一种抗几何攻击的数字水印算法

A blind watermarking algorithm against geometric attacks

根据人类视觉系统的频率掩蔽特性,提出了一种抗几何攻击的水印自适应嵌入方案。首先对宿主图像进行归一化并确定感兴趣区,然后对感兴趣区实施Contourlet变换,选取Contourlet域的低频子带进行分块。由于每个块的特性不尽相同,所以对所有的块进行分类,不同类的块选择不同的量化步长,置乱后的水印信息自适应地嵌入到每个分块奇异值分解后的最大的奇异值上。实验结果表明,该方法含水印图像不可见性好,且该方法对几何攻击及常规图像攻击都具有较强的鲁棒性。

According to human visual system,an image watermarking algorithm which has a good invisibility is proposed in this paper. Firstly, the region of interest（ROI） is obtained from the normalized image by utilizing the invariant centroid theory,the Contourlet transform is performed on the ROI. Then,the low-pass sub-band coefficients are selected to be divided into non-overlapping blocks.Singular value decomposition is performed on the blocks. According to the different characteristics of each block,the largest singular value of the selected low-pass sub-band coefficients block is adaptively operated by different quantization step for embedding watermark information. The experimental results show that the algorithm is effective on improving the invisibility of watermarked image and has strong robustness against the common image attacks.