Left image shows a cross section of the eye and the retina. The right image shows an example of lesions (red) detected by an Oak Ridge National Laboratory (ORNL) project. The white lines represent a model of the eye with the optic nerve and blood vessels used as reference points.

Researchers at the Oak Ridge National Laboratory, Oak Ridge, TN, have borrowed a machine-vision technique used to find defects in ICs to seek out retinal diseases in patients. In a test project in Memphis, digital photos of patients’ retinas are taken and relayed to a distant center housing a database containing thousands of images of diseased retinas. The computer automatically compares a patient’s image to the database, quickly determining if there are any similarities. If there are, and the computer calculates a risk, it also provides a follow-up plan in a matter of minutes. If retinal diseases are diagnosed early, treatment can prevent blindness and further damage.

This technology could be widely applied to other diseases of the eye and other organs. But even if only used on retinal diseases, it could make a significant impact. One of the leading causes of retinal diseases is diabetes, a disease that affects 21 million people in the U.S. today and is predicted to affect twice that many by 2050. Globally, over a million people per day could benefit from these automated screenings by 2050.

The engineering team plans to increase the screening system’s level of automation. It would also like to take advantage of the global connectivity offered by the Web.