I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
p
u
t
er
Science
Vo
l.
39
,
No
.
2
,
A
u
g
u
s
t
2025
,
p
p
.
1
013
~
1
0
2
6
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ee
cs.v
39
.i
2
.
p
p
1
0
1
3
-
1
0
2
6
1013
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Dev
elo
p
m
e
nt
o
f
m
o
bile
-
ba
sed
Ba
t
a
k
scri
pt
recog
nit
io
n
a
pplica
tion usin
g
YO
L
O
v
8
alg
o
rithm
I
us
t
i
s
ia
Na
t
a
lia
Si
m
bo
lo
n,
H
er
i
m
a
nto
,
Ra
nty
Dev
ia
na
Si
a
ha
a
n,
Sa
m
uel A
di
k
a
L
u
m
b
a
nto
bin
g
,
G
ra
ce
Na
t
a
lia
B
r
Site
pu
D
e
p
a
r
t
me
n
t
s
o
f
I
n
f
o
r
mat
i
c
s,
F
a
c
u
l
t
y
o
f
I
n
f
o
r
mat
i
c
s
a
n
d
El
e
c
t
r
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
I
n
st
i
t
u
t
T
e
k
n
o
l
o
g
i
D
e
l
,
T
o
b
a
,
I
n
d
o
n
e
si
a
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Au
g
6
,
2024
R
ev
i
s
ed
Mar
18
,
2
0
2
5
A
cc
ep
ted
J
u
l
1
,
2025
T
h
e
Ba
tak
p
e
o
p
le
a
re
o
n
e
o
f
t
h
e
e
th
n
ic
g
ro
u
p
s
th
a
t
p
a
ss
d
o
w
n
m
a
n
y
v
a
lu
e
s
a
n
d
tra
d
it
i
o
n
s
t
o
e
a
c
h
g
e
n
e
ra
ti
o
n
,
in
c
l
u
d
i
n
g
th
e
w
rit
ten
trad
it
i
o
n
k
n
o
w
n
a
s
th
e
Ba
tak
sc
rip
t.
T
h
e
Ba
tak
T
o
b
a
p
e
o
p
le,
in
p
a
rti
c
u
lar,
h
a
v
e
th
e
B
a
tak
T
o
b
a
sc
rip
t
a
s
p
a
rt
o
f
th
e
ir
lo
c
a
l
w
isd
o
m
th
a
t
n
e
e
d
s
to
b
e
p
re
se
rv
e
d
a
n
d
m
a
in
tain
e
d
.
Ho
w
e
v
e
r,
th
e
u
se
o
f
th
e
Ba
tak
sc
rip
t
h
a
s
sig
n
if
ica
n
tl
y
d
e
c
li
n
e
d
in
t
h
e
c
u
rre
n
t
e
ra
.
T
o
p
re
v
e
n
t
t
h
e
lo
ss
o
f
th
is
h
e
rit
a
g
e
,
p
re
se
rv
a
ti
o
n
t
h
ro
u
g
h
tec
h
n
o
l
o
g
y
is
n
e
c
e
ss
a
r
y
.
T
h
is
r
e
se
a
rc
h
u
ti
li
z
e
s
a
d
e
e
p
lea
rn
i
n
g
a
p
p
r
o
a
c
h
u
sin
g
t
h
e
YO
L
Ov
8
a
lg
o
rit
h
m
to
d
e
tec
t
im
a
g
e
s
o
f
sc
rip
t
o
b
jec
ts,
p
ro
v
id
e
t
h
e
c
o
o
rd
i
n
a
tes
o
f
th
e
sc
rip
t
lo
c
a
ti
o
n
s,
a
n
d
p
e
rf
o
rm
o
b
jec
t
re
c
o
g
n
it
io
n
b
a
se
d
o
n
th
e
d
a
tas
e
t.
T
h
e
f
in
a
l
re
su
lt
o
f
th
i
s
re
se
a
rc
h
is
a
n
A
n
d
ro
i
d
-
b
a
se
d
a
p
p
li
c
a
ti
o
n
th
a
t
c
a
n
d
e
tec
t
t
h
e
Ba
tak
T
o
b
a
s
c
rip
t
i
n
re
a
l
ti
m
e
a
n
d
u
p
l
o
a
d
ima
g
e
s.
T
h
e
re
se
a
rc
h
p
ro
c
e
ss
in
v
o
lv
e
s
e
x
p
e
ri
m
e
n
ts
o
n
se
v
e
ra
l
h
y
p
e
rp
a
ra
m
e
ter
s,
su
c
h
a
s
e
p
o
c
h
s
w
it
h
a
v
a
lu
e
o
f
2
0
0
,
c
o
n
f
id
e
n
c
e
t
h
re
sh
o
l
d
,
a
n
d
I
o
U
w
it
h
a
v
a
lu
e
o
f
0
.
5
.
T
h
e
m
o
d
e
l
e
v
a
lu
a
ti
o
n
sh
o
w
s
e
x
c
e
ll
e
n
t
re
su
lt
s,
w
it
h
a
p
re
c
isio
n
o
f
0
.
9
4
5
,
re
c
a
ll
o
f
0
.
9
0
2
,
m
A
P
@0
.
5
o
f
0
.
9
5
4
,
a
n
d
a
h
ig
h
c
o
n
f
id
e
n
c
e
sc
o
r
e
f
ro
m
th
e
a
p
p
li
c
a
ti
o
n
'
s d
e
tec
ti
o
n
.
K
ey
w
o
r
d
s
:
C
o
n
f
id
en
ce
s
co
r
e
I
m
ag
e
p
r
ep
r
o
ce
s
s
in
g
Ob
j
ec
t d
etec
tio
n
T
r
ain
in
g
YOL
O
v
8
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
I
u
s
ti
s
ia
Nata
lia
S
i
m
b
o
lo
n
Dep
ar
t
m
en
t
s
o
f
I
n
f
o
r
m
at
ics,
F
ac
u
lt
y
o
f
I
n
f
o
r
m
at
ic
s
an
d
E
lec
tr
ical
E
n
g
i
n
ee
r
i
n
g
,
I
n
s
tit
u
t T
ek
n
o
lo
g
i D
el
Sis
i
n
g
a
m
a
n
g
ar
aj
a
Stre
et,
Sit
o
lu
a
m
a,
L
a
g
u
b
o
ti 2
2
3
8
1
,
T
o
b
a
,
I
n
d
o
n
esia
E
m
ail: i
u
s
t
is
ia.
s
i
m
b
o
lo
n
@
d
el.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
B
atak
T
r
ib
e,
p
a
r
ticu
lar
ly
th
e
B
ata
k
T
o
b
a
,
is
r
en
o
w
n
ed
f
o
r
its
w
r
itte
n
tr
ad
itio
n
h
er
itag
e
,
s
p
ec
if
icall
y
th
e
B
ata
k
T
o
b
a
s
cr
ip
t.
T
h
e
B
atak
s
cr
ip
t
is
clas
s
if
ied
a
s
a
n
ab
u
g
id
a,
a
t
y
p
e
o
f
p
h
o
n
etic
w
r
i
tin
g
w
h
er
e
ea
ch
lan
g
u
ag
e
ca
n
b
e
ac
cu
r
atel
y
r
ep
r
esen
ted
.
T
h
e
w
r
itin
g
o
f
B
atak
letter
s
i
s
b
r
o
ad
l
y
d
i
v
id
ed
in
to
t
w
o
c
a
t
e
g
o
r
i
e
s
:
"
I
n
a
n
i
S
u
r
a
t
"
a
n
d
"
A
n
a
k
n
i
S
u
r
a
t
"
.
U
n
f
o
r
t
u
n
a
t
e
l
y
,
t
h
e
u
s
e
o
f
t
h
i
s
s
c
r
i
p
t
h
a
s
s
i
g
n
i
f
i
c
a
n
t
l
y
d
e
c
l
i
n
e
d
[
1
]
.
T
h
er
ef
o
r
e,
it
is
cr
u
c
ial
to
p
r
es
er
v
e
it
u
s
i
n
g
tec
h
n
o
lo
g
y
,
p
ar
ti
cu
lar
l
y
in
t
h
e
r
ap
id
l
y
e
v
o
lv
i
n
g
f
ield
o
f
m
ac
h
i
n
e
lear
n
in
g
.
T
h
is
r
esear
ch
h
o
ld
s
s
i
g
n
if
ica
n
t
i
m
p
o
r
tan
ce
f
o
r
s
ev
er
al
i
n
ter
r
elate
d
r
ea
s
o
n
s
.
First,
it
ai
m
s
t
o
p
r
ese
r
v
e
cu
lt
u
r
al
h
er
ita
g
e
b
y
d
ev
elo
p
i
n
g
a
m
o
b
ile
ap
p
licatio
n
u
s
in
g
m
ac
h
i
n
e
lear
n
i
n
g
al
g
o
r
ith
m
s
li
k
e
YO
L
O
v
8
to
f
ac
ilit
ate
t
h
e
lear
n
i
n
g
an
d
u
s
e
o
f
th
e
B
atak
T
o
b
a
s
cr
ip
t
am
o
n
g
y
o
u
n
g
er
g
en
er
atio
n
s
an
d
th
e
w
id
er
co
m
m
u
n
it
y
,
th
u
s
co
n
tr
ib
u
ti
n
g
to
t
h
e
co
n
tin
u
it
y
o
f
th
e
B
atak
p
eo
p
le's
cu
l
tu
r
al
a
n
d
h
is
to
r
ical
id
en
tit
y
.
Seco
n
d
,
th
e
ap
p
licatio
n
h
as
h
i
g
h
ed
u
ca
tio
n
al
v
al
u
e
b
y
p
r
o
v
id
in
g
an
in
t
er
ac
tiv
e
an
d
en
g
a
g
i
n
g
m
eth
o
d
to
lear
n
th
e
B
atak
T
o
b
a
s
cr
ip
t,
w
h
ich
m
a
y
n
o
t
b
e
av
ailab
le
in
tr
ad
itio
n
al
ed
u
ca
tio
n
al
r
eso
u
r
ce
s
,
th
er
eb
y
en
h
a
n
ci
n
g
c
u
lt
u
r
al
liter
ac
y
an
d
u
n
d
er
s
ta
n
d
i
n
g
o
f
t
h
e
B
atak
s
cr
ip
t
an
d
h
elp
in
g
u
s
er
s
u
n
d
er
s
ta
n
d
th
e
m
ea
n
in
g
an
d
s
ig
n
i
f
ica
n
ce
o
f
a
B
atak
s
cr
ip
t.
T
h
ir
d
,
th
e
u
s
e
o
f
YOL
O
v
8
i
n
t
h
is
co
n
tex
t d
e
m
o
n
s
tr
ate
s
tech
n
o
lo
g
ica
l a
d
v
a
n
ce
s
in
p
r
eser
v
in
g
t
h
e
B
atak
s
cr
ip
t,
as
t
h
e
ap
p
licatio
n
h
elp
s
u
s
er
s
u
n
d
er
s
ta
n
d
th
e
m
ea
n
in
g
o
f
a
B
ata
k
s
cr
ip
t,
wh
o
s
e
co
m
p
r
eh
e
n
s
io
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
A
u
g
u
s
t
20
25
:
1
0
1
3
-
1
026
1014
h
as
b
e
g
u
n
to
f
ad
e
i
n
th
e
c
u
r
r
e
n
t
er
a
[
2
]
,
[
3
]
.
Fo
u
r
th
,
t
h
e
p
r
a
ctica
l
i
m
p
le
m
e
n
tatio
n
o
f
m
ac
h
in
e
lear
n
i
n
g
in
r
ea
l
-
ti
m
e
o
b
j
ec
t
d
etec
tio
n
n
o
t
o
n
ly
p
r
o
v
id
es
a
r
elev
an
t
ca
s
e
s
t
u
d
y
in
A
I
ap
p
licatio
n
s
b
u
t
also
s
h
o
w
ca
s
e
s
th
e
p
o
ten
tial
o
f
tech
n
o
lo
g
y
f
o
r
cu
l
tu
r
al
ap
p
licatio
n
s
[
4
]
.
