关键词: 马尔可夫链模型, 隐写分析, 经验转移矩, 隐马尔可夫

Abstract: Aiming at steganography of the block-DCT domain，a new blind steganalysis algorithm based on Markov Chain Model（MCM） is proposed.This algorithm analizes and researches the distribution characteristic and the statistical features of the DCT coefficients，the Hiding Markov Tree（HMT） model is applied to generate the estimated cover images，and a two-order Markov empirical transition matrix model is built to capture both inter-block and intra-block dependencies between the block-DCT coefficients.And then，the diagonal elements are extracted from its empirical transition matrix as eigenvectors of the images，and the angle of the eigenvectors is constructed for detection the present or not of secrete messages in the digital images without knowledge of the original cover images.At the end，it takes an example to illustrate both the reliability and general performances of the algorithm is superior to the general detection algorithm.

Key words: Markov Chain Model（MCM）, steganalysis, empirical transition matrix, Hiding Markov Tree（HMT）