I
n
d
on
e
s
i
an
Jo
u
r
n
al
o
f
El
e
c
t
r
i
c
al
En
gi
n
e
e
r
i
n
g
an
d
C
o
m
p
u
te
r
S
c
i
e
n
c
e
V
o
l
.
22
,
N
o
.
1
,
A
p
r
i
l
2021
,
p
p.
222
~
231
IS
S
N
:
25
02
-
4752
,
D
O
I
:
10.
1
1591
/
i
j
e
e
c
s
.
v
22
.i
1
.
pp
222
-
231
222
Jou
r
n
al
h
o
m
e
pa
ge
:
ht
t
p:
/
/
i
j
e
e
c
s
.
i
a
e
s
c
or
e
.
c
om
E
y
e
b
l
i
n
k
d
e
t
e
c
t
i
o
n
u
s
i
n
g
CNN
to
d
e
t
e
c
t
d
r
o
w
s
i
n
e
ss
l
e
v
e
l
in
d
r
i
v
e
r
s
f
o
r
r
o
a
d
s
a
f
e
t
y
P
o
th
u
r
aj
u
V
i
s
h
e
s
h
1
,
R
agh
av
e
n
d
r
a
S
2
,
S
an
to
s
h
K
u
m
ar
Jan
k
att
i
3
,
R
e
k
h
a
V
4
1,
2
,
4
D
e
pa
r
t
m
e
n
t
of
C
o
m
put
e
r
S
c
i
e
nc
e
a
n
d
E
ng
i
ne
e
r
i
ng
,
C
H
R
I
S
T
D
e
e
m
e
d
to
be
U
n
i
v
e
r
s
i
t
y
,
I
ndi
a
3
D
e
pa
r
t
m
e
n
t
of
C
o
m
put
e
r
S
c
i
e
nc
e
a
nd
E
ng
i
n
e
e
r
i
ng
,
K
S
S
E
M
,
V
T
U
,
I
ndi
a
A
r
ti
c
l
e
I
n
fo
A
B
S
TR
A
C
T
Ar
t
i
c
l
e
h
i
s
t
or
y
:
R
e
c
e
i
v
e
d
O
c
t
30
,
2020
R
e
v
i
s
e
d
J
a
n
17
,
202
1
A
c
c
e
pt
e
d
M
a
r
3
,
202
1
B
l
i
n
ki
ng
is
a
r
e
g
ul
a
r
bo
di
l
y
f
unc
t
i
o
n
a
nd
it
is
t
he
s
e
m
i
a
u
t
o
m
a
t
i
c
f
a
s
t
c
l
o
s
i
ng
of
t
he
e
y
e
l
i
d.
A
s
pe
c
i
f
i
c
bl
i
nk
is
e
xa
m
i
ne
d
by
d
y
na
m
i
c
f
o
l
di
ng
of
t
he
e
y
e
l
i
d
.
It
is
a
v
i
t
a
l
f
unc
t
i
o
n
of
t
h
e
e
y
e
w
hi
c
h
he
l
ps
in
s
pr
e
a
d
of
t
e
a
r
s
a
c
r
o
s
s
a
nd
e
l
i
m
i
n
a
t
e
s
i
r
r
i
t
a
n
t
s
f
r
o
m
t
he
s
ha
l
l
o
w
of
c
o
r
ne
a
.
In
t
hi
s
r
e
s
e
a
r
c
h
w
o
r
k
we
m
a
de
us
e
of
c
o
nvo
l
ut
i
o
n
ne
ur
a
l
n
e
t
w
o
r
k
,
t
h
e
de
e
p
l
e
a
r
n
i
ng
c
o
n
c
e
pt
s
a
nd
i
m
a
g
e
pr
o
c
e
s
s
i
ng
to
de
t
e
c
t
d
r
o
w
s
i
n
e
s
s
l
e
v
e
l
in
dr
i
v
e
r
s
.
To
t
r
a
i
n
t
he
bl
i
nk
de
t
e
c
t
i
o
n
m
o
de
l
t
h
e
m
o
bi
l
e
n
e
t
V2
is
us
e
d
as
ba
s
e
.
T
h
e
l
o
s
s
f
unc
t
i
o
n
us
e
d
f
o
r
t
r
a
i
n
i
ng
w
a
s
R
M
S
pr
o
p
a
nd
t
h
e
o
pt
i
m
i
z
e
r
is
b
i
n
a
r
y
c
r
o
s
s
e
nt
r
o
py
.
T
he
dl
i
b
f
a
c
i
a
l
l
a
ndm
a
r
k
w
a
s
e
xpl
o
i
t
e
d
to
pe
r
c
e
i
v
e
a
nd
p
r
e
-
pr
o
c
e
s
s
t
he
d
e
t
e
c
t
e
d
f
a
c
e
s
.
T
he
da
t
a
s
e
t
us
e
d
f
o
r
t
he
t
r
a
i
n
i
ng
m
o
de
l
is
s
e
l
e
c
t
e
d
f
r
o
m
t
he
“
X
i
a
o
y
a
n
g
T
a
n”
of
na
nj
i
ng
uni
v
e
r
s
i
t
y
of
a
e
r
o
na
ut
i
c
s
a
nd
a
s
t
r
o
na
ut
i
c
s
.
B
a
s
e
d
on
t
he
e
xpe
r
i
m
e
nt
a
l
o
ut
c
o
m
e
t
he
p
r
o
j
e
c
t
e
d
m
e
t
ho
d
a
c
hi
e
v
e
s
an
a
c
c
ur
a
c
y
of
97%
.
T
he
pr
o
t
o
t
y
pe
de
v
e
l
o
pe
d
s
e
r
v
e
s
as
a
b
a
s
e
f
o
r
f
ur
t
he
r
de
v
e
l
o
pm
e
nt
of
t
hi
s
pr
o
c
e
s
s
to
a
c
hi
e
v
e
be
t
t
e
r
r
o
a
d
s
a
f
e
t
y
.
Ke
y
w
or
d
s
:
Co
n
v
o
l
ut
i
o
n
n
e
u
ra
l
n
e
t
w
o
r
k
Cr
o
s
s
e
nt
r
o
p
y
D
e
e
p
l
e
a
rni
n
g
D
l
i
b
E
y
e
b
l
i
n
k
M
ob
i
l
e
n
e
t
V2
R
M
S
P
r
o
p
T
hi
s
is
an
ope
n
ac
c
e
s
s
ar
t
i
c
l
e
u
nde
r
t
he
CC
BY
-
SA
l
i
c
e
ns
e
.
Cor
r
e
s
pon
di
n
g
Au
t
h
or
:
R
a
gha
v
e
n
d
r
a
S
D
e
pa
rt
m
e
n
t
of
C
o
m
put
e
r
S
c
i
e
n
c
e
a
nd
E
n
g
i
n
e
e
r
i
ng
S
c
h
o
o
l
of
E
ngi
n
e
e
ri
n
g
a
nd
T
e
c
hn
o
l
o
g
y
CH
R
IS
T
D
e
e
m
e
d
to
be
U
n
i
v
e
r
s
i
ty
M
y
s
o
r
e
R
o
a
d,
B
e
n
ga
l
u
r
u
,
K
a
rn
a
t
a
ka
,
I
ndi
a
E
m
a
i
l
:
r
a
g
ha
v
.
t
r
g
@
g
m
a
i
l
.
c
o
m
1.
I
N
TR
O
D
U
C
TI
O
N
A
c
c
o
r
di
n
g
to
W
H
O
s
t
a
t
i
s
t
i
c
s
,
m
i
l
l
i
o
n
s
of
pe
o
pl
e
s
a
r
e
l
o
s
i
n
g
t
h
e
i
r
v
a
l
ua
b
l
e
l
i
v
e
s
e
a
c
h
da
y
.
N
um
b
e
r
s
e
m
pha
s
i
z
e
t
h
a
t
t
h
e
m
a
xi
m
um
of
t
h
e
l
e
t
h
a
l
c
o
l
l
i
s
i
o
n
s
a
r
e
a
nt
i
c
i
pa
t
e
d
to
d
r
i
v
e
r
’s
e
x
h
a
us
t
i
o
n
a
nd
n
e
gl
i
ge
n
c
e
.
B
a
s
e
d
on
t
h
e
s
t
a
t
i
s
t
i
c
s
of
t
h
e
a
s
s
o
c
i
a
t
i
o
n
of
t
h
e
A
m
e
r
i
c
a
n
A
ut
o
m
o
b
i
l
e
s
it
w
a
s
fo
un
d
t
h
a
t
7
%
of
t
h
e
c
o
l
l
i
s
i
o
n
s
a
n
d
21
%
of
t
h
e
l
e
t
h
a
l
t
r
a
f
f
i
c
c
o
l
l
i
s
i
o
n
s
a
r
e
c
a
us
e
d
by
e
xh
a
us
t
e
d
d
r
i
v
e
r
s
[1
].
Co
m
put
e
r
v
i
s
i
o
n
(CV
)
is
an
i
nt
e
gra
t
e
d
a
r
e
na
t
h
a
t
de
m
o
n
s
t
ra
t
e
s
h
o
w
pr
o
c
e
s
s
o
r
s
m
i
g
ht
be
us
e
d
to
a
c
qui
r
e
a
dv
a
n
c
e
d
unde
r
s
t
a
n
d
i
n
g
f
r
o
m
v
i
de
os
/
i
m
a
ge
s
.
It
is
c
on
c
e
rn
e
d
w
i
t
h
s
po
n
t
a
n
e
o
us
m
i
n
i
ng,
i
n
v
e
s
t
i
ga
t
i
o
n,
a
n
d
de
m
o
n
s
t
ra
t
i
o
n
of
be
n
e
f
i
c
i
a
l
m
a
t
e
r
i
a
l
f
r
o
m
a
di
s
t
i
n
c
t
c
o
p
y
of
an
i
m
a
ge
or
a
s
e
r
i
e
s
of
i
m
a
ge
s
.
D
e
e
p
l
e
a
rn
i
ng
a
pp
r
o
a
c
h
e
s
l
i
ke
c
o
n
v
o
l
ut
i
o
n
n
e
u
r
a
l
n
e
t
w
o
r
ks
(CN
N
s
)
a
r
e
e
m
pl
oy
e
d
to
i
de
nt
i
fy
dr
o
w
s
i
n
e
s
s
[2
-
5]
.
E
y
e
B
l
i
n
ks
a
r
e
ge
n
e
r
a
l
l
y
a
c
c
o
m
pa
ni
e
d
by
t
h
e
de
s
i
r
e
to
c
l
e
a
n
s
e
t
h
e
e
y
e
s
by
r
e
m
ov
i
n
g
dus
t
a
nd
to
e
xt
e
n
d
t
e
a
r
s
k
i
n.
So
t
h
e
e
y
e
b
l
i
n
ks
t
a
ke
p
l
a
c
e
by
da
m
p
n
e
s
s
,
h
e
a
t
,
o
r
ga
ni
c
i
n
f
l
ue
n
c
e
s
,
a
nd
air
c
o
m
po
n
e
n
t
s
[6
,
7]
.
T
h
e
pe
r
i
o
d
a
m
o
ngs
t
b
l
i
n
ks
u
ps
u
r
ge
s
w
h
e
n
a
pe
r
s
o
n
v
i
e
w
s
a
pi
c
t
o
r
i
a
l
de
m
o
n
s
t
ra
t
i
o
n
uni
t
[8]
or
is
a
b
s
t
ra
c
t
e
d
by
a
c
h
a
l
l
e
n
g
i
n
g
t
a
s
k
.
T
h
e
pe
ri
o
d
a
m
o
n
gs
t
b
l
i
n
ks
is
c
o
n
v
e
y
e
d
to
di
m
i
nut
i
o
n
w
i
t
h
an
a
u
r
a
l
j
o
b
t
hr
o
ugh
o
ut
d
r
i
v
i
n
g
as
a
s
s
o
c
i
a
t
e
d
w
i
t
h
d
ri
v
i
n
g
w
i
t
h
o
ut
a
s
uppl
e
m
e
nt
a
r
y
j
ob
[9].
T
hi
s
r
e
c
o
m
m
e
n
ds
a
n
a
f
f
i
l
i
a
t
i
o
n
a
m
o
n
gs
t
t
h
e
c
a
t
e
go
r
y
of
t
h
e
j
o
b
a
n
d
i
t
s
c
o
n
s
e
que
nc
e
on
t
h
e
d
ri
v
e
r
’s
b
l
i
nki
n
g
o
c
c
urr
e
n
c
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
IS
S
N
:
2502
-
4752
E
y
e
b
l
i
n
k
de
t
e
c
t
i
on
us
i
ng
CNN
t
o
d
e
t
e
c
t
dr
ow
s
i
ne
s
s
l
e
v
e
l
i
n
dr
i
v
e
r
s
f
or
r
oad
s
a
f
e
t
y
(
P
ot
hur
aj
u
V
i
s
he
s
h
)
223
T
h
e
M
o
b
i
l
e
N
e
t
V
2
s
t
ruc
t
u
r
a
l
de
s
i
g
n
is
g
r
o
un
de
d
on
a
r
e
v
e
rs
e
d
o
ut
s
t
a
n
d
i
n
g
s
t
ruc
t
u
r
e
in
w
hi
c
h
t
h
e
i
n
put
a
n
d
o
ut
put
a
r
e
t
hi
n
l
a
y
e
r
s
as
o
ppos
e
d
to
o
l
d
-
s
t
y
l
e
r
e
s
i
dua
l
c
o
pi
e
s
w
h
i
c
h
m
a
ke
us
e
of
e
xt
e
n
de
d
i
l
l
us
t
r
a
t
i
o
n
s
in
t
h
e
i
n
pu
t
.
M
o
b
i
l
e
N
e
t
V
2
us
e
s
t
r
i
v
i
a
l
de
pt
h
-
w
i
s
e
c
o
m
pl
i
c
a
t
i
o
n
s
to
c
a
t
e
go
r
i
z
e
a
t
t
ri
b
ut
e
s
in
t
h
e
in
-
b
e
t
w
e
e
n
e
xt
e
n
s
i
o
n
l
a
y
e
r
s
[10].
R
M
S
P
r
o
p
e
m
pl
oy
s
t
h
e
c
on
c
e
pt
of
t
h
e
e
xpo
n
e
nt
i
a
l
b
i
a
s
e
d
m
e
a
n
of
all
t
h
e
i
n
c
l
i
n
e
s
a
na
l
o
go
us
to
g
r
a
d
i
e
n
t
de
s
c
e
n
t
t
hr
o
ug
h
m
o
t
i
o
n
.
