Diffusion models meet image counter-forensics
This addresses the problem of image forgery detection for forensic analysts, presenting an incremental improvement in counter-forensics methods.
The paper tackles the problem of deceiving image forensics methods by using diffusion models to erase tampering traces, achieving superior performance in both deception and preserving image naturalness compared to existing counter-forensics techniques.
From its acquisition in the camera sensors to its storage, different operations are performed to generate the final image. This pipeline imprints specific traces into the image to form a natural watermark. Tampering with an image disturbs these traces; these disruptions are clues that are used by most methods to detect and locate forgeries. In this article, we assess the capabilities of diffusion models to erase the traces left by forgers and, therefore, deceive forensics methods. Such an approach has been recently introduced for adversarial purification, achieving significant performance. We show that diffusion purification methods are well suited for counter-forensics tasks. Such approaches outperform already existing counter-forensics techniques both in deceiving forensics methods and in preserving the natural look of the purified images. The source code is publicly available at https://github.com/mtailanian/diff-cf.