Ind
o
n
es
ian Jou
r
n
al
o
f
E
le
ctric
a
l E
n
g
in
ee
r
ing
and
C
o
mp
u
t
er
S
c
ienc
e
V
ol
. 8
,
No.
3
,
Dec
em
be
r
20
17
, p
p
.
6
23
~
6
26
DO
I: 1
0.
11
5
91
/
i
j
ee
c
s
.v
8
.i
3
.
623
-
6
26
6
23
Rec
ei
v
ed
A
ug
us
t
2
, 2
01
7
;
Rev
i
s
ed
O
c
to
be
r
2
2,
2
01
7
;
A
c
c
ep
ted
Nov
e
mb
er
8,
2
01
7
Text
ure Clas
sifi
catio
n Base
d
O
n Emp
i
rical
W
a
v
e
l
et
Trans
fo
rm
Usin
g L
BP
Feat
ures
Ramesh
P
.
1
,
V
.M
ath
iv
anan
2
1
AM
E
T
Un
i
v
e
rs
i
ty
,
Che
n
n
a
i
2
ARM
c
o
l
l
e
g
e
o
f
En
g
i
n
e
e
ri
n
g
a
n
d
T
e
c
h
n
o
l
o
g
y
,
Ch
e
n
n
a
i
Ab
strac
t
Au
to
m
a
ti
c
i
n
s
p
e
c
t
i
o
n
s
y
s
te
m
s
b
e
c
o
m
e
m
o
re
i
m
p
o
rta
n
c
e
fo
r
i
n
d
u
s
tri
e
s
wit
h
h
i
g
h
p
ro
d
u
c
ti
v
e
p
l
a
n
s
e
s
p
e
c
i
a
l
l
y
i
n
te
x
tu
re
i
n
d
u
s
try
.
A
n
o
v
e
l
a
p
p
ro
a
c
h
to
L
o
c
a
l
Bi
n
a
r
y
Pa
tt
e
rn
(L
BP)
fe
a
t
u
re
fo
r
te
x
t
u
re
c
l
a
s
s
i
fi
c
a
ti
o
n
i
s
p
ro
p
o
s
e
d
i
n
t
h
i
s
s
y
s
te
m
.
At
t
h
e
f
i
r
s
t,
th
e
p
ro
p
o
s
e
d
Em
p
i
ri
c
a
l
W
a
v
e
l
e
t
Tr
a
n
s
f
o
rm
(EWT
)
b
a
s
e
d
te
x
tu
re
c
l
a
s
s
i
fi
c
a
ti
o
n
i
s
t
e
s
te
d
o
n
g
ra
y
s
c
a
l
e
a
n
d
c
o
l
o
r
i
m
a
g
e
s
b
y
u
s
i
n
g
Bro
d
a
tz
te
x
tu
re
i
m
a
g
e
s
.
Th
e
g
ra
y
s
c
a
l
e
a
n
d
c
o
l
o
r
i
m
a
g
e
i
s
d
e
c
o
m
p
o
s
e
d
b
y
E
W
T
a
t
2
a
n
d
3
l
e
v
e
l
o
f
d
e
c
o
m
p
o
s
i
ti
o
n
.
L
BP
fe
a
tu
r
e
s
a
r
e
c
a
l
c
u
l
a
te
d
f
o
r
e
a
c
h
e
m
p
i
ri
c
a
l
t
ra
n
s
f
o
rm
e
d
i
m
a
g
e
.
Ex
tr
a
c
t
e
d
f
e
a
tu
re
s
a
re
g
i
v
e
n
a
s
i
n
p
u
t
to
t
h
e
c
l
a
s
s
i
f
i
c
a
ti
o
n
s
ta
g
e
.
K
-
NN
c
l
a
s
s
i
fi
e
r
i
s
u
s
e
d
fo
r
c
l
a
s
s
i
fi
c
a
t
i
o
n
s
ta
g
e
.
Th
e
r
e
s
u
l
t
o
f
th
e
p
r
o
p
o
s
e
d
s
y
s
te
m
g
i
v
e
s
s
a
ti
s
fa
c
to
r
y
c
l
a
s
s
i
fi
c
a
ti
o
n
a
c
c
u
ra
c
y
o
f
o
v
e
r
9
8
%
f
o
r a
l
l
t
y
p
e
s
o
f
i
m
a
g
e
s
.
