Severity Grading of Diabetic Retinopathy Using Extreme Learning Machine

A computationally effective method for diagnosing the severity of Diabetic Retinopathy is proposed. The proposed approach follows 3 stages; the preprocessing, feature extraction and classification. The purpose of the first step is to make the image suitable for subsequent process. In feature extraction, number of micro aneurysms and texture features were extracted by morphological process and the last step is the classification, in which the images are categorized by these features with the help of ELM classifier. The above procedure was tested and analyzed using images in DIARETDB0 and DRIVE database and we were able to achieve accuracy of 95% with good training speed.