Research Of Spread Spectrum Steganalysis Method

Spread-spectrum steganography, as one of the most promising steganography technologies, has the advantages of perceptual transparency and robustness and so on. The detection of spread-spectrum steganography has attracted more and more attentions.In this paper, we give some research of steganalysis of direct sequence spread spectrum steganography, the detailed cotents is as follows:a) A detection method of DSSS steganography based on absolute cross correlation is presented in the paper. The absolute cross correlation value between covers that befor and after DSSS hiding information and the best estimated sequence in different length is different, so that, we can calculate the kurtosis of the histogram which is comprised of the absolute cross correlation values. And then the kurtosis is as feature for dection, support vector machines(SVM) play role of classifier. The presented approach has been implemented on WAV adio and BMP images on the spread-spectrum diction systems, and experimental results is good.b) An recovering method (secret message is not encrypted) of direct sequence spread-spectrum hidden information based on genetic algorithm is proposed . In this method the embedded sequence length is obtained firstly, and then the best estimated sequence with the sequence length is got by genetic algorithm, finally the confidential message hidden in stego-cover can be recovered. The presented approach has been implemented on WAV audio and BMP images on spread spectrum recovering systems, and experimental results show that the proposed method is effective.