I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
11
,
No
.
3
,
J
u
n
e
2
0
2
1
,
p
p
.
2595
~
2
6
0
3
I
SS
N:
2
0
8
8
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/
ijece
.
v
1
1
i
3
.
pp
2
5
9
5
-
2
6
0
3
2595
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Co
ntras
t
-
dis
torte
d ima
g
e quality
a
ss
ess
ment bas
ed o
n curvele
t
do
ma
in f
ea
tures
I
s
m
a
il T
a
ha
Ahm
ed
1
,
Chen
So
o
ng
Der
2
,
B
a
ra
a
T
a
re
q H
a
m
ma
d
3
,
No
rz
ia
na
J
a
m
il
4
1,
3
Co
ll
e
g
e
o
f
C
o
m
p
u
ter
S
c
ien
c
e
s a
n
d
In
f
o
rm
a
ti
o
n
Tec
h
n
o
l
o
g
y
,
U
n
i
v
e
rsity
o
f
An
b
a
r,
An
b
a
r,
Ira
q
2
Co
ll
e
g
e
o
f
G
ra
d
u
a
te S
tu
d
ies
,
Un
i
v
e
rsiti
Ten
a
g
a
Na
sio
n
a
l,
M
a
lay
si
a
4
Co
ll
e
g
e
o
f
Co
m
p
u
ti
n
g
a
n
d
I
n
fo
r
m
a
ti
c
s,
Un
iv
e
rsiti
Ten
a
g
a
Na
sio
n
a
l,
M
a
lay
sia
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
1
1
,
2
0
1
9
R
ev
is
ed
J
u
l
2
6
,
2
0
20
Acc
ep
ted
Sep
24
,
2
0
20
Co
n
tras
t
is o
n
e
o
f
th
e
m
o
st
p
o
p
u
l
a
r
fo
rm
s o
f
d
isto
r
ti
o
n
.
Re
c
e
n
tl
y
,
t
h
e
e
x
isti
n
g
ima
g
e
q
u
a
li
t
y
a
ss
e
ss
m
e
n
t
a
lg
o
ri
th
m
s
(IQA
s)
wo
rk
s
fo
c
u
sin
g
o
n
d
isto
rte
d
ima
g
e
s
b
y
c
o
m
p
re
ss
io
n
,
n
o
ise
a
n
d
b
lu
rr
in
g
.
Re
d
u
ced
-
re
fe
re
n
c
e
ima
g
e
q
u
a
li
ty
m
e
tri
c
fo
r
c
o
n
tras
t
-
c
h
a
n
g
e
d
ima
g
e
s
(RIQMC)
a
n
d
n
o
re
fe
re
n
c
e
-
ima
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t
(NR
-
IQA
)
fo
r
c
o
n
tras
t
-
d
isto
rted
ima
g
e
s
(NR
-
IQA
-
CDI)
h
a
v
e
b
e
e
n
c
re
a
ted
f
o
r
CDI.
NR
-
IQA
-
CDI
sh
o
we
d
p
o
o
r
p
e
rf
o
rm
a
n
c
e
in
t
wo
o
u
t
o
f
th
re
e
ima
g
e
d
a
t
a
b
a
se
s,
wh
e
re
t
h
e
P
e
a
rso
n
c
o
rre
latio
n
c
o
e
fficie
n
t
(P
LCC)
we
re
o
n
ly
0
.
5
7
3
9
a
n
d
0
.
7
6
2
3
in
TID2
0
1
3
a
n
d
CS
IQ
d
a
tab
a
se
,
re
sp
e
c
ti
v
e
ly
.
S
p
a
ti
a
l
d
o
m
a
in
fe
a
tu
re
s
a
re
t
h
e
b
a
sis
o
f
NR
-
IQA
-
CDI
a
r
c
h
it
e
c
tu
re
.
Th
e
re
fo
re
,
i
n
th
is
p
a
p
e
r,
t
h
e
sp
a
ti
a
l
d
o
m
a
in
fe
a
tu
re
s
a
re
c
o
m
p
lem
e
n
tary
wi
th
c
u
rv
e
let
d
o
m
a
in
fe
a
tu
re
s,
in
o
rd
e
r
to
tak
e
a
d
v
a
n
tag
e
o
f
t
h
e
p
o
ten
t
p
ro
p
e
rti
e
s
o
f
t
h
e
c
u
rv
e
let
in
e
x
trac
ti
n
g
i
n
fo
r
m
a
ti
o
n
fr
o
m
ima
g
e
s
su
c
h
a
s
m
u
lt
isc
a
le
a
n
d
m
u
lt
id
irec
ti
o
n
a
l.
Th
e
e
x
p
e
rime
n
t
a
l
o
u
tco
m
e
re
l
y
o
n
K
-
fo
ld
c
r
o
ss
v
a
li
d
a
ti
o
n
(K
ra
n
g
e
d
2
-
1
0
)
a
n
d
sta
ti
stica
l
te
st
sh
o
we
d
t
h
a
t
t
h
e
p
e
rfo
rm
a
n
c
e
o
f
NR
-
IQA
-
CDI
re
ly
o
n
c
u
rv
e
let
d
o
m
a
in
fe
a
tu
re
s
(NR
-
IQA
-
CDI
-
Cv
T)
sig
n
ifi
c
a
n
t
ly
su
rp
a
ss
e
s th
o
se
wh
ic
h
a
re
re
ly
o
n
fiv
e
sp
a
ti
a
l
d
o
m
a
in
fe
a
tu
re
s.
K
ey
w
o
r
d
s
:
C
o
n
tr
ast
-
d
is
to
r
ted
im
ag
e
I
QAs
NR
-
I
QA
NR
-
I
QA
-
C
DI
NR
-
I
QA
-
C
DI
-
C
v
T
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
I
s
m
ail
T
ah
a
Ah
m
ed
C
o
lleg
e
o
f
C
o
m
p
u
ter
Scien
ce
s
an
d
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
y
Un
iv
er
s
ity
o
f
An
b
ar
An
b
ar
,
I
r
aq
E
m
ail:
is
m
ail.
tah
a@
u
o
an
b
ar
.
e
d
u
.
iq
1.
I
NT
RO
D
UCT
I
O
N
Var
io
u
s
k
in
d
s
o
f
d
is
to
r
tio
n
s
u
ch
as
n
o
is
e,
b
lu
r
r
in
g
,
f
ast
f
ad
i
n
g
,
b
l
o
ck
in
g
ar
tifa
cts
an
d
co
n
t
r
ast
wh
ich
m
ay
ap
p
ea
r
b
ec
au
s
e
o
f
s
o
m
e
o
f
ce
r
tain
p
r
o
ce
s
s
es
o
n
th
e
im
ag
e
ca
n
d
eg
r
a
d
e
th
e
q
u
ality
o
f
im
ag
es.
Su
b
jectiv
e
an
d
o
b
jectiv
e
m
eth
o
d
s
ar
e
tw
o
ty
p
es
o
f
I
QA
o
r
v
id
e
o
q
u
ality
ass
ess
m
en
t
(
VQA)
u
s
ed
to
e
v
alu
ate
th
e
im
ag
e
q
u
ality
[
1
,
2
]
.
I
n
r
ea
l
-
tim
e
ap
p
licatio
n
s
is
im
p
r
ac
tical
to
u
s
e
s
u
b
jectiv
e
q
u
ality
ass
ess
m
e
n
t
b
ec
au
s
e
it
tak
es
tim
e
an
d
is
ex
p
en
s
iv
e.
T
h
er
ef
o
r
e,
o
b
jectiv
e
I
QA
alg
o
r
ith
m
s
ar
e
th
e
b
est
s
o
lu
tio
n
s
in
ce
th
e
r
o
le
o
f
h
u
m
an
is
lim
ited
in
o
r
d
er
to
p
r
ed
ict
t
h
e
q
u
ality
o
f
im
ag
e.
Ob
jectiv
e
I
QAs
ca
n
b
e
g
r
o
u
p
ed
i
n
to
f
u
ll
r
ef
er
en
ce
(
FR
)
,
r
ed
u
ce
r
e
f
er
en
ce
(
R
R
)
an
d
n
o
r
ef
er
en
ce
(
NR
)
.
T
h
ese
g
r
o
u
p
s
r
eq
u
ir
e
th
e
av
ailab
ilit
y
o
f
o
r
ig
i
n
al
im
ag
e
[3
-
5
]
.
Mo
s
t
o
f
FR
-
I
QA
an
d
RR
-
I
QA
ap
p
licatio
n
s
ar
e
co
n
s
tr
ain
ed
d
u
e
to
av
ailab
le
o
f
o
r
ig
i
n
al
im
ag
e.
T
h
er
ef
o
r
e,
NR
-
I
QA
is
t
h
e
b
e
s
t
s
o
lu
tio
n
f
o
r
th
is
ca
s
e.
R
ec
en
tly
,
th
e
ex
is
tin
g
NR
-
I
QA
wo
r
k
s
f
o
cu
s
in
g
o
n
d
is
to
r
ted
im
ag
es
b
y
c
o
m
p
r
e
s
s
io
n
,
n
o
is
e
an
d
b
lu
r
r
in
g
[
6
,
7
]
.
C
o
n
tr
ast
is
o
n
e
o
f
th
e
m
o
s
t
p
o
p
u
lar
f
o
r
m
s
o
f
d
is
to
r
tio
n
.
Fig
u
r
e
1
s
h
o
ws
th
e
co
n
tr
ast
-
d
is
to
r
ted
im
ag
e
is
lo
w
g
r
ay
s
ca
le
im
ag
e
[
8
]
.
p
o
o
r
lig
h
tin
g
co
n
d
itio
n
an
d
l
o
w
q
u
ality
im
a
g
e
ac
q
u
is
itio
n
d
e
v
ice
ca
n
p
r
o
d
u
ce
co
n
t
r
ast
d
is
to
r
tio
n
[
9
,
1
0
]
.
Un
f
o
r
tu
n
ately
,
th
e
c
u
r
r
e
n
t
wo
r
k
ca
r
r
ie
d
o
u
t i
n
th
e
ar
ea
o
f
NR
-
I
QA
f
o
r
C
DI
is
v
er
y
m
in
i
m
al.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
3
,
J
u
n
e
2
0
2
1
:
2
5
9
5
-
2603
2596
(
a)
(
b
)
Fig
u
r
e
1
.
C
o
n
tr
ast
-
d
is
to
r
ted
im
ag
e
,
(
a)
g
o
o
d
co
n
tr
ast im
ag
e
(
m
y
len
a.
b
m
p
)
,
an
d
(
b
)
p
o
o
r
c
o
n
tr
ast im
ag
e
[
8
]
R
ec
en
tly
,
th
e
ex
is
tin
g
I
QAs
wo
r
k
s
f
o
r
C
DI
ar
e
(
1
)
R
I
Q
MC
[
1
1
]
b
ased
o
n
e
n
tr
o
p
ies
an
d
im
ag
e
h
is
to
g
r
am
o
r
d
er
s
tatis
tics
;
an
d
(
2
)
NR
-
I
QA
-
C
DI
[
1
2
]
wh
i
ch
is
co
n
s
tr
u
cted
o
n
th
e
b
asis
o
f
NSS
p
r
in
cip
les
s
u
ch
th
at
th
e
r
e
ar
e
ce
r
tain
r
e
g
u
lar
ities
in
n
atu
r
al
s
ce
n
e
s
ta
tis
tics
th
at
m
ay
b
e
m
is
s
in
g
f
r
o
m
th
e
s
tatis
tics
o
f
d
is
to
r
ted
im
ag
es.
