Chinese researchers develop AI technology for screening diabetic retinopathy

CHANGSHA -- A research team with Hunan University has successfully developed an artificial intelligence (AI) algorithm that can swiftly detect diabetic retinopathy.

The algorithm enables doctors to find signs of diabetic retinopathy from pictures with the help of a supercomputer center.

"After analyzing more than 80,000 retinal images, we found that the AI algorithm had a 91 percent accuracy rate of detecting diabetic retinopathy, which is on par with an ophthalmologist," said Xi Ziwei, a member of the research team.

"AI detection takes only one thirtieth of the time used by an ophthalmologist, enabling it to screen more cases and reduce human error," Xi said, adding that ophthalmologists are in short supply in China, particularly at the community level.

Early detection of diabetic retinopathy improves outcomes in a disease that is a major cause of vision loss.