Video steganalysis using spatial and temporal redundancies

In this paper we present a novel video steganalysis method using neural networks and support vector machines to detect video steganograms with very limited a-prior knowledge about the steganogram embedding method. We apply temporal and spatial redundancies by averaging the frames in the sliding window to obtain an estimate of the frame and extract the merged discrete cosine features (DCT) and Markov features. MSU stegovideo tool by Moscow State University and the spread spectrum steganography tool are used for producing video steganograms. Results show that the features we use give the best accuracy to detect video steganograms. Our results thus demonstrate the potential of using learning machines and averaging temporal and spatial redundancies in detecting video steganograms.