Feature Based Audio Steganalysis (FAS)

Taxonomy of audio signals containing secret information or not is a security issue addressed in the context of steganalysis. A cover audio object can be converted into a stego-audio object via different steganographic methods. In this work the authors present a statistical method based audio steganalysis technique to detect the presence of hidden messages in audio signals. The conceptual idea lies in the difference of the distribution of various statistical distance measures between the cover audio signals and their denoised versions i.e. stego-audio signals. The design of audio steganalyzer relies on the choice of these audio quality measures and the construction of two-class classifier based on KNN (k nearest neighbor), SVM (support vector machine) and two layer Back Propagation Feed Forward Neural Network (BPN). Experimental results show that the proposed technique can be used to detect the small presence of hidden messages in digital audio data. Experimental results demonstrate the effectiveness and accuracy of the proposed technique.