PIZZARO project aims at three issues related to the digital images, their acquisition and analysis:
- identification of imaging device
- verification of acquired images
- insufficient quality of acquired data for further analysis
The ability to uniquely identify or to exclude certain types/models of imaging devices as an alleged imaging device is very important for the investigation process. The Module for Imaging Device Identification provides decision about the manufacturer, given model and/or specific unit based on the analysis of sensor noise characteristics. Noise characteristics were chosen for their discriminability, stability, and accuracy even under real conditions. Next to the matching of image and video data to a certain imaging device, it is important to be able to decide if the data were tampered. The second Module for Analysis of Additional Changes is devoted to the verification of image data. The methodology is based on the scientific results achieved in UTIA using image and video mathematical characteristics.
The last Module for Image and Video Reconstruction addresses the situation when the quality of acquired images or videos is not sufficient. Images can be blurred, noisy or degraded in some other unknown way. A methodology based on the blind deconvolution and super-resolution is the baseline for further development of image quality enhancement and restoration. Algorithms are further modified for the application in forensic research.
Proposed approaches are designed to be applicable on both images and video sequences. This improves the applicability of the video data as forensic evidence. The project was carried out in the close cooperation between UTIA and Institute of Criminalistics (Kriminalistický ústav) in Prague.