Spatio-Temporal Rich Model For Motion Vector Steganalysis

We propose a spatio-temporal rich model of motion vector planes as a part of a full steganalytic system against motion vector based steganography. Superior detection accuracy of the rich model over the previous methods has been lately demonstrated for digital images in both spatial and DCT domain. It has not been heretofore used for detection of motion vector steganography. We also introduced a transformation so as to extend the feature set with temporal residuals. We carried out the tests along with most recent motion vector steganalysis and steganography methods. Test results show that the proposed model delivers an outstanding performance compared to the previous methods.