I
n
t
e
r
n
at
ion
al
Jou
r
n
a
l
of
I
n
f
o
r
m
at
ics
an
d
Com
m
u
n
icat
ion
T
e
c
h
n
ol
ogy
(
I
J
-
I
CT
)
Vo
l
.
1
4
,
N
o
.
2
,
A
ugus
t
20
2
5
,
pp.
645
~
652
I
S
S
N:
2252
-
8776
,
DO
I
:
10
.
11591/i
ji
c
t
.
v
1
4
i
2
.
pp
64
5
-
652
645
Jou
r
n
al
h
o
m
e
page
:
ht
tp:
//
ij
ict
.
iaes
c
or
e
.
c
om
Hu
m
a
n
d
e
t
e
c
t
io
n
i
n
C
C
T
V
sc
r
e
e
n
sh
ot
u
si
n
g f
in
e
-
t
u
n
in
g
V
GG
-
19
F
i
r
d
au
s
Angga
De
wangga,
Abb
a
S
u
ga
n
d
a
Gi
r
s
an
g
D
e
pa
r
tm
e
nt
of
C
o
mput
e
r
,
B
I
N
U
S
G
r
a
dua
t
e
P
r
o
g
r
a
m, M
a
s
te
r
of
I
n
f
o
r
ma
ti
c
s
, B
in
a
N
us
a
nt
a
r
a
U
ni
ve
r
s
it
y
, J
a
ka
r
ta
, I
nd
o
ne
s
ia
Ar
t
ic
l
e
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
i
ve
d
J
u
n
28,
2024
R
e
vi
s
e
d
No
v
20,
2024
A
c
c
e
pt
e
d
De
c
15,
2024
Cl
o
s
e
d
-
c
i
r
cu
i
t
t
el
e
v
i
s
i
o
n
(CCT
V
)
s
y
s
t
em
s
h
a
v
e
g
en
e
rat
e
d
a
v
as
t
am
o
u
n
t
o
f
v
i
s
u
al
d
at
a
c
ru
c
i
a
l
fo
r
s
ecu
ri
t
y
an
d
s
u
r
v
e
i
l
l
a
n
ce
p
u
rp
o
s
e
s
.
E
ff
ec
t
i
v
el
y
c
at
eg
o
r
i
zi
n
g
s
ec
u
r
i
t
y
l
e
v
el
t
y
p
e
s
i
s
v
i
t
al
f
o
r
m
ai
n
t
ai
n
i
n
g
as
s
e
t
s
ec
u
r
i
t
y
e
ff
ec
t
i
v
el
y
.
T
h
i
s
s
t
u
d
y
p
ro
p
o
s
e
s
a
p
rac
t
i
c
a
l
ap
p
ro
a
c
h
f
o
r
cl
as
s
i
f
y
i
n
g
CCT
V
s
c
reen
s
h
o
t
i
m
a
g
e
s
u
s
i
n
g
v
i
s
u
al
g
eo
me
t
r
y
g
ro
u
p
(
V
G
G
-
19
)
t
ran
s
fe
r
l
e
ar
n
i
n
g
,
a
c
o
n
v
o
l
u
t
i
o
n
al
n
e
u
ra
l
n
e
t
w
o
rk
(C
N
N
)
cl
as
s
i
fi
c
at
i
o
n
mo
d
el
t
h
at
w
o
r
k
s
re
a
l
l
y
w
el
l
i
n
i
m
a
g
e
c
l
as
s
i
fi
c
at
i
o
n
.
T
h
e
t
as
k
i
n
cl
as
s
i
fi
c
at
i
o
n
c
o
m
p
ro
mi
s
e
o
f
c
at
eg
o
r
i
zi
n
g
s
c
r
e
en
s
h
o
t
s
i
n
t
o
t
w
o
cl
as
s
e
s
:
“
h
u
m
a
n
s
p
re
s
e
n
t
”
an
d
“
n
o
h
u
m
an
s
p
re
s
e
n
t
.
”
F
i
n
e
-
t
u
n
i
n
g
V
G
G
-
19
m
o
d
el
at
t
ai
n
ed
9
8
%
t
rai
n
i
n
g
a
cc
u
ra
cy
,
9
8
%
v
al
i
d
at
i
o
n
a
cc
u
ra
cy
,
a
n
d
8
5
%
t
e
s
t
accu
ra
cy
f
o
r
t
h
i
s
c
l
as
s
i
fi
c
at
i
o
n
.
T
o
e
v
a
l
u
at
e
i
t
s
p
e
rfo
r
m
a
n
ce
,
w
e
c
o
m
p
ar
e
d
fi
n
e
-
t
u
n
i
n
g
V
G
G
-
19
m
o
d
e
l
w
i
t
h
an
o
t
h
e
r
me
t
h
o
d
.
T
h
e
V
G
G
-
19
-
b
a
s
e
d
fi
n
e
-
t
u
n
i
n
g
mo
d
e
l
d
em
o
n
s
t
rat
e
s
e
ff
ec
t
i
v
en
e
s
s
i
n
h
an
d
l
i
n
g
i
m
a
g
e
s
c
r
een
s
h
o
t
s
,
p
re
s
en
t
i
n
g
a
v
al
u
ab
l
e
t
o
o
l
f
o
r
CCT
V
i
m
a
g
e
cl
as
s
i
fi
c
at
i
o
n
a
n
d
c
o
n
t
ri
b
u
t
i
n
g
t
o
t
h
e
en
h
an
cemen
t
o
f
as
s
e
t
s
e
cu
ri
t
y
s
t
rat
e
g
i
e
s
.
K
e
y
w
o
r
d
s
:
F
i
ne
-
t
uni
n
g
Hu
m
a
n
de
t
e
c
t
i
o
n
I
m
a
ge
c
l
a
s
s
i
f
i
c
a
t
i
o
n
T
r
a
n
s
f
e
r
l
e
a
r
ni
ng
VGG
-
19
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
cen
s
e.
C
or
r
e
s
pon
din
g
A
u
th
or
:
F
i
r
d
a
us
An
gga
De
wa
n
gga
De
pa
r
t
m
e
n
t
o
f
C
o
m
put
e
r
S
c
i
e
n
c
e
,
B
I
NU
S
Gr
a
dua
te
P
r
o
gr
a
m
,
M
a
s
t
e
r
o
f
C
o
m
put
e
r
S
c
i
e
n
c
e
B
i
n
a
Nu
s
a
n
t
a
r
a
Uni
v
e
r
s
i
t
y
J
a
ka
r
t
a
,
11480
,
I
n
do
n
e
s
i
a
E
m
a
i
l
:
f
i
r
da
u
s
.
de
wa
n
gga
@bi
nus
.
a
c
.
i
d
1.
I
NT
RODU
C
T
I
ON
C
l
o
s
e
d
-
c
i
r
c
u
i
t
t
e
l
e
vi
s
i
o
n
(
C
C
T
V)
s
e
r
v
e
s
a
s
a
s
i
t
ua
t
i
o
n
a
l
c
r
i
m
e
pr
e
ve
n
t
i
o
n
(
S
C
P
)
m
e
t
h
o
d
a
i
m
e
d
a
t
e
nh
a
nc
i
n
g
f
o
r
m
a
l
s
ur
v
e
il
l
a
n
c
e
i
n
s
pe
c
if
i
c
a
r
e
a
s
.
S
C
P
e
m
p
h
a
s
i
z
e
s
c
r
i
m
e
de
t
e
r
r
e
n
c
e
by
m
ini
mi
z
i
ng
o
pp
o
r
t
un
i
t
i
e
s
f
o
r
c
r
i
mi
na
l
a
c
t
s
a
n
d
e
l
e
v
a
t
i
n
g
t
h
e
pe
r
c
e
i
v
e
d
l
i
ke
li
h
o
o
d
o
f
b
e
i
ng
c
a
ug
h
t
[
1]
.
Vi
d
e
o
s
ur
v
e
i
ll
a
n
c
e
,
f
a
c
il
i
t
a
t
e
d
by
C
C
T
V
s
y
s
t
e
m
s
,
ha
s
pr
o
v
e
n
t
o
b
e
a
n
e
f
f
e
c
t
i
v
e
m
e
t
h
o
d
f
o
r
m
o
ni
t
o
r
i
n
g
s
pe
c
if
i
c
l
o
c
a
t
i
o
ns
.
A
r
e
a
s
c
a
n
b
e
c
o
n
t
i
n
uo
us
ly
m
o
ni
t
o
r
e
d
a
r
o
un
d
t
h
e
c
l
o
c
k
us
i
n
g
vi
de
o
s
ur
v
e
i
ll
a
nc
e
de
vi
c
e
s
,
w
i
t
h
f
o
ot
a
ge
a
c
c
e
s
s
i
b
l
e
a
s
n
e
e
d
e
d,
c
o
m
p
l
e
m
e
n
t
e
d
by
b
ur
g
l
a
r
a
l
a
r
m
s
y
s
t
e
ms
[
2]
.
B
ur
gl
a
r
a
l
a
r
m
s
y
s
t
e
m
s
ha
v
e
b
e
e
n
a
f
u
n
d
a
m
e
n
t
a
l
c
o
m
po
n
e
n
t
o
f
h
o
m
e
s
e
c
ur
i
t
y
s
y
s
t
e
m
s
f
o
r
m
a
ny
ye
a
r
s
.
H
o
we
v
e
r
,
i
m
p
l
e
m
e
n
t
i
n
g
i
n
t
r
ude
r
b
e
h
a
vi
o
r
de
t
e
c
t
i
o
n
t
e
c
h
ni
que
s
pr
e
s
e
n
t
s
c
h
a
ll
e
n
ge
s
due
to
i
nhe
r
e
n
t
l
im
i
t
a
t
i
o
n
s
s
uc
h
a
s
r
e
s
o
ur
c
e
c
o
n
s
t
r
a
i
n
t
s
a
n
d
m
o
t
i
o
n
a
r
t
i
f
a
c
t
s
[
3]
.
T
h
e
s
e
c
h
a
ll
e
n
g
e
s
i
n
m
o
de
r
n
s
ur
v
e
il
l
a
n
c
e
s
y
s
t
e
m
s
r
e
s
u
l
t
i
n
a
hi
g
h
n
u
m
be
r
o
f
f
a
l
s
e
a
l
a
r
m
s
[
4]
.
T
o
a
ddr
e
s
s
t
hi
s
i
s
s
ue
,
we
pr
o
p
o
s
e
t
h
e
ut
i
l
i
z
a
t
i
o
n
o
f
de
e
p
l
e
a
r
nin
g
a
l
go
r
i
t
hm
s
t
o
r
e
duc
e
f
a
l
s
e
a
l
a
r
m
s
i
n
C
C
T
V
s
c
r
e
e
ns
h
o
t
s
.
I
n
r
e
c
e
n
t
y
e
a
r
s
,
de
e
p
l
e
a
r
ni
ng,
pa
r
t
i
c
u
l
a
r
l
y
c
o
nv
o
l
ut
i
o
n
a
l
n
e
ur
a
l
n
e
t
wo
r
ks
(
C
NN
s
)
,
h
a
s
a
c
hi
e
ve
d
s
i
g
nif
i
c
a
n
t
a
dv
a
nc
e
m
e
n
t
s
i
n
o
bj
e
c
t
c
l
a
s
s
i
f
i
c
a
t
i
o
n
a
n
d
r
e
c
o
gni
t
i
o
n
[
5]
.
C
NN
f
r
a
m
e
wo
r
k
i
s
ut
i
li
z
e
d
i
n
n
u
m
e
r
o
us
a
s
pe
c
t
s
o
f
da
i
ly
l
if
e
,
s
uc
h
a
s
a
n
o
m
a
ly
d
e
t
e
c
t
i
o
n
,
n
a
t
ur
a
l
l
a
n
gua
ge
pr
o
c
e
s
s
i
n
g
(
NL
P
)
,
c
o
m
put
e
r
vi
s
i
o
n
,
t
i
m
e
-
s
e
r
i
e
s
pr
e
d
i
c
t
i
o
n
,
i
m
a
ge
i
de
n
t
i
f
i
c
a
t
i
o
n
,
dr
ug
de
v
e
l
o
pm
e
n
t
,
vi
de
o
e
v
a
l
ua
t
i
o
n
,
h
e
a
l
t
h
r
i
s
k
a
n
a
ly
s
i
s
,
a
n
d
r
e
c
o
m
m
e
n
da
t
i
o
n
s
y
s
t
e
m
s
.
T
hi
s
a
l
l
o
w
s
f
o
r
t
h
e
e
n
c
o
di
n
g
o
f
i
m
a
ge
f
e
a
t
ur
e
s
i
n
t
o
t
h
e
a
r
c
hi
t
e
c
t
ur
e
,
m
a
k
i
ng
t
h
e
s
t
r
uc
tur
e
e
s
pe
c
i
a
ll
y
e
f
f
e
c
t
i
v
e
f
o
r
t
a
s
ks
i
nv
o
l
vi
ng
i
m
a
g
e
s
whil
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8776
I
n
t
J
I
n
f
&
C
o
m
m
u
n
T
e
c
hn
o
l
,
Vo
l
.
1
4
,
N
o.
2
,
A
ugus
t
20
2
5
:
645
-
652
646
mi
n
im
i
z
i
ng
t
h
e
pa
r
a
m
e
t
e
r
s
r
e
qui
r
e
d
to
s
e
t
up
t
h
e
m
o
de
l
[
6]
.
C
NN
s
a
r
e
a
t
y
pe
o
f
f
e
e
d
f
o
r
wa
r
d
ne
t
w
o
r
ks
i
ns
p
i
r
e
d
f
r
o
m
t
h
e
s
t
r
uc
t
ur
e
o
f
a
nim
a
l
vi
s
ua
l
c
o
r
t
e
x
[
7]
.
T
h
e
y
a
r
e
v
a
r
i
a
t
i
o
n
s
o
f
m
u
l
t
il
a
y
e
r
pe
r
c
e
pt
r
on
s
,
w
i
t
h
n
e
ur
o
n
s
a
r
r
a
n
ge
d
i
n
t
h
r
e
e
d
i
m
e
ns
i
o
n
s
,
e
s
t
a
bl
i
s
h
i
n
g
a
l
o
c
a
l
c
o
n
ne
c
t
i
vi
t
y
pa
t
t
e
r
n
a
m
o
n
g
n
e
a
r
by
n
e
ur
o
n
s
a
n
d
s
ha
r
i
n
g
t
h
e
we
i
g
h
t
s
o
f
l
e
a
r
n
e
d
f
il
t
e
r
s
.
