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.
8
,
No
.
2
,
A
p
r
il
201
8
,
p
p
.
971
~
9
7
8
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v8
i
1
.
p
p
9
7
1
-
9
7
8
971
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
I
m
a
g
e Denois
ing
by
using
Mo
difie
d SGH
P
Alg
o
rith
m
Sree
dh
a
r
K
o
lle
m
1
,
K
.
Ra
m
a
l
ing
a
Reddy
2
,
D.
Sree
niv
a
s
a
Ra
o
3
1
De
p
a
rtme
n
t
o
f
El
e
c
tro
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
E
n
g
in
e
e
rin
g
,
S
R
E
n
g
in
e
e
rin
g
Co
ll
e
g
e
,
W
a
ra
n
g
a
l,
T
e
lan
g
a
n
a
,
In
d
ia
2
De
p
a
rtme
n
t
o
f
El
e
c
tro
n
ics
a
n
d
T
e
le
m
a
ti
c
s E
n
g
in
e
e
rin
g
,
G
.
Na
ra
y
a
n
a
m
m
a
In
stit
u
te o
f
T
e
c
h
n
o
lo
g
y
a
n
d
S
c
ien
c
e
Hy
d
e
ra
b
a
d
,
T
e
lan
g
a
n
a
,
In
d
ia
3
De
p
a
rtme
n
t
o
f
El
e
c
tro
n
ics
a
n
d
C
o
m
m
u
n
ica
ti
o
n
E
n
g
in
e
e
rin
g
,
JN
T
UH
CEH,
Ku
k
a
tp
a
ll
y
,
H
y
d
e
ra
b
a
d
,
T
e
lan
g
a
n
a
,
In
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
2
0
,
2
0
1
7
R
ev
i
s
ed
Dec
2
7
,
2
0
1
7
A
cc
ep
ted
J
an
7
,
2
0
1
8
In
re
a
l
ti
m
e
a
p
p
li
c
a
ti
o
n
s,
im
a
g
e
d
e
n
o
isi
n
g
is
a
p
re
d
o
m
in
a
n
t
tas
k
.
T
h
is
tas
k
m
a
k
e
s
a
d
e
q
u
a
te
p
re
p
a
ra
ti
o
n
f
o
r
im
a
g
e
s
lo
o
k
s
p
ro
m
in
e
n
t.
B
u
t
th
e
re
a
re
se
v
e
r
a
l
d
e
n
o
isin
g
a
lg
o
rit
h
m
s
a
n
d
e
v
e
r
y
a
l
g
o
rit
h
m
h
a
s
it
s
o
w
n
d
isti
n
c
ti
v
e
a
tt
rib
u
te
b
a
se
d
u
p
o
n
d
if
f
e
r
e
n
t
n
a
tu
ra
l
im
a
g
e
s.
In
th
is
p
a
p
e
r,
w
e
p
ro
p
o
se
d
a
p
e
rs
p
e
c
ti
v
e
th
a
t
is
m
o
d
if
ied
p
a
ra
m
e
ter
in
S
-
G
r
a
d
ien
t
Histo
g
ra
m
P
re
se
rv
a
ti
o
n
d
e
n
o
isi
n
g
m
e
th
o
d
.
S
-
G
ra
d
ien
t
Histo
g
ra
m
P
re
se
rv
a
ti
o
n
is
a
m
e
th
o
d
t
o
c
o
m
p
u
te
th
e
stru
c
tu
re
g
ra
d
ien
t
h
isto
g
ra
m
f
ro
m
th
e
n
o
isy
o
b
se
r
v
a
ti
o
n
b
y
tak
in
g
d
iff
e
re
n
t
n
o
ise
sta
n
d
a
rd
d
e
v
iatio
n
s
o
f
d
if
f
e
re
n
t
i
m
a
g
e
s.
T
h
e
p
e
rf
o
r
m
a
n
c
e
o
f
th
is
m
e
th
o
d
is
e
n
u
m
e
ra
ted
in
term
s
o
f
p
e
a
k
si
g
n
a
l
to
n
o
ise
ra
ti
o
a
n
d
str
u
c
tu
ra
l
sim
il
a
rit
y
in
d
e
x
o
f
a
p
a
rti
c
u
lar
i
m
a
g
e
.
In
th
is
p
a
p
e
r,
m
a
in
ly
f
o
c
u
s
o
n
p
e
a
k
sig
n
a
l
to
n
o
ise
ra
ti
o
,
stru
c
tu
ra
l
sim
il
a
rit
y
in
d
e
x
,
n
o
ise
e
sti
m
a
ti
o
n
a
n
d
a
m
e
a
su
re
o
f
stru
c
tu
re
g
ra
d
ien
t
h
isto
g
ra
m
o
f
a
g
i
v
e
n
im
a
g
e
.
K
ey
w
o
r
d
:
Gr
ad
ien
t h
i
s
to
g
r
a
m
No
is
e
esti
m
atio
n
P
r
in
cip
al
co
m
p
o
n
e
n
t
a
n
al
y
s
is
P
SNR
S
-
GHP
SS
I
M
Co
p
y
rig
h
t
©
2
0
1
8
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
:
Sre
ed
h
ar
Ko
lle
m
Dep
ar
t
m
en
t o
f
E
lectr
o
n
ics a
n
d
C
o
m
m
u
n
icat
io
n
E
n
g
i
n
ee
r
in
g
,
SR
E
n
g
i
n
ee
r
in
g
C
o
lle
g
e
,
W
ar
an
g
al,
T
elan
g
a
n
a
,
I
n
d
ia.
E
m
ail:
k
s
r
ee
d
h
ar
4
4
6
@
g
m
ail.
c
o
m
1.
I
NT
RO
D
UCT
I
O
N
I
m
ag
e
s
af
f
ec
ted
by
u
n
w
a
n
ted
n
o
is
e
f
r
o
m
d
if
f
er
en
t
s
o
u
r
ce
s
lik
e
tr
ad
itio
n
al
f
il
m
ca
m
er
as
an
d
d
ig
ital
ca
m
er
as.
T
h
ese
n
o
i
s
e
ele
m
en
t
s
w
ill
cr
ea
te
s
o
m
e
s
er
io
u
s
i
s
s
u
es
f
o
r
f
u
r
th
er
p
r
o
ce
s
s
i
n
g
o
f
i
m
a
g
es
i
n
p
r
ac
tical
ap
p
licatio
n
s
s
u
c
h
as
co
m
p
u
t
er
v
is
io
n
,
ar
tis
tic
w
o
r
k
o
r
m
ar
k
etin
g
a
n
d
also
i
n
m
an
y
f
ield
s
.
So
,
d
if
f
er
en
t
class
i
f
icatio
n
o
f
n
o
i
s
es
li
k
es
s
alt
an
d
p
ep
p
er
,
Gau
s
s
ian
,
s
h
o
t
an
d
q
u
an
tizat
io
n
.
I
n
s
al
t
an
d
p
ep
p
er
n
o
is
e,
all
th
e
i
m
a
g
es
ar
e
co
n
s
tr
u
cted
w
i
th
p
ix
els
i
n
a
t
w
o
-
d
i
m
e
n
s
io
n
a
l
ar
r
a
y
.
I
n
th
a
t
p
ix
e
l
to
p
i
x
el
,
th
e
d
i
f
f
er
en
ce
is
o
b
s
er
v
ed
w
h
e
n
th
e
i
m
ag
e
is
af
f
ec
ted
by
n
o
i
s
e
th
at
is
in
te
r
m
s
o
f
in
te
n
s
it
y
o
f
n
eig
h
b
o
u
r
in
g
p
ix
e
ls
.
So
,
it
is
id
en
ti
f
ied
p
ix
els
a
n
d
n
ei
g
h
b
o
u
r
in
g
p
ix
el
s
o
n
l
y
th
e
s
m
al
l
n
u
m
b
er
o
f
p
ix
el
s
is
a
f
f
ec
ted
in
an
i
m
a
g
e.
T
h
e
s
alt
an
d
p
ep
p
er
n
o
is
e
i
s
clea
r
l
y
id
en
ti
f
ied
i
n
a
n
i
m
a
g
e
b
y
it
co
n
tain
s
b
lack
an
d
w
h
i
te
s
p
ec
k
le
s
.
W
h
e
n
w
e
v
ie
w
ed
an
i
m
ag
e
w
h
ic
h
is
a
f
f
ec
ted
by
s
alt
an
d
p
ep
p
er
n
o
is
e,
th
e
i
m
ag
e
co
n
tai
n
s
b
lac
k
an
d
w
h
ite
d
o
ts
,
h
en
ce
it
ter
m
s
as salt an
d
p
ep
p
er
n
o
is
e.
