I.J.

Intelligent Systems and Applications, 2012, 1, 32-38

Published Online February 2012 in MECS (http://www.mecs-press.org/)
DOI: 10.5815/ijisa.2012.01.04

An Efficient Method of Steganography using

Matrix Approach

Nirmalya Chowdhury
Department of Computer Science and Engineering,Jadavpur University,Kolkata-700032,India
[email protected]

Puspita Manna
Department of OCLAN,Bharat Sanchar Nigam Limited,24 pgs(s) West Bengal,India
[email protected]

Abstract— A large number of the world business is going on stego-image will look identical to the cover image to
using “INTERNET” and the data over the internet which is human eyes.
vulnerable for attacks from the hackers. Thus, uses of highly Data security can also be achieved by cryptography.
efficient methods are required for sensitive data transmission Sometimes these two techniques are used in a combined
over the internet to ensure data security. One of the solutions to
data security is to use an efficient method of steganography. manner to increase the level of data security. The term
The goal of steganography is to hide messages inside other steganography means “cover writing” whereas
‘harmless’ messages in a way that does not allow any enemy to cryptography means “secret writing”.
even detect that there is a second message present.
Steganography can be used with a large number of file formats Cryptography is the study of methods of sending
most commonly used in the digital world of today. The different messages in distinct form so that only the intended
file formats popularly used are .bmp, .gif, .txt etc. Thus the recipients can only remove the disguise and read the
techniques of steganography are going to play a very important message. The message we want to send here is called
part in the future of data security and privacy on open systems plain text and disguised message is called cipher text.
such as the Internet.
This paper presents an efficient method for hiding data into an
The process of converting a plain text to a cipher text is
image and send to the destination in a safe manner. This called enciphering or encryption, and the reverse process
technique does not need any key for embedding and extracting i.e. recovering the plain text from the cipher text is called
data. Also, it allows hiding four bits in a block of size 5×5 with deciphering or decryption.
minimal distortion. The proposed algorithm ensures security
and safety of the hidden information. The experimental results The encryption method protects contents during
presented in this paper show the efficacy of the proposed transmission of the data from the sender to receiver.
method. However, after receipt and subsequent decryption, the
data is no longer protected and is clear. Steganography
Index Terms- Steganography; data security; data hiding; hides messages in plain sight rather than encrypting the
Stego-image message. The message is embedded in the data and does
not require a secret transmission. In fact, the message is
I. INTRODUCTION carried inside data. Steganography is therefore a better
approach to data security than cryptography.
Steganography is the art of hiding information in such This paper proposes an efficient algorithm which has
a way that prevents the detection of hidden messages been developed by modifying an existing algorithm of
[1,2]. In this technique, no one apart from the sender and steganography. The proposed algorithm has
the intended recipient even realize that there is a hidden experimentally been found to give better performance
message. compared to the existing one.
In steganography, the secret message is embedded II. A BRIEF REVIEW OF EXISTING METHODS
into an image (or any media) called cover image, and
then sent to the receiver who extracts the secret message Steganography helps to hide secret information which
from the cover message [3,4]. After embedding of the is to be protected using a digital object, also referred to as
secret message, the cover image is called a stego-image. cover object, in such a manner that the information
This image should not be distinguishable from the cover becomes a part of the cover object. The majority of
image, so that the attacker can not discover the presence today’s steganographic systems uses multimedia objects
of any embedded message[5]. Note that, the resulting like image[6], audio, video etc as the cover media

Copyright © 2012 MECS I.J. Intelligent Systems and Applications, 2012, 1, 32-38
 An Efficient Method of Steganography using Matrix Approach 33

because people often transmit digital pictures over email i. For every row in the first four rows of the
and through other internet communications[7]. There are block,
several methods of steganography[8,9]. Some of them exclusive-or all the bits of that row to
are described below. get
r1r2r3r4.
Least significant bit (LSB) insertion is a common and ii. For every column in the first four columns
simple approach to embed information in a cover object. of
For images as a cover object, the LSB of a pixel is the block, exclusive-or all the bits of
replaced by one bit from the message. If we use 24-bit that
image file as the cover object, we can store 3 bits of column to get c1c2c3c4.
information from the message in each pixel by modifying
the LSB of the R, G, B components of the pixel. This Step 3: Exclusive-or the results from step i and ii to
will not make any difference between the visual get s1s2s3s4 where s1 = r1⊕c1, s2 = r2⊕c2, and
appearance of the resulting stego image and that of the so on.
original cover image.
Step 4: Compare the results obtained from step 3 with
The Encrypt and Scatter technique is a method of the four embedded bits b1b2b3b4. If there is
image steganography to hide the data to be sent being no difference, no change of bits in F is
embedded in an image. This technique of embedding needed, otherwise, consider the following
message makes it appear more like noise and it is cases:
generally done using LSB modification. The demerits of
this approach is that even if the message (encrypted • if the difference is in one bit bi, the bit [F] i,5 or [F]5,i
version) bits are extracted successfully, it will be useless should be complemented
until we are able to decode the message using the
appropriate stego-key[10]. • else if difference in two bits bi and bj, then the bit [F]
i,j or [F]j,i should be complemented or bit [F] 5,i and
An interesting method, proposed by G Sahoo and R K [F]5,j should be complemented.
Tiwari in 2008, works on more than one image using the • else if difference in three bits bi, bj and bk, then the bits
concept of file hybridization. This particular method (( [F] i,j or [F] j,i) and ( [F]k,5 or [F]5,k)) or
implements the cryptographic technique to embed two (( [F] i,5 or [F] 5,i) and ( [F]k,i or [F]j,k)) or
information files using steganography. Thus they have (( [F] 5,j or [F] j,5) and ( [F]k,i or [F]i,k))
used a stego key for the embedding process[11]. should be complemented.

