I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
2
,
A
p
r
il
20
25
,
p
p
.
1
8
3
4
~
1
8
4
9
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
2
.
pp
1
8
3
4
-
1
8
4
9
1834
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Identifica
tion o
f
Andro
id A
PK ma
lwa
re t
hro
ug
h
lo
ca
l and
g
lo
ba
l f
ea
ture
e
x
t
ra
ction using
me
t
a
clas
sifier
Yo
g
a
H
er
a
w
a
n,
I
m
a
s
Su
k
a
es
ih Sita
ng
g
a
ng
,
Sh
elv
ie
Nidy
a
Ney
m
a
n
D
e
p
a
r
t
me
n
t
o
f
C
o
mp
u
t
e
r
S
c
i
e
n
c
e
,
F
a
c
u
l
t
y
o
f
M
a
t
h
e
ma
t
i
c
s
a
n
d
N
a
t
u
r
a
l
S
c
i
e
n
c
e
,
I
P
B
U
n
i
v
e
r
s
i
t
y
,
B
o
g
o
r
,
I
n
d
o
n
e
s
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
r
2
3
,
2
0
2
4
R
ev
is
ed
Oct
1
8
,
2
0
2
4
Acc
ep
ted
No
v
2
0
,
2
0
2
4
An
d
ro
id
,
t
h
e
m
o
st
wid
e
ly
u
se
d
m
o
b
il
e
o
p
e
ra
ti
n
g
s
y
ste
m
,
is
a
lso
th
e
m
o
st
v
u
l
n
e
ra
b
le
to
m
a
lwa
re
d
u
e
to
i
ts
h
ig
h
p
o
p
u
larit
y
.
Th
is
h
a
s
sig
n
ifi
c
a
n
tl
y
fo
c
u
se
d
o
n
An
d
r
o
id
m
a
lwa
re
d
e
tec
ti
o
n
in
m
o
b
il
e
se
c
u
rit
y
.
W
h
il
e
e
x
ten
siv
e
re
se
a
rc
h
h
a
s
b
e
e
n
c
o
n
d
u
c
ted
u
sin
g
v
a
rio
u
s
m
e
th
o
d
s,
n
e
w
m
a
lwa
re
’s
e
m
e
rg
e
n
c
e
u
n
d
e
rsc
o
re
s
th
is
fie
ld
’s
d
y
n
a
m
ic
n
a
tu
re
a
n
d
th
e
n
e
e
d
fo
r
c
o
n
ti
n
u
o
u
s
re
se
a
rc
h
.
T
h
e
m
o
ti
v
a
ti
o
n
th
a
t
d
ri
v
e
s
m
a
lwa
re
d
e
v
e
lo
p
e
rs
t
o
c
re
a
te
An
d
ro
i
d
m
a
lwa
re
c
o
n
sta
n
tl
y
is
th
e
p
o
ten
ti
a
l
t
o
a
c
c
e
ss
An
d
ro
i
d
d
e
v
ice
s,
th
e
re
b
y
g
a
in
i
n
g
a
c
c
e
ss
to
se
n
siti
v
e
u
se
r
i
n
fo
rm
a
ti
o
n
.
T
h
is
stu
d
y
,
wh
ich
is
a
c
o
m
p
lex
a
n
d
i
n
-
d
e
p
th
e
x
p
lo
ra
ti
o
n
,
a
ims
t
o
d
e
tec
t
An
d
r
o
i
d
m
a
lwa
re
u
sin
g
a
m
e
ta
-
c
las
sifier
th
a
t
c
o
m
b
in
e
s
t
h
e
sin
g
le
-
c
las
sifier
li
g
h
t
g
ra
d
ien
t
b
o
o
sti
n
g
m
a
c
h
in
e
,
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
i
n
e
,
a
n
d
ra
n
d
o
m
f
o
re
st.
T
h
e
p
ro
c
e
ss
in
v
o
lv
e
s
c
o
n
v
e
rti
n
g
d
isa
ss
e
m
b
led
m
a
lwa
re
c
o
d
e
s
in
to
g
r
e
y
ima
g
e
s
fo
r
g
l
o
b
a
l
a
n
d
l
o
c
a
l
fe
a
tu
re
e
x
tra
c
ti
o
n
.
Th
e
c
las
sifica
ti
o
n
a
c
c
u
ra
c
y
is
9
7
%
a
t
b
e
st
o
n
a
m
a
lwa
re
d
a
tas
e
t
o
f
3
,
9
6
3
sa
m
p
les
.
Th
e
m
a
in
c
o
n
t
rib
u
ti
o
n
o
f
t
h
is
p
a
p
e
r
is
t
o
p
ro
d
u
c
e
a
n
An
d
ro
i
d
APK
m
a
lwa
re
d
e
tec
to
r
m
o
d
e
l
th
a
t
wo
rk
s
b
y
c
o
m
b
in
i
n
g
m
u
lt
i
p
le
m
a
c
h
in
e
lea
rn
i
n
g
a
lg
o
rit
h
m
s
train
e
d
u
sin
g
t
h
e
d
a
tas
e
t
re
su
lt
in
g
fro
m
l
o
c
a
l
a
n
d
g
l
o
b
a
l
fe
a
tu
re
e
x
trac
ti
o
n
a
lg
o
rit
h
m
s.
K
ey
w
o
r
d
s
:
APK
An
d
r
o
id
v
is
u
aliza
tio
n
Glo
b
al
f
ea
tu
r
e
e
x
tr
ac
tio
n
L
o
ca
l f
ea
tu
r
e
e
x
tr
ac
tio
n
Ma
lwar
e
Me
ta
class
if
ier
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
I
m
as Su
k
ae
s
ih
Sit
an
g
g
an
g
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
Facu
lty
o
f
Ma
th
em
atics a
n
d
Natu
r
al
Scien
ce
,
I
PB
Un
iv
er
s
ity
W
in
g
2
0
L
ev
el
5
Me
r
an
ti R
o
a
d
,
Dr
am
ag
a
Dis
tr
ict,
B
o
g
o
r
1
6
6
8
0
,
I
n
d
o
n
esia
E
m
ail: im
as.si
tan
g
g
an
g
@
ap
p
s
.
ip
b
.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
An
d
r
o
id
is
th
e
m
o
s
t
wid
ely
u
s
ed
m
o
b
ile
o
p
e
r
atin
g
s
y
s
tem
to
d
ay
.
B
ased
o
n
in
f
o
r
m
atio
n
f
r
o
m
Stat
C
o
u
n
ter
Glo
b
al
Stats
,
u
n
t
il
J
u
n
e
2
0
2
4
,
An
d
r
o
id
b
ec
am
e
th
e
m
o
b
ile
o
p
er
atin
g
s
y
s
tem
with
th
e
lar
g
est
m
ar
k
et
s
h
ar
e
in
th
e
wo
r
ld
,
w
h
ich
am
o
u
n
te
d
to
7
2
.
1
5
%,
b
e
atin
g
i
OS
at
2
7
.
1
9
%.
T
h
e
o
p
en
-
s
o
u
r
ce
An
d
r
o
id
o
p
er
atin
g
s
y
s
tem
an
d
f
r
ee
licen
s
e
m
ak
e
it m
o
r
e
p
o
p
u
lar
.
T
h
e
h
ig
h
p
o
p
u
lar
ity
o
f
An
d
r
o
i
d
is
p
r
o
p
o
r
tio
n
al
to
th
e
in
cr
ea
s
ed
r
is
k
o
f
v
ir
u
s
an
d
m
alwa
r
e
attac
k
s
.
Fu
r
th
er
m
o
r
e,
th
e
tech
n
o
l
o
g
y
f
o
r
f
r
am
ewo
r
k
-
b
ased
a
p
p
licatio
n
d
ev
elo
p
m
e
n
t
is
ad
v
a
n
cin
g
r
a
p
id
ly
,
o
f
f
er
in
g
a
c
o
n
v
e
n
ien
t
way
to
cr
ea
te
An
d
r
o
i
d
-
b
ased
ap
p
licatio
n
s
.
T
h
o
s
e
wh
o
wis
h
to
tak
e
ad
v
an
tag
e
o
f
th
is
co
n
v
en
ien
ce
o
f
ten
r
ev
er
s
e
-
en
g
in
ee
r
co
m
p
leted
a
p
p
licatio
n
s
an
d
ad
d
p
r
o
g
r
a
m
lo
g
ic
to
cr
ea
te
n
ew
ap
p
licatio
n
s
q
u
ick
ly
.
Ho
wev
e
r
,
th
is
co
n
v
en
ien
ce
also
p
r
ese
n
ts
a
r
is
k
,
as
s
o
m
e
in
d
iv
id
u
als
m
a
y
u
p
lo
ad
m
alic
io
u
s
s
cr
ip
ts
th
at
ca
n
ex
p
lo
it
t
h
e
d
ata
o
n
An
d
r
o
id
h
an
d
s
ets.
No
wad
ay
s
,
u
s
er
s
f
r
eq
u
e
n
tly
d
o
wn
lo
ad
ap
p
s
f
r
o
m
an
y
lo
ca
tio
n
,
lead
i
n
g
to
is
s
u
es.
I
n
s
tallin
g
ap
p
s
an
d
p
e
r
m
itti
n
g
th
em
to
d
o
an
y
th
in
g
th
e
y
ask
with
o
u
t
u
n
d
er
s
tan
d
in
g
th
eir
p
u
r
p
o
s
e
is
w
h
at
lead
s
to
s
ec
u
r
ity
is
s
u
es
o
n
s
m
ar
tp
h
o
n
es,
s
u
ch
as
s
cr
ip
t
in
tr
u
s
io
n
.
I
t
ca
n
p
e
r
f
o
r
m
a
s
ec
u
r
ity
-
v
io
latin
g
o
p
er
atio
n
,
s
u
ch
as
s
tar
tin
g
a
c
o
v
er
t
d
ata
tr
an
s
f
er
,
allo
win
g
attac
k
er
s
to
o
b
tain
p
r
iv
ate
in
f
o
r
m
atio
n
f
r
o
m
d
ev
ic
es.
At
s
o
m
e
p
o
in
t,
attac
k
er
s
will
tak
e
co
n
tr
o
l
o
f
th
e
s
m
ar
tp
h
o
n
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
d
en
tifi
ca
tio
n
o
f A
n
d
r
o
id
A
P
K
ma
lw
a
r
e
th
r
o
u
g
h
lo
ca
l a
n
d
g
lo
b
a
l fe
a
t
u
r
e
…
(
Yo
g
a
Hera
w
a
n
)
1835
So
f
twar
e
th
at
s
ee
k
s
to
o
b
s
tr
u
ct
r
eg
u
lar
co
m
p
u
ter
ize
d
s
y
s
tem
o
p
er
atio
n
s
b
y
o
b
tain
i
n
g
p
r
iv
ate
in
f
o
r
m
atio
n
g
ai
n
in
g
illeg
al
ac
ce
s
s
to
co
m
p
u
ter
s
y
s
tem
s
,
an
d
m
o
s
tly
h
ar
ass
in
g
u
s
er
s
is
k
n
o
wn
as
m
alwa
r
e,
s
h
o
r
t
f
o
r
m
alicio
u
s
s
o
f
twar
e
[
1
]
.
