LG Uplus System with Multi-Speaker IDs and Discriminator-based Sub-Judges for the WildSpoof Challenge

TL;DR

This paper presents a spoof-aware speaker verification system for high-quality TTS attacks, utilizing multi-speaker IDs and discriminator-based sub-judges to improve detection accuracy in the WildSpoof Challenge.

Contribution

The paper introduces dual and multi-speaker ID strategies and discriminator-based sub-judges using internal GAN features for enhanced spoof detection.

Findings

Improved detection cost function (a-DCF) on SpoofCeleb corpus.

Effective use of discriminator features for spoof detection.

Enhanced speaker verification robustness against TTS attacks.

Abstract

This paper describes our submission to the WildSpoof Challenge Track 2, which focuses on spoof-aware speaker verification (SASV) in the presence of high-quality text-to-speech (TTS) attacks. We adopt a ResNet-221 back-bone and study two speaker-labeling strategies, namelyDual-Speaker IDs and Multi-Speaker IDs, to explicitly enlarge the margin between bona fide and generated speech in the embedding space. In addition, we propose discriminator-based sub-judge systems that reuse internal features from HiFi-GAN and BigVGAN discriminators, aggregated via multi-query multi-head attentive statistics pooling(MQMHA). Experimental results on the SpoofCeleb corpus show that our system design is effective in improving agnostic detection cost function (a-DCF).