AI Reads Faces to Predict Cancer Patient Survival, Trained on Nearly 35,000 Photos

Mass General Brigham trained the FAHR-FaceSurvival model on nearly 35,000 photos of cancer patients, and it focuses on the area under the eyes and the bridge of the nose. Co-author Raymond Mak links these features to Hippocrates, who described them as signs of wasting and a risk of death as early as 400 BCE.