RoI-level Breast Cancer Tumor Subtyping on BRACS (test)
90.78AUROC (Test)GrapHist
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GrapHistAggregation=add-ABMIL2026.02 | 90.78 | — | |
| GrapHistAggregation=conj-ABMIL2026.02 | 90.04 | — | |
| GrapHistAggregation=ABMIL2026.02 | 89.99 | — | |
| MAEAggregation=add-ABMIL2026.02 | 88.98 | — | |
| MAEAggregation=conj-ABMIL2026.02 | 86.24 | — | |
| MAEAggregation=ABMIL2026.02 | 86.23 | — | |
| DINOv2Aggregation=add-ABMIL2026.02 | 85.65 | — | |
| DINOv2Aggregation=ABMIL2026.02 | 83 | — | |
| DINOv2Aggregation=conj-ABMIL2026.02 | 81.91 | — | |
| ACM-bioAggregation=add-ABMIL2026.02 | 63.08 | — | |
| ACM-UNIAggregation=add-ABMIL2026.02 | 62.77 | — | |
| ACM-bioAggregation=ABMIL2026.02 | 57.13 | — | |
| ACM-bioAggregation=conj-ABMIL2026.02 | 55.42 | — | |
| ACM-UNIAggregation=conj-ABMIL2026.02 | 54.8 | — | |
| ACM-UNIAggregation=ABMIL2026.02 | 47.23 | — | |
| ACM-bioModel Category=Supervised graph models, MIL aggregation=best variant2026.02 | — | 21.55 | |
| ACM-UNIModel Category=Supervised graph models, MIL aggregation=best variant2026.02 | — | 19.95 | |
| DINOv2Model Category=Self-supervised vision models, MIL aggregation=best variant2026.02 | — | 52.68 | |
| GrapHistModel Category=Ours, MIL aggregation=best variant2026.02 | — | 60.3 | |
| MAEModel Category=Self-supervised vision models, MIL aggregation=best variant2026.02 | — | 56.33 |