Image Segmentation on UPENN-GBM
79.2WT Dice ScoreLLaBIT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| LLaBITModel Stage=Stage II2026.04 | 79.2 | 77.8 | 61.3 | 72.7 | |
| LLaBITModel Stage=Stage I2026.04 | 64.5 | 43.9 | 31.8 | 46.7 | |
| M3DTraining Status=Fine-tuned2026.04 | 61.9 | 41.1 | 32.2 | 45 | |
| M3DTraining Status=Pretrained2026.04 | 4.1 | 3.5 | 1.6 | 3 | |
| VoCo (Wu et al., 2024)Pre-training Method=Medical SSL, Backbone=Swin UNETR2026.05 | 0.8602 | 0.8192 | 0.877 | 0.8521 | |
| TACOPre-training Method=Medical SSL, Backbone=Swin UNETR2026.05 | 0.8598 | 0.8258 | 0.8784 | 0.8562 | |
| Swin UNETR (Tang et al., 2022)Pre-training Method=Medical SSL, Backbone=Swin UNETR2026.05 | 0.8591 | 0.8184 | 0.8736 | 0.8504 | |
| S2DC (Pan et al., 2025)Pre-training Method=Medical SSL, Backbone=Swin UNETR2026.05 | 0.8577 | 0.8138 | 0.8749 | 0.849 | |
| SwinMM (Wang et al., 2023c)Pre-training Method=Medical SSL, Backbone=Swin UNETR2026.05 | 0.8569 | 0.8253 | 0.8802 | 0.8541 | |
| M3AE (Liu et al., 2023)Pre-training Method=Medical SSL, Backbone=UNET3D2026.05 | 0.8568 | 0.8198 | 0.8811 | 0.8526 | |
| UniFormerPre-training Method=From Scratch, Backbone=UniFormer2026.05 | 0.8537 | 0.8043 | 0.8662 | 0.8414 | |
| UNET3DPre-training Method=From Scratch, Backbone=UNET3D2026.05 | 0.8504 | 0.8036 | 0.8714 | 0.8418 | |
| UNETRPre-training Method=From Scratch, Backbone=UNETR2026.05 | 0.8478 | 0.7977 | 0.845 | 0.8302 | |
| BrainMVP (Rui et al., 2025)Pre-training Method=Medical SSL, Backbone=UniFormer2026.05 | 0.8467 | 0.8181 | 0.8765 | 0.8471 | |
| Swin UNETRPre-training Method=From Scratch, Backbone=Swin UNETR2026.05 | 0.8418 | 0.8233 | 0.867 | 0.8402 |