I
n
Ga
u
s
s
ian
n
o
i
s
e,
n
o
i
s
y
p
ix
el
v
al
u
e
w
i
ll
b
e
a
s
m
all
c
h
a
n
g
e
o
f
th
e
o
r
ig
i
n
al
v
al
u
e
o
f
a
p
ix
el.
A
d
iag
r
a
m
co
n
s
i
s
ti
n
g
o
f
r
ec
tan
g
les
w
h
o
s
e
ar
ea
is
p
r
o
p
o
r
tio
n
al
to
th
e
f
r
eq
u
en
c
y
o
f
a
v
ar
iab
le
o
r
P
SNR
a
n
d
w
h
o
s
e
w
id
t
h
i
s
eq
u
al
to
t
h
e
d
if
f
er
e
n
t
n
o
is
e
s
ta
n
d
ar
d
d
ev
iatio
n
s
is
a
h
i
s
to
g
r
a
m
.
Ot
h
er
G
au
s
s
ian
m
o
d
els
ar
e
p
r
e
s
en
t
m
ai
n
l
y
d
ep
en
d
s
u
p
o
n
th
e
ce
n
tr
al
li
m
it
t
h
eo
r
e
m
s
h
o
w
s
th
at
ad
d
itio
n
o
f
d
i
f
f
er
e
n
t
n
o
is
es
f
r
o
m
d
i
f
f
er
e
n
t
s
o
u
r
ce
s
to
ass
o
ciate
d
w
it
h
Ga
u
s
s
ian
d
is
tr
ib
u
tio
n
.
Den
o
is
i
n
g
o
f
an
i
m
ag
e
i
n
v
o
l
v
es
t
h
e
m
a
n
ip
u
la
tio
n
o
f
t
h
e
im
ag
e
d
ata
to
p
r
o
d
u
ce
a
v
is
u
all
y
h
i
g
h
-
q
u
alit
y
i
m
a
g
e.
T
h
er
e
ar
e
n
u
m
er
o
u
s
m
o
d
els
th
at
h
av
e
b
ee
n
p
u
b
lis
h
ed
s
o
f
ar
w
h
ic
h
ar
e
u
s
ed
f
o
r
d
en
o
is
in
g
an
i
m
a
g
e
[
1
]
.
Sp
ar
s
e
r
ep
r
esen
tat
io
n
f
o
r
i
m
a
g
e
r
es
to
r
atio
n
[
2
]
,
[
3
]
,
T
o
tal
v
ar
iatio
n
m
o
d
el
[
4
]
,
W
av
elet
-
b
ased
m
o
d
el
[
5
]
,
B
M3
D
[
6
]
m
o
d
el
an
d
h
is
to
g
r
a
m
p
r
eser
v
atio
n
al
g
o
r
ith
m
[
7
]
ar
e
s
o
m
e
o
f
th
e
m
.
E
ac
h
m
et
h
o
d
h
as
it
s
o
w
n
ch
ar
ac
ter
is
tic
s
,
b
en
ef
it
a
n
d
also
d
em
er
it.
T
w
o
m
aj
o
r
c
la
s
s
es
o
f
d
en
o
is
in
g
m
et
h
o
d
s
ar
e
(
a)
m
o
d
el
b
ased
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.
8
,
No
.
2
,
A
p
r
il 2
0
1
8
:
9
7
1
–
9
7
8
972
an
d
(b
)
L
ea
r
n
in
g
-
b
ased
m
eth
o
d
.
I
n
th
e
m
o
d
el
,
b
ased
m
e
th
o
d
,
a
s
tatis
tical/
m
at
h
e
m
atica
l
m
o
d
el
w
il
l
b
e
u
s
ed
f
o
r
th
e
d
en
o
is
i
n
g
.
W
h
er
ea
s
i
n
L
ea
r
n
in
g
b
ased
m
et
h
o
d
,
an
alg
o
r
it
h
m
w
il
l
b
e
tr
ain
ed
b
y
u
s
i
n
g
s
u
f
f
icie
n
t
p
ar
am
eter
s
an
d
th
e
n
t
h
e
m
o
d
el
is
allo
w
ed
to
w
o
r
k
b
ased
o
n
it
s
w
e
ig
h
ta
g
e
f
u
n
ctio
n
[
8
]
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
I
n
th
e
p
r
esen
t
w
o
r
k
,
t
h
e
d
en
o
i
s
in
g
is
d
o
n
e
i
n
a
m
o
r
e
r
ea
lis
ti
c
w
a
y
a
s
in
p
r
ac
tical
s
it
u
atio
n
s
,
o
n
l
y
t
h
e
n
o
is
y
i
m
ag
e
w
ill b
e
a
v
ailab
le.
A
n
o
is
y
i
m
a
g
e
i
s
ta
k
en
a
s
i
n
p
u
t to
t
h
e
al
g
o
r
ith
m
i
s
s
h
o
w
n
i
n
F
ig
u
r
e
1
.
W
e
h
a
v
e
ad
o
p
ted
p
atch
-
b
ased
n
o
is
e
le
v
el
esti
m
a
tio
n
al
g
o
r
ith
m
b
y
Xi
n
h
ao
L
i
u
et
al
[
9
]
.
P
atch
e
s
ar
e
g
en
er
ated
f
r
o
m
t
h
e
s
in
g
le
n
o
is
y
i
m
a
g
e
a
n
d
it
s
w
ea
k
te
x
t
u
r
ed
p
atch
e
s
ar
e
id
en
ti
f
ied
.
T
h
e
No
is
e
lev
el
is
esti
m
ated
f
r
o
m
t
h
e
P
r
in
cip
al
C
o
m
p
o
n
en
t
A
n
al
y
s
is
[
1
0
]
,
[
1
1
]
.
Fig
u
r
e
1
.
Flo
w
c
h
ar
t
o
f
t
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
I
n
m
o
s
t
o
f
th
e
d
en
o
i
s
in
g
m
et
h
o
d
,
it
is
s
ee
n
th
at,
a
f
ter
its
i
m
p
le
m
en
ta
tio
n
,
t
h
e
i
m
a
g
e
w
ill
b
e
b
lu
r
r
ed
th
an
t
h
at
o
f
th
e
o
r
ig
i
n
al
i
m
ag
e.
A
ls
o
,
th
e
ed
g
e
o
f
th
e
d
en
o
i
s
ed
i
m
ag
e
g
ets
s
m
o
o
th
e
n
ed
an
d
w
il
l
h
av
e
le
s
s
er
d
etails
th
a
n
t
h
at
o
f
th
e
o
r
ig
in
a
l
i
m
a
g
e.
A
s
t
u
d
y
h
as
b
ee
n
co
n
d
u
cted
to
f
i
n
d
th
e
ed
g
e
o
f
t
h
e
o
r
ig
in
al
a
n
d
n
o
is
y
i
m
a
g
e
b
y
u
s
in
g
s
a
m
p
le
d
ata.
I
n
t
h
is
s
t
u
d
y
,
it
i
s
f
o
u
n
d
th
a
t
t
h
er
e
f
e
w
er
d
etail
s
o
f
ed
g
es
in
t
h
e
d
en
o
i
s
ed
i
m
a
g
e.
T
o
ad
d
r
ess
th
i
s
is
s
u
e,
w
e
h
av
e
e
m
p
lo
y
ed
f
u
zz
y
b
ased
ed
g
e
d
etec
tio
n
a
n
d
t
h
e
n
t
h
e
ed
g
e
is
e
n
h
a
n
ce
d
i
n
th
e
d
en
o
is
ed
i
m
a
g
e
th
at
w
e
h
av
e
r
ec
eiv
ed
b
y
u
s
i
n
g
o
u
r
m
et
h
o
d
.
No
w
th
e
d
en
o
i
s
in
g
is
p
er
f
o
r
m
ed
b
ased
o
n
th
e
m
o
d
i
f
ied
p
ar
a
m
eter
S
-
GHP
f
o
cu
s
o
n
s
m
o
o
t
h
in
g
o
f
t
h
e
i
m
ag
e
b
y
i
m
p
le
m
en
tin
g
t
h
e
g
r
ad
ien
t
h
is
to
g
r
a
m
p
r
eser
v
atio
n
.
2
.
1
.
