1. International Journal of Electronics and Communication Engineering & Technology (IJECET),
ISSN 0976 – 6464(Print), ISSN 0976 – 6472(Online) Volume 4, Issue 3, May – June (2013), © IAEME
154
MODIFIED WEIGHTED EMBEDDING METHOD FOR IMAGE
STEGANOGRAPHY
#1
Geetha C.R. #2
Dr.Puttamadappa C.
Research scholar, Professor and principal,
Jain University Sapthagiri College of Engineering,
Bangalore, Karnataka, INDIA Bangalore, Karnataka, INDIA
ABSTRACT
In this paper, proposing a unique steganography method for the digital images. In
steganography embedding the data inside an image before transmitting it for secure
communication. The basic method used in time domain for image steganography is LSB
substitution which can easily be detected. Hence instead of simple LSB substitution here
choosing an embedding algorithm which uses the decimal value of the data to be hidden. This
makes the data to be hidden as the key itself. With this method high security along with good
visual quality is obtained.
Keywords: Security, Cryptography, Multiple edge detection, Minimum error replacement,
Variable embedding ratio.
I. INTRODUCTION
Present communication system uses digital signals more than analog signals. Digital
communication system uses digital signals, which are sent through the communication
channel. Communication channels are susceptible for the intrusions. Internet is the major
channel for communication and is very much necessary to secure the data transmitted through
the internet. Many methods are developed to secure data. Data security is the major area
where many developments are required. Cryptography is used extensively to secure the data.
Steganography is another method which can be used to secure the data. Cryptography
modifies the data with respect to some fixed model, so that the information is scrambled. But
in Steganography a cover image is taken and the data is embedded into it in such a way that
cover image is least altered. So in Steganography, not only the information is hidden also the
existence of the information is hidden. It may be some time is required to hide the presence of
INTERNATIONAL JOURNAL OF ELECTRONICS AND
COMMUNICATION ENGINEERING & TECHNOLOGY (IJECET)
ISSN 0976 – 6464(Print)
ISSN 0976 – 6472(Online)
Volume 4, Issue 3, May – June, 2013, pp. 154-161
© IAEME: www.iaeme.com/ijecet.asp
Journal Impact Factor (2013): 5.8896 (Calculated by GISI)
www.jifactor.com
IJECET
© I A E M E

2. International Journal of Electronics and Communication Engineering & Technology (IJECET),
ISSN 0976 – 6464(Print), ISSN 0976 – 6472(Online) Volume 4, Issue 3, May – June (2013), © IAEME
155
secrete data rather than just making the data encrypted. Steganography is hence used to make
the data more secure and then its existence is made secrete as well. It’s better to use a digital
image for Steganography. Digital images can store large amount of data. For example let us
take an example of 512×512 grayscale image will have 262,144 pixels and each pixel is made
up of 8-bits. So the total numbers of bits in the image is 2,097,152. The total number of LSB
bits in the image is 262,144. If we embed only for LSB bits in the image we can still hide
262,144 bits of data. This makes digital images suitable for Steganography. Also digital
image is a common data which is shared in the internet and that makes image steganography
better than any other.
The simplest way for image steganography is LSB substitution method. In LSB
substitution method the information which is to be hidden is converted into bit stream. The
LSB bits of each pixel of the image is then substituted or replaced by the information bits.
Chan C.K and Chen L.M [1] proposed simple LSB substitution method. Simple LSB
substitution method is simple but it is not secure. Simply by extracting the LSB bits we can
get the information back. Wang R.Z, Lin C.F and Lin J.C [2] proposed moderately significant
–bit replacement method. Marghny Mohamed, Fadwa Al-Afari, and Mohammed Bamatraf
[3] proposed LSB substitution by genetical-optical key permutation.
Neighbor pixel information method is another way in which the number bits
substituted for a given pixel is dependent on its neighbor pixel values. MoazzamHossain,
SadiaAl Haque, FarhanaSharmin [4] proposed steganography method using neighborhood
pixel information. But the PSNR is obtained was about 43.144dB. For the less amount of data
hidden a PSNR of 43.144 dB was less. Other methods include Edge detection which along
with Variable embedding can be used to increase the amount of data hidden by a large margin
but the PSNR dropped drastically which went against the definition of steganography.
Hence in this paper we propose the Value Weighted embedding method which clearly
retains high visual quality as the cover image with little compromise in the amount of data
embedded. This method can also be combined with many error reducing methods such as
MER and AER to further increase visual quality.
II. PROPOSED METHOD
In this method while embedding the data the decimal value of the data is considered.
The data to be embedded is first converted into decimal values, then these decimal values are
used to embed the data. Here instead of directing replacing the LSB with the data bits we use
the magnitude of the input data to just compliment certain LSB. The exact process of
embedding is given in Fig.1
A. Embedding Algorithm
a) Let the data to be embedded be in the form of bit stream 'X'.
X= [X13 X12 X11 X10 X9 X8 X7 X6 X5 X4 X3 X2 X1 X0]
where, X0 to X13 take values 0 or 1.
Example:-
X= [00000110010101000111100101...]

