Extraction of Retinal Blood Vessels and Diagnosis of Proliferative Diabetic Retinopathy Using Extreme Learning Machine

In medical field, wide varieties of applications can be dealt using Image processing. Detection and screening of retinal diseases is one such application in image processing. Diabetic retinopathy is a complication of diabetes. The disease affects blood vessels inside the retina. Hence, in this work we have proposed a new method for extracting the retinal blood vessels from the color fundus images based on feature classification, to reduce the ophthalmologists' time for examining the retinal images. The blood vessels are extracted from the color fundus image by applying the preprocessing methods and segmentation techniques using matched filter and modified local entropy thresholding operation. The method was evaluated on the publicly available DRIVE and DIARETDB0 databases. In this paper, a recently developed machine learning algorithm called Extreme Learning Machine (ELM) is used for diagnosing the Proliferative Diabetic Retinopathy. Performance of ELM is compared in terms of classification accuracy with Support Vector Machine (SVM) Classifier. It is observed that ELM leads to 97.5% average accuracy in comparison with SVM classifier (87.5%) for DIARETDB0 images.