H
e
r
e
'
b
'
m
e
a
n
s
t
h
e
‘b
i
a
s
’
s
h
o
w
i
n
g
pe
r
pe
ndi
c
ul
a
r
f
l
uc
t
ua
t
i
o
n
s
a
nd
'
W
'
is
t
h
e
‘w
e
i
ght
’
s
h
o
w
i
n
g
t
h
e
p
r
o
gr
e
s
s
in
f
l
a
t
o
r
de
r
s
[11]
.
In
t
h
e
p
r
o
po
s
e
d
w
o
r
k,
we
do
e
y
e
b
l
i
n
k
de
t
e
c
t
i
o
n
us
i
n
g
CN
N
fo
r
pe
r
c
e
i
v
i
ng
t
h
e
d
r
o
w
s
i
n
e
s
s
l
e
ve
l
in
dri
v
e
r
s
fo
r
r
o
a
d
s
a
f
e
t
y
.
T
h
e
p
r
o
t
o
t
y
pe
d
e
v
e
l
o
p
e
d
s
e
r
v
e
s
as
a
b
a
s
e
fo
r
f
ur
t
h
e
r
de
v
e
l
o
pm
e
n
t
of
t
h
i
s
p
r
o
c
e
s
s
to
a
c
hi
e
v
e
be
t
t
e
r
r
o
a
d
s
a
f
e
t
y
.
V
a
r
i
o
us
t
e
c
hn
i
que
s
a
r
e
e
m
pl
oy
e
d
fo
r
dr
o
w
s
i
n
e
s
s
de
t
e
c
t
i
o
n
a
n
d
a
r
e
de
l
i
b
e
r
a
t
e
d
in
t
h
e
e
n
s
u
i
n
g
s
e
gm
e
nt
of
t
h
e
L
i
t
e
ra
t
u
r
e
R
e
v
i
e
w
.
2.
LI
TER
A
TU
R
E
R
EV
I
EW
A
m
e
t
h
o
d
to
de
t
e
c
t
dr
o
w
s
i
n
e
s
s
b
a
s
e
d
on
o
b
t
a
i
n
i
ng
b
l
o
o
d
vo
l
um
e
pul
s
a
t
i
o
n,
e
y
e
l
i
d
f
l
i
c
ke
r
i
n
g
,
a
nd
y
a
w
n
i
ng
s
i
g
na
l
s
w
a
s
i
m
pl
e
m
e
n
t
e
d
a
n
d
t
h
e
m
e
t
h
o
d
a
c
h
i
e
v
e
s
t
h
e
a
c
c
ura
c
y
of
90.
3%,
94
.
4%
,
a
n
d
9
3.
7
%
for
a
y
a
w
n
,
b
l
oo
d
vo
l
um
e
pr
e
s
s
ur
e
,
a
n
d
e
y
e
l
i
d
f
l
i
c
ke
r
i
n
g
r
e
s
pe
c
t
i
v
e
l
y
b
a
s
e
d
on
n
e
ga
t
i
v
e
pr
e
di
c
t
i
v
e
r
a
t
e
i
n
di
c
a
t
o
r
[12]
.
A
n
e
w
m
e
t
h
o
do
l
o
g
y
t
e
r
m
e
d
D
r
i
Ca
r
e
w
a
s
p
ro
pos
e
d,
w
h
i
c
h
pe
r
c
e
i
v
e
s
t
h
e
d
r
i
v
e
r
s
’
w
e
a
ri
n
e
s
s
po
s
i
t
i
o
n
i
de
nt
i
c
a
l
to
y
a
w
n
,
b
l
i
n
k
,
a
nd
e
xt
e
n
t
of
e
y
e
c
l
o
s
i
n
g
of
us
e
r
’s
pi
c
t
u
r
e
s
.
T
h
e
i
n
v
e
s
t
i
ga
t
i
o
na
l
c
o
n
s
e
que
n
c
e
s
e
xh
i
b
i
t
e
d
t
h
a
t
t
h
e
n
e
w
s
y
s
t
e
m
a
c
c
o
m
pl
i
s
h
e
d
a
ppr
o
xi
m
a
t
e
l
y
92%
e
xa
c
t
n
e
s
s
[13]
.
T
h
e
e
ye
b
l
i
n
k
c
o
unt
s
a
r
e
us
e
d
as
an
i
m
po
r
t
a
n
t
pa
ra
m
e
t
e
r
fo
r
d
r
o
w
s
i
n
e
s
s
de
t
e
c
t
i
o
n
of
t
h
e
d
ri
v
e
r
.
T
h
e
p
r
o
po
s
e
d
f
r
a
m
e
w
o
r
k
a
l
s
o
c
o
n
t
i
nuo
us
l
y
a
na
l
y
s
e
s
t
h
e
e
y
e
m
ov
e
m
e
n
t
of
t
h
e
dr
i
v
e
r
a
n
d
a
l
e
r
t
s
t
h
e
dri
v
e
r
[14]
.
T
h
e
f
a
c
i
a
l
l
a
ndm
a
r
k
de
t
e
c
t
i
o
n
(F
L
D
)
a
l
o
n
g
w
i
t
h
e
y
e
a
s
pe
c
t
r
a
t
i
o
(E
A
R
)
is
e
xe
r
c
i
s
e
d
to
de
t
e
c
t
e
y
e
b
l
i
n
k
.
T
h
e
t
i
m
e
of
t
h
e
c
l
o
s
e
d
s
t
a
t
e
of
t
h
e
e
y
e
is
t
h
e
m
a
i
n
p
a
ra
m
e
t
e
r
to
de
c
i
de
t
h
e
dr
o
w
s
i
n
e
s
s
.
T
h
e
p
r
o
po
s
e
d
m
e
t
h
o
d
a
c
h
i
e
v
e
d
a
c
c
ura
c
y
in
an
e
y
e
b
l
i
n
k
p
a
t
t
e
rn
of
92.
7%
[
15].
A
f
a
c
i
a
l
l
a
n
d
m
a
rk
b
l
i
n
k
de
t
e
c
t
o
r
(F
L
B
D
)
a
n
d
m
ul
t
i
-
l
a
y
e
r
pe
r
c
e
pt
r
o
n
(M
L
P
)
c
o
m
po
s
i
t
e
d
w
i
t
h
v
a
r
i
o
us
hi
dde
n
l
a
y
e
r
s
,
n
e
u
r
o
n
s
,
a
n
d
a
c
t
i
v
a
t
i
o
n
f
un
c
t
i
o
n
s
a
r
e
us
e
d
as
t
h
e
c
l
a
s
s
i
f
i
e
r
for
de
t
e
c
t
i
n
g
t
h
e
e
y
e
b
l
i
n
k
[16]
.
T
h
e
a
l
go
r
i
t
h
m
t
ha
t
m
a
i
n
l
y
a
na
l
y
s
e
s
t
h
e
e
y
e
b
l
i
n
k
pa
t
t
e
rn
(E
B
P
)
a
n
d
m
e
a
n
e
y
e
l
a
n
d
m
a
rks
’
di
s
t
a
n
c
e
(M
E
L
D
)
of
t
h
e
d
ri
v
e
r
s
to
s
po
t
t
h
e
s
l
e
e
pi
n
e
s
s
of
t
h
e
d
r
i
v
e
r
w
a
s
i
m
p
l
e
m
e
n
t
e
d.
T
h
e
du
r
a
t
i
o
n
of
time
of
e
y
e
c
l
o
s
ur
e
is
c
o
n
s
i
de
r
e
d
to
r
e
gul
a
t
e
t
h
e
d
r
o
w
s
i
n
e
s
s
c
o
n
di
t
i
o
n.
T
h
e
m
e
t
h
o
d
a
c
h
i
e
v
e
s
an
a
c
c
ura
c
y
of
93%
[17]
.
In
e
a
c
h
s
e
c
t
i
o
n
,
t
h
e
p
i
xe
l
’
s
w
o
r
t
h
is
c
a
l
c
ul
a
t
e
d
a
c
c
o
r
di
n
g
to
t
h
e
pi
xe
l
f
l
uc
t
ua
t
i
o
n
r
a
t
i
o
.
T
h
e
m
o
de
l
is
t
r
a
i
n
e
d
w
i
t
h
c
a
r
e
f
ul
l
y
c
h
o
s
e
n
t
ra
i
ni
n
g
t
r
i
a
l
s
a
l
o
ng
w
i
t
h
SVM
a
nd
a
c
hi
e
v
e
d
go
o
d
r
e
s
ul
t
s
[18]
.
T
h
e
c
ha
l
l
e
n
ge
s
f
a
c
e
d
w
e
r
e
to
di
s
c
r
i
m
i
na
t
e
a
m
o
n
gs
t
e
y
e
b
l
i
n
k
t
r
i
a
l
s
a
n
d
g
a
z
e
-
r
e
l
a
t
e
d
c
l
o
s
i
ng
of
t
h
e
e
y
e
l
i
d,
m
a
i
n
l
y
t
h
e
gl
i
m
ps
e
s
t
ow
a
r
ds
t
h
e
da
s
h
b
o
a
r
d.
T
h
i
s
us
e
d
t
h
e
t
hr
e
s
h
o
l
d
for
t
h
e
m
a
xi
m
u
m
v
e
l
o
c
i
t
y
of
t
h
e
e
y
e
l
i
ds
fo
r
d
i
s
c
r
i
m
i
na
t
i
o
n
[19]
.
A
d
r
i
v
e
r
f
a
t
i
gue
r
e
c
o
gn
i
t
i
o
n
s
c
h
e
m
e
in
r
e
a
l
-
t
i
m
e
g
r
o
u
n
de
d
on
t
h
e
SVM
a
l
go
ri
t
hm
is
p
r
o
j
e
c
t
e
d.
F
a
t
i
g
ue
di
s
c
ove
r
y
c
h
i
e
f
l
y
e
m
ph
a
s
i
z
e
s
dr
i
v
e
r
s
’
f
a
c
ial
a
ppe
a
ra
n
c
e
s
a
n
d
b
e
h
a
v
i
o
r
s
[20]
.
T
h
e
m
o
de
l
t
ha
t
e
m
pl
o
y
s
t
h
e
i
n
f
r
a
r
e
d
v
i
de
o
s
a
i
m
e
d
at
pe
r
c
e
i
v
i
n
g
a
nd
a
m
e
t
h
o
d
us
i
ng
CN
N
a
i
m
e
d
at
i
de
n
t
i
fy
i
n
g
e
y
e
s
t
a
t
e
w
a
s
pr
o
j
e
c
t
e
d.
T
h
e
m
o
de
l
ul
t
i
m
a
t
e
l
y
c
a
l
c
ul
a
t
e
s
t
h
e
p
r
o
po
r
t
i
o
n
of
e
y
e
l
i
d
c
l
o
s
i
n
g
c
o
n
c
e
rni
n
g
t
h
e
pu
pi
l
w
i
t
h
t
i
m
e
(P
E
R
CL
O
S
)
a
n
d
t
h
e
r
a
t
e
of
b
l
i
nk
to
pe
r
c
e
i
v
e
t
h
e
w
e
a
r
i
ne
s
s
[21].
An
e
f
f
i
c
i
e
n
t
m
e
t
h
o
d
for
de
t
e
c
t
i
n
g
t
h
e
e
y
e
-
b
l
i
n
k
is
CN
N
a
n
d
SVM.
T
h
e
m
a
i
n
a
i
m
w
a
s
to
de
t
e
c
t
t
h
e
e
y
e
-
b
l
i
n
k
r
e
l
i
a
b
l
y
.
F
r
o
m
t
h
e
e
xpe
r
i
m
e
nt
a
l
r
e
s
ul
t
s
,
it
w
a
s
fo
un
d
t
h
a
t
t
h
e
p
r
e
c
i
s
i
o
n
w
a
s
94
.
4%
[2
2].
T
h
e
us
e
of
e
y
e
-
b
l
i
n
k
s
e
n
s
o
r
s
for
t
h
e
de
t
e
c
t
i
o
n
of
e
y
e
-
b
l
i
n
k
du
r
i
ng
dri
v
i
n
g
w
a
s
de
v
e
l
o
pe
d.
In
t
h
i
s
r
e
s
e
a
r
c
h
,
t
h
e
s
pe
e
d
w
i
l
l
be
r
e
duc
e
d
or
t
h
e
v
e
h
i
c
l
e
w
i
l
l
be
s
t
o
ppe
d.
A
l
o
n
g
w
i
t
h
t
hi
s
,
t
h
e
o
w
n
e
r
w
i
l
l
be
i
n
f
o
r
m
e
d.
By
t
hi
s
,
t
h
e
d
r
i
v
e
r
a
n
d
t
h
e
o
w
n
e
r
w
i
l
l
be
n
o
t
i
f
i
e
d
a
n
d
c
a
n
a
v
o
i
d
f
ur
t
h
e
r
c
o
n
s
e
que
n
c
e
s
[23].
To
r
e
c
o
gn
i
z
e
t
h
e
m
e
nt
a
l
a
t
t
e
nt
i
o
n
l
e
v
e
l
of
t
h
e
d
ri
v
e
r
t
h
e
s
y
s
t
e
m
m
a
ke
s
us
e
of
a
b
r
a
i
n
-
c
o
m
put
e
r
i
n
t
e
r
f
a
c
e
a
l
o
n
g
w
i
t
h
t
w
o
s
e
n
s
o
r
s
.
A
m
i
c
r
o
c
o
n
t
r
o
l
l
e
r
is
us
e
d
to
f
i
n
d
t
h
e
di
f
f
e
r
e
n
c
e
b
e
t
w
e
e
n
t
h
e
e
y
e
b
l
i
n
k
of
a
n
o
r
m
a
l
pe
r
s
o
n
w
i
t
h
t
h
e
d
r
o
w
s
i
n
e
s
s
p
e
r
s
o
n
.
A
b
l
i
n
ki
ng
L
E
D
is
us
e
d
to
a
l
e
r
t
t
h
e
dri
v
e
r
[2
4].
D
a
t
a
s
e
t
s
a
r
e
ke
y
w
h
e
n
w
o
r
ki
n
g
o
ut
a
D
N
N
.
