Bat Optimization Assisted Diabetic Retinopathy Detection Framework

Abstract

Diabetic retinopathy (DR) is a disease that occurs among the persons, who were affected by diabetes for a long time. At such conditions, leakage of protein and fluid from the blood vessels occur. This work intends to establish a new automated DR recognition scheme that includes stages such as ―Feature extraction and Classification‖. At first, feature extraction is performed; where Local Vector Pattern (LVP) and spatial map based edge detection features are extracted. Further, the extracted features are subjected to the classification phase, for which Optimized Deep Convolutional Neural Network (DCNN) is deployed as the classifier. Moreover, to accomplish better accuracy, the weights of CNN are optimally selected by means of the Bat algorithm (BA). Finally, analysis is held to validate the efficacy of the proposed model over other models

References