Transferable Dual-Domain Feature Importance Attack against AI-Generated Image Detector

TL;DR

This paper introduces DuFIA, a novel adversarial attack method targeting AI-generated image detectors by combining spatial and frequency domain features, demonstrating high transferability and robustness.

Contribution

The paper proposes a dual-domain feature importance attack that enhances transferability and robustness against various AI-generated image detectors.

Findings

DuFIA effectively transfers across different models.

It significantly reduces detector accuracy under attack.

The method maintains transparency and robustness.

Abstract

Recent AI-generated image (AIGI) detectors achieve impressive accuracy under clean condition. In view of antiforensics, it is significant to develop advanced adversarial attacks for evaluating the security of such detectors, which remains unexplored sufficiently. This letter proposes a Dual-domain Feature Importance Attack (DuFIA) scheme to invalidate AIGI detectors to some extent. Forensically important features are captured by the spatially interpolated gradient and frequency-aware perturbation. The adversarial transferability is enhanced by jointly modeling spatial and frequency-domain feature importances, which are fused to guide the optimization-based adversarial example generation. Extensive experiments across various AIGI detectors verify the cross-model transferability, transparency and robustness of DuFIA.