Deep Image Restoration For Image Anti-Forensics
Eren Tahir, Mert Bal
TL;DR
This work examines how deep image restoration can be used to counteract image forgery detection by applying anti-forensic techniques that degrade forensic traces and then restoring image quality to mask tampering. It proposes a two-step pipeline (corruption followed by restoration) that uses both non-deep-learning and deep-learning restorers (e.g., Restormer, SwinIR, FBCNN) to reduce the detectability of manipulations on high-resolution datasets. Evaluations on the DSO-1 and COVERAGE datasets with state-of-the-art detectors (Trufor, Early Fusion) show that restoration-based anti-forensics can significantly lower detection accuracy, at times approaching random guessing, and that Defense methods must anticipate such restorations. The results underscore the need to incorporate anti-forensics considerations into dataset construction and detector training to maintain robustness in real-world scenarios. The work provides public code and demonstrates how restoration techniques can meaningfully impact forgery detection pipelines.
Abstract
While image forensics is concerned with whether an image has been tampered with, image anti-forensics attempts to prevent image forensics methods from detecting tampered images. The competition between these two fields started long before the advancement of deep learning. JPEG compression, blurring and noising, which are simple methods by today's standards, have long been used for anti-forensics and have been the subject of much research in both forensics and anti-forensics. Although these traditional methods are old, they make it difficult to detect fake images and are used for data augmentation in training deep image forgery detection models. In addition to making the image difficult to detect, these methods leave traces on the image and consequently degrade the image quality. Separate image forensics methods have also been developed to detect these traces. In this study, we go one step further and improve the image quality after these methods with deep image restoration models and make it harder to detect the forged image. We evaluate the impact of these methods on image quality. We then test both our proposed methods with deep learning and methods without deep learning on the two best existing image manipulation detection models. In the obtained results, we show how existing image forgery detection models fail against the proposed methods. Code implementation will be publicly available at https://github.com/99eren99/DIRFIAF .