University of Washington Researchers Create First Contactless Tool to Detect Cardiac Arrest Using AI
Researchers at the University of Washington developed an AI system that monitors people for cardiac arrest while they are sleeping. The algorithm detects gasping sounds most commonly heard during a cardiac arrest, and is the first of its kind that does not require touch. It was designed to ultimately function as an app or skill for smart speakers and smart phones.
The research team used agonal breathing recorded from real 911 calls to accurately depict the noises. Calls were captured over multiple smart devices including Amazon Alexa, iPhone 5S and Samsung Galaxy S4. According to UW, they utilized machine learning to create a tool that accurately identifies agonal breathing 97% of the time when a smart device is placed up to 6 meters or about 20 feet away from the user.
Each year 475,000 Americans die from cardiac arrest. Over 25% of cardiac arrests happen outside of the hospital and often inside someone’s home, particularly in the bedroom. “A lot of people have smart speakers in their homes, and these devices have amazing capabilities that we can take advantage of,” said Shyam Gollakota, an associate professor in the UW’s School of Computer Science & Engineering.
Research was funded by the National Science Foundation and the technology is planned to go to market through a UW spinout called Sound Life Sciences.
Monitoring Health Via Voice
In an effort to prevent heart attacks and strokes Omron Healthcare released an Alexa skill last year that provides blood pressure monitoring. Users can download the Omron Connect app to check their latest readings and set reminders to regularly track their blood pressure, receiving recommendations on how to correctly measure themselves.
Heart rate monitoring is common in conjunction with fitness apps such as Google Fit where they aim to “coach you to a healthier and more active lifestyle.” The health-tracking platform partnered with the American Heart Association (AHA) to create physical activity goals, like brisk walking 30 minutes a day, to reduce the risk of heart disease. Users can ask Google Assistant for insights like their heart rate or total steps.