No
is
e
esti
m
a
t
io
n
I
n
p
u
t i
m
ag
e
i
s
d
ec
o
m
p
o
s
ed
in
t
o
o
v
er
lap
p
in
g
p
atch
es b
y
y
z
n
i
i
i
(
1
)
W
h
er
e
z
i
h
as
r
ep
r
esen
ted
th
e
o
r
ig
in
al
i
m
a
g
e
p
atch
w
it
h
t
h
e
it
h
p
ix
el
at
its
ce
n
tr
e
an
d
y
i
is
th
e
o
b
s
er
v
ed
v
ec
to
r
ized
p
atch
co
r
r
u
p
ted
b
y
ze
r
o
-
m
ea
n
Ga
u
s
s
ian
n
o
i
s
e
[
1
2
]
v
ec
to
r
n
i
.
T
h
e
o
b
j
ec
tiv
e
o
f
t
h
e
n
o
is
e
lev
e
l
esti
m
atio
n
is
to
co
m
p
u
te
th
e
s
tan
d
ar
d
d
ev
iatio
n
σ
n
o
f
th
e
n
o
is
y
i
m
ag
e
is
g
iv
e
n
.
I
n
th
i
s
m
et
h
o
d
,
th
e
Ho
r
izo
n
tal
an
d
v
er
tica
l
d
er
iv
at
iv
e
(
h
Dy
an
d
v
Dy
ar
e
ca
lcu
lated
a
n
d
t
h
e
n
t
h
e
g
r
ad
ien
t
v
ec
to
r
G
y
i
s
o
b
tain
ed
b
y
ta
k
in
g
hv
D
y
D
y
.
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
I
ma
g
e
Den
o
is
in
g
b
y
u
s
in
g
Mo
d
ified
S
GHP
A
lg
o
r
ith
m
(
S
r
ee
d
h
a
r
K
o
llem)
973
No
w
t
h
e
co
v
ar
ian
ce
m
atr
i
x
C
o
v
y
is
ca
lc
u
lated
b
y
T
C
o
v
G
G
y
y
y
(
2
)
T
h
e
Dir
ec
tio
n
al
Der
iv
ati
v
e
in
b
o
th
Ho
r
izo
n
tal
Dir
ec
tio
n
an
d
Ver
tical
Dir
ec
tio
n
is
ca
lc
u
lated
an
d
tr
ac
e
o
f
Gr
ad
ien
t M
atr
ix
i
s
ca
l
cu
lated
b
y
D
t
r
D
D
D
D
vv
hh
(
3
)
No
w
t
h
e
i
n
itial
n
o
is
e
le
v
el
i
s
esti
m
ated
b
y
co
m
p
u
ti
n
g
t
h
e
First
co
m
p
o
n
e
n
t
o
f
E
ig
e
n
v
a
lu
e
o
f
t
h
e
co
v
ar
ian
t
m
atr
ix
.
T
h
is
i
s
tak
en
as
t
h
e
in
i
tial
v
a
lu
e
f
o
r
ca
lcu
lati
n
g
n
o
is
e
le
v
el
b
y
u
s
i
n
g
i
ter
ativ
e
n
o
is
e
esti
m
atio
n
[
1
3
]
,,
0
i
n
v
(
4
)
No
w
t
h
e
n
o
is
e
le
v
el
e
s
ti
m
atio
n
f
o
r
m
w
ea
k
te
x
t
u
r
ed
p
atch
is
p
er
f
o
r
m
ed
[
1
4
]
.
Fo
r
th
is
I
n
v
er
s
e
g
a
m
m
a
f
u
n
ctio
n
,,
0
i
n
v
w
it
h
th
e
s
h
ap
e
p
ar
am
eter
α
an
d
s
ca
le
p
ar
a
m
eter
β
is
u
s
ed
1
0
k
(
5
)
I
f
th
e
s
elec
ted
p
atc
h
s
ize
is
less
t
h
a
n
th
e
n
t
h
e
p
atch
i
s
s
elec
ted
a
s
a
W
ea
k
T
ex
t
u
r
e
P
atch
.
Ma
x
i
m
u
m
ei
g
e
n
v
al
u
e
s
o
f
t
h
e
g
r
ad
ien
t
co
v
ar
ia
n
ce
ar
e
co
m
p
u
ted
w
h
e
n
t
h
e
s
tr
e
n
g
t
h
o
f
i
m
a
g
e
p
atch
es
ar
e
to
b
e
esti
m
ated
.
No
w
th
e
No
is
e
L
ev
e
l
o
f
W
ea
k
T
ex
tu
r
e
P
atch
is
f
o
u
n
d
b
y
u
s
i
n
g
th
e
E
ig
e
n
Val
u
e
o
f
C
o
v
ar
ian
c
e
Ma
tr
ix
o
f
t
h
e
w
ea
k
te
x
tu
r
ed
p
atch
an
d
its
p
r
in
cip
al
co
m
p
o
n
en
t
[
1
5
]
,
[
1
6
]
.
T
h
e
iter
atio
n
is
co
n
tin
u
ed
u
n
til
th
e
d
if
f
er
e
n
ce
b
et
w
ee
n
s
ig
m
a
i
n
s
t
ep
n
-
1
an
d
n
i
s
les
s
th
a
n
1
0
-
4
.
2
.
2
.
I
m
a
g
e
deno
is
ing
f
ra
m
e
w
o
rk
T
h
e
n
o
is
y
i
m
ag
e
i
s
d
ef
i
n
ed
b
y
th
e
E
q
u
atio
n
(
6
)
t
h
at
i
s
y
=
x
+
v
(
6
)
W
h
er
e
th
e
n
o
is
y
i
m
ag
e
is
r
e
p
r
esen
ted
w
it
h
y
,
t
h
e
Or
i
g
i
n
al
i
m
a
g
e
i
s
r
ep
r
esen
ted
w
it
h
x
,
A
d
d
itiv
e
w
h
i
te
Gau
s
s
ia
n
n
o
is
e
(
A
W
GN)
w
it
h
ze
r
o
m
ea
n
is
r
ep
r
esen
ted
w
it
h
v
an
d
th
e
s
tan
d
ar
d
d
ev
iatio
n
i
s
d
en
o
ted
w
it
h
.
T
h
e
m
ai
n
p
u
r
p
o
s
e
o
f
i
m
a
g
e
d
en
o
is
i
n
g
is
to
co
m
p
u
te
th
e
cle
an
i
m
a
g
e
x
f
r
o
m
n
o
is
y
i
m
ag
e
y
.
T
h
e
v
ib
r
atio
n
al
m
et
h
o
d
is
th
e
b
est d
en
o
i
s
in
g
a
p
p
r
o
ac
h
is
o
b
tain
ed
b
y
1
2
ˆ
a
r
g
m
in
2
2
x
y
x
R
x
x
(
7
)
W
h
er
e
r
eg
u
lar
izat
io
n
ter
m
is
d
en
o
ted
w
it
h
R
(
x
)
a
n
d
p
o
s
itiv
e
co
n
s
ta
n
t
is
w
it
h
λ
.
T
h
e
R
(
x
)
r
elies
o
n
ex
i
s
ti
n
g
i
m
a
g
es.
I
m
ag
e
d
en
o
is
in
g
m
e
th
o
d
s
h
a
v
e
a
g
en
er
al
is
s
u
e
th
at
i
m
ag
e
q
u
alit
y
s
ca
le
c
h
ar
ac
ter
is
tic
s
s
u
c
h
a
s
s
tr
u
ct
u
r
es
li
k
e
te
x
t
u
r
e
w
ill
b
e
o
v
er
-
s
m
o
o
th
ed
.
T
h
e
o
r
ig
in
a
l
i
m
ag
e
h
a
s
s
u
b
s
ta
n
tial
g
r
ad
ie
n
t
s
th
a
n
t
h
e
g
r
ad
ie
n
ts
o
f
o
v
er
s
m
o
o
th
ed
i
m
ag
e.
I
n
h
er
en
tl
y
,
a
s
tr
u
ct
u
r
e
li
k
e
te
x
t
u
r
e
d
o
esn
’
t
d
ep
en
d
o
n
o
v
er
s
m
o
o
th
in
g
a
n
d
th
e
tex
t
u
r
e
h
a
v
e
an
i
n
d
is
ti
n
g
u
is
h
ab
le
g
r
ad
ien
t
d
is
tr
ib
u
t
io
n
o
f
x
f
o
r
g
o
o
d
ev
al
u
atio
n
o
f
x
.
