Image forgery detection using segmentation and swarm intelligent algorithm

Small or smooth cloned regions are difficult to be detected in image copy-move forgery (CMF) detection. Aiming at this problem, an effective method based on image segmentation and swarm intelligent (SI) algorithm is proposed. This method segments image into small nonoverlapping blocks. A calculation of smooth degree is given for each block. Test image is segmented into independent layers according to the smooth degree. SI algorithm is applied in finding the optimal detection parameters for each layer. These parameters are used to detect each layer by scale invariant features transform (SIFT)-based scheme, which can locate a mass of keypoints. The experimental results prove the good performance of the proposed method, which is effective to identify the CMF image with small or smooth cloned region.