基于模糊聚类分析的自适应图像空间域水印技术  

Adaptive Spatial Domain Watermarking Algorithm Based on Fuzzy Cluster Analysis

作　　者：王向阳[1,2] 杨红颖[1,2] 赵红[1,2] 陈利科[1,2] 

机构地区：[1]辽宁师范大学计算机与信息技术学院,辽宁大连116029 [2]中国科学院软件研究所信息安全国家重点实验室,北京100039

出　　处：《小型微型计算机系统》2005年第7期1255-1259,共5页Journal of Chinese Computer Systems

基　　金：信息安全国家重点实验室开放基金(03-02)资助;辽宁省自然科学基金(20032100)资助.

摘　　要：结合人眼的视觉感知特性与图像的局部相关特性,提出了一种基于模糊聚类分析的自适应图像空间域水印嵌入算法.该算法首先根据人眼视觉系统(HVS)的感知特性,结合模糊c-均值(FCM)聚类分析,自适应确定出数字水印嵌入位置;然后利用图像自身的局部相关性,通过修改像素值,将二值水印图像降维置乱后自适应嵌入到宿主图像内,并且具有水印嵌入深度智能调节特性.该算法在提取水印信息时,不需要原始载体图像.实验结果表明:该自适应空间域图像水印嵌入技术不仅具有较好的透明性,而且对诸如叠加噪声、JPEG压缩、平滑滤波、几何剪切、图像增强、马赛克效果等攻击均具有较好的鲁棒性.Combining the human visual system with the image local properties, an adaptive spatial domain image watermarking scheme based on fuzzy cluster analysis is proposed. First, the watermark embedding locations are identified adaptively by applying fuzzy c-mean clustering algorithm based on the human visual system. Then, the binary watermark sequence is embedded adaptively by modifying the pixel value according to the image local correlation, and the watermark embedding depth can be adjusted intelligently. The presented watermarking scheme can extract the watermark without the help from the original image. Experimental results show that the proposed adaptive spatial domain image watermarking scheme is invisible and robust against various signal processing such as cropping, compression, filtering, noise adding, image enhancement, and mosaic.

关 键 词：数字水印 人眼视觉系统 模糊聚类分析 空间域 图像局部相关性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]