Adaptive Frequency Learning in Two-branch Face Forgery Detection

TL;DR

This paper introduces AFD, an adaptive frequency learning framework for face forgery detection that dynamically learns frequency decomposition and integrates it with spatial features, outperforming fixed-transform methods.

Contribution

The paper proposes a novel adaptive frequency learning approach that automatically learns frequency decomposition and incorporates it into a two-branch face forgery detection network.

Findings

AFD outperforms existing fixed-transform methods.

Adaptive frequency learning improves detection accuracy.

Extensive experiments validate the effectiveness of AFD.

Abstract

Face forgery has attracted increasing attention in recent applications of computer vision. Existing detection techniques using the two-branch framework benefit a lot from a frequency perspective, yet are restricted by their fixed frequency decomposition and transform. In this paper, we propose to Adaptively learn Frequency information in the two-branch Detection framework, dubbed AFD. To be specific, we automatically learn decomposition in the frequency domain by introducing heterogeneity constraints, and propose an attention-based module to adaptively incorporate frequency features into spatial clues. Then we liberate our network from the fixed frequency transforms, and achieve better performance with our data- and task-dependent transform layers. Extensive experiments show that AFD generally outperforms.