Detection of hemorrhages for diagnosis of malarial retinopathy

Malarial retinopathy is an eye disease caused due to malaria and retinal changes like vessels discoloration, retinal whitening and hemorrhages happen in this disease. Hemorrhages are red lesions, which are the result of blood leakage and are signs of severe malaria. In this paper color feature based automatic hemorrhage detection method is proposed which comprise of preprocessing, illumination equalization, candidate hemorrhages detection, post processing, feature extraction and classification. The evaluation of proposed system is done on locally gathered retinal images of patients suffering from malarial retinopathy with manually labeled hemorrhages. Results show the significance of proposed system.