I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
201
9
,
p
p
.
5
2
5
3
~5
2
6
2
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v9
i
6
.
p
p
5
2
5
3
-
5262
5253
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e
.
co
m/
jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JE
C
E
O
pti
m
i
z
a
tion o
f
d
iscrete
w
a
v
elet
tr
a
nsfo
r
m
featur
es
using
a
rtif
icia
l be
e colo
ny
alg
o
rith
m
for t
ex
ture
i
m
a
g
e clas
sifica
tion
F
t
hi M
.
Albk
o
s
h
,
M
uh
a
m
m
a
d Suzuri H
it
a
m
,
Wa
n Nura
l
J
a
wa
hir H
j
Wa
n Yus
s
o
f
,
Abdu
l A
z
iz
K
.
Abdu
l H
a
m
i
d
,
Ro
zniza
Ali
S
c
h
o
o
l
o
f
In
f
o
rm
a
ti
c
s a
n
d
A
p
p
li
e
d
M
a
th
e
m
a
ti
c
s,
Un
iv
e
siti
M
a
ly
is
a
T
e
re
n
g
g
a
n
u
,
Ku
a
la Ne
ru
s,
T
e
re
n
g
g
a
n
u
,
M
a
lay
sia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
No
v
1
5
,
2
0
1
8
R
ev
i
s
ed
A
p
r
18
,
2
0
1
9
A
cc
ep
ted
J
u
n
1
0
,
2
0
1
9
S
e
lec
ti
o
n
o
f
a
p
p
ro
p
r
iate
ima
g
e
te
x
tu
re
p
ro
p
e
rti
e
s i
s o
n
e
o
f
th
e
m
a
j
o
r
issu
e
s in
tex
tu
re
c
las
si
f
ica
ti
o
n
.
T
h
is
p
a
p
e
r
p
re
se
n
ts
a
n
o
p
ti
m
iza
ti
o
n
tec
h
n
iq
u
e
f
o
r
a
u
to
m
a
ti
c
se
le
c
ti
o
n
o
f
m
u
lt
i
-
sc
a
l
e
d
isc
re
te
w
a
v
e
let
tran
sf
o
r
m
fe
a
t
u
re
s
u
sin
g
a
rti
f
icia
l
b
e
e
c
o
lo
n
y
a
lg
o
rit
h
m
f
o
r
ro
b
u
st
tex
tu
re
c
las
sif
ica
ti
o
n
p
e
rf
o
r
m
a
n
c
e
.
In
th
is p
a
p
e
r,
a
n
a
rti
f
icia
l
b
e
e
c
o
lo
n
y
a
lg
o
rit
h
m
h
a
s
b
e
e
n
u
se
d
to
f
i
n
d
t
h
e
b
e
st
c
o
m
b
in
a
ti
o
n
o
f
w
a
v
e
let
f
il
ters
w
it
h
th
e
c
o
rre
c
t
n
u
m
b
e
r
o
f
d
e
c
o
m
p
o
siti
o
n
lev
e
l
in
th
e
d
isc
re
te
wa
v
e
l
e
t
tran
sf
o
r
m
.
T
h
e
m
u
lt
i
-
la
y
e
re
d
p
e
rc
e
p
tro
n
n
e
u
ra
l
n
e
tw
o
rk
is
e
m
p
lo
y
e
d
a
s
a
n
i
m
a
g
e
tex
tu
re
c
las
sif
i
e
r.
T
h
e
p
ro
p
o
s
e
d
m
e
th
o
d
tes
ted
o
n
a
h
ig
h
-
re
so
lu
t
io
n
d
a
tab
a
se
o
f
UMD
te
x
tu
re
.
T
h
e
tex
tu
re
c
las
si
f
ica
ti
o
n
re
su
lt
s
sh
o
w
th
a
t
th
e
p
r
o
p
o
se
d
m
e
th
o
d
c
o
u
l
d
p
ro
v
id
e
a
n
a
u
to
m
a
ted
a
p
p
ro
a
c
h
f
o
r
f
in
d
in
g
t
h
e
b
e
st i
n
p
u
t
p
a
ra
m
e
ters
c
o
m
b
in
a
ti
o
n
s
e
tt
i
n
g
f
o
r
d
isc
re
te
w
a
v
e
let
tran
sf
o
r
m
fe
a
tu
re
s
th
a
t
lea
d
to
t
h
e
b
e
st
c
la
ss
if
ic
a
ti
o
n
a
c
c
u
ra
c
y
p
e
r
f
o
r
m
a
n
c
e
.
K
ey
w
o
r
d
s
:
A
r
ti
f
icial
b
ee
co
lo
n
y
a
lg
o
r
it
h
m
Dis
cr
ete
w
av
ele
t tr
an
s
f
o
r
m
m
u
lti
-
la
y
er
p
er
ce
p
tr
o
n
n
eu
r
al
n
et
w
o
r
k
T
ex
tu
r
e
class
i
f
icatio
n
Co
p
y
rig
h
t
©
201
9
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Mu
h
a
m
m
ad
S
u
zu
r
i H
ita
m
,
Sch
o
o
l o
f
I
n
f
o
r
m
atic
s
an
d
A
p
p
lied
Ma
th
e
m
at
ics
,
Un
i
v
esit
i M
al
y
i
s
a
T
er
en
g
g
a
n
u
,
Un
i
v
er
s
iti Ma
la
y
s
ia,
2
1
0
3
0
Ku
ala
T
er
en
g
g
a
n
u
,
T
er
en
g
g
a
n
u
,
M
a
lay
sia
.
E
m
ail:
s
u
z
u
r
i@
u
m
t.e
d
u
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
T
ex
tu
r
e
p
atter
n
p
lay
s
an
i
m
p
o
r
tan
t
r
o
le
in
i
m
ag
e
a
n
al
y
s
i
s
an
d
p
atter
n
r
ec
o
g
n
itio
n
.
Du
e
to
its
w
id
e
v
ar
iatio
n
o
f
p
atter
n
,
t
h
e
d
e
v
e
lo
p
m
e
n
t
o
f
a
n
au
to
m
ated
i
m
ag
e
te
x
t
u
r
e
a
n
al
y
s
is
f
ac
es
s
o
m
e
c
h
alle
n
g
es
[
1
]
.
I
n
g
e
n
er
al,
th
e
i
m
ag
e
te
x
t
u
r
e
an
al
y
s
is
co
m
p
o
s
ed
o
f
t
h
r
ee
m
a
in
i
s
s
u
es;
te
x
t
u
r
e
cla
s
s
i
f
icatio
n
,
tex
t
u
r
e
s
eg
m
e
n
tatio
n
an
d
s
h
ap
e
r
ec
o
v
er
y
f
r
o
m
te
x
t
u
r
e
[
2
]
.
T
ex
tu
r
e
c
lass
i
f
icatio
n
i
s
b
asical
l
y
d
ea
lt
w
it
h
claa
s
i
f
y
i
n
g
an
i
m
a
g
e
in
to
an
y
g
iv
e
n
s
et
o
f
t
ex
tu
r
e
cla
s
s
es
b
ased
o
n
te
x
tu
r
e
p
atter
n
p
r
esen
ted
in
th
e
g
i
v
en
i
m
a
g
e.
T
ex
tu
r
e
class
i
f
icatio
n
h
a
s
d
iv
er
s
e
ap
p
licatio
n
ar
ea
in
cl
u
d
in
g
r
e
m
o
te
s
en
s
in
g
,
m
ed
ical
i
m
a
g
e
d
iag
n
o
s
is
an
d
s
to
n
e
t
y
p
e
ap
p
licatio
n
[
3
]
.
T
ex
tu
r
e
an
al
y
s
i
s
an
d
class
i
f
ica
tio
n
tech
n
i
q
u
e
ca
n
b
e
ca
teg
o
r
ized
as
s
tatis
tical
ap
p
r
o
ac
h
es,
s
ig
n
al
-
p
r
o
ce
s
s
i
n
g
b
ased
ap
p
r
o
ac
h
es
an
d
s
tr
u
ct
u
r
al
ap
p
r
o
a
ch
es.
E
v
e
n
th
o
u
g
h
s
tati
s
tical
m
et
h
o
d
s
is
o
n
e
o
f
th
e
m
o
s
t
p
o
p
u
lar
m
et
h
o
d
s
f
o
r
tex
t
u
r
e
an
al
y
s
is
a
n
d
clas
s
if
i
ca
tio
n
,
th
e
y
h
a
v
e
a
m
aj
o
r
d
r
a
w
b
ac
k
to
ad
d
r
ess
th
e
s
p
atial
s
ca
le
p
r
o
b
le
m
[
1
]
.
Sig
n
al
p
r
o
ce
s
s
i
n
g
b
ased
m
et
h
o
d
co
u
ld
d
ea
l
w
i
th
t
h
i
s
p
r
o
b
lem
i
n
a
m
u
c
h
b
etter
f
o
r
m
an
d
t
h
e
w
a
v
ele
ts
tr
an
s
f
o
r
m
(
W
T
)
is
o
n
e
o
f
th
ese
m
eth
o
d
s
.
W
T
ca
n
b
e
ca
teg
o
r
ized
in
to
d
is
cr
ete
an
d
co
n
tin
u
o
u
s
f
o
r
m
.
I
n
co
n
ti
n
u
o
u
s
w
a
v
elet
tr
a
n
s
f
o
r
m
(
C
W
T
)
,
th
e
w
av
ele
t
tr
an
s
f
o
r
m
a
tio
n
c
h
a
n
g
e
s
co
n
ti
n
u
o
u
s
l
y
.
W
h
er
ea
s
,
th
e
d
is
cr
ete
w
a
v
el
et
tr
an
s
f
o
r
m
(
DW
T
)
d
ec
o
m
p
o
s
es
th
e
s
i
g
n
a
l
in
to
m
u
ltire
s
o
lu
tio
n
co
ef
f
icie
n
ts
u
s
i
n
g
h
i
g
h
p
ass
a
n
d
lo
w
p
ass
f
il
ter
s
[
4
]
.
T
h
e
m
ai
n
ad
v
an
ta
g
es
o
f
w
a
v
elets
ar
e
th
at
ar
e
ad
ap
ted
to
th
e
tr
an
s
ie
n
ts
o
f
ea
ch
s
ca
le
a
n
d
p
r
o
v
id
e
a
m
u
lti
-
r
eso
lu
t
io
n
an
d
n
o
n
-
r
ed
u
n
d
an
t
r
ep
r
ese
n
t
atio
n
o
f
s
ig
n
al
s
[
5
]
.
T
h
er
e
ar
e
m
a
n
y
p
r
ev
io
u
s
r
es
ea
r
ch
w
o
r
k
s
r
ec
o
m
m
e
n
d
ed
to
s
o
lv
e
te
x
t
u
r
e
clas
s
i
f
icatio
n
p
r
o
b
lem
b
ased
o
n
DW
T
.
W
T
h
as
t
h
e
ab
ilit
y
to
d
ec
o
m
p
o
s
e
co
m
p
le
x
i
n
f
o
r
m
atio
n
in
to
ele
m
e
n
tar
y
f
o
r
m
s
a
n
d
ca
n
co
v
er
th
e
f
r
eq
u
e
n
c
y
d
o
m
ain
m
o
r
e
a
cc
u
r
ate
as
co
m
p
ar
ed
to
o
th
er
tr
an
s
f
o
r
m
m
et
h
o
d
s
s
u
c
h
as
G
ab
o
r
Fil
ter
in
g
(
GF)
ap
p
r
o
ac
h
[
6
,
7
]
.
Vid
y
a
et
a
l
.
[
8
]
co
m
p
ar
ed
th
e
p
er
f
o
r
m
an
ce
o
f
c
o
m
p
u
ter
a
id
ed
d
iag
n
o
s
is
ap
p
r
o
ac
h
u
s
i
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
2
5
3
-
5
2
6
2
5254
DW
T
f
ea
tu
r
es
,
g
r
e
y
lev
el
co
-
o
cc
u
r
r
en
ce
m
atr
i
x
(
G
L
C
M
)
an
d
h
i
g
h
er
-
o
r
d
er
s
p
ec
tr
a
(
HOS
)
tex
tu
r
e
d
escr
ip
to
r
s
w
h
ile
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM
)
is
u
s
ed
as
a
te
x
t
u
r
e
class
i
f
ier
.
T
h
eir
tex
tu
r
e
cl
ass
i
f
icatio
n
r
es
u
lt
s
s
h
o
w
ed
th
at
t
h
e
s
u
p
er
io
r
it
y
o
f
DW
T
b
ased
ap
p
r
o
ac
h
as c
o
m
p
ar
ed
to
o
th
er
s
.
T
h
e
m
ai
n
p
r
o
b
le
m
i
n
th
is
a
p
p
r
o
ac
h
is
t
h
at
t
h
e
r
esear
ch
e
r
s
h
a
v
e
to
m
a
n
u
all
y
f
i
n
d
o
u
t
th
e
b
es
t
co
m
b
i
n
atio
n
o
f
DW
T
p
ar
am
e
t
er
s
f
o
r
o
p
ti
m
u
m
i
m
ag
e
te
x
tu
r
e
class
if
icatio
n
r
esu
lts
,
i.e
.
w
a
v
elet
d
ec
o
m
p
o
s
itio
n
f
ilter
a
n
d
th
e
n
u
m
b
er
o
f
w
a
v
el
et
d
ec
o
m
p
o
s
i
tio
n
[
9
]
.
T
h
er
ef
o
r
e,
to
r
ed
u
ce
th
e
ted
io
u
s
a
n
d
ti
m
e
co
n
s
u
m
i
n
g
tr
ial
an
d
o
b
s
er
v
atio
n
m
eth
o
d
o
f
f
i
n
d
in
g
t
h
e
o
p
ti
m
u
m
s
etti
n
g
,
t
h
is
p
ap
er
in
tr
o
d
u
ce
s
a
n
a
u
to
m
ated
ap
p
r
o
ac
h
f
o
r
b
in
ar
y
te
x
t
u
r
e
clas
s
i
f
icatio
n
wh
er
e
th
e
ar
tific
ial
b
ee
co
lo
n
y
(
A
B
C
)
al
g
o
r
ith
m
[
1
0
]
is
e
m
p
lo
y
ed
to
f
i
n
d
t
h
e
b
est
co
m
b
i
n
atio
n
o
f
w
a
v
elet
f
ilter
s
an
d
d
ec
o
m
p
o
s
it
io
n
le
v
els
f
o
r
a
g
iv
e
n
p
r
o
b
le
m
.
2.
