Curated Medical Datasets
- Harvard-FairVision30k: Largest 2D (SLO fundus) + 3D (OCT Bscans) medical imaging fairness dataset (30,000 patients), involving glaucoma, age-related macular degeneration (AMD) and diabeitic diabetic retinopathy (DR) diseases. Patient demographic information include gender, race, ethnicity, preferred lanague, and marital status. GitHub Download the Data
- FairSeg: A large-scale medical image segmentation dataset (10,000 patients) for fairness learning. Patient demographic information include gender, race, ethnicity, preferred lanague, and marital status. GitHub Download the Data
- FairCLIP: A large-scale vision-language medical dataset (10,000 patients) for fairness learning. Patient demographic information include gender, race, ethnicity, preferred lanague, and marital status. GitHub Download the Data
- FairDomain: A large-scale cross-domain medical imaging dataset (20,000 patients) for domain adaptation and fairness learning, including image classification and segmentation tasks. Patient demographic information include gender, race, ethnicity, preferred lanague, and marital status. GitHub Download the Data
Medical AI Tools
- EyeLearn: An AI-driven tool for artifact correction in the retinal nerve fiber layer thickness map. GitHub
Related Paper: Artifact-tolerant clustering-guided contrastive embedding learning for ophthalmic images in glaucoma
M Shi, A Lokhande, M Fazli, V Sharma, Y Tian, Y Luo, L Pasquale, T Elze, M Boland, N Zebardast, D Friedman, L Shen, M Wang. IEEE Journal of Biomedical and Health Informatics. 2023. [link]
- VFTransformer: An AI-driven tool for predicting 10-2 visual field map from 24-2 visual field data. GitHub
Related Paper: Transformer-based Deep Learning Prediction of 10-Degree Humphrey Visual Field Tests from 24-Degree Data
M Shi, A Lokhande, Y Tian, Y Luo, M Eslami, S Kazeminasab, T Elze, L Shen, L Pasquale, S Wellik, C Moraes, J Myers, N Zebardast, D Friedman, M Boland, M Wang. Translational Vision Science & Technology. 2024. [link]