Scene Flow Estimation on KITTIo (test)
0.033EPEGMSF
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GMSFTraining Dataset=F3Do, Testing Mode=Zero-shot Generalization2023.05 | 0.033 | — | 95.9 | 13.7 | 91.6 | |
| SCOOP+Supervision=Self, Train data=FT3D_o (1,800)2022.11 | 0.047 | 91.3 | 95 | 18.6 | — | |
| SCOOP+Training Dataset=F3Do, Testing Mode=Zero-shot Generalization2023.05 | 0.047 | — | 95 | 18.6 | 91.3 | |
| SCOOPSupervision=Self, Train data=FT3D_o (1,800)2022.11 | 0.063 | 79.7 | 91 | 24.4 | — | |
| BiPFNSupervision=Full, Train data=FT3D_o (18,000)2022.11 | 0.065 | 76.9 | 90.6 | 26.4 | — | |
| Bi-PointFlowTraining Dataset=F3Do, Testing Mode=Zero-shot Generalization2023.05 | 0.065 | — | 90.6 | 26.4 | 76.9 | |
| 3DFlowSupervision=Full, Train data=FT3D_o (18,000)2022.11 | 0.073 | 81.9 | 89 | 26.1 | — | |
| 3DFlowTraining Dataset=F3Do, Testing Mode=Zero-shot Generalization2023.05 | 0.073 | — | 89 | 26.1 | 81.9 | |
| SPFlowNetSupervision=Self-supervised, Training Data=KITTI_r2023.05 | 0.086 | 61.1 | 82.4 | 39.1 | — | |
| FESTASupervision=Full, Train data=FT3D_o (18,000)2022.11 | 0.094 | 44.9 | 83.4 | — | — | |
| RigidFlowSupervision=Self-supervised, Training Data=KITTI_r2023.05 | 0.102 | 48.4 | 75.6 | 44.2 | — | |
| Self-Point-FlowSup.=Self, Training data=KITTIt2021.05 | 0.105 | 41.7 | 72.5 | 50.1 | — | |
| Self-Point-FlowSupervision=Self-supervised, Training Data=KITTI_r2023.05 | 0.105 | 41.7 | 72.5 | 50.1 | — | |
| FLOTSup.=Full, Training data=FT3Do2021.05 | 0.107 | 45.1 | 74 | 46.3 | — | |
| FLOTSupervision=Fully-supervised, Training Data=FT3D_s2023.05 | 0.107 | 45.1 | 74 | 46.3 | — | |
| FLOTSupervision=Full, Train data=FT3D_o (18,000)2022.11 | 0.107 | 45.1 | 74 | 46.3 | — | |
| PointPWCTraining Dataset=F3Do, Testing Mode=Zero-shot Generalization2023.05 | 0.118 | — | 75.7 | 49.6 | 40.3 | |
| FLOTTraining Dataset=F3Do, Testing Mode=Zero-shot Generalization2023.05 | 0.13 | — | 66.7 | 52.9 | 27.8 | |
| FlowNet3DSup.=Full, Training data=FT3Do2021.05 | 0.173 | 27.6 | 60.9 | 64.9 | — | |
| FlowNet3DSupervision=Fully-supervised, Training Data=FT3D_s2023.05 | 0.173 | 27.6 | 60.9 | 64.9 | — | |
| FlowNet3DSupervision=Full, Train data=FT3D_o (18,000)2022.11 | 0.173 | 27.6 | 60.9 | 64.9 | — | |
| FlowNet3DTraining Dataset=F3Do, Testing Mode=Zero-shot Generalization2023.05 | 0.183 | — | 39.4 | 79.9 | 9.8 | |
| HPLFlowNetTraining Dataset=F3Do, Testing Mode=Zero-shot Generalization2023.05 | 0.343 | — | 38.6 | 81.4 | 10.3 |