A Study of Various Steganographic Techniques Used for Information Hiding

International Journal of Computer Science & Engineering Survey (IJCSES) Vol.4, No.6, December 2013

A Study of Various Steganographic Techniques
Used for Information Hiding
C.P.Sumathi1, T.Santanam2 and G.Umamaheswari3
1,3

Department of Computer Science, SDNB Vaishnav College For Women, Chennai,
India.
2
Department of Computer Science, DG Vaishnav College For Men, Chennai, India.

ABSTRACT
The art of information hiding has received much attention in the recent years as security of information has
become a big concern in this internet era. As sharing of sensitive information via a common communication
channel has become inevitable, Steganography – the art and science of hiding information has gained
much attention. We are also surrounded by a world of secret communication, where people of all types are
transmitting information as innocent as an encrypted credit card number to an online-store and as
insidious as a terrorist plot to hijackers. Steganography derives from the Greek word steganos, meaning
covered or secret, and graphy (writing or drawing) [1]. Steganography is a technology where modern data
compression, information theory, spread spectrum, and cryptography technologies are brought together to
satisfy the need for privacy on the Internet. This paper is an attempt to analyse the various techniques used
in steganography and to identify areas in which this technique can be applied, so that the human race can
be benefited at large.

KEYWORDS
Steganography, Covert- Communications, Carrier-Image, Stego-Key, Stego-Image.

1. INTRODUCTION
The main purpose of Steganography, which means ‘writing in hiding’ is to hide data in a cover
media so that others will not be able to notice it (Figure 1). While cryptography is about
protecting the content of messages, steganography is about concealing their very existence [2].
The applications of information hiding systems mainly range over a broad area from military,
intelligence agencies, online elections, internet banking, medical-imaging and so on. These
variety of applications make steganography a hot topic for study. The cover medium is usually
chosen keeping in mind the type and the size of the secret message and many different carrier file
formats can be used. In the current situation digital images are the most popular carrier/cover files
that can be used to transmit secret information.

DOI : 10.5121/ijcses.2013.4602

9


Figure1. Data Hiding Scheme (Intelligent Multimedia Data Hiding- Springer-ebook)

Steganography equation is ‘Stego-medium = Cover medium + Secret message + Stego key’. The
general model of data hiding can be described as follows. The embedded data is the message that
one wishes to send secretly. It is usually hidden in an innocuous message referred to as a covertext or cover-image or cover-audio as appropriate, producing the stego-text or other stego-object.
A stego-key is used to control the hiding process so as to restrict detection and /or recovery of the
embedded data to parties who know it [3].
Figure.2 depicts the different branches of steganography

Figure 2: Steganography in security domain
(Source: www.ijarcsse.com Vol. 1, issue. 1 Dec 2011 )

While steganography can be achieved using any cover media, we are concerned with hiding data
in digital images. The features expected of a stego-medium are imperceptibility and robustness,
so that the secret message is known only to the intended receiver and also the stego-medium
10


being able to withstand attacks from intruders. The amount of secret message embedded should
be such that it doesn’t reduce the quality of the stego image. This paper aims at studying the
different techniques used in steganography for embedding data, their advantages and their
differences. The goal of steganography is to embed secret data into a cover in such a way that no
one apart from the sender and intended recipients even realizes there is secret data.
A few key properties that must be considered when creating a digital data hiding system are
•
•
•

Imperceptibility: Imperceptibility is the property in which a person should be unable to
distinguish the original and the stego-image.
Embedding Capacity: Refers to the amount of secret information that can be embedded
without degradation of the quality of the image.
Robustness: Refers to the degree of difficulty required to destroy embedded information
without destroying the cover image.

2. STEGANOGRAPHY TECHNIQUES
2.1. Classification of Steganographic Categories
Steganography is classified into 3 categories,
•
•
•

Pure steganography where there is no stego key. It is based on the assumption that no
other party is aware of the communication.
Secret key steganography where the stego key is exchanged prior to communication. This
is most suspectible to interception.
Public key steganography where a public key and a private key is used for secure
communication.

2.2 Classification of Steganographic Methods
Steganography methods can be classified mainly into six categories, although in some cases exact
classification is not possible [2].
Steganography Methods

Substitution Transform domain Spread spectrum Statistical Distortion Cover generation
•
•
•
•
•

Substitution methods substitute redundant parts of a cover with a secret message (spatial
domain).
Transform domain techniques embed secret information in a transform space of the signal
(frequency domain)
Spread spectrum techniques adopt ideas from spread spectrum communication.
Statistical methods encode information by changing several statistical properties of a
cover and use hypothesis testing in the extraction process.
Distortion techniques store information by signal distortion and measure the deviation
from the original cover in the decoding step.
11


•

Cover generation methods encode information in the way a cover for secret
communication is created.

3. REVIEW
This paper analyses the various articles on steganography which help understand the topic in a
new perspective.

