Examinando por Autor "El-Baz, Ayman"

Mostrando 1 - 2 de 2
  • Miniatura
    Ítem
    Explainable AI-based approach for Age-related Macular Degeneration (AMD) detection via fundus imaging
    (Institute of Electrical and Electronics Engineers Inc., 2025) Osa Sánchez, Ainhoa; Magdy Balaha, Hossam; Mahmoud, Ali; Sewelam, Ashraf; Ghazal, Mohammed; García-Zapirain, Begoña; El-Baz, Ayman
    Age-related macular degeneration (AMD) is a leading cause of vision loss in older people and is characterized by subtle retinal changes that make early identification difficult. Previous studies have demonstrated the efficacy of Vision Transformers (ViTs) in classifying medical images by successfully detecting retinal disorders such as AMD. This paper addresses multiple shortcomings in conventional AMD diagnostic techniques by exploring the detection and explanation of various AMD subtypes from numerical features extracted with a ViT model from fundus images through cascaded artificial intelligence (AI) models using transformers, convolutional neural networks (CNNs), and multilayer perceptrons (MLPs). The data were preprocessed to recognize intricate disease-related patterns. The best test results using the cascade method for each model type show that the MLP model achieved an accuracy of 91.86% (with a sensitivity of 92.22% and a specificity of 95.74%). The Transformer model achieved its highest accuracy of 83.72% (with a sensitivity of 83.86% and a specificity of 89.74%). The CNN model demonstrated the best performance, with an accuracy of 94.19% (with a sensitivity of 93.84% and a specificity of 96.00%). This work helps clinicians interpret AMD cases and supports decision-making revealing hidden features of AMD that are not visible to the human eye. Future research will focus on improving these systems by expanding the databases in aggregate and incorporating multimodal data.
  • Miniatura
    Ítem
    Explainable multimodal foundational models for retinal disease stratification: a robustness study across 15+ heterogeneous datasets
    (Institute of Electrical and Electronics Engineers Inc., 2026-02-25) Osa Sánchez, Ainhoa ; El-Baz, Ayman; Oleagordia Ruiz, Ibon ; García-Zapirain, Begoña
    The automated stratification of retinal diseases remains a significant challenge due to data heterogeneity and the closed-box nature of deep learning models. Although foundational models have demonstrated remarkable success in general computer vision, their clinical reliability and interpretability in multimodal ophthalmology remain insufficiently explored. In this work, we introduce an Explainable Multimodal Foundational AI framework trained on a large-scale integrated corpus of 760,243 retinal images collected from over 15 heterogeneous repositories, encompassing both fundus photography and optical coherence tomography (OCT). We systematically evaluate self-supervised learning (SSL) paradigms DINO and iBOT across convolutional (ResNet) and Transformer-based (Vision Transformer, ViT) architectures. Our results show that ResNet-DINO achieves state-of-the-art performance, reaching 93.53% accuracy and a 0.935 F1-score in 6-class multimodal retinal disease classification, while exhibiting superior robustness under data-limited conditions, attributed to its inductive bias. Notably, we observe emergent clinical localization capabilities in Vision Transformer models (ViT-DINOv2 and ViT-iBOT). Using frozen pre-trained weights and without exposure to expert-labeled data or ground truth labels, these models autonomously highlight clinically relevant biomarkers, including subretinal fluid and drusen, demonstrating intrinsic pathological awareness. By bridging the semantic gap between unsupervised representation learning and targeted clinical diagnosis, this study establishes a benchmark for robust, explainable, and label-efficient AI in ophthalmology. Our findings indicate that large-scale foundational pre-training not only enhances diagnostic accuracy but also induces meaningful visual priors aligned with established clinical biomarkers, supporting the deployment of trustworthy AI systems in real-world clinical decision support.