New AI tool can diagnose COVID-19 just by the sound of your coughs


Scientists at the University of Essex have created a COVID-19 test tool that can accurately diagnose the virus by analyzing the sound of a person’s cough.

Researchers say it could be used in a smartphone app to provide a more comfortable form of virus detection than swab testing.

The tool, named DeepCough3D, uses AI to analyze cough audio samples at frequencies humans cannot hear.

Researchers have tested it on more than 8,000 samples from people coughing in hospitals in Spain and Mexico since April 2020. About 2,000 patients tested positive for COVID-19, while the rest had tested negative.

DeepCough3D was found to be 98% accurate in identifying whether samples were positive or negative.

[Read: How do you build a pet-friendly gadget? We asked experts and animal owners]

Lead researcher Dr Javier Andreu-Perez said the tool could ‘be a game changer’ in how we are fighting the pandemic:

It is much less invasive than most other COVID-19 tests and also offers rapid results, paving the way for point-of-need pre-screening testing solutions.

DeepCough3D is one of many attempts to diagnose COVID-19 by listening to a cough. Notably, MIT researchers recently developed an algorithm that successfully detected around 98% of COVID-19 infections by people with COVID-19.

However, the University of Essex team says their research stands out from other studies because it has been shown to be very accurate at detecting infection in thousands of clinically validated samples that have been tested by certified laboratories. .

They say previous studies mostly used crowdsourced samples found online or only a small amount of clinically validated samples.

The researchers also used the tool to classify the cough into three levels of severity, which could help healthcare professionals allocate resources such as ventilators.

They now plan to conduct intervention studies with the technology and work on a broader version and certification of the tool.

You can read the study article in the journal IEEE Transactions on Service Computing.

Published March 10, 2021 – 13:53 UTC


Leave a Reply

Your email address will not be published. Required fields are marked *