP
S)
ar
e
u
s
ed
to
ass
ess
ef
f
icie
n
cy
an
d
r
ea
l
-
tim
e
p
e
r
f
o
r
m
an
c
e.
Ad
d
itio
n
ally
,
th
e
m
o
d
el
’
s
weig
h
t
s
ize
(
Size)
i
s
co
n
s
id
er
ed
t
o
ev
al
u
ate
its
s
u
itab
ilit
y
f
o
r
e
d
g
e
d
ev
ice
i
m
p
lem
en
tatio
n
[
6
3
]
.
T
h
e
eq
u
atio
n
s
f
o
r
m
AP5
0
an
d
m
AP5
0
-
9
5
ar
e
s
h
o
wn
in
(
4
)
a
n
d
(
5
)
.
A
P50
=
1
∑
=
0
.
5
=
1
(
=
0
.
5
)
(
4
)
A
P50
−
95
=
1
10
(
50
+
55
+
⋯
+
95
)
(
5
)
W
h
er
e
r
ep
r
esen
ts
p
r
ec
is
io
n
,
th
e
r
atio
o
f
co
r
r
ec
tly
p
r
ed
ic
ted
p
o
s
itiv
e
s
am
p
les
to
all
p
r
ed
icted
p
o
s
itiv
e
s
am
p
les.
r
ep
r
esen
ts
r
ec
all,
th
e
r
atio
o
f
co
r
r
ec
tly
p
r
ed
icted
p
o
s
itiv
e
s
am
p
les
to
all
ac
tu
al
p
o
s
itiv
e
s
am
p
les.
AP5
0
r
ef
er
s
to
th
e
m
ea
n
AP
ac
r
o
s
s
ca
teg
o
r
ies
w
h
en
th
e
in
ter
s
ec
tio
n
o
v
e
r
u
n
io
n
(
I
o
U)
th
r
esh
o
ld
is
s
et
at
5
0
%.
AP5
0
-
9
5
r
ef
lect
s
th
e
av
er
ag
e
AP
as
th
e
I
o
U
th
r
esh
o
ld
in
c
r
ea
s
es
f
r
o
m
5
0
%
to
9
5
%
in
5
%
in
cr
em
en
ts
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
3
.
1
.
T
ra
ini
ng
v
a
lid
a
t
io
n
T
h
is
s
tu
d
y
tr
ain
ed
an
d
v
alid
a
ted
YOL
Ov
8
n
an
d
YOL
Ov
9
t
u
s
in
g
3
2
s
ce
n
ar
io
s
.
T
h
e
tr
ai
n
in
g
d
ata
in
clu
d
ed
1
n
o
n
-
au
g
m
en
ted
d
ataset
(
5
7
6
s
am
p
les),
1
0
s
in
g
le
-
au
g
m
e
n
tatio
n
d
atasets
(
1
,
7
2
8
s
am
p
les
ea
ch
)
,
an
d
5
m
u
ltip
le
-
au
g
m
en
tatio
n
d
atasets
(
6
,
3
3
6
s
am
p
les
ea
ch
)
.
Valid
atio
n
u
s
ed
1
4
4
s
am
p
les
(
2
0
%
o
f
t
h
e
in
itial
d
ataset)
an
d
m
ea
s
u
r
e
d
p
r
ec
is
i
o
n
,
r
ec
all
,
m
AP5
0
,
an
d
m
AP5
0
-
9
5
,
with
m
AP5
0
-
9
5
as
th
e
p
r
im
ar
y
m
etr
ic
f
o
r
p
er
f
o
r
m
an
ce
ev
al
u
atio
n
.
T
a
b
le
4
p
r
esen
ts
th
e
p
e
r
f
o
r
m
an
ce
o
f
ea
ch
m
o
d
el
with
its
r
esp
ec
tiv
e
au
g
m
en
tatio
n
s
.
T
ab
le
4
s
h
o
ws
th
at
th
e
YOL
O
m
o
d
el
with
o
u
t
au
g
m
en
tatio
n
ac
h
iev
ed
t
h
e
h
ig
h
est
m
AP5
0
-
9
5
(
0
.
3
9
0
)
in
th
e
YOL
Ov
8
n
test
,
wh
ile
th
e
b
est
s
in
g
le
au
g
m
en
tatio
n
p
er
f
o
r
m
a
n
ce
was
with
th
e
b
lu
r
tech
n
iq
u
e
(
0
.
4
6
5
)
,
an
d
m
u
ltip
le
au
g
m
e
n
tatio
n
s
co
m
b
in
in
g
s
ca
lin
g
,
zo
o
m
in
,
b
r
i
g
h
tn
ess
ad
ju
s
tm
en
t,
co
lo
r
jitt
er
,
an
d
n
o
is
e
in
jectio
n
r
ea
ch
ed
th
e
h
ig
h
est m
AP5
0
-
9
5
(
0
.
5
2
6
)
.
T
h
ese
r
esu
lts
co
n
f
ir
m
th
at
au
g
m
en
tatio
n
tech
n
iq
u
es e
n
h
an
c
e
m
o
d
el
p
er
f
o
r
m
a
n
ce
,
with
m
u
ltip
le
au
g
m
en
tatio
n
s
p
r
o
v
id
i
n
g
th
e
b
est
d
etec
tio
n
ac
c
u
r
ac
y
.
