Indonesian J
ournal of Ele
c
trical Engin
eering and
Computer Sci
e
nce
Vol. 2, No. 3,
Jun
e
201
6, pp. 712 ~ 71
9
DOI: 10.115
9
1
/ijeecs.v2.i3.pp71
2-7
1
9
712
Re
cei
v
ed Ma
rch 2
4
, 2016;
Re
vised Ma
y 10, 2016; Accepted Ma
y 25
, 2016
A Review on LSB Substitution and PVD Based Image
Steganography Techniques
Adit
y
a
Kumar Sahu*, Gan
dharba S
w
ai
n
CSE, KL Unive
r
sit
y
, Vad
des
w
a
ram-5
225
02,
Andhr
a Prad
es
h, India
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: adit
y
asa
hu.c
s
e@gma
il.com
A
b
st
r
a
ct
T
here
has
bee
n a tre
m
en
dou
s grow
th in
Informatio
n
a
nd
Co
mmun
icati
o
n tech
nol
og
ies
duri
ng t
h
e
last d
e
cad
e
. Int
e
rnet
has
bec
o
m
e
the
do
mina
nt medi
a fo
r
da
ta co
mmu
n
icati
on. But th
e sec
r
ecy of th
e d
a
t
a
is to be taken
care. Stegano
grap
hy is a techni
que fo
r ac
hievi
ng secr
ec
y for the data communic
a
ted
i
n
Internet. T
h
is paper pr
ese
n
ts a review
of the stegan
ogr
aphy
techniq
ues b
a
s
ed on l
east si
gnific
ant bit (L
SB)
substitutio
n
an
d p
i
xel
val
u
e
differenc
ing
(P
VD). T
he
v
a
ri
ous tec
hni
qu
e
s
pro
pose
d
in
the
literat
ure
are
discuss
ed
an
d
possi
ble
co
mp
ariso
n
is
do
ne
alo
ng w
i
th t
heir
resp
ective
me
rits. T
he co
mp
ariso
n
p
a
ra
met
e
rs
consi
dere
d
are
,
(i) hidin
g
cap
a
city, (ii) distortion
me
as
ure, (iii) security, and (iv)
comp
utati
ona
l co
mpl
e
xit
y
.
Ke
y
w
ords
: cryptogra
phy; ste
gan
ogra
phy; le
ast signific
ant
bit substitutio
n
;
pixel va
lue
differenc
ing
Copy
right
©
2016 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
Cryptog
r
a
phy
and Stegano
grap
hy are th
e two widely
use
d
tech
niq
ues to provid
e se
cret
comm
uni
cati
on [1]. But the way they f
unctio
n
s
are
different. Cryptogra
phy is
the tech
niqu
e
of
transfo
rmin
g any plaintext into ciphe
rtex
t. St
eganogra
phy focu
se
s on ke
eping th
e existen
c
e o
f
a
messag
e se
cret. It was ori
g
inated fro
m
Gree
k words
Stegano
s (co
v
ered
) and G
r
apto
s
(writin
g
),
calle
d it as “covered
writin
g” [3]. Stegan
ogra
phy is
a
bran
ch
of se
curity whi
c
h d
eals
with hidi
ng
of informatio
n in any cover mediu
m
like imag
e, audio, video, and text [40-41]. The orig
inal
image
used f
o
r hi
ding
data
is
called
cover-im
age,
an
d t
he ima
ge a
fter hidin
g
se
cret
data into
it is
calle
d stego
-i
mage. Stega
nography techniqu
es a
r
e
prop
osed by different re
se
arche
r
s in bo
th
spatial d
o
mai
n
and tran
sfo
r
m domai
n.
Steganalysi
s
is a
n
a
r
t an
d
sci
en
ce
of di
sc
overin
g
the
hidd
en data from stego
-i
mage
[5
,
35]. This is something li
ke
cryptanaly
s
is. Many
stega
nalytic attacks pre
s
e
n
t on spatial do
mai
n
s
are RS a
naly
s
is, Hi
stog
ra
m analysi
s
, Chi-squa
re
atta
ck, Weighte
d
-Stego (WS
) a
nalysi
s
etc. The
efficien
cy of any steg
ano
grap
hic te
ch
nique
dep
e
n
d
s u
pon va
ri
ous
paramet
ers. Som
e
of
the
important
parameters are, (i) capa
city, (i
i) di
stortion
measure, (
iii) security or
attack resistance,
and (iv)
com
p
utational com
p
lexity. Capa
city is
the ma
ximum amou
nt of data tha
t
can be
hidd
en
insid
e
the i
m
age. It is
usu
a
lly rep
r
e
s
ent
ed in te
rm
s
o
f
bits pe
r
pixe
l. The di
storti
on in th
e
ste
go-
image can b
e
mea
s
u
r
ed by
pea
k
si
gn
al-to-noi
se
ra
tio (PSNR).
High
er the P
S
NR m
ean
s l
e
sser
is the di
storti
on. A good
stegano
gra
phi
c techniq
ue
should
be resi
stant to vario
u
s
stega
naly
s
is
attacks. The
comp
utationa
l complexity refers
to the time requi
red
to hide the data insi
de th
e
cover im
age.
The
requi
rem
ents fo
r a
go
od
stegan
og
raphi
c techniq
ue a
r
e
as
sh
own i
n
T
able
1. The
r
e
are different
t
e
ch
niqu
es
i
n
spatial dom
ai
n.
Tho
s
e
are,
(i)
Lea
st Sig
n
ificant Bit (L
SB) su
bstituti
on,
(ii) Pixel Value Differencin
g (PVD), Expl
oiting modifica
tion dire
ctio
n (EMD) etc.
Table 1. Perf
orma
nce parameters
Parameters
Requirement
Capacit
y
Should be High
Distortion
Should be Lo
w
Securit
y
Should be High
Computational C
o
mplexit
y
Should be lo
w
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 3, Jun
e
2016 : 712
– 719
713
The re
st of the pape
r is orga
nize
d
as fo
llo
ws. In se
ction 2
the LSB sub
s
titution
techni
que
s a
r
e represent
ed, in sectio
n 3 the PV
D techniq
u
e
s
are
de
scrib
ed. In se
ctio
n 4,
LSB+PVD te
chni
que
s are descri
bed a
n
d
in se
ction 5
the paper i
s
con
c
lu
ded.
2. LSB Subs
titution T
ech
niques
To a
c
hieve
secu
re
co
mmu
nicatio
n
in
sp
atial dom
ain t
he
se
cret
dat
a is emb
edd
e
d
in th
e
cover me
diu
m
. It ca
n b
e
achi
eved
by
modifying th
e
bits at the
L
SB plane
s.
L
SB sub
s
titutio
n
can
be extende
d u to 4-LSB p
l
ane
s to achi
eve high cap
a
city. John
so
n and Jajodi
a [2] propo
se
d
variou
s steg
a
nographi
c software tool
s li
ke S-T
ool
s a
nd EzStego t
ool usi
ng lea
s
t signifi
cant
bit
(LSB)
su
bstit
u
tion. Wang
et al [3] p
r
o
p
o
se
d a
ne
w
way of
emb
e
dding
bit to
L
SB plane
by
usin
g
pixel adju
s
tm
ent pro
c
e
s
s. In their meth
o
d
they
introd
u
c
ed th
e co
ncept of Mode
rately-Significant-
Bit (MSB) i.e. they targete
d
the interme
d
iate bit
of the cover im
ag
e. So by this mean
s the se
cret
data ca
n be
se
cured if an
y damage to
the LSB bit
is cau
s
ed in fut
u
re. Fu
rtherm
o
re, to impro
v
e
the stego
-im
age qu
ality it use
s
the co
nce
p
t of
pixel adjustm
ent pro
c
e
ss.
