T
E
L
KO
MNIK
A
, V
ol
.
17
,
No.
6,
Dec
e
mb
er
20
1
9,
p
p.
2
95
9
~
2967
IS
S
N: 1
69
3
-
6
93
0
,
accr
ed
ited
F
irst
Gr
ad
e b
y K
em
en
r
istekdikti,
Decr
ee
No: 2
1/E/
K
P
T
/20
18
DOI:
10.12928/TE
LK
OM
N
IK
A
.v
1
7
i
6
.
11301
◼
29
59
Rec
ei
v
ed
S
ep
te
mb
er 26
,
2
01
8
;
Rev
i
s
ed
F
eb
r
u
ary
2
, 2
01
9
;
A
c
c
ep
te
d
M
arc
h
12
, 2
01
9
An
ov
ervie
w
of
th
e f
un
da
men
ta
l app
r
oach
e
s
th
at
yield
sever
al imag
e de
no
is
ing
t
ech
niq
ues
Bilal
Ch
ar
mo
u
t
i
*
1
, Ahm
ad
Kadri
Jun
o
h
2
, M
o
h
d
Y
u
s
o
f
f
Ma
sho
r
3
,
Naja
h
G
h
az
a
li
4
,
Ma
h
yun
Ab
W
ah
ab
5
,
W
an
Z
u
ki
Az
man
W
an Muh
am
ad
6
,
Z
ainab Y
ahya
7
,
Ab
d
es
se
l
am Be
r
o
u
al
8
1
,2
,4
,6
,7
In
s
t
i
tu
t
e
o
f
E
n
g
i
n
e
e
r
i
n
g
M
a
th
e
m
a
ti
c
s
,
Un
i
v
e
r
s
i
t
i
M
a
l
a
y
s
i
a
P
e
rl
i
s
(Un
i
M
AP),
Pe
r
l
i
s
,
M
a
l
a
y
s
i
a
3
Sc
h
o
o
l
o
f
M
e
c
h
a
tro
n
i
c
,
U
n
i
v
e
rs
i
ti
M
a
l
a
y
s
i
a
Pe
rl
i
s
(Un
i
M
AP),
Pe
rl
i
s
,
M
a
l
a
y
s
i
a
5
Sc
h
o
o
l
o
f
En
v
i
ro
n
m
e
n
ta
l
E
n
g
i
n
e
e
ri
n
g
,
Un
i
v
e
r
s
i
t
i
M
a
l
a
y
s
i
a
P
e
rl
i
s
(Un
i
M
AP),
Pe
r
l
i
s
,
M
a
l
a
y
s
i
a
8
Dep
a
rt
m
e
n
t
o
f
Co
m
p
u
t
e
r Sc
i
e
n
c
e
,
In
te
r
n
a
ti
o
n
a
l
I
s
l
a
m
i
c
Un
i
v
e
rs
i
t
y
M
a
l
a
y
s
i
a
,
M
a
l
a
y
s
i
a
*C
o
rre
s
p
o
n
d
i
n
g
a
u
th
o
r,
e
-
m
a
i
l
:
d
y
b
i
l
e
l
@y
a
h
o
o
.f
r
Ab
strac
t
Dig
i
ta
l
i
m
a
g
e
i
s
c
o
n
s
i
d
e
r
e
d
a
s
a
p
o
werf
u
l
to
o
l
to
c
a
rry
a
n
d
t
ra
n
s
m
i
t
i
n
fo
rm
a
ti
o
n
b
e
twe
e
n
p
e
o
p
l
e
.
Th
u
s
,
i
t
a
tt
ra
c
ts
th
e
a
tt
e
n
ti
o
n
o
f
l
a
r
g
e
n
u
m
b
e
r
o
f
re
s
e
a
rc
h
e
r
s
,
a
m
o
n
g
th
e
m
t
h
o
s
e
i
n
te
re
s
t
e
d
i
n
p
re
s
e
rv
i
n
g
th
e
i
m
a
g
e
fe
a
t
u
re
s
fro
m
a
n
y
f
a
c
to
r
s
t
h
a
t
m
a
y
r
e
d
u
c
e
th
e
i
m
a
g
e
q
u
a
l
i
t
y
.
O
n
e
o
f
th
e
s
e
fa
c
t
o
rs
i
s
th
e
n
o
i
s
e
whi
c
h
a
ff
e
c
t
s
th
e
v
i
s
u
a
l
a
s
p
e
c
t
o
f
th
e
i
m
a
g
e
a
n
d
m
a
k
e
s
o
th
e
rs
i
m
a
g
e
p
ro
c
e
s
s
i
n
g
m
o
r
e
d
i
ff
i
c
u
l
t.
Th
u
s
fa
r
,
s
o
l
v
i
n
g
th
i
s
n
o
i
s
e
p
r
o
b
l
e
m
re
m
a
i
n
s
a
c
h
a
l
l
e
n
g
e
fo
r
th
e
re
s
e
a
rc
h
e
rs
i
n
th
i
s
fi
e
l
d
.
A
l
o
t
o
f
i
m
a
g
e
d
e
n
o
i
s
i
n
g
te
c
h
n
i
q
u
e
s
h
a
v
e
b
e
e
n
i
n
tro
d
u
c
e
d
i
n
o
rd
e
r
t
o
re
m
o
v
e
th
e
n
o
i
s
e
b
y
ta
k
i
n
g
c
a
re
o
f
th
e
i
m
a
g
e
fe
a
t
u
re
s
;
i
n
o
th
e
r
word
s
,
g
e
t
ti
n
g
th
e
b
e
s
t
s
i
m
i
l
a
ri
t
y
to
th
e
o
ri
g
i
n
a
l
i
m
a
g
e
fro
m
th
e
n
o
i
s
y
o
n
e
.
Howe
v
e
r,
th
e
f
i
n
d
i
n
g
s
a
re
s
ti
l
l
i
n
c
o
n
c
l
u
s
i
v
e
.
Be
s
i
d
e
th
e
e
n
o
rm
o
u
s
a
m
o
u
n
t
o
f
r
e
s
e
a
r
c
h
e
s
a
n
d
s
tu
d
i
e
s
whi
c
h
a
d
o
p
t
s
e
v
e
r
a
l
m
a
th
e
m
a
ti
c
a
l
c
o
n
c
e
p
ts
(
s
ta
ti
s
ti
c
s
,
p
ro
b
a
b
i
l
i
ti
e
s
,
m
o
d
e
l
i
n
g
,
P
DEs
,
wa
v
e
l
e
t,
fu
z
z
y
l
o
g
i
c
,
e
t
c
.),
th
e
re
i
s
a
l
s
o
th
e
s
c
a
r
c
i
t
y
o
f
r
e
v
i
e
w
p
a
p
e
rs
whi
c
h
c
a
rr
y
a
n
i
m
p
o
rta
n
t
ro
l
e
i
n
th
e
d
e
v
e
l
o
p
m
e
n
t
a
n
d
p
r
o
g
r
e
s
s
o
f
re
s
e
a
r
c
h
.
Th
u
s
,
th
i
s
re
v
i
e
w
p
a
p
e
r
i
n
t
o
rd
u
c
e
a
n
o
v
e
r
v
i
e
w
o
f
t
h
e
d
i
ff
e
r
e
n
t
f
u
n
d
a
m
e
n
ta
l
a
p
p
r
o
a
c
h
e
s
th
a
t
y
i
e
l
d
th
e
s
e
v
e
ra
l
i
m
a
g
e
-
d
e
n
o
i
s
i
n
g
te
c
h
n
i
q
u
e
s
,
p
r
e
s
e
n
te
d
wit
h
a
n
e
w
c
l
a
s
s
i
f
i
c
a
ti
o
n
.
Fu
rth
e
rm
o
re
,
th
e
p
a
p
e
r
p
re
s
e
n
ts
th
e
d
i
ff
e
re
n
t
e
v
a
l
u
a
ti
o
n
to
o
l
s
n
e
e
d
e
d
o
n
th
e
c
o
m
p
a
ri
s
o
n
b
e
twe
e
n
th
e
s
e
te
c
h
n
i
q
u
e
s
i
n
o
r
d
e
r
t
o
fa
c
i
l
i
ta
te
t
h
e
p
ro
c
e
s
s
i
n
g
o
f
th
i
s
n
o
i
s
e
p
r
o
b
l
e
m
,
a
m
o
n
g
a
g
re
a
t
d
i
v
e
rs
i
ty
o
f
te
c
h
n
i
q
u
e
s
a
n
d
c
o
n
c
e
p
ts
.
Key
w
ords
:
d
e
n
o
i
s
i
n
g
m
e
th
o
d
s
,
d
i
g
i
t
a
l
i
m
a
g
e
,
i
m
a
g
e
d
e
n
o
i
s
i
n
g
,
i
m
a
g
e
p
ro
c
e
s
s
i
n
g
,
n
o
i
s
e
Copy
righ
t
©
2
0
1
9
Uni
v
e
rsi
t
a
s
Ahm
a
d
D
a
hl
a
n.
All
rig
ht
s
r
e
s
e
rve
d
.
1.
Int
r
o
d
u
ctio
n
A
mo
ng
the
di
ffe
r
e
nt
t
oo
l
s
of
c
om
mu
n
i
c
at
i
on
be
twee
n
pe
o
pl
e,
t
he
r
e
i
s
th
e
i
ma
ge
whi
c
h
c
arr
i
es
a
l
arge
am
o
un
t
o
f
i
nfo
r
ma
ti
o
n.
How
ev
er,
un
f
or
tun
at
el
y
,
th
e
i
ma
g
e
c
an
be
c
orr
up
ted
by
pa
r
as
i
t
i
c
i
nfo
r
m
ati
on
,
wh
i
c
h
i
s
c
al
l
ed
n
oi
s
e:
an
a
l
ter
ati
o
n
of
i
m
ag
e,
wh
i
c
h
ma
y
be
c
au
s
ed
by
the
i
ma
ge
ac
qu
i
s
i
t
i
on
proc
e
s
s
or tr
an
s
mi
s
s
i
o
n [
1,
2
]. T
he
m
ai
n c
on
c
ern
of
th
e
r
es
ea
r
c
he
r
s
i
n
th
i
s
s
ub
j
ec
t
i
s
to
s
uc
c
ee
d
i
n
s
ol
v
i
n
g
th
i
s
prob
l
em
by
r
em
ov
i
n
g
no
i
s
e
fr
om
t
he
n
o
i
s
y
i
ma
g
e
an
d
ac
hi
ev
e
th
e
be
s
t
r
es
torati
o
n
of
t
he
orig
i
n
al
o
ne
.
T
o
a
c
hi
ev
e
the
d
es
i
r
ed
a
i
m,
th
ey
ne
ed
am
pl
e
an
d
n
ec
es
s
ary
k
no
wl
e
dg
e
,
c
on
c
ep
ts
an
d
de
fi
n
i
t
i
on
s
,
whi
c
h
ar
e
i
ns
p
i
r
ed
fr
om
the
r
ev
i
e
w
of
r
el
ev
an
t
l
i
terat
ure.
Im
ag
e
p
r
oc
es
s
i
ng
i
s
qu
i
te
a
wi
de
fi
el
d
o
f
k
no
w
l
ed
ge
d
ue
to
i
t
s
i
mp
orta
nc
e,
an
d
wi
de
l
y
a
pp
l
i
e
d
i
n
s
ev
eral
f
i
el
ds
l
i
k
e
bi
ol
o
gy
,
as
tr
on
om
y
,
i
nd
us
tr
i
al
,
m
ed
i
c
al
an
d
de
fe
nc
e.
T
hu
s
,
a
l
arge
am
ou
nt
of
r
es
ea
r
c
he
s
an
d
s
tu
di
es
h
av
e
b
ee
n
do
ne
ab
ou
t
th
i
s
area.
W
he
r
e,
am
on
g
th
e
ex
i
s
ti
ng
b
arr
i
er
s
i
n
r
es
tori
n
g
the
c
orr
up
te
d
i
m
ag
e
i
s
the
s
c
arc
i
ty
of
r
ev
i
ew
pa
pe
r
s
whi
c
h
c
arr
y
an
i
mp
ortan
t
r
ol
e
i
n
t
he
de
v
e
l
o
pm
e
nt
an
d
progr
es
s
of
r
es
ea
r
c
h,
In
th
i
s
pa
pe
r
we
wi
l
l
tak
e
a
c
l
os
er
l
oo
k
i
nt
o
th
e
ne
c
es
s
ary
i
nfo
r
m
ati
on
r
es
tr
i
c
ted
i
n
the
pr
i
m
ary
ph
as
e
of
i
ma
ge
proc
es
s
i
ng
(
i
ma
g
e
pre
-
pr
oc
es
s
i
ng
)
,
c
al
l
ed
I
ma
ge
de
no
i
s
i
n
g,
p
arti
c
ul
arly
the
fun
d
a
me
nta
l
ap
pro
ac
he
s
of
i
m
ag
e
de
n
oi
s
i
n
g.