Key
w
ords
:
Te
x
tu
r
e
c
l
a
s
s
i
f
i
c
a
ti
o
n
,
E
W
T,
L
BP,
K
-
NN.
Copy
righ
t
©
2
0
1
7
I
ns
titu
t
e
o
f
Adv
a
nc
e
d
Eng
i
ne
e
ring
a
nd
Sc
ie
nc
e
.
All
righ
t
s
re
s
e
rve
d.
1.
Int
r
o
d
u
ctio
n
T
he
r
ev
i
e
w
of
l
i
t
erature
g
i
v
en
i
n
t
hi
s
s
ec
t
i
on
i
s
c
e
nt
ered
up
o
n
v
ari
ou
s
t
ec
hn
i
q
ue
s
f
or
tex
ture
c
l
as
s
i
f
i
c
at
i
on
.
Int
e
ns
i
v
e
r
es
ea
r
c
h
w
ork
ha
s
be
en
un
de
r
tak
en
i
n
the
d
ev
e
l
o
pm
en
t
of
au
tom
ate
d
i
m
ag
e
an
al
y
s
i
s
m
eth
od
s
to
c
l
as
s
i
f
y
t
ex
ture
i
m
ag
es
.
Cur
v
e
l
et
s
t
ati
s
ti
c
al
a
nd
c
o
-
oc
c
urr
en
c
e
f
ea
tures
f
or
tex
ture
c
l
as
s
i
f
i
c
ati
o
n
i
s
ex
pl
a
i
ne
d
i
n
[1]
.
T
he
tex
ture
c
l
as
s
i
f
i
c
ati
on
i
s
s
ue
us
i
ng
c
urv
e
l
et
tr
an
s
f
orm
i
s
an
a
l
y
z
e
d.
C
urv
el
e
t
ar
i
t
hm
eti
c
an
d
c
o
-
oc
c
urr
en
c
e
f
ea
tures
are
s
ub
s
eq
ue
nt
f
r
om
th
e s
ub
b
an
ds
of
th
e
c
urv
e
l
et
de
c
om
po
s
i
t
i
on
an
d
are
us
ed
f
or c
l
as
s
i
f
i
c
ati
on
.
W
av
el
et
T
r
an
s
f
or
m
s
ba
s
ed
on
g
au
s
s
i
a
n
m
ar
k
ov
r
an
do
m
f
i
el
d
b
as
ed
t
ex
ture
c
l
as
s
i
f
i
c
ati
on
i
s
d
i
s
c
us
s
ed
i
n
[2]
.
G
MRF
m
od
el
o
n
l
i
n
e
ar
wav
el
ets
i
s
us
e
d
f
or
f
ea
ture
ex
tr
ac
ti
on
f
or
the
pu
r
po
s
e
of
tex
ture
c
l
as
s
i
f
i
c
ati
on
i
n
th
i
s
pa
p
er.
Le
as
t
s
qu
are
err
or
es
t
i
m
ati
on
m
eth
od
i
s
us
ed
f
or ex
tr
ac
ti
o
n t
h
e s
ev
en
f
ea
tures
o
n t
h
i
r
d
order
m
ar
k
ov
ne
i
gh
bo
r
ho
od
.
T
ex
ture
c
l
as
s
i
f
i
c
ati
on
b
as
e
d
on
wav
el
e
t
f
ea
tures
i
s
pr
es
en
te
d
i
n
[3]
.
F
ea
ture
ex
tr
ac
ti
on
al
g
orit
hm
us
i
ng
wav
el
e
t
d
ec
om
po
s
ed
i
m
ag
es
of
an
i
m
ag
e
an
d
i
ts
c
om
pl
e
m
en
tar
y
i
m
ag
e
f
or
tex
ture
c
l
as
s
i
f
i
c
at
i
on
i
s
pre
s
en
ted
i
n
t
hi
s
p
ap
er.
E
uc
l
i
de
an
di
s
t
an
c
e
m
ea
s
ure
an
d
the
m
i
ni
m
u
m
di
s
tan
c
e
c
l
as
s
i
f
i
er to
c
l
as
s
i
f
y
th
e t
ex
t
ure are
us
ed
.
