Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Comparative Study
. 2025 Aug 1;20(8):e0327743.
doi: 10.1371/journal.pone.0327743. eCollection 2025.

A comparative study of machine learning models for automated detection and classification of retinal diseases in Ghana

Gifty Duah  1 , Eric Nyarko  1 , Anani Lotsi  1
Affiliations
Comparative Study

A comparative study of machine learning models for automated detection and classification of retinal diseases in Ghana

Gifty Duah et al. PLoS One. .

Abstract

Introduction: Retinal diseases, a significant global health concern, often lead to severe vision impairment and blindness, resulting in substantial functional and social limitations. This study explored a novel goal of developing and comparing the performance of multiple state-of-the-art convolutional neural network (CNN) models for the automated detection and classification of retinal diseases using optical coherence tomography (OCT) images.

Method: The study utilized several models, including DenseNet121, ResNet50, Inception V3, MobileNet, and OCT images obtained from the WATBORG Eye Clinic, to detect and classify multiple retinal diseases such as glaucoma, macular edema, posterior vitreous detachment (PVD), and normal eye cases. The preprocessing techniques employed included data augmentation, resizing, and one-hot encoding. We also used the Gaussian Process-based Bayesian Optimization (GPBBO) approach to fine-tune the hyperparameters. Model performance was evaluated using the F1-Score, precision, recall, and area under the curve (AUC).

Result: All the CNN models evaluated in this study demonstrated a strong capability to detect and classify various retinal diseases with high accuracy. MobileNet achieved the highest accuracy at 96% and AUC of 0.975, closely followed by DenseNet121, which had 95% accuracy and an AUC of 0.963. Inception V3 and ResNet50, while not as high in accuracy, showed potential in specific contexts, with 83% and 79% accuracy, respectively.

Conclusion: These results underscore the potential of advanced CNN models for diagnosing retinal diseases. With the exception of ResNet50, the other CNN models displayed accuracy levels that are comparable to other state-of-the-art deep learning models. Notably, MobileNet and DenseNet121 showed considerable promise for use in clinical settings, enabling healthcare practitioners to make rapid and accurate diagnoses of retinal diseases. Future research should focus on expanding datasets, integrating multi-modal data, exploring hybrid models, and validating these models in clinical environments to further enhance their performance and real-world applicability.

PubMed Disclaimer

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1. Sample retinal OCT images.
Fig 2
Fig 2. MobileNet ROC Curve per class.
MobileNet ROC Curve for each class performs exceptionally well. It achieves an AUC of 1.00 for glaucoma, macular edema, and PVD, indicating perfect performance. The model also shows higher performance with an AUC of 0.98 for normal eye conditions, indicating its effectiveness in distinguishing between normal and diseased states
Fig 3
Fig 3. Inception V3 ROC Curve per class.
Inception V3 ROC Curve for each class shows impressive results. It achieves an AUC of 1.00 for glaucoma, macular edema, and PVD, indicating perfect performance. The model also demonstrates high performance with an AUC of 0.97 for normal eye conditions, indicating its strong capability in classifying healthy eyes effectively.
Fig 4
Fig 4. DenseNet121 ROC Curve per class.
DenseNet121 ROC Curve for each class shows impressive results. It achieves an AUC of 1.00 for glaucoma, macular edema, and PVD, indicating perfect performance. The model also performs strongly with an AUC of 0.97 for normal eye conditions.
Fig 5
Fig 5. ResNet50 ROC Curve per class.
ResNet50 ROC Curve for each class achieves an AUC of 1.00 for glaucoma and 0.91 for macular edema, indicating strong performance. However, it struggles to identify normal cases (AUC=0.79) and PVD (AUC=0.82).
Fig 6
Fig 6. Model accuracy vs model loss per Epoch for MobileNet.
Fig 7
Fig 7. Model accuracy vs model loss per Epoch for Inception V3.
Fig 8
Fig 8. Model accuracy vs model loss per Epoch for DenseNet121.
Fig 9
Fig 9. Model accuracy vs model loss per Epoch for ResNet50.
Fig 10
Fig 10. Gradcam and saliency map for MobileNet.
Fig 11
Fig 11. Gradcam and saliency map for MobileNet.
Fig 12
Fig 12. Gradcam and saliency map for DenseNet121.
Fig 13
Fig 13. Gradcam and saliency map for DenseNet121.
Fig 14
Fig 14. Gradcam and saliency map for Inception V3.
Fig 15
Fig 15. Grad-cam and saliency map for Inception V3.
Fig 16
Fig 16. Grad-cam and saliency map for ResNet50.
Fig 17
Fig 17. Grad-cam and saliency map for ResNet50.
See this image and copyright information in PMC

Similar articles

See all similar articles

References

    1. Flaxman SR, Bourne RRA, Resnikoff S, Ackland P, Braithwaite T, Cicinelli MV, et al. Global causes of blindness and distance vision impairment 1990 -2020: a systematic review and meta-analysis. Lancet Glob Health. 2017;5(12):e1221–34. doi: 10.1016/S2214-109X(17)30393-5 - DOI - PubMed
    1. Thapa R, Bajimaya S, Paudyal G, Khanal S, Tan S, Thapa SS, et al. Population awareness of diabetic eye disease and age related macular degeneration in Nepal: the Bhaktapur Retina Study. BMC Ophthalmol. 2015;15:188. doi: 10.1186/s12886-015-0175-z - DOI - PMC - PubMed
    1. Zhou C, Li S, Ye L, Chen C, Liu S, Yang H, et al. Visual impairment and blindness caused by retinal diseases: a nationwide register-based study. J Glob Health. 2023;13:04126. doi: 10.7189/jogh.13.04126 - DOI - PMC - PubMed
    1. Ilesanmi AE, Ilesanmi T, Gbotoso GA. A systematic review of retinal fundus image segmentation and classification methods using convolutional neural networks. Healthc Analyt. 2023;4:100261.
    1. Ansu-Mensah M, Bawontuo V, Kuupiel D, Ginindza TG. Sustainable solutions to barriers of point-of-care diagnostic testing services in health facilities without laboratories in the bono region, Ghana: a qualitative study. BMC Prim Care. 2024;25(1):179. doi: 10.1186/s12875-024-02406-4 - DOI - PMC - PubMed

Publication types