Spatial Reconstructed Local Attention Res2Net with F0 Subband for Fake Speech Detection

TL;DR

Fake speech detection faces a challenge because the fundamental frequency ($F0$) of synthetic speech often differs from real speech. The authors introduce a dedicated $F0$ subband and a Spatial Reconstructed Local Attention Res2Net (SR-LA Res2Net) backbone to model this subband with multiscale features and local attention, aided by a spatial reconstruction mechanism to preserve information across channel groups. The proposed approach achieves state-of-the-art performance among single systems on the ASVspoof 2019 LA dataset, with an EER of 0.47% and a min t-DCF of 0.0159. This work demonstrates the discriminative potential of $F0$-based cues and showcases how combining Res2Net with local attention for subband modeling can improve robustness in fake speech detection.

Abstract

The rhythm of bonafide speech is often difficult to replicate, which causes that the fundamental frequency (F0) of synthetic speech is significantly different from that of real speech. It is expected that the F0 feature contains the discriminative information for the fake speech detection (FSD) task. In this paper, we propose a novel F0 subband for FSD. In addition, to effectively model the F0 subband so as to improve the performance of FSD, the spatial reconstructed local attention Res2Net (SR-LA Res2Net) is proposed. Specifically, Res2Net is used as a backbone network to obtain multiscale information, and enhanced with a spatial reconstruction mechanism to avoid losing important information when the channel group is constantly superimposed. In addition, local attention is designed to make the model focus on the local information of the F0 subband. Experimental results on the ASVspoof 2019 LA dataset show that our proposed method obtains an equal error rate (EER) of 0.47% and a minimum tandem detection cost function (min t-DCF) of 0.0159, achieving the state-of-the-art performance among all of the single systems.