A novel segmentation based copy-move forgery detection in digital images

Copy-move forgery is a class of forgery that is usually used for duplicating regions in an image in which a portion of the image is replicated and is placed on some other part of the same image. In this paper an efficient method to detect copy-move forgery in digital images is proposed. This segmentation based method detects the forged area in an image using two stages. In the first stage, the input image is segmented into independent patches and the features of these patches are compared with other patches to find the matching areas. Using the suspicious pair of patches from the first stage, a second stage of matching is done to confirm the existence of forgery using oversegmentation. The experimental analysis proves that the proposed method can effectively detect copy-move forgery.