Open research questions in Retinal Imaging and Analysis

What the literature leaves open

• Future work will focus on patient-level resampling-based validation, external multi-center testing, prospective collection of standardized severity labels, and the inclusion of more diverse keratoconus stages to obtain more robust confidence estimates and further enhance model robustness, particularly for extreme corneal morphologies.

  Multimodal learning for clinically consistent RGP fitting in keratoconus · 2026 · DOI

• The field remains early: most systems are retrospective or proof-of-concept, prospective clinical validation is scarce, and FL does not by itself guarantee privacy, fairness, safety…

  Federated learning for privacy-preserving ophthalmic artificial intelligence: clinical applications and translational challenges · 2026 · DOI

• The rapid advancement of multimodal NLP in ophthalmology is paving the way for transformative clinical applications, addressing Future multimodal NLP systems are expected to evolve from static analysis tools to dynamic, real-time assistants integrated into clinical workflows. Potential applications include intraoperative voice-to-text documentation, where AI converts surgical narration into structured operative notes synchronized with microscope video (109, 110). Another promising avenue is context-aware clinical deci- sion support, which retrieves relevant literature or prior imaging studies by parsing live physician-patient dialogs. Achieving these capabilities will require advancements in speech recognition, low- latency processing, and seamless interoperability with electronic health records. 5.4 Multimodal foundation models for ophthalmology A promising future direction involves the development of oph- thalmology-specific foundation models pretrained on diverse mul- timodal datasets. These models could integrate cross-modal corpora, aligning text-image pairs from published literature, EHRs, and public datasets. Such foundation models could serve as versatile platforms for transfer learning, reducing reliance on task-specific annotations and accelerating deployment across various clinical applications.

  Multimodal natural language processing in ophthalmology: bridging clinical text and medical imaging · 2026 · DOI

• 6.1 Summary of key findings The key findings in this study can be presented as follow: • A model for DR identification called Compact Convolutional Transformer for Automated Diagnosis (CCTAD) is proposed in this study. • An improvement of feature extraction, by integrating convolutional tokenization with self-attention mechanisms, that effectively capturing both local and global image features. • According to obtained results, we can ensure that the proposed CCTAD have achieved high precision, recall, and F1-scores across different severity levels, outperforming conventional CNN-based models in terms of classification accuracy and efficiency. • Additionally, preprocessing techniques such as contrast enhancement and noise reduction significantly contributed and greatly helped to enhance model performance. 6.2 Future research directions To further improve CC TAD’s functionality and applicability, future studies will concentrate on interesting directions as opti- mizing computational efficiency, integrating the system into real-world deployment in healthcare applications. • Model Optimization:Various techniques, like Pruning, quantization, and knowledge distillation, can be used as methods of reducing model complexity and enabling deployment on resource-constrained devices. • Clinical Validation: The generalizability and reliability of CCTAD in real healthcare settings will be evaluated through extensive validation on varied clinical datasets and real-world patient data. • Real-world feature: In real-world screening programs, a significant portion of fundus images may be ungradable due to blur, low illumination, or motion artifacts. Although the datasets used in this study predominantly contained high-quality, expert-verified images, future implementations of the proposed CCTAD system will integrate an auto- mated image-quality assessment module. This component will use sharpness metrics (e.g., variance of Laplacian), contrast evaluation, and deep-learning–based IQA networks to identify and exclude ungradable images before anal- ysis. Incorporating this module will enhance the reliability and safety of automated DR screening in large-scale or teleophthalmology settings. • Deployment in Healthcare Settings:Ensure that AI systems and medical professionals can work together seamlessly by investigating the integration of CCTAD into automated screening systems and telemedicine platforms.

  Artificial intelligence for early detection of diabetic retinopathy: A vision transformer-based approach · 2026 · DOI

• While deep learning (DL) has recently emerged as a transformative approach for automated stratification, a comprehensive synthesis of evidence regarding its diagnostic performance and clinical application remains lacking.

  Diagnostic accuracy and clinical performance of deep learning models for grading diabetic retinopathy: a systematic review and meta-analysis · 2026 · DOI

• Retinopathy of prematurity (ROP) is a preventable cause of childhood blindness, with rising burden in low- and middle-income countries where ROP-trained ophthalmologists are scarce.