Fo
r
th
is
r
ea
s
o
n
,
w
e
p
r
o
p
o
s
e
a
m
o
d
i
f
ied
p
ar
a
m
eter
in
S
-
GHP
m
eth
o
d
b
y
ta
k
in
g
d
if
f
er
en
t
d
atab
ase
i
m
ag
e
s
.
T
h
e
g
r
ad
ien
t
h
i
s
to
g
r
a
m
o
f
th
e
d
en
o
is
ed
i
m
a
g
e
ˆ
x
v
er
y
clo
s
e
to
th
e
r
ef
er
en
ce
h
is
to
g
r
a
m
h
r
b
ased
o
n
th
e
co
m
p
u
t
e
o
f
th
e
g
r
ad
ien
t
h
is
to
g
r
a
m
o
f
x
,
d
en
o
te
h
r.
T
h
e
f
o
llo
w
in
g
p
r
o
p
o
s
ed
S
-
GHP
d
en
o
is
i
n
g
m
et
h
o
d
is
d
ef
i
n
ed
as
2
1
2
ˆ
a
r
g
m
in
,
2
2
x
y
x
R
x
F
x
x
xF
s
.
t.
h
F
=h
r
(
8
)
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.
8
,
No
.
2
,
A
p
r
il 2
0
1
8
:
9
7
1
–
9
7
8
974
W
h
er
e
th
e
o
d
d
f
u
n
ctio
n
is
F
u
n
i
f
o
r
m
l
y
n
o
n
-
d
escen
d
i
n
g
,
h
F
is
h
i
s
to
g
r
a
m
o
f
th
e
tr
an
s
f
o
r
m
ed
g
r
ad
ien
t
i
m
ag
e
|
F
(
∇
x)
|
,
∇
is
g
r
ad
ien
t
o
p
er
ato
r
an
d
p
o
s
itiv
e
co
n
s
ta
n
t
is
µ.
T
h
e
p
r
o
p
o
s
ed
m
o
d
if
ied
p
ar
a
m
eter
in
S
-
GHP
m
et
h
o
d
ac
q
u
ir
es
t
h
e
a
lter
n
ati
n
g
o
p
ti
m
izat
io
n
ap
p
r
o
ac
h
.
Fo
r
g
i
v
en
F,
th
e
n
0
x
F
x
an
d
u
p
d
ate
t
o
x
.
Fo
r
g
iv
e
n
x
,
b
ased
on
eq
u
at
io
n
0
x
F
x
F
is
u
p
d
ated
b
y
u
s
i
n
g
m
o
d
i
f
i
ed
p
ar
am
eter
S
-
GHP
s
p
ec
if
ic
atio
n
o
p
er
ato
r
.
An
o
th
er
ca
s
e
i
n
th
e
S
-
GHP
m
et
h
o
d
is
w
h
a
t
w
a
y
to
p
er
ce
iv
e
t
h
e
r
e
f
er
en
ce
h
i
s
to
g
r
a
m
h
r
o
f
u
n
s
p
ec
if
ied
i
m
a
g
e
x
.
C
o
m
p
u
t
atio
n
o
f
h
r
d
ep
en
d
s
o
n
t
h
e
n
o
i
s
y
o
b
s
er
v
atio
n
y
.
Fo
r
f
i
n
d
in
g
h
r,
n
e
w
m
et
h
o
d
s
ar
e
p
r
o
p
o
s
ed
f
ir
s
t
o
n
e
i
s
a
r
eg
u
la
r
ized
d
ec
o
n
v
o
lu
tio
n
m
et
h
o
d
an
d
th
e
s
ec
o
n
d
o
n
e
is
a
n
iter
at
iv
e
d
ec
o
n
v
o
l
u
tio
n
m
et
h
o
d
f
r
o
m
t
h
e
n
o
is
y
i
m
a
g
e
[
1
7
]
d
ep
en
d
s
u
p
o
n
d
if
f
er
en
t
n
o
is
e
le
v
els
[
1
8
]
.
Af
ter
r
ef
e
r
en
ce
h
i
s
to
g
r
a
m
is
attain
ed
,
th
e
n
m
o
d
i
f
ied
p
ar
a
m
eter
in
S
-
GHP
m
et
h
o
d
is
ap
p
lied
f
o
r
im
a
g
e
d
en
o
i
s
in
g
.
3.
S
-
G
RAD
I
E
NT
H
I
ST
O
G
RA
M
P
RE
SE
RV
AT
I
O
N
DE
NO
I
SI
N
G
M
E
T
H
O
D
S
-
GHP
i
s
a
p
r
o
p
o
s
ed
m
et
h
o
d
b
ased
on
th
e
p
atc
h
m
eth
o
d
.
L
et
ii
x
R
x
is
a
p
atch
tak
e
o
u
t
a
t
p
o
s
itio
n
i
=
1
,
2
.
.
.
N
,
w
h
er
e
p
atch
ex
tr
ac
tio
n
o
p
er
ato
r
is
R
i
an
d
N
in
d
icate
s
p
ix
el
s
in
th
e
i
m
a
g
e.
Giv
e
n
a
d
ictio
n
ar
y
D,
i
n
f
r
eq
u
en
tl
y
en
co
d
e
th
e
p
atch
x
i
o
v
er
D,
g
i
v
es
t
h
e
s
p
ar
s
e
co
d
in
g
v
ec
to
r
i
.
I
m
ag
e
p
atc
h
es
h
av
i
n
g
co
d
in
g
v
ec
to
r
s
ar
e
attai
n
ed
,
th
e
i
m
ag
e
x
ca
n
b
e
r
en
o
v
ated
by
1
11
NN
TT
x
D
R
R
R
D
i
i
i
i
ii
(
9
)
W
h
er
e
co
n
ca
ten
atio
n
b
elo
n
g
s
to
α
f
o
r
all
th
e
v
alu
e
s
o
f
i
.
I
m
ag
e
s
ar
e
tak
en
f
r
o
m
d
atab
ases
ar
e
test
in
g
m
o
d
i
f
ied
p
ar
am
eter
S
-
GHP
Me
th
o
d
.
So
,
th
e
co
m
b
i
n
atio
n
r
eg
ar
d
i
n
g
id
en
tic
al
p
r
io
r
s
r
ef
i
n
es
th
e
m
o
d
i
f
ied
p
ar
a
m
eter
S
-
GHP
.
Fo
r
e
x
a
m
p
le,
th
e
e
s
ti
m
atio
n
p
r
o
ce
d
u
r
es
in
[
1
9
]
-
[
2
3
]
m
er
g
e
i
m
a
g
e
n
o
n
-
lo
ca
l
N
SS
p
r
io
r
to
i
m
a
g
e
l
o
ca
l
s
p
ar
s
it
y
p
r
io
r
an
d
w
e
h
av
e
b
etter
d
en
o
is
in
g
r
es
u
lt
s
.
I
n
th
e
m
eth
o
d
m
o
d
if
ied
p
ar
a
m
eter
i
n
S
-
GHP
,
th
e
R
(
x
)
,
w
h
ic
h
is
s
p
ar
s
e
n
o
n
-
lo
ca
l
r
eg
u
lar
izatio
n
ter
m
p
r
o
p
o
s
ed
in
th
e
n
o
n
-
lo
ca
ll
y
ce
n
tr
alize
d
s
p
ar
s
e
r
ep
r
esen
tatio
n
(
NC
S
R
)
m
o
d
el
[
2
4
]
is
1
Rx
ii
i
(
1
0
)
W
h
er
e
w
ei
g
h
ted
av
er
ag
e
o
f
q
i
is
i
th
en
qq
w
i
i
i
q
(
1
1
)
an
d
co
d
in
g
v
ec
to
r
o
f
th
e
q
th
n
ea
r
est
p
atch
(
q
x
i
)
to
x
i
is
q
i
.
W
eig
h
t
is
d
en
o
ted
as
2
11
ˆˆ
e
x
p
qq
w
x
x
i
i
i
Wh
,
w
h
er
e
th
e
p
r
ed
ef
in
ed
co
n
s
tan
t
is
h
an
d
n
o
r
m
aliza
t
io
n
f
ac
to
r
is
W
.