Li
n
ea
r
r
eg
r
es
s
i
on
m
od
el
ba
s
ed
on
wav
el
et
tr
a
ns
form
f
or
tex
ture
c
l
as
s
i
f
i
c
ati
on
i
s
ap
pro
ac
he
d
i
n
[4]
.
F
ea
t
ure
s
are
ex
tr
ac
ted
f
r
o
m
th
e
l
i
n
ea
r
r
eg
r
es
s
i
on
m
od
el
an
d
c
orr
el
ati
on
to
o
tha
t c
h
arac
teri
z
e
the
s
am
pl
es
. P
y
r
am
i
d s
tr
uc
tured
wav
el
et
trans
f
orm
an
d t
r
ee
s
tr
u
c
tured
w
a
v
e
l
et
tr
an
s
f
or
m
i
s
us
ed
.
G
ab
or
f
i
l
ters
an
d
c
o
-
oc
c
urr
en
c
e
pro
ba
b
i
l
i
ti
es
ba
s
ed
de
s
i
gn
f
or
t
ex
ture
f
ea
ture
f
us
i
on
i
s
d
i
s
c
us
s
ed
i
n
[5]
[
8].
T
he
f
us
ed
f
ea
ture
s
et
us
ed
G
ab
or
f
i
l
ters
pro
ba
b
l
e
of
ac
c
uratel
y
c
ap
turi
n
g
l
es
s
er
an
d
c
en
te
r
-
fr
eq
ue
nc
y
t
ex
ture
de
t
ai
l
.
K
-
m
ea
ns
,
prin
c
i
pa
l
c
om
po
ne
nt
an
al
y
s
i
s
i
s
us
ed
.
Mu
l
t
i
m
od
al
i
nv
aria
nt
l
oc
al
b
i
na
r
y
pa
t
tern
b
as
ed
tex
t
ure
c
l
as
s
i
f
i
c
ati
on
i
s
pres
en
t
ed
i
n
[6]
.
Ho
w
t
hi
s
s
y
s
tem
de
s
c
r
i
pto
r
c
an
be
bu
i
l
t
ef
f
i
c
i
en
t
l
y
i
s
d
es
c
r
i
be
d
i
n
th
i
s
p
ap
er.
A
l
s
o
d
em
on
s
tr
ate
em
pi
r
i
c
al
l
y
th
at
c
om
pa
r
ed
to
al
l
th
e
s
tat
e
of
the
art
L
B
P
-
b
as
ed
d
es
c
r
i
pto
r
s
.
Cur
v
e
l
et
tr
a
ns
f
or
m
ba
s
ed
t
ex
ture
c
l
as
s
i
f
i
c
ati
o
n
i
s
ap
pr
oa
c
he
d
i
n
[7]
.
O
ne
group
f
ea
t
ure
v
ec
tor
c
an
b
e
de
v
e
l
op
ed
b
y
the
m
ea
n
an
d
v
aria
nc
e
of
the
c
urv
e
l
et
s
tat
i
s
ti
c
a
l
f
ea
tures
,
whi
c
h
are
r
es
u
l
ti
ng
f
r
o
m
the
s
ub
-
ba
nd
s
of
the
c
ur
v
e
l
et
de
c
o
m
po
s
i
ti
on
a
nd
are
us
e
d
f
or
c
l
as
s
i
f
i
c
ati
o
n
p
urpos
e.
W
a
v
e
l
et
tr
an
s
f
or
m
an
d
ne
ural
n
et
w
o
r
k
ba
s
ed
c
ol
or
tex
ture
c
l
as
s
i
f
i
c
at
i
on
i
s
de
s
c
r
i
be
d
i
n
[9
].
F
ea
tur
e
ex
tr
ac
tor
us
i
n
g
wav
el
e
t
d
om
ai
n
an
d
ne
ura
l
ne
t
w
ork
s
en
s
em
bl
es
c
l
as
s
i
f
i
er
are
us
e
d
i
n
th
i
s
s
c
h
em
e.
F
ea
ture
ex
tr
ac
tor us
i
n
g
w
a
v
e
l
et
do
m
ai
n i
nc
l
u
de
s
en
tr
op
y
a
nd
en
erg
y
f
ea
tures
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N:
25
02
-
4
75
2
IJE
E
CS
V
ol
.
8
,
N
o.
3
,
Dec
em
be
r
2017
:
623
–
6
26
624
2.
