Video steganalysis based on intra prediction mode calibration

Abstract

Currently, some H.264 video steganographic algorithms based on intra prediction mode (IPM) have been proposed. By modulating IPMs over blocks during the intra-coding process, they achieve low computational complexity, considerable capacity and high security. Up to now, there is no existing steganalytic methods found for effective detection. However, we have observed that the existing IPM-based steganography usually modifies IPMs with non-optimal selection rules which violate the encoding principles. By exploiting these weaknesses, we propose a calibration-based video steganalytic scheme, which extracts IPM calibration (IPMC) features to detect IPM-based steganography. The proposed features are of a very low dimension and sensitive to the presences of non-optimal IPMs. Particularly, calibration parameters are directly obtained from the video stream. Therefore, we avoid estimating erroneous calibration parameters which can lead to the decline of detection performance. Experimental results demonstrate that the proposed IPMC features can be used to effectively detect the IPM modulations even at a low embedding rate, and the IPMC makes the new steganalytic scheme outperform the existing ones.