k-NN Classification on Synthetic (test)
71.8AccuracyΦeat
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ΦeatSupervision Signal=Self-supervised2025.11 | 71.8 | 75 | 66.6 | |
| DINOv3Supervision Signal=Self-supervised2025.11 | 66.9 | 60 | 55.4 | |
| DINOv2Supervision Signal=Self-supervised2025.11 | 64.3 | 56.9 | 53.3 | |
| PE SpatialSupervision Signal=Agglomerative2025.11 | 50.3 | 44 | 39 | |
| PE-coreSupervision Signal=Weakly supervised2025.11 | 44.2 | 37 | 33.9 | |
| CLIPSupervision Signal=Weakly supervised2025.11 | 39.9 | 36.6 | 31.2 | |
| RADIOv3Supervision Signal=Agglomerative2025.11 | 35.9 | 32.9 | 28.8 | |
| SigLIP 2Supervision Signal=Weakly supervised2025.11 | 33.5 | 30.8 | 25.1 |