Accessibility links

Breaking News

Smartphone stroke detection breakthrough announced by Australian team


PhD scholar Guilherme Camargo de Oliveira demonstrates the face screening tool with Visiting Associate Professor Nemuel Daniel Pah, left, from RMIT University. (Credit: Seamus Daniel, RMIT University)
PhD scholar Guilherme Camargo de Oliveira demonstrates the face screening tool with Visiting Associate Professor Nemuel Daniel Pah, left, from RMIT University. (Credit: Seamus Daniel, RMIT University)

A new technology that allows smartphones to identify strokes far quicker than existing methods has been developed by researchers in Australia.

The new technology uses artificial intelligence as it scans a patient’s face for symmetry and certain muscle movements, which are called action units. People who have suffered a stroke often have one side of their face looking different from the other.

The biomedical engineers at Melbourne’s RMIT University say the smartphone technology can detect facial asymmetry, potentially identifying strokes within seconds - much sooner and more precisely than current technologies.

Professor Dinesh Kumar, who led the research team, explained to Australian Broadcasting Corp. how the AI-driven device works.

“It takes a video of a person who is doing a smile, and the model determines whether this particular smile is indicative of (a) person who has had a stroke," Kumar said. "We then inform the paramedic or the clinician who is aware of the very high risk of this person having a stroke and, thus, can be treated immediately.”

Strokes affect millions of people around the world. They occur when the supply of blood to part of the brain is interrupted or reduced, which stops brain tissue from receiving oxygen and nutrients. Experts say that if treatment is delayed by even a few minutes, the brain can suffer permanent damage.

Symptoms of stroke include confusion, speech impairments and reduced facial expressions.

The RMIT team reports that the smartphone tool has an accuracy rating of 82% for detecting stroke. They stress that it would not replace comprehensive medical diagnostic tests for stroke, but instead would guide initial treatment by first responders by quickly identifying patients who need urgent care.

The Australian study, which was a collaboration with São Paulo State University in Brazil, is published in the journal, Computer Methods and Programs in Biomedicine.

XS
SM
MD
LG