T
h
e
a
r
c
hi
t
e
c
t
ur
e
o
f
vi
s
ua
l
ge
o
m
e
t
r
y
gr
o
up
(
VGG
-
19
)
wa
s
i
nf
l
ue
n
c
e
d
b
y
Al
e
x
Ne
t
,
a
C
NN
pr
e
s
e
n
t
e
d
i
n
pr
e
vi
o
us
y
e
a
r
s
’
I
mage
Ne
t
c
o
m
pe
t
i
t
i
o
ns
[
8]
.
VGG
-
19
i
s
a
C
NN
t
h
o
r
o
ugh
19
l
a
y
e
r
s
,
i
nc
l
ud
i
ng
3
f
u
ll
y
c
o
nn
e
c
t
e
d
l
a
y
e
r
s
a
n
d
16
c
o
n
v
o
l
ut
i
o
n
l
a
y
e
r
s
,
de
s
i
g
n
e
d
f
o
r
i
m
a
g
e
s
c
l
a
s
s
if
i
c
a
t
i
o
n
i
n
t
o
1
,
000
o
bj
e
c
t
c
a
t
e
g
or
i
e
s
[
9]
.
An
i
n
-
de
pt
h
i
l
l
u
s
t
r
a
t
i
on
o
f
t
h
e
VGG
-
19
a
r
c
hi
t
e
c
t
ur
e
hi
g
hli
g
h
t
s
t
h
e
s
i
g
nif
ica
n
c
e
o
f
t
h
e
i
ni
t
i
a
l
c
o
nv
o
l
ut
i
o
n
a
l
l
a
y
e
r
s
i
n
c
a
pt
ur
i
n
g
f
u
nda
m
e
n
t
a
l
a
t
tr
i
b
ut
e
s
s
uc
h
a
s
s
ha
pe
s
a
n
d
e
dge
s
,
of
f
e
r
i
ng
a
c
o
m
pr
e
h
e
ns
i
ve
un
de
r
s
t
a
n
d
i
n
g
o
f
t
h
e
m
o
de
l
’
s
e
a
r
l
y
da
t
a
pr
o
gr
e
s
s
i
o
n
f
r
o
m
i
nput
to
o
u
t
pu
t
[
10]
.
T
h
e
f
i
e
l
d
o
f
de
e
p
l
e
a
r
ni
ng,
pa
tt
e
r
n
r
e
c
o
gni
t
i
o
n
,
a
n
d
h
u
m
a
n
-
c
o
m
put
e
r
i
n
t
e
r
a
c
t
i
o
n
h
a
s
ga
r
n
e
r
e
d
s
i
g
nif
i
c
a
n
t
i
n
t
e
r
e
s
t
f
r
o
m
r
e
s
e
a
r
c
h
s
c
i
e
n
t
i
s
t
s
,
wi
t
h
a
m
a
j
o
r
f
o
c
us
o
n
i
m
a
ge
c
l
a
s
s
i
f
i
c
a
t
i
o
n
[
11]
.
T
hi
s
r
e
s
e
a
r
c
h
a
l
s
o
de
l
v
e
s
i
n
t
o
i
m
a
ge
c
l
a
s
s
i
f
i
c
a
t
i
o
n
us
i
ng
t
r
a
n
s
f
e
r
l
e
a
r
ni
ng,
whi
c
h
i
nv
o
l
v
e
s
l
e
v
e
r
a
g
i
ng
kn
o
w
l
e
d
ge
f
r
o
m
a
n
o
t
h
e
r
t
a
s
k
t
h
a
t
h
a
v
e
u
s
e
d
pr
e
-
t
r
a
i
n
e
d
m
o
de
l
.
T
r
a
n
s
f
e
r
l
e
a
r
ni
ng
s
i
g
ni
f
i
c
a
n
t
l
y
e
nha
n
c
e
s
l
e
a
r
ni
ng
pe
r
f
o
r
m
a
n
c
e
by
b
o
r
r
o
wi
n
g
kn
o
w
l
e
dge
a
n
d
l
a
b
e
l
da
t
a
f
r
o
m
r
e
l
a
t
e
d
d
o
m
a
i
ns
a
n
d
ge
t
e
x
t
r
a
c
t
e
d
to
a
s
s
i
s
t
a
m
a
c
hi
ne
l
e
a
r
ni
ng
a
l
go
r
i
t
hm
i
n
a
c
hi
e
vi
ng
b
e
t
t
e
r
pe
r
f
o
r
m
a
n
c
e
i
n
t
h
e
t
a
r
ge
t
d
o
m
a
i
n
[
12]
.
B
y
t
r
a
n
s
f
e
r
l
e
a
r
ni
ng,
we
e
x
t
r
a
c
t
i
n
f
o
r
m
a
t
i
o
n
pa
t
t
e
r
n
s
f
r
o
m
s
c
r
e
e
ns
h
ot
i
m
a
ge
s
,
i
mpr
o
vi
n
g
o
ur
c
l
a
s
s
i
f
i
c
a
t
i
o
n
m
e
t
h
o
d
a
n
d
o
f
f
e
r
i
n
g
v
a
l
ua
bl
e
i
ns
i
g
h
t
s
to
b
oo
s
t
s
e
c
ur
i
t
y
m
a
n
a
ge
m
e
n
t
a
n
d
a
s
s
e
t
pr
e
s
e
r
v
a
t
i
o
n
.
B
y
ut
i
li
z
i
ng
t
r
a
n
s
f
e
r
l
e
a
r
ni
n
g
a
n
d
f
in
e
-
t
uni
ng
t
e
c
h
ni
que
s
,
o
u
r
m
o
de
l
e
f
f
e
c
t
i
ve
l
y
c
a
t
e
go
r
i
z
e
s
i
mage
s
i
n
t
o
t
w
o
c
l
a
s
s
e
s
:
h
u
m
a
n
a
n
d
n
o
n
-
h
u
m
a
n
,
r
e
a
c
hi
ng
a
r
e
m
a
r
ka
bl
e
85%
a
c
c
ur
a
c
y
.
T
hi
s
s
ur
pa
s
s
e
s
a
n
o
t
h
e
r
m
e
t
h
o
d,
s
uc
h
a
s
VG
G
-
16,
whi
c
h
a
c
hi
e
v
e
60%
a
c
c
ur
a
c
y
[
13]
.
T
hi
s
de
m
o
n
s
t
r
a
t
e
s
VGG
-
19
’
s
pr
o
f
i
c
i
e
nc
y
i
n
im
a
ge
c
l
a
s
s
if
i
c
a
t
i
o
n
a
n
d
i
t
s
s
i
g
nif
i
c
a
n
t
pot
e
n
t
i
a
l
t
o
e
nh
a
nc
e
s
a
f
e
t
y
m
a
n
a
ge
m
e
n
t
s
t
r
a
t
e
gi
e
s
.
2.
RE
S
E
AR
CH
M
E
T
HO
D
A
s
y
s
t
e
m
a
t
i
c
s
i
x
-
s
t
e
p
pr
o
c
e
s
s
wa
s
us
e
d
i
n
r
e
s
e
a
r
c
h
m
e
t
h
o
d.
F
i
r
s
t
,
a
n
e
x
t
e
ns
i
ve
s
ur
ve
y
i
s
c
o
n
duc
t
e
d
to
c
o
l
l
e
c
t
a
di
v
e
r
s
e
s
e
t
o
f
C
C
T
V
s
c
r
e
e
n
s
h
o
t
r
e
po
r
t
s
a
n
d
r
e
l
e
va
n
t
i
nf
o
r
m
a
t
i
o
n
.
T
h
e
ga
t
h
e
r
e
d
d
a
t
a
t
h
e
n
un
de
r
go
e
s
de
t
a
i
l
e
d
pr
e
pr
o
c
e
s
s
i
ng
to
i
m
pr
o
v
e
i
t
s
s
u
i
t
a
bil
i
t
y
a
n
d
qua
li
t
y
f
o
r
a
n
a
ly
s
i
s
.
Ne
x
t
,
t
h
e
da
t
a
s
e
t
i
s
c
a
r
e
f
u
ll
y
d
i
v
i
de
d
i
n
t
o
t
r
a
i
ni
ng,
v
a
li
da
t
i
o
n
,
a
n
d
t
e
s
t
da
t
a
s
e
t
,
e
n
s
ur
i
ng
b
a
l
a
nc
e
d
r
e
pr
e
s
e
n
t
a
t
i
o
n
.
T
o
s
i
mp
l
i
f
y
t
h
e
pr
o
c
e
s
s
,
e
a
c
h
da
t
a
s
e
t
i
s
c
a
t
e
go
r
i
z
e
d
i
n
t
o
t
w
o
c
l
a
s
s
e
s
:
h
u
m
a
n
a
n
d
n
o
nh
u
m
a
n
.
M
o
de
l
i
s
t
he
n
t
r
a
i
n
e
d
o
n
t
h
e
l
a
b
e
l
e
d
t
r
a
i
ni
ng
da
t
a
s
e
t
,
f
o
l
l
o
w
e
d
by
a
t
h
o
r
o
ugh
v
a
l
i
da
t
i
o
n
o
f
i
t
s
pe
r
f
o
r
m
a
n
c
e
.
Af
t
e
r
tr
a
i
ni
ng
t
h
e
m
o
de
l
,
we
s
e
t
a
n
d
c
h
a
n
ge
hy
pe
r
pa
r
a
m
e
t
e
r
o
f
t
h
e
m
o
de
l
f
o
r
f
i
ne
-
t
uni
n
g
a
n
d
r
e
t
r
a
i
n
t
h
e
m
o
de
l
w
i
t
h
v
a
li
da
t
i
o
n
da
t
a
s
e
t
.
L
a
s
t
l
y
,
a
n
e
v
a
l
ua
t
i
o
n
p
h
a
s
e
e
x
a
mi
ne
s
t
h
e
m
o
de
l
’
s
pr
e
c
i
s
i
o
n
i
n
c
l
a
s
s
i
f
yi
ng
a
n
d
pr
e
d
i
c
t
i
n
g
h
u
m
a
n
de
t
e
c
t
i
o
n
b
a
s
e
d
o
n
C
C
T
V
s
c
r
e
e
n
s
h
o
t
s
wi
t
h
t
e
s
t
da
t
a
s
e
t
.
T
h
e
c
o
n
c
e
pt
ua
l
f
r
a
m
e
wo
r
k
o
f
t
hi
s
r
e
s
e
a
r
c
h
i
s
il
l
u
s
t
r
a
t
e
d
i
n
F
i
gur
e
1
.
F
i
gur
e
1.
C
o
n
c
e
pt
ua
l
f
r
a
m
e
wo
r
k
h
u
m
a
n
de
t
e
c
t
i
o
n
i
n
C
C
T
V
s
c
r
e
e
n
s
h
o
t
us
i
n
g
f
i
ne
-
t
uni
ng
VGG
-
19
2.
1.
Dat
a
c
ol
l
e
c
t
ion
T
h
e
da
t
a
s
e
t
s
i
z
e
i
s
c
o
n
s
i
de
r
e
d
a
c
r
i
t
i
c
a
l
f
a
c
t
or
a
f
f
e
c
t
i
n
g
t
h
e
pe
r
f
o
r
m
a
n
c
e
o
f
m
o
de
l
i
n
m
a
c
hi
ne
l
e
a
r
ni
ng
[
14]
.
Ou
r
s
t
udy
t
a
ke
s
a
pr
a
c
t
i
c
a
l
a
pp
r
o
a
c
h
to
b
u
i
l
d
i
ng
a
h
u
m
a
n
de
t
e
c
t
i
o
n
c
l
a
s
s
i
f
i
c
a
t
i
o
n
m
o
de
l
.
W
e
s
t
a
r
t
by
c
o
l
l
e
c
t
i
n
g
hi
s
t
o
r
i
c
a
l
r
e
a
l
-
li
f
e
da
t
a
a
nd
da
t
a
f
r
o
m
t
h
e
K
a
gg
l
e
we
bs
i
t
e
,
s
pe
c
if
i
c
a
ll
y
t
he
“
Hu
m
a
n
De
t
e
c
t
i
o
n
Da
t
a
s
e
t
”
by
Ve
r
n
e
r
[
15]
,
f
o
c
us
i
ng
on
C
C
T
V
i
m
a
ge
s
c
r
e
e
n
s
h
o
t
s
r
e
l
a
t
e
d
to
h
u
m
a
n
de
t
e
c
t
i
o
n
.
W
e
t
h
e
n
r
e
f
i
ne
t
h
e
da
t
a
s
e
t
to
i
n
c
l
ude
o
nly
s
c
r
e
e
ns
h
o
t
s
pe
r
t
i
ne
n
t
to
h
u
m
a
n
de
t
e
c
t
i
o
n
.
R
e
c
o
gniz
i
n
g
t
h
e
im
po
r
t
a
n
c
e
o
f
da
t
a
s
e
t
s
i
z
e
,
o
u
r
g
o
a
l
i
s
to
a
s
s
e
s
s
i
t
s
pr
a
c
t
i
c
a
l
i
n
f
l
ue
n
c
e
o
n
t
h
e
a
c
c
ur
a
c
y
o
f
c
l
a
s
s
if
i
c
a
t
i
o
n
by
o
ur
m
a
c
hi
ne
l
e
a
r
ni
ng
m
o
de
l
f
o
r
de
t
e
c
t
i
n
g
h
u
m
a
n
s
i
n
s
c
r
e
e
ns
h
o
t
s
.
T
h
e
da
t
a
c
o
l
l
e
c
t
i
o
n
pr
o
c
e
s
s
b
e
ga
n
by
a
c
c
e
s
s
i
ng
t
h
e
K
a
gg
l
e
w
e
b
s
i
t
e
,
s
pe
c
if
i
c
a
ll
y
t
a
r
ge
t
i
n
g
413
da
t
a
s
e
t
c
l
us
t
e
r
s
r
e
l
a
t
e
d
to
h
u
m
a
n
d
e
t
e
c
t
i
o
n
.
