One of the biggest causes of permanent blindness in the world today is glaucoma, and effective treatment depends on early detection. Although artificial intelligence (AI) has demonstrated potential in automating glaucoma screening, there is still a significant obstacle in transferring research datasets to actual clinical settings. When applied to clinical data, current models show an 8% - 18% performance loss, which is mostly caused by out-of-distribution samples, demographic bias, and poor imaging quality. We provide a clinically-translatable glaucoma detection paradigm that makes use of multi-modal fusion, uncertainty quantification, and the EyePACS-AIROGS-light-V2 dataset in order to overcome these difficulties. Our method, called H-UQ-MFF (Hybrid Uncertainty-Aware Multi-Feature Fusion), combines structural and texture characteristics from optic disc analysis with deep features from ResNet50 while dynamically weighting contributions according to prediction uncertainty. With an AUC of 0.9969, sensitivity of 0.9811, and specificity of 0.9717, internal validation outperforms ResNet50, EfficientNet-B0, and Deep Ensemble baselines. Generalizability is confirmed by external validation on REFUGE and PAPILA datasets, where H-UQ-MFF outperforms cutting-edge models and lowers calibration error. Beyond technical performance, the framework ensures clinical safety and regulatory preparedness by incorporating drift tracking techniques, bias analysis, and FDA-compliant evaluation processes. This work bridges the gap between research innovation and clinical deployment by establishing a repeatable baseline for AI translation in ophthalmology.
References
[1]
Roy, A., etal. (2025) DeepEyeNet: Adaptive Genetic Bayesian Algorithm Based Hybrid ConvNeXtTiny Framework for Multi-Feature Glaucoma Eye Diagnosis. arXiv: 2501.11168.
[2]
Ming, H.K., Tze, V.W., Shiong, B.L.C. and Then, P.H.H. (2025) Evaluation of Post-Hoc Explainability Methods for Glaucoma Classification Using Fundus Images. In: Arai, K., Ed., Intelligent Computing, Springer, 650-667. https://doi.org/10.1007/978-3-031-92605-1_40
[3]
Kumar, S., Dixit, V. and Gupta, M. (2025) Deep Learning Approaches for Retinal Disease Diagnosis: Insights from Fundus and OCT Analysis. EPJWebofConferences, 328, Article ID: 01041. https://doi.org/10.1051/epjconf/202532801041
Feng, Y., Wu, C. and Zhou, Y. (2024) DG2Net: A MLP-Based Dynamixing Gate and Depthwise Group Norm Network for Classification of Glaucoma. In: Antonacopoulos, A., Chaudhuri, S., Chellappa, R., Liu, C.L., Bhattacharya, S. and Pal, U., Eds., Pattern Recognition, Springer, 295-308. https://doi.org/10.1007/978-3-031-78383-8_20
[6]
Wang, H. (2025) Regulating AI Medical Devices: A Clinically Anchored, Adaptive Approach. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5575550
[7]
Roy, M. (2025) Effect of Screen Time on Glaucoma. https://openreview.net/pdf?id=BuK3fZNS9U
[8]
Nguyen, D.M., etal. (2025) Deep Learning for Ophthalmology: The State-Of-The-Art and Future Trends. arXiv: 2501.04073.
[9]
Alvi, M.D.M., Ubaid, D., Ul Wara, K., Zaidi, F.R.S., Anwar, S.M., Javed, A., et al. (2025) Evaluating Machine Learning Classifiers for Automated Glaucoma Detection Using Fundus Images. Biomedical Materials & Devices. https://doi.org/10.1007/s44174-025-00569-x
[10]
Luo, X., etal. (2025) A Survey of Multimodal Ophthalmic Diagnostics: From Task-Specific Approaches to Foundational Models. arXiv: 2508.03734.