T
h
e
ex
p
o
n
e
n
tial
g
r
o
wth
o
f
m
alwa
r
e
f
o
r
ce
s
co
m
p
u
ter
s
ec
u
r
ity
r
esear
ch
er
s
to
d
ev
elo
p
n
ew
m
eth
o
d
s
f
o
r
p
r
o
tectin
g
n
etwo
r
k
s
an
d
s
y
s
tem
s
.
Ma
lwar
e
id
en
tific
atio
n
is
ess
e
n
tial
b
ec
au
s
e
o
f
its
ex
p
o
n
e
n
tial
g
r
o
wth
an
d
ab
ilit
y
to
in
f
lict
ex
ten
s
iv
e
d
am
ag
e
o
n
a
wid
e
s
u
r
f
ac
e
ar
ea
[
2
]
.
R
esear
ch
er
s
h
av
e
cr
ea
ted
m
eth
o
d
s
f
o
r
d
etec
tin
g
m
alwa
r
e
as
a
r
esu
lt
o
f
th
e
g
r
o
wth
o
f
m
alwa
r
e
[
3
]
.
T
h
er
e
ar
e
th
r
ee
m
alwa
r
e
d
etec
tio
n
ap
p
r
o
a
ch
es:
s
tatic,
d
y
n
am
ic,
an
d
h
y
b
r
id
.
Static
an
aly
s
is
ex
am
in
es
th
e
al
l
eg
ed
co
d
e
with
o
u
t
lau
n
ch
in
g
th
e
p
r
o
g
r
am
[
4
]
.
Fo
r
f
ea
tu
r
e
ex
tr
ac
tio
n
,
th
e
s
o
u
r
ce
co
d
e
m
u
s
t
b
e
d
is
ass
e
m
b
led
b
y
a
n
aly
zin
g
p
er
m
is
s
io
n
u
s
ag
e
[
5
]
,
m
in
in
g
th
e
co
d
e
s
tr
u
ctu
r
es
[
6
]
,
an
al
y
z
in
g
th
e
u
s
ed
co
m
p
o
n
en
ts
[
7
]
,
an
d
m
o
n
ito
r
in
g
t
h
e
ap
p
licatio
n
p
r
o
g
r
am
m
in
g
i
n
ter
f
ac
e
(
API
)
in
v
o
k
e
d
[
7
]
,
[
8
]
.
I
t
is
a
g
atew
ay
with
in
th
e
An
d
r
o
id
a
r
ch
itectu
r
e
to
ac
ce
s
s
s
en
s
itiv
e
s
er
v
ices
[
9
]
.
Un
lik
e
s
tatic
an
aly
s
is
,
d
y
n
am
ic
an
aly
s
is
ex
am
in
es
th
e
a
p
p
l
icatio
n
'
s
attr
ib
u
tes
an
d
s
cr
u
tin
izes th
e
ev
e
n
ts
th
at
tr
an
s
p
ir
e
d
u
r
in
g
its
ex
ec
u
tio
n
[
4
]
,
[
1
0
]
,
[
1
1
]
.
Har
m
f
u
l
ap
p
s
m
u
s
t
b
e
s
tr
ip
p
e
d
o
f
th
eir
i
n
tr
icate
f
ea
tu
r
es
an
d
h
id
d
e
n
s
tr
u
ctu
r
es
to
id
e
n
tif
y
m
alwa
r
e
u
s
in
g
n
ew,
ef
f
ec
tiv
e
d
etec
tio
n
tech
n
i
q
u
es
[
1
2
]
.
T
h
e
ad
v
a
n
ce
m
en
t
o
f
th
e
in
ter
n
et
an
d
te
ch
n
o
lo
g
y
will
m
a
k
e
th
e
p
o
te
n
tial
th
r
ea
t
o
f
An
d
r
o
id
APK
m
alwa
r
e
e
v
en
m
o
r
e
s
ig
n
if
ican
t,
m
ak
in
g
it
n
ec
ess
ar
y
to
d
ev
elo
p
a
m
ac
h
in
e
lear
n
in
g
-
b
ased
m
alw
ar
e
d
etec
tio
n
s
y
s
tem
.
T
h
e
d
i
v
er
s
e
an
ato
m
ical
s
tr
u
ctu
r
e
o
f
An
d
r
o
id
APK
f
iles
m
ak
es
th
e
co
m
b
in
atio
n
o
f
lo
ca
l
an
d
g
lo
b
al
f
ea
tu
r
e
ex
tr
ac
tio
n
tech
n
iq
u
es
an
o
p
p
o
r
tu
n
it
y
to
h
elp
p
r
o
d
u
ce
ac
cu
r
ate
an
d
f
ast
m
o
d
els.
Ma
l
war
e
class
if
icatio
n
u
s
in
g
m
ac
h
in
e
lear
n
in
g
(
ML
)
is
o
n
e
o
f
t
h
e
tech
n
o
lo
g
ies
th
at
ar
tific
ial
in
tellig
en
ce
(
AI
)
h
as
m
ad
e
av
ailab
le
f
o
r
u
s
e
in
m
alwa
r
e
r
esear
ch
.
Data
m
in
in
g
w
as
f
ir
s
t
p
r
o
p
o
s
ed
in
2
0
0
1
to
id
en
tif
y
m
alwa
r
e
o
n
W
in
d
o
ws
o
p
er
atin
g
s
y
s
tem
s
[
1
3
]
.
T
h
e
r
esear
ch
em
p
l
o
y
ed
R
I
PP
E
R
,
n
aïv
e
B
ay
es,
an
d
m
u
lti
-
class
if
ier
s
y
s
tem
as
a
m
ac
h
in
e
lear
n
in
g
m
eth
o
d
.
T
h
e
m
u
lti
-
n
aï
v
e
B
ay
es
ap
p
r
o
ac
h
y
ield
s
th
e
b
est
ac
cu
r
ac
y
o
f
9
7
.
7
6
% in
test
in
g
.
I
n
p
lace
o
f
s
tatic
an
d
d
y
n
am
i
c
ap
p
r
o
ac
h
es,
r
esear
ch
e
r
s
h
av
e
s
tu
d
ied
m
alwa
r
e
u
s
in
g
v
is
u
aliza
tio
n
-
b
ased
tech
n
iq
u
es.
L
ar
g
e
-
s
ca
le
m
alwa
r
e
d
etec
tio
n
ca
n
b
e
ac
h
iev
ed
m
o
r
e
e
f
f
ec
tiv
ely
b
y
v
i
s
u
al
an
aly
s
is
,
wh
ich
u
s
es
th
e
s
tr
u
ctu
r
e
o
f
m
alw
ar
e
im
ag
es
r
ath
er
th
an
s
tati
c
o
r
d
y
n
am
ic
a
p
p
r
o
ac
h
es.
C
o
m
p
u
ter
m
alwa
r
e
class
if
icatio
n
m
eth
o
d
b
y
v
is
u
alizin
g
b
in
ar
y
m
alwa
r
e
in
g
r
ay
s
ca
le
im
ag
es
[
1
4
]
.
T
h
is
ap
p
r
o
ac
h
d
if
f
er
s
f
r
o
m
m
alwa
r
e
ex
p
e
r
ts
'
wid
esp
r
ea
d
s
tatic
an
d
d
y
n
a
m
ic
d
etec
tio
n
a
p
p
r
o
ac
h
es.
T
h
ese
tech
n
iq
u
es
a
r
e
b
ein
g
in
tr
o
d
u
ce
d
to
p
r
o
v
id
e
a
b
etter
alter
n
ativ
e
to
th
e
d
r
awb
ac
k
s
o
f
t
h
e
s
tatic
an
d
d
y
n
a
m
ic
tech
n
iq
u
es.
Usi
n
g
th
e
k
-
n
ea
r
es
t
n
eig
h
b
o
r
(
KNN)
tech
n
iq
u
e,
9
4
5
8
s
am
p
les we
r
e
em
p
lo
y
ed
in
h
is
r
esear
ch
,
y
ield
in
g
an
ac
c
u
r
ac
y
r
ate
o
f
9
8
%.
T
h
e
ex
ec
u
tab
le
f
ile
i
n
m
alwa
r
e
was
co
n
v
er
ted
i
n
to
a
g
r
a
y
s
ca
le
im
ag
e
f
o
r
f
ea
tu
r
e
ex
tr
ac
t
io
n
u
s
in
g
wav
elet
tr
an
s
f
o
r
m
an
d
Gab
o
r
f
ilter
[
1
5
]
.
T
h
e
m
ac
h
in
e
lea
r
n
in
g
al
g
o
r
ith
m
u
s
ed
is
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e
(
SVM
)
,
wh
ich
h
as
an
ac
cu
r
ac
y
o
f
9
5
%.
A
lig
h
tweig
h
t
m
alwa
r
e
d
etec
tio
n
m
eth
o
d
th
at
m
ak
es
it
p
o
s
s
ib
le
to
d
etec
t
m
alwa
r
e
f
r
o
m
s
m
ar
tp
h
o
n
es
[
7
]
.
I
n
th
eir
r
esear
ch
,
ea
ch
f
ea
tu
r
e
is
tak
en
f
r
o
m
th
e
co
m
p
o
n
en
ts
o
f
th
e
.
f
ile
an
d
th
e
.
f
ile
is
co
llected
i
n
to
a
s
p
ac
e
v
ec
to
r
f
o
r
a
n
SVM
ap
p
r
o
ac
h
with
9
4
%
ac
cu
r
ac
y
.
A
n
d
r
o
i
d
n
etwo
r
k
f
lo
ws
an
d
API
ca
lls
wer
e
test
ed
f
o
r
m
alwa
r
e
d
etec
tio
n
with
An
d
r
o
i
d
p
er
m
is
s
io
n
s
an
d
in
ten
t f
ea
tu
r
e
s
[
1
6
]
.
I
n
ad
d
itio
n
to
p
r
o
v
id
in
g
ac
cu
r
ac
y
r
esu
lts
o
f
9
5
.
3
% f
o
r
s
tatic
an
aly
s
is
an
d
8
3
.
3
%
f
o
r
d
y
n
am
ic
an
al
y
s
is
o
n
m
alwa
r
e
d
etec
tio
n
o
n
th
e
A
n
d
r
o
id
s
y
s
tem
,
th
e
s
tu
d
y
also
o
f
f
er
ed
th
e
p
u
b
licly
ac
ce
s
s
ib
le
C
I
C
An
d
Ma
l2
0
1
7
d
ataset
wh
ile
p
er
f
o
r
m
in
g
d
y
n
am
ic
an
aly
s
is
o
n
ac
tu
al
s
m
ar
tp
h
o
n
es.