L
a
s
tl
y
,
b
y
m
a
k
i
n
g
t
h
e
B
atak
T
o
b
a
s
cr
ip
t
m
o
r
e
ac
ce
s
s
ib
le
th
r
o
u
g
h
a
m
o
b
ile
ap
p
licatio
n
,
th
is
r
esear
ch
co
n
tr
ib
u
te
s
to
t
h
e
in
cr
ea
s
ed
ac
ce
s
s
ib
ilit
y
a
n
d
u
s
e,
s
u
p
p
o
r
tin
g
lo
n
g
-
ter
m
g
o
als i
n
t
h
e
p
r
eser
v
atio
n
an
d
r
ev
italizatio
n
o
f
B
atak
c
u
l
tu
r
e
[
5
]
,
[
6
]
.
B
y
ad
d
r
ess
i
n
g
th
e
s
e
asp
ec
ts
,
th
is
r
esear
c
h
n
o
t
o
n
l
y
ai
m
s
to
p
r
eser
v
e
an
i
m
p
o
r
tan
t
p
ar
t
o
f
B
atak
h
er
itag
e
b
u
t
also
co
n
tr
ib
u
tes
t
o
th
e
b
r
o
ad
e
r
d
is
co
u
r
s
e
o
n
th
e
r
o
le
o
f
tech
n
o
lo
g
y
i
n
cu
lt
u
r
al
p
r
eser
v
atio
n
.
T
h
e
in
te
g
r
atio
n
o
f
YO
L
O
v
8
in
t
h
i
s
co
n
te
x
t
h
ig
h
li
g
h
ts
t
h
e
i
n
n
o
v
ativ
e
u
s
e
o
f
m
ac
h
i
n
e
lear
n
in
g
in
tac
k
li
n
g
c
u
lt
u
r
al
an
d
ed
u
ca
tio
n
al
c
h
alle
n
g
es.
T
h
e
o
b
j
ec
tiv
es
o
f
t
h
is
r
esear
c
h
in
cl
u
d
e
i
)
i
m
p
le
m
en
tin
g
t
h
e
YOL
O
v
8
al
g
o
r
ith
m
in
a
r
ea
l
-
ti
m
e
B
atak
s
cr
ip
t r
ec
o
g
n
itio
n
ap
p
licatio
n
a
n
d
i
m
a
g
e
i
m
p
o
r
t to
f
ac
i
litate
t
h
e
r
ec
o
g
n
itio
n
o
f
r
o
o
t
w
o
r
d
s
an
d
af
f
i
x
es;
ii
)
m
ea
s
u
r
in
g
t
h
e
ac
cu
r
ac
y
o
f
t
h
e
YO
L
O
v
8
al
g
o
r
ith
m
i
n
d
etec
t
i
n
g
o
b
j
ec
ts
t
h
r
o
u
g
h
t
h
e
e
v
al
u
atio
n
o
f
t
h
e
B
atak
T
o
b
a
s
cr
ip
t
m
o
d
el;
an
d
iii
)
i
n
te
g
r
ati
n
g
th
e
B
atak
T
o
b
a
s
cr
ip
t
o
b
j
ec
t
d
et
ec
tio
n
r
es
u
lts
w
it
h
a
m
o
b
ile
ap
p
licatio
n
u
s
in
g
T
en
s
o
r
Flo
w
L
ite.
T
h
e
YOL
O
alg
o
r
it
h
m
i
s
ab
l
e
to
d
etec
t
o
b
j
ec
ts
in
r
ea
l
-
ti
m
e
w
el
l.
T
h
e
d
ev
e
lo
p
m
e
n
t
o
f
YO
L
Ov
4
in
cr
ea
s
es
t
h
e
A
P
an
d
FP
S
v
alu
es
o
f
YO
L
O
v
3
b
y
1
0
%
an
d
1
2
%
[
3
]
.
T
h
is
is
d
if
f
er
en
t
f
r
o
m
YO
L
O
an
d
Y
O
L
O
v
2
w
h
i
c
h
a
r
e
n
o
t
e
f
f
e
c
t
i
v
e
i
n
d
e
t
e
c
t
i
n
g
s
m
a
l
l
t
a
r
g
e
t
s
s
o
t
h
a
t
m
u
l
t
i
-
s
c
a
l
e
d
e
t
e
c
t
i
o
n
i
s
a
d
d
e
d
t
o
Y
O
L
O
v
3
[
7
]
.
T
h
e
YOL
Ov
4
alg
o
r
it
h
m
i
n
tr
o
d
u
ce
d
b
y
A
le
x
e
y
B
o
ch
k
o
v
s
k
i
y
h
as
a
s
p
ee
d
o
f
u
p
to
4
5
f
p
s
.
Ho
w
ev
er
,
th
e
YOL
O
v
7
alg
o
r
it
h
m
i
s
o
n
e
o
f
th
e
o
b
j
ec
t
d
etec
tio
n
alg
o
r
ith
m
s
th
a
t
ca
n
b
e
d
o
n
e
in
r
ea
l
-
ti
m
e
w
it
h
h
i
g
h
ef
f
icien
c
y
a
n
d
ac
c
u
r
ac
y
.
T
h
e
YOL
O
v
7
al
g
o
r
ith
m
h
as
an
ac
cu
r
ac
y
o
f
5
6
.
8
%
an
d
h
as
a
h
i
g
h
d
etec
ti
o
n
s
p
ee
d
,
r
ea
ch
in
g
5
-
1
6
0
FP
S
[
8
]
ev
en
in
s
it
u
atio
n
s
w
h
er
e
th
er
e
is
m
o
r
e
th
an
o
n
e
o
b
j
ec
t
in
th
e
im
ag
e.
L
i
u
et
a
l.
[
9
]
YOL
O
v
8
,
FP
S c
o
n
s
i
s
te
n
tl
y
ab
o
v
e
3
0
0
.
T
h
er
ef
o
r
e,
th
is
s
t
u
d
y
w
il
l
co
n
d
u
ct
an
al
y
s
i
s
an
d
i
m
p
le
m
en
ta
tio
n
o
f
r
ea
l
-
ti
m
e
o
b
j
ec
t
d
etec
t
io
n
u
s
in
g
th
e
YO
L
O
v
8
alg
o
r
ith
m
.
I
n
t
h
i
s
s
tu
d
y
,
th
e
o
b
j
ec
t
to
b
e
d
etec
t
ed
is
B
atak
s
cr
ip
t.
T
h
is
s
tu
d
y
w
il
l
i
m
p
le
m
e
n
t
th
e
YOL
O
v
8
alg
o
r
ith
m
in
o
b
j
ec
t
d
etec
tio
n
an
d
is
ex
p
ec
ted
to
b
e
ca
r
r
ied
o
u
t
in
r
ea
l
-
ti
m
e
w
ith
h
ig
h
ef
f
icie
n
c
y
a
n
d
ac
cu
r
ac
y
,
a
n
d
allo
w
s
t
h
e
s
y
s
t
e
m
to
q
u
ic
k
l
y
id
en
t
if
y
s
cr
ip
t
o
b
j
ec
ts
.
T
h
is
s
tu
d
y
w
il
l
also
an
al
y
ze
YO
L
O
v
8
h
y
p
er
p
ar
a
m
eter
s
to
en
s
u
r
e
th
a
t
th
e
r
esu
lti
n
g
m
o
d
el
p
r
o
v
id
es
o
p
ti
m
al
d
etec
tio
n
.
H
y
p
er
p
ar
a
m
eter
s
e
tti
n
g
s
w
i
ll
b
e
ad
j
u
s
ted
to
th
e
d
ataset
u
s
ed
in
th
e
e
x
p
er
i
m
e
n
t
to
m
a
x
i
m
ize
m
o
d
el
p
er
f
o
r
m
a
n
ce
in
o
b
j
ec
t
d
etec
tio
n
.
H
y
p
er
p
ar
am
e
ter
tu
n
i
n
g
is
ca
r
r
ied
o
u
t
to
f
i
n
d
t
h
e
m
o
s
t
o
p
t
i
m
al
m
o
d
el
ac
cu
r
ac
y
v
alu
e
i
n
r
ea
l
-
ti
m
e
o
b
j
ec
t
d
etec
tio
n
.
T
h
e
m
o
d
el
w
i
th
o
p
ti
m
al
h
y
p
er
p
ar
a
m
e
ter
s
w
i
ll
b
e
i
m
p
le
m
en
ted
in
to
a
m
o
b
ile
-
b
a
s
ed
ap
p
licatio
n
a
n
d
is
ex
p
ec
ted
to
au
to
m
ate
o
b
j
ec
t
d
etec
tio
n
.
2.
RE
L
AT
E
D
R
E
SE
ARCH
I
n
p
r
ev
io
u
s
r
esear
ch
,
th
e
d
e
tectio
n
o
f
th
e
B
atak
T
o
b
a
s
cr
ip
t
w
as
co
n
d
u
cted
u
s
i
n
g
th
e
C
NN
alg
o
r
ith
m
s
to
d
etec
t
ea
ch
g
i
v
en
i
m
ag
e
in
p
u
t
a
n
d
ass
i
g
n
a
class
n
a
m
e
to
ea
ch
d
etec
ted
o
b
j
ec
t
[
1
]
.
Ho
w
e
v
er
,
th
is
r
esear
c
h
o
n
l
y
r
ea
ch
ed
th
e
s
tag
e
o
f
s
cr
ip
t
d
etec
tio
n
w
i
th
o
u
t
i
m
p
le
m
e
n
tatio
n
i
n
an
ap
p
l
icatio
n
,
an
d
it
o
n
l
y
d
et
ec
ted
"
in
a
n
i
s
u
r
at"
(
r
o
o
t
w
o
r
d
s
)
,
w
it
h
o
u
t
d
etec
ti
n
g
"
an
ak
n
i
s
u
r
at"
(
af
f
i
x
es).
P
r
atam
a
et
a
l.
[
1
0
]
,
th
e
r
ec
o
g
n
itio
n
an
d
tr
an
s
latio
n
o
f
th
e
B
atak
T
o
b
a
s
cr
ip
t
w
er
e
s
u
cc
ess
f
u
l
l
y
ca
r
r
ied
o
u
t
u
s
in
g
s
i
n
g
le
s
h
o
t
d
etec
tio
n
b
y
i
m
p
le
m
e
n
ti
n
g
t
h
e
C
NN
a
r
ch
itect
u
r
e.
Ho
w
e
v
er
,
th
er
e
i
s
th
e
d
etec
tio
n
o
f
r
ep
ea
ted
o
b
j
ec
ts
in
th
e
i
m
a
g
e
u
s
i
n
g
s
i
n
g
le
s
h
o
t d
etec
tio
n
w
h
ich
af
f
ec
t
s
th
e
d
etec
tio
n
r
esu
l
t
s
.
T
h
is
r
esear
ch
u
s
e
s
th
e
YO
L
Ov
8
alg
o
r
ith
m
,
w
h
ich
i
s
m
o
r
e
ef
f
icie
n
t
a
n
d
ac
cu
r
ate
in
d
etec
tin
g
th
e
B
atak
T
o
b
a
s
cr
ip
t,
in
clu
d
in
g
af
f
i
x
es.
YO
L
Ov
8
i
m
p
le
m
en
ts
a
co
n
v
o
l
u
tio
n
al
n
e
u
r
al
n
et
w
o
r
k
(
C
NN)
-
b
ased
d
etec
tio
n
m
eth
o
d
[
1
1
]
.
A
s
th
e
m
o
s
t
co
m
m
o
n
r
ep
r
esen
tat
io
n
o
f
s
in
g
le
-
s
ta
g
e
o
b
j
ec
t
d
e
tectio
n
alg
o
r
ith
m
s
,
YOL
O
v
8
is
a
n
e
u
r
al
n
et
w
o
r
k
-
b
ased
alg
o
r
ith
m
u
s
ed
to
id
en
ti
f
y
a
n
d
d
eter
m
in
e
o
b
j
ec
t
lo
ca
ti
o
n
s
.
YO
L
Ov
8
u
s
e
s
a
s
in
g
le
C
NN
m
o
d
el
to
d
etec
t
en
d
-
to
-
en
d
o
b
j
ec
ts
.