3. International Journal of Electronics and Communication Engineering & Technology (IJECET),
ISSN 0976 – 6464(Print), ISSN 0976 – 6472(Online) Volume 4, Issue 3, May – June (2013), © IAEME
156
b) Next convert this input bit stream into a form of 1D array of 3 digits decimal numbers
'Y'. Hence divide the input bit stream into groups of 13 bits and convert
Fig.1: Block diagram for Data Embedding
them into decimal by multiplying in powers of 2. This step can be avoided if the input is
characters by considering their ASCII values. This step can be avoided if the input is
characters by considering their ASCII values.
Y= [ X8*28
+ X7*27
+ X6*26
+ X5*25
+ X4*24
+ X3*23
+ X2*22
+ X1*21
+ X0*20
]
Example: Y = [0405, 0485..]
c) Consider the original image called the cover image. Find the edges. Now divide the array
of edges into 1x5 blocks and consider these blocks as 1D arrays of pixels with 5 elements
each.
d) Now consider the first decimal data to be embedded, divide that number into 3 digits
according to their position weight i.e. 1000th
, 100th,10th and units place. Each digit is
embedded separately depending upon their position value and decimal value.
Example Z = [0 ,4 , 0, 5 ]
e) Now move to the next decimal value and the next image block and repeat the steps d)
and e) till all the data is

4. International Journal of Electronics and Communication Engineering & Technology (IJECET),
ISSN 0976 – 6464(Print), ISSN 0976 – 6472(Online) Volume 4, Issue 3, May – June (2013), © IAEME
157
Let First Image block = [ 100, 105, 103, 102, 100 ]
Digits to be embedded Z = [ 0, 4, 0, 5 ]
f) Now move to the next decimal value and the next image block and repeat the steps d)
and e) till all the data is embedded or the entire image is used. The image obtained at the end
is called the Stego image.
g) The above table given here is the normal order without any randomness. But this order can
changed like the LSB complimented for unit and tenth places can be interchanged. This
increases security and these tables act as the key for the embedding and retrieval process.
Ref. Fig.3
Table 1: Change of Pixel as per Digits Value
Digits Value Pixels to be Changed in 1x5 Matrix
0 None
1 First
2 Second
3 Third
4 Fourth
5 Fifth
6 First and Fifth
7 Second and Fifth
8 Third and Fifth
9 Fourth and Fifth
The above table given here is the normal order without any randomness. But this
order can changed like the LSB complimented for unit and tenth places can be interchanged.
This increases security and these tables act as the key for the embedding and retrieval
process.
Table 2: Key for Embedding
Digits Position LSB to be
Complimented
Unit Last LSB
Tenth 2nd
Last LSB
Hundredth 3rd
Last LSB
Thousand 4th
Last LSB

5. International Journal of Electronics and Communication Engineering & Technology (IJECET),
ISSN 0976 – 6464(Print), ISSN 0976 – 6472(Online) Volume 4, Issue 3, May – June (2013), © IAEME
158
Fig.2: Retrieval block diagram
Experiments are carried out to support our theory and the following results are found.
Experiments are done in MATLAB 2011a version. Experiment is carried out on different
grayscale images and a comparative study is presented.
B. Retrieval
Retrieval is actually easier; the receiver first subtracts the cover image and the stego
image. Then uses absolute function on the difference to convert all 1's and -1's to 1's. Again
divides the image into many 1 x 5 blocks. Then using the key i.e. the embedding table the
retriever obtains the original data.
a) Edge detection is first applied on an untouched cover image present in the database and its
edges are detected.
b) These pixels are stored in the form of 1D array for both the cover image and also the
absolute function on it. This is done to detect the pixels whose value is changed and also
make sure that the array has no negative values. Stegoimage, hence there are now two 1D
arrays of pixels.
c) Another array is formed by subtracting these both the arrays an applying absolute
d) The array is divided into 1x5 blocks again, then with the help of the value present in the
array and the key i.e. the table the data is extracted

