A Comparison of SSL-Based Feature Extractors and Back-End Classifiers for Spoofing Detection: A Multi-Corpus Training and Cross-Linguistic Analysis

2026-06-07

AI Summary

This study addresses the vulnerability of speaker verification systems to spoofing attacks and the inconsistent performance of existing countermeasures under cross-dataset and cross-lingual conditions. The authors systematically evaluate four self-supervised feature extractors combined with four backend classifiers across multilingual and multi-corpus settings, uncovering significant domain bias in the ASVspoof 2021 (ASVspoof5) dataset. To mitigate these issues, they propose a global sequence modeling approach integrating attention mechanisms with graph neural networks. Their experiments demonstrate that improper data scaling adversely affects performance, while fine-tuning on as little as eight hours of target-language data substantially enhances cross-lingual robustness. The method is validated across six evaluation datasets, underscoring the critical importance of domain-aware design and language-specific adaptation in anti-spoofing systems.

Abstract

Voice biometric systems face growing threats from spoofing attacks, yet the evaluation of detection models remains inconsistent across datasets. To investigate these unpredictable fluctuations, we conduct a comprehensive benchmark of four self-supervised learning feature extractors paired with four back-end classifiers. We compare the hierarchical local feature extraction of ResNet with the global sequence and relational modeling of attention and graph-based back-ends. Through multi-corpus training across three scenarios and six evaluation datasets, our empirical analysis yields two critical findings. First, we expose a domain bias within the ASVspoof 5 dataset, showing that naive data scaling actively degrades performance. Second, our cross-linguistic analysis reveals that fine-tuning with just 8 hours of target-language data enhances detection robustness. Together, these findings emphasize the critical need for domain-aware and language-specific adaptation in spoofing detection.

Problem

• spoofing detection
• voice biometrics
• domain bias
• cross-linguistic analysis
• multi-corpus training

Innovation

• self-supervised learning
• spoofing detection
• cross-linguistic analysis
• multi-corpus training
• domain bias