R
eg
ar
d
in
g
m
alwa
r
e
d
etec
tio
n
u
s
in
g
APK
im
ag
e
v
is
u
aliza
tio
n
,
th
e
r
an
d
o
m
f
o
r
est
(
R
F)
alg
o
r
ith
m
p
r
o
v
id
es
th
e
b
est
p
er
f
o
r
m
an
ce
with
a
n
ac
cu
r
ac
y
o
f
9
2
.
8
1
%
[
1
7
]
.
A
m
alwa
r
e
d
etec
tio
n
f
r
am
ewo
r
k
b
ased
o
n
a
s
tack
o
f
en
s
em
b
le
lear
n
in
g
alg
o
r
ith
m
s
ca
lled
MFDr
o
id
[
1
8
]
.
T
h
e
s
tu
d
y
u
s
es
s
ev
en
f
ea
tu
r
e
ex
tr
ac
tio
n
al
g
o
r
ith
m
s
b
ased
o
n
Per
m
is
s
io
n
,
API
C
all
s
,
an
d
Op
co
d
e
d
atasets
an
d
th
e
n
p
r
o
d
u
c
es a
n
F1
Sco
r
e
o
f
9
6
% u
s
in
g
a
m
eta
-
class
if
ier
.
A
m
ac
h
in
e
-
lear
n
in
g
a
p
p
r
o
ac
h
b
ased
o
n
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
g
r
ay
s
ca
le
p
ictu
r
e
ca
teg
o
r
izatio
n
is
u
s
ed
to
id
e
n
tify
A
n
d
r
o
i
d
A
PK
m
alwa
r
e
[
1
9
]
.
T
h
is
s
tu
d
y
tu
r
n
s
m
o
s
t
APK
f
iles
—
in
clu
d
in
g
m
u
ltiDex
,
r
eso
u
r
ce
s
,
ce
r
tific
ates,
an
d
m
an
if
est
f
iles
—
in
to
g
r
ay
s
ca
le
p
ictu
r
es
to
d
etec
t
m
alwa
r
e.
Su
b
s
eq
u
en
tly
,
lo
ca
l
m
alwa
r
e
p
ictu
r
e
f
ea
tu
r
es
ar
e
ex
tr
ac
ted
u
s
in
g
lo
ca
l
f
ea
tu
r
e
ex
tr
ac
tio
n
tec
h
n
iq
u
es.
T
h
e
ac
cu
r
ac
y
o
f
th
e
m
o
d
el
y
ield
s
o
u
tco
m
es
th
at
a
r
e
9
6
.
8
6
%.
C
o
n
v
en
tio
n
al
a
n
d
e
n
s
em
b
le
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
ar
e
co
m
p
a
r
ed
t
o
d
eter
m
in
e
t
h
e
ac
c
u
r
ac
y
o
f
t
h
e
m
atr
ices
c
o
m
p
ar
is
o
n
[
2
0
]
.
T
h
e
Dr
eb
in
d
ataset,
wh
ic
h
h
ad
2
1
5
s
tatically
ev
alu
ated
f
ea
tu
r
es,
was
em
p
lo
y
ed
in
h
is
r
esear
ch
.
T
h
e
in
p
u
t
d
ata
s
h
o
ws
th
e
attr
ib
u
te'
s
av
ailab
ilit
y
in
th
e
d
ataset
an
d
is
ac
ce
s
s
ib
le
in
0
an
d
1
f
o
r
m
ats.
Acc
o
r
d
in
g
to
t
h
e
in
v
esti
g
atio
n
,
th
e
en
s
em
b
l
e
m
eth
o
d
p
er
f
o
r
m
s
b
etter
at
p
r
ed
ictin
g
m
alwa
r
e
th
an
co
n
v
e
n
tio
n
al
m
ac
h
in
e
le
ar
n
in
g
m
eth
o
d
s
.
R
an
d
o
m
f
o
r
est
y
ield
ed
9
9
.
1
%
ac
cu
r
ac
y
,
wh
e
r
ea
s
lig
h
t g
r
a
d
ie
n
t b
o
o
s
tin
g
m
ac
h
in
e
(
L
GB
M)
p
r
o
d
u
ce
d
9
9
.
5
%.
N
u
m
e
r
o
u
s
r
e
s
e
a
r
c
h
w
o
r
k
s
o
n
A
n
d
r
o
i
d
m
a
l
w
a
r
e
d
e
t
e
c
t
i
o
n
A
P
Ks
h
a
v
e
b
e
e
n
p
u
b
l
i
s
h
e
d
i
n
a
c
a
d
e
m
i
c
j
o
u
r
n
a
l
s
.
E
a
r
li
e
r
s
t
u
d
i
e
s
e
m
p
l
o
y
e
d
d
i
f
f
e
r
e
n
t
A
P
K
c
o
m
p
o
n
e
n
t
s
a
s
d
at
a
s
e
t
s
a
n
d
t
h
e
n
a
n
a
l
y
z
ed
t
h
e
m
s
t
a
ti
c
a
ll
y
o
r
d
y
n
a
m
i
c
a
l
l
y
u
s
i
n
g
m
a
c
h
i
n
e
-
l
ea
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
.
T
h
e
p
r
i
m
a
r
y
m
o
t
i
v
a
t
i
o
n
o
f
t
h
i
s
s
t
u
d
y
i
s
t
o
k
n
o
w
t
h
e
i
m
p
a
c
t
o
f
c
o
m
b
i
n
i
n
g
d
i
f
f
e
r
e
n
t
m
a
c
h
i
n
e
-
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
es
o
n
t
h
e
o
u
t
c
o
m
e
s
o
f
l
o
ca
l
a
n
d
g
l
o
b
al
e
x
t
r
a
c
t
i
o
n
i
n
t
h
e
v
i
s
u
a
l
iz
a
t
i
o
n
o
f
A
n
d
r
o
i
d
A
PK
s
.
T
h
is
s
t
u
d
y
d
e
t
e
ct
s
m
a
l
wa
r
e
i
n
A
n
d
r
o
i
d
AP
K
f
il
es
u
s
in
g
a
m
e
t
a
-
c
l
ass
i
f
i
er
t
e
c
h
n
i
q
u
e
c
o
m
b
i
n
i
n
g
l
o
c
a
l
an
d
g
l
o
b
a
l
d
a
t
a
s
e
t
f
e
a
t
u
r
es
,
i
n
t
e
g
r
a
t
i
n
g
t
h
e
r
a
n
d
o
m
f
o
r
e
s
t
,
S
V
M
,
a
n
d
L
GB
M
m
e
t
h
o
d
s
.
T
h
e
r
e
m
a
i
n
i
n
g
s
e
ct
i
o
n
s
o
f
t
h
e
p
a
p
e
r
a
r
e
s
t
r
u
c
t
u
r
e
d
a
s
f
o
l
l
o
ws
.
T
h
e
m
et
h
o
d
o
l
o
g
y
u
s
e
d
i
s
e
x
p
l
a
i
n
e
d
i
n
s
e
c
ti
o
n
2
.
T
h
e
a
n
a
l
y
s
is
r
es
u
l
ts
a
n
d
d
i
s
c
u
s
s
i
o
n
s
a
r
e
p
r
e
s
e
n
te
d
in
s
e
c
t
i
o
n
3
,
a
n
d
t
h
e
c
o
n
c
l
u
s
i
o
n
i
n
s
e
c
ti
o
n
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
8
3
4
-
1
8
4
9
1836
2.
T
H
E
P
RO
P
O
SE
D
M
E
T
H
O
D
T
h
e
m
a
i
n
o
b
j
e
c
t
i
v
e
s
o
f
t
h
is
s
t
u
d
y
a
r
e
t
o
a
s
s
e
s
s
a
n
d
a
n
a
l
y
z
e
r
aw
d
a
t
a
A
PK
a
s
l
o
c
al
a
n
d
g
l
o
b
a
l
f
e
a
t
u
r
e
s
t
o
i
m
p
r
o
v
e
t
h
e
d
et
e
c
ti
o
n
a
c
cu
r
a
c
y
a
n
d
p
e
r
f
o
r
m
a
n
c
e
o
f
m
e
t
a
-
c
l
as
s
i
f
i
e
r
s
.
T
h
e
s
t
u
d
y
w
as
c
o
n
d
u
c
t
e
d
i
n
t
h
r
e
e
p
r
i
m
a
r
y
p
h
a
s
e
s
,
e
a
c
h
m
e
ti
c
u
lo
u
s
l
y
d
e
s
i
g
n
e
d
a
n
d
e
x
ec
u
t
e
d
:
d
a
t
a
p
r
e
p
r
o
c
e
s
s
i
n
g
,
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
,
a
n
d
m
o
d
e
l
e
v
a
l
u
a
t
i
o
n
.
T
w
o
m
e
t
h
o
d
s
f
o
r
ex
t
r
a
c
t
i
n
g
f
e
at
u
r
e
s
a
r
e
a
v
ai
l
a
b
l
e
d
u
r
i
n
g
t
h
e
p
r
e
p
r
o
c
e
s
s
i
n
g
s
t
a
g
e
:
l
o
c
al
a
n
d
g
l
o
b
a
l.
T
h
e
l
e
a
r
n
i
n
g
p
r
o
c
e
s
s
e
m
p
l
o
y
s
a
s
i
n
g
l
e
-
c
l
as
s
i
f
i
e
r
o
r
m
e
t
a
-
c
la
s
s
i
f
i
e
r
a
t
t
h
e
m
a
c
h
i
n
e
-
le
a
r
n
i
n
g
s
t
e
p
.
I
n
o
r
d
e
r
t
o
c
o
n
s
t
r
u
c
t
a
m
e
t
a
-
c
l
ass
i
f
i
e
r
m
o
d
e
l
,
t
h
e
s
i
n
g
l
e
-
c
l
ass
i
f
i
e
r
m
o
d
e
l'
s
d
e
t
e
ct
i
o
n
r
e
s
u
l
ts
w
i
ll
s
e
r
v
e
a
s
a
d
a
t
a
s
o
u
r
c
e
.
F
i
g
u
r
e
1
is
a
b
i
g
p
i
ct
u
r
e
o
f
c
r
e
a
t
i
n
g
a
s
y
s
te
m
t
o
i
d
e
n
ti
f
y
w
h
e
t
h
e
r
A
n
d
r
o
i
d
A
PK
f
i
le
s
a
r
e
m
a
l
w
a
r
e
o
r
b
e
n
i
g
n
.
Mo
d
el
p
e
r
f
o
r
m
an
ce
m
ea
s
u
r
e
m
en
ts
ar
e
m
ea
s
u
r
ed
b
ased
o
n
f
o
u
r
p
r
im
ar
y
cr
iter
ia:
ac
c
u
r
a
cy
,
F1
s
co
r
e
,
r
ec
all,
an
d
p
r
ec
is
io
n
.
