Audio Steganalysis based on collaboration of fractal dimensions and convolutional neural networks

Steganography is the art of concealing a message within a cover media with the least understandable changes. On the other hand, steganalysis algorithms try to distinguish information-carrying signals from clean signals. This paper proposes a new approach to audio steganalysis that uses fractal dimensions as features and convolutional neural network (CNN) as a classifier. Fractal dimensions are extracted using Higuchi's, Katz's, and Petrosian's algorithms. Hide4PGP and StegHide are the two steganography tools employed at different embedding rates. In order to evaluate the proposed audio steganalysis system, we use 10 audio samples, consisting of 4000 clean and steganographic frames. The proposed system has been compared with several audio steganalysis systems based on MFCC, Wavelet, 2D-MFCC, R-MFCC and LPC as well as classifiers LDA, SVM and KNN. According to the experiment results, the proposed audio steganalysis system shows a so better performance than other systems and brings above 99.5%.