Another method of steganography has been proposed • else (if difference in four bits bi, bj, bk and bm) then
by Ahmed Al-Jaber and Khair Eddin Sabri. This the bits
algorithm allows hiding four bits in a block of size 5 * 5 (( [F] i,j or [F] j,i) and ( [F]k,m or [F]m,k)) or
by changing a maximum of two bits. The selection of the (( [F] i,m or [F] m,i) and ( [F]k,i or [F]j,k)) or
bits to be changed in the block depends on the number of (( [F] m,j or [F] j,m) and ( [F]k,i or [F]i,k))
adjacent bits with the same value. The bit that has the should be complemented.
least number of adjacent bits is selected[12].
Note : The bit(s) to be changed is selected in such a
III. PROPOSED TECHNIQUE way that the MSB and next two significant bits of the
RGB values of each pixel of the cover image remain
In this paper, we have presented an efficient method unaffected so that there is minimal distortion in the cover
of steganography. This method is based on taking pixel image.
information from the cover image and forming a matrix,
each of size 5×5. In each matrix, 4 bits from the secret Steps for Extracting the data from the stego image :
message can be embedded.
Algorithm 2
For embedding of data the following steps are
adopted : Input : Stego-image

Algorithm 1 Output : Secret message

Input : Cover image and secret message The algorithm used for extracting the embedded data
is similar to that used for embedding. The following steps
Output : Stego-image. are carried out to obtain the embedded data.

Step1 : Divide the cover image into blocks (F) each of Step1 : Divide the cover image into blocks (F) each of
size 5×5. size 5×5.

Step2 : For each block, we proceed as follows Step2 : For each block, we proceed as follows:

Copyright © 2012 MECS I.J. Intelligent Systems and Applications, 2012, 1, 32-38
34 An Efficient Method of Steganography using Matrix Approach

i) For every row in the first four rows of the It can be seen that the bits number 2 and 4 in the
block, exclusive-or all the bits of that row result are different form those in the embedded data. So
to get r1r2r3r4. there should be a change in either the bit [F]2,4 or [F]4,2.
ii) For every column in the first four To minimize its effect on the cover image, bit [F]4,2
columns of the block, exclusive-or all the should be changed. In the other block, if the same
bits of that column to get c1c2c3c4. operation is repeated, the following results are obtained:

Step 3 : Exclusive-or the results in 2.i. and 2.ii. to get r1r2r3r4 = 0101 c1c2c3c4 = 0010 s1s2s3s4 = 0111
the embedded bits s1s2s3s4 where s1=r1⊕c1, The embedded data is 0101.
s2=r2⊕c2, and so on.
Therefore, the bit that should be changed is either
Step 4 : Stop. [F]3,5 or [F]5,3. Again due to the same reason as stated
above [F]3,5 is to be changed.
Example : Suppose that we have the following cover
image: IV. EXPERIMENTAL RESULTS
The proposed algorithm has been applied on different
1111110010 images for different data. The same image-data
1100010101 combinations have been used for experimentation with
0101000000 the existing algorithm for comparision. The following
1010011010 computations were performed for each stego-image:
1111111000
Average : It is computed for each pixel depending on
and the embedded data is : 11100101 its neighbors. Then the average of pixel average values is
also computed to test the consistency between each pixel
and its neighbors.

