Detecting Image Forgery Using XOR and Determinant of Pixels for Image Forensics

Abstract

© 2015 IEEE. Using images in case investigation may doubt that the image are forged or not. Image forensics is an approach for image validity. In this paper, we propose a new method for detecting forgery of image that use of the truth from XOR comparison between two images, and the determinant of 3x3 pixels for more performance instead of comparing pixel by pixel. Euclidean of a pixel is computed to reduce the dimensions from RGB vectors. In this experiment, XOR with pixel-by-pixel and XOR with determinant of 3x3 pixels were compared. Fifty images were used: 25 real images and 25 forged images. The result showed that our new method increased speed 36.42% and the error was at 13%.