Communications of the ACM
Computer Spots Rare Diseases in Family Photos
After looking at photos of U.S. president Abraham Lincoln, the software ranked Marfan syndrome a disorder resulting in unusually large features, which some believe he had as the seventh most likely diagnosis out of 91 syndromes.
Credit: Alexander Gardner/Hulton Archive/Getty
University of Oxford researchers have developed software that can learn to identify rare medical conditions by analyzing a face from a digital photograph. The researchers say the software also should be able to identify unknown genetic disorders if groups of photos in the database share specific facial features.
"The idea is to offer it to health systems right across the world because all you need is a computer and a digital photo," says University of Oxford researcher Christoffer Nellaker.
The researchers developed the software by feeding a computer-vision algorithm 1,363 publicly available pictures of people with eight genetic disorders, including Down's syndrome, fragile X syndrome, and progeria. The program learned to identify each condition from a pattern of 36 facial features in each photo, such as the shapes of the eyes, brows, lips, and noses. "It automatically analyzes the picture and annotates key feature points, producing from that a description of the face which expands the features that are important for distinctiveness," Nellaker says. The features then are compared with those from patients with confirmed disorders, enabling the software to suggest patterns in new patients.
During testing, the software proved to be 93-percent correct in predicting disorders based on photographs of the patient.
From New Scientist
View Full Article
Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA
No entries found