I
A
E
S
I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
r
t
if
ic
ia
l
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
V
ol
.
14
, N
o.
5
,
O
c
to
be
r
2025
, pp.
3563
~
3575
I
S
S
N
:
2252
-
8938
,
D
O
I
:
10.11591/
ij
a
i.
v
14
.i
5
.pp
3563
-
3575
3563
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
ai
.
ia
e
s
c
or
e
.c
om
A
h
yb
r
i
d
st
e
gan
ogr
ap
h
y sc
h
e
m
e
w
i
t
h
r
e
d
u
c
e
d
d
i
f
f
e
r
e
n
c
e
e
xp
an
si
on
a
n
d
p
i
xe
l
-
val
u
e
or
d
e
r
i
n
g
I
K
ad
e
k
A
gu
s
A
r
ie
s
t
a P
u
t
r
a
1
, N
t
iv
u
gu
r
u
z
w
a Je
an
D
e
L
a C
r
oi
x
1,
2
, T
oh
ar
i
A
h
m
ad
1
1
D
e
pa
r
t
m
e
nt
of
I
nf
or
m
a
t
i
c
s
,
F
a
c
ul
t
y of
E
l
e
c
t
r
i
c
a
l
E
ngi
ne
e
r
i
ng a
nd I
nt
e
l
l
i
ge
nt
I
n
f
or
m
a
t
i
c
s
,
I
ns
t
i
t
ut
T
e
knol
ogi
S
e
pul
uh N
ope
m
be
r
,
S
ur
a
ba
ya
, I
ndone
s
i
a
2
C
ol
l
e
ge
of
S
c
i
e
nc
e
a
nd T
e
c
hnol
ogy, U
ni
ve
r
s
i
t
y of
R
w
a
nda
, K
i
g
a
l
i
, R
w
a
nda
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
J
un 12, 2024
R
e
vi
s
e
d
J
un 30,
2025
A
c
c
e
pt
e
d
J
ul
13
,
2025
Stegano
graph
y embe
ds sec
ret me
ssages
into pu
blic
media
while
ensuri
ng the
stego
content
remains
visually
indis
tinguishable
from
the
origina
l.
The
primary
challenge
lies
in
maximizing
embedding
capacity
and
image
quality
without
introducing
noticeable
distortions
.
This
research
proposes
a
novel
reversible
data
hiding
(RDH)
scheme
that
integrates
reduced
diff
erence
expansio
n
(RDE)
with
four
directio
nal
pixel
-
value
ordering
(PVO)
sc
hemes,
horizontal,
vertical,
diagonal
-
right,
and
diagonal
-
left,
to
enhance
emb
edding
efficiency
and
visual
fidelit
y.
Unlike
existin
g
RDH
methods
that
appl
y
RDE
with
fixed
or
limited
PVO
directions,
the
proposed
scheme
dyn
a
mically
selects
the
optimal
PVO
orientation
based
on
pixel
pair
characte
ristics,
effectively
improvi
ng
local
predicti
on
accuracy
and
reducing
embe
dding
-
induced disto
rtion. Previou
s studies hav
e large
ly overlooke
d this relati
onship
between
pixel
pair
selection
and
embedding
performance.
Experi
mental
evaluatio
n
on
medical
images
with
secret
data
sizes
ranging
from
5
kb
to
100
kb
demonstrates
significant
gains
over
recent
PVO
-
based
methods.
The
proposed
method
increases
the
average
embedding
capacity
from
0.8
315
to
0.9781
bit
per
pixel
(bpp)
(a
17.6%
improvement)
and
raises
the
a
verage
peak
signal
-
to
-
noise
ratio
(PSNR)
from
49.44
to
53.40
dB,
re
ducing
distortion by approximately 3.96 dB.
K
e
y
w
o
r
d
s
:
I
nf
or
m
a
ti
on s
e
c
ur
it
y
P
ix
e
l
-
va
lu
e
or
de
r
in
g
R
e
duc
e
d di
f
f
e
r
e
nc
e
e
xpa
n
s
io
n
R
e
ve
r
s
ib
le
da
ta
hi
di
ng
S
te
ga
nogr
a
phy
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
T
oha
r
i
A
hm
a
d
D
e
pa
r
tm
e
nt
of
I
nf
or
m
a
ti
c
s
,
F
a
c
ul
ty
of
E
le
c
tr
ic
a
l
E
ng
in
e
e
r
in
g a
nd I
nt
e
ll
ig
e
nt
I
nf
or
m
a
ti
c
s
I
ns
ti
tu
t
T
e
knol
ogi
S
e
pul
uh N
ope
m
be
r
S
ur
a
ba
ya
, I
ndone
s
ia
E
m
a
il
:
to
ha
r
i@it
s
.a
c
.i
d
1.
I
N
T
R
O
D
U
C
T
I
O
N
S
t
e
ga
n
o
g
r
a
ph
y
is
t
he
a
r
t
o
f
c
o
nc
e
a
l
i
ng
s
e
c
r
e
t
i
nf
o
r
m
a
ti
o
n
w
it
h
i
n
o
t
he
r
da
ta
s
o
th
a
t
t
he
pr
e
s
e
nc
e
o
f
t
he
hi
dd
e
n
m
e
s
s
a
ge
r
e
m
a
i
ns
u
nd
e
te
c
ta
b
le
to
un
i
nt
e
n
de
d
p
a
r
t
i
e
s
[
1
]
.
S
t
e
g
a
n
og
r
a
p
hy
ha
s
be
e
n
us
e
d
s
in
c
e
a
nc
ie
n
t
t
i
m
e
s
,
w
he
r
e
s
e
c
r
e
t
m
e
s
s
a
ge
s
w
e
r
e
h
i
dd
e
n
i
n
v
a
r
i
ous
ph
ys
ic
a
l
f
o
r
m
s
o
f
m
e
d
ia
[
2
]
,
a
n
d
h
a
s
s
in
c
e
e
v
ol
ve
d
i
n
t
he
d
i
gi
t
a
l
e
r
a
t
o
e
m
be
d
d
a
t
a
w
it
h
in
f
i
l
e
s
s
uc
h
a
s
im
a
ge
s
[
3
]
–
[
5
]
,
a
ud
i
o
[
6
]
–
[
8
]
,
vi
de
o
[
9
]
,
[
10
]
,
a
n
d
e
le
c
t
r
oc
a
r
d
io
g
r
a
m
(
E
C
G
)
s
i
gn
a
ls
[
11
]
.
I
ts
a
p
pl
ic
a
ti
ons
s
p
a
n
s
e
c
r
e
t
c
om
m
un
ic
a
t
i
on
[
1
2
]
,
d
i
gi
ta
l
w
a
te
r
m
a
r
k
i
ng
[
13
]
,
a
n
d
c
op
y
r
i
g
ht
p
r
o
te
c
ti
on
[
1
4]
,
e
na
b
li
n
g
th
e
s
e
c
u
r
e
e
m
b
e
d
d
in
g
o
f
i
n
f
o
r
m
a
t
io
n
w
i
th
o
ut
p
e
r
c
e
p
t
ib
l
y
a
l
te
r
in
g
t
he
h
os
t
m
e
d
i
a
.
A
s
i
g
ni
f
ic
a
n
t
s
u
b
f
i
e
l
d
w
i
t
h
in
s
te
ga
no
g
r
a
ph
y
is
r
e
v
e
r
s
i
b
le
d
a
t
a
hi
d
in
g
(
R
D
H
)
,
w
h
ic
h
e
n
a
b
le
s
th
e
e
x
a
c
t
r
e
c
o
ve
r
y
o
f
t
he
h
i
dd
e
n
d
a
t
a
a
n
d
th
e
o
r
ig
in
a
l
c
o
ve
r
m
e
d
ia
a
f
te
r
e
x
t
r
a
c
t
i
on
[
3
]
.
O
ne
o
f
t
he
f
o
un
da
t
io
na
l
te
c
h
n
iq
ue
s
in
R
D
H
is
d
i
f
f
e
r
e
nc
e
e
x
pa
ns
io
n,
w
h
ic
h
e
m
be
ds
i
n
f
o
r
m
a
t
io
n
by
e
x
pa
nd
in
g
th
e
d
i
f
f
e
r
e
nc
e
b
e
t
w
e
e
n
pa
i
r
e
d
p
i
xe
l
v
a
l
ue
s
[
4
]
.
A
l
t
ho
ug
h
di
f
f
e
r
e
n
c
e
e
x
pa
ns
i
o
n
o
f
f
e
r
s
h
ig
h
e
m
be
dd
i
ng
c
a
pa
c
i
ty
,
i
t
te
nds
t
o
in
t
r
od
uc
e
n
o
ti
c
e
a
b
le
d
is
to
r
ti
o
n
,
pa
r
ti
c
u
la
r
l
y
in
s
m
oo
t
h
im
a
g
e
r
e
g
io
ns
[
5
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
5
,
O
c
to
be
r
2025
:
3563
-
3575
3564
T
o
m
i
ti
ga
te
th
is
,
r
e
d
uc
e
d
di
f
f
e
r
e
n
c
e
e
xp
a
n
s
i
o
n
(
R
D
E
)
w
a
s
i
n
t
r
o
d
uc
e
d
t
o
c
ons
t
r
a
i
n
t
he
d
i
f
f
e
r
e
n
c
e
e
x
pa
ns
i
o
n
a
n
d
li
m
i
t
d
is
to
r
t
io
n
[
6
]
.
P
i
xe
l
-
va
lu
e
o
r
de
r
in
g
(
P
V
O
)
,
a
n
ot
he
r
ke
y
c
o
nc
e
p
t
in
R
D
H
,
s
o
r
t
s
p
i
xe
l
g
r
o
ups
ba
s
e
d
on
in
t
e
ns
i
ty
a
n
d
p
r
e
d
ic
ts
d
i
f
f
e
r
e
nc
e
s
m
o
r
e
p
r
e
c
is
e
l
y,
w
h
ic
h
h
e
l
ps
im
p
r
ove
e
m
be
dd
i
ng
a
c
c
u
r
a
c
y
a
n
d
r
e
du
c
e
v
is
ua
l
de
g
r
a
da
t
io
n
[
7
]
.
H
o
w
e
ve
r
,
m
o
s
t
e
xi
s
t
in
g
a
pp
r
oa
c
he
s
a
p
p
ly
P
V
O
i
n
a
f
ix
e
d
o
r
i
e
n
ta
t
io
n,
o
f
te
n
o
ve
r
l
oo
ki
n
g
t
he
i
m
p
a
c
t
of
d
i
r
e
c
t
io
na
l
p
ix
e
l
pa
i
r
s
e
le
c
t
i
on
o
n
o
ve
r
a
l
l
s
te
ga
no
g
r
a
ph
ic
p
e
r
f
o
r
m
a
nc
e
[
8
]
.
O
ne
o
f
t
he
m
a
i
n
c
ha
ll
e
nge
s
in
s
te
ga
no
gr
a
ph
y
is
e
ns
u
r
i
ng
th
a
t
th
e
m
od
i
f
ie
d
o
r
i
gi
na
l
m
e
d
ia
(
s
te
go
)
doe
s
n
ot
r
a
is
e
s
us
pi
c
io
n.
H
ig
h
d
is
t
or
ti
on
in
th
e
o
r
ig
in
a
l
m
e
di
a
due
to
th
e
da
t
a
h
id
i
ng
pr
oc
e
s
s
c
a
n
b
e
a
c
ons
p
ic
u
ous
s
ig
n
th
a
t
th
e
m
e
di
a
is
c
a
r
r
yi
ng
s
e
c
r
e
t
in
f
o
r
m
a
t
io
n
.
W
he
n
i
m
a
ge
s
,
a
ud
io
,
o
r
vi
d
e
o
un
de
r
go
s
ig
n
i
f
ic
a
nt
c
ha
n
ge
s
in
q
ua
li
ty
o
r
vi
s
ua
l
/a
u
di
to
r
y
s
t
r
uc
tu
r
e
,
t
hi
s
c
a
n
a
t
t
r
a
c
t
t
he
a
t
te
n
t
io
n
o
f
t
hi
r
d
pa
r
t
ie
s
obs
e
r
v
in
g
.