So
fo
r
i
m
p
r
o
v
i
n
g
t
h
e
pe
r
f
o
r
m
a
n
c
e
of
dr
o
w
s
i
n
e
s
s
de
t
e
c
t
i
o
n
fo
r
v
a
r
y
i
n
g
c
ul
t
u
r
a
l
a
s
s
e
m
b
l
i
e
s
a
f
r
a
m
e
w
o
r
k
b
a
s
e
d
on
CN
N
a
n
d
t
ra
i
n
i
n
g
b
a
s
e
d
on
ge
n
e
r
a
t
i
v
e
a
dv
e
r
s
a
r
i
a
l
n
e
t
w
o
r
k
s
(G
A
N
)
w
a
s
pr
o
po
s
e
d.
T
h
i
s
w
a
s
pr
i
m
a
ri
l
y
di
r
e
c
t
e
d
on
da
t
a
e
xt
e
n
s
i
o
n
g
r
o
unde
d
on
t
h
e
pub
l
i
c
b
i
a
s
c
o
n
c
e
pt
i
o
n
a
pp
r
o
a
c
h
t
h
a
t
c
l
us
t
e
r
s
a
p
pe
a
r
a
n
c
e
s
w
i
t
h
c
o
m
pa
r
a
b
l
e
f
a
c
i
a
l
c
h
a
ra
c
t
e
r
i
s
t
i
c
s
.
T
h
e
pr
o
po
s
e
d
f
r
a
m
e
w
o
r
k
r
e
s
ul
t
e
d
in
an
a
c
c
u
r
a
c
y
of
96.
75%
a
n
d
91.
62
%
for
t
r
a
i
n
i
ng
a
n
d
t
e
s
t
i
ng
r
e
s
pe
c
t
i
v
e
l
y
[25].
T
h
e
c
r
uc
i
a
l
h
y
po
t
h
e
s
i
s
w
a
s
t
h
a
t
t
h
e
s
i
t
t
i
n
g
po
s
i
t
i
o
n
c
o
rr
e
l
a
t
e
d
i
n
de
x
can
s
i
g
n
po
s
t
f
e
e
b
l
e
dr
ow
s
i
n
e
s
s
,
w
h
i
c
h
t
h
e
d
r
i
v
e
r
s
c
a
nn
o
t
f
e
e
l
.
At
f
i
r
s
t
,
t
h
e
s
e
n
s
i
t
i
v
i
t
y
of
t
h
e
s
i
t
t
i
ng
po
s
i
t
i
o
n
a
nd
c
o
n
v
e
n
t
i
o
n
a
l
i
ndi
c
e
s
a
r
e
de
t
e
r
m
i
n
e
d
a
nd
a
m
e
t
h
o
d
t
ha
t
can
de
t
e
c
t
t
h
e
d
r
o
w
s
i
n
e
s
s
w
a
s
pr
o
j
e
c
t
e
d
by
c
o
m
b
i
ni
n
g
t
h
e
v
a
ri
o
us
i
ndi
c
e
s
r
e
l
a
t
e
d
to
t
h
e
s
e
n
s
i
t
i
v
i
t
y
of
t
h
e
f
e
e
b
l
e
a
n
d
t
h
e
r
o
b
us
t
d
row
s
i
n
e
s
s
.
T
h
e
p
r
o
j
e
c
t
e
d
m
e
t
h
o
d
i
m
p
r
o
v
e
d
t
h
e
a
c
c
ur
a
c
y
of
t
h
e
f
e
e
b
l
e
dr
o
w
s
i
n
e
s
s
de
t
e
c
t
i
o
n
w
i
t
h
an
R
M
S
E
of
62%
[2
6
].
A
m
e
t
h
o
d
b
a
s
e
d
on
m
ul
t
i
p
l
e
fe
a
t
u
r
e
f
us
i
o
n
w
a
s
de
ve
l
o
pe
d
by
c
o
m
b
i
n
i
ng
t
h
e
dri
v
e
r
’s
a
t
t
ri
b
ut
e
s
w
h
e
r
e
Y
O
L
O
v
3
CN
N
w
a
s
us
e
d
to
t
a
ke
t
h
e
f
a
c
i
a
l
a
r
e
a
s
b
e
n
e
a
t
h
di
f
f
i
c
ul
t
d
r
i
v
i
n
g
c
i
r
c
um
s
t
a
n
c
e
s
.
B
a
s
e
d
on
t
h
e
dl
i
b
t
o
o
l
ki
t
,
E
F
U
a
n
d
MFU
a
r
e
us
e
d
as
a
s
s
e
s
s
m
e
n
t
l
i
m
i
t
a
t
i
o
n
s
for
t
h
e
e
y
e
s
t
a
t
e
a
nd
m
o
ut
h
s
t
a
t
e
of
t
h
e
dr
i
v
e
r
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
,
V
o
l
.
22
,
N
o
.
1
,
A
p
r
i
l
20
21
:
2
2
2
-
2
3
1
224
c
o
r
r
e
s
po
n
di
ngl
y
.
A
l
i
b
r
a
r
y
w
a
s
b
ui
l
t
for
t
h
e
i
de
n
t
i
f
i
c
a
t
i
o
n
of
t
h
e
d
ri
v
e
r
w
i
t
h
i
n
f
o
r
m
a
t
i
o
n
s
uc
h
as
e
y
e
s
t
a
t
e
,
m
o
ut
h
s
t
a
t
e
,
a
n
d
d
ri
v
e
r
b
i
o
m
e
t
r
i
c
.
At
l
a
s
t
,
t
h
e
v
e
r
i
f
i
c
a
t
i
o
n
m
o
de
l
fo
r
t
h
e
d
r
i
v
e
r
w
a
s
b
ui
l
t
a
l
o
ng
w
i
t
h
t
h
e
f
a
t
i
gue
v
a
l
ua
t
i
o
n
m
o
de
l
.
U
s
i
ng
v
i
r
t
ua
l
a
pp
l
i
c
a
t
i
o
n
s
,
t
h
e
m
e
t
h
o
d
w
a
s
c
a
pa
b
l
e
of
de
t
e
c
t
i
ng
t
h
e
f
a
t
i
gue
s
t
a
t
e
w
i
t
h
an
a
c
c
u
r
a
c
y
of
95.
10%
[27]
.
A
n
e
w
t
e
c
hn
i
que
to
n
o
t
i
c
e
D
e
e
pf
a
ke
s
us
i
n
g
t
h
e
GANs
m
o
de
l
t
hr
o
ug
h
t
h
e
D
e
e
pV
i
s
i
o
n
a
l
go
r
i
t
h
m
w
a
s
de
ve
l
o
pe
d.
D
e
e
pV
i
s
i
o
n
w
a
s
c
a
pa
b
l
e
of
a
n
a
l
y
z
i
n
g
a
n
y
m
a
j
o
r
m
o
di
f
i
c
a
t
i
o
n
in
t
h
e
b
l
i
n
ki
ng
pa
t
t
e
rn.
T
h
e
b
l
i
n
k
i
n
g
p
a
t
t
e
rn
of
h
u
m
a
n
’s
c
h
a
nge
s
f
r
o
m
o
n
e
to
a
n
o
t
h
e
r
a
n
d
t
hi
s
is
m
a
i
nl
y
b
a
s
e
d
on
b
i
o
l
o
gi
c
a
l
f
e
a
t
ur
e
s
,
ph
y
s
i
c
a
l
c
i
r
c
um
s
t
a
n
c
e
s
,
c
o
gn
i
t
i
v
e
a
c
t
i
o
n
s
,
e
t
c
.
U
s
i
n
g
a
h
e
u
ri
s
t
i
c
m
o
de
l
,
a
n
y
c
h
a
n
ge
s
in
t
h
e
e
y
e
b
l
i
n
k
pa
t
t
e
rn
s
c
a
n
be
de
t
e
r
m
i
n
e
d
us
i
ng
D
e
e
pf
a
ke
s
.
D
e
e
pV
i
s
i
o
n
w
a
s
c
a
pa
b
l
e
of
de
t
e
c
t
i
ng
t
h
e
D
e
e
pf
a
ke
s
w
i
t
h
an
a
c
c
ura
c
y
of
87.
5%
[28
].
EEG
s
i
g
n
a
l
o
b
j
e
c
t
s
l
e
a
d
to
s
o
m
e
c
o
m
pl
i
c
a
t
i
o
n
s
in
t
h
e
i
n
v
e
s
t
i
ga
t
i
o
n
a
n
d
m
a
n
a
ge
m
e
nt
of
EEG
s
i
g
n
a
l
s
.
E
v
e
n
t
h
o
ug
h
m
a
n
y
m
e
t
h
o
ds
a
r
e
a
v
a
i
l
a
b
l
e
to
i
de
n
t
i
f
y
t
h
e
s
e
obj
e
c
t
s
,
t
h
e
a
c
c
ura
c
y
a
c
h
i
e
v
e
d
by
t
h
i
s
is
n
o
t
s
a
t
i
s
f
a
c
t
o
r
y
.
S
o
,
t
h
e
r
e
is
a
r
e
qu
i
r
e
m
e
n
t
for
de
v
e
l
o
pi
n
g
n
e
w
m
e
t
h
o
ds
for
i
m
p
r
o
v
i
n
g
a
c
c
ura
c
y
.
To
de
t
e
c
t
t
h
e
e
y
e
b
l
i
n
k
o
b
j
e
c
t
s
in
EEG
s
i
g
n
a
l
s
,
t
h
e
s
e
m
a
nt
i
c
s
e
gm
e
n
t
a
t
i
o
n
m
e
t
h
o
ds
a
r
e
us
e
d
a
n
d
t
h
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
a
c
c
ura
c
y
a
c
hi
e
v
e
d
w
a
s
f
o
un
d
to
be
94.
4
%
[2
9].
O
n
e
of
t
h
e
m
a
j
o
r
s
o
ur
c
e
s
fo
r
t
h
e
l
e
t
h
a
l
r
o
a
d
a
c
c
i
de
nt
is
t
h
e
d
ri
v
e
r
’s
dr
o
w
s
i
n
e
s
s
.
O
n
e
of
t
h
e
s
o
l
ut
i
o
n
s
to
a
vo
i
d
r
o
a
d
a
c
c
i
de
n
t
s
is
a
ut
o
m
a
t
i
c
d
r
o
w
s
i
n
e
s
s
de
t
e
c
t
i
o
n
a
n
d
is
c
o
n
s
i
de
r
e
d
a
gi
f
t
e
d
s
o
l
ut
i
o
n
.
An
e
m
b
e
dde
d
CN
N
b
a
s
e
d
s
o
l
ut
i
o
n
w
a
s
pr
o
po
s
e
d
w
h
e
r
e
t
h
e
e
y
e
b
l
i
n
k
is
de
t
e
c
t
e
d
t
hr
o
ug
h
s
m
a
r
t
c
o
nn
e
c
t
e
d
gl
a
s
s
e
s
.
F
r
o
m
t
h
e
e
xpe
r
i
m
e
nt
a
l
r
e
s
ul
t
s
,
it
w
a
s
fo
un
d
t
h
a
t
CN
N
gi
v
e
s
be
t
t
e
r
r
e
s
ul
t
s
w
h
e
n
c
o
m
pa
r
e
d
to
t
h
e
a
l
go
ri
t
hm
s
b
a
s
e
d
on
t
h
e
t
hr
e
s
h
o
l
d
m
e
t
h
o
d
a
nd
t
h
e
a
c
c
u
r
a
c
y
a
c
h
i
e
v
e
d
w
a
s
90.
8%
[
30].
A
l
o
t
of
c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
s
uc
h
as
y
a
w
n
i
n
g
,
m
o
v
e
m
e
n
t
of
t
h
e
h
e
a
d
,
e
y
e
b
l
i
n
ks
can
be
e
xt
r
a
c
t
e
d
f
r
o
m
t
h
e
f
a
c
e
t
h
a
t
can
be
us
e
d
to
de
t
e
c
t
t
h
e
a
m
o
unt
of
dr
o
w
s
i
n
e
s
s
.
B
ut
de
ve
l
o
pi
n
g
a
s
y
s
t
e
m
t
h
a
t
can
de
t
e
c
t
t
h
i
s
a
n
d
t
h
a
t
can
y
i
e
l
d
c
o
n
s
i
s
t
e
n
t
a
n
d
e
xa
c
t
o
ut
c
o
m
e
s
is
an
i
nt
e
r
e
s
t
i
ng
t
a
s
k
a
nd
it
n
e
e
ds
pr
e
c
i
s
e
a
n
d
r
o
b
us
t
m
e
t
h
o
ds
.
An
a
na
l
y
s
i
s
,
of
m
a
n
y
t
e
c
h
ni
que
s
us
e
d
by
di
ffe
r
e
nt
r
e
s
e
a
r
c
he
r
s
for
dr
o
w
s
i
n
e
s
s
de
t
e
c
t
i
o
n
,
w
a
s
do
n
e
a
n
d
f
r
o
m
t
h
e
a
n
a
l
y
s
i
s
,
it
w
a
s
i
de
n
t
i
f
i
e
d
t
ha
t
SVM
is
t
h
e
t
e
c
hn
i
q
ue
t
ha
t
is
us
e
d
m
o
s
t
of
t
e
n
,
b
ut
t
h
e
CN
N
s
ga
v
e
t
h
e
b
e
s
t
r
e
s
ul
t
s
c
o
m
pa
r
e
d
to
o
t
h
e
r
m
e
t
h
o
ds
[31].
E
v
e
n
t
h
o
ug
h
m
a
n
y
t
e
c
hn
i
q
ue
s
ha
v
e
b
e
e
n
p
r
o
po
s
e
d
t
o
da
t
e
a
n
d
t
h
e
r
e
is
s
ub
s
t
a
nt
i
a
l
s
c
o
pe
to
pr
o
gr
e
s
s
in
t
h
e
de
t
e
c
t
i
o
n
s
y
s
t
e
m
s
,
t
h
e
m
a
i
n
c
ha
l
l
e
nge
h
e
r
e
is
t
ha
t
di
f
f
e
r
e
n
t
r
e
s
e
a
r
c
h
e
r
s
us
e
di
v
e
r
s
e
da
t
a
s
e
t
s
a
n
d
t
h
e
o
ut
c
o
m
e
s
c
a
nn
o
t
be
s
t
r
a
i
g
ht
f
o
r
w
a
r
dl
y
a
s
s
e
s
s
e
d.
A
n
d
a
l
s
o
t
h
e
da
t
a
s
e
t
s
ut
i
l
i
z
ed
a
r
e
fo
un
d
to
be
i
n
a
de
qu
a
t
e
a
n
d
a
r
e
c
o
l
l
e
c
t
e
d
f
r
o
m
t
h
e
c
o
n
t
ro
l
l
e
d
e
n
v
i
r
o
n
m
e
n
t
a
n
d
may
f
a
i
l
in
r
e
a
l
-
w
o
r
l
d
c
i
r
c
um
s
t
a
n
c
e
s
.