Glo
b
al
s
p
at
ial
s
tatis
tic
s
f
ea
tu
r
es
ar
e
u
s
ed
in
NR
-
I
QA
-
C
DI
in
clu
d
i
n
g
t
h
e
m
ea
n
,
s
tan
d
ar
d
d
ev
iatio
n
,
en
tr
o
p
y
,
k
u
r
to
s
is
an
d
s
k
ewn
ess
.
R
eg
r
ettab
ly
,
N
R
-
I
QA
-
C
DI
p
er
f
o
r
m
a
n
ce
a
r
e
n
o
t
p
o
s
itiv
e
in
s
o
m
e
o
f
test
im
ag
e
d
atab
ases
,
T
I
D2
0
1
3
an
d
C
SIQ
,
wh
er
e
th
e
PLC
C
ar
e
o
n
ly
ar
o
u
n
d
0
.
5
7
an
d
0
.
7
6
,
r
esp
ec
tiv
ely
.
T
h
e
cu
r
r
e
n
t
NR
-
I
QA
-
C
DI
is
b
ased
o
n
g
l
o
b
al
s
p
atial
s
tatis
t
ics
wh
ich
s
u
f
f
er
s
f
r
o
m
th
r
ee
p
r
o
b
lem
s
.
Seco
n
d
ly
,
s
p
atial
d
o
m
ain
is
lack
in
g
o
f
in
f
o
o
n
s
ca
le
a
n
d
o
r
ien
tatio
n
wh
ic
h
ar
e
es
s
en
tial
f
o
r
co
m
p
ac
t
r
ep
r
esen
t
atio
n
o
f
cu
r
v
e
wh
ich
is
a
co
m
m
o
n
s
tr
u
ctu
r
e
in
n
atu
r
al
im
ag
e.
C
o
m
p
ac
t
r
e
p
r
esen
tatio
n
is
ess
en
tial
f
o
r
an
aly
s
is
to
b
e
f
o
cu
s
ed
o
n
r
el
ev
an
t
in
f
o
r
m
atio
n
t
o
r
ed
u
ce
c
o
n
f
u
s
io
n
b
y
ir
r
elev
an
t
in
f
o
o
r
n
o
is
e.
I
t
h
as
b
ee
n
p
r
o
v
e
n
u
s
ef
u
l in
im
ag
e
co
n
tr
a
s
t e
n
h
an
ce
m
en
t a
n
d
I
QA.
Py
r
am
id
an
d
wav
elet
tr
an
s
f
o
r
m
is
o
n
e
o
f
m
u
ltis
ca
le
d
ec
o
m
p
o
s
itio
n
tr
an
s
f
o
r
m
(
MSD)
h
as
b
ee
n
u
s
ed
s
u
cc
ess
f
u
lly
in
m
an
y
im
ag
e
p
r
o
ce
s
s
in
g
a
p
p
licatio
n
s
.
Ho
wev
er
,
W
T
lack
s
d
ir
ec
ti
o
n
ality
.
m
u
ltis
ca
le
g
eo
m
etr
ical
an
al
y
s
is
(
MG
A)
tr
an
s
f
o
r
m
s
[
1
3
]
wer
e
p
r
o
p
o
s
ed
to
r
eso
lv
e
th
is
is
s
u
e.
C
u
r
v
e
let
tr
an
s
f
o
r
m
(
C
T
)
d
ec
o
m
p
o
s
es
th
e
o
r
ig
in
al
im
ag
e
in
to
s
et
o
f
f
r
eq
u
e
n
cy
co
e
f
f
ic
ien
ts
o
f
s
u
b
b
an
d
s
with
v
ar
io
u
s
s
ca
les,
o
r
ien
tatio
n
an
d
lo
ca
tio
n
[
1
4
]
.
I
n
C
o
ar
s
e
p
ar
t,
L
o
w
-
f
r
e
q
u
en
c
y
co
ef
f
icie
n
ts
ar
e
d
is
tr
ib
u
ted
.
I
n
Fin
e
p
a
r
t,
h
ig
h
f
r
e
q
u
en
cies
co
ef
f
icien
ts
ar
e
d
is
tr
ib
u
ted
.
An
d
th
e
m
id
d
le
lay
e
r
ar
e
d
is
tr
ib
u
ted
to
d
etail
.
C
u
r
v
el
et
h
as
a
co
m
m
o
n
ch
ar
ac
ter
is
tics
th
at
d
is
tin
g
u
is
h
es
it
f
r
o
m
th
e
o
t
h
er
ty
p
es
o
f
T
r
an
s
f
o
r
m
s
ar
e
Hig
h
er
d
i
r
ec
tio
n
al
s
en
s
itiv
ity
an
d
lo
wer
r
ed
u
n
d
an
c
y
[
1
5
]
.
T
h
e
f
ir
s
t
NR
-
I
Q
A
[
1
6
]
u
s
e
d
f
e
a
t
u
r
e
s
d
e
r
i
v
e
d
f
r
o
m
c
u
r
v
e
l
e
t
t
r
a
n
s
f
o
r
m
.
L
i
u
e
t
a
l
.
[
1
7
]
an
d
Sh
en
et
a
l
.
[
1
8
]
in
tr
o
d
u
ce
d
a
n
ew
NR
-
I
QA
b
ased
o
n
C
T
.
T
h
e
s
u
cc
e
s
s
o
f
all
o
f
th
ese
wo
r
k
s
r
ely
o
n
th
e
s
tr
en
g
th
o
f
cu
r
v
elet,
in
wh
ich
C
T
p
r
o
v
id
e
s
a
r
ich
s
o
u
r
ce
o
f
o
r
ien
tatio
n
i
n
f
o
r
m
atio
n
o
n
im
ag
es.
Su
ch
i
n
f
o
co
u
l
d
b
e
u
s
ed
to
d
etec
t
th
e
p
r
esen
ce
o
f
d
is
to
r
t
io
n
in
th
e
im
ag
e.
L
iu
et
a
l
.
[
1
7
]
p
r
esen
t
an
ef
f
ec
tiv
e
NR
-
I
QA
u
s
in
g
en
er
g
y
f
ea
tu
r
es d
er
iv
ed
f
r
o
m
th
e
cu
r
v
elet
d
o
m
ain
.
T
h
er
e
f
ea
tu
r
es we
r
e
f
o
u
n
d
clo
s
ely
r
elate
d
th
r
o
u
g
h
v
ar
io
u
s
ty
p
es o
f
d
is
to
r
tio
n
to
th
e
n
atu
r
al
im
a
g
e
q
u
ality
.
Ho
wev
e
r
,
all
o
f
th
ese
wo
r
k
s
p
r
o
p
o
s
ed
s
p
ec
ially
f
o
r
q
u
ality
ev
alu
atio
n
o
f
d
is
to
r
ted
im
ag
es
d
u
e
to
c
o
m
p
r
ess
io
n
,
n
o
is
e,
an
d
b
l
u
r
r
i
n
g
.
T
h
is
s
tu
d
y
is
m
o
tiv
ate
d
b
y
th
e
s
tr
en
g
t
h
o
f
cu
r
v
elet
tr
an
s
f
o
r
m
in
ef
f
icien
t
r
ep
r
esen
tatio
n
o
f
cu
r
v
e,
a
b
asic
s
tr
u
ctu
r
e
o
f
ten
s
ee
n
in
n
atu
r
al
s
ce
n
e
im
ag
es.
I
t
b
eg
in
s
with
ex
p
lo
r
in
g
th
e
u
s
e
o
f
cu
r
v
elet
d
o
m
ai
n
f
ea
tu
r
es
i
n
ch
ar
ac
ter
izin
g
co
n
tr
ast
-
d
is
to
r
ted
im
ag
es
(
C
DI
)
.
T
h
er
ef
o
r
e,
th
is
p
ap
e
r
en
h
a
n
ce
s
th
e
ex
is
tin
g
NR
-
I
QA
-
C
DI
u
s
in
g
cu
r
v
elet
d
o
m
ai
n
f
ea
tu
r
es.
2.
CURV
E
L
E
T
T
RANSF
O
R
M
B
y
ap
p
ly
in
g
an
e
f
f
ec
tiv
e
p
a
r
a
b
o
lic
s
ca
lin
g
law:
wid
th
≅
(
le
n
g
th
)
2
,
t
h
e
C
T
ca
n
b
etter
r
e
p
r
e
s
en
t
ed
g
es
an
d
o
th
er
s
in
g
u
lar
ities
alo
n
g
c
u
r
v
es
[
1
3
]
.
Ho
r
iz
o
n
tal,
v
er
tica
l
an
d
d
iag
o
n
al
a
r
e
lim
ited
d
ir
e
ctio
n
al
in
f
o
r
m
atio
n
o
n
ly
ca
p
tu
r
e
d
in
W
av
elet
tr
an
s
f
o
r
m
.
Ho
we
v
er
,
cu
r
v
elet
tr
an
s
f
o
r
m
d
ec
o
m
p
o
s
e
th
e
im
ag
e
w
ith
g
eo
m
et
r
ic
b
ases
in
m
u
ltip
le
d
ir
ec
tio
n
s
an
d
p
o
s
itio
n
s
[
1
3
,
1
5
]
.
Sin
ce
th
e
cu
r
v
elet
co
ef
f
icien
ts
ar
e
ca
teg
o
r
ized
b
ased
o
n
o
r
ien
tatio
n
an
d
s
ca
le,
th
e
cu
r
v
elet
ed
g
e
is
s
m
o
o
th
er
th
a
n
th
e
W
av
elet
ed
g
e.
Fig
u
r
e
2
s
h
o
ws
th
e
ed
g
e
r
ep
r
esen
tatio
n
in
c
u
r
v
elet
an
d
wav
elet
tr
an
s
f
o
r
m
s
.
I
t
ca
n
b
e
o
b
s
er
v
e
th
at
cu
r
v
elet
ed
g
e
ca
n
b
e
r
ep
r
esen
ted
b
y
s
m
aller
am
o
u
n
t
o
f
la
r
g
e
c
o
ef
f
icien
ts
,
wh
ile
wav
elet
ed
g
e
ca
n
b
e
r
ep
r
esen
ted
b
y
a
b
ig
g
er
am
o
u
n
t
o
f
lar
g
e
co
ef
f
icien
ts
in
th
e
f
i
n
e
s
ca
le.
