Researchers at MIT in the US have created an AI model that can identify people with COVID-19 with 98.5% accuracy based on their cough.
As countries around the world struggle with intermittant lockdowns and implement various track-and-trace programmes to help slow the spread of COVID-19, researchers have been investigating alternative ways of detecting the virus in the general public. Thermal cameras have been used to help detect anyone with a fever, for example, but that's not the only telltale sign of infection that we can watch for.
Researchers at Massachusetts Institute of Technology (MIT) in the US have now developed an AI that can detect people infected with COVID-19 by another of its primary symptoms: The cough. The virus causes a persistent cough and there is evidence that it can change the way you make sound even if you don't show symptoms.
RELATED: NI firm creates 15-minute rapid Covid-19 test
The new system used artificial intelligence to analyse the sounds of thousands of people infected with the virus and compare them to healthy breathing and coughing sounds. The final model achieved a 98.5% accuracy rating in differentiating the sound of those with COVID-19 from healthy people. A similar project in Cambridge in the UK reported an 80% accuracy rating with a lower sample number.
Many businesses have already installed heat cameras at their entrances to detect people with fevers, so this new technology could be used in a similar manner. It can detect audio cues of an infected person much more accurately than a human would be able to, and in cases could be able to detect asymptomatic individuals.
Source: BBC News