Automated Extraction of Retinal Blood Vessels for Early Diagnosis of Diabetic Retinopathy Using Enhancement Filters and Adaboost

Abstract

This paper proposes an automated extraction and segmentation of blood vessels to diagnose the symptoms of diabetic retinopathy at early stage. The proposed method has been evaluated on two commonly used, publicly available benchmark datasets DRIVE, and STARE. The results show that the proposed method attains the best trade-off results in terms of effectiveness and efficiency and performs comparably better than state-of-the-art methods. The average accuracy on DRIVE, and STARE database turns out to be 0.973, and 0.952, respectively.