Steganalysis Based on Feature in DCT Domain for JPEG Images
MA Chunb · Computer and Digital Engineering · 2015
A steganalysis algorithm based on DCT feature and support vector machine for JPEG image was proposed.In DCT domain for JPEG images,the Markov model was used to calculate intrablock correlations,and the joint probability density matrix was used to calculate interblock correlations,the calibration techniques was applied to extract 89-dimensional features,and then the support vector machine was utilized to train and classify.The experimental results for three kinds of popular JPEG steganographic schemes(F5,Outguess,MB1)had demonstrated that the detection rate of the proposed algorithm is better than the other algorithms in the case of low embedding rate.