一种基于三维小波的视频水印空时算法

针对数字视频的版权保护,以及基于数字视频的信息隐藏,提出了一种基于三维小波的视频水印空时算法。它利用了视频场景镜头分割技术和三维小波变换的空时多分辨水印嵌入策略,以及根据最大似然准则的水印检测算法。针对视频的版权保护,作为版权标志的水印图像先用两个序列扩频预处理成一个图像序列,扩频后的水印图像序列自适应地嵌入到视频镜头的三维小波系数上。该基于三维小波域的水印策略充分利用了视频序列良好的空时多分辨特点,水印算法可靠。针对视频水印的几种特殊攻击,仿真实验证明了该算法具有很强的鲁棒性。

Spatial-temporal approach based on 3D wavelet for video watermarking

For the copyright protection of digital video and the information hiding based on digital video, a spatial-temporal approach based on 3D wavelet for video watermarking is presented. It adopts video shot segmentation and 3D wavelet transform to watermarking embedding, as well as watermarking detection arithmetic based on maximum likelihood criteria. For the copyright protection, a sequence of watermarks as copyright information is generated from a watermark image by direct spread spectrum technique and adaptively added into the 3D wavelet coefficients of the video shot. The video watermarking algorithm based on 3D wavelet fully utilizes the good spatial-temporal multi-resolution of the video sequence, so the watermarking algorithm is reliable. The experimental results demonstrate the robustness of the proposed video watermarking method to several video degradations and watermarking attacks.