Wa
ng et al [4] again
came
up
with a uni
que i
dea to hi
de t
he data f
r
om
the interce
p
tor. They u
s
e
d
the con
c
ep
t of
geneti
c
algo
rithm to hide the valuable data in
the k-LSB
s
of the cover image. Although it
incr
ea
se
s t
h
e
comput
at
io
n
a
l t
i
me,
but provides g
r
eate
r
se
cu
rity.
To de
cre
a
se the co
mputati
onal time an
d
incr
ea
se the
hiding
cap
a
cit
y
, a novel techniqu
e
has
be
en pro
posed by Ch
ang et al [6].
This meth
o
d
use
s
the co
n
c
ept of dyna
mic pro
g
ram
m
ing
to identify the
optimal bit in LSB for substitution.
It uses the co
nce
p
t of “Principl
e
of optimality”.
This p
r
in
ciple
ensu
r
e
s
that
each su
b seq
uen
ce
of a so
lution mu
st be optimal. By experim
ent the
authors
have
prove
d
that,
the computati
onal time
is less
comp
are
d
to the
meth
od p
r
op
osed
by
Wan
g
et. al. [4] and also q
uality of the stego image i
s
better.
Cha
n
an
d Ch
eng [7] exten
ded the
wo
rk of Wa
ng et
al [4], they experime
n
tally sho
w
e
d
that the image quality of the stego
-i
mage
can
b
e
gre
a
tly enhan
ced
with low computat
iona
l
compl
e
xity. T
he autho
rs h
a
ve proved t
hat t
he worst
mean-sq
uare-e
rro
r (WM
SE) betwee
n
the
cover
ima
ge and steg
o-im
age ca
n
b
e
redu
ced
to
h
a
l
f
as compute
d
by
simpl
e
l
east
sig
n
ifica
n
t bit
method prop
ose
d
by Wan
g
et al. [4]. Swain an
d Len
ka [8] prop
osed a techni
qu
e, RGB cha
n
nel
based ste
gan
ogra
phy with
two levels
of se
curity.
In order to en
crypt the plaintext, RSA algorith
m
has b
een u
s
ed. The cip
h
e
rtext is emb
edde
d in any
two cha
nnel
s out of the available Re
d (R),
Gree
n
(G),
a
nd Blue
(B) chann
els. S
w
a
i
n and
Le
nka
[9] prop
osoed
a dyna
mic
st
egan
ography,
it
is a uniqu
e a
nd se
cu
red t
e
ch
niqu
e by combi
n
i
ng
cryptograp
hy and steg
anog
raphy. In orde
r to
encrypt
the messag
e
the block ciphe
r cryptog
r
a
phi
c algorithm
wit
h
block le
ngt
h of 128
key
has
been
use
d
. T
h
ree
LSB bits (6
th
, 7th, a
nd 8th)
are
u
s
ed to
hide t
w
o bit
s
of cip
her text in ea
ch
pixel of the i
m
age. Expe
ri
mental results reveal
th
at the ste
go-im
age q
uality will be p
r
e
s
e
r
ved
.
Gutub et
al.
10
propo
se
d a
novel te
chni
q
ue
calle
d pix
e
l indi
cato
r te
chni
que.
The
last two
LSB
bits
from any on
e of the available R, G,
and B ch
ann
el is ch
osen
as indi
cato
r to indicate t
he
pre
s
en
ce
of
data. The
ch
oice
of
choo
sing
the
in
dicator bit i
s
do
ne rand
omly. The i
ndi
cato
r i
s
sele
cted o
n
the ba
sis of seque
nce
of R, G, and B channel
s. If R is indicato
r in the first pixel the
n
G and B ch
a
nnel are the
cha
nnel 1 a
n
d
2 re
spe
c
tively. Similarly G and B ch
annel
s are th
e
indicators for
se
con
d
and t
h
ird pixel
s
.
To increa
se t
he ca
pa
city, a better tech
nique h
a
s b
e
en propo
sed
by Parvez a
n
d
Gutub
[11]. The co
nce
p
t behi
nd
this meth
od
is, embe
d
data bits i
n
l
o
w valu
e ch
annel i
n
ste
a
d
o
f
embed
ding in
high value chann
els. Sup
pose ea
ch of
the values of
the R, G, and B chan
nel
s is
255, and the
n
cha
nge in
any cha
nnel
s will lead to distortio
n
. So the chan
nel
which ha
s less
value
will be
approp
riate f
o
r d
a
ta e
m
be
dding
rath
er t
han
highe
r v
a
lue
ch
annel.
As thi
s
m
e
th
od
inse
rts differe
nt numbe
r of bits depe
ndi
ng upo
n the
cha
nnel valu
es, this can
be also call
e
d
as
variable
-
bit
s
per
ch
ann
el. The
se
cu
rity and
cap
a
city is incre
a
sed
comp
are
d
to the e
a
rli
e
r
prop
osed pix
e
l indicator te
chni
que by G
u
tub et al. [10].
The pixel
in
dicato
r te
chn
i
que
althoug
h in
cre
a
ses
the security
but capa
city is
not
predi
ctabl
e.
Furthe
rmo
r
e
to in
crea
se
the
se
cu
rity and to
en
han
ce th
e
capa
city, a
more
rand
omi
z
atio
n tech
niqu
e i
n
sel
e
ctin
g th
e pixel to
sto
r
e data h
a
s
b
een p
r
op
ose
d
by Gutub
e
t
al.
[12]. This
alg
o
rithm i
s
call
ed tripl
e
– A
algorith
m
. Th
e algo
rithm
u
s
e
s
the
simil
a
r p
r
o
c
ed
ure
as
use
d
by LSB method but
with more ra
ndomi
z
ation.
At first the data is en
crypte
d by using A
ES
algorith
m
to
prod
uce
ciph
ertext. By usi
ng a
p
s
eu
do
rand
om
num
ber ge
nerato
r
(P
RNG)
wh
ich
gene
rate
s two rand
om nu
mbers pe
r ea
ch num
ber
of
iterations, let
it be seed1 and se
ed2. T
he
see
d1 rand
o
m
numb
e
r i
s
use
d
to locate the
chann
el
whe
r
e the
ci
phertext can
be emb
edd
ed
and
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
A Revie
w
on
LSB Substitution and PVD
Based Im
age
Steganog
rap
h
y Te
chni
que
s (Aditya K.S.)
714
see
d2 i
s
use
d
to find
ho
w many
bits
ca
n be
em
bedd
ed. Thi
s
algo
rithm p
r
ovide
s
b
e
tter
se
cu
rity
and capa
city.
In ord
e
r to
compa
r
e the
para
m
eters
such
as
security and capa
city provid
ed
by pixel
indicator a
nd
triple-A al
gori
t
hm, Tiwari
a
nd Shan
dilya
[13] experim
entally
ob
served that there is
an increa
se
of 14% capa
city if we use more n
u
m
ber of bits in
an image in
case of triple-A
algorith
m
. Also hi
gh security can
be
achi
eved b
y
using p
s
e
udorand
om
numbe
r ge
n
e
rato
r
(PRNG) i
n
th
e triple
-A alg
o
rithm. So a
s
per the
re
sul
t
the triple-A
algorith
m
is
more
efficient
in
terms of capa
city compa
r
e
d
to pixel indicator m
e
thod.
