基于三类公式与似然比的集成抗欺骗自动说话人验证
音频与语音处理
2026-03-19 v2 声音
摘要
抗欺骗自动说话人验证(SASV)旨在集成自动说话人验证(ASV)与反欺骗措施(CM)。流行的解决方案是对独立的 ASV 和 CM 分数进行融合。为更好地建模 SASV,一些框架将 ASV 和 CM 集成到单个网络中。然而,这些解决方案通常基于双编码器,解释性有限,且无法在无需重新训练的情况下适应新的评估参数。基于此,我们提出了一个通过三类公式实现的统一端到端框架,使其能够从类别 logits 中进行似然比(LLR)推断,以实现更可解释的决策流程。实验表明,在 ASVSpoof5 和 SpoofCeleb 上的性能与现有方法相当或更好。可视化和分析也证明,三类改造提供了更好的解释性。
引用
@article{arxiv.2603.13780,
title = {Integrated Spoofing-Robust Automatic Speaker Verification via a Three-Class Formulation and LLR},
author = {Kai Tan and Lin Zhang and Ruiteng Zhang and Johan Rohdin and Leibny Paola García-Perera and Zexin Cai and Sanjeev Khudanpur and Matthew Wiesner and Nicholas Andrews},
journal= {arXiv preprint arXiv:2603.13780},
year = {2026}
}
备注
Submitted to Interspeech 2026; put on arxiv based on requirement from Interspeech: "Interspeech no longer enforces an anonymity period for submissions." and "For authors that prefer to upload their paper online, a note indicating that the paper was submitted for review to Interspeech should be included in the posting."