Binary image steganalysis based on pixel mesh Markov transition matrix

摘要

Binary image stego systems have already been well developed, which raises the requirement of a steganalytic method that detects these stego systems reliably. In this paper, a steganalytic method based on the pixel mesh Markov transition matrix (PMMTM) is presented to detect binary image steganography in the spatial domain. The proposed scheme measures the embedding distortion on the texture consistency. Further, the dependence among texture structures is organized as the Markov transition of pixel meshes. The final dimensionality-reduced feature set is formed by shrinking the obtained PMMTM according to its detection performance on the embedding simulators, which are developed to simulate practical stego systems. In the end, experimental results are reported, demonstrating that the proposed approach can effectively and reliably detect state-of-the-art binary image stego systems.