Enhanced matching method for copy-move forgery detection by means of zernike moments

Abstract

Copy-move is one of the most popular and efficient operations to create image forgery. Many passive detection techniques have been proposed to detect such a forgery in digital images. The performance of the detection algorithms depends mainly on the features used for matching image blocks or keypoints and the matching method as well. Among the existing detection algorithms, those which employ Zernike moments as features provide remarkable detection accuracy. The robustness of Zernike moments comes from the fact that they are invariant to rotation and scaling. However, Zernike moments-based algorithms can be improved further by adopting more efficient matching methods. In this paper, we propose a new matching method in order to enhance the detection accuracy. Compared to the lexicographical sorting-based matching method, the proposed method improved the detection accuracy by 40 %.