基于图像复杂度的数字水印算法  

Watermarking Algorithm Based on Complexity of Image

作　　者：胡学刚[1] 王月[2] 

机构地区：[1]重庆邮电大学数理学院,重庆400065 [2]重庆邮电大学计算机学院,重庆400065

出　　处：《小型微型计算机系统》2012年第5期1149-1152,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(11071266)资助;重庆市教委科研基金项目(KJ100505)资助

摘　　要：对图像复杂度进行了分析,结合人眼视觉系统,提出一种基于图像复杂度的数字水印算法.该算法根据载体图像灰度共生矩阵的图像熵对载体图像的纹理特征区域进行划分,按照不同的纹理特征将载体图像的DCT(discrete cosine transform)块分成3类分别进行嵌入,并利用Watson视觉模型控制水印嵌入强度.仿真实验结果表明,本文算法不仅具有较大的水印容量和良好的遮蔽性,而且对噪声、滤波、有损压缩、剪切等常见攻击具有较好的鲁棒性.In this paper,complexity of the image is analyzed and a watermarking algorithm which uses human visual system and based on complexity of the image is proposed.First this algorithm separates the texture regions of the carrier image into different parts by using the image entropy of gray level co-occurence matrix,then it divides the discrete cosine transform（DCT） blocks of images into three categories and embeds them individually,and controls the watermark embedding strength by using the Watson visual model.Experimental results show that the present scheme has large watermark capacity,and performs excellent concealment properties,and has good robustness for noise,filtering,lossy compression,cutting and other common attacks.

关 键 词：图像复杂度 纹理特征 离散余弦变换(DCT) 鲁棒性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]