T
h
e
tim
e
c
o
m
p
ar
is
o
n
p
r
o
ce
d
u
r
e
(
m
ea
s
u
r
ed
in
s
ec
o
n
d
s
)
an
d
th
e
am
o
u
n
t
o
f
r
an
d
o
m
-
ac
ce
s
s
m
em
o
r
y
(
R
AM
)
u
s
ed
d
u
r
in
g
t
h
e
test
in
g
p
r
o
ce
s
s
ar
e
two
ad
d
itio
n
al
co
m
p
lem
en
tin
g
cr
iter
ia
in
ad
d
itio
n
to
th
e
f
o
u
r
p
r
im
ar
y
m
ea
s
u
r
em
e
n
t c
r
iter
ia
f
o
r
th
e
m
o
d
el
p
er
f
o
r
m
an
c
e
an
aly
s
is
.
Fig
u
r
e
1
.
Step
s
o
f
cr
ea
tin
g
m
a
lwar
e
d
etec
tio
n
s
y
s
tem
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
d
en
tifi
ca
tio
n
o
f A
n
d
r
o
id
A
P
K
ma
lw
a
r
e
th
r
o
u
g
h
lo
ca
l a
n
d
g
lo
b
a
l fe
a
t
u
r
e
…
(
Yo
g
a
Hera
w
a
n
)
1837
3.
M
E
T
H
O
D
3
.
1
.
Resea
rc
h
s
ce
na
rio
A
s
er
ies
o
f
e
x
p
er
im
e
n
ts
n
ee
d
to
b
e
ca
r
r
ied
o
u
t
t
o
ac
h
ie
v
e
th
e
m
ain
o
b
jectiv
es
o
f
t
h
is
r
e
s
ea
r
ch
.
I
n
ad
d
itio
n
,
it
ca
n
h
elp
o
b
tain
a
m
o
d
el
with
th
e
m
o
s
t
o
p
tim
al
p
er
f
o
r
m
an
ce
.
T
ab
le
1
d
escr
ib
es
th
e
ex
p
er
im
en
tal
s
ettin
g
s
th
at
wer
e
co
n
d
u
cted
in
th
is
r
esear
ch
.
T
ab
le
1
.
E
x
p
er
im
en
tal
s
ettin
g
s
S
c
e
n
a
r
i
o
Ex
p
l
a
n
a
t
i
o
n
F
i
r
st
e
x
p
e
r
i
me
n
t
D
e
f
a
u
l
t
e
x
p
e
r
i
me
n
t
u
t
i
l
i
z
i
n
g
a
s
-
i
s
d
a
t
a
set
s
S
e
c
o
n
d
e
x
p
e
r
i
me
n
t
Th
e
F
i
r
st
e
x
p
e
r
i
m
e
n
t
u
si
n
g
t
h
e
b
a
l
a
n
c
e
d
d
a
t
a
s
e
t
Th
i
r
d
e
x
p
e
r
i
me
n
t
S
e
c
o
n
d
e
x
p
e
r
i
me
n
t
w
i
t
h
g
r
i
d
s
e
a
r
c
h
h
y
p
e
r
p
a
r
a
me
t
e
r
c
o
n
f
i
g
u
r
a
t
i
o
n
a
p
p
l
i
e
d
t
o
t
h
e
b
a
se
-
c
l
a
ssi
f
i
e
r
a
l
g
o
r
i
t
h
m
F
o
u
r
t
h
e
x
p
e
r
i
me
n
t
S
e
c
o
n
d
e
x
p
e
r
i
me
n
t
w
i
t
h
r
a
n
d
o
m
se
a
r
c
h
h
y
p
e
r
p
a
r
a
met
e
r
c
o
n
f
i
g
u
r
a
t
i
o
n
a
p
p
l
i
e
d
t
o
t
h
e
b
a
se
-
c
l
a
ss
i
f
i
e
r
a
l
g
o
r
i
t
h
m
3
.
2
.
Da
t
a
prepro
ce
s
s
ing
T
h
is
r
esear
ch
u
s
es
o
p
en
d
ata
s
ets
th
at
ca
n
b
e
f
r
ee
ly
d
o
wn
l
o
ad
ed
[
1
6
]
.
T
h
e
d
ataset
is
d
o
wn
lo
ad
ed
u
s
in
g
th
e
f
o
llo
win
g
lin
k
:
h
ttp
:
//2
0
5
.
1
7
4
.
1
6
5
.
8
0
/C
I
C
Da
ta
s
et/C
I
C
I
n
ve
s
A
n
d
Ma
l2
0
1
9
/
Da
ta
s
et/A
P
K
s
/
an
d
h
ttp
:
//2
0
5
.
1
7
4
.
1
6
5
.
8
0
/C
I
C
Da
ta
s
et/Ma
lDr
o
id
-
2
0
2
0
/Da
ta
s
et/A
P
K
s
/
.
T
h
e
1
5
.
9
GB
d
ataset
u
s
ed
in
th
is
r
esear
ch
in
clu
d
es
3
,
9
6
3
APKs
class
if
i
ed
as
b
en
ig
n
an
d
m
alicio
u
s
in
Z
I
P
Fo
r
m
at.
T
h
is
an
al
y
s
is
o
n
ly
h
a
n
d
led
th
e
f
o
llo
win
g
ca
te
g
o
r
ies
o
f
m
alw
ar
e:
ad
war
e,
r
an
s
o
m
war
e
,
SM
S
m
alwa
r
e
,
an
d
s
ca
r
ewa
r
e
.
T
h
e
co
llected
d
ata
is
th
en
p
r
ep
r
o
ce
s
s
ed
to
p
r
o
d
u
ce
th
e
.
f
ile
ty
p
e.
Sin
ce
all
J
av
a
ex
ec
u
tab
le
co
d
e
is
s
to
r
ed
in
f
iles
with
th
e
ex
ten
s
io
n
“
.
,
”
th
e
DE
X
f
ile
s
h
a
r
es
m
an
y
c
h
ar
ac
ter
is
tics
with
an
An
d
r
o
id
a
p
p
licatio
n
[
2
1
]
.
F
ig
u
r
e
2
s
h
o
ws
th
e
p
r
o
ce
d
u
r
e
f
o
r
c
r
ea
tin
g
g
r
a
y
s
ca
le
im
ag
es f
r
o
m
zip
d
ata
s
o
u
r
ce
s
.
Fig
u
r
e
2
.
Pro
ce
d
u
r
e
f
o
r
cr
ea
ti
n
g
g
r
a
y
s
ca
le
im
ag
es f
r
o
m
zip
d
ata
s
o
u
r
ce
s
T
h
e
d
ataset
d
is
tr
ib
u
tio
n
is
s
p
lit in
to
tr
ain
in
g
,
test
in
g
,
a
n
d
v
alid
atio
n
,
with
th
e
p
e
r
ce
n
tag
es b
ein
g
7
2
%,
2
0
%,
an
d
8
%,
r
esp
ec
tiv
ely
.
T
h
e
b
ase
-
class
if
ier
m
o
d
el
is
co
n
s
tr
u
cted
u
s
in
g
th
e
tr
ain
in
g
d
ataset.
T
h
e
test
in
g
d
ataset,
wh
ich
ac
c
o
u
n
ts
f
o
r
2
0
%
o
f
th
e
d
is
tr
ib
u
tio
n
,
is
c
r
u
cial
in
tr
ain
in
g
th
e
m
eta
-
cl
ass
if
ier
m
o
d
el.
T
h
e
b
ase
-
class
if
ier
m
o
d
el
d
eter
m
in
es
th
is
m
o
d
el'
s
ac
cu
r
ac
y
.
T
h
e
v
alid
atio
n
d
ataset
is
th
en
u
s
ed
to
ev
alu
ate
th
e
m
eta
-
class
if
ier
m
o
d
el
an
d
ass
ess
its
ac
cu
r
ac
y
.
3
.
3
.
AP
K
v
is
ua
liza
t
io
n
T
h
e
h
ea
d
er
s
ec
tio
n
,
id
s
s
ec
tio
n
s
,
_
s
ec
tio
n
s
,
an
d
d
ata
s
ec
tio
n
s
ar
e
am
o
n
g
t
h
e
s
ec
tio
n
s
th
at
m
ak
e
u
p
a
DE
X
f
ile.
T
h
e
h
ea
d
er
s
ec
tio
n
co
n
tain
s
to
p
-
lev
e
l
d
ata,
s
u
ch
as
th
e
s
u
b
s
eq
u
en
t
s
ec
tio
n
s
'
f
ile
s
ize,
s
ig
n
atu
r
e,
o
f
f
s
et,
an
d
s
ize.
T
h
e
s
y
s
tem
m
u
s
t
g
ath
er
th
e
f
ile
s
ize
to
d
eter
m
in
e
th
e
im
ag
e'
s
w
id
th
b
ef
o
r
e
co
n
v
er
tin
g
th
e
d
ata
in
to
im
a
g
es,
u
s
in
g
th
e
r
ef
er
e
n
ce
s
in
T
ab
le
2
.
T
h
e
d
ata
s
ec
tio
n
is
a
cr
u
cial
p
ar
t
o
f
.
co
n
tain
s
class
d
ata
a
n
d
ex
ec
u
tio
n
c
o
d
e,
r
ep
r
esen
tin
g
th
e
ap
p
licatio
n
'
s
b
eh
av
io
r
,
wh
eth
er
it
is
class
if
ied
as
m
alwa
r
e
o
r
b
en
ig
n
ap
p
licatio
n
[
2
2
]
.
T
h
e
d
esire
d
in
f
o
r
m
atio
n
is
k
ep
t
in
th
e
_
ℎ
o
f
th
e
.
f
ile
m
ay
b
e
r
etr
iev
ed
b
y
p
ar
s
in
g
th
e
_
ℎ
to
o
b
tain
th
e
_
an
d
_
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
8
3
4
-
1
8
4
9
1838
T
ab
le
2
.
I
m
ag
e
wid
th
r
ec
o
m
m
en
d
atio
n
[
1
4
]
F
i
l
e
si
z
e
r
a
n
g
e
I
mag
e
’
s Wi
d
t
h
<
1
0
k
B
32
1
0
k
B
–
3
0
k
B
64
3
0
k
B
–
6
0
k
B
1
2
8
6
0
k
B
–
1
0
0
k
B
2
5
6
1
0
0
k
B
–
2
0
0
k
B
3
8
4
2
0
0
k
B
–
5
0
0
k
B
5
1
2
5
0
0
k
B
–
1
0
0
0
k
B
7
6
8
>
1
0
0
0
k
B
1
0
2
4
T
h
e
d
ata
s
ize
an
d
o
f
f
s
ets
m
u
s
t
b
e
d
eter
m
in
ed
b
ef
o
r
e
s
ec
tio
n
d
ata
ca
n
b
e
ex
tr
ac
te
d
f
r
o
m
t
h
e
.
f
ile,
ar
e
lo
ca
ted
at
s
p
ec
if
ic
o
f
f
s
ets.
T
h
is
in
f
o
r
m
atio
n
is
o
b
tain
ed
f
r
o
m
th
e
.
f
ile
h
ea
d
er
.
