International Journal of Power Control Signal and Computation(IJPCSC) Vol. 4. No.2. pp.

102 -108 April - June 2012 ISSN: 0976-268X www.ijcns.com

RobustDiscrete Wavelet Transform based Steganography

T. Narasimmalou, R. Allen Joseph Department of Electronics and Communication Dr. Pauls Engineering College, Villupuram, India E-mail:nr_lou@yahoo.co.in, allenjoseph1985@gmail.com

Abstract:Steganography is the art of hiding information andthe detection of steganographically encoded package is called as steganalysis. In this paper we propose a robustdiscrete wavelet transform(DWT) based steganography for hiding images.First a three level DWT decomposition is done on a host image and the secret information is hidden by manipulating the approximation coefficients of the decomposed image. After embedding, the stego image is subjected to various types of image processing attacks such asGaussian white noise, salt and pepper noise, blurring and sharpening. Experiments show that thepeak signal noise ratio (PSNR) generated by the proposed method is better. Keywords:DWT, Peak signal noise ratio, Steganalysis, Steganography. 1. INTRODUCTION Steganography is a technique used to embedmessages into a medium, such as images and videos, to communicate secretly without arousing any suspicion.The new file carrying the embedded-objectis called the stego-object. The embedded-object is the message that will be hidden using a stego-key or algorithm. A cover-object is the data file that will hold the secret message. Steganalysis is the art and science of discrimination of stego-objects and coverobjects. Without any knowledge of secret key used for embedding steganalysis needs to be carried out.The simplest method to detect

modified files, however, is to compare them to the originals. To detect information being moved through the graphics on a website, for example, an analyst can maintain known-clean copies of these materials and compare them against the current contents of the site. The differences, assuming, that the carrier is error free, will compose the payload. The goal of modern steganography is to keep its mere presence undetectable [36]Several forms of digital media may be used as cover for hidden information in computer based steganography. Photos, documents, web pages, and even music files may all serve as innocuous-looking hosts for secret messages. In covert communications through the Internet, digital images are possibly the most practical type of steganographic medium primarily due to their sheer abundance in the Web. However, one common problem with using digital images is the use of insufficient hiding capacities.[37] In this paper a robust steganographic tech-nique using Discrete Wavelet Transform for transmitting pictures is proposed.We illustrate our technique by considering Vladimir Bnoci et al [1] for three level wavelet transform and then subject the imageto various image processing attacks likeGaussian white noise, salt and peppernoise, blurring and sharpening. The PSNR generated by the proposed method are

Robust Discrete Wavelet Transform based Steganography

103 A. Procedure for Embedding using Three Level Wavelet Decomposition 1.Perform three level 2D-Haar DWT decomposition as follows: a) Take the cover image (JPEG) (512 x512) and its green plane alone and perform first level 2D-DWT on the image to obtain approximation 1coefficient (LL1), horizontal 1coefficient (HL1), vertical 1coefficient (LH1), diagonal 1 coefficient (HH1) respectively. b) Take the approximation 1 coefficient (LL1) and perform second level 2D-DWT on the image to obtain approximation 2 coefficient (LL2), horizontal 2 coefficient (HL2), vertical 2 coefficient (LH2), diagonal 2 coefficient (HH2) respectively. c) Take the approximation 2 coefficient (LL2) and perform third level 2D-DWT on the image to obtain approximation 3 coefficient (LL3), horizontal 3 coefficient (HL3), vertical 3 coefficient (LH3), diagonal 3 coefficient (HH3) respectively. 2. Take the secret image and turn it into black and white. 3. Perform Embedding as follows: a) Assume an embedding coefficient of value of 0.05. b) Process LL3 block by block (4x4). c) Process the secret image block by block (4X4). d) The following formula is used to obtain the secret image block (4x4) which is basically swapping, Secret image block = ((1- Embedding Coefficient) x LL3 intensity value) + (Embedding coefficient x secret image intensity value) 4. Perform three level 2D-Haar inverse DWT for reconstruction which is the inverse process of three level 2D-Haar DWT decomposition in order to obtain the stego image

better than those generated by the reported schemes. 2. DISCRETE WAVELET TRANSFORM The extension of one-dimensional analysis is multi-resolution analysis based on two-dimensional wavelet transform. If (x) and (x) stand for one dimension scale function and wavelet function respectively, the following one two-dimensional scale function and three two dimensional wavelet functions comprise the foundation of twodimensional wavelet trans-form. (x,y)=(x)(y) ( , )= ( , )= ( , )= ( ) ( ) ( ) ( ) ( ) ( ) (1)

