A forgery detection algorithm for exemplar-based inpainting images using multi-region relation

摘要

The identification of image authenticity has received much attention because of the increasing power of image editing methods. This paper proposes a novel forgery detection algorithm to recognize tampered inpainting images, which is one of the effective approaches for image manipulation. The proposed algorithm contains two major processes: suspicious region detection and forged region identification. Suspicious region detection searches the similarity blocks in an image to find the suspicious regions and uses a similarity vector field to remove the false positives caused by uniform area. Forged region identification applies a new method, multi-region relation (MRR), to identify the forged regions from the suspicious regions. The proposed approach can effectively recognize if an image is a forged one and identify the forged regions, even for the images containing the uniform background. Moreover, we propose a two-stage searching algorithm based on weight transformation to speed up the computation speed. The experimental results show that the proposed approach has good performance with fast speed under different kinds of inpainting images.