[11]
Baggu, A., Polisetti, V.A. and Nidamanuri, J. (2024) A Novel Attention Informed Voting Ensemble Framework for Retinal Disease Classification. 2024 International Conference on Emerging Techniques in Computational Intelligence (ICETCI), Hyderabad, 22-24 August 2024, 230-237. https://doi.org/10.1109/icetci62771.2024.10704188
[12]
Remyes, D., Nasef, D., Remyes, S., Tawfellos, J., Sher, M., Nasef, D., etal. (2025) Clinical Applicability and Cross-Dataset Validation of Machine Learning Models for Binary Glaucoma Detection. Information, 16, Article 432. https://doi.org/10.3390/info16060432
[13]
de Sousa, J.G., Brito, L.E.S., Dalília, A., Veloso, R.R., de Carvalho Filho, A.O. and de Araújo, F.H.D. (2025) GlaucoAware: Desenvolvimento de uma ferramenta de classificação de glaucoma em imagens do fundo ocular utilizando CNNs. Revista de Sistemas e Computação-RSC, 15, 55-62.
[14]
Das, P., Nirmala, S.R. and Medhi, J.P. (2015) Diagnosis of Glaucoma Using CDR and NRR Area in Retina Images. NetworkModelingAnalysisinHealthInformaticsandBioinformatics, 5, Article No. 3. https://doi.org/10.1007/s13721-015-0110-5
[15]
MacCormick, I.J.C., Williams, B.M., Zheng, Y., Li, K., Al-Bander, B., Czanner, S., etal. (2019) Accurate, Fast, Data Efficient and Interpretable Glaucoma Diagnosis with Automated Spatial Analysis of the Whole Cup to Disc Profile. PLOS ONE, 14, e0209409. https://doi.org/10.1371/journal.pone.0209409
[16]
Law, S.K., Kornmann, H.L., Nilforushan, N., Moghimi, S. and Caprioli, J. (2016) Evaluation of the “IS” Rule to Differentiate Glaucomatous Eyes from Normal. JournalofGlaucoma, 25, 27-32. https://doi.org/10.1097/ijg.0000000000000072
[17]
Das, P., Nirmala, S.R. and Medhi, J.P. (2016) Detection of Glaucoma Using Neuroretinal Rim Information. 2016 International Conference on Accessibility to Digital World (ICADW), Guwahati, 16-18 December 2016, 181-186. https://doi.org/10.1109/icadw.2016.7942538
[18]
MacCormick, I.J.C., et al. (2019) Accurate, Fast, Data Efficient and Interpretable Glaucoma Diagnosis with Automated Spatial Analysis of the Whole Cup to Disc Profile. PLOS ONE, 14, e0209409.
[19]
Neto, A., Camara, J. and Cunha, A. (2022) Evaluations of Deep Learning Approaches for Glaucoma Screening Using Retinal Images from Mobile Device. Sensors, 22, Article 1449. https://doi.org/10.3390/s22041449
[20]
Choudhary, K. and Tiwari, S. (2015) ANN Glaucoma Detection Using Cup-to-Disk Ratio and Neuroretinal Rim. InternationalJournalofComputerApplications, 111, 8-14. https://doi.org/10.5120/19581-1233
[21]
Kiefer, R. (2024) Glaucoma Dataset: EyePACS-AIROGS-Light-V2. Kaggle.