(2)

Following L-level decomposition of the image f(x,y), we obtain approximation and three detail transform coefficients AL(x,y) = (x, y), (x, y) D (x,y) = (x, y), (x, y) D (x,y) = (x, y), (x, y) D (x,y) = (x, y), (x, y) (3) (4) (5) (6)

The cover image decomposition represented byapproximation and detail coefficients is shown on figure 1[1].

Figure 1. Image decomposition by DWT 3. IMPLEMENTATION OF THE PROPOSED METHOD

Published: goplex Publishing

Robust Discrete Wavelet Transform based Steganography

104 5. Assume an embedding co efficient of 0.05 6. Use the formula which follows to get the image blocks of the secret image. Secret image block = (LL3 intensity value of the stego image-((1-embedding coefficient) x LL3 intensity value of the cover image))/Embedding coefficient. 4. EXPERIMENTAL RESULTS 1 A. Embedding and extraction results using 2D- Haar DWT Host image before embedding isshown in figure 2 and after three level decomposition using 2D Haar DWT is shown in figure 3.

5. Calculate the PSNR value in order to check for the visual quality of the stego image. PSNR = 10log 2552 MSE (7)

where MSE (Mean Square Error) stands for the mean-squared difference between the cover-image and the stego-image. The mathematical definition for MSE is: = 1 ( ) ___ (8)

In Equation 8, aijmeans the pixel value at position (i, j) in the cover-image and bij means the pixel value at the same position in the corresponding stego-image. The calculated PSNR usually adopts dB value for quality judgment. The larger PSNR is, the higher the image quality is (which means there is only little difference between the cover-image and the stego-image). On the contrary, a small dB value of PSNR means there is great distortion between the cover-image and the stego-image. The above mentioned procedure can be used for Daubechies and Symlet wavelets also. B. Procedure for Extracting using Three Level Wavelet Decomposition 1. Perform three level 2D-Haar DWT decomposition on the stego image as well the cover image as we did in the embedding procedure. 2. Calculate the PSNR for the decomposed stego image as well as the cover image for finding out in which sub-band the secret image has been embedded. On examining we will find it out to be LL3. 3. Process LL3 of the stego image block by block (4x4). 4. Process LL3 of the cover image block by block (4x4).
Published: goplex Publishing

Figure 2. Host image before embedding

Figure3. Host Image after three level decomposition using 2D Haar DWT

Robust Discrete Wavelet Transform based Steganography

105 PSNR value is found by using host image and stego image, for finding which subband has been used for embedding.The extracted secret image is shown in figure 7.The extracted secret image is intact and it has very high visual quality.

Figure4. Stego Image with Secret Image embedded on Original Image Stego image with secret image embedded on the original Image is shown in figure 4.PSNR value after embedding for the proposed techniqueis 32.9220 db which is for the cover image mountain and secret image mail. Figure 7: Extracted secret image The table 1 below lists the image quality (PSNR) in dB for some of the cover images after embedding.These results were obtained using MATLAB 7. Table 1: The Image Quality (PSNR) In Db Of The Cover Images After Embedding. Secret Cover Image(512 x 512) Image (128 x Paddy Mountain Flower 128) Mail 31.3969 32.9220 32.7621 Checkbox 31.3964 32.9217 32.7620 Telephone 31.3919 31.3964 32.7582 For testing the image robustness, the image was subjected to various types of attacks like Gaussian white noise, salt and pepper noise, blurring and sharpening. Figure 8 shows the stego image after it has been subjected image processing attacks.

Figure 5. Stego Image The stego image is shown in figure 5 and the stego image after three level decomposition using 2D Haar DWT is shown in figure 6.

