Artificial intelligence (AI) algorithm can detect COVID-19 cough

Researchers at the Massachusetts Institute of Technology (MIT) say they have developed a method to discover if someone has COVID-19.

The method involves opening an app on your smartphone and simply coughing. You will then be notified immediately if you have COVID-19.

ARTIFICIAL INTELLIGENCE ALGORITHM BREAKTHROUGH

The breakthrough research involved analysing about 70,000 submitted audio samples of people’s coughs. This also included around 2,500 recordings of people confirmed to have the virus at the time.

The research team created an artificial intelligence (AI) algorithm using the data which is capable of telling if someone has the virus from the sound of their cough even if they have no symptoms.

The research team at MIT said: “AI techniques can produce a free, non-invasive, real-time, anytime, instantly distributable, large-scale COVID-19 asymptomatic screening tool to augment current approaches in containing the spread of COVID-19.”

HIGH SUCCESS RATE

The algorithm had a success rate of 98.5% when tests were conducted on a group of people already diagnosed with the virus. The rate increased to 100% when tests were conducted on an asymptomatic group whose coughs were forced.

A MIT news report said the cough of a COVID-infected individual was not the same as a cough by a healthy individual. The difference was so subtle that it would be inaudible to the human ear.

Technology using AI is able to spot the difference and determine if the person has the virus with a high degree of accuracy.

If the US Food and Drug Administration allows MIT to incorporate the technology into a mobile app, the software could play a vital role in helping to quickly identify asymptomatic COVID-infected individuals as well as those with symptoms.

It would mean something as simple as a daily cough into your smartphone could help to slow the spread of the virus. It would allow identified carriers to immediately isolate and also protect others from infection.

Source: thesouthafrican.com

Source