I
n
last
g
en
er
atio
n
o
f
cu
r
v
el
et
tr
an
s
f
o
r
m
,
C
an
d
'
es
et
al
.
p
r
o
p
o
s
ed
a
two
f
ast
d
is
cr
e
te
cu
r
v
elet
tr
an
s
f
o
r
m
s
(
FDC
T
)
i.e
:
E
q
u
atio
n
(
1
)
u
n
eq
u
ally
-
s
p
ac
ed
f
ast
f
o
u
r
ier
tr
an
s
f
o
r
m
(
USFF
T
)
b
ased
cu
r
v
elet
an
d
(
2
)
f
r
eq
u
e
n
cy
wr
ap
p
in
g
b
ased
cu
r
v
elet
[
1
5
]
.
T
h
e
DC
T
o
f
a
2
-
D
f
u
n
ctio
n
ƒ
[
t1
,
t2
]
is
f
o
r
m
u
late
d
as:
(
,
ℓ
,
)
≔
∑
[
1
,
2
]
0
≤
1
,
2
<
,
ℓ
,
[
1
,
2
]
̅
̅
̅
̅
̅
̅
̅
̅
̅
̅
̅
̅
̅
̅
̅
,
(
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
-
8
7
0
8
C
o
n
tr
a
s
t
-
d
is
to
r
ted
ima
g
e
q
u
a
l
ity
a
s
s
es
s
men
t b
a
s
ed
o
n
c
u
r
ve
let
d
o
ma
in
fea
t
u
r
es
(
I
s
ma
il Ta
h
a
A
h
med
)
2597
wh
er
e
φ
,
,
ℓ
,
ar
e
c
u
r
v
elet
f
u
n
ctio
n
s
,
s
ca
le,
o
r
ien
tatio
n
an
d
p
o
s
itio
n
r
esp
ec
tiv
ely
.
t1
,
t2
d
e
n
o
te
c
o
o
r
d
in
ates
in
th
e
s
p
atial
d
o
m
ain
:
0
≤
t1
,
t2
<
n
.
(
,
ℓ
,
)
d
en
o
tes
cu
r
v
elet
co
e
f
f
ici
en
t.
I
n
o
u
r
w
o
r
k
,
FDC
T
b
y
wr
ap
p
in
g
is
ap
p
lied
.
Fig
u
r
e
2
.
W
av
elet
an
d
cu
r
v
elet
tr
an
s
f
o
r
m
s
ed
g
e
r
ep
r
esen
tati
o
n
s
[
1
6
]
3.
AP
P
L
I
CA
T
I
O
N
O
F
CUR
V
E
L
E
T
DO
M
AI
N
F
E
A
T
UR
E
S IN NR
-
I
Q
A
-
CD
I
Her
e,
we
p
r
esen
ts
h
o
w
cu
r
v
el
et
d
o
m
ain
f
ea
t
u
r
es
ar
e
ap
p
lie
d
in
NR
-
I
QA
-
C
DI
.
I
t
b
eg
in
s
with
tr
y
in
g
to
ad
d
r
ess
in
g
th
e
p
r
o
b
lem
s
o
f
cu
r
r
e
n
t
NR
-
I
QA
-
C
DI
f
o
llo
w
ed
b
y
f
o
r
m
u
latin
g
cu
r
v
elet
d
o
m
ain
f
ea
t
u
r
es
an
d
p
r
ed
ictin
g
im
a
g
e
q
u
ality
u
s
in
g
th
e
f
ea
tu
r
es.
3
.
1
.
CDI
di
s
t
ing
uis
h
C
T
is
an
ef
f
ec
tiv
e
im
ag
e
r
ep
r
esen
tatio
n
b
y
s
m
aller
am
o
u
n
t o
f
lar
g
e
co
ef
f
icien
ts
as c
o
m
p
ar
ed
to
th
o
s
e
in
th
e
s
p
atial
d
o
m
ain
as m
e
n
tio
n
in
[
1
9
]
.
L
ar
g
e
-
m
ag
n
itu
d
e
c
o
ef
f
icien
ts
o
n
ly
o
cc
u
r
i
n
r
eg
io
n
s
o
f
th
e
im
ag
e
th
at
co
n
tain
f
in
e
d
etails;
f
in
e
s
ca
l
es
co
ef
f
icien
ts
s
h
o
w
th
e
e
x
is
ten
ce
lo
ca
l
in
f
o
in
an
im
a
g
e.
B
u
t
in
th
e
co
ar
s
est
s
ca
le
co
n
s
is
t
s
o
f
m
ain
ly
lar
g
e
-
m
ag
n
itu
d
e
co
ef
f
icien
ts
an
d
m
atch
es
th
e
im
ag
e
in
lo
w
s
p
atial
r
eso
lu
tio
n
.
Ov
er
all
co
n
tr
ast
an
d
b
r
ig
h
t
n
e
s
s
o
f
an
im
ag
e
ar
e
d
ep
en
d
s
o
n
th
e
d
is
tr
ib
u
tio
n
o
f
th
e
co
ef
f
icien
ts
.
I
n
an
im
ag
e
with
g
o
o
d
co
n
t
r
ast
m
u
s
t
b
e
in
clu
d
ed
a
p
o
wer
f
u
l
o
v
er
all
co
n
tr
ast,
r
elativ
ely
b
r
i
g
h
t
ap
p
ea
r
an
ce
an
d
f
in
e
d
etails
.
I
n
o
r
d
er
to
r
ep
r
esen
t
th
e
c
h
ar
ac
ter
is
tics
,
ea
s
ily
ca
n
b
e
ac
co
m
p
lis
h
ed
b
y
u
s
in
g
d
is
tr
ib
u
tio
n
o
f
cu
r
v
elet
co
ef
f
icien
ts
.
I
n
g
o
o
d
o
v
er
all
co
n
tr
ast
an
d
b
r
ig
h
t
a
p
p
ea
r
a
n
c
e,
a
d
is
tr
ib
u
tio
n
o
f
co
e
f
f
icien
ts
at
co
ar
s
est
s
ca
l
e
with
h
ig
h
d
is
p
er
s
io
n
a
n
d
h
ig
h
ce
n
tr
al
ten
d
en
cy
.
W
h
ile
in
p
r
esen
ce
o
f
f
in
e
d
etails,
a
d
is
tr
ib
u
tio
n
o
f
co
ef
f
icien
ts
at
f
in
er
s
ca
le
with
h
ig
h
c
en
tr
al
ten
d
e
n
cy
.
B
r
ie
f
l
y
,
f
r
o
m
th
e
o
b
s
er
v
atio
n
o
f
t
h
e
d
is
tr
ib
u
tio
n
s
o
f
cu
r
v
elet
co
ef
f
icien
ts
at
th
e
c
o
ar
s
est
an
d
f
in
est
s
ca
les
ca
n
b
e
ac
h
iev
e
d
b
y
g
en
e
r
ate
p
o
s
s
ib
le
f
ea
tu
r
es
in
ch
ar
ac
ter
izin
g
g
o
o
d
co
n
t
r
ast
im
ag
es
an
d
also
C
DI
[
1
9
]
.
T
h
er
ef
o
r
e,
in
th
e
n
ex
t
s
ec
tio
n
,
th
e
d
etails
ab
o
u
t
f
o
r
m
u
latin
g
th
e
f
ea
tu
r
es f
r
o
m
th
e
cu
r
v
elet
co
e
f
f
icien
ts
will b
e
p
r
esen
ted
.
3
.
2
.
F
o
rm
ula
t
ing
curv
elet
do
m
a
i
n f
ea
t
ures
T
h
is
s
ec
tio
n
p
r
esen
ts
th
e
two
m
ain
p
r
o
ce
s
s
es
in
th
e
p
r
o
p
o
s
ed
NR
-
I
QA
-
C
DI
r
ely
o
n
c
u
r
v
el
et
d
o
m
ain
f
ea
tu
r
es
(
NR
-
I
QA
-
C
DI
-
C
v
T
)
.
T
h
e
f
ir
s
t
p
r
o
ce
s
s
is
f
o
r
co
m
p
u
tin
g
cu
r
v
elet
f
ea
tu
r
es
an
d
th
e
s
ec
o
n
d
o
n
e
is
f
o
r
p
r
ed
ictin
g
im
a
g
e
q
u
ality
u
s
in
g
th
e
s
tatis
ti
c
s
as e
x
p
lain
ed
in
t
h
e
s
u
b
-
s
ec
tio
n
s
.
3
.
2
.
1
.
Co
m
pu
t
ing
curv
elet
f
e
a
t
ures
T
h
er
e
ar
e
3
s
tep
s
in
co
m
p
u
tin
g
cu
r
v
elet
f
ea
tu
r
es a
s
s
h
o
wn
i
n
Fig
u
r
e
3
.
Step
1
:
Pre
-
p
r
o
ce
s
s
in
g
I
n
th
is
s
tep
,
C
o
lo
u
r
im
ag
es
ar
e
co
n
v
er
ted
to
g
r
ay
s
ca
le
im
ag
e
u
s
in
g
(
2
)
,
wh
er
e
I
gray
,
I
red
I
gr
een
an
d
I
blue
ar
e
th
e
g
r
ay
s
ca
le
im
ag
e
,
an
d
r
ed
,
g
r
ee
n
a
n
d
b
lu
e
c
h
an
n
el
o
f
th
e
o
r
ig
in
al
R
GB
im
ag
e,
r
esp
ec
tiv
ely
.
MA
T
L
AB
©
f
u
n
ctio
n
r
g
b
2
g
r
a
y
(
)
is
u
s
ed
to
p
e
r
f
o
r
m
th
e
c
o
n
v
er
s
io
n
.
I
gray
=
0
.
2
9
8
9
I
red
+
0
.
5
8
7
0
I
green
+
0
.
1
1
4
0
I
blue
(
2)
I
n
ad
d
itio
n
,
r
esize
s
tep
was
u
s
ed
to
n
o
r
m
alize
th
e
s
ize
o
f
th
e
in
p
u
t
im
ag
e
t
o
h
av
e
t
h
e
s
ize
o
f
m
in
im
u
m
d
im
en
s
io
n
f
ix
e
d
to
2
5
6
i
n
o
r
d
er
to
av
o
i
d
m
is
m
atch
o
f
r
eso
lu
tio
n
d
u
r
i
n
g
co
m
p
ar
is
o
n
.
Step
2
:
Fas
t
d
is
cr
ete
cu
r
v
elet
tr
an
s
f
o
r
m
Usi
n
g
FDC
T
,
T
h
e
g
r
ay
s
ca
le
im
ag
e
is
p
ar
titi
o
n
ed
in
to
g
r
o
u
p
s
o
f
cu
r
v
elet
co
e
f
f
icien
ts
C
(
j
,
l,
k
1
,
k
2
)
as
d
ef
in
ed
b
y
(
3
)
.
T
h
e
cu
r
v
elet
tr
an
s
f
o
r
m
atio
n
was
d
o
n
e
u
s
in
g
th
e
cu
r
v
elet
2
.
1
.
