Phase-Aware Spoof Speech Detection Based on Res2Net with Phase Network

TL;DR

This paper introduces a phase-aware spoof speech detection system using Res2Net with a dedicated phase network, improving detection accuracy by effectively integrating magnitude and phase features for diverse spoofing attacks.

Contribution

The paper proposes a novel phase network to reduce feature randomness, enabling effective fusion of magnitude and phase features in spoof speech detection.

Findings

Significant performance improvement in spoof detection, especially for attacks where phase info is crucial.

Effective handling of feature randomness enhances generalization to unknown spoofing attacks.

Demonstrated robustness in both known- and unknown-kind spoofing scenarios.

Abstract

The spoof speech detection (SSD) is the essential countermeasure for automatic speaker verification systems. Although SSD with magnitude features in the frequency domain has shown promising results, the phase information also can be important to capture the artefacts of certain types of spoofing attacks. Thus, both magnitude and phase features must be considered to ensure the generalization ability to diverse types of spoofing attacks. In this paper, we investigate the failure reason of feature-level fusion of the previous works through the entropy analysis from which we found that the randomness difference between magnitude and phase features is large, which can interrupt the feature-level fusion via backend neural network; thus, we propose a phase network to reduce that difference. Our SSD system: phase network equipped Res2Net achieved significant performance improvement, specifically in the spoofing attack for which the phase information is considered to be important. Also, we demonstrate our SSD system in both known- and unknown-kind SSD scenarios for practical applications.