6. International Journal of Electronics and Communication Engineering & Technology (IJECET),
ISSN 0976 – 6464(Print), ISSN 0976 – 6472(Online) Volume 4, Issue 3, May – June (2013), © IAEME
159
Table 3: Methods proposed by MoazzomHosain, Sadia Al Haque and FarhanaSharmin
Fig.3: Embedding Method
III. EXPERIMENTAL ANALYSIS
Table 4: Value weighted embedding method for whole image
Bits Hidden PSNR in dB
262144 (1 Bpp) 43.21
470025(Max) 40.65
Method Maximum Hiding
Capacity(In Bits)
PSNR(In
dB)
Four Neighbors 392208 41.1468
Diagonal
Neighbors
395680 40.6505

7. International Journal of Electronics and Communication Engineering & Technology (IJECET),
ISSN 0976 – 6464(Print), ISSN 0976 – 6472(Online) Volume 4, Issue 3, May – June (2013), © IAEME
160
Fig.4: Baboon a)Original b)Stego
Table 5: Value weighted embedding only in edge pixels
Image Bits Hidden PSNR in dB
Lena 96785 42.6446
Baboon 155220 40.2119
Peppers 84708 42.8209
By observing the two tables 4 and 5 it can seen that in our method not only the hiding
capacity has increased but also the visual quality is high compare to the Four neighbors and
Diagonal neighbors method.
IV CONCLUSION
With this method the security increases significantly with little distortion of the visual
quality, also the data hiding capacity is good. For future work the visual quality can be
improved by employing some error reducing techniques such as MER.
Fig.5: Lena a)Original b)Stego

8. International Journal of Electronics and Communication Engineering & Technology (IJECET),
ISSN 0976 – 6464(Print), ISSN 0976 – 6472(Online) Volume 4, Issue 3, May – June (2013), © IAEME
161
REFERENCES
[1] Chan K. and Cheng M. “Hiding Data in Images using Simple LSB Substitution”
Computer Journal of Pattern Recognition Letters, volume 37,No.3, PP . 469-474, 2004.
[2] Wang R. Z , Lin C. F and Lin J. C “Hiding Data in Images Optimal Moderately
Significant –bit replacement”, IEE Electron Letter, vol. 36, no. 25, PP. 2069-2070, 2000.
[3] Marghny Mohamed, Fadwa Al-Afari, and Mohammed. Bamatrf “Data Hiding by LSB
substitution using Genetic Optimal Key Permutation”, International Arab Journal of
e-technology, vol 2, No.1 January 2011.
[4] MoazzamHossain, Sadia Al Haque, FarhanaSharmin. “Variable Rate Steganography
International Arab Journal of information technology, vol. 7, No. 1, January 2010.
[5] Li Li, Bin Luo, and Qiang Li Xiaojun Fang., “A color Images steganography method by
multiple embedding strategy based on Sobel operator conference on multimedia
information networking.
[6] Sonali Patil, Kapil Tajane and Janhavi Sirdeshpande, “Analysing Secure Image Secret
Sharing Schemes Based on Steganography”, International Journal of Computer
Engineering & Technology (IJCET), Volume 4, Issue 2, 2013, pp. 172 - 178, ISSN Print:
0976 – 6367, ISSN Online: 0976 – 6375.
[7] Vismita Nagrale, Ganesh Zambre and Aamir Agwani, “Image Stegano-Cryptography
Based on LSB Insertion & Symmetric Key Encryption”, International Journal of
Electronics and Communication Engineering & Technology (IJECET), Volume 2,
Issue 1, 2011, pp. 35 - 42, ISSN Print: 0976- 6464, ISSN Online: 0976 –6472.