Diabetic Retinopathy Detection and Stage Classification Using Lenet-5Architecture

Diabetic retinopathy diagnosis and treatment are made possible through retinal screening. To process the retinal screening, we are developing a deep learning system to detect and classify diabetic retinopathy.If it is not diagnosed treated it may lead to other serious eye infection like macular edema and glaucoma ultimately causing loss of eye sight. In contrasting to manually created features, we employed the convolutional neural networks (CNN) model i.e Lenet-5 architecture to automatically extract features. Methodologies like edge detection, thresholding and data augmentation are applied to the system during error checking. The efficiency of the system is assessed using statistical metrics like sensitivity (SE), specificity (SP), F-measure, and classification accuracy (ACC).The model achieved an average accuracy of 97%,recall of 22% and f1-score of 65%.