LiDAR Semantic Segmentation on SemanticPOSS (test)
53.5mIoUFRNet
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FRNetRepresentation=Frustum-Range, Param=10.0M, FPS=29.12023.12 | 53.5 | 68.1 | |
| MinkUNetRepresentation=Sparse Voxel, Param=21.7M, FPS=9.12023.12 | 53.1 | 68.3 | |
| Cylinder3DRepresentation=Sparse Voxel, Param=55.9M, FPS=6.22023.12 | 52.9 | 64.9 | |
| Fast-FRNetRepresentation=Frustum-Range, Param=7.5M, FPS=33.82023.12 | 52.5 | 67.1 | |
| CENetRepresentation=Range View, Param=6.8M, FPS=33.42023.12 | 50.3 | — | |
| SPVNASRepresentation=Sparse Voxel, Param=21.8M, FPS=8.92023.12 | 48.4 | 61.5 | |
| FIDNetRepresentation=Range View, Param=6.1M, FPS=31.82023.12 | 46.4 | — | |
| RangeNet++Representation=Range View, Param=50.4M, FPS=12.02023.12 | 30.9 | — |