Google's AI System Detects Heart Disease
Brain and Heart AI Google's Brain Team has developed a new AI system that accurately predicts heart disease by scanning images of people's retina. The researchers believe this may lead to the discovery of more ways to diagnose health issues from retinal images.
Deep Learning Algorithms
They used deep learning algorithms trained on data from more than 284,000 patients. The AI system is able to predict cardiovascular disease risk factors from retinal images with very high accuracy. The accuracy rate to actually pick out patients in danger of a heart attack or stroke is 70%.
Eye and Heart Connections
According to lead researcher Lily Peng, their approach is to use deep learning to draw connections between changes in human anatomy and disease. In this case, using the eyes to reflect what's going on with the heart.