Figure 6. Stego Image after three level decomposition of 2D Haar DWT

Published: goplex Publishing

Robust Discrete Wavelet Transform based Steganography

106 be aware of the existence of the secret-image. The secret image extracted is similar to the original secret image. . Experimental results show that the proposed technique after it has been subjected to image processing attacks has very high visual quality. REFERENCES 1. Banoci, V.; Bugar, G.; Levicky, D.A novel method of image steganography in DWT domain, Radioelektronika, 2011 21st International Conference 2. Kumar, V., Kumar, D,Performance evaluation of DWT based image steganography, Advance Computing Conference (IACC), 2010 IEEE 2nd International 3. Po-Yueh Chen, Hung-Ju Lin, A DWT Based Approach for Image Steganography, International Journal of Applied Science and Engineering 2006. 4, 3: 275-290 4. Wang, H, Wang, S, Cyber warfare: steganography vs. Steganalysis, Communications of the ACM, 47:10, October 2004 5. Dunbar, B., Steganographic techniques and their use in an Open-Systems environment, SANS Institute, January 2002 6. Artz, D., Digital Steganography: Hiding Data within Data, IEEE Internet Computing Journal, June 2001 7. Bret Dunbardetailed look atSteganographicTechniques and their usein an Open-Systems SANS Institute 2002, as part of the Information Security Reading Room. 8. Goel, R., Garuba, M. , Liu, C. , Nguyen, T. . The Security Threat Posed by Steganographic Content on the Internet. Information Technology, 2007. ITNG '07. Fourth International Conference on. 794798. DOI: 10.1109/ITNG.2007.192

Figure 8. Stego image after it has been subjected to image processing attacks. The PSNR of the proposed technique after attacks on the stego image proves to be better which is listed in table 2. Table 2: The Image Quality (Psnr) In Db Of The Stego Images After Image Processing Attacks Image Processing Attacks Images after attack Stegoimage Paddy Stegoimage Mounta in Stegoimage Flower Gaussi an white noise 20.129 1 20.350 6 21.039 9 Salt & peppe r noise 22.20 75 21.94 52 21.21 18 Blur Sharpeni ng

25.53 91 21.33 35 20.32 20

27.4977

20.0069

21.3810

5. CONCLUSION In this paper, a new image data hiding technique based on three level DWT is proposed. The stego-image has high PSNR value even after subjecting the image to various attacks. Hence, the observer will not
Published: goplex Publishing

Robust Discrete Wavelet Transform based Steganography

107 19. van Schyndel, R. G., A. Tirkel, C. F. Osborne, "A Digital Watermark," in Proceedings of the IEEE International Conference on Image Processing,vol. 2, 1994, pp. 8690. 20. Johnson, N. F., S. Jajodia, "Exploring Steganography: Seeing the Unseen,"IEEE Computer, vol. 31, no. 2, 1998, pp. 2634. 21. Gerzon, M. A., P. G. Graven, "A HighRate Buried-Data Channel for Audio CD,"Journal of the AudioEngineering Society,vol. 43, no. 1/2, 1995, pp. 322. 22. "StegoDosBlack Wolf's Picture Encoder v0.90B," <ftp://ftp.csua.berkeley.edu/pub/cypherpu nks/steganography/stegodos.zip>, 1993. 23. Brown,A.,"S-Toolsfor indows,"<ftp://idea.sec.dsi.unimi 24. .unimi.it/pub/security/crypt/code/stools4.zip>, 1996. 25. Hastur,H.,"Mandelsteg,"<ftp://idea.sec.dsi. unimi.it/pub/security/crypt/code/steg.tar.Z >, 1994. 26. Machado, R., "EzStego, Stego Online, Stego," <stego.com>, 1997. 27. Arachelian, R., "White Noise Storm," <ftp://ftp.csua.berkeley.edu/pub/cypherpu nks/steganography/wns210.zip>, 1994. 28. Hansmann, F., "Steganos, Deus Ex Machina Communi -cations," < steganography.com/>, 1996. 29. Stefan Katzenbeisser and Fabien A. P. Petitcolas. Information hiding techniques for steganography and digital watermarking. Artech House Books, 1999. 30. Kenneth R. Castleman. Digital Image Processing. Prentice Hall, 1996. 31. Bo Yang and Beixing Deng, Oriented data hiding scheme for restoring the original host images," The J. of Sys. and Software 79, 2006, pp.1754-1766. 32. Mller, S., A. Pfitzmann, and I. Stirand, "Computer Based Steganography: How It Works and Why Therefore Any Restrictions on Cryptography Are Nonsense, At Best," in Information

