Deep Learning Solution for Pathological Voice Detection using LSTM-based Autoencoder Hybrid with Multi-Task Learning

Dávid Sztahó, Kiss Gábor, Tulics Miklós Gábriel

2021

Abstract

In this paper, a deep learning approach is introduced to detect pathological voice disorders from continuous speech. Speech as bio-signal is getting more and more attention as a discriminant for different diseases. To exploit information in speech, a long-short term memory (LSTM) autoencoder hybrid with multi-task learning solution is proposed with spectrogram as input feature. Different speech databases (voice disorders, depression, Parkinson’s disease) are applied as evaluation datasets. Applicability of the method is demonstrated by obtaining accuracies 85% for Parkinson’s disease, 86% for dysphonia, and 90% for depression on test datasets. The advantage of this method is that it is fully data-driven, in the sense that it does not require special acoustic-phonetic preprocessing separately for the types of disease to be recognized. We believe that the applied method in this article can be used to other diseases as well and can be used for other languages also.

Paper Citation

in Harvard Style

Sztahó D., Gábor K. and Gábriel T. (2021). Deep Learning Solution for Pathological Voice Detection using LSTM-based Autoencoder Hybrid with Multi-Task Learning. In Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 4: BIOSIGNALS; ISBN 978-989-758-490-9, SciTePress, pages 135-141. DOI: 10.5220/0010193100002865

in Bibtex Style

@conference{biosignals21,
author={Dávid Sztahó and Kiss Gábor and Tulics Miklós Gábriel},
title={Deep Learning Solution for Pathological Voice Detection using LSTM-based Autoencoder Hybrid with Multi-Task Learning},
booktitle={Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 4: BIOSIGNALS},
year={2021},
pages={135-141},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010193100002865},
isbn={978-989-758-490-9},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 4: BIOSIGNALS
TI - Deep Learning Solution for Pathological Voice Detection using LSTM-based Autoencoder Hybrid with Multi-Task Learning
SN - 978-989-758-490-9
AU - Sztahó D.
AU - Gábor K.
AU - Gábriel T.
PY - 2021
SP - 135
EP - 141
DO - 10.5220/0010193100002865
PB - SciTePress