I
n
u
r
b
a
n
tr
af
f
ic
du
r
in
g
r
u
s
h
h
o
u
r
s
,
wh
e
r
e
v
eh
icles
ap
p
ea
r
s
m
all
a
n
d
a
r
e
af
f
ec
ted
b
y
s
u
n
lig
h
t
r
ef
le
ctio
n
s
,
YOL
Ov
8
n
o
u
tp
er
f
o
r
m
ed
YOL
Ov
9
t,
d
em
o
n
s
tr
atin
g
h
i
g
h
er
ac
cu
r
ac
y
in
co
m
p
le
x
co
n
d
itio
n
s
.
T
h
e
e
v
a
lu
atio
n
m
etr
ics
f
o
r
ea
ch
class
u
s
in
g
m
u
ltip
le
au
g
m
en
tatio
n
s
ar
e
d
etailed
in
T
ab
le
5
.
T
ab
le
5
p
r
esen
ts
th
e
ev
al
u
atio
n
m
etr
ics
f
o
r
d
if
f
er
en
t
o
b
je
ct
class
es,
s
h
o
win
g
th
at
th
e
‘
bus
’
class
ac
h
iev
es
th
e
h
ig
h
est
m
AP5
0
-
9
5
(
0
.
6
2
6
)
an
d
p
r
ec
is
io
n
(
0
.
9
7
7
)
,
in
d
icatin
g
s
tr
o
n
g
d
etec
tio
n
p
er
f
o
r
m
a
n
ce
with
m
in
im
al
f
alse
d
etec
tio
n
s
d
u
e
to
its
lar
g
e,
ea
s
ily
r
ec
o
g
n
izab
le
f
ea
tu
r
es.
T
h
e
‘
ca
r
’
class
h
a
s
th
e
h
ig
h
est
r
ec
all
(
0
.
7
6
3
)
,
s
u
g
g
esti
n
g
ef
f
e
ctiv
e
d
etec
tio
n
,
lik
ely
b
ec
au
s
e
o
f
it
s
f
r
eq
u
en
t
a
p
p
ea
r
a
n
ce
in
th
e
d
ataset,
wh
ile
also
lead
in
g
in
m
AP5
0
(
0
.
8
5
)
,
r
ein
f
o
r
cin
g
its
r
eliab
ilit
y
at
an
I
o
U
th
r
esh
o
ld
o
f
0
.
5
0
.
C
o
n
v
e
r
s
ely
,
th
e
‘
m
o
to
r
c
y
cle
’
class
h
as
th
e
lo
west
m
AP5
0
-
9
5
(
0
.
3
5
9
)
an
d
r
elativ
ely
lo
w
p
r
ec
is
io
n
(
0
.
8
3
1
)
,
in
d
icati
n
g
h
ig
h
er
d
etec
tio
n
er
r
o
r
s
,
lik
ely
d
u
e
to
its
s
m
all
er
s
ize
an
d
r
ef
lectio
n
s
.
Fig
u
r
e
5
h
ig
h
lig
h
ts
co
m
m
o
n
m
is
class
if
icatio
n
s
,
wi
th
m
o
to
r
cy
cles
o
f
te
n
m
is
tak
en
f
o
r
ca
r
s
(
1
,
3
0
3
tim
es)
an
d
ca
r
s
m
is
id
en
tifie
d
as
b
ac
k
g
r
o
u
n
d
(
3
9
4
tim
es).
Desp
ite
st
r
o
n
g
o
v
er
all
p
er
f
o
r
m
an
ce
,
th
e
m
o
d
el
n
ee
d
s
im
p
r
o
v
em
e
n
ts
in
h
a
n
d
lin
g
s
m
all
o
b
jects
an
d
c
h
allen
g
in
g
lig
h
tin
g
co
n
d
itio
n
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
B
o
o
s
tin
g
r
ea
l
-
time
ve
h
icle
d
et
ec
tio
n
in
u
r
b
a
n
tr
a
ffic u
s
in
g
a
n
o
ve
l m
u
lti
-
a
u
g
men
ta
tio
n
(
I
m
a
m
A
h
ma
d
A
s
h
a
r
i
)
663
T
ab
le
4
.
C
o
m
p
a
r
is
o
n
o
f
p
er
f
o
r
m
an
ce
m
etr
ics o
f
th
e
test
ed
m
o
d
els
No
A
u
g
m
e
n
t
a
t
i
o
n
M
o
d
e
l
s
B
e
st
e
p
o
c
h
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
mA
P
5
0
mA
P
5
0
-
95
1
W
i
t
h
o
u
t
a
u
g
me
n
t
a
t
i
o
n
Y
O
LO
v
8
n
64
0
.
8
1
9
0
.
5
5
0
0
.
6
7
3
0
.
3
9
0
2
W
i
t
h
o
u
t
a
u
g
me
n
t
a
t
i
o
n
Y
O
LO
v
9
t
64
0
.
8
2
5
0
.
5
5
6
0
.
6
6
9
0
.
3
7
8
3
B
l
u
r
Y
O
LO
v
8
n
64
0
.
8
2
5
0
.
