MPEG-4自适应视频水印算法  

Adaptive Video Watermarking Algorithm Based on MPEG-4 Streams

作　　者：曲蕴慧[1] 

机构地区：[1]西安医学院公共课部计算机教研室,陕西西安710021

出　　处：《现代电子技术》2010年第10期99-101,117,共4页Modern Electronics Technique

摘　　要：为了更好地保证水印的鲁棒性和不可感知性,结合MPEG-4视频压缩原理以及人类视觉模型系统,提出一种自适应的视频水印算法。根据I帧图像离散余弦变换(DCT)后的直流系数(DC)以及中低频系数个数,设计自适应因子,对不同的图像块采用不同的嵌入强度,将水印嵌入到DCT变换后的低频系数中。实验证明,该算法能够很好地与人类视觉系统相适应,并对常见的攻击有很强的鲁棒性。A self-adaptive video watermarking algorithm is proposed based on MPEG-4 video compression principle and the human visual system model,to ensure the robustness and invisibility.The adaptive factor is designed according to the direct current coefficient and the number of low and intermediate frequency coefficients after discrete cosine transform（DCT） of I frame.The different embedded strength is used for different image blocks to embed the watermark into the low frequency coefficients.The experiments show that the algorithm is compatible with the human visual system,and is very robust to the common attack.

关 键 词：视觉系统模型 离散余弦变换 自适应因子 视频水印算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]