Semantic Segmentation on CHAOS
97DiceUniversal Model
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Universal Model# pre-trained data=2.1k, Evaluation Protocol=Linear Probing2026.06 | 97 | — | — | |
| SuPreM# pre-trained data=2.1k, Evaluation Protocol=Linear Probing2026.06 | 97 | — | — | |
| VoCo-B# pre-trained data=160k, Evaluation Protocol=Linear Probing2026.06 | 97 | — | — | |
| VoCo-H# pre-trained data=160k, Evaluation Protocol=Linear Probing2026.06 | 97 | — | — | |
| CoralBayU96B# pre-trained data=11k, Evaluation Protocol=Linear Probing2026.06 | 97 | — | — | |
| CoralBayU96H# pre-trained data=11k, Evaluation Protocol=Linear Probing2026.06 | 97 | — | — | |
| CoralBayU96H (FT)# pre-trained data=11k, Evaluation Protocol=Fine-tuned2026.06 | 97 | — | — | |
| SwinUNETR# pre-trained data=—, Evaluation Protocol=Fine-tuned2026.06 | 96 | — | — | |
| Fully Supervised Learning BaselineType=FSL2023.06 | 84.7 | 76.5 | 3.6 | |
| CG-CDMType=DM2023.06 | 31.1 | 18.6 | 23.3 | |
| CDMType=DM2023.06 | 26.3 | 16.7 | 26.6 | |
| CG-DiffType=DM2023.06 | 23.5 | 15.2 | 27.1 | |
| LayerCAMType=CAM2023.06 | 19.4 | 13.1 | 29.7 | |
| GradCAM++Type=CAM2023.06 | 14.7 | 8.5 | 28.5 | |
| ScoreCAMType=CAM2023.06 | 13.5 | 7.8 | 32.1 | |
| GradCAMType=CAM2023.06 | 10.5 | 5.9 | 33.9 |