T
he
m
a
in
pr
ob
le
m
li
e
s
i
n
t
he
tr
a
de
-
o
f
f
be
twe
e
n
e
m
be
d
di
ng
c
a
pa
c
it
y
a
n
d
d
is
t
o
r
t
io
n
.
I
nc
r
e
a
s
in
g
pa
yl
oa
d
s
iz
e
o
f
te
n
r
e
qu
i
r
e
s
m
o
r
e
a
g
g
r
e
s
s
i
ve
m
od
if
ic
a
ti
ons
to
th
e
m
e
d
ia
,
w
h
ic
h
e
xa
c
e
r
ba
te
s
de
t
e
c
ta
bl
e
a
nom
a
l
ie
s
a
n
d
r
is
ks
e
x
pos
in
g
hi
d
de
n
da
ta
[
1
5]
.
F
ur
th
e
r
m
o
r
e
,
r
e
ve
r
s
ib
il
it
y
i
ts
e
l
f
p
r
e
s
e
nt
s
a
n
ot
he
r
c
ha
ll
e
nge
.
I
n
c
e
r
ta
i
n
c
o
nt
e
xt
s
,
s
uc
h
a
s
m
e
di
c
a
l
i
m
a
gi
n
g,
le
ga
l
d
oc
u
m
e
n
t
s
,
or
di
gi
ta
l
m
e
d
ia
f
o
r
e
ns
ic
s
,
th
e
in
te
gr
it
y
o
f
th
e
o
r
ig
in
a
l
m
e
di
a
m
us
t
be
m
a
i
nt
a
in
e
d
a
f
te
r
c
o
n
f
id
e
n
ti
a
l
da
ta
h
a
s
be
e
n
e
x
tr
a
c
te
d.
B
a
s
e
d
o
n
th
is
r
e
ve
r
s
ib
il
it
y
a
s
pe
c
t,
s
te
ga
n
og
r
a
p
hy
is
di
v
id
e
d
i
nt
o
r
e
v
e
r
s
ib
le
a
nd
i
r
r
e
ve
r
s
ib
l
e
te
c
hni
que
s
.
R
D
H
t
e
c
h
ni
q
ue
s
a
ll
ow
f
u
ll
r
e
c
o
ve
r
y
of
th
e
c
ove
r
m
e
d
ia
a
f
te
r
t
he
s
e
c
r
e
t
d
a
ta
e
xt
r
a
c
ti
on
p
r
oc
e
s
s
,
w
he
r
e
a
s
ir
r
e
ve
r
s
ib
le
te
c
h
ni
que
s
do
no
t
a
l
lo
w
f
ul
l
r
e
c
ove
r
y
o
f
th
e
c
o
ve
r
m
e
d
ia
.
V
a
r
i
o
us
s
te
ga
no
g
r
a
ph
y
s
c
h
e
m
e
s
ha
ve
be
e
n
p
r
o
p
os
e
d
to
a
d
d
r
e
s
s
th
e
i
s
s
u
e
o
f
h
ig
h
d
is
t
o
r
ti
o
n.
O
n
e
s
uc
h
s
c
he
m
e
is
R
D
E
ba
s
e
d
o
n
di
f
f
e
r
e
n
c
e
e
x
pa
ns
io
n,
a
s
in
t
r
od
u
c
e
d
i
n
[
16
]
,
w
h
ic
h
e
m
b
e
ds
s
e
c
r
e
t
d
a
t
a
b
y
e
x
pl
o
it
i
ng
d
i
f
f
e
r
e
n
c
e
s
b
e
t
w
e
e
n
p
a
i
r
s
o
f
p
ix
e
ls
.
H
ow
e
v
e
r
,
di
f
f
e
r
e
nc
e
e
x
pa
ns
i
o
n
c
a
n
le
a
d
to
s
ig
n
i
f
i
c
a
nt
d
is
t
o
r
t
i
on
w
he
n
la
r
ge
d
i
f
f
e
r
e
nc
e
s
a
r
e
r
e
q
u
ir
e
d
to
s
t
o
r
e
d
a
t
a
.
T
hi
s
dr
a
w
ba
c
k
i
s
m
i
t
ig
a
t
e
d
in
R
D
E
,
a
m
or
e
a
d
va
nc
e
d
m
e
t
ho
d
p
r
o
p
os
e
d
i
n
[
17
]
,
[
18
]
t
ha
t
r
e
d
uc
e
s
t
he
s
iz
e
o
f
th
e
d
i
f
f
e
r
e
nc
e
s
us
in
g
l
oga
r
i
th
m
ic
a
nd
e
x
po
ne
nt
ia
l
f
u
nc
ti
o
ns
.
P
r
e
vi
ou
s
r
e
s
e
a
r
c
h
o
n
di
f
f
e
r
e
n
c
e
e
x
p
a
ns
i
on
a
nd
R
D
E
h
a
s
f
oc
us
e
d
on
d
a
t
a
e
m
be
dd
i
ng
s
c
he
m
e
a
l
go
r
i
th
m
s
,
w
h
ic
h
r
e
q
u
i
r
e
p
i
xe
l
pa
i
r
s
t
o
p
e
r
f
o
r
m
t
he
e
m
be
d
di
n
g
p
r
o
c
e
s
s
.
T
h
e
s
e
l
e
c
ti
o
n
o
f
t
he
s
e
p
ix
e
l
pa
i
r
s
te
n
ds
t
o
be
do
ne
b
y
s
o
r
t
in
g
th
e
pi
xe
ls
i
n
a
f
la
t
s
e
qu
e
n
c
e
[
19
]
,
w
h
ic
h
c
a
n
r
e
s
u
lt
in
t
he
s
e
le
c
t
i
on
o
f
p
a
i
r
s
w
i
th
hi
gh
d
is
t
o
r
t
i
on
.
T
o
a
d
d
r
e
s
s
t
h
is
is
s
ue
,
t
he
s
t
ud
y
i
n
[
20
]
,
[
21
]
p
r
o
pos
e
d
t
h
e
c
onc
e
p
t
o
f
P
V
O
,
w
h
ic
h
f
oc
us
e
s
o
n
s
o
r
t
in
g
pi
xe
l
v
a
l
ue
s
w
i
th
i
n
b
l
oc
ks
o
f
pi
xe
ls
in
da
ta
e
m
be
dd
i
ng
.
W
h
il
e
th
is
c
o
nc
e
p
t
s
ha
r
e
s
t
he
s
a
m
e
ob
je
c
t
i
ve
,
t
he
P
V
O
te
c
hn
i
q
ue
'
s
li
m
i
ta
ti
on
is
t
ha
t
i
t
i
s
o
n
ly
l
im
i
te
d
to
s
pe
c
i
f
ic
p
i
xe
l
b
lo
c
ks
.
T
hi
s
r
e
s
e
a
r
c
h
p
r
e
s
e
nt
s
a
n
R
D
H
s
c
he
m
e
i
nt
e
gr
a
t
in
g
a
n
e
nh
a
nc
e
d
R
D
E
t
e
c
h
ni
q
ue
w
it
h
a
da
p
ti
ve
P
V
O
t
o
a
c
h
ie
ve
h
ig
h
e
m
b
e
dd
in
g
c
a
pa
c
it
y
a
nd
l
ow
d
is
t
or
ti
on.
T
he
ke
y
c
on
c
e
p
t
of
th
is
s
c
he
m
e
is
th
e
c
om
p
ut
a
t
io
n
o
f
r
e
du
c
e
d
d
if
f
e
r
e
n
c
e
va
l
ue
s
,
w
hi
c
h
a
l
lo
w
s
e
a
c
h
pi
xe
l
pa
ir
t
o
s
to
r
e
u
p
to
2
b
it
s
o
f
s
e
c
r
e
t
da
ta
.
T
he
p
r
o
pos
e
d
P
V
O
s
c
he
m
e
a
r
r
a
n
ge
s
pi
xe
l
in
di
c
e
s
b
a
s
e
d
on
i
nt
e
ns
it
y
to
op
ti
m
iz
e
e
m
be
dd
in
g
,
g
ui
di
ng
th
e
s
e
le
c
t
io
n
o
f
pi
xe
l
pa
i
r
s
.
U
nl
ik
e
p
r
e
vi
ous
m
e
th
ods
t
ha
t
r
e
ly
on
f
i
xe
d
p
ix
e
l
b
lo
c
ks
,
t
hi
s
a
p
pr
oa
c
h a
pp
li
e
s
f
o
ur
di
r
e
c
t
io
na
l
P
V
O
s
c
he
m
e
s
,
ho
r
iz
o
nt
a
l,
v
e
r
ti
c
a
l
,
r
i
gh
t
-
di
a
g
ona
l
,
a
n
d
le
f
t
-
d
ia
g
ona
l,
to
e
x
pl
oi
t
s
pa
t
ia
l
c
or
r
e
la
t
io
ns
a
c
r
os
s
d
i
f
f
e
r
e
nt
o
r
ie
nt
a
t
io
ns
.
T
h
is
a
da
pt
a
bi
li
ty
r
e
duc
e
s
d
is
t
or
ti
on
a
nd
i
nc
r
e
a
s
e
s
c
a
pa
c
i
ty
by
a
li
gni
ng
pi
xe
l
o
r
de
r
i
ng
w
i
th
lo
c
a
l
im
a
ge
s
t
r
uc
t
ur
e
s
.
T
h
is
m
e
th
o
d
is
pa
r
t
ic
u
la
r
l
y
s
u
it
a
bl
e
f
o
r
r
e
a
l
-
w
or
ld
a
ppl
ic
a
ti
ons
s
uc
h
a
s
s
e
c
r
e
t
c
om
m
un
ic
a
ti
on,
c
op
yr
ig
ht
p
r
o
te
c
ti
on,
a
nd
m
e
di
c
a
l
d
a
ta
s
e
c
ur
it
y,
w
he
r
e
f
id
e
li
ty
a
nd
c
on
f
id
e
nt
ia
l
it
y a
r
e
e
s
s
e
n
ti
a
l.
T
hus
,
th
e
m
a
in
c
on
t
r
i
but
io
ns
of
t
h
is
r
e
s
e
a
r
c
h a
r
e
:
i)
I
n
tr
od
uc
t
io
n
o
f
a
nov
e
l
R
D
E
-
ba
s
e
d
R
D
H
m
e
t
ho
d
:
t
hi
s
s
tu
d
y
p
r
o
pos
e
s
a
n
a
d
va
nc
e
d
R
D
H
te
c
hn
iq
ue
th
a
t
in
te
g
r
a
te
s
a
n
e
nha
nc
e
d
r
e
v
e
r
s
ib
le
d
if
f
e
r
e
nc
e
e
x
pa
ns
io
n
m
e
c
ha
ni
s
m
.
T
he
in
nova
t
io
n
s
ig
n
i
f
ic
a
nt
ly
m
in
im
iz
e
s
t
he
di
s
to
r
t
io
n
ty
pi
c
a
l
ly
in
t
r
od
uc
e
d
du
r
i
ng
da
ta
e
m
be
dd
in
g,
t
he
r
e
by
p
r
e
s
e
r
v
in
g
th
e
o
r
ig
in
a
l
im
a
ge
'
s
v
is
ua
l
qua
li
ty
. U
nl
ik
e
tr
a
di
ti
ona
l
R
D
E
s
c
he
m
e
s
,
th
is
a
pp
r
oa
c
h c
a
n
e
m
be
d
up
t
o
2
bi
ts
pe
r
pi
x
e
l
pa
i
r
w
i
th
o
ut
c
a
us
i
ng
no
ti
c
e
a
b
le
a
r
ti
f
a
c
ts
,
m
a
ki
ng
i
t
s
ui
ta
b
le
f
o
r
s
e
ns
i
ti
ve
i
m
a
ge
a
p
pl
ic
a
ti
ons
s
uc
h
a
s
m
e
d
ic
a
l
i
m
a
g
in
g
or
s
e
c
u
r
e
c
o
m
m
uni
c
a
t
io
ns
.
ii
)
D
e
ve
lo
p
m
e
n
t
o
f
a
n
a
da
pt
iv
e
P
V
O
s
c
he
m
e
to
i
m
p
r
o
ve
e
m
be
d
d
in
g
e
f
f
ic
ie
n
c
y:
th
e
r
e
s
e
a
r
c
h
a
ls
o
p
r
e
s
e
nt
s
a
r
e
f
in
e
d
P
V
O
m
e
c
h
a
ni
s
m
th
a
t
a
r
r
a
n
ge
s
p
ix
e
l
in
d
ic
e
s
ba
s
e
d
on
i
nt
e
ns
i
ty
le
v
e
ls
to
gu
id
e
e
f
f
ic
ie
nt
e
m
be
dd
in
g
.
