Countering Anti-Forensics of Lateral Chromatic Aberration

Research has shown that lateral chromatic aberrations (LCA), an imaging fingerprint, can be anti-forensically modified to hide evidence of cut-and-paste forgery. In this paper, we propose a new technique for securing digital images against anti-forensic manipulation of LCA. To do this, we exploit resizing differences between color channels, which are induced by LCA anti-forensics, and define a feature vector to quantitatively capture these differences. Furthermore, we propose a detection method that exposes anti-forensically manipulated image patches. The technique algorithm is validated through experimental procedure, showing dependence on forgery patch size as well as anti-forensic scaling factor.