Harvard-GF (Harvard Glaucoma Fairness) is a retinal nerve disease dataset for fairness learning in glaucoma detection, featuring both 2D and 3D OCT imaging data with balanced racial groups. It contains 3,300 samples from 3,300 patients with equal representation across Asian, Black, and White racial groups — a unique design addressing the doubled glaucoma prevalence observed in Black patients compared to other races.
This dataset… See the full description on the dataset page: https://huggingface.co/datasets/harvardairobotics/Harvard-GF.
