Chain of evidence generation for contrast enhancement in digital image forensics
Sebastiano Battiato, Giuseppe Messina, Daniela Strano · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2010
The quality of the images obtained by digital cameras has improved a lot since digital cameras early days. Unfortunately, it is not unusual in image forensics to find wrongly exposed pictures. This is mainly due to obsolete techniques or old technologies, but also due to backlight conditions. To extrapolate some invisible details a stretching of the image contrast is obviously required. The forensics rules to produce evidences require a complete documentation of the processing steps, enabling the replication of the entire process. The automation of enhancement techniques is thus quite difficult and needs to be carefully documented. This work presents an automatic procedure to find contrast enhancement settings, allowing both image correction and automatic scripting generation. The technique is based on a preprocessing step which extracts the features of the image and selects correction parameters. The parameters are thus saved through a JavaScript code that is used in the second step of the approach to correct the image. The generated script is Adobe Photoshop compliant (which is largely used in image forensics analysis) thus permitting the replication of the enhancement steps. Experiments on a dataset of images are also reported showing the effectiveness of the proposed methodology.