T
h
e
f
o
r
m
u
la
f
o
r
m
o
d
if
ied
p
ar
am
eter
S
-
GHP
m
et
h
o
d
is
d
ef
i
n
ed
as b
y
u
s
i
n
g
Eq
u
atio
n
(
3
)
is
1
2
ˆ
a
r
g
m
in
,
2
2
2
1
x
y
x
x
i
x
Fx
i
i
F
(
1
2
)
Su
c
h
th
at
xD
,
Fr
hh
(
1
3
)
Fro
m
t
h
e
S
-
GHP
m
e
th
o
d
,
u
s
i
n
g
E
q
u
atio
n
(
7
)
,
F
(
∇
x
)
i
s
ap
p
r
o
x
i
m
ate
to
∇
x
w
h
e
n
h
i
s
to
g
r
a
m
p
ar
am
e
ter
lead
s
to
lar
g
er
an
d
w
e
ca
n
ac
h
iev
e
r
eq
u
ir
ed
h
is
to
g
r
a
m
p
ar
a
m
eter
f
o
r
S
-
GHP
.
W
h
en
th
e
h
i
s
to
g
r
a
m
h
F
o
f
|
F
(
∇
x)
|
i
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:
2
0
8
8
-
8708
I
ma
g
e
Den
o
is
in
g
b
y
u
s
in
g
Mo
d
ified
S
GHP
A
lg
o
r
ith
m
(
S
r
ee
d
h
a
r
K
o
llem)
975
r
eq
u
ir
ed
an
d
ap
p
r
o
x
im
ate
to
th
e
h
r
,
(
h
is
to
g
r
a
m
o
f
∇
x
=
h
r
)
th
en
ac
q
u
ir
e
th
e
r
eq
u
ir
ed
g
r
ad
ien
t
h
is
to
g
r
a
m
p
ar
am
eter
f
o
r
S
-
GHP
.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
4
.
1
.
P
er
f
o
r
m
a
nce
a
na
ly
s
is
T
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
is
v
er
if
i
ed
b
y
u
s
i
n
g
th
r
ee
d
if
f
er
en
t
i
m
ag
es
li
k
e
i
m
a
g
e
-
3
,
i
m
ag
e
-
4
an
d
im
a
g
e
-
5
.
Her
e,
th
r
ee
i
m
a
g
es
ar
e
g
r
e
y
-
s
ca
le
i
m
a
g
es
h
a
v
i
n
g
a
r
an
g
e
b
et
w
ee
n
0
to
2
5
5
.
Fo
r
i
m
a
g
e
-
3
,
i
m
a
g
e
-
4
an
d
i
m
a
g
e
-
5
ar
e
tak
in
g
f
i
v
e
d
i
f
f
er
e
n
t
n
o
is
e
lev
e
ls
ar
e
2
0
,
2
5
,
3
0
,
3
5
an
d
4
0
w
ith
r
esp
ec
t
to
t
h
at
d
i
f
f
er
en
t
P
SNR
an
d
SS
I
M
v
al
u
es
ar
e
o
b
tain
ed
.
I
n
Fig
u
r
e
2
,
Fig
u
r
e
3
a
n
d
Fig
u
r
e
4
,
th
er
e
is
o
r
ig
in
a
l
im
ag
e
a
n
d
d
if
f
er
e
n
t
en
h
a
n
ce
d
i
m
a
g
es
w
it
h
d
i
f
f
er
e
n
t
n
o
is
e
lev
el
s
.
I
n
F
i
g
u
r
e
5
,
n
u
m
b
er
s
o
f
iter
a
t
io
n
s
ar
e
i
n
c
r
ea
s
ed
th
e
n
P
SN
R
v
alu
e
i
n
cr
ea
s
e
s
.
W
h
en
n
o
is
e
s
tan
d
ar
d
d
ev
iatio
n
is
i
n
cr
ea
s
ed
th
en
t
h
e
s
tr
u
ct
u
r
al
s
i
m
ilar
it
y
in
d
ex
i
s
d
ec
r
ea
s
ed
.
Fro
m
th
e
F
i
g
u
r
e
5
,
i
m
ag
e
-
3
h
av
in
g
m
o
r
e
s
tr
u
ctu
r
al
s
i
m
ilar
it
y
i
n
d
ex
.
I
n
T
ab
le
1
,
T
a
b
le
2
a
n
d
T
ab
le
3
g
iv
e
th
e
s
tr
u
ct
u
r
al
s
i
m
ilar
it
y
i
n
d
ex
an
d
P
SNR
v
al
u
es
o
f
i
m
a
g
e
-
3
,
i
m
a
g
e
-
4
a
n
d
i
m
a
g
e
-
5
b
y
u
s
in
g
a
m
o
d
i
f
ied
p
ar
a
m
eter
in
S
-
GHP
m
et
h
o
d
.
(
a)
Or
ig
in
al
I
m
a
g
e
(
b
)
20
(
c)
25
(d
)
30
(e
)
35
(f
)
40
Fig
u
r
e
2
.
Den
o
is
ed
i
m
a
g
e
-
3
u
n
d
er
d
if
f
er
e
n
t n
o
i
s
e
lev
el
s
T
ab
le
1
.
Stru
ctu
r
al
s
i
m
ilar
it
y
i
n
d
ex
(
SS
I
M)
an
d
P
SNR
(
d
B
)
r
esu
lt
s
o
f
s
-
g
r
ad
ien
t h
is
to
g
r
a
m
p
r
eser
v
atio
n
o
f
i
m
a
g
e
-
3
N
o
.
o
f
I
t
e
r
a
t
i
o
n
s
S
i
g
ma
=
2
0
S
i
g
ma
=
2
5
S
i
g
ma=
3
0
S
i
g
ma=
3
5
S
i
g
ma=
4
0
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
1
2
7
.
7
6
6
2
6
.
4
4
0
2
5
.
2
9
7
2
4
.
8
4
2
2
4
.
0
3
1
0
.
7
3
8
0
.
6
7
8
0
.
6
2
1
0
.
6
0
0
0
.
5
5
6
2
2
7
.
9
5
7
2
6
.
7
7
1
2
5
.
7
7
9
2
5
.
4
0
0
2
4
.
7
6
6
0
.
7
4
8
0
.
6
9
8
0
.
6
5
1
0
.
6
4
0
0
.
6
0
8
3
2
8
.
0
9
3
2
7
.
0
1
5
2
6
.
1
3
8
2
5
.
6
8
4
2
5
.
1
2
7
0
.
7
5
5
0
.
7
1
3
0
.
6
7
5
0
.
6
6
0
0
.
6
3
7
4
2
8
.
1
6
1
2
7
.
1
4
7
2
6
.
3
34
2
5
.
7
3
6
2
5
.
1
9
0
0
.
7
5
7
0
.
7
1
9
0.
6
8
7
0
.
6
6
0
0
.
6
3
8
5
2
8
.
1
7
4
2
7
.
1
9
3
2
6
.
4
0
6
2
5
.
7
2
7
2
5
.
1
8
6
0
.
7
5
5
0
.
7
1
9
0
.
6
8
9
0
.
6
5
6
0
.
6
3
4
6
2
8
.
1
5
7
2
7
.
1
9
2
2
6
.
4
1
7
2
5
.
7
0
8
2
5
.
1
6
9
0
.
7
5
2
0
.
7
1
6
0
.
6
8
7
0
.
6
5
4
0
.
6
3
2
A
v
e
r
a
g
e
P
S
N
R
2
8
.
0
5
1
2
6
.
9
5
9
2
6
.
0
6
1
2
5
.
5
1
6
2
4
.
9
1
1
a
n
d
S
S
I
M
0
.
7
5
0
0
.
7
0
7
0
.
6
6
8
0
.
6
4
5
0
.
6
1
7
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.
8
,
No
.
2
,
A
p
r
il 2
0
1
8
:
9
7
1
–
9
7
8
976
(
a)
Or
ig
in
al
I
m
ag
e
(
b
)
20
(
c)
25
(
d
)
30
(
e)
35
(
f
)
40
Fig
u
r
e
3
.
Den
o
is
ed
i
m
a
g
e
-
4
u
s
in
g
u
n
d
er
d
if
f
er
en
t
n
o
is
e
le
v
e
ls
(
a)
Or
ig
in
al
I
m
a
g
e
(
b
)
20
(
c)
25
(
d
)
30
(
e)
35
(
f
)
40
Fig
u
r
e
4
.
Den
o
is
ed
i
m
a
g
e
-
5
u
s
in
g
u
n
d
er
d
if
f
er
en
t
n
o
is
e
le
v
e
ls
T
ab
le
2
.
