The University of Washington researchers are creating a new AI technology that could potentially be used on smartphones and smart speakers in order to detect when someone is having a heart attack while asleep by just listening out for unique sounds.
People suffering from a cardiac arrest have to go through irregular gasps of breath called agonal breathing. The devices will be trained to detect these specific sounds and also immediately contact emergency services to send for help.
The AI smart speaker is under testing with 236 real-life agonal breathing recordings from emergency calls. These sounds are played back with additional background noise and from various distances to ensure that the technology is able to pick out the specific breathing amongst other sounds, such as snoring.
Up till now, the new technology has accurately detected the agonal breathing 97% of the time by ‘listening’ through a smart device that was placed up to 20ft apart from the speaker giving out the audio.
The team says that the algorithm requires more work in order to prevent any kind of unnecessary calls to the emergency services. “We don’t want to alert either emergency services or loved ones unnecessarily, so it’s important that we reduce our false positive rate,” said researcher Justin Chan.