P
r
o
p
o
se
d
S
y
st
em
T
hi
s
s
ec
ti
on
c
on
c
e
ntrat
es
on
c
l
as
s
i
f
i
c
at
i
on
of
gra
y
i
m
ag
es
b
as
ed
on
E
W
T
an
d
K
NN.
In
c
o
m
m
on
,
a
t
y
pi
c
a
l
c
l
as
s
i
f
i
c
ati
on
s
y
s
tem
ge
ne
r
a
l
l
y
c
on
s
i
s
ts
of
tw
o
ph
as
es
;
f
i
r
s
t
i
s
f
ea
ture
ex
tr
ac
ti
on
a
nd
s
ec
on
d
i
s
c
l
as
s
i
f
i
c
ati
on
p
ha
s
e.
A
l
l
the
s
tag
es
are
ex
p
l
a
i
ne
d
i
n
f
ac
t
i
n
the
f
ol
l
o
w
i
ng
s
ub
s
ec
ti
on
s
.
P
r
op
os
e
d
gr
a
y
tex
t
ure
c
l
as
s
i
f
i
c
ati
on
s
y
s
tem
us
i
ng
em
pi
r
i
c
al
b
an
d
s
i
gn
at
ure
i
s
s
ho
w
n
i
n
F
i
gu
r
e
1.
T
ab
l
es
an
d
F
i
gu
r
es
are
pres
e
nte
d
c
en
ter,
as
s
ho
w
n
b
el
o
w
a
nd
c
i
te
d
i
n
t
he
m
an
us
c
r
i
pt.
F
i
gu
r
e
1
.
P
r
op
os
e
d g
r
a
y
t
e
x
ture c
l
as
s
i
f
i
c
a
ti
on
s
y
s
t
em
ba
s
ed
on
D
S
T
an
d
K
NN
2.1
.
E
WT
&
LBP
T
he
ai
m
of
the
E
W
T
i
s
to
de
c
a
y
an
i
m
ag
e
or
s
i
gn
al
on
wav
el
e
t
r
i
gi
d
f
r
a
m
es
w
h
i
c
h
ar
e
bu
i
l
t
ef
f
i
c
i
en
tl
y
.
F
i
n
di
n
g
t
h
e
c
ha
i
ns
of
s
om
e
"
m
od
es
"
i
n
t
he
F
ou
r
i
er
s
pe
c
tr
u
m
us
i
ng
the
s
e
s
up
po
r
ts
to
c
on
s
tr
uc
t
Li
tt
l
e
w
oo
d
-
P
a
l
e
y
t
y
pe
wav
el
e
ts
i
n
1D
i
s
the
pr
oc
ed
ur
e.
A
C
ur
v
e
l
et
tr
an
s
f
or
m
,
Li
ttl
e
w
oo
d
-
P
al
e
y
tr
a
ns
f
or
m
,
an
d
t
he
Ri
d
ge
l
et
tr
a
ns
f
or
m
.
T
o
k
ee
p
m
utu
al
l
y
s
om
e
de
ta
i
l
s
i
s
th
e
ad
v
an
tag
e
of
thi
s
em
pi
r
i
c
al
ad
v
a
nc
e
tha
t
or
el
s
e
w
ou
l
d
b
e
s
pl
i
t
i
n
t
h
e
c
on
ta
i
ne
r
of
d
y
a
di
c
f
i
l
t
ers
.
(a)
(b)
F
i
g
ure
2
.
(
a) I
np
ut
Im
ag
e (b)
Dete
c
ted
Four
i
er S
up
p
ort
s
B
ac
k
ground
of
the
pi
x
e
l
s
a
r
e
ex
pl
a
i
ne
d
i
n
bi
n
ar
y
p
att
erns
i
n
the
r
e
gi
on
s
i
s
de
f
i
n
ed
as
LB
P
f
ea
tures
.
T
he
ob
tai
ne
d
f
ea
tures
fr
o
m
the
r
eg
i
on
s
are
i
nc
l
ud
e
d
i
nt
o
an
ex
ac
ti
ng
f
ea
ture
hi
s
tog
r
am
. Im
ag
es
c
an
be
c
o
m
pa
r
ed
b
y
m
ea
s
urin
g t
h
e c
om
pa
r
i
s
on
be
t
wee
n t
h
ei
r
hi
s
tog
r
am
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJE
E
CS
IS
S
N:
2
50
2
-
4
75
2
T
ex
ture Cl
as
s
i
fi
c
at
i
o
n B
as
e
d On E
mp
i
r
i
c
a
l
W
av
el
et
T
r
a
ns
form Us
i
ng
…
(
Ram
es
h
P
.
