基于HVS的自适应鲁棒视频水印算法

视频水印与静止图像比较，包含有运动部分，具有变化的特性。根据人眼对快速运动的物体敏感度会下降的特点及DCT域视觉掩蔽模型，提出一种自适应的鲁棒视频水印算法。算法首先通过比较I帧与相邻帧的同一区域的DCT直流系数，得到一个快速运动图像块集，然后在该图像块集中根据视觉掩蔽模型计算图像块的DCT中频系数，选择区域自适应嵌入水印。仿真实验表明，该水印算法不仅能很好地抵抗MPEG压缩，而且能较好地协调不可见性和鲁棒性，并能抵抗一定的几何攻击。

Robust Adaptive Video Watermarking Algorithm Based on HVS

A robust adaptive Video Watermarking algorithm based on Human Visual System(HVS)was proposed.By comparing the direct current DCT coefficients of the I frame and its neighbor frame of the video,a fast moving image block set was chosen.The new algorithm embedded watermarking bits adaptively in proper place for some image block,which was decided based on HVS.Experiments show that the proposed algorithm is robust against MPEG compression,also the attacks of RST.