T
hr
o
ugh
a
t
h
o
r
o
u
gh
c
o
l
l
e
c
t
i
o
n
e
f
f
o
r
t,
405
da
t
a
p
o
i
n
t
s
we
r
e
g
a
t
h
e
r
e
d,
r
e
s
u
l
t
i
n
g
i
n
a
c
o
m
pr
e
h
e
ns
i
ve
da
t
a
s
e
t
c
o
n
s
i
s
t
i
n
g
o
f
2,
820
i
n
s
t
a
n
c
e
s
.
T
o
f
ur
t
h
e
r
e
n
h
a
nc
e
t
h
e
da
t
a
s
e
t
,
hi
s
to
r
i
c
a
l
r
e
a
l
-
li
f
e
da
t
a
wa
s
a
dde
d,
b
r
i
n
g
i
ng
t
h
e
tot
a
l
to
3
,
046
im
a
ge
s
.
T
hi
s
e
n
r
i
c
h
e
d
da
t
a
s
e
t
s
e
r
v
e
s
a
s
a
s
t
r
o
n
g
f
o
un
da
t
i
o
n
f
o
r
i
n
-
de
pt
h
a
n
a
ly
s
i
s
a
n
d
e
x
p
l
o
r
a
t
i
o
n
o
f
v
a
r
i
o
us
a
s
pe
c
t
s
o
f
h
u
m
a
n
de
t
e
c
t
i
o
n
,
o
f
f
e
r
i
ng
va
l
u
a
bl
e
i
ns
i
g
h
t
s
f
o
r
p
r
e
di
c
t
i
ve
de
c
i
s
i
o
n
-
m
a
k
i
ng.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t
J
I
n
f
&
C
o
m
m
u
n
T
e
c
hn
o
l
I
S
S
N:
2252
-
8776
Hum
an
de
tec
ti
on
in
C
C
T
V
s
c
r
e
e
ns
hot
us
ing
f
ine
-
tu
ning
V
GG
-
19
(
F
ir
daus
A
ngga
De
w
angga
)
647
2.
2.
Dat
a
p
r
e
p
r
o
c
e
s
s
in
g
A
c
r
uc
i
a
l
pr
e
li
m
i
na
r
y
s
t
e
p
b
e
f
o
r
e
e
m
p
l
o
yi
ng
m
a
c
hi
ne
l
e
a
r
ni
ng
t
e
c
hni
que
s
i
s
u
s
i
n
g
d
a
t
a
pr
e
pr
o
c
e
s
s
i
ng
[
16]
.
Du
r
i
n
g
da
t
a
pr
e
pa
r
a
t
i
o
n
,
we
un
de
r
t
a
ke
f
u
n
d
a
m
e
n
t
a
l
pr
e
pr
o
c
e
s
s
i
n
g
t
a
s
ks
.
T
hi
s
i
nv
o
l
ve
s
e
l
im
i
na
t
i
n
g
i
r
r
e
l
e
v
a
n
t
C
C
T
V
s
c
r
e
e
n
s
h
o
t
s
to
e
n
s
ur
e
i
m
a
ge
c
l
e
a
nl
i
ne
s
s
a
n
d
s
t
r
uc
t
ur
a
l
i
n
t
e
gr
i
t
y
.
W
e
a
l
s
o
i
nc
l
ude
C
C
T
V
i
m
a
ge
s
c
a
pt
ur
e
d
un
de
r
v
a
r
i
o
us
l
i
g
h
t
i
n
g
c
o
n
d
i
t
i
o
n
s
,
r
a
n
g
i
n
g
f
r
o
m
l
o
w
b
r
i
g
h
t
n
e
s
s
a
t
ni
g
ht
to
f
u
l
l
b
r
i
g
h
t
n
e
s
s
i
n
da
yli
g
h
t
,
t
h
e
r
e
by
e
nh
a
nc
i
n
g
o
v
e
r
a
l
l
da
t
a
c
o
n
s
i
s
t
e
n
c
y
.
A
dd
i
t
i
o
n
a
ll
y
,
we
f
il
t
e
r
o
ut
bl
ur
r
y
,
c
r
o
ppe
d,
a
n
d
i
r
r
e
l
e
v
a
n
t
i
m
a
ge
s
,
e
n
s
ur
i
ng
t
h
a
t
our
da
t
a
s
e
t
i
s
de
v
o
i
d
o
f
e
x
t
r
a
n
e
o
us
e
l
e
m
e
n
t
s
t
h
a
t
c
o
ul
d
pot
e
n
t
i
a
ll
y
d
i
s
r
upt
o
u
r
a
n
a
l
y
s
i
s
.
A
c
r
uc
i
a
l
p
a
r
t
o
f
t
h
e
da
t
a
a
n
a
l
y
s
i
s
pr
o
c
e
dur
e
i
s
d
a
t
a
pr
o
c
e
s
s
i
n
g,
whi
c
h
o
f
t
e
n
de
m
a
n
d
s
s
u
b
s
t
a
n
t
i
a
l
e
f
f
o
r
t
a
n
d
t
i
m
e
[
17]
.
I
n
t
h
e
pr
e
pr
o
c
e
s
s
i
n
g
ph
a
s
e
,
t
h
e
da
t
a
s
e
t
wa
s
pr
e
pa
r
e
d
us
i
n
g
t
h
e
i
m
a
ge
da
t
a
ge
n
e
r
a
t
o
r
li
b
r
a
r
y
f
r
o
m
T
e
n
s
o
r
F
l
o
w.
E
a
c
h
p
i
xe
l
v
a
l
ue
i
n
t
h
e
im
a
ge
s
wa
s
s
c
a
l
e
d
t
o
1/255
,
e
n
s
ur
i
n
g
u
nif
o
r
m
i
t
y
a
c
r
o
s
s
t
h
e
da
t
a
s
e
t.
A
dd
i
t
i
o
n
a
ll
y
,
t
h
e
i
m
a
ge
s
we
r
e
r
e
s
i
z
e
d
to
150
×
150
×
3
di
m
e
ns
i
o
n
s
,
a
s
t
e
p
a
ppl
i
e
d
c
o
n
s
i
s
t
e
n
t
l
y
t
o
t
h
e
t
r
a
i
ni
ng,
v
a
l
i
d
a
t
i
o
n
,
a
n
d
t
e
s
t
da
t
a
s
e
t
s
.
2.
3.
Dat
as
e
t
d
ivi
s
ion
I
n
t
hi
s
pha
s
e
o
f
r
e
s
e
a
r
c
h
,
we
pa
r
t
i
t
i
o
n
da
t
a
s
e
t
i
n
t
o
t
h
r
e
e
da
t
a
s
e
t
s
:
a
tr
a
i
ni
ng
da
t
a
s
e
t
,
c
o
m
pr
i
s
i
ng
77%
da
t
a
,
a
v
a
l
i
da
t
i
o
n
da
t
a
s
e
t
,
c
o
n
s
t
i
t
ut
i
n
g
19%
da
t
a
a
nd
t
e
s
t
da
t
a
s
e
t
wi
t
h
4%
da
t
a
.
T
h
e
tr
a
i
ni
ng
da
t
a
s
e
t
a
i
ds
i
n
t
h
e
m
o
de
l
’
s
l
e
a
r
ni
ng
pr
o
c
e
s
s
a
n
d
i
t
s
pe
r
f
o
r
m
a
n
c
e
t
o
a
c
c
ur
a
t
e
l
y
c
l
a
s
s
if
y
im
a
ge
[
18]
.
T
h
e
a
c
c
ur
a
c
y
o
f
t
h
e
m
o
de
l
c
a
n
b
e
o
bj
e
c
t
i
v
e
ly
e
va
l
ua
t
e
d
us
i
n
g
t
h
e
v
a
li
da
t
i
o
n
da
t
a
s
e
t
,
pr
o
vi
d
i
ng
i
ns
i
g
h
t
s
i
n
t
o
r
e
a
l
-
wo
r
l
d
c
l
a
s
s
if
i
c
a
t
i
o
n
r
e
s
u
l
t
s
[
19]
.
A
n
d
t
e
s
t
da
t
a
s
e
t
wi
ll
be
us
e
d
a
s
t
e
s
t
i
ng
t
h
e
a
c
c
ur
a
c
y
o
f
t
h
e
m
o
de
l
i
n
r
e
a
l
wo
r
l
d
s
c
e
n
a
r
i
o
.
T
hi
s
d
i
vi
s
i
o
n
i
s
i
m
po
r
t
a
n
t
f
o
r
s
t
a
t
i
s
t
i
c
a
l
a
n
a
ly
s
is
a
n
d
m
a
c
hi
ne
l
e
a
r
ni
ng,
a
s
t
h
e
m
o
de
l
u
s
e
d
t
h
e
t
r
a
i
ni
ng
da
t
a
to
b
u
i
l
d
a
n
d
t
r
a
i
n
t
h
e
m
o
de
l
s
,
whil
e
t
h
e
m
o
de
l
us
e
d
t
h
e
v
a
li
da
t
i
o
n
da
t
a
to
e
n
c
o
ur
a
ge
m
o
de
l
’
s
pa
r
a
m
e
t
e
r
s
i
n
f
i
ne
-
t
uni
ng
[
20]
.
T
hi
s
pa
r
t
i
t
i
o
ni
ng
e
ns
ur
e
s
a
c
c
ur
a
c
y
a
n
d
i
n
t
e
gr
i
t
y
o
f
t
h
e
f
o
l
l
o
w
i
ng
m
o
de
li
ng
pr
o
c
e
s
s
e
s
a
n
d
a
n
a
ly
s
i
s
,
r
e
s
u
l
t
i
n
g
i
n
r
e
l
i
a
bl
e
o
u
t
c
o
m
e
s
a
n
d
i
n
s
i
g
h
t
s
.
E
a
c
h
da
t
a
s
e
t
c
o
n
t
a
i
ni
ng
t
w
o
c
l
a
s
s
e
s
,
t
r
a
i
ni
ng
h
a
ve
h
u
m
a
n
c
l
a
s
s
(1
,
135
i
m
a
ge
s
)
a
n
d
n
o
nh
u
m
a
n
c
l
a
s
s
(
1,
222
i
m
a
ge
s
)
,
v
a
li
da
t
i
o
n
h
a
ve
h
u
m
a
n
c
l
a
s
s
(
288
i
m
a
ge
s
)
a
n
d
n
o
nh
u
m
a
n
c
l
a
s
s
(
297
i
m
a
ge
s
)
,
a
n
d
t
e
s
t
h
a
ve
h
u
m
a
n
c
l
a
s
s
(
52
i
m
a
ge
s
)
a
n
d
n
o
nh
u
man
c
l
a
s
s
(
52
i
m
a
ge
s
)
.
T
hi
s
da
t
a
s
p
l
i
t
g
i
v
e
s
a
b
a
l
a
nc
e
d
r
e
pr
e
s
e
n
t
a
t
i
o
n
f
o
r
m
o
de
l
t
r
a
i
ni
ng,
v
a
li
da
t
i
o
n
a
n
d
t
e
s
t
i
n
g,
e
nh
a
nc
i
n
g
t
h
e
m
o
de
l
’
s
pr
e
d
i
c
t
i
v
e
c
a
pa
bil
i
t
i
e
s
a
n
d
th
e
r
o
b
us
t
n
e
s
s
o
f
t
h
e
r
e
s
e
a
r
c
h
f
i
nd
i
n
g
s
.
2.
4.
M
od
e
l
t
r
ain
in
g
I
n
o
ur
tr
a
i
ni
ng
pr
o
c
e
s
s
,
we
ut
i
li
z
e
VGG
-
19
,
a
C
NN
m
o
de
l
t
r
a
i
ne
d
o
n
o
bj
e
c
t
de
t
e
c
t
i
o
n
a
n
d
im
a
ge
c
l
a
s
s
if
i
c
a
t
i
o
n
da
t
a
,
whi
c
h
pr
e
s
e
n
t
s
n
u
m
e
r
o
us
b
e
ne
f
i
t
s
f
o
r
pr
o
c
e
s
s
i
n
g
i
m
a
ge
da
t
a
s
e
t
s
[
21]
.
W
i
t
h
pr
e
t
r
a
i
ni
ng
a
n
d
f
i
ne
-
t
uni
ng,
t
hi
s
m
o
de
l
i
s
c
a
pa
bl
e
o
f
d
i
s
c
e
r
ni
ng
i
n
t
r
i
c
a
t
e
pa
tt
e
r
n
s
a
n
d
f
e
a
t
ur
e
s
w
i
t
hi
n
i
m
a
g
e
s
[
22]
.
T
h
e
t
r
a
i
ni
ng
pr
o
c
e
s
s
s
t
a
r
t
s
wi
t
h
tr
a
n
s
f
e
r
l
e
a
r
ni
ng
,
u
t
i
li
z
i
ng
pr
e
-
t
r
a
i
n
e
d
I
m
a
ge
Ne
t
we
i
g
h
t
s
a
n
d
f
r
e
e
z
i
n
g
a
l
l
l
a
y
e
r
s
e
x
c
e
pt
t
h
e
f
u
ll
y
c
o
nne
c
t
e
d
l
a
y
e
r
.
T
h
e
m
o
de
l
i
s
t
h
e
n
t
r
a
i
ne
d
o
n
t
h
e
t
r
a
i
ni
ng
da
t
a
s
e
t
wi
t
h
a
l
e
a
r
ni
ng
r
a
t
e
o
f
0.
001,
b
a
t
c
h
s
i
z
e
64
a
n
d
25
e
po
c
h
.
A
f
t
e
r
t
hi
s
,
v
a
l
i
da
t
i
o
n
i
s
pe
r
f
o
r
m
e
d
us
i
n
g
20%
s
p
l
i
t
o
f
t
h
e
t
r
a
i
ni
ng
da
t
a
s
e
t
to
e
v
a
l
ua
t
e
t
h
e
m
o
de
l
’
s
pe
r
f
o
r
m
a
n
c
e
a
n
d
to
r
e
f
i
ne
i
t
s
a
bi
li
t
y
i
n
pr
e
d
i
c
t
i
n
g
h
u
m
a
n
de
t
e
c
t
i
o
n
w
i
t
hin
s
c
r
e
e
n
s
h
o
t
s
.