6
7
4
0
.
7
6
7
0
.
4
6
5
4
B
l
u
r
Y
O
LO
v
9
t
64
0
.
8
3
0
0
.
6
4
2
0
.
7
4
6
0
.
4
4
3
5
B
r
i
g
h
t
n
e
ss
a
d
j
u
s
t
me
n
t
Y
O
LO
v
8
n
64
0
.
7
7
4
0
.
7
0
1
0
.
7
5
7
0
.
4
6
2
6
B
r
i
g
h
t
n
e
ss
a
d
j
u
s
t
me
n
t
Y
O
LO
v
9
t
62
0
.
8
8
0
0
.
6
1
9
0
.
7
4
9
0
.
4
4
9
7
C
o
n
t
r
a
st
a
d
j
u
st
m
e
n
t
Y
O
LO
v
8
n
49
0
.
5
9
4
0
.
0
7
1
0
.
3
2
5
0
.
2
0
3
8
C
o
n
t
r
a
st
a
d
j
u
st
m
e
n
t
Y
O
LO
v
9
t
10
0
.
4
8
1
0
.
0
9
7
0
.
2
8
0
0
.
1
6
7
9
C
o
l
o
r
j
i
t
t
e
r
Y
O
LO
v
8
n
61
0
.
8
6
1
0
.
6
4
8
0
.
7
5
5
0
.
4
6
1
10
C
o
l
o
r
j
i
t
t
e
r
Y
O
LO
v
9
t
46
0
.
8
7
8
0
.
5
6
5
0
.
7
1
3
0
.
4
1
1
11
C
r
o
p
p
i
n
g
Y
O
LO
v
8
n
60
0
.
7
8
4
0
.
5
4
0
0
.
6
4
6
0
.
3
5
6
12
C
r
o
p
p
i
n
g
Y
O
LO
v
9
t
64
0
.
7
4
0
0
.
5
5
9
0
.
6
3
5
0
.
3
4
9
13
F
l
i
p
p
i
n
g
Y
O
LO
v
8
n
64
0
.
8
5
3
0
.
6
0
3
0
.
7
3
2
0
.
4
3
6
14
F
l
i
p
p
i
n
g
Y
O
LO
v
9
t
62
0
.
7
8
5
0
.
6
3
3
0
.
7
1
9
0
.
4
1
1
15
N
o
i
se
i
n
j
e
c
t
i
o
n
Y
O
LO
v
8
n
51
0
.
8
7
0
0
.
6
1
9
0
.
7
3
9
0
.
4
3
3
16
N
o
i
se
i
n
j
e
c
t
i
o
n
Y
O
LO
v
9
t
64
0
.
8
6
4
0
.
6
2
5
0
.
7
1
5
0
.
4
2
2
17
R
o
t
a
t
i
o
n
Y
O
LO
v
8
n
62
0
.
7
7
9
0
.
5
5
9
0
.
6
6
3
0
.
3
7
1
18
R
o
t
a
t
i
o
n
Y
O
LO
v
9
t
63
0
.
7
6
6
0
.
5
4
5
0
.
6
5
3
0
.
3
6
6
19
S
c
a
l
i
n
g
Y
O
LO
v
8
n
61
0
.
8
3
6
0
.
6
6
9
0
.
7
5
3
0
.
4
6
4
20
S
c
a
l
i
n
g
Y
O
LO
v
9
t
64
0
.
8
1
2
0
.
6
6
3
0
.
7
4
1
0
.
4
5
9
21
Zo
o
m
-
In
Y
O
LO
v
8
n
64
0
.
7
8
6
0
.
6
4
0
0
.
7
2
2
0
.
4
2
6
22
Zo
o
m
-
In
Y
O
LO
v
9
t
64
0
.
7
9
2
0
.
6
2
5
0
.
7
1
6
0
.
4
2
1
23
S
c
a
l
i
n
g
+
C
r
o
p
p
i
n
g
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
st
me
n
t
+
N
o
i
se
i
n
j
e
c
t
i
o
n
+
B
l
u
r
Y
O
LO
v
8
n
63
0
.
8
9
3
0
.
6
7
7
0
.
7
7
7
0
.
5
1
1
24
S
c
a
l
i
n
g
+
C
r
o
p
p
i
n
g
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
st
me
n
t
+
N
o
i
se
i
n
j
e
c
t
i
o
n
+
B
l
u
r
Y
O
LO
v
9
t
64
0
.
8
8
8
0
.
6
8
1
0
.
7
8
4
0
.
5
0
3
25
R
o
t
a
t
i
o
n
+
F
l
i
p
p
i
n
g
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
st
me
n
t
+
C
o
n
t
r
a
s
t
a
d
j
u
s
t
me
n
t
+
C
o
l
o
r
j
i
t
t
e
r
Y
O
LO
v
8
n
64
0
.
8
6
7
0
.
6
5
2
0
.
7
6
2
0
.
4
9
1
26
R
o
t
a
t
i
o
n
+
F
l
i
p
p
i
n
g
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
st
me
n
t
+
C
o
n
t
r
a
s
t
a
d
j
u
s
t
me
n
t
+
C
o
l
o
r
j
i
t
t
e
r
Y
O
LO
v
9
t
63
0
.
