RST Invariant Zero-watermarking Scheme Based on Matching Pursuit  

作　　者：WANG Na LI Xia 

机构地区：[1]College of Information Engineering, Shenzhen University, Shenzhen 518060, China

出　　处：《Chinese Journal of Electronics》2006年第2期269-272,共4页电子学报（英文版）

基　　金：This work is supported by the National Natural Science Foundation of China （No.60372087） and the Natural Science Foundation of Guangdong Province （No.04300851）.

摘　　要：Digital watermarking techniques have been proposed as an an effective method for authentication and copyright protection in recent years. In this regard, the embedded watermark must satisfy at least two fundamental requirements： imperceptivity and robustness. However, many existing watermarking methods are vulnerable to geometric attacks and while modifying the original image in the process of embedding the watermark information, cause contradiction between imperceptivity and robustness. This paper presents a RST （Rotation, scale and translation） invariant image zero-watermarking scheme based on matching pursuit which perfectly balances between the two conflicting requirements. Robustness against geometric attacks is achieved with the use of a descriptormatching pursuit whose dictionary is built by a set of affine transformation of Gaussian functions. To guarantee imperceptivity, we construct zero-watermark through RST invariant matching pursuit coefficients which represent the significant characteristics of image, and register the zero-watermark in IPR （Information protection registration） database instead of embedding it into the original image. Experimental results show that it is robust against popular watermarking test software-Stirmark.

关 键 词：Zero-watermark Matching pursuit RST invariant Geometric attacks. 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]