T
hi
s
v
a
li
da
t
i
o
n
he
l
p
s
i
n
gu
i
d
i
ng
t
h
e
m
o
de
l
’
s
d
i
r
e
c
t
i
o
n
a
n
d
i
m
pr
o
vi
ng
i
t
s
pr
e
d
i
c
t
i
v
e
a
c
c
ur
a
c
y
.
T
h
e
VGG
-
19
m
o
de
l
,
pa
r
t
o
f
t
h
e
VG
G
f
a
mi
ly
,
c
o
ns
i
s
t
s
o
f
19
l
a
y
e
r
s
i
n
tot
a
l
:
16
c
o
n
v
o
l
ut
i
o
n
a
l
l
a
y
e
r
s
,
5
M
a
x
P
o
o
l
l
a
y
e
r
s
,
1
f
u
ll
y
c
o
nn
e
c
t
e
d
l
a
y
e
r
,
a
n
d
1
S
o
f
t
M
a
x
l
a
y
e
r
.
I
t
i
nv
o
l
v
e
s
19.
6
bil
li
o
n
f
l
o
a
t
i
n
g
po
i
n
t
o
pe
r
a
t
i
o
n
s
(
F
L
OPs
)
.
T
h
e
VGG
-
19
m
o
de
l
i
s
s
i
m
p
li
f
i
e
d
a
n
d
m
a
de
m
o
r
e
e
f
f
e
c
t
i
ve
by
s
t
a
c
ki
ng
3
×
3
c
o
n
v
o
l
ut
i
o
n
a
l
l
a
y
e
r
s
t
o
i
nc
r
e
a
s
e
i
t
s
de
pt
h
[
23]
,
[
2
4]
.
I
n
VGG
-
19
,
M
a
x
P
oo
l
i
ng
l
a
y
e
r
s
a
r
e
us
e
d
t
o
r
e
duc
e
t
h
e
s
i
z
e
o
f
v
o
l
u
m
e
,
a
n
d
o
nly
o
ne
f
u
ll
y
c
o
nn
e
c
t
e
d
(
F
C
)
l
a
y
e
r
i
s
e
m
p
l
o
y
e
d.
T
h
e
m
o
de
l
o
b
t
a
i
n
s
a
150
×
15
0
×
3
r
e
d,
gr
e
e
n
,
a
n
d
bl
ue
(
R
G
B
)
i
m
a
ge
a
s
i
nput
.
I
n
C
NN
,
t
h
e
r
e
a
r
e
f
o
ur
pr
i
m
a
r
y
l
a
y
e
r
s
:
c
o
nv
o
l
ut
i
o
n
,
poo
l
i
ng,
r
e
c
t
i
f
i
e
d
li
ne
a
r
uni
t
(
R
e
L
U
)
,
a
n
d
f
u
ll
y
c
o
nn
e
c
t
e
d
to
e
x
t
r
a
c
t
i
nf
o
r
m
a
t
i
o
n
f
r
o
m
a
n
im
a
ge
.
T
h
e
c
o
n
v
o
l
ut
i
o
n
l
a
y
e
r
us
e
s
s
e
v
e
r
a
l
f
e
a
t
ur
e
f
il
t
e
r
s
t
o
pe
r
f
o
r
m
c
o
nv
o
l
ut
i
o
n
,
c
o
m
pa
r
i
ng
s
m
a
ll
s
e
c
t
i
o
n
s
o
f
l
a
r
ge
r
i
m
a
ge
s
t
o
c
l
a
s
s
if
y
t
h
e
m
.
T
hi
s
pr
o
c
e
s
s
i
nv
o
l
v
e
s
a
l
i
g
ni
ng
t
h
e
f
e
a
t
ur
e
f
il
t
e
r
w
i
t
h
i
m
a
ge
,
m
u
l
t
i
p
lyi
ng
c
o
r
r
e
s
po
n
d
i
n
g
p
i
xe
l
s
,
s
u
m
mi
n
g
t
h
e
s
e
v
a
l
ue
s
,
a
n
d
d
i
v
i
d
i
ng
by
t
h
e
t
ot
a
l
n
u
m
be
r
o
f
p
i
xe
l
s
to
o
b
t
a
i
n
f
il
t
e
r
e
d
im
a
ge
’
s
f
i
na
l
v
a
l
u
e
.
T
hi
s
pr
o
c
e
s
s
i
s
r
e
c
i
pr
o
c
a
t
e
d
a
c
r
o
s
s
i
m
a
ge
to
o
b
t
a
i
n
c
o
n
v
o
l
ut
i
o
n
o
u
t
pu
t
f
o
r
e
a
c
h
f
e
a
t
ur
e
f
i
l
t
e
r
.
R
e
L
U,
a
r
e
c
t
i
f
i
e
d
l
i
ne
a
r
a
c
t
i
v
a
t
i
o
n
f
u
nc
t
i
o
n
,
o
u
t
pu
t
s
t
h
e
i
n
put
di
r
e
c
t
l
y
if
i
t
i
s
po
s
i
t
i
v
e
a
n
d
z
e
r
o
ot
h
e
r
wi
s
e
.
I
t
i
s
t
h
e
de
f
a
u
l
t
a
c
t
i
v
a
t
i
o
n
f
u
n
c
t
i
o
n
f
o
r
m
a
ny
n
e
ur
a
l
n
e
t
wo
r
ks
due
to
i
t
s
s
upe
r
i
o
r
s
pe
e
d
a
n
d
pe
r
f
o
r
m
a
n
c
e
.
R
e
L
U
i
s
a
pp
l
i
e
d
to
a
l
l
r
e
c
t
i
f
i
e
d
f
e
a
t
ur
e
m
a
ps
(
f
e
a
t
ur
e
i
m
a
ge
s
)
t
o
c
r
e
a
t
e
t
h
e
f
i
na
l
o
u
t
pu
t.
Dur
i
n
g
po
o
l
i
ng,
a
n
i
n
i
t
i
a
l
w
i
ndo
w
s
i
z
e
i
s
s
e
l
e
c
t
e
d,
a
n
d
t
hi
s
w
i
ndo
w
m
o
v
e
s
a
c
r
o
s
s
i
m
a
ge
t
h
a
t
h
a
v
e
b
e
e
n
f
il
t
e
r
e
d,
g
i
vi
ng
m
a
xim
u
m
v
a
l
ue
f
r
o
m
e
a
c
h
s
e
c
t
i
o
n
.
F
l
a
t
t
e
n
l
a
y
e
r
t
h
e
n
f
l
a
t
t
e
n
e
d
t
h
e
poo
l
e
d
f
e
a
t
ur
e
m
a
p
a
n
d
s
e
n
t
t
h
e
i
nf
o
r
m
a
t
i
o
n
t
o
de
n
s
e
l
a
y
e
r
o
r
f
u
l
ly
c
o
nn
e
c
t
e
d
l
a
y
e
r
.
T
hi
s
l
a
y
e
r
pe
r
f
o
r
m
s
t
h
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
pr
o
c
e
s
s
a
n
d
ge
ne
r
a
t
e
s
t
h
e
f
i
na
l
o
u
t
pu
t
us
i
n
g
t
h
e
S
o
f
t
M
a
x
a
c
t
i
v
a
t
i
o
n
f
u
n
c
t
i
o
n
.
S
o
ft
M
a
x
,
a
t
y
pe
o
f
pr
o
b
a
bil
i
s
t
i
c
l
o
g
i
s
t
i
c
r
e
gr
e
s
s
i
o
n
,
pr
o
duc
e
s
a
pr
o
b
a
bil
i
t
y
d
i
s
t
r
i
b
ut
i
o
n
b
a
s
e
d
o
n
a
gi
ve
n
s
e
t
o
f
va
l
u
e
s
.
VGG
-
19
l
a
y
e
r
s
m
o
de
l
a
r
e
i
ll
us
t
r
a
t
e
d
i
n
T
a
bl
e
1.
On
c
e
t
h
e
m
o
de
l
ge
n
e
r
a
t
e
s
pr
e
d
i
c
t
i
o
n
s
f
o
r
h
u
m
a
n
de
t
e
c
t
i
o
n
,
we
us
e
t
e
s
t
da
t
a
s
e
t
to
m
e
a
s
ur
e
t
h
e
a
c
c
ur
a
c
y
o
f
t
h
e
t
r
a
n
s
f
e
r
l
e
a
r
ni
ng
a
ppr
o
a
c
h
.
T
hi
s
e
v
a
l
ua
t
i
o
n
h
e
l
p
s
de
t
e
r
m
i
ne
h
o
w
we
l
l
t
he
m
o
de
l
p
e
r
f
o
r
m
s
in
r
e
a
l
-
wo
r
l
d
s
c
e
n
a
r
i
o
s
,
pr
o
v
i
d
i
ng
i
ns
i
g
h
t
i
n
t
o
i
t
s
pr
a
c
t
i
c
a
l
e
f
f
e
c
t
i
ve
n
e
s
s
.
A
dd
i
t
i
o
n
a
ll
y
,
t
hi
s
pr
o
c
e
s
s
e
s
t
a
bli
s
h
e
s
a
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8776
I
n
t
J
I
n
f
&
C
o
m
m
u
n
T
e
c
hn
o
l
,
Vo
l
.
1
4
,
N
o.
2
,
A
ugus
t
20
2
5
:
645
-
652
648
b
a
s
e
li
ne
a
c
c
ur
a
c
y
,
w
hi
c
h
s
e
r
v
e
s
a
s
a
r
e
f
e
r
e
n
c
e
f
o
r
f
ur
t
h
e
r
f
i
ne
-
t
uni
ng
t
h
e
m
o
de
l
.
B
y
a
s
s
e
s
s
i
ng
t
h
e
m
o
de
l
i
n
t
hi
s
wa
y
,
we
c
a
n
i
de
n
t
i
f
y
a
r
e
a
s
f
o
r
i
m
pr
o
v
e
m
e
n
t
a
n
d
o
p
t
i
mi
z
e
i
t
s
o
v
e
r
a
ll
pe
r
f
o
r
m
a
n
c
e
.
T
a
bl
e
1.
A
r
c
hi
t
e
c
t
ur
e
l
a
y
e
r
s
VGG
-
19
m
o
de
l
L
a
y
e
r
O
ut
put
s
ha
pe
I
nput
_1 (
I
nput
L
a
y
e
r
)
(
[
N
o
ne
, 150, 150, 3
)
]
B
l
oc
k1_c
o
n
v
1
(
C
o
n
v
2D
)
(
N
o
n
e
, 150, 150, 64)
B
l
oc
k1_c
o
n
v
1
(
C
o
n
v
2D
)
(
N
o
n
e
, 150, 150, 64)
B
l
oc
k1_p
oo
l
(
M
a
xP
oo
li
ng2D)
(
N
o
n
e
, 75, 75, 64)
B
l
oc
k2_c
o
n
v
1
(
C
o
n
v
2D
)
(
N
o
n
e
, 75, 75, 128)
B
l
oc
k2_c
o
n
v
2
(
C
o
n
v
2D
)
(
N
o
n
e
, 75, 75, 128)
B
l
oc
k2_p
oo
l
(
M
a
xP
oo
li
ng2D)
(
N
o
n
e
, 37, 37, 128)
B
l
oc
k3_c
o
n
v
1
(
C
o
n
v
2D
)
(
N
o
n
e
, 37, 37, 256)
B
l
oc
k3_c
o
n
v
2
(
C
o
n
v
2D
)
(
N
o
n
e
, 37, 37, 256)
B
l
oc
k3_c
o
n
v
3
(
C
o
n
v
2D
)
(
N
o
n
e
, 37, 37, 256)
B
l
oc
k3_c
o
n
v
4
(
C
o
n
v
2D
)
(
N
o
n
e
, 37, 37, 256)
B
l
oc
k3_p
oo
l
(
M
a
xP
oo
li
ng2D)
(
N
o
n
e
, 18, 18, 256)
B
l
oc
k4_c
o
n
v
1
(
C
o
n
v
2D
)
(
N
o
n
e
, 18, 18, 512)
B
l
oc
k4_c
o
n
v
2
(
C
o
n
v
2D
)
(
N
o
n
e
, 18, 18, 512)
B
l
oc
k4_c
o
n
v
3
(
C
o
n
v
2D
)
(
N
o
n
e
, 18, 18, 512)
B
l
oc
k4_c
o
n
v
4
(
C
o
n
v
2D
)
(
N
o
n
e
, 18, 18, 512)
B
l
oc
k4_p
oo
l
(
M
a
xP
oo
li
ng2D)
(
N
o
n
e
, 9, 9, 512)
B
l
oc
k5_c
o
n
v
1
(
C
o
n
v
2D
)
(
N
o
n
e
, 9, 9,
512)
B
l
oc
k5_c
o
n
v
2
(
C
o
n
v
2D
)
(
N
o
n
e
, 9, 9, 512)
B
l
oc
k5_c
o
n
v
3
(
C
o
n
v
2D
)
(
N
o
n
e
, 9, 9, 512)
B
l
oc
k5_c
o
n
v
4
(
C
o
n
v
2D
)
(
N
o
n
e
, 9, 9, 512)
B
l
oc
k5_p
oo
l
(
M
a
xP
oo
li
ng2D)
(
N
o
n
e
, 4, 4, 512)
F
la
tt
e
n (
F
la
tt
e
n)
(
N
o
n
e
, 25088)
D
e
ns
e
(
D
e
ns
e
)
(
N
o
n
e
, 256)
D
e
ns
e
(
D
e
ns
e
)
(
N
o
n
e
,
2)
2.
5.
F
in
e
-
t
u
n
in
g
VG
G
-
19
Af
t
e
r
t
h
e
i
ni
t
i
a
l
t
r
a
i
ni
ng
pr
o
c
e
s
s
,
s
e
v
e
r
a
l
pa
r
a
m
e
t
e
r
s
a
r
e
a
dj
us
t
e
d
to
f
i
ne
-
t
un
e
t
h
e
m
o
de
l
.
T
h
e
f
i
ne
-
t
uni
n
g
pr
o
c
e
s
s
s
t
a
r
t
s
by
u
nf
r
e
e
z
i
ng
t
h
e
l
a
s
t
t
w
o
l
a
ye
r
s
,
i
nc
l
ud
i
ng
t
h
e
f
u
ll
y
c
o
nn
e
c
t
e
d
l
a
y
e
r
.