2
t
o
o
lb
o
x
,
av
ailab
le
f
r
o
m
h
ttp
://www.
cu
r
v
elet.
o
r
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
3
,
J
u
n
e
2
0
2
1
:
2
5
9
5
-
2603
2598
(
,
,
1
,
2
)
=
∑
[
,
]
,
,
1
,
2
[
,
]
0
≤
<
0
≤
<
(
3
)
wh
er
e
,
,
1
,
2
is
a
cu
r
v
elet
co
ef
f
icien
ts
at
d
if
f
er
en
t scale
s
,
j
,
o
r
ien
ta
tio
n
s
,
l
an
d
lo
ca
tio
n
,
(
k
1,
k
2
)
[
2
0
]
.
Step
3
: E
x
tr
ac
t
f
ea
tu
r
es f
r
o
m
c
u
r
v
elet
co
ef
f
icien
ts
First,
th
e
m
ag
n
itu
d
e
o
f
co
ef
f
icien
ts
at
ea
ch
s
ca
le
j
,
|
|
wer
e
co
m
p
u
ted
.
C
o
ef
f
icien
ts
f
r
o
m
s
am
e
s
ca
le
b
u
t
d
if
f
e
r
en
t
o
r
ien
tatio
n
s
wer
e
g
r
o
u
p
ed
to
g
eth
er
.
Nex
t
,
th
e
m
ea
n
,
μ
(
)
an
d
s
tan
d
ar
d
d
ev
iatio
n
,
σ
(
)
o
f
th
e
cu
r
v
elet
co
ef
f
icien
ts
at
th
e
two
co
ar
s
est
s
ca
les
j
={
1
,
2
}
wer
e
ex
tr
ac
ted
as
th
e
f
o
u
r
f
ea
tu
r
e
s
,
f
1,
f
2,
f
3
an
d
f
4
as
d
ef
in
ed
b
y
(
4
)
-
(
7
)
.
1
=
(
|
1
|
)
(
4
)
2
=
(
|
2
|
)
(
5
)
3
=
(
|
1
|
)
(
6)
4
=
(
|
2
|
)
(
7)
T
h
e
last
f
ea
tu
r
e,
f
5
,
is
th
e
m
ea
n
,
μ
(
)
o
f
th
e
h
ig
h
est 1
% c
u
r
v
el
et
co
ef
f
icien
ts
,
′
at
th
e
f
in
est s
ca
le,
j
=5
as d
ef
in
ed
b
y
(
8
)
:
5
=
(
|
′
5
|
)
(
8
)
Fig
u
r
e
3
.
Flo
wch
ar
t
o
f
c
o
m
p
u
t
in
g
cu
r
v
elet
d
o
m
ain
f
ea
tu
r
es
3
.
2
.
2
.
P
re
dict
ing
im
a
g
e
qu
a
lity
T
h
e
s
tep
s
f
o
r
p
r
ed
ictin
g
im
a
g
e
q
u
ality
ar
e
s
h
o
wn
in
Fig
u
r
e
4
.
Step
1
:
Mo
d
ellin
g
u
s
in
g
cu
r
v
elet
d
o
m
ain
f
ea
tu
r
es
Step
1
.
1
:
C
o
m
p
u
te
cu
r
v
elet
d
o
m
ain
f
e
atu
r
es f
r
o
m
tr
ai
n
in
g
im
a
g
es
T
h
e
s
tep
to
co
m
p
u
te
cu
r
v
elet
d
o
m
ain
f
ea
tu
r
es
is
as
ex
p
lain
e
d
in
s
ec
tio
n
3
.
2
.
T
h
e
im
ag
es
u
s
ed
wer
e
f
r
o
m
tr
ain
i
n
g
s
et
f
o
r
m
o
d
ellin
g
p
u
r
p
o
s
es.
Step
1
.
2
: Per
f
o
r
m
f
ea
t
u
r
e
n
o
r
m
aliza
tio
n
an
d
r
eg
r
ess
io
n
I
n
f
ea
tu
r
e
n
o
r
m
aliza
tio
n
,
th
e
v
alu
es
o
f
ea
ch
f
ea
tu
r
e
ar
e
n
o
r
m
alize
d
ag
ain
s
t
th
eir
m
ea
n
,
an
d
s
tan
d
ar
d
d
e
v
iatio
n
,
s
u
ch
th
at
th
e
n
o
r
m
alize
d
v
alu
es,
will
h
av
e
ze
r
o
m
ea
n
a
n
d
u
n
it
s
tan
d
a
r
d
d
ev
iatio
n
as
d
ef
in
ed
in
(
9
)
.
Featu
r
e
n
o
r
m
al
izatio
n
is
im
p
o
r
tan
t
b
ef
o
r
e
p
er
f
o
r
m
in
g
an
y
m
ac
h
in
e
lear
n
in
g
s
u
ch
as
r
eg
r
ess
io
n
b
ec
au
s
e
it
ten
d
s
to
in
cr
ea
s
e
th
e
ac
cu
r
ac
y
.
=
−
(
9
)
T
h
e
m
ea
n
,
an
d
s
tan
d
ar
d
d
ev
i
atio
n
,
an
d
th
e
r
eg
r
ess
io
n
f
u
n
ctio
n
d
er
iv
ed
f
r
o
m
t
h
is
s
tep
will
b
e
u
s
ed
later
in
Step
2
.
2
.
R
eg
r
ess
io
n
aim
s
to
f
in
d
th
e
m
a
p
p
in
g
f
u
n
ctio
n
,
o
r
b
etter
k
n
o
wn
as
r
eg
r
ess
io
n
f
u
n
ctio
n
,
wh
ich
m
ap
in
d
e
p
en
d
e
n
t
v
ar
i
ab
les
to
d
e
p
en
d
e
n
t
v
a
r
iab
les.
I
n
th
is
wo
r
k
,
th
e
in
d
e
p
en
d
e
n
t
v
ar
iab
les
a
r
e
th
e
f
ea
tu
r
es
an
d
th
e
d
ep
e
n
d
en
t
v
a
r
iab
le
is
th
e
s
u
b
jectiv
e
m
ea
n
o
p
in
io
n
s
co
r
e
(
MO
S).
R
eg
r
es
s
io
n
is
im
p
o
r
tan
t
in
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
-
8
7
0
8
C
o
n
tr
a
s
t
-
d
is
to
r
ted
ima
g
e
q
u
a
l
ity
a
s
s
es
s
men
t b
a
s
ed
o
n
c
u
r
ve
let
d
o
ma
in
fea
t
u
r
es
(
I
s
ma
il Ta
h
a
A
h
med
)
2599
th
is
wo
r
k
to
r
em
o
v
e
n
o
n
-
lin
e
ar
ity
to
im
p
r
o
v
e
th
e
lin
ea
r
co
r
r
elatio
n
b
etwe
en
th
e
f
ea
t
u
r
es
an
d
MO
S.
I
n
th
is
wo
r
k
,
s
u
p
p
o
r
t
v
ec
to
r
r
eg
r
ess
io
n
(
SVR
)
(
v
ia
L
I
B
SVM
-
3
.
1
2
p
ac
k
a
g
e
[
2
1
]
)
is
u
s
ed
to
f
in
d
th
e
r
eg
r
ess
io
n
f
u
n
ctio
n
,
s
im
ilar
to
wh
at
h
as
b
ee
n
u
s
ed
in
th
e
cu
r
r
en
t
NR
-
I
QA
-
C
DI
f
o
r
f
air
co
m
p
ar
is
o
n
.
I
n
SVR
,
r
eg
r
ess
io
n
f
u
n
ctio
n
is
d
eter
m
in
ed
th
r
o
u
g
h
th
e
ap
p
r
o
ac
h
o
f
s
u
p
er
v
is
ed
m
ac
h
in
e
lear
n
in
g
.
Step
2
:
C
o
m
p
u
tin
g
im
ag
e
q
u
ality
Step
2
.
1
:
C
o
m
p
u
te
cu
r
v
elet
d
o
m
ain
f
e
atu
r
es f
r
o
m
i
n
p
u
t im
a
g
e
T
h
e
s
tep
to
co
m
p
u
te
cu
r
v
elet
d
o
m
ain
f
ea
tu
r
es
is
as
ex
p
lain
ed
in
s
ec
tio
n
3
.
2
.
T
h
e
im
ag
e
u
s
ed
is
th
e
in
p
u
t im
ag
e
w
h
ich
th
e
q
u
ality
is
to
b
e
ass
ess
ed
.
Step
2
.
2
:
C
o
m
p
u
te
im
ag
e
q
u
ality
u
s
in
g
th
e
m
ea
n
a
n
d
s
tan
d
ar
d
d
ev
iatio
n
f
o
r
f
ea
t
u
r
e
n
o
r
m
aliz
atio
n
an
d
r
eg
r
ess
io
n
f
u
n
ctio
n
o
b
tain
ed
f
r
o
m
s
tep
1
.
2
I
n
th
is
s
tep
,
th
e
cu
r
v
elet
d
o
m
ain
f
ea
tu
r
es
wer
e
n
o
r
m
aliz
ed
u
s
in
g
th
e
m
ea
n
,
an
d
s
tan
d
ar
d
d
ev
iatio
n
,
,
o
b
tain
ed
f
r
o
m
Ste
p
1
.
2
.
Nex
t,
th
e
y
wer
e
p
l
u
g
g
ed
in
to
th
e
r
eg
r
ess
io
n
f
u
n
ctio
n
o
b
tain
ed
f
r
o
m
Step
1
.
2
to
c
o
m
p
u
te
th
e
f
in
al
i
m
ag
e
q
u
ality
s
c
o
r
e.
Fig
u
r
e
4
.
Flo
wch
ar
t
o
f
th
e
p
r
e
d
ictin
g
im
ag
e
q
u
ality
4.
E
XP
E
R
I
M
E
N
T
A
L
RE
SUL
T
S AN
D
D
I
SC
USS
I
O
N
Her
e,
t
h
e
NR
-
I
QA
-
C
DI
r
el
y
o
n
c
u
r
v
e
let
f
e
at
u
r
es
(
NR
-
I
QA
-
C
DI
-
C
v
T
)
p
e
r
f
o
r
m
a
n
ce
is
e
v
a
lu
a
te
d
.
T
h
e
ev
al
u
at
io
n
p
r
o
c
ed
u
r
e
,
d
is
c
u
s
s
i
o
n
s
,
a
n
d
co
n
c
l
u
s
io
n
s
was
d
es
c
r
i
b
e
d
.
4
.
1
.
E
v
a
lua
t
io
n
pro
ce
du
re
T
o
e
v
al
u
ate
o
u
r
p
r
o
p
o
s
e
d
,
th
e
s
a
m
e
test
im
ag
e
d
a
ta
b
as
es
u
s
e
d
b
y
NR
-
I
Q
A
-
C
D
I
we
r
e
u
s
e
d
.