Following the said necessary steps, for every row in Standard Deviation : Compute the average for each
the first four rows of the block, we exclusive-or all the pixel depending on its neighbors, and then compare it
bits of that row to get r1r2r3r4 on the given block and the with the original image.
following result is obtained :
Here we have considered three .bmp files as the cover
1⊕ 1 ⊕ 1 ⊕ 1 ⊕ 1 = 1 images and three text files to be embedded in those cover
1⊕ 1 ⊕ 0 ⊕ 0 ⊕ 0 = 0 images. The output consists of three stego images which
0⊕ 1 ⊕ 0 ⊕ 1 ⊕ 0 = 0 are .bmp files.
1⊕ 0 ⊕ 1 ⊕ 0 ⊕ 0 = 0
Experiment No. 1:
The result is 1000 In this experiment we have used the following cover
image as shown in Fig1.1. The text to be embedded is
Following the said necessary steps for every column presented in Table 1. The stego image is shown in Fig
in the first four columns of the block, we exclusive-or all 1.2 . It is found that the proposed method can
the bits of that column to get c1c2c3c4 on the same block successfully embed the given text into the cover image
to get the following result: and is also able to extract the given text (secret message)
from the stego image. It may be noted that here all the
1⊕1⊕0⊕1⊕1=0 words of the secret message are successfully extracted in
1⊕1⊕1⊕0⊕1=0 the right order without any error.
1⊕0⊕0⊕1⊕1=1
1⊕0⊕1⊕0⊕1=1

The result is 0011

Now we exclusive-or the results obtained in 1 and 2

to get the embedded bits s1s2s3s4 where s1=r1⊕c1,
s2=r2⊕c2, and so on for the same block to get the
following result:

1⊕0=1
0⊕0=0
0⊕1=1
0⊕1=1

The result is 1011 and the embedded data is 1110.

Copyright © 2012 MECS I.J. Intelligent Systems and Applications, 2012, 1, 32-38
 An Efficient Method of Steganography using Matrix Approach 35

Fig 1.1 Original Image

Fig. 2.1 Original Image

Table 1 Text to be embedded
India is our motherland,
Its the land of our Table 2 Text to be embedded
love, We are proud
to be an Indian.
// bmp in C
We serve our life
#include <stdio.h>
for our beloved
#include<stdlib.h>
country.
#include<math.h>
#include<ctype.h>
#include<conio.h>

typedef struct tagBITMAP

{
unsigned short bfType; //might
need to be a char
unsigned long size;
unsigned short bfReserved1;
unsigned short reserved2;
unsigned long offset;
unsigned long sizeofstruct;
unsigned long width;
unsigned long height;
unsigned short planes;
unsigned short bits;
unsigned long compression;
Figure 1.2 Stego Image unsigned long imagesize;
unsigned long xresolution;
Experiment No. 2: unsigned long yresolution;
In this experiment we have used the following cover unsigned long ncolors;
image as shown in Fig2.1 . The text to be embedded is unsigned long impcols;
presented in Table 2. The stego image is shown in Fig } BITMAP;
2.2 . It is found that the proposed method can
successfully embed the given text into the cover image
and is also able to extract the given text (secret message)
from the stego image. It may be noted that here all the
words of the secret message are successfully extracted in
the right order without any error.

Copyright © 2012 MECS I.J. Intelligent Systems and Applications, 2012, 1, 32-38
36 An Efficient Method of Steganography using Matrix Approach

Table 3 Text to be embedded

260320 Nirmalendu mondal Near

kalikapur hospital
Haanpur,Champahati,pin-743330
261803 Smt Bijoli Mondal
Sahebpur Champahati,
Pin-743330
260545 Sri Debabrata Chatterjee
vill_p.o.-Champahati,
Pin-743330
260317 Sri Netai Charan Maity
Kalikapur
Pin-743330
260163 Sri someswar Banerjee Vill+
P.O.-South Garia Pin-
743618
260855 Sri Sovan Chatterjee
Vill+P.O. South Garia
260856 Sri Barendra nath Roychoudhury
Figure 2.2 Stego Image Vill+P.O. South Garia
260447 Sri Subal Ghosh Champahati
Experiment No. 3: Main Road,Champahati
In this experiment we have used the following cover 261726 Sri Lalit Bhattacharjee
image as shown in Fig3.1 . The text to be embedded is Vill+P.O. South Garia
presented in Table 3. The stego image is shown in Fig 260821 Sri Anil Kr Naskar
3.2 . It is found that the proposed method can Vill+P.O. South Garia
successfully embed the given text into the cover image 261651 Sri Ranjan Banerjee South
and is also able to extract the given text (secret message) Garia PIN-73613
from the stego image. Here also all the words of the 260432 Sri Sujan Chakraborty O/O
secret message are successfully extracted in the right The CAO(TR)
order without any error. GM,CAL SSA

Txt

Figure 3.1 Original Image

Figure 3.2 Stego Image

It may be noted that in all of the above three

experiments, no difference in the visual appearance
between the cover image and the stego image is
observed. This is important since if any difference in
appearance between the cover image and the stego
image is introduced due to the embedding of the
secret message, then someone may doubt or even
decipher the embedded text from the stego image.

