一种基于奇异值量化的鲁棒性水印算法

提出了一种基于小波变换和奇异值分解的数字水印算法,将水印信息分为有意义水印(比如公司标志等等)和无意义水印两部分,根据水印信息以及设定的密钥,通过量化奇异值,实现数字水印的嵌入;水印检测时只考虑有意义的水印信息部分,利用奇异值的量化步长,实现水印盲检测。实验结果表明,该方法对于原始图像上一般的干扰,能完全提取出水印信息,对于较大的攻击,也能证明存在水印。

Robust Watermarking Algorithm Based on Quantization of Sigular Value

A new watermarking algorithm based on DWT and SVD is presented,it divides the watermarking into significant one (for example sign of company) and insignificant one,based on the watermarking and the given key to quantization sigular value,sequentially to embed the watermarking.In the recover progress,only to consider the significant watermarking.Depending on the step of quantizationing sigular value to achieve the blind detection of watermarking.Extensive experimental results show that this method can absolutely recover the watermarking for commonly change on original image,it also can prove existence of the watermarking for conspicuous change.