)
625
F
i
gu
r
e
3
.
L
B
P
O
pe
r
at
or
2.2
.
K
-
NN
Cla
ss
if
ie
r
T
he
c
on
c
l
ud
i
n
g
s
tag
e
of
the
propos
ed
a
pp
r
o
ac
h
i
s
c
l
as
s
i
f
i
c
ati
on
or
v
al
i
da
t
i
o
n
s
tag
e.
I
n
the
c
l
as
s
i
f
i
c
ati
on
s
tag
e,
th
e
t
y
p
e
of
the
un
k
no
w
n
tex
tur
e
i
m
ag
e
i
s
m
en
ti
o
ne
d
to
a
k
no
w
n
t
y
p
e
b
y
us
i
ng
t
he
f
ea
tures
ex
tr
ac
te
d.
T
he
c
l
as
s
i
f
i
er
us
ed
to
as
s
i
gn
a
tar
ge
t
c
l
as
s
to
a
n
un
k
no
w
n
tex
tur
e
i
m
ag
e
i
s
K
NN
c
l
as
s
i
f
i
er.
T
he
prop
os
ed
em
pi
r
i
c
al
b
an
d
f
ea
tures
are
ex
tr
ac
ted
f
or
the
un
k
no
wn
tex
ture
i
m
ag
e
at
f
i
r
s
t.
B
y
us
i
ng
t
he
f
ea
ture
r
ef
erenc
e
l
i
s
t
a
l
on
g
w
i
t
h
the
f
ea
t
ures
of
un
k
no
w
n
tex
ture
i
m
ag
e,
K
N
N
c
l
as
s
i
f
i
er
as
s
i
gn
s
the
r
ef
erenc
e
t
e
x
ture
i
m
ag
e
c
l
as
s
to
t
he
un
k
no
w
n
t
ex
ture
i
m
ag
e.
3.
E
xper
i
men
t
al
R
es
u
lt
s
In
th
i
s
c
ha
pte
r
,
t
he
ex
pe
r
i
m
en
tal
r
es
ul
ts
an
d
the
i
r
i
m
pl
i
c
a
ti
o
n
ar
e
d
i
s
c
us
s
ed
.
A
l
s
o,
th
e
pe
r
f
or
m
an
c
e
of
the
propos
ed
c
l
as
s
i
f
i
c
ati
on
s
y
s
t
em
s
for
gra
y
t
ex
ture
i
m
ag
es
an
d
c
ol
or
tex
t
ure
i
m
ag
es
ba
s
ed
on
th
e
prop
o
s
ed
E
W
T
an
d
LB
P
f
ea
tures
us
i
ng
K
NN
c
l
as
s
i
f
i
er
i
s
ex
p
l
ai
ne
d
.
T
ab
l
e
1
s
ho
w
s
the
pro
po
s
ed
s
y
s
t
em
c
o
m
pa
r
a
ti
v
e
an
al
y
s
i
s
w
i
th
oth
er
m
eth
od
s
i
n
the
l
i
ter
atu
r
e
i
n
term
s
of
c
l
as
s
i
f
i
c
ati
on
ac
c
urac
y
.
A
l
s
o,
the
pro
po
s
ed
s
y
s
tem
ou
tpe
r
f
or
m
s
al
l
f
or
ms
of
w
a
v
el
et
tr
an
s
f
or
m
s
i
nc
l
ud
i
n
g T
S
W
T
, P
S
W
T
an
d Gabor tr
a
ns
f
orm
s
.
T
ab
l
e 1
C
om
pa
r
ati
v
e a
n
al
y
s
i
s
Tec
h
n
iqu
e
s
C
las
s
i
f
ier
A
c
c
u
r
a
c
y
(
%
)
Ga
b
o
r
w
it
h
GL
C
M
[
7
]
4
8
.
9
9
5
C
u
r
v
e
let
Tr
a
n
s
f
o
r
m
[
2
]
8
7
.
1
C
L
B
P
[
8
]
8
9
.
4
L
ine
a
r
R
e
g
r
e
s
s
ion
M
o
d
e
l
[
6
]
9
7
.
