IMAGE STEGANOGRAPHIC TECHNIQUES

WITH IMPROVED EMBEDDING CAPACITY

AND ROBUSTNESS
J.Anita Christaline
1
, D. Vaishali
2

Assitant Professors ECE,
SRM University, Chennai. India.
1
[email protected] ,
2
[email protected]



Abstract
Secured data transmission over computer networks
can be achieved through steganography. In
specific, Image Steganography entails the
opportunity of hide any secret information into
images. This paper presents the implementation of
two image steganographic techniques in MATLAB.
The first is a filter method to embed text
information into image and new methods have
been demonstrated to increase the information
embedding capacity in the same domain. The
second method is the wavelet transform method
which proves to be more secured than any other
method of image steganography.

Keywords Information hiding, steganography,
masking and filtering method, transform method,
stego image, carrier, key.

I. INTRODUCTION
Recent years have witnessed the rapid
development of the Internet and
telecommunication techniques. In such a
scenario, information security is becoming more
and more important. Applications such as covert
communication, copyright protection, etc,
stimulate the research of information hiding
techniques. Traditionally, cryptography was
used to realize secured communication. In
cryptography, important information is not
protected as the cipher text message can arouse
suspicion. This necessitated the need for hiding
information rather than modifying the
information. The secret information is initially
hidden in a host data, such as digital image,
video or audio, etc. and then transmitted secretly
to the receiver. Generally, information hiding
techniques can be classified under two
categories, namely, information hiding
techniques in time domain or spatial domain and
those in transform domain. Information hiding
algorithms in time domain or spatial domain
have high capacity and relatively lower
robustness. In contrast, the algorithms in
transform domain, such as DCT [1], Wavelet
domain [2][3], have certain robustness against
other multimedia processing.
This paper, intends to implement two
methods of image steaganography, namely the
filter method (time domain) for high capacity
text communication and the wavelet method
(transform domain) for highly secured image
communication. While the first method provides
high embedding capacity, the second method
focuses on high robustness.

II. IMAGE STEGANOGRAPHIC
APPROACHES
There are many techniques for hiding
information or messages in images in such a
manner that the alterations made to the image are
perceptually indiscreminable. The most common
image steganographic approaches include :
Least significant bit insertion
(LSB).
Masking and filtering.
Transform techniques.

A. Masking and Filter method of Image
Steganography:
The masking and filtering method is
usually restricted to 24 bit images. This method
uses specific masking algorithms or
mathematical formulae to select specific pixels
from the cover image. The chosen pixels can be
used to embed the secret information after due
pre-processing. Thus the embedded secret
information looks as an integral part of the cover
image [4]. By using proper filter (mathematical
expression) to select the pixels, the security of
IEEE-International Conference on Recent Trends in Information Technology, ICRTIT 2011
978-1-4577-0590-8/11/$26.00 2011 IEEE
MIT, Anna University, Chennai. June 3-5, 2011
97
the image steganograpophic method is enhanced
and the embedding capacity can be increased by
choosing many pixels.
B. Algorithm for Embedding Logic in Filter
method:
Input : Secret Text message, Carrier image.
Output : Embedded (Stego) Image.
1. Choose a set of pixels from the carrier
image using a mathematical function
(KEY).
2. The gray / red level values of these
pixels are processed.
3. Convert secret text information into bit
stream (ASCII coding).
4. If the first bit in the text bit stream is
zero, then the first pixel value is not
modified.
5. If the first bit happens to be one, then
the first pixel is made to be odd by
adding one to the pixel value.
6. Repeat above logic till the text bit
stream expires.
7. The modified pixels are put back into
the original cover image to get the
Stego Image.
C. Algorithm for Debedding Logic in Filter
method:
Input : Embedded (Stego) Image, Key.
Output : Secret text message.
1. Identify the specific pixels in stego
image by using the same mathematical
function (KEY).
2. Examine the pixel values. If the pixel
value is odd, then the embedded text bit
was a one.
3. If the pixel value is even, then the
embedded text was a zero.
4. Based on the above interpretation,
retrieve the text bit stream.
5. Convert bit stream into ASCII and then
to the respective character to get the
secret text.
D. Algorithm to Increase the Embedding
Capacity:
Embedding capacity is the maximum
number of characters that can be embedded in
an image. It depends on size of the image, the
key used to select pixels, type of coding. Thus
ways to increase embedding capacity could be to
use an user defined coding (either 5 or 6 bit
coding) rather than the usual 8 bit ASCII. Also
some word compression logic can be used to
reduce the number of characters per word to
increase the embedding capacity.
This paper intends to implement 5 bit user
coding to increase the embedding capacity and a
4D dictionary of words to work with a specific
word compression and retrieval logic.
E. The Dictionary Approach for increased
embedding capacity:
This method has two parts, namely the
word compression logic in the sender side and
the word retrieval logic in the receiver side.
Word compressing logic
Input: Text information.
Output: Compressed text.
1. Read the first (or next) word.
2. Keep the first and the last letter intact.
3. Delete the alternate letters from the remaining
word.
5. Repeat the above for all words.
Word Retrieval logic
Input: Compressed text.
Output: original text
1. Read the first (or next) compressed word.
2. Keep the first and the last letter intact.
3. Calculate the number of letters in the word
4. Perform the search in the 4 D Dictionary.
5. Identify the word.
6. Repeat for all compressed words.