I
ns
te
a
d
o
f
r
e
l
yi
n
g
o
n
s
ta
ti
c
p
ix
e
l
b
lo
c
ks
,
t
hi
s
m
e
t
hod
a
da
p
ts
to
lo
c
a
l
i
m
a
g
e
c
ha
r
a
c
te
r
is
ti
c
s
by
e
m
pl
o
yi
ng
f
ou
r
d
ir
e
c
t
io
na
l
P
V
O
s
t
r
a
te
g
ie
s
:
ho
r
iz
on
ta
l
,
v
e
r
t
ic
a
l,
r
ig
ht
-
d
ia
gon
a
l,
a
nd
le
f
t
-
di
a
go
na
l
.
T
hi
s
or
ie
n
ta
ti
on
-
a
w
a
r
e
e
m
be
d
di
ng
f
r
a
m
e
w
o
r
k
u
t
il
iz
e
s
s
pa
t
ia
l
c
or
r
e
la
t
io
ns
w
it
hi
n
th
e
i
m
a
g
e
,
le
a
di
ng
t
o
hi
g
he
r
d
a
ta
c
a
pa
c
i
ty
a
nd
r
e
du
c
e
d
v
is
ua
l
d
is
t
or
ti
on
.
T
hi
s
pa
pe
r
is
o
r
ga
ni
z
e
d
i
nt
o
4
s
e
c
ti
o
ns
.
S
e
c
t
i
on
1
p
r
ov
i
de
s
a
n
o
ve
r
v
i
e
w
o
f
th
e
r
e
s
e
a
r
c
h
p
r
o
b
le
m
.
I
n
s
e
c
t
i
on
2
,
s
e
v
e
r
a
l
R
D
E
a
n
d
P
V
O
s
c
h
e
m
e
s
a
r
e
i
nt
r
od
uc
e
d
.
S
e
c
ti
on
3
i
n
t
r
o
du
c
e
s
a
nd
e
x
p
la
in
s
th
e
p
r
o
pos
e
d
m
e
t
h
od
i
n
d
e
t
a
i
l
,
i
nc
l
ud
i
ng
h
ow
th
is
m
e
t
ho
d
is
i
nt
e
n
de
d
to
s
o
lv
e
th
e
r
e
s
e
a
r
c
h
p
r
ob
le
m
.
S
e
c
t
i
on
4
p
r
e
s
e
nt
s
a
n
e
va
lu
a
t
i
on
o
f
t
he
pr
o
pos
e
d
m
e
th
od
a
nd
c
o
m
p
a
r
e
s
t
he
e
xp
e
r
i
m
e
n
ta
l
r
e
s
u
l
ts
o
f
t
h
is
m
e
th
o
d
w
i
t
h
p
r
e
v
io
us
m
e
th
od
s
.
S
e
c
t
io
n
5
,
i
n
c
on
c
l
us
io
n
,
s
um
m
a
r
i
z
e
s
t
h
is
r
e
s
e
a
r
c
h
a
n
d
d
is
c
us
s
e
s
t
he
p
o
te
n
ti
a
l
f
u
t
u
r
e
w
o
r
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
hy
br
id
s
te
ganogr
aphy
s
c
he
m
e
w
it
h r
e
du
c
e
d di
ff
e
r
e
nc
e
e
x
pan
s
io
n and
…
(
I
K
ade
k
A
gus
A
r
ie
s
ta
P
ut
r
a
)
3565
2.
R
E
L
A
T
E
D
WORKS
R
e
s
e
a
r
c
h r
e
la
te
d t
o R
D
E
w
a
s
f
ir
s
t
in
tr
oduc
e
d i
n
[
17]
. T
he
a
ut
hor
s
i
de
nt
if
ie
d l
im
it
a
ti
ons
t
o t
r
a
di
ti
ona
l
di
f
f
e
r
e
nc
e
e
xpa
ns
io
n
te
c
hni
que
s
,
pa
r
ti
c
ul
a
r
ly
th
e
de
gr
a
da
ti
on
of
vi
s
ua
l
qua
li
ty
w
he
n
us
e
d
f
or
m
ul
ti
la
ye
r
e
m
be
ddi
ng
a
nd
th
e
r
is
k
of
ove
r
f
lo
w
or
unde
r
f
lo
w
.
A
n
R
D
E
s
c
he
m
e
w
a
s
pr
opos
e
d
to
r
e
duc
e
th
e
di
f
f
e
r
e
nc
e
va
lu
e
in
tr
a
di
ti
ona
l
di
f
f
e
r
e
nc
e
e
xp
a
ns
io
n
m
e
th
ods
by
im
pl
e
m
e
nt
in
g
tr
a
ns
f
or
m
a
ti
on
f
unc
ti
ons
ba
s
e
d
on
e
xpone
nt
ia
l
a
nd l
oga
r
it
hm
ic
ope
r
a
ti
ons
t
o a
ddr
e
s
s
t
he
s
e
i
s
s
ue
s
.
F
ur
th
e
r
de
ve
lo
pm
e
nt
s
w
e
r
e
c
onduc
te
d i
n
[
22]
,
in
tr
oduc
in
g
a
ne
w
t
r
a
ns
f
or
m
a
ti
on
f
unc
ti
on
w
it
h
a
w
id
e
r
r
a
nge
.
T
hi
s
f
unc
ti
on
pr
ove
d
m
o
r
e
e
f
f
e
c
ti
ve
i
n
im
pr
ovi
ng
im
a
ge
qua
li
ty
da
m
a
ge
d
by
m
ul
ti
la
ye
r
e
m
be
ddi
ng.
H
ow
e
ve
r
,
ke
y
c
ha
ll
e
nge
s
s
uc
h
a
s
lo
w
e
m
be
ddi
ng c
a
pa
c
it
y a
r
e
s
ti
ll
a
s
ig
ni
f
ic
a
nt
l
im
it
a
ti
on i
n a
ppl
yi
ng s
uc
h m
e
th
ods
.
T
o
im
pr
ove
th
e
s
e
c
ur
it
y
a
nd
qua
li
ty
of
da
ta
e
m
be
ddi
ng
in
m
e
di
c
a
l
im
a
ge
s
,
s
tu
di
e
s
in
[
23]
,
[
24]
pr
opos
e
d
a
n
R
D
E
m
e
th
od
th
a
t
ope
r
a
te
s
on
2×
2
pi
xe
l
bl
oc
ks
(
p_1,
p_2,
p_3,
p_4)
.
T
he
s
e
bl
oc
ks
a
r
e
c
a
te
gor
iz
e
d
in
to
th
r
e
e
ty
p
e
s
:
e
xp
a
nda
bl
e
,
c
ha
nge
a
bl
e
,
a
nd
unc
ha
nge
a
bl
e
.
T
o
ov
e
r
c
om
e
th
e
pr
obl
e
m
of
ove
r
f
lo
w
or
unde
r
f
lo
w
,
onl
y
bl
oc
ks
w
it
h
e
xpa
nda
bl
e
a
nd
c
ha
nge
a
bl
e
c
a
te
gor
ie
s
a
r
e
u
s
e
d
to
e
m
be
d
1
-
bi
t
s
e
c
r
e
t
da
t
a
.
T
hi
s
m
e
th
od
a
dj
us
t
s
th
e
ba
s
e
poi
nt
u
s
e
d
to
c
a
lc
ul
a
te
th
e
di
f
f
e
r
e
nc
e
va
lu
e
f
or
th
e
ot
he
r
th
r
e
e
pi
xe
ls
,
s
uc
h
a
s
a
t
p_2
f
or
th
e
m
e
th
od
in
[
23]
a
nd
in
or
de
r
p_3
f
or
th
e
m
e
th
od
in
[
24]
.
I
n
e
xpa
nda
bl
e
bl
oc
ks
,
a
bs
ol
ut
e
di
f
f
e
r
e
nc
e
va
lu
e
s
gr
e
a
te
r
th
a
n
1 a
r
e
r
e
duc
e
d
u
s
in
g
a
tr
a
ns
f
or
m
a
ti
on
f
unc
ti
on
be
f
or
e
th
e
s
e
c
r
e
t
da
ta
is
e
m
be
dde
d,
w
hi
le
va
lu
e
s
out
s
id
e
th
is
r
a
nge
a
r
e
ha
ndl
e
d
us
in
g
tr
a
di
ti
ona
l
di
f
f
e
r
e
nc
e
e
xpa
n
s
io
n
. T
he
s
e
c
r
e
t
da
ta
is
e
m
be
dde
d
in
th
e
le
a
s
t
s
ig
ni
f
ic
a
nt
bi
t
or
us
in
g
tr
a
di
ti
ona
l
d
if
f
e
r
e
nc
e
e
xpa
ns
io
n
[
23]
f
or
m
ut
a
bl
e
bl
oc
ks
.
A
lt
h
ou
gh
t
hi
s
R
D
E
c
o
nc
e
p
t
s
u
c
c
e
s
s
f
ul
ly
a
dd
r
e
s
s
e
s
th
e
pr
obl
e
m
of
hi
gh
d
i
s
to
r
ti
o
n
a
nd
im
pr
ov
e
s
im
a
g
e
q
ua
li
t
y,
th
e
pi
x
e
l
bl
o
c
k
s
e
l
e
c
ti
on
p
r
o
c
e
s
s
r
e
m
a
in
s
a
c
ha
ll
e
ng
e
,
e
s
p
e
c
i
a
ll
y
in
f
in
d
in
g
th
e
o
pt
im
a
l
ba
s
e
p
oi
n
t
to
a
c
hi
e
ve
m
in
i
m
a
l
d
if
f
e
r
e
nc
e
va
lu
e
s
a
nd
s
ti
l
l
r
e
s
ul
ts
i
n
m
a
n
y u
nc
h
a
ng
e
a
bl
e
bl
o
c
ks
a
f
f
e
c
ti
ng
t
h
e
e
m
be
ddi
ng
c
a
p
a
c
it
y.
S
hi
f
ti
ng
th
e
f
oc
us
to
P
V
O
-
ba
s
e
d
a
ppr
oa
c
he
s
,
L
e
e
e
t
al
.
[
21]
pr
e
s
e
nt
a
n
a
ppr
oa
c
h
f
or
R
D
H
s
c
he
m
e
s
by
ut
il
iz
in
g
a
c
om
bi
na
ti
on
of
P
V
O
a
nd
two
-
la
ye
r
e
m
be
ddi
ng
t
o
im
pr
ove
th
e
e
f
f
ic
ie
nc
y
of
hi
di
ng
s
e
c
r
e
t
da
ta
in
di
gi
ta
l
im
a
ge
s
.
B
y
di
vi
di
ng
th
e
da
ta
e
m
be
ddi
ng
pr
oc
e
s
s
in
to
two
la
ye
r
s
,
th
e
f
ir
s
t
a
nd
s
e
c
ond
la
ye
r
s
ut
il
iz
e
di
f
f
e
r
e
nt
bl
oc
k
pa
r
ti
ti
on
m
ode
ls
to
a
voi
d
th
e
ove
r
us
e
of
pi
xe
ls
,
th
us
im
pr
ovi
ng
th
e
ove
r
a
ll
pe
r
f
or
m
a
nc
e
.
T
hi
s
s
c
he
m
e
ut
il
iz
e
s
th
e
P
V
O
te
c
hni
que
to
pr
e
di
c
t
a
nd
e
m
be
d
s
e
c
r
e
t
da
ta
by
s
or
ti
ng
th
e
pi
x
e
l
va
lu
e
s
w
it
hi
n
e
a
c
h
bl
oc
k.