  Complementary Roles of Image Classification and Vessel Segmentation in AI-Based Screening for Retinopathy of Prematurity Plus Disease in a Kenyan Preterm Cohort · 2026

• While there has been much research in glaucoma detection, most has relied on discrete CNNs and simplified clustering techniques, leading to inconsistent results due to their limited generalizability across different imaging modes and their inability to preserve the anatomical state of the optic nerve head.

  Development and Comparative Evaluation of Hybrid EfficientNet-CapsNet Architecture for Glaucoma Diagnosis in Resource-Limited Settings · 2026 · DOI

• The limitation of this study is it focuses on fundus images for DR classification; future work will explore the integration of additional imaging modalities such as fundus fluorescein angiography and optical coherence tomography angiography to enhance clinical accuracy, particularly in advanced stages of the disease. Looking ahead, future research could focus on enhancing interpretability through explainable artificial intelligence (AI) techniques and fostering collaboration for data sharing and model validation. The limitation of this study is it focuses on fundus images for DR classification; future work will explore the integration of additional imaging modalities such as fundus fluorescein angiography and optical coherence tomography angiography to enhance clinical accuracy, particularly in advanced stages of the disease. One key limitation is its reliance solely on fundus images, potentially overlooking important information from other diagnostic modalities.

  Enhanced diagnosis of diabetic retinopathy: integrating advanced algorithms for automated detection and classification · 2026 · DOI

• Background Pretrained foundation models are increasingly adopted for diabetic retinopathy (DR) screening, yet it remains unclear how much of their performance derives from the learned representations versus the adaptation procedure.

  How well do frozen foundation models transfer? A calibration-focused benchmark for diabetic retinopathy grading · 2026 · DOI

• However, the association between comprehensive fundus fluorescein angiography (FFA) biomarkers and the total CSVD burden score remains to be fully elucidated.

  Retinal hemodynamic and caliber biomarkers for the assessment of cerebral small vessel disease burden: insights from fundus fluorescein angiography · 2026 · DOI

• Certain limitations should be acknowledged at the outset. The narrative review design itself carries some risk of selection bias. PubMed served as the primary source database, which may have restricted the inclusion of some relevant literature. The field is also evolving rapidly; as a result, conclusions drawn today may shift relatively quickly as newer systems and validation studies emerge.

  Integrating Artificial Intelligence into Eye Care: Diagnostic Performance, Workflow Impact, and Ethical Guardrails (2015–2025) · 2026 · DOI

• Several areas are likely to shape the next phase of ophthalmic AI development: • Federated learning • Prospective clinical trials • • Development of ethical frameworks within education and clinical…

  Integrating Artificial Intelligence into Eye Care: Diagnostic Performance, Workflow Impact, and Ethical Guardrails (2015–2025) · 2026 · DOI

• Multimodal retinal imaging using AI capabilities marks a paradigm change from late-stage disease diagnosis to early risk assessment, real-time surveillance, and large- scale preventive medicine for systemic vascular and neu- rodegenerative disorders. By synergistically combining structural, vascular, and metabolic patterns in the retina with sophisticated learning paradigms, the retina can be leveraged as a feasible and noninvasive biomarker plat- form for cardiovascular, metabolic, and brain health. While existing literature indicates great promise in mul- tiple diseases, further clinical translation in a broader context is still needed to ensure rigorous multicenter val- idation, prospective outcome-driven studies, standard- ized data collection, and effective bias, uncertainty, and regulatory considerations. Future directions will neces- sitate multimodal fusion with clinical, genomic, and longitudinal health information, as well as digital health infrastructure and cost-effectiveness studies to ensure real-world applicability. In the short term, retinal AI is most likely to find applications as a scalable triage and surveillance platform that will complement, rather than substitute for, existing diagnostic capabilities. In the long term, future advances in foundation models, explain- able AI, and federated learning may enable precision screening frameworks that are equitable and effective in improving early intervention, alleviating healthcare bur- den, and advancing preventive neurometabolic medicine worldwide.

