Feelix uses AI to cancel out external noise to allow clinicians to more accurately diagnose lung abnormalities
AI digital stethoscope eliminates noise for more accurate diagnosis of lung ailments
December 14, 2020
by John R. Fischer
, Senior Reporter
A new AI-powered digital stethoscope with noise-cancelling abilities may enable more accurate diagnoses of lung abnormalities, including COVID-19 — even by people with no training.
The FDA-approved solution is known as Feelix and suppresses noise automatically to enable users to hear clear body sounds. It was designed by Sonavi Labs and James West, a professor of electrical engineering and mechanical engineering at Johns Hopkins University.
“You can speak or cough into this stethoscope and it will tell you immediately if the individual is contaminated,” West told HCB News.
With the device, healthcare workers in highly populated areas or noisy clinics are expected to be better able to distinguish sounds that indicate lung ailments, while maintaining social distancing. They also can share recordings virtually with care teams in different rooms to help reduce exposure.
Feelix has an algorithm that can detect pneumonia in infants with accuracy comparable to physicians. The company is also looking at another algorithm for assessing COPD and other respiratory diseases. Testing showed physicians to be in favor of the device more than 95% of the time over traditional techniques, according to West.
In addition, the stethoscope is easier, faster, non-disposable and noninvasive, compared to current COVID-19 testing, which West describes as "lengthy, wet and often inaccurate."
He adds that "we need a better screening method. We need an instant detection of the virus in people and if this holds up, I think it goes a long way to meeting that goal."
The researchers plan to continue refining the stethoscope before distributing it for use.
The solution is currently available in a research capacity, with a commercial release expected in Q2 2021.