Medical Image Segmentation on MSD Prostate 5-fold CV
63.78Dice (Label 1)ensemble 2D U-Net + 3D U-Net
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ensemble 2D U-Net + 3D U-NetEnsemble=2D U-Net + 3D U-Net2018.09 | 63.78 | 85.31 | — | — | — | |
| 2D U-NetArchitecture=2D U-Net2018.09 | 61.98 | 84.31 | — | — | — | |
| 3D U-NetArchitecture=3D U-Net2018.09 | 60.77 | 83.73 | — | — | — | |
| nnUnet-2dParams=29.97M, Throughput=0.209s/vol2021.02 | — | — | 62.85 | 83.8 | 73.33 | |
| nnUnet-3d_fullParams=44.80M, Throughput=0.381s/vol2021.02 | — | — | 66.63 | 84.1 | 75.37 | |
| TransFuse-SParams=26.30M, Throughput=0.192s/vol2021.02 | — | — | 67.38 | 85.39 | 76.39 |