Ne
x
t
,
t
h
e
l
e
a
r
ni
ng
r
a
t
e
i
s
r
e
duc
e
d
to
10%
o
f
i
t
s
or
i
g
i
na
l
v
a
l
ue
,
a
n
d
da
t
a
a
ugm
e
n
t
a
t
i
o
n
i
s
a
pp
l
i
e
d
t
o
t
h
e
v
a
li
da
t
i
o
n
da
t
a
s
e
t
to
e
nh
a
nc
e
i
t
s
d
i
v
e
r
s
i
t
y
s
u
c
h
a
s
h
o
r
i
z
o
n
t
a
l
f
l
i
p,
r
e
s
c
a
l
i
ng,
30%
r
ot
a
t
i
o
n
r
a
n
ge
,
20%
wi
dt
h
s
hif
t
r
a
nge
,
20%
h
e
i
g
h
t
s
hi
f
t
r
a
n
ge
,
20%
z
oo
m
i
ng
,
a
n
d
20%
s
h
e
a
r
in
g
[
25]
.
T
h
e
m
o
de
l
i
s
t
h
e
n
r
e
t
r
a
i
n
e
d
us
i
ng
t
h
e
a
ug
m
e
n
t
e
d
v
a
li
da
t
i
o
n
da
t
a
s
e
t
wi
t
h
128
ba
t
c
h
s
i
z
e
a
n
d
100
e
po
c
hs
.
Dur
i
n
g
t
hi
s
r
e
t
r
a
i
ni
n
g
p
h
a
s
e
,
t
h
e
m
o
de
l
a
l
s
o
un
de
r
go
e
s
v
a
li
da
t
i
o
n
t
h
r
o
ugh
20%
s
pl
i
t
v
a
li
da
t
i
o
n
da
t
a
s
e
t
to
f
ur
t
h
e
r
a
s
s
e
s
s
i
t
s
pe
r
f
o
r
m
a
n
c
e
.
Af
t
e
r
f
i
ne
-
t
uni
n
g
t
h
e
m
o
de
l
a
n
d
o
b
t
a
i
ni
ng
t
h
e
pr
e
di
c
t
i
o
n
r
e
s
u
l
t
s
,
we
pr
o
c
e
e
d
to
e
v
a
l
ua
t
e
i
t
s
a
c
c
ur
a
c
y
us
i
n
g
a
de
s
i
g
n
a
t
e
d
t
e
s
t
da
t
a
s
e
t
.
T
hi
s
s
t
e
p
i
s
c
r
uc
i
a
l
b
e
c
a
u
s
e
i
t
a
l
l
o
ws
us
to
a
s
s
e
s
s
h
o
w
we
l
l
t
h
e
f
in
e
-
t
un
e
d
m
o
de
l
pe
r
f
o
r
m
s
w
h
e
n
a
pp
li
e
d
t
o
un
s
e
e
n
da
t
a
,
e
f
f
e
c
t
i
v
e
ly
s
im
u
l
a
t
i
n
g
r
e
a
l
-
wo
r
l
d
c
o
n
d
i
t
i
o
n
s
,
a
n
d
s
c
e
na
r
i
o
s
.
B
y
e
v
a
l
ua
t
i
n
g
i
t
s
pe
r
f
o
r
m
a
n
c
e
o
n
t
hi
s
i
n
de
p
e
n
de
n
t
da
t
a
s
e
t
,
we
c
a
n
i
de
n
t
i
f
y
a
ny
d
i
s
c
r
e
pa
n
c
i
e
s
b
e
t
we
e
n
t
h
e
m
o
de
l
’
s
t
r
a
i
ni
ng
a
n
d
t
e
s
t
i
n
g
c
a
p
a
bi
li
t
i
e
s
.
T
h
e
pr
i
mar
y
go
a
l
o
f
t
hi
s
e
v
a
l
ua
t
i
o
n
i
s
t
o
e
n
s
ur
e
t
h
a
t
t
h
e
m
o
de
l
i
s
n
o
t
o
nl
y
e
f
f
e
c
t
i
ve
o
n
t
h
e
t
r
a
i
ni
ng
da
t
a
b
ut
a
l
s
o
ge
n
e
r
a
l
i
z
e
s
we
ll
t
o
pr
a
c
t
i
c
a
l
a
pp
li
c
a
t
i
o
ns
.
T
hi
s
pr
o
c
e
s
s
pr
o
vi
de
s
i
ns
i
g
h
t
s
i
n
t
o
t
h
e
m
o
de
l
’
s
r
o
b
us
t
n
e
s
s
a
n
d
r
e
l
i
a
bil
i
t
y
,
w
hi
c
h
a
r
e
e
s
s
e
n
t
i
a
l
f
o
r
i
t
s
de
p
l
o
ym
e
n
t
i
n
r
e
a
l
-
wo
r
l
d
s
e
tt
i
n
g
s
.
Ul
t
i
m
a
t
e
l
y
,
a
c
c
ur
a
t
e
e
v
a
l
ua
t
i
o
n
h
e
l
ps
i
n
b
u
il
d
i
ng
c
o
nf
i
d
e
n
c
e
i
n
t
h
e
m
o
de
l
’
s
pr
e
d
i
c
t
i
v
e
a
bil
i
t
i
e
s
a
n
d
i
t
s
po
t
e
n
t
i
a
l
im
pa
c
t
i
n
r
e
l
e
v
a
n
t
f
i
e
l
d
s
.
2.
6.
Re
s
u
l
t
an
d
e
val
u
at
ion
I
n
t
hi
s
s
e
c
t
i
o
n
,
t
h
e
pe
r
f
o
r
m
a
n
c
e
o
f
t
h
e
VGG
-
19
m
o
de
l
,
w
hi
c
h
ha
s
b
e
e
n
t
e
s
t
e
d
,
i
s
c
o
m
pa
r
e
d
w
i
t
h
ot
h
e
r
m
o
de
l
s
to
ga
i
n
de
e
pe
r
i
n
s
i
g
h
t
s
i
n
to
i
t
s
a
bi
li
t
y
to
c
l
a
s
s
if
y
im
a
ge
s
f
o
r
h
u
m
a
n
de
t
e
c
t
i
o
n
.
T
h
e
e
v
a
l
ua
t
i
o
n
f
o
c
us
e
s
o
n
c
o
m
pa
r
i
ng
pe
r
f
o
r
m
a
nc
e
m
e
t
r
i
c
s
,
pa
r
t
i
c
u
l
a
r
ly
a
c
c
ur
a
c
y
.
A
dd
i
t
i
o
n
a
ll
y
,
t
h
e
f
i
na
l
m
o
de
l
p
r
e
s
e
n
t
s
a
c
o
nf
us
i
o
n
m
a
t
r
i
x
ba
s
e
d
o
n
t
h
e
t
e
s
t
da
t
a
s
e
t
,
i
l
l
u
s
t
r
a
ti
ng
t
h
e
t
r
ue
p
o
s
i
t
i
ve
,
t
r
ue
n
e
ga
t
i
v
e
,
f
a
l
s
e
po
s
i
t
i
v
e
,
a
n
d
f
a
l
s
e
n
e
ga
t
i
v
e
pr
e
d
i
c
t
i
o
n
s
.
T
h
e
a
c
c
ur
a
c
y
o
f
t
h
e
pe
r
f
o
r
m
a
n
c
e
m
e
t
r
i
c
s
i
s
c
o
m
put
e
d
us
i
n
g
t
h
e
(
1)
.
=
+
(
1)
3.
RE
S
UL
T
S
AN
D
DI
S
CU
S
S
I
ON
3.
1.
T
r
ain
in
g
m
od
e
l
r
e
s
u
l
t
T
hi
s
ph
a
s
e
,
we
w
i
l
l
d
i
s
c
u
s
s
t
h
e
r
e
s
u
l
t
s
o
f
tr
a
i
ni
ng
VGG
-
19
tr
a
n
s
f
e
r
l
e
a
r
ni
ng
m
o
de
l
.
R
e
s
e
a
r
c
h
e
r
s
ut
i
li
z
e
d
t
h
e
P
y
t
h
o
n
pr
o
gr
a
m
mi
ng
l
a
n
gua
g
e
to
c
r
e
a
t
e
t
h
e
m
o
de
l
s
,
l
e
v
e
r
a
g
i
ng
t
h
e
T
e
n
s
o
r
F
l
o
w
l
i
b
r
a
r
y
,
whi
c
h
i
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t
J
I
n
f
&
C
o
m
m
u
n
T
e
c
hn
o
l
I
S
S
N:
2252
-
8776
Hum
an
de
tec
ti
on
in
C
C
T
V
s
c
r
e
e
ns
hot
us
ing
f
ine
-
tu
ning
V
GG
-
19
(
F
ir
daus
A
ngga
De
w
angga
)
649
r
e
n
o
wn
e
d
f
o
r
b
u
i
l
d
i
ng
de
e
p
l
e
a
r
ni
ng
m
o
de
l
s
.
M
o
d
e
l
de
v
e
l
o
p
m
e
n
t
wa
s
c
o
n
duc
t
e
d
us
i
n
g
t
h
e
J
up
y
t
e
r
Not
e
b
oo
k
I
DE
i
n
G
o
o
gl
e
C
o
l
l
a
b
o
r
a
tor
y
.
De
s
p
i
t
e
s
o
m
e
l
im
i
t
a
t
i
o
n
s
,
Go
o
gl
e
C
o
l
l
a
b
o
r
a
to
r
y
’
s
GPU
r
e
s
o
ur
c
e
s
we
r
e
a
de
qua
t
e
f
o
r
t
hi
s
r
e
s
e
a
r
c
h
.
C
r
e
a
t
i
n
g
t
h
e
m
o
de
l
i
n
T
e
n
s
o
r
F
l
o
w
i
nv
o
l
v
e
s
s
e
ve
r
a
l
pa
r
a
m
e
t
e
r
s
.
VGG
-
19
m
o
de
l
we
r
e
tr
a
i
ne
d
w
i
t
h
2
,
097
,
922
t
r
a
i
n
a
bl
e
pa
r
a
m
e
t
e
r
s
a
n
d
20
,
024
,
384
n
o
n
-
tr
a
i
n
a
bl
e
pa
r
a
m
e
t
e
r
s
.
L
a
s
t
e
po
c
h
or
e
po
c
h
25
i
n
t
hi
s
t
r
a
i
ni
ng
r
e
s
u
l
t
i
ng
i
n
1.
0000
i
n
tr
a
i
ni
ng
a
c
c
ur
a
c
y
a
n
d
1.
0000
i
n
v
a
li
da
t
i
o
n
a
c
c
ur
a
c
y
.
L
a
s
t
e
po
c
h
a
l
s
o
r
e
s
u
l
t
i
n
g
i
n
0.
0029
l
o
s
s
a
n
d
0.
0045
v
a
li
da
t
i
o
n
l
o
s
s
.
T
r
a
i
ni
ng
a
n
d
v
a
li
da
t
i
o
n
r
e
s
u
l
t
o
f
t
h
e
tr
a
i
ni
n
g
m
o
de
l
a
r
e
i
ll
us
t
r
a
t
e
d
i
n
F
i
gur
e
2.
F
i
gur
e
2.
A
c
c
ur
a
c
y
a
n
d
l
o
s
s
t
r
a
i
ni
ng
VGG
-
19
m
o
de
l
Af
t
e
r
t
h
e
m
o
de
l
h
a
v
e
pr
e
d
i
c
t
i
o
n
i
n
h
u
m
a
n
de
t
e
c
t
i
o
n
s
c
r
e
e
ns
h
o
t,
t
h
e
n
we
t
e
s
t
t
h
e
pr
e
d
i
c
t
i
o
n
us
i
ng
t
e
s
t
da
t
a
s
e
t
wi
t
h
tot
a
l
o
f
52
h
u
m
a
n
i
m
a
ge
s
a
n
d
52
n
o
nh
u
m
a
n
i
m
a
ge
s
w
i
t
h
64
b
a
t
c
h
s
i
z
e
a
n
d
f
a
l
s
e
s
h
u
f
f
l
e
.
F
o
r
a
c
hi
e
vi
ng
a
n
o
p
t
i
m
a
l
c
l
a
s
s
i
f
i
e
r
dur
i
n
g
t
r
a
i
nin
g,
e
v
a
l
ua
t
i
o
n
m
e
t
r
i
c
s
a
r
e
e
s
s
e
n
t
i
a
l
[
26]
.
T
h
e
m
e
t
r
i
c
s
f
o
r
e
v
a
l
ua
t
i
o
n
we
r
e
a
c
c
ur
a
c
y
,
w
hi
c
h
we
r
e
c
a
l
c
u
l
a
t
e
d
by
e
x
a
mi
ni
ng
t
r
ue
p
o
s
i
t
i
v
e
s
,
t
r
ue
n
e
ga
t
i
v
e
s
,
f
a
l
s
e
po
s
i
t
i
v
e
s
,
a
n
d
f
a
l
s
e
n
e
ga
t
i
v
e
s
b
e
t
we
e
n
t
h
e
c
l
a
s
s
e
s
a
n
d
pr
e
d
i
c
t
i
o
n
r
e
s
u
l
t
s
o
f
m
o
de
l
.
T
hi
s
t
e
s
t
i
n
g
r
e
s
u
l
t
i
n
g
i
n
0.
7596
o
r
75%
a
c
c
ur
a
c
y
w
i
t
h
45
t
r
ue
p
o
s
i
t
i
v
e
s
,
7
f
a
l
s
e
po
s
i
t
i
v
e
s
,
18
f
a
l
s
e
n
e
ga
t
i
v
e
s
a
n
d
34
t
r
ue
n
e
ga
t
i
v
e
s
.
3.
2.
F
in
e
-
t
u
n
in
g
VG
G
-
19
r
e
s
u
l
t
Af
t
e
r
t
r
a
i
ni
ng
t
h
e
m
o
de
l
,
we
w
il
l
d
i
s
c
u
s
s
t
h
e
r
e
s
u
l
t
s
o
f
f
i
ne
-
t
uni
ng
VGG
-
19
m
o
de
l
.
