一种基于Contourlet域的块均值关系的盲水印算法  

A BLIND WATERMARKING ALGORITHM OF BLOCK MEAN VALUE RELATION BASED ON CONTOURLET TRANSFORM DOMAIN

作　　者：崔壮[1] 吕俊白[1] 

机构地区：[1]华侨大学计算机学院,福建厦门361000

出　　处：《计算机应用与软件》2012年第12期318-321,共4页Computer Applications and Software

摘　　要：从Contourlet变换的特点出发,提出一种基于分块均值关系的Contourlet域水印算法。首先将水印图像利用Arnold混沌映射进行置乱,之后对载体图像进行2级的Contourlet变换,选择能量最大的低频子带作为嵌入区域,通过修改选定位置的系数数值嵌入水印,在水印提取时不需要原始图像,实现了水印的盲提取。实验表明所提的算法在抵抗JPEG、椒盐噪声等攻击时有较好的鲁棒性。In this paper, proceeding from the characteristics of Contourlet transform, we propose a watermarking algorithm ot t^ontourlet transform domain which is based on block mean value relation. First, we scramble the watermarking image using Arnold chaos mapping, then conduct level 2 Contourlet transform on the carrier image. We choose low frequency sub-band with maximum energy as the embedding region, embed the watermarking by modifying the value of location-selected coefficient. In addition, no primary image is needed in watermarking ex- traction, which achieves the blind watermarking extraction. Experiments show that this algorithm has better robustness against the attacks of JPEG, salt and pepper noise and so on.

关 键 词：CONTOURLET变换 块均值 盲水印 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]