1
5
1
P
r
o
p
o
s
e
d
m
e
t
h
o
d
98
4.
Co
n
clus
ion
In
th
i
s
s
c
he
m
e,
an
ef
f
i
c
i
en
t
an
d
no
v
e
l
tex
tur
e
c
l
as
s
i
f
i
c
ati
o
n
s
y
s
tem
i
s
prop
os
ed
f
or
bo
th
gra
y
an
d
c
ol
or
i
m
ag
es
.
A
t
the
f
i
r
s
t,
th
e
pro
po
s
ed
D
S
T
ba
s
ed
tex
ture
c
l
as
s
i
f
i
c
ati
on
i
s
tes
ted
on
gra
y
s
c
a
l
e
i
m
ag
es
b
y
us
i
n
g
40
B
r
od
at
z
tex
ture
i
m
ag
es
.
T
he
gra
y
s
c
al
e
i
m
ag
e
i
s
de
c
om
po
s
ed
us
i
ng
E
W
T
m
eth
od
.
C
l
as
s
i
f
i
c
ati
on
pa
r
t
i
s
do
n
e
us
i
ng
the
f
ea
t
ures
f
r
om
LB
P
.
K
-
NN
c
l
as
s
i
f
i
er
i
s
us
ed
f
or
c
l
as
s
i
f
i
c
ati
on
pu
r
p
os
e.
E
x
pe
r
i
m
en
tal
r
es
u
l
ts
s
ho
w
tha
t
th
e
propos
e
d
em
pi
r
i
c
al
ba
s
ed
gra
y
t
ex
ture
c
l
as
s
i
f
i
c
ati
o
n
s
y
s
t
em
ou
tpe
r
f
or
m
s
the
ex
i
s
ti
ng
tex
ture
c
l
as
s
i
f
i
c
ati
on
te
c
hn
i
q
ue
s
s
uc
h
as
wav
el
et
a
nd
G
a
bo
r
trans
f
or
m
.
Ref
er
en
ce
s
[1]
S.Ari
v
a
z
h
a
g
a
n
,
e
t
a
l
.
Te
x
tu
re
c
l
a
s
s
i
f
i
c
a
ti
o
n
u
s
i
n
g
Curv
e
l
e
t
Sta
ti
s
ti
c
a
l
a
n
d
Co
-
o
c
c
u
rre
n
c
e
Fe
a
tu
re
s
.
Pro
c
e
e
d
i
n
g
s
o
f
1
8
th
I
n
te
rn
a
ti
o
n
a
l
Co
n
fe
re
n
c
e
o
n
Pa
tt
e
rn
Re
c
o
g
n
i
ti
o
n
(I
CPR
2
0
0
6
).
2
0
0
6
.
[2]
B.V.
Ram
a
n
a
Re
d
d
y
,
e
t
a
l
.
Te
x
tu
r
e
Cla
s
s
i
fi
c
a
ti
o
n
M
e
th
o
d
u
s
i
n
g
Wa
v
e
l
e
t
Tra
n
s
fo
rm
s
Ba
s
e
d
o
n
G
a
u
s
s
i
a
n
M
a
rk
o
v
Ra
n
d
o
m
F
i
e
l
d
.
In
te
rn
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
Si
g
n
a
l
a
n
d
I
m
a
g
e
Pro
c
e
s
s
i
n
g
.
2
0
1
0
;(
1
)
.1
.
[3]
P.S.
Hir
e
m
a
th
,
S.S
h
i
v
a
s
h
a
n
k
a
r
.
W
a
v
e
l
e
t
Ba
s
e
d
Fe
a
t
u
re
s
fo
r
T
e
x
tu
re
Cla
s
s
i
fi
c
a
t
i
o
n
.
J
o
u
rn
a
l
o
n
G
ra
p
h
i
c
s
,
Vi
s
i
o
n
a
n
d
I
m
a
g
e
Pr
o
c
e
s
s
i
n
g
.
De
c
e
m
b
e
r,
2
0
0
6
;
6
(
3
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N:
25
02
-
4
75
2
IJE
E
CS
V
ol
.
8
,
N
o.
3
,
Dec
em
be
r
2017
:
623
–
6
26
626
[4]
Zh
i
-
Zh
o
n
g
W
a
n
g
,
J
u
n
-
Hai
Y
o
n
g
.