8
9
0
0
.
6
5
1
0
.
7
6
8
0
.
4
8
0
27
S
c
a
l
i
n
g
+
Z
o
o
m
i
n
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
s
t
me
n
t
+
C
o
l
o
r
j
i
t
t
e
r
+
N
o
i
se
i
n
j
e
c
t
i
o
n
Y
O
LO
v
8
n
61
0
.
8
7
2
0
.
7
1
5
0
.
7
9
2
0
.
5
2
6
28
S
c
a
l
i
n
g
+
Z
o
o
m
i
n
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
s
t
me
n
t
+
C
o
l
o
r
j
i
t
t
e
r
+
N
o
i
se
i
n
j
e
c
t
i
o
n
Y
O
LO
v
9
t
64
0
.
8
3
8
0
.
6
9
5
0
.
7
7
6
0
.
5
0
6
29
C
r
o
p
p
i
n
g
+
Z
o
o
m
i
n
+
C
o
n
t
r
a
s
t
a
d
j
u
s
t
men
t
+
N
o
i
se
i
n
j
e
c
t
i
o
n
+
B
l
u
r
Y
O
LO
v
8
n
63
0
.
8
7
2
0
.
6
6
8
0
.
7
7
4
0
.
4
9
8
30
C
r
o
p
p
i
n
g
+
Z
o
o
m
i
n
+
C
o
n
t
r
a
s
t
a
d
j
u
s
t
men
t
+
N
o
i
se
i
n
j
e
c
t
i
o
n
+
B
l
u
r
Y
O
LO
v
9
t
62
0
.
8
3
5
0
.
6
4
5
0
.
7
4
7
0
.
4
6
9
31
S
c
a
l
i
n
g
+
R
o
t
a
t
i
o
n
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
s
t
me
n
t
+
C
o
n
t
r
a
st
a
d
j
u
st
m
e
n
t
+
N
o
i
se
i
n
j
e
c
t
i
o
n
Y
O
LO
v
8
n
64
0
.
8
8
2
0
.
6
7
0
0
.
7
8
4
0
.
5
1
1
32
S
c
a
l
i
n
g
+
R
o
t
a
t
i
o
n
+
B
r
i
g
h
t
n
e
ss
a
d
j
u
s
t
me
n
t
+
C
o
n
t
r
a
st
a
d
j
u
st
m
e
n
t
+
N
o
i
se
i
n
j
e
c
t
i
o
n
Y
O
LO
v
9
t
63
0
.
8
9
7
0
.
6
4
3
0
.
7
7
7
0
.
4
9
3
Fig
u
r
e
5
.
C
o
n
f
u
s
io
n
m
atr
i
x
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
1
,
J
u
ly
20
25
:
656
-
6
6
8
664
3
.
2
.
M
o
del
t
esting
T
h
e
aim
o
f
th
is
test
in
g
is
t
o
e
v
alu
ate
th
e
d
ev
el
o
p
ed
m
o
d
el
t
o
d
eter
m
in
e
wh
et
h
er
it
wo
r
k
s
ef
f
icien
tly
an
d
ac
cu
r
ately
in
co
u
n
tin
g
a
n
d
class
if
y
in
g
ty
p
es
o
f
v
eh
icle
s
in
tr
af
f
ic
with
m
u
lti
-
lab
els.
I
n
th
is
ex
p
er
im
en
t,
we
u
s
e
th
e
YOL
O
f
r
am
ewo
r
k
an
d
m
u
ltip
le
au
g
m
en
tatio
n
s
,
wh
ich
ac
h
iev
e
d
th
e
h
ig
h
est
v
a
lu
es
in
th
e
tr
ai
n
in
g
d
ata.
T
h
e
1
0
test
in
g
s
ce
n
ar
io
s
ar
e
lis
ted
in
T
ab
le
6
.
T
ab
le
6
ev
alu
ates
th
e
o
b
ject
d
etec
tio
n
m
o
d
el
in
v
ar
io
u
s
tr
af
f
ic
s
ce
n
ar
io
s
,
in
clu
d
in
g
co
n
g
ested
tr
af
f
i
c
with
lig
h
t
r
ef
lectio
n
,
o
p
tim
al
l
ig
h
tin
g
,
s
m
o
o
th
tr
a
f
f
ic
with
li
g
h
t
r
ef
lectio
n
,
an
d
r
ai
n
,
to
ass
ess
its
ad
ap
tab
ilit
y
to
r
ea
l
-
wo
r
ld
c
h
allen
g
es
lik
e
d
etec
ti
o
n
er
r
o
r
s
f
r
o
m
lig
h
t
r
ef
lectio
n
s
o
r
r
e
d
u
ce
d
v
is
ib
ilit
y
d
u
e
to
r
ain
.
B
y
s
im
u
latin
g
th
ese
co
n
d
itio
n
s
,
r
esear
ch
er
s
ca
n
ev
alu
ate
th
e
m
o
d
el
’
s
r
esil
ien
ce
an
d
f
u
n
ctio
n
ality
u
n
d
e
r
d
if
f
er
en
t
wea
th
er
a
n
d
lig
h
tin
g
co
n
d
itio
n
s
.
