Research
Research areas, themes, and current/past funding.
Multi-modal Learning
Developing models that fuse heterogeneous data sources — retinal images, OCT, angiography — to improve diagnostic and predictive accuracy in ophthalmology and beyond.
Longitudinal Deep Learning
Designing temporal and continuous-time neural architectures (LatiM, LMT, L-MAE, Neural ODEs) for disease progression modelling from sequential patient data.
Predictive Medicine
Building end-to-end deep learning pipelines for early prediction of diabetic retinopathy progression and age-related macular degeneration.
Explainable AI for Health
Applying saliency maps, attention visualisation, and XAI frameworks to make clinical deep learning models transparent and trustworthy.
Agentic AI for Health
Exploring LLM-based autonomous agents (LangGraph, AutoGen, CrewAI) orchestrated for clinical decision support and biomedical research automation.
Self-Supervised Learning
Designing pre-training strategies (masked autoencoders, contrastive learning) that learn rich representations from unlabelled medical imaging data, enabling efficient fine-tuning for downstream clinical tasks.
Medical Image Analysis
Processing and analysing ophthalmology images — fundus photographs, OCT volumes, fluorescein angiography — to extract clinically relevant biomarkers for disease monitoring and grading.
Benchmark & Challenges
Organising and contributing to competitive international challenges (MARIO AMD Progression Challenge, MICCAI) to drive reproducible progress in automated retinal disease assessment.
Funding & Grants
Evired Project — MARIO: Monitoring Age-related Macular Degeneration Progression in OCT
Evired · $5,000 (2024)
Funding to design and run the MARIO MICCAI challenge on monitoring AMD progression in OCT imaging.
completedED SVS Doctoral School Conference Grant
ED SVS Doctoral School · $500 (2024)
Grant for conference attendance and research dissemination at international venues.
completedIBSAM Research Grant
IBSAM · $500 (2024)
Grant supporting medical image analysis research and challenge organisation.
completed