[22]
Valdez-Rodríguez, J.E., Felipe-Riverón, E.M. and Calvo, H. (2021) Optic Disc Preprocessing for Reliable Glaucoma Detection in Small Datasets. Mathematics, 9, Article 2237. https://doi.org/10.3390/math9182237
[23]
Veena, H.N., Muruganandham, A. and Senthil Kumaran, T. (2022) A Novel Optic Disc and Optic Cup Segmentation Technique to Diagnose Glaucoma Using Deep Learning Convolutional Neural Network over Retinal Fundus Images. JournalofKingSaudUniversity—ComputerandInformationSciences, 34, 6187-6198. https://doi.org/10.1016/j.jksuci.2021.02.003
[24]
Kumar, S. and Kumar, B. (2022) Automatic Early Glaucoma Detection by Extracting Parapapillary Atrophy and Optic Disc from Fundus Image Using SVM. MultimediaToolsandApplications, 81, 13513-13535. https://doi.org/10.1007/s11042-021-11023-7
[25]
Elmannai, H., Hamdi, M., Meshoul, S., Alhussan, A.A., Ayadi, M. and Ksibi, A. (2024) An Improved Deep Learning Framework for Automated Optic Disc Localization and Glaucoma Detection. ComputerModelinginEngineering&Sciences, 140, 1429-1457. https://doi.org/10.32604/cmes.2024.048557
[26]
Joshi, S., Partibane, B., Hatamleh, W.A., Tarazi, H., Yadav, C.S. and Krah, D. (2022) Glaucoma Detection Using Image Processing and Supervised Learning for Classification. JournalofHealthcareEngineering, 2022, Article ID: 2988262. https://doi.org/10.1155/2022/2988262
[27]
Virbukaitė, S. and Bernatavičienė, J. (2023) Impact of Eye Fundus Image Preprocessing on Key Objects Segmentation for Glaucoma Identification. NonlinearAnalysis: ModellingandControl, 29, 96-110. https://doi.org/10.15388/namc.2024.29.33669
[28]
Khalil, T., Akram, M.U., Raja, H., Jameel, A. and Basit, I. (2018) Detection of Glaucoma Using Cup to Disc Ratio from Spectral Domain Optical Coherence Tomography Images. IEEEAccess, 6, 4560-4576. https://doi.org/10.1109/access.2018.2791427
[29]
Mayya, V., S, S.K., Kulkarni, U., Surya, D.K. and Acharya, U.R. (2022) An Empirical Study of Preprocessing Techniques with Convolutional Neural Networks for Accurate Detection of Chronic Ocular Diseases Using Fundus Images. AppliedIntelligence, 53, 1548-1566. https://doi.org/10.1007/s10489-022-03490-8
[30]
Zhu, Y., Salowe, R., Chow, C., Li, S., Bastani, O. and O’Brien, J.M. (2024) Advancing Glaucoma Care: Integrating Artificial Intelligence in Diagnosis, Management, and Progression Detection. Bioengineering, 11, Article 122. https://doi.org/10.3390/bioengineering11020122
[31]
Goldmann, N., Skalicky, S.E., Weinreb, R.N., Paletta Guedes, R.A., Baudouin, C., Zhang, X., etal. (2023) Defining Functional Requirements for a Patient-Centric Computerized Glaucoma Treatment and Care Ecosystem. JournalofMedicalArtificialIntelligence, 6, Article 3. https://doi.org/10.21037/jmai-22-33
[32]
Wong, K. (2025) Innovative Vision Glasses for Glaucoma Detection and Management. In: Rath, M. and Samal, T., Eds., KeyIssuesinNetworkProtocolsandSecurity, IntechOpen. https://doi.org/10.5772/intechopen.1006973
[33]
Zeppieri, M., Gagliano, C., Tognetto, D., Musa, M., Avitabile, A., D’Esposito, F., etal. (2025) Augmented Decisions: AI-Enhanced Accuracy in Glaucoma Diagnosis and Treatment. JournalofClinicalMedicine, 14, Article 6519. https://doi.org/10.3390/jcm14186519
[34]
Tappeiner, C. (2025) Artificial Intelligence in Ophthalmology: Acceptance, Clinical Integration, and Educational Needs in Switzerland. JournalofClinicalMedicine, 14, Article 6307. https://doi.org/10.3390/jcm14176307
[35]
Pathmakumara, H.C. and Perera, G. (2025) Explainable Deep Learning for Glaucoma Detection: A DenseNet121-Based Classification with Grad-CAM Visualization. medRxiv. https://doi.org/10.1101/2025.10.08.25337634
[36]
Santone, A., Cesarelli, M., Colasuonno, E., Bevilacqua, V. and Mercaldo, F. (2024) A Method for Ocular Disease Diagnosis through Visual Prediction Explainability. Electronics, 13, Article 2706. https://doi.org/10.3390/electronics13142706
[37]
Ennab, M. and Mcheick, H. (2025) Advancing AI Interpretability in Medical Imaging: A Comparative Analysis of Pixel-Level Interpretability and Grad-Cam Models. MachineLearningandKnowledgeExtraction, 7, Article 12. https://doi.org/10.3390/make7010012
[38]
Lemij, H.G., de Vente, C., Sánchez, C., Cuadros, J., Jaccard, N. and Vermeer, K. (2022) Glaucomatous Features in Fundus Photographs of Eyes with ‘Referable Glaucoma’ of a Large Population Based Labeled Data Set for Training an Artificial Intelligence (AI) Algorithm for Glaucoma Screening. InvestigativeOphthalmology&VisualScience, 63, 2041-A0482.
