LLMs for Alzheimer's Prediction
LLMs for Alzheimer's Prediction
We present an explainable OCTA-based framework for studying Alzheimer’s-related retinal vascular changes without requiring diagnostic labels. It combines vessel segmentation, vascular biomarker extraction, exploratory subject grouping, and LLM-generated reports to provide transparent, measurement-based interpretations while avoiding diagnostic claims.
Publications:
P. P. Dutta, J. Maliha and M. R. Kabir, “Retinal OCTA Phenotyping with LLM Reporting for Alzheimer’s Disease,” 4th IEEE International Conference on Artificial Intelligence, Blockchain, and Internet of Things (AIBThings), 2026