The
pixel in
di
cator te
chni
q
ue a
nd t
r
iple
-A algo
rithm
u
s
e
s
the
lea
s
t
two
LSB bit
s
of a
n
y
one
of the
chann
els u
s
ed
as in
dicato
r,
by thi
s
the
cap
a
city of
the
stego
ima
ge i
s
limite
d
. T
o
further imp
r
o
v
e the hi
ding
ca
pa
city Kaur et
al
. [14] i
n
trodu
ce
d a
novel te
chniq
ue of
embe
d
d
ing
data in the 4
LSB bits of a
n
image. On
e
of the
three
cha
nnel
s is
u
s
ed a
s
in
dica
tor. The data
is
store
d
i
n
tha
t
cha
nnel
wh
ose
colo
r val
ue i
s
small.
The i
ndi
cator is sele
cted
depe
nding
o
n
spe
c
ific ra
ng
e. Three
ran
ge of
colo
r v
a
lue
s
su
ch a
s
{
0
, 85},
{8
5, 170
}, {17
0
,
255}
ha
s b
een
spe
c
ified. If the colo
r valu
e of a chann
el falls in the first ran
ge {0,
85} then we
can st
ore 4 bi
t o
f
se
cret
data in
the ch
ann
el. Similarly 2 bit
and n
o
bi
t of
data can b
e
hidde
n if the
colo
r value i
s
in
the se
co
nd
ra
nge a
nd thi
r
d
ran
ge, which
indicates tha
t
more
numb
e
r of
data bit
s
can
be hi
dd
en
in the cha
n
n
e
l with lower range. So this metho
d
increa
se
s the
security and
capa
city of an
image.
To in
crea
se t
he
capa
city a
nd imp
e
rcepti
b
ility
of the st
ego im
age, i
n
[15] the a
u
th
or h
a
s
prop
osed
an
ada
ptive LS
B bit method
. The diffe
re
nt ran
ge
of the
st
ego key
determine
s the
amount of bits to hide in the LSB plan
e of an
image. The uniqu
ene
ss of this techniqu
e is to
verify the inte
grity of the
hi
dden
data
by usi
ng
se
le
ct
ed bit
key. In
addition to
th
at this te
chni
que
hide
s 4.15 bi
ts in each pi
xel of an image. This
m
e
thod achiev
es better
se
curity and hig
her
cap
a
city. Swain and
Len
ka [37] pro
p
o
s
ed
a ne
w tech
niqu
e ba
sed on LSB a
rray. Depe
nd
ing
upon
the le
n
g
th of the
se
cret
me
ssag
e the
re
spe
c
tive LSB array is
cho
s
e
n
.
This te
ch
ni
que
provide
s
two
levels of se
curity with a
c
c
eptabl
e hid
i
ng ca
pa
city. The com
parison am
ong t
he
variou
s LSB tech
niqu
es h
a
s
bee
n don
e in the Table 2.
Table 2. Co
m
pari
s
on a
m
on
g the variou
s LSB techniq
u
e
s
Ref
No.
Capacit
y Securit
y
Distortion
Complexit
y
Advantage
3
Moderate
Moderate
High
Lo
w
Data can
be e
m
bedded in
mo
deratel
y significant-bit
(MSB). So less chance of loss of data in future.
4
High
Moderate
Moderate
High
Capacit
y
is high.
6
Moderate
Moderate
Lo
w
Lo
w
Qualit
y
of stego
-
i
mages obtained is much better w
i
th
low
computation
a
l time than simple LSB-substitution
method.
7
Moderate
Moderate
Lo
w
Lo
w
Lo
w
computatio
nal complexit
y
and high quality of
stego-images obt
ained Compar
ed
to method [3] [4
].
8
Moderate
High
Lo
w
High
T
w
o levels of
securit
y
is provided t
o
secret data.
9
Moderate
High
Moderate
Moderate
Additional se
curit
y
can be achieve
d
to secret data.
10
Moderate
Moderate
High
Lo
w
Ne
w
direction of
data embedding.
11
Moderate
High
Lo
w
Lo
w
Securit
y
to secre
t
data is high.
12
High
High
Lo
w
Lo
w
More capacit
y
a
nd se
curit
y
comp
ared to pixel indicator
technique [11].
14
Moderate
Moderate
Moderate
Moderate
Capacit
y
is increased due to 4 bit
s
LSB embedding.
15
High
High
Moderate
High
High securit
y
to
hidden data.
37
Moderate
High
Moderate
High
Securit
y
is high.
3. PVD Tech
niques
A ne
w
con
c
e
p
t call
ed
pixel value
diffe
ren
c
ing
(PV
D
) h
a
s be
en
p
r
opo
se
d by
Wu
and
Tsai
[16] in
th
e field
of ima
ge
stega
nog
raphy fo
r g
r
ay
image
s. T
he
PVD metho
d
divides the to
tal
image i
n
to
smooth a
nd
e
dge a
r
e
a
s. T
he differen
c
e
value d i
s
cal
c
ulate
d
bet
ween the
two
p
i
xels.
A rang
e table
has
bee
n sp
ecified fo
r th
e value d. A
new
differen
c
e value d'
re
places th
e ol
d d
value to emb
ed the secret
data. The wi
dth of t
he ran
ge table d
e
ci
des the
num
ber of bit that
is
allowed to e
m
bed. Thi
s
method give
s better result
s in te
rms
of
imperce
ptibility and capa
cit
y
as
comp
ared to LSB techniq
u
e
s.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 3, Jun
e
2016 : 712
– 719
715
Zhang a
nd
Wan
g
[17] noticed that, the origi
nal PVD techni
qu
e prop
osed b
y
Wu and
Tsai [16]
i
s
vulnera
b
le to
histog
ram
-
b
a
se
d ste
gan
alysis. Th
at i
s
the
stego
-i
mage exhi
bit an
abno
rmal
be
havior i
n
the
histog
ram. T
hey propo
se
d
a p
s
eu
do-ra
ndom
ditheri
n
g to get
dyna
mic
rang
e of values inste
ad of
static or fixed r
ang
e for the blocks. This tech
niqu
e
preserve
s the
advantag
es li
ke
cap
a
city o
f
origin
al PVD an
d
al
so a
v
oids
the
un
usu
a
l
be
havi
o
r sho
w
n by the
histog
ram.
Hence the
security is
an ad
ded advantag
e.
Ch
ang
et
a
l
.
[20] found
that the
ca
pa
city
of the PVD tech
niqu
e pre
s
ente
d
by Wu and T
s
ai
[1
6] can
be in
crea
sed fu
rthe
r for a g
r
ay-l
evel
image. The
r
e
is a gain of
84.16% on averag
e hi
di
ng ca
pa
city by the
overlappin
g
co
ncept
prop
osed by the autho
rs
while maintaini
ng sati
sfacto
ry image quali
t
y.
A new
stega
nographi
c m
e
thod h
a
s
be
en propo
se
d
by
Wan
g
et
al [21] to min
i
mize the
distortio
n
on
the stego-i
m
age. The
prop
osed
m
e
thod is the
combin
ation
of pixel va
lue
differen
c
ing
(PVD) an
d m
odulu
s
fun
c
ti
on. At fi
rst the differe
nce
value is
co
mputed from
two
con
s
e
c
utive pixels by app
lying PVD method. The
di
fference valu
e sugg
est
s
the numbe
r of bits
to hide.
The
n
by u
s
in
g
modulo
op
eration the
re
mainde
r of
the
two con
s
ecutive pixel
s
i
s
derieve
d, an
d the
secret
data a
r
e
hid
d
en in
the
pix
e
ls
by alte
rin
g
the
re
main
der. Expe
rim
ental
results
reveal
that the u
s
e
of modul
o o
peratio
n
g
r
ea
tly minimize
s the di
stortio
n
and
incre
a
s
e
s
the attack resistan
ce. To i
n
cre
a
se the
ca
pacity of o
r
igi
nal PVD te
ch
nique
pro
p
o
s
ed by Wu a
n
d
Tsai
16
, a
steg
anog
rap
h
ic te
chni
que
calle
d Tri
-
Way PV
D h
a
s be
en
p
r
opo
se
d by
Chang
et al. [2
2]
by usi
ng
2×2 pixel
blocks with
mult
i-directio
nal
differen
c
e
s
.
