An Effective Keypoints Extraction Scheme for Image Tampering Detection

Abstract

Digital image security has become a key and challenging social issue with the advent of sophisticated image editing tools. Copy move forgery is an ordinary and malicious image manipulation technique. Numerous keypoints-based forgery detection approaches have been proposed, however some algorithms report the difficulty of matched keypoints generation competency and performance. In this approach, an effective keypoints generation is offered centred on SIFT features. A robust feature point generator is presented to extract enough keypoints. Moreover, an active extracting step is applied to estimate keypoints of small or smooth textured forged regions. Experimental results illustrate that the proposed keypoints generation scheme can generate more keypoints in terms of different scales and filter sizes.