一种基于混沌变换的自适应盲水印  

A Wavelet-Based Blind Digital Watermarking Using Self-Adaptive Threshold

作　　者：王树梅[1] 赵卫东[1] 王志成[1] 

机构地区：[1]同济大学CAD研究中心

出　　处：《计算机仿真》2007年第11期129-133,共5页Computer Simulation

基　　金：国家863高技术研究发展计划(2005AA414020);上海市科学技术委员会科研计划项目(052912065)

摘　　要：提出一种基于混沌变换技术的双密钥小波域图像数字水印算法,首先利用其中一个密钥产生一个随机数子,用来决定对小波变换后的载体图像的哪一个方向做进一步多分辨分解,确定嵌入水印的位置;然后利用另一密钥对二值水印图像进行随机置乱;根据嵌入位置的自带的自身特点,计算出三个阈值s、s1和s2,将经过随机置乱处理的含版权信息的二值水印图像嵌入到将水印嵌入到水平、垂直和对角方向的子带。在水印检测过程中,提出一种不需要载体图像相关信息的盲检测水印算法。首先根据其中一个密钥找到水印所在的位置,同样计算出相应的s*、s*1和s2*,然后再用另一个密钥重构水印图像。实验证明,提出的水印方案具有较好的抗攻击性,使用双密钥具有很好的安全性。In this paper, a robust digital watermarking algorithm using wavelet - packet transform is proposed. In the first stage of the proposed method, the original image is converted to frequency domain with wavelet packet transform by using one key that decides which direction will be further transformed, so that the sub bands where watermark is embedded are determined. After that, the binary watermarking image with copyright information is scrambled by a pseudorandom sequence generated from another key. According to the sub bands＇ own characteristics, three thresholds s ,s1 and s2 are calculated ,then the watermark is embedded to the selected sub bands coefficients by using different scale thresholds depending on the level decomposition. In addition, in the watermarking image extraction process, the original image is not needed. Firstly, the positions are found by using one key; secondly, three thresh- olds s^ ＊ ,s1^＊ and s2^＊ are generated, then the watermarking information is withdrawn. Finally, the watermarking image is reconstructed by using another key. The experiments＇ results have shown that the proposed watermark is invisible to human eyes and very robust to various attacks .

关 键 词：小波变换 数字水印 自适应阈值 盲检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]