计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (13): 91-94.

• 网络、通信、安全 • 上一篇    下一篇

基于空域的水印图像几何校正和零水印算法

廖琪男   

  1. 广西财经学院 计算机与信息管理系,南宁 530003
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-05-01 发布日期:2011-05-01

New spatial-domain based algorithm of watermarked image geometric correction and zero-watermarking

LIAO Qinan   

  1. Department of Computer & Information Management,Guangxi University of Finance & Economics,Nanning 530003,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-05-01 Published:2011-05-01

摘要: 为了有效实现对数字图像的版权保护,提出一种通用的水印图像几何校正和零水印算法。,用两个稳定而又相距最远的图像SIFT特征点的变化估计图像所经历的旋转和缩放参数,校正之后,再用一个稳定的SIFT特征点的变化校正图像的平移变换。根据零水印特点,随机从载体图像提取特征像素与水印像素按位异或运算构造零水印。理论分析和实验结果表明,图像几何校正算法能有效恢复水印同步,使水印算法能够正确检测水印;零水印算法具有较好的安全性,比基于DWT的同类零水印算法具有更强的抗剪切、几何变换、图像处理和常规信号处理攻击的能力。

关键词: 数字水印, 几何攻击, 尺度不变特征变换(SIFT), 参数估计, 零水印

Abstract: In order to achieve the copyright protection of digital image,a universal watermarked image geometric correction and zero-watermarking method is proposed.Firstly,test image rotation and scaling parameters can be estimated by the difference between the stable and the farther SIFT feature point’s coordinates of it and those of the original image.After revising,the test image translation can be corrected by the change of one stable SIFT feature point’s coordinates of it and those of the original image.Then,According to the zero watermark characteristic,the zero-watermarking is constructed by performing the XOR operation on the pixel of watermark with another pixel randomly extracted from the original image.Theoretical analysis and experimental results show that the watermarked image geometric correction algorithm can effectively recover watermark synchronization,so that the watermark can be correctly detected.The zero-watermarking method has a better security performance,and compared with the same class zero watermarking algorithm based on DWT,it has a stronger ability of resisting shear,geometric distortions,image process and signal process attacks.

Key words: digital watermarking, geometric attacks, Scale Invariant Feature Transform(SIFT), estimate affine factors, zero-watermarking