T
h
e
h
ea
d
er
s
ize
o
f
th
e
.
f
ile
is
1
1
2
b
y
tes,
an
d
th
e
d
at
a
s
ec
tio
n
's
s
ize
an
d
th
e
d
ata
s
e
ctio
n
'
s
o
f
f
s
et
v
alu
e
ar
e
r
esp
ec
tiv
ely
l
o
ca
ted
at
o
f
f
s
ets
1
0
4
an
d
1
0
8
,
as
m
an
y
as
4
b
y
tes.
Fig
u
r
e
3
p
r
o
v
id
es
a
v
is
u
al
g
u
id
e
f
o
r
d
i
g
itally
v
is
u
alizin
g
a
r
aw
APK
f
ile
.
Fig
u
r
e
1
.
Pro
ce
d
u
r
es f
o
r
tu
r
n
i
n
g
an
APK
f
ile
in
to
a
d
ig
ital p
ictu
r
e
3
.
4
.
E
x
t
r
a
ct
io
n
f
ea
t
ure
Featu
r
e
ex
tr
ac
tio
n
is
ap
p
lied
to
th
e
.
f
ile
co
n
v
er
ted
in
to
a
d
ig
ital
im
ag
e
to
g
et
th
e
r
eq
u
ir
ed
in
f
o
r
m
atio
n
.
T
h
is
r
esear
ch
u
s
es
two
f
ea
tu
r
e
ex
tr
ac
tio
n
m
eth
o
d
s
:
lo
ca
l
an
d
g
lo
b
al.
L
o
ca
l
f
ea
tu
r
es
ch
ar
ac
ter
ize
a
n
ar
r
o
w
p
o
r
tio
n
o
f
th
e
im
ag
e
,
wh
er
ea
s
g
lo
b
al
f
ea
tu
r
es
d
escr
ib
e
th
e
en
tire
p
ictu
r
e
[
2
3
]
.
M
ea
n
wh
ile,
th
e
lo
ca
l
f
ea
tu
r
e
m
et
h
o
d
r
ep
r
esen
ts
an
i
m
ag
e
b
ased
o
n
l
y
o
n
a
f
ew
p
r
o
m
in
en
t
ar
ea
s
,
wh
ich
a
r
e
c
o
n
s
tan
tly
u
n
af
f
ec
ted
b
y
v
iewp
o
in
t o
r
ch
an
g
es in
illu
s
io
n
[
2
4
]
.
3
.
4
.
1
.
G
lo
ba
l
f
e
a
t
ure
ex
t
r
a
ct
i
o
n
T
h
e
g
lo
b
al
f
ea
tu
r
es
p
r
o
v
id
e
an
o
v
er
all
d
escr
ip
tio
n
o
f
th
e
im
ag
e
to
r
ep
r
esen
t
th
e
co
m
p
l
ete
o
b
ject
u
n
iv
er
s
ally
[
2
5
]
.
T
h
e
GI
ST
a
lg
o
r
ith
m
was
u
s
ed
i
n
th
is
r
esear
ch
to
ex
tr
ac
t
th
e
g
lo
b
al
f
e
atu
r
es
o
f
an
APK
v
is
u
aliza
tio
n
.
A
d
e
f
icien
t
d
i
m
en
s
io
n
al
r
ep
r
esen
tatio
n
o
f
a
g
iv
en
im
ag
e
ca
lled
th
e
s
p
a
tial
en
v
elo
p
e
is
th
e
f
o
u
n
d
atio
n
f
o
r
th
e
GI
ST
im
ag
e
d
escr
ip
to
r
c
r
ea
ted
b
y
T
o
r
r
al
b
a
an
d
Oliv
a
[
2
6
]
.
T
h
e
p
r
im
a
r
y
s
p
atial
s
tr
u
ctu
r
e
o
f
a
s
ce
n
e
ca
n
b
e
r
ep
r
esen
ted
u
s
in
g
a
s
er
ies
o
f
p
e
r
ce
p
t
u
al
d
im
en
s
io
n
s
,
n
am
el
y
n
atu
r
aln
ess
,
o
p
en
n
ess
,
r
o
u
g
h
n
ess
,
ex
p
a
n
s
io
n
,
a
n
d
r
u
g
g
ed
n
ess
.
T
h
e
ef
f
icien
t
esti
m
atio
n
o
f
th
ese
d
im
en
s
io
n
s
was
d
em
o
n
s
tr
ated
u
s
in
g
co
ar
s
ely
lo
ca
lized
an
d
s
p
ec
tr
al
in
f
o
r
m
atio
n
.
T
h
is
led
to
th
e
cr
ea
tio
n
o
f
a
m
u
ltid
im
en
s
io
n
a
l
s
p
ac
e
f
o
r
s
ce
n
es.
Ho
wev
er
,
it
is
ess
en
tial
to
n
o
t
e
th
at
wh
ile
th
e
GI
ST
ch
a
r
ac
te
r
is
tics
o
f
f
er
ed
n
u
m
er
o
u
s
b
en
ef
its
,
th
er
e
wer
e
also
ce
r
tain
d
r
awb
ac
k
s
,
s
u
c
h
as
i
n
f
o
r
m
atio
n
lo
s
s
in
s
p
ar
s
e
g
r
i
d
c
o
m
p
u
tin
g
[
2
7
]
.
T
h
e
am
p
litu
d
e
s
o
f
t
h
e
o
u
tp
u
t
o
f
K
Gab
o
r
f
ilter
s
at
v
ar
io
u
s
s
ca
les
(
S)
an
d
o
r
ien
tatio
n
s
(
O)
ar
e
co
m
b
in
ed
t
o
f
o
r
m
th
e
g
lo
b
al
d
escr
ip
to
r
,
a
s
ig
n
if
ican
t a
s
p
ec
t in
th
e
co
n
tex
t o
f
m
alwa
r
e
im
ag
es.
E
ac
h
f
a
cial
p
ictu
r
e
in
th
e
f
ilter
o
u
tp
u
t
is
r
ed
u
ce
d
to
a
s
ize
N
×
N
b
lo
ck
,
y
ield
in
g
a
v
ec
to
r
with
S
×
O
×
N
×
N
d
im
en
s
io
n
s
to
r
ed
u
ce
th
e
f
ea
tu
r
e
v
ec
to
r
'
s
s
ize
[
2
8
]
.
Fig
u
r
e
4
illu
s
tr
ates th
e
o
p
er
ati
o
n
o
f
GI
ST
o
n
m
alwa
r
e
im
ag
e
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
d
en
tifi
ca
tio
n
o
f A
n
d
r
o
id
A
P
K
ma
lw
a
r
e
th
r
o
u
g
h
lo
ca
l a
n
d
g
lo
b
a
l fe
a
t
u
r
e
…
(
Yo
g
a
Hera
w
a
n
)
1839
Fig
u
r
e
4
.
GI
ST
f
ea
t
u
r
e
ex
t
r
ac
tio
n
in
m
alwa
r
e
im
a
g
e
3
.
4
.
2
.
L
o
ca
l
f
ea
t
ure
ex
t
ra
c
t
io
n
T
h
e
KAZ
E
alg
o
r
ith
m
was
u
s
ed
in
th
is
r
esear
ch
to
ca
r
r
y
o
u
t
lo
ca
l
f
ea
tu
r
e
ex
tr
ac
tio
n
.
T
h
e
im
ag
e
i
s
d
escr
ib
ed
as
b
atch
es
o
f
an
o
b
ject
b
y
th
e
lo
ca
l
f
ea
tu
r
es.
T
h
ese
f
ea
tu
r
es
ar
e
b
ased
o
n
id
en
tify
in
g
n
u
m
er
o
u
s
u
n
iq
u
e
f
o
ca
l p
o
in
ts
with
in
th
e
p
ictu
r
e,
ea
ch
d
escr
ib
e
d
with
a
s
u
itab
le
an
d
r
eliab
le
d
escr
ip
to
r
[
2
5
]
.
I
n
2
0
1
2
,
th
e
non
-
lin
ea
r
s
ca
le
-
s
p
ac
e
KAZ
E
alg
o
r
ith
m
was
p
r
esen
ted
.
T
h
e
d
escr
ip
to
r
in
KAZ
E
is
b
ased
o
n
th
e
l
o
ca
l
d
if
f
er
en
ce
b
in
ar
y
(
L
DB
)
d
esc
r
ip
to
r
,
wh
e
r
ea
s
th
e
d
etec
to
r
i
s
b
ased
o
n
t
h
e
d
eter
m
i
n
an
t
o
f
a
Hess
ian
m
atr
ix
[
2
9
]
.
W
h
ile
lo
wer
in
g
n
o
is
e
in
th
e
im
ag
e,
n
o
n
-
lin
ea
r
d
if
f
u
s
io
n
f
ilter
in
g
p
r
eser
v
es
th
e
b
o
r
d
e
r
s
o
f
th
e
r
eg
io
n
s
at
v
ar
io
u
s
s
izes
[
3
0
]
.
T
h
e
KAZ
E
alg
o
r
ith
m
u
s
es
th
e
p
ar
am
et
er
s
in
T
ab
le
3
to
d
eter
m
in
e
t
h
e
d
escr
ip
to
r
v
alu
es
an
d
cr
u
cial
p
o
in
ts
.
T
ab
le
3
.
KAZ
E
p
a
r
am
eter
s
u
s
ed
in
th
e
r
esear
c
h
P
a
r
a
me
t
e
r
Ex
p
l
a
n
a
t
i
o
n
V
a
l
u
e
s
Ex
t
e
n
d
e
d
To
a
l
l
o
w
e
x
t
e
n
d
e
d
d
e
s
c
r
i
p
t
o
r
e
x
t
r
a
c
t
i
o
n
(
1
2
8
-
b
y
t
e
)
F
a
l
se
U
p
r
i
g
h
t
To
a
l
l
o
w
t
h
e
u
se
o
f
p
e
r
p
e
n
d
i
c
u
l
a
r
(
n
o
n
-
r
o
t
a
t
i
o
n
-
i
n
v
a
r
i
a
n
t
)
d
e
s
c
r
i
p
t
o
r
s.
F
a
l
se
Th
r
e
s
h
o
l
d
Th
e
r
e
c
e
i
v
i
n
g
p
o
i
n
t
t
h
r
e
sh
o
l
d
f
o
r
t
h
e
d
e
t
e
c
t
o
r
r
e
s
p
o
n
se
0
.
0
0
1
n
O
c
t
a
v
e
s
M
a
x
i
m
u
m
o
c
t
a
v
e
e
v
o
l
u
t
i
o
n
o
f
t
h
e
i
m
a
g
e
4
n
O
c
t
a
v
e
L
a
y
e
r
s
D
e
f
a
u
l
t
n
u
m
b
e
r
o
f
s
u
b
l
e
v
e
l
s
p
e
r
s
c
a
l
e
l
e
v
e
l
4
D
i
f
f
u
si
v
i
t
y
Ty
p
e
s
o
f
d
i
f
f
u
si
v
i
t
i
e
s.