In
our
p
r
o
po
s
e
d
w
o
r
k
t
h
e
d
a
t
a
s
e
t
us
e
d
for
t
he
t
ra
i
ni
n
g
m
o
de
l
is
s
e
l
e
c
t
e
d
f
r
o
m
t
h
e
“
X
i
a
oy
a
n
g
T
a
n”
of
N
a
n
j
i
n
g
U
n
i
v
e
r
s
i
t
y
of
A
e
r
o
n
a
ut
i
c
s
a
n
d
A
s
t
r
o
n
a
u
t
i
c
s
.
W
i
t
h
t
h
i
s
p
r
o
po
s
e
d
r
e
s
e
a
r
c
h
w
o
r
k
,
an
e
f
fo
r
t
is
m
a
de
to
b
e
t
t
e
r
t
h
e
a
c
c
u
r
a
t
e
n
e
s
s
of
t
h
e
f
o
r
e
c
a
s
t
.
3.
F
R
A
M
EWO
R
K
OF
T
H
E
P
R
O
P
O
S
ED
WO
R
K
T
h
i
s
w
o
r
k
is
a
b
o
ut
de
v
e
l
o
pi
n
g
an
a
rt
i
f
i
c
i
a
l
i
nt
e
l
l
i
ge
n
c
e
(A
I
)
m
o
de
l
t
ha
t
is
c
a
pa
b
l
e
of
i
de
n
t
i
f
y
i
n
g
a
pe
r
s
o
n
a
nd
t
e
l
l
i
ng
w
h
e
t
h
e
r
hi
s
/
h
e
r
e
y
e
s
a
r
e
o
pe
n
or
c
l
o
s
e
.
T
he
m
o
de
l
de
ve
l
o
p
e
d
is
a
de
e
p
l
e
a
r
ni
n
g
m
o
de
l
a
n
d
s
h
o
ul
d
t
a
ke
t
h
e
i
m
a
ge
s
as
i
n
pu
t
a
n
d
r
e
s
ul
t
in
an
o
ut
pu
t
s
pe
c
i
fy
i
n
g
t
h
e
e
y
e
s
in
t
h
e
i
m
a
ge
a
r
e
o
pe
n
or
c
l
o
s
e
d.
T
h
i
s
w
o
r
k
aims
to
p
r
o
v
i
de
a
s
o
l
ut
i
o
n
for
e
y
e
b
l
i
n
k
de
t
e
c
t
i
o
n
in
a
ut
o
m
o
t
i
v
e
s
a
fe
t
y
.
T
h
e
a
pp
r
o
a
c
h
f
o
l
l
ow
e
d
in
t
h
i
s
w
o
r
k
f
o
l
l
ow
s
r
e
t
ra
i
ni
n
g
on
t
h
e
p
r
e
v
i
o
us
l
y
b
ui
l
d
m
o
de
l
s
,
to
ut
i
l
i
z
e
t
h
e
m
e
t
ri
c
s
of
t
h
o
s
e
m
o
de
l
s
to
a
c
h
i
e
v
e
b
e
t
t
e
r
pe
r
f
o
r
m
a
n
c
e
.
T
h
e
g
e
n
e
r
a
l
f
l
o
w
of
CN
N
is
gi
v
e
n
in
F
i
g
ur
e
1.
F
i
gu
r
e
1.
E
y
e
b
l
i
n
k
de
t
e
c
t
i
o
n
m
o
de
l
Evaluation Warning : The document was created with Spire.PDF for Python.
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
IS
S
N
:
2502
-
4752
E
y
e
b
l
i
n
k
de
t
e
c
t
i
on
us
i
ng
CNN
t
o
d
e
t
e
c
t
dr
ow
s
i
ne
s
s
l
e
v
e
l
i
n
dr
i
v
e
r
s
f
or
r
oad
s
a
f
e
t
y
(
P
ot
hur
aj
u
V
i
s
he
s
h
)
225
T
h
e
b
a
s
i
c
c
o
m
po
n
e
nt
s
m
e
nt
i
o
n
e
d
in
F
i
g
u
r
e
1
c
o
n
s
i
s
t
s
of
t
h
e
o
pe
r
a
t
i
o
n
s
as
f
o
l
l
ow
s
:
T
h
e
c
o
n
v
o
l
ut
i
o
n
s
t
e
p
To
m
i
n
e
,
t
h
e
c
h
a
ra
c
t
e
r
i
s
t
i
c
s
of
t
h
e
e
nt
e
r
e
d
i
m
a
ge
Co
n
v
N
e
t
a
r
e
e
m
pl
o
y
e
d.
Co
n
v
o
l
ut
i
o
n
gi
v
e
s
t
h
e
s
pa
t
i
a
l
a
s
s
o
c
i
a
t
i
o
n
a
m
o
ng
pi
xe
l
s
by
un
de
r
s
t
a
n
d
i
n
g
i
m
a
ge
c
ha
r
a
c
t
e
r
i
s
t
i
c
s
e
m
pl
oy
i
n
g
s
m
a
l
l
s
ha
pe
s
of
e
n
t
e
r
e
d
f
a
c
t
s
.
T
h
e
Co
n
v
o
l
ut
i
o
n
p
h
a
s
e
is
s
h
o
w
n
in
F
i
gu
r
e
2.
D
e
s
i
gn
(o
pt
i
m
i
z
e
r,
l
o
s
s
f
un
c
t
i
o
n
a
nd
l
a
y
e
r
s
)
CN
N
c
o
n
s
i
s
t
s
of
an
i
n
p
ut
l
a
y
e
r
,
an
o
ut
put
l
a
y
e
r
,
a
n
d
v
a
r
i
o
us
h
i
dde
n
l
a
y
e
r
s
.
E
a
c
h
s
e
que
n
c
e
of
l
a
y
e
r
s
in
t
h
e
CN
N
s
h
i
dde
n
l
a
y
e
r
s
‘Co
n
v
o
l
v
e
’
w
i
t
h
e
xpo
n
e
n
t
i
a
t
i
o
n.
R
e
L
U
is
t
h
e
v
e
r
y
f
r
e
que
n
t
l
y
e
m
pl
oy
e
d
a
c
t
i
v
a
t
i
o
n
f
un
c
t
i
o
n
in
CN
N
a
n
d
is
s
h
o
w
n
in
F
i
g
u
r
e
3.
A
c
t
i
v
a
t
i
o
n
f
un
c
t
i
o
n
s
a
r
e
m
a
t
h
e
m
a
t
i
c
a
l
e
s
t
i
m
a
t
e
s
t
ha
t
r
e
gul
a
t
e
t
h
e
CN
N
s
o
ut
put
.
F
i
gu
r
e
2.
Co
n
v
o
l
ut
i
o
n
s
t
e
p
F
i
gu
r
e
3.
R
e
L
U
a
c
t
i
v
a
t
i
o
n
f
i
n
c
t
i
o
n
O
pt
i
m
i
z
e
r
O
pt
i
m
i
z
e
r
s
a
r
e
m
e
t
h
o
ds
t
ha
t
a
r
e
e
m
pl
oy
e
d
to
t
r
a
n
s
f
o
r
m
t
h
e
c
h
a
ra
c
t
e
ri
s
t
i
c
s
of
a
n
y
CN
N
in
t
e
rm
s
of
l
e
a
rn
i
ng
p
r
o
po
r
t
i
o
n
s
a
nd
w
e
i
gh
t
s
to
l
e
s
s
e
n
t
h
e
s
h
o
rt
f
a
l
l
s
.
T
he
fe
a
t
u
r
e
e
xt
ra
c
t
i
o
n
l
a
y
e
r
of
t
h
i
s
w
o
r
k
m
a
ke
s
us
e
of
R
MSP
r
o
p.
O
pt
i
m
i
z
a
t
i
o
n
po
l
i
c
i
e
s
a
r
e
a
c
c
o
un
t
a
b
l
e
fo
r
d
r
o
ppi
n
g
t
h
e
l
o
s
s
e
s
a
n
d
to
de
l
i
v
e
r
t
h
e
ut
m
o
s
t
pr
e
c
i
s
e
c
o
n
s
e
que
n
c
e
s
p
o
s
s
i
b
l
e
.
T
h
e
R
M
S
P
r
o
p
us
e
s
dW
a
n
d
db
for
e
a
c
h
e
poc
h
a
n
d
it
c
o
n
s
i
de
r
s
t
h
e
e
xpo
n
e
nt
i
a
l
l
y
w
e
i
ght
e
d
m
e
a
n
s
of
s
qua
r
e
s
of
dW
a
n
d
db
as
s
h
o
w
n
in
(
1
)
a
nd
(
2
)
:
=
.
+
(
1
−
)
.
2
(1)
=
.
+
(
1
−
)
.
2
(2)
H
e
r
e
,
v
a
r
i
e
s
f
r
o
m
0
to
1
a
n
d
is
a
h
y
pe
r
pa
r
a
m
e
t
e
r
.
It
us
e
s
t
h
e
w
e
i
gh
t
e
d
m
e
a
n
of
pr
e
v
i
o
us
e
s
t
i
m
a
t
e
s
a
n
d
t
h
e
s
qua
r
e
s
of
t
h
e
pr
e
s
e
n
t
e
s
t
i
m
a
t
e
s
to
f
i
gur
e
o
ut
t
h
e
f
ut
u
r
e
w
e
i
ght
e
d
m
e
a
n
a
n
d
t
h
e
pa
ra
m
e
t
e
r
s
a
r
e
up
da
t
e
d
us
i
n
g
(
3
)
a
nd
(
4
)
:
+
1
=
−
.
2
√
(3)
+
1
=
−
.
2
√
(4)
L
o
s
s
f
un
c
t
i
o
n
L
o
s
s
f
un
c
t
i
o
n
s
uppo
rt
s
i
m
p
r
o
v
i
n
g
t
h
e
l
i
m
i
t
a
t
i
o
n
s
of
t
h
e
C
N
N
s
by
c
a
l
c
ul
a
t
i
ng
t
h
e
l
o
s
s
.
T
h
i
s
w
o
r
k
m
a
ke
s
us
e
of
B
i
n
a
r
y
c
r
o
s
s
-
e
n
t
r
o
py
fo
r
its
b
i
na
r
y
n
a
t
u
r
e
,
it
m
e
a
s
ur
e
s
t
h
e
p
r
e
di
c
t
i
o
n
f
r
o
m
t
h
e
t
r
ue
v
a
l
ue
(w
h
i
c
h
is
e
i
t
h
e
r
0
or
1)
is
for
e
v
e
r
y
c
l
a
s
s
a
n
d
t
h
e
n
f
i
n
ds
t
h
e
m
e
a
n
of
t
h
e
s
e
c
l
a
s
s
-
w
i
s
e
f
a
ul
t
s
to
a
c
qui
r
e
t
h
e
u
l
t
i
m
a
t
e
l
o
s
s
.
T
h
e
b
i
n
a
r
y
c
r
o
s
s
-
e
n
t
r
o
py
l
o
s
s
is
de
f
i
n
e
d
us
i
n
g
(
5
)
.
=
∑
l
o
g
(
(
)
)
=
−
1
l
o
g
(
(
1
)
)
−
(
1
−
1
)
′
=
2
=
1
l
o
g
(
1
−
(
1
)
)
(5)
H
e
r
e
,
C′
=
2,
C
1
,
a
n
d
C
2
a
r
e
t
w
o
c
l
a
s
s
e
s
a
s
s
um
e
d,
t
1
a
nd
s
1
a
r
e
t
h
e
v
a
l
ue
fo
r
C
1
,
t
2
=1
-
t
1
,
a
nd
s
2
=1
-
s
1
is
t
h
e
v
a
l
ue
of
C
2
.
T
h
e
l
o
s
s
can
be
r
e
p
r
e
s
e
n
t
e
d
as
s
h
o
w
n
in
(
6
)
.
=
{
−
l
o
g
(
(
1
)
)
1
=
1
−
l
o
g
(
1
−
(
1
)
)
1
=
0
(6)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
,
V
o
l
.
22
,
N
o
.
1
,
A
p
r
i
l
20
21
:
2
2
2
-
2
3
1
226
P
oo
l
i
n
g
CN
N
may
c
o
m
pr
i
s
e
ge
n
e
r
a
l
or
uni
v
e
r
s
a
l
po
o
l
i
n
g
l
a
y
e
r
s
to
ra
t
i
o
n
a
l
i
z
e
t
h
e
f
un
d
a
m
e
n
t
a
l
c
a
l
c
ul
a
t
i
o
n.
P
oo
l
i
n
g
l
a
y
e
r
s
di
m
i
n
i
s
h
t
h
e
m
a
g
n
i
t
ude
s
of
t
h
e
f
a
c
t
s
by
m
e
r
g
i
n
g
t
h
e
y
i
e
l
ds
of
n
e
ur
o
n
g
r
o
ups
at
o
n
e
l
a
y
e
r
w
i
t
h
pa
r
t
i
c
ul
a
r
n
e
u
r
o
n
in
t
h
e
s
ub
s
e
que
n
t
l
a
y
e
r
.
G
e
n
e
r
i
c
po
o
l
i
n
g
un
i
t
e
s
m
i
n
o
r
c
l
us
t
e
r
s
,
n
a
t
u
r
a
l
l
y
2x2.
U
n
i
v
e
r
s
a
l
po
o
l
i
n
g
p
r
o
c
e
e
ds
on
e
ve
r
y
o
t
h
e
r
n
e
u
r
o
n
in
t
h
e
c
o
n
v
o
l
ut
i
o
n
a
l
l
a
y
e
r
.
P
o
o
l
i
n
g
m
i
g
h
t
c
a
l
c
ul
a
t
e
a
m
a
xi
m
u
m
or
a
m
e
a
n
.
T
h
e
po
o
l
i
n
g
p
ha
s
e
is
s
h
o
w
n
in
F
i
gu
r
e
4.
F
i
gu
r
e
4.
P
o
o
l
i
ng
s
t
e
p
F
ul
l
y
c
o
nn
e
c
t
e
d
l
a
y
e
r
T
h
i
s
is
t
y
pi
c
a
l
l
y
t
h
e
l
a
y
e
r
in
a
n
y
CN
N
upo
n
w
h
i
c
h
t
h
e
F
l
a
t
t
e
n
a
n
d
D
e
n
s
e
l
a
y
e
r
s
a
r
e
a
pp
l
i
e
d.