It has
bee
n
experi
m
ent
ally
con
c
lu
ded
th
at the
cap
a
cit
y
and
se
cu
rit
y
can
be
furth
e
r
enha
nced
comp
ared to
the o
r
iginal
P
V
D
method.
A novel lossl
e
ss data hi
ding tech
niqu
e
has b
een p
r
opo
sed by Li
n et al [23] by taking
three non overlap
p
ing pixel
blo
c
ks hav
ing
two
ab
sol
u
te
differe
nces calle
d
a
s
block
diffe
ren
c
e.
Experimentall
y
it has b
een
prove
d
by th
e autho
rs
tha
t
the avera
g
e
embe
dding
capa
city can
b
e
increa
sed. T
h
is h
a
s b
een
observed th
at the PVD
method int
r
o
duces
disto
r
tion to steg
o-i
m
ag
e
no matter ho
w much the cap
a
city is re
duced. So
ke
eping thi
s
in view by avoiding mo
re da
ta
embed
ding to
smooth re
gi
ons Lu
o et al [24] pr
opo
se
d a new way of embeddi
ng
secret data to
cover imag
e.
At first the im
age i
s
pa
rtitio
ned into
sm
al
l squ
a
res. T
h
e sq
ua
re
s are furthe
r ma
de
rotation
of 0,
90,18
0, and
270
deg
re
e
s
. The
two
d
i
fference val
ue of th
ree
non-overl
appi
ng
con
s
e
c
utive
pixels i
s
cal
c
ulated
and
mi
ddle
pixel i
s
use
d
to
hide
t
he
se
cret dat
a. The
am
ou
nt of
se
cret
bits
wil
l
be e
m
be
dd
ed de
pen
ding
upon
the differen
c
e
s
amo
ng
the
th
re
e non-overl
appi
ng
con
s
e
c
utive p
i
xels. The pro
posed techni
que re
si
sts to
PVD histogram analysi
s
.
Wan
g
et al [21] in 2008
prop
osed a
new di
re
ction
for data em
beddi
ng whi
c
h is the
combi
nation
of PVD and
modulu
s
fun
c
tion. Althoug
h good
ca
pa
city is achieve
d
but the security
is not imp
r
ov
ed. So Jo
o et
al [25]
propo
sed
a novel
method to im
prove
the ste
go-im
age qu
ality
whi
c
h
will ensure the security for the secret
data.
A
c
cording to them the al
gorithm is
divided i
n
to
four
step
s. T
he first on
e i
s
the
pixel p
a
irin
g
ste
p
where
the
cov
e
r im
age i
s
divided into
two
con
s
e
c
utive p
i
xels of no
n-
overlap
p
ing
sub-bl
o
cks. Se
con
d
ly in the
embed
ding
st
ep by u
s
ing t
h
e
modulu
s
fu
nction the
pixel
value i
s
in
cre
a
se
d o
r
red
u
c
ed
to m
a
tch
with th
e m
e
ssag
e. Thi
r
dly
in
the
adj
ustin
g
step
it solve
s
the
out
-of-su
b-rang
e p
r
obl
em
so th
at th
ere i
s
no va
ri
ation in
the P
V
D
histogram. Fi
nally in the last step
, if the
pixel value
goes
beyond th
e range of 0 t
o
255 then it will
bring b
a
ck int
o
the rang
e. The re
sult
s sugge
st
that this method p
r
o
v
es to be better co
mpa
r
ed
to
Wang et al. [
21] in terms
of s
e
c
u
rity and c
apac
i
ty.
Hon
g
et al [26]
propo
sed
a new te
ch
nique of d
a
ta embe
ddin
g
using th
e d
i
amond
encodin
g
(DE
)
techniq
ue.
A multiple-ba
s
e n
o
ta
tional
system
(MBNS) ha
s b
een i
n
trodu
ce
d u
s
i
ng
modified di
a
m
ond
encodi
ng. The
pro
p
o
se
d meth
od
m
odifies the
diamon
d en
coding to
emb
ed in
multiple ba
se
s and
solves the overflow and
unde
rfl
o
w problem.
Experimenta
lly it has been
sho
w
n
that t
he p
r
op
osed
tech
niqu
e h
a
s
bette
r em
beddi
ng cap
a
city
an
d
tol
e
rant agai
nst
RS
sc
heme a
nd
hist
og
ram a
n
a
ly
sis
st
ega
n
a
ly
sis at
t
a
ck.
A new
way o
f
data e
m
be
d
d
ing
pro
p
o
s
e
d
by M
and
al and
Das [27
]
whi
c
h
exten
d
s th
e
PVD techni
q
ue to colo
r im
age
s. Each pi
xel have 24
b
i
ts contai
ns
R, G, B compo
nents. All the
3
colo
r co
mpon
ents are use
d
for data em
beddi
ng. Initially the difference value
for each blo
c
k
can
be fou
n
d
by
|
|, where
and
are t
he two
co
nsecutive n
on-overlap
p
ing
pixels of a
n
cover-im
age.
The differen
c
e value d
e
termines ho
w m
any bits
will
be em
bedd
e
d
in
whi
c
h
com
p
o
net of a pixel.
Basin
g
up
on
the co
nt
rib
u
tion of
R, G, B
comp
one
nts i
n
a
colo
r ima
ge
the maximum
secret bits th
at can be e
m
bed in ea
ch
of R, G, B compon
ent of a pixel will be
5
,
3, 7 bits re
spectively. Ag
ain for emb
e
dding
of se
cret bits it uses th
e o
r
igin
al PVD
con
c
ept
prop
osed
by Wu
and
Tsai [16]. Thi
s
results
of
thi
s
schema
re
veals that
be
tter steg
o-im
age
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
A Revie
w
on
LSB Substitution and PVD
Based Im
age
Steganog
rap
h
y Te
chni
que
s (Aditya K.S.)
716
quality an
d secu
rity co
mp
ared
to o
r
igin
al PVD
co
ncept propo
se
d
by Wu a
nd
Tsai [1
6] al
so
the
falling-off bou
ndary p
r
obl
e
m
can be avo
i
ded.
In 2012, Le
e
et al [28]
pro
p
o
se
d a meth
o
d
whi
c
h in
cre
a
se
s the
cap
a
city of steg
o
-
imag
e.
This te
ch
niq
ue u
s
e
s
the
JPEG2
000
comp
re
ssion
and t
r
i-way
pixel value
differen
c
in
g
for
embed
ding t
he secret im
age. The
pro
posed meth
o
d
is u
s
eful fo
r sendin
g
la
rge secret im
age
without any d
i
stortion. To i
n
cr
ea
se the capa
city and secu
rity
of stego-im
age, Ch
ang an
d Tse
n
g
[29] propo
se
d
a novel tech
nique
called t
w
o, thre
e,
four side
d sid
e
match meth
o
d
. The pixels
are
visited in ra
ster scan o
r
de
r. In t
he two sided si
de mat
c
h ste
gan
ography, let
P
be the target pixe
l
whe
r
e secret data will be e
m
bedd
ed an
d
g
be the gray value for
P
. Let
g
and
g
be the gray
values for the upper pixe
l
P
and left pixel
P
of a
target pixel
P
. Th
e difference value d is
cal
c
ulate
d
as d= (
g
+
g
)/2 -
g
. If the difference value are in the range
of -1 to 1 the
n
there is
only 1 bit allowe
d to embed in the LSB bit of th
e target pixe
l
P
, otherwise, if d >1 then
b=
log
|d|
, bits are
allowed to e
m
bed. A n
e
w value is a
ssi
gned to
the
differen
c
e val
ue d
an
d
target pixel
g
.
