A New Steganalysis Method for BPCS

Abstract

Propose a New Steganalysis Method for BPeS which base on Markov model. According to BPeS steganography algorithm destroy the continuity of images, divide the least significant bit plane of images to sub-blocks. Then calculate the complexity of the various blocks to form Markov matrix, and scan the matrix to construct the Markov model, extracting features to classification test. In the experiment use support vector machine (SVM) classifier to train and c1assity.Experimental results show that the method performs well for BPeS steganography algorithm.

A Method of Universal Steganalysis Using Independence with Image Format

“…Formula (2.9) is used to count distribution of coefficient pair numbered (d1, d2) in 63 positions of each DCT block. Absolute value of this coefficient pair is set to (0,0), (0,1), (1,1), (1,2), (2,3). Of whichi and j can't be equal to 0 at same time.…”

“…Therefore, steganalysis is gaining more and more attention. Steganalysis methods are mainly divided into two categories: specific steganalysis [1][2][3][4] and universal steganalysis. Universal steganalysis algorithms highlight the prospects for future development from a practical point of view [5].…”