Vessel enhancement algorithm in digital retinal fundus microaneurysms filter for nonproliferative diabetic retinopathy classification

Abstract

Diabetic Retinopathy is one of a complication of diabetes which can cause blindness. There are two levels of diabetic retinopathy which are nonproliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR). The presence of microaneurysms in the eye is one of the early signs of diabetic retinopathy. Automated detection of microaneurysms in digital color fundus photographs is developed to help opthalmologist to detect the emergence of its early symptoms and determine the next action step for the patient. In this study we developed microaneurysms filter algorithm with the concept of vessel enhancement in order to extract the structure of microaneurysms in medical imagery, especially in the retinal image. Microaneurysms filter algorithm has been able to extract part of microaneurysms in the retinal image and detect pretty well. In addition, microaneurysms filters can also be developed to detect the next symptoms of diabetic retinopathy-hemorraghes (bleeding) that has features in common with microaneurysms which are larger and have irregular shape. Furthermore, the nonproliferative diabetic retinopathy (NPDR) classification system was developed based on the microaneurysms and hemorraghes detection. The system was tested against the patient data. Microanuerisyms detection results on DiaRetDB data has sensitivity value by 57,4% to ophthalmlogist 1, 35,37% for opthalmologist 2 and 62,68% for opthalmologist 3.