D
e
s
p
it
e
i
t
s
a
d
va
nt
a
g
e
s
, t
he
r
e
is
a
li
m
i
ta
ti
o
n of
th
e
P
V
O
te
c
hn
iq
ue
u
s
e
d:
i
t
i
s
o
nl
y
l
im
it
e
d
to
c
e
r
ta
in
bl
o
c
k
s
of
pi
x
e
l
s
.
T
hi
s
li
m
it
a
ti
o
n
r
e
du
c
e
s
i
ts
f
l
e
xi
b
il
it
y
a
nd
e
m
b
e
dd
in
g
c
a
p
a
c
it
y
a
s
c
o
m
pl
e
x
b
lo
c
k
s
w
it
h
h
ig
h
va
r
ia
ti
o
ns
in
p
ix
e
l
v
a
lu
e
s
a
r
e
of
te
n
s
k
ip
p
e
d
,
th
u
s
l
im
it
i
ng
t
h
e
a
m
oun
t
o
f
s
e
c
r
e
t
da
ta
t
h
a
t
c
a
n
b
e
e
m
be
dd
e
d
i
n
t
h
e
im
a
g
e
.
A
dva
nc
in
g
P
V
O
-
ba
s
e
d
m
e
th
odol
ogi
e
s
,
R
e
n
e
t
al
.
[
25]
pr
e
s
e
n
ts
a
dyna
m
ic
hybr
id
R
D
H
te
c
hni
que
de
s
ig
ne
d
to
e
nha
nc
e
da
ta
e
m
be
ddi
ng
c
a
pa
c
it
y
a
nd
im
a
ge
qua
li
ty
.
T
he
a
ppr
oa
c
h
r
e
f
in
e
s
th
e
tr
a
di
ti
ona
l
P
V
O
m
e
th
od
by
c
la
s
s
if
yi
ng
im
a
ge
bl
oc
ks
in
to
th
r
e
e
c
a
te
gor
ie
s
:
e
xt
r
e
m
e
ly
s
m
oot
h,
s
m
oot
h,
a
nd
r
ough,
ba
s
e
d
on
th
e
ir
c
om
pl
e
xi
ty
.
A
bl
oc
k
s
ubdi
vi
s
io
n
a
lg
or
it
hm
is
a
ppl
ie
d
f
or
s
m
oot
h
bl
oc
ks
, a
ll
ow
in
g
up
to
6
bi
ts
of
da
ta
to
be
e
m
be
dde
d.
I
n
e
xt
r
e
m
e
ly
s
m
oot
h
bl
oc
ks
,
a
m
e
di
a
n
pi
xe
l
p
r
e
di
c
ti
on
te
c
hni
que
is
ut
il
iz
e
d
to
pr
e
di
c
t
th
e
r
e
m
a
in
in
g pi
xe
ls
, e
na
bl
in
g t
he
e
m
be
ddi
ng of
up t
o 8 bi
ts
of
da
t
a
. R
ough bloc
ks
a
r
e
e
xc
lu
de
d f
r
om
e
m
be
ddi
ng
to
a
voi
d
di
s
to
r
ti
on.
T
he
e
xpe
r
im
e
nt
a
l
r
e
s
ul
t
s
de
m
on
s
tr
a
te
t
ha
t
th
is
m
e
th
od
of
f
e
r
s
s
ig
ni
f
ic
a
nt
ly
hi
ghe
r
e
m
be
ddi
ng
c
a
pa
c
it
y
a
nd
be
tt
e
r
im
a
ge
qu
a
li
ty
th
a
n
tr
a
di
ti
ona
l
P
V
O
te
c
hni
que
s
.
H
ow
e
v
e
r
,
th
e
m
e
th
od’
s
e
f
f
e
c
ti
ve
ne
s
s
is
in
f
lu
e
nc
e
d
by
th
e
te
xt
ur
e
c
h
a
r
a
c
te
r
is
ti
c
s
of
t
he
im
a
ge
,
p
a
r
ti
c
ul
a
r
ly
th
e
s
m
oot
hne
s
s
of
th
e
bl
oc
ks
be
in
g pr
oc
e
s
s
e
d.
F
ur
th
e
r
m
or
e
,
va
r
io
us
s
tu
di
e
s
ha
ve
b
e
e
n
c
onduc
t
e
d
on
a
ppl
yi
n
g
P
V
O
in
pr
e
di
c
ti
on
e
r
r
or
e
xpa
ns
io
n
(
P
E
E
)
s
c
he
m
e
s
.
K
a
ur
e
t
al
.
[
20]
pr
opos
e
d
a
n
R
D
H
s
c
h
e
m
e
in
vol
vi
ng
P
V
O
in
th
e
f
or
m
of
a
s
or
ti
ng
s
tr
a
te
gy
c
om
bi
ne
d
w
it
h
P
E
E
by
e
m
be
ddi
ng s
e
c
r
e
t
da
t
a
in
to
th
e
le
a
s
t
s
ig
ni
f
ic
a
nt
pi
xe
ls
s
or
te
d
by
th
e
ir
pr
e
di
c
ti
on
e
r
r
or
.
S
im
il
a
r
ly
,
C
hi
e
t
al
.
[
26
]
e
xpl
or
e
d
P
V
O
a
nd
P
E
E
,
a
s
c
he
m
e
th
a
t
de
r
iv
e
s
pr
e
di
c
te
d
va
lu
e
s
f
r
om
th
e
s
or
te
d
pi
xe
l
va
lu
e
s
,
w
hi
c
h
ut
il
iz
e
s
two
s
te
ps
of
pr
e
di
c
ti
on
e
r
r
or
tr
ip
le
t
e
xpa
ns
io
n
ba
s
e
d
on
P
V
O
.
O
ve
r
a
ll
,
pr
e
vi
ous
r
e
s
e
a
r
c
h
on
th
e
P
V
O
c
onc
e
pt
ha
s
s
im
il
a
r
it
ie
s
in
s
or
ti
ng
pi
xe
l
va
lu
e
s
w
it
hi
n
bl
oc
ks
of
pi
xe
ls
in
da
t
a
e
m
be
ddi
ng. How
e
ve
r
, t
he
f
unda
m
e
nt
a
l
di
f
f
e
r
e
nc
e
l
ie
s
i
n t
he
a
p
pr
oa
c
h pr
opos
e
d i
n t
hi
s
m
e
th
od, whe
r
e
P
V
O
i
s
m
or
e
f
oc
us
e
d
on
s
or
ti
ng
th
e
pi
xe
l
in
di
c
e
s
to
s
e
le
c
t
th
e
r
ig
ht
pi
xe
l
bl
oc
k
r
a
th
e
r
th
a
n
s
or
ti
ng
th
e
pi
xe
l
va
lu
e
s
w
it
hi
n t
he
bl
oc
k.
3.
P
R
O
P
O
S
E
D
M
E
T
H
O
D
T
r
a
di
ti
ona
l
di
f
f
e
r
e
nc
e
e
xpa
ns
io
n
,
a
s
pr
opos
e
d
in
[
27]
,
e
m
be
ds
s
e
c
r
e
t
da
ta
by
e
xpa
ndi
ng
th
e
di
f
f
e
r
e
nc
e
va
lu
e
s
of
a
dj
a
c
e
nt
pi
xe
l
pa
ir
s
.
S
pe
c
if
ic
a
ll
y,
gi
ve
n
a
pa
ir
of
pi
xe
ls
(
,
)
,
di
f
f
e
r
e
nc
e
e
xpa
ns
io
n
f
ir
s
t
c
a
lc
ul
a
te
s
t
he
a
ve
r
a
ge
va
lu
e
=
⌊
(
+
)
/
2
⌋
a
nd t
he
di
f
f
e
r
e
nc
e
va
lu
e
ℎ
=
−
. N
e
xt
, t
he
s
e
c
r
e
t
bi
t
w
il
l
be
e
m
be
dde
d
by
m
odi
f
yi
ng
th
e
di
f
f
e
r
e
nc
e
va
lu
e
ℎ
,
w
hi
c
h
w
il
l
a
ls
o
be
us
e
d
to
pe
r
f
or
m
r
e
ve
r
s
ib
le
r
e
c
ons
tr
uc
ti
on.
A
lt
hough
di
f
f
e
r
e
nc
e
e
xpa
ns
io
n
i
s
qui
te
s
im
pl
e
a
nd
e
f
f
e
c
ti
ve
,
it
te
nds
to
ha
v
e
hi
gh
di
s
to
r
ti
on
e
s
p
e
c
ia
ll
y
w
he
n
it
ha
s
a
l
a
r
ge
di
f
f
e
r
e
nc
e
va
lu
e
.
T
o
a
ddr
e
s
s
th
is
li
m
it
a
ti
on,
R
D
E
is
in
tr
oduc
e
d
a
s
a
n
im
pr
ove
m
e
nt
th
a
t
r
e
duc
e
s
th
e
di
f
f
e
r
e
nc
e
va
lu
e
by a
ppl
yi
ng t
r
a
ns
f
or
m
a
ti
on
f
unc
ti
ons
s
uc
h a
s
l
oga
r
it
hm
ic
or
e
xpone
nt
ia
l
s
c
a
li
ng
[
17]
. B
y l
im
it
in
g
th
e
r
e
duc
e
d
di
f
f
e
r
e
nc
e
va
lu
e
,
R
D
E
is
not
onl
y
a
bl
e
to
r
e
duc
e
di
s
to
r
ti
on,
but
is
a
ls
o
a
bl
e
to
in
c
r
e
a
s
e
th
e
e
m
be
ddi
ng
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
5
,
O
c
to
be
r
2025
:
3563
-
3575
3566
c
a
pa
c
it
y
a
s
m
or
e
s
m
oot
h
pi
xe
l
p
a
ir
s
w
il
l
be
e
m
be
dda
bl
e
.
T
o
a
c
hi
e
ve
th
is
,
R
D
E
r
e
qui
r
e
s
two
a
ddi
ti
ona
l
c
om
pone
nt
s
:
a
s
e
c
r
e
t
ta
bl
e
(
S
T
)
a
nd
a
lo
c
a
ti
on
m
a
p
(
L
M
)
.
T
he
ST
s
to
r
e
s
th
e
s
e
le
c
te
d
pi
xe
l
pa
ir
s
'
lo
c
a
ti
on
f
or
da
ta
hi
di
ng,
e
ns
ur
in
g
th
a
t
onl
y
s
ui
ta
bl
e
lo
c
a
ti
ons
(
w
it
h
lo
w
vi
s
ua
l
im
pa
c
t)
a
r
e
us
e
d.
T
he
LM
r
e
c
or
ds
th
e
in
f
or
m
a
ti
on
r
e
qui
r
e
d
dur
in
g
th
e
e
xt
r
a
c
ti
on
a
nd
r
e
s
to
r
a
ti
on
pr
oc
e
s
s
.
B
a
s
e
d
on
th
e
s
e
pr
in
c
ip
le
s
,
w
e
d
e
s
ig
n
a
nove
l
e
m
be
ddi
ng
a
lg
or
it
hm
th
a
t
in
te
g
r
a
te
s
R
D
E
w
it
h
a
n
im
pr
o
ve
d
pi
xe
l
s
e
le
c
ti
on
m
e
c
ha
ni
s
m
us
in
g
th
e
P
V
O
s
c
he
m
e
, t
hus
e
na
bl
in
g e
f
f
ic
ie
nt
R
D
E
w
it
h m
in
im
iz
e
d di
s
to
r
ti
on.
3.1. T
h
e
p
r
op
os
e
d
d
at
a h
id
in
g s
c
h
e
m
e
u
s
in
g r
e
d
u
c
e
d
d
if
f
e
r
e
n
c
e
e
xp
an
s
io
n
3.1.1. Dat
a e
m
b
e
d
d
in
g
T
he
pr
opos
e
d
s
c
he
m
e
r
e
qui
r
e
s
a
pa
ir
of
a
nd
pi
xe
ls
to
e
m
be
d
2
bi
ts
of
s
e
c
r
e
t
da
ta
.
A
s
s
e
e
n
in
F
ig
ur
e
1, t
hi
s
pr
oc
e
s
s
be
gi
ns
w
it
h
c
a
lc
ul
a
ti
ng t
he
di
f
f
e
r
e
nc
e
va
l
ue
ba
s
e
d on the
v
a
lu
e
s
of
a
nd
us
in
g (
1)
.