Sin
g
le
-
s
ta
g
e
d
etec
to
r
s
tr
ea
t
o
b
ject
d
etec
tio
n
as
a
r
eg
r
ess
io
n
/clas
s
i
f
icatio
n
p
r
o
b
le
m
u
s
i
n
g
a
u
n
i
f
ied
f
r
a
m
e
w
o
r
k
to
o
b
tain
lab
els
a
n
d
lo
ca
tio
n
s
d
ir
ec
tl
y
.
T
h
ese
d
etec
to
r
s
lin
k
i
m
ag
e
p
ix
e
ls
d
ir
ec
tl
y
to
b
o
u
n
d
in
g
b
o
x
co
o
r
d
in
ates
an
d
class
p
r
o
b
ab
ilit
ies
[
1
2
]
.
T
h
is
is
d
o
n
e
b
y
p
r
o
p
o
s
in
g
p
r
ed
ictio
n
b
o
x
es
d
ir
ec
tl
y
f
r
o
m
t
h
e
in
p
u
t
i
m
a
g
e
w
ith
o
u
t
a
r
eg
io
n
p
r
o
p
o
s
a
l
s
t
e
p
[
1
3
]
.
T
h
e
Y
OL
Ov
8
a
l
g
o
r
ith
m
ta
k
e
s
th
e
en
t
i
r
e
im
ag
e
a
s
in
p
u
t
in
t
o
th
e
n
etw
o
r
k
s
t
r
u
ct
u
r
e
an
d
d
i
r
e
c
t
ly
r
eg
r
es
s
es
t
h
e
b
o
u
n
d
in
g
b
o
x
l
o
c
a
t
i
o
n
s
w
h
i
l
e
c
la
s
s
if
y
in
g
o
b
je
c
t
s
in
t
o
a
p
p
r
o
p
r
i
a
t
e
c
a
t
eg
o
r
i
es
i
n
th
e
o
u
t
p
u
t
l
ay
e
r
.
YOL
Ov
8
w
as
ch
o
s
en
b
e
c
au
s
e
i
t
c
an
d
e
t
e
c
t
s
m
a
ll
e
r
o
b
je
c
t
s
w
i
t
h
h
i
g
h
e
r
r
es
o
lu
t
i
o
n
(
6
0
8
x
6
0
8
p
i
x
e
ls
)
c
o
m
p
a
r
e
d
t
o
Y
O
L
Ov
3
,
a
n
d
i
t
p
r
o
c
e
s
s
es
im
ag
es
a
t
a
s
p
e
e
d
o
f
1
5
5
f
r
am
e
s
p
e
r
s
e
c
o
n
d
[
1
4
]
.
T
h
e
Y
O
L
Ov
8
a
p
p
r
o
a
c
h
a
l
l
o
w
s
f
o
r
t
h
e
d
et
ec
t
i
o
n
o
f
r
o
o
t
w
o
r
d
s
a
n
d
af
f
ix
es
a
n
d
r
e
s
u
lt
s
in
an
A
n
d
r
o
i
d
a
p
p
li
c
a
ti
o
n
f
o
r
r
e
a
l
-
tim
e
d
e
t
e
c
t
i
o
n
.
T
h
e
Y
O
L
O
v
8
a
l
g
o
r
i
t
h
m
u
s
e
s
n
i
n
e
a
n
c
h
o
r
b
o
x
e
s
t
o
d
e
t
e
c
t
a
w
i
d
e
r
v
a
r
i
e
t
y
o
f
o
b
j
e
c
t
s
h
a
p
e
s
a
n
d
s
i
z
e
s
[
1
5
]
.
3.
M
E
T
H
O
D
T
h
is
p
h
ase
i
n
v
o
l
v
es
e
x
p
lain
in
g
th
e
r
esear
c
h
ch
r
o
n
o
lo
g
i
ca
ll
y
,
i
n
clu
d
i
n
g
th
e
r
esear
c
h
d
esig
n
,
r
esear
ch
p
r
o
ce
d
u
r
es,
test
in
g
m
eth
o
d
s
,
an
d
d
ata
ac
q
u
is
itio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
Dev
elo
p
men
t o
f m
o
b
ile
-
b
a
s
ed
B
a
ta
k
s
crip
t reco
g
n
itio
n
a
p
p
li
ca
tio
n
u
s
in
g
…
(
I
u
s
tis
ia
N
a
ta
lia
S
imb
o
lo
n
)
1015
3.
1
.
YO
L
O
v
8
m
o
del
Ob
j
ec
t
d
etec
tio
n
is
o
n
e
o
f
th
e
cr
u
cial
task
s
in
th
e
f
ield
o
f
co
m
p
u
ter
v
i
s
io
n
,
w
it
h
b
r
o
ad
a
p
p
licatio
n
s
r
an
g
i
n
g
f
r
o
m
s
ec
u
r
it
y
s
u
r
v
eill
an
ce
to
au
to
n
o
m
o
u
s
v
e
h
icle
s
.
A
p
o
p
u
lar
ap
p
r
o
ac
h
f
o
r
o
b
j
e
ct
d
etec
tio
n
is
th
e
y
o
u
o
n
l
y
lo
o
k
o
n
ce
(
YOL
O)
m
o
d
el
[
1
6
]
,
k
n
o
w
n
f
o
r
its
ab
ilit
y
to
d
etec
t
o
b
j
ec
ts
in
r
ea
l
ti
m
e
w
it
h
h
i
g
h
ac
cu
r
ac
y
.
YO
L
Ov
3
,
i
n
tr
o
d
u
ce
d
b
y
R
ed
m
o
n
a
n
d
Far
h
ad
i,
i
s
o
n
e
o
f
t
h
e
m
o
s
t
w
id
el
y
u
s
ed
v
er
s
io
n
s
o
f
YO
L
O.
T
h
is
m
o
d
el
u
s
e
s
r
esid
u
al
b
lo
ck
s
,
m
u
l
ti
-
s
ca
le
p
r
ed
ictio
n
s
,
a
n
d
a
d
ee
p
er
Dar
k
n
et
-
5
3
b
ac
k
b
o
n
e
n
et
w
o
r
k
[
1
7
]
.
T
h
ese
en
h
a
n
ce
m
e
n
t
s
allo
w
Y
OL
O
v
3
to
ac
h
iev
e
a
b
etter
b
alan
ce
b
et
w
ee
n
s
p
ee
d
an
d
ac
cu
r
ac
y
co
m
p
ar
ed
to
YOL
O
v
2
[
2
]
.
A
s
tu
d
y
b
y
C
h
en
et
a
l.
[
1
2
]
d
em
o
n
s
tr
ated
t
h
at
YO
L
O
v
3
is
ef
f
ec
ti
v
e
in
d
etec
tin
g
o
b
j
ec
ts
in
m
ed
ical
i
m
ag
e
s
,
s
u
c
h
as t
u
m
o
r
s
in
u
ltra
s
o
u
n
d
i
m
a
g
es [
1
8
]
.
YOL
O
v
5
,
alt
h
o
u
g
h
n
o
t
r
elea
s
ed
b
y
th
e
o
r
ig
i
n
al
a
u
t
h
o
r
s
o
f
YOL
O,
h
a
s
attr
ac
ted
s
ig
n
i
f
ica
n
t
atte
n
tio
n
in
th
e
r
esear
c
h
co
m
m
u
n
it
y
.
R
esear
ch
b
y
Z
h
a
n
g
et
a
l
.
[
1
9
]
h
ig
h
lig
h
t
s
s
e
v
er
al
ad
v
an
ta
g
es
o
f
YO
L
O
v
5
,
s
u
c
h
as
s
m
al
ler
m
o
d
el
s
ize,
f
a
s
ter
in
f
e
r
en
ce
s
p
ee
d
,
an
d
ea
s
e
o
f
u
s
e
w
it
h
t
h
e
P
y
T
o
r
ch
f
r
a
m
e
w
o
r
k
[
1
9
]
.
YOL
O
v
5
also
in
tr
o
d
u
ce
s
s
e
v
er
al
m
o
d
el
s
ize
v
ar
ian
ts
(
s
,
m
,
l,
x
)
t
h
at
ca
n
b
e
tailo
r
e
d
to
s
p
ec
if
ic
n
ee
d
s
.
An
o
th
er
s
t
u
d
y
b
y
Sh
e
n
et
al.
[
2
0
]
s
h
o
w
ed
t
h
at
YOL
O
v
5
ca
n
b
e
u
s
ed
f
o
r
o
b
j
e
ct
d
etec
tio
n
i
n
co
m
p
le
x
tr
a
f
f
ic
en
v
ir
o
n
m
en
t
s
w
it
h
h
ig
h
ac
c
u
r
ac
y
[
2
0
]
.
A
cc
o
r
d
in
g
to
r
esear
ch
b
y
W
an
g
et
a
l.
[
8
]
YOL
Ov
7
is
t
h
e
f
aste
s
t
an
d
m
o
s
t
ac
c
u
r
ate
YOL
O
m
o
d
el
to
d
ate.
T
h
is
m
o
d
el
ad
o
p
ts
tech
n
o
lo
g
ical
i
n
n
o
v
atio
n
s
s
u
c
h
as
m
o
d
el
s
ca
li
n
g
,
r
e
-
p
ar
a
m
eter
ized
co
n
v
o
lu
tio
n
,
a
n
d
th
e
u
s
e
o
f
a
n
ef
f
ic
ien
t
la
y
er
a
g
g
r
e
g
atio
n
n
et
w
o
r
k
(
E
L
A
N)
.
T
h
e
s
t
u
d
y
r
esu
lt
s
in
d
icate
t
h
a
t
YOL
O
v
7
ac
h
ie
v
es
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
v
ar
io
u
s
o
b
j
ec
t
d
etec
tio
n
b
en
c
h
m
ar
k
s
,
w
ith
an
o
p
ti
m
al
b
alan
ce
b
et
w
ee
n
s
p
ee
d
an
d
ac
cu
r
ac
y
[
8
]
.
R
e
s
ea
r
ch
b
y
C
h
e
n
et
a
l
.
[
1
2
]
d
em
o
n
s
tr
ated
t
h
at
YOL
O
v
7
co
u
ld
b
e
u
s
ed
f
o
r
o
b
j
ec
t d
etec
tio
n
in
in
d
u
s
tr
ial
en
v
ir
o
n
m
e
n
ts
,
s
u
c
h
as
d
ef
ec
t d
etec
tio
n
o
n
p
r
o
d
u
ctio
n
lin
e
s
[
8
]
.
YOL
O
v
8
is
th
e
latest
v
er
s
io
n
o
f
th
e
YOL
O
o
b
j
ec
t
d
etec
tio
n
m
o
d
el.
T
h
is
v
er
s
io
n
in
tr
o
d
u
ce
s
a
n
e
w
n
eu
r
al
n
et
w
o
r
k
ar
c
h
itect
u
r
e
th
at
lev
er
ag
e
s
t
h
e
f
ea
tu
r
e
p
y
r
a
m
id
n
et
w
o
r
k
(
FP
N)
an
d
p
ath
a
g
g
r
eg
atio
n
n
e
t
w
o
r
k
(
P
A
N)
[
2
1
]
.
A
d
d
itio
n
all
y
,
Y
OL
O
v
8
co
m
es
w
i
th
n
e
w
lab
elin
g
to
o
ls
th
at
s
i
m
p
li
f
y
t
h
e
an
n
o
tatio
n
p
r
o
ce
s
s
,
in
cl
u
d
in
g
f
ea
t
u
r
es
li
k
e
a
u
to
-
l
ab
elin
g
,
lab
eli
n
g
s
h
o
r
tcu
t
s
,
a
n
d
cu
s
to
m
izab
le
h
o
t
k
e
y
s
.
T
h
ese
i
m
p
r
o
v
e
m
e
n
t
s
m
ak
e
th
e
i
m
a
g
e
a
n
n
o
tatio
n
p
r
o
ce
s
s
f
o
r
m
o
d
el
tr
ai
n
i
n
g
e
a
s
ie
r
.
Ho
w
ev
er
,
f
u
r
t
h
er
r
esear
ch
i
s
n
ee
d
ed
to
ev
alu
ate
YOL
O
v
8
'
s
p
er
f
o
r
m
a
n
ce
i
n
r
ea
l
-
w
o
r
ld
ap
p
licatio
n
s
.
