【摘要】 基于傅里叶-梅林变换特性和相位相关性,提出了一种新的抗几何失真数字水印算法.这种算法的特点是:在水印系统的检测端对水印图像进行基于傅里叶-梅林变换的图像匹配分析,并对水印图像的几何变化做出判断,然后在对水印图像进行逆变换后,再进行相似性计算.在线性相关检测水印系统中应用该算法进行对比实验,表明该算法能够有效改善水印系统抗几何失真的性能,不仅降低了嵌入匹配模板引入的图像失真,而且减少了计算的复杂度.更多还原

【Abstract】 A new watermarking detecting algorithm which resists geometric distortion is proposed,based on Fourier-Mellin transform and phase correlation features.In this algorithm,the image matching of watermarking was analyzed by using Fourier-Mellin transform on the watermarking system′s detection end,the geometric changes of the watermarked images were judged,and the similarity was computed after the watermarked images were transformed inversely.Comparative test was carried out by this algorithm in the linear correlation detection systems.Experiment results show that the new detection algorithm could obviously improve the watermarking system to resist geometric distortion.The distortion of the images caused by the embedded match template was lightened and the computing complicacy was reduced.更多还原