W
e
s
e
ek
to
ga
th
e
r
an
d
i
ntro
du
c
e
t
he
s
e
ap
proac
he
s
a
nd
c
on
c
ep
ts
s
mo
oth
l
y
,
w
i
th
g
oo
d
arr
an
g
em
e
nt,
wher
e
t
he
r
e
ad
ers
c
an
p
en
etr
ate
i
nto
t
he
s
ub
j
ec
t,
i
n
order
to
f
ac
i
l
i
tat
e
the
p
at
h
tha
t
l
ea
ds
to
th
e
go
al
,
by
ma
k
i
n
g
thi
s
pa
pe
r
a
s
tarti
n
g
po
i
nt
of
ot
he
r
s
tud
i
es
c
on
c
ern
i
ng
th
e
i
m
a
ge
de
n
oi
s
i
ng
.
T
hu
s
,
i
ma
g
e,
di
gi
t
al
i
ma
g
e,
no
i
s
e,
no
i
s
y
i
ma
ge
,
i
ma
g
e
de
no
i
s
i
n
g,
a
pp
r
oa
c
h
es
for
i
ma
ge
d
en
o
i
s
i
n
g,
as
we
l
l
as
me
th
od
c
om
pa
r
i
s
o
n,
ar
e
pres
en
t
ed
i
n
the
r
es
t
of
th
e
pa
p
er
by
the
fo
l
l
owi
ng
ma
i
n
s
ec
t
i
on
s
.
S
ec
ti
o
n
2:
d
i
s
c
us
s
i
on
a
bo
ut
r
e
l
ev
an
t
ma
ter
i
al
s
a
nd
to
ol
s
,
wh
i
c
h
pres
en
t
de
f
i
n
i
ti
on
s
an
d
c
on
c
ep
ts
ab
ou
t
i
ma
ge
,
n
oi
s
e
an
d
f
i
l
t
erin
g.
S
ec
ti
on
3:
Di
s
c
us
s
i
on
ab
ou
t
th
e
de
no
i
s
i
n
g
me
t
ho
d
s
,
al
s
o
th
e
de
s
c
r
i
pt
i
o
n
of
the
b
as
i
s
of
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
MNIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
b
er 20
19
:
29
5
9
-
2967
2960
de
no
i
s
i
n
g
ap
proac
he
s
.
S
ec
ti
on
4
:
bo
t
h
the
ev
a
l
u
ati
o
n
an
d
c
o
mp
ar
i
s
on
of
the
de
n
oi
s
i
ng
me
th
od
s
. Fi
na
l
l
y
,
s
ec
ti
on
5
off
ers
th
e
c
on
c
l
us
i
on
.
2.
Ma
t
er
ial
s
and
T
o
o
ls
2
.1.
Im
age
A
n
i
m
ag
e
i
s
a
s
c
en
e
r
e
pres
en
tat
i
o
n
thro
ug
h
pa
i
n
ti
n
g,
s
c
ul
ptu
r
es
,
draw
i
ng
,
p
ho
to
gr
ap
hy
,
an
d
s
o
on
.
It
i
s
al
s
o
a
s
tr
u
c
tured
s
et
of
i
nf
orma
ti
o
n
th
at,
aft
er
th
e
di
s
p
l
ay
on
th
e
s
c
r
ee
n,
gi
v
es
me
an
i
n
g
to
hu
ma
n
.
It
ma
y
be
de
fi
ne
d
as
a
two
-
di
me
ns
i
on
al
fu
nc
ti
o
n
I
(
x
,
y
)
an
al
o
g
brig
htn
es
s
c
on
ti
n
uo
us
,
d
efi
ne
d
i
n
a
b
ou
nd
ed
do
ma
i
n,
x
a
nd
y
as
the
s
pa
ti
a
l
c
oo
r
d
i
na
t
es
of
a
po
i
nt
o
f
the
i
m
ag
e
a
nd
I
i
s
a
f
un
c
ti
on
of
l
um
i
n
an
c
e
i
n
ten
s
i
ty
an
d
c
o
l
or.
In
th
i
s
as
p
ec
t,
the
i
ma
ge
i
s
un
us
ab
l
e
by
th
e
ma
c
h
i
n
e,
whi
c
h re
qu
i
r
es
di
gi
t
i
z
at
i
on
[
3
].
T
he
di
g
i
ta
l
i
m
ag
e
,
i
n
i
ts
br
oa
de
s
t
s
en
s
e,
r
ef
ers
to
an
y
i
m
ag
e
ac
q
ui
r
ed
,
proc
es
s
ed
an
d
s
tored
i
n
a
c
od
e
d
f
orm
w
hi
c
h
ma
y
b
e
r
e
pres
en
te
d
by
nu
mb
ers
(
nu
m
eric
a
l
v
al
ue
s
)
.
Di
g
i
ti
z
a
ti
o
n
i
s
the
proc
es
s
wh
i
c
h
pe
r
m
i
ts
the
i
ma
g
e
t
o
mo
v
e
fr
om
t
he
ph
y
s
i
c
a
l
s
ta
te
(
o
pti
c
a
l
i
ma
ge
)
r
ep
r
es
en
te
d
by
c
on
t
i
n
uo
us
s
i
gn
a
l
(
a
n
i
nf
i
n
i
ty
of
i
nt
en
s
i
t
y
v
al
ue
)
,
to
t
he
di
gi
t
al
i
ma
g
e
s
tat
e
w
hi
c
h
i
s
c
ha
r
ac
teri
z
e
d
by
th
e
d
i
s
c
r
ete
as
pe
c
t
(
i
nte
ns
i
ty
tak
e
the
i
r
v
al
u
es
i
n
a
f
i
n
i
te
nu
mb
er
o
f
po
i
nts
)
.
T
hi
s
i
s
the
di
g
i
ta
l
for
m
whi
c
h
pe
r
mi
ts
a
f
urther
ex
p
l
o
i
tat
i
on
by
c
om
p
ute
r
s
oft
war
e
too
l
s
.
T
hu
s
,
the
di
g
i
ta
l
i
m
ag
e
i
s
a
b
ou
nd
ed
s
et
of
e
l
e
me
n
ts
,
wh
ere
ea
c
h
o
ne
of
t
he
m
i
s
de
term
i
n
ed
by
l
oc
at
i
on
c
o
ordi
na
te
an
d
v
al
ue
.
T
he
s
e
el
em
e
nts
are
n
am
ed
i
m
ag
e
el
em
e
nts
,
p
i
c
ture
el
em
en
ts
,
pi
x
el
s
,
an
d
pe
l
s
,
wh
er
ea
s
,
the
m
os
t
wi
de
l
y
us
ed
term
to
d
en
ot
e
th
e
d
i
g
i
ta
l
i
ma
g
e
el
em
e
nts
i
s
''
pi
x
el
''
.
A
c
c
ordi
n
g
to
th
e
l
u
mi
n
an
c
e
i
nte
ns
i
ty
v
al
u
e,
di
gi
ta
l
i
ma
g
e
i
s
c
l
as
s
i
f
i
ed
i
nt
o
three
ty
pe
s
(
c
ol
or
-
c
od
i
n
g): b
l
ac
k
an
d w
hi
te
gray
s
c
al
es
an
d c
o
l
o
ured s
c
al
es
.
T
he
b
i
n
ary
i
ma
g
es
are
a
n
arr
ay
of
i
nte
ge
r
s
(
k
=
1
bi
t)
,
pi
x
e
l
tha
t
c
an
tak
e
o
ne
of
the
v
a
l
u
es
:
0
or
1.
W
h
i
te
i
s
de
no
te
d
by
0
an
d
B
l
ac
k
b
y
1.
T
h
i
s
i
s
th
e
s
i
mp
l
es
t
ty
pe
s
of
i
ma
ge
s
an
d
us
u
al
l
y
thi
s
k
i
nd
i
s
us
ed
to
s
c
an
t
ex
ts
when
t
he
l
att
er
i
s
c
om
po
s
ed
of
on
e
c
ol
or.
T
he
gray
l
ev
el
i
s
the
v
al
u
e
of
th
e
l
u
mi
n
an
c
e
i
nte
ns
i
ty
at
on
e
p
oi
nt
.
T
hi
s
i
m
ag
e
c
on
s
i
s
ts
of
on
l
y
8
by
tes
,
an
d
the
p
i
x
el
c
o
l
or
c
a
n
t
ak
e
v
al
ue
s
r
an
g
i
ng
fr
o
m
bl
ac
k
t
o
w
hi
t
e
throu
gh
v
a
r
i
ou
s
l
ev
e
l
s
of
brig
ht
ne
s
s
[0,
1,
2,
...,
25
5].
Us
ua
l
l
y
c
ol
or
i
ma
ge
s
a
r
e
ba
s
ed
o
n
three
prim
ar
y
c
ol
ou
r
s
:
Red
,
G
r
ee
n
a
nd
B
l
ue
(
RG
B
)
,
an
d
ea
c
h
of
th
es
e
c
o
l
ou
r
s
us
e
8
bi
ts
for
on
e
p
i
x
el
wh
i
c
h
tak
es
a
v
al
ue
fr
om
th
e
r
a
ng
e
[0,
1,
2,
..
.,2
55
],
s
o
ea
c
h
p
i
x
el
i
n
the
c
ol
or
i
m
ag
e
r
eq
u
i
r
es
3×
8=
2
4
bi
ts
to
en
c
od
e
three c
o
mp
o
ne
nts
.
2
.
2
.
No
i
se
Noi
s
e
i
s
a
n
a
l
terat
i
o
n
of
i
ma
ge
(
pa
r
as
i
tes
i
nfo
r
m
at
i
o
n)
tha
t
ma
y
be
c
r
ea
ted
i
n
on
e
of
the
s
e
ph
as
es
,
du
r
i
ng
th
e
ac
qu
i
s
i
ti
o
n
proc
es
s
(
c
on
v
ers
i
on
op
erati
on
fr
o
m
op
t
i
c
al
s
i
gn
al
s
to
the
el
ec
tr
i
c
al
th
en
fr
om
el
e
c
tr
i
c
al
t
o
d
i
g
i
ta
l
s
i
gn
al
)
thro
ug
h
the
tr
a
ns
mi
s
s
i
o
n,
s
e
ns
or
s
tat
us
an
d
en
v
i
r
on
me
n
tal
c
on
di
t
i
o
ns
[
1
,
4
].
Th
e
no
i
s
e
ha
s
di
ff
erent
ori
gi
ns
,
b
ut
i
t
c
au
s
es
s
i
mi
l
a
r
ef
fec
ts
s
uc
h
as
the
l
os
s
of
s
h
arpnes
s
i
n
the
de
ta
i
l
s
or
the
ap
pe
ara
nc
e
of
grai
ns
,
where
the
a
mo
un
t
of
no
i
s
e
i
n
no
i
s
y
i
ma
ge
i
s
as
s
i
gn
e
d
by
th
e
n
um
be
r
o
f
c
orr
up
t
e
d
p
i
x
el
s
.
In
t
erms
of
r
em
ov
i
ng
th
i
s
no
i
s
e
fr
om
t
he
c
orr
up
te
d
i
ma
ge
,
the
r
es
ea
r
c
he
r
s
of
thi
s
fi
el
d
ha
v
e
m
ad
e
s
om
e
i
n
-
d
ep
th
s
tud
i
es
t
o
k
no
w
the
na
ture
an
d
ty
p
e
s
of
no
i
s
e.
T
h
e
i
ma
g
e
i
s
c
orr
up
ted
d
ue
to
t
he
fac
t
tha
t
th
ere
are
v
ario
us
ty
pe
s
of
n
oi
s
e
s
uc
h
as
th
e
G
au
s
s
i
a
n
n
oi
s
e,
P
oi
s
s
on
no
i
s
e,
S
pe
c
k
l
e
n
o
i
s
e,
S
al
t
an
d
P
ep
pe
r
n
oi
s
e
an
d
ma
ny
mo
r
e f
un
d
am
en
ta
l
n
oi
s
e
ty
pe
s
i
n t
he
c
as
e
of
d
i
g
i
ta
l
i
ma
ge
s
[
5
].
Noi
s
e
i
s
i
nt
eg
r
at
ed
i
n
t
he
i
ma
ge
(
,
)
throug
h
three
form
s
of
no
i
s
e
(
,
)
,
na
me
l
y
ad
d
i
ti
v
e,
m
ul
t
i
p
l
i
c
at
i
v
e,
an
d
i
mp
u
l
s
e
no
i
s
e
[
6
,
7
],
to
gi
v
e
a
no
i
s
y
i
ma
ge
de
no
ted
by
(
,
)
.
T
he
ad
di
t
i
v
e
no
i
s
e
fo
r
m
ul
a
i
s
s
ho
wn b
y
(
1) :
v
(
x
,
y
)
=
u
(
x
,
y
)
+
n
(
x
,
y
)
(
1)
t
he
m
ul
t
i
p
l
i
c
a
ti
v
e
no
i
s
e
fo
r
mu
l
a i
s
mo
d
el
e
d b
y
(
2
)
:
v
(
x
,
y
)
=
u
(
x
,
y
)
×
n
(
x
,
y
)
(
2)
3.