T
h
e
po
o
l
e
d
f
e
a
t
ur
e
s
a
r
e
t
r
a
n
s
f
o
r
m
e
d
i
nt
o
a
s
pe
c
i
f
i
c
c
o
l
um
n
by
t
he
F
l
a
t
t
e
n
f
un
c
t
i
o
n
.
T
h
i
s
c
o
l
um
n
is
di
s
p
a
t
c
h
e
d
to
t
h
e
f
ul
l
y
c
o
nn
e
c
t
e
d
l
a
y
e
r
.
T
h
e
f
ul
l
y
c
o
nn
e
c
t
e
d
l
a
y
e
r
is
i
n
s
e
rt
e
d
in
to
CN
N
w
i
t
h
t
h
e
h
e
l
p
of
D
e
n
s
e
.
A
f
ul
l
y
c
o
n
n
e
c
t
e
d
l
a
y
e
r
l
i
n
ks
e
a
c
h
n
e
ur
o
n
in
a
s
i
ngl
e
l
a
y
e
r
w
i
t
h
e
a
c
h
o
t
h
e
r
n
e
u
r
o
n
in
t
h
e
n
e
xt
l
a
y
e
r
a
n
d
is
s
i
m
i
l
a
r
to
M
L
P
.
T
h
e
f
l
a
t
t
e
n
e
d
m
a
t
r
i
x
m
o
ve
s
a
r
o
u
n
d
t
hi
s
l
a
y
e
r
to
c
a
t
e
go
ri
z
e
t
h
e
i
m
a
ge
s
.
4.
D
ES
I
G
N
AND
I
M
P
LEM
EN
TA
TI
O
N
CN
N
’
s
a
r
e
us
e
d
to
i
m
pl
e
m
e
nt
t
h
i
s
p
r
o
j
e
c
t
e
d
w
o
r
k.
CN
N
’
s
a
r
e
v
e
r
s
i
o
n
s
of
M
L
P
s
a
n
d
us
e
s
f
ul
l
y
c
o
n
n
e
c
t
e
d
l
a
y
e
r
.
T
h
e
D
e
s
i
gn
of
t
h
i
s
w
o
r
k
is
gi
v
e
n
b
r
i
e
f
l
y
in
t
h
e
h
i
g
h
-
l
e
v
e
l
de
s
i
gn
di
a
g
ra
m
in
F
i
gu
r
e
5,
w
h
e
r
e
t
h
e
l
o
a
di
n
g
of
t
h
e
c
urr
e
nt
da
t
a
s
e
t
(
c
l
o
s
e
d
ey
es
in
t
h
e
w
i
l
d
–
(
CE
W
)
)
a
n
d
a
ddi
ng
of
fe
a
t
ur
e
e
xt
r
a
c
t
i
o
n
l
a
y
e
r
on
t
h
e
CN
N
a
r
e
m
e
nt
i
o
n
e
d.
F
i
gu
r
e
5.
H
i
g
h
l
e
v
e
l
di
a
g
ra
m
T
h
e
s
t
e
ps
f
o
l
l
ow
e
d
a
r
e
:
S
t
e
p
1:
S
t
a
rt
.
S
t
e
p
2:
L
o
a
d
t
h
e
D
a
t
a
s
e
t
(
c
l
o
s
e
d
e
y
e
s
i
n
t
h
e
w
i
l
d
).
-
In
t
hi
s
w
o
r
k,
we
h
a
v
e
e
m
pl
oy
e
d
t
h
e
o
pe
n
da
t
a
s
e
t
pub
l
i
s
h
e
d
by
’X
i
a
oy
a
n
g
T
a
n
’
f
r
o
m
N
a
n
j
i
n
g
U
n
i
v
e
r
s
i
t
y
of
A
e
r
o
n
a
ut
i
c
s
a
n
d
A
s
t
r
o
na
ut
i
c
s
.
A
s
a
m
p
l
e
of
t
he
i
m
a
ge
s
b
e
l
o
n
gi
n
g
to
t
h
e
s
e
t
c
l
o
s
e
d
e
y
e
s
or
o
pe
n
e
y
e
s
is
s
h
o
w
n
in
F
i
gu
r
e
6
a
nd
F
i
gu
r
e
7.
S
t
e
p
3:
R
e
s
i
z
e
a
n
d
R
e
s
h
a
pe
e
v
e
r
y
i
m
a
ge
i
n
t
h
e
D
a
t
a
s
e
t
.
S
t
e
p
4:
R
o
t
a
t
e
a
n
d
f
l
i
p
e
v
e
r
y
i
m
a
ge
f
o
r
b
e
t
t
e
r
pe
r
f
o
r
m
a
n
c
e
d
uri
n
g
t
ra
i
ni
n
g
.
S
t
e
p
5:
S
e
l
e
c
t
t
h
e
m
o
de
l
a
n
d
a
r
c
hi
t
e
c
t
u
r
e
a
c
c
o
r
di
n
g
t
o
t
h
e
p
r
ob
l
e
m
s
t
a
t
e
m
e
nt
.
-
T
h
e
CN
N
us
e
d
fo
r
t
hi
s
m
o
de
l
i
s
’M
o
b
i
l
e
N
e
t
V
2’
a
n
d
t
h
e
o
ut
put
o
f
t
h
e
m
o
de
l
i
s
b
i
n
a
r
y
i
.
e
.
e
y
e
s
c
l
o
s
e
d
o
r
e
y
e
s
o
p
e
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
IS
S
N
:
2502
-
4752
E
y
e
b
l
i
n
k
de
t
e
c
t
i
on
us
i
ng
CNN
t
o
d
e
t
e
c
t
dr
ow
s
i
ne
s
s
l
e
v
e
l
i
n
dr
i
v
e
r
s
f
or
r
oad
s
a
f
e
t
y
(
P
ot
hur
aj
u
V
i
s
he
s
h
)
227
S
t
e
p
6:
W
r
i
t
e
t
h
e
t
o
p
F
e
a
t
u
r
e
E
xt
ra
c
t
i
o
n
L
a
y
e
r
t
o
de
t
e
c
t
t
h
e
s
t
a
t
e
o
f
t
h
e
e
y
e
s
.
S
t
e
p
7:
T
ra
i
ni
n
g
-
T
h
e
l
a
s
t
l
a
y
e
r
s
a
r
e
f
ul
l
y
c
o
n
n
e
c
t
e
d
n
e
t
w
o
r
k
l
a
y
e
r
s
f
o
l
l
ow
e
d
by
‘s
of
t
m
a
x
r
e
g
r
e
s
s
i
o
n
’
f
o
r
c
l
a
s
s
i
f
i
c
a
t
i
o
n
i
n
t
h
e
o
ut
put
l
a
y
e
r
.
T
h
e
w
e
i
ght
s
i
n
t
h
e
t
o
p
l
a
y
e
r
s
o
f
t
h
e
pr
e
-
t
ra
i
n
e
d
p
r
o
t
o
t
y
p
e
s
a
r
e
a
dj
us
t
e
d
f
ur
t
h
e
r
t
o
ups
ur
ge
t
h
e
f
un
c
t
i
o
n
i
n
g
.
T
hi
s
w
i
l
l
a
dj
u
s
t
t
h
e
w
e
i
gh
t
s
t
o
a
s
s
o
c
i
a
t
e
w
i
t
h
t
h
e
f
e
a
t
ur
e
r
e
l
a
t
e
d
t
o
t
h
e
d
a
t
a
s
e
t
.
S
t
e
p
8:
T
e
s
t
i
ng
t
h
e
m
o
de
l
a
n
d
t
u
ni
n
g
o
f
h
y
pe
r
-
pa
r
a
m
e
t
e
r
s
.
R
e
pe
a
t
s
t
e
p
7
.
-
T
h
e
f
i
n
e
-
t
u
n
i
ng
a
pp
r
o
a
c
h
w
a
s
e
m
pl
oy
e
d
i
n
t
ra
i
ni
n
g
t
h
e
m
o
d
e
l
.
CN
N
’s
a
r
e
s
i
m
i
l
a
r
t
o
n
o
rm
a
l
N
N
.
T
h
e
y
c
o
n
s
i
s
t
of
n
e
ur
o
n
s
w
i
t
h
l
e
a
rn
a
b
l
e
b
i
a
s
e
s
a
n
d
w
e
i
gh
t
s
.
E
a
c
h
n
e
ur
o
n
t
a
k
e
s
t
h
e
da
t
a
a
n
d
do
e
s
t
h
e
do
t
p
r
o
duc
t
.
T
h
e
c
o
m
pl
e
t
e
n
e
t
w
o
r
k
gi
v
e
s
o
n
e
di
f
fe
r
e
nt
i
a
l
s
c
o
r
e
f
un
c
t
i
o
n
a
nd
s
t
i
l
l
h
a
s
a
l
o
s
s
f
un
c
t
i
o
n.
S
t
e
p
9:
M
a
ke
P
r
e
di
c
t
i
o
n
s
o
n
c
o
m
pl
e
t
e
l
y
n
e
w
u
n
s
e
e
n
d
a
t
a
.
S
t
e
p
10:
E
nd.
S
a
v
e
t
h
e
m
o
de
l
a
nd
de
pl
o
y
i
t
.
F
i
gu
r
e
6.
C
l
o
s
e
d
e
y
e
s
F
i
gu
r
e
7.
O
pe
n
e
y
e
s
5.
P
R
O
TO
T
Y
P
I
N
G
AND
TES
TI
N
G
T
h
e
t
e
s
t
i
n
g
p
ha
s
e
of
t
h
e
pr
o
j
e
c
t
w
a
s
i
n
i
t
i
a
l
l
y
fo
c
us
e
d
on
de
v
e
l
o
pi
n
g
a
m
o
b
i
l
e
a
pp
to
t
e
s
t
t
h
e
m
o
de
l
t
r
a
i
n
e
d
.
F
i
n
a
l
l
y
,
a
w
e
b
c
a
m
-
b
a
s
e
d
pr
o
t
o
t
y
pe
w
a
s
de
ve
l
o
p
e
d
to
t
e
s
t
t
h
e
m
o
de
l
t
ra
i
n
e
d
for
t
h
e
p
r
o
b
l
e
m
s
t
a
t
e
m
e
n
t
,
us
i
ng
d
lib
f
a
c
i
a
l
l
a
n
d
m
a
rks
to
de
t
e
c
t
f
a
c
e
s
in
t
h
e
v
i
de
o
s
t
r
e
a
m
a
nd
t
h
e
n
p
r
e
-
p
r
o
c
e
s
s
t
h
e
de
t
e
c
t
e
d
f
a
c
e
for
i
n
f
e
r
e
n
c
e
w
i
t
h
t
h
e
m
o
de
l
.
T
h
e
r
e
s
ul
t
s
a
f
t
e
r
t
h
e
t
r
a
i
ni
n
g
s
h
o
w
gr
a
p
h
s
a
b
o
ut
t
h
e
l
o
s
s
a
n
d
a
c
c
u
r
a
t
e
n
e
s
s
de
t
a
i
l
s
t
hr
o
ug
h
o
ut
t
h
e
t
r
a
i
n
i
ng
a
n
d
v
a
l
i
d
a
t
i
o
n
p
ha
s
e
.
5.
1
.
Th
e
w
e
b
c
am
ap
p
r
o
ac
h
T
h
e
w
e
b
c
a
m
a
pp
r
o
a
c
h
t
a
ke
s
t
h
e
l
i
v
e
v
i
de
o
i
n
pu
t
f
r
o
m
t
h
e
w
e
bc
a
m
a
nd
w
a
s
us
e
d
for
t
e
s
t
i
n
g
t
h
e
m
o
de
l
.
S
i
n
c
e
t
h
e
m
o
de
l
w
a
s
t
r
a
i
n
e
d
w
i
t
h
o
n
l
y
i
m
a
ge
s
of
f
a
c
e
da
t
a
w
i
t
h
e
y
e
s
c
l
o
s
e
d
or
o
pe
n
,
t
h
e
m
o
de
l
c
a
n
o
n
l
y
i
n
f
e
r
e
n
c
e
on
f
a
c
e
i
m
a
ge
s
.
To
de
t
e
c
t
a
n
d
c
r
o
p
f
a
c
e
s
f
r
om
t
h
e
i
n
p
ut
s
t
r
e
a
m
,
dl
i
b
f
a
c
i
a
l
de
t
e
c
t
o
r
w
a
s
us
e
d
to
de
t
e
c
t
f
a
c
e
s
in
t
h
e
i
nput
f
e
e
l
.
T
h
e
dl
i
b
f
a
c
i
a
l
de
t
e
c
t
o
r
w
o
r
ks
on
t
h
e
f
a
c
i
a
l
l
a
ndm
a
r
ks
.
T
h
e
d
l
i
b
f
a
c
i
a
l
l
a
n
d
m
a
rk
is
a
c
o
l
l
e
c
t
i
o
n
of
po
i
n
t
s
a
nd
is
us
e
d
to
de
t
e
c
t
v
a
ri
o
us
pa
rt
s
of
t
h
e
f
a
c
e
us
i
n
g
68
po
i
n
t
s
on
t
h
e
f
a
c
e
.
T
h
e
s
e
f
a
c
i
a
l
l
a
ndm
a
r
ks
us
e
d
by
dl
i
b
f
a
c
i
a
l
de
t
e
c
t
o
r
a
r
e
gi
v
e
n
in
F
i
gu
r
e
8
s
h
o
w
i
n
g
h
o
w
t
h
e
po
i
nt
s
m
a
ke
us
t
h
e
f
a
c
e
of
a
n
y
h
um
a
n.
T
h
e
w
h
o
l
e
f
l
o
w
of
t
h
e
de
m
o
a
p
pl
i
c
a
t
i
o
n
is
gi
v
e
n
in
F
i
gu
r
e
9.
F
i
gu
r
e
8.
68
-
po
i
n
t
f
a
c
i
a
l
l
a
n
d
m
a
r
ks
F
i
gu
r
e
9.
F
l
o
w
of
de
m
o
a
ppl
i
c
a
t
i
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
,
V
o
l
.
22
,
N
o
.