RE
L
AT
E
D
WO
RK
S
I
n
s
i
g
n
al
p
r
o
ce
s
s
i
n
g
b
ased
te
x
tu
r
e
a
n
al
y
s
i
s
m
e
th
o
d
s
,
t
h
e
i
m
ag
e
is
t
y
p
icall
y
f
i
lter
ed
w
i
t
h
a
b
an
k
o
f
f
ilter
s
o
f
d
if
f
er
i
n
g
s
ca
le
s
a
n
d
o
r
ien
tatio
n
s
i
n
o
r
d
er
to
ca
p
tu
r
e
th
e
ch
a
n
g
es
b
et
w
ee
n
s
p
ec
i
f
i
c
f
r
eq
u
e
n
c
y
b
an
d
s
i
n
th
e
an
al
y
ze
d
i
m
a
g
e
[1
1
]
.
T
h
e
w
av
e
let
tr
an
s
f
o
r
m
(
W
T
)
is
p
er
f
o
r
m
ed
o
n
th
e
f
r
eq
u
e
n
c
y
d
o
m
ai
n
o
f
th
e
i
m
a
g
e
s
th
at
ch
ar
ac
ter
izes
m
u
l
tis
ca
le
f
r
eq
u
en
c
y
co
n
te
n
t
ca
lled
w
a
v
elet
co
ef
f
icie
n
ts
at
ea
ch
s
p
atial
lo
ca
tio
n
o
f
an
i
m
a
g
e.
T
h
e
d
is
cr
im
in
at
iv
e
p
o
w
er
o
f
te
x
tu
r
e
d
escr
ip
to
r
s
ca
n
b
e
s
ig
n
i
f
ica
n
tl
y
i
m
p
r
o
v
ed
if
d
if
f
er
e
n
t
s
ca
le
s
ar
e
co
n
s
id
er
ed
a
m
o
n
g
th
e
i
m
ag
e
s
d
u
r
in
g
th
e
d
escr
ip
to
r
ex
tr
ac
tio
n
[
1
]
.
T
h
e
b
asic
id
ea
o
f
m
u
lti
-
r
eso
lu
tio
n
a
n
al
y
s
i
s
is
to
r
ep
r
esen
t
a
n
i
m
ag
e
o
n
s
ev
er
al
s
u
b
-
i
m
ag
e
s
,
f
r
o
m
c
o
ar
s
e
to
f
in
e
r
eso
l
u
tio
n
,
an
d
an
al
y
ze
th
e
m
i
n
th
e
r
esp
ec
ti
v
e
d
o
m
ain
.
DW
T
h
av
e
b
ee
n
s
u
cc
e
s
s
f
u
ll
y
i
m
p
le
m
e
n
ted
i
n
m
a
n
y
ap
p
licatio
n
s
i
n
te
x
tu
r
e
s
class
i
f
icatio
n
.
A
w
av
ele
t
-
b
as
ed
s
o
il
te
x
tu
r
e
f
ea
t
u
r
e
e
x
tr
a
c
tio
n
m
et
h
o
d
h
ad
b
ee
n
in
tr
o
d
u
ce
d
b
y
Z
h
a
n
g
et
al.
[
1
2
]
an
d
s
u
cc
e
s
s
f
u
ll
y
ap
p
lied
to
s
o
il
tex
t
u
r
e
cla
s
s
i
f
ica
tio
n
s
y
s
te
m
.
A
m
ax
i
m
u
m
li
k
e
lih
o
o
d
class
i
f
ier
i
s
u
s
ed
f
o
r
clas
s
i
f
icatio
n
,
w
h
ic
h
s
h
o
w
s
g
o
o
d
r
es
u
lts
b
ased
o
n
t
h
e
w
a
v
el
e
t
te
x
tu
r
e
f
ea
t
u
r
es.
R
aj
p
o
o
t
an
d
R
aj
p
o
o
t
[
1
3
]
p
r
o
d
u
ce
d
p
r
o
m
i
s
in
g
r
esu
lts
o
f
te
x
t
u
r
e
class
i
f
i
ca
tio
n
b
y
u
s
in
g
Dau
b
ec
h
ies
f
ilter
s
an
d
a
3
lev
el
DW
T
d
ec
o
m
p
o
s
itio
n
w
it
h
SV
M
class
i
f
ier
.
Sid
h
u
a
n
d
R
aa
h
e
m
i
f
ar
[
1
4
]
in
v
es
tig
a
ted
th
e
w
a
v
elet
tr
an
s
f
o
r
m
a
n
d
SVM
clas
s
i
f
ier
f
o
r
te
x
t
u
r
e
cla
s
s
i
f
icatio
n
.
B
y
ap
p
l
y
i
n
g
th
e
p
r
o
p
o
s
ed
m
et
h
o
d
,
th
e
y
o
b
tain
ed
g
o
o
d
class
i
f
icatio
n
r
esu
lt
s
b
ased
o
n
t
h
e
B
r
o
d
atz
tex
t
u
r
e
d
at
ab
ase.
Si
n
g
h
et
al.
[
1
5
]
u
s
ed
a
Haa
r
w
av
elet
tr
an
s
f
o
r
m
a
n
d
m
u
lti
-
la
y
er
ed
p
er
ce
p
tr
o
n
n
eu
r
al
n
et
w
o
r
k
(
M
L
P
NN)
f
o
r
tex
t
u
r
e
i
m
ag
e
clas
s
if
icatio
n
.
T
h
e
ex
p
er
i
m
en
t
r
esu
lts
s
h
o
w
ed
th
at
t
h
e
ac
cu
r
ac
y
o
f
cla
s
s
i
f
ica
tio
n
is
i
n
t
h
e
r
an
g
e
b
et
w
ee
n
8
6
.
2
%
to
9
9
.
0
6
%
b
ased
o
n
th
e
s
elec
ted
s
a
m
p
le
s
.
J
ay
a
s
u
d
h
a
an
d
P
u
g
az
h
e
n
th
i
[
1
6
]
ap
p
ly
i
n
g
DW
T
f
o
r
co
lo
r
tex
tu
r
e
s
clas
s
if
icatio
n
u
s
i
n
g
K
-
n
ea
r
est
n
e
ig
h
b
o
r
(
KNN)
class
i
f
ie
r
.
T
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
ac
h
iev
ed
8
5
.
8
3
%
class
i
f
icat
io
n
ac
cu
r
ac
y
at
th
e
f
ir
s
t
lev
el
o
f
d
ec
o
m
p
o
s
itio
n
u
s
in
g
Vi
s
T
ex
d
atab
ase.
Am
o
n
g
t
h
e
m
ai
n
p
ar
a
m
eter
s
th
at
a
f
f
ec
t
DW
T
class
i
f
icatio
n
p
er
f
o
r
m
a
n
ce
s
ar
e
t
h
e
t
y
p
e
o
f
w
av
e
let
d
ec
o
m
p
o
s
itio
n
f
i
lter
an
d
th
e
n
u
m
b
er
o
f
w
av
e
let
d
ec
o
m
p
o
s
itio
n
[9
,
17
]
.
Mo
j
s
ilo
v
ić
et
al.
[
1
7
]
in
v
es
tig
a
ted
w
h
et
h
er
th
e
p
r
o
p
er
ties
o
f
d
ec
o
m
p
o
s
i
tio
n
w
a
v
elet
f
ilter
s
p
lay
a
n
i
m
p
o
r
ta
n
t
r
o
le
in
tex
t
u
r
e
d
escr
ip
tio
n
.
T
h
ey
p
er
f
o
r
m
ed
clas
s
i
f
icatio
n
ex
p
er
i
m
e
n
t
s
u
s
i
n
g
B
r
o
d
atz
tex
tu
r
e
s
.
T
h
e
ex
p
e
r
i
m
en
ta
l
r
es
u
lts
i
n
d
icate
d
th
at
th
e
s
elec
tio
n
o
f
t
h
e
d
ec
o
m
p
o
s
itio
n
f
ilter
h
as
a
s
ig
n
i
f
ica
n
t
i
n
f
l
u
en
ce
o
n
t
h
e
r
esu
lt
o
f
te
x
t
u
r
e
ch
ar
ac
ter
izatio
n
.
B
u
s
ch
an
d
B
o
les
[
1
8
]
co
m
p
ar
ed
th
e
r
esp
o
n
s
e
b
o
t
h
t
h
e
Haa
r
an
d
B
io
r
th
o
g
o
n
al
w
av
ele
ts
u
s
in
g
B
r
o
d
atz
alb
u
m
in
te
x
t
u
r
es
c
lass
if
ica
tio
n
.
T
h
e
ex
p
er
i
m
e
n
tal
r
e
s
u
l
ts
i
n
d
ica
ted
th
at
b
o
th
w
a
v
elet
s
p
r
o
d
u
ce
a
s
i
m
ilar
r
esp
o
n
s
e
i
n
m
o
s
t
ca
s
e
s
at
t
h
e
f
ir
s
t
le
v
el
o
f
d
etail,
h
o
w
e
v
er
,
ar
e
g
en
er
all
y
d
if
f
er
en
t
a
t
th
e
s
ec
o
n
d
lev
e
l
.
T
h
e
r
esu
lts
s
h
o
w
th
at
t
h
e
c
lass
if
ica
tio
n
ac
c
u
r
a
cies
ar
e
8
8
%
a
n
d
8
7
%
b
y
H
aa
r
an
d
B
io
r
th
o
g
o
n
al
w
a
v
ele
ts
,
r
esp
e
cti
v
el
y
an
d
p
r
o
d
u
cin
g
9
5
%
class
i
f
icatio
n
ac
cu
r
ac
y
b
y
t
h
e
co
m
b
in
at
io
n
o
f
Haa
r
an
d
B
io
r
th
o
g
o
n
al
w
a
v
elets.
Hir
e
m
at
h
an
d
Sh
i
v
as
h
a
n
k
ar
[
9
]
in
v
esti
g
ated
th
e
ef
f
ec
t
o
f
u
s
i
n
g
d
if
f
er
e
n
t
f
ilter
s
o
n
th
e
tex
t
u
r
es
clas
s
i
f
ic
atio
n
u
s
i
n
g
B
r
o
d
atz
tex
t
u
r
es
an
d
d
if
f
er
en
t
w
av
ele
t
f
ilter
b
an
k
s
i.e
.
,
Haa
r
,
Dau
b
ec
h
ies
(
d
b
2
,
d
b
3
,
d
b
4
,
d
b
6
,
d
b
8
,
d
b
1
0
,
d
b
1
6
)
,
b
io
r
th
o
g
o
n
al
(
b
io
r
2
.
2
,
b
io
r
3
.
3
,
b
io
r
4
.
4
,
b
i
o
r
5
.
5
)
w
a
v
elet
f
ilte
r
s
an
d
also
th
e
Gab
o
r
f
ea
t
u
r
es
.
T
h
e
ex
p
er
im
e
n
tal
r
esu
lt
s
d
e
m
o
n
s
tr
ated
th
at
t
h
e
b
est
cla
s
s
i
f
icatio
n
r
es
u
lt
w
a
s
at
9
6
.
8
4
%
clas
s
i
f
icatio
n
ac
cu
r
ac
y
ac
h
ie
v
ed
b
y
Haa
r
w
av
e
let.
B
ased
o
n
th
ese
r
ev
ie
w
s
o
f
DW
T
p
ar
am
e
ter
s
s
elec
tio
n
a
n
d
its
i
m
p
o
r
tan
ce
t
h
at
af
f
ec
t
t
h
e
to
tal
class
i
f
icatio
n
ac
cu
r
ac
y
,
it
ca
n
b
e
co
n
clu
d
ed
th
at
s
o
m
e
s
p
ec
i
f
ic
w
a
v
elet
p
ar
a
m
eter
s
n
ee
d
to
b
e
o
p
tim
al
l
y
tu
n
ed
to
ac
h
iev
e
t
h
e
b
est cla
s
s
i
f
icati
o
n
p
er
f
o
r
m
an
ce
.
Op
ti
m
izatio
n
tec
h
n
iq
u
e
is
a
wa
y
to
f
i
n
d
t
h
e
m
o
s
t
co
s
t
e
f
f
ec
t
iv
e,
h
i
g
h
e
s
t
p
er
f
o
r
m
an
ce
s
y
s
te
m
u
n
d
er
a
g
iv
e
n
co
n
s
tr
ai
n
ts
.
B
y
m
a
x
i
m
i
zin
g
d
esire
d
f
ac
to
r
s
an
d
m
i
n
i
m
izi
n
g
t
h
e
u
n
d
esire
d
o
n
e
s
.
Ar
tif
icial
b
ee
co
lo
n
y
alg
o
r
ith
m
i
s
a
s
to
ch
a
s
tic,
n
a
t
u
r
e
-
i
n
s
p
ir
ed
,
s
w
ar
m
in
telli
g
e
n
t
al
g
o
r
ith
m
p
r
o
p
o
s
ed
b
y
Ka
r
ab
o
g
a
[
1
9
]
w
h
ic
h
m
i
m
ics
t
h
e
f
o
r
ag
in
g
b
eh
a
v
io
r
o
f
h
o
n
e
y
b
ee
s
an
d
u
s
ed
f
o
r
s
o
lv
in
g
o
p
ti
m
iza
tio
n
p
r
o
b
le
m
s
[
2
0
]
.
A
B
C
alg
o
r
ith
m
is
a
m
o
r
e
r
ec
en
t
g
lo
b
al
o
p
ti
m
i
za
tio
n
alg
o
r
it
h
m
t
h
at
h
as
b
ee
n
s
u
cc
ess
f
u
ll
y
e
m
p
lo
y
ed
to
s
o
lv
e
v
a
r
io
u
s
t
y
p
e
o
f
o
p
tim
izatio
n
p
r
o
b
le
m
s
in
v
ar
io
u
s
f
ie
ld
o
f
s
tu
d
y
.
A
ta
s
ev
er
et
al.
[
1
0
]
co
m
p
ar
ed
th
e
p
er
f
o
r
m
a
n
ce
o
f
a
f
e
w
ev
o
lu
tio
n
ar
y
co
m
p
u
tatio
n
alg
o
r
i
th
m
s
an
d
f
o
u
n
d
o
u
t
th
a
t
th
e
A
B
C
al
g
o
r
ith
m
s
p
er
f
o
r
m
b
etter
th
an
o
th
e
r
ar
tif
icial
i
n
telli
g
e
n
ce
to
o
ls
.
Sh
a
n
t
h
i
an
d
B
h
as
k
ar
an
[
2
1
]
co
m
p
ar
ed
th
e
p
er
f
o
r
m
a
n
ce
o
f
m
o
d
if
ied
A
B
C
alg
o
r
ith
m
w
it
h
p
ar
ticle
s
w
ar
m
o
p
ti
m
izatio
n
(
P
SO)
a
n
d
g
e
n
et
ic
alg
o
r
it
h
m
(
G
A
)
f
o
r
f
ea
t
u
r
e
s
elec
tio
n
tech
n
iq
u
e
to
s
elec
t
th
e
p
r
ed
o
m
i
n
an
t
f
ea
t
u
r
e
s
et
o
f
b
r
ea
s
t
lesi
o
n
in
m
a
m
m
o
g
r
a
m
i
m
ag
e
s
.