O
n
l
y
co
n
t
r
as
t
d
is
to
r
t
ed
im
ag
es
(
1
1
6
,
2
5
0
,
a
n
d
4
0
0
)
w
er
e
c
h
o
s
e
n
f
r
o
m
C
S
I
Q
[
2
2
]
,
T
I
D
2
0
1
3
[
2
3
]
,
an
d
C
I
D
2
0
1
3
[
1
1
]
,
r
es
p
e
cti
v
e
ly
.
M
ea
n
o
p
i
n
i
o
n
s
c
o
r
e
(
MO
S
)
o
r
d
i
f
f
er
en
tia
l
m
ea
n
o
p
i
n
i
o
n
s
c
o
r
e
(
DM
OS
)
a
r
e
u
s
ed
t
o
r
e
p
r
ese
n
t
t
h
e
Su
b
je
cti
v
e
s
co
r
es
.
I
n
o
r
d
er
t
o
ev
al
u
at
e
t
h
e
I
QA
a
cc
u
r
ac
y
,
K
-
f
o
ld
cr
o
s
s
v
al
id
ati
o
n
(
C
V
)
w
a
s
u
s
e
d
t
o
d
ete
r
m
i
n
e
h
o
w
we
ll
t
h
e
I
QA
c
o
u
ld
b
e
g
en
er
ali
ze
d
t
o
i
n
d
e
p
e
n
d
e
n
t
d
a
t
a
g
r
o
u
p
s
w
h
i
le
d
e
cr
ea
s
e
b
ias
,
s
i
n
c
e
r
e
g
r
ess
i
o
n
is
ess
e
n
ti
all
y
a
le
ar
n
i
n
g
al
g
o
r
i
th
m
t
h
at
r
eq
u
i
r
es
t
r
a
in
in
g
.
T
h
r
e
e
d
at
ab
ases
we
r
e
s
p
li
t
r
an
d
o
m
l
y
i
n
t
o
1
0
s
u
b
g
r
o
u
p
s
w
h
e
n
p
e
r
f
o
r
m
in
g
t
h
e
K
-
f
o
l
d
c
r
o
s
s
v
al
id
ati
o
n
.
T
o
t
est
t
h
e
p
r
o
p
o
s
ed
m
e
tr
ic
,
t
h
e
m
et
h
o
d
n
am
e
d
10
-
f
o
l
d
l
ea
v
e
-
o
n
e
-
o
u
t
C
V
w
as
u
s
e
d
.
B
y
u
s
i
n
g
K
ti
m
es
,
t
h
e
r
es
u
lt
o
f
t
h
e
ass
ess
m
e
n
t
we
r
e
av
er
ag
e
d
.
T
o
d
e
cr
ea
s
e
t
h
e
v
a
r
i
ab
i
lit
y
,
r
e
p
e
ate
d
r
o
u
n
d
s
o
f
cr
o
s
s
-
v
ali
d
ati
o
n
(
k
=
2
to
1
0
)
w
er
e
ac
co
m
p
lis
h
e
d
t
h
r
o
u
g
h
v
ar
io
u
s
p
ar
titi
o
n
s
[
8
,
2
4
]
.
T
h
e
a
b
o
v
e
cr
o
s
s
-
v
ali
d
ati
o
n
h
as
b
ee
n
it
er
ate
d
h
u
n
d
r
e
d
ti
m
es
(
t
o
p
r
e
v
en
t
b
ias
)
.
T
a
b
le
s
1
a
n
d
2
s
h
o
w
t
h
e
av
e
r
a
g
e
r
es
u
lts
.
T
o
e
v
al
u
at
e
I
QA
p
e
r
f
o
r
m
a
n
c
e,
t
h
r
e
e
m
et
r
i
c
w
er
e
u
s
e
d
b
etwe
e
n
t
h
e
esti
m
at
ed
o
b
jec
ti
v
e
s
c
o
r
es
a
n
d
th
e
s
u
b
je
cti
v
e
m
ea
n
o
p
i
n
i
o
n
s
co
r
es
(
MO
S
)
.
T
h
e
m
et
r
i
cs
a
r
e
(
1
)
SR
OC
C
,
(
2
)
PLCC
a
n
d
(
3
)
R
MSE
.
T
h
e
e
f
f
ec
t
iv
e
p
e
r
f
o
r
m
a
n
c
e
i
n
c
ase
o
f
c
o
r
r
e
lat
io
n
w
it
h
h
u
m
an
a
wa
r
e
n
ess
i
f
SR
OC
C
~
1
,
PLCC
~
1
a
n
d
R
M
SE~
0
[
8
,
2
4
]
.
4
.
2
.
NR
-
I
Q
A
-
CDI
-
Cv
T
ev
a
lua
t
io
n
T
ab
le
s
1
a
n
d
2
lis
ts
th
e
th
r
ee
-
p
er
f
o
r
m
an
ce
m
etr
ics
u
s
ed
to
e
v
alu
ate
th
e
cu
r
r
en
t
NR
-
I
QA
-
C
DI
an
d
th
e
p
r
o
p
o
s
ed
NR
-
I
QA
-
C
DI
-
C
v
T
with
th
e
th
r
ee
test
im
ag
e
d
at
ab
ases
an
d
k
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
with
k
r
an
g
e
f
r
o
m
2
to
1
0
.
Fig
u
r
e
5
s
h
o
ws
th
e
b
ar
ch
ar
t
g
r
ap
h
c
o
m
p
ar
in
g
th
e
a
v
er
ag
e
v
alu
es
o
f
ea
ch
o
f
th
e
t
h
r
ee
-
p
er
f
o
r
m
an
ce
m
etr
ics
o
f
t
h
e
t
wo
NR
-
I
QA
-
C
DI
f
o
r
ea
c
h
o
f
th
e
th
r
ee
d
atab
ases
.
At
a
g
la
n
ce
,
NR
-
I
QA
-
C
DI
-
C
v
T
o
u
tp
er
f
o
r
m
ed
th
e
cu
r
r
en
t
NR
-
I
QA
-
C
DI
in
all
th
e
th
r
ee
-
p
er
f
o
r
m
an
ce
m
etr
ics u
s
in
g
C
SIQ
an
d
T
I
D
2
0
1
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
3
,
J
u
n
e
2
0
2
1
:
2
5
9
5
-
2603
2600
T
ab
le
1
.
T
h
e
PLCC
,
SR
OC
C
an
d
R
MSE
ac
r
o
s
s
1
0
0
tr
ain
-
test
iter
atio
n
v
ia
NR
-
I
QA
-
C
DI
[
1
2
]
(
f
iv
e
g
lo
b
al
NSS f
ea
tu
r
es)
k
C
S
I
Q
TI
D
2
0
1
3
C
I
D
2
0
1
3
P
LC
C
S
R
O
C
C
R
M
S
E
P
LC
C
S
R
O
C
C
R
M
S
E
P
LC
C
S
R
O
C
C
R
M
S
E
2
0
.
6
8
4
9
0
.
7
0
9
3
0
.
1
2
3
3
0
.
5
1
1
1
0
.
4
4
5
0
0
.
8
6
0
8
0
.
8
6
8
5
0
.
8
7
1
1
0
.
3
0
9
0
3
0
.
7
5
1
8
0
.
7
6
1
9
0
.
1
1
2
1
0
.
5
6
9
8
0
.
4
8
8
6
0
.
8
1
1
2
0
.
8
8
4
4
0
.
8
8
7
4
0
.
2
9
4
5
4
0
.
7
3
8
9
0
.
7
5
2
2
0
.
1
1
3
3
0
.
5
7
8
5
0
.
4
9
9
2
0
.
8
2
6
6
0
.
8
7
9
9
0
.
8
7
9
7
0
.
2
9
7
0
5
0
.
7
9
2
3
0
.
7
8
1
8
0
.
1
0
9
2
0
.
5
7
6
2
0
.
4
9
5
7
0
.
8
0
9
1
0
.
8
7
3
8
0
.
8
7
4
3
0
.
3
0
3
5
6
0
.
7
6
0
1
0
.
7
6
8
3
0
.
1
0
9
7
0
.
5
9
1
1
0
.
5
1
7
3
0
.
7
9
2
3
0
.
8
8
0
4
0
.
8
8
2
9
0
.
2
9
7
9
7
0
.
7
8
4
0
0
.
7
7
7
5
0
.
1
0
4
5
0
.
5
8
1
2
0
.
5
0
2
4
0
.
8
1
4
1
0
.
8
8
5
7
0
.
8
8
1
6
0
.
2
9
1
7
8
0
.
7
7
2
4
0
.
7
8
4
6
0
.
1
0
7
0
0
.
5
9
0
0
0
.
5
0
2
8
0
.
8
0
8
9
0
.
8
8
1
8
0
.
8
8
1
7
0
.
2
9
6
1
9
0
.
7
7
8
8
0
.
7
8
0
3
0
.
1
0
5
4
0
.
5
8
7
0
0
.
4
9
3
2
0
.
8
0
6
4
0
.
8
8
4
4
0
.
8
8
3
3
0
.
2
9
6
2
10
0
.
7
9
7
8
0
.
7
7
5
1
0
.
1
0
7
6
0
.
5
7
9
9
0
.
5
0
4
8
0
.
8
0
1
9
0
.
8
7
8
9
0
.
8
7
6
5
0
.
2
9
7
4
Av
0
.
7
6
2
3
0
.
7
6
5
7
0
.
1
1
0
2
0
.
5
7
3
9
0
.
4
9
4
3
0
.
8
1
4
6
0
.
8
7
9
7
0
.
8
7
9
8
0
.
2
9
8
1
T
ab
le
2
.
T
h
e
av
er
a
g
e
PLCC
,
S
R
OC
C
an
d
R
MSE
ac
r
o
s
s
1
0
0
tr
ain
-
test
iter
atio
n
v
ia
th
e
NR
-
I
QA
-
C
DI
-
C
v
T
(
5
f
ea
tu
r
es)
k
C
S
I
Q
TI
D
2
0
1
3
C
I
D
2
0
1
3
P
LC
C
S
R
O
C
C
R
M
S
E
P
LC
C
S
R
O
C
C
R
M
S
E
P
LC
C
S
R
O
C
C
R
M
S
E
2
0
.
8
0
7
8
0
.
8
0
7
6
0
.
1
0
2
4
0
.
6
6
8
1
0
.
5
8
8
7
0
.
7
3
6
0
0
.
8
5
2
2
0
.
8
5
6
6
0
.
3
2
8
0
3
0
.
8
3
8
4
0
.
8
3
5
4
0
.
0
9
6
2
0
.
6
8
7
5
0
.
6
2
5
9
0
.
7
1
4
4
0
.
8
8
4
8
0
.
8
8
8
2
0
.
2
9
3
6
4
0
.
8
3
0
1
0
.
8
3
5
8
0
.
0
9
5
0
0
.
6
8
8
1
0
.
6
2
1
1
0
.
7
2
1
3
0
.
8
7
6
9
0
.
8
8
2
3
0
.
3
0
2
3
5
0
.
8
5
4
7
0
.
8
4
1
7
0
.
0
9
0
3
0
.
6
8
8
2
0
.
6
2
2
0
0
.