=
{
−
,
>
−
,
ℎ
(
1)
O
nc
e
th
e
va
lu
e
is
obt
a
in
e
d,
w
e
th
e
n
n
e
e
d
to
c
he
c
k
w
he
th
e
r
t
hi
s
va
lu
e
c
a
n
be
e
xpa
nde
d
or
not
.
I
n
th
is
c
ont
e
xt
,
th
e
va
lu
e
is
c
on
s
id
e
r
e
d
e
xp
e
nda
bl
e
if
it
e
xc
e
e
d
s
3.
I
f
th
is
c
ondi
ti
on
is
m
e
t,
th
e
n
th
e
va
lu
e
c
a
n
be
e
xpa
nde
d,
a
nd
it
w
il
l
be
a
dj
us
te
d
a
c
c
or
di
ng
to
(
2)
.
C
onve
r
s
e
ly
,
if
th
e
va
lu
e
is
le
s
s
th
a
n
or
e
qua
l
to
3,
it
is
de
e
m
e
d non
-
e
xpa
nda
bl
e
. I
n s
u
c
h i
ns
ta
nc
e
s
, t
he
va
lu
e
i
s
upda
te
d t
o 2.
′
=
{
⌊
4
⌋
−
⌊
√
⌊
l
o
g
2
⌋
⌋
,
>
3
,
ℎ
(
2)
T
he
ne
xt
s
ta
ge
in
th
is
pr
oc
e
dur
e
in
vol
ve
s
e
m
be
ddi
ng
th
e
s
e
c
r
e
t
da
ta
(
)
.
T
hi
s
is
a
c
c
om
pl
is
he
d
by
ut
il
iz
in
g
(
3)
.
O
nc
e
th
e
s
e
c
r
e
t
da
ta
i
s
s
uc
c
e
s
s
f
ul
ly
e
m
be
dde
d,
it
'
s
ne
c
e
s
s
a
r
y
to
upda
te
th
e
m
o
s
t
r
e
c
e
nt
a
nd
va
lu
e
s
a
s
′
a
nd
′
. T
hi
s
upda
te
i
s
don
e
us
in
g (
4)
.
′′
=
′
×
4
+
S
(
3)
′
,
′
=
{
′
=
+
′′
′
=
,
>
′
=
+
′′
′
=
,
ℎ
(
4)
A
f
te
r
s
uc
c
e
s
s
f
ul
ly
obt
a
in
in
g
th
e
ne
w
va
lu
e
s
of
a
nd
,
th
e
s
ubs
e
que
nt
s
te
p
in
vol
ve
s
va
li
da
ti
ng
th
e
s
e
va
lu
e
s
.
T
hi
s
va
li
da
ti
on
a
im
s
to
e
ns
ur
e
th
a
t
th
e
va
lu
e
s
of
′
a
nd
′
r
e
m
a
in
w
it
hi
n
th
e
a
c
c
e
pt
a
bl
e
r
a
nge
of
0
to
255
,
c
or
r
e
s
ponding
to
th
e
va
li
d
c
ol
or
r
a
nge
.
T
hi
s
s
te
p
is
c
r
uc
ia
l
a
s
it
pr
e
ve
nt
s
ove
r
f
lo
w
or
unde
r
f
lo
w
,
w
hi
c
h
c
oul
d
c
a
us
e
e
r
r
or
s
i
n t
he
s
ub
s
e
que
nt
pr
oc
e
s
s
e
s
. P
r
e
ve
nt
in
g t
hi
s
ove
r
f
lo
w
or
unde
r
f
lo
w
f
ol
lo
w
s
w
ha
t
is
s
how
n i
n (
5)
.
o
v
e
r
flo
w
/
un
d
e
r
flo
w
=
{
,
0
≤
≤
255
,
ℎ
(
5)
I
f
th
e
upda
te
d
va
lu
e
s
of
a
nd
f
a
ll
out
s
id
e
th
e
va
li
d
c
ol
or
r
a
nge
,
it
im
pl
ie
s
th
a
t
th
is
pi
x
e
l
pa
ir
is
un
s
ui
ta
bl
e
f
or
e
m
be
ddi
ng
s
e
c
r
e
t
da
ta
.
I
n
s
uc
h
a
c
a
s
e
,
th
e
va
lu
e
is
th
e
n
upda
te
d
to
0
a
s
a
n
in
di
c
a
to
r
of
a
non
-
e
m
be
dda
bl
e
pa
ir
. T
he
r
e
f
or
e
, w
e
ne
e
d t
o l
ook f
or
ot
he
r
pi
xe
l
pa
ir
s
a
s
a
n a
lt
e
r
na
ti
ve
. H
ow
e
v
e
r
, i
f
t
hi
s
pi
xe
l
pa
ir
c
a
n e
m
be
d t
he
s
e
c
r
e
t
m
e
s
s
a
ge
(
ST
≠
0
)
, t
he
n t
he
ne
xt
s
te
p i
s
t
o de
te
r
m
in
e
t
he
va
lu
e
by us
in
g t
he
(
6)
.
=
4
(
6)
3.1.2. Dat
a e
xt
r
ac
t
io
n
p
r
oc
e
s
s
I
n
th
e
s
e
c
r
e
t
da
ta
e
xt
r
a
c
ti
on
s
ta
ge
,
th
e
f
ir
s
t
s
te
p
th
a
t
m
us
t
be
t
a
ke
n
is
to
obt
a
in
pi
xe
l
pa
ir
s
f
r
om
th
e
s
te
go
im
a
ge
. T
he
s
te
go
im
a
ge
is
a
n
im
a
ge
th
a
t
ha
s
be
e
n
us
e
d
to
hi
de
s
e
c
r
e
t
da
t
a
w
it
hi
n
it
.
A
f
te
r
obt
a
in
in
g
th
e
pi
xe
l
pa
ir
s
a
nd
th
e
lo
c
a
ti
ons
w
he
r
e
th
e
s
e
c
r
e
t
da
ta
is
hi
dde
n,
th
e
ne
xt
s
te
p
is
pe
r
f
or
m
in
g
a
s
e
r
ie
s
of
c
a
lc
ul
a
ti
ons
.
T
he
s
e
c
a
lc
ul
a
ti
ons
a
r
e
th
e
r
e
v
e
r
s
e
of
th
e
d
a
ta
in
s
e
r
ti
on
pr
oc
e
s
s
,
in
te
ndi
ng
to
r
e
ve
a
l
a
g
a
in
th
e
s
e
c
r
e
t
da
ta
t
ha
t
ha
s
be
e
n hi
dde
n.
T
he
pr
oc
e
s
s
of
e
xt
r
a
c
ti
ng t
hi
s
s
e
c
r
e
t
da
ta
i
s
c
a
r
r
ie
d out, a
s
s
e
e
n i
n F
ig
ur
e
2.
T
o
c
a
r
r
y
out
th
e
e
xt
r
a
c
ti
on
of
s
e
c
r
e
t
da
ta
,
th
e
r
e
a
r
e
two
im
por
ta
nt
va
lu
e
s
ne
e
de
d,
na
m
e
ly
th
e
va
lu
e
a
nd
th
e
LM
va
lu
e
.
T
he
s
e
two
va
lu
e
s
a
r
e
us
ua
ll
y
tr
a
ns
m
it
te
d
di
f
f
e
r
e
nt
ly
a
nd
a
r
e
r
e
qui
r
e
d
f
o
r
e
xt
r
a
c
ti
on.
T
he
f
ir
s
t
s
te
p
in
th
e
e
xt
r
a
c
ti
on
pr
oc
e
s
s
is
to
c
a
lc
ul
a
te
th
e
di
f
f
e
r
e
nc
e
va
lu
e
′′
.
T
hi
s
va
lu
e
is
obt
a
in
e
d us
in
g (
7)
. I
f
t
he
va
lu
e
i
s
-
1, 1, or
2, t
he
n s
e
c
r
e
t
da
ta
i
s
hi
dde
n i
n t
ha
t
pi
xe
l
pa
ir
. C
onve
r
s
e
ly
, i
f
t
he
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
hy
br
id
s
te
ganogr
aphy
s
c
he
m
e
w
it
h r
e
du
c
e
d di
ff
e
r
e
nc
e
e
x
pan
s
io
n and
…
(
I
K
ade
k
A
gus
A
r
ie
s
ta
P
ut
r
a
)
3567
va
lu
e
i
s
0,
th
e
n
th
e
r
e
is
no
s
e
c
r
e
t
da
t
a
hi
dde
n
in
th
a
t
pi
x
e
l
pa
ir
.
A
f
te
r
obt
a
in
in
g
th
e
′′
va
lu
e
,
th
e
n
e
xt
s
t
e
p
is
t
o e
xt
r
a
c
t
th
e
s
e
c
r
e
t
d
a
ta
a
nd r
e
ga
in
t
he
or
ig
in
a
l
va
lu
e
. T
he
s
e
c
r
e
t
da
ta
S
va
lu
e
i
s
obt
a
in
e
d u
s
in
g t
he
(
8)
.
′′
=
{
′
−
′
,
=
−
1
′
−
′
,
∈
{
1
,
2
}
(
7)
S
=
′′
4
(
8)
T
o
r
e
s
to
r
e
th
e
di
f
f
e
r
e
nc
e
va
lu
e
′′
to
it
s
or
ig
in
a
l
va
lu
e
,
s
e
v
e
r
a
l
s
t
e
ps
m
us
t
b
e
ta
ke
n
d
e
pe
ndi
ng
on
th
e
va
lu
e
.
I
f
∈
{
1
,
2
}
,
th
e
′
va
lu
e
c
a
n
be
obt
a
in
e
d
us
in
g
(
9)
,
a
nd
th
e
or
ig
in
a
l
va
lu
e
c
a
n
be
obt
a
in
e
d
us
in
g
(
10)
.
I
f
th
e
va
lu
e
is
2,
th
e
n
th
e
or
ig
in
a
l
va
lu
e
c
a
n
be
obt
a
in
e
d
us
in
g
(
11)
.
A
f
te
r
a
ll
th
e
s
e
s
te
ps
a
r
e
c
om
pl
e
te
d,
th
e
in
it
ia
l
va
lu
e
s
of
a
nd
c
a
n
be
c
a
lc
ul
a
te
d
u
s
in
g
th
e
(
12)
.
T
hus
,
th
e
s
e
c
r
e
t
da
ta
e
xt
r
a
c
ti
on
pr
oc
e
s
s
is
c
om
pl
e
te
.
T
h
e
pr
e
vi
ous
ly
hi
dde
n
s
e
c
r
e
t
da
ta
w
it
hi
n
t
he
im
a
ge
ha
s
now
b
e
e
n
s
u
c
c
e
s
s
f
ul
ly
obt
a
in
e
d,
a
nd w
e
c
a
n a
l
s
o r
e
tr
ie
ve
t
he
or
ig
in
a
l
c
ove
r
i
m
a
ge
.
′
=
⌊
′′
−
4
⌋
(
9)
=
4
×
(
′
+
⌊
√
′
⌊
l
o
g
2
′
⌋
⌋
)
+
LM
(
10)
d
=
⌊
′′
−
4
⌋
(
11)
,
=
{
=
′
=
+
,
=
−
1
=
′
=
+
,
∈
{
1
,
2
}
(
12)
I
n
ge
ne
r
a
l,
th
e
c
om
pa
r
is
on
be
tw
e
e
n
th
e
pr
opos
e
d
m
e
th
od
a
n
d
th
e
pr
e
vi
ous
m
e
th
ods
[
24]
,
[
28]
is
il
lu
s
tr
a
te
d
in
T
a
bl
e
1.
A
not
he
r
ke
y
di
f
f
e
r
e
nc
e
i
s
th
a
t
m
e
th
ods
[
24]
,
[
28]
e
m
be
d
1
s
e
c
r
e
t
m
e
s
s
a
ge
f
or
e
a
c
h
di
f
f
e
r
e
nc
e
va
lu
e
d
.
I
n
c
ont
r
a
s
t,
th
e
pr
opos
e
d
m
e
th
od
e
m
b
e
ds
2
bi
ts
of
s
e
c
r
e
t
m
e
s
s
a
ge
f
or
e
a
c
h
di
f
f
e
r
e
nc
e
va
lu
e
d, r
e
s
ul
ti
ng i
n a
l
a
r
ge
r
e
m
be
ddi
ng c
a
pa
c
it
y.