T
h
e
y
r
ep
o
r
ted
th
at
th
e
A
B
C
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:
2
0
8
8
-
8708
Op
timiz
a
tio
n
o
f d
is
crete
w
a
ve
l
et
tr
a
n
s
fo
r
m
fea
tu
r
es u
s
in
g
a
r
t
ificia
l b
ee
co
lo
n
y
a
lg
o
r
ith
m…
(
F
th
i M.
A
lb
ko
s
h
)
5255
alg
o
r
ith
m
m
a
n
ag
ed
to
s
elec
t
o
n
l
y
4
2
f
ea
tu
r
e
s
as
co
m
p
ar
ed
to
5
0
an
d
5
6
f
ea
tu
r
es
o
b
tain
e
d
b
y
G
A
a
n
d
P
SO,
r
esp
ec
tiv
el
y
w
h
ile
m
ai
n
tai
n
s
t
h
e
h
i
g
h
ac
cu
r
ac
y
o
f
clas
s
i
f
icat
io
n
.
L
i
a
n
d
L
i
[
2
2
]
ap
p
lied
m
u
lti
-
o
b
j
ec
tiv
e
A
B
C
alg
o
r
ith
m
to
ca
m
el
lia
f
r
u
it
r
ec
o
g
n
itio
n
s
y
s
te
m
in
a
f
r
u
it
p
ick
i
n
g
r
o
b
o
t.
T
h
eir
a
p
p
r
o
ac
h
u
s
ed
eig
h
t
ch
ar
ac
ter
is
tic
p
ar
a
m
eter
s
o
f
th
e
co
lo
r
f
ea
tu
r
e,
th
e
m
o
r
p
h
o
lo
g
y
f
ea
t
u
r
e
an
d
th
e
tex
tu
r
e
f
ea
tu
r
e
as
th
e
p
r
ef
er
en
ce
r
eg
io
n
to
ca
m
e
llia
f
r
u
it
r
ec
o
g
n
itio
n
.
I
n
a
n
o
th
er
r
esear
ch
,
B
an
er
j
ee
et
al.
[
2
3
]
p
r
o
p
o
s
ed
t
o
u
s
e
A
B
C
al
g
o
r
ith
m
to
s
o
l
v
e
th
e
la
n
d
-
co
v
er
p
r
o
b
le
m
b
y
i
m
a
g
e
cl
ass
i
f
icatio
n
.
B
ased
o
n
th
e
r
ev
ie
w
o
f
t
h
e
AB
C
alg
o
r
ith
m
a
n
d
its
s
u
cc
ess
to
p
r
o
d
u
ce
im
p
r
es
s
iv
e
r
e
s
u
l
ts
in
s
o
m
e
o
f
th
e
ap
p
licatio
n
s
as
w
e
ll
as
b
ased
o
n
th
e
li
m
itatio
n
s
o
f
t
h
e
DW
T
m
eth
o
d
,
th
is
p
ap
er
p
r
o
p
o
s
ed
to
u
s
e
A
B
C
alg
o
r
ith
m
to
o
p
ti
m
ize
t
h
e
s
ele
ctio
n
o
f
p
ar
a
m
eter
s
o
f
DW
T
f
o
r
b
in
ar
y
te
x
t
u
r
e
class
i
f
icat
io
n
.
3.
RE
S
E
ARCH
M
E
T
H
O
D
T
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
u
s
ed
in
th
is
s
t
u
d
y
co
n
s
is
t
s
o
f
t
h
r
ee
m
a
in
co
m
p
o
n
e
n
ts
;
f
ea
t
u
r
e
ex
tr
ac
tio
n
u
s
in
g
DW
T
m
et
h
o
d
,
class
if
y
i
n
g
te
x
tu
r
e
u
s
i
n
g
M
L
P
NN
a
n
d
A
B
C
alg
o
r
it
h
m
is
u
s
ed
to
a
u
to
m
ati
c
all
y
f
i
n
d
t
h
e
m
o
s
t
s
u
itab
le
co
m
b
in
at
io
n
o
f
w
a
v
el
et
f
il
ter
s
an
d
d
ec
o
m
p
o
s
i
tio
n
le
v
el
f
o
r
ac
h
ie
v
i
n
g
th
e
m
o
s
t
ac
c
u
r
ate
clas
s
i
f
icatio
n
r
esu
lt
s
.
3
.
1
.
Dis
cr
et
e
w
a
v
ele
t
t
ra
ns
f
o
r
m
DW
T
is
d
ec
o
m
p
o
s
ed
an
i
m
a
g
e
in
to
f
o
u
r
s
u
b
-
b
an
d
s
.
T
h
ese
s
u
b
-
b
an
d
s
ar
e
m
ar
k
ed
as
ap
p
r
o
x
i
m
at
io
n
co
ef
f
icie
n
t
(
L
L
)
w
h
ic
h
is
t
h
e
o
r
ig
in
a
l
i
m
a
g
e
at
lo
w
er
r
eso
lu
tio
n
a
n
d
th
r
ee
h
i
g
h
f
r
e
q
u
en
c
y
s
u
b
-
b
an
d
s
co
r
r
esp
o
n
d
in
g
to
h
o
r
izo
n
tal
(
HL
)
w
h
ic
h
h
i
g
h
li
g
h
t
s
t
h
e
h
o
r
i
zo
n
tal
ed
g
es
o
f
t
h
e
o
r
ig
i
n
al
i
m
ag
e,
v
er
tical
(
L
H)
w
h
ic
h
h
ig
h
li
g
h
ts
th
e
v
er
tical
ed
g
es
a
n
d
d
iag
o
n
al
d
etail
s
(
HH)
w
h
ic
h
h
i
g
h
l
ig
h
ts
t
h
e
d
iag
o
n
al
ed
g
es
[
2
4
]
.
On
ce
t
h
e
i
m
ag
e
is
d
i
v
id
ed
in
to
s
u
b
-
b
a
n
d
,
an
y
n
u
m
b
er
o
f
f
ea
t
u
r
es
ca
n
b
e
e
x
tr
ac
ted
f
r
o
m
t
h
e
tr
an
s
f
o
r
m
ed
i
m
a
g
e.
T
h
e
L
L
i
m
a
g
e
k
ee
p
s
m
o
s
t
d
etail
s
a
n
d
it
i
s
u
s
e
d
to
p
r
o
d
u
ce
th
e
f
o
llo
w
i
n
g
le
v
el
o
f
d
ec
o
m
p
o
s
itio
n
as
s
h
o
w
n
in
F
ig
u
r
e
1
.
Fig
u
r
e
1
.
W
av
elet
d
ec
o
m
p
o
s
it
io
n
o
f
an
i
m
a
g
e
T
h
e
f
o
llo
w
in
g
te
x
t
u
r
e
f
ea
tu
r
es
w
er
e
t
h
en
co
m
p
u
ted
f
o
r
d
ec
id
ed
r
eso
lu
ti
o
n
le
v
el
o
f
t
h
e
w
a
v
elet
d
etails;
m
ea
n
s
ta
n
d
ar
d
d
ev
iat
io
n
,
en
er
g
y
an
d
en
tr
o
p
y
.
T
h
e
m
ea
n
v
al
u
e
ca
n
b
e
m
at
h
e
m
atica
ll
y
d
ef
in
ed
as
f
o
llo
w
s
:
=
1
∑
∑
(
,
)
−
1
=
0
−
1
=
0
(
1
)
W
h
ile
th
e
s
tan
d
ar
d
d
ev
iatio
n
(
ST
D)
ca
n
b
e
co
m
p
u
ted
b
y
u
s
i
n
g
t
h
e
f
o
llo
w
in
g
eq
u
atio
n
:
STD
=
√
∑
∑
(
(
,
)
−
)
2
−
1
=
0
−
1
=
0
R
×
S
(
2
)
w
h
er
e
f(
r
,
s
)
is
t
h
e
v
al
u
e
o
f
t
h
e
p
ix
el
in
t
h
is
p
o
s
itio
n
f
o
r
an
y
s
u
b
-
b
an
d
o
f
s
ize
R
×
S
On
t
h
e
o
th
er
h
a
n
d
,
th
e
w
a
v
elet
en
tr
o
p
y
a
n
d
en
er
g
y
w
er
e
co
m
p
u
ted
u
s
i
n
g
th
e
f
o
llo
w
in
g
d
ef
i
n
itio
n
:
E
n
tr
opy
=
∑
∑
(
,
)
(
(
,
)
)
=
1
=
1
(
3
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
2
5
3
-
5
2
6
2
5256
w
h
er
e
M
an
d
N
ar
e
t
h
e
d
i
m
e
n
s
io
n
s
o
f
t
h
e
co
ef
f
icie
n
t
m
atr
ix
.
E
n
e
r
gy
=
∑
∑
(
,
)
2
−
1
=
0
−
1
=
0
(
4
)
an
d
N
is
t
h
e
w
id
th
o
f
th
e
c
h
a
n
n
el
an
d
i
an
d
j
ar
e
th
e
r
o
w
s
a
n
d
co
lu
m
n
s
o
f
t
h
e
ch
a
n
n
el
.
3
.
2
.
Art
if
ici
a
l neura
l net
w
o
r
k
cla
s
s
if
ier
T
h
e
Mu
lti
-
la
y
er
P
er
ce
p
tr
o
n
Neu
r
al
Net
w
o
r
k
(
M
L
P
NN)
is
p
er
h
ap
s
th
e
m
o
s
t
p
o
p
u
lar
n
eu
r
al
n
et
w
o
r
k
th
at
h
as
b
ee
n
u
s
ed
f
o
r
clas
s
i
f
icatio
n
r
eq
u
ir
e
m
en
ts
[
2
5
]
.
ML
P
NN
is
a
s
et
o
f
co
n
n
ec
ted
i
n
p
u
t/o
u
tp
u
t
u
n
its
in
w
h
ic
h
ea
ch
co
n
n
ec
tio
n
h
as
a
w
ei
g
h
t
as
s
o
ciate
d
w
it
h
it.
Du
r
in
g
t
h
e
lea
r
n
in
g
p
h
ase,
t
h
e
n
et
w
o
r
k
lear
n
s
b
y
ad
j
u
s
tin
g
th
e
w
eig
h
t
s
,
s
o
as
t
o
b
e
ab
le
to
p
r
e
d
ict
th
e
co
r
r
ec
t
class
lab
el
o
f
t
h
e
in
p
u
t
t
u
p
les.
Neu
r
al
n
et
w
o
r
k
lear
n
in
g
is
a
ls
o
r
e
f
er
r
ed
to
as
co
n
n
ec
t
io
n
is
t
lear
n
in
g
d
u
e
to
t
h
e
co
n
n
ec
tio
n
s
b
et
wee
n
u
n
its
.
I
t
g
iv
e
s
th
e
r
eq
u
ir
ed
o
u
tp
u
t
f
o
r
a
p
ar
ticu
lar
in
p
u
t.
3
.
3
.
K
-
f
o
ld
cr
o
s
s
v
a
lid
a
t
io
n
C
r
o
s
s
-
v
al
id
atio
n
i
s
a
s
ta
tis
t
ica
l
tech
n
iq
u
e
w
h
ic
h
i
n
v
o
lv
e
s
p
ar
titi
o
n
i
n
g
t
h
e
d
ata
i
n
to
s
u
b
s
et
s
,
tr
ain
i
n
g
th
e
d
ata
o
n
a
s
u
b
s
et
an
d
u
s
e
t
h
e
o
th
er
s
u
b
s
et
to
e
v
al
u
ate
t
h
e
m
o
d
el’
s
p
er
f
o
r
m
a
n
ce
.
O
n
e
o
f
t
h
e
co
m
m
o
n
t
y
p
e
o
f
cr
o
s
s
v
alid
atio
n
t
h
at
is
w
id
el
y
u
s
ed
i
s
k
-
f
o
ld
cr
o
s
s
v
alid
atio
n
tec
h
n
iq
u
e
[
2
6
]
.
I
n
a
k
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
alg
o
r
ith
m
,
a
s
a
m
p
le
s
e
t
i
s
d
iv
i
d
ed
in
to
k
s
u
b
s
et
s
o
f
eq
u
al
s
iz
e.
T
h
e
n
et
w
o
r
k
w
ill
b
e
tr
ain
ed
k
ti
m
es,
ea
c
h
ti
m
e
leav
i
n
g
o
u
t o
n
e
o
f
th
e
s
u
b
s
e
ts
f
r
o
m
tr
ai
n
i
n
g
a
n
d
co
n
s
id
er
it
f
o
r
test
in
g
.
T
h
e
av
er
ag
e
o
f
k
r
e
s
u
lt
s
f
r
o
m
t
h
e
f
o
ld
s
g
iv
e
t
h
e
te
s
t a
cc
u
r
ac
y
o
f
th
e
al
g
o
r
ith
m
.
3
.
4
.
Art
if
ic
i
a
l bee
co
lo
ny
(
A
B
C
)
Gen
er
all
y
,
t
h
e
a
r
ti
f
icial
b
ee
c
o
lo
n
y
as i
n
tr
o
d
u
ce
d
b
y
b
y
Kar
a
b
o
g
a
[
1
8
]
c
an
b
e
d
escr
ib
ed
as
f
o
llo
w
s
:
Th
e
Gen
era
l A
s
p
ec
ts
I
n
th
e
A
B
C
al
g
o
r
ith
m
,
ea
ch
f
o
o
d
s
o
u
r
ce
p
o
s
itio
n
r
ep
r
esen
ts
a
s
o
l
u
tio
n
to
a
s
p
ec
if
ic
p
r
o
b
lem
a
n
d
th
e
a
m
o
u
n
t
o
f
n
ec
tar
in
a
f
o
o
d
s
o
u
r
ce
r
ep
r
esen
ts
t
h
e
o
b
j
ec
tiv
e
f
u
n
ctio
n
(
th
e
f
i
tn
e
s
s
)
o
f
th
e
s
o
lu
tio
n
.
I
n
th
e
h
i
v
e,
th
r
ee
t
y
p
e
s
o
f
b
ee
s
ar
e
co
n
s
id
er
ed
:
em
p
lo
y
ed
b
ee
s
,
o
n
lo
o
k
er
b
ee
s
an
d
s
co
u
t
b
ee
s
.
T
h
e
A
B
C
alg
o
r
ith
m
co
n
s
i
s
ts
o
f
a
n
u
m
b
er
o
f
c
y
cles.
Du
r
i
n
g
ea
ch
c
y
cle,
t
h
er
e
ar
e
th
r
ee
m
ai
n
p
ar
ts
:
s
en
d
i
n
g
th
e
e
m
p
lo
y
ed
b
ee
s
to
th
e
f
o
o
d
s
o
u
r
ce
s
an
d
m
ea
s
u
r
in
g
t
h
ei
r
n
ec
tar
q
u
an
titi
es
;
s
elec
ti
n
g
t
h
e
f
o
o
d
s
o
u
r
ce
s
b
y
th
e
o
n
lo
o
k
er
s
;
d
eter
m
i
n
i
n
g
th
e
s
co
u
t
b
ee
s
a
n
d
e
x
p
lo
r
in
g
n
e
w
p
o
s
s
ib
le
f
o
o
d
s
o
u
r
ce
s
[
2
7
]
.