D
I
F
F_
P
M
_
G
1
,
D
I
FF_
P
M
_
G
2
,
D
I
FF
_
WE
I
C
K
ERT
o
r
D
I
FF
_
C
H
ARBON
N
I
ER
D
I
FF_
P
M_
G
2
An
im
ag
e'
s
v
is
u
al
in
f
o
r
m
atio
n
ca
n
b
e
f
o
u
n
d
in
th
e
r
etr
iev
ed
lo
ca
l
f
ea
tu
r
es.
A
q
u
a
n
tizatio
n
ap
p
r
o
ac
h
s
u
ch
as
k
-
m
ea
n
s
co
m
p
r
ess
es
th
e
f
ea
tu
r
e
s
p
ac
e
t
o
clu
s
ter
s
f
o
r
a
co
m
p
ac
t
p
ictu
r
e
r
ep
r
esen
tatio
n
.
“
Vis
u
al
wo
r
d
s
”
r
ef
er
to
th
e
clu
s
ter
ce
n
ter
s
[
3
1
]
.
T
h
e
KAZ
E
d
etec
tio
n
p
r
o
ce
s
s
p
r
o
d
u
ce
s
a
v
ec
to
r
t
h
a
t c
o
n
tain
s
a
v
ar
iety
o
f
d
ata.
Ho
we
v
er
,
th
e
class
if
ier
alg
o
r
ith
m
o
n
ly
ac
ce
p
ts
n
d
a
ta
an
d
n
-
f
ea
tu
r
e
v
ec
to
r
s
as
in
p
u
t.
T
h
is
m
ea
n
s
th
at
th
e
o
u
tp
u
t
f
r
o
m
KAZ
E
is
n
o
t
d
ir
ec
tly
co
m
p
atib
le
with
th
e
c
lass
if
ier
alg
o
r
ith
m
.
T
h
e
u
r
g
en
cy
o
f
th
is
p
r
o
b
lem
n
ec
ess
itates
a
s
o
lu
tio
n
,
wh
ich
is
to
ap
p
ly
th
e
b
ag
o
f
v
is
u
al
wo
r
d
s
(
B
OVW)
to
cr
ea
te
a
s
i
n
g
le
f
ea
tu
r
e
v
ec
to
r
f
r
o
m
th
e
m
an
y
d
escr
ip
to
r
s
o
f
l
o
ca
l f
ea
tu
r
es
[
3
2
]
.
Fig
u
r
e
5
ex
p
lain
s
th
e
KAZ
E
f
ea
tu
r
e
ex
tr
ac
tio
n
p
r
o
ce
s
s
.
B
OVW
em
p
lo
y
s
a
clu
s
ter
in
g
a
lg
o
r
ith
m
to
class
if
y
th
e
v
ec
to
r
s
p
r
o
d
u
ce
d
b
y
th
e
d
escr
ip
to
r
in
to
th
e
m
o
s
t
s
u
itab
le
n
-
clu
s
ter
s
.
T
h
e
alg
o
r
ith
m
p
r
e
d
icts
th
e
class
es
o
f
ea
ch
d
escr
ip
to
r
d
at
a,
an
d
an
ess
en
tial
s
tep
in
t
h
e
p
r
o
ce
s
s
is
th
e
f
o
r
m
atio
n
o
f
a
h
is
to
g
r
am
.
T
h
is
h
is
to
g
r
am
is
u
s
ed
to
ca
lcu
late
th
e
f
r
eq
u
e
n
cy
o
f
th
e
class
r
esu
lts
in
d
icate
d
b
y
th
e
clu
s
ter
in
g
alg
o
r
ith
m
.
T
h
e
f
in
al
r
esu
lt
is
a
n
o
r
m
alize
d
f
ea
tu
r
e
v
ec
to
r
(
lo
ca
l
f
ea
tu
r
es)
af
ter
ap
p
ly
i
n
g
B
OVW,
wh
ich
ca
n
b
e
u
s
ed
as
a
f
ea
tu
r
e
d
ataset
f
o
r
th
e
class
if
ier
.
T
h
e
K
-
m
ea
n
s
clu
s
ter
in
g
al
g
o
r
ith
m
,
a
wid
ely
u
s
ed
m
eth
o
d
in
u
n
s
u
p
er
v
is
ed
lear
n
in
g
,
is
th
e
alg
o
r
ith
m
o
f
c
h
o
ice
in
th
is
r
esear
ch
.
3
.
5
.
Vis
ua
liza
t
io
n
-
ba
s
ed
m
a
chine le
a
rning
m
o
del
I
n
b
u
ild
in
g
a
m
o
d
el,
th
e
b
a
s
e
-
class
if
ier
alg
o
r
ith
m
ac
ts
as
th
e
in
itial
clas
s
if
ier
.
Nex
t,
th
e
m
eta
-
class
if
ier
tech
n
iq
u
e,
a
cr
u
cial
s
tep
in
th
e
p
r
o
ce
s
s
,
i
s
u
s
ed
t
o
co
n
s
tr
u
ct
th
e
f
in
al
p
r
ed
ictio
n
s
b
y
u
tili
zin
g
th
e
p
r
ed
ictio
n
r
esu
lts
f
r
o
m
th
e
b
a
s
e
-
class
if
ier
m
eth
o
d
.
C
o
m
b
in
i
n
g
s
ev
er
al
class
if
icatio
n
m
o
d
e
ls
is
th
en
u
s
ed
as
a
n
ew
f
ea
tu
r
e
[
3
3
]
.
T
h
e
r
esu
lts
o
f
th
e
b
ase
class
if
ier
ar
e
co
m
b
in
ed
an
d
u
s
ed
t
o
tr
ain
th
e
m
eta
-
class
if
ier
task
ed
with
p
r
o
v
id
i
n
g
f
in
al
p
r
ed
icti
o
n
r
esu
lts
[
3
4
]
.
S
VM
,
RF
,
an
d
L
GB
M
wer
e
am
o
n
g
th
e
s
u
p
er
v
is
ed
lear
n
in
g
alg
o
r
ith
m
s
em
p
lo
y
e
d
in
th
is
r
esear
ch
as
b
ase
-
clas
s
if
ier
s
.
M
o
r
eo
v
e
r
,
th
e
lo
g
is
tic
r
eg
r
ess
io
n
tech
n
iq
u
e,
wh
ich
s
er
v
es a
s
a
m
eta
-
class
if
ier
,
is
e
m
p
lo
y
ed
in
th
is
s
tu
d
y
m
o
d
el.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
8
3
4
-
1
8
4
9
1840
Fig
u
r
e
5
.
T
h
e
KAZ
E
f
ea
tu
r
es e
x
tr
ac
tio
n
p
r
o
ce
s
s
3
.
5
.
1
.
B
a
s
e
-
cla
s
s
if
ier
T
h
e
d
ataset
is
p
r
o
ce
s
s
ed
u
s
in
g
s
u
p
er
v
is
ed
lear
n
in
g
tech
n
iq
u
es
d
u
r
i
n
g
th
e
tr
ai
n
in
g
an
d
te
s
tin
g
s
tag
es
o
f
cr
ea
tin
g
th
e
b
ase
-
class
if
ier
m
o
d
el,
wh
ich
will
p
r
o
d
u
ce
ea
ch
p
r
ed
ictio
n
m
o
d
el.
T
h
e
in
itial
s
tep
in
b
u
ild
in
g
th
e
m
o
d
el
is
to
r
u
n
th
e
tr
ai
n
in
g
d
ataset
f
o
r
ev
e
r
y
alg
o
r
ith
m
to
cr
ea
te
a
b
ase
-
class
if
ier
m
o
d
el,
a
f
o
u
n
d
atio
n
al
m
o
d
el
th
at
will
b
e
u
s
ed
to
m
a
k
e
p
r
e
d
ictio
n
s
.
T
h
e
th
ir
d
a
n
d
f
o
u
r
th
ex
p
er
im
en
ts
test
ed
h
y
p
er
p
ar
am
eter
tu
n
in
g
o
n
th
e
b
ase
-
class
if
ier
m
o
d
el.
T
h
e
s
elec
tio
n
o
f
ap
p
r
o
p
r
iate
h
y
p
er
p
ar
am
eter
s
ca
n
im
p
r
o
v
e
th
e
ac
cu
r
ac
y
an
d
p
er
f
o
r
m
an
ce
o
f
th
e
alg
o
r
ith
m
u
s
ed
[
3
5
]
.
Hy
p
e
r
p
ar
am
eter
tu
n
in
g
d
eter
m
in
es
t
h
e
id
ea
l
v
al
u
e
f
o
r
t
h
e
class
if
ier
h
y
p
er
p
ar
am
eter
s
to
en
h
a
n
ce
t
h
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
.
Hy
p
er
p
ar
am
eter
s
lo
wer
th
e
to
tal
n
u
m
b
er
o
f
iter
atio
n
s
,
wh
ich
ca
n
h
el
p
to
im
p
r
o
v
e
t
h
e
m
o
d
e'
s
ef
f
icien
cy
[
3
6
]
.
3
.
5
.
2
.
M
et
a
-
cla
s
s
if
ier
T
h
e
p
er
f
o
r
m
an
ce
o
f
ea
ch
b
ase
clas
s
if
ier
p
r
o
d
u
ce
s
v
ar
y
i
n
g
lev
els
o
f
ac
cu
r
ac
y
.
C
lass
if
ier
s
ar
e
in
teg
r
ated
in
to
o
n
e
to
in
c
r
ea
s
e
th
e
ac
cu
r
ac
y
o
f
th
e
s
in
g
le
class
if
ier
an
d
im
p
r
o
v
e
its
p
r
e
d
icted
p
er
f
o
r
m
an
ce
[
1
8
]
.
Me
ta
-
class
if
ier
s
co
m
b
i
n
e
d
ec
is
io
n
s
f
r
o
m
s
ev
er
al
co
m
b
in
atio
n
m
eth
o
d
s
[
3
7
]
.
L
o
g
is
tic
r
eg
r
ess
io
n
tech
n
iq
u
e
is
u
tili
ze
d
in
t
h
e
c
o
n
s
tr
u
ctio
n
o
f
t
h
e
m
eta
-
class
if
ier
m
o
d
el.
T
h
e
m
eta
-
class
if
ier
o
f
th
e
lo
g
is
tic
r
eg
r
ess
io
n
alg
o
r
ith
m
an
d
a
c
o
m
b
in
atio
n
o
f
b
ase
-
class
if
ier
s
h
av
e
ac
h
ie
v
ed
th
e
h
ig
h
est
ac
c
u
r
ac
y
[
3
8
]
.
Fig
u
r
e
6
ex
p
lain
s
h
o
w
th
e
m
eta
-
class
i
f
ier
was
im
p
lem
en
ted
in
th
is
r
esear
ch
to
in
cr
ea
s
e
th
e
ac
c
u
r
ac
y
o
f
th
e
b
ase
-
class
if
ier
m
o
d
el.