3.1 Substitution Methods (Spatial-Domain)
In Mamta Juneja et. al’s [4] research paper a secured robust approach of information security is
proposed. It presents two component based LSB ( Least Significant Bit )methods for embedding
secret data in the LSB’s of blue components and partial green components of random pixel
locations in the edges of images. An adaptive LSB based steganography is proposed for
embedding data based on data available in MSB’s of red, green, and blue components of
randomly selected pixels across smooth areas. It is more robust as it is integrated with an
Advanced Encryption Standard(AES).
In P.Thiyagarajan et.al’s [5] scheme a new high capacity Steganographic scheme using 3D
geometric models is proposed. The algorithm re-triangulates a part of a triangular mesh and
embeds the secret information into newly added position of triangular meshes. This algorithm
also resists against uniform affine transformations such as cropping, rotation and scaling. The
stego key is generated from the message to be embedded. The vertices of the triangle are used for
embedding.
In Shamim Ahmed Laskar et.al’s[6] method data is embedded into the red plane of the image and
the pixel is selected using a random number generator. It is almost impossible to notice the
changes in the image. A stego key is used to seed the PRNG (Pseudo Random Number
Generator) to select pixel locations. This paper focuses on increasing the security of the message
and reducing distortion rate.
In S.Shanmuga Priya et. al’s [7] article the authors propose a novel method based on LSB. Data
embedding is performed using a pair of pixels as a unit, where LSB of the first pixel carries one
bit of information and a function to two pixel values carries another bit of information. The
proposed method shows better performance in terms of distortion and resistance against existing
steganalysis. Embedding is done in the sharper edge regions using a threshold. PSNR value is
compared for adaptive and non-adaptive techniques of data hiding in gray scale & color images.
In B.Sharmila et. al.’s [8] article, the authors propose an algorithm which works on colour
images (JPEG). The edges are chosen for data hiding to improve robustness. The regions located
at the sharper edges present more complicated statistical features and are highly dependent on the
image contents. It is also more difficult to observe changes at the sharper edges than in smooth
regions. In the embedding procedure, the RGB components are separated, and based on a shared
key, one/more components are selected. The cover image is divided into non-overlapping
blocks. Each block is rotated by a random degree determined by a secret key. The resulting image
is rearranged as a row vector V by raster scanning. The secret message is encrypted and by using
LSBMR, 2 secret bits can be embedded into each embedding unit. The message is embedded
after calculating the capacity estimation using a threshold.
In Shweta Singhal et.al’s [9] paper a new image steganography scheme is proposed in the spatial
domain. In the technique, one byte of blue factor of pixels of an image have been replaced with
12


secret bits of text data, which results in better image quality. A stego key is used for security
purposes.
In Fahim Irfan Alam et. al’s paper [10] the authors suggest noise filtering in the beginning before
embedding. After extraction at receiving end, ARQ (Automatic Repeat Request) is used for error
detection & correction. For secure transmission of data, encryption & data hiding are combined in
a single step. Host image and secret data are converted into bit stream. Before encryption of
secret data median filtering is used. The input values are converted to ASCII and then to binary,
the host image RGB values are converted to binary. Substitution is performed character by
character using encryption key. The LSB of every pixel octet is replaced by secret bit stream.
Error detection and correction ensures correct transmission of data.
In Rajkumar Yadav et. al.’s paper [11], the authors present a study of a new method for insertion
of message in an image. The last two bits of pixel value are used for insertion and retrieval of
message. If the last two bits of pixel value are 00 or 10 , we can insert 0, else by adding
/subtracting 1 at that pixel value we can insert 0. Similarly 1 is inserted if last two bits are 01 or
11. For increased security, message is embedded at pseudo random locations. The message is
retrieved similarly based on the pixel values of the last two bits.
In M.B.Ould MEDENI et.al.’s article [12], the authors propose a novel method for hiding
information within the spatial domain of the gray scale image. The Pixel Value Differencing
(PVD) method segments the cover image into nonoverlapping blocks containing two connecting
pixels and modifies the pixel difference in each block (pair) for data embedding. While
embedding secret data, each pixel is split into two equal parts. The number of 1’s in the most
significant part is counted and the secret message is embedded in the least part according to the
number of corresponding bits. The proposed method is based on four-pixel differencing and LSB
substitution.
In Weiqi Luo et. al.’s paper [13], the authors propose an edge adaptive scheme which can select
the embedding regions according to the size of the secret message and the difference between two
consecutive pixels in the cover image. In the data embedding stage, the scheme first initializes
some parameters, which are used for estimating the capacity of the selected regions. Finally stego
image is obtained after pre-processing. A region adaptive scheme is applied to the spatial LSB
domain and the difference between two adjacent pixels is used as a criterion for region selection
and LSBMR (LSB Matching Revisited) as the data hiding algorithm.
In C.H.Yang et. al.’s article [14], a predictive method to enhance the histogram-based reversible
data hiding approach is proposed. Two interleaving predictive stages are used. Most pixels are
predicted by their two neighbourhood pixels and four neighbouring pixels in the column-based
and chess-board based approach. The difference value of each pixel between the original image
and the stego-image remains within ± 1. In interleaving predictions, pixels in odd columns will be
predicted by pixels in even columns or vice versa. In the embedding process predictive error
values of odd columns are used to generate a histogram to embed secret data. The predictive error
values are converted to get the stego-image.
In Venkata Abhiram.M et. al.’s [15] article , the authors propose a randomization technique that
makes use of RGB values of colour images to enhance imperceptibility. In the three channels
RED, BLUE, GREEN the LSB of any one of the 3 channels is used as a pointer to decide
embedding capacity in the other two channels. In the randomization technique, the LSB of any
one of the channels (RGB) are used to indicate how data has to be hidden in the remaining 2
channels. If the last two bits of the channel are 00 there is no hidden data, if it is 01 data is
13