Deno
i
sing
Me
t
h
o
d
s
Ima
g
e
d
en
o
i
s
i
n
g
i
s
a
ph
as
e,
wh
i
c
h
pr
ec
ed
es
the
i
ma
g
e
pr
oc
es
s
i
ng
(
i
ma
ge
pre
-
proc
es
s
i
ng
)
;
i
n
ot
he
r
wor
ds
,
a
s
et
of
op
er
ati
on
s
oc
c
urin
g
on
th
e
i
ma
ge
,
wh
i
c
h
a
i
ms
to
am
é
l
i
orate
the
v
i
s
ua
l
as
p
ec
t
of
t
he
i
m
ag
e
by
m
ea
ns
of
r
ed
uc
i
ng
or
r
em
ov
i
ng
no
i
s
e
fr
om
i
ma
g
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
MNIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
A
n o
v
erv
i
e
w o
f th
e f
u
nd
a
m
en
ta
l
a
pp
r
oa
c
h
es
tha
t y
i
e
l
d
s
ev
eral
i
ma
g
e..
. (
B
i
l
a
l
Ch
armout
i
)
2961
T
hi
s
i
s
i
n
ord
er
to
fac
i
l
i
t
ate
oth
er
proc
es
s
es
s
uc
h
as
the
c
om
pres
s
i
on
,
a
na
l
y
s
es
,
c
l
as
s
i
fi
c
at
i
on
,
s
eg
me
n
tat
i
on
,
an
d
ex
tr
ac
ti
on
of
i
nf
orma
ti
o
n
[
8
].
De
no
i
s
i
ng
m
eth
od
s
are
ba
s
ed
on
the
ma
i
n
i
de
a
whi
c
h
i
s
t
o
r
ep
l
ac
e
ea
c
h
c
orr
up
ted
p
i
x
el
v
al
u
e
(
n
oi
s
e
)
i
n
the
no
i
s
y
i
ma
g
e
w
i
th
an
ot
he
r
v
a
l
ue
,
ma
k
i
n
g
t
ha
t
i
ma
g
e
c
l
os
er
to
th
e
orig
i
n
al
on
e
(
n
oi
s
e
fr
ee
i
ma
ge
)
as
m
uc
h
as
p
os
s
i
bl
e
,
us
i
ng
the
ne
c
es
s
ary
k
no
w
l
e
dg
e
a
bo
ut
no
i
s
e
an
d
i
m
ag
e
c
i
t
ed
i
n
t
he
prev
i
o
us
s
ec
ti
on
.
In
oth
er
w
ords
,
the
de
n
oi
s
i
ng
m
eth
o
ds
a
i
m
t
o
ac
h
i
ev
e
th
e
v
ers
i
on
c
l
os
es
t
to
th
e
or
i
g
i
na
l
i
ma
ge
.
S
ev
era
l
me
th
od
s
ha
v
e
b
ee
n
d
ev
el
op
ed
i
n
th
i
s
way
,
al
t
ho
ug
h
ea
c
h
me
tho
d
ha
s
i
ts
a
dv
an
ta
ge
s
an
d
di
s
ad
v
a
nta
g
es
.
T
he
s
e
de
no
i
s
i
ng
me
t
ho
ds
ma
y
be
c
l
as
s
i
fi
ed
i
nto
two
ma
i
n
c
ate
g
orie
s
,
the
s
p
ati
al
do
m
ai
n
a
nd
tr
an
s
form
do
ma
i
n
f
i
l
teri
ng
w
hi
c
h
me
r
ge
d
tog
eth
er
i
n
s
om
e
c
as
es
to
y
i
el
d
hy
brid
f
i
l
t
ers
.
3.1
.
S
p
atial
Do
main
Fi
lt
er
i
n
g
S
pa
t
i
a
l
do
m
ai
n
f
i
l
t
erin
g
i
s
c
on
s
i
de
r
e
d
as
a
tr
a
di
t
i
o
na
l
way
to
r
em
ov
e
c
orr
up
t
ed
pi
x
e
l
s
(
no
i
s
e)
fr
om
the
i
ma
ge
[
9
].
It
i
s
a
s
et
of
ma
t
he
m
ati
c
a
l
op
erat
i
o
ns
th
at
de
a
l
di
r
ec
t
l
y
wi
th
th
e
pi
x
e
l
i
n
t
he
i
m
ag
e
p
l
an
e,
b
ec
au
s
e
i
n
th
i
s
do
ma
i
n,
t
he
s
i
g
na
l
i
s
r
ep
r
es
en
te
d
by
pi
x
e
l
s
,
or
i
n
oth
er
wor
ds
,
i
ma
ge
e
l
e
me
nts
are
pi
x
el
s
.
T
he
f
i
l
t
erin
g
op
era
ti
on
s
are
ap
p
l
i
ed
i
nto
th
es
e
pi
x
el
s
,
b
as
ed
on
ma
t
he
ma
t
i
c
s
c
on
c
ep
ts
,
throug
h
whi
c
h
w
e
ap
pl
y
the
fi
l
t
erin
g
wi
nd
o
w
t
o
ea
c
h
pi
x
e
l
i
n
t
he
who
l
e
no
i
s
y
i
ma
ge
(
ov
e
r
al
l
proc
es
s
i
n
g),
i
n
th
e
c
as
e
th
at
the
d
ete
c
ti
on
ph
as
e
i
s
ab
s
en
t
,
or
to
s
om
e
p
i
x
el
s
of
no
i
s
y
i
ma
g
e
(
pa
r
ti
al
proc
es
s
i
ng
)
,
or
i
n
the
c
as
e
t
ha
t
w
e
fi
r
s
tl
y
pa
s
s
throug
h
the
de
t
ec
ti
o
n
ph
as
e
(
No
i
s
e
d
ete
c
ti
on
)
us
i
n
g
s
ev
er
al
de
t
ec
tor
t
ec
hn
i
qu
es
s
uc
h
as
t
he
fuz
z
y
tec
hn
i
qu
es
as
s
ho
wn
i
n
F
i
gu
r
e
1
.
G
e
ne
r
a
l
l
y
,
th
e
fi
l
ter
tr
ea
ts
the
n
oi
s
y
i
m
ag
e
thr
ou
gh
t
wo
way
s
,
l
i
n
ea
r
or
no
n
-
l
i
n
ea
r
as
s
h
o
wn i
n
F
i
g
ure
1
.
F
i
g
ure
1
.
B
as
i
c
p
ath
of
s
pa
t
i
al
fi
l
teri
ng
proc
es
s
3.1
.1
.
L
inea
r
F
ilte
r
ing
T
he
l
i
ne
ar
fi
l
ter
tr
an
s
f
orms
an
i
n
pu
t
da
ta
s
e
t
i
n
to
a
s
et
of
ou
t
pu
t
da
ta
ac
c
ordi
ng
to
a
m
ath
em
a
ti
c
a
l
op
era
ti
o
n
c
al
l
e
d
c
o
nv
ol
uti
on
.
It
a
l
l
o
w
s
for
ea
c
h
pi
x
e
l
i
n
t
he
ar
ea
t
o
wh
i
c
h
i
t
ap
p
l
i
es
,
t
o
c
ha
ng
e
i
ts
v
al
u
e
throug
h
a
l
i
ne
ar
c
om
bi
n
ati
o
n
of
i
ts
n
ei
g
hb
ors
.
T
he
l
i
n
ea
r
fi
l
ter
i
ng
op
erat
i
o
n
i
s
e
as
y
to
i
mp
l
em
en
t
for
r
e
mo
v
i
ng
no
i
s
e
bu
t
us
ua
l
l
y
i
t
produc
es
b
l
urr
ed
i
ma
ge
s
,
wh
i
c
h
are th
e m
ai
n
de
fec
t o
f th
i
s
f
i
l
t
er [
10
,
11
].
Ma
ny
w
el
l
k
no
wn l
i
n
ea
r
f
i
l
t
ers
ha
v
e
be
e
n i
ntrodu
c
e
d l
i
k
e
the
m
ea
n
fi
l
ter
an
d
W
i
en
er
fi
l
t
er.
3.1
.2
.
No
n
-
linea
r
F
ilte
r
ing
T
he
y
are
d
es
i
g
ne
d
to
s
o
l
v
e
th
e
prob
l
e
ms
of
th
e
l
i
ne
ar
fi
l
ter,
es
p
ec
i
a
l
l
y
wi
t
h
r
e
ga
r
d
to
the
po
or
pres
erv
at
i
on
of
e
d
ge
s
[
1
0,
11
];
T
h
ei
r
prin
c
i
pl
e
i
s
t
he
s
am
e
as
the
l
i
n
ea
r
f
i
l
ters
,
w
he
r
e
i
t
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
MNIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
b
er 20
19
:
29
5
9
-
2967
2962
al
way
s
r
e
pl
ac
es
th
e
v
a
l
u
e
of
e
ac
h
pi
x
e
l
wi
t
h
t
he
c
al
c
u
l
ate
d
v
al
ue
fr
om
a
m
ath
em
a
ti
c
al
o
pe
r
at
i
on
ap
p
l
i
e
d
t
o
i
ts
n
ei
g
hb
ors
’
v
al
ue
s
,
i
nc
l
u
di
ng
i
ts
e
l
f.
T
he
di
f
ferenc
e
i
s
t
ha
t
thi
s
op
erati
o
n
i
s
no
l
o
ng
er
l
i
n
ea
r
.
T
he
r
e
are
m
an
y
po
pu
l
ar
no
n
-
l
i
n
ea
r
fi
l
ters
s
uc
h
as
the
m
ed
i
an
an
d
i
ts
ex
ten
s
i
o
ns
(
order
s
tat
i
s
ti
c
f
i
l
t
ers
)
,
bi
l
ate
r
a
l
f
i
l
t
er,
T
ota
l
v
aria
t
i
o
n
(
T
V
)
,
m
o
r
ph
ol
og
i
c
al
f
i
l
t
er,
a
ni
s
otro
pi
c
fi
l
ter
an
d
s
o
on.
O
t
he
r
c
h
arac
teri
s
ti
c
fi
l
t
erin
g
m
ay
d
i
ffe
r
be
tw
ee
n
s
ev
eral
d
en
o
i
s
i
ng
ap
pro
ac
he
s
,
s
uc
h
as
s
tat
i
s
ti
c
a
l
/d
ete
r
m
i
n
i
s
ti
c
,
a
d
ap
ti
v
e
/no
n
-
ad
a
pti
v
e
an
d
l
o
c
al
/no
n
-
l
oc
a
l
.
In
[
1
2
-
1
5
],
s
o
me
ex
am
pl
es
are pres
en
ted
.
3.2
. T
r
ansf
o
r
m
Do
main
Fil
t
er
ing
T
he
tr
an
s
f
orm
d
om
a
i
n
fi
l
ter
i
ng
i
s
th
e
s
ec
on
d
c
ate
g
ory
of
i
ma
g
e
d
en
o
i
s
i
n
g
a
pp
r
o
ac
he
s
,
whi
c
h
g
ai
n
ed
e
no
r
m
ou
s
i
n
t
eres
t
fr
om
s
c
ho
l
ars
an
d
s
tu
di
es
.
It
m
ay
be
d
efi
ne
d
as
a
s
et
of
fi
l
teri
ng
op
erat
i
o
ns
t
h
at
tr
ea
ts
t
he
i
ma
ge
i
n
an
o
the
r
form
(
oth
er
do
ma
i
ns
)
s
uc
h
as
fr
e
qu
en
c
y
do
ma
i
n
(
F
i
gu
r
e
2
)
,
r
ath
er
tha
n
t
he
orig
i
na
l
f
orm
(
ori
gi
na
l
do
ma
i
n),
i
n
ord
er
to
ga
i
n
f
urt
he
r
i
nfo
r
m
ati
on
fr
om
th
e
s
i
g
na
l
an
d
r
ea
c
h
a
s
uc
c
es
s
ful
fi
l
ter
i
ng
throu
gh
th
i
s
ne
w
form.
T
he
ne
w
do
m
ai
n
ma
y
di
ff
er
fr
om
the
orig
i
na
l
on
e
i
n
th
e
di
me
ns
i
on
s
(
e.g
.
fr
om
2D
to
3
D)
or
i
n
th
e
c
ha
r
ac
teri
s
ti
c
s
(
e.g
.
fr
o
m
s
pa
ti
a
l
t
o
fr
e
qu
en
c
y
do
m
ai
n).
T
h
i
s
tr
an
s
f
orm
i
s
a
m
ath
em
a
ti
c
a
l
c
o
nv
ers
i
on
t
ha
t
i
s
ac
hi
ev
e
d
thro
ug
h
a
c
l
ev
er
ma
th
em
ati
c
a
l
to
ol
(
ba
s
i
s
fun
c
ti
on
)
.