T
esti
n
g
ac
r
o
s
s
all
v
eh
icle
class
es
en
s
u
r
es
ac
cu
r
ate
d
etec
tio
n
,
ev
en
f
o
r
s
m
all
o
b
jects,
h
elp
in
g
to
id
en
tify
wea
k
n
ess
es
an
d
im
p
r
o
v
e
th
e
m
o
d
el
’
s
r
eliab
ilit
y
an
d
ac
cu
r
ac
y
in
r
ea
l
-
wo
r
ld
ap
p
licatio
n
s
.
T
h
e
test
r
e
s
u
lts
ar
e
s
u
m
m
ar
ized
in
T
ab
le
7
.
T
h
e
m
AP
ca
lcu
latio
n
s
f
o
r
e
ac
h
test
in
g
s
ce
n
ar
io
ar
e
illu
s
tr
ated
in
th
e
g
r
a
p
h
.
T
h
e
g
r
a
p
h
clea
r
ly
d
em
o
n
s
tr
ates
h
o
w
th
e
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
v
a
r
ies
ac
r
o
s
s
d
if
f
er
en
t
tr
af
f
ic
an
d
lig
h
tin
g
co
n
d
itio
n
s
.
T
h
e
m
AP
ca
lcu
latio
n
s
f
o
r
ea
ch
test
in
g
s
ce
n
ar
io
ar
e
p
r
esen
ted
in
Fig
u
r
e
6
.
T
ab
le
5
.
E
v
alu
atio
n
m
et
r
ics cla
s
s
f
r
o
m
Yo
lo
v
8
n
u
s
in
g
m
u
lti
-
au
g
m
en
tatio
n
C
l
a
s
s
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
mA
P
5
0
mA
P
5
0
-
59
A
l
l
0
.
8
7
2
0
.
7
1
5
0
.
7
9
2
0
.
5
2
6
B
u
s
0
.
9
7
7
0
.
7
8
8
0
.
8
5
0
0
.
6
2
6
Car
0
.
8
6
6
0
.
7
6
3
0
.
8
3
4
0
.
5
3
8
M
o
t
o
r
c
y
c
l
e
0
.
8
3
1
0
.
5
5
9
0
.
6
7
2
0
.
3
5
9
Tr
u
c
k
0
.
8
1
4
0
.
7
4
9
0
.
8
1
0
0
.
5
8
2
T
ab
le
6
.
Mo
d
el
test
in
g
s
ce
n
ar
i
o
s
No
C
o
n
d
i
t
i
o
n
1
C
o
n
g
e
st
e
d
t
r
a
f
f
i
c
w
i
t
h
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
2
C
o
n
g
e
st
e
d
t
r
a
f
f
i
c
w
i
t
h
p
r
o
p
e
r
l
y
e
x
p
o
s
e
d
3
S
mo
o
t
h
t
r
a
f
f
i
c
w
i
t
h
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
4
S
mo
o
t
h
t
r
a
f
f
i
c
w
i
t
h
p
r
o
p
e
r
l
y
e
x
p
o
se
d
5
C
o
n
g
e
st
e
d
t
r
a
f
f
i
c
i
n
r
a
i
n
6
S
mo
o
t
h
t
r
a
f
f
i
c
i
n
r
a
i
n
7
Tr
a
f
f
i
c
w
i
t
h
a
l
l
v
e
h
i
c
l
e
c
l
a
sses
,
C
o
n
g
e
st
e
d
w
i
t
h
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
8
Tr
a
f
f
i
c
w
i
t
h
a
l
l
v
e
h
i
c
l
e
c
l
a
sses
,
C
o
n
g
e
st
e
d
w
i
t
h
p
r
o
p
e
r
l
y
e
x
p
o
se
d
9
Tr
a
f
f
i
c
w
i
t
h
a
l
l
v
e
h
i
c
l
e
c
l
a
sses
,
S
m
o
o
t
h
w
i
t
h
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
10
Tr
a
f
f
i
c
w
i
t
h
a
l
l
v
e
h
i
c
l
e
c
l
a
sses
,
S
m
o
o
t
h
w
i
t
h
p
r
o
p
e
r
l
y
e
x
p
o
se
d
T
ab
le
7
.
