Investigating feature selection and explainability for COVID-19 diagnostics from cough sounds

Avila, F and Poorjam, AH and Mittal, D and Dognin, C and Muguli, A and Kumar, R and Chetupalli, SR and Ganapathy, S and Singh, M (2021) Investigating feature selection and explainability for COVID-19 diagnostics from cough sounds. In: 22nd Annual Conference of the International Speech Communication Association, INTERSPEECH 2021, 30 Aug - 03 Sep 2021, Brno, pp. 4246-4250.

PDF INTERSPEECH_2021.pdf - Published Version Restricted to Registered users only Download (457kB) | Request a copy

Abstract

In this paper, we propose an approach to automatically classify COVID-19 and non-COVID-19 cough samples based on the combination of both feature engineering and deep learning models. In the feature engineering approach, we develop a support vector machine classifier over high dimensional (6373D) space of acoustic features. In the deep learning-based approach, on the other hand, we apply a convolutional neural network trained on the log-mel spectrograms. These two methodologically diverse models are then combined by fusing the probability scores of the models. The proposed system, which ranked 9th on the 2021 Diagnosing COVID-19 using Acoustics (Di- COVA) challenge leaderboard, obtained an area under the receiver operating characteristic curve (AUC) of 0:81 on the blind test data set, which is a 10:9 absolute improvement compared to the baseline. Moreover, we analyze the explainability of the deep learning-based model when detecting COVID-19 from cough signals. Copyright © 2021 ISCA.

Item Type:	Conference Paper
Publication:	Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
Publisher:	International Speech Communication Association
Additional Information:	The copyright for this article belongs to International Speech Communication Association
Keywords:	Convolutional neural networks; Deep learning; Speech communication; Statistical tests; Support vector machines, Cough sounds; COVID-19; Explainability; Feature engineerings; Features selection; High-dimensional; Higher-dimensional; Learning models; Respiratory diagnosis; Support vector machine classifiers, Vector spaces
Department/Centre:	Division of Electrical Sciences > Electrical Engineering
Date Deposited:	03 Dec 2021 08:50
Last Modified:	03 Dec 2021 08:50
URI:	http://eprints.iisc.ac.in/id/eprint/70635

Actions (login required)

View Item