embedded only in channel 2, if it is 10 data is embedded in channel 1 and if it is 11 data is
embedded in both the channels. Three methodologies are used. They are, 1. RED is used as
default pointer. 2. User selects any channel as pointer. 3. Pointers are chosen based on a cyclic
sequence and data is embedded. Images were taken and same size data is embedded using all
methodologies. Based on the histogram study and the values of MSE and PSNR (Mean Square
Error and Peak Signal to Noise Ratio) the 3rd method i.e. the randomized method has better
secrecy and performance with enhanced embedding capacities.
In G.Sahoo et. al.’s [16] article the authors recommend the use of a movie clip as carrier file to
increase the capacity of secret data. The methodology works on the concept of replacement of
entire non-sensitive pixel and the substitution of some part of the sensitive pixel with secret data.
A movie clip is a temporal sequence of two dimensional samples of visual field with each sample
being a frame of the movie. The parts of a movie clip can be divided into moving and static parts.
The static and the dynamic parts can be obtained through Pixel Level Analysis, Likelihood
Analysis or Colour Histogram Technique and stored in a static and dynamic buffer. In static
portion embedding process one pixel is used to store three characters using the formula xij = i+(j–
1)*d where i is the initial location, j is character of the secret data and d is the distance between
two embedding pixels. In dynamic portion embedding MSB method is used. A different stegokey is used for the dynamic portion. Main advantage of this method is more hiding capacity.
In Jasvinder Kaur et.al.’s article[17] the authors analyse different steganographic techniques
based on digital logic and proposes a new enhanced steganographic technique based on it. The
carrier image is selected depending on the information to carry. This technique uses digital
operations based on logic gates and shift operators to embed/derive the hidden information from
image data. Depending on the size of the information to embed the carrier image is divided into
rows and data is embedded using digital operations.
In Hao-Tian et. al.’s paper [18] the authors suggest a method using adjacent bin mapping that is
applied to 3D geometrical models. With the popularity of 3D scanning and modelling techniques
steganography in 3D geometrics using adjacent bin mapping is analysed. Elements in two
adjacent bins are mapped to each other for data embedding. Elements in the original object are
bijectively mapped to those in the stego object. To apply the ABM method to digital images, in
which the pixel values are represented by integers, the bin size ∆ is set to 1 to minimize
distortion. Since it is a kind of LSB method it is named LSB+ algorithm. Given a gray scale
image, its histogram is calculated by counting the pixels with the same value, i.e. the amount of
pixels in each bin. The histogram of the cover image is preserved by restricting the amounts of
1’s and 0’s.
In Bawankar Chetan.D et. al.’s article [19] a prioritized sub blocks by pattern matching scheme
is used to embed the code and a micro controller used for sake of security where it transmits a
pre-programmed key at the beginning of each process. The steganalysis algorithm is continued
only if the received key is correct at the destination; otherwise retrieving secure information from
cover image is not accomplished. Perceptual multimedia sources enables data embedding as well
as lossy compression either imperceptibly or with a controllable amount of perceptual
degradation, whereas non-perceptual sources like text and executable codes require loss-less
processing, transmission and storage. The image is divided into blocks. A block is selected for
embedding based on a pattern. Several blocks are rejected due to various reasons like high
visibility, non reversible at receiver side etc. In each sub block only the middle pixel is selected
for hiding information. The secret message is encrypted to enhance security. In this paper the
authors propose a steganographic model in which the hidden message can be composed and
inserted in the cover in real-time. This is realized by designing and implementing a secret key
14


steganographic micro architecture employing Field Programmable Gate Arrays FPGA. Four
steganographic algorithms were found to be suitable for hardware implementation.
In Tanmay Bhattacharya et. al.’s article [20] the authors use a session based encryption and cross
fold transposition for embedding. The secret text is converted to its binary form and cross fold
transposition is performed. This binary form is perturbed by genetically generated session-key
and embedded within the host image. For extraction both the stego image and the original image
along with the session-key is sent.
In Chin-Chen Chang et.al.’s paper[21] an adaptive method is proposed. Data is hidden based on
codeword grouping. A set of code words generated using palette generation algorithm is
employed in index-based images. A code word grouping based steganographic scheme for index
encoding images is presented. The relationship of code words is explored to group different
member sub-clusters. The size of the sub-cluster determines the hiding capacity. To enhance
hiding capacity sub-clusters with larger members are grouped together & sub-clusters with
smaller members are grouped together. In the embedding procedure the sub-cluster to which the
closest searched codeword belongs is identified, and the original encoded codeword is modified
to hide secret message. The number of sub-cluster members indicates how many bits of secret
message can be embedded. A set of thresholds is used to determine members of sub-cluster.
Therefore choosing an adequate threshold is important. To improve security the sequence of
embedding pixels is reorganised using a pseudo random generator.
In R.Chandramouli et.al.’s work [22] the authors analyse adaptive steganographic techniques
which deal with hiding a message without causing statistically significant modifications. It is esecure if the relative entropy of the probability distribution of cover objects and stego-objects is <
or = e. LSB steganography is most simple and straight forward. The steganographic capacities of
LSB based images are analysed using probability techniques & performance of steganalysis is
also measured. Only framework of a passive warden is analysed. For security, message is
encrypted before it is embedded. LSB based technique changes pixel value by ± 1 or leave them
unchanged. The goal of a steganalyst is to estimate if I has hidden data. (I - Index set that denote
the mean subtracted cover image)
Table 1. Substitution Based Steganographic Methods listed in chronological order starting from latest.

S.N
o
1.

2.

Author
Mamta Juneja
et.al.[4]

P.Thiyagarajan
et.al [5]

Year

Method used

Advantage

2013

Two component
based LSB

More robust as
algorithm is
integrated with
Advanced
Encryption
Standard (AES)

2013

Scheme using 3D
geometric models.

Resistance
against uniform
affine
transformations
such as cropping,
rotation &
scaling.