[39]
Farhad, M.A., Razaque, A., Mukhanov, S.B., Hassan, D.S.M. and Mohan Rai, H. (2025) Enhanced Lesion Localization and Classification in Ocular Tumor Detection Using Grad-Cam and Transfer Learning. IEEEAccess, 13, 167762-167777. https://doi.org/10.1109/access.2025.3610183
[40]
Scientific, L.L. (2024) Enhancing Glaucoma Diagnosis: Deep Learning Models for Automated Identification and Explainability Using Fundus IMAGES. Journal of TheoreticalandAppliedInformationTechnology, 102, 5346-5363.
[41]
Gandhi, V.C., Gandhi, P., Ogundiran, J.O., Tshibola, M.S.S. and Kapuya Bulaba Nyembwe, J. (2025) Computational Modeling and Optimization of Deep Learning for Multi-Modal Glaucoma Diagnosis. AppliedMath, 5, Article 82. https://doi.org/10.3390/appliedmath5030082
[42]
Faria, F.T.J., Moin, M.B., Debnath, P., Fahim, A.I. and Shah, F.M. (2024) Explainable Convolutional Neural Networks for Retinal Fundus Classification and Cutting-Edge Segmentation Models for Retinal Blood Vessels from Fundus Images. arXiv: 2405.07338.
[43]
Eladawi, N., Sabry, D. and Salaheldin, H. (2025) Enhancing Glaucoma Detection Using Convolutional Neural Networks: A Comparative Study of Multi-Class and Binary Classification Approaches. Alfarama Journal of Basic & Applied Sciences, 6, 75-95.
[44]
Morales, P.H. (2023) Deep Transfer Learning Strategy to Diagnose Eye-Related Conditions and Diseases: An Approach Based on Low-Quality Fundus Images. IEEE Access, 11, 37403-37411.
[45]
Lemij, H.G., Vente, C.D., Sánchez, C.I. and Vermeer, K.A. (2023) Characteristics of a Large, Labeled Data Set for the Training of Artificial Intelligence for Glaucoma Screening with Fundus Photographs. OphthalmologyScience, 3, Article ID: 100300. https://doi.org/10.1016/j.xops.2023.100300
[46]
Tajerian, A., Keshtkar, M., Almasi-Hashiani, A. and Tajerian, M. (2023) Enhancing Eye Diseases Diagnosis through Transfer Learning: Study of Deep Convolutional Neural Networks for Accurate Classification of Glaucoma and Diabetic Retinopathy from Healthy Eye Using Fundus Images. https://www.researchsquare.com/article/rs-3120228/v1
[47]
Chen, R., etal. (2024) EyeDiff: Text-To-Image Diffusion Model Improves Rare Eye Disease Diagnosis. arXiv: 2411.10004.
[48]
Tohye, T.G., Qin, Z., Al-antari, M.A., Ukwuoma, C.C., Lonseko, Z.M. and Gu, Y.H. (2024) Ca-Vit: Contour-Guided and Augmented Vision Transformers to Enhance Glaucoma Classification Using Fundus Images. Bioengineering, 11, Article 887. https://doi.org/10.3390/bioengineering11090887
[49]
Li, Y., Carrillo-Perez, F., Al Owaifeer, A.M., Alawad, M. and Gevaert, O. (2025) Enhancing Glaucoma Detection through Supervised Pre-Training with Intermediate Phenotypes: A Multi-Institutional Study. medRxiv. https://doi.org/10.1101/2025.04.22.25326210
