Document Sample
					  International Journal of JOURNAL OF COMPUTER (IJCET), ISSN 0976-
 INTERNATIONALComputer Engineering and Technology ENGINEERING
  6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME
                            & TECHNOLOGY (IJCET)

ISSN 0976 – 6367(Print)
ISSN 0976 – 6375(Online)                                                    IJCET
Volume 4, Issue 2, March – April (2013), pp. 102-107
Journal Impact Factor (2013): 6.1302 (Calculated by GISI)


                    Prabakaran. G 1, Dr. Bhavani.R 2 and Kanimozhi.K 3
                          Assistant Professor , 2 Professor, 3 PG student
                 Dept. of CSE, FEAT, Annamalai University, Tamilnadu, India


          Steganography is an art of hiding the information in the cover such as a way that, it
  looks like simple cover, though it has hidden information. In this paper, we presented a two
  secret image hiding using Singular Value Decomposition (SVD) and Discrete Wavelet
  Transform (DWT) technique. First applied SVD on cover image and adds singular value of
  diagonal component value (∑) with secret image1 to get a SVD cover image. Both secret
  image2 and SVD cover image are decomposed into 8x8 matrix. Scrambling the secret image2
  by using a key to get scrambled secret image. Extract DWT coefficients on both SVD cover
  image and scrambled secret image, then embedded by using image reverse fusion
  technique.We achieved good visual quality stego image with excellent PSNR values,
  provides high level security and more robustness.

  Keywords: DWT, SVD, Fusion Process, Statistical Parameters.


          The amazing developments in the field of network communications during the past
  years have created a great requirement for secure image transmission over the Internet. The
  Internet is a public network and is not so can recover the message. The two methods can be
  combined by encrypting the message then hiding it using steganography, so even if the
  hidden message is discovered, its meaning can remain secret . Steganography is the science
  of hiding messages in a medium called cover object in such a way that the existence of the
  message is concealed. The cover object could be an audio file, image file or a video file. The
  message to be hidden called the payload. The payload could be plain text, audio, video or an
  image. The cover object along with the hidden message is known as the stego object or
  steganogram. Steganography is in contrast to cryptography where the existence of the hidden
  message is known, but the content is intentionally obscured.

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME


        Review of literature survey has been conducted on discrete wavelet transformation
combined with singular value decomposition techniques for hiding information in digital
color images. The powerful numerical analysis SVD and DWT transformation that is widely
applied to digital image applications is first introduced. Ray-Shine Run et al., [1] has
proposed reliable SVD-based image watermarking. It was solving the ambiguities situation
and the false positive problem and he gets the best PSNR value. Ahmad A. Mohammad et al.,
[2] has proposed SVD-based watermarking algorithm for ownership protection. This
proposed algorithm was more robust and solves the false-positive detection flaw in most
SVD-based techniques.
        Bai Ying Lei et al., [3] has proposed a blind and robust audio watermarking technique
combined with SVD, Discrete Cosine Transform (DCT) and synchronization code technique
achieves very low error probability rates. Veysel Aslantas et al., [4] has presented optimal
robust image watermarking technique based on SVD using the differential evolution
algorithm effectively to improve the quality of the watermarked image. Khasim T, et al., [5]
has presented a dual transform technique for robust steganography for secret and secure
communication. Ali Al-Ataby et al., [6] proposed a modified high capacity image
steganography technique that depends on wavelet transform with acceptable levels of
imperceptibility and distortion on the cover image with high levels of overall security. Sushil
Kumar et al., [4] presented a multilayered secure, robust and high capacity image
steganography algorithm. This algorithm achieved three layers of security, better in terms of
imperceptibility, robustness and embedding capacity compared with corresponding
algorithms based on DWT. Tanmay Bhattacharya et al., [8] proposed a DWT based
steganographic technique. In the proposed method second image is embedded in the HL sub
band of the cover image.
        The paper is organized as following sections. The steganography method is described
in section3. Section4 introduced the proposed model with illustration of each phase. Testing
and Result analysis is illustrated in section5. Conclusion discussed in section6. References
are given in the last section.


3.1 Singular value decomposition
        Singular Value Decomposition (SVD) is one of a number of effective numerical
analysis tools used to analyze matrices. In SVD transformation, a matrix can be decomposed
into three matrices that are the same size as the original matrix. Although SVD works for any
n*n matrix A, and without loss of generality, our discussion will be limited for the n*n
matrix. The SVD of the n*n matrix A is
       A=    U∑ VT         (1)
Where, U and V€R n*n are unitary, and ∑€R n*n is a diagonal matrix and the superscript T
denotes matrix transposition. The SVD can be looked at from three mutually compatible
points of view. On the one hand, it is a method for transforming correlated variables into a set
of uncorrelated ones that better expose the original data items.

