SEVERITY LEVEL CLASSIFIERS FOR DIABETIC DISEASE CLASSIFICATION USING DEEP LEARNING AND FUZZY DECISION MAPPING

Abstract

In diagnosing and treating Diabetic Retinopathy, segmenting and classifying retinal images is a difficult job. Fundus Retinal Imaging is used to diagnose diabetics and it provides additional detail for obtaining Retinal Image sequences. The proposed study uses a Fractional Jaya Optimizer-based Deep Convolutional Neural Network (FJO-DCNN) to classify blood vessels in retinal images segmented by the optic disc. The segments are created using a clustering mechanism known as Particle Swarm Optimization (PSO), which ensures the efficacy of optimum segment placement, allowing for more precise detection of the optic disc. Finally, using this hybrid algorithm, the intensity degree is determined, and a better output score is obtained, despite the Fuzzy judgement mapping. The proposed study computed the best values for accuracy, sensitivity, and specificity in FJO-DCNN for blood vessel classification.