1
,
A
p
r
i
l
20
21
:
2
2
2
-
2
3
1
228
T
h
e
s
t
e
p
f
o
l
l
o
w
e
d
a
r
e
:
W
e
b
c
a
m
:
a
n
y
w
e
b
c
a
m
w
i
t
h
a
pi
c
t
u
r
e
qu
a
l
i
t
y
of
720p
or
g
r
e
a
t
e
r
.
dl
i
b
:
t
h
e
f
a
c
i
a
l
de
t
e
c
t
o
r
to
de
t
e
c
t
f
a
c
e
s
in
t
h
e
w
e
b
c
a
m
f
e
e
d,
to
l
o
c
a
t
e
t
h
e
f
a
c
e
in
t
h
e
w
i
l
d.
Cr
o
p:
T
h
e
l
o
c
a
t
i
o
n
of
f
a
c
e
s
is
ob
t
a
i
n
e
d
f
r
o
m
d
l
i
b
f
a
c
i
a
l
de
t
e
c
t
o
r
a
nd
is
c
r
o
ppe
d
c
o
n
s
i
s
t
i
n
g
of
o
n
l
y
t
h
e
f
a
c
e
a
n
d
n
o
t
hi
n
g
e
l
s
e
.
R
e
s
i
z
e
:
T
h
e
m
o
de
l
is
m
o
r
e
e
ff
i
c
i
e
n
t
to
pr
o
c
e
s
s
i
m
a
ge
s
of
di
m
e
n
s
i
o
n
s
244*24
4,
so
t
h
e
i
m
a
ge
of
t
h
e
f
a
c
e
is
r
e
s
i
z
e
d
a
c
c
o
r
di
n
g
l
y
w
h
a
t
e
v
e
r
t
h
e
s
i
z
e
m
a
y
b
e
.
In
f
e
r
e
n
c
e
:
T
h
e
i
m
a
ge
is
c
o
n
v
e
r
t
e
d
i
nt
o
t
e
n
s
o
r
s
a
nd
i
n
f
e
r
e
n
c
e
s
r
e
a
dy
fo
r
p
r
e
di
c
t
i
o
n
.
P
r
e
di
c
t
:
T
h
e
i
m
a
ge
is
pa
s
s
e
d
for
p
r
e
di
c
t
i
o
n
w
i
t
h
a
m
o
de
l
t
ra
i
n
e
d
for
E
B
D
.
O
ut
put
:
T
h
e
o
ut
pu
t
g
i
v
e
n
o
ut
by
t
h
e
m
o
de
l
is
t
ra
n
s
l
a
t
e
d
i
n
t
o
o
pe
n
or
c
l
o
s
e
a
n
d
d
i
s
pl
a
y
e
d
on
t
h
e
s
c
r
e
e
n.
6.
R
ES
U
LTS
AND
DISCUSSIO
N
R
e
s
ul
t
s
ob
t
a
i
n
e
d
du
r
i
n
g
t
h
e
t
ra
i
ni
n
g
of
t
h
e
n
e
t
w
o
r
k
is
gi
v
e
n
in
F
i
gu
r
e
10
.
T
he
a
c
c
ur
a
c
y
m
e
n
t
i
o
n
e
d
h
e
r
e
is
c
o
n
c
e
r
ni
n
g
t
h
e
v
a
l
i
da
t
i
o
n
da
t
a
gi
v
e
n
d
u
r
i
n
g
t
h
e
t
r
a
i
ni
n
g
t
i
m
e
a
n
d
is
s
ub
j
e
c
t
o
n
l
y
to
t
h
e
da
t
a
in
t
h
e
da
t
a
s
e
t
.
T
h
e
o
ut
put
of
t
h
e
p
a
r
t
i
a
l
l
y
de
v
e
l
o
pe
d
m
ob
i
l
e
a
pp
is
gi
v
e
n
in
F
i
gu
r
e
11
.
T
h
e
f
a
c
e
s
a
r
e
de
t
e
c
t
e
d
by
F
i
r
e
b
a
s
e
V
i
s
i
o
n
ML
a
n
d
t
h
e
n
c
r
o
ppe
d
a
nd
r
e
s
i
z
e
d
a
c
c
o
r
di
ng
to
t
h
e
m
o
de
l
a
n
d
t
h
e
n
i
n
f
e
r
e
n
c
e
.
F
i
gu
r
e
10
.
A
c
c
u
r
a
c
y
a
n
d
l
o
s
s
du
r
i
ng
t
r
a
i
n
i
ng
a
n
d
v
a
l
i
da
t
i
o
n
F
i
gu
r
e
11
.
O
b
j
e
c
t
de
t
e
c
t
i
o
n
A
pp
T
h
e
a
n
d
r
o
i
d
a
p
p
w
a
s
e
xpe
r
i
m
e
n
t
e
d
w
i
t
h
us
i
n
g
v
a
r
i
o
us
a
pp
ro
a
c
h
e
s
us
i
n
g
A
n
d
r
o
i
d’s
f
i
r
e
b
a
s
e
v
i
s
i
o
n
f
a
c
e
de
t
e
c
t
o
r
to
de
t
e
c
t
a
nd
t
r
a
c
k
f
a
c
e
s
.
-
T
h
e
i
m
a
ge
c
l
a
s
s
i
f
i
c
a
t
i
o
n
a
pp
r
o
a
c
h
:
T
hi
s
a
pp
r
o
a
c
h
us
e
d
t
h
e
no
r
m
a
l
c
l
a
s
s
i
f
i
c
a
t
i
o
n
m
o
de
l
for
c
l
a
s
s
i
fy
i
n
g
t
h
e
de
t
e
c
t
e
d
f
a
c
e
s
as
o
pe
n
or
c
l
o
s
e
.
B
ut
t
h
e
o
b
s
t
a
c
l
e
w
a
s
t
h
a
t
t
h
e
a
pp
w
a
s
n
o
t
a
b
l
e
to
de
t
e
c
t
a
n
d
c
r
o
p
f
a
c
e
b
e
fo
r
e
r
u
nn
i
ng
i
n
f
e
r
e
n
c
e
on
t
h
e
i
m
a
ge
.
T
h
e
i
s
s
ue
t
ha
t
w
a
s
i
de
n
t
i
f
i
e
d
w
a
s
r
e
l
a
t
e
d
to
s
y
n
c
hr
o
n
o
us
a
n
d
a
s
y
n
c
hr
o
n
o
us
w
a
y
s
of
e
xe
c
ut
i
n
g
t
a
s
ks
.
-
T
h
e
ob
j
e
c
t
d
e
t
e
c
t
i
o
n
a
pp
r
o
a
c
h:
T
h
e
ob
j
e
c
t
de
t
e
c
t
i
o
n
a
pp
r
o
a
c
h
f
o
l
l
ow
s
a
t
o
t
a
l
l
y
n
e
w
w
a
y
of
a
ppr
o
a
c
hi
n
g
t
h
e
pr
o
b
l
e
m
s
t
a
t
e
m
e
nt
.
T
h
e
r
e
s
e
a
r
c
h
in
t
hi
s
f
i
e
l
d
is
s
t
i
l
l
in
t
h
e
p
r
o
c
e
s
s
.
W
h
e
r
e
t
h
e
p
r
e
-
p
r
o
c
e
s
s
i
n
g
of
t
h
e
i
m
a
g
e
to
de
t
e
c
t
f
a
c
e
s
is
n
o
t
r
e
qui
r
e
d
.
T
h
e
t
r
a
i
n
i
ng
p
r
o
c
e
s
s
i
t
s
e
l
f
l
e
a
rn
s
to
de
t
e
c
t
t
h
e
f
a
c
e
s
in
t
h
e
w
i
l
d
a
n
d
t
h
e
n
r
u
n
i
n
f
e
r
e
n
c
e
on
t
h
e
de
t
e
c
t
e
d
f
a
c
e
s
.
-
W
i
n
k
de
t
e
c
t
i
o
n
:
T
h
e
m
o
de
l
w
a
s
n
o
t
t
r
a
i
n
e
d
w
i
t
h
i
m
a
ge
s
of
s
ub
j
e
c
t
s
w
i
t
h
o
n
e
e
y
e
c
l
o
s
e
d
a
n
d
t
h
e
o
t
h
e
r
e
ye
o
pe
n
.
A
da
t
a
-
s
e
t
of
a
t
o
t
a
l
of
400
i
m
a
ge
s
f
r
o
m
di
f
f
e
r
e
n
t
a
ngl
e
s
c
o
n
s
i
s
t
i
n
g
of
t
w
o
c
a
t
e
go
r
i
e
s
w
a
s
m
a
de
.
F
i
r
s
t
l
y
,
t
h
e
l
e
f
t
e
y
e
c
l
o
s
e
d
a
n
d
t
h
e
r
i
g
ht
e
y
e
o
pe
n
.
S
e
c
o
n
dl
y
,
t
h
e
l
e
f
t
e
y
e
o
pe
n
a
n
d
t
h
e
r
i
g
h
t
e
y
e
c
l
o
s
e
d.
T
h
e
H
a
r
dw
a
r
e
n
e
e
de
d
w
a
s
n
o
t
a
v
a
i
l
a
b
l
e
to
c
a
rr
y
o
ut
t
h
e
e
xpe
ri
m
e
n
t
.
S
o
,
t
h
e
us
e
-
c
a
s
e
w
a
s
s
us
pe
n
de
d
f
r
o
m
w
o
r
k.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
IS
S
N
:
2502
-
4752
E
y
e
b
l
i
n
k
de
t
e
c
t
i
on
us
i
ng
CNN
t
o
d
e
t
e
c
t
dr
ow
s
i
ne
s
s
l
e
v
e
l
i
n
dr
i
v
e
r
s
f
or
r
oad
s
a
f
e
t
y
(
P
ot
hur
aj
u
V
i
s
he
s
h
)
229
T
h
e
n
u
m
b
e
r
of
i
m
a
ge
s
a
nd
t
h
e
s
e
l
e
c
t
i
o
n
of
h
y
p
e
r
-
pa
ra
m
e
t
e
r
s
pl
a
y
a
v
i
t
a
l
pa
rt
in
de
c
i
di
ng
t
h
e
a
c
c
ur
a
c
y
of
t
h
e
m
o
de
l
.
T
h
e
r
e
s
ul
t
s
of
us
i
n
g
d
i
f
fe
r
e
nt
h
y
pe
r
-
p
a
r
a
m
e
t
e
r
s
a
r
e
gi
v
e
n
b
e
l
o
w
;
-
D
a
t
a
s
e
t
us
e
d:
c
l
o
s
e
d
e
y
e
s
in
t
h
e
w
i
l
d
(CE
W
)
-
No
of
i
m
a
ge
s
:
1194(c
l
o
s
e
d
e
y
e
s
),
1239
(o
pe
n
e
y
e
s
)
-
T
o
t
a
l
:
24
33i
m
a
ge
s
(
100X
100
)
JPG
-
T
o
t
a
l
n
o
.
of
e
poc
h
s
:
5
-
T
r
a
i
n
i
ng
i
m
a
ge
s
:
21
81
-
T
e
s
t
i
n
g
i
m
a
ge
s
:
252
T
h
e
pr
o
c
e
s
s
of
i
m
pr
o
v
i
n
g
t
h
e
m
o
de
l
pe
r
f
o
r
m
a
n
c
e
a
nd
n
e
a
ri
n
g
t
h
e
o
bj
e
c
t
i
v
e
is
gi
v
e
n
in
t
h
e
t
a
b
l
e
b
e
l
ow
in
T
a
b
l
e
1
to
s
h
o
w
t
h
e
r
e
s
ul
t
o
b
t
a
i
n
e
d
du
r
i
ng
t
r
a
i
ni
n
g
.
F
i
g
u
r
e
12
d
i
s
pl
a
y
s
t
h
e
a
s
s
e
s
s
m
e
nt
of
t
h
e
pr
o
j
e
c
t
e
d
w
o
r
k
w
i
t
h
s
o
m
e
of
t
h
e
p
r
e
v
a
i
l
i
n
g
w
o
r
k
m
e
nt
i
o
n
e
d
in
t
h
e
l
i
t
e
r
a
t
u
r
e
s
u
r
v
e
y
.
T
a
b
l
e
1.
T
h
e
R
e
s
ul
t
s
o
b
t
a
i
n
e
d
du
ri
n
g
t
ra
i
ni
n
g
on
t
h
e
da
t
a
s
e
t
(
CE
W
)
S
l
.
No
M
o
d
e
l
T
ra
i
n
i
n
g
T
i
m
e
O
p
t
i
m
i
z
e
r
A
c
c
u
ra
c
y
1
3
L
a
y
e
r
(
S
e
q
u
e
n
t
i
a
l
)
1
m
i
n
s
p
e
r
e
p
o
c
h
A
d
a
m
34%
2
3
L
a
y
e
r
(
S
e
q
u
e
n
t
i
a
l
)
1
.
3
m
i
n
s
p
e
r
e
p
o
c
h
A
d
a
m
57%
3
2
L
a
y
e
r
(
S
e
q
u
e
n
t
i
a
l
)
1
m
i
n
s
p
e
r
e
p
o
c
h
A
d
a
m
60%
4
M
o
b
i
l
e
N
e
t
V
2
8
m
i
n
s
p
e
r
e
p
o
c
h
A
d
a
m
81%
5
M
o
b
i
l
e
N
e
t
V
2
8
.
3
m
i
n
s
p
e
r
e
p
o
c
h
RM
S
P
r
o
p
95%
6
M
o
b
i
l
e
N
e
t
V
2
8
.
7
m
i
n
s
p
e
r
e
p
o
c
h
RM
S
P
r
o
p
L
o
s
s
:
b
i
n
a
ry
c
ro
s
s
e
n
t
r
o
p
y
97%
7
In
c
e
p
t
i
o
n
40
m
i
n
s
p
e
r
e
p
o
c
h
RM
S
P
r
o
p
82%
F
i
gu
r
e
12
.
Co
m
pa
ri
s
o
n
of
pr
o
po
s
e
d
w
o
r
k
w
i
t
h
t
h
e
p
r
o
j
e
c
t
e
d
w
o
r
k
7.
C
O
N
C
LU
S
I
O
N
T
h
e
e
y
e
b
l
i
n
k
de
t
e
c
t
i
o
n
w
o
r
k
f
i
n
a
l
l
y
s
e
r
v
e
d
its
pu
r
po
s
e
of
de
t
e
c
t
i
n
g
t
h
e
s
t
a
t
e
of
t
h
e
e
y
e
s
in
a
g
i
v
e
n
i
m
a
ge
.
T
h
i
s
w
o
r
k
aims
to
pr
o
v
i
de
a
s
o
l
ut
i
o
n
fo
r
r
e
duc
i
n
g
ro
a
d
a
c
c
i
de
n
t
s
c
a
us
e
d
by
h
um
a
n
e
rr
o
r.
