Audio steganalysis with content-independent distortion measures

We first propose a novel content-independent distortion measurement method and use this methodology for digital audio steganalysis. Content-independent distortion measures are utilized as features for the classifier (steganalyzer) design. Experimental results show that the removal of content dependency from features enhances their discriminatory power.