Application of Large-Scale AI Models in Diabetic Retinopathy

ABSTRACT

Diabetic Retinopathy (DR) is one of the most common and severe complications of diabetes, ranking among the leading causes of blindness worldwide. With the increasing prevalence of diabetes, the demand for early diagnosis and effective management of DR has grown significantly. In recent years, artificial intelligence (AI) tools based on large deep learning models have been trained on vast amounts of data, enabling them to efficiently extract pathological features from fundus images. These AI-driven technologies play a crucial role in the early detection, accurate diagnosis, treatment monitoring, and patient management of DR. This article systematically reviews the research achievements of large AI models in multiple aspects of DR management, including diagnosis, treatment, and prognosis evaluation. The aim is to provide comprehensive scientific guidance for clinical practice.