Towards single integrated spoofing-aware speaker verification embeddings
Abstract
This study aims to develop a single integrated spoofing-aware speaker verification (SASV) embeddings that satisfy two aspects. First, rejecting non-target speakers' input as well as target speakers' spoofed inputs should be addressed. Second, competitive performance should be demonstrated compared to the fusion of automatic speaker verification (ASV) and countermeasure (CM) embeddings, which outperformed single embedding solutions by a large margin in the SASV2022 challenge. We analyze that the inferior performance of single SASV embeddings comes from insufficient amount of training data and distinct nature of ASV and CM tasks. To this end, we propose a novel framework that includes multi-stage training and a combination of loss functions. Copy synthesis, combined with several vocoders, is also exploited to address the lack of spoofed data. Experimental results show dramatic improvements, achieving a SASV-EER of 1.06% on the evaluation protocol of the SASV2022 challenge.
Keywords
Cite
@article{arxiv.2305.19051,
title = {Towards single integrated spoofing-aware speaker verification embeddings},
author = {Sung Hwan Mun and Hye-jin Shim and Hemlata Tak and Xin Wang and Xuechen Liu and Md Sahidullah and Myeonghun Jeong and Min Hyun Han and Massimiliano Todisco and Kong Aik Lee and Junichi Yamagishi and Nicholas Evans and Tomi Kinnunen and Nam Soo Kim and Jee-weon Jung},
journal= {arXiv preprint arXiv:2305.19051},
year = {2023}
}
Comments
Accepted by INTERSPEECH 2023. Code and models are available in https://github.com/sasv-challenge/ASVSpoof5-SASVBaseline