The IDLAB VoxSRC-20 Submission: Large Margin Fine-Tuning and Quality-Aware Score Calibration in DNN Based Speaker Verification

Abstract

In this paper we propose and analyse a large margin fine-tuning strategy and a quality-aware score calibration in text-independent speaker verification. Large margin fine-tuning is a secondary training stage for DNN based speaker verification systems trained with margin-based loss functions. It enables the network to create more robust speaker embeddings by enabling the use of longer training utterances in combination with a more aggressive margin penalty. Score calibration is a common practice in speaker verification systems to map output scores to well-calibrated log-likelihood-ratios, which can be converted to interpretable probabilities. By including quality features in the calibration system, the decision thresholds of the evaluation metrics become quality-dependent and more consistent across varying trial conditions. Applying both enhancements on the ECAPA-TDNN architecture leads to state-of-the-art results on all publicly available VoxCeleb1 test sets and contributed to our winning submissions in the supervised verification tracks of the VoxCeleb Speaker Recognition Challenge 2020.

arXiv commentproceedings of ICASSP 2021

References in corpus (12)

• Adam: A Method for Stochastic Optimization
• SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
• Res2Net: A New Multi-scale Backbone Architecture
• VoxCeleb2: Deep Speaker Recognition
• Squeeze-and-Excitation Networks
• VoxCeleb: a large-scale speaker identification dataset
• ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification
• Attentive Statistics Pooling for Deep Speaker Embedding
• The BOSARIS Toolkit: Theory, Algorithms and Code for Surviving the New DCF
• VoxSRC 2020: The Second VoxCeleb Speaker Recognition Challenge
• Cross-Lingual Speaker Verification with Domain-Balanced Hard Prototype Mining and Language-Dependent Score Normalization
• The IDLAB VoxCeleb Speaker Recognition Challenge 2020 System Description

Cited by in corpus (11)

• WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing
• Integrating Frequency Translational Invariance in TDNNs and Frequency Positional Information in 2D ResNets to Enhance Speaker Verification
• The SpeakIn System for VoxCeleb Speaker Recognition Challange 2021
• Leveraging ASR Pretrained Conformers for Speaker Verification through Transfer Learning and Knowledge Distillation
• The VoxCeleb Speaker Recognition Challenge: A Retrospective
• The IDLAB VoxCeleb Speaker Recognition Challenge 2020 System Description
• Tackling the Score Shift in Cross-Lingual Speaker Verification by Exploiting Language Information
• Beijing ZKJ-NPU Speaker Verification System for VoxCeleb Speaker Recognition Challenge 2021
• North America Bixby Speaker Diarization System for the VoxCeleb Speaker Recognition Challenge 2021
• HeightCeleb - an enrichment of VoxCeleb dataset with speaker height information
• Two Methods for Spoofing-Aware Speaker Verification: Multi-Layer Perceptron Score Fusion Model and Integrated Embedding Projector