基于人眼HVS模型的改进型小波水印算法

An Improved DWT-Watermarking Algorithm based on HVS Model

现如今为了实现多媒体版权的保护和认证等功能,视频水印技术研究在很多方面都取得了巨大的进步,而原始域的视频水印算法由于其极高复杂度和广阔的应用前景引起了很多研究者的关注。该文利用最能体现水印不可见性的HVS系统模型,首先对视频帧做四次小波变换,找到视频帧中人眼能够接受的不同嵌入强度,然后选取嵌入强度最大的前N个位置,在相应位置的二次小波系数中写入水印。实验结果表明,该文提出的视频水印算法对MPEG编解码、视频剪切等攻击都具有良好的抵抗性。

Today we have made rapid progress to develop algorithm of video watermarked in order to protect the copyright of media.Because of Potential applications and High complexity more developers are attracted to research on the algorithm of original domain.In this paper,we developed an improved algorithm based on HVS model which can best reflect the invisibility of watermarked video.First,we do four wavelet transforms and use the HVS formula to calculate the embedding strength of every point in the whole picture.Then we choose the first N biggest point and map them to the coefficients of second wavelet.The result shows that this algorithm can resist the compression of MPEG,cut and so on.