T
e
x
tu
re
An
a
l
y
s
i
s
a
n
d
Cla
s
s
i
f
i
c
a
ti
o
n
w
i
th
L
i
n
e
a
r
Reg
re
s
s
i
o
n
M
o
d
e
l
Ba
s
e
d
o
n
W
a
v
e
l
e
t
T
ra
n
s
fo
r
m
.
IEEE
tra
n
s
a
c
ti
o
n
s
o
n
i
m
a
g
e
p
ro
c
e
s
s
i
n
g
.
Au
g
u
s
t
2
0
0
8
;
17
(
8
):
1
4
2
1
-
1430.
[5]
D.
A.
Cla
u
s
i
,
H.
Den
g
,
De
s
i
g
n
-
b
a
s
e
d
te
x
tu
re
fe
a
t
u
re
fu
s
i
o
n
u
s
i
n
g
G
a
b
o
r
fi
l
te
r
s
a
n
d
c
o
-
o
c
c
u
rre
n
c
e
p
ro
b
a
b
i
l
i
ti
e
s
.
IEEE
T
ra
n
s
a
c
ti
o
n
s
o
n
I
m
a
g
e
Pr
o
c
e
s
s
i
n
g
.
J
u
l
y
2
0
0
5
;
14
(
7
)
:
9
2
5
–
9
3
6
.
[6]
Rafi
M
d
.
Naj
m
u
s
S
a
d
a
t
,
Sh
y
h
W
e
i
T
e
n
g
,
T
e
x
tu
re
Cla
s
s
i
f
i
c
a
ti
o
n
U
s
i
n
g
M
u
l
ti
m
o
d
a
l
In
v
a
ri
a
n
t
L
o
c
a
l
Bi
n
a
ry
Pa
tt
e
r
n
.
IEEE
Wo
rk
s
h
o
p
o
n
Ap
p
l
i
c
a
t
i
o
n
s
o
f
Co
m
p
u
te
r
Vi
s
i
o
n
(
W
ACV)
.
2
0
1
1
.
3
1
5
-
3
2
0
.
[7]
Don
g
,
e
t
a
l
.
W
a
v
e
l
e
t
-
b
a
s
e
d
i
m
a
g
e
te
x
tu
re
c
l
a
s
s
i
fi
c
a
t
i
o
n
u
s
i
n
g
l
o
c
a
l
e
n
e
rg
y
h
i
s
to
g
ra
m
s
.
Si
g
n
a
l
Pro
c
e
s
s
i
n
g
L
e
tt
e
r
s
,
IEEE
.
2
0
1
1
;
1
8
(
4
):
2
4
7
-
2
5
0
.
[8]
Pa
n
d
a
,
S
.S.
,
J
e
n
a
,
G
.
a
n
d
Sa
h
u
,
S.K.
,
2
0
1
5
.
I
m
a
g
e
Su
p
e
r
Res
o
l
u
ti
o
n
Rec
o
n
s
tru
c
ti
o
n
Us
i
n
g
It
e
r
a
ti
v
e
A
d
a
p
ti
v
e
Re
g
u
l
a
ri
z
a
ti
o
n
M
e
th
o
d
a
n
d
G
e
n
e
t
i
c
A
l
g
o
r
i
th
m
.
In
Com
p
u
ta
ti
o
n
a
l
In
t
e
l
l
i
g
e
n
c
e
i
n
D
a
ta
M
i
n
i
n
g
.
Sp
ri
n
g
e
r
I
n
d
i
a
.
2
:6
7
5
-
6
8
1
.
[9]
Ab
d
u
l
k
a
d
i
r
S.
Col
o
r
T
e
x
tu
re
Cla
s
s
i
fi
c
a
ti
o
n
U
s
i
n
g
W
a
v
e
l
e
t
T
ra
n
s
fo
r
m
a
n
d
N
e
u
ra
l
Net
w
o
rk
E
n
s
e
m
b
l
e
s
.
Th
e
Ara
b
i
a
n
J
o
u
rn
a
l
f
o
r S
c
i
e
n
c
e
a
n
d
En
g
i
n
e
e
ri
n
g
.
O
c
t
o
b
e
r
2
0
0
9
;
3
4
(
2B
):
492
-
502
.
Evaluation Warning : The document was created with Spire.PDF for Python.