Other parameters

Ave. PSNR >52

PSNR > 55 db
MSE 0 or
between 0 & 0.1

15


3.

Shamim Ahmed
Laskar et.al [6]

2013

Data embedding
in the red plane of
the image selected
using PRNG

Increases security
with reduced
distortion rate

PSNR 58.8 db
MSE 0.0854 %

4.

S.Shanmuga
Priya et. al [7]

2012

Embedding done
in the sharper edge
regions using a
threshold

Better
performance in
terms of
distortion and
resistance against
existing
steganalysis.

Non adaptive
technique has
more PSNR &
less MSE than
adaptive
technique.

5.

B.Sharmila et.
al.[8]

2012

Edge regions
selected for
embedding using
LSBMR ( LSB
Matching
Revisited)

Embedding
capacity is more.

MSE 0.013168

1 byte of blue
factor of pixels are
replaced with
secret bits.

Better image
quality & secure.

6.

Shweta Singhal
et.al [9]

2011

PSNR 66.96 dB
Improved
security due to
encryption.

630 pixels
embedded in a
256x256 image.
PSNR 47.94 db
MSE 1.056

7.

Fahim
Irfan
Alam et al[10]

2011

Noise filtering
before embedding
combined with
encryption.

Error detection & Success rate of
Noise free
>83%
with
transmission.
different types of
images.

8.

Rajkumar
Yadav et al.[11]

2011

A novel approach
for image
steganography In
spatial domain
using last two bits
of pixel value

Limited changes Chance of no
in cover-image.
change in pixel
value is 50%

A novel
steganographic
method based on
Pixel Value
Differencing(PVD
)

Imperceptibility
as there is very
less difference
between original
and stego image
40017 bits

42.68 dB PSNR

An Edge adaptive
scheme for Region
selection and
LSBMR for data

Improved Visual
Quality and
security of secret

54.1 PSNR

9.

10.

M.B.Ould
MEDENI
et.al.[12]

2010

Weiqi Luo

2010

et. al.[13]

(7th & 8th bit)

40017 bits
embedded in a
512x512 image.

Embedding
capacity 50% of
16


embedding
11.

C.-H.Yang

2010

et. al.[14]

message

cover image

Improving
histogram based
reversible data
hiding by
interleaving
predictions

Larger
embedding
capacity & better
image quality.

48.82 dB PSNR
99947bits
embedding
capacity
(512x512 Image)

12.

Venkata
Abhiram.M

.
2009

et. al.[15]

13.

G.Sahoo
al.[16]

et. 2009

Pixel Intensity
based
steganography
with improved
randomness.

Imperceptibility

Data embedded in
static & dynamic
portions after
place analysis

Very high
embedding
capacity with
security

Embedding
capacity is more.

55dB app PSNR
0.1790 app MSE

Embedding
position is
calculated using
Xij=i+(j-1)*d

14.

Jasvinder Kaur
et.al.[17]

2009

Embedding using More embedding Embedding
digital operations capacity
capacity
are compared
1165084 bits
.Image size 512
x 512

15.

Hao-Tian

2009

Steganography in
3D geometrics &
images using
adjacent Bin
Mapping

et. al.[18]

Statistics of
image preserved

PSNR>60 dB
Capacity
bits

1200

(512x512 image)
(LSB+ algorithm)
16.

Bawankar
Chetan.D

2009

et. al.[19]

17.

Tanmay
Bhattacharya
et. al.[20]

2009

Steganography
Algorithm using
Pattern Matching
with External
Hardware

External security
using Hardware

35.56dB

¼ the size of the
host image

0.0013 MSE

A hiding
technique using
bit level cross fold
transposition and
genetic algorithm

Embedding
capacity &
security of text is
more compared to
S-Tools.

Embedding
capacity
108,242bytes
Image size
732x486

17


18.

Chin-Chen
Chang
et.al.[21]

2004

19.

R.Chandramoul
i et.al.[22]

2001

Code word
grouping –palette
generation
algorithm
Encoded
codeword is
modified to hide
secret message
Adaptive
Steganography

Good Image
Quality & Better
Hiding Capacity

PSNR 36.34
Capacity
355588 bits
Cover
image:512x512

Imperceptibility

Analysis of pd
& pf.
probability of
correct detection
probability of
false detection

3.2. Transform Domain Methods
In Hemalatha.S et.al’s [23] paper, the authors propose a method that uses two gray scale images
of size 128 x 128 that are used as secret images and embedding is done in RGB and YCbCr
domains. The quality of stego images are good in RGB domain by comparing the PSNR values.
The authors have used Integer Wavelet Transform (IWT) to hide secret images in the color cover
image. The authors have compared the PSNR values and image quality when embedding is done
in the RGB and YCbCr domains.
In another article by Hemalatha .S et. al. [24] Integer Wavelet Transform (IWT) have been
suggested to hide multiple secret images and keys in a color cover image which is more efficient.
The cover image is represented in the YCbCr color space. Two keys are obtained, encrypted and
hidden in the cover image using IWT.
In Keith.L. Haynes ’s article [25] the author studies the use of image steganography to breach an
organization’s physical and cyber defences. The proposed method utilizes computer vision and
machine learning techniques to produce messages that are undetectable and if intercepted cannot
be decrypted without key compromise. To avoid detection DWT (Discrete Wavelet Transform) is
used. The goal of a computer vision system is to allow machines to analyze an image and make a
decision as to the content of that image. The computer vision can be categorized as Model-Based
& Appearance Based which uses example images and machine learning techniques to identify
significant areas or aspects of images that are important for discrimination of objects contained
within the image. Machine learning is different from human knowledge/ learning. A computer
has to make decision of the presence of a face based on the numbers contained in a 2D matrix.
The feature is identified by using Haar feature selection. The goal is to identify the set of features
that best distinguishes between images in the different classes. In the proposed method the cover
image does not contain a secret message, rather the classification of the image yields the hidden
message. Since the proposed algorithm utilizes ordinary unmodified images, there are no inherent
indicators of covert communication taking place.
In S.Arivazhagan et. al.’s work [26] the authors propose a method that works in the transform
domain and attempts to extract the secret almost as same as the embedded one, maintaining
minimal changes to cover image by using techniques like median maintenance, offset &
quantization. A modified approach for embedding colour images within colour images is
proposed and it overcomes the limitations in embedding. Arnold Transform is applied on the
secret image to increase robustness. This transformed image is then split into the three colour
18