T
h
e
d
ev
elo
p
m
e
n
t
o
f
v
ar
io
u
s
YOL
O
v
er
s
io
n
s
d
e
m
o
n
s
tr
ates
s
ig
n
i
f
ica
n
t
p
r
o
g
r
ess
in
o
b
j
ec
t
d
etec
tio
n
,
w
it
h
co
n
ti
n
u
o
u
s
i
m
p
r
o
v
e
m
e
n
ts
i
n
s
p
ee
d
an
d
ac
cu
r
ac
y
.
YOL
O
v
8
,
as
th
e
latest
v
er
s
io
n
,
o
f
f
er
s
v
ar
io
u
s
en
h
a
n
ce
m
en
ts
t
h
at
m
ak
e
i
t
a
m
o
r
e
e
f
f
icie
n
t
to
o
l
f
o
r
o
b
j
ec
t d
etec
tio
n
.
Fu
r
t
h
er
r
esear
ch
i
s
n
ec
es
s
ar
y
to
ev
al
u
at
e
YOL
O
v
8
'
s
p
er
f
o
r
m
a
n
ce
i
n
r
ea
l
-
w
o
r
ld
ap
p
licatio
n
s
an
d
co
m
p
ar
e
it
w
it
h
p
r
ev
io
u
s
v
er
s
io
n
s
.
3
.
2
.
YO
L
O
v
8
a
rc
hite
ct
ure
YOL
O
v
8
co
m
p
r
is
e
s
th
r
ee
m
a
in
m
o
d
u
les:
t
h
e
b
ac
k
b
o
n
e,
n
e
ck
,
an
d
h
ea
d
,
as
ill
u
s
tr
ated
in
Fig
u
r
e
1
.
YOL
O
v
8
m
a
in
ta
in
s
th
e
b
asic
s
tr
u
ct
u
r
e
o
f
YO
L
O
v
5
b
u
t
r
ep
lace
s
th
e
C
3
m
o
d
u
le
w
it
h
th
e
C
2
f
m
o
d
u
le,
w
h
ic
h
in
te
g
r
ates
t
h
e
C
SP
an
d
E
L
A
N
co
n
ce
p
ts
f
r
o
m
YO
L
O
v
7
[
2
1
]
.
T
h
e
C
2
f
m
o
d
u
le
en
h
a
n
ce
s
g
r
ad
ien
t
f
lo
w
,
i
m
p
r
o
v
i
n
g
th
e
ca
p
tu
r
e
an
d
r
ep
r
esen
tatio
n
o
f
i
m
a
g
e
f
ea
t
u
r
e
s
.
YOL
O
v
8
co
n
tin
u
es
to
u
s
e
th
e
SP
P
F
m
o
d
u
le
at
th
e
e
n
d
o
f
th
e
b
ac
k
b
o
n
e,
co
n
s
is
ti
n
g
o
f
th
r
ee
la
y
er
s
o
f
5
×
5
Ma
x
p
o
o
l,
en
s
u
r
i
n
g
ac
cu
r
ate
o
b
j
ec
t
d
etec
tio
n
at
v
ar
io
u
s
s
ca
les.
T
h
e
in
te
g
r
ati
o
n
o
f
C
2
f
an
d
SP
P
F
k
ee
p
s
YOL
O
v
8
li
g
h
t
w
eig
h
t
a
n
d
s
u
itab
le
f
o
r
r
ea
l
-
ti
m
e
ap
p
licatio
n
s
[
2
1
]
.
Fig
u
r
e
1
.
YOL
O
v
8
ar
ch
itect
u
r
e
[
2
0
]
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
A
u
g
u
s
t
20
25
:
1
0
1
3
-
1
026
1016
I
n
th
e
n
ec
k
s
ec
tio
n
,
YO
L
Ov
8
u
s
es
P
A
N
-
FP
N
to
co
m
b
in
e
an
d
u
tili
ze
in
f
o
r
m
atio
n
f
r
o
m
d
if
f
er
en
t
f
ea
t
u
r
e
s
ca
les
[
2
1
]
.
T
h
e
C
2
f
m
o
d
u
le
en
s
u
r
es
th
at
i
m
p
o
r
t
an
t
in
f
o
r
m
atio
n
is
n
o
t
m
is
s
i
n
g
d
u
r
i
n
g
m
er
g
i
n
g
.
YOL
O
v
8
also
ad
o
p
ts
a
d
ec
o
u
p
led
h
ea
d
s
tr
u
ct
u
r
e,
as
s
ee
n
in
YOL
Ox
,
to
i
m
p
r
o
v
e
ac
cu
r
ac
y
b
y
s
ep
ar
atin
g
class
i
f
icatio
n
a
n
d
r
eg
r
ess
io
n
t
ask
s
.
T
h
is
ap
p
r
o
ac
h
en
s
u
r
es
t
h
at
YO
L
O
v
8
r
e
m
ain
s
e
f
f
icie
n
t
a
n
d
ac
cu
r
ate
f
o
r
r
ea
l
-
ti
m
e
o
b
j
ec
t d
etec
tio
n
.
B
y
s
ep
ar
ati
n
g
t
h
e
clas
s
if
icatio
n
an
d
lo
ca
lizatio
n
b
r
an
c
h
es,
Y
OL
O
v
8
r
ed
u
ce
s
co
n
f
licts
b
et
w
ee
n
th
e
s
e
task
s
,
en
h
a
n
ci
n
g
o
v
er
all
ac
cu
r
ac
y
a
n
d
ef
f
icien
c
y
i
n
h
an
d
li
n
g
v
ar
io
u
s
t
y
p
e
s
o
f
o
b
j
ec
ts
i
n
i
m
a
g
es
[
2
2
]
.
T
h
is
d
ec
o
u
p
led
h
ea
d
ap
p
r
o
ac
h
en
s
u
r
es c
o
n
s
is
ten
t a
n
d
ac
cu
r
ate
r
es
u
lts
[
2
1
]
.
3
.
3
.
B
a
t
a
k
T
o
ba
s
cr
i
pt
det
ec
t
io
n
m
ec
ha
nis
m
T
h
e
d
ev
elo
p
m
e
n
t
o
f
a
B
atak
T
o
b
a
s
cr
ip
t
r
ec
o
g
n
itio
n
ap
p
licatio
n
in
v
o
lv
e
s
s
e
v
er
al
s
tag
e
s
:
p
r
ep
r
o
ce
s
s
in
g
,
lab
elin
g
,
an
d
tr
ain
in
g
d
ata
to
cr
ea
te
a
m
o
d
el
u
s
i
n
g
t
h
e
YO
L
O
v
8
alg
o
r
ith
m
.
T
h
e
d
etec
tio
n
ap
p
licatio
n
w
ill
b
e
in
te
g
r
ated
w
it
h
t
h
e
YO
L
O
v
8
m
o
d
el
a
n
d
w
il
l
in
c
lu
d
e
f
ea
tu
r
e
s
f
o
r
ca
p
tu
r
in
g
i
m
a
g
es
v
ia
a
ca
m
er
a
an
d
u
p
lo
ad
in
g
i
m
a
g
e
s
.
W
h
en
y
o
u
p
o
in
t
th
e
ca
m
er
a
at
th
e
B
atak
T
o
b
a
s
cr
ip
t,
t
h
e
ap
p
licatio
n
w
ill
au
to
m
at
icall
y
d
etec
t
t
h
e
ch
a
r
ac
ter
s
an
d
p
r
o
v
id
e
th
eir
co
r
r
esp
o
n
d
in
g
s
p
elli
n
g
.
T
h
e
ap
p
licatio
n
o
p
er
ates
th
r
o
u
g
h
s
e
v
er
al
s
tag
e
s
to
id
en
t
if
y
s
cr
ip
t
o
b
j
ec
ts
,
d
etec
t
b
o
u
n
d
in
g
b
o
x
es,
an
d
class
i
f
y
s
cr
ip
t
ch
ar
ac
ter
s
.
Deta
ils
o
f
th
e
s
e
s
ta
g
es a
r
e
s
h
o
w
n
i
n
Fi
g
u
r
e
2
.
I
n
Fig
u
r
e
2
,
th
e
f
ir
s
t
co
n
f
id
en
ce
s
co
r
e
is
f
o
u
n
d
in
th
e
f
ir
s
t
ch
an
n
el,
an
d
th
e
f
ir
s
t
b
o
u
n
d
i
n
g
b
o
x
w
i
t
h
ce
n
ter
co
o
r
d
in
ates
(
x
,
y
)
,
w
id
t
h
(
w
)
,
an
d
h
eig
h
t
(
h
)
ar
e
s
h
o
w
n
in
ch
an
n
el
s
t
w
o
th
r
o
u
g
h
f
iv
e,
m
ar
k
ed
w
it
h
p
u
r
p
le
g
r
id
s
.
Me
an
w
h
i
le,
th
e
s
ec
o
n
d
co
n
f
id
en
ce
s
co
r
e
is
i
n
th
e
s
ix
t
h
c
h
an
n
el,
a
n
d
th
e
s
e
co
n
d
b
o
u
n
d
in
g
b
o
x
w
it
h
ce
n
ter
co
o
r
d
in
ates
(
x
,
y
)
,
w
id
t
h
(
w
)
,
an
d
h
ei
g
h
t
(
h
)
ar
e
s
h
o
w
n
i
n
c
h
an
n
el
s
s
e
v
en
th
r
o
u
g
h
ten
,
m
ar
k
ed
w
it
h
g
r
ee
n
g
r
id
s
,
an
d
th
e
w
h
it
e
g
r
id
in
d
icate
s
t
h
e
n
u
m
b
er
o
f
d
etec
ted
o
b
j
ec
t c
lass
es.
Fig
u
r
e
2
.
Dete
ctio
n
o
f
B
atak
T
o
b
a
s
cr
ip
t
3
.
3
.
1
.
B
o
un
din
g
bo
x
predict
i
o
n
m
ec
ha
nis
m
T
h
e
m
o
d
el
p
r
ed
icts
b
o
u
n
d
in
g
b
o
x
es u
s
in
g
t
h
e
f
o
llo
w
i
n
g
m
ec
h
an
i
s
m
:
a)
T
h
e
in
p
u
t
i
m
a
g
e
i
s
d
i
v
id
ed
in
to
a
g
r
id
o
f
s
ize
S×S.
E
ac
h
c
ell
in
th
e
g
r
id
i
s
r
esp
o
n
s
ib
le
f
o
r
p
r
ed
ictin
g
w
h
et
h
er
an
o
b
j
ec
t’
s
ce
n
ter
is
w
it
h
i
n
t
h
at
ce
ll
.
b)
Featu
r
e
e
x
tr
ac
tio
n
i
s
p
er
f
o
r
m
ed
u
s
i
n
g
C
o
n
v
o
lu
tio
n
al
L
a
y
er
s
,
w
h
er
e
th
e
in
p
u
t
i
m
ag
e
i
s
p
r
o
ce
s
s
ed
th
r
o
u
g
h
a
s
er
ie
s
o
f
co
n
v
o
lu
t
io
n
al
la
y
er
s
to
ex
tr
ac
t
its
f
ea
t
u
r
es.
Fig
u
r
e
3
illu
s
tr
ates
h
o
w
t
h
e
o
u
tp
u
t
f
r
o
m
th
e
n
e
u
r
al
n
et
w
o
r
k
is
o
r
g
a
n
iz
ed
in
to
a
g
r
id
,
w
it
h
ea
ch
ce
ll
s
to
r
in
g
in
f
o
r
m
atio
n
n
ee
d
ed
to
d
etec
t
o
b
j
ec
ts
,
in
cl
u
d
in
g
b
o
u
n
d
in
g
b
o
x
p
r
ed
i
cti
o
n
s
,
co
n
f
id
e
n
ce
s
co
r
es,
a
n
d
class
ca
te
g
o
r
y
p
r
ed
ictio
n
s
.
