Research On Video Watermarking Algorithm For Video Streaming

With the rapid development of network and multimedia technology, various industries based on streaming media technology, such as network TV, remote teaching and VOD, successively appear and are developing gradually. Along with this trend, the security and copyright protection of the streaming media are becoming an urgent problem. It promotes the advent of digital watermarking techniques for streaming media. By embedding some watermark, which can reflects the video itself or user information, the illegal spreading of streaming media can be tracked to effectively prevent malicious behaviour such as pirate.Based on the analysis of the existed digital video watermarking techniques and an overview of several typical video watermarking schemes, this thesis presents two watermarking algorithms of practical use for streaming media. It can satisfy the requirements of the P2P streaming media network. The main work is summarized as follows:First, a DC component based adaptive video watermarking algorithm is proposed for P2P streaming media. The original video frames are selected by secret key, and then the two dimensional integer transform is performed on the luminance component of those selected frames. The DC coefficients are grouped for 1D integer transform. To tradeoff between transparency and robustness, the watermark intensity transformed coefficients are adaptively selected for watermark embedding. Watermark is permuted and coded by LDPC to improve the robustness against attacks. Experimental results show that the proposed approach is robust to common video attacking. Especially, it can effectively extract watermark under various P2P streaming media compressions.Second, for the low bit-rate video stream of H.264/AVC, a dual-watermark algorithm is proposed for video-on-demand (VOD) system. Because of the excellent performance of CDMA communication system, the watermark information is encoded by CDMA spread spectrum. The copyright information is embedded at the encoder side. To reduce the errors caused by quantization and watermarking, an efficient error compensation mechanism is introduced. At the server side, the watermark is embedded by a XOR based mapping. It fully exploits the characteristic of fewness of nonzero coefficients in P frames after quantization. The impacts on bit rate and video quality are very slight. Moreover, it can achieve fast embedding and extraction and satisfy the requirement of the real-time watermarking for streaming media. Both copyright protection and the private tracking are achieved.