A
g
e
n
er
al
al
g
o
r
ith
m
f
o
r
th
e
A
B
C
o
p
ti
m
iza
tio
n
ap
p
r
o
ac
h
[
2
8
]
ca
n
b
e
ex
p
r
ess
ed
as f
o
llo
w
s
:
1.
I
n
itialize
f
o
o
d
s
o
u
r
ce
p
o
s
itio
n
s
Repea
t
2.
Sen
d
i
n
g
e
m
p
lo
y
ed
b
ee
s
to
f
o
o
d
s
o
u
r
ce
p
o
s
itio
n
s
an
d
ca
lc
u
lat
in
g
t
h
e
p
r
o
b
ab
ilit
y
v
al
u
es.
3.
Selectio
n
o
f
f
o
o
d
s
o
u
r
ce
p
o
s
itio
n
s
b
y
o
n
lo
o
k
er
b
ee
s
b
y
co
n
s
i
d
er
in
g
t
h
e
p
r
o
b
ab
ilit
y
v
alu
e
s
.
4.
A
b
a
n
d
o
n
in
g
s
o
u
r
ce
s
w
it
h
les
s
p
r
o
b
a
b
ilit
y
an
d
p
r
o
d
u
cin
g
n
e
w
f
o
o
d
s
o
u
r
ce
in
n
ei
g
h
b
o
u
r
h
o
o
d
o
f
o
ld
s
o
u
r
ce
.
5.
T
h
e
b
est f
o
o
d
s
o
u
r
ce
f
o
u
n
d
s
o
f
ar
is
r
eg
i
s
ter
ed
Unt
il
(
r
eq
u
ir
e
m
e
n
ts
ar
e
m
e
t)
4.
T
E
X
T
UR
E
C
L
ASS
I
F
I
C
AT
I
O
N
T
h
is
p
ap
er
f
o
cu
s
es
o
n
a
b
in
ar
y
tex
t
u
r
e
clas
s
if
icatio
n
a
n
d
th
e
Un
iv
er
s
it
y
o
f
Ma
r
y
la
n
d
,
C
o
lleg
e
P
ar
k
i
m
a
g
e
tex
t
u
r
ed
d
atab
ase
(
UM
D
d
atab
a
s
e)
h
as
b
ee
n
e
m
p
lo
y
ed
.
T
h
e
tex
tu
r
e
class
i
f
icat
io
n
p
r
o
ce
s
s
co
n
s
is
ts
o
f
th
r
ee
m
ai
n
s
ta
g
e
s
: i
m
a
g
e
te
x
t
u
r
e
p
r
ep
a
r
atio
n
,
f
ea
tu
r
e
e
x
tr
ac
ti
o
n
an
d
class
i
f
ica
tio
n
.
4
.
1
.
I
m
a
g
e
t
ex
t
ure
prepa
ra
t
io
n
I
m
ag
e
tex
t
u
r
e
p
r
ep
ar
atio
n
i
s
t
h
e
s
tag
e
w
h
er
e
UM
D
te
x
t
u
r
ed
i
m
a
g
e
d
at
ab
ase
is
p
r
ep
ar
ed
,
in
o
r
d
er
to
g
et
a
s
u
f
f
icie
n
t
n
u
m
b
er
o
f
i
m
a
g
e
s
a
m
p
le
s
w
it
h
r
eq
u
ir
ed
d
iv
er
s
it
y
.
UM
D
te
x
tu
r
ed
i
m
a
g
e
d
atab
ase
co
n
tain
s
2
5
g
r
o
u
p
s
;
ea
ch
g
r
o
u
p
co
n
tai
n
s
4
0
s
am
p
le
s
w
it
h
s
ize
1
2
8
0
×
9
6
0
p
ix
els.
E
ac
h
i
m
a
g
e
s
a
m
p
le
f
r
o
m
th
e
UM
D
d
atab
ase
w
it
h
s
ize
1
28
0
×
960
p
ix
el
w
er
e
s
u
b
s
eq
u
e
n
tl
y
s
e
g
m
en
ted
in
to
8
0
s
a
m
p
le
s
,
w
h
e
r
e
7
0
s
am
p
le
s
w
it
h
s
ize
1
2
8
×
1
2
8
an
d
1
0
s
a
m
p
les
w
it
h
t
h
e
s
ize
1
2
8
×
6
4
.
I
n
th
is
p
ap
er
,
w
e
u
s
e
t
h
e
s
a
m
p
le
s
w
it
h
s
ize
1
2
8
×
1
2
8
,
an
d
ig
n
o
r
e
th
e
s
ize
1
2
8
×
6
4
.
A
t
th
e
e
n
d
o
f
t
h
is
s
tep
,
w
e
o
b
tain
ed
2
8
0
0
i
m
a
g
e
s
f
r
o
m
t
h
e
s
a
m
e
s
a
m
p
le
w
i
th
s
ize
128
×
1
2
8
.
T
h
u
s
t
h
e
to
tal
n
u
m
b
er
o
f
test
i
m
ag
e
s
f
r
o
m
2
5
s
a
m
p
le
s
ar
e
7
0
,
0
0
0
im
a
g
e
s.
A
t
o
tal
o
f
5
0
0
tex
tu
r
e
i
m
a
g
es
w
er
e
r
an
d
o
m
l
y
s
elec
te
d
f
r
o
m
th
e
2
8
0
0
i
m
ag
e
s
.
L
astl
y
b
i
n
ar
y
i
m
a
g
e
te
x
t
u
r
e
g
r
o
u
p
is
es
tab
lis
h
ed
f
r
o
m
d
if
f
er
e
n
t
i
m
a
g
e
g
r
o
u
p
.
E
ac
h
b
in
ar
y
g
r
o
u
p
co
n
tai
n
s
1
0
0
0
tex
tu
r
es
i
m
a
g
e
s
w
h
er
e
t
w
o
s
a
m
p
les
f
o
r
ea
ch
g
r
o
u
p
o
f
500
im
a
g
es
w
er
e
r
a
n
d
o
m
l
y
s
elec
ted
.
So
m
e
s
a
m
p
les
o
f
th
e
i
m
a
g
e
tex
t
u
r
e
p
r
ep
ar
atio
n
p
r
o
ce
s
s
o
f
t
h
e
b
in
ar
y
i
m
a
g
e
g
r
o
u
p
i
s
il
lu
s
tr
ated
in
F
ig
u
r
e
2.
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:
2
0
8
8
-
8708
Op
timiz
a
tio
n
o
f d
is
crete
w
a
ve
l
et
tr
a
n
s
fo
r
m
fea
tu
r
es u
s
in
g
a
r
t
ificia
l b
ee
co
lo
n
y
a
lg
o
r
ith
m…
(
F
th
i M.
A
lb
ko
s
h
)
5257
Fig
u
r
e
2
.
Sa
m
p
le
i
m
ag
e
s
o
f
d
i
f
f
er
en
t b
in
ar
y
i
m
ag
e
cla
s
s
e
s
4
.
2
.
F
ea
t
ures e
x
t
ra
ct
io
n
a
nd
cla
s
s
if
ica
t
io
n
Featu
r
e
e
x
tr
ac
tio
n
is
th
e
s
ec
o
n
d
m
aj
o
r
s
tag
e
o
f
th
is
s
t
u
d
y
.
A
t
th
i
s
s
ta
g
e,
r
ele
v
an
t i
m
ag
e
f
ea
t
u
r
es
w
er
e
ex
tr
ac
ted
f
r
o
m
th
e
i
m
ag
e
te
x
t
u
r
e
to
m
a
k
e
it
r
ea
d
y
f
o
r
class
if
icatio
n
p
r
o
ce
s
s
.
W
e
u
s
e
th
e
A
B
C
al
g
o
r
ith
m
to
f
i
n
d
th
e
s
u
itab
le
co
m
b
i
n
atio
n
o
f
DW
T
d
ec
o
m
p
o
s
itio
n
le
v
el
an
d
th
e
b
est
f
ilter
o
f
DW
T
.
T
h
e
f
i
n
al
s
tag
e
in
th
is
s
tu
d
y
is
t
h
e
class
if
ica
tio
n
m
et
h
o
d
s
.
I
n
th
is
p
ap
er
,
th
e
M
L
P
NN
u
p
d
ates
its
w
ei
g
h
ts
t
h
r
o
u
g
h
b
ac
k
-
p
r
o
p
ag
atio
n
lear
n
in
g
al
g
o
r
ith
m
.
I
t
u
s
e
s
t
h
e
g
en
er
alize
d
lea
s
t
m
ea
n
s
q
u
ar
e
(
L
MS)
r
u
le
a
n
d
e
m
p
lo
y
s
th
e
g
r
ad
ien
t
s
ea
r
c
h
lear
n
in
g
m
et
h
o
d
to
r
ed
u
ce
th
e
av
er
ag
e
d
i
f
f
er
e
n
ce
b
et
w
ee
n
th
e
o
u
tp
u
t
an
d
t
h
e
tar
g
et
v
a
lu
e
o
n
t
h
e
n
eu
r
al
n
et
w
o
r
k
,
4
0
n
eu
r
o
n
s
i
n
th
e
h
i
d
d
en
lay
er
s
is
u
s
ed
.
A
h
y
p
er
b
o
lic
tan
g
e
n
t
an
d
li
n
ea
r
ac
tiv
at
io
n
f
u
n
ct
io
n
s
w
er
e
u
s
ed
i
n
t
h
e
h
id
d
en
an
d
o
u
tp
u
t
la
y
er
,
r
esp
ec
ti
v
el
y
.
T
o
ev
alu
ate
t
h
e
c
lass
if
ier
p
er
f
o
r
m
a
n
ce
,
a
5
-
f
o
ld
cr
o
s
s
v
alid
atio
n
s
tr
ate
g
y
is
ap
p
lied
w
h
er
e
2
0
0
i
m
a
g
es
u
s
ed
f
o
r
tes
tin
g
w
h
er
ea
s
t
h
e
r
e
m
ain
in
g
8
0
0
i
m
ag
e
s
w
er
e
u
s
ed
f
o
r
tr
ain
i
n
g
.
Af
ter
ap
p
l
y
i
n
g
a
n
y
f
il
ter
w
ith
s
p
ec
if
ic
d
ec
o
m
p
o
s
itio
n
lev
el,
th
e
f
o
llo
w
i
n
g
tex
t
u
r
e
d
escr
ip
to
r
s
w
er
e
ca
lcu
lated
f
o
r
p
ar
ticu
lar
r
eso
lu
t
io
n
lev
e
l
o
f
t
h
e
w
a
v
elet
:
m
ea
n
,
s
ta
n
d
ar
d
d
ev
iatio
n
,
e
n
er
g
y
a
n
d
en
tr
o
p
y
.
T
h
e
e
n
er
g
y
f
o
r
th
e
ap
p
r
o
x
i
m
atio
n
s
u
b
-
b
an
d
an
d
th
e
m
ea
n
o
f
en
er
g
y
f
o
r
th
e
d
etails
s
u
b
-
b
an
d
ar
e
u
tili
ze
d
.
E
n
tr
o
p
y
,
m
ea
n
an
d
s
ta
n
d
ar
d
d
ev
iatio
n
ar
e
ca
lc
u
lated
f
o
r
e
ac
h
s
u
b
-
b
an
d
(
ap
p
r
o
x
i
m
at
io
n
,
d
etail
r
eg
io
n
s
)
o
f
b
est d
ec
o
m
p
o
s
itio
n
lev
e
l.
B
y
t
h
at
it
y
ield
s
a
to
tal
o
f
1
4
d
escr
ip
to
r
s
.
4
.
3
.
O
pti
m
iza
t
io
n o
f
DWT
pa
ra
m
et
er
s
Desig
n
i
n
g
o
f
DW
T
d
ep
en
d
s
o
n
s
o
m
e
p
ar
a
m
eter
s
s
u
ch
a
s
w
a
v
elet
d
ec
o
m
p
o
s
itio
n
f
ilter
f
o
r
an
al
y
s
i
s
an
d
n
u
m
b
er
o
f
w
a
v
elet
d
ec
o
m
p
o
s
itio
n
.
T
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
is
d
ev
elo
p
ed
f
o
r
d
ete
r
m
in
in
g
th
e
m
o
s
t
ef
f
icien
t
f
il
ter
o
u
t
o
f
4
9
d
if
f
er
en
t
w
a
v
elet
f
ilter
s
,
in
ad
d
itio
n
to
ch
o
o
s
in
g
th
e
m
o
s
t
a
p
p
r
o
p
r
iate
w
a
v
elet
d
ec
o
m
p
o
s
itio
n
f
i
lter
f
r
o
m
ei
g
h
t d
ec
o
m
p
o
s
itio
n
f
i
lter
.
T
ab
le
1
s
h
o
w
th
e
f
ilter
s
u
s
ed
.
a.
I
m
p
le
m
e
n
tatio
n
o
f
A
B
C
A
l
g
o
r
ith
m
to
Op
ti
m
ize
DW
T
I
n
th
i
s
w
o
r
k
,
th
e
p
o
s
itio
n
s
o
f
f
o
o
d
s
s
o
u
r
c
es
in
A
B
C
al
g
o
r
it
h
m
ar
e
co
n
s
id
er
ed
as
th
e
p
ar
a
m
eter
s
o
f
DW
T
th
at
ca
n
p
r
o
v
id
e
s
p
ec
if
i
c
s
o
lu
tio
n
s
.
Fo
r
i
n
s
ta
n
ce
,
t
h
e
p
ar
am
eter
s
th
a
t
ar
e
b
ee
n
co
n
s
i
d
er
ed
f
o
r
th
e
DW
T
ar
e
f
ilter
t
y
p
e
an
d
its
d
ec
o
m
p
o
s
itio
n
lev
e
l.
T
h
e
p
ar
am
eter
s
w
er
e
m
o
d
i
f
ied
w
ith
t
h
e
ti
m
e
b
y
f
i
n
d
in
g
o
u
t
s
o
m
e
p
lace
s
w
ith
h
i
g
h
n
ec
tar
s
as
s
t
ated
in
A
B
C
al
g
o
r
ith
m
.
Her
e
,
th
e
h
ig
h
a
m
o
u
n
t
o
f
th
e
n
ec
t
ar
is
co
n
s
id
er
ed
as
th
e
h
i
g
h
ac
cu
r
ac
y
in
b
i
n
ar
y
tex
t
u
r
es
clas
s
i
f
icatio
n
.
T
h
e
p
s
eu
d
o
co
d
e
o
f
th
e
A
B
C
alg
o
r
ith
m
to
o
p
ti
m
ize
th
e
DW
T
p
ar
am
eter
s
to
g
et
t
h
e
m
a
x
i
m
u
m
ac
cu
r
ac
y
o
f
b
in
ar
y
tex
t
u
r
es c
la
s
s
i
f
icat
io
n
is
il
lu
s
tr
ated
in
Fi
g
u
r
e
3
.
