Deep learning‐based detection and stage grading for optimising diagnosis of diabetic retinopathy

Abstract

The herein described automated model employed DR lesions and stage information to identify referable DR and displayed better diagnostic value than models built without this information.