Copy-move forgery detection exploiting statistical image features

Copy-move forgery is a form of forgery on digital images, where an area of an image is copied and pasted to a different position of the same image. Majority of the image forgery detection techniques try to find dissimilarities or variation in natural image statistics. However such techniques fail in case of region duplication forgery detection, since here the forged area originates from the image itself. In this paper, we introduce a technique to find duplicate regions in an image, which exploits statistical features of an image. We use mean and variance for this purpose here, by splitting the image into pixel blocks. Mean is used to find the contribution of each individual block with respect to pixel intensity of the entire image, and variance is used to find how each pixel varies from its neighbors in a block. We evaluate the presented algorithm and compare with others copy-move forgery detection methods. According to our experimental results it is clear that presented algorithms is better perform to the existing techniques.