Research map / Point clouds / Point cloud classification
Bi-PointFlowNet: Bidirectional Learning for Point Cloud Based Scene Flow Estimation
Wencan Cheng, Jong Hwan Ko
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add bidirectional feature propagation (BFP) (ablation) | Bi-PointFlowNet · FlyingThings3D non-occluded (FT3D_s) | EPE3D (m): 0.042 -> 0.028 | Improved |
| add bidirectional feature propagation (BFP) (ablation) | Bi-PointFlowNet · KITTI non-occluded (KITTI_s, trained on FT3D_s) | EPE3D (m): 0.053 -> 0.030 | Improved |
| use decomposed form of bidirectional flow embedding (ablation) | Bi-PointFlowNet · FlyingThings3D non-occluded (FT3D_s) | GFLOPs: 23.8 -> 13.3 | Improved |
| use decomposed form of bidirectional flow embedding (ablation) | Bi-PointFlowNet · FlyingThings3D non-occluded (FT3D_s) | runtime (ms): 61.2 -> 40.5 | Improved |
| use decomposed BFE (ablation) | Bi-PointFlowNet · FlyingThings3D non-occluded (FT3D_s) | EPE3D (m): 0.029 -> 0.028 | No clear effect |
| use decomposed BFE (ablation) | Bi-PointFlowNet · KITTI non-occluded (KITTI_s, trained on FT3D_s) | ACC3DS: 0.925 -> 0.920 | Hurt |
| add BFP to PointPWC (ablation) | PointPWC · FlyingThings3D non-occluded (FT3D_s) | EPE3D (m): 0.059 -> 0.051 | Improved |
| add BFP to PointPWC (ablation) | PointPWC · KITTI non-occluded (KITTI_s, trained on FT3D_s) | EPE3D (m): 0.069 -> 0.059 | Improved |
| add BFP to FlowNet3D (ablation) | FlowNet3D · FlyingThings3D occluded (FT3D_o) | EPE3D (m): 0.157 -> 0.138 | Improved |
| add BFP to FlowNet3D (ablation) | FlowNet3D · KITTI occluded (KITTI_o, trained on FT3D_o) | EPE3D (m): 0.173 -> 0.118 | Improved |
| make FlowNet3D deeper to match BFP parameters (ablation) | FlowNet3D · FlyingThings3D occluded (FT3D_o) | EPE3D (m): 0.157 -> 0.160 | Hurt |
| make FlowNet3D deeper to match BFP parameters (ablation) | FlowNet3D · KITTI occluded (KITTI_o, trained on FT3D_o) | EPE3D (m): 0.173 -> 0.197 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- FlowNet3D — FlowNet3D: Learning Scene Flow in 3D Point Clouds (CVPR 2019) · better on 4 of 4 shared results
- HPLFlowNet — HPLFlowNet: Hierarchical Permutohedral Lattice FlowNet for Scene Flow Estimation on Large-Scale Point Clouds (CVPR 2019) · better on 4 of 4 shared results
- HCRF-Flow — HCRF-Flow: Scene Flow From Point Clouds With Continuous High-Order CRFs and Position-Aware Flow Embedding (CVPR 2021) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Multi-Scale Bidirectional Recurrent Network with Hybrid Correlation for Point Cloud Based Scene Flow Estimation (ICCV 2023 · similarity 0.91)
- What Matters for 3D Scene Flow Network (ECCV 2022 · similarity 0.89)
- PointFlowNet: Learning Representations for Rigid Motion Estimation From Point Clouds (CVPR 2019 · similarity 0.89)
- Deep Scene Flow Learning: From 2D Images to 3D Point Clouds (TPAMI 2023 · similarity 0.88)
- FlowMamba: Learning Point Cloud Scene Flow with Global Motion Propagation (AAAI 2025 · similarity 0.87)
- DeltaFlow: An Efficient Multi-frame Scene Flow Estimation Method (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
