Motion vector modification distortion analysis-based payload allocation for video steganography

摘要

Video steganography forms a covert communication channel by data embedding in cover elements. To consider inter-frame mutual embedding impacts, this paper proposes a payload allocation strategy in video steganography based on motion vector modification distortion analysis. Firstly, the motion vector modification distortion caused by data embedding is analyzed. Then, a rate-distortion model reflecting the residue deviation propagation in successive inter-coded frames is derived. According to this model, the residue deviation propagation weight of each inter-coded frame can be computed. Finally, an inter-frame payload allocation strategy is designed in order to restrain the residue deviation propagation. Experimental results demonstrate that the proposed payload allocation strategy can enhance existing motion vector-based video steganographic methods in terms of undetectability and video coding performance. Besides, the lower computational complexity can be achieved.