摘要
为了寻求有效的数字媒体保护方法,该文提出了一种在三维小波变换域内基于二值运算的数字视频水印嵌入算法。为了实现水印的获取,提出了一种相应的水印盲提取算法。水印的嵌入与提取算法每次只对4帧图像进行小波分解,不仅充分利用了小波变换良好的空时多分辨特性,而且占用内存空间少、时延低,适合视频流的在线水印嵌入和提取,并易于硬件的实时实现。实验结果表明,该文所提出的视频水印算法对高斯噪声、MPEG编码、帧平均、帧抽取和帧置乱具有良好的鲁棒性。
To develop an effective digital media protection method, an algorithm for digital video watermarking is proposed based on binary operation in 3D wavelet transform domain. For realizing the extraction of watermark image, a blind detection and restoration algorithm is presented, which does not require the original video. The proposed watermarking embedding and extracting algorithm uses only four image frames for performing wavelet decomposition per operation with lower memory demands and short delay, which suits for watermark online embedding and extracting and real-time implementation with hardware. The experimental results illustrate that the proposed watermarking algorithm is robust to video degradation and distortion attacks, such as Gaussian noise MPEG coding, frame averaging, frame dropping and frame swapping.
出处
《电子与信息学报》
EI
CSCD
北大核心
2007年第2期447-450,共4页
Journal of Electronics & Information Technology
基金
教育部留学归国人员实验室
新世纪优秀人才支持计划(NCET-04-0948)
教育部重点项目(104173)资助课题
关键词
数字视频水印
三维小波变换
实时嵌入
盲检测
Digital video watermarking
3D wavelet transform
Real-time embedding
Blind detection