T
h
e
i
ma
g
e
f
i
l
t
eri
ng
proc
es
s
i
n
the
tr
an
s
form
d
om
a
i
n
c
an
s
pl
i
t
i
nto
tw
o
branc
h
es
ac
c
or
di
n
g
t
o
t
he
ty
p
e
of
th
i
s
ba
s
i
s
fun
c
ti
on
[
9
]
;
ad
ap
ti
v
e
to
proc
es
s
ed
da
t
a
an
d
no
n
-
a
da
p
ti
v
e.
F
or
t
he
ad
ap
t
i
v
e
ap
proac
he
s
t
he
r
e
are
tw
o
eff
ec
ti
v
e
tec
hn
i
q
ue
s
na
me
l
y
the
P
r
i
nc
i
pa
l
c
om
po
ne
n
t
an
a
l
y
s
i
s
(
P
C
A
)
[1
6
]
,wh
i
c
h
tr
ea
ts
o
nl
y
the
da
ta
i
nfo
r
m
ati
on
gi
v
e
n
by
th
e
s
ec
on
d
-
order
s
ta
t
i
s
ti
c
s
,
an
d
t
he
I
nd
e
pe
nd
e
nt
c
om
po
n
en
t
an
a
l
y
s
i
s
(
ICA
)
[
17
] w
hi
c
h
c
om
es
as
a
n
ex
ten
s
i
on
to
P
CA
t
o
gi
v
e
b
ett
er
pe
r
f
orma
nc
e,
by
l
i
v
i
n
g
up
to
hi
g
h
order
s
tat
i
s
ti
c
s
(
the
c
as
e
of
the
m
os
t
na
tu
r
al
i
m
ag
es
)
[
18
]
.
T
he
ma
i
n
i
de
a
of
bo
t
h
s
tat
i
s
ti
c
a
l
t
ec
hn
i
qu
es
,
P
CA
an
d
IC
A
,
i
s
t
o
us
e
a
n
ortho
go
na
l
de
c
o
mp
os
i
ti
on
to
s
e
pa
r
ate
l
i
n
ea
r
l
y
as
mu
c
h
as
p
os
s
i
b
l
e
t
he
c
orr
el
at
ed
d
ata
i
nt
o
i
nd
e
pe
n
de
nt
s
u
b
-
s
ets
[
1
9]
.
In
t
he
no
n
-
a
da
pt
ati
v
e
ap
r
oa
c
h
es
the
r
e
are
m
an
y
me
th
od
s
d
ue
to
the
v
arie
ty
of
b
as
i
c
fu
nc
ti
o
ns
s
uc
h
as
,
wav
el
e
ts
[2
0
]
,
wav
e
at
om
s
[
21
],
c
urv
el
et
s
[2
2
],
c
o
nto
urle
ts
[2
3
]
,
w
ed
ge
l
ets
[2
4
]
,
a
nd
b
an
de
l
ets
[2
5
]
,
wh
i
c
h
tr
an
s
form
the
i
ma
ge
to
the
fr
eq
u
en
c
y
d
om
ai
n
.
F
r
o
m
t
ho
s
e,
the
o
l
de
s
t,
th
e
mo
s
t
po
pu
l
ar,
an
d
the
d
om
i
na
nt
o
ne
(
''
wav
e
l
et'')
[2
6
],
are
hi
gh
l
i
gh
te
d
i
n
t
he
f
ol
l
ow
i
ng
s
ec
ti
o
ns
.
Rec
en
t
l
y
i
n
the
tr
an
s
for
m
do
m
ai
n,
a
n
ew
eff
i
c
i
en
t
me
th
od
(
B
M3
D
)
i
s
de
v
el
op
ed
by
Dab
ov
et
al
.
[2
7
]
us
i
ng
s
pa
r
s
e
3D
tr
a
ns
form
by
gr
ou
p
i
ng
s
i
m
i
l
ar
2D
-
B
l
oc
k
s
i
n
the
i
ma
ge
i
nto
a
3D
-
arr
ay
s
(
group
ed
)
,
i
n
thi
s
c
as
e
the
i
ma
ge
i
s
r
ep
r
es
en
t
ed
i
n
th
e
tr
a
n
s
fo
r
m
do
ma
i
n
by
m
an
y
3D
-
groups
.
T
he
n
,
the
s
pe
c
tr
um
i
s
s
hrunk
en
to
s
ep
arate
t
he
n
oi
s
e
fr
o
m
oth
er
f
ea
tures
[
2
8].
T
h
i
s
me
tho
d
ha
s
s
ho
wn
great
ef
fi
c
i
en
c
y
i
n
r
em
ov
i
ng
i
m
ag
e
c
o
mp
are
d
to
s
ev
era
l
t
ec
hn
i
qu
es
,
b
ut
i
n
the
c
as
e
of
hi
g
h
l
ev
e
l
of
no
i
s
e
i
t
gi
v
es
l
es
s
pe
r
forma
nc
e,
an
d
for
t
ha
t,
i
n
[
29
]
a
bo
u
nd
e
d
B
M3
D
me
th
od
w
as
propos
e
d
ba
s
ed
on
th
e
ba
s
i
s
B
ME
D
to
ex
c
ee
d
thi
s
l
i
mi
ta
ti
o
n.
F
urth
ermor
e
,
ma
ny
tec
hn
i
qu
es
wer
e
d
eriv
ed
fr
om
th
e
B
M3
D
s
uc
h
as
i
n
[
29
]
,
wh
i
c
h
f
o
c
us
es
on
t
he
op
t
i
ma
l
c
h
oi
c
e
of
s
hrin
k
a
ge
op
erat
or,
an
d
t
he
t
wo
a
l
go
r
i
thm
s
i
n
[
30
],CD
-
B
M3
D
a
nd
i
t
erati
v
e
CD
-
B
M
3D
us
e
d
the
c
om
pl
ex
do
ma
i
n.
T
h
e
s
ev
era
l
f
un
d
am
en
tal
tec
h
ni
qu
es
of
i
m
ag
e
de
no
i
s
i
ng
i
n
tr
a
ns
for
m
d
om
a
i
n
are
pres
en
te
d b
y
t
he
b
el
ow d
i
a
gram
i
n Fi
gu
r
e
3.
F
i
g
ure
2
. Th
e a
b
ov
e
i
ma
ge
s
trans
forme
d
to
t
he
fr
eq
u
e
nc
y
do
ma
i
n
be
l
ow
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
MNIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
A
n o
v
erv
i
e
w o
f th
e f
u
nd
a
m
en
ta
l
a
pp
r
oa
c
h
es
tha
t y
i
e
l
d
s
ev
eral
i
ma
g
e..
. (
B
i
l
a
l
Ch
armout
i
)
2963
F
i
g
ure
3
. I
ma
g
e d
en
o
i
s
i
n
g
ap
pro
ac
he
s
3.3
. H
ybrid
F
ilt
er
ing
Hy
brid
f
i
l
t
erin
g
m
ay
be
c
on
s
i
de
r
ed
as
th
e
th
i
r
d
i
m
ag
e
-
de
no
i
s
i
n
g
c
at
eg
ory
.
It
i
s
a
di
v
ers
e
c
om
bi
na
t
i
on
of
s
ev
eral
a
p
proac
he
s
fr
om
the
s
pa
t
i
a
l
or
tr
a
ns
form
do
ma
i
n
or
bo
th
to
ge
th
er
as
s
ho
wn
i
n
F
i
gu
r
e
3
.
T
hi
s
k
i
nd
of
i
m
ag
e
fi
l
teri
ng
,
ex
pl
o
i
ts
the
ad
v
an
tag
es
o
f
d
i
f
ferent
ex
i
s
ti
n
g
fi
l
t
ers
to
b
ui
l
d
a
n
ew
f
i
l
t
er
whi
c
h
i
nc
l
ud
es
a
m
i
x
ture
o
f
the
s
e
ad
v
a
nta
g
es
,
i
n
ord
er
to
ov
erc
o
me
the
l
i
m
i
tat
i
o
ns
of
c
on
v
e
nti
o
na
l
t
ec
hn
i
qu
es
an
d
gi
v
e
be
tte
r
pe
r
for
ma
nc
e.
T
hi
s
ha
s
be
en
us
ed
by
ma
ny
r
es
ea
r
c
h
ers
an
d
s
tu
d
i
es
i
n
th
e
i
ma
g
e
d
e
no
i
s
i
n
g.
T
hu
s
,
nu
me
r
o
us
hy
brid
fi
l
t
ers
ha
v
e
be
en
propos
e
d,
w
he
r
eb
y
t
he
po
i
nt
wh
i
c
h
att
r
ac
ts
m
uc
h
at
t
en
ti
on
,
i
s
t
he
c
o
ns
i
de
r
ab
l
e
att
e
nd
a
nc
e
o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
MNIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
b
er 20
19
:
29
5
9
-
2967
2964
the
wav
e
l
et
i
n
t
he
s
e
hy
bri
d
f
i
l
t
ers
.
F
or
ex
a
mp
l
e,
i
n
[31
]
the
f
i
l
t
er
d
ec
om
po
s
es
the
i
ma
ge
by
(
DT
-
CW
T
)
the
n,
i
t
a
pp
l
i
es
an
ad
ap
t
i
v
e
wav
e
l
et
t
hres
ho
l
di
n
g
to
t
he
d
eta
i
l
e
d
c
oe
ffi
c
i
e
nts
,
wh
i
l
e
the
ap
prox
i
ma
ti
o
n
c
o
eff
i
c
i
e
nts
are
tr
ea
t
ed
by
N
on
-
Lo
c
al
M
ea
ns
(
NLM)
tec
hn
i
q
u
e,
a
nd
fi
na
l
l
y
the
de
no
i
s
ed
i
ma
g
e
i
s
ob
tai
n
ed
by
th
e
i
nv
ers
e
DT
-
CW
T
.
A
s
we
l
l
,
i
n
[
32
]
th
e
un
de
c
i
ma
t
ed
wav
el
e
t
tr
a
ns
form
(
UD
W
T
)
an
d
NL
M
i
s
pe
r
f
ormed
.
I
n
[3
3],
a
d
en
o
i
s
i
n
g
t
ec
hn
i
qu
e
ba
s
ed
o
n
the
c
o
mb
i
na
t
i
on
of
wav
e
l
et
tr
an
s
form
an
d
an
i
s
otrop
i
c
fi
l
teri
ng
w
as
pres
e
nte
d
.
M
an
y
oth
er
hy
br
i
d
fi
l
t
ers
us
e
the
m
i
x
ture
o
f
wav
el
e
t
an
d
t
ota
l
v
ari
ati
o
n
ha
s
be
en
i
ntrod
uc
ed
i
n
[
34
-
37
].
T
he
prin
c
i
pl
e c
o
mp
o
ne
n
t a
na
l
y
s
i
s
(
P
CA
)
was
me
r
g
ed
wi
th
th
e w
av
el
et
an
d
N
LM
i
n [
3
8]
an
d [
39
],
r
es
pe
c
ti
v
el
y
.
A
hy
bri
d
ne
i
g
hb
orh
oo
d
fi
l
ter
i
s
ap
p
l
i
ed
i
n
[4
0],
ba
s
e
d
on
the
G
au
s
s
i
an
fi
l
ter
wi
t
h
the
c
orr
el
at
i
o
n
of
wav
e
l
et
c
oe
ffi
c
i
e
nts
.
W
i
en
er
fi
l
t
er
i
ng
was
c
om
bi
n
ed
w
i
th
DT
-
CW
T
an
d
c
on
ti
n
uo
us
wav
el
et
tr
an
s
f
orm
(
CW
T
)
(
t
o
a
s
ei
s
m
i
c
s
i
gn
al
)
i
n
[
41
]
an
d
[4
2],
r
es
pe
c
ti
v
e
l
y
.
T
he
area
i
n
th
e
i
m
ag
e
ma
y
be
de
f
i
n
ed
as
a
tex
t
ure,
ed
ge
or
s
mo
ot
h
(
ho
m
og
en
e
o
us
)
r
eg
i
on
[
43
].
If
for
ea
c
h
of
t
he
s
e
areas
we
c
ho
s
e
the
f
i
l
t
er
wh
i
c
h
g
i
v
es
go
od
r
es
ul
t
a
mo
n
g
t
he
ex
i
s
ti
ng
f
i
l
ters
,
ev
en
tu
al
l
y
we
ge
t
three
di
v
ers
e
fi
l
t
ers
for
the
i
ma
ge
.
T
hi
s
i
s
the
c
o
nc
ep
t
w
hi
c
h
i
s
ad
op
ted
i
n
[4
3]
by
us
i
n
g
t
he
c
om
bi
n
ati
on
of
thre
e
d
i
ff
erent
m
ul
t
i
r
e
s
ol
ut
i
on
tec
h
ni
qu
es
,
na
me
l
y
wav
e
at
om
s
(
tex
tures
)
,
c
urv
el
et
(
ed
g
es
)
an
d
wav
e
l
et
(
s
mo
oth
r
eg
i
o
ns
)
.
Me
an
w
hi
l
e,
th
es
e
l
as
t
t
wo
tec
hn
i
qu
es
wer
e
bl
e
nd
ed
i
n
[
44
].
