Forgery detection using statistical features

Digital images are present everywhere on magazine covers, in newspapers, in courtrooms as evidences, and all over the Internet signifying one of the major ways for communication nowadays. Easy availability of image editing tools has made it very simple to tamper the digital images thus putting the authenticity of these images under suspicion. There are a number of type of forgeries that can be carried out on digital images most common being the copy-move and splicing forgeries. This paper proposes a new method for image copy-move and splicing detection based on statistical features of the digital image. Copy-move involves copying of a region in an image and pasting it somewhere in the same image to hide any important detail and Splicing involves merging of two or more images to form a composite image that is significantly different from the original image. The proposed approach calculates grey level run length matrix (GLRLM) texture features for the forged images and original images. Support vector machine is used for classification. Results show that the proposed algorithm is very effective in detection of forgery.