Robust data hiding in MPEG-2 video against transcoding

Embedding watermark in compressed video is a challenging problem, and moreover transcoding the watermarked video due to bandwidth requirement makes the watermark more difficult to survive. As we know, transcoding sometimes involves numerical process as well as geometrical process, such as aspect ratio changing. In this paper, synthetical techniques including redundant embedding and scene-based embedding are presented to increase the watermark robustness against video transcoding, especially heterogeneous transcoding. Experimental results show that the proposed strategies are perfect in solving the attack.