介绍了矩阵奇异值分解的基本原理,分析了现有基于SVD的图像水印算法的缺陷。这种缺陷导致了提取出的水印结果并不是先前嵌入的水印,这样就无法应用于版权保护.
This paper has introduced the fundamental principles of matrix SVD (singular value decomposition) and analyzed the existing drawback in the SVD-based image watermarking algorithm. The flaw makes the extracted watermark not the embedded one and thus can not be applied to copyright protection.