At times the new valu
e of the pixel
P
may fall off the boun
dary of the ran
ge {0,
255}. Any pix
e
l that suffers with fall off bound
ary p
r
obl
em (F
OBP) will be not con
s
idere
d
for d
a
t
a
embed
ding.
The three
sid
ed si
de mat
c
h method
ha
ve
three va
ri
ants. In first variant the th
ree
neigh
bori
ng p
i
xels su
ch
as
uppe
r, left and right a
r
e
u
s
ed. In variant
2 instea
d of ri
ght the botto
m
pixels
tak
e
n
along with upper
and
left. Left-upper, right-upper,
le
f
t
-bottom
and right-bottom are
taken to find
the difference value in
case of
last variant. Similarly upper, left, right and bottom
neigh
bori
ng
pixels are exploited for se
cret data
e
m
beddi
ng in a target pixel in four sid
ed
side
match
metho
d
. Thi
s
met
h
od h
a
s the
cl
ear advanta
g
e
of m
o
re
ste
go-im
age
cap
a
city an
d b
e
tter
se
curity co
m
pare
d
to LSB techni
que
s.
The
cap
a
city
and
qu
ality of the
stego
-image
plays
a vital rol
e
f
o
r a
ste
g
o
-
i
m
age i
n
se
cret d
a
ta communi
catio
n
, In this reg
a
r
d Lia
o
et al [30], have pro
posed a te
ch
nique
calle
d four
pixel differen
c
ing a
nd mod
i
fied LSB sub
s
titution. In this wo
rk the
cover image i
s
sepa
rated in
to
non-overl
appi
ng four pixel
blocks having
gray va
lue
s
. The average
differen
c
e va
lue (k) i
s
use
d
to locate the
rang
e. Th
e
concept of
mo
dified LSB
su
bstitution i
s
u
s
ed
to e
m
be
d k-bit
s
of
da
ta
bits in the p
i
xels located
in that block. As
this te
chni
que i
s
h
i
ghly inclin
ed
towards
LSB
sub
s
titution, so the stego
-i
mage ha
s le
ss attack
re
si
stance, but the hiding
capa
cit
y
is more.
The capa
city of stego-i
m
age an
d se
curity of
secre
t
data have major
role b
ehind th
e
su
ccess of any stegano
graphi
c algo
rith
m. Yang et al [31] s
ugge
sted a new te
chniqu
e to ach
i
eve
this.
In contra
st
to Wu and
Tsai
[16] wh
ere a
p
a
ir
of
pixels
are
pro
c
e
s
sed
at a ti
me, the a
u
th
ors
con
s
id
ere
d
two pai
r of pixels for pro
c
e
ssi
ng. The
r
e are three
ways the fo
ur pixels
ca
n be
grou
ped. T
h
e groupin
g
o
f
pairs of two pixels is
d
one by ta
kin
g
the verti
c
a
l
, hori
z
ontal
and
diago
nal p
a
irs. Also
to p
r
event the fall
-off bou
nda
ry pro
b
lem th
e
y
prop
osed
a
shifting
sche
ma.
The p
r
opo
se
d method
avoids the
Frid
rich et al.’s
de
tection [5] wit
h
improve
d
h
i
ding capa
city. In
ca
se
of PVD
techni
que
the
more th
e diff
eren
ce
value
s
the
mo
re
th
e data
that
ca
n be
em
bed
d
ed.
But to em
be
d mo
re
data
so
as to i
n
crease the
c
a
pac
i
ty sometimes the
pixel values
c
r
oss
the
boun
dary
values. If th
e p
i
xel value
s
e
x
ceed
the
b
ound
ary valu
es th
en, thi
s
is called
fal
l
-off
boun
dary p
r
o
b
lem. Swai
n
and L
e
n
k
a [3
2]
marked thi
s
issu
e an
d then p
r
op
ose
d
revised va
riants
of two, three
and four
side
d side mat
c
h
with high
er e
m
beddi
ng ca
pacity.
Swain
33
ha
s
prop
osed a
stegano
gra
phi
c techniq
ue u
s
ing
pixel val
ue differen
c
in
g. There
are fou
r
diffe
rent metho
d
s and ea
ch h
a
s thei
r uniq
ue idea to fi
nd the differe
nce valu
e. In five
neigh
bors dif
f
eren
cing me
thod the differen
c
e valu
e
is calculate
d
by taking the differen
c
e
of
maximum to
minimum of
gray values
among fi
ve pixels nam
ely right, upper,
left, upper-right,
bottom an
d u
pper-ri
ght
co
rner
pixel of
a targ
et pixel
.
In six n
e
igh
bors diffe
ren
c
ing
metho
d
the
differen
c
e val
ue is cal
c
ulat
ed by same
way as in five neighbo
r differen
c
in
g method with on
e extra
pixel a
s
u
ppe
r-left
co
rne
r
.
Similarly for
seven
nei
ghb
ors differen
c
i
ng meth
od
o
ne extra
pixe
l as
bottom left corne
r
an
d for eight neigh
b
o
rs
differen
c
i
ng method o
ne more extra pixel su
ch
as
bottom left co
rne
r
. Experim
ental study
shows that the
quality of the im
age is
better in
case of five
neigh
bors differen
c
in
g and
the cap
a
city is hi
gh
er in ei
ght neigh
bors differenci
ng.
Pradh
an et a
l
[38] propo
sed a pixel va
lue differe
nci
ng tech
niqu
e
base
d
on P
V
D call
ed
two neigh
bo
r method, thre
e neighb
or m
e
thod and fo
ur
neig
hbo
r method. The
result reveal
s that
the ca
pa
city is go
od in
fou
r
nei
ghbo
r m
e
thod
with
a
c
cepta
b
le
steg
o-ima
ge q
uali
t
y. An adaptive
pixel value dif
f
eren
cing
met
hod u
s
ing ve
rtical
an
d
ho
ri
zontal edge
s has been pro
posed
by
Swain
[39]. Two te
chniqu
es i
s
gi
ven a
s
first o
ne u
s
e
s
2
×
2
pixel blo
c
ks
a
nd
se
cond
on
e u
s
e
s
3
×
3
p
i
xel
blocks. The f
i
rst te
chniq
u
e
offers go
o
d
ca
pa
city and second
o
ne provide
s
good
steg
o-i
m
age
quality.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 3, Jun
e
2016 : 712
– 719
717
The co
mpa
r
ison between v
a
riou
s PVD tech
niqu
es h
a
s
bee
n don
e in the Table 3.
Table 3. Co
m
pari
s
on b
e
tween vario
u
s P
V
D method
s.
Ref
No
Capacit
y
Securit
y
Distortion
Complexit
y
Advantage
16
Moderate
Moderate
Moderate
Lo
w
Provides a ne
w
method for
hiding secret data.
17
Moderate
High
Lo
w
Lo
w
Avoids the histo
g
ram steganal
y
s
is occurred in Wu
and Tsai [16] me
thod.
20
High
Moderate
Moderate
Moderate
An increase in 8
4
.16%
ofcapacity compa
r
ed to
P
V
D
method pro
pose
d
b
y
Wu an
d Tsa
i
[16].
21
High
Moderate
Moderate
High
Better stego-ima
ge qualit
y
and securit
y
compare
d
to
Wu and Tsai [16]
.
