基于分块DCT的自适应视频水印算法  

An Adaptive Video Watermarking Algorithm based on DCT Block

作　　者：蒋凯来[1] 康志伟[1] 

机构地区：[1]湖南大学信息科学与工程学院,湖南长沙410082

出　　处：《微计算机信息》2012年第4期137-139,共3页Control & Automation

基　　金：国家自然科学基金(60702065);湘潭市科技计划重点项目资助课题(ZJ20071008)

摘　　要：在以往的离散余弦变换(DCT)水印嵌入算法中,选择直流系数嵌入能增加稳定性,但也会导致水印的不可见性较差。本文在DCT直流系数嵌入法的基础上进行改进,结合人类视觉系统模型(HVS),优先选择纹理强度大的DCT分块并自适应控制嵌入强度,能较好兼顾稳定性和不可见性。改进后的算法与改进前算法相比,数字水印的安全性、鲁棒性、嵌入量、不可见性、稳定性等各方面都有较大提高。The existing algorithm based on DC coefficient embedding has better robustness but poor quality of watermarked image., An improved watermarking algorithm with the features of human visual system （HVS） taken into account during embedding is proposed in the paper. It gives preference to the blocks with close texture and adaptively adjusts watermarking intensity according to the texture features of each block, can guarantee both of the robustness and the invisibility of watermark. Compared to the conventional DCT algorithm, the improved algorithm has better security, robustness, watermark canacitv, invisibility, stability.

关 键 词：信息隐藏 视频水印 离散余弦变换 人类视觉系统 

分 类 号：TN918[电子电信—通信与信息系统]