U
nf
r
e
e
z
i
ng
2
l
a
s
t
l
a
y
e
r
i
n
VGG
-
19
m
o
de
l
o
r
i
n
bl
o
c
k
4
l
a
y
e
r
c
r
e
a
t
i
n
g
19,
796,
738
t
r
a
i
n
a
bl
e
pa
r
a
m
e
t
e
r
s
a
n
d
2,
325,
5
68
n
o
n
-
t
r
a
i
n
a
bl
e
pa
r
a
m
e
t
e
r
s
.
T
h
e
o
p
t
i
mi
z
e
r
us
e
d
wa
s
A
d
a
m
w
i
t
h
a
l
e
a
r
ni
ng
r
a
t
e
o
f
0.
0001,
b
a
t
c
h
s
i
z
e
128,
a
n
d
t
h
e
l
o
s
s
f
u
n
c
t
i
o
n
wa
s
c
a
t
e
gor
i
c
a
l
c
r
o
s
s
-
e
n
t
r
o
py
.
T
h
e
f
i
ne
-
t
uni
n
g
t
r
a
i
ni
ng
pha
s
e
c
o
n
s
i
s
t
e
d
o
f
100
e
po
c
h
s
us
i
ng
v
a
li
da
t
i
o
n
da
t
a
s
e
t
to
e
v
a
l
ua
t
e
m
o
de
l
pe
r
f
o
r
m
a
n
c
e
.
L
a
s
t
e
po
c
h
o
r
e
p
o
c
h
100
i
n
t
hi
s
t
r
a
i
ni
ng
r
e
s
u
l
t
i
ng
i
n
0.
9829
i
n
t
r
a
i
ni
ng
a
c
c
ur
a
c
y
a
n
d
0.
9828
i
n
v
a
li
da
t
i
o
n
a
c
c
ur
a
c
y
.
L
a
s
t
e
p
o
c
h
a
l
s
o
r
e
s
ul
t
i
n
g
i
n
0.
0440
l
o
s
s
a
nd
0.
0313
v
a
li
da
t
i
o
n
l
o
s
s
.
T
r
a
i
ni
n
g
a
n
d
v
a
li
da
t
i
o
n
r
e
s
u
l
t
o
f
t
h
e
f
i
ne
-
t
uni
ng
m
o
de
l
a
r
e
i
ll
u
s
t
r
a
t
e
d
i
n
F
i
gur
e
3.
Af
t
e
r
t
h
e
m
o
de
l
m
a
d
e
pr
e
d
i
c
t
i
o
n
s
f
o
r
h
u
m
a
n
de
t
e
c
t
i
o
n
i
n
t
h
e
s
c
r
e
e
ns
h
o
t
s
,
we
t
e
s
t
e
d
i
t
s
pe
r
f
o
r
m
a
n
c
e
us
i
n
g
a
t
e
s
t
da
t
a
s
e
t
c
o
n
s
i
s
t
i
n
g
o
f
52
h
u
m
a
n
i
m
a
ge
s
a
n
d
52
n
o
n
-
h
u
m
a
n
im
a
ge
s
.
T
h
e
b
a
t
c
h
s
i
z
e
us
e
d
wa
s
64,
a
n
d
t
h
e
da
t
a
wa
s
pr
o
c
e
s
s
e
d
w
i
t
h
o
ut
s
h
u
f
f
li
ng.
F
o
r
e
v
a
l
ua
t
i
o
n
,
a
c
c
ur
a
c
y
w
a
s
m
e
a
s
ur
e
d
by
c
o
m
pa
r
i
n
g
t
h
e
t
r
ue
po
s
i
t
i
v
e
s
,
t
r
ue
n
e
ga
t
i
v
e
s
,
f
a
l
s
e
po
s
i
t
i
v
e
s
,
a
n
d
f
a
l
s
e
n
e
ga
t
i
v
e
s
b
e
t
we
e
n
t
h
e
pr
e
d
i
c
t
e
d
r
e
s
u
l
t
s
a
n
d
t
he
a
c
t
ua
l
c
l
a
s
s
e
s
.
T
h
e
t
e
s
t
y
i
e
l
de
d
a
n
a
c
c
ur
a
c
y
o
f
85%
(
0.
8557)
,
wi
t
h
50
tr
ue
p
o
s
i
t
i
v
e
s
,
2
f
a
l
s
e
po
s
i
t
i
v
e
s
,
13
f
a
l
s
e
n
e
ga
t
i
v
e
s
,
a
n
d
39
t
r
ue
n
e
ga
t
i
ve
s
.
F
i
gur
e
3.
A
c
c
ur
a
c
y
a
n
d
l
o
s
s
f
i
ne
-
t
un
e
d
VGG
-
19
m
o
de
l
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8776
I
n
t
J
I
n
f
&
C
o
m
m
u
n
T
e
c
hn
o
l
,
Vo
l
.
1
4
,
N
o.
2
,
A
ugus
t
20
2
5
:
645
-
652
650
3.
3.
E
val
u
at
ion
I
n
o
u
r
s
t
udy
,
we
m
e
a
s
ur
e
pe
r
f
o
r
m
a
n
c
e
f
i
ne
t
uni
n
g
VGG
-
19
m
o
de
l
w
i
t
h
a
n
o
t
h
e
r
m
o
de
l
f
o
r
h
u
m
a
n
im
a
ge
c
l
a
s
s
if
i
c
a
t
i
o
n
.
T
h
e
r
e
s
u
l
t
s
s
h
o
w
t
h
a
t
V
GG
-
19
a
c
hi
e
v
e
d
b
e
t
t
e
r
a
c
c
ur
a
c
y
r
a
t
e
85%
i
n
t
e
s
t
i
n
g,
o
u
t
pe
r
f
o
r
m
i
ng
a
n
o
t
h
e
r
m
e
t
h
o
d
a
s
s
h
o
wn
i
n
T
a
bl
e
2.
T
hi
s
hi
g
hl
i
g
h
t
s
t
h
e
e
f
f
e
c
t
i
v
e
n
e
s
s
o
f
VG
G
-
19
i
n
c
l
a
s
s
if
yi
ng
i
m
a
ge
c
a
t
e
g
o
r
i
e
s
,
pa
r
t
i
c
u
l
a
r
ly
w
i
t
hi
n
th
e
c
o
n
t
e
x
t
o
f
C
C
T
V
s
c
r
e
e
n
s
h
o
t
s
.
Al
t
h
o
ugh
ot
h
e
r
m
e
t
h
o
d
h
a
s
i
t
s
a
d
v
a
n
t
a
ge
s
i
n
c
e
r
t
a
i
n
s
c
e
n
a
r
i
o
s
,
VG
G
-
19
out
pe
r
f
o
r
m
s
i
t
i
n
m
a
na
g
i
ng
t
h
e
c
o
m
p
l
e
xi
t
i
e
s
a
n
d
c
o
n
d
i
t
i
o
n
s
o
f
C
C
T
V
im
a
ge
s
,
m
a
k
i
ng
VG
G
-
19
i
de
a
l
o
pt
i
o
n
f
o
r
h
u
m
a
n
de
t
e
c
t
i
o
n
i
n
i
m
a
g
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
pr
o
c
e
s
s
.
I
t
s
de
e
p
l
e
a
r
ni
ng
c
a
pa
bil
i
t
i
e
s
a
l
l
o
w
i
t
to
c
a
t
c
h
s
pe
c
i
f
i
c
de
t
a
i
l
s
a
n
d
n
ua
n
c
e
s
i
n
C
C
T
V
i
m
a
ge
s
,
r
e
s
u
l
t
i
n
g
i
n
i
m
pr
o
v
e
d
a
c
c
ur
a
c
y
a
n
d
pe
r
f
o
r
m
a
n
c
e
.
T
a
bl
e
2.
C
o
m
pa
r
i
s
o
n
w
i
t
h
a
n
o
t
h
e
r
m
e
t
h
o
d
M
e
th
o
d
A
c
c
u
r
a
c
y
F
in
e
-
tu
ni
ng
VGG
-
19
85%
T
r
a
ns
f
e
r
l
e
a
r
n
in
g
VGG
-
19
75%
V
G
G
16 [
13]
60%
4.
CONC
L
USI
ON
T
hi
s
r
e
s
e
a
r
c
h
pr
o
po
s
e
d
a
di
f
f
e
r
e
n
t
a
ppr
o
a
c
h
f
o
r
h
u
m
a
n
de
t
e
c
t
i
o
n
i
n
s
c
r
e
e
ns
h
o
t
s
,
u
t
i
l
i
z
i
ng
VGG
-
19
,
a
c
us
to
m
i
z
e
d
o
bj
e
c
t
de
t
e
c
t
i
o
n
m
o
de
l
f
o
r
i
m
a
ge
c
l
a
s
s
i
f
i
c
a
t
i
o
n
.
A
c
hi
e
vi
ng
a
r
e
m
a
r
ka
bl
e
85%
a
c
c
ur
a
c
y
r
a
t
e
,
t
hi
s
m
e
t
h
o
d
s
ur
pa
s
s
e
s
ot
h
e
r
m
e
t
h
o
d
i
n
c
a
t
e
gor
i
z
i
n
g
s
c
r
e
e
ns
h
o
t
s
i
n
t
o
t
w
o
di
s
t
i
nc
t
c
l
a
s
s
e
s
,
s
h
o
wc
a
s
i
ng
i
t
s
pr
o
f
i
c
i
e
nc
y
i
n
c
a
pt
ur
i
n
g
im
a
ge
i
n
t
r
i
c
a
c
i
e
s
.
T
his
b
r
e
a
kt
h
r
o
ugh
r
e
pr
e
s
e
n
t
s
a
s
i
g
nif
i
c
a
n
t
a
dv
a
n
c
e
m
e
n
t
i
n
b
ur
g
l
a
r
y
pr
e
v
e
n
t
i
o
n
a
n
d
s
e
c
ur
i
t
y
m
a
n
a
ge
m
e
n
t
,
w
i
t
h
pot
e
n
t
i
a
l
t
o
l
e
s
s
e
n
f
a
t
a
l
i
t
i
e
s
,
i
nj
ur
i
e
s
,
a
n
d
e
c
o
n
o
m
i
c
l
o
s
s
e
s
a
s
s
o
c
i
a
t
e
d
w
i
t
h
b
ur
g
l
a
r
y
i
nc
i
de
n
t
s
.
F
ut
ur
e
r
e
s
e
a
r
c
h
s
h
o
u
l
d
f
o
c
us
o
n
e
nha
n
c
i
ng
t
h
e
m
o
de
l
’
s
c
a
pa
bil
i
t
y
to
h
a
n
d
l
e
d
i
ve
r
s
e
i
m
a
g
e
f
e
a
t
ur
e
s
.
F
ur
t
h
e
r
m
o
r
e
,
ke
y
go
a
l
s
i
nc
l
ude
e
nh
a
n
c
i
ng
t
h
e
c
l
a
s
s
i
f
i
c
a
t
i
o
n
s
y
s
t
e
m
f
o
r
m
o
r
e
c
o
m
pr
e
h
e
ns
i
ve
o
bj
e
c
t
c
a
t
e
g
o
r
i
z
a
t
i
o
n
a
n
d
e
x
p
l
o
r
i
n
g
r
e
a
l
-
t
i
m
e
i
m
p
l
e
m
e
n
t
a
t
i
o
n
f
o
r
i
mm
e
d
i
a
t
e
h
u
m
a
n
de
t
e
c
t
i
o
n
.
E
nh
a
n
c
i
ng
t
h
e
da
t
a
s
e
t,
c
o
l
l
a
b
o
r
a
t
i
n
g
w
i
t
h
j
ur
i
s
d
i
c
t
i
o
n
,
a
n
d
i
nc
o
r
por
a
t
i
n
g
t
h
e
m
o
de
l
i
n
t
o
de
c
i
s
i
o
n
s
uppo
r
t
s
y
s
t
e
m
s
a
n
d
de
t
e
c
t
i
o
n
r
e
po
r
t
i
n
g
p
l
a
t
f
o
r
m
s
a
r
e
i
m
po
r
t
a
n
t
a
c
t
i
o
n
s
to
m
a
xi
mi
z
e
i
t
s
r
e
a
l
-
wo
r
l
d
e
f
f
e
c
t
a
n
d
e
nh
a
nc
e
s
a
f
e
t
y
i
n
h
o
m
e
s
ur
r
o
un
d
i
n
gs
.
AC
K
NOWL
E
DGE
M
E
NT
S
W
e
s
i
nc
e
r
e
ly
t
h
a
n
k
M
r
.
K
o
ns
t
a
n
t
i
n
Ve
r
n
e
r
f
o
r
ge
n
e
r
o
us
l
y
pr
o
vi
d
i
n
g
im
po
r
t
a
n
t
Hum
a
n
De
t
e
c
t
i
o
n
Da
t
a
s
e
t
vi
a
K
a
gg
l
e
.
H
i
s
da
t
a
s
e
t
wa
s
c
r
uc
i
a
l
t
o
o
u
r
s
t
udy
,
a
l
l
o
w
i
n
g
us
to
c
o
n
duc
t
o
ur
r
e
s
e
a
r
c
h
a
n
d
a
c
hi
e
ve
s
i
g
nif
i
c
a
n
t
r
e
s
u
l
t
s
i
n
h
u
m
a
n
de
t
e
c
t
i
o
n
c
l
a
s
s
i
f
i
c
a
t
i
o
n
.
W
e
de
e
p
ly
h
o
n
o
r
hi
s
w
il
li
ng
n
e
s
s
a
n
d
s
uppo
r
t
to
s
h
a
r
e
t
hi
s
i
m
po
r
t
a
n
t
d
a
t
a
c
o
l
l
e
c
t
i
o
n
,
whi
c
h
h
a
s
gr
e
a
t
l
y
a
dv
a
n
c
e
d
o
ur
c
o
m
pr
e
h
e
ns
i
o
n
o
f
h
u
m
a
n
de
t
e
c
t
i
o
n
i
n
C
C
T
V
s
c
r
e
e
n
s
h
o
t
s
.
W
e
a
l
s
o
w
i
de
n
o
ur
gr
a
t
i
t
ude
to
t
h
e
or
ga
ni
z
a
t
i
o
n
s
a
n
d
a
ut
h
o
r
s
wh
o
s
e
publi
c
a
t
i
o
ns
a
n
d
r
e
s
e
a
r
c
h
h
a
v
e
b
e
e
n
e
s
s
e
n
t
i
a
l
t
o
o
ur
r
e
s
e
a
r
c
h
.
