Pattern Analysis Applied on Steganalysis for Binary Text Images

This paper studied the application of pattern analysis on steganalysis for binary text images, and proposed a novel steganalysis algorithm based on pattern analysis. The idea of image "pattern" was proposed originally through deep investigating all kinds of information hiding and steganalysis algorithms. Firstly the patterns were divided into three categories according to human visual effect for flipping center pixel of the patterns, so established the best embedding pattern storehouse for binary images. Then statistically analyzed the patternpsilas changing characteristics before and after embedding message, thus proposed a steganalysis algorithm based on pattern analysis. The experiments prove the algorithm is an effective steganalysis algorithm for detecting secret message in binary text images.

[1]  Nasir D. Memon,et al.  Steganalysis of boundary-based steganography using autoregressive model of digital boundaries , 2004, 2004 IEEE International Conference on Multimedia and Expo (ICME) (IEEE Cat. No.04TH8763).

[2]  Tung-Shou Chen,et al.  A new data hiding method in binary image , 2003, Fifth International Symposium on Multimedia Software Engineering, 2003. Proceedings..

[3]  Min Wu,et al.  Data hiding in digital binary image , 2000, 2000 IEEE International Conference on Multimedia and Expo. ICME2000. Proceedings. Latest Advances in the Fast Changing World of Multimedia (Cat. No.00TH8532).

[4]  Tieniu Tan,et al.  Steganalysis of data hiding in binary images , 2004, Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004..

[5]  Alex ChiChung Kot,et al.  Data hiding for text document image authentication by connectivity-preserving , 2005, Proceedings. (ICASSP '05). IEEE International Conference on Acoustics, Speech, and Signal Processing, 2005..

[6]  Nasir D. Memon,et al.  Data hiding in binary text documents , 2001, IS&T/SPIE Electronic Imaging.