Detection of Diabetic Retinopathy Using Convolutional Neural Network

EasyChair Preprint 10127

Abstract

Images of the retina taken by a fundus camera are used to diagnose diabetic retinopathy which requires experienced optometrist to recognize the level of severity and significant features to reduce the time consumption and difficulty using complex grading. We suggest a convolutional neural network  architecture in this paper to diagnose the diabetic retinopathy and accurately classify its  severity by data augmentation which can recognize the characteristics like micro-aneurysm, hard exudates and haemorrhages. We train the data which is available on kaggle. We have a data set of 2755 images which is used  in our proposed method to achieve an accuracy of 91.67% .

Keyphrases: Convolutional Neural Networks, Diabetic Retinopathy, deep learning, image classification

@booklet{EasyChair:10127,
  author    = {Mukesh Raj and Pallavi Singh and Kunwar Randhir Singh and Nishant Malik},
  title     = {Detection of Diabetic Retinopathy Using Convolutional Neural Network},
  howpublished = {EasyChair Preprint 10127},
  year      = {EasyChair, 2023}}