摘要
借鉴Hartung视频水印嵌入方法,提出了一种新的基于MPEG-2的局部控制水印算法;将空域扩展谱水印DCT变换后嵌入到宿主视频的DCT系数中达到嵌入目的.为了克服一般视频水印局部算法运算量巨大的缺点,将码流中运动向量解码,基于逐块的思想,通过运动预测、权重分配,设计出一种局部权值获取控制算法.同时这种算法可以根据图像局部特征控制水印鲁棒性,改善画质.实验证明了这用方法的有效性和自适应性.
In this paper, we advanced a new watermarking method based on local gain control algorithm for MPEG-2 bit-streams according to Hartung's approach of video watermarking. We embeded the DCT of a spatial spread watermark directly into the DCT coefficients of host video. In order to resolve the problem of huge operation quantity of current local watermarking approach, a gain control algorithm, weighted on block-by-block basis, was designed according to decoded motion vector from the bit-streams, activity estimation and weight assignment. Furthermore, this algorithm, adapted to the local image characteristics, can control the robustness of embedded watermark and improve visual quality of the watermarked video. Finally, in the experiment, we proved the efficiency and adaptation of the method.
出处
《华中师范大学学报(自然科学版)》
CAS
CSCD
北大核心
2004年第2期165-169,共5页
Journal of Central China Normal University:Natural Sciences
基金
国家自然科学基金资助项目(60072041).
关键词
MPEG-2视频
扩展谱水印
压缩码流
MPEG-2 video
spread spectrum watermark
compressed bit-streams