Adebisi Abraham Owodunni* and Yetunde Mary Balogun

Computer Science (Artificial Intelligence and Data Science), University of Hull, England, United Kingdom

*Corresponding Author: Adebisi Abraham Owodunni, Computer Science (Artificial Intelligence and Data Science), University of Hull, England, United Kingdom.

Received: February 19, 2025; Published: August 29, 2025

Abstract

To avoid irreparable vision loss, it is imperative to diagnose eye diseases early. Ophthalmologists typically manually screen the images of the diseases. An increase in the number of patients and a shortage of skilled ophthalmologists are detrimental to the treatment of patients. In this study, the classification of the images into cataract, glaucoma, normal, and diabetic retinopathy-infected eyes is the main objective. In a number of issues including disease classification, the Convolutional Neural Network – CNN model has been shown to be successful. The aim of this paper is to use CNN and transfer learning (MobileNetV2 and ResNet50) models to classify images of eye diseases into four categories: normal, cataract, glaucoma, and diabetic retinopathy through Data and image pre-processing and data augmentation, CNN will be used to extract features from the dataset, using CNN and pre-trained models for the classification of eye diseases, comparison of CNN and pre-trained models and with models used for multiclass classification by some other authors. The suggested CNN models combined with Transfer Learning models (names written above) extract a variety of distinct properties from the images. The dataset for this research was gotten from Kaggle. The performance of the models improved after introducing several hyper-tuning approaches on the parameters and the highest accuracy gotten for the CNN model was 88% while 94% and 93% were gotten for ResNet50 and MobileNetV2 respectively.

Keywords:Eye Disease; Blind; ResNet50; MobileNetV2

References

1. Abràmoff MD., et al. “Improved Automated Detection of Diabetic Retinopathy on a Publicly Available Dataset Through Integration of Deep Learning”. Investigative Ophthalmology and Visual Science13 (2016): 5200.
2. Burton MJ., et al. “The Lancet Global Health Commission on Global Eye Health: vision beyond 2020”. The Lancet Global Health4 (2021): e489-e551.
3. GRUEN R. “To Err Is Human”. Eds. L. T. Kohn, J. M. Corrigan and M. S. Donaldson. National Academy Press Washington, D.C (2000): 287.
4. Gulshan V., et al. “Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs”. JAMA22 (2016): 2402.
5. Hanaa Mohsin Ahmed and Hameed SR. “Eye Diseases Classification Using Hierarchical MultiLabel Artificial Neural Network”. (2020).
6. Hanung Adi Nugroho., et al. “Analysis of Texture-based Features for Image Classification of Retinal Exudates”. (2017).
7. Kabari LG and Nwachukwu EO. “Neural Networks and Decision Trees For Eye Diseases Diagnosis”. InTech eBooks (2012).
8. Kanagasingam Y., et al. “Evaluation of Artificial Intelligence–Based Grading of Diabetic Retinopathy in Primary Care”. JAMA Network Open5 (2018): e182665.
9. Louis P., et al. “Global initiative for the elimination of avoidable blindness”. (2000).
10. Malik S., et al. “Data Driven Approach for Eye Disease Classification with Machine Learning”. Applied Sciences14 (2019): 2789.
11. “Glaucoma” (2017).
12. Pascolini D and Mariotti SP. “Global estimates of visual impairment: 2010”. The British Journal of Ophthalmology5 (2012): 614–618.
13. Sahana M and Gowrishankar SN. “Identification and Classification of Cataract Stages in Old Age People Using Deep Learning Algorithm”. 8.10 (2019): 2767-2772.
14. Zhang Z., et al. “A survey on computer aided diagnosis for ocular diseases”. BMC Medical Informatics and Decision Making 1 (2014).
15. Ruoan X., et al. “Deep Learning for Eye Disease Classification: Diabetic Retinopathy, Macular Disease, and Glaucoma”. Journal of Medical Imaging4 (2022): 123-135.

Citation

Citation: Adebisi Abraham Owodunni and Yetunde Mary Balogun. “Eye Disease Classification using Integrated Deep Learning Approaches - CNN and Transfer Learning (ResNet50 and MobileNetV2)".Acta Scientific Computer Sciences 7.6 (2025): 41-53.

Copyright

Copyright: © 2025 Adebisi Abraham Owodunni and Yetunde Mary Balogun. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.