一种结合数据融合与树状小波的鲁棒数字水印算法  

A ROBUST DIGITAL WATERMARKING ALGORITHM COMBINING DATA FUSION AND TREE-STRUCTURED WAVELET

作　　者：佘维[1] 孙星明[2] 杨恒伏[2] 易叶青[2] 

机构地区：[1]湖南大学软件学院,湖南长沙410082 [2]湖南大学计算机与通信学院,湖南长沙410082

出　　处：《计算机应用与软件》2009年第1期28-31,共4页Computer Applications and Software

基　　金：国家自然科学基金(60373062;60573045);湖南省教育厅青年项目(04B016)资助

摘　　要：利用树状小波分解结合人类视觉系统(HVS)的特性,提出一种基于数据融合的鲁棒性数字水印算法,向载体图像中自适应的嵌入多个数字水印副本;在提取水印时,在每个块中使用独立分量分析ICA的方法提取水印,并对提取出的多个水印副本图像进行融合操作,以提高水印的鲁棒性。仿真试验表明了该方法的有效性,嵌入的水印具有较高的透明性,同时对常见图像处理攻击有很强的鲁棒性。By the use of tree-structured wavelet decomposition and in unification with the character of human visual system （ HVS）, a robust digital watermarking algorithm based on data fusion was proposed in this paper. Multiple copies of digital watermarking were adaptively embedded into the carrier image, and then ICA method was used in each block to extract these copies when picking up the watermarking. These extracted copies were fused by PCA （principal component analysis） image fusion algorithm so as to improve the robustness of the proposed algorithm. Simulated experimental results show that this scheme is effective, and achieves higher transparency in the embedded watermarking and has strong robustness to common image processing attacks.

关 键 词：数字水印 树状小波分解 独立分量分析 数据融合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP274.2[自动化与计算机技术—计算机科学与技术]