Detection of Hard Exudates using K-Mean Clustering for Screening of Diabetic Retinopathy

Abstract

Diabetic Retinopathy (DR) is the prolong complication of Diabetes and is widely spread all over the world. Prolongation DR may lead to blindness, hence prevention at early stage of DR should be taken. Exudates are the primary sign of DR. Thus early detection of exudates may help in diagnosis of DR. In this paper, we have used k-mean clustering for detection of hard exudates with the optic disc , followed by mathematical morphological operation to eliminate the optic disc. A set of 30 digital retinal fundus images were used in this experiment. The performance parameter such as sensitivity, specificity and accuracy obtained as 84.7%, 99.62% and 99.57% respectively. The proposed method takes 34 seconds to automatically detect the hard exudates and thus reducing the time consumption of screening process by ophthalmologists.

4 Figures and Tables

Cite this paper

@inproceedings{NGaikwad2013DetectionOH, title={Detection of Hard Exudates using K-Mean Clustering for Screening of Diabetic Retinopathy}, author={Neha N.Gaikwad and V. M. Mane}, year={2013} }