Audio-Based Deep Learning Frameworks for Detecting COVID-19

TL;DR

This study presents an audio-based deep learning framework that effectively detects COVID-19 from breathing, cough, and speech sounds, outperforming existing methods in key diagnostic metrics.

Contribution

The paper introduces a novel combination of spectrogram feature extraction, high-level embedding, and LightGBM classification for COVID-19 detection from audio data.

Findings

Achieved highest AUC of 89.03% in the Second DiCOVA Challenge.

Outperformed state-of-the-art systems with improvements in AUC, F1 score, and sensitivity.

Demonstrated the effectiveness of combining spectrogram features with deep learning and LightGBM.

Abstract

This paper evaluates a wide range of audio-based deep learning frameworks applied to the breathing, cough, and speech sounds for detecting COVID-19. In general, the audio recording inputs are transformed into low-level spectrogram features, then they are fed into pre-trained deep learning models to extract high-level embedding features. Next, the dimension of these high-level embedding features are reduced before finetuning using Light Gradient Boosting Machine (LightGBM) as a back-end classification. Our experiments on the Second DiCOVA Challenge achieved the highest Area Under the Curve (AUC), F1 score, sensitivity score, and specificity score of 89.03%, 64.41%, 63.33%, and 95.13%, respectively. Based on these scores, our method outperforms the state-of-the-art systems, and improves the challenge baseline by 4.33%, 6.00% and 8.33% in terms of AUC, F1 score and sensitivity score, respectively.