7
2
4
2
0
.
8
8
0
8
0
.
8
8
6
6
0
.
2
9
6
5
6
0
.
8
4
8
7
0
.
8
3
9
6
0
.
0
9
4
3
0
.
6
9
5
5
0
.
6
3
0
9
0
.
7
0
9
5
0
.
8
8
6
3
0
.
8
8
9
6
0
.
2
9
0
8
7
0
.
8
4
7
5
0
.
8
3
5
1
0
.
0
9
1
0
0
.
7
0
4
4
0
.
6
2
9
5
0
.
7
0
7
3
0
.
8
8
6
0
0
.
8
8
8
0
0
.
2
9
2
2
8
0
.
8
5
6
4
0
.
8
4
7
3
0
.
0
8
9
3
0
.
6
9
1
4
0
.
6
2
6
8
0
.
7
1
7
7
0
.
8
8
5
0
0
.
8
8
9
0
0
.
2
9
4
8
9
0
.
8
5
3
5
0
.
8
3
9
6
0
.
0
9
2
2
0
.
6
9
9
1
0
.
6
3
1
3
0
.
7
0
2
4
0
.
8
9
0
0
0
.
8
9
2
6
0
.
2
8
2
1
10
0
.
8
6
9
3
0
.
8
5
3
6
0
.
0
8
7
0
0
.
6
9
2
8
0
.
6
1
6
4
0
.
7
1
9
7
0
.
8
8
5
1
0
.
8
9
5
4
0
.
2
8
5
6
Av
0
.
8
4
5
1
0
.
8
3
7
3
0
.
0
9
3
1
0
.
6
9
0
6
0
.
6
2
1
4
0
.
7
1
6
9
0
.
8
8
0
8
0
.
8
8
5
4
0
.
2
9
6
2
(
a)
(
b
)
(
c)
Fig
u
r
e
5
.
C
o
m
p
a
r
is
o
n
o
f
SR
OC
C
,
PL
C
C
,
an
d
R
MSE
o
f
NR
-
I
QA
-
C
DI
,
NR
-
I
QA
-
C
DI
-
C
v
T
on
;
(
a)
C
SIQ
d
atab
ase
,
(
b
)
T
I
D2
0
1
3
d
ata
b
ase
,
(
c)
C
I
D2
0
1
3
d
atab
ase
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
-
8
7
0
8
C
o
n
tr
a
s
t
-
d
is
to
r
ted
ima
g
e
q
u
a
l
ity
a
s
s
es
s
men
t b
a
s
ed
o
n
c
u
r
ve
let
d
o
ma
in
fea
t
u
r
es
(
I
s
ma
il Ta
h
a
A
h
med
)
2601
4
.
2
.
1
.
Sta
t
is
t
ica
l per
f
o
r
m
a
nc
e
a
na
ly
s
is
a.
P
er
ce
nta
g
e
o
f
diff
er
ence
E
ac
h
k
in
ea
c
h
o
f
th
e
d
atab
as
es,
th
e
d
if
f
er
en
ce
b
etwe
en
th
e
two
-
p
er
f
o
r
m
a
n
ce
m
e
t
r
ics
ar
e
ca
lcu
lated
by
(
1
0
)
:
=
−
(
1
0
)
W
h
er
e
ci
co
r
r
esp
o
n
d
s
to
th
e
f
ir
s
t
m
etr
ic
v
alu
es
with
o
u
t
u
s
in
g
f
ea
tu
r
es
in
cu
r
v
elet
d
o
m
ain
an
d
cv
tci
co
r
r
esp
o
n
d
s
to
th
e
s
ec
o
n
d
m
et
r
ic
v
alu
es b
y
u
s
in
g
f
ea
tu
r
es in
cu
r
v
elet
d
o
m
ain
.
T
h
en
th
e
av
e
r
ag
e
p
er
ce
n
tag
e
o
f
d
if
f
er
en
ce
s
is
ca
lcu
lated
f
o
r
all
th
e
k
v
alu
es
an
d
d
ata
b
ase
s
.
T
h
e
p
er
ce
n
tag
e
is
ca
lcu
lat
ed
b
y
d
iv
id
i
n
g
th
e
p
er
f
o
r
m
an
ce
d
i
f
f
er
en
ce
b
y
t
h
e
ab
s
o
lu
te
f
ir
s
t m
etr
ic
v
alu
e
o
f
ci.
=
1
(
∑
=
1
/
(
)
)
(
)
W
h
er
e
n
co
r
r
esp
o
n
d
s
to
th
e
to
tal
n
u
m
b
er
k
ac
r
o
s
s
all
d
atab
ases
.
T
h
e
ab
s
o
lu
te
v
alu
e
is
u
s
ed
to
k
ee
p
th
e
p
er
ce
n
tag
e
(
in
c
r
em
en
t
o
r
d
ec
r
em
en
t)
s
ig
n
o
f
d
if
f
e
r
en
ce
in
p
er
f
o
r
m
a
n
ce
.
T
h
e
p
er
ce
n
tag
e
o
f
d
if
f
er
en
ce
is
s
h
o
wn
in
T
ab
le
3
.
T
ab
le
3
.
Per
ce
n
ta
g
e
d
if
f
e
r
en
ce
r
esu
lts
f
o
r
NR
-
I
QA
-
C
DI
-
C
v
T
-
NR
-
I
QA
-
C
DI
I
mag
e
D
a
t
a
b
a
se
P
LC
C
S
R
O
C
C
R
M
S
E
TI
D
2
0
1
3
2
0
.
4
9
%
2
5
.
8
1
%
-
1
1
.
9
6
%
C
I
D
2
0
1
3
0
.
1
1
%
0
.
6
3
%
-
0
.
6
7
%
C
S
I
Q
1
0
.
9
9
%
9
.
4
1
%
-
1
5
.
5
3
%
A
l
l
D
a
t
a
b
a
s
e
s
1
0
.
5
3
%
1
1
.
9
5
%
-
9
.
3
9
%
b.
Sta
t
is
t
ica
l
s
ig
nifica
nce
A
p
air
ed
T
-
test
h
y
p
o
th
esis
test
[
2
5
,
2
6
]
is
im
p
lem
en
ted
t
o
t
h
e
p
er
f
o
r
m
a
n
ce
m
etr
ic
v
alu
e
ca
lcu
lated
b
ef
o
r
e
an
d
af
ter
ad
d
in
g
c
u
r
v
e
let
f
ea
tu
r
es
to
p
r
o
d
u
ce
th
e
p
-
v
alu
e
as
s
h
o
wn
in
T
ab
le
4
.
G
en
er
ally
,
p
-
v
alu
e
o
f
less
th
an
0
.
0
5
im
p
lies
th
at
a
s
ig
n
if
ican
t
d
if
f
er
en
ce
a
p
p
ea
r
with
in
th
e
v
alu
es.
T
ab
les
4
an
d
5
d
is
p
lay
s
t
h
e
P
-
v
alu
es o
f
th
e
p
air
ed
T
-
test
s
f
o
r
ad
d
in
g
c
u
r
v
elet
f
ea
t
u
r
e.
T
ab
le
4
.
P
-
v
al
u
es
o
f
d
if
f
er
en
c
es b
etwe
en
NR
-
I
QA
-
C
DI
&
NR
-
I
QA
-
C
DI
-
C
v
T
I
mag
e
D
a
t
a
b
a
se
PLCC
S
RO
C
C
RM
S
E
TI
D
2
0
1
3
1
.
2
4
×
10
-
08
2
.
5
4
×
10
-
10
1
.
4
8
×
10
-
08
C
I
D
2
0
1
3
3
.
4
1
×
10
-
01
5
.
4
3
×
10
-
02
2
.
8
8
×
10
-
01
C
S
I
Q
4
.
3
4
×
10
-
07
1
.
2
5
×
10
-
07
4
.
3
2
×
10
-
08
A
l
l
D
a
t
a
b
a
s
e
s
1
.
8
5
×
10
-
07
1
.
4
9
×
10
-
07
4
.
7
3
×
10
-
05
T
ab
le
5
.
P
-
v
al
u
es
o
f
d
if
f
er
en
c
es b
etwe
en
NR
-
I
QA
-
C
DI
&
NR
-
I
QA
-
C
DI
-
C
v
T
I
f
p
-
v
alu
e
≤
0
.
0
5
: th
e
o
b
s
er
v
e
d
d
if
f
er
en
ce
is
“sig
n
if
ican
t”
I
mag
e
D
a
t
a
b
a
se
P
LC
C
S
R
O
C
C
R
M
S
E
TI
D
2
0
1
3
S
i
g
n
i
f
i
c
a
n
t
S
i
g
n
i
f
i
c
a
n
t
S
i
g
n
i
f
i
c
a
n
t
C
I
D
2
0
1
3
I
n
si
g
n
i
f
i
c
a
n
t
I
n
si
g
n
i
f
i
c
a
n
t
I
n
si
g
n
i
f
i
c
a
n
t
C
S
I
Q
S
i
g
n
i
f
i
c
a
n
t
S
i
g
n
i
f
i
c
a
n
t
S
i
g
n
i
f
i
c
a
n
t
A
l
l
D
a
t
a
b
a
s
e
s
S
i
g
n
i
f
i
c
a
n
t
S
i
g
n
i
f
i
c
a
n
t
S
i
g
n
i
f
i
c
a
n
t
4.
2
.
2.
Dis
cus
s
io
n
T
h
e
d
is
cu
s
s
io
n
o
n
th
e
r
esu
lts
in
T
ab
les 3
,
4
,
an
d
5
ar
e
as
:
−
T
a
b
le
3
i
n
d
ic
ates
t
h
a
t
t
h
e
f
i
n
d
i
n
g
s
u
s
in
g
t
h
e
T
I
D
2
0
1
3
h
av
e
i
m
p
r
o
v
e
d
,
t
h
a
t
w
as
o
u
r
m
ai
n
ai
m
f
o
r
en
h
a
n
ce
m
e
n
t
.
T
h
e
r
e
was
a
s
ig
n
i
f
ic
a
n
t
i
n
c
r
e
ase
i
n
P
L
C
C
a
n
d
SR
OC
C
b
y
2
0
.
4
9
%
a
n
d
2
5
.
8
1
%
,
r
es
p
e
cti
v
e
ly
.
T
h
e
R
MSE
d
e
cr
e
ase
d
n
o
ti
ce
a
b
l
y
b
y
1
1
.
9
6
%
.
T
h
e
3
p
-
v
al
u
es
f
o
r
T
I
D2
0
1
3
w
e
r
e
b
el
o
w
t
h
a
n
0
.
0
5
,
m
ea
n
i
n
g
s
i
g
n
if
i
ca
n
t
d
i
f
f
e
r
e
n
c
es
in
t
h
o
s
e
t
h
r
ee
p
e
r
f
o
r
m
a
n
c
e
m
e
asu
r
es
as
s
h
o
wn
i
n
T
ab
l
e
s
4
an
d
5
.