E
ac
h
g
r
id
ce
l
l
co
n
tain
s
2
b
o
u
n
d
in
g
b
o
x
es a
n
d
2
co
n
f
id
en
ce
v
al
u
es.
c)
I
n
Fig
u
r
e
3
,
th
e
f
ir
s
t
co
n
f
id
en
ce
s
co
r
e
is
in
th
e
f
ir
s
t
ch
an
n
e
l,
an
d
f
o
r
th
e
f
ir
s
t
b
o
u
n
d
in
g
b
o
x
,
th
e
ce
n
ter
co
o
r
d
in
ates
(
x
,
y
)
,
w
id
th
(
w
)
,
an
d
h
ei
g
h
t
(
h
)
ar
e
s
h
o
w
n
i
n
ch
a
n
n
els
t
w
o
t
h
r
o
u
g
h
f
i
v
e,
m
ar
k
ed
w
it
h
p
u
r
p
le
g
r
id
s
.
T
h
e
s
ec
o
n
d
co
n
f
id
en
ce
s
co
r
e
is
in
t
h
e
s
i
x
th
c
h
an
n
el,
a
n
d
f
o
r
th
e
s
ec
o
n
d
b
o
u
n
d
in
g
b
o
x
,
th
e
ce
n
ter
co
o
r
d
in
ates
(
x
,
y
)
,
w
i
d
th
(
w
)
,
an
d
h
eig
h
t
(
h
)
ar
e
s
h
o
w
n
in
ch
a
n
n
els
s
e
v
e
n
th
r
o
u
g
h
ten
,
also
m
ar
k
ed
w
it
h
p
u
r
p
le
g
r
id
s
.
d)
B
o
u
n
d
in
g
b
o
x
p
r
ed
ictio
n
s
ar
e
m
ad
e
as
f
o
llo
w
s
:
E
ac
h
ce
ll
i
n
t
h
e
g
r
id
p
r
ed
icts
s
e
v
er
al
b
o
u
n
d
i
n
g
b
o
x
es
(
e.
g
.
,
N
b
o
u
n
d
in
g
b
o
x
es
p
e
r
ce
ll).
E
ac
h
b
o
u
n
d
in
g
b
o
x
is
d
escr
ib
ed
b
y
f
iv
e
p
ar
a
m
eter
s
:
ce
n
ter
co
o
r
d
in
ates
(
x
,
y
)
,
w
id
th
(
w
)
,
h
eig
h
t
(
h
)
,
an
d
co
n
f
id
en
ce
s
co
r
e.
T
h
e
co
o
r
d
in
ates
(
x
,
y
)
ar
e
n
o
r
m
alize
d
r
elativ
e
to
th
e
ce
ll
s
ize,
w
h
i
le
(
w
,
h
)
ar
e
n
o
r
m
alize
d
r
elativ
e
to
th
e
s
ize
o
f
t
h
e
en
tir
e
i
m
ag
e.
T
h
e
co
n
f
id
e
n
ce
s
co
r
e
r
e
f
lects
th
e
m
o
d
el
’
s
b
elie
f
i
n
t
h
e
p
r
ese
n
c
e
o
f
a
n
o
b
j
ec
t
w
it
h
i
n
t
h
e
b
o
u
n
d
in
g
b
o
x
a
n
d
th
e
ac
cu
r
ac
y
o
f
t
h
e
b
o
u
n
d
i
n
g
b
o
x
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
Dev
elo
p
men
t o
f m
o
b
ile
-
b
a
s
ed
B
a
ta
k
s
crip
t reco
g
n
itio
n
a
p
p
li
ca
tio
n
u
s
in
g
…
(
I
u
s
tis
ia
N
a
ta
lia
S
imb
o
lo
n
)
1017
e)
C
o
n
f
id
en
ce
s
co
r
e
f
o
r
m
u
la
tio
n
I
n
o
b
j
ec
t
d
etec
tio
n
,
th
e
m
o
d
el
o
u
tp
u
ts
b
o
u
n
d
in
g
b
o
x
p
r
ed
ict
io
n
s
m
ar
k
in
g
th
e
o
b
j
ec
t'
s
lo
ca
tio
n
alo
n
g
w
it
h
a
co
n
f
id
en
ce
s
co
r
e.
T
h
e
c
o
n
f
id
e
n
ce
s
co
r
e
is
ca
lcu
la
ted
as f
o
llo
w
s
:
=
(
)
×
ℎ
(
1
)
P
r
(
Ob
j
ec
t)
r
ep
r
esen
ts
th
e
p
r
o
b
ab
ilit
y
t
h
at
t
h
er
e
is
an
o
b
j
ec
t
w
it
h
i
n
th
e
p
r
ed
icted
b
o
u
n
d
i
n
g
b
o
x
.
T
h
is
p
r
o
b
a
b
ilit
y
r
ef
lect
s
th
e
m
o
d
el
's co
n
f
id
en
ce
th
a
t th
e
b
o
u
n
d
in
g
b
o
x
in
d
ee
d
co
n
tain
s
a
n
o
b
j
ec
t
.
Fig
u
r
e
3
.
B
o
u
n
d
in
g
b
o
x
p
r
ed
i
ctio
n
ℎ
is
a
m
etr
ic
u
s
ed
to
m
ea
s
u
r
e
th
e
o
v
er
lap
b
et
w
ee
n
t
w
o
b
o
u
n
d
i
n
g
b
o
x
es,
w
h
ic
h
is
cr
u
ci
al
f
o
r
i
m
p
r
o
v
i
n
g
d
etec
tio
n
ac
c
u
r
ac
y
.
T
h
e
p
r
esen
ce
o
f
m
u
ltip
le
b
o
u
n
d
in
g
b
o
x
e
s
in
d
icate
s
i
n
itial
d
etec
tio
n
r
esu
lt
s
b
y
th
e
m
o
d
el,
an
d
I
o
U
h
elp
s
ev
a
lu
ate
ac
c
u
r
ac
y
b
y
co
m
p
ar
in
g
p
r
ed
ictio
n
s
w
it
h
g
r
o
u
n
d
tr
u
t
h
b
o
u
n
d
in
g
b
o
x
es.
I
o
U
is
u
s
ed
in
t
h
e
n
o
n
-
m
ax
i
m
u
m
s
u
p
p
r
es
s
io
n
(
NM
S)
p
r
o
ce
s
s
to
r
e
m
o
v
e
o
v
er
lap
p
in
g
b
o
x
es
b
y
s
elec
ti
n
g
th
e
b
o
x
w
i
th
th
e
h
i
g
h
est
co
n
f
id
en
ce
s
co
r
e,
an
d
in
s
etti
n
g
th
r
e
s
h
o
ld
s
to
f
ilter
o
u
t
le
s
s
ac
cu
r
ate
p
r
ed
ictio
n
s
.
T
h
e
r
ef
o
r
e,
I
o
U
p
lay
s
a
v
ital
r
o
le
in
ac
h
iev
in
g
m
o
r
e
ac
cu
r
ate
o
b
j
ec
t
d
etec
tio
n
an
d
r
ed
u
cin
g
d
u
p
licate
b
o
u
n
d
i
n
g
b
o
x
es.
f)
T
h
r
esh
o
ld
an
d
NM
S
A
t
h
r
esh
o
ld
is
ap
p
lied
to
th
e
co
n
f
id
e
n
ce
s
co
r
e
o
f
ea
ch
p
r
e
d
ictio
n
.
P
r
ed
ictio
n
s
w
it
h
co
n
f
id
en
ce
s
co
r
es
b
elo
w
th
e
th
r
es
h
o
ld
ar
e
ig
n
o
r
ed
,
r
e
d
u
cin
g
t
h
e
n
u
m
b
er
o
f
b
o
u
n
d
in
g
b
o
x
es
co
n
s
id
er
ed
.
A
f
ter
ap
p
ly
in
g
th
e
th
r
es
h
o
ld
to
f
ilter
o
u
t
p
r
ed
ictio
n
s
w
it
h
lo
w
co
n
f
id
en
ce
s
c
o
r
es,
NM
S
is
u
s
ed
to
r
e
m
o
v
e
o
v
er
lap
p
in
g
b
o
u
n
d
in
g
b
o
x
es.
Am
o
n
g
m
u
lti
p
le
b
o
u
n
d
in
g
b
o
x
es
i
n
d
icati
n
g
th
e
s
a
m
e
o
b
j
ec
t,
NM
S
r
etain
s
o
n
l
y
t
h
e
b
o
x
w
it
h
t
h
e
h
i
g
h
e
s
t
co
n
f
id
en
ce
s
co
r
e
an
d
d
is
ca
r
d
s
th
e
o
th
er
s
.
T
h
is
p
r
o
ce
s
s
r
esu
lts
in
m
o
r
e
ac
cu
r
ate
o
b
j
ec
t
d
etec
tio
n
an
d
r
ed
u
ce
s
d
u
p
licat
io
n
,
p
r
o
d
u
cin
g
clea
n
er
an
d
cle
ar
er
r
esu
lts
,
as ill
u
s
tr
ated
in
Fi
g
u
r
e
4
.
Fig
u
r
e
4
.
Dete
r
m
in
a
tio
n
o
f
b
o
u
n
d
i
n
g
b
o
x
s
ize
an
d
co
o
r
d
in
ates
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
A
u
g
u
s
t
20
25
:
1
0
1
3
-
1
026
1018
3
.
3
.
2
.
Cha
ra
ct
er
identif
ica
t
io
n pre
dict
io
n
m
ec
ha
nis
m
Af
ter
t
h
e
b
o
u
n
d
i
n
g
b
o
x
d
ete
ctio
n
p
r
o
ce
s
s
is
co
m
p
lete,
t
h
e
n
ex
t
s
tep
is
c
h
ar
ac
ter
id
en
tif
icat
io
n
.
C
h
ar
ac
ter
id
en
ti
f
icatio
n
is
t
h
e
p
r
o
ce
s
s
o
f
d
eter
m
i
n
i
n
g
th
e
class
ca
te
g
o
r
y
o
f
th
e
o
b
j
ec
t
d
etec
ted
w
it
h
i
n
t
h
e
b
o
u
n
d
in
g
b
o
x
.
a)
C
o
n
d
itio
n
a
l
clas
s
p
r
o
b
ab
ilit
y
:
e
ac
h
g
r
id
ce
ll
al
s
o
p
r
ed
icts
a
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
t
io
n
f
o
r
th
e
p
o
ten
t
ial
o
b
j
ec
t c
lass
es
w
it
h
i
n
t
h
e
b
o
u
n
d
in
g
b
o
x
.
Fo
r
ex
a
m
p
le,
i
f
th
er
e
ar
e
N
p
o
s
s
ib
le
cla
s
s
e
s
,
ea
c
h
ce
ll p
r
ed
icts
an
N
-
s
ized
v
ec
to
r
w
h
er
e
ea
ch
el
e
m
en
t
r
ep
r
esen
t
s
t
h
e
p
r
o
b
ab
ilit
y
o
f
t
h
e
o
b
j
ec
t
b
elo
n
g
i
n
g
to
th
at
c
lass
.
I
n
Fig
u
r
e
2
,
th
e
w
h
ite
g
r
id
in
d
i
ca
tes
th
e
n
u
m
b
er
o
f
d
etec
ted
o
b
j
ec
t
class
es,
w
h
ich
i
s
1
8
1
class
es
in
t
h
i
s
s
tu
d
y
.
b)
C
las
s
p
r
ed
ictio
n
in
ea
ch
ce
ll
:
i
n
Fig
u
r
e
5
,
class
p
r
o
b
ab
ilit
ies
ar
e
ca
lcu
lated
u
s
in
g
th
e
s
o
f
t
m
ax
f
u
n
ctio
n
o
n
th
e
clas
s
o
u
tp
u
t
to
o
b
tain
a
v
a
lid
p
r
o
b
a
b
ilit
y
d
is
tr
ib
u
tio
n
.
T
h
ese
p
r
o
b
ab
ilit
ies
ar
e
co
n
d
itio
n
ed
o
n
th
e
g
r
id
ce
ll c
o
n
tain
in
g
t
h
e
o
b
j
ec
t
.
c)
Mu
ltip
l
y
i
n
g
cla
s
s
p
r
o
b
ab
ilit
y
an
d
co
n
f
id
en
ce
s
co
r
e
:
d
u
r
in
g
i
n
f
er
en
ce
,
th
e
cla
s
s
p
r
o
b
ab
ilit
y
o
f
ea
c
h
b
o
u
n
d
in
g
b
o
x
is
m
u
ltip
lied
b
y
th
e
co
n
f
id
e
n
ce
s
co
r
e
to
o
b
t
ain
th
e
f
i
n
al
co
n
f
id
en
ce
s
co
r
e
f
o
r
ea
ch
class
,
h
ig
h
li
g
h
ted
b
y
th
e
t
h
ic
k
es
t b
o
u
n
d
i
n
g
b
o
x
at
th
e
en
d
o
f
Fig
u
r
e
5
.