Fig
u
r
e
6
.
Flo
w
d
ia
g
r
am
f
o
r
th
e
m
eta
-
class
if
ier
f
r
am
ewo
r
k
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
d
en
tifi
ca
tio
n
o
f A
n
d
r
o
id
A
P
K
ma
lw
a
r
e
th
r
o
u
g
h
lo
ca
l a
n
d
g
lo
b
a
l fe
a
t
u
r
e
…
(
Yo
g
a
Hera
w
a
n
)
1841
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
ec
tio
n
s
u
m
m
ar
izes
an
d
d
is
cu
s
s
es
th
e
m
ain
f
in
d
in
g
s
o
f
th
e
wo
r
k
.
W
ith
a
R
y
ze
n
7
5
8
0
0
X
C
PU
an
d
6
4
GB
o
f
R
AM
,
I
P
B
'
s
co
m
p
u
tin
g
s
er
v
er
en
v
ir
o
n
m
e
n
t
is
u
s
ed
to
r
u
n
an
ex
p
er
im
en
t.
T
h
is
tech
n
ical
f
o
u
n
d
atio
n
,
co
m
b
in
ed
with
th
e
Py
th
o
n
p
r
o
g
r
am
m
in
g
lan
g
u
a
g
e
an
d
th
e
m
ain
lib
r
a
r
y
p
ac
k
a
g
es
f
o
r
d
ata
s
cien
ce
an
d
m
ath
em
atica
l
co
m
p
u
tatio
n
,
in
clu
d
i
n
g
J
o
b
lib
,
Ma
tp
lo
tl
ib
,
Nu
m
P
y
,
Scik
it_
lear
n
,
Mlx
ten
d
,
L
ig
h
tGB
M
,
Op
en
CV
-
p
y
th
o
n
,
an
d
I
m
b
ala
n
ce
d
-
lear
n
,
allo
wed
u
s
to
p
r
e
-
p
r
o
ce
s
s
an
d
v
is
u
alize
th
e
r
aw
An
d
r
o
id
APK,
ex
tr
ac
tin
g
l
o
ca
l
an
d
g
lo
b
al
f
ea
tu
r
es.
T
h
e
m
ac
h
i
n
e
-
lear
n
in
g
p
r
o
ce
s
s
led
u
s
to
c
r
ea
te
th
e
b
as
e
-
class
if
ier
an
d
t
h
e
m
eta
-
class
if
ier
m
o
d
els.
4
.
1
.
I
ma
g
e
t
ra
ns
f
o
rma
t
io
n
A
h
ex
ad
ec
im
al
r
ea
d
in
g
o
n
th
e
tr
o
jan
APK
k
n
o
wn
as
_
r
ev
ea
l
ed
th
at
th
e
_
is
0
0
6
1
E
1
0
0
b
y
tes
(
r
ev
er
s
e
en
d
ian
)
.
T
h
e
d
ec
im
al
r
e
p
r
esen
ta
tio
n
o
f
th
is
v
alu
e
is
6
4
1
4
5
9
2
.
On
th
e
o
th
er
h
an
d
,
_
is
1
3
5
9
4
0
0
in
d
ec
im
al
n
o
tatio
n
o
r
0
0
1
4
B
E
2
8
i
n
h
ex
a
d
ec
im
a
l.
Data
wer
e
tak
en
f
r
o
m
6
4
1
4
5
9
2
'
s
o
f
f
s
et
1
3
5
9
4
0
0
an
d
p
r
o
ce
s
s
ed
b
y
th
e
s
y
s
tem
to
cr
ea
te
an
8
-
b
it
g
r
ay
s
ca
le
p
ictu
r
e.
T
h
e
p
r
o
ce
s
s
o
f
cr
ea
tin
g
8
-
b
it
g
r
ay
s
ca
le
im
ag
es
is
as
f
o
llo
ws:
a
p
ix
el
is
b
lack
if
t
h
e
d
ata
v
a
lu
e
p
er
b
y
te
is
0
×
0
0
0
0
0
0
0
(
0
)
,
wh
ite
if
th
e
v
alu
e
is
0
×
1
1
1
1
1
1
1
1
(
2
5
5
)
,
an
d
g
r
ey
if
th
e
v
alu
e
is
b
etwe
en
0
an
d
2
5
5
.
Fig
u
r
e
7
s
h
o
ws
th
e
o
u
tco
m
e
o
f
im
a
g
e
tr
an
s
f
o
r
m
atio
n
o
f
ad
war
e
f
am
ily
f
o
r
_
in
Fig
u
r
e
7
(
a)
an
d
_
ℎ
i
n
Fig
u
r
e
7
(
b
)
.
Fig
u
r
e
8
s
h
o
ws th
e
o
u
tco
m
e
o
f
im
ag
e
tr
an
s
f
o
r
m
atio
n
o
f
SM
S f
am
ily
f
o
r
_
in
Fig
u
r
e
8
(
a)
an
d
_
in
Fig
u
r
e
8
(
b
)
.
(
a)
(
b
)
Fig
u
r
e
7
.
I
m
ag
e
tr
a
n
s
f
o
r
m
atio
n
r
esu
lts
f
o
r
m
alwa
r
e
(
a
)
_
an
d
(
b
)
_
ℎ
(
a)
(
b
)
Fig
u
r
e
8
.
I
m
ag
e
tr
a
n
s
f
o
r
m
atio
n
r
esu
lts
f
o
r
m
alwa
r
e
(
a
)
_
an
d
(
b
)
_
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
25
:
1
8
3
4
-
1
8
4
9
1842
4
.
2
.
G
l
o
ba
l
f
ea
t
ure
e
x
t
ra
c
t
io
n
T
h
e
G
I
S
T
f
e
a
t
u
r
e
e
x
t
r
a
ct
i
o
n
te
c
h
n
i
q
u
e
e
x
t
r
a
c
t
s
g
l
o
b
a
l
f
e
at
u
r
e
s
f
r
o
m
a
s
i
n
g
l
e
g
r
a
y
s
c
al
e
im
a
g
e
.
T
h
i
s
r
e
s
e
a
r
c
h
p
r
o
c
es
s
es
u
p
t
o
8
o
r
i
en
t
a
t
i
o
n
s
u
s
i
n
g
a
f
o
u
r
-
s
ca
l
e
G
ab
o
r
f
i
l
t
e
r
,
d
i
v
i
d
es
e
a
c
h
f
ea
t
u
r
e
m
a
p
i
n
t
o
1
6
r
e
g
i
o
n
s
u
s
i
n
g
a
4
×
4
g
r
i
d
,
a
n
d
t
h
e
n
t
a
k
es
th
e
a
v
e
r
a
g
e
o
f
e
a
c
h
r
e
g
i
o
n
'
s
f
ea
t
u
r
e
v
a
l
u
e
s
.
C
o
n
c
at
e
n
a
ti
n
g
t
h
e
1
6
a
v
e
r
a
g
e
d
v
a
l
u
e
s
o
f
a
l
l
3
2
f
e
a
t
u
r
e
m
a
p
s
y
i
e
l
d
s
4
×
8
×
4
×
4
=
512
G
I
S
T
d
e
s
c
r
i
p
t
o
r
s
a
f
t
er
d
e
t
e
r
m
i
n
i
n
g
e
a
c
h
r
e
g
i
o
n
's
a
v
e
r
a
g
e
f
e
a
t
u
r
e
v
a
l
u
es
.
D
e
ci
m
al
n
u
m
b
e
r
s
a
r
e
k
e
p
t
i
n
te
x
t
f
i
l
es
f
o
r
e
v
e
r
y
f
e
a
t
u
r
e
,
a
s
s
h
o
w
n
i
n
F
i
g
u
r
e
9
.
Fig
u
r
e
9
.
Gr
a
y
s
ca
le
im
ag
e
f
ea
tu
r
es p
r
o
ce
s
s
ed
ag
ain
s
t th
e
GI
ST
alg
o
r
ith
m
4
.
3
.
L
o
ca
l
f
ea
t
ure
ex
t
ra
ct
io
n
T
h
e
KAZ
E
alg
o
r
ith
m
is
in
s
tr
u
m
en
tal
in
id
en
tify
i
n
g
th
e
cr
u
cial
p
o
in
ts
o
f
th
e
p
r
o
ce
s
s
ed
g
r
ay
s
ca
le
im
ag
e.
W
h
ile
n
o
t
d
ir
ec
tly
u
s
ab
le
as
f
ea
tu
r
e
s
elec
tio
n
o
u
t
p
u
ts
f
o
r
an
im
a
g
e,
th
ese
id
e
n
tifie
d
cr
itical
s
p
o
ts
p
r
o
d
u
ce
th
e
f
ea
tu
r
e
e
x
tr
ac
tio
n
v
alu
e.
T
h
is
v
alu
e
is
g
en
er
ate
d
b
y
co
n
s
tr
u
ctin
g
a
co
d
e
b
o
o
k
u
s
in
g
th
e
d
escr
ip
to
r
v
alu
e
to
d
escr
ib
e
th
e
i
d
en
tifi
ed
cr
itical
p
o
in
t.
T
h
e
f
ea
tu
r
e
s
cr
ea
ted
r
esu
lt
f
r
o
m
th
e
b
a
g
o
f
v
is
u
al
wo
r
d
s
(
B
OVW)
co
n
s
tr
u
ctio
n
p
r
o
ce
d
u
r
e
u
s
in
g
th
e
s
ize
ten
d
ictio
n
ar
y
(
n
u
m
b
er
o
f
class
es*
5
)
ch
o
s
en
f
o
r
t
h
is
r
esear
ch
.
C
r
ea
tin
g
a
co
d
eb
o
o
k
is
a
co
m
p
r
eh
en
s
iv
e
p
r
o
ce
s
s
in
v
o
lv
i
n
g
an
aly
zin
g
th
e
g
r
ay
s
ca
le
im
ag
e
to
id
en
tify
an
y
p
o
ten
tial d
escr
ip
to
r
s
.
As we
f
in
d
ea
ch
ch
ar
ac
ter
izer
,
th
e
n
u
m
b
er
o
f
r
ec
o
r
d
s
f
o
r
th
at
d
e
s
cr
ip
to
r
in
cr
ea
s
es in
th
e
co
d
eb
o
o
k
.
T
h
e
cu
lm
in
ati
o
n
o
f
th
is
p
r
o
ce
s
s
is
a
h
is
to
g
r
am
v
alu
e
th
at
s
u
cc
in
ctly
d
escr
ib
es
th
e
u
n
iq
u
e
ch
ar
ac
ter
is
tics
o
f
th
e
g
r
ay
s
ca
l
e
im
ag
e.
Fig
u
r
e
1
0
v
is
u
ally
r
ep
r
esen
ts
th
e
KAZ
E
f
ea
tu
r
e
e
x
tr
ac
tio
n
,
a
p
r
o
ce
s
s
we
ex
ec
u
te
u
s
in
g
th
e
p
o
wer
f
u
l
Py
th
o
n
Data
f
r
a
m
e
p
ac
k
a
g
e.
