Abstract: To solve the computationally intensive problems of large amount of calculation and difficult algorithm realization in the digital watermark scheme based on Singular Value Decomposition(SVD), this paper proposes a blind watermark scheme in Contourlet domain based on Schur factorization. It transforms the cover image by Contourlet transform and obtains block symmetric matrices using matrix splitting method, decomposes the selected symmetric blocks by adopting Schur factorization and qualifies the diagonal elements to complete the watermark embedding, and watermark extraction is the inverse process of watermark embedding. Experimental results show that the scheme has fast speed of embedding and extraction, and it is efficiently invisible and robust.

Key words: digital watermark, Contourlet transform, Schur decomposition, blind detection, robustness

摘要： 基于奇异值分解的数字水印方案计算量大、算法实现困难，为此，提出一种基于Schur分解的Contourlet域数字水印方案。将载体图像进行Contourlet变换，利用矩阵分裂理论得到分块对称矩阵，对每个对称矩阵进行Schur分解得到对角矩阵，通过量化对角元素的方法实现水印嵌入。水印提取是嵌入的逆过程。实验结果表明，该水印方案嵌入和提取的速度快，具有较好的不可见性和鲁棒性。

关键词: 数字水印, Contourlet变换, Schur分解, 盲检测, 鲁棒性

CLC Number: 

• TP309