Fig
u
r
e
5
.
C
h
ar
ac
ter
o
b
j
ec
t id
e
n
ti
f
icatio
n
p
r
o
ce
s
s
d)
Dete
r
m
i
n
atio
n
o
f
cla
s
s
an
d
o
b
j
ec
t
class
s
co
r
e
:
t
h
is
clas
s
s
co
r
e
r
ef
lects
th
e
p
r
o
b
ab
ilit
y
o
f
th
e
clas
s
o
cc
u
r
r
in
g
w
it
h
in
th
e
b
o
u
n
d
i
n
g
b
o
x
a
n
d
h
o
w
w
el
l
t
h
e
b
o
u
n
d
in
g
b
o
x
p
r
ed
ictio
n
m
atc
h
e
s
t
h
e
o
b
j
ec
t.
T
h
e
class
w
it
h
t
h
e
h
i
g
h
est s
co
r
e
is
th
en
s
elec
ted
as t
h
e
f
i
n
al
cla
s
s
o
f
th
e
o
b
j
ec
t
w
it
h
in
t
h
at
b
o
u
n
d
in
g
b
o
x
.
3
.
4
.
Dev
el
o
p
m
ent
o
f
B
a
t
a
k
s
cr
ipt
det
ec
t
io
n
m
o
del w
i
t
h T
ens
o
rF
lo
w
li
te
T
en
s
o
r
Flo
w
L
ite
is
o
n
e
o
f
t
h
e
m
o
s
t p
o
p
u
lar
f
r
a
m
e
w
o
r
k
s
f
o
r
m
ac
h
in
e
lear
n
i
n
g
[
2
3
]
.
T
en
s
o
r
Flo
w
i
s
an
in
ter
f
ac
e
a
n
d
an
i
m
p
le
m
e
n
t
atio
n
f
o
r
ex
ec
u
ti
n
g
m
ac
h
in
e
lear
n
in
g
al
g
o
r
ith
m
s
p
r
esen
t
ed
in
2
0
1
5
[
2
4
]
.
T
en
s
o
r
f
lo
w
L
ite
i
s
a
f
r
ee
d
e
ep
lear
n
in
g
f
r
a
m
e
w
o
r
k
t
h
at
e
n
ab
les
d
ev
elo
p
er
s
to
b
u
ild
a
n
d
d
ep
lo
y
m
ac
h
in
e
lear
n
in
g
m
o
d
els
[
2
5
]
.
Dev
elo
p
in
g
an
A
n
d
r
o
id
ap
p
licatio
n
r
eq
u
ir
es
f
ea
t
u
r
es
th
at
en
ab
le
u
s
er
s
to
c
o
n
n
ec
t
w
it
h
th
e
ca
m
er
a
a
n
d
ca
p
tu
r
e
i
m
a
g
e
s
f
o
r
th
e
s
y
s
te
m
to
test
.
T
h
is
a
p
p
licatio
n
w
i
ll
b
e
b
u
ilt
u
s
in
g
An
d
r
o
id
Stu
d
io
an
d
Ko
tlin
,
w
i
th
T
en
s
o
r
Flo
w
L
i
t
e
as
t
h
e
f
r
a
m
e
w
o
r
k
to
i
m
p
l
e
m
en
t
t
h
e
m
ac
h
i
n
e
lear
n
i
n
g
m
o
d
el
f
o
r
i
m
a
g
e
r
ec
o
g
n
itio
n
.
3
.
5
.
E
x
peri
m
ent
s
T
h
e
r
esear
ch
b
eg
in
s
w
it
h
s
t
u
d
y
i
n
g
r
elev
a
n
t
liter
at
u
r
e
an
d
co
llectin
g
i
m
a
g
e
d
ata
o
f
t
h
e
B
atak
T
o
b
a
s
cr
ip
t
to
b
e
u
s
ed
.
On
ce
th
e
i
m
ag
e
d
ata
is
g
ath
er
ed
,
th
e
n
e
x
t
s
tep
is
i
m
a
g
e
p
r
o
ce
s
s
in
g
ai
m
e
d
at
im
p
r
o
v
i
n
g
t
h
e
q
u
alit
y
a
n
d
v
ar
iet
y
o
f
th
e
d
at
aset
o
r
p
r
ep
ar
in
g
th
e
d
ata
i
n
a
s
u
itab
le
f
o
r
m
at,
i
n
cl
u
d
in
g
a
u
g
m
e
n
tatio
n
:
i
m
ag
e
r
o
tatio
n
an
d
b
r
ig
h
t
n
es
s
ad
j
u
s
t
m
en
t,
r
esizin
g
,
an
d
d
ata
lab
eli
n
g
.
Af
ter
p
r
o
ce
s
s
in
g
th
e
i
m
a
g
es,
th
e
d
ata
is
d
iv
id
ed
in
to
th
r
ee
s
ets:
v
al
id
atio
n
(
2
0
%),
tr
ain
in
g
(
7
0
%),
an
d
test
i
n
g
(
1
0
%).
I
n
t
h
is
s
tu
d
y
[
2
2
]
,
th
e
d
ataset
d
iv
is
io
n
u
s
ed
w
a
s
7
0
:
2
0
:
1
0
,
w
h
ich
i
s
th
e
o
p
ti
m
al
d
ata
s
et
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
Dev
elo
p
men
t o
f m
o
b
ile
-
b
a
s
ed
B
a
ta
k
s
crip
t reco
g
n
itio
n
a
p
p
li
ca
tio
n
u
s
in
g
…
(
I
u
s
tis
ia
N
a
ta
lia
S
imb
o
lo
n
)
1019
d
iv
is
io
n
r
atio
ac
c
o
r
d
in
g
to
r
e
s
ea
r
ch
t
h
at
h
as
b
ee
n
co
n
d
u
c
t
ed
,
w
it
h
A
cc
u
r
ac
y
(
m
A
P
)
r
e
s
u
lts
o
f
0
.
9
9
5
1
an
d
tr
ain
i
n
g
ti
m
e
r
esu
lts
o
f
1
5
m
i
n
u
te
s
5
5
s
ec
o
n
d
s
.
T
h
e
tr
ain
in
g
d
ata
is
u
s
ed
to
tr
ain
th
e
o
b
ject
d
etec
tio
n
m
o
d
el
u
s
i
n
g
th
e
YO
L
O
v
8
alg
o
r
it
h
m
.
On
ce
tr
ain
i
n
g
i
s
co
m
p
lete,
t
h
e
m
o
d
el
is
e
v
alu
a
ted
u
s
i
n
g
t
h
e
v
alid
atio
n
s
et
to
ass
es
s
its
in
i
tial
p
er
f
o
r
m
an
ce
.
Mo
d
el
ev
alu
atio
n
en
s
u
r
es
th
e
m
o
d
el
w
o
r
k
s
w
ell
b
ef
o
r
e
f
i
n
al
test
i
n
g
.
T
esti
n
g
u
s
e
s
th
e
test
s
e
t
to
ass
ess
t
h
e
m
o
d
el
'
s
o
v
er
all
p
er
f
o
r
m
an
ce
an
d
s
elec
t
th
e
b
est
m
o
d
el
b
ased
o
n
ev
alu
at
io
n
r
esu
lt
s
.
T
h
e
o
p
tim
a
l
m
o
d
el
is
s
u
b
s
eq
u
e
n
tl
y
co
n
v
er
ted
to
th
e
T
FL
ite
f
o
r
m
a
t
f
o
r
in
teg
r
ati
o
n
an
d
d
ep
lo
y
m
e
n
t
w
it
h
i
n
th
e
ap
p
licatio
n
.
T
h
e
f
in
al
s
tep
is
d
ev
elo
p
in
g
an
ap
p
licatio
n
th
at
u
s
es
t
h
e
T
FL
it
e
m
o
d
el
f
o
r
o
b
j
ec
t
d
etec
tio
n
,
co
m
p
letin
g
t
h
e
p
r
o
ject
w
o
r
k
f
lo
w.
3
.
5
.
1
.
Da
t
a
s
et
I
n
th
i
s
s
t
u
d
y
,
t
h
e
d
ataset
co
lle
cted
co
n
s
is
t
s
o
f
1
8
1
class
es
w
i
th
2
,
4
0
8
an
n
o
tatio
n
s
.
T
h
er
e
ar
e
1
,
8
5
7
lab
els f
o
r
tr
ain
i
n
g
d
ata,
5
3
2
lab
els f
o
r
v
alid
atio
n
d
ata,
an
d
2
5
4
lab
els f
o
r
test
in
g
d
ata.
T
h
e
co
llected
d
ataset
u
n
d
er
g
o
e
s
p
r
ep
r
o
ce
s
s
in
g
,
lab
e
lin
g
,
an
d
d
ataset
s
p
litt
in
g
.
a)
I
m
ag
e
p
r
ep
r
o
ce
s
s
in
g
T
h
e
co
llected
s
cr
ip
t
i
m
ag
e
s
ar
e
r
esized
to
6
4
0
×
6
4
0
p
ix
els.
Fig
u
r
e
6
s
h
o
w
s
th
e
r
es
u
lt
s
o
f
th
e
r
esizin
g
p
r
o
ce
s
s
.
Af
ter
r
esizi
n
g
,
t
h
e
n
e
x
t
s
tep
i
s
au
g
m
en
tatio
n
:
r
o
tati
o
n
w
it
h
s
ca
le
s
o
f
-
2
°,
-
4
°,
-
6
°,
-
8
°,
-
1
0
°,
1
0
°,
8
°,
6
°,
4
°,
2
° as sh
o
w
n
i
n
Fi
g
u
r
e
7
,
an
d
b
r
ig
h
tn
e
s
s
ad
j
u
s
t
m
en
t w
it
h
a
s
ca
le
o
f
0
.
7
as sh
o
w
n
i
n
Fig
u
r
e
8
.
Fig
u
r
e
6
.
R
esizi
n
g
o
f
s
cr
ip
t i
m
ag
e
d
i
m
e
n
s
io
n
s
R
o
t
a
t
e
-
2°
R
o
t
a
t
e
2
°
R
o
t
a
t
e
4
°
R
o
t
a
t
e
-
4°
R
o
t
a
t
e
-
6°
R
o
t
a
t
e
6
°
R
o
t
a
t
e
-
8°
R
o
t
a
t
e
8
°
R
o
t
a
t
e
-
10°
R
o
t
a
t
e
1
0
°
Fig
u
r
e
7
.
R
o
tatio
n
o
f
B
atak
T
o
b
a
s
cr
ip
t m
a
g
es
Fig
u
r
e
8
.
B
r
ig
h
tn
e
s
s
ad
j
u
s
t
m
e
n
t o
f
B
atak
T
o
b
a
s
cr
ip
t im
a
g
e
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
A
u
g
u
s
t
20
25
:
1
0
1
3
-
1
026
1020
b)
Data
s
et
lab
eli
n
g
I
n
th
i
s
s
t
u
d
y
,
lab
elin
g
w
as
p
er
f
o
r
m
ed
u
s
i
n
g
t
h
e
R
o
b
o
f
l
o
w
p
latf
o
r
m
,
i
n
v
o
lv
i
n
g
th
e
cr
ea
tio
n
o
f
r
ec
tan
g
u
lar
b
o
u
n
d
i
n
g
b
o
x
e
s
ar
o
u
n
d
ea
ch
s
cr
ip
t
o
b
j
ec
t.
U
p
o
n
co
m
p
letio
n
,
a
f
ile
co
n
ta
in
i
n
g
all
t
h
e
lab
el
class
es,
ca
lled
clas
s
es.t
x
t,
w
as
g
en
er
ated
.
T
h
e
r
esu
lts
f
r
o
m
t
h
is
s
ta
g
e
w
ill p
r
o
ce
ed
to
th
e
tr
ain
in
g
p
r
o
ce
s
s
.
c)
Data
s
e
t
s
p
litt
in
g
T
h
e
d
etails
o
f
t
h
e
d
ata
s
et
d
i
v
i
s
io
n
i
n
t
h
i
s
s
tu
d
y
ar
e
s
h
o
w
n
i
n
T
ab
le
1
.