In
[4
5],
a
hy
brid
f
i
l
t
er
i
nc
l
u
di
ng
a
m
i
x
of
HMF
(
hy
brid
m
ed
i
an
f
i
l
t
er)
an
d
tot
a
l
v
aria
t
i
o
n (T
V
)
i
s
propo
s
ed
.
4.
E
va
luat
ion
and
Co
mp
ar
ison
T
he
c
om
m
on
ob
j
ec
t
i
v
e
of
t
he
s
e
i
n
tr
od
uc
e
d
de
no
i
s
i
ng
me
th
od
s
i
s
to
r
em
ov
e
no
i
s
e
fr
om
i
ma
ge
by
w
ay
of
pres
erv
i
ng
i
ts
fea
t
ures
,
or
i
n
ot
he
r
wor
ds
,
r
ea
c
hi
ng
th
e
r
es
tore
d
i
ma
ge
wh
i
c
h
i
s
the
mo
s
t
s
i
m
i
l
ar
to
the
or
i
gi
na
l
on
e,
tak
i
n
g
i
nt
o
ac
c
ou
n
t
c
ertai
n
s
tr
i
c
t
an
d
c
r
uc
i
a
l
c
r
i
teri
a
s
uc
h
as
the
s
i
mp
l
i
c
i
ty
of
f
i
l
t
er
i
mp
l
em
en
tat
i
on
a
nd
the
c
os
t
me
r
i
t
[4
6].
S
e
ek
i
n
g
t
o
ov
e
r
c
om
e
f
i
l
t
erin
g
l
i
m
i
ta
ti
o
ns
a
nd
to
ac
hi
ev
e
the
op
t
i
ma
l
i
ty
,
a
gre
at
n
u
mb
er
of
de
n
oi
s
i
ng
tec
hn
i
q
ue
s
h
av
e
b
ee
n
i
ntrod
uc
ed
i
n
t
he
s
e
l
as
t
y
ea
r
s
.
T
hu
s
,
ma
ny
d
i
s
pa
r
i
t
i
es
a
nd
d
i
ffe
r
e
nc
es
c
an
be
ob
s
e
r
v
ed
be
tw
ee
n
the
s
e
fi
l
ters
.
G
ett
i
ng
a
c
om
pa
r
i
s
o
n
an
d
ev
al
ua
t
i
on
f
r
om
thi
s
l
arge
n
um
b
er
prov
es
to
be
an
arduo
us
tas
k
,
es
pe
c
i
al
l
y
as
i
t
s
h
ares
ma
ny
c
ha
r
ac
teri
s
ti
c
s
s
uc
h
as
l
i
n
ea
r
/no
n
-
l
i
ne
ar,
s
tat
i
s
ti
c
a
l
/d
ete
r
m
i
n
i
s
ti
c
,
etc
.
(
ne
tw
ork
of
f
ea
t
ures
)
,
th
erefore
i
t
i
s
q
ui
t
e
n
ec
es
s
ary
t
o
r
el
y
on
s
om
e
po
i
nts
l
i
nk
ed
wi
t
h t
h
e t
hre
e
pa
r
ts
of
t
he
f
i
l
t
erin
g p
r
oc
es
s
(
T
ab
l
e
1) wh
i
c
h a
r
e:
a.
T
he
no
i
s
y
i
ma
ge
(
i
n
pu
t)
:
t
he
ty
pe
of
no
i
s
e
(
ad
d
i
ti
v
e
,
mu
l
t
i
p
l
i
c
at
i
v
e,
i
mp
ul
s
e
n
oi
s
e,
mi
x
ed
no
i
s
e),
a
mo
u
nt
of
n
oi
s
e
(
hi
g
h,
l
o
w)
,
s
tr
uc
ture
of
i
ma
ge
(
t
ex
ture,
s
mo
oth
,
e
dg
e),
p
i
x
el
s
i
nte
ns
i
ty
.
b.
F
i
l
ter
(
d
en
o
i
s
i
n
g
to
ol
)
:
c
om
pu
tat
i
o
na
l
c
os
t
(
ac
c
ep
tab
l
e
,
hi
g
h),
fi
l
ter
i
mp
l
e
me
nt
ati
o
n
(
s
i
mp
l
e,
c
om
pl
ex
)
.
c.
Res
tored
i
ma
g
e
(
o
utp
ut):
t
he
i
m
ag
e
qu
al
i
ty
i
s
o
ne
o
f
th
e
ev
a
l
u
ati
on
c
r
i
teri
a
of
de
no
i
s
i
ng
tec
hn
i
qu
es
p
erfor
m
an
c
e,
s
o
the
qu
es
t
i
on
i
s
ho
w
do
we
as
s
es
s
thi
s
qu
al
i
ty
?
T
he
way
by
whi
c
h t
he
i
m
ag
e
qu
a
l
i
ty
i
s
e
v
al
ua
t
ed
(
ev
a
l
u
ati
on
c
r
i
t
eria
)
ma
y
be
di
v
i
de
d
i
n
to
tw
o w
ay
s
:
-
T
he
fi
r
s
t
on
e
i
s
th
e
v
i
s
ua
l
ev
al
ua
ti
o
n
de
ter
m
i
ne
d
by
the
ob
s
erv
er,
wh
ere
th
e
hu
ma
n
j
ud
gm
e
nt
i
s
i
nte
r
es
t
ed
i
n
the
i
ma
g
e
c
o
mp
o
ne
n
ts
ap
pe
ara
nc
e,
i
f
i
t
c
on
ta
i
ns
an
y
de
gra
da
t
i
on
fac
tors
or ot
he
r
wi
s
e,
s
uc
h a
s
art
i
fac
ts
,
di
s
c
on
ti
n
ui
t
i
es
a
nd
bl
ur [47]
.
-
T
he
s
ec
on
d
o
ne
i
s
th
e
q
ua
nti
t
ati
v
e
ev
al
ua
t
i
on
(
qu
al
i
ty
me
tr
i
c
)
by
us
i
n
g
th
e
m
ea
s
u
r
em
en
t
pa
r
am
ete
r
s
,
w
hi
c
h
i
nc
l
ud
e
:
1)
S
i
gn
a
l
t
o
N
oi
s
e
Ra
ti
o
(
S
NR)
,
whi
c
h
m
ea
s
ures
the
am
ou
nt
of
no
i
s
e
n
i
n
the
no
i
s
y
i
ma
ge
I
(
i
,
j
)
us
i
ng
th
e
s
tan
da
r
d
d
e
v
i
ati
on
of
t
he
n
oi
s
e
σ
(
n
)
an
d
i
ma
ge
σ
(
I
)
(
σ
(
I
)
=
60
i
n
di
c
at
i
n
g
g
oo
d
i
ma
ge
q
ua
l
i
ty
)
[4
8],
i
t
i
s
g
i
v
en
by
(
3)
;
2)
M
ea
n
S
qu
ared
E
r
r
or
(
MS
E
)
,
tha
t
me
as
ures
th
e
di
s
s
i
mi
l
ari
ty
be
twe
en
th
e
r
es
tored
i
m
a
ge
I
̂
(
i
,
j
)
an
d
th
e
orig
i
n
al
o
ne
I
(
i
,
j
)
as
s
ho
wn
b
el
ow
i
n
(
4),
th
us
when
ev
er
th
e
M
S
E
i
s
l
ower,
the
i
ma
ge
de
no
i
s
i
n
g
ac
hi
ev
es
mo
r
e
s
uc
c
es
s
[49
];
3)
P
ea
k
S
i
g
na
l
t
o
Noi
s
e
R
ati
o
(
P
S
NR)
[50
],
a we
l
l
-
k
no
w
n p
arame
ter th
at
ha
s
an
i
nv
ers
e rel
ati
on
s
h
i
p w
i
th
MS
E
,
as
de
no
te
d i
n
(
5).
It
i
s
no
t
ne
c
es
s
ary
t
o
b
e
an
en
ta
i
l
me
nt
r
e
l
at
i
on
b
et
ween
th
e
v
i
s
u
al
an
d
q
ua
n
ti
tat
i
v
e
as
s
es
s
me
nt,
b
ec
au
s
e
s
o
m
eti
m
es
an
i
ma
g
e,
ev
e
n
w
i
t
h
hi
gh
P
S
NR
or
l
o
w
M
S
E
,
do
es
n
ot
s
ee
m
c
l
ea
n.
T
hi
s
me
a
ns
th
at
th
e
s
e
i
nd
ex
es
(
M
S
E
,
P
S
NR)
are
no
t
we
l
l
ma
tc
he
d
to
pe
r
c
ei
v
ed
v
i
s
ua
l
qu
a
l
i
ty
[5
1].
T
hu
s
,
t
he
r
e
a
r
e
ma
ny
qu
al
i
ty
as
s
es
s
me
nt
(
Q
A
)
me
th
od
s
de
v
e
l
op
ed
i
n
t
he
l
as
t
de
c
ad
e
s
,
wh
i
c
h
tak
e
i
nto
ac
c
ou
nt
t
he
hu
ma
n
v
i
s
ua
l
s
y
s
tem
(
HV
S
)
,
s
uc
h
as
t
he
S
tr
uc
tura
l
S
IM
i
l
arit
y
(
S
S
IM)
i
nd
ex
[5
1
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
MNIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
A
n o
v
erv
i
e
w o
f th
e f
u
nd
a
m
en
ta
l
a
pp
r
oa
c
h
es
tha
t y
i
e
l
d
s
ev
eral
i
ma
g
e..
. (
B
i
l
a
l
Ch
armout
i
)
2965
SN
R
=
σ
(
I
)
σ
(
n
)
(
3)
M
SE
=
1
N
∑
(
I
[
i
,
j
]
−
I
̂
[
i
,
j
]
)
2
i
,
j
(
4)
PS
N
R
=
10
l
og
10
(
255
2
MS
E
)
(
5)
the
nu
m
be
r
of
pi
x
e
l
s
i
n i
ma
ge
.
B
y
us
i
ng
th
es
e
ab
ov
e
t
hree
gu
i
d
an
c
e
po
i
nts
,
th
e
s
ev
eral
tec
hn
i
q
ue
s
of
i
m
ag
e
de
n
oi
s
i
n
g
c
an
be
ev
a
l
u
ate
d
an
d
c
o
m
pa
r
ed
br
i
ef
l
y
ac
c
ordi
n
g
to
t
he
i
r
c
l
as
s
i
f
i
c
at
i
on
i
n
T
a
bl
e
1
(
fi
e
l
d
''
ty
pe
s
of
fi
l
t
ers
''
)
,
as
fol
l
o
ws
:
T
he
pe
r
forman
c
e
of
f
i
l
teri
ng
me
th
od
r
i
s
es
fr
o
m
t
he
l
ef
t
c
ol
u
mn
t
o
th
e
r
i
gh
t
c
ol
um
n
an
d
fr
o
m
the
f
i
r
s
t
r
ow
to
the
l
as
t
r
ow,
wher
e
the
pe
r
f
ormanc
e
o
f
de
n
oi
s
i
ng
me
t
ho
d
i
s
de
no
ted
by
th
e
bl
ue
tr
i
an
g
l
e,
wh
ere
i
t
i
s
r
i
s
es
fr
om
t
he
he
ad
to
t
he
ba
s
e
of
th
i
s
tr
i
an
gl
e.
F
or
ex
am
p
l
e,
t
he
ad
a
pti
v
e
de
n
oi
s
i
n
g
m
eth
o
ds
i
n
th
e
tr
an
s
form
d
om
a
i
n
are
m
ore
e
ffi
c
i
en
t
c
om
p
ared
to
the
n
on
-
a
da
pt
i
v
e
m
eth
o
ds
i
n
the
tr
an
s
form
do
m
ai
n
or
wi
th
ad
a
pti
v
e
me
th
od
s
i
n
the
s
pa
ti
a
l
do
ma
i
n.
I
n
ad
di
t
i
o
n,
the
l
i
ne
ar
fi
l
ters
are
s
i
m
pl
e
to
i
mp
l
em
e
nt
th
an
no
n
-
l
i
ne
ar
on
es
,
wh
ereas
the
s
tat
i
s
ti
c
a
l
f
i
l
t
ers
are ex
p
en
s
i
v
e a
nd
c
o
mp
l
ex
c
om
pa
r
ed
to
t
he
de
ter
mi
ni
s
ti
c
f
i
l
te
r
s
.
5.
Co
n
clus
ion
T
hi
s
p
ap
er
pres
e
nts
a
n
ov
e
r
v
i
ew
o
f
t
he
fun
da
m
en
t
al
a
pp
r
oa
c
h
es
fr
om
w
hi
c
h
the
s
ev
eral
de
no
i
s
i
n
g
a
l
g
orit
h
ms
are
d
eriv
ed
,
as
a
s
tart
i
ng
po
i
nt
of
fut
ure
s
tu
di
es
an
d
r
es
e
arc
he
s
i
n
thi
s
s
ub
j
ec
t.
