UC San Diego

The COVID-19 pandemic has overwhelmed health care systems around the world as a very serious pulmonary ailment that often leads to coughing. Cough sounds contain underutilized pulmonary health information that can be analyzed. Using signal analysis methods for audio,

this paper explores the analysis and classification of cough sounds and audio samples from individuals diagnosed with COVID-19. Using features extracted through signal analysis, a classifier is developed to evaluate whether an audio sample is likely to have COVID-19 symptoms. This can potentially be used for remote and early diagnostic efforts to help the world initiatives tackling this pandemic.