How to (possibly) foil multimedia security?

Abstract

Multimedia security can be foiled thanks to Slepian's permutation modulation. Originally proposed in 1965 for standard problems of channel and source coding in communications, permutation codes can also provide optimum solutions in two relevant fields: steganography (foiling hidden information detection tests) and counterforensics (foiling forensic detection tests). In the first scenario, permutation codes have been shown to implement optimum perfect universal steganography (that is to say, steganography with maximum information embedding rate, undetectable and only relying on the empirical distribution of the host) for histogram-based hidden information detectors. In the second scenario, permutation codes have been shown to implement minimum-distortion perfect counterforensics (that is to say, forgeries which are undetectable and as close as possible to a target forgery) for histogram-based forensic detectors. Interestingly, both of these developments have revealed connections with compression through theoretical bounds from the mathematical theory of information. In steganography, the long-acknowledged duality between perfect steganography and lossless compression has been made explicit by permutation coding. On the other hand, a connection between counterforensics, lossy compression and histogram specification is also shown.