T
h
e
i
r
e
n
do
w
m
e
n
t
h
a
s
b
r
o
a
d
e
ne
d
o
ur
un
de
r
s
t
a
n
d
i
n
g
a
n
d
s
ha
pe
d
t
h
e
m
e
t
h
o
do
l
o
g
i
e
s
we
ut
i
l
i
z
e
d.
A
dd
i
t
i
o
n
a
ll
y
,
we
v
a
l
u
e
t
h
e
v
a
r
i
e
t
y
o
f
s
o
ur
c
e
s
t
h
a
t
pr
o
vi
de
d
e
s
s
e
n
t
i
a
l
i
n
s
i
g
h
t
a
n
d
c
o
n
t
e
x
t
f
o
r
o
ur
r
e
s
e
a
r
c
h
.
F
UN
DI
NG
I
N
F
ORM
AT
I
ON
A
ut
h
o
r
s
s
t
a
t
e
n
o
f
u
n
d
i
ng
i
nv
o
l
v
e
d.
AU
T
HO
R
CONT
RI
B
UT
I
ONS
S
T
AT
E
M
E
NT
Nam
e
of
Aut
h
or
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
F
i
r
d
a
us
An
gga
De
wa
n
gga
Abb
a
S
uga
n
da
G
i
r
s
a
n
g
C
:
C
o
n
c
e
pt
ua
li
z
a
ti
o
n
M
:
M
e
th
o
d
o
l
o
g
y
So
:
So
f
twa
r
e
Va
:
Va
li
da
ti
o
n
Fo
:
Fo
r
ma
l
a
na
l
y
s
is
I
:
I
nve
s
ti
ga
ti
o
n
R
:
R
e
s
o
u
r
c
e
s
D
:
D
a
ta
C
ur
a
ti
o
n
O
:
W
r
it
in
g
-
O
r
ig
in
a
l
D
r
a
f
t
E
:
W
r
it
in
g
-
R
e
v
i
e
w
&
E
di
ti
ng
Vi
:
Vi
s
ua
li
z
a
ti
o
n
Su
:
Su
pe
r
v
is
io
n
P
:
P
r
o
j
e
c
t
a
dmi
ni
s
tr
a
ti
o
n
Fu
:
Fu
ndi
ng a
c
qui
s
it
i
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t
J
I
n
f
&
C
o
m
m
u
n
T
e
c
hn
o
l
I
S
S
N:
2252
-
8776
Hum
an
de
tec
ti
on
in
C
C
T
V
s
c
r
e
e
ns
hot
us
ing
f
ine
-
tu
ning
V
GG
-
19
(
F
ir
daus
A
ngga
De
w
angga
)
651
CONF
L
I
CT
OF
I
NT
E
RE
S
T
S
T
AT
E
M
E
NT
A
ut
h
o
r
s
s
t
a
t
e
n
o
c
o
nf
li
c
t
o
f
i
n
t
e
r
e
s
t
.
E
T
HI
CA
L
AP
P
ROV
AL
T
h
e
r
e
s
e
a
r
c
h
r
e
l
a
t
e
d
to
h
u
m
a
n
us
e
h
a
s
b
e
e
n
c
o
m
p
li
e
d
w
i
t
h
a
ll
t
h
e
r
e
l
e
va
n
t
n
a
t
i
o
n
a
l
r
e
gu
l
a
t
i
o
ns
a
n
d
i
ns
t
i
t
ut
i
o
n
a
l
po
l
i
c
i
e
s
i
n
a
c
c
o
r
da
n
c
e
w
i
t
h
t
h
e
t
e
n
e
t
s
o
f
t
h
e
He
l
s
i
nk
i
D
e
c
l
a
r
a
t
i
o
n
a
n
d
h
a
s
b
e
e
n
a
ppr
o
v
e
d
by
t
h
e
a
ut
h
o
r
s
'
i
n
s
t
i
t
ut
i
o
n
a
l
r
e
vi
e
w
b
o
a
r
d
o
r
e
qui
v
a
l
e
n
t
c
omm
i
t
t
e
e
.
DA
T
A
AV
AI
L
AB
I
L
I
T
Y
T
h
e
da
t
a
t
h
a
t
s
upp
o
r
t
t
h
e
f
i
nd
i
ng
s
o
f
t
hi
s
s
t
udy
a
r
e
o
pe
nly
a
va
i
l
a
bl
e
i
n
Ka
gg
l
e
a
t
h
t
t
ps
:/
/www
.
ka
ggl
e
.
c
o
m
/dat
a
s
e
t
s
/co
n
s
t
a
n
t
i
n
we
r
n
e
r
/h
u
m
a
n
-
de
t
e
c
t
i
o
n
-
da
t
a
s
e
t
.
RE
F
E
R
E
NC
E
S
[
1]
E
.
L
.
P
i
z
a
,
B
.
C
.
W
e
ls
h,
D
.
P
.
F
a
r
r
in
gt
o
n,
a
nd
A
.
L
.
T
h
o
ma
s
,
“
C
C
T
V
s
ur
v
e
il
la
nc
e
f
o
r
c
r
im
e
pr
e
ve
nt
i
o
n:
a
40
-
y
e
a
r
s
y
s
t
e
m
a
ti
c
r
e
v
i
e
w
w
it
h
m
e
ta
-
a
na
l
y
s
is
,”
C
r
imi
nol
ogy
and
P
ubl
ic
P
ol
ic
y
,
vo
l.
18,
n
o
.
1,
pp.
135
–
1
59,
F
e
b.
20
19,
do
i:
10.1111/1745
-
9133.12419.
[
2]
P
.
L
.
C
ho
ng,
S
.
G
a
n
e
s
a
n,
Y
.
Y
.
T
ha
n,
a
nd
P
.
R
a
v
i,
“
D
e
s
ig
ni
ng
a
n
a
ut
o
n
om
o
us
tr
ig
g
e
r
in
g
c
o
nt
r
o
l
s
y
s
te
m
v
ia
m
ot
i
o
n
d
e
t
e
c
ti
o
n
f
o
r
I
oT
ba
s
e
d
s
ma
r
t
h
o
m
e
s
ur
ve
il
la
nc
e
C
C
T
V
c
a
m
e
r
a
,”
M
al
ay
s
ia
n
J
our
nal
o
f
Sc
ie
nc
e
and
A
dv
anc
e
d
T
e
c
hnol
ogy
,
pp.
80
–
88,
M
a
r
. 2023, do
i:
10.56532/m
js
a
t.
v
2i
s
1.120.
[
3]
P
.
N
e
ti
na
nt
,
A
.
A
ma
t
y
a
kul
,
a
nd
M
.
R
ukhi
r
a
n,
“
A
le
r
t
in
tr
ud
e
r
de
t
e
c
t
i
o
n
s
y
s
te
m
us
in
g
pa
s
s
iv
e
in
f
r
a
r
e
d
mot
i
o
n
de
t
e
c
t
o
r
ba
s
e
d
o
n
in
te
r
n
e
t
of
th
in
gs
,”
i
n
A
C
M
I
nt
e
r
nat
io
nal
C
onf
e
r
e
n
c
e
P
r
oc
e
e
di
ng Se
r
ie
s
,
J
a
n. 2022, pp. 171
–
175, do
i:
10.1145/3520084.3
520112.
[
4]
C
. M
ur
o
i
e
t
al
.
, “
A
ut
o
ma
t
e
d
f
a
ls
e
a
la
r
m
r
e
du
c
ti
o
n i
n
a
r
e
a
l
-
li
f
e
in
te
ns
i
v
e
c
a
r
e
s
e
tt
in
g us
in
g m
o
ti
o
n
d
e
te
c
ti
o
n,”
N
e
ur
oc
r
it
ic
al
C
ar
e
,
vo
l.
32, n
o
. 2, pp. 419
–
426, Apr
. 2020, d
o
i
:
10.1007/s
12028
-
01
9
-
00711
-
w.
[
5]
M
.
T
.
B
ha
tt
i,
M
.
G
.
K
ha
n,
M
.
A
s
la
m,
a
nd
M
.
J
.
F
ia
z
,
“
W
e
a
po
n
d
e
t
e
c
ti
o
n
in
r
e
a
l
-
ti
m
e
C
C
T
V
v
id
e
o
s
us
in
g
de
e
p
le
a
r
ni
ng,”
I
E
E
E
A
c
c
e
s
s
, v
ol
. 9, pp. 34366
–
34382, 2021, d
o
i:
10.1109/AC
C
E
S
S
.2021.3059170.
[
6]
A
.
A
.
E
ln
ga
r
e
t
al
.
,
“
I
ma
ge
c
la
s
s
if
i
c
a
ti
o
n
ba
s
e
d
o
n
C
N
N
:
a
s
ur
ve
y
,”
J
our
nal
of
C
y
be
r
s
e
c
ur
it
y
and
I
nf
o
r
m
at
io
n
M
anage
m
e
nt
,
p. P
P
. 18
-
50, 2021, d
o
i:
10.54216/j
c
im
.060102.
[
7]
M
.
L
a
guna
s
a
nd
E
.
G
a
r
c
e
s
,
“
T
r
a
ns
f
e
r
l
e
a
r
ni
ng
f
o
r
il
lu
s
tr
a
ti
o
n
c
l
a
s
s
if
ic
a
ti
o
n,”
27t
h
Spanis
h
C
om
put
e
r
G
r
aph
i
c
s
C
on
f
e
r
e
nc
e
,
C
E
I
G
2017
, pp. 77
–
85, 2017, do
i:
10.2312/
c
e
ig
.20171213.
[
8]
O
.
R
us
s
a
ko
v
s
k
y
e
t
al
.
,
“
I
ma
g
e
N
e
t
la
r
g
e
s
c
a
l
e
v
is
ua
l
r
e
c
o
gni
ti
on
c
ha
ll
e
ng
e
,”
I
nt
e
r
nat
io
nal
J
our
nal
of
C
om
put
e
r
V
is
io
n
,
vo
l.
1
15,
no
. 3, pp. 211
–
252, 2015, d
o
i:
10.1007/s
11263
-
015
-
0816
-
y.
[
9]
M
. B
a
ns
a
l,
M
. K
uma
r
, M
.
S
a
c
hd
e
v
a
, a
nd A
. M
it
ta
l,
“
T
r
a
ns
f
e
r
l
e
a
r
ni
ng
f
or
i
ma
g
e
c
la
s
s
if
ic
a
ti
o
n
us
in
g
VGG
-
19
:
C
a
lt
e
c
h
-
101 im
a
ge
da
ta
s
e
t,
”
J
our
nal
o
f
A
m
bi
e
nt
I
nt
e
ll
ig
e
nc
e
and
H
um
a
ni
z
e
d
C
om
put
in
g
,
v
o
l.
14,
n
o
.
4,
pp.
3609
–
3620,
2
023,
do
i:
10.1007/s
12652
-
021
-
03488
-
z.
[
10]
D
.
H
in
da
r
t
o
,
N
.
A
f
a
r
in
i,
a
nd
E
.
T
.
E
s
th
i
H
,
“
C
o
mpa
r
is
o
n
e
f
f
ic
a
c
y
of
V
G
G
16
a
nd
VGG
-
19
in
s
e
c
t
c
la
s
s
if
i
c
a
ti
o
n
m
o
d
e
ls
,”
J
I
K
O
(
J
ur
nal
I
nf
o
r
m
at
ik
a dan K
om
put
e
r
)
, vo
l.
6, n
o
. 3, pp. 189
–
195, 2023, do
i:
10.33387/j
ik
o
.
v
6
i3
.7008.
[
11]
D
.
J
a
s
w
a
l,
S
.
V
,
a
nd
K
.
P
.
S
oma
n,
“
I
ma
ge
c
la
s
s
if
ic
a
ti
o
n
us
in
g
c
o
n
vo
lu
ti
o
na
l
n
e
ur
a
l
n
e
tw
o
r
ks
,”
I
nt
e
r
nat
io
nal
J
our
nal
of
Sc
ie
n
ti
f
ic
and E
ngi
ne
e
r
in
g R
e
s
e
ar
c
h
, v
o
l.
5, n
o
. 6, pp. 1661
–
1668, 2014,
do
i:
10.14299/i
js
e
r
.2014.06.002.
[
12]
S
.
T
a
mm
in
a
,
“
T
r
a
ns
f
e
r
l
e
a
r
ni
ng
us
in
g
V
G
G
-
16
w
it
h
d
e
e
p
c
onvo
lu
ti
o
na
l
ne
u
r
a
l
n
e
tw
o
r
k
f
or
c
la
s
s
if
y
in
g
im
a
ge
s
,”
I
n
te
r
nat
io
nal
J
our
nal
of
S
c
ie
nt
if
i
c
and R
e
s
e
a
r
c
h P
ubl
ic
at
io
ns
(
I
J
SR
P
)
,
vo
l.
9,
no
. 10, p. p9420, 2019, d
o
i:
10.29322/i
js
r
p.9.10.2019.p9420.
[
13]
S
r
il
a
x
mi
,
S
.
K
a
ma
th
,
a
nd
V
.
M
a
y
y
a
,
“
A
ut
o
ma
t
e
d
huma
n
d
e
t
e
c
t
i
o
n
in
im
a
g
e
s
us
in
g
de
e
p
l
e
a
r
ni
ng,”
J
our
nal
of
P
r
opul
s
io
n
T
e
c
hnol
ogy
, vo
l.
44, n
o
. 5, pp. 1897
–
1902, 2023.
[
14]
A
.
A
lt
hni
a
n
e
t
al
.
,
“
I
mpa
c
t
of
da
ta
s
e
t
s
iz
e
o
n
c
la
s
s
if
i
c
a
ti
o
n
pe
r
f
or
ma
nc
e
:
a
n
e
mpi
r
i
c
a
l
e
v
a
lu
a
ti
o
n
in
th
e
m
e
di
c
a
l
d
o
ma
in
,”
A
ppl
ie
d Sc
ie
nc
e
s
(
Sw
it
z
e
r
la
nd)
, vo
l.
11, n
o
. 2, pp. 1
–
18, J
a
n. 2021, do
i:
10.3390/app110207
96.
[
15]
K
.
