An Image Hided-Data Detection Method Combining Markov Chain and Support Vector Machines

Abstract:

An image hided-data detection method is proposed combining 2-D Markov chain model and Support Vector Machines (SVM) by the paper, in which image pixels are predicted with their neighboring pixels, and the prediction-error image is generated by subtracting the prediction value from the pixel value. Support vector machines are utilized as classifier. As embedding data rate being 0.1 bpp, experimental investigation utilizing spread spectrum (SS) and a Quantization Index Modulation (QIM) method data hiding method respectively , correction detection rates are all above 90% . For optimum LSB method ,the method achieves a detection rate from 50% to 90% above with 0.01bpp-0.3bpp various embedding data rates.

• Periodical: Applied Mechanics and Materials (Volumes 128-129)
• Pages: 520-524
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.128-129.520
• Online since: October 2011
• Authors: Hui Min Zhao, Li Zhu
• Keywords: Data Detection, Markov Chain (MC), Prediction Error, Support Vector Machine (SVM)

References

[1] Joseph J. K, Ruanaidh O, Thierry P, in: Rotation, scale and translation invariant spread spectrum digital image watermarking[J], Signal Processing, 1998, 66(5): 303– 317.

DOI: 10.1016/s0165-1684(98)00012-7

[2] Cox J, Miller L, and Mckellips L, in: Watermarking as Communications with Side Information[J], Proceedings of the IEEE, 1999, 7: 1127-1140.

[3] Moulin P, in: The role of information theory in watermarking and its application to image watermarking[J], Signal Processing, Special Issue on Information Theoretic issues in Digital Watermarking, June 2001, 81: 1121-1139.

DOI: 10.1016/s0165-1684(01)00037-8

[4] Moulin P, and Liu J, in: Analysis of multiresolution image denoising schemes using generalized-gaussian and complexity priors[J], In IEEE Trans. on Information Theory, April 2010, 45: 909–919.

DOI: 10.1109/18.761332

[5] Sullivan K, Madhow U, Chandrasekaran S, and Manjunath B. S, in: Steganalysis of spread spectrum data hiding exploiting cover memory[C], SPIE2005, 5681: 38-46.

DOI: 10.1117/12.588121

[6] Wang L. Y, Cui L. R, et al, in : Availability of Markov Repairable Systems with Changeable Operating Regimes [C], IEEE the 17th International Conference on Industrial Engineering and Engineering Management, 2010, 1: 924-926.

DOI: 10.1109/icieem.2010.5646473

[7] Batu T. G, Guha S, and Kannan S, in: Inferring Mixtures of Markov Chains[R], In ACM Symposium on Theory of Computing, 2006: 247-257.

[8] Zou D, Shi Q, Su W, Xuan G, in : Steganalysis based on Maerkov model of thresholded prediction-error image[C], IEEE International Conference on Multimedia Engineering( ICME 2006): 1365-1368.

DOI: 10.1109/icme.2006.262792

[9] Information on http: /vision. ece. ucsb. edu/~sullivak/Research_imgs.

[10] Cox J, Kilian J, Leighton T, and Shamoon T, in: Secure spread spectrum watermarking for multimedia [J], IEEE Trans. on Image Processing, Nov. 1997, 6: 1673-1687.

DOI: 10.1109/83.650120

[11] Zhao Huimin, Chi Yanli, Zhang Huile, in: Study of An Information Hiding Method Research of Joint Adaptive Quantization Index Modulation and Channel-Coding[J], Journal of Circuits and System, 2007, 12(2): 69-72.

[12] Zhao Huimin, Guo Yizhen，Ding Xiaoyan, in : An Optimum Method Research Based on Bit Plane Combination of Digital Image and Fingerprint Watermarking [J], Acta Scientiarum Naturalium Universitatis Sunyatseni, 2010, 49(6): 64-67.