22
High
High
High
High
Capacit
y
and securi
t
y
is in
creased compared to Wu
and Tsai [16].
24
Moderate
High
Moderate
High
Securit
y
to steg
o-image is bette
r compared
to
Ref.
[17] and Ref.
[18]
.
25 High
High
Moderate
High
Minimum
differe
nces in the PVD
histograms bet
ween
the cover and st
ego-images com
pared to Wan
g
e
t
al
[21].
26
High
High
Lo
w
High
Better embedd
ing performan
ce compared to
previous PVD-based methods in terms of pay
l
oad
and image quality.
27
Moderate
High
Lo
w
Moderate
A ne
w
direction f
o
r data emb
eddi
ng in color images.
28
High
High
Lo
w
High
Capacit
y
and Se
curit
y
are high.
29
High
Moderate
Moderate
Lo
w
Better capacity
a
nd securit
y
compared
to
conventional LSBs s
ubstitution m
e
thod.
30
High
Moderate
Moderate
Lo
w
Capacit
y
is more
compared to LS
B methods.
31
High
Moderate
Moderate
High
More edge area
s present compared to Wu and Tsai
[16], So more ca
pacity
.
32
High
Moderate
Moderate
Moderate
Falling-off bound
ary
p
r
oblem cau
s
ed in Ref. [29]
is
addressed.
33 Moderate
Moderate
Moderate
Moderate
Adds
more flex
i
b
ility
in choosing data embeddi
ng
method.
4. LSB+PVD
Techniqu
es
The
LSB techniqu
e p
r
ovi
des the
larg
er
ca
pa
city and PV
D te
chni
que
offers hi
ghe
r
se
curity. In t
h
is di
re
ction,
Wu
et al [1
8]
in
20
05 h
a
s pro
p
o
s
ed a
techniqu
e by
com
b
inin
g
the
advantag
e of LSB and PVD to increa
se
the cap
a
city and ste
g
o
-
im
age qu
ality. The ra
nge val
ue
0 to 25
5 is p
a
rtitioned
into
two p
a
rts,
o
ne is
lo
wer level ra
nge if
the value i
s
0 to 15, a
nd
the
other on
e is
uppe
r level ra
nge 16 to 25
5. The differe
nce valu
e is
obtaine
d by using
simple P
V
D
strategy
16
. As opp
osed to
Wu
and
Tsai
16
, here th
e
authors
have
targete
d
the
smo
o
th a
r
e
a
to
embed
mo
re
data by
usi
n
g
LSB sub
s
titution with
p
r
e
s
erving th
e
ste
go-im
age
qu
ality. Whe
r
e
as
in edge a
r
e
a
s
sim
p
le PVD tech
niqu
e is used. The
r
e can
be 6 bi
ts hidde
n in l
o
we
r level of
two
con
s
e
c
utive conne
cted
pix
e
ls by
u
s
ing
LSB
su
bsti
tuti
on
in
sm
ooth area
s, whi
c
h inturn
i
n
cre
a
ses
the cap
a
city of the prop
osed metho
d
co
mpared
to Wu and T
s
ai [1
6]. But the sugge
sted meth
od
has the d
e
me
rit that the stego-im
age
can
be detecte
d by RS analysi
s
.
The
LSB+PV
D m
e
thod
p
r
opo
sed
by
Wu et
al [18]
is highly
identi
c
al to
LSB te
chni
que.
An image ca
n have majo
r portion of a
r
ea having
sm
all differen
c
e
value. So the LSB embedd
ing
is used for th
e major p
a
rt
of the image
to hide t
he bi
ts. It brings a
conflict
scen
ario to Wu a
n
d
Tsai [1
6]
whe
r
e la
rg
e nu
m
ber
of secret
data bit
s
sho
u
ld b
e
hid
den
in e
dge
area
and l
e
ss n
u
m
ber
of bits
sho
u
ld
be em
bed
de
d in
smooth
a
r
ea. Althou
gh
the u
s
e of
L
SB increa
se
s the capa
city
o
f
the steg
o-im
a
ge but i
s
can
easily p
e
rc
eiv
able to F
r
id
rich et al.’s m
e
thod
5
. In this
context Yang
et
al. [19] prop
o
s
ed
an imp
r
o
v
ed LSB+PV
D techniq
ue
wh
i
c
h e
n
sure
s the
se
cu
rity of se
cret dat
a by
pre
s
e
r
ving th
e same
cap
a
c
ity as offere
d by the
LSB+PVD metho
d
prop
osed b
y
Wu et al [18].
Swain [34]
p
r
opo
sed a n
o
v
el stegan
og
raphi
c techni
que to provide better
ca
pacity to
stego
-ima
ge. The autho
r combine
d
the PVD method
with modified
LSB substitu
tion. The ima
g
e
is p
a
rtitione
d
into 3
×
3
pixel
blo
c
ks. T
he
differen
c
e val
ue i
s
d
e
termi
ned i
n
e
a
ch b
l
ock. Acco
rdi
n
g
to the autho
r
the differen
c
e
may fall in any of the
four levels spe
c
ified. If the differen
c
e val
ue
is
in lower level
then 2 bit L
SB sub
s
titution is
ca
rri
ed
out and
3, 4
and 5
bit LSB sub
s
titution
has
been d
one if
the value fa
lls in lo
wer-middle, hig
h
e
r-mi
ddle
an
d highe
r leve
ls. The fallin
g-off
boun
dary val
ue also ha
s b
een taken ca
re. The expe
ri
mental re
sult
reveals th
at highe
r ca
pa
city
Evaluation Warning : The document was created with Spire.PDF for Python.
IJEECS
ISSN:
2502-4
752
A Revie
w
on
LSB Substitution and PVD
Based Im
age
Steganog
rap
h
y Te
chni
que
s (Aditya K.S.)
718
and le
ss
dist
ortion i
s
achi
eved co
mpa
r
ed to Wu
et al [18]. The comp
ari
s
o
n
betwe
en vari
ous
LSB+PVD te
chni
que
s ha
s been do
ne in
the Table 4.
Table 4. Co
m
pari
s
on b
e
tween vario
u
s L
SB+PVD met
hod
s.
Ref
No
Capacit
y Securit
y
Distortion
Complexit
y
Advantage
18
High
Lo
w
Lo
w
Lo
w
Better embe
dding capacit
y
co
mpared to
Wu and Tsai [16]
method
19
High
Moderate
Moderate
Moderate
Securit
y
is increased compared to
Wu et al
[18].
34
High
Moderate
Moderate
Lo
w
Better capacit
y
,
security
a
nd less distortion
compared to Wu
et al [18].
5. Conclusio
n
and Futu
r
e
Scope
This arti
cle re
views the vari
ous resear
ch
pape
rs b
a
sed
on LSB and PVD and co
mpares
them with
regard to
em
beddi
ng
cap
a
city, disto
r
tion,
attack resi
stan
ce an
d
computatio
nal
compl
e
xity. Lea
st Signifi
cant Bit
(LS
B
) an
d Pixel
Value
Diffe
ren
c
ing
(PV
D
) are the
most
gene
rally utilized
strategi
e
s
for steg
ano
grap
hy. The LSB sub
s
titution can give h
i
gher em
bed
d
i
ng
cap
a
city whe
n
extende
d
up to 4
LSB plane
s.
Th
e
PVD ste
gan
ogra
phy
can
provid
e hig
her
se
curity. A combinatio
n of PVD and LSB can offer b
o
th highe
r e
m
beddi
ng ca
pacity and b
e
tter
s
e
c
u
rity.
