基于DTCWT纹理分析的视频水印算法

A Video Watermarking Algorithm Based on DTCWT Texture Analysis

该文根据MPEG-4引入的视频对象概念以及对视频对象纹理的编码原理,利用双树复小波变换实现视频图像的纹理分析,将水印嵌入与视频编码结合起来,提出了基于DTCWT和MPEG-4纹理对象的视频水印算法。仿真实验结果表明,该算法使水印和视频对象一起具有灵活多样的存取和随机操作性,对于加性Gaussian噪声、高压缩比的MPEG压缩编码、帧重组和丢帧等攻击具有良好的鲁棒性。

In this paper, we propose a video watermarking algorithm based on texture object in MPEG - 4 and the dual- tree complex wavelet transform（DTCWT）. The watermark is embedded in the texture areas extracted by .DTCWT, which is used to classify video texture. Simulation results show that the proposed algorithm realizes the watermark has the same properties of stochastic operational with video object, and is robust to the additive Gaussian noise, MPEG coding at high compression ratio, frame swap and frame drop.