T
ab
le
1
.
Dif
f
er
en
t
w
a
v
elet
f
ilt
er
s
u
s
ed
W
a
v
e
l
e
t
F
a
mi
l
i
e
s
W
a
v
e
l
e
t
s
H
a
a
r
w
a
v
e
l
e
t
'
h
a
a
r
'
D
a
u
b
e
c
h
i
e
s
‘
d
b
1
’
,
‘
d
b
2
’
,
‘
d
b
3
’
,
‘
d
b
4
’
,
‘
d
b
5
’
,
‘
d
b
6
’
,
‘
d
b
7
’
,
‘
d
b
8
’
,
‘
d
b
9
’
C
o
i
f
l
e
t
s
'
‘
c
o
i
f
1
’
,
‘
c
o
i
f
2
’
,
‘
c
o
i
f
3
’
,
‘
c
o
i
f
4
’
,
‘
c
o
i
f
5
’
S
y
m
l
e
t
s
‘
sym1
’
,
s
y
m
2
,
s
y
m
3
,
sym
4
,
sym5
,
s
y
m
6
,
‘
s
y
m
7
’
D
i
scre
t
e
M
e
y
e
r
'
d
m
e
y
'
B
i
o
r
t
h
o
g
o
n
a
l
‘
b
i
o
r1
.
1
’
,
‘
b
i
o
r
1
.
3
’
,
‘
b
i
o
r
1
.
5
’
,
‘
b
i
o
r2
.
2
’
,
‘
b
i
o
r
2
.
4
’
,
‘
b
i
o
r
2
.
6
’
,
‘
b
i
o
r2
.
8
’
,
‘
b
i
o
r
3
.
1
’
,
‘
b
i
o
r
3
.
3
’
,
‘
b
i
o
r
3
.
5
’
,
‘
b
i
o
r
3
.
7
’
,
‘
b
i
o
r3
.
9
’
,
‘
b
i
o
r
4
.
4
’
R
e
v
e
r
s
e
B
i
o
r
t
h
o
g
o
n
a
l
'
rb
i
o
1
.
1
'
,
'
r
b
i
o
1
.
3
'
,
'
r
b
i
o
1
.
5
'
,
'
rb
i
o
2
.
2
'
,
'
r
b
i
o
2
.
4
'
,
'
r
b
i
o
2
.
6
'
,
'
r
b
i
o
2
.
8
’
,
'
r
b
i
o
3
.
1
'
,
'
rb
i
o
3
.
3
'
,
'
r
b
i
o
3
.
5
'
,
'
r
b
i
o
3
.
7
'
,
'
r
b
i
o
3
.
9
'
,
'
rb
i
o
4
.
4
'
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
2
5
3
-
5
2
6
2
5258
In
i
t
i
a
l
i
z
a
t
i
o
n
Pha
se
:
DO
R
=
r
a
n
d
o
m (f
i
l
t
e
r
(
y
1
)
a
n
d
d
e
c
o
mp
o
s
i
t
i
o
n
l
e
v
e
l
(
y
2
)
)
,
R
i
s
t
h
e
n
u
m
b
e
r
o
f
i
n
i
t
i
a
l
f
o
o
d
so
u
r
c
e
s
D
=
i
mr
e
a
d
(
f
i
l
e
n
a
me
)
;
f
i
l
e
n
a
me
c
o
n
si
s
t
o
f
a
g
r
o
u
p
s
t
h
a
t
c
o
n
t
a
i
n
s
t
w
o
c
l
a
sse
s o
f
i
mag
e
s
d
w
t
=
w
a
v
e
d
e
c
2
(
D
,
y
1
,
y
2
)
;
w
a
v
e
d
e
c
2
i
s f
u
n
c
t
i
o
n
o
f
w
a
v
e
l
e
t
t
r
a
n
sf
o
r
m
Ex
t
r
a
c
t
i
o
n
1
4
F
e
a
t
u
r
e
s u
s
i
n
g
(
E
n
t
r
o
p
y
,
M
e
a
n
,
E
n
e
r
g
y
a
n
d
S
t
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
)
Ev
a
l
u
a
t
e
t
h
e
i
n
i
t
i
a
l
i
z
e
s re
su
l
t
o
f
d
w
t
b
y
se
l
e
c
t
e
d
f
i
l
t
e
r
a
n
d
d
e
c
o
mp
o
si
t
i
o
n
l
e
v
e
l
u
si
n
g
M
L
P
N
N
b
a
se
d
o
n
:
1
.
A
c
c
u
r
a
c
y
2
.
S
p
e
c
i
f
i
c
i
t
y
3
.
S
e
n
s
i
t
i
v
i
t
y
W
HILE
(
I
n
i
t
i
a
l
i
z
e
st
a
g
e
a
p
p
l
i
e
d
)
,
E
m
p
l
o
y
e
d
p
h
a
se:
DO
R
=
r
a
n
d
o
m (f
i
l
t
e
r
(
y
1
)
a
n
d
d
e
c
o
mp
o
si
t
i
o
n
l
e
v
e
l
(
y
2
)
)
,
R
i
s
t
h
e
n
u
m
b
e
r
o
f
e
mp
l
o
y
e
d
f
o
o
d
so
u
r
c
e
s
D
=
i
mr
e
a
d
(
f
i
l
e
n
a
me
)
;
f
i
l
e
n
a
me
c
o
n
si
s
t
o
f
a
g
r
o
u
p
s
t
h
a
t
c
o
n
t
a
i
n
s
t
w
o
c
l
a
sse
s o
f
i
mag
e
s
d
w
t
=
w
a
v
e
d
e
c
2
(
D
,
y
1
,
y
2
)
;
w
a
v
e
d
e
c
2
i
s f
u
n
c
t
i
o
n
o
f
w
a
v
e
l
e
t
t
r
a
n
sf
o
r
m
Ex
t
r
a
c
t
i
o
n
1
4
F
e
a
t
u
r
e
s u
si
n
g
(
E
n
t
r
o
p
y
,
M
e
a
n
,
E
n
e
r
g
y
a
n
d
S
t
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
)
Ev
a
l
u
a
t
e
t
h
e
e
mp
l
o
y
e
d
r
e
su
l
t
o
f
d
w
t
b
y
se
l
e
c
t
e
d
f
i
l
t
e
r
a
n
d
d
e
c
o
mp
o
si
t
i
o
n
l
e
v
e
l
u
si
n
g
M
L
P
N
N
b
a
se
d
o
n
:
1
.
A
c
c
u
r
a
c
y
2
.
S
p
e
c
i
f
i
c
i
t
y
3
.
S
e
n
s
i
t
i
v
i
t
y
Ev
a
l
u
a
t
e
d
E
mp
l
o
y
e
d
d
i
st
a
n
c
e
a
n
d
a
n
g
l
e
b
y
M
L
P
N
N
5
-
f
o
l
d
IF
A
c
c
u
r
a
c
y
M
L
P
N
N
-
I
n
i
t
i
a
l
i
z
e
>
A
c
c
u
r
a
c
y
M
L
P
N
N
-
Emp
l
o
y
e
d
S
e
l
e
c
t
I
n
i
t
i
a
l
i
z
e
f
i
l
t
e
r
a
n
d
d
e
c
o
mp
o
si
t
i
o
n
l
e
v
e
l
E
L
S
E
S
e
l
e
c
t
Emp
l
o
y
e
d
f
i
l
t
e
r
a
n
d
d
e
c
o
m
p
o
si
t
i
o
n
l
e
v
e
l
E
N
D
W
HILE
(
Em
p
l
o
y
e
d
st
a
g
e
a
p
p
l
i
e
d
),
F
o
u
n
d
t
h
e
G
l
o
b
a
l
so
l
u
t
i
o
n
-
e
mp
l
o
y
e
d
st
a
g
e
(
T
h
e
b
e
st
so
l
u
t
i
o
n
(
F
i
l
t
e
r
a
n
d
d
e
c
o
mp
o
si
t
i
o
n
l
e
v
e
l
t
h
a
t
p
r
o
v
i
d
e
t
h
e
b
e
st
a
c
c
u
r
a
c
y
)
)
On
l
o
o
k
e
r
p
h
a
se:
L
o
c
a
l
se
a
r
c
h
f
o
r
t
h
e
b
e
st
so
l
u
t
i
o
n
-
a
n
g
l
e
s a
n
d
d
i
s
t
a
n
c
e
s
b
y
l
o
o
k
i
n
g
f
o
r
t
h
e
n
e
i
g
h
b
o
r
h
o
o
d
d
e
c
o
m
p
o
si
t
i
o
n
l
e
v
e
l
w
i
t
h
same
f
i
l
t
e
r
Ev
a
l
u
a
t
e
t
h
e
n
e
w
so
l
u
t
i
o
n
s
-
F
i
l
t
e
r
a
n
d
d
e
c
o
mp
o
si
t
i
o
n
l
e
v
e
l
b
y
M
L
P
N
N
IF
Ev
a
l
u
a
t
e
d
M
L
P
N
N
Em
p
l
o
y
e
d
>
Ev
a
l
u
a
t
e
d
M
L
P
N
N
O
n
l
o
o
k
e
r
S
e
l
e
c
t
Emp
l
o
y
e
d
f
i
l
t
e
r
a
n
d
d
e
c
o
mp
o
si
t
i
o
n
l
e
v
e
l
E
L
S
E
S
e
l
e
c
t
O
n
l
o
o
k
e
r
f
i
l
t
e
r
a
n
d
d
e
c
o
m
p
o
si
t
i
o
n
l
e
v
e
l
E
N
D
M
e
mo
r
i
z
e
t
h
e
o
p
t
i
mal
f
i
l
t
e
r
a
n
d
d
e
c
o
m
p
o
si
t
i
o
n
l
e
v
e
l
Fig
u
r
e
3
.
T
h
e
p
s
eu
d
o
co
d
e
o
f
A
DW
T
b.
P
er
f
o
r
m
a
n
ce
Me
as
u
r
es
T
h
e
ac
cu
r
ac
y
i
s
m
ea
s
u
r
ed
u
s
i
n
g
f
o
llo
w
i
n
g
p
ar
a
m
eter
s
:
C
o
n
fu
s
io
n
Ma
tr
ix:
T
h
e
co
n
f
u
s
io
n
m
atr
i
x
s
h
o
w
s
t
h
e
p
er
ce
n
ta
g
es
o
f
co
r
r
ec
t
an
d
in
co
r
r
ec
t
clas
s
if
icatio
n
s
.
I
t
in
c
lu
d
es
f
o
u
r
class
i
f
i
ca
tio
n
p
er
f
o
r
m
an
ce
i
n
d
ices:
tr
u
e
p
o
s
itiv
e
(
T
P)
,
f
a
ls
e
p
o
s
itiv
e
(
FP
)
,
f
alse
n
e
g
at
iv
e
(
FN)
,
an
d
tr
u
e
n
eg
at
iv
e
(
T
N)
[
2
9
]
.
1)
Sen
s
iti
v
it
y
a
n
d
Sp
ec
if
icit
y
Sen
s
iti
v
it
y
a
n
d
s
p
ec
if
ici
t
y
ar
e
a
m
ea
s
u
r
e
o
f
tex
t
u
r
e
clas
s
i
f
icatio
n
ac
cu
r
ac
y
an
d
p
r
o
b
ab
ilit
y
t
h
at
a
tex
t
u
r
e
class
i
f
icat
io
n
tec
h
n
iq
u
e
w
ill i
n
co
r
r
ec
tl
y
r
ej
ec
t o
r
ac
c
ep
t a
n
in
p
u
t a
s
a
n
e
g
ati
v
e
o
r
p
o
s
itiv
e
cla
s
s
.
Sen
s
iti
v
it
y
,
S
N
, i
t i
s
t
h
e
p
r
o
p
o
r
tio
n
s
o
f
co
r
r
ec
t c
lass
i
f
icatio
n
f
r
o
m
f
ir
s
t c
las
s
.
S
N
v
al
u
e
i
s
co
m
p
u
ted
as
f
o
llo
w
s
:
SN
=
T
r
ue
Po
s
i
t
i
v
e
T
r
ue
Po
s
i
t
i
v
e
+
F
al
s
e
N
eg
at
i
v
e
(
9
)
Sp
ec
if
icit
y
,
SP
is
th
e
p
r
o
p
o
r
tio
n
s
o
f
co
r
r
ec
t c
las
s
i
f
icatio
n
f
r
o
m
t
h
e
s
ec
o
n
d
class
.
SP
is
co
m
p
u
ted
as
f
o
llo
w
s
:
SP
=
T
r
ue
N
eg
at
i
v
e
(
F
al
s
e
Po
s
i
t
i
v
e
+
T
r
ue
N
eg
at
i
v
e
)
(
1
0
)
2)
A
cc
u
r
ac
y
:
A
cc
u
r
ac
y
,
A
cc
is
t
h
e
g
lo
b
al
r
ep
r
esen
tatio
n
o
f
c
lass
if
ier
p
er
f
o
r
m
an
ce
a
n
d
is
d
ef
i
n
ed
as f
o
llo
w
s
:
SP
=
T
r
ue
N
eg
at
i
v
e
(
F
al
s
e
Po
s
i
t
i
v
e
+
T
r
ue
N
eg
at
i
v
e
)
(
1
1
)
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:
2
0
8
8
-
8708
Op
timiz
a
tio
n
o
f d
is
crete
w
a
ve
l
et
tr
a
n
s
fo
r
m
fea
tu
r
es u
s
in
g
a
r
t
ificia
l b
ee
co
lo
n
y
a
lg
o
r
ith
m…
(
F
th
i M.
A
lb
ko
s
h
)
5259
5.
RE
SU
L
T
S
A
ND
D
I
SCU
SS
I
O
N
R
es
u
lts
o
f
DW
T
an
d
A
DW
T
m
e
th
o
d
s
ar
e
s
h
o
w
n
i
n
T
ab
le
2
.
I
t
s
h
o
w
s
th
at
t
h
e
ML
P
N
N
class
i
f
ier
ac
h
iev
ed
d
if
f
er
e
n
t
ac
cu
r
ac
y
,
s
en
s
i
ti
v
it
y
a
n
d
s
p
ec
if
icit
y
.
T
h
e
r
esu
l
ts
o
b
tain
ed
clea
r
l
y
s
h
o
w
n
t
h
at
d
if
f
er
en
t
p
ar
am
eter
s
ar
e
ab
le
to
p
ick
u
p
m
in
u
te
ch
a
n
g
es
i
n
th
e
i
m
ag
e
s
class
i
f
icatio
n
ev
al
u
atio
n
.