T
hrough
t
hi
s
s
tu
dy
,
i
t
i
s
fou
n
d
tha
t
the
r
e
i
s
an
en
orm
ou
s
nu
mb
er
of
de
no
i
s
i
n
g
al
g
orit
h
ms
whi
c
h
ma
k
es
t
he
ev
a
l
ua
t
i
o
n
an
d
the
c
o
mp
aris
o
n
be
twee
n
th
em
di
ff
i
c
ul
t
t
o
be
att
a
i
na
bl
e
.
In
order
to
ov
e
r
c
om
e
thi
s
s
i
tu
ati
on
by
fac
i
l
i
t
ati
ng
th
e
un
d
ers
tan
d
i
n
g,
the
ev
al
ua
t
i
o
n
an
d
the
c
om
pa
r
i
s
o
n
b
etwe
e
n
t
he
s
i
g
ni
f
i
c
an
t
nu
mb
er
of
di
v
ers
e
tec
hn
i
qu
es
of
i
ma
ge
de
n
oi
s
i
ng
,
t
hi
s
pa
p
er
prop
os
es
s
e
v
en
c
ate
g
orie
s
t
o
c
l
as
s
i
fy
the
s
e
al
g
orit
h
ms
(
Li
n
ea
r
/No
nl
i
n
ea
r
-
Dete
r
m
i
n
i
s
ti
c
/
S
tat
i
s
ti
c
a
l
-
Non
-
A
d
ap
t
i
v
e/A
da
pti
v
e
-
L
o
c
al
/No
n
-
Lo
c
a
l
-
w
i
th
ou
t
N
oi
s
e
Dete
c
t
i
on
(
t
r
ea
t
a
l
l
pi
x
e
l
s
)
/
w
i
th
No
i
s
e
Dete
c
ti
on
(
tr
e
at
c
or
r
up
t
ed
pi
x
el
s
)
-
S
pa
t
i
al
Doma
i
n/Tr
a
ns
form
Do
ma
i
n
-
Unmi
x
e
d/Hy
bri
d),
where
t
he
c
ha
r
ac
teri
s
t
i
c
s
of
an
y
c
ate
g
ory
aff
ec
t
the
eff
i
c
i
en
c
y
of
th
e
d
en
o
i
s
i
ng
fi
l
ter.
T
o
ac
hi
ev
e
th
i
s
g
oa
l
,
we
a
l
s
o
pres
e
nt
t
he
e
v
al
ua
t
i
o
n
to
ol
s
whi
c
h
pe
r
m
i
t t
he
es
ti
m
ati
on
of
t
he
eff
i
c
i
e
nc
y
a
nd
t
he
di
s
pa
r
i
ty
a
mo
n
g t
h
es
e t
ec
hn
i
qu
es
.
A
l
l
t
he
s
e
c
ate
go
r
i
es
an
d e
v
a
l
ua
ti
o
n t
oo
l
s
are
s
um
ma
r
i
s
ed
i
n
T
a
bl
e
1.
T
ab
l
e
1
.
T
he
C
om
p
aris
on
b
etwe
e
n S
ev
eral
Fun
da
m
en
t
al
D
en
o
i
s
i
n
g M
eth
o
ds
E
v
a
lua
t
ion
Too
l
s
Fil
t
e
r
ing
Ty
p
e
s
The
n
o
is
y
im
a
g
e
(
inp
u
t
)
Fil
t
e
r
(
d
e
n
o
is
ing
t
o
o
l)
R
e
s
t
o
r
e
d
i
mag
e
(
o
u
t
p
u
t
)
-
Ty
p
e
o
f
n
o
i
s
e
(
a
d
d
it
iv
e
,
mult
ipl
ic
a
t
iv
e
,
im
p
u
ls
e
n
o
is
e
,
mi
x
e
d
n
o
is
e
)
-
A
mou
n
t
o
f
n
o
is
e
(
h
igh
,
low
)
-
S
t
r
u
c
t
u
r
e
o
f
i
mag
e
(
t
e
x
t
u
r
e
,
s
moo
t
h
,
e
d
g
e
)
,
p
ix
e
l
s
int
e
n
s
i
t
y
-
C
o
mpu
t
a
t
ion
a
l
c
o
s
t
(
a
c
c
e
p
t
a
b
le
,
h
igh
)
-
Fil
t
e
r
im
p
lemen
t
a
t
ion
(
s
im
p
le
,
c
o
mple
x
)
-
Mea
s
u
r
e
men
t
o
f
n
o
is
e
(
S
N
R
,
MS
E
,
P
S
N
R
,
e
t
c
.
)
-
V
is
u
a
l
q
u
a
li
t
y
(
b
lur
,
a
r
t
if
a
c
t
s
,
inf
o
r
mat
ion
los
s
,
e
t
c
.
)
L
ine
a
r
N
o
n
-
L
ine
a
r
D
e
t
e
r
mi
n
is
t
i
c
S
t
a
t
is
t
i
c
a
l
N
o
n
-
A
d
a
p
t
iv
e
A
d
a
p
t
i
v
e
L
o
c
a
l
N
o
n
-
L
o
c
a
l
W
it
h
o
u
t
N
o
i
s
e
D
e
t
e
c
t
ion
(
t
r
e
a
t
a
ll
p
ix
e
ls
)
W
it
h
N
o
i
s
e
D
e
t
e
c
t
ion
(
t
r
e
a
t
c
o
r
r
u
p
t
e
d
p
ix
e
l
s
)
S
p
a
t
ial
D
o
main
Tr
a
n
s
f
o
r
m
D
o
main
U
n
mi
x
e
d
H
y
b
r
id
ACKN
O
W
L
E
DG
E
ME
NT
T
hi
s
wor
k
was
fi
na
nc
i
al
l
y
s
up
po
r
te
d
by
M
i
n
i
s
tr
y
of
E
du
c
at
i
o
n
Ma
l
ay
s
i
a
(
MO
E
)
un
de
r
F
un
da
me
n
tal
Res
e
arc
h Gr
an
t
S
c
he
m
e (F
RG
S
)
(
(
Ref:
F
RG
S
/1/
20
1
9/S
T
G
0
6/UNI
MA
P
/02
/3).
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
MNIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
b
er 20
19
:
29
5
9
-
2967
2966
Ref
er
en
ce
s
[1
]
Zh
u
L
.
A
p
p
l
i
c
a
ti
o
n
o
f
St
a
c
k
Fi
l
te
r
o
n
Im
a
g
e
Pr
o
c
e
s
s
i
n
g
.
Ap
p
l
i
e
d
M
e
c
h
a
n
i
c
s
a
n
d
M
a
te
ri
a
l
s
.
2
0
1
4
;
5
4
3
:
2
1
6
3
-
2166
.
[2
]
Ya
n
M
.
Res
t
o
ra
t
i
o
n
o
f
i
m
a
g
e
s
c
o
rru
p
t
e
d
b
y
i
m
p
u
l
s
e
n
o
i
s
e
a
n
d
m
i
x
e
d
G
a
u
s
s
i
a
n
i
m
p
u
l
s
e
n
o
i
s
e
u
s
i
n
g
b
l
i
n
d
i
n
p
a
i
n
ti
n
g
.
SIAM
J
o
u
rn
a
l
o
n
I
m
a
g
i
n
g
S
c
i
e
n
c
e
s
.
2
0
1
3
;
6
(
3
):
1
2
2
7
-
1
2
4
5
.
[3
]
An
d
e
r
M
.
In
tr
o
d
u
c
ti
o
n
a
u
x
t
e
c
h
n
i
q
u
e
s
d
e
t
ra
i
te
m
e
n
t
d
’
i
m
a
g
e
s
.
Ey
ro
l
l
e
s
.
1
9
8
7
.
[4
]
Liu
J
,
T
a
i
XC
,
Hua
n
g
H,
Hua
n
Z.
A
wei
g
h
te
d
d
i
c
ti
o
n
a
r
y
-
l
e
a
r
n
i
n
g
m
o
d
e
l
fo
r
d
e
n
o
i
s
i
n
g
i
m
a
g
e
s
c
o
rru
p
te
d
b
y
m
i
x
e
d
n
o
i
s
e
.
IEE
E T
ra
n
s
a
c
t
i
o
n
s
o
n
I
m
a
g
e
Pr
o
c
e
s
s
i
n
g
.
2
0
1
3
;
2
2
(3
):
1
1
0
8
-
1
1
2
0
.
[5
]
Bo
y
a
t
AK,
J
o
s
h
i
BK.
A
r
e
v
i
e
w
p
a
p
e
r
:
n
o
i
s
e
m
o
d
e
l
s
i
n
d
i
g
i
ta
l
i
m
a
g
e
p
r
o
c
e
s
s
i
n
g
.
ArXi
v
p
re
p
ri
n
t
.
2
0
1
5
:
1
5
0
5
.
0
3
4
8
9
.
[6
]
Sc
h
u
l
te
S
,
De
Wi
tt
e
V,
K
e
rre
EE.
A
f
u
z
z
y
n
o
i
s
e
re
d
u
c
ti
o
n
m
e
th
o
d
fo
r
c
o
l
o
r
i
m
a
g
e
s
.
IEEE
Tra
n
s
a
c
t
i
o
n
s
o
n
I
m
a
g
e
Pro
c
e
s
s
i
n
g
.
2
0
0
7
;
1
6
(5
):
1
4
2
5
-
1
4
3
6
.
[7
]
Sa
e
e
d
i
J
,
M
o
ra
d
i
M
H,
Fa
e
z
K.
A
n
e
w
wav
e
l
e
t
-
b
a
s
e
d
f
u
z
z
y
s
i
n
g
l
e
a
n
d
m
u
l
ti
-
c
h
a
n
n
e
l
i
m
a
g
e
d
e
n
o
i
s
i
n
g
.
Im
a
g
e
a
n
d
Vi
s
i
o
n
Co
m
p
u
ti
n
g
.
2
0
1
0
;
2
8
(
1
2
):
1
6
1
1
-
1
6
2
3
.
[8
]
L
i
u
L
,
Che
n
L
,
Che
n
CP,
T
a
n
g
YY,
Pu
n
CM
.
We
i
g
h
te
d
j
o
i
n
t
s
p
a
rs
e
r
e
p
re
s
e
n
t
a
ti
o
n
f
o
r re
m
o
v
i
n
g
m
i
x
e
d
n
o
i
s
e
i
n
i
m
a
g
e
.
IEEE
tra
n
s
a
c
ti
o
n
s
o
n
c
y
b
e
rn
e
t
i
c
s
.
2
0
1
6
;
4
7
(
3
):
6
0
0
-
6
11
.
[9
]
M
o
twa
n
i
M
C,
G
a
d
i
y
a
M
C,
M
o
twa
n
i
RC
,
Harri
s
FC.
S
u
rv
e
y
o
f
i
m
a
g
e
d
e
n
o
i
s
i
n
g
t
e
c
h
n
i
q
u
e
s
.
Pro
c
e
e
d
i
n
g
s
o
f
G
SPX.
2
0
0
4
:
27
-
30
.
[1
0
]
L
i
X.
I
m
a
g
e
re
s
to
ra
t
i
o
n
:
Fu
n
d
a
m
e
n
t
a
l
s
a
n
d
a
d
v
a
n
c
e
s
:
CR
C
Pre
s
s
.
2
0
1
2
.
[1
1
]
We
i
s
s
B
.
F
a
s
t
m
e
d
i
a
n
a
n
d
b
i
l
a
te
r
a
l
fi
l
te
ri
n
g
.
Ac
m
Tr
a
n
s
a
c
ti
o
n
s
o
n
G
ra
p
h
i
c
s
(
TOG
)
.
2
0
0
6
;
2
5
(3
)
:
519
-
5
2
6
.
[1
2
]
Che
n
CLP,
L
i
u
L
,
C
h
e
n
L
,
Ta
n
g
YY,
Zh
o
u
Y.
We
i
g
h
te
d
c
o
u
p
l
e
s
p
a
rs
e
re
p
r
e
s
e
n
ta
t
i
o
n
wit
h
c
l
a
s
s
i
fi
e
d
re
g
u
l
a
ri
z
a
ti
o
n
fo
r
i
m
p
u
l
s
e
n
o
i
s
e
re
m
o
v
a
l
.
IEEE
Tr
a
n
s
a
c
t
i
o
n
s
o
n
Im
a
g
e
Pro
c
e
s
s
i
n
g
.
2
0
1
5
;
2
4
(1
1
):
4014
-
4
0
2
6
.
[1
3
]
Zh
o
u
Y
,
L
i
n
M
,
Xu
S,
Za
n
g
H
,
He
H
,
L
i
Q
,
G
u
o
J
.
An
i
m
a
g
e
d
e
n
o
i
s
i
n
g
a
l
g
o
ri
t
h
m
fo
r
m
i
x
e
d
n
o
i
s
e
c
o
m
b
i
n
i
n
g
n
o
n
l
o
c
a
l
m
e
a
n
s
fi
l
te
r
a
n
d
s
p
a
rs
e
re
p
re
s
e
n
ta
ti
o
n
te
c
h
n
i
q
u
e
.
