Improving the detection and localization of duplicated regions in copy-move image forgery

Using keypoint-based features, such as SIFT features, for detecting copy-move image forgeries has yielded promising results. In this paper, our emphasis is on improving the detection and localization of duplicated regions using more powerful keypoint-based features. In this context, we have adopted a more powerful set of keypoint-based features, called MIFT, which share the properties of SIFT features but also are invariant to mirror reflection transformations. To improve localization, we propose estimating the parameters of the affine transformation between copied and pasted regions more accurately using an iterative scheme which finds additional keypoint matches incrementally. To reduce the number of false positives and negatives, we propose using “dense” MIFT features, instead of standard pixel correlation, along with hystereresis thresholding and morphological operations. The proposed approach has been evaluated and compared with competitive approaches through a comprehensive set of experiments using a large dataset of real images. Our results indicate that our method can detect duplicated regions in copy-move image forgery with higher accuracy, especially when the size of the duplicated region is small.