Copyright © 2012 MECS I.J. Intelligent Systems and Applications, 2012, 1, 32-38
 An Efficient Method of Steganography using Matrix Approach 37

Table 4. and Table 5. show the comparision of the

existing and the proposed method in terms of the
average and standard deviation respectively. The said
Average
comparisons are also graphically demonstrated in Fig.
4 and Fig. 5. It can be clearly seen that the proposed 45
algorithm shows a better performance in terms of both 40
35
the average and standard deviations between the 30 Average of Existing
original image and stego-image.

Average
25 Algorithm
20 Average of Proposed
15 Algorithm
10
5
Table 4 Average using existing and proposed algorithm 0
0 5 10 15 20
File size
File Average using Average using
size(KB) Existing Proposed
Algorithm Algorithm Figure 4 Average using existing and proposed algorithm

1 25.80 23.59
Standard Deviation
2 39.42 26.99
45
3 22.69 21.16 40
Standard Deviation
35
4 25.80 21.91 30 Standard Deviation of
25 Existing Algorithm
6 28.07 21.16 20 Standard Deviation of
15 Proposed Algorithm
7 23.59 15.86 10
5
9 22.69 18.87 0
0 5 10 15 20
10 20.43 14.98 File size

15 28.07 18.87
16 29.00 22.69
Figure 5 Standard deviation using existing and
proposed algorithm

V. CONCLUSIONS
Table 5 Standard deviation using existing and In this paper, the technique which has been proposed
proposed algorithm to hide data within a color image is secure and the hidden
information is quite invisible. The advantages of the
File Standard Standard proposed technique are:
size(KB) Deviation Deviation using ƒ The proposed algorithm shows a better
using Existing Proposed performance in terms of both the average and
standard deviations between the original image
Algorithm Algorithm and stego-image.
1 24.37 21.60 ƒ The proposed algorithm does not need a secret
key. It needs only an agreement between the
2 35.10 34.18 embedding and extracting agents.
ƒ The proposed algorithm ensures that the MSB
4 35.33 34.47 and the next two bits of any pixel value of the
6 37.91 37.10 original image will not be modified. So the
distortion of the image will be less compared to
7 36.83 35.92 that of the existing algorithm.
9 39.55 39.19
A comparison of the proposed algorithm with the
10 31.62 30.92 existing scheme shows that the proposed algorithm gives
a better performance in terms of the average and standard
16 33.78 32.32 deviation factors.

Copyright © 2012 MECS I.J. Intelligent Systems and Applications, 2012, 1, 32-38
38 An Efficient Method of Steganography using Matrix Approach

References

[1] Y. Chen, H. Pan, and Y. Tseng, “A secure Data

Hiding Scheme for Two-Color Images,” IEEE
Symposium On Computers and Communications, 2000.
[2] C.Cachin, ”An Information-Theoritic Model
for Steganography”, Proc. of the 2nd International
Workshop on Information Hiding, Portland, Oregon,
USA,15-17 Apr 1998, Lecture notes in CS,
Vol.1525, Springer-Verlag.
[3] N. Johnson and S. Jajodia, “Exploring steganography:
seeing the unseen,” IEEE Computer, pp. 26-34, February
1998.
[4] R. Chandramouli and Nasir Memon, "Analysis of
LSB Based Image Steganography Techniques",2001
International Conference on Image Processing, October
7-10, 2001, Thessaloniki, Greece, Vol. 3, pp. 1019-1022.
[5] 1. Johnson, N. F. and Jajodia, S. (1998).
Exploring steganography:Seeing the unseen.
Computer,31(2):26–34.
[6] C.T.Hsu and J.L.Wu., ”Hidden Singatures in
Images”, Proc. IEEE International Conf. on Image
Processing, ICIP-96, Vol.3, pp.223-226.
[7] Niels-Provos, Peter Honeyman, Hide and Seek:
Introduction to steganography(2003).
[8] K. Curran and K. Bailey, "An Evaluation of
Image Based Steganography Methods," International
Journal of Digital Evidence, vol. 2, issue 2, pp. 1-40,
Fall 2003.
[9] Luis von Ahn, Nicholas J. Hopper., Public Key
Steganography.
[10] Robert Krenn. Steganography and steganalysis.
[11] G Sahoo, R K Tiwari “Designing an Embedded
Algorithms for Data Hiding using Steganographic
Technique by File Hybridization” , IJCSNS, vol 8, No
1,pp-226-233, January 2008
[12] Ahmed Al-Jaber , Khair Eddin Sabri “Data Hiding
in a Binary Image”

Copyright © 2012 MECS I.J. Intelligent Systems and Applications, 2012, 1, 32-38