Medical Image Segmentation on QUBIQ Kidney
98BDiceSAMRIbox
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SAMRIboxPrompting Strategy=box-only2025.10 | 98 | 5.43 | 1.37 | |
| SAMRIbpPrompting Strategy=box+point2025.10 | 98 | 4.47 | 1.25 | |
| SAM-ViTB2025.10 | 95 | 19.53 | 1.94 | |
| MedSAM2025.10 | 95 | 9.18 | 1.82 | |
| Dice Semimetric LossesBackbone=UNet-ResNet502023.03 | 75.67 | — | — | |
| SoftSegBackbone=UNet-ResNet502023.03 | 69.83 | — | — | |
| MRNetBackbone=UNet-ResNet502023.03 | 68.56 | — | — | |
| Multi-headBackbone=UNet-ResNet502023.03 | 68 | — | — | |
| DropoutBackbone=UNet-ResNet502023.03 | 66.53 | — | — |