Stru
ctu
r
al
s
i
m
ilar
it
y
i
n
d
ex
(
SS
I
M)
an
d
P
SNR
(
d
B
)
r
esu
lt
s
o
f
s
-
g
r
ad
ien
t h
is
to
g
r
a
m
p
r
eser
v
atio
n
o
f
i
m
a
g
e
-
4
N
o
.
o
f
I
t
e
r
a
t
i
o
n
s
S
i
g
ma
=
2
0
S
i
g
ma
=
2
5
S
i
g
ma=
3
0
S
i
g
ma=
3
5
S
i
g
ma=
4
0
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
1
2
6
.
4
4
9
2
5
.
1
2
0
2
4
.
0
3
3
2
3
.
4
9
5
2
2
.
7
7
0
0
.
7
7
2
0
.
7
0
8
0
.
6
4
8
0
.
6
1
5
0
.
5
6
8
2
2
6
.
5
4
7
2
5
.
2
5
6
2
4
.
2
1
3
2
3
.
7
1
5
2
3
.
0
6
8
0
.
7
8
0
0
.
7
2
0
0
.
6
6
3
0
.
6
3
7
0
.
5
9
5
3
2
6
.
6
3
8
2
5
.
3
9
6
2
4
.
4
0
5
2
3
.
9
2
7
2
3
.
3
2
6
0
.
7
8
8
0
.
7
3
3
0
.
6
8
1
0
.
6
6
2
0
.
6
2
7
4
2
6
.
7
0
4
2
5
.
5
0
6
2
4
.
5
5
6
2
4
.
0
2
0
2
3
.
4
2
1
0
.
7
9
5
0
.
7
4
5
0
.
6
9
9
0
.
6
7
3
0
.
6
3
9
5
2
6
.
7
4
1
2
5
.
5
7
7
2
4
.
6
5
4
2
4
.
0
3
6
2
3
.
4
2
8
0
.
8
0
0
0
.
7
5
3
0
.
7
1
1
0
.
6
7
4
0
.
6
3
8
6
2
6
.
7
4
4
2
5
.
6
0
0
2
4
.
6
8
9
2
4
.
0
1
8
2
3
.
3
9
8
0
.
8
0
1
0
.
7
5
7
0
.
7
1
5
0
.
6
7
2
0
.
6
3
7
A
v
e
r
a
g
e
P
S
N
R
2
6
.
6
3
7
2
5
.
4
0
9
2
4
.
4
2
5
2
3
.
8
6
8
2
3
.
2
3
5
a
n
d
S
S
I
M
0
.
7
8
9
0
.
7
3
6
0
.
6
8
6
0
.
6
5
5
0
.
6
1
7
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
I
ma
g
e
Den
o
is
in
g
b
y
u
s
in
g
Mo
d
ified
S
GHP
A
lg
o
r
ith
m
(
S
r
ee
d
h
a
r
K
o
llem)
977
(a
)
P
SNR
o
f
i
m
a
g
e
-
3
(
b
)
P
SNR
o
f
i
m
ag
e
-
4
(
c)
P
SNR
o
f
i
m
a
g
e
-
5
(
d
)
C
o
m
p
ar
is
o
n
o
f
Si
g
m
a
an
d
SS
I
M
Fig
u
r
e
5
.
Var
iatio
n
o
f
P
SNR
o
f
i
m
ag
e
-
3
,
i
m
a
g
e
-
4
,
i
m
a
g
e
-
5
u
s
in
g
d
if
f
er
en
t si
g
m
a
v
al
u
es a
n
d
its
SS
I
M
T
ab
le
3
.
Stru
ctu
r
al
s
i
m
ilar
it
y
i
n
d
ex
(
SS
I
M)
an
d
P
SNR
(
d
B
)
r
esu
lt
s
o
f
s
-
g
r
ad
ien
t h
is
to
g
r
a
m
p
r
eser
v
atio
n
o
f
i
m
a
g
e
-
5
N
o
.
o
f
I
t
e
r
a
t
i
o
n
s
S
i
g
ma
=
2
0
S
i
g
m
a
=
2
5
S
i
g
m
a
=
3
0
S
i
g
ma
=
3
5
S
i
g
m
a
=
4
0
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
S
-
G
H
P
1
2
9
.
4
5
1
2
7
.
9
3
6
2
6
.
6
1
7
2
6
.
4
2
1
2
5
.
4
6
9
0
.
7
2
8
0
.
6
5
2
0
.
5
8
0
0
.
5
6
8
0
.
5
1
3
2
2
9
.
9
7
2
2
8
.
7
0
3
2
7
.
6
3
7
2
7
.
6
0
6
2
6
.
9
3
9
0
.
7
6
0
0
.
7
0
1
0
.
6
4
7
0
.
6
5
5
0
.
6
2
0
3
3
0
.
3
6
1
2
9
.
2
8
0
2
8
.
3
9
8
2
8
.
2
4
3
2
7
.
6
9
0
0
.
7
8
5
0
.
7
4
2
0
.
7
0
3
0
.
7
0
9
0
.
6
8
7
4
3
0
.
5
7
0
2
9
.
5
8
5
2
8
.
7
7
9
2
8
.
4
0
3
2
7
.
8
6
3
0
.
7
9
9
0
.
7
6
4
0
.
7
3
3
0
.
7
2
2
0
.
7
0
1
5
3
0
.
6
6
7
2
9
.
7
1
1
2
8
.
9
3
0
2
8
.
4
5
4
2
7
.
9
1
4
0
.
8
0
4
0
.
7
7
2
0
.
7
4
4
0
.
7
2
4
0
.
7
0
4
6
3
0
.
7
0
7
2
9
.
7
6
1
2
8
.
9
8
6
2
8
.
4
7
3
2
7
.
9
3
0
0
.
8
0
6
0
.
7
7
4
0
.
7
4
7
0
.
7
2
6
0
.
7
0
6
A
v
e
r
a
g
e
P
S
N
R
3
0
.
2
8
8
2
9
.
1
6
3
2
8
.
2
2
4
2
7
.
9
3
3
2
7
.
3
0
0
a
n
d
S
S
I
M
0
.
7
8
0
0
.
7
3
4
0
.
6
9
2
0
.
6
8
4
0
.
6
5
5
4
.
2
.
Co
m
pa
ra
t
iv
e
a
na
ly
s
is
T
h
e
ex
is
ti
n
g
m
et
h
o
d
s
an
d
p
r
o
p
o
s
ed
m
et
h
o
d
v
er
if
ied
b
y
u
s
in
g
t
h
r
ee
d
if
f
er
e
n
t
i
m
ag
e
s
l
ik
e
i
m
ag
e
-
3
,
i
m
a
g
e
-
4
an
d
i
m
ag
e
-
5
w
it
h
f
i
v
e
d
if
f
er
en
t
n
o
is
e
le
v
els
ar
e
2
0
,
2
5
,
3
0
,
3
5
an
d
4
0
.
P
er
f
o
r
m
a
n
ce
o
f
th
ese
m
e
th
o
d
s
is
m
e
n
tio
n
ed
in
ter
m
s
o
f
P
ea
k
s
i
g
n
a
l
to
n
o
is
e
r
atio
an
d
s
t
r
u
ctu
r
al
s
i
m
ilar
it
y
i
n
d
ex
[
2
5
]
,
[
2
6
]
as
s
h
o
w
n
i
n
T
ab
le
4.
T
ab
le
4
.
C
o
m
p
ar
is
o
n
o
f
E
x
i
s
ti
n
g
m
et
h
o
d
s
an
d
p
r
o
p
o
s
ed
m
et
h
o
d
in
ter
m
s
o
f
P
SNR
(
d
B
)
r
e
s
u
lt
s
I
mag
e
N
o
Ex
i
s
t
i
n
g
M
e
t
h
o
d
s
P
r
o
p
o
se
d
M
e
t
h
o
d
B
-
G
H
P
A
P
B
S
M
o
d
i
f
i
e
d
S
-
G
H
P
P
S
N
R
P
S
N
R
P
S
N
R
3
2
7
.
0
1
2
6
.
0
5
2
8
.
0
5
1
4
2
5
.
4
9
2
5
.
1
1
2
6
.
6
3
7
5
2
9
.
9
0
2
8
.
6
6
3
0
.
2
8
8
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.
8
,
No
.
2
,
A
p
r
il 2
0
1
8
:
9
7
1
–
9
7
8
978
5.
