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
. 2025 Aug 13;25(16):5019.
doi: 10.3390/s25165019.

RetinoDeep: Leveraging Deep Learning Models for Advanced Retinopathy Diagnostics

Sachin Kansal  1 , Bajrangi Kumar Mishra  2 , Saniya Sethi  3 , Kanika Vinayak  3 , Priya Kansal  1 , Jyotindra Narayan  4
Affiliations

RetinoDeep: Leveraging Deep Learning Models for Advanced Retinopathy Diagnostics

Sachin Kansal et al. Sensors (Basel). .

Abstract

Diabetic retinopathy (DR), a leading cause of vision loss worldwide, poses a critical challenge to healthcare systems due to its silent progression and the reliance on labor-intensive, subjective manual screening by ophthalmologists, especially amid a global shortage of eye care specialists. Addressing the pressing need for scalable, objective, and interpretable diagnostic tools, this work introduces RetinoDeep-deep learning frameworks integrating hybrid architectures and explainable AI to enhance the automated detection and classification of DR across seven severity levels. Specifically, we propose four novel models: an EfficientNetB0 combined with an SPCL transformer for robust global feature extraction; a ResNet50 ensembled with Bi-LSTM to synergize spatial and sequential learning; a Bi-LSTM optimized through genetic algorithms for hyperparameter tuning; and a Bi-LSTM with SHAP explainability to enhance model transparency and clinical trustworthiness. The models were trained and evaluated on a curated dataset of 757 retinal fundus images, augmented to improve generalization, and benchmarked against state-of-the-art baselines (including EfficientNetB0, Hybrid Bi-LSTM with EfficientNetB0, Hybrid Bi-GRU with EfficientNetB0, ResNet with filter enhancements, Bi-LSTM optimized using Random Search Algorithm (RSA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), and a standard Convolutional Neural Network (CNN)), using metrics such as accuracy, F1-score, and precision. Notably, the Bi-LSTM with Particle Swarm Optimization (PSO) outperformed other configurations, achieving superior stability and generalization, while SHAP visualizations confirmed alignment between learned features and key retinal biomarkers, reinforcing the system's interpretability. By combining cutting-edge neural architectures, advanced optimization, and explainable AI, this work sets a new standard for DR screening systems, promising not only improved diagnostic performance but also potential integration into real-world clinical workflows.

Keywords: EfficientNetB0; SHAP explainability; SPCL transformer; ant colony optimization; bidirectional LSTM; data augmentation; diabetic retinopathy; particle swarm optimization.

PubMed Disclaimer

Conflict of interest statement

The authors do not have any conflicts of interest. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Figures

Figure 1
Figure 1
Fundus Images in seven classes of the dataset class (a–g).
Figure 2
Figure 2
Bi-LSTM with SHAP explainability architecture.
Figure 3
Figure 3
EfficientNetB0 with SPCL transformer architecture.
Figure 4
Figure 4
Bi-LSTM with genetic algorithm hyperparameter optimization architecture.
Figure 5
Figure 5
ResNet50 ensembled with Bi-LSTM architecture.
Figure 6
Figure 6
Model architecture.
Figure 7
Figure 7
Accuracy over epochs and box plots of accuracy/loss for regular models.
Figure 7
Figure 7
Accuracy over epochs and box plots of accuracy/loss for regular models.
Figure 7
Figure 7
Accuracy over epochs and box plots of accuracy/loss for regular models.
Figure 8
Figure 8
Accuracy over epochs and box plots of accuracy/loss for proposed models.
Figure 8
Figure 8
Accuracy over epochs and box plots of accuracy/loss for proposed models.
Figure 9
Figure 9
SHAP explanability feature graphs.
Figure 10
Figure 10
SHAP heatmaps.
Figure 11
Figure 11
Comparison of the highest accuracy between models.
See this image and copyright information in PMC

References

    1. Grzybowski A., Singhanetr P., Nanegrungsunk O., Ruamviboonsuk P. Artificial Intelligence for Diabetic Retinopathy Screening Using Color Retinal Photographs: From Development to Deployment. Ophthalmol. Ther. 2023;12:1419–1437. doi: 10.1007/s40123-023-00691-3. - DOI - PMC - PubMed
    1. Venkatesh R., Gandhi P., Choudhary A., Kathare R., Chhablani J., Prabhu V., Bavaskar S., Hande P., Shetty R., Reddy N.G., et al. Evaluation of Systemic Risk Factors in Patients with Diabetes Mellitus for Detecting Diabetic Retinopathy with Random Forest Classification Model. Diagnostics. 2024;14:1765. doi: 10.3390/diagnostics14161765. - DOI - PMC - PubMed
    1. Abushawish I.Y., Modak S., Abdel-Raheem E., Mahmoud S.A., Hussain A.J. Deep Learning in Automatic Diabetic Retinopathy Detection and Grading Systems: A Comprehensive survey and comparison of methods. IEEE Access. 2024;12:84785–84802. doi: 10.1109/ACCESS.2024.3415617. - DOI
    1. Paranjpe M.J., Kakatkar M.N. Review of methods for diabetic retinopathy detection and severity classification. Int. J. Res. Eng. Technol. 2014;3:619–624. doi: 10.15623/ijret.2014.0303115. - DOI
    1. van der Heijden A.A., Nijpels G., Badloe F., Lovejoy H.L., Peelen L.M., Feenstra T.L., Moons K.G., Slieker R.C., Herings R.M., Elders P.J., et al. Prediction models for development of retinopathy in people with type 2 diabetes: Systematic review and external validation in a Dutch primary care setting. Diabetologia. 2020;63:1110–1119. doi: 10.1007/s00125-020-05134-3. - DOI - PMC - PubMed

LinkOut - more resources