T
h
e
d
ataset
is
s
p
li
t
in
to
7
0
%
f
o
r
tr
ain
i
n
g
,
2
0
%
f
o
r
v
al
id
ati
o
n
,
an
d
1
0
%
f
o
r
te
s
tin
g
.
T
h
is
d
is
tr
ib
u
tio
n
w
ill
p
r
o
ce
ed
to
th
e
tr
ai
n
in
g
p
h
a
s
e
to
o
b
tain
th
e
o
p
ti
m
al
m
o
d
el.
T
ab
le
1
.
Data
s
et
s
p
litt
in
g
D
a
t
a
se
t
s
p
l
i
t
t
i
n
g
N
u
mb
e
r
o
f
i
m
a
g
e
s
T
r
a
i
n
S
e
t
7
0
%
1
8
5
7
V
a
l
i
d
S
e
t
2
0
%
5
3
2
T
e
st
S
e
t
1
0
%
2
5
4
T
o
t
a
l
2
6
4
3
3
.
5
.
2
.
H
y
perpa
ra
m
et
er
ev
a
lu
a
t
io
n
T
h
e
d
ata
v
alid
atio
n
p
r
o
ce
s
s
in
th
is
s
t
u
d
y
i
n
v
o
l
v
es
d
eter
m
i
n
i
n
g
cr
u
cia
l
h
y
p
er
p
ar
a
m
eter
s
,
s
u
ch
as
t
h
e
I
o
U
th
r
esh
o
ld
,
at
th
e
in
i
tial
s
ta
g
e.
Fo
llo
w
in
g
th
is
,
th
e
v
alid
ati
o
n
r
esu
lt
s
ar
e
ev
alu
ated
,
an
d
a
th
o
r
o
u
g
h
an
al
y
s
i
s
is
co
n
d
u
cted
d
u
r
in
g
th
e
d
ec
is
io
n
-
m
a
k
i
n
g
p
h
ase.
I
f
t
h
e
m
ea
n
av
er
ag
e
p
r
ec
is
io
n
(
m
A
P
)
ac
h
iev
e
s
a
h
ig
h
an
d
s
tab
le
p
er
f
o
r
m
an
ce
lev
el
ac
co
r
d
in
g
to
t
h
e
s
ta
n
d
ar
d
v
alu
e,
a
r
ee
v
alu
a
tio
n
w
i
th
d
i
f
f
er
en
t
h
y
p
er
p
ar
am
eter
v
alu
e
s
is
co
n
d
u
cted
to
ac
h
iev
e
o
p
ti
m
al
r
esu
lts
.
A
s
tan
d
ar
d
v
alu
e
is
co
n
s
id
er
ed
g
o
o
d
at
0
.
5
,
an
d
v
alu
es
ab
o
v
e
0
.
5
a
r
e
d
ee
m
ed
v
er
y
g
o
o
d
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
4.
1
.
T
ra
ini
ng
da
t
a
re
s
ults a
nd
dis
c
us
s
io
n
Af
ter
g
at
h
er
in
g
i
m
ag
e
d
ata
an
d
s
p
litt
in
g
it,
th
e
d
ataset
u
n
d
er
g
o
es
th
e
tr
ain
i
n
g
p
r
o
ce
s
s
.
T
h
e
r
esu
lts
ar
e
s
h
o
w
n
in
T
ab
les
2
an
d
3
.
B
ased
o
n
T
ab
le
3
,
th
e
ex
p
er
i
m
e
n
t
r
es
u
lts
w
it
h
v
ar
io
u
s
ep
o
ch
s
s
h
o
w
t
h
at
t
h
e
p
ar
am
eter
w
it
h
2
0
0
ep
o
ch
s
h
a
s
a
h
ig
h
er
m
A
P
v
al
u
e
co
m
p
ar
ed
to
o
th
er
ep
o
ch
p
ar
am
eter
s
,
s
p
ec
if
icall
y
0
.
9
5
4
.
A
h
ig
h
m
A
P
v
al
u
e
in
d
icate
s
a
g
o
o
d
b
alan
ce
b
etw
ee
n
p
r
ec
is
io
n
an
d
r
ec
all.
T
h
is
m
ea
n
s
th
e
m
o
d
el
i
s
g
en
er
all
y
b
etter
at
d
etec
tin
g
a
n
d
r
ec
o
g
n
izi
n
g
T
o
b
a
s
cr
ip
t
o
b
j
ec
ts
u
n
d
er
v
ar
io
u
s
co
n
d
itio
n
s
.
B
esid
es
m
A
P
,
f
u
r
th
e
r
i
m
p
r
o
v
e
m
en
t
s
in
p
r
ec
is
io
n
an
d
r
ec
all
d
em
o
n
s
tr
ate
th
at
t
h
e
m
o
d
el
i
s
m
o
r
e
r
eliab
le
in
d
ete
ctin
g
T
o
b
a
s
cr
ip
t.
Fig
u
r
e
9
s
h
o
w
s
t
h
e
tr
ain
i
n
g
r
e
s
u
lt
s
f
r
o
m
T
ab
le
3
w
it
h
th
e
x
-
ax
is
s
h
o
w
i
n
g
t
h
e
n
u
m
b
er
o
f
e
p
o
ch
s
an
d
th
e
y
-
a
x
is
o
f
t
h
e
f
ir
s
t
i
m
ag
e
t
h
e
r
an
g
e
o
f
m
A
P
v
a
lu
e
s
,
th
e
s
e
co
n
d
i
m
ag
e
t
h
e
p
r
ec
is
io
n
an
d
th
e
th
ir
d
th
e
r
ec
all.
T
h
e
T
o
b
a
s
cr
ip
t
m
o
d
el
s
h
o
ws
s
i
g
n
i
f
ica
n
t
i
m
p
r
o
v
e
m
en
t
s
i
n
p
r
ec
is
io
n
,
r
ec
all,
an
d
m
ea
n
Av
er
ag
e
P
r
ec
is
io
n
(
m
A
P
)
m
etr
ics
at
a
t
h
r
es
h
o
ld
o
f
0
.
5
d
u
r
in
g
tr
ai
n
in
g
.
P
r
ec
is
i
o
n
an
d
r
ec
all
v
al
u
es
in
cr
ea
s
ed
f
r
o
m
0
.
0
to
n
ea
r
l
y
0
.
9
at
ep
o
ch
1
5
0
,
th
en
s
tab
iliz
ed
u
n
til
ep
o
ch
2
0
0
,
in
d
icati
n
g
th
e
m
o
d
el
r
ea
ch
ed
o
p
ti
m
al
ac
cu
r
ac
y
i
n
av
o
id
i
ng
f
alse
p
o
s
it
iv
e
s
an
d
d
etec
ti
n
g
al
m
o
s
t
al
l
o
b
j
ec
ts
.
T
h
e
m
A
P
v
al
u
e
also
i
n
cr
ea
s
ed
f
r
o
m
0
.
0
to
n
ea
r
l
y
0
.
9
5
at
ep
o
ch
1
5
0
an
d
s
tab
ilized
u
n
til
ep
o
ch
2
0
0
,
r
ef
lectin
g
g
o
o
d
o
v
er
all
m
o
d
el
p
er
f
o
r
m
an
ce
i
n
o
b
j
ec
t
d
etec
tio
n
.
T
h
ese
r
esu
lt
s
i
n
d
icate
t
h
e
m
o
d
el
ef
f
ec
t
iv
e
l
y
lear
n
ed
o
v
er
2
0
0
ep
o
ch
s
,
w
i
th
i
m
p
r
o
v
ed
ac
cu
r
ac
y
a
n
d
g
o
o
d
p
atter
n
r
ec
o
g
n
itio
n
in
t
h
e
d
ata.
T
ab
le
2
.
T
r
ain
in
g
ti
m
e
p
r
o
ce
s
s
b
ased
o
n
ep
o
ch
s
ize
I
mag
e
si
z
e
Ep
o
c
h
T
r
a
i
n
i
n
g
d
u
r
a
t
i
o
n
D
e
scri
p
t
i
o
n
6
4
0
x
6
4
0
50
0
.
7
h
o
u
r
s
T
r
a
i
n
i
n
g
su
c
c
e
ssf
u
l
.
6
4
0
x
6
4
0
1
0
0
1
.
5
h
o
u
r
s
T
r
a
i
n
i
n
g
su
c
c
e
ssf
u
l
6
4
0
x
6
4
0
1
5
0
2
.
9
h
o
u
r
s
T
r
a
i
n
i
n
g
su
c
c
e
ssf
u
l
6
4
0
x
6
4
0
2
0
0
2
.
9
h
o
u
r
s
T
r
a
i
n
i
n
g
su
c
c
e
ssf
u
l
T
ab
le
3
.
E
x
p
er
im
e
n
t r
esu
lts
with
v
ar
io
u
s
ep
o
ch
s
Ep
o
c
h
P
r
e
c
i
si
o
n
R
e
c
a
l
l
mA
P
@
.
5
50
0
.
7
6
3
0
.
8
0
3
0
.
8
5
8
1
0
0
0
.
8
4
1
0
.
8
7
4
0
.
9
2
5
1
5
0
0
.
8
5
6
0
.
8
9
6
0
.
9
5
2
2
0
0
0
.
9
4
5
0
.
9
0
2
0
.
9
5
4
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4752
Dev
elo
p
men
t o
f m
o
b
ile
-
b
a
s
ed
B
a
ta
k
s
crip
t reco
g
n
itio
n
a
p
p
li
ca
tio
n
u
s
in
g
…
(
I
u
s
tis
ia
N
a
ta
lia
S
imb
o
lo
n
)
1021
Fig
u
r
e
9
.
H
y
p
er
p
ar
am
eter
tr
ai
n
in
g
d
ataset
e
v
al
u
atio
n
r
es
u
lt
s
4
.
2
.
M
o
del
ev
a
lua
t
i
o
n r
esu
lt
s
a
nd
dis
c
us
s
io
n
P
r
ec
is
io
n
,
r
ec
all,
an
d
m
A
P
v
alu
es
d
u
r
i
n
g
tr
ai
n
i
n
g
ar
e
in
f
l
u
en
ce
d
b
y
v
ar
io
u
s
f
ac
to
r
s
,
in
clu
d
in
g
th
e
n
u
m
b
er
o
f
ep
o
ch
s
,
t
h
e
q
u
alit
y
an
d
q
u
an
tit
y
o
f
tr
ai
n
i
n
g
d
ata,
p
r
ep
r
o
ce
s
s
in
g
a
n
d
au
g
m
e
n
tat
io
n
tech
n
iq
u
e
s
,
an
d
h
y
p
er
p
ar
a
m
eter
s
.
Mo
d
els
tr
ain
ed
w
it
h
a
h
i
g
h
er
n
u
m
b
er
o
f
ep
o
ch
s
ten
d
to
h
av
e
b
etter
p
r
ec
is
io
n
,
r
ec
all,
an
d
m
A
P
b
ec
a
u
s
e
t
h
e
y
h
av
e
m
o
r
e
o
p
p
o
r
tu
n
itie
s
to
lear
n
an
d
o
p
ti
m
ize
t
h
eir
p
er
f
o
r
m
an
ce
f
r
o
m
t
h
e
av
ailab
le
d
ata.
T
h
e
T
o
b
a
s
cr
ip
t
o
b
j
ec
t
d
etec
tio
n
ex
p
er
i
m
e
n
ts
u
s
i
n
g
m
o
d
els
w
it
h
5
0
,
1
0
0
,
1
5
0
,
an
d
2
0
0
e
p
o
ch
s
ar
e
p
r
esen
ted
in
Fi
g
u
r
e
1
0
.
T
h
e
m
o
d
el
u
s
ed
in
Fi
g
u
r
e
1
0
(
a)
h
as
5
0
ep
o
ch
s
.
T
h
e
r
esu
lts
s
h
o
w
in
co
r
r
ec
t
o
b
j
ec
t
d
etec
tio
n
,
w
h
er
e
th
e
co
r
r
ec
t
lab
el
s
h
o
u
ld
b
e
"P
U
"
in
s
tead
o
f
"J
U
"
.