F
ig
ur
e
1. D
a
ta
e
m
be
ddi
ng pr
oc
e
s
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
5
,
O
c
to
be
r
2025
:
3563
-
3575
3568
F
ig
ur
e
2. D
a
ta
e
xt
r
a
c
ti
on pr
oc
e
s
s
T
a
bl
e
1. C
om
pa
r
is
on b
e
twe
e
n t
he
m
e
th
od i
n
[
24]
, [
28]
, a
nd t
he
pr
opos
e
d m
e
th
od
M
e
t
hod
P
i
xe
l
bl
oc
k
(
)
D
i
f
f
e
r
e
nc
e
(
)
R
e
duc
e
d di
f
f
e
r
e
nc
e
e
xpa
ns
i
on (
′
)
N
e
w
pi
xe
l
c
om
put
a
t
i
on (
′
)
M
e
t
hod i
n
[
24]
2
×
2
(
0
,
1
,
2
,
3
)
{
0
=
0
−
2
1
=
1
−
2
2
=
0
3
=
3
−
2
{
−
(
2
+
2
⌊
2
(
)
⌋
+
⌊
√
⌊
2
(
)
⌋
⌋
)
,
>
1
+
(
2
+
2
⌊
2
(
)
⌋
+
⌊
√
⌊
2
(
)
⌋
⌋
)
,
<
1
{
0
′
=
0
′′
+
2
1
′
=
1
+
2
2
′
=
2
3
′
=
ℎ
3
′′
+
2
M
e
t
hod i
n
[
28]
3
×
3
(
0
,
1
,
…
,
8
)
{
0
=
0
−
4
1
=
1
−
4
.
.
.
3
=
3
−
4
{
−
2
⌊
2
(
)
⌋
−
1
,
2
×
2
⌊
2
(
)
⌋
−
1
≤
≤
3
×
2
⌊
2
(
)
⌋
−
1
−
2
⌊
2
(
)
⌋
,
<
3
×
2
⌊
2
(
)
⌋
−
1
≤
≤
4
×
2
⌊
2
(
)
⌋
−
1
{
0
′
=
0
′′
+
2
1
′
=
1
+
2
.
.
.
3
′
=
ℎ
3
′′
+
1
P
r
opos
e
d
m
e
t
hod
1
×
2
(
0
,
1
)
|
1
−
0
|
{
⌊
4
⌋
−
⌊
√
⌊
l
o
g
2
⌋
⌋
,
>
3
,
ℎ
{
{
0
′
=
1
+
′′
1
′
=
1
,
0
≥
1
{
0
′
=
0
1
′
=
1
+
′′
,
0
<
1
3.1.3. E
m
b
e
d
d
in
g c
ap
ac
it
y an
d
d
is
t
or
t
io
n
m
e
as
u
r
e
m
e
n
t
E
m
be
ddi
ng
c
a
pa
c
it
y
is
m
e
a
s
ur
e
d
in
te
r
m
s
of
bi
t
pe
r
pi
xe
l
(
bpp)
,
qua
nt
if
yi
ng
how
m
uc
h
s
e
c
r
e
t
da
ta
c
a
n
be
e
m
be
dde
d
in
e
a
c
h
c
ove
r
im
a
ge
pi
xe
l.
T
h
e
e
m
be
ddi
ng
c
a
pa
c
it
y
f
or
a
gi
ve
n
im
a
ge
c
a
n
be
c
a
l
c
ul
a
te
d
us
in
g t
he
(
13)
.
Em
b
e
d
d
in
g
c
a
p
a
c
it
y
=
t
o
t
a
l
e
m
b
e
d
d
e
d
b
i
t
s
h
e
ig
h
t
×
w
id
th
(1
3
)
D
is
to
r
ti
on
is
e
va
lu
a
te
d
u
s
in
g
pe
a
k
s
ig
na
l
-
to
-
noi
s
e
r
a
ti
o
(
P
S
N
R
)
a
s
s
e
e
n
in
(
14)
a
nd
s
tr
uc
tu
r
a
l
s
im
il
a
r
it
y
in
de
x
m
e
a
s
ur
e
m
e
nt
(
S
S
I
M
)
.
M
e
a
n
s
qu
a
r
e
d
e
r
r
or
(
M
S
E
)
is
c
a
lc
ul
a
te
d by
a
ve
r
a
gi
ng
th
e
s
qua
r
e
d
di
f
f
e
r
e
nc
e
s
be
twe
e
n
th
e
pi
xe
l
va
lu
e
s
of
th
e
or
ig
in
a
l
i
m
a
ge
a
nd
th
e
s
te
go
im
a
ge
′
a
t
th
e
pos
it
io
n
(
,
)
,
w
he
r
e
a
nd
r
e
pr
e
s
e
nt
t
he
i
m
a
ge
di
m
e
ns
io
n
s
a
s
s
e
e
n i
n (
15)
.
=
10
l
o
g
10
(
255
2
)
(1
4
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
hy
br
id
s
te
ganogr
aphy
s
c
he
m
e
w
it
h r
e
du
c
e
d di
ff
e
r
e
nc
e
e
x
pan
s
io
n and
…
(
I
K
ade
k
A
gus
A
r
ie
s
ta
P
ut
r
a
)
3569
=
1
⋅
∑
∑
(
−
′
)
2
=
1
=
1
(1
5
)
(
,
)
=
(
2
μ
μ
+
1
)
(
2
σ
+
2
)
(
μ
2
+
μ
2
+
1
)
(
σ
2
+
σ
2
+
2
)
(1
6
)
S
S
I
M
,
a
s
s
e
e
n
in
f
or
m
ul
a
(
16)
,
m
e
a
s
ur
e
s
th
e
pe
r
c
e
pt
ua
l
s
im
il
a
r
it
y
be
twe
e
n
th
e
or
ig
in
a
l
a
nd
s
te
go
im
a
ge
s
by
c
ons
id
e
r
in
g
s
tr
uc
tu
r
a
l
in
f
or
m
a
ti
on,
lu
m
in
a
nc
e
,
a
nd
c
ont
r
a
s
t.
S
S
I
M
is
ba
s
e
d
on
c
om
pa
r
in
g
lo
c
a
l
pi
xe
l
in
te
ns
it
y
pa
tt
e
r
ns
nor
m
a
li
z
e
d
f
or
lu
m
in
a
nc
e
a
nd
c
ont
r
a
s
t.
W
he
r
e
μ
a
nd
μ
a
r
e
th
e
m
e
a
n
in
te
ns
it
ie
s
,
σ
2
a
nd
σ
2
a
r
e
t
he
va
r
ia
nc
e
s
a
nd
σ
is
t
he
c
ova
r
ia
nc
e
be
twe
e
n t
he
t
w
o i
m
a
ge
s
. T
he
c
on
s
ta
nt
s
1
a
nd
2
a
r
e
us
e
d
to
s
ta
bi
li
z
e
th
e
di
vi
s
io
n
w
it
h
w
e
a
k
d
e
nom
in
a
to
r
va
lu
e
s
.
S
S
I
M
va
lu
e
s
r
a
nge
f
r
om
-
1
to
1,
1
in
di
c
a
ti
ng
pe
r
f
e
c
t
s
tr
uc
tu
r
a
l
s
im
il
a
r
it
y.
3.2. P
ix
e
l
-
val
u
e
or
d
e
r
in
g s
c
h
e
m
e
M
os
t
d
e
ve
l
op
e
d
s
t
e
ga
n
og
r
a
p
hy
m
e
t
ho
ds
f
oc
u
s
s
o
le
l
y
o
n
e
m
b
e
d
d
in
g
a
s
e
c
r
e
t
m
e
s
s
a
g
e
i
n
to
a
p
r
e
-
d
e
t
e
r
m
i
ne
d
p
a
i
r
o
f
p
i
xe
ls
,
k
no
w
n
a
s
f
ix
e
d
p
ix
e
l
pa
i
r
p
os
i
t
io
n.
I
n
th
e
m
e
t
h
ods
c
i
te
d
in
[
2
9]
,
th
e
p
os
i
t
io
n
o
f
t
he
p
ix
e
l
p
a
ir
s
is
de
t
e
r
m
i
ne
d
be
f
o
r
e
a
s
s
e
s
s
i
ng
w
he
t
he
r
t
he
p
i
xe
l
pa
i
r
c
a
n
be
e
m
be
d
de
d
w
i
t
h
a
s
e
c
r
e
t
m
e
s
s
a
ge
.
T
he
r
e
is
c
on
c
e
r
n
t
ha
t
th
is
f
ix
e
d
pi
xe
l
p
a
ir
p
l
a
c
e
m
e
n
t
c
ou
l
d
be
p
r
o
b
le
m
a
t
ic
w
he
n
f
a
c
e
d
w
i
t
h
pi
xe
ls
w
ho
s
e
p
os
it
i
ons
a
r
e
no
t
a
dj
a
c
e
n
t,
a
s
s
ho
w
n
in
F
ig
u
r
e
3.
T
h
is
f
ig
u
r
e
i
l
lu
s
t
r
a
t
e
s
t
he
s
e
le
c
t
io
n
o
f
p
i
xe
l
pa
i
r
s
us
in
g
a
c
o
m
m
o
n
f
u
nc
t
io
n
us
ua
ll
y
us
e
d
t
o
c
ha
n
ge
a
2
-
d
i
m
e
n
s
i
o
na
l
a
r
r
a
y
i
nt
o
a
1
-
d
i
m
e
ns
i
o
na
l
a
r
r
a
y
,
a
s
f
o
u
nd
i
n
[
30
]
,
th
us
f
i
nd
i
ng
p
i
xe
l
p
a
ir
s
th
a
t
a
r
e
n
o
t
a
d
j
a
c
e
n
t,
a
s
s
e
e
n
i
n
p
a
ir
3
,
w
h
ic
h
ha
s
a
l
a
r
ge
d
i
f
f
e
r
e
n
c
e
va
lu
e
.
T
he
r
e
f
o
r
e
,
p
i
xe
l
p
a
i
r
s
l
ik
e
"
p
a
i
r
3
"
t
e
n
d
to
h
a
ve
t
he
po
te
n
ti
a
l
t
o
b
e
c
o
m
e
p
ix
e
l
s
t
ha
t
c
a
n
n
ot
b
e
e
m
be
d
de
d
w
i
t
h s
e
c
r
e
t
m
e
s
s
a
ge
s
.
F
ig
ur
e
3. A
pr
obl
e
m
i
n f
la
tt
e
n
-
a
r
r
a
y
P
V
O
I
n
t
hi
s
s
t
ud
y,
a
m
e
th
o
d
w
a
s
p
r
o
po
s
e
d
t
ha
t
e
x
p
lo
r
e
s
f
ou
r
di
f
f
e
r
e
n
t
w
a
ys
of
s
e
l
e
c
t
in
g
p
ix
e
l
pa
i
r
s
.
T
he
in
s
p
i
r
a
t
io
n
f
o
r
t
hi
s
m
e
t
ho
d
c
a
m
e
f
r
o
m
th
e
w
a
y
a
p
a
i
n
te
r
w
o
r
ks
.
W
it
h
hi
s
u
n
iq
ue
s
k
il
ls
a
n
d
s
ty
le
,
a
pa
in
te
r
c
a
n
m
ov
e
h
is
b
r
us
h
i
n
va
r
i
ous
di
r
e
c
ti
o
ns
,
w
h
e
t
he
r
h
o
r
iz
on
ta
l
ly
,
v
e
r
t
i
c
a
ll
y,
o
r
d
ia
go
na
l
ly
,
to
th
e
r
i
g
h
t
o
r
l
e
f
t
.
B
y
a
p
p
ly
i
ng
t
he
s
a
m
e
p
r
in
c
i
p
le
,
th
e
p
os
i
ti
o
n
of
c
o
lo
r
s
th
a
t
a
r
e
n
ot
m
uc
h
d
if
f
e
r
e
nt
c
a
n
be
f
o
u
nd
by
s
c
a
n
n
in
g
t
he
im
a
g
e
a
c
c
o
r
d
i
ng
to
th
e
a
s
s
um
e
d
b
r
us
h
m
ov
e
m
e
n
t
d
i
r
e
c
t
i
on
.
