Applied Image Processing Algorithms for Forensic Applications

Abstract

The research activities described in this thesis have been mainly focused on the analysis and development of techniques able to detect image forgeries. The first part of the thesis has been devoted on detecting tampering by using compression artifacts analysis. Driven by an application aiming to retrieve the original camera model used to shoot the image, two research topics have been targeted: the cropping detection, i.e. the operation that allows obtaining a subpart of the image, and the double quantization detection, i.e. the fact that an image, previously compressed, is opened, modified and hence re-compressed. Moreover techniques aiming to retrieve the first quantization coefficients were also been analyzed. A second part of the thesis is devoted to describe a new system (both hardware and software) developed in the IPLab (Image Processing Lab) of the University of Catania for banknote counterfeit detection.