Abstract:

 A universal steganalysis method was proposed by using the superior property of the contourlet transform with representation of an image. It merged the highorder statistics model of coefficient moments statistics, noise residual moments statistics, and characteristic function moments in the high frequency subband of the contourlet domain. At the same time, a nonlinear support vector machine(SVM) classifier was used to classify JSteg, Jphide, F5 and Outguess with different embedding rates. Experimental results showed that the proposed method has the superion discriminative performance  for  most of steganography methods. Compared with the classical wavelet, the contourlet transform has better detection effect to capture  slight differences during embedding messages.
 

Key words: steganalysis, contourlet transform, statistics characteristic