F
i
gu
r
e
4
i
l
lu
s
t
r
a
te
s
t
h
e
s
e
f
o
u
r
w
a
y
s
of
s
or
t
i
ng
p
ix
e
l
s
in
d
e
t
a
i
l
.
H
o
r
iz
on
ta
l
p
ix
e
l
-
va
lu
e
or
de
r
i
ng
(
H
-
P
V
O
)
c
a
n
b
e
s
e
e
n
in
A
l
go
r
i
th
m
1,
ve
r
ti
c
a
l
p
ix
e
l
-
va
l
ue
o
r
de
r
i
ng
(
V
-
P
V
O
)
in
A
l
g
or
i
t
h
m
2
,
d
ia
go
na
l
-
r
ig
h
t
p
ix
e
l
-
va
l
ue
o
r
de
r
in
g
(R
-
P
V
O
)
in
A
l
g
o
r
i
t
hm
3
,
a
nd
d
ia
go
na
l
-
l
e
f
t
p
ix
e
l
-
va
l
ue
o
r
de
r
i
ng
(
L
-
P
V
O
)
in
A
l
g
o
r
i
t
hm
4
.
B
y
a
pp
ly
i
ng
t
h
is
m
e
t
h
od
,
w
e
c
a
n
e
ns
u
r
e
th
a
t
th
e
s
e
q
ue
nc
e
o
f
pi
xe
ls
o
bt
a
in
e
d
w
i
ll
be
a
d
ja
c
e
n
t.
U
s
i
ng
t
h
is
m
e
t
ho
d
,
w
e
c
a
n
e
ns
u
r
e
th
e
s
e
le
c
te
d
p
i
xe
ls
w
i
l
l
be
a
d
ja
c
e
nt
,
t
hus
f
a
c
i
li
t
a
t
i
ng
a
n
e
f
f
ic
ie
n
t
p
i
xe
l
pa
i
r
s
e
l
e
c
ti
on
a
nd
a
m
o
r
e
e
f
f
e
c
t
iv
e
s
te
ga
no
g
r
a
p
hy
p
r
oc
e
s
s
.
A
lg
or
it
hm
1:
H
or
iz
ont
a
l
pi
xe
l
-
va
lu
e
or
de
r
in
g
I
nput
:
P
a
s
N
×
pi
xe
l
a
r
r
a
y
O
ut
put
:
P
_i
dx
a
s
a
l
is
t
of
pi
xe
l
pa
ir
pos
it
io
n
1
r
o
w
,
c
o
l
←
0
;
2
c
o
l
_
in
c
←
1
;
3
P
_
id
x
←
[];
4
w
hi
le
l
e
n
g
t
h
(
_
)
<
N
×
do
5
P
_
id
x
[]
←
(
,
)
6
c
o
l
=
c
o
l
+
c
o
l
_
in
c
A
lg
or
it
hm
2:
V
e
r
ti
c
a
l
pi
xe
l
-
va
lu
e
or
de
r
in
g
I
nput
:
P
a
s
N
×
pi
xe
l
a
r
r
a
y
O
ut
put
:
P
_i
dx
a
s
a
l
is
t
of
pi
xe
l
pa
ir
pos
it
io
n
1
r
o
w
,
c
o
l
←
0
;
2
r
o
w
_
in
c
←
1
;
3
P
_
id
x
←
[];
4
w
hi
le
l
e
n
g
t
h
(
_
)
<
N
×
do
5
P
_
id
x
[]
←
(
,
)
6
r
o
w
=
r
o
w
+
r
o
w
_
in
c
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
5
,
O
c
to
be
r
2025
:
3563
-
3575
3570
7
if
c
o
l
≥
th
e
n
8
c
o
l
=
c
o
l
+
1
;
r
o
w
=
r
o
w
+
1
;
10
c
o
l
_
in
c
=
−
1
;
11
e
ls
e
i
f
c
o
l
<
0
th
e
n
12
c
o
l
=
0
;
r
o
w
=
r
o
w
+
1
;
14
c
o
l
_
in
c
=
1
;
15
e
nd i
f
16
e
nd w
hi
le
7
if
r
o
w
≥
th
e
n
8
c
o
l
=
c
o
l
+
1
;
r
o
w
=
r
o
w
−
1
;
10
r
o
w
_
in
c
=
−
1
;
11
e
ls
e
i
f
r
o
w
<
0
th
e
n
12
c
o
l
=
c
o
l
+
1
;
r
o
w
=
0
;
14
r
o
w
_
in
c
=
1
;
15
e
nd i
f
16
e
nd w
hi
le
A
lg
or
it
hm
3:
D
ia
gona
l
-
r
ig
ht
pi
xe
l
-
va
lu
e
or
de
r
in
g
I
nput
:
P
a
s
N
×
pi
xe
l
a
r
r
a
y
O
ut
put
:
P
_i
dx
a
s
a
l
is
t
of
pi
xe
l
pa
ir
pos
it
io
n
1
r
o
w
,
c
o
l
←
0
;
r
o
w
_
in
c
←
−
1
;
3
c
o
l
_
in
c
←
1
;
P
_
id
x
←
[];
5
w
hi
le
l
e
n
g
t
h
(
_
)
<
N
×
do
6
P
_
id
x
[]
←
(
,
)
7
r
o
w
=
r
o
w
+
r
o
w
_
in
c
8
c
o
l
=
c
o
l
+
c
o
l
_
in
c
9
if
r
o
w
<
0
th
e
n
10
r
o
w
=
0
;
11
r
o
w
_
in
c
=
1
;
c
o
l
_
in
c
=
−
1
;
13
e
ls
e
i
f
r
o
w
≥
th
e
n
14
c
o
l
=
c
o
l
+
2
;
r
o
w
=
r
o
w
−
1
;
16
r
o
w
_
in
c
=
−
1
;
c
o
l
_
in
c
=
1
;
18
e
ls
e
i
f
c
o
l
<
0
th
e
n
19
c
o
l
=
0
;
20
r
o
w
_
in
c
=
−
1
;
c
o
l
_
in
c
=
1
;
22
e
ls
e
i
f
c
o
l
≥
th
e
n
23
r
o
w
=
r
o
w
+
2
;
c
o
l
=
c
o
l
−
1
;
25
r
o
w
_
in
c
=
1
;
c
o
l
_
in
c
=
−
1
;
27
e
nd i
f
28
e
nd w
hi
le
A
lg
or
it
hm
4:
D
ia
gona
l
-
le
f
t
pi
xe
l
-
va
lu
e
or
de
r
in
g
I
nput
:
P
a
s
N
×
pi
xe
l
a
r
r
a
y
O
ut
put
:
P
_i
dx
a
s
a
l
is
t
of
pi
xe
l
pa
ir
pos
it
io
n
1
r
o
w
,
c
o
l
←
0
;
r
o
w
_
in
c
←
−
1
;
3
c
o
l
_
in
c
←
−
1
;
P
_
id
x
←
[];
5
w
hi
le
l
e
n
g
t
h
(
_
)
<
N
×
do
6
P
_
id
x
[]
←
(
,
)
7
r
o
w
=
r
o
w
+
r
o
w
_
in
c
8
c
o
l
=
c
o
l
+
c
o
l
_
in
c
9
if
r
o
w
<
0
th
e
n
10
c
o
l
=
c
o
l
+
2
;
r
o
w
=
0
;
12
r
o
w
_
in
c
=
1
;
c
o
l
_
in
c
=
1
;
14
e
ls
e
i
f
r
o
w
≥
th
e
n
15
r
o
w
=
r
o
w
−
1
;
16
r
o
w
_
in
c
=
−
1
;
c
o
l
_
in
c
=
−
1
;
18
e
ls
e
i
f
c
o
l
<
0
th
e
n
19
c
o
l
=
0
;
20
r
o
w
_
in
c
=
1
;
c
o
l
_
in
c
=
1
;
22
e
ls
e
i
f
c
o
l
≥
th
e
n
23
c
o
l
=
c
o
l
−
1
;
r
o
w
=
r
o
w
−
1
;
24
r
o
w
_
in
c
=
−
1
;
c
o
l
_
in
c
=
−
1
;
27
e
nd i
f
28
e
nd w
hi
le
F
ig
ur
e
4. T
he
pr
opos
e
d
P
V
O
s
c
h
e
m
e
F
ur
th
e
r
m
or
e
,
th
e
pr
opos
e
d
m
e
th
od’
s
s
c
he
m
e
e
f
f
e
c
ts
on
th
e
e
m
be
ddi
ng
c
a
pa
c
it
y
a
r
e
il
lu
s
tr
a
te
d
in
F
ig
ur
e
5.
T
he
f
ig
ur
e
s
how
s
th
e
e
m
be
ddi
ng
of
6
bi
t
s
of
s
e
c
r
e
t
da
ta
(
11,
11,
a
nd
00)
in
to
a
3
×
5
gr
a
ys
c
a
le
im
a
ge
. T
he
f
ig
ur
e
c
om
pa
r
e
s
t
he
f
la
tt
e
ne
d pi
xe
l
a
r
r
a
y o
r
de
r
in
g
w
it
h our
P
V
O
s
c
he
m
e
s
. T
he
pi
xe
l
pa
ir
s
us
e
d i
n
e
a
c
h
s
c
he
m
e
a
r
e
hi
ghl
ig
ht
e
d
w
it
h
a
gr
e
e
n
bor
de
r
in
th
e
r
e
s
ul
ti
ng
s
te
go
im
a
ge
.
T
he
P
S
N
R
r
e
s
ul
ts
s
how
th
a
t
th
e
f
la
tt
e
ne
d
pi
xe
l
a
r
r
a
y
or
de
r
in
g
ha
s
th
e
lo
w
e
s
t
P
S
N
R
. T
hi
s
is
due
to
th
e
hi
gh
di
s
to
r
ti
on
w
he
n
a
ppl
ie
d
to
our
pr
opos
e
d
R
D
E
s
c
he
m
e
.
M
e
a
n
w
hi
le
,
th
e
hi
gh
e
s
t
P
S
N
R
r
e
s
ul
t
o
f
38.888
dB
is
a
c
hi
e
v
e
d
w
it
h
th
e
di
a
gona
l
-
le
f
t
pi
xe
l
or
de
r
in
g.
T
hi
s
in
di
c
a
te
s
th
a
t
th
e
c
ove
r
im
a
ge
c
h
a
r
a
c
te
r
is
t
ic
s
a
r
e
be
tt
e
r
s
ui
te
d
f
or
th
e
di
a
gona
l
-
le
f
t
pi
xe
l
or
de
r
in
g s
c
he
m
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
A
hy
br
id
s
te
ganogr
aphy
s
c
he
m
e
w
it
h r
e
du
c
e
d di
ff
e
r
e
nc
e
e
x
pan
s
io
n and
…
(
I
K
ade
k
A
gus
A
r
ie
s
ta
P
ut
r
a
)
3571
F
ig
ur
e
5. I
ll
us
tr
a
ti
on e
xe
m
pl
if
yi
ng t
he
e
m
be
ddi
ng pr
oc
e
s
s
a
nd i
ts
di
s
to
r
ti
on i
m
pa
c
t
4.
R
E
S
U
L
T
S
A
N
D
D
I
S
C
U
S
S
I
O
N
I
n
th
is
s
e
c
ti
on,
w
e
w
il
l
di
s
c
u
s
s
th
e
e
xp
e
r
im
e
nt
a
l
r
e
s
ul
ts
of
th
e
m
e
th
od
w
e
pr
opos
e
d.
T
h
e
e
xpe
r
im
e
nt
in
vol
ve
d
m
e
di
c
a
l
im
a
ge
pi
c
tu
r
e
s
s
iz
e
d
512
×
512
pi
xe
ls
a
s
a
m
e
di
um
f
or
da
ta
hi
di
ng,
in
c
lu
di
ng
a
bdomi
na
l
,
c
he
s
t,
ha
nd,
he
a
d,
a
nd
le
g
im
a
ge
s
f
r
om
th
e
publ
ic
da
ta
s
e
t
[
31]
.