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME

3.2 Discrete wavelet transform

        The wavelet transform has the ability of reconstructing, so there is no information loss
and redundancy in the process of decomposition and reconstruction. Wavelets convert the
image into a series of wavelets that can be stored more efficiently than pixel blocks. In
numerical analysis and functional analysis, a DWT any wavelet transform for which the
wavelets are discretely sampled. Discrete wavelet transforms map data from the time domain
to the wavelet domain. The result is a vector of the same size. Wavelet transforms are linear
and they can be defined by matrices of dimension n X n if they are applied to inputs of size n.
DWT is applied to discrete data sets and produce discrete outputs.


4.1 Proposed steganography encoding process

        In non-blind steganography,embedding process first read one cover image and two
secret images are secret image 1 and secret image 2. Apply singular value decomposition of
the cover image and get components such as U, ∑ , VT. Multiply ∑ component with secret
image 1 to get a SVD cover image. Both SVD cover image and secret image 2 are
decomposing into 8x8 matrix. Scrambling the secret image 2 by using a key to get scrambled
secret image SS. Perform DWT on both SVD cover image and scrambled secret image to
get fused image. Take the inverse DWT on the fused to get a stego image. The decoding
process is the reverse process of encoding process and the schematic diagram of the proposed
embedding and decoding process is shown in figure 1 and 2.

                SECRET          SECRET IMAGE2
  IMAGE                                                        STEGO IMAGE                 COVER IMAGE
                MAGE 1

 Apply                          Decompose                     Decompose into             Decompose into
 SVD                +            into 8x8                       8x8 blocks                 8x8 blocks
                   U∑’ VT
                                                                DWT                -           DWT           KEY
 U∑ VT
               Decompose into                                                    IDWT               SS
                                                              Apply SVD
                 8x8 blocks

                                                                                SECRET              SECRET
                                                               U∑ VT
                    DWT         FUSION PROCESS                                 IMAGE 1             IMAGE 2

     STEGO IMAGE            IDWT

 Fig 1: Shows proposed encoding model                     Fig 2: Shows proposed decoding model.
      (scrambled secret image-SS)                         (secret image1-W, scrambled secret image-SS)

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME


        MATLAB is a numerical computing environment and fourth-generation programming
language. Created by Math Works. We implement the proposed method by using Matlab R2010a
and 7.10 version. In this paper , we are embedding and extracting secret image through a
standard encoding and decoding with SVD and DWT methods. We use an original cover image
and both secret images of size (512x512 pixels). To retain the image quality and provide a
stronger robustness and security of a two secret image hiding method, the quality metrics are
further considered. The values of quality metrics not only reduces the image perceptibility but
also enhances the robustness to resist attacks. The equations of the image quality metrics with
corresponding formulas used in our study has been illustrated above table1.

                                                 Table 1:
                             Shows the proposed quality metrics with its formulas.

        QUALITY METRICS                                                 FORMULAS

1.Mean Square Error (MSE)                                                                  ′

2.Peak Signal to Noise Ratio (PSNR)

3.Normalized Cross Correlation (NCC)                                            ′

4. Structural Content (SC)                                                                     ′

          (a)                  (b)              (c)               (d)                (e)           (f)

                              (g)                     (h)                 (i)
  Fig 3: Shows (a) Cover Image (b) Secret Image 1 (c) SVD cover image (d ) Srambled secret
  Image (e) Cover Image DWT (f) Secret Image 2 DWT (g) Stego Image (h) extracted Secret
                            Image 1 (i) recovered Secret Image 2.

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME

        In our approach, we proposed a new method, and their experimental results are shown
in the fig 3. Cover image ,secret image 1 and scramble secret image are shown in the
corresponding figures 3(a), 3(b), 3(d). Next we applied SVD transform on cover image that
correspondent result without any significant difference shown in the fig 3(a) and fig 3(c).
DWT transform on fig 3(c) and fig (d) separately and obtained the DWT coefficients and
fused the images. Take the inverse transformation on the fig 3(g) represents the stego images.
By using SVD transform and extract ∑ component is shown in fig 3(h). The fig 3(h) shows
the corresponding results of recovered secret image 1.We compare with the other existing
method, our proposed algorithm calculated the quality metrics ratio values which gets better
acceptable ratio that was shown in table 1.

                          Shows the performance of quality metric factors.

