Learning Emotion-Invariant Speaker Representations for Speaker Verification

TL;DR

This paper introduces a novel approach to enhance speaker verification by developing emotion-invariant speaker representations through data augmentation, cosine similarity loss, and emotion-aware masking, significantly reducing error rates.

Contribution

The paper presents a combined method including data augmentation, loss function modification, and masking to improve emotion robustness in speaker verification systems.

Findings

Achieved a 19.29% relative reduction in EER.

Demonstrated the effectiveness of emotion-aware masking.

Validated improvements through comprehensive ablation studies.

Abstract

In recent years, the rapid progress in speaker verification (SV) technology has been driven by the extraction of speaker representations based on deep learning. However, such representations are still vulnerable to emotion variability. To address this issue, we propose multiple improvements to train speaker encoders to increase emotion robustness. Firstly, we utilize CopyPaste-based data augmentation to gather additional parallel data, which includes different emotional expressions from the same speaker. Secondly, we apply cosine similarity loss to restrict parallel sample pairs and minimize intra-class variation of speaker representations to reduce their correlation with emotional information. Finally, we use emotion-aware masking (EM) based on the speech signal energy on the input parallel samples to further strengthen the speaker representation and make it emotion-invariant. We conduct a comprehensive ablation study to demonstrate the effectiveness of these various components. Experimental results show that our proposed method achieves a relative 19.29\% drop in EER compared to the baseline system.