The DKU-OPPO System for the 2022 Spoofing-Aware Speaker Verification Challenge

TL;DR

This paper presents a dual-system approach for the 2022 SASV Challenge, employing separate optimized speaker verification and spoofing countermeasure modules, with novel augmentation and loss techniques, achieving significant performance improvements.

Contribution

We introduce two new methods, ERSA and OCCL, for spoofing detection, and explore embedding-based improvements, achieving state-of-the-art results on the challenge dataset.

Findings

Cascaded system reduces SASV-EER from 1.71% to 0.21%.

Embedding techniques can enhance spoofing countermeasure performance.

Proposed methods outperform baseline on challenge dataset.

Abstract

This paper describes our DKU-OPPO system for the 2022 Spoofing-Aware Speaker Verification (SASV) Challenge. First, we split the joint task into speaker verification (SV) and spoofing countermeasure (CM), these two tasks which are optimized separately. For ASV systems, four state-of-the-art methods are employed. For CM systems, we propose two methods on top of the challenge baseline to further improve the performance, namely Embedding Random Sampling Augmentation (ERSA) and One-Class Confusion Loss(OCCL). Second, we also explore whether SV embedding could help improve CM system performance. We observe a dramatic performance degradation of existing CM systems on the domain-mismatched Voxceleb2 dataset. Third, we compare different fusion strategies, including parallel score fusion and sequential cascaded systems. Compared to the 1.71% SASV-EER baseline, our submitted cascaded system obtains a 0.21% SASV-EER on the challenge official evaluation set.