CO
NCLU
SI
O
N
I
n
th
is
p
ap
er
,
th
e
p
r
o
p
o
s
ed
m
et
h
o
d
m
o
d
if
ied
Stru
ct
u
r
e
g
r
ad
ien
t
h
i
s
to
g
r
a
m
p
r
eser
v
at
io
n
u
s
ed
f
o
r
en
h
a
n
ci
n
g
th
e
d
i
f
f
er
en
t
i
m
a
g
es
b
y
ta
k
in
g
d
i
f
f
er
e
n
t
n
o
is
e
l
ev
els
l
ik
e
2
0
,
2
5
,
3
0
an
d
4
0
.
B
ased
on
th
e
n
o
i
s
e
lev
els,
t
h
e
P
SNR
a
n
d
SS
I
M
v
alu
es
ar
e
i
m
p
r
o
v
ed
co
m
p
ar
ed
to
o
th
er
m
et
h
o
d
s
lik
e
A
P
B
S
a
n
d
B
-
GHP
.
All
th
e
ab
o
v
e
-
m
en
t
io
n
ed
r
esu
l
ts
p
r
o
v
ed
th
at
th
e
m
o
d
i
f
ied
p
ar
a
m
eter
S
-
GHP
is
b
etter
co
m
p
ar
ed
to
B
-
GHP
an
d
A
P
B
S.
RE
F
E
R
E
NC
E
S
[1
]
Bu
a
d
e
s,
B.
C
o
ll
,
a
n
d
J.
M
o
re
l,
“
A
re
v
ie
w
o
f
i
m
a
g
e
d
e
n
o
isin
g
m
e
t
h
o
d
s
,
w
it
h
a
n
e
w
o
n
e
”
,
M
u
l
ti
sc
a
l
e
M
o
d
e
l.
S
imu
l.
,
v
o
l.
4
,
n
o
.
2
,
p
p
.
4
9
0
5
3
0
,
2
0
0
5
.
[2
]
K.
Da
b
o
v
,
A
.
F
o
i
,
V
.
Ka
tk
o
v
n
ik
,
a
n
d
K.
Eg
iaz
a
rian
,
“
Im
a
g
e
d
e
n
o
isi
n
g
b
y
sp
a
rse
3
-
D
tra
n
sf
o
rm
-
d
o
m
a
in
c
o
ll
a
b
o
ra
ti
v
e
f
il
terin
g
”
,
IEE
E
T
r
a
n
s.
Ima
g
e
Pro
c
e
ss
.
,
v
o
l.
1
6
,
n
o
.
8
,
p
p
.
2
0
8
0
-
2
0
9
5
,
A
u
g
.
2
0
0
7
.
[3
]
W.
Do
n
g
,
L
.
Zh
a
n
g
,
G
.
S
h
i,
a
n
d
X
.
L
i,
“
No
n
lo
c
a
ll
y
c
e
n
tralize
d
sp
a
rse
re
p
re
s
e
n
tatio
n
f
o
r
ima
g
e
re
st
o
ra
ti
o
n
”
,
IEE
E
T
ra
n
s.
Im
a
g
e
Pr
o
c
e
ss
,
v
o
l.
2
2
,
n
o
.
4
,
p
p
.
1
6
2
0
-
1
6
3
0
,
A
p
r.
2
0
1
3
.
[4
]
L
.
Ru
d
in
,
S
.
Os
h
e
r,
a
n
d
E
.
F
a
tem
i,
“
No
n
li
n
e
a
r
to
tal
v
a
riatio
n
b
a
se
d
n
o
ise
re
m
o
v
a
l
a
lg
o
rit
h
m
s
”
,
Ph
y
s.
D
,
v
o
l.
6
0
,
n
o
s.
1
4
,
p
p
.
2
5
9
-
2
6
8
,
No
v
.
1
9
9
2
.
[5
]
P
a
n
k
a
j
He
d
a
o
o
a
n
d
S
w
a
ti
S
Go
d
b
o
le,
“
W
a
v
e
let
T
h
re
sh
o
ld
i
n
g
A
p
p
ro
a
c
h
f
o
r
Im
a
g
e
D
e
n
o
isin
g
”
,
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
Ne
two
rk
S
e
c
u
rity I
ts
A
p
p
li
c
a
ti
o
n
s (
IJ
NS
A)
, v
o
l.
3
,
n
o.
4
,
Ju
ly
2
0
1
1
.
[6
]
H.
C.
Bu
r
g
e
r,
C.
J.
S
c
h
u
ler,
a
n
d
S
.
Ha
r
m
e
li
n
g
,
“
I
m
a
g
e
d
e
n
o
isin
g
:
Ca
n
p
lain
n
e
u
ra
l
n
e
tw
o
rk
s
c
o
m
p
e
t
e
w
it
h
BM
3
D
”,
in
Pro
c
.
In
t
.
Co
n
f.
C
VP
R
,
J
u
n
.
2
0
1
2
,
p
p
.
2
3
9
2
-
2
3
9
9
.
[7
]
S
.
S
u
b
h
a
,
I
.
Je
su
d
a
ss
a
n
d
K.
T
h
a
n
u
sh
k
o
d
i
,
“
Im
a
g
e
d
e
n
o
isi
n
g
b
y
u
sin
g
it
e
ra
ti
v
e
h
ist
o
g
ra
m
a
n
d
p
re
se
rv
a
ti
v
e
a
lg
o
rit
h
m
”
,
AR
PN
J
o
u
rn
a
l
o
f
E
n
g
in
e
e
rin
g
a
n
d
A
p
p
l
ied
S
c
ie
n
c
e
s
,
v
o
l.
1
0
,
n
o
.
1
1
,
Ju
n
e
2
0
1
5
.
[8
]
P
.
Ko
lt
so
v
,
“
Co
m
p
a
ra
ti
v
e
stu
d
y
o
f
te
x
tu
re
d
e
tec
ti
o
n
a
n
d
c
las
sif
ic
a
ti
o
n
a
lg
o
rit
h
m
s,”
Co
mp
u
t.
M
a
t
h
.
M
a
th
.
Ph
y
s.
,
v
o
l.
5
1
,
n
o
.
8
,
p
p
.
1
4
6
0
-
1
4
6
6
,
2
0
1
1
.
[9
]
L
iu
,
X
i
n
h
a
o
,
M
i
tsu
r
u
T
a
n
a
k
a
,
a
n
d
M
a
sa
to
sh
i
Ok
u
to
m
i,
“
No
ise
lev
e
l
e
sti
m
a
ti
o
n
u
sin
g
w
e
a
k
te
x
tu
re
d
p
a
tch
e
s
o
f
a
sin
g
le n
o
isy
ima
g
e
”
,
Ima
g
e
Pro
c
e
ss
in
g
(
ICIP)
,
2
0
1
2
1
9
t
h
IEE
E
In
t
e
rn
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
.
IEE
E
,
2
0
1
2
.
[1
0
]
S
.
P
y
a
t
y
k
h
,
J.
He
ss
e
r
a
n
d
L
.
Zh
e
n
g
,
"
I
m
a
g
e
No
ise
Lev
e
l
Esti
m
a
ti
o
n
b
y
P
rin
c
ip
a
l
C
o
m
p
o
n
e
n
t
A
n
a
l
y
sis,"
in
IEE
E
T
ra
n
sa
c
ti
o
n
s o
n
Im
a
g
e
P
ro
c
e
ss
in
g
,
v
o
l.
2
2
,
n
o
.
2
,
p
p
.
6
8
7
-
6
9
9
,
F
e
b
.
2
0
1
3
.
d
o
i:
1
0
.
1
1
0
9
/T
IP
.
2
0
1
2
.
2
2
2
1
7
2
8
.
[1
1
]
Jo
ll
if
f
e
,
“
P
rin
c
i
p
a
l
c
o
m
p
o
n
e
n
t
a
n
a
l
y
sis,”
in
En
c
y
c
lo
p
e
d
ia
o
f
S
t
a
ti
s
ti
c
s
in
Beh
a
v
io
r
a
l
S
c
i
e
n
c
e
.
Ne
w
Yo
rk
:
S
p
rin
g
e
r
-
V
e
rlag
,
2
0
0
2
.
[1
2
]
D.
P
a
sto
r
,
“
A
th
e
o
re
ti
c
a
l
re
su
lt
f
o
r
p
ro
c
e
ss
in
g
sig
n
a
ls
th
a
t
h
a
v
e
u
n
k
n
o
w
n
d
istri
b
u
ti
o
n
s
a
n
d
p
rio
rs
i
n
w
h
it
e
G
a
u
ss
ian
n
o
ise
,
”
Co
m
p
u
t
.
S
t
a
t.
Da
t
a
A
n
a
l.
,
v
o
l.
5
2
,
n
o
.