T
h
e
av
er
ag
e
ac
cu
r
ac
y
o
f
ea
ch
g
r
o
u
p
is
co
m
p
ar
ed
w
it
h
t
h
e
av
er
ag
e
ac
cu
r
ac
y
b
y
u
s
i
n
g
DW
T
w
it
h
d
b
4
an
d
d
ec
o
m
p
o
s
itio
n
o
f
lev
el
t
w
o
.
Fig
u
r
e
4
ill
u
s
tr
ates
t
h
e
r
elatio
n
b
et
w
ee
n
p
ar
a
m
eter
s
w
h
ich
ar
e
f
ilter
t
y
p
e
w
it
h
d
ec
o
m
p
o
s
itio
n
le
v
el
an
d
th
e
p
er
f
o
r
m
a
n
ce
f
o
r
th
e
g
r
o
u
p
n
u
m
b
er
1
4
.
T
h
e
h
o
r
in
zo
n
tal
a
x
is
,
x
r
ep
r
ese
n
ts
r
a
n
d
o
m
d
i
s
tr
ib
u
tio
n
o
f
p
ar
am
eter
s
an
d
th
e
v
er
tical
a
x
is
,
y
r
ep
r
esen
t
s
th
e
p
er
f
o
r
m
a
n
ce
o
f
ea
ch
p
ar
a
m
eter
.
As
ca
n
b
e
o
b
s
er
v
ed
f
r
o
m
th
e
g
r
ap
h
i
n
F
ig
u
r
e
4
,
al
m
o
s
t
d
if
f
er
e
n
t
f
ilter
s
w
it
h
d
i
f
f
e
r
en
t
d
ec
o
m
p
o
s
itio
n
le
v
els
g
i
v
e
d
if
f
er
e
n
t
te
s
ti
n
g
class
i
f
icatio
n
p
er
f
o
r
m
a
n
ce
s
.
B
ased
o
n
th
e
b
est
te
s
tin
g
ac
c
u
r
ac
y
an
d
u
s
i
n
g
ar
ti
f
icial
b
ee
co
lo
n
y
,
all
g
r
ad
atio
n
o
f
test
in
g
ac
c
u
r
ac
y
f
r
o
m
4
9
.
5
%
f
o
r
b
io
r
3
.
1
f
ilter
w
i
th
d
ec
o
m
p
o
s
itio
n
lev
e
l
6
u
n
t
il
9
8
.
2
0
%
f
o
r
h
aa
r
a
n
d
d
m
e
y
f
ilter
s
a
n
d
d
ec
o
m
p
o
s
itio
n
le
v
e
l 2
f
o
r
ea
ch
A
DW
T
o
b
tain
ed
th
e
h
i
g
h
e
s
t
r
ate
o
f
ac
cu
r
ac
y
r
ec
o
r
d
ed
w
i
th
p
ar
a
m
eter
b
io
r
3
.
1
f
ilter
an
d
at
d
ec
o
m
p
o
s
itio
n
le
v
el
1
w
i
th
m
ax
i
m
u
m
ac
c
u
r
ac
y
o
f
9
9
.
4
0
% f
o
r
th
e
g
r
o
u
p
n
u
m
b
er
9
.
W
h
ile
th
e
o
t
h
er
2
9
g
r
o
u
p
s
d
atab
ases
r
ec
o
r
d
ed
th
eir
h
i
g
h
e
s
t
a
cc
u
r
ac
y
v
ar
iatio
n
s
at
d
if
f
er
en
t p
ar
a
m
eter
s
.
T
h
e
p
ar
am
eter
v
al
u
es r
ep
lace
s
th
e
ac
cu
r
ac
y
o
f
th
e
o
u
tp
u
t
b
as
ed
o
n
th
e
g
i
v
en
te
x
t
u
r
es
g
r
o
u
p
.
a.
Fo
r
s
o
m
e
te
x
t
u
r
e
g
r
o
u
p
s
,
s
o
m
e
ti
m
es
th
e
al
g
o
r
ith
m
p
r
o
v
id
e
m
u
lti
p
ar
a
m
eter
v
al
u
es
w
it
h
eq
u
i
v
ale
n
t
test
i
n
g
p
er
f
o
r
m
a
n
ce
s
i
m
ilar
to
G1
4
to
g
et
t
w
o
p
ar
a
m
e
t
er
v
al
u
es
w
h
ich
ar
e
haar
a
n
d
d
mey
a
n
d
d
ec
o
m
p
o
s
itio
n
le
v
e
l 2
f
o
r
ea
ch
,
w
it
h
test
i
n
g
p
er
f
o
r
m
an
ce
9
8
.
2
0
% a
s
ca
n
o
b
s
er
v
ed
in
Fig
u
r
e
4
.
b.
So
m
e
tex
tu
r
e
t
y
p
e
s
w
h
ic
h
g
et
d
if
f
er
en
t
p
er
f
o
r
m
an
ce
a
n
d
s
a
m
e
p
ar
a
m
eter
v
al
u
es
o
f
f
ilter
a
n
d
d
ec
o
m
p
o
s
itio
n
lev
el
(
G1
an
d
G
6
)
o
b
tain
ed
th
e
b
est
p
ar
a
m
eter
v
al
u
es
(
d
m
e
y
f
i
lter
an
d
d
ec
o
m
p
o
s
itio
n
lev
el
2
)
w
i
th
t
h
e
h
ig
h
er
test
i
n
g
p
er
f
o
r
m
a
n
ce
8
9
.
7
0
%
an
d
9
6
.
2
0
%,
r
esp
ec
tiv
el
y
.
T
h
e
b
eh
av
io
r
o
f
th
e
al
g
o
r
ith
m
s
v
ar
ies
o
v
er
d
if
f
er
e
n
t
g
r
o
u
p
s
as
t
h
e
ch
ar
ac
ter
is
tics
o
f
th
e
tex
tu
r
e
i
m
a
g
es
d
i
f
f
er
.
T
h
is
is
also
a
f
f
ec
ted
b
y
t
h
e
d
etail
s
i
n
clu
d
ed
in
t
h
e
i
m
a
g
es
a
n
d
t
h
e
cl
ar
it
y
i
n
q
u
al
i
t
y
an
d
th
e
r
ep
ea
tin
g
o
r
r
an
d
o
m
p
atter
n
tex
t
u
r
es d
is
tr
ib
u
tio
n
.
T
ab
le
2
.
R
esu
lts
o
f
A
B
C
alg
o
r
ith
m
w
it
h
DW
T
G
ro
u
p
n
o
D
W
T
A
cc
u
s
i
n
g
d
b
4
an
d
D
ec
l
ev
el
2
Param
et
e
r
S
el
ect
i
o
n
E
v
al
u
at
i
o
n
p
erf
o
rman
ce
ADWT
F
i
l
t
er
T
y
p
e
D
ec
-
l
ev
el
A
CC
SN
SP
G1
8
8
.
4
0
d
m
ey
2
8
9
.
7
0
9
0
.
5
2
8
8
.
6
8
G2
9
2
.
7
0
b
i
o
r
3
.
5
1
9
3
.
9
0
9
3
.
5
7
9
4
.
1
8
G3
7
6
.
9
0
d
b
5
1
7
8
.
7
0
7
8
.
3
0
7
8
.
9
0
G4
8
0
.
8
0
r
b
i
o
4
.
4
1
8
4
.
6
0
7
9
.
7
2
8
9
.
3
2
G5
8
4
.
5
0
s
ym
7
1
8
5
.
7
0
8
8
.
1
8
8
3
.
4
7
G6
9
4
.
3
0
d
m
ey
2
9
6
.
2
0
9
5
.
7
7
9
6
.
6
6
G7
6
9
.
9
0
b
i
o
r
1
.
5
1
8
1
.
8
0
7
7
.
2
4
8
6
.
4
5
G8
8
2
.
8
0
s
ym
7
1
8
6
.
3
0
8
6
.
2
2
8
6
.
4
6
G9
9
8
.
4
0
b
i
o
r
3
.
1
1
9
9
.
4
0
9
9
.
2
1
9
9
.
6
3
G
1
0
9
7
.
2
0
b
i
o
r
4
.
4
2
9
7
.
9
0
9
7
.
3
7
9
8
.
4
0
G
1
1
9
2
.
4
0
r
b
i
o
3
.
1
1
9
7
.
3
0
9
6
.
4
9
9
7
.
9
6
G
1
2
9
4
.
9
0
d
b
3
1
9
6
.
3
0
9
5
.
8
0
9
6
.
8
0
G
1
3
9
6
.
2
0
s
ym
1
2
9
9
.
0
0
9
8
.
9
7
9
8
.
9
6
G
1
4
9
7
.
5
0
h
a
a
r
2
9
8
.
2
0
9
8
.
41
9
7
.
9
4
d
m
ey
2
9
7
.
6
9
8
.
7
7
G
1
5
9
3
.
9
0
r
b
i
o
1
.
5
1
9
5
.
2
0
9
5
.
0
0
9
5
.
4
1
G
1
6
9
0
.
3
0
r
b
i
o
3
.
3
2
9
3
.
0
0
8
4
.
0
5
7
6
.
3
8
G
1
7
7
4
.
9
0
d
b
5
1
8
0
.
3
0
9
3
.
0
3
9
0
.
9
1
G
1
8
8
8
.
2
0
r
b
i
o
2
.
2
1
9
1
.
9
0
9
2
.
3
9
8
9
.
2
2
G
1
9
8
7
.
8
0
r
b
i
o
2
.
6
2
9
0
.
7
0
9
9
.
5
9
9
9
.
8
0
G
2
0
9
7
.
3
0
b
i
o
r
3
.
9
1
9
8
.
1
0
9
8
.
2
2
9
8
.
0
4
co
i
f
4
1
9
7
.
8
3
9
8
.
3
9
G
2
1
9
6
.
7
0
d
b
4
3
9
7
.
4
0
9
7
.
4
2
9
7
.
3
2
G
2
2
9
6
.
9
0
b
i
o
r
1
.
1
1
9
7
.
9
0
9
6
.
9
2
9
8
.
8
1
G
2
3
8
4
.
8
0
d
b
2
2
8
8
.
4
0
8
8
.
7
6
8
7
.
9
4
G
2
4
8
3
.
0
0
b
i
o
r
1
.
1
2
8
4
.
5
0
8
4
.
7
2
8
4
.
2
6
G
2
5
9
5
.
3
0
co
i
f
4
1
9
7
.
4
0
9
6
.
9
5
9
7
.
7
6
G
2
6
9
2
.
0
0
r
b
i
o
2
.
4
2
95
.
00
9
2
.
4
6
8
6
.
4
9
G2
7
8
5
.
5
0
s
ym
6
2
8
9
.
4
0
9
5
.
8
8
9
3
.
8
7
G
2
8
6
6
.
8
0
r
b
i
o
1
.
3
1
7
2
.
0
0
9
6
.
7
6
9
2
.
4
9
G
2
9
9
2
.
4
0
r
b
i
o
2
.
2
2
9
4
.
8
0
9
6
.
2
0
9
8
.
8
7
G
3
0
9
1
.
6
0
s
ym
2
1
9
4
.
6
0
9
7
.
4
2
9
7
.
3
2
A
v
erag
e
8
8
.
8
1
9
1
.
5
2
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
2
5
3
-
5
2
6
2
5260
Fig
u
r
e
4.
C
lass
i
f
icat
io
n
ac
cu
r
ac
y
f
o
r
i
m
a
g
e
Gr
o
u
p
1
4
B
ased
o
n
th
e
s
en
s
iti
v
it
y
a
n
d
s
p
ec
if
icit
y
p
er
f
o
r
m
a
n
ce
m
ea
s
u
r
e,
it c
an
b
e
co
n
clu
d
ed
as f
o
llo
w
s
:
a.
Gr
o
u
p
1
(
Scr
e
w
s
,
B
u
c
k
et
s
)
.
A
DW
T
w
i
th
p
ar
a
m
e
ter
d
m
e
y
f
ilter
w
i
th
d
ec
o
m
p
o
s
i
tio
n
le
v
el
2
p
r
o
v
id
es
ac
cu
r
ac
y
o
f
8
9
.
7
0
%.
A
t
th
e
s
a
m
e
ti
m
e
an
d
f
o
r
th
e
s
a
m
e
g
r
o
u
p
,
th
e
s
en
s
iti
v
it
y
a
n
d
s
p
ec
i
f
icit
y
o
b
tain
ed
w
a
s
9
0
.
5
2
%
an
d
8
8
.
6
8
%,
r
es
p
ec
tiv
el
y
.
T
h
is
m
ea
n
s
t
h
e
ac
cu
r
ac
y
is
i
n
c
r
ea
s
ed
i
n
class
o
n
e
(
Scr
e
w
s
)
o
f
th
e
g
r
o
u
p
a
n
d
d
ec
r
ea
s
ed
w
it
h
t
h
e
class
t
w
o
(
B
u
c
k
ets).
b.
Gr
o
u
p
6
(
T
r
ee
s
an
d
leav
es).
ADW
T
w
it
h
p
ar
a
m
eter
d
m
e
y
f
il
ter
an
d
d
ec
o
m
p
o
s
itio
n
le
v
el
2
,
g
iv
e
ac
c
u
r
ac
y
o
f
9
6
.
2
0
%.
T
h
e
s
en
s
iti
v
it
y
an
d
s
p
ec
if
icit
y
o
f
w
a
s
9
5
.
7
7
% a
n
d
9
6
.
6
6
%,
r
esp
ec
tiv
el
y
.
c.
Gr
o
u
p
7
(
leav
es1
an
d
leav
es2
)
.
A
DW
T
w
it
h
p
ar
a
m
eter
b
io
r
1
.
5
f
ilter
an
d
d
ec
o
m
p
o
s
itio
n
l
ev
el
1
p
r
o
v
id
e
ac
cu
r
ac
y
o
f
8
1
.
8
0
%.
T
h
e
s
en
s
it
iv
i
t
y
a
n
d
s
p
ec
i
f
icit
y
o
f
A
DW
T
o
b
t
ain
ed
is
7
7
.
2
4
%
an
d
8
6
.
4
5
%,
r
esp
ec
tiv
el
y
.
L
ast
l
y
,
all
t
h
e
g
r
ap
h
s
o
f
t
h
e
e
x
p
er
i
m
e
n
t
r
es
u
lts
s
h
o
w
ed
th
a
t
a
v
ar
iab
le
cu
r
v
e
w
h
en
ac
c
u
r
ac
y
v
al
u
es
w
er
e
p
lo
tted
a
g
ain
s
t
p
ar
a
m
et
er
v
al
u
es.
I
t
s
h
o
u
ld
b
e
h
ig
h
li
g
h
ted
th
at
A
DW
T
au
to
m
at
ic
all
y
s
elec
t
t
h
e
b
es
t
co
m
b
i
n
atio
n
o
f
p
ar
a
m
eter
s
t
o
g
et
m
ax
i
m
u
m
cla
s
s
i
f
icatio
n
ac
cu
r
ac
y
.