planes R, G, B and are subjected to DWT individually, converted to bit stream and then
concatenated to be embedded in the cover image which is also subjected to DWT.
In Anindya Sarkar et. al.’s paper [27] the authors propose a Matrix Embedding with Repeat
Accumulate (ME-RA) based steganography in which the host coefficients are minimally
perturbed such that the transmitted bits fall in a coset of a linear code, with the syndrome
conveying the hidden bits. The hiding blocks are pseudo-randomly chosen. A powerful repeat
accumulate code is used for error correction. The authors have compared QIM (Quantization
Index Modulation) and ME-RA methods. The comparisons with a slight modification of the MERA (puncture and non-shrinkage) methods with different decoding methods are also tabulated.
The authors highlight that the use of ME instead of QIM within the YASS (Yet another
Steganographic Scheme) that provides improved steganalysis performance but software
complexity is more.
In Prosanta Gope et. al.’s article [28], the authors introduce an enhanced JPEG steganography
along with a suitable encryption methodology using a symmetric key cryptographic algorithm.
The JPEG cover image is broken into 8 x 8 blocks of pixel. DCT is applied to each block and
quantization is done and data is encrypted using a new encryption method which uses CRC
checking.
In Po-Chyi et.al.’s article [29] the authors compare the advantage of embedding in JPEG 2000
images with the previous approach of embedding in JPEG images. Most of the steganographic
methods are based on JPEG because as a block DCT codec JPEG lends itself a good candidate for
information hiding due to its fixed block structure. JPEG 2000 which is an upcoming still image
coding standard can be used to hide high volume data. If information is embedded in the output of
tier-2 coding, i.e. the JPEG 200 packets, it can be guaranteed that all the embedded information
will be received without error and in correct order. But, difficulty lies in the modification of
packets for embedding, since the bit-streams are compactly compressed by the arithmetic coder.
Careless modification would result in failure of expanding compressed image. In the embedding
process the image is decomposed using wavelet transform. (Number of wavelet decomposing
levels & image size should be related to the host image), Lazy Mode Coding (Magnitude
Refinement pass is suitable for steganographic purposes) is used for embedding.
In Hideki Noda et.al.’s paper [30] the authors propose a method that is based on a seamless
integration of JPEG2000 lossy compression scheme and bit-plane complexity segmentation
(BPCS) steganography. In bit-plane decomposition an n bit image is decomposed into a set of n
binary images by bit slicing operations, combined with replacing binary data in LSB bit planes
with secret data. The BPCS steganography uses bit-plane decomposition and characteristics of
human vision. In JPEG 2000, wavelet coefficients of an image are quantized into a bit-plane
structure. Each bit plane of the cover image is segmented into small size 8x8 blocks and are
classified into informative / noise like blocks, using a threshold of the complexity α0 (e.g. value of α0 0.3
αmax) αmax is the possible complexity value. The secret file is segmented into a series of blocks
containing 8 bytes of data that are regarded as 8x8 binary images. If secret block is less complex
than the threshold α0, conjugate (XOR) it to make more complex. (α=αmax – α). The image will
now be a conjugated image. Replace each noise like block in the bit planes with a block of secret
data. If block is conjugated store it in the conjugation map. Blocks can be randomly selected by
using a random-number generator. Also embed the conjugation map with secret data (usually the
first noise like block). Secret data is embedded after tier-2 encoding.

19

Table 2. Transform Domain Steganographic Methods in chronological order starting from latest

S.No

Author

Year

Method used

Advantage

Other parameters

1.

Hemalatha.S
et.al [23]

2013

Integer Wavelet
Transform is
used to compare
embedding in
two different
domains.

Quality of image
is compared by
embedding in
RGB and YCbCr
domains.

PSNR in RGB 47
PSNR in YCbCr 41

2.

Hemalatha .S et.
al. [24]

2013

Multiple secret
images are
hidden in a
cover image
using IWT.

More efficient
and secure as
encryption is
used.

Ave PSNR 44.7db

3.

Keith.L. Haynes
et. al.[25]

2011

Using Image
Steganography
to Establish
Covert
Communication
Channels

Covert
communication

Min. Image Size
(128x128)

Security by added
encryption

10 Haar features
depicted
Final class
recognition 11.5%

4.

S.Arivazhagan
et. al. [26]

2011

Colour Image
Steganography
Using Median
Maintenance

Increased
Security due to
random selection

PSNR 60.23 dB
(app)

Minimal Changes
in the cover
image
5.

