Publication Details
How To Improve Your Speaker Embeddings Extractor in Generic Toolkits
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Rohdin Johan Andréas, M.Sc., Ph.D. (DCGM)
Stafylakis Themos
Černocký Jan, prof. Dr. Ing. (DCGM)
Deep neural network, speaker embedding, xvector,Tensorflow, Kaldi.
Recently, speaker embeddings extracted with deep neural networksbecame the state-of-the-art method for speaker verification. In thispaper we aim to facilitate its implementation on a more generictoolkit than Kaldi, which we anticipate to enable further improvementson the method. We examine several tricks in training, suchas the effects of normalizing input features and pooled statistics, differentmethods for preventing overfitting as well as alternative nonlinearitiesthat can be used instead of Rectifier Linear Units. In addition,we investigate the difference in performance between TDNNand CNN, and between two types of attention mechanism. Experimentalresults on Speaker in the Wild, SRE 2016 and SRE 2018datasets demonstrate the effectiveness of the proposed implementation.
@inproceedings{BUT158087,
author="Hossein {Zeinali} and Lukáš {Burget} and Johan Andréas {Rohdin} and Themos {Stafylakis} and Jan {Černocký}",
title="How To Improve Your Speaker Embeddings Extractor in Generic Toolkits",
booktitle="Proceedings of 2019 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)",
year="2019",
pages="6141--6145",
publisher="IEEE Signal Processing Society",
address="Brighton",
doi="10.1109/ICASSP.2019.8683445",
isbn="978-1-5386-4658-8",
url="https://ieeexplore.ieee.org/abstract/document/8683445"
}