   CI           SI 1             SI 2        ST I       MSE     PSNR         NCC       SC

Home.jpg      fruit.bmp        liliyr.jpg   Home
  SVD           DWT             DWT         IDWT      0.17615   55.6724      0.9984   1.0032

Chilly.tiff   House.tiff       Car.bmp      Chilly
  SVD           DWT             DWT         IDWT      0.6220    50.1932      0.9997   1.0005

Home.jpg      Bliss.bmp       Winter.jpg    home
  SVD           DWT                         IDWT      0.3150    53.1477      1.0019   0.9962

  Water       Tree.tiff      lavender.Jpg Water
lilies.jpg                                lilies
                DWT             DWT                   0.6494    50.0362      0.9956   1.0086
  SVD                                       IDWT

      (CI=Cover image ),(SI 1=secret image 1) (SI 2= secret image 2), (ST I= stego
    image),(MSE= Mean Square Error), (PSNR= Peak Signal To Noise Ratio),(NCC=
   Normalized Cross Correlation), (AD=Average Difference) , (SC=Structural Content) ,
            (MD=Maximum Difference) ,(NAE=Normalized Absolute Error).


         In this paper, two secret image hiding method based on SVD and DWT technique
was proposed. The quality of steganography scheme should satisfy the requirements of
robustness and resist distortions due to common image manipulations such as image
suppression, rotation, cropping and other attacks. In addition to this feature one important
benefits are our technique, that shows the dual secret image steganography scheme with dual
transformation. In this two secret image hiding method achieved high PSNR ratio values
nearly 55dB and less MSE values. Future work will concentrate on perfecting the visual
effect of the stego image and the more security and robustness against the various attacks.

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-
6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME


[1]  Ahmad A. Mohammad, Ali Alhaj, Sameer Shaltaf,(2008)”An improved SVD-based
     watermarking scheme for protecting rightful ownership” Signal Processing Vol-88 pp:
     2158– 2180.
[2] Ray-Shine Run, Shi-Jinn Horng, et al.,(2012)” An improved SVD-based watermarking
     technique for copyright protection”, Expert Systems with Applications, Vol-39, pp: 673–
[3] BaiYingLei n, IngYannSoon,ZhenLi,(2011),“ Blind and robust audio watermarking
     schemes based on SVD–DCT” Signal Processing,Vol-91 ,pp:1973–1984.
[4] Veysel Aslantas et al.,(2009) ”An optimal robust digital image watermarking based on
     SVD using the differential evolution algorithm” Optics Communications, Vol-282, pp-
[5] K B Shiva Kumar and Khasim T, K B Raja , (2011)‘ Dual Transform Technique for
     Robust steganography”, International conference on Computational Intelligence and
     Communication Systems, pp. 173-178.
[6] Ali Al-Ataby and Fawzi Al-Naima,(2010) “A Modified High Capacity Image
     Steganography Technique Based on Wavelet Transform “ , International Arab Journal of
     Information Technology, Vol 7 (4) , pp . 1-7.
[7] Mutoo S.K. and Sushil Kumar, (2011),“ A Multilayered Secure, Robust and High
     Capacity Image Steganographic Algorithm ” , World of Computer Science and
     Information Technology Journal, Vol 6 , pp . 239-246 , 2011.
[8] Tanmay Battacharya, Nilanjan Dey and Bhadra Chauduri S.R. (2012),’A Session Based
     Multiple Image Hiding Technique using DWT and DCT”, International Journal of
     Computer Applications, Vol 38 (5), pp. 18-21.
[9] Nagham Hamid, Abid Yahya ,R. Badlishah Ahmad and Osamah M. Al-Qershi, “An
     Improved Robust and Secured Image Steganographic Scheme” International journal of
     Electronics and Communication Engineering &Technology (IJECET), Volume 3, Issue 3,
     2012, pp. 22 - 33, ISSN Print: 0976- 6464, ISSN Online: 0976 –6472.
[10] P. Prasanth Babu, L.Rangaiah and D.Maruthi Kumar, “Comparison and Improvement of
     Image Compression using Dct, Dwt & Huffman Encoding Techniques” International
     journal of Computer Engineering & Technology (IJCET), Volume 4, Issue 1, 2013,
     pp. 54 - 60, ISSN Print: 0976 – 6367, ISSN Online: 0976 – 6375.


      Prabakaran G holds the position of Assistant Professor of Computer Science and
      Engineering at Annamalai University. He is pursing his Image Steganography
      from Annamalai University. He authored over 25 papers on data embedding and

      Dr.Bhavani R holds the position of Professor of Computer Science and Engineering at
      Annamalai University. She received her Image Processing from Annamalai
      University. She authored over 60 papers on Image processing and steganalysis.

      Kanimozhi K, PG student, Department of Computer Science, Annamalai University.


Shared By: