基于DCT域多通道彩色图像盲检水印算法  

Multi-channel color image blind watermarking algorithm based on DCT

作　　者：刘志军[1] 

机构地区：[1]河北沧州师范学院计算机系,河北沧州061001

出　　处：《山东大学学报（工学版）》2011年第3期31-35,共5页Journal of Shandong University（Engineering Science）

基　　金：河北省教育厅自然科学基金资助项目(Z2010204)

摘　　要：提出了基于人类视觉系统(human visual system,HVS)特性计算载体图像块的水印强度因子,应用发送分集技术嵌入加密水印在载体图像离散余弦变换(discrete cosine transform,DCT)域的3种颜色信道中的嵌入算法;从3种颜色信道提取3幅水印,采用图像融合方法获得最后的水印以提高水印的抗攻击能力的提取算法。利用Matlab平台实现了本文的算法,并进行了图像的容量实验,JPEG压缩、剪切操作、噪声和滤波的攻击实验。实验表明该算法嵌入容量大、误码率低;当载体图像被攻击时具有较强的不可见性和鲁棒性。The embedding algorithm,based on the property of human visual system（HVS） calculating watermarking intensity factor of carrier image blocks,applied transmitting diversity technique embedding encryption watermarking in 3 color channels of carrier image DCT and the extraction algorithm,which extracted 3 watermarking from 3 color channels,made use of image fusion methods achieving the last watermarking to heighten the resisting attack ability of watermarking was proposed.The algorithm,image capacity experiments,JPEG compression,cutting operation,noise and filter attacking experiments were achieved in the matlab.Experimental results showed that the algorithm has larger embedding capacity,lower bit error rate and stronger invisibility and robustness when images are attacked.

关 键 词：人类视觉系统 多通道 离散余弦变换 水印算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]