AI Successfully Detects More Than 50 Different Eye Diseases


DeepMind AI, in cooperation with Moorfields Eye Hospital made an AI that can identify 50 eye problems as good as the best eye specialist can do. / Photo by: osde8info via Flickr


Computer programs developer DeepMind, in collaboration with the University College London and the Moorfields Eye Hospital NHS Foundation Trust, has recently developed an artificial intelligence that correctly identifies more than 50 different eye diseases from patients. This ability and accuracy is as good as that of world-leading eye specialists, reports The Guardian.

The AI system takes a more intelligent and human-like approach in analyzing the complex optical coherence tomography scans of retinas. OCT is an imaging technique that utilizes coherent light to capture the micrometer-resolution images from within the optical scattering. The OCT scans are usually used for triage patients who have sight problems that fall into four clinical categories: observation only, routine, semi-urgent, and urgent.

A total of five separate machine-learning systems were trained to use clinical OCT scans and then create maps out of it. The maps are analyzed and interpreted for a referral decision. Moorfields Eye Hospital’s consultant ophthalmologist Dr. Pearse Keane said that today, using the technology, they were able to perform much faster eye scanning than what human experts can achieve.

“The AI technology we’re developing is designed to prioritize patients who need to be seen and treated urgently by a doctor or eye care professional,” added Dr. Keane.

Moorfields Eye Charity’s Chief Executive Robert Dufton added that the need for treatment for eye diseases is expected to grow based on a forecast. He also pointed out that with AI, there’s a potential to further improve the speed of treating and diagnosing eye diseases.