Tamper Detection and Localization in Forensic Images

Tamper Detection and Localization in Forensic Images

연구 분야: Analysis

학회: CompSysTech '23: Proceedings of the 24th International Conference on Computer Systems and Technologies

Malicious users may tamper forensic digital images for various reasons including among others the creation of false evidences. In this work we propose a tamper detection and localization technique for forensic digital images detecting whether a forensic image has been tampered and locating the area that has been modified. The methodology deploys the partition of the image into blocks and the calculation of the hash values of the red channel in particular areas of the image, the so-called integrity-blocks, embedding these values into the green channel of other blocks based on a permutation of the integrity-blocks. Moreover, once a tampering has been detected into a forensic image the derived system that implements the proposed method provides a visual evidence of the tampering by coloring the tampered area with red color identifying the extent and point of the tampering, making it easier to understand the issue and resolve it. We provide an experimental evaluation of the proposed technique by performing a set of attacks by both inserting and removing elements from forensic images proving the potentials of our model in detecting and precisely locating the tampered areas.