6
,
p
p
.
3
1
6
7
–
3
1
8
6
,
2
0
0
8
.
[1
3
]
L
iu
,
“
A
f
a
st
m
e
th
o
d
o
f
e
stim
a
ti
n
g
G
a
u
ss
ian
n
o
ise
,
”
in
Pr
o
c
.
1
st I
n
t
.
Co
n
f
.
In
f.
S
c
i.
E
n
g
.
,
2
0
0
9
,
p
p
.
4
4
1
–
4
4
4
.
[1
4
]
C.
Ke
rv
ra
n
n
a
n
d
J.
Bo
u
lan
g
e
r,
“
Op
ti
m
a
l
sp
a
ti
a
l
a
d
a
p
tatio
n
f
o
r
p
a
tch
-
b
a
se
d
im
a
g
e
d
e
n
o
isin
g
,
”
IEE
E
T
ra
n
s.
Ima
g
e
Pro
c
e
ss
.
,
v
o
l.
1
5
,
n
o
.
1
0
,
p
p
.
2
8
6
6
–
2
8
7
8
,
Oc
t.
2
0
0
6
.
[1
5
]
R.
Bil
c
u
a
n
d
M
.
V
e
h
v
il
a
in
e
n
,
“
Ne
w
m
e
th
o
d
f
o
r
n
o
ise
e
sti
m
a
ti
o
n
in
im
a
g
e
s,”
in
Pro
c
.
IEE
E
-
Eu
ra
sip
No
n
li
n
e
a
r
S
ig
n
a
l
Ima
g
e
Pro
c
e
ss
.
,
M
a
y
2
0
0
5
.
[1
6
]
G
.
S
m
it
h
a
n
d
I.
Bu
r
n
s,
“
M
e
a
su
ri
n
g
tex
tu
re
c
las
sif
ica
ti
o
n
a
lg
o
rit
h
m
s,”
Pa
tt
e
rn
Rec
o
g
n
it
.
L
e
tt
.
,
v
o
l
.
1
8
,
n
o
.
1
4
,
p
p
.
1
4
9
5
–
1
5
0
1
,
1
9
9
7
.
[1
7
]
D.
M
a
k
o
v
o
z
,
“
No
ise
v
a
rian
c
e
e
stim
a
ti
o
n
i
n
sig
n
a
l
p
r
o
c
e
ss
in
g
,
”
in
Pro
c
.
I
EE
E
I
n
t.
S
y
mp
.
S
ig
n
a
l
Pr
o
c
e
ss
.
In
f.
T
e
c
h
n
o
l
.
,
Ju
l.
2
0
0
6
,
p
p
.
3
6
4
–
3
6
9
.
[1
8
]
M
.
Us
s,
B.
V
o
z
e
l,
V
.
L
u
k
in
,
S
.
A
b
ra
m
o
v
,
I.
Ba
r
y
sh
e
v
,
a
n
d
K.
Ch
e
h
d
i,
“
Im
a
g
e
in
f
o
r
m
a
ti
v
e
m
a
p
s
f
o
r
e
sti
m
a
ti
n
g
n
o
i
se
sta
n
d
a
rd
d
e
v
iatio
n
a
n
d
tex
tu
re
p
a
ra
m
e
ters
,
”
EURA
S
IP
J
.
A
d
v
.
S
ig
n
a
l
Pr
o
c
e
ss
.
,
v
o
l.
2
0
1
1
,
p
p
.
8
0
6
5
1
6
-
1
–
8
0
6
5
1
6
-
1
2
,
M
a
r.
2
0
1
1
.
[1
9
]
M
.
El
a
d
a
n
d
M
.
A
h
a
ro
n
,
“
Im
a
g
e
d
e
n
o
isi
n
g
v
ia
sp
a
rse
a
n
d
re
d
u
n
d
a
n
t
re
p
re
se
n
tatio
n
s
o
v
e
r
lea
rn
e
d
d
icti
o
n
a
ries
”
,
IEE
E
T
ra
n
s.
Im
a
g
e
Pr
o
c
e
ss
.
,
v
o
l.
1
5
,
n
o
.
1
2
,
p
p
.
3
7
3
6
-
3
7
4
5
,
De
c
.
2
0
0
6
.
[2
0
]
J.
Ja
n
c
sa
r
y
,
S
.
No
w
o
z
in
,
a
n
d
C.
Ro
th
e
r,
“
L
o
ss
-
sp
e
c
if
ic
train
in
g
o
f
n
o
n
p
a
ra
m
e
tri
c
i
m
a
g
e
re
sto
ra
ti
o
n
m
o
d
e
ls:
a
n
e
w
sta
te o
f
th
e
a
rt,
”
in
Pr
o
c
.
E
u
r.
Co
n
f.
C
o
mp
u
t.
Vi
s
.
,
2
0
1
2
.
[2
1
]
M
.
Ha
sh
e
m
i
a
n
d
S
.
Be
h
e
sh
ti
,
“
A
d
a
p
ti
v
e
n
o
ise
v
a
rian
c
e
e
sti
m
a
ti
o
n
i
n
Ba
y
e
s
sh
rin
k
”
,
IEE
E
S
i
g
n
a
l
Pro
c
e
ss
.
L
e
tt
.
,
v
o
l.
1
7
,
n
o
.
1
,
p
p
.
1
2
-
1
5
,
Ja
n
.
2
0
1
0
.
[2
2
]
T
.
T
a
sd
ize
n
,
“
P
rin
c
i
p
a
l
c
o
m
p
o
n
e
n
ts
f
o
r
n
o
n
-
l
o
c
a
l
m
e
a
n
s i
m
a
g
e
d
e
n
o
isi
n
g
,
”
in
Pro
c
.
IEE
E
I
n
t.
Co
n
f.
Ima
g
e
Pro
c
e
ss
.
,
F
e
b
.
2
0
0
8
,
p
p
.
1
7
2
8
-
1
7
3
1
.
[2
3
]
V
.
Ra
jan
e
sh
a
n
d
Ko
ll
e
m
S
re
e
d
h
a
r,
“
A
Ne
w
A
p
p
ro
a
c
h
to
Im
a
g
e
De
n
o
isin
g
b
y
P
a
tch
-
Ba
se
d
A
lg
o
rit
h
m
”
,
in
In
ter
n
a
t
io
n
a
l
Jo
u
r
n
a
l
o
f
A
d
va
n
ced
R
esea
r
ch
in
C
o
mp
u
ter
a
n
d
C
o
mmu
n
ica
ti
o
n
E
n
g
i
n
eerin
g
,
v
o
l.
5
,
n
o
.
1
2
,
pp.
2
1
2
-
2
1
8
,
2
0
1
6
.
[2
4
]
Ya
s
m
in
,
S
a
b
i
n
a
,
a
n
d
M
d
M
a
su
d
Ra
n
a
,
“
P
e
rf
o
rm
a
n
c
e
S
tu
d
y
o
f
S
o
f
t
L
o
c
a
l
Bin
a
ry
P
a
tt
e
rn
o
v
e
r
L
o
c
a
l
Bin
a
ry
P
a
tt
e
r
n
u
n
d
e
r
N
o
isy
I
m
a
g
e
s
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
ter
E
n
g
i
n
e
e
rin
g
,
v
o
l.
6
,
n
o
.
3
(
2
0
1
6
),
p
.
1
1
6
1
.
[2
5
]
O
m
id
io
ra
,
E.
O.,
S
.
O.
Ola
b
iy
isi,
J.
A
.
Ojo
,
A
b
a
y
o
m
i
-
A
ll
i
A
d
e
b
a
y
o
,
O.
A
b
a
y
o
m
i
-
A
ll
i,
a
n
d
K.
B.
Eram
e
h
,
“
F
a
c
ial
Im
a
g
e
V
e
ri
f
ica
ti
o
n
a
n
d
Q
u
a
li
ty
A
s
se
ss
m
e
n
t
S
y
ste
m
-
F
a
c
e
IV
QA
”
,
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
i
n
e
e
rin
g
,
v
o
l.
3
,
n
o
.
6
(2
0
1
3
),
p
.
8
6
3
.
[2
6
]
S
u
b
ra
m
a
n
y
a
m
,
M
.
V
.
,
a
n
d
G
iri
P
ra
sa
d
,
“
A
N
e
w
A
p
p
ro
a
c
h
f
o
r
S
A
R
I
m
a
g
e
De
n
o
isin
g
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
,
v
o
l.
5
,
n
o
.
5
(
2
0
1
5
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.