IJCSNS International Journal of Computer Science and Network Security, VOL.6 No.2B, February 2006 147
Manuscript revised January 2006.
A Lossless Watermarking Scheme for Halftone Image
Authentication
Jeng-Shyang Pan†, Hao Luo††, and Zhe-Ming Lu††,†††
†Department of Electronic Engineering, National Kaohsiung University of Applied Sciences, Kaohsiung, Taiwan ††Harbin Institute of Technology Shenzhen Graduate School, Shenzhen, China
†††University of Freiburg, Freiburg, Germany
Summary
Nowadays, halftone images appear routinely in books, magazines, printer outputs, and fax documents. It is desirable to embed data in halftone images for copyright protection and content authentication. This paper proposes a novel watermarking scheme for halftone image authentication, exploiting image hash as a fragile watermark. After pixel blocking and ordering, a lookup table is constructed according to blocks’ frequency of occurrence. Watermark embedding is to displace the original blocks with the corresponding similar blocks in the lookup table, and in the reverse process watermark is extracted. Some extra blocks are randomly selected with a secret key for the lookup table embedding, and the original data of these blocks are also
inserted into the image. In image authentication, the
lookup
table is reconstructed first with the secret key, and then a simple table-look-up procedure is employed to extract the watermark hash, finally we only need to compare the watermark hash with the hash of recovered image: if they are equal, the original image suffers no alteration; otherwise it is changed. As a lossless technique, the original image can be perfectly recovered by performing the reverse process of the watermark embedding if the watermarked image is intact. It is necessary to keep the content of original host image unchanged in some specific applications, where content accuracy of the host image must be guaranteed, e.g. military maps, medical images, great works of art, etc. As a fragile watermarking, even one pixel toggling can be detected. Because of the small quantity of watermark, low quality distortion is introduced to the halftone image. Experiment results demonstrate the effectiveness of the scheme.
Key words:
Halftone image, Lossless watermarking, Hash sequence, Image authentication
1. Introduction
Digital halftoning is a technique to transform multi-tone images into two-tone images, e.g. 8-bit grayscale images into 1-bit binary images. The halftone images can resemble the original images when viewing from a distance by the low-pass filtering in the human visual system. Most popular halftoning methods can be divided into three categories: ordered dithering [1], error diffusion [2], and direct binary search [3]. Among these, error
diffusion achieves a preferable tradeoff between good visual quality and reasonable computational complexity. With halftone images widely used, content authentication, changes localization and copyright protection for this kind of images are receiving an increasing interest among researchers with digital watermarking techniques. Many watermarking techniques are proposed for multi-tone images, and those for halftone images are developed in recent years. In contrast, quite a small number of authentication watermarking approaches are available for halftone images.
In the literature, most methods cannot perfectly restore the original image in watermark extraction or detection. Existing watermarking usually introduces irreversible degradation to the host medium. Although the degradation is slight, it cannot satisfy the requirement of some specific situations, where content accuracy of the host image must be guaranteed, e.g., military maps, medical images, great works of art, etc. Therefore, it is quite necessary to develop a lossless watermarking method for authenticating halftone images. However, till the present time, there has been little attention paid to the lossless watermarking techniques for halftone images.
This paper proposes a watermarking scheme for halftone images combining the above two characteristics, namely, lossless and for authentication purpose. In image authentication, we only need to compare two hash sequences: the extracted watermark hash and the hash of the restored image. If the two sequences are equal, the halftone image suffers no alteration; otherwise it is changed intentionally or unintentionally.
The rest of the paper is organized as follows. Section 2 reviews the previous work on halftone image watermarking, in particular for secure authentication. Section 3 extensively describes the proposed scheme including watermark embedding and extraction, and meanwhile the authentication process. In Section 4, experimental results are presented for the demonstration of its effectiveness. Section 5 concludes the paper.
IJCSNS International Journal of Computer Science and Network Security, VOL.6 No.2B, February 2006 151
sequences are equal, as shown in Fig. 5(g) and Fig. 5(h), we can confirm the watermarked Lena suffers no alteration. Both of them are equal to the original watermark, as shown in Fig. 5(e). While if the watermarked Lena is tampered by a mark “KUAS HIT” (Fig. 5(d)), the two sequences are different, as shown in Fig. 5(i) and Fig. 5(j). Therefore, we can make a judgment by virtue of the two sequences are equal or not: if equal, the image suffers no alteration; otherwise changed.
