摘要
以非下采样Contourlet变换、奇异值分解以及Zernike矩知识为基础,结合图像不变质心在几何攻击前后相对位置不变的特性和Zernike矩对旋转、缩放攻击的不变性以及对噪声不敏感的特性,提出了一种利用Zernike矩对图像几何校正的NSCT-SVD几何鲁棒盲水印算法。该方案中宿主图像通过非下采样Contourlet变换分解提取出低频区域,通过量化每块奇异值矩阵的欧氏范数来嵌入水印。水印检测时,先利用Zernike矩和不变质心等几何参数对被检测图像进行几何攻击校正,恢复水印的同步信息后再提取水印。实验结果表明,该算法对噪声、滤波、压缩以及各类几何攻击具有较好的鲁棒性。
In this paper, we proposed a NSCT-SVD blind geometric robust watermarking algorithm by using the geo- metric correction to image that based on non-subsampled contourlet transform, sigular valve decomposition and Zernike moment,which combimes both the invariance of Zernike moment to attack and rotate, the fixation of invariant centorid relative place both before and after attack and insensitive characteristics to noise. In the algorithm, the host image ex- tracts low-frequency region by NSCT decomposition, the watermarking is embedded with quantizing the Euclideam norms of every Sigular of Zernike moments and invariant centorid, then the watermarking is extracted after recovering the synchronization information. The experimental results show that the algorithm has better robustness for various processing such as noise, filtering, compression, and all kinds of geometric attack.
作者
陈盈
郑洪源
丁秋林
CHEN Ying ZHENG Hong-yuan DING Qiu-lin(College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China)
出处
《计算机科学》
CSCD
北大核心
2016年第8期84-88,共5页
Computer Science
基金
江苏省产学研联合创新资金项目(SBY201320423)资助