Mo
d
el
test
in
g
r
esu
lts
No
C
o
n
d
i
t
i
o
n
Te
st
r
e
su
l
t
s
1
C
o
n
g
e
st
e
d
t
r
a
f
f
i
c
w
i
t
h
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
4
1
c
a
r
s
,
3
9
mo
t
o
r
c
y
c
l
e
s,
4
t
r
u
c
k
s,
0
b
u
ses
2
C
o
n
g
e
st
e
d
t
r
a
f
f
i
c
w
i
t
h
p
r
o
p
e
r
l
y
e
x
p
o
s
e
d
3
6
c
a
r
s
,
1
0
7
m
o
t
o
r
c
y
c
l
e
s
,
1
t
r
u
c
k
,
0
b
u
ses
3
S
mo
o
t
h
t
r
a
f
f
i
c
w
i
t
h
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
6
c
a
r
s,
2
6
m
o
t
o
r
c
y
c
l
e
s,
0
t
r
u
c
k
s,
0
b
u
s
e
s
4
S
mo
o
t
h
t
r
a
f
f
i
c
w
i
t
h
p
r
o
p
e
r
l
y
e
x
p
o
se
d
9
c
a
r
s,
2
0
m
o
t
o
r
c
y
c
l
e
s,
0
t
r
u
c
k
s,
0
b
u
s
e
s
5
C
o
n
g
e
st
e
d
t
r
a
f
f
i
c
i
n
r
a
i
n
4
3
c
a
r
s
,
8
6
mo
t
o
r
c
y
c
l
e
s,
7
t
r
u
c
k
s,
0
b
u
ses
6
S
mo
o
t
h
t
r
a
f
f
i
c
i
n
r
a
i
n
5
c
a
r
s,
9
m
o
t
o
r
c
y
c
l
e
s
,
0
t
r
u
c
k
s
,
0
b
u
s
e
s
7
Tr
a
f
f
i
c
w
i
t
h
a
l
l
v
e
h
i
c
l
e
c
l
a
sses
,
C
o
n
g
e
st
e
d
w
i
t
h
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
1
5
c
a
r
s
,
3
7
mo
t
o
r
c
y
c
l
e
s,
9
t
r
u
c
k
s,
0
b
u
ses
8
Tr
a
f
f
i
c
w
i
t
h
a
l
l
v
e
h
i
c
l
e
c
l
a
sses
,
C
o
n
g
e
st
e
d
w
i
t
h
p
r
o
p
e
r
l
y
e
x
p
o
se
d
1
5
c
a
r
s
,
5
7
mo
t
o
r
c
y
c
l
e
s,
3
t
r
u
c
k
s,
0
b
u
ses
9
Tr
a
f
f
i
c
w
i
t
h
a
l
l
v
e
h
i
c
l
e
c
l
a
sses
,
S
m
o
o
t
h
w
i
t
h
l
i
g
h
t
r
e
f
l
e
c
t
i
o
n
1
0
c
a
r
s
,
1
6
mo
t
o
r
c
y
c
l
e
s,
2
t
r
u
c
k
s,
1
b
u
s
10
Tr
a
f
f
i
c
w
i
t
h
a
l
l
v
e
h
i
c
l
e
c
l
a
sses
,
S
m
o
o
t
h
w
i
t
h
p
r
o
p
e
r
l
y
e
x
p
o
se
d
6
c
a
r
s,
2
1
m
o
t
o
r
c
y
c
l
e
s,
4
t
r
u
c
k
s,
1
b
u
s
T
h
e
test
r
esu
lts
in
d
icate
th
a
t
th
e
m
o
d
el
’
s
d
etec
tio
n
p
e
r
f
o
r
m
an
ce
v
ar
ies
d
e
p
en
d
in
g
o
n
tr
af
f
ic
co
n
d
itio
n
s
.
I
n
d
en
s
e
tr
af
f
ic
with
lig
h
t
r
ef
lectio
n
s
,
th
e
m
AP
was
0
.
4
6
4
,
in
d
icatin
g
th
at
lig
h
t
r
ef
lectio
n
s
m
ak
e
o
b
ject
d
etec
tio
n
m
o
r
e
ch
all
en
g
in
g
.
Un
d
er
o
p
tim
al
lig
h
tin
g
co
n
d
itio
n
s
,
th
e
m
AP
in
cr
ea
s
ed
to
0
.
6
2
3
,
s
u
g
g
esti
n
g
th
at
b
etter
lig
h
tin
g
en
h
an
ce
s
d
etec
tio
n
ac
cu
r
ac
y
.
I
n
s
m
o
o
th
tr
af
f
ic
with
lig
h
t
r
ef
lectio
n
s
,
th
e
m
AP
in
cr
ea
s
e
d
to
0
.
8
2
5
,
an
d
u
n
d
e
r
s
m
o
o
th
tr
af
f
ic
with
o
p
tim
al
lig
h
tin
g
,
th
e
m
AP r
ea
c
h
ed
0
.
9
3
2
,
d
em
o
n
s
tr
atin
g
t
h
e
b
est p
er
f
o
r
m
an
ce
.
Du
r
in
g
r
ain
a
n
d
d
e
n
s
e
tr
af
f
ic,
th
e
m
AP
d
r
o
p
p
e
d
to
0
.
7
2
4
,
i
n
d
icatin
g
th
at
r
ain
d
im
in
is
h
es
d
etec
tio
n
ac
cu
r
ac
y
.
I
n
s
m
o
o
th
tr
af
f
ic
d
u
r
in
g
r
ai
n
,
th
e
m
AP
r
o
s
e
to
0
.
8
1
7
.
I
n
d
e
n
s
e
co
n
d
itio
n
s
with
lig
h
t
r
ef
lectio
n
s
an
d
v
ar
io
u
s
ty
p
es
o
f
v
eh
icles,
th
e
m
AP
was
0
.
8
8
7
,
s
u
g
g
esti
n
g
th
at
th
e
m
o
d
el
s
till
p
er
f
o
r
m
ed
well
in
th
ese
co
n
d
itio
n
s
.
Un
d
e
r
o
p
tim
al
lig
h
tin
g
,
th
e
m
AP
s
lig
h
tly
d
ec
r
e
ased
to
0
.
