An ensemble approach to detect exudates in digital fundus images

摘要: Fundus Image analysis is a major concern with respect to various disease detection. Diabetic retinopathy (DR) seen in patients suffering from diabetes mellitus type 2 which leads blindness. images are used identify abnormalities like microaneurysms, haemorrhages, cotton wool spots, exudates, venous beading, and optic disc oedema that cause DR. Automated diagnosis of DR gives first-hand information about the presence, save diabetic vision loss. This paper presents novel ensemble approach automatically detect exudates fundus images. Normal background features removed initially. Morphological operations combined logical has enhanced detection marking exudates. Publicly available standard database DIARETDB1 Forus Health experiment algorithm. 89.6% specificity, 100% sensitivity obtained evaluated logistic regression classifier. Also, 89.13% positive predictive value negative this approach. The AUC ROC plot 0.969.