  AI-driven multimodal retinal imaging for early detection and risk stratification of vascular and neurodegenerative diseases · 2026 · DOI

• The use of AI-assisted retinal imaging is fraught with potential pitfalls, as the retinal biomarkers are subject to several non-systemic ophthalmic confounders. Changes in OCT parameters that may constitute valid biomarkers include RNFL, GCIPL/GCC, macular thickness, thickness of outer retinal layers, and choroidal thickness. However, any of these changes are not specific to a particular disease. Hence, retinal nerve fiber layer or ganglion cell complex thinning may occur not only due to neurodegeneration but also as a result of glaucoma, optic neuritis, high myopia, aging, or retinal vascular disease. Additionally, measurement results obtained with OCT are vulnerable to segmentation errors caused by the presence of poor signal strength, macular edema, epiretinal membrane, drusen, high myopia, or an unusual retinal contour. This may cause misinterpretation of thickness and subsequent erroneous conclusions regarding the course of a disease. OCTA biomarkers like vessel density, perfusion density, capillary dropout, and foveal avascular zone may also be distorted by numerous imaging artifacts caused by motion, blink, projection, shadowing, defocus, or slab-segmentation. Patient-level confounders may affect AI’s performance by altering image quality, changing magnification, distorting vascular or structural measurements, or introducing systemic bias. The list includes cataracts, corneal opacity, vitreous haze, small pupil size, poor fixation, ocular media opacity, axial length, refractive error, high myopia, and local ocular diseases like diabetic retinopathy, age-related macular degeneration, retinal vein occlusion, glaucoma, or prior ocular surgeries. Thus, in further retinal-AI research (see Table 5), one should consider incorporating image-quality control, artifact detection, segmentation validation, axial length/ ocular refraction normalization, ocular comorbidities registration, standardization of devices/vendors, uncertainty estimates calculation, and human-in-the-loop review [68, 195–197].

  AI-driven multimodal retinal imaging for early detection and risk stratification of vascular and neurodegenerative diseases · 2026 · DOI

• The ophthalmological assessment within the NAKO provides a large, standardized, prospective, population-based resource integrating retinal imaging, visual acuity, and comprehensive systemic health measures. We demonstrated its potential as a modern, accessible platform for methodological development and validation in eye health research. Future analyses including follow-up data will provide the opportunity to investigate longitudinal developments among NAKO participants and use the results of the ophthalmological assessment in NAKO for early detection of ocular or systemic diseases.

  The German National Cohort: Ophthalmological Assessment, Baseline Profile and Potential for AI-based Eye Research · 2026 · DOI

• Key strengths of NAKO are its large scale, nationwide coverage, and longitudinal design. Standardized operating procedures across study centers, together with high-quality nonmydriatic color fundus photography and visual acuity assessment, support robust population-based analyses.

  The German National Cohort: Ophthalmological Assessment, Baseline Profile and Potential for AI-based Eye Research · 2026 · DOI

• The integration of artificial intelligence (AI) with advanced ocular drug delivery systems is expected to significantly transform ophthalmic therapeutics. With increasing computational power and availability of large clinical datasets, AI models are likely to evolve from supportive analytical tools to integrated systems that assist therapeutic optimization. One promising development is the use of AI-driven digital twins for ocular pharmacokinetics. These virtual patient models can simulate processes such as corneal permeability, vitreous diffusion, and drug clearance, enabling researchers to predict therapeutic testing. Such predictive outcomes before clinical modeling may accelerate formulation development and improve dose optimization for both anterior and posterior segment diseases.[30,31] AI is also expected to enhance advanced therapeutic approaches, including gene and cell-based therapies. Intelligent algorithms can assist in optimizing viral vector design, predicting off-target genetic effects, and modeling long-term expression in retinal tissues. In addition, AI-guided optimization of sustained-release implants and nano-ophthalmic carriers may allow longer dosing intervals and improved patient adherence. Another emerging area is the integration of oculomics with AI-driven precision medicine, where retinal imaging data are combined with genomic and clinical information to guide personalized treatment strategies for as glaucoma, diabetic chronic retinopathy, and macular degeneration. Furthermore, AIenabled smart delivery systems, including biosensorbased implants and intelligent contact lenses, may enable real-time monitoring and controlled drug release.[33,34] conditions such transparency, Despite these advancements, challenges such as data standardization, algorithm regulatory approval, and ethical considerations remain important.

