中文摘要：

提出一种鲁棒自适应差分能量视频水印算法，采用差分能量原理并利用Watson视觉模型，自适应地控制视频流中交流次低频系数的水印嵌入强度因子，根据能量可调阈值与差分能量的关系，有选择地嵌入水印，从而使得水印不仅满足视觉上不可感知性，而且拥有更好的鲁棒性．实验证明，算法对常见的视频攻击（噪声、滤波、帧攻击等）具有较强的鲁棒性，算法安全性高．

英文摘要：

This paper proposed a robust adaptive differential energy video watermarking algorithm based on Watson visual model, which uses the differential energy theory to insert robust watermark into video stream. This algorithm can control the watermark＇s embedding intensity of sub-low AC coefficients in the video stream adaptively through Watson visual model. And it also can choose the region to embed watermark according to the relationship between the energy＇s adjustable threshold and differential energy. So watermark not only meets the non-visual perception, but also has better robustness. The experiments show that this algorithm has strong robustness and security against the common video attacks such as noise, filter and frame attack etc.