Fig
u
r
e
10
.
Gr
ay
s
ca
le
im
ag
e
f
e
atu
r
es p
r
o
ce
s
s
ed
ag
ain
s
t th
e
K
AZ
E
alg
o
r
ith
m
4
.
4
.
H
y
perpa
ra
m
e
t
er
t
un
ing
R
an
d
o
m
s
ea
r
ch
(
R
S)
an
d
g
r
i
d
s
ea
r
ch
(
GS)
ar
e
h
y
p
e
r
p
ar
a
m
eter
s
ea
r
ch
m
eth
o
d
o
lo
g
ies
u
s
ed
in
th
is
r
esear
ch
.
T
h
e
f
o
u
r
th
ex
p
er
im
en
t
em
p
lo
y
ed
th
e
r
an
d
o
m
s
ea
r
ch
tech
n
iq
u
e,
wh
er
ea
s
th
e
th
ir
d
u
s
ed
th
e
g
r
id
s
ea
r
ch
tech
n
iq
u
e.
I
n
b
o
th
ca
s
e
s
,
th
e
m
o
d
el
was
tr
ain
ed
o
n
th
e
tr
ain
in
g
s
et,
a
cr
u
cial
s
tep
th
at
d
eter
m
in
es
th
e
m
o
d
el'
s
p
er
f
o
r
m
a
n
ce
,
u
s
in
g
o
p
tim
al
h
y
p
er
p
a
r
am
eter
s
f
o
r
e
ac
h
m
o
d
el.
O
p
tim
al
p
a
r
am
ete
r
s
id
en
tifie
d
u
s
in
g
Gr
id
Sear
ch
an
d
R
an
d
o
m
Sear
ch
ar
e
p
r
esen
ted
in
T
a
b
le
4
.
4
.
5
.
M
a
chine
-
lea
rning
m
o
de
l
Fo
u
r
ex
p
er
im
e
n
tal
s
ch
em
es h
av
e
b
ee
n
ca
r
r
ied
o
u
t u
s
in
g
f
ea
t
u
r
es f
r
o
m
g
r
ay
s
ca
le
im
ag
e
d
at
asets
.
T
h
i
s
r
esear
ch
is
g
en
er
ally
d
iv
id
ed
i
n
to
th
r
ee
s
ec
tio
n
s
ap
p
lied
to
f
o
u
r
ex
p
e
r
im
en
tal
s
ch
em
es.
T
h
e
f
ir
s
t
s
ec
t
io
n
u
s
ed
lo
ca
l
f
ea
tu
r
e
ex
tr
ac
tio
n
(
KAZ
E
)
r
esu
lts
to
tr
ain
m
o
d
els
f
o
r
all
ex
p
er
im
e
n
ts
.
Glo
b
al
f
ea
tu
r
e
ex
tr
ac
tio
n
(
GI
ST)
r
esu
lts
wer
e
u
s
ed
in
th
e
s
ec
o
n
d
s
ec
tio
n
to
tr
ain
th
e
m
o
d
el
f
o
r
all
ex
p
er
im
en
ts
.
T
h
e
th
ir
d
s
ec
tio
n
co
m
b
in
es
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
I
d
en
tifi
ca
tio
n
o
f A
n
d
r
o
id
A
P
K
ma
lw
a
r
e
th
r
o
u
g
h
lo
ca
l a
n
d
g
lo
b
a
l fe
a
t
u
r
e
…
(
Yo
g
a
Hera
w
a
n
)
1843
class
if
icatio
n
r
esu
lt
s
o
f
th
r
ee
m
ac
h
in
e
-
lear
n
i
n
g
alg
o
r
ith
m
s
,
in
clu
d
in
g
b
u
ild
in
g
m
o
d
els u
s
in
g
lo
ca
l,
g
lo
b
al,
o
r
a
co
m
b
in
atio
n
o
f
f
ea
tu
r
es.
T
h
e
last
s
ec
tio
n
is
ca
lled
th
e
ME
T
A
m
o
d
el.
T
h
e
r
esu
lts
o
f
t
h
e
m
o
d
el
p
e
r
f
o
r
m
an
ce
co
m
p
ar
is
o
n
b
etwe
en
KAZ
E
an
d
GI
ST
alg
o
r
ith
m
f
ea
tu
r
es
ar
e
s
h
o
w
n
in
T
ab
le
4
,
d
em
o
n
s
tr
atin
g
o
u
r
co
n
f
id
en
ce
in
o
u
r
r
esear
ch
p
r
o
ce
s
s
an
d
r
esu
lts
.
T
ab
le
4
.
T
h
e
b
est p
ar
am
ete
r
s
u
s
in
g
g
r
id
s
ea
r
c
h
(
GS)
an
d
r
an
d
o
m
s
ea
r
ch
(
R
S)
t
ec
h
n
iq
u
es
M
e
t
o
d
e
F
e
a
t
u
r
e
s
e
x
t
r
a
c
t
i
o
n
P
a
r
a
me
t
e
r
LG
B
M
RF
S
V
M
GS
G
I
S
T
B
o
o
st
i
n
g
_
t
y
p
e
G
O
S
S
-
-
Le
a
r
n
i
n
g
_
r
a
t
e
0
.
2
-
-
M
a
x
_
d
e
p
t
h
5
-
-
R
a
n
d
o
m
_
st
a
t
e
41
-
-
D
e
t
e
r
m
i
n
i
st
i
c
Tr
u
e
-
-
F
o
r
c
e
_
c
o
l
_
w
i
s
e
Tr
u
e
-
-
M
a
x
_
d
e
p
t
h
-
6
-
M
a
x
_
f
e
a
t
u
r
e
s
-
N
o
n
e
-
M
a
x
_
l
e
a
d
_
n
o
d
e
s
-
9
-
N
_
e
st
i
ma
t
o
r
s
-
25
-
C
-
-
1
0
0
G
a
mm
a
-
-
0
.
1
k
e
r
n
e
l
-
-
Li
n
e
a
r
K
A
ZE
B
o
o
st
i
n
g
_
t
y
p
e
G
B
D
T
-
-
Le
a
r
n
i
n
g
_
r
a
t
e
0
.
1
-
-
M
a
x
_
d
e
p
t
h
-
10
-
-
R
a
n
d
o
m
_
st
a
t
e
40
-
-
D
e
t
e
r
m
i
n
i
st
i
c
Tr
u
e
-
-
F
o
r
c
e
_
c
o
l
_
w
i
s
e
Tr
u
e
-
-
M
a
x
_
d
e
p
t
h
-
6
-
M
a
x
_
f
e
a
t
u
r
e
s
-
N
o
n
e
-
M
a
x
_
l
e
a
d
_
n
o
d
e
s
-
9
-
N
_
e
st
i
ma
t
o
r
s
-
50
-
C
-
-
1
0
0
G
a
mm
a
-
-
0
.
0
0
0
1
k
e
r
n
e
l
-
-
RBF
RS
G
I
S
T
B
o
o
st
i
n
g
_
t
y
p
e
G
B
D
T
-
-
Le
a
r
n
i
n
g
_
r
a
t
e
0
.
1
-
-
M
a
x
_
d
e
p
t
h
-
5
-
-
R
a
n
d
o
m
_
st
a
t
e
42
-
-
D
e
t
e
r
m
i
n
i
st
i
c
Tr
u
e
-
-
F
o
r
c
e
_
c
o
l
_
w
i
s
e
Tr
u
e
-
-
M
a
x
_
d
e
p
t
h
-
9
-
M
a
x
_
f
e
a
t
u
r
e
s
-
S
Q
R
T
-
M
a
x
_
l
e
a
d
_
n
o
d
e
s
-
9
-
N
_
e
st
i
ma
t
o
r
s
-
1
5
0
-
C
-
-
1
0
0
G
a
mm
a
-
-
0
.
1
k
e
r
n
e
l
-
-
Li
n
e
a
r
K
A
ZE
B
o
o
st
i
n
g
_
t
y
p
e
G
B
D
T
-
-
Le
a
r
n
i
n
g
_
r
a
t
e
0
.
0
5
-
-
M
a
x
_
d
e
p
t
h
-
5
-
-
R
a
n
d
o
m
_
st
a
t
e
41
-
-
D
e
t
e
r
m
i
n
i
st
i
c
Tr
u
e
-
-
F
o
r
c
e
_
c
o
l
_
w
i
s
e
Tr
u
e
-
-
M
a
x
_
d
e
p
t
h
-
9
-
M
a
x
_
f
e
a
t
u
r
e
s
-
N
o
n
e
-
M
a
x
_
l
e
a
d
_
n
o
d
e
s
-
9
-
N
_
e
st
i
ma
t
o
r
s
-
1
0
0
-
C
-
-
1
0
0
G
a
mm
a
-
-
0
.
0
0
0
1
k
e
r
n
e
l
-
-
RBF
4
.
5
.
1
.
M
o
del per
f
o
rm
a
nce
ba
s
ed
o
n lo
ca
l f
ea
t
ure
ex
t
r
a
ct
io
n
T
h
is
s
ec
tio
n
u
s
es
a
d
ataset
f
r
o
m
lo
ca
l
f
ea
tu
r
e
e
x
tr
ac
tio
n
(
K
AZ
E
)
f
o
r
b
u
ild
in
g
m
o
d
els.
B
ased
o
n
d
ata
f
r
o
m
T
ab
le
5
,
th
e
b
est
m
o
d
el
p
er
f
o
r
m
an
ce
b
ased
o
n
KAZ
E
f
ea
tu
r
e
d
ata
was
o
b
tain
ed
i
n
th
e
f
ir
s
t
ex
p
er
im
en
t
f
o
r
t
h
e
L
GB
M
class
if
ier
.
T
h
e
L
GB
M
m
o
d
el
d
em
o
n
s
tr
ated
e
x
ce
p
tio
n
al
p
er
f
o
r
m
an
ce
with
ac
cu
r
ac
y
,
F1
s
co
r
e
,
r
ec
all,
an
d
p
r
ec
is
io
n
v
alu
es
o
f
9
5
%.
Fig
u
r
e
1
1
co
m
p
a
r
es
th
e
ex
ec
u
tio
n
tim
e
o
f
th
e
b
ase
class
if
ier
alg
o
r
ith
m
o
n
KAZ
E
f
ea
tu
r
e
d
ata.
B
ased
o
n
th
is
d
ata,
th
e
L
GB
M
m
o
d
e
l
r
eq
u
ir
es
th
e
f
astes
t
ex
ec
u
tio
n
tim
e
c
o
m
p
ar
e
d
to
o
th
er
b
ase
-
class
if
ier
alg
o
r
ith
m
s
.
I
n
ad
d
itio
n
,
L
GB
M
p
er
f
o
r
m
ed
s
tab
lely
in
all
ex
p
e
r
im
en
ts
.
Evaluation Warning : The document was created with Spire.PDF for Python.