8
0
9
,
in
d
icati
n
g
th
at
v
eh
icle
co
m
p
lex
ity
af
f
ec
ts
p
er
f
o
r
m
a
n
ce
.
I
n
s
m
o
o
th
tr
af
f
ic
with
v
ar
io
u
s
v
eh
icl
e
ty
p
es
an
d
lig
h
t
r
ef
lectio
n
s
,
th
e
m
AP
r
ea
ch
ed
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
B
o
o
s
tin
g
r
ea
l
-
time
ve
h
icle
d
et
ec
tio
n
in
u
r
b
a
n
tr
a
ffic u
s
in
g
a
n
o
ve
l m
u
lti
-
a
u
g
men
ta
tio
n
(
I
m
a
m
A
h
ma
d
A
s
h
a
r
i
)
665
0
.
8
1
8
,
an
d
u
n
d
e
r
o
p
tim
al
lig
h
tin
g
,
it
i
n
cr
ea
s
ed
to
0
.
8
9
.
O
v
er
all,
th
e
b
est
p
er
f
o
r
m
a
n
ce
was
ac
h
iev
ed
u
n
d
er
o
p
tim
al
lig
h
tin
g
a
n
d
s
m
o
o
th
tr
af
f
ic,
wh
ile
r
ain
o
r
lig
h
t
r
ef
lec
tio
n
s
r
ed
u
ce
d
ac
cu
r
ac
y
.
Fig
u
r
e
6
.
m
AP
ca
lcu
latio
n
f
o
r
v
ar
io
u
s
tr
af
f
ic
a
n
d
lig
h
tin
g
co
n
d
itio
n
s
3
.
3
.
Cha
lleng
es
T
h
e
p
r
im
ar
y
ch
allen
g
es
in
o
b
ject
d
etec
tio
n
with
in
t
h
is
d
ataset
ar
e
th
e
v
ar
iatio
n
s
in
lig
h
tin
g
co
n
d
itio
n
s
an
d
t
r
af
f
ic
d
e
n
s
ity
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
o
b
ject
d
et
ec
tio
n
v
ar
ies
s
ig
n
if
ican
tly
d
e
p
en
d
in
g
o
n
wh
eth
er
th
e
im
ag
es
ar
e
af
f
ec
ted
b
y
lig
h
t
r
ef
lectio
n
s
o
r
a
r
e
p
r
o
p
er
ly
ex
p
o
s
ed
.
Ob
ject
d
etec
ti
o
n
b
ec
o
m
es
m
o
r
e
ch
allen
g
in
g
u
n
d
er
s
u
b
o
p
tim
al
lig
h
tin
g
o
r
d
is
r
u
p
tiv
e
lig
h
t
r
e
f
lectio
n
s
,
as we
ll a
s
in
co
n
g
ested
tr
af
f
ic
co
n
d
itio
n
s
co
m
p
ar
ed
to
s
m
o
o
th
e
r
tr
a
f
f
ic.
T
h
is
s
u
g
g
ests
th
at
p
o
o
r
lig
h
t
in
g
an
d
h
ig
h
tr
af
f
ic
d
e
n
s
ity
h
in
d
er
t
h
e
m
o
d
el
’
s
ab
ilit
y
to
ac
cu
r
ately
d
etec
t o
b
j
ec
ts
.
Mo
r
eo
v
er
,
th
e
p
r
esen
ce
o
f
al
l
o
b
ject
class
es
in
a
s
in
g
le
i
m
ag
e
f
u
r
th
er
c
o
m
p
licates
th
e
d
etec
tio
n
ch
allen
g
e.
I
m
ag
es
co
n
tain
in
g
all
ty
p
es
o
f
o
b
jects
(
ca
r
s
,
m
o
t
o
r
cy
cles,
tr
u
c
k
s
,
an
d
b
u
s
es)
s
h
o
w
th
at
th
e
m
o
d
el
s
tr
u
g
g
les
to
id
en
tify
an
d
d
if
f
er
en
tiate
o
b
jects
wh
en
t
h
er
e
ar
e
m
u
ltip
le
o
v
er
lap
p
in
g
cl
ass
es.
Sp
ec
if
ically
,
class
es
s
u
ch
as
b
u
s
es
an
d
tr
u
ck
s
ex
h
ib
it
lo
wer
d
etec
tio
n
p
er
f
o
r
m
a
n
ce
co
m
p
ar
ed
to
ca
r
s
an
d
m
o
to
r
cy
cles,
s
u
g
g
esti
n
g
th
at
th
e
m
o
d
el
m
ay
b
e
less
ef
f
ec
tiv
e
in
d
etec
tin
g
less
co
m
m
o
n
o
b
jects
o
r
th
o
s
e
with
less
d
is
tin
ctiv
e
v
is
u
al
f
ea
tu
r
es.
I
m
p
r
o
v
e
m
en
t
ef
f
o
r
ts
s
h
o
u
ld
co
n
ce
n
tr
ate
o
n
ad
ju
s
tin
g
f
o
r
lig
h
tin
g
v
ar
iat
io
n
s
,
ac
co
m
m
o
d
atin
g
d
if
f
e
r
en
t
r
o
a
d
co
n
d
itio
n
s
,
an
d
en
h
a
n
cin
g
th
e
m
o
d
el
’
s
ab
ilit
y
to
d
etec
t
m
o
r
e
ch
allen
g
in
g
class
es.
