A robust detection algorithm for image Copy-Move forgery in smooth regions

Abstract

Region duplication or simply Copy-Move forgery is one of the most common forgery types in digital images where a region of the image is copied and then pasted into another location in the same image. Many forgery detection algorithms have been proposed to detect such type of forgery. However, most of the existing algorithms fail when the smooth regions are used to cover up the forgery. In this paper, we propose a robust region duplication forgery detection method based on extracting local extrema points from Difference of Gaussians (DoG) operator. DoG is used because it is a good approximation for the Laplacian of Gaussian (LoG) and much faster to calculate. To extract the descriptive features and hence improve the matching performance, Multi-support Region Order-based Gradient Histogram (MROGH) descriptor is adopted. Experimental results show the robustness of the proposed method compared with the state-of-the-art methods.