Copy-Rotation-Move Forgery Detection Using the MROGH Descriptor

Abstract

Due to its simplicity and effectiveness, copy-move forgery has become a commonly used approach to forging images and detection of this behavior has attracted much attention in recent years. Usually integrated in the copy-move forgery process is the rotation operation, which makes rotation invariance a critical property of the copy-move forgery detection methods. In this paper, we propose to use screened Harris Corner Detector and the MROGH descriptor as feature detector and descriptor respectively, to gain better feature coverage and robustness against rotation. Experimental results illustrates the efficacy of our method.