摘要
现有的大多数基于奇异值分解的图像水印算法,对于宿主图像施加轻微扰动时其奇异值不会发生明显的改变,然而当宿主图像发生几何变换,例如旋转和缩放之后其奇异值也发生了相应的改变,所以现有的算法缺乏抵抗几何攻击的能力。最主要的原因是目前大部分奇异值分解的水印算法都是基于笛卡尔坐标系下的,而笛卡尔坐标系下宿主图像缺乏对旋转和缩放的不变性,这样就会使得水印的嵌入和检测失去同步。为了解决笛卡尔坐标系下水印不同步问题,文章提出了一种对数极坐标系下的使用SVD抗几何信号变换的图像水印算法。实验结果表明,该算法能有效抵抗旋转、缩放等几何变换处理,对于一些压缩、滤波等常规信号处理,也显示出较好的鲁棒性。
The most existing image watermarking algorithms based on svd are robust for host image changed little. However, the singular value would be changed when host image is rotated and scaled ,so the existing watermarking algorithms based on svd is insufficient for reisisting to geometric transform abilities. The most important reason is the existing watermarking algorithms are based on Cartesian coordinate, which is insufficient to rotation and scaling transform. Therefore, watermark embedding and extracting would not be synchronized. In order to solve the watermark non-synchronization on Cartesian coordinate, this paper proposes a robust watermarking algorithm resisting to geometric transform based on SVD in logarithm polar coordinate. The simulation experiment results demonstrate the proposed method is robust enough to rotation and scaling geometric degradation. Meanwhile, it is also robust for commonplace signal processing such as：adding noise, filtering and JPEG compression etc.
出处
《信息网络安全》
2012年第11期71-73,共3页
Netinfo Security