V
e
r
n
e
r
,
“
H
uma
n
d
e
t
e
c
t
i
o
n
da
ta
s
e
t,
”
K
aggl
e
,
2022.
ht
t
ps
:/
/ww
w
.ka
ggl
e
.c
o
m/
da
ta
s
e
ts
/
c
o
ns
ta
nt
in
w
e
r
n
e
r
/h
u
ma
n
-
de
te
c
ti
o
n
-
da
ta
s
e
t.
[
16]
Z.
X
ia
o
e
t
al
.
,
“
I
mpa
c
ts
of
da
ta
pr
e
pr
oc
e
s
s
in
g
a
nd
s
e
le
c
ti
o
n
o
n
e
n
e
r
g
y
c
o
ns
umpt
i
o
n
pr
e
di
c
ti
o
n
m
o
d
e
l
of
H
V
A
C
s
y
s
t
e
ms
ba
s
e
d
o
n
de
e
p l
e
a
r
n
in
g,”
E
ne
r
gy
and B
ui
ld
in
gs
, vo
l.
258, p. 111832, M
a
r
. 2022, do
i:
10.1016/j
.
e
nbui
ld
.2022.111832.
[
17]
S
.
R
a
mí
r
e
z
-
G
a
ll
e
g
o
,
B
.
K
r
a
w
c
z
y
k,
S
.
G
a
r
c
ía
,
M
.
W
oź
ni
a
k,
a
nd
F
.
H
e
r
r
e
r
a
,
“
A
s
ur
ve
y
o
n
da
ta
pr
e
p
r
oc
e
s
s
in
g
f
o
r
da
ta
s
tr
e
a
m
mi
ni
ng:
C
ur
r
e
nt
s
ta
tu
s
a
nd
f
ut
ur
e
di
r
e
c
ti
o
ns
,”
N
e
ur
oc
om
put
in
g
,
v
o
l.
239,
pp.
39
–
57,
M
a
y
2
017,
do
i:
10.1016/j
.n
e
u
c
o
m.2017.01.078.
[
18]
B
.
G
e
nç
a
nd
H
.
T
unç
, “
O
pt
im
a
l
tr
a
in
in
g a
nd t
e
s
t
s
e
ts
d
e
s
ig
n
f
or
ma
c
hi
n
e
l
e
a
r
ni
ng,”
T
ur
k
is
h J
our
nal
of
E
le
c
tr
ic
al
E
ngi
ne
e
r
in
g and
C
om
put
e
r
Sc
ie
n
c
e
s
, v
o
l.
27, n
o
. 2, pp. 1534
–
1545, M
a
r
. 2019, do
i:
10.3906/
e
lk
-
1807
-
212.
[
19]
M
.
R
a
f
a
ło
,
“
C
r
o
s
s
v
a
li
da
ti
o
n
m
e
th
o
ds
:
a
na
l
y
s
is
ba
s
e
d
o
n
d
ia
g
no
s
ti
c
s
of
th
y
r
oi
d
c
a
nc
e
r
me
ta
s
ta
s
is
,”
I
C
T
E
x
pr
e
s
s
,
vo
l.
8,
n
o.
2,
pp. 183
–
188, J
un. 2022, do
i:
10.1016/j
.i
c
t
e
.2021.05.001.
[
20]
V
.
P
r
a
de
e
p,
A
.
B
.
J
a
y
a
c
ha
ndr
a
,
S
.
S
.
A
s
ka
r
,
a
nd
M
.
A
bo
uha
w
w
a
s
h,
“
H
y
p
e
r
pa
r
a
m
e
t
e
r
tu
ni
ng
us
in
g
L
é
vy
f
li
ght
a
nd
in
te
r
a
c
ti
ve
c
r
o
s
s
ove
r
-
ba
s
e
d
r
e
pt
i
le
s
e
a
r
c
h
a
lg
or
it
hm
f
or
e
y
e
m
ov
e
me
nt
e
ve
nt
c
la
s
s
i
f
ic
a
ti
o
n,”
F
r
ont
ie
r
s
in
P
hy
s
io
lo
gy
,
vo
l.
15,
M
a
y
2
024,
do
i:
10.3
389/
f
ph
y
s
.2024.1366910.
[
21]
M
.
A
.
N
a
uf
a
l
a
nd
A
.
S
.
G
ir
s
a
ng,
“
T
r
a
f
f
ic
a
c
c
id
e
nt
c
la
s
s
if
i
c
a
ti
o
n
us
in
g
I
nd
oB
E
R
T
,”
I
nt
e
r
nat
io
nal
J
our
nal
of
I
nf
o
r
m
at
ic
s
and
C
om
m
uni
c
at
io
n T
e
c
hnol
ogy
, v
o
l.
13, n
o
. 1, pp. 42
–
49, Apr
. 20
24, do
i:
10.11591/i
ji
c
t.
v
13i
1.pp42
-
49.
[
22]
A
.
S
ta
te
c
z
n
y
,
G
.
U
.
K
ir
a
n,
G
.
B
in
du,
K
.
R
.
C
h
y
th
a
n
y
a
,
a
nd
K
.
A
.
S
w
a
m
y
,
“
S
pi
r
a
l
s
e
a
r
c
h
gr
a
s
s
ho
pp
e
r
f
e
a
tu
r
e
s
s
e
l
e
c
ti
o
n
w
it
h
VGG
-
19
-
R
e
s
N
e
t5
0
f
or
r
e
m
o
t
e
s
e
ns
in
g
o
b
je
c
t
de
t
e
c
ti
o
n,”
R
e
m
ot
e
Se
ns
in
g
,
vo
l.
14,
n
o
.
21,
p.
5398,
O
c
t.
20
22,
do
i:
10.3390/
r
s
14215398.
[
23]
Y
.
Z
he
ng,
C
.
Y
a
ng,
a
nd
A
.
M
e
r
ku
l
ov
,
“
B
r
e
a
s
t
c
a
nc
e
r
s
c
r
e
e
ni
ng
us
in
g
c
o
n
vo
lu
ti
o
na
l
n
e
ur
a
l
ne
twor
k
a
nd
f
ol
l
o
w
-
up
di
gi
ta
l
ma
mm
o
gr
a
ph
y
,”
i
n
C
om
put
at
io
nal
I
m
agi
ng
I
I
I
, M
a
y
2018, p. 4
, do
i:
10.1117/12.230456
4.
[
24]
L
.
S
in
gh,
R
.
R
.
J
a
nghe
l,
a
nd
S
.
P
.
S
a
hu,
“
A
n
e
mpi
r
ic
a
l
r
e
v
i
e
w
o
n
e
v
a
lu
a
ti
ng
th
e
im
pa
c
t
of
im
a
ge
s
e
gm
e
nt
a
ti
o
n
o
n
th
e
c
la
s
s
if
i
c
a
ti
o
n
p
e
r
f
or
ma
nc
e
f
or
s
ki
n
l
e
s
io
n
d
e
t
e
c
ti
o
n,”
I
E
T
E
T
e
c
hni
c
al
R
e
v
ie
w
(
I
ns
ti
tu
ti
on
of
E
le
c
tr
oni
c
s
and
T
e
le
c
om
m
uni
c
a
ti
on
E
ngi
ne
e
r
s
, I
ndi
a)
, vo
l.
40, n
o
. 2, pp. 190
–
201, M
a
r
. 2023
, d
o
i:
10.1080/02564602.2
022.2068681.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8776
I
n
t
J
I
n
f
&
C
o
m
m
u
n
T
e
c
hn
o
l
,
Vo
l
.
1
4
,
N
o.
2
,
A
ugus
t
20
2
5
:
645
-
652
652
[
25]
F
.
C
ho
ll
e
t,
“
D
e
e
p
l
e
a
r
ni
ng
w
it
h
P
y
th
o
n,”
M
anni
ng
P
ubl
ic
at
io
n
C
o
,
2017.
ht
tp
s
:/
/ww
w
.ma
nni
ng.c
o
m/
b
oo
ks
/d
e
e
p
-
le
a
r
n
in
g
-
w
it
h
-
p
y
th
o
n.
[
26]
S
.
S
he
d
r
ik
o
a
nd
M
.
F
i
r
da
us
,
“
C
o
mpa
r
is
o
n
of
a
da
gr
a
d,
a
da
de
lt
a
,
a
nd
a
da
m
o
pt
im
i
z
e
r
s
in
im
a
g
e
c
la
s
s
if
i
c
a
ti
o
n
us
in
g
d
e
e
p
l
e
a
r
ni
ng
(
in
I
ndo
n
e
s
ia
n)
,”
ST
R
I
N
G
(
Sat
uan
T
ul
is
an
R
is
e
t
dan
I
nov
as
i
T
e
k
nol
ogi
)
,
v
o
l.
8,
no
.
1,
p.
103,
A
ug.
2023,
do
i:
10.30998/s
tr
in
g.
v
8
i1
.16564.
B
I
OG
RA
P
HI
E
S
OF
AU
T
HO
RS
F
i
rda
u
s
A
n
g
g
a
D
ew
a
n
g
g
a
rece
i
v
e
d
h
i
s
fi
rs
t
d
e
g
r
ee
fro
m
Ba
k
ri
e
U
n
i
v
e
rs
i
t
y
,
Co
m
p
u
t
e
r
E
n
g
i
n
ee
ri
n
g
,
J
ak
art
a,
i
n
2
0
1
7
.
H
e
i
s
cu
rren
t
l
y
a
s
e
n
i
o
r
I
T
s
t
aff
i
n
Cro
w
e
I
n
d
o
n
e
s
i
a
.
H
i
s
re
s
p
o
n
s
i
b
i
l
i
t
i
e
s
i
n
t
h
i
s
ro
l
e
en
c
o
m
p
as
s
a
w
i
d
e
ran
g
e
o
f
t
as
k
s
re
l
at
ed
t
o
I
T
an
d
ap
p
l
i
c
at
i
o
n
d
ev
e
l
o
p
men
t
.
H
e
c
o
n
t
i
n
u
e
s
t
o
e
x
ce
l
i
n
t
h
i
s
p
o
s
i
t
i
o
n
,
b
ri
n
g
i
n
g
h
i
s
s
k
i
l
l
s
an
d
e
x
p
e
rt
i
s
e
t
o
c
o
n
t
ri
b
u
t
e
t
o
t
h
e
i
n
d
u
s
t
r
y
.
H
e
c
an
b
e
co
n
t
a
c
t
ed
at
em
ai
l
:
Fi
rd
au
s
.
d
ew
a
n
g
g
a@
b
i
n
u
s
.
a
c
.
i
d
.
A
bba
S
u
g
a
n
da
Gi
rs
a
n
g
i
s
c
u
rr
en
t
l
y
l
ec
t
u
r
e
r
at
m
as
t
er
i
n
f
o
r
m
at
i
o
n
t
ec
h
n
o
l
o
g
y
a
t
Bi
n
a
N
u
s
an
t
ara
U
n
i
v
e
rs
i
t
y
,
J
ak
art
a.
H
e
o
b
t
ai
n
e
d
P
h
.
D
.
d
eg
r
ee
i
n
t
h
e
I
n
s
t
i
t
u
t
e
o
f
Co
m
p
u
t
e
r
an
d
C
o
mm
u
n
i
c
at
i
o
n
E
n
g
i
n
ee
r
i
n
g
,
D
e
p
art
me
n
t
o
f
E
l
ec
t
r
i
c
a
l
E
n
g
i
n
ee
ri
n
g
an
d
N
at
i
o
n
al
C
h
e
n
g
K
u
n
g
U
n
i
v
e
rs
i
t
y
,
T
ai
n
an
,
T
ai
w
a
n
,
i
n
2
0
1
4
.
H
e
g
rad
u
at
e
d
b
ach
e
l
o
r
fr
o
m
t
h
e
D
e
p
art
men
t
o
f
E
l
ec
t
ri
c
al
E
n
g
i
n
ee
ri
n
g
,
G
ad
j
ah
Ma
d
a
U
n
i
v
e
rs
i
t
y
(
U
G
M),
Y
o
g
y
a
k
art
a
I
n
d
o
n
e
s
i
a,
i
n
2
0
0
0
.
H
e
t
h
e
n
c
o
n
t
i
n
u
e
d
h
i
s
m
as
t
e
r
d
eg
ree
i
n
t
h
e
D
e
p
art
m
en
t
o
f
Co
m
p
u
t
e
r
S
ci
e
n
ce
i
n
t
h
e
s
am
e
u
n
i
v
e
rs
i
t
y
i
n
2
0
0
6
–
2
0
0
8
.
H
e
w
as
a
s
t
aff
c
o
n
s
u
l
t
an
t
p
ro
g
ramme
r
i
n
B
e
t
h
e
s
d
a
H
o
s
p
i
t
al
,
Y
o
g
y
ak
art
a,
i
n
2
0
0
1
an
d
a
l
s
o
w
o
r
k
e
d
as
a
w
e
b
d
ev
e
l
o
p
e
r
i
n
2
0
0
2
–
2
0
0
3
.
H
e
t
h
e
n
j
o
i
n
e
d
t
h
e
fac
u
l
t
y
o
f
D
e
p
art
me
n
t
o
f
In
fo
r
m
at
i
c
s
E
n
g
i
n
ee
r
i
n
g
i
n
J
an
ab
ad
ra
U
n
i
v
e
rs
i
t
y
as
a
l
ec
t
u
r
e
r
i
n
2
0
0
3
-
2
0
1
5
.
H
e
al
s
o
t
au
g
h
t
s
o
me
s
u
b
j
ec
t
s
at
s
o
m
e
u
n
i
v
e
rs
i
t
i
e
s
i
n
2
0
0
6
–
2
0
0
8
.
H
i
s
re
s
e
ar
ch
i
n
t
e
r
e
s
t
s
i
n
cl
u
d
e
s
w
ar
m
i
n
t
el
l
i
g
en
ce
,
co
m
b
i
n
at
o
r
i
al
o
p
t
i
mi
zat
i
o
n
,
an
d
d
e
ci
s
i
o
n
s
u
p
p
o
rt
s
y
s
t
em
.
H
e
c
a
n
b
e
c
o
n
t
ac
t
e
d
at
em
ai
l
:
ag
i
rs
an
g
@
b
i
n
u
s
.
ed
u
.
Evaluation Warning : The document was created with Spire.PDF for Python.