Title: Diabetic retinopathy using image processing and deep learning

Authors: Debabrata Swain; Sanket Sanjay Bijawe; Prasanna Prasad Akolkar; Aditya Shinde; Mihir Vijay Mahajani

Addresses: Department of Information Technology, Vishwakarma Institute of Technology, Pune, 411037, India ' Department of Information Technology, Vishwakarma Institute of Technology, Pune, 411037, India ' Department of Information Technology, Vishwakarma Institute of Technology, Pune, 411037, India ' Department of Information Technology, Vishwakarma Institute of Technology, Pune, 411037, India ' Department of Information Technology, Vishwakarma Institute of Technology, Pune, 411037, India

Abstract: Diabetic retinopathy is one of the most non-uniform and confront regions to diagnose as it is exceptionally perplexing. In the circle of retinopathy, the number of times intensive assessments are required to be done to determine upon the diabetes mellitus or blindness that patient might be facing. Various professionals may take different amount of time to recognise diabetic retinopathy. So, a framework is required that can effectively and precisely analyse the retinal conditions with no of such limitations. This paper presents a two- stage method to effectively predict the level grading of diabetic retinopathy. The first stage involves preprocessing the retinal image and reducing the noise from an image. The second stage involves building a convolutional neural network architecture for predicting diabetic retinopathy level. It is a hurdle of diabetes that can affect the retinal nervous and lead to total or partial loss of vision.

Keywords: CNN; convolutional neural network; diabetic retinopathy; disease diagnoses; deep learning; machine intervention; image processing; artificial neural networks; Gaussian blur filter; Gabor filter; K-means clustering; ReLu; Softmax.

DOI: 10.1504/IJCSM.2021.120686

International Journal of Computing Science and Mathematics, 2021 Vol.14 No.4, pp.397 - 409

Received: 08 May 2020
Accepted: 06 Jul 2020

Published online: 03 Feb 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article