基于HVS和分形自相似的数字水印算法  

Robust image watermarking scheme based on HVS and fractal self-similarity

作　　者：邵晓根[1] 孙天凯[1] 王兴元[2] 

机构地区：[1]徐州工程学院信电工程学院,江苏徐州221008 [2]大连理工大学电子与信息工程学院,辽宁大连116024

出　　处：《计算机工程与设计》2010年第14期3336-3338,F0003,共4页Computer Engineering and Design

基　　金：国家自然科学基金项目(60573172);高等学校博士学科点专项科研基金项目(20070141014);徐州工程学院青年基金项目(XKY2008107)

摘　　要：在传统分形图像编码的基础上,借用人类视觉系统(HVS)的掩蔽特性,对分形域上的水印方法进行了研究。在空间域上对原始图像进行预处理,使原始图像的纹理区得到了突出,对纹理区图像进行分形压缩编码的同时,实现了水印的嵌入和提取操作。水印提取时运用概率统计知识,提高了水印检测的准确度和水印抵抗攻击的鲁棒性。实验结果表明,该算法对剪切、噪声、滤波、JPEG等图像处理方法具有较好的鲁棒性。A novel image digital watermarking algorithm based on the fractal self-similarity is presented. The proposed scheme considers the features of human vision system （HVS） very richly. By dealing with the original image,the texture domains are separated from the original image firstly. Then using the theory of local iterated function systems,the texture image＇s encoding and decoding are achieved,meanwhile,the embeding and extracting of the watermarks is finished. The original image is not required for detecting the watermarks and using the statistical method improves the veracity. Simulation results show the superior performance of the methods and demonstrate the robustness against such signal distortion as JPEG compression,low pass filtering,additive noise,scaling and so on.

关 键 词：分形自相似 人类视觉系统 数字水印 局部函数迭代系统 鲁棒性 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]