  ARTIFICIAL INTELLIGENCE (AI) ENABLED PERSONALIZED MEDICINE - SHAPING THE FUTURE OF OCULAR THERAPEUTICS · 2026 · DOI

• Despite these contributions, our review acknowledges several limitations. First, the significant heterogeneity (I² > 90%) arises from variations in OCT platforms, scan resolutions, preprocessing methods, annotation protocols, and hyperparameter tuning, making direct comparisons of pooled metrics challenging. Second, nearly half of the included studies are retrospective or based on specialized datasets, which raises concerns about the performance of AI in real-world clinical settings. Additionally, reporting gaps, such as inadequate blinding, reliance on single-grader reference standards, and overlapping training/test cohorts, may introduce bias, inflating AI performance estimates. Furthermore, our meta- analyses were limited by the small number of studies providing comparable Dice or ICC statistics, restricting statistical power and subgroup analyses by device type or pathology. Lastly, the potential for publication bias toward positive AI findings should not be overlooked, given the absence of registry data for diagnostic- accuracy studies in ophthalmic imaging. Addressing these limitations will require prospective, multicenter validations with standardized, publicly available benchmarks and transparent reporting guidelines.

  Performance of Artificial Intelligence Systems for Automated Segmentation and Quantification of Retinal Fluid and Pathology in Optical Coherence Tomography Scans: A Systematic Review and Meta-Analysis · 2026 · DOI

• Future research and development opportunities saturate when artificial intelligence (AI) is applied to the diagnosis of ocular disorders. Improving hybrid models which fuse deep learning and conventional machine learning remains a primary goal to achieve improved generalizability and performance. While the two-centre validation provides an impor- tant step toward clinical applicability, further large-scale, multi-centre investigations outside Kerala are still required. Expanding validation to include more diverse popula- tions, imaging devices, and clinical workflows will be crucial to establish EyeDiagNet’s robustness, scalability, and readiness for widespread real-world deployment. Optimized models designed for specific eye diseases, including cataract, diabetic retinopathy (DR) or glaucoma, may lead to more accurate diagnoses [46].

  Evaluating the impact of preprocessing on CNN architectures: comparative analysis of a novel proposed model EyeDiagNet and existing models for eye disease detection · 2026 · DOI

• This article describes a rule-based, AI-inspired identification system to enable screening for childhood glaucoma from retinal images. This system does this through preprocessing, identification of the optic disc and cup, and calculation of the vertical Cup-to- Disc Ratio (CDR) to assess risk categories of Low, Medium, or High. By utilizing 747 retinal images, this study demonstrates that this system can reliably distinguish between all risk categories. By automating CDR measurement and classification, the amount of observer variability is reduced; therefore, speeding up the process of screening. The web-based interface ISSN: 2582-2012 118 Muthulakshmi I., Parameshwari V., Mubitha J., Vigneshwari K. of the system is also capable of real-time analysis with the assistance of webcams or by uploading retinal images. Therefore, it is readily applicable in schools, rural clinics, and tele- healthcare. This system provides a cost-effective and non-invasive tool for early screening of childhood glaucoma and will assist healthcare professionals with identifying children who are at risk of developing childhood glaucoma and will allow for timely interventions to be made. Some improvements that can be made to the system may include improving the accuracy of segmentation through more advanced techniques; enlarging the size of the training data to test on children with an even greater variety of conditions; and including additional retinal characteristics to describe the optic nerve head structure and vascular patterns. Further enhancements such as providing mobile-compatible systems, utilizing the processing power of the cloud for the image processing, and integrating with healthcare record systems will help to add to the scalability of the system allowing for wider usage which will assist in the prevention of loss of vision in children.

  AI Based Early Childhood Glaucoma Risk Screening System Using Eye Image Analysis · 2026 · DOI

• Future work will focus on expanding the suggested model to address the issue with third-party datasets that were TAU-PSO: A Transferable Attention U-Net with Particle Swarm… Informatica 50 (2026) 23–38 37 previously noted.

  TAU-PSO: A Transferable Attention U-Net with Particle Swarm Optimization for Optic Cup-Disc Segmentation in Fundus Images · 2026 · DOI

