Detection of tampered region for JPEG images by using mode-based first digit features

Abstract

Abstract With the widespread availability of image editing software, digital images have been becoming easy to manipulate and edit even for non-professional users. For a tampered Joint Photographic Experts Group (JPEG) image, the tampered region usually has different JPEG compression history from the authentic region, which can be used to detect and locate the tampered region. In this article, we propose to apply the statistical features of the first digits of individual alternate current modes and support vector machine to detect and locate the tampered region. Experimental results show that our proposed method is effective for detecting three popularly used image manipulations. Its expectation of the percentage of overlap between the detected tampered region and the truth tampered region is higher than the existing algorithms.