Fro
m
T
ab
le
2
,
i
t
ca
n
b
e
s
ee
n
t
h
at
th
e
p
r
o
p
o
s
ed
A
DW
T
m
o
d
el,
w
h
ic
h
is
a
n
o
p
ti
m
izat
io
n
o
f
t
h
e
DW
T
p
a
r
am
eter
s
u
s
in
g
AB
C
alg
o
r
ith
m
,
h
a
s
a
m
o
r
e
ac
c
u
r
ate
ca
ls
s
ı
f
ıcatıo
n
p
er
f
o
r
m
a
n
ce
w
h
ic
h
ac
h
ie
v
ed
an
av
er
a
g
e
ac
cu
r
ac
y
o
f
9
1
.
5
2
%
w
h
e
n
co
m
p
ar
ed
ag
ain
s
t
30
g
r
o
u
p
s
f
r
o
m
UM
D
d
atab
ase.
I
t
is
b
etter
th
an
th
e
a
v
er
ag
e
ac
cu
r
ac
y
o
f
th
e
s
a
m
e
3
0
g
r
o
u
p
s
w
it
h
d
b
4
f
ilter
an
d
d
ec
o
m
p
o
s
itio
n
lev
el
2
w
h
e
n
u
s
i
n
g
DW
T
w
h
ic
h
o
b
tain
ed
8
8
.
8
1
% c
lass
if
icatio
n
ac
cu
r
ac
y
.
6.
CO
NCLU
SI
O
N
C
las
s
i
f
y
i
n
g
t
h
e
te
x
t
u
r
es
cla
s
s
e
s
is
o
n
e
o
f
t
h
e
r
ec
en
t
r
esear
ch
is
s
u
es
in
t
h
e
f
ield
o
f
i
m
a
g
e
p
r
o
ce
s
s
in
g
.
T
h
e
class
if
icatio
n
ac
c
u
r
ac
y
c
an
b
e
i
m
p
r
o
v
ed
if
th
e
f
ea
t
u
r
e
s
elec
tio
n
i
s
p
r
o
p
er
.
T
h
e
b
est
DW
T
p
er
f
o
r
m
a
n
ce
ac
h
iev
ed
a
f
ter
a
s
er
ies
o
f
o
p
tim
izatio
n
o
n
t
h
e
s
u
itab
le
s
elec
t
io
n
o
f
w
a
v
elet
f
ilter
an
d
d
ec
o
m
p
o
s
itio
n
f
u
n
c
tio
n
s
in
v
o
l
v
ed
.
T
h
is
p
ap
er
s
h
o
w
s
t
h
at
a
u
to
m
at
ic
s
elec
tio
n
o
f
D
W
T
p
ar
am
eter
s
co
u
ld
b
e
ac
h
i
ev
ed
b
y
u
s
i
n
g
A
B
C
o
p
tim
izatio
n
al
g
o
r
ith
m
f
o
r
b
in
ar
y
tex
t
u
r
e
i
m
a
g
e
clas
s
if
icatio
n
.
T
h
e
ted
io
u
s
tr
ial
an
d
er
r
o
r
p
r
o
ce
s
s
o
f
p
ar
am
eter
s
elec
tio
n
f
o
r
f
i
n
d
in
g
t
h
e
b
est
p
ar
a
m
eter
co
m
b
i
n
a
tio
n
h
as
b
ee
n
a
v
o
id
ed
w
h
er
e
A
B
C
o
p
ti
m
izat
io
n
alg
o
r
ith
m
w
i
ll c
h
o
o
s
e
t
h
e
o
p
ti
m
al
p
ar
a
m
eter
s
f
o
r
t
h
e
b
in
ar
y
tex
t
u
r
e
class
if
ica
tio
n
.
F
u
t
u
r
e
c
h
alle
n
g
e
s
w
o
u
ld
b
e
i
m
p
le
m
en
t
in
g
t
h
e
s
i
m
ilar
p
r
o
ce
s
s
f
o
r
m
u
l
tip
le
class
i
m
a
g
e
te
x
tu
r
e
cla
s
s
i
f
icatio
n
.
RE
F
E
R
E
NC
E
S
[1
]
F
.
R.
d
e
S
i
q
u
e
ira,
W
.
R.
S
c
h
w
a
rtz,
a
n
d
H
.
P
e
d
ri
n
i,
"
M
u
lt
i
-
sc
a
le
g
ra
y
lev
e
l
c
o
-
o
c
c
u
rre
n
c
e
m
a
tri
c
e
s
f
o
r
tex
tu
re
d
e
sc
rip
ti
o
n
,
"
Ne
u
ro
c
o
m
p
u
ti
n
g
,
v
o
l.
1
2
0
,
p
p
.
3
3
6
-
3
4
5
,
2
0
1
3
.
[2
]
S
.
A
riv
a
z
h
a
g
a
n
a
n
d
L
.
Ga
n
e
sa
n
,
"
T
e
x
tu
re
se
g
m
e
n
tatio
n
u
sin
g
wa
v
e
let
tran
sf
o
r
m
,
"
Pa
tt
e
rn
Rec
o
g
n
it
i
o
n
L
e
tt
e
rs
,
v
o
l.
2
4
,
p
p
.
3
1
9
7
-
3
2
0
3
,
2
0
0
3
.
[3
]
G
.
M
u
rth
y
a
n
d
T
.
V
e
e
rra
ju
,
"
A
n
o
v
e
l
a
p
p
ro
a
c
h
b
a
se
d
o
n
d
e
c
re
a
se
d
d
im
e
n
sio
n
a
n
d
re
d
u
c
e
d
g
ra
y
le
v
e
l
ra
n
g
e
m
a
tri
x
f
e
a
t
u
re
s
f
o
r
sto
n
e
tex
tu
re
c
las
sif
i
c
a
ti
o
n
,
"
In
ter
n
a
t
io
n
a
l
J
o
u
r
n
a
l
o
f
El
e
c
trica
l
a
n
d
Co
mp
u
ter
En
g
in
e
e
rin
g
,
v
o
l.
7
,
p
p
.
2
5
0
2
,
2
0
1
7
.
[4
]
J.
T
o
o
,
A
.
A
b
d
u
ll
a
h
,
N.
M
o
h
d
S
a
a
d
,
N.
M
o
h
d
A
li
,
a
n
d
H.
M
u
sa
,
"
A
d
e
tail
stu
d
y
o
f
wa
v
e
let
f
a
m
il
ies
f
o
r
EM
G
p
a
tt
e
rn
re
c
o
g
n
it
i
o
n
,
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
,
v
o
l
.
8
,
p
p
.
4
2
2
1
-
4
2
2
9
,
2
0
1
8
.
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:
2
0
8
8
-
8708
Op
timiz
a
tio
n
o
f d
is
crete
w
a
ve
l
et
tr
a
n
s
fo
r
m
fea
tu
r
es u
s
in
g
a
r
t
ificia
l b
ee
co
lo
n
y
a
lg
o
r
ith
m…
(
F
th
i M.
A
lb
ko
s
h
)
5261
[5
]
E.
A
v
c
i,
I.
T
u
rk
o
g
lu
,
a
n
d
M
.
P
o
y
ra
z
,
"
In
telli
g
e
n
t
targ
e
t
re
c
o
g
n
it
io
n
b
a
se
d
o
n
w
a
v
e
let
p
a
c
k
e
t
n
e
u
ra
l
n
e
tw
o
rk
,
"
Exp
e
rt S
y
ste
ms
wit
h
A
p
p
l
ica
ti
o
n
s,
v
o
l.
2
9
,
p
p
.
1
7
5
-
1
8
2
,
2
0
0
5
.
[6
]
J.
A
.
R.
Re
c
io
,
L
.
A
.
R.
F
e
rn
á
n
d
e
z
,
a
n
d
A
.
F
e
rn
á
n
d
e
z
-
S
a
rriá,
"
Us
e
o
f
G
a
b
o
r
f
il
ters
f
o
r
te
x
tu
re
c
las
si
f
ica
ti
o
n
o
f
d
ig
it
a
l
im
a
g
e
s,"
Fí
sic
a
d
e
la
T
ier
ra
,
v
o
l.
1
7
,
p
.
4
7
,
2
0
0
5
.
[7
]
S
.
E.
G
rig
o
re
sc
u
,
N.
P
e
tk
o
v
,
a
n
d
P
.
Kru
izi
n
g
a
,
"
Co
mp
a
riso
n
o
f
tex
tu
re
fea
tu
re
s
b
a
se
d
o
n
G
a
b
o
r
fi
lt
e
rs
,
"
Im
a
g
e
P
r
o
c
e
ss
in
g
,
IEE
E
T
ra
n
s
a
c
ti
o
n
s o
n
,
v
o
l.
1
1
,
p
p
.
1
1
6
0
-
1
1
6
7
,
2
0
0
2
.
[8
]
K.
S
.
V
i
d
y
a
,
E.
Ng
,
U.
R.
A
c
h
a
ry
a
,
S
.
M
.
Ch
o
u
,
R.
S
a
n
T
a
n
,
a
n
d
D.
N.
G
h
ista,
"
Co
m
p
u
ter
-
a
id
e
d
d
iag
n
o
sis
o
f
M
y
o
c
a
rd
ial
In
fa
rc
ti
o
n
u
sin
g
u
lt
r
a
s
o
u
n
d
im
a
g
e
s
w
it
h
D
WT
,
GLCM
a
n
d
HO
S
m
e
th
o
d
s:
A
c
o
m
p
a
ra
ti
v
e
stu
d
y
,
"
Co
mp
u
ter
s i
n
b
io
l
o
g
y
a
n
d
me
d
ici
n
e
,
v
o
l.
6
2
,
p
p
.
8
6
-
9
3
,
2
0
1
5
.
[9
]
P
.
Hire
m
a
th
a
n
d
S
.
S
h
iv
a
sh
a
n
k
a
r,
"
Wav
e
l
e
t
b
a
se
d
c
o
-
o
c
c
u
rre
n
c
e
h
isto
g
ra
m
f
e
a
tu
re
s
f
o
r
tex
tu
re
c
la
ss
if
ic
a
ti
o
n
w
it
h
a
n
a
p
p
li
c
a
ti
o
n
to
sc
rip
t
id
e
n
ti
f
ica
ti
o
n
i
n
a
d
o
c
u
m
e
n
t
i
m
a
g
e
,
"
Pa
tt
e
rn
Rec
o
g
n
it
io
n
L
e
tt
e
rs
,
v
o
l.
2
9
,
p
p
.
1
1
8
2
-
1
1
8
9
,
2
0
0
8
.
[1
0
]
Ü.
H.
A
tas
e
v
e
r,
C.
Öz
k
a
n
,
a
n
d
F
.
S
u
n
a
r,
"
T
h
e
u
se
o
f
a
rti
f
icia
l
in
tel
li
g
e
n
c
e
o
p
ti
m
iza
ti
o
n
a
lg
o
rit
h
m
s
in
u
n
su
p
e
rv
ise
d
c
las
si
f
ica
ti
o
n
,
"
e
d
:
EA
RS
e
L
,
2
0
1
1
.
[1
1
]
X
.
L
iu
a
n
d
D.
W
a
n
g
,
"
Tex
tu
re
c
las
sif
ic
a
ti
o
n
u
si
n
g
sp
e
c
tral
h
isto
g
ra
m
s,"
IEE
E
tra
n
sa
c
ti
o
n
s
o
n
ima
g
e
p
ro
c
e
ss
in
g
,
v
o
l.
1
2
,
p
p
.
6
6
1
-
6
7
0
,
2
0
0
3
.
[1
2
]
X
.
Zh
a
n
g
,
N.
H.
Y
o
u
n
a
n
,
a
n
d
R.
Kin
g
,
"
S
o
il
tex
tu
re
c
la
ss
if
i
c
a
ti
o
n
u
si
n
g
wa
v
e
let
tra
n
sfo
rm
a
n
d
ma
x
i
m
u
m
li
k
e
li
h
o
o
d
a
p
p
ro
a
c
h
,
"
in
G
e
o
sc
ie
n
c
e
a
n
d
Re
m
o
te
S
e
n
sin
g
S
y
m
p
o
siu
m
,
2
0
0
3
.
IG
A
RS
S
'
0
3
.
P
ro
c
e
e
d
i
n
g
s.
2
0
0
3
IEE
E
In
tern
a
ti
o
n
a
l
,
p
p
.
2
8
8
8
-
2
8
9
0
,
2
0
0
3
.
[1
3
]
K.
M
.
Ra
jp
o
o
t
a
n
d
N.
M
.
Ra
j
p
o
o
t
,
"
W
a
v
e
lets
a
n
d
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
s
f
o
r
tex
tu
re
c
la
ss
if
ica
ti
o
n
,
"
in
M
u
lt
i
to
p
ic
Co
n
f
e
re
n
c
e
,
2
0
0
4
.
P
ro
c
e
e
d
in
g
s o
f
INMI
C
2
0
0
4
.
8
th
In
tern
a
ti
o
n
a
l
,
p
p
.
3
2
8
-
3
3
3
,
2
0
0
4
.
[1
4
]
S
.
S
i
d
h
u
a
n
d
K.
Ra
a
h
e
m
i
f
a
r,
"
T
e
x
tu
re
c
las
sif
i
c
a
ti
o
n
u
si
n
g
w
a
v
e
let
tran
sf
o
r
m
a
n
d
su
p
p
o
r
t
v
e
c
to
r
m
a
c
h
in
e
s,"
in
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
,
Ca
n
a
d
ia
n
Co
n
f
e
re
n
c
e
o
n
,
2
0
0
5
,
p
p
.
9
4
1
-
9
4
4
,
2
0
0
5
.
[1
5
]
A
.
K.
S
in
g
h
,
S
.
T
i
w
a
ri,
a
n
d
V
.
S
h
u
k
la,
"
W
a
v
e
let
b
a
se
d
m
u
lt
i
c
l
a
ss
i
m
a
g
e
c
las
si
f
ica
ti
o
n
u
sin
g
n
e
u
ra
l
n
e
tw
o
rk
,
"
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
C
o
mp
u
ter
Ap
p
l
ica
ti
o
n
s,
v
o
l.
3
7
,
p
p
.
2
1
-
2
5
,
2
0
1
2
.
[1
6
]
A
.
Ja
y
a
su
d
h
a
a
n
d
D.
P
u
g
a
z
h
e
n
t
h
i,
"
Co
l
o
u
r
tex
tu
re
c
la
ss
if
ica
t
io
n
u
sin
g
w
a
v
e
let
tra
n
sf
o
rm
fro
m
i
ts
g
ra
y
sc
a
le,
"
in
Cu
rre
n
t
T
re
n
d
s
in
E
n
g
in
e
e
rin
g
a
n
d
T
e
c
h
n
o
l
o
g
y
(ICC
T
E
T
),
2
0
1
4
2
n
d
I
n
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
,
p
p
.
3
7
6
-
379
,
2
0
1
4
.