F. The 5 bit User Coding for increased
embedding capacity:
Text Coding logic
Input: Compressed Text
Output: Binary Stream based on User
Coding.
1. Read the first (or next) word.
2. Convert it into binary based on the user
coding.
3. Construct the binary stream.
Text decoding logic
Input: Binary Stream
Output: Compressed Textual Data
1. Read the first (next) 5 bits from text bit
stream.
2. convert to equivalent character based on User
Coding.

III. WAVELET BASED IMAGE
STEGANOGRAPHY
Wavelet transform is used to convert a
spatial domain into frequency domain [5]. The
use of wavelet in image stenographic model lies
in the fact that the wavelet transform clearly
separates the high frequency and low frequency
information on a pixel by pixel basis. Discrete
Wavelet Transform (DWT) is preferred over
Discrete Cosine Transforms (DCT) because
image in low frequency at various levels can
offer corresponding resolution needed. The
DWT decomposition results into four classes or
IEEE-ICRTIT 2011
98
band coefficients, namely, approximate band
(LL), Vertical Band (LH), Horizontal band (HL),
and diagonal detail band (HH). The
approximation band consists of low frequency
wavelet coefficients, which contain significant
part of the spatial domain image. The other
bands also called as detail bands consists of high
frequency coefficients, which contain the edge
details of the spatial domain image[6]. It is
therefore expected that use of the DWT will
allow independent processing of the resulting
components without significant perceptible
interaction between them, and hence makes the
process imperceptibility marking more effective.

A. Embedding logic for DWT image
steganography:
Input : Carrier image (KEY) , Secret
image.
Output : Stego image.
1.Apply DWT ( db4) on to the carrier image to
get the decomposed carrier image (Four bands)
2.Apply DWT (db4) on to the secret Image to get
the decomposed secret image (Four bands)
3. Apply wavelet fusion of DWT coefficients of
the carrier and secret images.
4. Take Inverse DWT with the four sub bands of
the fused image to get the Stego image.

B. Debedding logic for DWT image
steganography:
Input : Stego image, carrier image
(KEY)
Output : Secret image.
1. Apply DWT (db4 ) on to the received Stego
image to get the four sub bands ( wavelet
coefficients )
2. Apply DWT (db4) on to the carrier image
(KEY) , to get the four sub bands ( wavelet
coefficients ).
3. Subtract DWT coefficients of carrier image
from that of the Stego image to get the DWT
image of the secret image.
4.Apply Inverse DWT to get the secret image.

IV. EXPERIMENTAL RESULTS
A. Experimental results for Filter method:











MATLAB GUI showing Embedding Logic


MATLAB GUI showing comparison of original carrier
image and embedded stego image


MATLAB GUI showing the retrieval in receiver end


MATLAB GUI showing the implementation of Dictionary
and user coding to increase embedding capacity


MATLAB GUI showing the retrieval with Dictionary and
user coding in receiver end







Image Steganographic Techniques with Improved Embedding Capacity and Robustness
99

B. Experimental results for Wavelet method:

Original secret image to be transmitted




Carrier image





Stego image that is transmitted










Retrieved image


V. PERFORMANCE EVALUATION
A. Filter Method
For the filter method, the performance is
evaluated by the Comparison of maximum
embedding capacities as in following tables.

