Investigation of Specaugment for Deep Speaker Embedding Learning
Shuai Wang, Johan Rohdin, Oldrich Plchot, Lukás Burget, Kai Yu, Jan Cernocký
Abstract
SpecAugment is a newly proposed data augmentation method for speech recognition. By randomly masking bands in the log Mel spectogram this method leads to impressive performance improvements. In this paper, we investigate the usage of SpecAugment for speaker verification tasks. Two different models, namely 1-D convolutional TDNN and 2-D convolutional ResNet34, trained with either Softmax or AAM-Softmax loss, are used to analyze SpecAugment's effectiveness. Experiments are carried out on the Voxceleb and NIST SRE 2016 dataset. By applying SpecAugment to the original clean data in an on-the-fly manner without complex off-line data augmentation methods, we obtained 3.72% and 11.49% EER for NIST SRE 2016 Cantonese and Tagalog, respectively. For Voxceleb1 evaluation set, we obtained 1.47% EER.
BibTeX
@inproceedings{icassp2020_investigationofs,
title = {Investigation of Specaugment for Deep Speaker Embedding Learning},
author = {Shuai Wang and Johan Rohdin and Oldrich Plchot and Lukás Burget and Kai Yu and Jan Cernocký},
booktitle = {ICASSP 2020},
year = {2020}
}