Cambridge University and Carnegie Mellon are running independent initiatives to develop apps that can listen to users’ coughs and voice to predict whether they are infected with the coronavirus. Both the Cambridge University’s COVID-19 Sounds Project and Carnegie Mellon’s Covid Voice Detector are being built on machine learning and are crowdsourcing the sounds of coughs from the public – both of people who are infected and those who’re not.
The technology is not intended to replace other more thorough medical tests, but as a tool to help differentiate COVID-19 coughs from other illnesses. Many challenges still remain, including collecting diverse and large amounts of data as well as the quality of recordings from people.