• Despite the promising performance of the proposed deep learning–based glaucoma detection framework, several limitations need to be acknowledged. One of the primary challenges is dataset bias, which arises due to variations in image acquisition conditions, demographic distribution, and class imbalance across different datasets. Models trained on limited or non-diverse datasets may not settings. generalize well Additionally, the dependency on image quality remains a critical concern, as factors such as poor illumination, noise, blur, and occlusions can significantly affect model accuracy. Another important limitation is the lack of interpretability in deep learning models. Although these models achieve high performance, they often function as “black boxes,” making it difficult for clinicians to fully trust and understand the decision-making process, which is crucial in medical applications. To address these challenges, future research will focus on (XAI) integrating explainable artificial techniques, such as attention maps and visualization methods, to improve transparency and build clinical trust. including Optical Incorporating multimodal data, Page 174 intelligence real-world clinical to A Deep Learning–Driven Framework for Automated Glaucoma Screening and Detection in deployed real-world enhance diagnostic information beyond fundus Coherence Tomography (OCT), visual field tests, and patient clinical history, is another promising direction that accuracy by providing can complementary images. Furthermore, expanding the framework to include largescale and diverse clinical datasets will be essential for improving robustness and generalizability. Future work may also explore advanced data augmentation, domain adaptation, and federated learning approaches to overcome data scarcity and privacy concerns. Ultimately, the goal is to develop a more reliable, interpretable, and clinically that can be applicable glaucoma detection system effectively healthcare environments, including resource-limited settings. 9. CONCLUSION This study presents a comprehensive evaluation of glaucoma detection approaches by integrating traditional image processing techniques with advanced deep learning architectures, particularly the FixEfficientNet framework. The experimental findings indicate that deep learning– based models are highly effective in extracting discriminative features from retinal fundus images, enabling accurate classification of glaucomatous and nonglaucomatous cases. While conventional methods such as Cup-to-Disc Ratio (CDR) estimation provide useful clinical their they are often dependency on precise segmentation and sensitivity to image quality.

  A Deep Learning–Driven Framework for Automated Glaucoma Screening and Detection · 2026 · DOI

• Future work will focus on extending the model for multi-class and multi-modal glaucoma classification, integrating additional clinical data, and exploring more adaptive optimization strategies. Variability in the explanations produced by different Explainable AI (XAI) techniques also presents challenges, as consistency in interpret- ability remains an open issue.

  Explainable AI-driven diagnosis model for early glaucoma detection using grey-wolf optimized extreme learning machine approach · 2026 · DOI

• Our proposed manuscript presents an improved explainability framework called GlaucoXAI (Glaucoma Explainable Artificial Intelligence). GlaucoXAI aims to enhance our understanding of deep learning networks’ behavior by utilizing advanced visualization techniques, such as attention maps. As a post hoc tool, GlaucoXAI can be applied to any existing deep neural models, offering significant insights into their operations. Our two case studies highlight the importance of integrating explainable AI (XAI) techniques in medical image analysis. Additionally, GlaucoXAI facilitates the Extreme Learning Machine (ELM) classifier for glaucoma detection. Our findings underscore the crucial role of XAI in medical imaging tasks, helping to improve the comprehensibility of machine learning models and accelerating their adoption by medical professionals. This paper introduces a sophisticated computer-aided design (CAD) model specifically designed for categorizing glaucoma and healthy images. The model effectively identifies relevant features in fundus images by utilizing a fast discrete curvelet transform with a wrapping (FDCT-WRP) process. To enhance feature reduction, we apply a combination of Principal Compo- nent Analysis (PCA) and Linear Discriminant Analysis (LDA), resulting in a set of reduced and more prominent features. Subsequently, the CAD model employs the IMGWO-ELM, a faster learning algorithm, to train the Single-Layer Feed- forward Network (SLFN). We rigorously evaluate the CAD model’s classification performance across two standard fundus image datasets. The experimental results demonstrate that the proposed CAD model achieves superior classification performance with fewer features compared to existing models. PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1013139 May 04, 2026 30 / 34 In future research, we plan to test the efficacy of our approach for generalization across various imaging modalities. Another potential avenue of exploration is the hybridization of ELM with a less parameter-based optimization algorithm, assessing its effectiveness in multi-class classification tasks. Additionally, the paper suggests considering deep learning algorithms as viable alternatives to the proposed model. Our future research will also focus on quantitatively evaluating XAI methods. This evaluation aims to assess the effectiveness of sensitivity maps generated by these methods and their correlation with deep learning accuracy metrics. We plan to conduct further experiments in the realm of multi-modal glaucoma detection to enhance our understanding. Moreover, we aim to investigate the potential for extracting quantitative features, such as tumor volume and centroid, from these explanation methods.