J
o
u
rn
a
l
o
f
Vi
s
u
a
l
Com
m
u
n
i
c
a
ti
o
n
a
n
d
I
m
a
g
e
Re
p
re
s
e
n
ta
t
i
o
n
.
2
0
1
6
;
41
:
74
-
8
6
.
[1
4
]
L
a
n
R,
Zh
o
u
Y,
Ta
n
g
YY,
Che
n
CP.
Im
a
g
e
d
e
n
o
i
s
i
n
g
u
s
i
n
g
n
o
n
-
l
o
c
a
l
fu
z
z
y
m
e
a
n
s
.
IE
EE
Chi
n
a
Su
m
m
i
t
a
n
d
In
te
rn
a
t
i
o
n
a
l
Con
fe
re
n
c
e
o
n
Si
g
n
a
l
a
n
d
In
fo
r
m
a
t
i
o
n
Pro
c
e
s
s
i
n
g
(
Chi
n
a
SIP)
.
2015
:
1
9
6
-
2
0
0
.
[1
5
]
Wu
Y
,
Tra
c
e
y
B
,
N
a
ta
ra
j
a
n
P
,
Noo
n
a
n
J
P.
Pr
o
b
a
b
i
l
i
s
t
i
c
n
o
n
-
l
o
c
a
l
m
e
a
n
s
.
IEEE
Si
g
n
a
l
P
ro
c
e
s
s
i
n
g
L
e
tt
e
r
s
.
2
0
1
3
;
2
0
(8
)
:
7
6
3
-
7
6
6
.
[1
6
]
M
u
re
s
a
n
DD
,
Pa
r
k
s
TW
.
Ad
a
p
ti
v
e
p
ri
n
c
i
p
a
l
c
o
m
p
o
n
e
n
ts
a
n
d
i
m
a
g
e
d
e
n
o
i
s
i
n
g
.
I
m
a
g
e
Pro
c
e
s
s
i
n
g
,
ICIP
.
2
0
0
3
:
101
-
104
.
[1
7
]
Hy
v
a
ri
n
e
n
A,
O
j
a
E,
Hoy
e
r
P,
Hurri
J
.
Im
a
g
e
fe
a
t
u
re
e
x
tra
c
ti
o
n
b
y
s
p
a
r
s
e
c
o
d
i
n
g
a
n
d
i
n
d
e
p
e
n
d
e
n
t
c
o
m
p
o
n
e
n
t
a
n
a
l
y
s
i
s
.
P
a
tt
e
rn
R
e
c
o
g
n
i
ti
o
n
,
Pro
c
e
e
d
i
n
g
s
o
f
Fo
u
rte
e
n
t
h
I
n
te
r
n
a
ti
o
n
a
l
Co
n
fe
re
n
c
e
.
1
9
9
8
;
2
:
1
2
6
8
-
1
2
7
3
.
[1
8
]
Va
s
e
g
h
i
S
,
J
e
te
l
o
v
á
H.
Pri
n
c
i
p
a
l
a
n
d
i
n
d
e
p
e
n
d
e
n
t
c
o
m
p
o
n
e
n
t
a
n
a
l
y
s
i
s
i
n
i
m
a
g
e
p
ro
c
e
s
s
i
n
g
.
Pro
c
e
e
d
i
n
g
s
o
f
th
e
1
4
th
AC
M
I
n
te
rn
a
ti
o
n
a
l
Con
f
e
re
n
c
e
o
n
M
o
b
i
l
e
Com
p
u
ti
n
g
a
n
d
Net
work
i
n
g
(M
O
BICO
M
’
0
6
).
2
0
0
6
: 1
-
5
.
[1
9
]
Q
u
a
n
J
.
I
m
a
g
e
De
n
o
i
s
i
n
g
o
f
G
a
u
s
s
i
a
n
a
n
d
Po
i
s
s
o
n
Noi
s
e
Ba
s
e
d
o
n
Wa
v
e
l
e
t
Th
re
s
h
o
l
d
i
n
g
.
Do
c
to
r
a
l
d
i
s
s
e
rta
ti
o
n
.
O
h
i
o
:
Uni
v
e
rs
i
ty
o
f
Cin
c
i
n
n
a
ti
.
2
0
1
3
.
[2
0
]
M
a
l
l
a
t
S.
A
wa
v
e
l
e
t
to
u
r
o
f
s
i
g
n
a
l
p
r
o
c
e
s
s
i
n
g
:
th
e
s
p
a
rs
e
wa
y
:
Ac
a
d
e
m
i
c
p
re
s
s
.
2
0
0
8
.
[2
1
]
L
e
Pe
n
n
e
c
E
,
M
a
l
l
a
t
S.
Sp
a
r
s
e
g
e
o
m
e
tri
c
i
m
a
g
e
re
p
re
s
e
n
t
a
ti
o
n
s
wit
h
b
a
n
d
e
l
e
t
s
.
IEEE
Tr
a
n
s
a
c
ti
o
n
s
o
n
I
m
a
g
e
Pr
o
c
e
s
s
i
n
g
.
2
0
0
5
;
1
4
(4
):
4
2
3
-
4
3
8
.
[2
2
]
Hy
v
a
ri
n
e
n
A,
O
j
a
E,
Hoy
e
r
P,
Hurri
J
.
Im
a
g
e
fe
a
t
u
re
e
x
tra
c
ti
o
n
b
y
s
p
a
r
s
e
c
o
d
i
n
g
a
n
d
i
n
d
e
p
e
n
d
e
n
t
c
o
m
p
o
n
e
n
t
a
n
a
l
y
s
i
s
.
Pr
o
c
e
e
d
i
n
g
s
.
Fo
u
rte
e
n
th
I
n
te
rn
a
ti
o
n
a
l
Con
f
e
re
n
c
e
o
n
Pa
tt
e
rn
R
e
c
o
g
n
i
ti
o
n
.
1998
;
2
:
1
2
6
8
-
1273
.
[2
3
]
Va
s
e
g
h
i
S
,
J
e
te
l
o
v
á
H.
Pri
n
c
i
p
a
l
a
n
d
i
n
d
e
p
e
n
d
e
n
t
c
o
m
p
o
n
e
n
t
a
n
a
l
y
s
i
s
i
n
i
m
a
g
e
p
ro
c
e
s
s
i
n
g
.
Pro
c
e
e
d
i
n
g
s
o
f
th
e
1
4
th
AC
M
I
n
te
rn
a
ti
o
n
a
l
Con
f
e
re
n
c
e
o
n
M
o
b
i
l
e
Com
p
u
ti
n
g
a
n
d
Net
work
i
n
g
(M
O
BICO
M
’
0
6
).
2
0
0
6
:
1
-
5
.
[2
4
]
Don
o
h
o
,
D.
L
.
We
d
g
e
l
e
ts
:
Ne
a
rl
y
m
i
n
i
m
a
x
e
s
t
i
m
a
ti
o
n
o
f
e
d
g
e
s
.
T
h
e
A
n
n
a
l
s
o
f
Sta
ti
s
ti
c
s
.
1
9
9
9
;
2
7
(3
)
:
859
-
8
9
7
.
[2
5
]
Sh
a
o
L
,
Ya
n
R,
L
i
X,
L
i
u
Y.
Fro
m
h
e
u
r
i
s
t
i
c
o
p
ti
m
i
z
a
t
i
o
n
to
d
i
c
t
i
o
n
a
ry
l
e
a
rn
i
n
g
:
A
r
e
v
i
e
w
a
n
d
c
o
m
p
re
h
e
n
s
i
v
e
c
o
m
p
a
r
i
s
o
n
o
f
i
m
a
g
e
d
e
n
o
i
s
i
n
g
a
l
g
o
r
i
th
m
s
.
I
EEE
tra
n
s
a
c
t
i
o
n
s
o
n
c
y
b
e
r
n
e
ti
c
s
.
2
0
1
4
;
4
4
(7
):
1001
-
1
0
1
3
.
[2
6
]
Dab
o
v
K,
Fo
i
A,
Ka
t
k
o
v
n
i
k
V,
Eg
i
a
z
a
ri
a
n
K.
Im
a
g
e
re
s
t
o
ra
ti
o
n
b
y
s
p
a
rs
e
3
D
tr
a
n
s
fo
rm
-
d
o
m
a
i
n
c
o
l
l
a
b
o
r
a
ti
v
e
f
i
l
te
r
i
n
g
.
El
e
c
tro
n
i
c
I
m
a
g
i
n
g
.
2
0
0
8
;
6812
:
6
8
1
2
0
7
.
[2
7
]
L
v
X
G
,
So
n
g
YZ,
L
i
F.
A
n
e
f
fi
c
i
e
n
t
n
o
n
c
o
n
v
e
x
re
g
u
l
a
ri
z
a
t
i
o
n
fo
r
wa
v
e
l
e
t
fr
a
m
e
a
n
d
to
t
a
l
v
a
ri
a
t
i
o
n
b
a
s
e
d
i
m
a
g
e
re
s
to
ra
t
i
o
n
.
J
o
u
r
n
a
l
o
f
Com
p
u
ta
t
i
o
n
a
l
a
n
d
Ap
p
l
i
e
d
M
a
th
e
m
a
t
i
c
s
.
2
0
1
5
;
290
:
5
5
3
-
5
6
6
.
[2
8
]
Zh
a
n
g
L
,
Do
n
g
W
,
Zh
a
n
g
D
,
Sh
i
G
.
Two
-
s
ta
g
e
i
m
a
g
e
d
e
n
o
i
s
i
n
g
b
y
p
ri
n
c
i
p
a
l
c
o
m
p
o
n
e
n
t
a
n
a
l
y
s
i
s
wit
h
l
o
c
a
l
p
i
x
e
l
g
ro
u
p
i
n
g
.
Pa
tt
e
rn
Re
c
o
g
n
i
ti
o
n
.
2
0
1
0
;
4
3
(
4
):
1
5
3
1
-
1
5
4
9
.
[2
9
]
Che
n
Q
,
Wu
D.
I
m
a
g
e
d
e
n
o
i
s
i
n
g
b
y
b
o
u
n
d
e
d
b
l
o
c
k
m
a
tc
h
i
n
g
a
n
d
3
D
f
i
l
t
e
ri
n
g
.
Si
g
n
a
l
P
ro
c
e
s
s
i
n
g
.
2
0
1
0
;
9
0
(9
)
:
2
7
7
8
-
2
7
8
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
MNIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
A
n o
v
erv
i
e
w o
f th
e f
u
nd
a
m
en
ta
l
a
pp
r
oa
c
h
es
tha
t y
i
e
l
d
s
ev
eral
i
ma
g
e..
. (
B
i
l
a
l
Ch
armout
i
)
2967
[3
0
]
Ka
tk
o
v
n
i
k
V
,
Eg
i
a
z
a
ri
a
n
K.
Sp
a
rs
e
p
h
a
s
e
i
m
a
g
i
n
g
b
a
s
e
d
o
n
c
o
m
p
l
e
x
d
o
m
a
i
n
n
o
n
l
o
c
a
l
B
M
3
D
te
c
h
n
i
q
u
e
s
.
Dig
i
t
a
l
S
i
g
n
a
l
Pro
c
e
s
s
i
n
g
.
2
0
1
7
;
6
3
:
72
-
85
.
[3
1
]
J
a
i
n
A.
M
e
d
i
c
a
l
Im
a
g
e
Noi
s
e
Su
p
p
re
s
s
i
o
n
U
s
i
n
g
Dua
l
T
re
e
Co
m
p
l
e
x
Wa
v
e
l
e
t
Tr
a
n
s
f
o
rm
s
a
n
d
Non
-
L
o
c
a
l
M
e
a
n
s
.
Im
p
e
ri
a
l
J
o
u
rn
a
l
o
f
In
te
rd
i
s
c
i
p
l
i
n
a
r
y
Re
s
e
a
rc
h
.
2
0
1
6
;
2
(1
2
)
:
1
8
8
9
-
1892
.
[3
2
]
Ya
n
g
F,
Yu
Z,
L
i
C.
An
a
d
a
p
ti
v
e
SAR
i
m
a
g
e
s
p
e
c
k
l
e
re
d
u
c
ti
o
n
a
l
g
o
ri
th
m
b
a
s
e
d
o
n
u
n
d
e
c
i
m
a
t
e
d
wav
e
l
e
t
tra
n
s
f
o
rm
a
n
d
n
o
n
-
l
o
c
a
l
m
e
a
n
s
.
G
e
o
s
c
i
e
n
c
e
a
n
d
Rem
o
te
Se
n
s
i
n
g
Sy
m
p
o
s
i
u
m
(IGAR
SS).
2016
:
1
0
3
0
-
1033
.
[3
3
]
Ko
u
r
S,
K
a
u
r
B.