M
a
n
y
e
ye
b
l
i
n
k
de
t
e
c
t
i
o
n
a
ppl
i
c
a
t
i
o
n
s
de
v
e
l
o
pe
d
a
r
e
s
e
r
v
i
n
g
t
hi
s
pu
rpo
s
e
,
t
h
i
s
is
a
b
e
t
t
e
r
m
o
de
l
to
h
e
l
p
i
m
p
r
o
v
e
t
h
e
pe
r
f
o
r
m
a
n
c
e
of
t
h
e
de
t
e
c
t
i
o
n
ra
t
e
a
n
d
i
de
n
t
i
f
y
p
o
s
s
i
b
l
e
t
hr
e
a
t
s
.
T
hi
s
w
o
r
k
p
r
o
v
i
de
s
a
s
o
l
ut
i
o
n
by
i
m
p
l
e
m
e
nt
i
ng
t
h
e
r
e
t
r
a
i
n
i
ng
of
t
h
e
G
oo
gl
e
M
ob
i
l
e
N
e
t
V
2
m
o
de
l
to
i
de
n
t
i
f
y
e
y
e
b
l
i
n
k
.
T
hi
s
w
o
r
k
aims
to
m
a
ke
t
h
e
w
o
r
l
d
a
s
a
f
e
r
pl
a
c
e
on
r
o
a
ds
fo
r
t
ra
v
e
l
e
r
s
a
nd
t
h
e
t
o
t
a
l
a
c
c
ur
a
c
y
a
c
h
i
e
ve
d
w
a
s
97%.
In
f
ut
ur
e
w
o
r
k,
we
t
r
y
to
c
o
n
c
e
n
t
ra
t
e
on
de
t
e
c
t
i
ng
t
h
e
d
r
o
w
s
i
n
e
s
s
in
t
h
e
f
a
c
e
s
w
i
t
h
s
u
n
gl
a
s
s
e
s
,
o
n
l
y
o
n
e
e
y
e
c
l
os
e
d,
a
nd
s
i
de
v
i
e
w
.
A
C
K
N
O
WL
ED
G
E
M
EN
TS
I
w
o
ul
d
l
i
ke
to
t
ha
n
k
CH
R
IS
T
(D
e
e
m
e
d
to
be
U
n
i
v
e
r
s
i
t
y
),
B
e
n
a
gl
u
r
u
in
s
u
ppo
r
t
i
n
g
us
by
pr
o
v
i
di
n
g
t
h
e
e
n
v
i
r
o
nm
e
n
t
to
ge
t
t
h
e
w
o
r
k
do
n
e
.
R
EF
ER
EN
C
ES
[
1]
B
.
C.
T
e
f
f
t
,
"
A
c
ut
e
S
l
e
e
p
D
e
p
r
i
v
a
t
i
o
n
a
nd
R
i
s
k
of
M
o
t
o
r
,
"
V
e
hi
c
l
e
C
r
as
h
I
nv
o
l
v
e
m
e
nt
,
20
16
.
[
2]
D.
H.
B
a
l
l
a
r
d
a
nd
C
h
r
i
s
t
o
ph
e
r
M.
B
r
o
w
n,
"
C
o
m
put
e
r
V
i
s
i
o
n,
"
P
r
e
nt
i
c
e
H
al
l
,
1
982
.
[
3]
T.
H
ua
ng
,
et
a
l
.
,
"
C
o
m
put
e
r
V
i
s
i
o
n:
E
vo
l
ut
i
o
n
a
n
d
P
r
o
m
i
s
e
,
"
19
t
h
C
E
R
N
Sc
ho
ol
of
C
om
pu
t
i
n
g.
G
e
ne
v
a:
C
E
R
,
pp.
21
-
25
,
199
6
,
do
i
:
10.
5
170
/
C
E
R
N
-
1996
-
00
8.
2
1
.
[
4]
M.
S
o
nka
,
et
a
l
.
,
"
I
m
a
g
e
P
r
o
c
e
s
s
i
ng
,
A
na
l
y
s
i
s
a
nd
M
a
c
hi
ne
V
i
s
i
o
n,
"
T
ho
m
s
o
n.
20
08
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
,
V
o
l
.
22
,
N
o
.
1
,
A
p
r
i
l
20
21
:
2
2
2
-
2
3
1
230
[
5]
F.
Y
o
u,
et
al
.,
"
A
R
e
a
l
-
t
i
m
e
D
r
i
v
i
ng
D
r
o
w
s
i
ne
s
s
D
e
t
e
c
t
i
o
n
A
l
g
o
r
i
t
hm
w
i
t
h
I
ndi
v
i
dua
l
D
i
f
f
e
r
e
nc
e
s
C
o
ns
i
de
r
a
t
i
o
n,
"
Spe
c
i
a
l
Se
c
t
i
on
on
A
r
t
i
f
i
c
i
al
I
n
t
e
l
l
i
ge
nc
e
(
ai
)
-
e
m
pow
e
r
e
d
I
nt
e
l
l
i
ge
nt
T
r
ans
por
t
a
t
i
o
n
Sy
s
t
e
m
s
,
I
E
E
E
A
c
c
e
s
s
,
v
o
l
.
7,
pp.
17
9396
-
17
9408
,
D
e
c
20
19
,
do
i
:
10
.
110
9/
A
C
C
E
S
S
.
2
019
.
295
86
67
.
[
6]
J.
A.
S
t
e
r
n
,
et
al
.
,
"
B
l
i
nk
r
a
t
e
:
A
po
s
s
i
b
l
e
M
e
a
s
u
r
e
of
F
a
t
i
g
ue
,
"
H
um
an
F
ac
t
or
s
,
v
o
l
.
36
,
pp.
28
5
-
297
,
19
94
,
do
i
:
10.
1177
/
0
0187
2089
4036
0020
9
.
[
7]
P.
W
o
l
ko
f
f
,
et
al
.
,
"
E
y
e
C
o
m
pl
a
i
n
t
s
in
t
he
O
f
f
i
c
e
E
nv
i
r
o
nm
e
n
t
:
P
r
e
c
o
r
ne
a
l
T
e
a
r
F
i
l
m
I
nt
e
g
r
i
t
y
I
nf
l
ue
nc
e
d
by
E
y
e
B
l
i
n
ki
ng
E
f
f
i
c
i
e
nc
y
,
"
O
c
c
upat
i
ona
l
and
E
n
v
i
r
onm
e
nt
a
l
M
e
di
c
i
ne
,
v
o
l
.
62,
pp.
4
-
12
,
2005
,
do
i
:
10.
1136
/
o
e
m
.
2
004
.
016
030
.
[
8]
S.
P
a
t
e
l
,
et
a
l
.
,
"
E
f
f
e
c
t
of
V
i
s
ua
l
D
i
s
p
l
a
y
U
ni
t
U
s
e
on
B
l
i
nk
R
a
t
e
a
nd
T
e
a
r
S
t
a
bi
l
i
t
y
,
”
O
pt
o
m
e
t
r
y
a
nd
V
i
s
i
o
n
S
c
i
e
nc
e
,
v
o
l
.
6
8,
pp
.
888
-
89
2,
19
91
,
do
i
:
10
.
109
7/
0
0006
324
-
199
11
1000
-
000
10
.
[
9]
Y.
F.
T
s
a
i
,
et
al
.
,
"
T
a
s
k
P
e
r
f
o
r
m
a
nc
e
a
nd
E
y
e
A
c
t
i
v
i
t
y
:
P
r
e
di
c
t
i
ng
B
e
ha
v
i
o
r
R
e
l
a
t
i
ng
to
C
og
ni
t
i
v
e
W
o
r
kl
o
a
d,
"
A
v
i
a
t
i
on
,
S
pac
e
,
an
d
E
nv
i
r
o
nm
e
nt
al
M
e
di
c
i
ne
,
v
o
l
.
7
8,
pp
.
B
176
-
B
185,
2
007
.
[
10]
M.
S
a
nd
l
e
r
,
et
al
.
,
"
M
o
bi
l
e
N
e
t
V
2:
I
n
v
e
r
t
e
d
R
e
s
i
du
a
l
s
a
nd
L
i
ne
a
r
B
o
t
t
l
e
n
e
c
ks
,
”
I
E
E
E
C
on
f
e
r
e
nc
e
on
C
om
put
e
r
V
i
s
i
on
an
d
P
at
t
e
r
n
R
e
c
o
gn
i
t
i
on
(
C
V
P
R
)
,
2018
,
pp
.
4510
-
45
20
,
do
i
:
10.
1
109
/
C
V
P
R
.
2018
.
0
0474
.
[
11]
S.
V
a
ni
a
nd
T.
V.
M.
R
a
o
,
"
A
n
E
xpe
r
i
m
e
nt
a
l
A
ppr
o
a
c
h
T
o
w
a
r
d
s
t
he
P
e
r
f
o
r
m
a
nc
e
A
s
s
e
s
s
m
e
n
t
of
V
a
r
i
o
us
O
pt
i
m
i
z
e
r
s
on
C
o
nvo
l
ut
i
o
na
l
N
e
ur
a
l
N
e
t
w
o
r
k
,
"
3
rd
I
n
t
e
r
na
t
i
on
al
C
on
f
e
r
e
nc
e
on
T
r
e
nds
in
E
l
e
c
t
r
on
i
c
s
and
I
nf
or
m
a
t
i
c
s
(
I
C
O
E
I
)
,
2
019
,
pp.
3
31
-
336
,
do
i
:
10.
11
09/
I
C
O
E
I
.
2019
.
88626
86
.
[
12]
C.
Z
h
a
ng
,
et
a
l
.
,
"
D
r
i
v
e
r
D
r
o
w
s
i
n
e
s
s
D
e
t
e
c
t
i
o
n
U
s
i
ng
M
u
l
t
i
-
C
ha
n
ne
l
S
e
c
o
nd
O
r
de
r
B
l
i
n
d
I
de
nt
i
f
i
c
a
t
i
o
ns
"
,
I
E
E
E
A
c
c
e
s
s
,
v
o
l
.
7,
p
p.
11
829
-
118
43
,
2
019
,
do
i
:
10.
11
09/
A
C
C
E
S
S
.
2019
.
289
1971
.
[
13]
W.
D
e
ng
a
nd
R.
W
u,
"
R
e
a
l
-
T
i
m
e
D
r
i
v
e
r
-
D
r
o
w
s
i
ne
s
s
D
e
t
e
c
t
i
o
n
S
y
s
t
e
m
U
s
i
ng
F
a
c
i
a
l
F
e
a
t
ur
e
s
,
"
I
E
E
E
A
c
c
e
s
s
,
v
o
l
.
7,
pp
.
1187
27
-
1187
38,
2
019
,
do
i
:
10.
11
09/
A
C
C
E
S
S
.
2019
.
29
3666
3
.
[
14]
F
o
uz
i
a
,
et
al
.
,
"
D
r
i
v
e
r
D
r
o
w
s
i
n
e
s
s
D
e
t
e
c
t
i
o
n
S
y
s
t
e
m
B
a
s
e
d
on
V
i
s
ua
l
F
e
a
t
u
r
e
s
,
"
2
nd
i
nt
e
r
n
at
i
on
al
C
o
nf
e
r
e
nc
e
on
I
nv
e
nt
i
v
e
C
om
m
un
i
c
a
t
i
on
and
C
om
pu
t
at
i
on
al
T
e
c
hn
ol
o
gi
e
s
(
I
C
I
C
C
T
)
,
2
018
,
pp
.
1344
-
13
47
,
do
i
:
10.
1109
/
I
C
I
C
C
T
.
2018.
8473
203
.
[
15]
A.
I
s
l
a
m
,
et
al
.
,
"
A
S
t
udy
on
T
i
r
e
dn
e
s
s
A
s
s
e
s
s
m
e
n
t
by
U
s
i
ng
E
y
e
B
l
i
n
k
D
e
t
e
c
t
i
o
n,
"
J
o
ur
n
al
of
E
ng
i
ne
e
r
i
ng
-
U
K
M
,
v
o
l
.
31,
no
.
2,
pp.
2
09
-
214
,
2019
.
[
16]
C.
C
ho
u
,
et
al
.
,
"
B
l
i
nk
D
e
t
e
c
t
i
o
n
U
s
i
ng
F
a
c
i
a
l
L
a
ndm
a
r
ks
B
l
i
nk
D
e
t
e
c
t
o
r
a
nd
M
u
l
t
i
-
L
a
y
e
r
P
e
r
c
e
p
t
r
o
n,
"
N
a
t
i
ona
l
C
om
put
e
r
Sy
m
p
os
i
um
,
2
019
,
ht
t
p:
/
/
dx
.
do
i
.
o
r
g
/
10.
6927
%
2f
N
C
S
.
201
911.
0105
.
[
17]
A.
A.
M
i
a
h
,
et
al
.
,
"
D
r
o
w
s
i
n
e
s
s
D
e
t
e
c
t
i
o
n
U
s
i
ng
E
y
e
-
B
l
i
nk
P
a
t
t
e
r
n
a
nd
M
e
a
n
E
y
e
L
a
ndm
a
r
ks
’
D
i
s
t
a
nc
e
,
"
I
nt
e
r
n
at
i
on
al
J
o
i
n
t
C
onf
e
r
e
nc
e
on
C
om
pu
t
at
i
o
nal
I
nt
e
l
l
i
ge
nc
e
,
Spr
i
nge
r
,
201
9
,
pp
.
111
-
12
1
.
[
18]
D.
D
o
u
a
n
d
Z.
Z
h
a
ng
,
"
B
l
i
n
k
D
e
t
e
c
t
i
o
n
B
a
s
e
d
on
P
i
xe
l
F
l
uc
t
ua
t
i
o
n
R
a
t
i
o
of
E
y
e
I
m
a
g
e
,
"
J
our
n
al
of
P
hy
s
i
c
s
:
C
onf
e
r
e
nc
e
Se
r
i
e
s
,
v
o
l
.
14
63,
no
.
01
2073
,
202
0
,
pp
.
1
-
5
,
do
i
:
10.
10
88/
1
742
-
659
6%
2F
14
53
%
2F
1
%
2F
0120
73
.
[
19]
M.
H.
B
a
c
o
o
ur
,
et
al
.
,
"
C
a
m
e
r
a
-
B
a
s
e
d
E
y
e
B
l
i
nk
D
e
t
e
c
t
i
o
n
A
l
gor
i
t
hm
f
o
r
A
s
s
e
s
s
i
ng
D
r
i
v
e
r
D
r
o
w
s
i
ne
s
s
,
"
I
E
E
E
I
nt
e
l
l
i
ge
nt
V
e
hi
c
l
e
s
S
y
m
po
s
i
um
(
I
V
)
,
pp
.