Anindya Sarkar
et. al.[27]

2010

Matrix
Embedding with
Repeat
accumulate
codes

Robust against
various global
attacks

LLR computation
(Log of 0 or 1
transmitted ) &
bpnc (hiding rate)
is compared with
different decoding
& hiding methods

6.

Prosanta Gope
et. al.[28]

2010

Enhanced JPEG
steganography
with suitable
encryption

Added security
using encryption

Bit Error Rate,
MSE and PSNR
values are
computed

20


7.

Po-Chyi

2003

et.al.[29]

Steganography
in JPEG 2000
compressed
images

Controllable
distortion
Can embed high
volume data.

PSNR is 38.87 dB
for 0.75 bit rate and
increases for
increasing bit rate.
Lazy mode coding

8.

Hideki Noda
et.al.[30]

2002

BPCS
Image Quality is
steganography
Good
with compressed
images

Embedding Rate
11.2%

3.3. Statistical Methods
In Tomas Filler et. al.’s work [31], the authors propose a practical methodology for minimizing
additive distortion in steganography with general embedding operation which is more flexible and
easy. Syndrome-Trellis Codes (STC) are used to improve the security of the system. STC divides
the samples into different bins (binning) which is a common tool used for solving many
information-theoretic and also data-hiding problems. The proposed method can be used in both
spatial & transform domain. A proper distortion function is chosen which makes statistical
detection difficult. Once the steganographer specifies the distortion function, the proposed
framework provides all tools for constructing practical embedding schemes. The distortion
function or the embedding operations need not be shared with the recipient.
In Jessica Fridrich et.al.’s research paper [32] the authors propose a reversible embedding
scheme for VQ-compressed images that is based on side matching and relocation. The new
method achieves reversibility without using the location map. Even a tiny distortion of the
original content is not applicable in some sensitive applications such as military, medical / fine art
data. Therefore the value of reversible methods of steganography is increasing. VQ (Vector
Quantization) is a popular compression technique because of its simple encoding and decoding
procedures. To achieve better imperceptibility the codebook is partitioned into several clusters
before embedding. The input needed will be a VQ compressed image, a stream of secret bits, a
super codebook SC, clusters of the super codebook SC and multiple hit maps. The output will be
a VQ stego image. Block X in the cover image will fall into one of the three following cases. If X
is equal to the ith codeword of Go, the embedding process is invoked. If X is equal to the ith
codeword of G1, no secret bit can be embedded and a compensation procedure is needed to avoid
conflicting with case 1. If X does not belong to G0 U G1, no secret bit can be embedded and X is
skipped. Secret bits can be embedded only in case 1.
In Chin-Chen Chang et.al.’s article [33] a new approach to wet paper codes using random linear
codes of small co-dimension is used which improves embedding efficiency is proposed. To
prevent from attack, the selection channel should not be publicly available even in any partial
form. A possible remedy is to select it according to some side information that is in principle
unavailable to the attacker (e.g.) random or that cannot be well estimated from the stego image.
Steganography with non shared selection channels requires codes for memories with defective
cells also called wet paper codes. This paper provides a new tool for steganography a coding
method that empowers the steganographer with the ability to use arbitrary selection channels
while substantially decreasing the number of embedding changes. The algorithm combines wet
paper codes with matrix embedding arbitrary selection channels and improved embedding
efficiency using random linear codes of small co-dimension.

21


In Zhicheng Ni et.al.’s article [34] the authors present a lossless data hiding which is robust
against JPEG / JPEG 2000 compression. The image is split into 8 x 8 blocks and each block is
split into two subsets (A, B). For each block the difference value α is calculated where α is the
arithmetic average of differences of pixel pairs within the block. This α is selected as a robust
quantity for embedding the information bit. Each bit of the secret message is associated with a
group of pixels eg. A block in an image. The bit embedding strategy used is as follows, If α is
located within a threshold & to embed bit 1, shift α to right/left beyond a threshold by
adding/subtracting a fixed number from each pixel value within one subset. To embed 0, the
block is intact. If α is located outside the threshold, always embed 1 thus shifting the value α
away beyond a threshold. Then error correction code is applied.
Table 3 Statistical Steganographic Methods in chronological order starting from latest

S.No

Author

Year

Method used

Advantage

Other parameters

1.

Tomas Filler

2010

Additive
distortion
function in
Steganography
using Syndrome
Trellis codes

Statistically
Undetectable.Can
be implemented
in spatial &
transform
domains.

Relative payload
α compared with
relative wetness τ
from the no. of
changed wet
elements

2006

Matrix
embedding with
wet paper codes

Embedding
Efficiency with
minimal changes
to the cover

Embedding
efficiency
e(p)=p/Ra(p)

et. al.[31]

2.

Jessica Fridrich
et.al.[32]

Efficiency
increases for
shorter messages
reaches a
maximum &
decreases
3.

4.

