Abstract
Rapid and affordable methods of testing for COVID-19 infection are essential to manage infection rates and prevent medical facilities from becoming overwhelmed. This study demonstrates that crowdsourced cough audio samples acquired on smartphones across the world and paired with COVID-19 status labels can be used to develop an AI algorithm that accurately predicts COVID-19 infection with an ROC-AUC of 77.1% (75.2%-78.3%). Furthermore, this AI algorithm is able to generalize to crowdsourced samples from Latin America and clinical samples from South Asia, without further training using the specific samples. As more crowdsourced data is collected, further development can be implemented using various respiratory audio samples to create a cough analysis-based AI solution for COVID-19 detection that can likely generalize globally to all demographic groups in both clinical and non-clinical settings.
Abstract (translated)
URL
https://arxiv.org/abs/2011.13320