9. Chunhua Chen, Yun Q. Shi, Wen Chen, Guorong Xuan, Statistical moments based universal steganalysis using JPEG 2-D array and 2-D characteristic function 2006 IEEE, ICIP 10. Petitcolas, Fabien A.P., Information Hiding: Techniques for steganography and Digital Watermarking., 2000. 11. Sellars, D., An Introduction to Steganography, URL:www.cs.uct.ac.za/courses/CS400W/ NIS/papers99/dsellars/stego.html 12. Bender, W., Techniques for Data Hiding, IBM Systems Journal, Vol. 35, Nos 3+4, Pgs 313-336, 1996 13. Nan-I Wu, Min-Shiang Hwang Data Hiding: Current Status and Key Issues, International Journal of Network Security, Vol.4, No.1, PP.19, Jan. 2007 14. Foley, J., et al.,Computer Graphics, Principles and Practice,Reading, MA: Addison Wesley, 1990. 15. Bender, W., D. Gruhl, N. Morimoto, "Techniques for data hiding," IBM Systems Journal,vol. 35, no. 3/4, 1996, pp. 131336. 16. Mller, S., A. Pfitzmann, I. Stirand, "Computer Based Steganography: How It Works and Why Therefore Any Restrictions on Cryptography Are Nonsense, At Best," inInformation Hiding: First International Workshop, Proceedings,vol. 1174 of Lecture Notes in Computer Science,Springer, 1996, pp. 7 21. 17. Gruhl, D., A. Lu, W. Bender, "Echo Hiding," inInformation Hiding: First International Workshop, Proceedings, vol. 1174 of Lecture Notes in Computer Science,Springer, 1996, pp. 295316. 18. Kurak, C., J. McHughes, "A Cautionary Note On Image Downgrading," inIEEE Computer SecurityApplications Conference 1992, Proceedings,IEEE Press, 1992, pp. 153159.

Published: goplex Publishing

Robust Discrete Wavelet Transform based Steganography

108 R. Allen Josephwas born in Pondicherry, India in 1985. He received his Bachelor of Engineering degree in Electronics and Instrumentation Engineering, from Anna University, Chennai and Master of Technology degree in Information Security from Pondicherry University, Pondicherry, India. Currently he is working as lecturer in Dr. Pauls Engineering College, Villupuram and affiliated to Anna University Chennai, Tamil Nadu, India. His area of interest includes Information Security and Computer Networks.

Hiding: First International Workshop, Proceedings, vol. 1174 of Lecture Notes in Computer Science, Springer, 1996, pp. 7 21. 33. Cachin, An Information-Theoretic Model for steganography, Proceedings of 2nd Workshop on Information Hiding, MIT Laboratory for Computer Science, May 1998 34. Anderson, R.J. & Petitcolas, F.A.P., On the limits of steganography, IEEE Journal of selected Areas in Communications, May 1998 35. Neils Provos and Peter Honeyman, Hide and Seek: An Introduction to Steganography, IEEE Security and Privacy, 2003 36. Mark David Gan, Chameleon Image Steganography Technical Paper, April 2003 37. Zhi-Min He, Wing W.Y. Ng, Patrick P.K.Chan, Daniel S. Yeung. Feature selection for blind steganalysis using localized generalization error model, Proceedings of the Ninth International Conference on Machine Learning and Cybernetics, Qingdao, 11-14 July 2010. T. Narasimmalou was born in Pondicherry, India in 1976. He received his Bachelor of Technology in Electronics and Communications Engineering, and Master of Technology degree in Electronics and Communications Engineering from Pondicherry University, Pondicherry, India. Currently he is working as Professor in Dr. Pauls Engineering College, Villupuram and affiliated to Anna University Chennai, Tamil Nadu, India. He is a life member of Indian Society for Technical Education, His area of interest includes Wireless Communications andNetwork Security.

Published: goplex Publishing