T
he
e
xpe
r
im
e
nt
w
a
s
c
onduc
te
d
a
g
a
in
s
t
th
r
e
e
di
f
f
e
r
e
nt
pa
yl
oa
d
s
iz
e
s
,
5
kb
,
10
kb
,
a
nd
20
kb
,
c
om
pa
r
e
d
to
th
e
m
e
th
od
in
[
28]
w
hi
c
h
h
a
s
a
ls
o
s
how
n
th
e
im
pr
ove
m
e
nt
of
th
e
m
e
th
od
in
[
24]
.
A
ddi
ti
ona
ll
y,
th
e
e
xpe
r
im
e
nt
s
e
xt
e
nd
e
d
up
to
100
kb
to
de
m
on
s
tr
a
te
th
e
e
f
f
e
c
ts
of
P
V
O
on e
a
c
h i
m
a
ge
.
T
o
e
va
lu
a
te
t
he
qua
li
ty
of
th
e
s
te
go
im
a
ge
,
w
e
m
e
a
s
u
r
e
th
e
P
S
N
R
va
l
ue
i
n
de
c
i
be
ls
(
d
B
)
,
a
m
e
t
r
i
c
c
om
m
on
ly
us
e
d
to
e
va
lu
a
te
im
a
ge
qua
l
it
y
a
f
te
r
da
ta
in
s
e
r
ti
o
n
.
W
e
a
ls
o
in
c
lu
d
e
a
c
o
m
pa
r
is
on
be
twe
e
n
ou
r
m
e
t
hod
a
nd
th
e
e
xi
s
ti
ng
R
D
E
-
ba
s
e
d
d
a
ta
h
id
in
g
[
2
8]
,
to
s
how
t
he
s
tr
e
ng
th
s
o
f
ou
r
m
e
th
od.
W
e
a
ls
o
in
c
lu
d
e
d
P
S
N
R
m
e
a
s
u
r
e
m
e
n
t
r
e
s
ul
ts
ba
s
e
d
on
f
ou
r
w
a
ys
o
f
pe
r
f
o
r
m
i
ng
P
V
O
:
H
-
P
V
O
,
V
-
P
V
O
,
R
-
P
V
O
,
a
nd
L
-
P
V
O
.
T
o
e
va
lu
a
te
th
e
pe
r
f
or
m
a
nc
e
of
th
e
p
r
opos
e
d
m
e
th
od
a
ga
in
s
t
s
ta
te
-
of
-
th
e
-
a
r
t
te
c
hni
que
s
,
c
om
pa
r
a
ti
ve
r
e
s
ul
ts
a
r
e
pr
e
s
e
nt
e
d
in
T
a
bl
e
s
2
to
4,
c
or
r
e
s
ponding
to
pa
yl
oa
d
c
a
pa
c
it
ie
s
of
5
kb
,
10
kb
,
a
n
d
20
kb
,
r
e
s
pe
c
ti
ve
ly
.
T
he
e
xpe
r
im
e
nt
a
l
r
e
s
ul
ts
s
how
th
a
t
th
e
p
r
opos
e
d
m
e
th
od
c
ons
is
te
nt
ly
a
c
hi
e
ve
s
hi
ghe
r
P
S
N
R
va
lu
e
s
t
ha
n t
he
e
xi
s
ti
ng me
th
od
[
30]
. F
or
e
xa
m
pl
e
, a
t
a
p
a
yl
oa
d of
10
kb
, t
he
pr
opos
e
d s
c
he
m
e
r
e
c
or
ds
P
S
N
R
va
lu
e
s
of
58.42
dB
(
a
bdomi
na
l
)
,
59.75
dB
(
c
he
s
t
)
,
a
nd
60.03
dB
(
ha
nd
)
,
w
he
r
e
a
s
th
e
m
e
th
od
in
[
28]
a
c
hi
e
ve
s
56
.18
dB
,
57.33
dB
,
a
nd
58.07
dB
,
r
e
s
pe
c
ti
ve
ly
.
T
hi
s
im
pr
ove
m
e
nt
of
a
ppr
oxi
m
a
te
ly
2
to
3
dB
de
m
ons
tr
a
te
s
th
e
e
f
f
e
c
ti
ve
ne
s
s
of
our
pi
xe
l
di
f
f
e
r
e
nc
e
r
e
duc
ti
on
s
tr
a
te
gy
in
pr
e
s
e
r
vi
ng
im
a
ge
qua
li
ty
.
S
im
il
a
r
tr
e
nds
a
r
e
ob
s
e
r
ve
d
a
c
r
os
s
a
ll
te
s
te
d
pa
yl
o
a
d
s
i
z
e
s
.
R
e
ga
r
di
ng
e
m
be
ddi
ng
c
a
pa
c
it
y,
T
a
bl
e
5
s
ho
w
s
th
a
t
L
-
P
V
O
c
ons
is
te
nt
ly
a
c
hi
e
ve
s
th
e
hi
ghe
s
t
va
lu
e
s
,
0.959
bpp
f
or
a
bdomi
na
l
,
0.991
bpp
f
or
c
he
s
t
,
1.0
bpp
f
o
r
ha
nd
,
a
nd
0.997
bpp
f
or
le
g
im
a
ge
s
.
F
or
th
e
he
a
d
im
a
ge
,
R
-
P
V
O
de
li
ve
r
s
th
e
be
s
t
r
e
s
ul
t
w
it
h
0.958
bpp
.
T
he
s
e
f
in
di
ngs
hi
ghl
ig
ht
th
e
im
por
ta
nc
e
of
c
hoos
in
g
th
e
a
ppr
opr
ia
te
P
V
O
ba
s
e
d
on
im
a
ge
s
tr
uc
tu
r
e
to
m
a
xi
m
iz
e
qua
li
ty
a
nd c
a
pa
c
it
y.
A
ddi
ti
ona
ll
y,
th
e
pr
opos
e
d
s
c
he
m
e
e
nh
a
nc
e
s
bot
h
P
S
N
R
a
nd
e
m
be
ddi
ng
c
a
pa
c
it
y
by
a
da
pt
in
g
pi
xe
l
pa
ir
or
de
r
in
g
to
th
e
s
pe
c
if
ic
c
ha
r
a
c
te
r
is
ti
c
s
of
e
a
c
h
m
e
di
c
a
l
im
a
ge
.
H
-
P
V
O
a
c
hi
e
ve
s
th
e
hi
ghe
s
t
P
S
N
R
f
or
a
bdomi
na
l
a
nd
c
he
s
t
im
a
g
e
s
,
w
hi
le
R
-
P
V
O
is
m
os
t
e
f
f
e
c
ti
v
e
f
or
th
e
le
g
im
a
ge
.
H
a
nd
a
nd
he
a
d
im
a
ge
s
pe
r
f
or
m
be
s
t
w
it
h
e
it
he
r
R
-
P
V
O
or
V
-
P
V
O
.
A
lt
hough
ve
r
ti
c
a
l
or
de
r
in
g
s
how
s
lo
w
e
r
P
S
N
R
c
om
pa
r
e
d
to
ot
he
r
s
c
he
m
e
s
,
it
s
ti
ll
out
pe
r
f
or
m
s
th
e
m
e
th
od
in
[
28]
.
A
s
s
ho
w
n
in
F
ig
ur
e
6,
th
e
m
e
th
od
m
a
in
ta
in
s
s
tr
ong
P
S
N
R
pe
r
f
or
m
a
nc
e
a
c
r
os
s
pa
yl
oa
d
s
r
a
ngi
ng f
r
om
40 kb to 100 kb.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
5
,
O
c
to
be
r
2025
:
3563
-
3575
3572
T
a
bl
e
2. P
S
N
R
(
dB
)
a
nd S
S
I
M
va
lu
e
s
of
e
m
be
ddi
ng 5 kb
s
e
c
r
e
t
m
e
s
s
a
ge
I
m
a
ge
M
e
t
hod i
n
[
28]
P
r
opos
e
d
H
-
P
V
O
V
-
P
V
O
R
-
P
V
O
L
-
P
V
O
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
A
bdom
i
na
l
51.3193
0.9990
51.7980
0.9991
58.4020
0.9988
51.3800
0.9984
53.0570
0.9983
C
he
s
t
51.2390
0.9988
51.2460
0.9991
58.7630
0.9988
53.3610
0.9985
53.3760
0.9982
H
a
nd
52.1842
0.9987
51.2840
0.9994
57.3200
0.9997
51.0900
0.9989
57.2320
0.9992
H
e
a
d
53.9669
0.9990
53.8890
0.9990
59.2610
0.9988
53.8530
0.9985
59.8250
0.9983
L
e
g
51.4348
0.9990
50.6110
0.9992
55.8280
0.9995
58.4660
0.9995
49.2040
0.9979
T
a
bl
e
3. P
S
N
R
(
dB
)
a
nd S
S
I
M
va
lu
e
s
of
e
m
be
ddi
ng 10 kb
s
e
c
r
e
t
m
e
s
s
a
g
e
I
m
a
ge
M
e
t
hod i
n
[
28]
P
r
opos
e
d
H
-
P
V
O
V
-
P
V
O
R
-
P
V
O
L
-
P
V
O
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
A
bdom
i
na
l
48.1542
0.9974
51.7980
0.9980
58.4020
0.9969
51.3800
0.9967
53.0570
0.9961
C
he
s
t
50.3226
0.9981
51.2460
0.9977
58.7630
0.9969
53.3610
0.9968
53.3760
0.9961
H
a
nd
47.0188
0.9968
51.2840
0.9988
57.3200
0.9991
51.0900
0.9978
57.2320
0.9985
H
e
a
d
50.0749
0.9975
53.8890
0.9980
59.2610
0.9969
53.8530
0.9967
59.8250
0.9965
L
e
g
48.8994
0.9970
50.6110
0.9981
55.8280
0.9988
58.4660
0.9989
49.2040
0.9958
T
a
bl
e
4. P
S
N
R
(
dB
)
a
nd S
S
I
M
va
lu
e
s
of
e
m
be
ddi
ng 20 kb
s
e
c
r
e
t
m
e
s
s
a
g
e
I
m
a
ge
M
e
t
hod i
n
[
28]
P
r
opos
e
d
H
-
P
V
O
V
-
P
V
O
R
-
P
V
O
L
-
P
V
O
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
P
S
N
R
S
S
I
M
A
bdom
i
na
l
45.8774
0.9940
47.9570
0.9945
55.2730
0.9931
48.1260
0.9927
48.9220
0.9915
C
he
s
t
49.1203
0.9941
49.3070
0.9942
55.3900
0.9931
51.8170
0.9930
49.3900
0.9919
H
a
nd
45.3506
0.9947
48.3620
0.9974
54.0360
0.9980
49.0150
0.9960
54.6430
0.9972
H
e
a
d
48.6546
0.9941
50.1810
0.9946
55.6700
0.9930
50.6010
0.9934
55.6660
0.9926
L
e
g
48.0024
0.9960
48.5530
0.9956
52.6400
0.9975
52.7860
0.9970
46.9050
0.9924
T
a
bl
e
5. C
om
pa
r
is
on of
t
he
m
a
xi
m
um
e
m
be
ddi
ng
c
a
pa
c
it
y (
bp
p)
be
twe
e
n t
he
pr
opos
e
d m
e
th
od
a
nd t
he
s
ta
te
-
of
-
th
e
-
a
r
t
I
m
a
ge
M
e
t
hod i
n
[
24]
M
e
t
hod i
n
[
28]
P
r
opos
e
d
H
-
P
V
O
V
-
P
V
O
R
-
P
V
O
L
-
P
V
O
A
bdom
i
na
l
0.742
0.832
0.955
0.953
0.957
0.959
C
he
s
t
0.737
0.842
0.985
0.984
0.988
0.991
H
a
nd
0.750
0.881
1
1
1
1
H
e
a
d
0.711
0.740
0.952
0.952
0.958
0.954
L
e
g
0.746
0.862
0.992
0.991
0.994
0.997
F
ig
ur
e
6. T
he
obt
a
in
e
d P
S
N
R
va
lu
e
(
dB
)
f
or
40
-
100 kb s
e
c
r
e
t
m
e
s
s
a
ge
Evaluation Warning : The document was created with Spire.PDF for Python.