Chin-Chen
Chang et.al.[33]

Zhicheng Ni
et.al.[34]

2006

2004

Reversible
embedding
scheme for VQcompressed
images based on
side matching
and re-location.
Uses location
map.
Lossless Data
Hiding

Suitable for VQ
compressed
images due to its
high utilization of
local block
properties

PSNR 31.16 dB

No salt pepper
noise

>38 dB

Payload 8707
bits
(128x128)

1024/512 bits
Can resist JPEG
compression
1024/512 bits

22


3.4. Distortion Methods
In M.B.Ould MEDENI et.al.’s article [35], the authors use error correcting codes in
steganographic protocols. An optimal code is one that makes most of the maximum embeddable
(MLE). The method referred to as matrix encoding requires the sender and recipient to agree in
advance on a parity check matrix H. The cover medium is processed to extract a sequence of
symbols ѵ, which is modified into s to embed the message m, s is sometimes called the stegodata, and modifications on s are translated on the cover-medium to obtain the stego-medium.
Relation between steganographic algorithms and error correcting codes are discussed.
In D.P.Gaikwad et. al.’s paper [36]the authors propose image restoration technique in
steganography. The image is blurred before hiding the message image using special point spread
function and randomly generated key. Sequential LSB embedding in the R plane is done in this
project. The number of rows and columns of the message image is encrypted in the first row of
the cover image. Before inserting, the original message image is blurred using the specific PSF
(Point Spread Function). The parameters used for blurring with PSF are used as keys during deblurring. The secret key values are sent through a secure channel (Tunnelling). The secret image
is recovered using the two keys and a third key, which is randomly generated and depends on the
content of the hiding message.
Table 4 Distortion Steganographic Methods in chronological order starting from latest

S.No

Author

Year

Method used

Advantage

Other parameters

1.

M.B.Ould
MEDENI
et.al.[35]

2010

Use of errorcorrecting codes
in
steganography

Improved
embedding
efficiency

Performance
evaluation by
using D, average
distortion, E
embedding rate
& R expected
number of
changes in
cover.

Security

2.

D.P.Gaikwad
et. al.[36]

2010

Image blurring
with sequential
LSB embedding

More secure and
robust than
simple LSB

7.3064e-008
random
key
value for a png
image of 24 KB

4. CONCLUSION & FUTURE WORK
In this paper different steganographic articles were studied and were categorized into different
techniques. As many new application areas are identified like internet banking, mobile
communication security, cloud security etc., the insight into the steganographic principles will
definitely guide us to identify new areas and to improve its applications in the already existing
application areas also.

23


ACKNOWLEDGEMENTS
We thank everyone who gave support and ideas in bringing out this paper.

REFERENCES
[1]
[2]
[3]
[4]
[5]

[6]

[7]

[8]

[9]

[10]
[11]
[12]

[13]

[14]
[15]

[16]

[17]

[18]

Eric Cole, Ronald D. Krutz, Consulting Editor (2003), Hiding in Plain Sight, Steganography and the
Art of Covert Communication, Wiley Publishing, Inc.
Stefan Katzenbeiser & Fabien A.P.Petitcolas(1999), Information Hiding Techniques for
Steganography and Digital Watermarking, Artech House, Computer Security series, Boston, London.
Fabien A.P.Petitcolas, Ross J.Anderson and Markus G.Kuhn, (1999) “Information Hiding – A
Survey”, Proceedings of the IEEE, special issue on protection of multimedia content, pp.1062-1078.
Mamta Juneja and Parvinder Singh Sandhu, (2013) “A New Approach for Information security using
an Improved Steganography Technique”, Journal of Info.Pro.Systems, Vol 9, No:3, pp.405-424.
P.Thiyagarajan, V.Natarajan, G.Aghila, V.Pranna Venkatesan, R.Anitha, (2013) “Pattern Based 3D
Image Steganography”, 3D Research center, Kwangwoon University and Springer 2013, 3DR
Express., pp.1-8.
Shamim Ahmed Laskar and Kattamanchi Hemachandran, (2013)
“Steganography Based On
Random Pixel Selection For Efficient Data Hiding”, International Journal of Computer Engineering
and Technology, Vol.4, Issue 2, pp.31-44.
S.Shanmuga Priya, K.Mahesh and Dr.K.Kuppusamy, (2012) “Efficient Steganography Method To
Implement Selected Least Significant Bits in Spatial Domain”, International Journal of Engineering
Research and Applications,, Vol2, Issue 3, pp. 2632-2637.
B. Sharmila and R.Shanthakumari, (2012) “Efficient Adaptive Steganography For Colour Images
Based on LSBMR Algorithm”, ICTACT Journal on Image and Video Processing, Vol. 2, Issue:03,
pp.387-392.
Shweta Singhal, Dr.Sachin Kumar and Manish Gupta, (2011) “A New Steganography Technique
Based on Amendment in Blue Factor ”, International Journal of Electronics Communication and
Computer Engineering, Vol.2, Issue 1, pp.52-56.
Fahim Irfan et. Al. ‘s (2011) “An Investigation into Encrypted Message Hiding through Images
Using LSB ”, International Journal of EST,
Rajkumar Yadav, (2011) “A Novel Approach For Image Steganography In Spatial Domain Using
Last Two Bits of Pixel Values”, International Journal of Security, Vol.5 Iss. 2 pp. 51-61.
M.B.Ould MEDENI and El Mamoun SOUIDI, (2010) “A Generalization of the PVD Steganographic
Method”, International Journal of Computer Science and Information Security, Vol.8.No.8, pp156159
Weiqi Luo, Member, IEEE, Fangjun Huang, Member, IEEE, and Jiwu Huang, Senior Member,
IEEE, (2010) “Edge Adaptive Image Steganography Based on LSB Matching Revisited”, IEEE
Transactions on Information Forensics and Security, Vol.5.No.2, pp.201-214.
C.-H. Yang and M.-H. Tsai, (2010) “Improving Histogram-based Reversible Data Hiding by
Interleaving Predictions”, IET Image Processing, Vol.4. Iss. 4 pp. 223-234.
Venkata Abhiram.M, Sasidhar Imadabathuni, U.Padmalochini, Maheedhar Imadabathuni and Ramya
Ramnath (2009), “Pixel Intensity Based Steganography with Improved Randomness”, International
Journal of Computer Science and Information Technology, Vol 2, No 2, pp.169-173.
G.Sahoo & Rajesh Kumar Tiwari (2009) “Hiding Secret Information in Movie Clip: A
Steganographic Approach”, International Journal of Computing and Applications, Vol. 4, No.1, pp
103-110.
Jaswinder Kaur, Inderjeet & Manoj Duhan, (2009) “A Comparative Analysis of Steganographic
Techniques”, International Journal of Information Technology and Knowledge Management, Vol.
2, No. 1, pp 191-194.
Hao-Tian Wu and Jean-Luc Dugelay , (2009) “Steganography in 3D Geometrics and Images by
Adjacent Bit Mapping”, EURASIP Journal on Information Security, Vol. 2009, Article ID 317165,
pp1-10.

