Improved analysis of Diabetic Maculopathy using level set spatial fuzzy clustering

Abstract

Patients suffering from Diabetic Retinopathy are at a high risk of sight threatening disease, the Diabetic Maculopathy. It gets initiated with the deposition of lesions formed from blood constituents, in a region of one optic disc diameter centered at fovea of retina. The effect becomes vision threatening when the deposition of the lesions spread close to fovea. These lesions are of two types, namely bright lesions such as soft and hard Exudates and dark lesions including Microaneurysms and Hemorrhages. The detection of the lesions become difficult when they overlap or lie close to each other. In this paper, we have presented a novel method for improving the detection of bright and dark lesions in positive Diabetic Maculopathy images. The algorithm consists of two stages. Initially the fovea is detected and the region for analysis of Maculopathy is marked. Secondly, level set spatial fuzzy clustering is performed over the region to enhance the detection of lesions and hence analysis of the disease. The performance evaluation of the proposed method is carried out by comparing the result with manually segmented ground truth images, obtained with the help of ophthalmologists. The results show improvement of the analysis as compared to present methodologies.