De
s
p
ite
th
ese
ch
allen
g
es,
th
e
p
r
o
p
o
s
ed
m
o
d
el
ex
h
i
b
its
r
ea
s
o
n
ab
ly
g
o
o
d
p
er
f
o
r
m
an
c
e.
4.
CO
NCLU
SI
O
N
T
h
is
s
tu
d
y
s
u
cc
ess
f
u
lly
d
em
o
n
s
tr
ates
th
at
u
s
in
g
YOL
O
a
n
d
m
u
lti
-
au
g
m
en
tatio
n
tech
n
iq
u
es
f
o
r
r
ea
l
-
tim
e
v
eh
icle
d
etec
tio
n
i
n
u
r
b
an
tr
af
f
ic
ca
n
en
h
a
n
ce
ac
cu
r
a
cy
an
d
e
f
f
icien
cy
i
n
id
en
tif
y
in
g
an
d
class
if
y
in
g
v
ar
io
u
s
ty
p
es o
f
v
eh
icles.
T
h
e
u
tili
za
tio
n
o
f
m
u
lti
-
lab
el
im
ag
es e
n
ab
les th
e
d
etec
tio
n
o
f
m
u
ltip
le
v
eh
icle
ty
p
es
with
in
a
s
in
g
le
f
r
am
e
,
o
f
f
er
i
n
g
a
m
o
r
e
c
o
m
p
r
e
h
en
s
iv
e
v
iew
o
f
tr
af
f
ic
co
n
d
itio
n
s
.
I
m
p
lem
en
tin
g
m
u
lti
-
au
g
m
en
tatio
n
in
cr
ea
s
es
th
e
d
i
v
er
s
ity
o
f
th
e
tr
ain
in
g
d
ata,
m
ak
in
g
th
e
m
o
d
el
m
o
r
e
r
o
b
u
s
t
to
v
ar
y
i
n
g
lig
h
ti
n
g
co
n
d
itio
n
s
,
an
g
les,
an
d
o
b
ject
o
b
s
tr
u
ctio
n
s
.
T
h
e
test
r
esu
lts
in
d
icate
th
at
YOL
Ov
8
n
with
m
u
lti
-
au
g
m
en
tatio
n
(
s
ca
lin
g
,
z
o
o
m
in
,
b
r
ig
h
tn
ess
ad
ju
s
tm
en
t,
co
lo
r
jitt
er
,
an
d
n
o
is
e
in
jectio
n
)
d
eliv
er
s
th
e
b
est
p
er
f
o
r
m
an
ce
,
ac
h
iev
i
n
g
a
p
r
ec
is
io
n
o
f
0
.
8
7
2
,
r
ec
all
o
f
0
.
7
1
5
,
m
AP5
0
o
f
0
.
7
9
2
,
an
d
m
A
P5
0
-
9
5
o
f
0
.
5
2
6
,
s
u
r
p
ass
in
g
o
th
e
r
au
g
m
en
tatio
n
tech
n
iq
u
es.
T
r
a
f
f
ic
an
aly
s
is
s
h
o
ws
th
at
o
p
tim
al
lig
h
tin
g
an
d
s
m
o
o
th
tr
af
f
ic
y
ield
th
e
b
est
d
etec
tio
n
p
e
r
f
o
r
m
a
n
ce
,
with
a
m
ax
im
u
m
m
AP
o
f
0
.
9
3
2
,
wh
er
ea
s
p
o
o
r
lig
h
tin
g
(
g
lar
e
)
an
d
h
ig
h
tr
af
f
ic
d
en
s
ity
r
ed
u
ce
d
ete
ctio
n
ac
cu
r
ac
y
,
r
esu
ltin
g
in
a
m
in
im
u
m
m
AP
o
f
0
.
4
6
4
.
I
t
was
also
s
h
o
wn
t
h
at
m
u
lti
-
au
g
m
en
tatio
n
s
ig
n
if
i
ca
n
tly
o
u
tp
er
f
o
r
m
s
s
in
g
le
au
g
m
en
tatio
n
a
n
d
n
o
au
g
m
en
tatio
n
.
I
n
th
is
s
tu
d
y
,
YOL
Ov
8
n
ex
h
ib
ited
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
o
v
er
YOL
Ov
9
t,
p
ar
ticu
lar
ly
u
n
d
e
r
co
m
p
lex
tr
a
f
f
ic
co
n
d
itio
n
s
an
d
ch
allen
g
in
g
lig
h
tin
g
e
n
v
ir
o
n
m
en
ts
.
ACK
NO
WL
E
DG
E
M
E
NT
S
W
e
wo
u
ld
lik
e
t
o
ex
p
r
ess
o
u
r
g
r
atitu
d
e
to
t
h
e
Sem
ar
an
g
C
ity
T
r
an
s
p
o
r
tatio
n
De
p
ar
tm
en
t
f
o
r
th
ei
r
ass
is
tan
ce
an
d
s
u
p
p
o
r
t in
p
r
o
v
id
in
g
th
e
v
alu
ab
le
d
ataset
u
s
ed
in
th
is
r
esear
ch
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.