Comparative Analysis of University of Auckland Diabetic Retinopathy Database

Diabetic retinopathy is a micro-vascular disease that affects the vision. People with diabetes are more likely to get affected by diabetic retinopathy. There are many screening techniques available to diagnose this pathology. Majority of this screening methods use the retinal image of the patient for the diagnosis. Recently many automatic retinal image analysis methods are being developed to automate the diagnosis of this pathology, thereby eliminating the need for a specialist ophthalmologist for such a screening. In order to evaluate the performance of any such automatic method, there is a need for a database of retinal images which could act as a reference for bench-marking the performance of the methods developed. In this paper, we discuss a new retinal image database known as University of Auckland Diabetic Retinopathy database (UoA-DR), prepared specifically for bench-marking the performance of various automatic systems developed for the diagnosis of diabetic retinopathy. This database consists of 200 retinal images mostly affected with diabetic retinopathy. This database also contains the manually segmented versions of various features obtained from these images which could act as a ground truth for algorithm's improvement to segment various features from retinal images automatically.