987
-
99
3,
20
19
,
do
i
:
10
.
11
09/
I
V
S
.
2019
.
88
1387
1
.
[
20]
B.
K.
S
a
v
a
s
,
a
n
d
Y.
B
e
c
e
r
i
kl
i
,
"
R
e
a
l
T
i
m
e
D
r
i
v
e
r
F
a
t
i
g
ue
D
e
t
e
c
t
i
o
n
B
a
s
e
d
on
S
V
M
A
l
g
o
r
i
t
hm
,
"
6
th
I
n
t
e
r
na
t
i
ona
l
C
onf
e
r
e
nc
e
on
C
on
t
r
ol
E
ng
i
ne
e
r
i
n
g
a
nd
I
n
f
o
r
m
at
i
on
T
e
c
hn
ol
o
gy
(
C
E
I
T
)
,
201
9
,
do
i
:
10.
1
109
/
C
E
I
T
.
2018
.
87
5188
6
.
[
21]
F.
Z
ha
ng
,
et
al
.
,
"
D
r
i
v
e
r
F
a
t
i
g
ue
D
e
t
e
c
t
i
o
n
B
a
s
e
d
on
E
y
e
S
t
a
t
e
R
e
c
o
g
ni
t
i
o
n
"
,
I
n
t
e
r
na
t
i
o
nal
C
onf
e
r
e
nc
e
on
M
ac
hi
ne
V
i
s
i
on
an
d
I
n
f
or
m
a
t
i
on
T
e
c
hno
l
og
y
(
C
M
V
I
T
)
,
2
017
,
pp
.
1
05
-
110
,
d
o
i
:
0.
11
09
/
C
M
V
I
T
.
2017
.
25
.
[
22]
Y.
J.
H
a
n
,
e
t
.
a
l
.
,
"
E
f
f
i
c
i
e
nt
E
y
e
-
bl
i
nk
D
e
t
e
c
t
i
o
n
U
s
i
ng
S
m
a
r
t
pho
ne
s
:
A
h
y
br
i
d
A
ppr
o
a
c
h
B
a
s
e
d
on
D
e
e
p
L
e
a
r
ni
ng
,
"
M
obi
l
e
I
nf
or
m
a
t
i
on
Sy
s
t
e
m
,
H
i
n
daw
i
,
v
o
l
.
2
018
,
pp.
1
-
8,
201
8
,
do
i
:
10.
1155
/
20
18
/
692
9762
.
[
23]
B.
M.
K.
K
um
a
r
i
,
et
al
.
,
"
D
e
t
e
c
t
i
o
n
of
D
r
i
v
e
r
D
r
o
w
s
i
ne
s
s
U
s
i
n
g
E
y
e
B
l
i
nk
S
e
n
s
o
r
,
"
I
n
t
e
r
nat
i
o
nal
J
our
n
al
of
E
ngi
ne
e
r
i
n
g
and
T
e
c
hno
l
og
y
,
v
o
l
.
7,
no
.
3.
12
,
p
p.
49
8
-
504
,
2018
,
d
o
i
:
10.
1
4419
/
i
j
e
t
.
v
7i
3
.
12
.
16
167
.
[
24]
O.
S
i
n
ha
,
et
al
.
,
"
D
e
v
e
l
o
pm
e
n
t
of
a
D
r
o
w
s
y
D
r
i
v
e
r
D
e
t
e
c
t
i
o
n
S
y
s
t
e
m
B
a
s
e
d
on
EEG
a
nd
IR
-
B
a
s
e
d
E
y
e
B
l
i
nk
D
e
t
e
c
t
i
o
n
A
na
l
y
s
i
s
,
"
A
dv
an
c
e
s
in
C
om
m
uni
c
a
t
i
o
n,
D
e
v
i
c
e
s
an
d
N
e
t
w
or
k
i
n
g,
v
o
l
.
46
2,
pp
.
313
-
319
,
2018
,
do
i
:
10.
1007
/
9
78
-
981
-
10
-
790
1
-
6_34
.
[
25]
M.
N
g
xa
nde
,
et
al
.,
"
B
i
a
s
R
e
m
e
di
a
t
i
o
n
in
D
r
i
v
e
r
D
r
o
w
s
i
n
e
s
s
D
e
t
e
c
t
i
o
n
S
y
s
t
e
m
s
U
s
i
ng
G
e
ne
r
a
t
i
v
e
A
dv
e
r
s
a
r
i
a
l
N
e
t
w
o
r
ks
,
"
I
E
E
E
A
c
c
e
s
s
,
v
o
l
.
8,
pp
.
5559
2
-
55601
,
202
0
,
do
i
:
10.
11
09/
A
C
C
E
S
S
.
2
020
.
298
1912
.
[
26]
M.
S
u
na
g
a
w
a
,
et
al
.
,
"
C
o
m
pr
e
he
ns
i
v
e
D
r
o
w
s
i
ne
s
s
L
e
v
e
l
D
e
t
e
c
t
i
o
n
M
o
de
l
C
o
m
bi
n
i
ng
M
ul
t
i
m
o
da
l
I
nf
o
r
m
a
t
i
o
n,
"
I
E
E
E
Se
ns
or
s
J
ou
r
na
l
,
v
o
l
.
20,
no
.
7,
pp
.
3
709
-
371
7,
A
pr
i
l
2020
,
d
o
i
:
10.
1
109
/
J
S
E
N
.
20
19
.
296
0158
.
[
27]
K.
L
i
,
et
a
l
.
,
"
A
F
a
t
i
g
ue
D
r
i
v
i
ng
D
e
t
e
c
t
i
o
n
A
l
g
o
r
i
t
hm
B
a
s
e
d
on
F
a
c
i
a
l
M
u
l
t
i
-
F
e
a
t
u
r
e
F
us
i
o
n,
"
I
E
E
E
A
c
c
e
s
s
,
v
o
l
.
8,
pp.
10
1244
-
10
1259
,
J
un
e
2020
,
do
i
:
10.
1
109
/
A
C
C
E
S
S
.
202
0.
2998
3
63
.
[
28]
T.
J
ung
,
et
al
.
,
"
D
e
e
pV
i
s
i
o
n:
D
e
e
pf
a
ke
s
D
e
t
e
c
t
i
o
n
U
s
i
ng
H
um
a
n
E
y
e
B
l
i
nki
ng
P
a
t
t
e
r
n
,
"
I
E
E
E
A
c
c
e
s
s
,
v
o
l
.
8,
pp.
83
144
-
831
54,
A
pr
i
l
202
0
,
h
t
t
p
s
:
/
/
do
i
.
o
r
g
/
10
.
1
109
/
A
C
C
E
S
S
.
202
0.
29
8866
0
.
[
29]
M.
T
o
s
un
a
nd
Ö.
K
a
s
ı
m
,
"
N
o
v
e
l
e
y
e
-
bl
i
nk
a
r
t
e
f
a
c
t
de
t
e
c
t
i
o
n
a
l
g
o
r
i
t
hm
f
r
o
m
r
a
w
EEG
s
i
g
na
l
s
u
s
i
ng
F
C
N
-
ba
s
e
d
s
e
m
a
n
t
i
c
s
e
g
m
e
n
t
a
t
i
o
n
m
e
t
ho
d
,
"
I
E
T
S
i
g
na
l
P
r
o
c
e
s
s
i
ng
,
v
o
l
.
14
,
no
.
8,
pp
.
489
-
4
94,
O
c
t
o
be
r
2
020
,
do
i
:
10.
1049
/
i
e
t
-
s
pr
.
20
19
.
060
2
.
[
30]
A.
A.
J
o
r
da
n,
et
al
.
,
"
D
e
e
p
L
e
a
r
ni
ng
f
o
r
E
y
e
B
l
i
nk
D
e
t
e
c
t
i
o
n
I
m
pl
e
m
e
nt
e
d
at
t
he
E
dg
e
,
"
I
E
E
E
E
m
be
dde
d
Sy
s
t
e
m
s
L
e
t
t
e
r
s
,
p
p.
1
-
1,
O
c
t
o
be
r
2
020
,
do
i
:
10.
110
9/
L
E
S
.
2020
.
30
2931
3
.
[
31]
M.
N
g
xa
n
de
,
et
al
.
,
"
D
r
i
v
e
r
dr
o
w
s
i
ne
s
s
d
e
t
e
c
t
i
o
n
us
i
ng
be
h
a
v
i
o
r
a
l
m
e
a
s
u
r
e
s
a
nd
m
a
c
hi
ne
l
e
a
r
ni
ng
t
e
c
hni
q
ue
s
:
A
r
e
v
i
e
w
of
s
t
a
t
e
-
of
-
a
r
t
t
e
c
hni
q
ue
s
,
"
P
at
t
e
r
n
R
e
c
ogn
i
t
i
o
n
A
s
s
oc
i
a
t
i
o
n
of
Sou
t
h
A
f
r
i
c
a
a
nd
R
ob
ot
i
c
s
and
M
e
c
ha
t
r
on
i
c
s
(
P
R
A
SA
-
R
obM
e
c
h)
,
pp
.
156
-
16
1,
20
17
,
do
i
:
10
.
110
9/
R
o
bo
M
e
c
h.
2
0
17.
8
2611
40
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
do
n
e
s
i
a
n
J
E
l
e
c
E
ng
&
Co
m
p
S
c
i
IS
S
N
:
2502
-
4752
E
y
e
b
l
i
n
k
de
t
e
c
t
i
on
us
i
ng
CNN
t
o
d
e
t
e
c
t
dr
ow
s
i
ne
s
s
l
e
v
e
l
i
n
dr
i
v
e
r
s
f
or
r
oad
s
a
f
e
t
y
(
P
ot
hur
aj
u
V
i
s
he
s
h
)
231
B
I
O
G
R
A
P
H
I
ES
OF
A
U
T
H
O
R
S
M
r
.
P
o
t
h
u
r
aj
u
V
i
s
h
e
s
h
is
c
ur
r
e
nt
l
y
s
t
ud
y
i
ng
in
F
i
na
l
y
e
a
r
B
.
T
e
c
h
in
C
o
m
put
e
r
S
c
i
e
nc
e
a
nd
E
ng
i
ne
e
r
i
ng
in
C
H
R
I
S
T
D
e
e
m
e
d
to
be
U
ni
v
e
r
s
i
t
y
,
B
e
na
g
a
l
u
r
u
.
He
c
o
m
pl
e
t
e
d
I
nt
e
r
s
h
i
p
in
LG
S
o
f
t
,
I
ndi
a
.
H
i
s
i
n
t
e
r
e
s
t
s
i
nc
l
ud
e
A
I
,
D
a
t
a
M
i
n
i
ng
a
nd
D
a
t
a
.
D
r
.
R
agh
av
e
n
d
r
a
S
is
c
ur
r
e
n
t
l
y
w
o
r
ki
ng
as
A
s
s
oc
i
a
t
e
P
r
o
f
e
s
s
o
r
in
t
he
D
e
p
a
r
t
m
e
n
t
of
C
o
m
put
e
r
S
c
i
e
nc
e
a
nd
E
ng
i
ne
e
r
i
ng
at
C
H
R
I
S
T
D
e
e
m
e
d
to
be
U
ni
v
e
r
s
i
t
y
,
B
e
ng
a
l
u
r
u
.
He
c
om
pl
e
t
e
d
hi
s
P
h
.
D
.
d
e
g
r
e
e
in
C
o
m
put
e
r
S
c
i
e
nc
e
a
n
d
E
ng
i
ne
e
r
i
ng
f
r
o
m
V
T
U
,
B
e
l
a
ga
vi
,
I
ndi
a
in
2017
a
nd
ha
s
15
y
e
a
r
s
of
t
e
a
c
hi
ng
e
xpe
r
i
e
nc
e
.
H
i
s
i
n
t
e
r
e
s
t
s
i
nc
l
u
de
A
I
,
D
a
t
a
M
i
n
i
ng
.
s.
M
r
.
S
an
t
o
s
h
K
u
m
ar
J
is
c
ur
r
e
n
t
l
y
w
o
r
ki
m
g
as
A
s
s
o
c
i
a
t
e
P
r
o
f
e
s
s
o
r
in
t
he
de
p
a
r
t
m
e
n
t
of
C
o
m
put
e
r
S
c
i
e
nc
e
a
n
d
e
ng
i
ne
e
r
i
ng
at
K
S
S
E
M
B
e
ng
a
l
ur
u
a
f
f
i
l
i
a
t
e
d
to
V
T
U
B
E
l
a
g
a
v
i
.
He
pur
s
ui
ng
P
h.
D
.
In
C
S
E
at
r
e
s
e
a
r
c
h
c
e
nt
r
e
B
G
S
I
T
BG
N
A
G
A
R
V
T
U
B
e
l
a
g
a
v
i
a
nd
ha
v
e
13
Y
e
a
r
s
of
e
xp
e
r
i
e
nc
e
a
n
d
hi
s
a
r
e
a
of
i
nt
e
r
e
s
t
a
r
e
B
i
g
da
t
a
A
na
l
y
t
i
c
s
.
D
r
.
R
e
k
h
a
V
is
c
u
r
r
e
nt
l
y
w
o
r
ki
ng
as
A
s
s
i
s
t
a
n
t
P
r
o
f
e
s
s
o
r
in
t
he
D
e
pa
r
t
m
e
nt
of
C
o
m
put
e
r
S
c
i
e
nc
e
a
nd
E
ng
i
ne
e
r
i
ng
at
C
H
R
I
S
T
D
e
e
m
e
d
to
be
U
ni
v
e
r
s
i
t
y
,
B
a
ng
a
l
o
r
e
.
S
h
e
c
o
m
pl
e
t
e
d
he
r
P
h.
D
.
d
e
g
r
e
e
in
C
o
m
put
e
r
S
c
i
e
nc
e
a
nd
E
ng
i
n
e
e
r
i
ng
f
r
o
m
V
T
U
,
B
e
l
a
ga
vi
,
I
ndi
a
in
201
7
a
nd
ha
s
15
y
e
a
r
s
of
t
e
a
c
hi
ng
e
xpe
r
i
e
nc
e
.
H
e
r
i
n
t
e
r
e
s
t
s
i
nc
l
ude
C
o
m
put
e
r
N
e
t
w
o
r
ks
a
nd
C
y
be
r
S
e
c
ur
i
t
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.