24

[19] Bawankar Chetan.D, Hande.K.N, Jaiswal .A.A & Bute. A (2009) “ Pattern Matching With External
Hardware For Steganography Algorithm”, International Journal of Information Technology and
Knowledge Management, Vol. 2 No.2, pp 289-295.
[20] Tanmay Bhattacharya, Manas Paul & Arindam Dasgupta (2009) “A Novel session Based Text
Encryption & Hiding Technique Using Bit Level Cross Fold Transposition & Genetic Algorithm”,
International Journal of Information Technology and Knowledge Management, Vol. 2, No.2,
pp.419-423.
[21] Chin Chen Chang, Piyu Tsai & Min-Hui Lin (2004) “An Adaptive Steganography for Index-Based
Images using Codeword Grouping”, Springer-Verlag Berlin Heidelberg 2004, pp.731- 738.
[22] R.Chandramouli & Nasir Memon (2001) “Analysis of LSB Based Image Steganography Techniques”,
IEEE Article.p.p 1019-1022.
[23] Hemalatha.S, U.Dinesh Acharya and Renuka.A, (2013) “Comparison of Secure and High Capacity
Color Image Steganography Techniques in RGB and YCBCR domains”, International Journal of
Advanced Information Technology, Vol.3, No.3, pp.1-9.
[24] Hemalatha.S, U.Dinesh Acharya and Renuka.A, Priya.R Kamnath, (2013) “A Secure and High
Capacity Image Steganography Technique”, Signal & Image Processing – An International Journal,
Vol.4, No.1, pp.83-89.
[25] Keith L.Haynes, (2011) “Using Image Steganography to Establish Covert Communication
Channels”, International Journal of Computer Science and Information Security, Vol 9, No.9, pp. 17.
[26] S.Arivazhagan, W.Sylvia Lilly Jebarani, and S.Bagavath (2011) “Colour Image Steganography
Using Median Maintenance”, ICTACT Journal on Image and Video Processing, Vol. 2, Iss:01,
pp.246-253.
[27] Anindya Sarkar, Member, IEEE, Upamanyu Madhow, Fellow,IEEE, and B.S.Manjunath, Fellow,
IEEE, (2010) “Matrix Embedding With Pseudorandom Coefficient Selection and Error Correction
for Robust and Secure Steganography”, IEEE Transactions on Information Forensics and Security,
Vol.5.No.2, pp.225-239.
[28] Prosanta Gope, Anil Kumar and Gaurav Luthra , (2010) “An Enhanced JPEG Steganography Scheme
with Encryption Technique”, International Journal of Computer and Electrical Engineering ,
Vol.2.No.5, pp924-930.
[29] Po-Chyi & C.-C.Jay Kuo, Fellow, IEEE(2003) “Steganography in JPEG 2000 Compressed Images”,
IEEE Transactions on Consumer Electronics, Vol. 49, No. 4, pp 824-832.
[30] Hideki Noda, Jeremiah Spaulding, Mahdad.NShirazi & Eiji Kawaguchi (2002) “Application of BitPlane Decomposition Steganography to JPEG 2000 Encoded Images”, IEEE Article.
[31] Tomas Filler, Student Member, IEEE, Jan Judas and Jessica Fridrich, Member, IEEE, (2010)
“Minimizing Additive Distortion in Steganography using Syndrome Trellis Codes”, IEEE Article,
pp.1-17.
[32] Jessica Fridrich, Miroslav Goljan, David Soukal (2006) “Wet Paper Codes With Improved
Embedding Efficiency ”, IEEE Transactions on Information Forensics and Security, Vol 1. No.1, pp
102-110.
[33] Chin-Chen Chang & Chih-Yang Lin (2006) “Reversible Steganography for VQ-Compressed Images
Using Side Matching and Relocation ”, IEEE Transactions on Information Forensics and Security,
Vol. 1. No.4, pp 493-501.
[34] Zhicheng Ni, Yun Q.Shi, Nirwan Ansari, Wei Su, Qibin Sun & Xiao Lin (2004) “Robust Lossless
Image Data Hiding”, IEEE Article.
[35] M.B.Ould MEDENI and El Mamoun SOUIDI, (2010) “Steganography and Error Correcting Codes”,
International Journal of Computer Science and Information Security, Vol.8.No.8, pp147-149.
[36] D.P.Gaikwad and S.J.Wagh, (2010) “Colour Image Restoration For An Effective Steganography”, Imanager’s Journal on Software Engineering, Vol.4 .No.3, pp.65-71.

25

A Study of Various Steganographic Techniques Used for Information Hiding

More Related Content

What's hot (20)

Similar to A Study of Various Steganographic Techniques Used for Information Hiding (20)

More from ijcses (14)

Recently uploaded (20)

A Study of Various Steganographic Techniques Used for Information Hiding