tqhuyen/harvard-oct-glaucoma-200
Harvard-GF 200^3 consolidated dataset Canonical raw 200^3 Harvard-GF OCT volumes as consolidated .npy (uint8), split-matched with the 96^3/128^3 convenience copies. Source: harvardairobotics/Harvard-GF (per-scan .npz[oct_bscans]), IEEE TMI 2024 (Luo et al.) store_shape: [1, 200, 200, 200] uint8 | source_shape: [200, 200, 200] | antialias: n/a (raw) Splits (volumes, pos=glaucoma): Training: 2100 (pos 1083 / neg 1017) Validation: 300 (pos 176 / neg 124) Test: 900 (pos 489 / neg… See the full description on the dataset page: https://huggingface.co/datasets/tqhuyen/harvard-oct-glaucoma-200.
Harvard-GF 200^3 consolidated dataset
Canonical raw 200^3 Harvard-GF OCT volumes as consolidated .npy (uint8), split-matched with the 96^3/128^3 convenience copies.
- Source:
harvardairobotics/Harvard-GF(per-scan.npz[oct_bscans]), IEEE TMI 2024 (Luo et al.) - storeshape: `[1, 200, 200, 200]` uint8 | sourceshape:
[200, 200, 200]| antialias: n/a (raw) - Splits (volumes, pos=glaucoma):
- Training: 2100 (pos 1083 / neg 1017)
- Validation: 300 (pos 176 / neg 124)
- Test: 900 (pos 489 / neg 411)
- Layout:
{Training,Validation,Test}_{volumes,labels}.npy(volumes(N,1,200,200,200)uint8, labels(N,)int64) - Downstream normalization:
x / 255.0.
This is the canonical resolution; 96^3 and 128^3 are resized copies for fast sweeps.