  Explainable AI-driven diagnosis model for early glaucoma detection using grey-wolf optimized extreme learning machine approach · 2026 · DOI

• challenges. There are concerns regarding bias, reproducibility, and equity because most of the studies rely on datasets which are small or not diverse enough. The unequal distribution of disease prevalence and the underrepresentation of specific ethnic groups intensify these issues. To address these challenges, it is essential to develop larger and more varied datasets, preferably utilizing collaborative methods like Federated Learning to uphold patient privacy. Furthermore, specific techniques like fair identity normalization are being developed to ensure equitable screening performance across all populations. excel Ultimately, each method has its own advantages and disadvantages. at Imaging-focused models assessing structure but often miss functional elements. EHR-driven systems facilitate large-scale prescreening but lack ophthalmic specificity. Sensor-based approaches offer real-time physiological insights but face limitations due to device availability and cost. This underscores the necessity for multimodal integration, where structural, functional, and systemic data are merged to better represent actual clinical practice. Such approaches hold the greatest promise for enhancing diagnostic precision, predicting disease progression, and fostering the clinical acceptance of AI-driven glaucoma systems.

  Artificial Intelligence in Opthamology:A study on different AIML approaches for Glaucoma prediction · 2026 · DOI

Most-cited papers in Retinal Imaging and Analysis

• A deep learning system for predicting time to progression of diabetic retinopathy · Nature Medicine · 2024 · 283 citations
• OCTA-500: A retinal dataset for optical coherence tomography angiography study · Medical Image Analysis · 2024 · 156 citations
• OCTDL: Optical Coherence Tomography Dataset for Image-Based Deep Learning Methods · Scientific Data · 2024 · 125 citations
• Improved Support Vector Machine based on CNN-SVD for vision-threatening diabetic retinopathy detection and classification · PLoS ONE · 2024 · 123 citations
• Automated Tool Support for Glaucoma Identification With Explainability Using Fundus Images · IEEE Access · 2024 · 102 citations
• Autonomous artificial intelligence increases screening and follow-up for diabetic retinopathy in youth: the ACCESS randomized control trial · Nature Communications · 2024 · 100 citations
• Retinal photograph-based deep learning predicts biological age, and stratifies morbidity and mortality risk · Age and Ageing · 2022 · 97 citations
• Ensemble deep learning and EfficientNet for accurate diagnosis of diabetic retinopathy · Scientific Reports · 2024 · 91 citations
• NIMEQ-SACNet: A novel self-attention precision medicine model for vision-threatening diabetic retinopathy using image data · Computers in Biology and Medicine · 2024 · 89 citations
• Retinal age gap as a predictive biomarker of future risk of Parkinson’s disease · Age and Ageing · 2022 · 85 citations

Most recent work

• Artificial intelligence for detection of age-related macular degeneration based on fundus images: A systematic review · Survey of Ophthalmology · 2026
• A multi-instance learning framework combining CNN and transformer for noninvasive coronary heart disease diagnosis using scleral images · Journal of Innovative Optical Health Sciences · 2026
• FastSCVM: A lightweight hybrid Mamba–convolution network for efficient glaucoma image segmentation · Biomedical Signal Processing and Control · 2026
• Construction and Interpretability of a Multimodal Deep Learning Model of Electronystagmography-Optical Coherence Tomography Angiography for Early Screening of Alzheimer’s Disease · American Journal of Alzheimer s Disease & Other Dementias® · 2026
• An intelligent segmentation method for leakage points in central serous chorioretinopathy based on fluorescein angiography images · International Journal of Ophthalmology · 2026
• Evaluating the impact of preprocessing on CNN architectures: comparative analysis of a novel proposed model EyeDiagNet and existing models for eye disease detection · Discover Artificial Intelligence · 2026
• The Early Diagnosis of Retinal Diseases for Cataracts and Glaucoma Detection Using a Deep Learning-based Approach. · Zanin Journal of Science and Engineering · 2026
• Diagnostic Performance of Artificial Intelligence and Deep Learning for Diabetic Retinopathy Screening: A Systematic Review and Meta-analysis · Journal of Optometry and Ophthalmology Research · 2026
• FIG-GAN: Fundus Image Generation via Deep Learning Based Generative Adversarial Network for AMD Disease Diagnosis · International Journal of Computational Intelligence Systems · 2026
• Deep Learning-Based Classification of Eye Diseases Using Convolutional Neural Network for OCT Images · International Journal for Research in Applied Science and Engineering Technology · 2026

Related topics in Medicine