中文摘要：

本文提出了一种基于分块离散余弦变换（DCT）、奇异值分解（SVD）和粒子群优化算法（PSO）的鲁棒视频水印方案.在其嵌入方案中,利用DC系数和奇异值的稳定性,首先将帧内所有块的DC系数置乱,并进行4×4奇异值分解,调制第3大奇异值实现水印的嵌入,由粒子群优化算法得到合适的水印嵌入强度.实验结果表明,相对于固定的嵌入强度,采用粒子群优化算法可以在保持原有视频序列视觉质量的基础上,提高水印自身的鲁棒性.同时算法能够有效抵抗H.264压缩.

英文摘要：

A robust video watermarking algorithm based on block Discrete Cosine Transform （DCT）, Singular Value Decomposition （SVD） and Particle Swarm Optimization （PSO） is proposed in this paper. The stability of DC coefficient and singular values is utilized in the embed process. First, all the DC co- efficients in luminance blocks are scrambled, and then SVD with block size 4 × 4 of the scrambled DC co- efficients is processed. The watermarking is embedded by modulating the third singular value. The PSO is used for selecting the appropriate embedding strength. Experimental results indicate that proposed algorithm makes a good tradeoff between invisibility and robustness compared with the algorithm with fixed embedding strength. In addition, this algorithm is robust against H. 264 compression