Copy move forgery detection using DCT, PatchMatch and cellular automata

摘要

Copy-move image forgery is a type of digital image forgery where content is copied and pasted within the same image, either by hiding foreground objects within an image by copying a background region over them or by emphasizing some foreground objects of the image through duplication. To prevent any consequences arising from the consumption of these forged images, many Copy-Move Forgery Detection (CMFD) techniques have been proposed in the past that analyze a suspected input image for a possible copy-move forgery. However, the existing detectors show limited detection accuracy in presence of post-processing manipulations like noise addition, compression, blur etc., which, in turn, are often used to hide the traces of tampering and produce convincing forgeries. In this paper, we propose a block-based CMFD technique which works well under these post-processing manipulations. We use Discrete Cosine Transform and Cellular Automata to extract features from the image blocks, which are subsequently matched using the patch match algorithm. Also, to extract the cloned regions corresponding to the matched features in a reliable way, we propose a simple and efficient CA-based post-processing procedure. The experimental results on the standard dataset demonstrate the effectiveness of our method for detecting copy-move forgeries under diverse post-processing manipulations of noise, compression, blur, brightness change, contrast change and their combinations.