The Baboon image is also divided into 4×4 blocks. As shown in Fig. 6, experimental results also verify effectiveness of the scheme.
4. Conclusion
This paper presents a lossless watermarking scheme for halftone image authentication. The hash sequence of the image is embedded as an adaptive fragile watermark. To judge weather the original image is changed or not, we only need to compare the extracted watermark and the hash sequence of the restored image. If they are exactly the same, the image suffers no alteration; otherwise it is changed. As long as the watermarked image is not unauthorized changed, the original image can be perfected recovered. Besides, no information needs to be saved except a secret key.
Acknowledgment
The authors would like to express their thanks to Dr. Bian Yang for his valuable advice.
References
[1] R. A. Ulichney, “Digital Halftoning,” Cambridge, MA: MIT Press, 1987.
[2] R. W. Floyd and L. Steinberg, “An adaptive algorithm for spatial gray scale,” Proc. SID 75 Digest. Society for information Display, pp. 36-37, 1975.
[3] D. Lieberman and J. Allebach, “Digital halftoning using direct binary search,” Proc. of 1996 1st IEEE Int. Conf. on High technology, pp. 114-124, September 1996
[4] M.S. Fu, and O.C. Au, “Data Hiding by Smart Pair Toggling for Halftone Images,” IEEE Int Conf Acoustics Speech and Signal Processing, vol. 4, pp. 2318-2321, 2000
[5] M.S. Fu, and O.C. Au, “Data Hiding Watermarking for Halftone Images,” IEEE Trans Image Processing, pp. 477- 484. 2002
[6] Z. Baharav, and D. Shaked, “Watermarking of Dither Halftone Images,” Hewlett-Packard Labs Tech Rep, HPL-98-32, 1998.
[7] H. Z. Hel-Or, “Watermarking and Copyright Labeling of Printed Images,” Journal of Electronic Imaging, pp. 794-803. 2001
[8] S.C. Pei, and J.M. Guo, “Hybrid Pixel-Based Data Hiding and Block-Based Watermarking for Error-Diffused Halftone Images,” IEEE Trans Circuits and Systems for Video Technology, pp. 867-884, 2003
[9] H.Y. Kim, and A. Afif, “Secure Authentication Watermarking for Binary Images,” Proc Brazilian Symp on Computer Graphics and Image Processing, pp. 199-206, 2003 [10] J. S. Pan, H. C. Huang and L. C. Jain, “Intelligent Watermarking Techniques,” World Scientific, 2004.
[11] P. S. Liao, J. S. Pan, Y. H. Chen and B. Y. Liao, “A Lossless Watermarking Technique for Halftone Image,” International Workshop on Intelligent Information Hiding and Multimedia Signal Processing, Melbourne, Australia, May 15, 2005.
Jeng-Shyang Pan received the B. S. degree in Electronic Engineering from the National Taiwan University of Science and Technology, Taiwan in 1986, the M. S. degree in Communication Engineering from the National Chiao Tung University, Taiwan in 1988, and the Ph.D. degree in Electrical Engineering from the University of Edinburgh, U.K. in 1996. Currently, he is a Professor in the Department of Electronic Engineering, National Kaohsiung University of Applied Sciences, Taiwan. Professor Pan has published more than 50 journal papers and 120 conference papers. He joints the editorial board for LNCS Transactions on Data Hiding and Multimedia Security, Springer, International Journal of Knowledge-Based Intelligent Engineering Systems, IOS Press, and International Journal of Hybrid Intelligent System, Advanced Knowledge International. He is the Co-Editors-in-Chief for International Journal of Innovative Computing, Information and Control. His current research interests include data mining, information security and image processing.
Hao Luo received the B. S. degree and the M. S. degree from Harbin Institute of Technology (HIT), Harbin, China in 2002 and 2004, respectively. He is a Ph.D candidate in the School of Electronic Engineering in HIT. His research interests are mainly in information security and mesh retrieval.
Zhe-Ming Lu received the B.S., M.S. and Ph. D. degrees in Electrical Engineering from Harbin Institute of Technology in 1995, 1997 and 2001, respectively. He was the Alexander von Humboldt Research Fellow in University of Freiburg in Germany, from Oct., 2004 to Jan. 2006. He has published more than 110 papers and four books. He has been program committee members in several international conferences. He is now the Professor and Director of the Visual Information Analysis and Processing Research Center, Harbin Institute of Technology Shenzhen Graduate School. His research interests are image processing, pattern recognition, information hiding and visual information retrieval.
