A spread spectrum steganalysis algorithm based on discontinuity detection

A novel audio steganalysis method is proposed.. the audio signal is denoised with wavelet transform. Then, a part of noise signal with different length is intercepted circularly and is used to calculate the cross correlation sequence with the rest of the noise signal. With the wavelet discontinuity detection technique, the feature is extracted from the cross correlation sequence for steganalysis and find out the steg audio. The detection rate is determined by the embedding strength of the secret message other than the embedding capacity. Experimental results show that the more embedding intensity of PN sequence is, the higher the detection rate will be. The detection rate of the algorithm is above 80% when the strength of the PN sequence is about 0.002, which demonstrates that the proposed algorithm has good detection performance.