Ackn
o
w
l
e
dg
ement
This
wo
rk i
s
sup
porte
d by
the De
partm
ent of Scien
c
e and T
e
chn
o
logy, India t
h
rou
gh
the fund
san
c
tione
d for i
m
provem
ent
of Scie
nce
&
Tech
nolo
g
y i
n
frast
r
u
c
ture,
at de
pa
rtme
nt of
CSE, by orde
r numb
e
r SR/
F
ST/ESI-332/2013.
KL Uni
v
ersity, Vadd
eswaram, Andhra Prade
sh
.
Referen
ces
[1]
Anders
on RJ, Petitcolas F
A
P
.
On the limits of stegano
gra
p
h
y
.
IEEE Journal on Selec
t
ed Areas in
Co
mmun
icati
o
ns
. 1998; 1
6
(4)
:
474-48
1.
[2]
Johns
on
NF
, Jajo
dia
S. Expl
o
r
ing ste
g
a
nogr
aph
y: s
eei
ng t
he u
n
se
en.
IE
EE Co
mp
uter
Journ
a
l
. 1
998
;
31(2): 26-
34.
[3]
W
ang
RZ
, Lin
CF
, Lin J
C
. Hi
din
g
d
a
ta
in
im
ages
b
y
optim
al mo
der
atel
y
signific
ant-b
it r
epl
aceme
n
t.
IEE Electron. Lett
. 2000; 36(2
5
): 2069
–2
070.
[4]
W
ang RZ
, Li
n
CF
, Lin JC. Image
hid
i
n
g
b
y
optima
l
LSB s
ubstitutio
n
a
n
d
gen
etic al
gor
ithm.
Pattern
Reco
gniti
on
. 2
001; 3: 67
1–6
8
3
.
[5]
Fridrich J, Golj
an M, Du R.
Relia
bl
e detecti
o
n
of LSB stega
nogr
aphy
in gr
ayscal
e
an
d co
lor i
m
a
ges
.
in: ACM W
o
rkshop o
n
Multim
edi
a an
d Secur
i
t
y
. 20
01: 27
–3
0.
[6]
Cha
ng CC, Hs
iao JY, Ch
an
CS. F
i
ndin
g
o
p
timal l
east-sig
n
ifica
n
t-bit sub
s
titution in im
a
ge hi
din
g
b
y
d
y
nam
ic progr
amming strate
g
y
.
Pattern Recognition
. 2
003
; 36: 1583 – 1
5
95.
[7]
Cha
n
CK,
Che
ng
LM. Hi
din
g
data
in
imag
es
b
y
sim
p
le
LS
B substituti
on.
Pattern R
e
cog
n
itio
n
. 2
004;
37: 469-
47
4.
[8]
S
w
a
i
n G, L
e
n
k
a SK. A
nov
el
appr
oac
h to
RGB ch
an
nel
bas
ed
ima
g
e
stega
no
grap
h
y
t
e
chn
i
qu
e
.
Internatio
na
l Arab Jour
nal
of e-T
e
chno
logy
. 2
012; 2(4): 1
81-
186.
[9]
S
w
a
i
n G, Lenk
a SK. A tech
ni
que for
secret
communic
a
tio
n
usin
g a
ne
w
block c
i
ph
er
w
i
th d
y
nam
ic
stegan
ogr
aph
y.
Internation
a
l
Journ
a
l of Sec
u
rity and Its Applicati
ons
. 2
0
1
2
; 6(4): 12-24.
[10]
Gutub A, Anke
er M, G
hali
oun
MA, Shahe
en
A, Alvi A.
Pixel
indic
a
tor h
i
g
h
capac
ity techni
q
ue for R
G
B
imag
e bas
ed
stegan
ogr
aphy
. in: F
i
fth IEEE Internatio
na
l
W
o
rkshop o
n
Sign
al Proc
e
ssing a
nd
its
Appl
icatio
ns, Universit
y
of
Sh
arja
h, U.A.E. 2
008: 56-
64.
[11]
Parvez MT
, Gutub AAA.
RGB intensity ba
sed varia
b
l
e
-bi
t
s imag
e steg
ano
grap
hy
. in:
IEEE Asia
pacific Serv
ice
s
Computi
ng C
onfere
n
ce. 20
0
8
: 1322-
13
27.
[12]
Gutub A, Qahtani AA,
T
abakh A. T
r
iple-A secure
RGB image
stegan
ogr
aph
y based
o
n
rand
omizati
on.
in:
IEEE/ACS International Conferenc
e on
Computer Syst
em
s and
Applic
ations
. 20
09
:
400-
403.
[13]
T
i
w
a
ri N, Sha
n
d
il
ya M. Se
cur
e
RGB image
stegan
ogr
aph
y from pixe
l in
di
cator to triple
alg
o
rithm-a
n
incremental gro
w
t.
Intern
atio
nal Jo
urna
l of Security an
d Its Applic
ations
.
201
0; 4(4): 53-
62.
[14]
Kaur M, Gupta
S, Sandh
u
PS, Kaur J. A d
y
namic RG
B intensit
y b
a
s
ed steg
ano
gr
aph
y sch
eme.
W
o
rld Acad
e
m
y of Science, Engi
neer
in
g and
T
e
chnol
ogy
. 2
010; 67: 8
33-8
36.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 25
02-4
752
IJEECS
Vol.
2, No. 3, Jun
e
2016 : 712
– 719
719
[15]
Jain YK, Ahir
w
a
l RR. A Novel
Image Stega
n
ogra
p
h
y
meth
o
d
w
i
t
h
ad
aptive
number of le
a
s
t significa
nt
bits modific
a
ti
on b
a
sed
on
private ste
go-
ke
y
s
.
Inter
nati
ona
l Jour
nal
of Co
mput
er Scienc
e an
d
Security (IJCSS)
. 2010; 4: 39
-49.
[16]
W
u
DC, T
s
ai W
H
. A steg
ano
grap
hic m
e
t
hod
for ima
ges b
y
pi
xel-
valu
e differ
e
n
c
ing.
Patter
n
Reco
gniti
on L
e
tters
. 2003; 24:
1613-
16
26.
[17]
Z
hang
X, W
a
n
g
S. V
u
ln
erab
il
it
y
of p
i
xe
l-val
u
e d
i
fferenc
ing
stegan
ogr
aph
y to h
i
stogr
am
ana
l
y
sis
an
d
modific
a
tion for
enha
nce
d
sec
u
rit
y
.
Pattern R
e
cog
n
itio
n Lett
e
rs
. 2004; 2
5
: 331-
339.
[18]
Wu HC, Wu
NI, T
s
ai CS, H
w
ang MS.
Ima
g
e
steg
a
nogr
aph
ic sc
h
e
me
base
d
o
n
p
i
xel-v
a
lu
e
differenc
ing
a
nd
LSB r
epl
a
c
ement
metho
d
s
. IEEE Proceedings
on
Vi
sion, Image and Signal
Processi
ng. 20
05; 152(
5) 61
1
-
615.
[19]
Yang C
H
, W
eng CY, SJ. Wang, Su
n HM. Varie
d
PVD+
LSB eva
d
in
g d
e
tection
progr
a
m
s to spatia
l
doma
i
n in d
a
ta
embed
din
g
s
ystems.
T
he Journa
l of Systems an
d Softw
are
. 2010; 8
3
: 1635
–1
643.
[20]
Cha
ng
CC, C
h
uan
g JC,
Hu Y
C
. Spati
a
l
Dom
a
in
im
a
ge hi
din
g
sch
eme usin
g
pi
xel-v
a
lu
es differenc
in
g
.
F
unda
menta In
formatic
ae
. 20
06; 70: 17
1–1
8
4
.
