Image forgery detection via forensic similarity graphs
Resumen:
In the article 'Exposing Fake Images with Forensic Similarity Graphs', O. Mayer and M. C. Stamm introduce a novel image forgery detection method. The proposed method is built on a graph-based representation of images, where image patches are represented as the vertices of the graph, and the edge weights are assigned in order to reflect the forensic similarity between the connected patches. In this representation, forged regions form highly connected subgraphs. Therefore, forgery detection and localization can be cast as a cluster analysis problem on the similarity graph. The authors present two graph clustering methods to detect and localize image forgeries. In this paper, we present briefly the method and offer an online executable version allowing everyone to test it on their own suspicious images.
| 2022 | |
|
Projecto ANR-16-DEFA-0004 Proyecto vera.ai (101070093) |
|
|
Image forensics Forgery detection Graph clustering |
|
| Inglés | |
| Universidad de la República | |
| COLIBRI | |
|
http://www.ipol.im/pub/art/2022/432/
https://hdl.handle.net/20.500.12008/39790 |
|
| Acceso abierto | |
| Licencia Creative Commons Atribución - No Comercial - Compartir Igual (CC - By-NC-SA 4.0) |
| Sumario: | In the article 'Exposing Fake Images with Forensic Similarity Graphs', O. Mayer and M. C. Stamm introduce a novel image forgery detection method. The proposed method is built on a graph-based representation of images, where image patches are represented as the vertices of the graph, and the edge weights are assigned in order to reflect the forensic similarity between the connected patches. In this representation, forged regions form highly connected subgraphs. Therefore, forgery detection and localization can be cast as a cluster analysis problem on the similarity graph. The authors present two graph clustering methods to detect and localize image forgeries. In this paper, we present briefly the method and offer an online executable version allowing everyone to test it on their own suspicious images. |
|---|
