基于奇异值分解的量化水印盲检算法

提出了一种基于奇异值分解的量化水印盲检算法.利用图像矩阵的奇异值稳定性好且最大奇异值远远大于其他奇异值的特点,首先时分块图像最大奇异值进行量化,再结合其奇偶性将二值水印嵌入量化后的奇异值中.实验结果表明,水印算法在满足较好的不可见性和较强的鲁棒性的同时;还可实现水印的盲提取.

Quantization Watermarking Blind Extraction Algorithm Using Singular Value Decomposition

This paper presents an efficient binary watermarking algorithm for digital image by using singular value decomposition. Because the singular values of an image matrix are steady and the maximal one is far bigger than any one of the others, the binary watermark could be embedded and blindly extracted by singular values decomposition of tiled image, where the parity characteristic is used to embed watermarking. The experimental results show that the proposed watermarking scheme achieves better transparency and robustness, and the original image is not used for watermark extraction show that blind detection of watermark is achieved.