[21]
W
ang
CM, W
u
NI, T
s
ai CS
, H
w
a
ng MS.
A hi
gh
qu
alit
y stega
no
grap
h
i
c meth
od
w
i
t
h
p
i
xel-v
a
lu
e
differenc
ing
an
d modu
lus func
tion.
T
he Jour
n
a
l of Systems
and Softw
are
. 200
8; 81: 150.-
158.
[22]
Cha
ng K
C
, Ch
ang
CP, H
u
a
n
g
PS, T
u
T
M
. A nov
el
imag
e
stegan
ogr
aph
i
c
metho
d
us
in
g tri-
w
a
y
pi
xel-
valu
e differe
nci
ng.
Journ
a
l of
Multi
m
ed
ia
. 20
08; 3(2): 37-4
4
.
[23]
Lin
CC, Hs
ue
h
NL. A
lossl
es
s data
hi
din
g
scheme
bas
ed
on thr
ee-
pi
xe
l
block
differ
e
n
c
es.
Pattern
Reco
gniti
on
. 2
008; 41: 1
415
– 142
5.
[24]
Luo W
,
H
uan
g
F
,
Huang
J. A
more sec
u
re st
ega
no
grap
h
y
b
a
sed on ad
apti
v
e
pi
xel-va
lu
e differenc
in
g
scheme.
Mu
l
t
im
ed
To
o
l
s Appl
.
DOI 10.1007/
s110
42-0
09-
04
40-3.
20
10:
40
7-43
0
.
[25]
Joo JC, L
ee
HY, Lee
HK. Improve
d
Steg
ano
gr
ap
hic M
e
thod Pr
eservi
ng Pi
xel-Va
lue
Differenci
n
g
Histogr
am
w
i
t
h
Modu
lus F
unctio
n
.
EUR
ASIP Journal
on Advanc
e
s
in Signa
l Processi
ng
.
doi:1
0.11
55/2
0
10/24
98
26. 20
10: 1-13.
[26]
Hon
g
W
,
C
hen
T
S
, Luo CW
.
Data
emb
edd
i
ng
usin
g
pi
xel
valu
e d
i
fferenc
i
ng
an
d d
i
am
on
d e
n
co
din
g
w
i
t
h
multi
p
le-
b
ase notati
o
n
a
l
s
y
stem.
T
he Jo
urna
l of Systems an
d Softw
are
. 2012; 8
5
: 1166-
117
5.
[27]
Mand
al JK,
Da
s D. Co
lor
ima
ge ste
gan
ogr
a
p
h
y
base
d
on
pi
xel
va
lue
diff
erenc
ing
in
sp
atial
dom
ain
.
Internatio
na
l Journ
a
l of Infor
m
ati
on Sci
enc
es and T
e
ch
ni
ques
. 20
12; 2(
4): 83-93.
[28]
Lee YP, Le
e J
C
, Chen W
K
, Cha
ng KC, Su
IJ, Chang CP
. High-p
a
y
l
oad
image
hid
i
ng
w
i
t
h
qu
ali
t
y
recover
y
us
ing tri-
w
a
y
pi
xe
l-va
lue
d
i
fferenci
n
g
.
Informatio
n Scienc
es
. 201
2; 191: 21
4-2
25.
[29]
Cha
ng
CC
an
d T
s
eng HW
.
A stega
nogr
ap
hic meth
od
for
dig
i
tal
ima
ges
usin
g si
de
match.
Pattern
Reco
gniti
on L
e
tters
. 2004; 25:
1431-
14
37.
[30]
Lia
o
X, W
en
QY, Z
hang J.
A stegano
gra
p
h
ic metho
d
for digita
l imag
es
w
i
th fo
ur-pi
x
el
differenci
n
g
and mo
difie
d
L
SB substitutio
n
.
J. V
is. Commun. Im
age. R
. 201
1; 22: 1-8.
[31]
Yang
CH, W
e
ng CY, T
s
o HK, W
ang SJ. A data h
i
d
i
ng scheme usin
g the
vari
eties o
f
pixel-v
a
lu
e
differenc
ing i
n
multimed
ia im
a
ge.
T
he Jour
na
l of Systems a
nd Softw
are
. 2011; 84: 6
69-6
78.
[32]
S
w
a
i
n G, Lenka SK. Steganogr
aph
y
usin
g t
w
o si
de
d, three si
ded, a
nd four sid
e
d
side mat
c
h
methods.
CSI Transactions on ICT
. 2013; 1(2): 127-1
33.
[33]
S
w
a
i
n
G.
Steg
ano
grap
h
y
in digit
a
l imag
es us
in
g ma
ximu
m differe
nce
o
f
nei
ghb
ori
ng
pi
xel
va
lue
s
.
Internatio
na
l Journ
a
l of Secur
i
ty and Its Appli
c
ations
. 20
13;
7(6): 285-
29
4.
[34]
S
w
a
i
n G. Digit
al ima
ge ste
g
a
nogr
aph
y
usi
n
g ni
ne-p
i
xel dif
f
erenci
ng
a
nd modifi
ed
LSB
substitutio
n
.
India
n
Jour
nal
of Science a
n
d
T
e
chnol
ogy
. 2
014; 7(9): 1
444
–14
50.
[35]
Subh
ed
ar MS, Mankar VH. Current status
and ke
y issu
es in ima
ge s
t
egan
ogr
aph
y:
A surve
y
.
Com
p
uter science review
. 2014: 95-1
13.
[36]
S
w
a
i
n G, Le
nk
a SK. Cl
assific
a
tion
of im
age
st
ega
nogr
ap
h
y
tec
h
n
i
qu
es i
n
sp
atial
dom
a
i
n: A stud
y.
Internatio
na
l Journ
a
l of Co
mputer Scie
nce &
Engin
eeri
ng T
e
chno
logy
(IJ
C
SET
)
. 2014; 5: 219 2
32.
[37]
S
w
a
i
n G, Lenka SK. A novel steg
an
og
raph
y tech
niq
ue b
y
m
a
p
p
i
ng
w
o
r
d
s
w
i
t
h
LSB arra
y.
Internatio
na
l Journ
a
l of Sig
n
a
l
and I
m
ag
in
g Systems En
gin
eeri
n
g
. 20
15; 8
(
1): 115-1
22.
[38]
Pradhan A, S
h
arma DS, S
w
ain G. Variable r
a
te
steganography
in digit
a
l images
using t
w
o,
three and
four nei
gh
bor
Pixels.
Indi
an J
ourn
a
l of Co
mput
er Scie
nce
and En
gi
neer
in
g
. 2012; 3(
3): 457-4
63.
[39]
S
w
a
i
n G. Ada
p
tive pi
xel v
a
lu
e differe
ncin
g
st
egan
ogr
aph
y using
both v
e
rtical an
d h
o
riz
ontal
edg
es.
Multi
m
ed
ia T
o
ols an
d App
lica
t
ions
. DOI: 10.1007/s11042-
015-
2937-
2, 2015: 1-16.
[40]
W
ang H, Ch
e
n
G, Z
hang M. E
dge steg
ano
grap
h
y
for
binar
y im
ag
e
.
T
E
LKOMNIKA Indon
esia
n
Journ
a
l of Elec
trical Eng
i
ne
eri
n
g
. 201
3; 11(5)
: 2822-2
8
2
9
.
[41]
Li Y, Liu Q. Breaki
ng t
he di
git
a
l vide
o stega
n
ogra
p
h
y
.
T
E
LKOMNIKA Indon
esia
n J
ourn
a
l of
Electrical
Engi
neer
in
g
. 2013; 11(
3): 169
1-16
96.
Evaluation Warning : The document was created with Spire.PDF for Python.