Fo
r
im
a
g
es
w
i
th
1
0
0
ep
o
ch
s
Fig
u
r
e
1
0
(
b
)
an
d
1
5
0
ep
o
ch
s
Fig
u
r
e1
0
(
c)
,
th
e
d
etec
tio
n
r
esu
lt
s
b
o
th
s
h
o
w
a
co
n
f
id
e
n
ce
v
alu
e
o
f
9
4
%,
b
u
t
w
i
th
d
if
f
er
en
t
d
e
t
e
c
t
i
o
n
s
p
e
e
d
s
,
w
i
t
h
t
h
e
m
o
d
e
l
t
r
a
i
n
e
d
f
o
r
1
5
0
e
p
o
c
h
s
d
e
t
e
c
t
i
n
g
f
a
s
t
e
r
t
h
a
n
t
h
e
1
0
0
-
e
p
o
c
h
m
o
d
e
l
.
F
o
r
t
h
e
i
m
a
g
e
w
i
t
h
t
h
e
2
0
0
e
p
o
c
h
F
i
g
u
r
e
1
0
(
d
)
m
o
d
e
l
,
t
h
e
d
e
t
e
c
t
i
o
n
r
e
s
u
l
t
s
s
h
o
w
a
v
e
r
y
g
o
o
d
d
e
t
e
c
t
i
o
n
a
c
c
u
r
a
c
y
o
f
9
9
%
.
(
a)
(
b
)
(
c)
(
d
)
Fig
u
r
e
1
0
.
E
x
p
er
im
e
n
t r
es
u
lt
s
w
it
h
v
ar
io
u
s
ep
o
ch
s
: (
a)
5
0
ep
o
ch
,
(
b
)
1
0
0
ep
o
ch
,
(
c)
1
5
0
e
p
o
ch
,
an
d
(
d
)
2
0
0
ep
o
ch
T
h
e
r
esear
ch
ex
p
er
i
m
en
ts
u
s
in
g
co
n
f
id
en
ce
th
r
e
s
h
o
ld
p
ar
a
m
eter
s
o
f
0
.
7
an
d
0
.
4
ar
e
s
h
o
w
n
in
T
ab
les
4
an
d
5
.
T
h
e
co
m
p
ar
is
o
n
o
f
d
etec
tio
n
r
es
u
lt
s
w
it
h
t
h
ese
co
n
f
id
e
n
ce
th
r
e
s
h
o
ld
v
a
l
u
es
p
r
o
v
es
t
h
at
t
h
e
h
ig
h
er
t
h
e
th
r
es
h
o
ld
,
th
e
m
o
r
e
ac
cu
r
ate
an
d
b
etter
th
e
m
o
d
el
d
etec
ts
B
atak
T
o
b
a
s
cr
ip
t.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4752
I
n
d
o
n
esia
n
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
39
,
No
.
2
,
A
u
g
u
s
t
20
25
:
1
0
1
3
-
1
026
1022
T
ab
le
4
.
E
x
p
er
im
e
n
t
w
it
h
co
n
f
id
en
ce
th
r
es
h
o
ld
p
ar
a
m
eter
0
.
4
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
R
e
mark
s
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
b
y
t
h
e
2
0
0
-
e
p
o
c
h
m
o
d
e
l
w
i
t
h
a
c
o
n
f
i
d
e
n
c
e
t
h
r
e
sh
o
l
d
p
a
r
a
me
t
e
r
o
f
0
.
4
.
T
h
e
r
e
a
r
e
st
i
l
l
t
w
o
b
o
u
n
d
i
n
g
b
o
x
e
s w
i
t
h
d
i
f
f
e
r
e
n
t
c
l
a
sse
s,
TA
a
n
d
T
.
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
b
y
t
h
e
2
0
0
-
e
p
o
c
h
m
o
d
e
l
w
i
t
h
a
c
o
n
f
i
d
e
n
c
e
t
h
r
e
sh
o
l
d
p
a
r
a
me
t
e
r
o
f
0
.
4
.
T
h
e
r
e
a
r
e
st
i
l
l
t
w
o
b
o
u
n
d
i
n
g
b
o
x
e
s w
i
t
h
d
i
f
f
e
r
e
n
t
c
l
a
sse
s,
N
A
a
n
d
D
O
.
T
ab
le
5
.
Var
io
u
s
ca
s
es e
x
p
er
im
en
t
w
it
h
YO
L
O
v
8
m
o
d
el
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
R
e
mark
s
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
u
s
i
n
g
t
h
e
2
0
0
-
e
p
o
c
h
mo
d
e
l
.
T
h
e
mo
d
e
l
su
c
c
e
ssf
u
l
l
y
d
e
t
e
c
t
s a
s
i
n
g
l
e
scri
p
t
c
h
a
r
a
c
t
e
r
B
EN
G
w
i
t
h
a
c
o
n
f
i
d
e
n
c
e
o
f
9
7
%.
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
u
s
i
n
g
t
h
e
2
0
0
-
e
p
o
c
h
mo
d
e
l
.
T
h
e
mo
d
e
l
su
c
c
e
ssf
u
l
l
y
d
e
t
e
c
t
s t
h
e
c
h
a
r
a
c
t
e
r
U
w
i
t
h
a
c
o
n
f
i
d
e
n
c
e
o
f
9
7
%
a
n
d
i
g
n
o
r
e
s o
t
h
e
r
n
o
n
-
T
o
b
a
s
c
r
i
p
t
o
b
j
e
c
t
s
.
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
u
s
i
n
g
t
h
e
2
0
0
-
e
p
o
c
h
mo
d
e
l
.
T
h
e
mo
d
e
l
a
c
c
u
r
a
t
e
l
y
d
e
t
e
c
t
s b
o
t
h
c
h
a
r
a
c
t
e
r
s,
TI
w
i
t
h
a
c
o
n
f
i
d
e
n
c
e
o
f
9
3
%
a
n
d
T
O
w
i
t
h
a
c
o
n
f
i
d
e
n
c
e
o
f
9
1
%.
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
u
s
i
n
g
t
h
e
2
0
0
-
e
p
o
c
h
mo
d
e
l
.
T
h
e
mo
d
e
l
d
o
e
s n
o
t
d
i
s
p
l
a
y
b
o
u
n
d
i
n
g
b
o
x
e
s a
n
d
c
l
a
ss
n
a
me
s
f
o
r
t
h
e
d
e
t
e
c
t
e
d
a
l
p
h
a
b
e
t
b
e
c
a
u
se
t
h
e
t
e
st
e
d
i
mag
e
i
s
n
o
t
B
a
t
a
k
T
o
b
a
scr
i
p
t
.
D
e
t
e
c
t
i
o
n
r
e
su
l
t
s
u
s
i
n
g
t
h
e
2
0
0
-
e
p
o
c
h
mo
d
e
l
.
T
h
e
mo
d
e
l
d
o
e
s n
o
t
d
i
s
p
l
a
y
b
o
u
n
d
i
n
g
b
o
x
e
s a
n
d
c
l
a
ss
n
a
me
s
f
o
r
t
h
e
d
e
t
e
c
t
e
d
p
a
t
t
e
r
n
b
e
c
a
u
se
t
h
e
t
e
st
e
d
i
mag
e
i
s
n
o
t
B
a
t
a
k
T
o
b
a
scri
p
t
.
4
.
3
.
Resul
t
s
a
nd
dis
cu
s
s
io
n
o
f
s
cr
ipt
det
ec
t
io
n w
it
h t
he
a
nd
ro
id a
pp
lica
t
io
n
T
h
e
m
o
d
el
o
b
tain
ed
f
r
o
m
th
e
p
r
ev
io
u
s
i
m
p
le
m
e
n
tatio
n
w
a
s
co
n
v
er
ted
f
r
o
m
th
e
.
p
t
f
o
r
m
at
to
th
e
tf
lite
f
o
r
m
at.
T
h
is
m
o
d
el
was
th
e
n
i
n
te
g
r
ated
i
n
to
an
An
d
r
o
id
ap
p
licatio
n
to
p
er
f
o
r
m
r
ea
l
-
t
i
m
e
o
b
j
ec
t
d
etec
tio
n
an
d
i
m
ag
e
u
p
lo
ad
s
.
Du
r
in
g
test
i
n
g
,
th
e
b
o
u
n
d
i
n
g
b
o
x
es
p
r
o
d
u
ce
d
b
y
th
e
m
o
d
el
w
er
e
co
n
s
is
te
n
tl
y
ac
cu
r
ate
in
lo
ca
ti
n
g
t
h
e
d
esir
ed
s
cr
ip
t
p
o
s
itio
n
s
.
T
h
e
d
ete
cted
class
n
a
m
es
m
atc
h
ed
th
e
an
n
o
tated
s
cr
ip
ts
,
in
d
icati
n
g
t
h
at
th
e
m
o
d
el
h
a
s
g
o
o
d
class
if
icat
io
n
ca
p
ab
ilit
y
.
T
h
e
d
etec
tio
n
p
r
o
ce
s
s
tim
e
r
an
g
ed
f
r
o
m
1
to
2
s
ec
o
n
d
s
p
er
i
m
a
g
e,
d
ep
en
d
in
g
o
n
th
e
i
m
ag
e
co
m
p
le
x
it
y
(
n
u
m
b
er
a
n
d
t
y
p
e
o
f
s
cr
ip
t
f
o
r
m
at
s
)
.
T
h
e
d
etec
tio
n
r
esu
lts
ar
e
d
is
p
l
a
y
ed
as
an
n
o
tatio
n
s
o
n
t
h
e
ca
m
er
a
s
cr
ee
n
.
R
ea
l
-
ti
m
e
d
etec
tio
n
r
esu
l
ts
ar
e
s
h
o
w
n
in
F
ig
u
r
e
1
1
.
Fig
u
r
e
1
1
(
a
)
illu
s
tr
ates
t
h
e
d
etec
tio
n
o
f
th
e
B
atak
T
o
b
a
s
cr
ip
t
in
d
ig
ital
f
o
r
m
at
w
ith
ac
cu
r
ac
y
,
d
etec
tin
g
"
B
A
"
w
i
t
h
a
co
n
f
id
e
n
ce
o
f
8
1
%
an
d
id
en
ti
f
y
in
g
th
e
co
m
b
i
n
atio
n
o
f
"
B
E
NG
"
an
d
"T
U
"
w
it
h
co
n
f
id
e
n
ce
v
al
u
es
o
f
9
7
%
an
d
9
3
%,
r
esp
ec
tiv
el
y
.
F
ig
u
r
e
1
1
(
b
)
s
h
o
w
s
t
h
e
d
etec
tio
n
o
f
th
e
B
atak
T
o
b
a
s
cr
ip
t in
h
a
n
d
w
r
itte
n
f
o
r
m
at,
ac
cu
r
atel
y
id
e
n
ti
f
y
in
g
t
h
e
co
m
b
in
atio
n
s
"
MI
,
"
"T
I
,
"
"T
O,
"
an
d
"T
U
.
"
.
B
atak
T
o
b
a
s
cr
ip
t
ca
n
also
b
e
d
etec
ted
u
s
in
g
t
h
e
i
m
ag
e
u
p
lo
ad
f
u
n
ctio
n
o
n
th
e
p
h
o
n
e,
as
s
h
o
w
n
i
n
Fig
u
r
e
1
2
.
User
s
ca
n
also
s
el
ec
t
i
m
a
g
es
f
r
o
m
th
eir
p
h
o
n
e
's
g
aller
y
,
an
d
t
h
e
ap
p
l
icatio
n
w
i
ll
p
r
o
ce
s
s
t
h
es
e
i
m
a
g
es
to
d
etec
t
an
d
r
ec
o
g
n
iz
e
th
e
T
o
b
a
s
cr
ip
t
as
s
h
o
w
n
i
n
Fig
u
r
e
1
2
(
a)
s
h
o
w
s
t
h
e
d
ete
ctio
n
o
f
“
R
U”
an
d
“RU NY
U”
as
f
o
r
Fig
u
r
e
1
2
(
b
)
s
h
o
w
s
t
h
e
d
etec
tio
n
o
f
“
DO”
an
d
“
DO
M
A
.
”
Evaluation Warning : The document was created with Spire.PDF for Python.