3D Semantic Segmentation on SensatUrban (val)
91.45OAUNet
Evaluation results
| Method | Links | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| UNetBackbone=ResNet-34, Evaluation Protocol=BEV Projection2021.09 | 91.45 | 67.76 | 59.9 | 78.45 | 95.39 | 97.89 | 60.72 | 59.8 | 40.45 | 40.35 | 67.32 | 51.42 | 82.11 | 41.94 | 0 | 62.87 | |
| OCRNetBackbone=HRNet, Evaluation Protocol=BEV Projection2021.09 | 91.37 | 71.87 | 61.17 | 83.21 | 92.16 | 94.4 | 54.84 | 28.61 | 52.25 | 36.55 | 74.46 | 50.91 | 80.1 | 48.19 | 0 | 65.37 | |
| Deeplabv3Backbone=ResNet-101, Evaluation Protocol=BEV Projection2021.09 | 90.53 | 72.32 | 60.28 | 81.27 | 90.09 | 93.98 | 52.28 | 59.82 | 49.32 | 15.88 | 72.81 | 48.72 | 76.86 | 46.23 | 0 | 61.51 |