Abstract:
Diabetes mellitus is an enduring disease related with significant morbidity and mortality. The main pathogenesis event behind disease is its numerous micro and microvascular complications. In developing countries, diabetic retinopathy is one of the prominent source of vision impairment in working age population. Diabetic retinopathy has been classified into two categories; proliferative diabetic retinopathy (PDR) and non-proliferative diabetic retinopathy (NPDR). Non-proliferative diabetic retinopathy is further classified as mild, moderate, and severe, while Proliferative diabetic retinopathy further classified as mild, moderate and advance diabetic eye disease. Diabetic Retinopathy is a disease caused due to high blood glucose level which result in vision loss or permanent blindness. High level advancement in the field of Bio-medical image processing speed up the automated process of disease diagnoses and analysis. There are number of researches conducted and computerized systems being designed to detect and analyze retinal diseases through image processing. Similarly there are number of algorithms designed to detect and grade Diabetic Retinopathy by analyzing different symptoms including Microaneurysms, soft exudates, hard exudates, cotton wool spots, Fibrotic bands, NVDs, NVEs, hemorrhages tractional bands etc. The primary objective of this research is to diagnose the disease and classify its stages which facilitate the Ophthalmologist in screening and detection of diabetic retinopathy disease in Diabetic patients. The visual examination of retina is vital test in order to diagnose DR related complications. However, all the diabetic retinopathy computer aided diagnoses systems requires a standard dataset for the estimation of their efficiency, performance and accuracy. The secondary objective of this research is to develop a benchmark for the evaluation of computer- based DR diagnostic systems. Existing Diabetic Retinopathy benchmarks are small in size and do not cover all the DR stages and categories. The dataset contains 1447 high quality fundus photographs of retinal images, acquired in two years from the patient record presented to Eye department Holy Family Hospital, Rawalpindi. This benchmark provides the evaluation platform for the medical image analysis researchers. Furthermore, it provides evaluation data for all the stages of Diabetic Retinopathy.
