In Belarus, scientists at The United Institute of Informatics Problems of the National Academy of Sciences of Belarus are developing a software package based on neural networks for the early diagnosis of diabetic retinopathy, a dangerous complication of diabetes mellitus that leads to blindness. The work is being carried out under the guidance of Valery Starovoitov, Doctor of Technical Sciences, Professor, and winner of the State Prize of the Republic of Belarus.

The goal of the development is the automated detection of microaneurysms (key signs of early-stage retinopathy) in digital images of the retina obtained using fundus cameras and optical coherence tomography (OCT).

At the current stage, the system is already capable of recognizing microaneurysms. Plans include integrating data from digital images from fundus cameras and optical coherence tomography of the eye.

The software package under development does not replace the doctor, but serves as a support tool: it speeds up analysis, reduces workload, and minimizes the risk of human error.

In the future, developers and doctors plan to expand the system's telemedicine functions: to enable remote transmission of retinal images, their automated analysis using AI, and the organization of remote consultations, including to support doctors in rural areas.

This development has high potential for exporting technology to countries with high morbidity and a shortage of ophthalmologists (India, Pakistan, Vietnam).