An explainable COVID-19 detection system based on human sounds

Huining Li; Xingyu Chen; Xiaoye Qian; Huan Chen; Zhengxiong Li; Soumyadeep Bhattacharjee; Hanbin Zhang; Ming-Chun Huang; Wenyao Xu

doi:10.1016/j.smhl.2022.100332

An explainable COVID-19 detection system based on human sounds

Smart Health (Amst). 2022 Dec:26:100332. doi: 10.1016/j.smhl.2022.100332. Epub 2022 Oct 19.

Authors

Huining Li¹, Xingyu Chen², Xiaoye Qian³, Huan Chen³, Zhengxiong Li², Soumyadeep Bhattacharjee¹, Hanbin Zhang¹, Ming-Chun Huang^{4

5}, Wenyao Xu¹

Affiliations

¹ Department of Computer Science and Engineering, University at Buffalo, United States.
² Department of Computer Science and Engineering, University of Colorado Denver, United States.
³ Department of Electrical, Computer, and Systems Engineering, Case Western Reserve University, United States.
⁴ Department of Data and Computational Science, Duke Kunshan University, China.
⁵ Suzhou Huanmu Intelligence Technology Co., Ltd., China.

Abstract

Acoustic signals generated by the human body have often been used as biomarkers to diagnose and monitor diseases. As the pathogenesis of COVID-19 indicates impairments in the respiratory system, digital acoustic biomarkers of COVID-19 are under investigation. In this paper, we explore an accurate and explainable COVID-19 diagnosis approach based on human speech, cough, and breath data using the power of machine learning. We first analyze our design space considerations from the data aspect and model aspect. Then, we perform data augmentation, Mel-spectrogram transformation, and develop a deep residual architecture-based model for prediction. Experimental results show that our system outperforms the baseline, with the ROC-AUC result increased by 5.47%. Finally, we perform an interpretation analysis based on the visualization of the activation map to further validate the model.

Keywords: Accurate; Acoustic; COVID-19; Explainable.