Frequency Masking for Universal Deepfake Detection

TL;DR

This paper introduces a novel frequency masking approach for universal deepfake detection, leveraging frequency domain analysis to improve generalization across unseen generative AI methods.

Contribution

It pioneers the use of frequency masking in deepfake detection, demonstrating significant performance improvements over existing spatial domain methods.

Findings

Frequency masking enhances detection accuracy on unseen deepfake methods.

Frequency domain focus outperforms spatial domain approaches.

Proposed method achieves state-of-the-art results in universal deepfake detection.

Abstract

We study universal deepfake detection. Our goal is to detect synthetic images from a range of generative AI approaches, particularly from emerging ones which are unseen during training of the deepfake detector. Universal deepfake detection requires outstanding generalization capability. Motivated by recently proposed masked image modeling which has demonstrated excellent generalization in self-supervised pre-training, we make the first attempt to explore masked image modeling for universal deepfake detection. We study spatial and frequency domain masking in training deepfake detectors. Based on empirical analysis, we propose a novel deepfake detector via frequency masking. Our focus on frequency domain is different from the majority, which primarily target spatial domain detection. Our comparative analyses reveal substantial performance gains over existing methods. Code and models are publicly available.