Authors:
Muhammad Simul Hasan Talukder, Ajay Kirshno Sarkar, Sharmin Akter, Muhammad Nuhi-Alamin, Rejwan Bin Sulaiman

Addresses:
1Department of Atomic Energy Regulatory, Bangladesh Atomic Energy Regulatory Authority, Bangladesh. 2,4Department of Electrical and Electronic Engineering, Rajshahi University of Engineering and Technology, Rajshahi, Bangladesh. 3Department of Biomedical Engineering, Jashore University of Science and Technology, Jashore, Bangladesh. 5Department of Artificial Intelligence and Cybersecurity, Northumbria University, North East of England, United Kingdom.  simulhasantalukder@gmail.com1, aksarkar@eee.ruet.ac.bd2, sharmintalukder120@gmail.com3, nuhialamin@eee.ruet.ac.bd4, Rejwan.sulaiman@northumbria.ac.uk5

Abstract:

Diabetic Retinopathy (DR) is an ocular condition caused by a sustained high blood sugar level, which causes the retinal capillaries to block and bleed, causing retinal tissue damage. It usually results in blindness. Early detection can help in lowering the risk of DR and its severity. The robust and accurate prediction and detection of diabetic retinopathy is challenging. This paper develops a machine-learning model for detecting Diabetic Retinopathy that is entirely accurate. Pre-trained models such as ResNet50, InceptionV3, Xception, DenseNet121, VGG19, NASNetMobile, MobileNetV2, DensNet169, and DenseNet201 with pooling layer, dense layer, and appropriate dropout layer at the bottom of them were carried out in transfer learning (TL) approach. Data augmentation and regularization were performed to reduce overfitting. Transfer Learning model of DenseNet121, Average and weighted ensemble of DenseNet169 and DenseNet201 TL architectures contribute the highest accuracy of 100%, the highest precision, recall, F-1 score of 100%, 100%, and 100% individually.

Keywords: Diabetic Retinopathy; Transfer Learning; Ensemble Learning; Augmentation of Retinal Images; Introduction Regularization; Convolution Neural Network; Diabetes Mellitus; Diabetic Image of Fundus.

Received on: 27/12/2022, Revised on: 01/03/2023, Accepted on: 04/06/2023, Published on: 02/12/2023

FMDB Transactions on Sustainable Health Science Letters, 2023 Vol. 1 No. 2, Pages: 92-106

• Views : 178
• Downloads : 5

Download PDF