[1
7
]
A
.
M
o
jsil
o
v
ić,
M
.
V
.
P
o
p
o
v
ić,
a
n
d
D.
M
.
Ra
c
k
o
v
,
"
On
th
e
se
l
e
c
ti
o
n
o
f
a
n
o
p
ti
m
a
l
w
a
v
e
let
b
a
sis
f
o
r
te
x
tu
re
c
h
a
ra
c
teriz
a
ti
o
n
,
"
Ima
g
e
Pro
c
e
ss
i
n
g
,
IEE
E
T
ra
n
s
a
c
ti
o
n
s o
n
,
v
o
l
.
9
,
p
p
.
2
0
4
3
-
2
0
5
0
,
2
0
0
0
.
[1
8
]
A
.
Bu
sc
h
a
n
d
W
.
W
.
Bo
les
,
"
T
e
x
tu
re
c
la
ss
if
ica
ti
o
n
u
sin
g
mu
lt
ip
l
e
wa
v
e
let
a
n
a
lys
is,"
in
P
ro
c
e
e
d
i
n
g
s
o
f
th
e
S
ix
th
Dig
it
a
l
Im
a
g
e
Co
m
p
u
ti
n
g
:
T
e
c
h
n
iq
u
e
s a
n
d
A
p
p
li
c
a
ti
o
n
s Co
n
f
e
re
n
c
e
,
p
p
.
3
4
1
-
3
4
5
,
2
0
0
2
.
[1
9
]
D.
Ka
ra
b
o
g
a
,
"
A
n
id
e
a
b
a
se
d
o
n
h
o
n
e
y
b
e
e
s
w
a
r
m
f
o
r
n
u
m
e
ric
a
l
o
p
ti
m
iza
ti
o
n
,
"
T
e
c
h
n
ica
l
re
p
o
rt
-
tr0
6
,
Erc
iye
s
u
n
ive
rs
it
y
,
e
n
g
in
e
e
rin
g
fa
c
u
lt
y
,
c
o
mp
u
ter
e
n
g
in
e
e
rin
g
d
e
p
a
rtme
n
t,
2
0
0
5
.
[2
0
]
D.
Ka
ra
b
o
g
a
,
B.
G
o
rk
e
m
li
,
C.
Oz
tu
rk
,
a
n
d
N.
Ka
ra
b
o
g
a
,
"
A
c
o
m
p
re
h
e
n
siv
e
su
rv
e
y
:
a
rti
f
ici
a
l
b
e
e
c
o
lo
n
y
(A
BC)
a
lg
o
rit
h
m
a
n
d
a
p
p
l
ica
ti
o
n
s,"
Arti
f
icia
l
In
telli
g
e
n
c
e
Rev
iew,
v
o
l.
4
2
,
p
p
.
2
1
-
5
7
,
2
0
1
4
.
[2
1
]
S
.
S
h
a
n
t
h
i
a
n
d
V
.
M
.
B
h
a
sk
a
ra
n
,
"
M
o
d
if
ied
a
rti
f
icia
l
b
e
e
c
o
lo
n
y
b
a
se
d
f
e
a
tu
re
se
lec
ti
o
n
:
a
n
e
w
m
e
th
o
d
i
n
th
e
a
p
p
l
ica
ti
o
n
o
f
m
a
m
m
o
g
ra
m
i
m
a
g
e
c
l
a
ss
i
f
ica
ti
o
n
,
"
In
t.
J
.
S
c
i.
E
n
g
.
T
e
c
h
n
o
l.
Res
,
v
o
l.
3
,
p
p
.
1
6
6
4
-
1
6
6
7
,
2
0
1
4
.
[2
2
]
X
.
L
i
a
n
d
L
.
L
i,
"
P
re
f
e
re
n
c
e
m
u
lt
i
-
o
b
jec
ti
v
e
a
rti
f
ici
a
l
b
e
e
c
o
lo
n
y
a
n
d
it
s
a
p
p
li
c
a
ti
o
n
in
c
a
m
e
ll
ia
f
ru
it
ima
g
e
re
c
o
g
n
it
io
n
,
"
Ap
p
li
c
a
ti
o
n
Res
e
a
rc
h
o
f
Co
mp
u
ter
s,
v
o
l.
1
2
,
p
.
1
0
0
,
2
0
1
2
.
[2
3
]
S
.
Ba
n
e
rjee
,
A
.
Bh
a
ra
d
w
a
j,
D.
G
u
p
ta,
a
n
d
V
.
P
a
n
c
h
a
l,
"
Re
m
o
te
se
n
sin
g
i
m
a
g
e
c
las
si
f
ic
a
ti
o
n
u
si
n
g
a
rti
f
ici
a
l
b
e
e
c
o
lo
n
y
a
lg
o
rit
h
m
,
"
In
t.
J
.
Co
m
p
u
t
.
S
c
i.
I
n
f,
v
o
l.
2
,
p
p
.
6
7
-
7
2
,
2
0
1
2
.
[2
4
]
M
.
A
.
S
h
n
a
n
a
n
d
T
.
H.
Ra
ss
e
m
,
"
F
a
c
ial
I
m
a
g
e
Re
tri
e
v
a
l
o
n
S
e
m
a
n
ti
c
F
e
a
tu
re
s
Us
in
g
A
d
a
p
ti
v
e
Ge
n
e
ti
c
A
l
g
o
rit
h
m
,
"
In
fo
rm
a
ti
c
a
Eco
n
o
mi
c
a
,
v
o
l.
2
2
,
2
0
1
8
.
[2
5
]
D.
Krie
se
l,
"
A
b
rief
In
tro
d
u
c
ti
o
n
o
n
Ne
u
ra
l
Ne
tw
o
rk
s,"
2
0
0
7
.
[2
6
]
A
.
M
o
n
a
d
jem
i,
"
T
o
w
a
rd
s e
ff
icie
n
t
tex
tu
re
c
las
sif
ic
a
ti
o
n
a
n
d
a
b
n
o
r
m
a
li
t
y
d
e
tec
ti
o
n
,
"
Un
iv
e
rsity
o
f
B
risto
l,
2
0
0
4
.
[2
7
]
O.
S
a
li
m
a
,
A
.
T
a
leb
-
A
h
m
e
d
,
a
n
d
B.
M
o
h
a
m
e
d
,
"
S
p
a
ti
a
l
in
f
o
r
m
a
ti
o
n
b
a
se
d
im
a
g
e
c
lu
ste
rin
g
w
it
h
a
s
w
a
r
m
a
p
p
ro
a
c
h
,
"
I
a
e
s In
t.
J
.
Arti
f.
In
tell
.
(
Ij
-
Ai
),
v
o
l.
1
,
p
p
.
1
4
9
-
1
6
0
,
2
0
1
2
.
[2
8
]
S
.
Ba
n
e
rjee
,
A
.
Bh
a
ra
d
w
a
j,
D.
G
u
p
ta,
a
n
d
V
.
P
a
n
c
h
a
l,
"
Re
m
o
te
se
n
sin
g
i
m
a
g
e
c
las
si
f
ic
a
ti
o
n
u
si
n
g
a
rti
f
ici
a
l
b
e
e
c
o
lo
n
y
a
lg
o
rit
h
m
,
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
In
fo
rm
a
ti
c
s,
v
o
l.
2
,
p
p
.
6
7
-
7
2
,
2
0
1
2
.
[2
9
]
M
.
S
.
Uz
e
r,
N.
Yilm
a
z
,
a
n
d
O.
In
a
n
,
"
F
e
a
tu
re
se
lec
ti
o
n
m
e
th
o
d
b
a
se
d
o
n
a
rti
f
icia
l
b
e
e
c
o
lo
n
y
a
lg
o
rit
h
m
a
n
d
su
p
p
o
r
t
v
e
c
to
r
m
a
c
h
in
e
s f
o
r
m
e
d
ica
l
d
a
tas
e
ts cla
ss
i
f
ica
ti
o
n
,
"
T
h
e
S
c
ien
t
if
ic
W
o
rld
J
o
u
rn
a
l,
v
o
l
.
2
0
1
3
,
2
0
1
3
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
Fth
i
M
.
Alb
k
o
s
h
o
b
tain
e
d
h
is
Ba
c
h
e
lo
r
o
f
Co
m
p
u
ter
E
n
g
in
e
e
rin
g
f
ro
m
th
e
Hig
h
In
stit
u
te
o
f
In
d
u
stry
,
L
ib
y
a
,
a
n
d
M
a
ste
r
d
e
g
re
e
in
th
e
Co
m
m
u
n
ica
ti
o
n
a
n
d
Co
m
p
u
ter
E
n
g
in
e
e
rin
g
f
ro
m
Un
iv
e
rsiti
Ke
b
a
n
g
sa
a
n
M
a
la
y
sia
.
He
is
c
u
rre
n
tl
y
a
P
h
.
D
stu
d
e
n
t
a
t
th
e
S
c
h
o
o
l
o
f
In
f
o
r
m
a
ti
c
s
a
n
d
A
p
p
li
e
d
M
a
th
e
m
a
ti
c
s,
Un
iv
e
rsiti
M
a
la
y
sia
T
e
r
e
n
g
g
a
n
u
.
His
re
se
a
r
c
h
in
tere
sts
in
th
e
a
re
a
o
f
i
m
a
g
e
p
ro
c
e
ss
in
g
a
n
d
p
a
tt
e
rn
re
c
o
g
n
it
io
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
2
5
3
-
5
2
6
2
5262
M
u
h
a
m
m
a
d
S
u
z
u
r
i
H
it
a
m
o
b
tain
e
d
h
is
Ba
c
h
e
lo
r
o
f
In
d
u
strial
T
e
c
h
n
o
l
o
g
y
(B.
Tec
h
(Ho
n
s.)
f
r
o
m
Un
iv
e
rsiti
S
a
in
s
M
a
lay
sia
a
n
d
P
h
.
D
d
e
g
re
e
f
ro
m
Un
iv
e
rsit
y
o
f
L
e
e
d
s,
Un
it
e
d
Kin
g
d
o
m
.
He
is
c
u
rre
n
tl
y
a
P
ro
f
e
so
r
a
t
th
e
S
c
h
o
o
l
o
f
In
f
o
r
m
a
ti
c
s
a
n
d
A
p
p
li
e
d
M
a
th
e
m
a
ti
c
s,
Un
iv
e
rsiti
M
a
l
a
y
sia
T
e
re
n
g
g
a
n
u
.
His
re
se
a
rc
h
in
tere
s
ts
a
re
f
o
c
u
se
d
o
n
im
a
g
e
p
ro
c
e
ss
in
g
,
c
o
m
p
u
tatio
n
a
l
in
tell
ig
e
n
c
e
a
n
d
ro
b
o
ti
c
s.
Wa
n
N
u
r
a
l
J
a
w
a
h
ir
H
j
W
a
n
Yu
ss
o
f
re
c
e
iv
e
d
h
e
r
B.
IT
in
S
o
ftw
a
r
e
En
g
in
e
e
rin
g
a
n
d
M
.
S
c
.
i
n
A
rti
f
icia
l
In
telli
g
e
n
c
e
f
ro
m
Ko
lej
Un
iv
e
rsiti
S
a
in
s
d
a
n
T
e
k
n
o
lo
g
i
M
a
la
y
sia
.
In
2
0
1
4
,
sh
e
o
b
tain
e
d
h
e
r
P
h
.
D.
f
ro
m
Un
iv
e
rsit
i
M
a
lay
sia
T
e
re
n
g
g
a
n
u
.
S
h
e
is
c
u
rre
n
tl
y
a
se
n
io
r
lec
tu
re
r
a
t
S
c
h
o
o
l
o
f
In
f
o
rm
a
ti
c
s
a
n
d
A
p
p
li
e
d
M
a
th
e
m
a
ti
c
s,
Un
iv
e
rsiti
M
a
la
y
sia
T
e
r
e
n
g
g
a
n
u
.
He
r
re
s
e
a
rc
h
in
tere
sts
a
re
in
2
D/
3
D i
m
a
g
e
a
n
a
l
y
sis a
n
d
u
n
d
e
rwa
ter v
id
e
o
p
ro
c
e
ss
in
g
.
Abd
u
l
Az
iz
K
Abd
u
l
H
a
m
i
d
i
s
a
se
n
io
r
l
e
c
tu
re
r
at
th
e
S
c
h
o
o
l
o
f
In
f
o
rm
a
ti
c
s
a
n
d
A
p
p
li
e
d
M
a
th
e
m
a
ti
c
s
,
Un
iv
e
rsiti
M
a
la
y
sia
Tere
n
g
g
a
n
u
.
He
re
c
e
i
v
e
d
h
is
P
h
.
D
in
c
o
m
p
u
ter
sc
ien
c
e
f
ro
m
Un
iv
e
rsiti
Ke
b
a
n
g
sa
a
n
M
a
la
y
sia
in
2
0
1
4
a
n
d
h
e
is
a
c
ti
v
e
l
y
in
v
o
lv
e
d
in
a
m
u
lt
it
u
d
e
o
f
tec
h
n
ica
l
p
ro
jec
ts
a
c
ro
ss
a
b
ro
a
d
ra
n
g
e
o
f
se
c
to
rs,
in
c
lu
d
in
g
f
in
a
n
c
ial,
m
a
n
u
f
a
c
tu
rin
g
,
p
o
rt
m
a
n
a
g
e
m
e
n
t
a
n
d
e
d
u
c
a
ti
o
n
.
His res
e
a
rc
h
in
tere
st i
n
c
lu
d
e
im
a
g
e
p
ro
c
e
ss
in
g
,
p
a
tt
e
rn
r
e
c
o
g
n
it
io
n
a
n
d
d
a
ta s
c
ien
c
e
.
Ro
z
n
iza
Ali
o
b
tain
e
d
B.
S
c
f
ro
m
Ko
lej
Un
iv
e
rsiti
S
u
lt
a
n
Zai
n
a
l
A
b
id
in
,
M
S
c
.
f
ro
m
Un
iv
e
rsiti
T
e
k
n
o
lo
g
i
M
A
RA
a
n
d
P
h
.
D
i
n
C
o
m
p
u
ter
S
c
ien
c
e
f
ro
m
S
ti
rli
n
g
U
n
iv
e
rsity
,
Un
it
e
d
Kin
g
d
o
m
.
No
w
sh
e
se
rv
e
d
a
t
Un
iv
e
rsiti
M
a
la
y
sia
Tere
n
g
g
a
n
u
a
s
se
n
io
r
lec
tu
re
r
a
t
th
e
S
c
h
o
o
l
o
f
In
f
o
r
m
a
ti
c
s
a
n
d
A
p
p
li
e
d
M
a
th
e
m
a
ti
c
s.
He
r
re
se
a
r
c
h
a
re
a
o
f
in
tere
st i
s o
n
im
a
g
e
p
ro
c
e
ss
in
g
a
n
d
p
a
tt
e
rn
re
c
o
g
n
i
ti
o
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.