基于奇异值分解的Contourlet域图像水印算法

针对小波变换由于方向性缺乏而无法有效稀疏地表示图像本身几何结构的问题,结合Contourlet变换和奇异值分解的优点,提出了一种Contourlet变换域内基于奇异值分解的数字水印算法。宿主图像经过两层Contourlet分解后,生成一系列多尺度多方向的子带,选取第二层分解中能量最大的子带进行大小为4×4的互不重叠的分块,并进行奇异值分解,再将置乱后的原始二值水印信息嵌入到奇异值矩阵中。实验结果表明,该算法实现简单,透明性良好,对常规攻击和几何攻击具有较强的鲁棒性,且优于一般的小波域算法。

Contourlet Domain Image Watermarking Algorithm Based on Singular Value Decomposition

Wavelet transform can not represent the geometrical structure of the image effectively and sparsely because of the lack of direction.A digital watermarking algorithm based on singular value decomposition in Contourlet domain is proposed to combine the advantages of Contourlet transform and SVD.After two layers of Contourlet transformation,the host image is decomposed into series of multi-scale and multi direction subbands.The subband of the second layer with the maximum energy is selected to be subdivided into non-overlapping blocks of size 4×4 and then the blocks are decomposed with singular value decomposition (SVD).The watermark information is embedded into the singular value matrix.Experimental results show that the algorithm is simple to implement with good transparency,strong robustness against the common attacks and geometric attacks and is superior to the general algorithms in Wavelet domain.