−
Fo
r
C
I
D2
0
1
3
,
PLCC
a
n
d
SR
OC
C
in
cr
ea
s
ed
v
e
r
y
m
a
r
g
in
al
ly
b
y
0
.
1
1
%
a
n
d
0
.
6
3
%
,
r
es
p
ec
ti
v
el
y
.
T
h
e
R
MSE
als
o
d
ec
r
e
ase
d
v
e
r
y
m
a
r
g
in
all
y
b
y
0
.
6
7
%
.
T
h
e
3
p
-
v
al
u
es
f
o
r
C
I
D
2
0
1
3
i
n
d
icat
e
t
h
at
t
h
e
d
i
f
f
er
e
n
ce
s
i
n
t
h
ese
t
h
r
e
e
m
e
as
u
r
es
we
r
e
n
o
t
s
t
atis
tic
all
y
s
ig
n
i
f
ic
a
n
t
as s
h
o
w
n
in
T
a
b
le
s
4
an
d
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
3
,
J
u
n
e
2
0
2
1
:
2
5
9
5
-
2603
2602
−
Fo
r
th
e
C
SIQ
,
t
h
e
r
e
w
as
a
m
o
d
e
r
a
te
i
n
c
r
e
ase
i
n
PLCC
a
n
d
SR
OC
C
b
y
1
0
.
9
9
%
a
n
d
9
.
4
1
%
,
r
esp
ec
t
iv
el
y
.
T
h
e
R
MSE
d
e
c
r
ea
s
e
d
m
o
d
e
r
at
ely
b
y
1
5
.
5
3
%
.
T
h
e
3
p
-
v
al
u
es
f
o
r
C
S
I
Q
b
el
o
w
t
h
a
n
0
.
0
5
,
in
d
ic
at
e
th
at
t
h
e
d
i
f
f
er
e
n
ce
s
i
n
t
h
ese
t
h
r
e
e
m
e
as
u
r
es
we
r
e
s
ta
tis
ti
ca
ll
y
s
i
g
n
i
f
i
ca
n
t
as
s
h
o
w
n
i
n
T
a
b
l
es
4
a
n
d
5
.
Fo
r
th
e
av
er
ag
e
r
esu
lts
o
f
th
e
th
r
ee
d
atab
ases
,
th
er
e
wer
e
m
o
d
er
ate
in
cr
ea
s
e
in
PLCC
an
d
SR
OC
C
by
1
0
.
5
3
%
an
d
1
1
.
9
5
%
,
r
esp
e
ctiv
ely
.
T
h
e
R
M
SE
d
e
cr
ea
s
ed
m
o
d
er
ately
b
y
9
.
3
9
%.
T
h
e
th
r
ee
p
-
v
alu
es
f
o
r
all
d
atab
ases
b
elo
w
th
an
0
.
0
5
,
in
d
icate
th
at
th
e
d
if
f
er
e
n
ce
s
in
t
h
ese
th
r
ee
p
er
f
o
r
m
an
c
e
m
atr
ic
es
wer
e
s
tatis
tically
s
ig
n
if
ican
t
as sh
o
wn
in
T
ab
les
4
an
d
5
.
5.
CO
NCLU
SI
O
N
I
n
th
is
p
ap
er
,
th
e
ex
is
tin
g
NR
-
I
QA
-
C
DI
was
en
h
an
c
ed
u
s
in
g
cu
r
v
elet
d
o
m
ain
f
ea
tu
r
es.
I
n
ch
ar
ac
ter
izin
g
a
g
o
o
d
co
n
t
r
ast
im
ag
e
a
n
d
c
o
n
tr
ast
-
d
is
to
r
ted
im
ag
e
at
v
ar
io
u
s
s
ca
les,
d
is
tr
i
b
u
tio
n
s
o
f
cu
r
v
elet
co
ef
f
icien
ts
wer
e
f
o
u
n
d
to
b
e
ac
cu
r
ate.
T
h
e
f
iv
e
cu
r
v
elet
d
o
m
ain
f
ea
tu
r
es
p
r
o
p
o
s
ed
wer
e
d
er
iv
ed
f
r
o
m
th
e
d
is
tr
ib
u
tio
n
o
f
th
e
cu
r
v
elet
co
ef
f
icien
ts
ac
r
o
s
s
th
e
two
co
ar
s
est
s
ca
le
s
an
d
th
e
f
in
est
s
ca
le.
T
h
e
p
e
r
f
o
r
m
an
c
e
ev
alu
atio
n
i
n
d
icate
d
t
h
at
N
R
-
I
QA
-
C
DI
-
C
v
T
s
ig
n
if
ican
tly
o
u
t
p
er
f
o
r
m
ed
th
e
e
x
is
tin
g
NR
-
I
QA
-
C
DI
in
d
atab
ase
T
I
D2
0
1
3
an
d
C
SIQ
,
wh
ic
h
wer
e
th
e
p
r
im
ar
y
tar
g
et
f
o
r
im
p
r
o
v
em
en
t
o
f
th
is
wo
r
k
,
alth
o
u
g
h
th
e
r
e
wasn
’
t m
u
ch
p
e
r
f
o
r
m
an
ce
d
i
f
f
er
en
ce
in
d
ata
b
ase
C
I
D2
0
1
3
.
ACK
NO
WL
E
DG
E
M
E
NT
S
T
h
is
r
esear
ch
is
s
u
p
p
o
r
ted
b
y
UNI
T
E
N
R
esear
ch
Fu
n
d
f
o
r
J
o
u
r
n
al
Pu
b
licatio
n
2
0
2
0
.
RE
F
E
R
E
NC
E
S
[1
]
N.
Th
a
k
u
r
a
n
d
S
.
De
v
i
,
“
A
n
e
w
m
e
th
o
d
fo
r
c
o
lo
r
ima
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
Co
mp
u
ter
Ap
p
li
c
a
ti
o
n
s
,
v
o
l
.
1
5
,
n
o
.
2
,
p
p
.
1
0
-
1
7
,
2
0
1
1
.
[2
]
I.
T.
Ah
m
e
d
C.
S
.
De
r,
a
n
d
B.
T.
Ha
m
m
a
d
,
“
A
S
u
rv
e
y
o
f
Re
c
e
n
t
Ap
p
r
o
a
c
h
e
s
o
n
N
o
-
Re
fe
re
n
c
e
Im
a
g
e
Qu
a
li
ty
As
se
ss
m
e
n
t
with
M
u
l
ti
sc
a
le
Ge
o
m
e
tri
c
An
a
ly
sis
Tran
sfo
rm
s,”
In
t
e
rn
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
S
c
ien
ti
fi
c
&
En
g
in
e
e
rin
g
Res
e
a
rc
h
,
v
o
l.
7
,
n
o
.
1
2
,
p
p
.
1
1
4
6
-
1
1
5
6
,
2
0
1
6
.
[3
]
S
.
G
.
G
.
M
ra
k
a
n
d
o
t
h
e
rs,
“
Re
li
a
b
il
it
y
o
f
o
b
jec
ti
v
e
p
ictu
re
q
u
a
li
t
y
m
e
a
su
re
s,”
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
in
e
e
rin
g
,
v
o
l.
5
5
,
n
o
.
1
-
2
,
p
p
.
3
-
1
0
,
2
0
0
4
.
[4
]
I.
T.
Ah
m
e
d
,
C
.
S
.
De
r
a
n
d
B.
T.
Ha
m
m
a
d
,
“
Re
c
e
n
t
a
p
p
r
o
a
c
h
e
s
o
n
n
o
-
re
fe
re
n
c
e
ima
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t
fo
r
c
o
n
tras
t
d
ist
o
rti
o
n
ima
g
e
s
with
m
u
lt
isc
a
le
g
e
o
m
e
tri
c
a
n
a
ly
sis
tra
n
sfo
rm
s:
a
su
rv
e
y
,
”
J
o
u
r
n
a
l
o
f
T
h
e
o
re
ti
c
a
l
a
n
d
Ap
p
li
e
d
In
fo
rm
a
t
io
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l.
9
5
,
n
o
.
3
,
p
p
.
5
6
1
-
5
6
9
,
2
0
1
7
.
[5
]
S
.
TM
a
n
d
K.
B
.
Ra
m
e
sh
,
“
An
e
fficie
n
t
c
o
m
p
u
tatio
n
a
l
a
p
p
ro
a
c
h
t
o
b
a
lan
c
e
t
h
e
trad
e
-
o
ff
b
e
twe
e
n
i
m
a
g
e
fo
re
n
sic
s
a
n
d
p
e
r
c
e
p
t
u
a
l
ima
g
e
q
u
a
li
ty
,
”
I
n
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
(I
J
ECE
)
,
v
o
l.
9
,
n
o
.
5
,
p
p
.
3
4
7
4
-
3
4
7
9
,
2
0
1
9
.
[6
]
T.
M
.
Ku
s
u
m
a
,
R.
Ra
h
m
a
n
to
a
n
d
E.
Ha
ry
a
tmi,
“
Ad
a
p
ti
v
e
p
o
we
r
li
n
k
a
d
a
p
tati
o
n
o
n
DV
B
-
T
sy
ste
m
b
a
se
d
o
n
p
ict
u
re
q
u
a
li
t
y
fe
e
d
b
a
c
k
,
”
I
n
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
in
e
e
rin
g
(IJ
ECE
)
,
v
o
l.
9
,
n
o
.
4
,
p
p
.
3
1
2
1
-
3
1
2
9
,
2
0
1
9
.
[7
]
B.
Bh
a
t
k
a
lk
a
r,
A.
Jo
sh
i
,
S
.
P
r
a
b
h
u
,
a
n
d
S
.
Bh
a
n
d
a
ry
,
“
Au
to
m
a
ted
fu
n
d
u
s
ima
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t
a
n
d
se
g
m
e
n
tatio
n
o
f
o
p
ti
c
d
isc
u
si
n
g
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
s
,
”
I
n
ter
n
a
ti
o
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
i
n
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
1
0
,
n
o
.
1
,
p
p
.
8
1
6
-
8
2
7
,
2
0
2
0
.
[8
]
I
.
T.
Ah
m
e
d
,
C
.
S
.
De
r,
N.
Ja
m
il
a
n
d
M
.
A.
M
o
h
a
m
e
d
,
“
Im
p
ro
v
e
o
f
c
o
n
tras
t
-
d
ist
o
rted
ima
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t
b
a
se
d
o
n
c
o
n
v
o
lu
ti
o
n
a
l
n
e
u
ra
l
n
e
two
rk
s
,
”
In
ter
n
a
ti
o
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
(IJ
ECE
)
,
v
o
l.
9
,
n
o
.
6
,
p
p
.
5
6
0
4
-
5
6
1
4
,
2
0
1
9
.
[9
]
R.
C.
G
o
n
z
a
lez
a
n
d
R.
E.
W
o
o
d
s,
“
Dig
it
a
l
ima
g
e
p
r
o
c
e
ss
in
g
,
”
U
p
p
e
r S
a
d
d
le R
ive
r,
NJ
:
Pre
n
ti
c
e
Ha
l
l
,
2
0
1
2
.