Hy
b
r
i
d
fi
l
t
e
r
wit
h
wa
v
e
l
e
t
d
e
n
o
i
s
i
n
g
a
n
d
a
n
i
s
o
t
ro
p
i
c
d
i
ff
u
s
i
o
n
f
i
l
t
e
r
fo
r
i
m
a
g
e
d
e
s
p
e
c
k
l
i
n
g
.
Com
p
u
te
r Grap
h
i
c
s
,
V
i
s
i
o
n
a
n
d
I
n
fo
rm
a
ti
o
n
Se
c
u
ri
ty
(CG
VIS).
2
0
1
5
:
17
-
21
.
[3
4
]
L
v
XG
,
So
n
g
YZ
,
L
i
F.
A
n
e
f
fi
c
i
e
n
t
n
o
n
c
o
n
v
e
x
re
g
u
l
a
ri
z
a
t
i
o
n
fo
r
wa
v
e
l
e
t
fr
a
m
e
a
n
d
to
t
a
l
v
a
ri
a
t
i
o
n
b
a
s
e
d
i
m
a
g
e
re
s
to
ra
t
i
o
n
.
J
o
u
r
n
a
l
o
f
Com
p
u
ta
t
i
o
n
a
l
a
n
d
Ap
p
l
i
e
d
M
a
th
e
m
a
t
i
c
s
.
2
0
1
5
;
2
9
0
,
5
5
3
-
5
6
6
.
[3
5
]
El
y
a
s
i
I,
Po
u
rm
i
n
a
M
A.
Red
u
c
ti
o
n
o
f
s
p
e
c
k
l
e
n
o
i
s
e
u
l
tra
s
o
u
n
d
i
m
a
g
e
s
b
a
s
e
d
o
n
TV
re
g
u
l
a
ri
z
a
t
i
o
n
a
n
d
m
o
d
i
f
i
e
d
b
a
y
e
s
s
h
ri
n
k
t
e
c
h
n
i
q
u
e
s
.
O
p
t
i
k
-
In
te
rn
a
ti
o
n
a
l
J
o
u
rn
a
l
fo
r
L
i
g
h
t
a
n
d
El
e
c
tr
o
n
O
p
ti
c
s
.
2016;
1
2
7
(
2
4
):
1
1
7
3
2
-
1
1
7
4
4
.
[3
6
]
Din
g
Y
,
Se
l
e
s
n
i
c
k
I
W.
Arti
f
a
c
t
-
fre
e
wav
e
l
e
t
d
e
n
o
i
s
i
n
g
:
N
o
n
-
c
o
n
v
e
x
s
p
a
rs
e
re
g
u
l
a
r
i
z
a
ti
o
n
,
c
o
n
v
e
x
o
p
ti
m
i
z
a
ti
o
n
.
IEEE
S
i
g
n
a
l
Pro
c
e
s
s
i
n
g
L
e
tt
e
rs
.
2
0
1
5
;
2
2
(
9
):
1
3
6
4
-
1
3
6
8
.
[3
7
]
He
W
,
Yi
C
,
L
i
Y
,
X
i
a
o
H.
R
e
s
e
a
r
c
h
o
n
M
e
c
h
a
n
i
c
a
l
F
a
u
l
t
Dia
g
n
o
s
i
s
Sc
h
e
m
e
B
a
s
e
d
o
n
I
m
p
ro
v
e
d
Wa
v
e
l
e
t
T
o
ta
l
Va
ri
a
ti
o
n
Den
o
i
s
i
n
g
.
Sh
o
c
k
a
n
d
V
i
b
ra
ti
o
n
.
2
0
1
6
;
2016
.
[3
8
]
Kh
m
a
g
A
,
Ra
m
l
i
AR
,
Al
-
Had
d
a
d
S
,
Ha
s
h
i
m
S
,
N
o
h
Z
M
,
Na
j
i
h
AA.
Des
i
g
n
o
f
n
a
tu
r
a
l
i
m
a
g
e
d
e
n
o
i
s
i
n
g
fi
l
te
r
b
a
s
e
d
o
n
s
e
c
o
n
d
-
g
e
n
e
ra
ti
o
n
wa
v
e
l
e
t
tr
a
n
s
fo
rm
a
ti
o
n
a
n
d
p
ri
n
c
i
p
l
e
c
o
m
p
o
n
e
n
t
a
n
a
l
y
s
i
s
.
J
o
u
rn
a
l
o
f
M
e
d
i
c
a
l
Im
a
g
i
n
g
a
n
d
He
a
l
th
I
n
fo
rm
a
ti
c
s
.
2
0
1
5
;
5
(
6
):
1
2
6
1
-
1266
.
[3
9
]
Ba
n
s
a
l
M
,
De
v
i
M
,
J
a
i
n
N
,
K
u
k
re
j
a
C.
A
Pro
p
o
s
e
d
A
p
p
ro
a
c
h
fo
r
Bi
o
m
e
d
i
c
a
l
Im
a
g
e
D
e
n
o
i
s
i
n
g
Us
i
n
g
PCA_NL
M
.
In
te
rn
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
B
i
o
-
S
c
i
e
n
c
e
a
n
d
B
i
o
-
Te
c
h
n
o
l
o
g
y
.
2
0
1
4
;
6
(
6
):
13
-
20.
[4
0
]
Hus
s
a
i
n
SA,
G
o
ra
s
h
i
SM
.
Im
a
g
e
d
e
n
o
i
s
i
n
g
a
l
g
o
ri
th
m
b
a
s
e
d
o
n
h
y
b
ri
d
n
e
i
g
h
b
o
r
h
o
o
d
fi
l
te
r
.
Si
g
n
a
l
a
n
d
Im
a
g
e
Pro
c
e
s
s
i
n
g
Ap
p
l
i
c
a
ti
o
n
s
(I
CSIPA).
2
0
1
3
:
25
-
30
.
[4
1
]
Nai
m
i
H
,
Ad
a
m
o
u
-
M
i
t
i
c
h
e
A
BH
,
M
i
ti
c
h
e
L
.
M
e
d
i
c
a
l
i
m
a
g
e
d
e
n
o
i
s
i
n
g
u
s
i
n
g
d
u
a
l
tre
e
c
o
m
p
l
e
x
th
re
s
h
o
l
d
i
n
g
wa
v
e
l
e
t
tra
n
s
f
o
r
m
a
n
d
Wi
e
n
e
r
fi
l
te
r.
J
o
u
r
n
a
l
o
f
Ki
n
g
S
a
u
d
Un
i
v
e
r
s
i
t
y
-
Co
m
p
u
te
r
a
n
d
In
fo
rm
a
ti
o
n
Sc
i
e
n
c
e
s
.
2
0
1
5
;
2
7
(1
):
40
-
4
5
.
[4
2
]
M
o
u
s
a
v
i
SM
,
L
a
n
g
s
t
o
n
CA.
Hy
b
ri
d
S
e
i
s
m
i
c
Den
o
i
s
i
n
g
Us
i
n
g
Hig
h
e
r
‐
O
rd
e
r
St
a
ti
s
ti
c
s
a
n
d
Im
p
r
o
v
e
d
Wa
v
e
l
e
t
Bl
o
c
k
Th
re
s
h
o
l
d
i
n
g
.
Bu
l
l
e
t
i
n
o
f
th
e
Se
i
s
m
o
l
o
g
i
c
a
l
So
c
i
e
t
y
o
f
Am
e
ri
c
a
.
2
0
1
6
;
1
0
6
(4
)
:
1380
-
1
3
8
9
.
[4
3
]
Swam
i
PD
,
J
a
i
n
A.
I
m
a
g
e
d
e
n
o
i
s
i
n
g
b
y
s
u
p
e
rv
i
s
e
d
a
d
a
p
ti
v
e
fu
s
i
o
n
o
f
d
e
c
o
m
p
o
s
e
d
i
m
a
g
e
s
r
e
s
t
o
re
d
u
s
i
n
g
wa
v
e
a
to
m
,
c
u
r
v
e
l
e
t
a
n
d
wa
v
e
l
e
t
tr
a
n
s
fo
r
m
.
S
i
g
n
a
l
,
I
m
a
g
e
a
n
d
Vi
d
e
o
Pro
c
e
s
s
i
n
g
.
2
0
1
4
;
8
(3
):
4
4
3
-
4
5
9
.
[4
4
]
L
e
Po
g
a
m
A
,
Han
z
o
u
l
i
H
,
Ha
t
t
M
,
L
e
Re
s
t
CC
,
Vi
s
v
i
k
i
s
D.
Den
o
i
s
i
n
g
o
f
PET
i
m
a
g
e
s
b
y
c
o
m
b
i
n
i
n
g
wav
e
l
e
ts
a
n
d
c
u
r
v
e
l
e
ts
fo
r
i
m
p
ro
v
e
d
p
re
s
e
r
v
a
ti
o
n
o
f
re
s
o
l
u
ti
o
n
a
n
d
q
u
a
n
t
i
ta
t
i
o
n
.
M
e
d
i
c
a
l
i
m
a
g
e
a
n
a
l
y
s
i
s
.
2
0
1
3
;
1
7
(
8
):
8
7
7
-
8
9
1
.
[4
5
]
Ba
p
p
y
D
,
J
e
o
n
I.
C
o
m
b
i
n
a
t
i
o
n
o
f
h
y
b
ri
d
m
e
d
i
a
n
fi
l
te
r
a
n
d
to
ta
l
v
a
ri
a
ti
o
n
m
i
n
i
m
i
s
a
ti
o
n
f
o
r
m
e
d
i
c
a
l
X
-
ra
y
i
m
a
g
e
re
s
to
ra
t
i
o
n
.
IET
I
m
a
g
e
Pro
c
e
s
s
i
n
g
.
2
0
1
6
;
1
0
(4
)
:
261
-
2
7
1
.
[4
6
]
O
k
te
m
R
,
Eg
i
a
z
a
ri
a
n
K
,
L
u
k
i
n
V
,
Po
n
o
m
a
re
n
k
o
N
,
Ts
y
m
b
a
l
O
.
L
o
c
a
l
l
y
a
d
a
p
t
i
v
e
DC
T
fi
l
te
ri
n
g
fo
r
s
i
g
n
al
-
d
e
p
e
n
d
e
n
t
n
o
i
s
e
r
e
m
o
v
a
l
.
EUR
ASIP
J
o
u
r
n
a
l
o
n
A
d
v
a
n
c
e
s
i
n
Si
g
n
a
l
Pro
c
e
s
s
i
n
g
.
2
0
0
7
;
2007
(1
)
:
0
4
2
4
7
2
.
[4
7
]
Ba
rte
n
PG
.
Con
tra
s
t
s
e
n
s
i
ti
v
i
ty
o
f
th
e
h
u
m
a
n
e
y
e
a
n
d
i
ts
e
ff
e
c
ts
o
n
i
m
a
g
e
q
u
a
l
i
ty
.
S
PIE
p
re
s
s
.
1
9
9
9
;
7
2
.
[4
8
]
Bu
a
d
e
s
A
,
Col
l
B
,
M
o
re
l
J
M
.
A
re
v
i
e
w
o
f
i
m
a
g
e
d
e
n
o
i
s
i
n
g
a
l
g
o
ri
t
h
m
s
,
w
i
th
a
n
e
w
o
n
e
.
M
u
l
ti
s
c
a
l
e
M
o
d
e
l
i
n
g
&
Si
m
u
l
a
ti
o
n
.
2
0
0
5
;
4
(2
):
4
9
0
-
5
3
0
.
[4
9
]
G
ra
p
s
A.
An
i
n
tr
o
d
u
c
ti
o
n
t
o
wav
e
l
e
ts
.
IEEE
c
o
m
p
u
t
a
ti
o
n
a
l
s
c
i
e
n
c
e
a
n
d
e
n
g
i
n
e
e
r
i
n
g
.
1
9
9
5
;
2
(2
)
:
50
-
61.
[5
0
]
Sa
k
ri
s
o
n
D.
O
n
th
e
r
o
l
e
o
f
t
h
e
o
b
s
e
rv
e
r
a
n
d
a
d
i
s
to
rti
o
n
m
e
a
s
u
re
i
n
i
m
a
g
e
tra
n
s
m
i
s
s
i
o
n
.
IEE
E
Tra
n
s
a
c
t
i
o
n
s
o
n
Co
m
m
u
n
i
c
a
ti
o
n
s
.
1
9
7
7
;
2
5
(
1
1
):
1
2
5
1
-
1
2
6
7
.
[5
1
]
Wa
n
g
Z,
B
o
v
i
k
AC,
Sh
e
i
k
h
HR
,
Si
m
o
n
c
e
l
l
i
EP.
Im
a
g
e
q
u
a
l
i
t
y
a
s
s
e
s
s
m
e
n
t:
fro
m
e
r
ro
r
v
i
s
i
b
i
l
i
ty
to
s
tru
c
tu
ra
l
s
i
m
i
l
a
ri
ty
.
IEEE
Tra
n
s
a
c
t
i
o
n
s
o
n
I
m
a
g
e
Pr
o
c
e
s
s
i
n
g
.
2
0
0
4
;
1
3
(
4
):
6
0
0
-
6
1
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.