基于小波零树的图像自适应水印算法  

Image-adaptive Watermarking Algorithm Based on Zero Tree Structure of Wavelet Transform

作　　者：鲍政[1] 张建伟[1] 夏德深[2] 董兵[1] 高尚彬[1] 

机构地区：[1]南京信息工程大学数学系,南京210044 [2]南京理工大学计算机系,南京210094

出　　处：《计算机工程与应用》2006年第32期72-76,共5页Computer Engineering and Applications

基　　金：香港特区政府研究资助局研究项目(CUHK/4180/01E;CUHK/1/00C)。

摘　　要：基于小波零树特性的视觉感知度模型的优化方案,给出了两种水印算法:一种算法嵌入的是高斯序列水印,通过相关检测实现盲检测;另一种算法嵌入的是二值图像水印,水印的提取是非盲提取。这两种算法在所有重要小波系数(包括最低频系数)中嵌入水印,以达到最大化水印嵌入量的目的,并结合感知度模型在水印的透明性和鲁棒性之间实现了较好的平衡,对于常见的图像处理操作,特别是对于JPEG和小波压缩均有较好的鲁棒性。Based on an optimized perceptual model of the zero tree structure,two kinds of watermarking algorithms are proposed：one embeds a watermark of Gaussian sequence,and the blind detector is based on correlation;the other embeds a watermark of binary image,and the extractor is not blind.In order to maximum embedding watermark bits,the two proposed algorithms embed watermark bits in all of the significant wavelet coefficients,including the coefficients of lowest pass sub-band,and provide a good tradeoff between the transparency and the robustness of watermarks based on the perceptual model,Experiments show that both of the proposed algorithms are effective and robust to common image processing operations,especially to JPEG and wavelet compression.

关 键 词：数字水印 小渡零树 感知度模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]