Identification of source-enhanced composite images

Abstract

Digital images have an important function in several fields like journalism, film industry and forensic investigations. Several image editing softwares can change the content of an image very easily. Attackers use contrast enhancement for avoiding the traces left by image forgery. So it is necessary to perform contrast enhancement detection for detecting an image forgery. In the proposed system, there are three algorithms that detect the global contrast enhancement, the type of contrast enhancement operation applied and local contrast enhancement in digital images. The global contrast enhancement is detected with the identification of zero-height gap bins present in the histogram. Then the type of contrast enhancement operation applied to each image is predicted using artificial neural network and support vector machine. Later on, for the detection of local contrast enhancement mappings, the positions of detected blockwise peak or gap bins are combined and for discovering the cut and paste image forgeries, the consistency between regional artifacts is checked. To verify the effectiveness and efficiency of the proposed technique extensive experiments have verified.