[1
0
]
T.
Aric
i,
S
.
Di
k
b
a
s
a
n
d
Y.
Alt
u
n
b
a
sa
k
,
“
A
h
ist
o
g
ra
m
m
o
d
ifi
c
a
ti
o
n
fra
m
e
wo
rk
a
n
d
it
s
a
p
p
li
c
a
ti
o
n
f
o
r
ima
g
e
c
o
n
tras
t
e
n
h
a
n
c
e
m
e
n
t,
”
IEE
E
T
ra
n
s
a
c
ti
o
n
s o
n
Im
a
g
e
Pr
o
c
e
ss
in
g
,
v
o
l
.
1
8
,
n
o
.
9
,
p
p
.
1
9
2
1
-
1
9
3
5
,
2
0
0
9
.
[1
1
]
K.
G
u
,
G
.
Zh
a
i,
X.
Ya
n
g
,
W.
Z
h
a
n
g
a
n
d
M
.
Li
u
,
“
S
u
b
jec
ti
v
e
a
n
d
o
b
jec
ti
v
e
q
u
a
li
t
y
a
ss
e
ss
m
e
n
t
fo
r
ima
g
e
s
wit
h
c
o
n
tras
t
c
h
a
n
g
e
,
”
2
0
1
3
IEE
E
In
te
rn
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Ima
g
e
P
ro
c
e
ss
in
g
,
M
e
l
b
o
u
rn
e
,
VIC,
2
0
1
3
,
p
p
.
3
8
3
-
3
8
7
.
[1
2
]
Y.
F
a
n
g
,
K.
M
a
,
Z.
Wan
g
,
W.
L
in
,
Z.
F
a
n
g
a
n
d
G
.
Zh
a
i
,
“
No
-
re
fe
re
n
c
e
q
u
a
li
t
y
a
ss
e
ss
m
e
n
t
o
f
c
o
n
tras
t
-
d
ist
o
rte
d
ima
g
e
s b
a
se
d
o
n
n
a
tu
ra
l
sc
e
n
e
sta
ti
stics
,
”
IEE
E
S
ig
n
a
l
Pro
c
e
ss
in
g
L
e
tt
e
rs
,
v
o
l
.
2
2
,
n
o
.
7
,
p
p
.
8
3
8
-
842
,
2
0
1
5
.
[1
3
]
H.
F
ü
h
r
,
L.
De
m
a
re
t
a
n
d
F
.
F
ri
e
d
rich
,
“
Be
y
o
n
d
wa
v
e
lets:
Ne
w
ima
g
e
re
p
re
se
n
tatio
n
p
a
ra
d
ig
m
s
,
”
Do
c
.
ima
g
e
c
o
mp
re
ss
io
n
,
vol
.
7
,
p
p
.
1
7
9
-
2
0
6
,
2
0
0
6
.
[1
4
]
E.
J.
Ca
n
d
è
s
a
n
d
D.
L
.
Do
n
o
h
o
,
“
Ne
w
ti
g
h
t
fra
m
e
s
o
f
c
u
rv
e
l
e
ts
a
n
d
o
p
t
ima
l
re
p
re
se
n
tatio
n
s
o
f
o
b
jec
ts
with
p
iec
e
wise
C2
sin
g
u
lariti
e
s,”
Co
m
mu
n
ica
t
io
n
s o
n
Pu
re
a
n
d
A
p
p
l
ied
M
a
t
h
e
ma
ti
c
s
,
v
o
l
.
5
7
,
n
o
.
2
,
p
p
.
2
1
9
-
2
6
6
,
2
0
0
4
.
[1
5
]
E.
Ca
n
d
e
s,
L.
De
m
a
n
e
t,
D.
Do
n
o
h
o
a
n
d
L.
Yin
g
,
“
F
a
st
d
isc
re
te
c
u
rv
e
let
tran
sfo
rm
s,”
S
IA
M
J
o
u
r
n
a
l
o
n
M
u
lt
isc
a
le
M
o
d
e
li
n
g
a
n
d
S
im
u
la
ti
o
n
,
v
o
l.
5
,
n
o
.
3
,
p
p
.
8
6
1
-
8
9
9
,
2
0
0
6
.
[1
6
]
J.
S
h
e
n
,
Q.
Li
a
n
d
G
.
Erl
e
b
a
c
h
e
r,
“
Cu
rv
e
let
b
a
se
d
n
o
-
re
fe
re
n
c
e
o
b
j
e
c
ti
v
e
ima
g
e
Qu
a
li
ty
As
se
ss
m
e
n
t
,”
2
0
0
9
Pi
c
tu
re
Co
d
i
n
g
S
y
mp
o
si
u
m,
C
h
ica
g
o
,
IL,
2
0
0
9
,
p
p
.
1
-
4.
[1
7
]
L.
Li
u
,
H
.
Do
n
g
,
H.
Hu
a
n
g
a
n
d
A.
C.
Bo
v
i
k
,
“
No
-
re
fe
re
n
c
e
ima
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t
in
c
u
rv
e
let
d
o
m
a
in
,
”
S
ig
n
a
l
Pro
c
e
ss
in
g
:
Ima
g
e
Co
mm
u
n
ica
t
i
o
n
,
v
o
l.
2
9
,
n
o
.
4
,
p
p
.
4
9
4
-
5
0
5
,
2
0
1
4
.
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
-
8
7
0
8
C
o
n
tr
a
s
t
-
d
is
to
r
ted
ima
g
e
q
u
a
l
ity
a
s
s
es
s
men
t b
a
s
ed
o
n
c
u
r
ve
let
d
o
ma
in
fea
t
u
r
es
(
I
s
ma
il Ta
h
a
A
h
med
)
2603
[1
8
]
J.
S
h
e
n
,
Q.
Li
,
a
n
d
G
.
Erl
e
b
a
c
h
e
r,
“
Hy
b
ri
d
n
o
-
re
fe
re
n
c
e
n
a
tu
r
a
l
ima
g
e
q
u
a
li
t
y
a
ss
e
ss
m
e
n
t
o
f
n
o
is
y
,
b
lu
rr
y
,
JPE
G
2
0
0
0
,
a
n
d
JPE
G
ima
g
e
s,”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Ima
g
e
Pro
c
e
ss
in
g
,
v
o
l.
2
0
,
n
o
.
8
,
p
p
.
2
0
8
9
-
2
0
9
8
,
2
0
1
1
.
[1
9
]
I.
T.
A.
Ah
m
e
d
,
C.
S
.
De
r
Ch
e
n
a
n
d
B.
T.
H.
Ha
m
m
a
d
,
“
Im
p
a
c
t
o
f
Co
n
tras
t
-
Disto
rte
d
I
m
a
g
e
o
n
Cu
r
v
e
let
Co
e
fficie
n
ts,”
2
0
1
8
1
st A
n
n
u
a
l
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
I
n
fo
rm
a
ti
o
n
a
n
d
S
c
ien
c
e
s (A
iCIS
)
,
F
a
ll
u
j
a
h
,
Ira
q
,
2
0
1
8
,
p
p
.
2
8
-
32
.
[2
0
]
T.
G
u
h
a
a
n
d
Q.
M
.
J.
W
u
,
“
Cu
r
v
e
let
b
a
se
d
fe
a
tu
re
e
x
trac
ti
o
n
,
”
i
n
F
a
c
e
Rec
o
g
n
i
ti
o
n
,
I
n
T
e
c
h
,
2
0
1
0
.
[2
1
]
C.
Ch
a
n
g
a
n
d
C.
L
in
,
“
LIBS
VM:
a
Li
b
ra
ry
f
o
r
S
u
p
p
o
r
t
Ve
c
to
r
M
a
c
h
in
e
s (Ve
rsio
n
2
.
3
),
”
L
IBS
V
M
,
p
p
.
1
-
39,
2
0
0
1
.
[2
2
]
E.
C.
Lars
o
n
a
n
d
D.
M
.
C
h
a
n
d
le
r,
“
M
o
st
a
p
p
a
re
n
t
d
ist
o
rti
o
n
:
f
u
ll
-
re
fe
re
n
c
e
ima
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t
a
n
d
t
h
e
ro
le
o
f
stra
teg
y
,
"
J
o
u
rn
a
l
o
f
El
e
c
tro
n
i
c
Ima
g
in
g
,
v
o
l
.
19
,
n
o
.
1
,
2
0
1
0
.
[2
3
]
N.
P
o
n
o
m
a
re
n
k
o
,
e
t
a
l.
,
“
Co
l
o
r
ima
g
e
d
a
tab
a
se
TID2
0
1
3
:
P
e
c
u
li
a
rit
ies
a
n
d
p
re
li
m
i
n
a
ry
re
su
lt
s,”
Eu
ro
p
e
a
n
W
o
rk
sh
o
p
o
n
Vi
su
a
l
In
fo
rm
a
t
io
n
Pro
c
e
ss
in
g
(EUV
IP)
,
P
a
ris,
2
0
1
3
,
p
p
.
1
0
6
-
1
1
1
.
[2
4
]
I.
T.
Ah
m
e
d
,
C.
S
.
De
r,
N.
Ja
m
il
,
a
n
d
B.
T.
Ha
m
m
a
d
,
“
An
a
ly
sis
o
f
G
lo
b
a
l
S
p
a
ti
a
l
S
tatisti
c
s
F
e
a
tu
re
s
in
Ex
isti
n
g
Co
n
tras
t
Im
a
g
e
Q
u
a
li
ty
As
se
ss
m
e
n
t
Alg
o
r
it
h
m
,
”
2
0
1
9
7
t
h
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
In
fo
rm
a
ti
o
n
a
n
d
Co
mm
u
n
ica
ti
o
n
T
e
c
h
n
o
l
o
g
y
(IC
o
ICT
)
,
Ku
a
la L
u
m
p
u
r
,
M
a
lay
sia
,
2
0
1
9
,
p
p
.
1
-
4.
[2
5
]
D.
J.
S
h
e
s
k
i
n
,
“
H
a
n
d
b
o
o
k
o
f
p
a
r
a
m
e
t
r
i
c
a
n
d
n
o
n
p
a
r
a
m
e
t
r
i
c
s
t
a
t
i
s
t
i
c
a
l
p
r
o
c
e
d
u
r
e
s
,”
C
h
a
p
m
a
n
&
H
a
l
l
/
C
R
C
,
p
.
1
7
7
6
,
2
0
0
3
.
[2
6
]
Y.
F
u
a
n
d
S
.
Wan
g
,
“
A
n
o
re
fe
re
n
c
e
ima
g
e
q
u
a
li
ty
a
ss
e
ss
m
e
n
t
m
e
tri
c
b
a
se
d
o
n
v
isu
a
l
p
e
rc
e
p
ti
o
n
,
”
Al
g
o
rit
h
ms
,
v
o
l.
9
,
n
o
.
4
,
p
p
.
8
7
-
1
0
8
,
2
0
1
6
.
Evaluation Warning : The document was created with Spire.PDF for Python.