针对MPEG-2视频压缩标准,提出一种具备快速实现能力的鲁棒性视频水印嵌入方案。利用视频帧分块DCT系数与整体DCT系数间的映射关系,在水印嵌入和提取过程中避免对视频进行完全解码,减小计算量,提高水印算法实时嵌入、实时检测能力。实验结果表明,该方法能抵抗缩小攻击、高斯白噪声攻击、MPEG-2重编码压缩攻击,而且能够快速实现。
A robust video watermarking algorithm based on MPEG-2 video compression standard is proposed, which can be implemented efficiently. By using the mapping relationship between video frames' block DCT coefficients and the full frame DCT coefficients, the proposed scheme does not need to completely decode the host video in watermark embedding and extraction, and thus reduces the operation cost. Experiments show that the method can effectively resist down-scaling attack and Gaussian white noise contamination and MPEG-2 compression.