基于奇异值分解的鲁棒性和脆弱性兼具图像水印方法

Image Watermarking Scheme with Robustness and Fragility Based on Singular Value Decomposition

针对图像安全水印算法功能单一的问题,亟待设计一种兼具鲁棒性或脆弱性的图像水印算法。基于此论文利用奇异值分解提出一种同时具有鲁棒性和脆弱性的图像水印方法。首先通过比较相邻两个子块的最大奇异值大小关系产生鲁棒零水印;然后利用子块奇异值的小数位奇偶性产生脆弱水印;最后将脆弱水印嵌入在对应偏移子块的空域LSB。实验结果表明:算法具备很强的鲁棒性和很敏感的脆弱性,能够在实现版权保护和内容认证中有一定的借鉴意义。

At present, some watermarking algorithms merely have single function. They possess either robustness or fragility. An image watermarking scheme with robustness and fragility is proposed based on singular value decomposition. It divides the original image into blocks, and derives robust zero-watermark by comparing the value of the biggest singular val- ues in two adjacent blocks. Afterwards, the fragile watermark is derived by judging the parity of the decimals of singular val- ues in every block. Finally, the fragile watermark is inserted into its corresponding shifted block＇s LSB. The experimental results show that it has strong robustness and sensitive fragility, and is able to accomplish both copyright protection and con- tent authentication.