A JPEG blocking artifact detector for image forensics

Abstract

Detection of JPEG blocking artifacts in an image plays an important role in the forensic analysis. JPEG anti-forensic techniques try to remove these artifacts to disguise forensic detectors. This paper proposes a new technique for detecting JPEG blocking artifacts. Unlike in uncompressed images, when a JPEG compressed image is cropped, there is a change in the inter-block correlation of DCT coefficients due to the shifting of blocking artifacts. We propose to use this change in correlation to detect the blockiness in an image. Experiments conducted on a large set of images show that the proposed detector outperforms the existing detectors in detecting blocking artifacts in anti-forensically modified images. © 2018 Elsevier B.V.