Frequency Bias Matters: Diving into Robust and Generalized Deep Image Forgery Detection

TL;DR

This paper identifies frequency bias as a key factor affecting the robustness and generalization of deep image forgery detectors, proposing a frequency alignment method to enhance detection reliability and counteract forgeries.

Contribution

It introduces a frequency perspective analysis revealing bias causes, and proposes a novel two-step frequency alignment technique for improved forgery detection and anti-forensic attacks.

Findings

Frequency bias impacts detector robustness and generalization.

The proposed frequency alignment improves detection accuracy.

The method is effective across multiple detectors and forgery models.

Abstract

As deep image forgery powered by AI generative models, such as GANs, continues to challenge today's digital world, detecting AI-generated forgeries has become a vital security topic. Generalizability and robustness are two critical concerns of a forgery detector, determining its reliability when facing unknown GANs and noisy samples in an open world. Although many studies focus on improving these two properties, the root causes of these problems have not been fully explored, and it is unclear if there is a connection between them. Moreover, despite recent achievements in addressing these issues from image forensic or anti-forensic aspects, a universal method that can contribute to both sides simultaneously remains practically significant yet unavailable. In this paper, we provide a fundamental explanation of these problems from a frequency perspective. Our analysis reveals that the frequency bias of a DNN forgery detector is a possible cause of generalization and robustness issues. Based on this finding, we propose a two-step frequency alignment method to remove the frequency discrepancy between real and fake images, offering double-sided benefits: it can serve as a strong black-box attack against forgery detectors in the anti-forensic context or, conversely, as a universal defense to improve detector reliability in the forensic context. We also develop corresponding attack and defense implementations and demonstrate their effectiveness, as well as the effect of the frequency alignment method, in various experimental settings involving twelve detectors, eight forgery models, and five metrics.