Generalizable Audio Deepfake Detection via Latent Space Refinement and Augmentation

TL;DR

This paper introduces a novel deepfake detection method that enhances generalization by refining and augmenting the latent space, effectively capturing diverse spoofing patterns and improving detection accuracy across multiple datasets.

Contribution

The paper proposes Latent Space Refinement and Augmentation techniques to improve deepfake detection generalization, a novel approach in the field.

Findings

The combined LSR and LSA approach outperforms existing methods.

The method achieves state-of-the-art results on multiple datasets.

Latent space augmentation increases spoofing pattern diversity.

Abstract

Advances in speech synthesis technologies, like text-to-speech (TTS) and voice conversion (VC), have made detecting deepfake speech increasingly challenging. Spoofing countermeasures often struggle to generalize effectively, particularly when faced with unseen attacks. To address this, we propose a novel strategy that integrates Latent Space Refinement (LSR) and Latent Space Augmentation (LSA) to improve the generalization of deepfake detection systems. LSR introduces multiple learnable prototypes for the spoof class, refining the latent space to better capture the intricate variations within spoofed data. LSA further diversifies spoofed data representations by applying augmentation techniques directly in the latent space, enabling the model to learn a broader range of spoofing patterns. We evaluated our approach on four representative datasets, i.e. ASVspoof 2019 LA, ASVspoof 2021 LA and DF, and In-The-Wild. The results show that LSR and LSA perform well individually, and their integration achieves competitive results, matching or surpassing current state-of-the-art methods.