COMPARISON OF DATA FOR FILTER METHOD
Image
File
Name
Key used

No. of Pixels
Selected
Max. no. of
Characters that can be
Embedded.
ASCII 5 bit UC


Peppers.
png
50x50 88 11 17
10x10 2028 253 405
Alternate
pixels
98304

12288 19660
All pixels 196608 24576 39321


rice.png
50x50 36 4 7
10x10 676 84 135
Alternate
pixels
32768 4096 6553
All pixels 65536 8192 13107

COMPARISON FOR DIFFERENT CODING METHODS
FOR HIGH EMBEDDING CAPACITY
Type of Steganography % Increase in
Embedding capacity
8 -Bit Steganography 0 ( ref.)
6 Bit Steganography 25 %
5 Bit Steganography 38 %
8 Bit Steganography
with Dictionary
Compression
37 %
6 Bit Steganography
with Dictionary
Compression
62 %
5 Bit Steganography
with Dictionary
Compression
75 %

IEEE-ICRTIT 2011
100


B. Wavelet Method
Performance evaluation for the DWT ( db4)
method has been done in terms of MSE, PSNR
and Entropy [7] of the images.

COMPARISON OF MSE
Images Type Size MSE
Between
Carrier
and
stego
image
Between
secret and
retrieved
image
Bust
(secret)
mask
(carrier)
JPEG 256X256 1.000 2.32 e -25
Woman
(secret)
mask
(carrier)
JPEG 256X256 1.000 -2.65 e -
27

COMPARISON OF PSNR AND ENTROPY
images PSNR Entropy
Stego
image
Retrieved
image
Stego
image
Retrieved
image
Bust
(secret)
mask
(carrier)
21.45 24.45 0.0 0.0
Woman
(secret)
mask
(carrier)
22.17 26.036 0.0012 0.0012


VI. CONCLUSION AND FUTURE SCOPE
This paper has implemented two
different steganographic methods. The filter
method shows that the embedding capacity can
be increased and the wavelet method shows that
the security of data is highlightened.
In future, this paper can be extended to
improve the robustness of the filter method by
improving the KEY used for pixel selection. For
the wavelet method, the future work may include
considering different mother wavelets to enhance
the robustness. Also various class of images can
be considered which will improve the MSE and
PSNR.


REFERENCES
[1] Chiou-Ting Hsu, Ja-Ling Wu. Hidden Digital
Watermarks In Images, IEEE Transaction on Image
Processing, 8(1): 56-58,January 1999.

[2] Nedeljko Cvejic, Tapio Seppanen. A Wavelet
Domain LSB Insertion Algorithm For High Capacity Audio
Steganography, Digital Signal Processing,
workshop 2002 and the 2nd signal processing
education workshop. Proceedings of 2002 IEEE,
10th, 13-16, Oct, 2002, Pages:
53-55.

[3] NI Ronggong, RUAN Qiuqi. Embedding
Information into Color Images Using Wavelet,
TENCON02, Proceedings, 2002 IEEE Region
10 Conference on Computers, Communications,
Control and Power Engineering, Volume:1, 28-
31 Oct, 2002, pages: 589-601, Vol.1.

[4]Sahoo1 and R. K. Tiwari, Designing an Embedded
Algorithm for Data Hiding using Steganographic Technique
by File Hybridization, IJCSNS International Journal
of Computer Science and Network Security,
VOL.8 No.1, January 2008.

[6] H S Majunatha Reddy, & K B Raja, HIGH
CAPACITY AND SECURITY STEGANOGRAPHY USING
DISCRETE WAVELET TRANSFORM, International
Journal of Computer Science and Security
(IJCSS), Volume (3): Issue (6)

[5] Bo Yang and Beixing Deng, Steganography In
Gray Images Using Wavelet,
http://online.redwoods.cc.ca.us/instruct/darnold/
maw/catmap.htm

[7] FRANK Y. SHIH , Image Processing and Pattern
Recognition, Fundamentals and Techniques, John Wiley
& Sons, Inc., Hoboken, New Jersey., 2010,
PP.71.

Image Steganographic Techniques with Improved Embedding Capacity and Robustness
101