Detection of hemorrhages in diabetic retinopathy analysis using color fundus images

Diabetic retinopathy (DR), a complication due to diabetes, is the leading cause of blindness. Diagnosis of DR at its earliest is thus very necessary. In this paper, we present an automated computer aided method for detection of hemorrhages using fundus images. In the first stage, blood vessel (BV) is detected and eliminated. In the following stage hemorrhage candidate is detected. The method is applied on HRF, DIARETDB0, DIARETDB1, MESSIDOR and Local databases which have both normal and pathological images. The proposed method shows an overall sensitivity of 97.3% and specificity of 98.92% in hemorrhage detection.