Research map / Point clouds / Point cloud classification
PointPWC-Net: Cost Volume on Point Clouds for (Self-)Supervised Scene Flow Estimation
Wenxuan Wu, Zhi Yuan Wang, Zhuwen Li, Wei Liu, Li Fuxin
Read the paperCheck your idea against 137,217 papers
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 |
|---|---|---|---|
| use the proposed learnable cost volume instead of PWC-Net inner product cost volume (ablation) | PointPWC-Net · FlyingThings3D | EPE3D: 0.0821 -> 0.0588 | Improved |
| use the proposed cost volume instead of MLP+Maxpool flow embedding (ablation) | PointPWC-Net · FlyingThings3D | EPE3D: 0.0741 -> 0.0588 | Improved |
| add the warping layer (ablation) | PointPWC-Net · FlyingThings3D | EPE3D: 0.0984 -> 0.0588 | Improved |
| add smoothness constraint to Chamfer loss (self-supervised) (ablation) | PointPWC-Net · FlyingThings3D | EPE3D: 0.2112 -> 0.1304 | Improved |
| add Laplacian regularization to Chamfer + smoothness (ablation) | PointPWC-Net · FlyingThings3D | EPE3D: 0.1304 -> 0.1213 | Improved |
| self-supervised fine-tuning on KITTI after self-supervised pretraining (Self+Self) | PointPWC-Net · KITTI | EPE3D: 0.2549 -> 0.0461 | Improved |
| self-supervised fine-tuning on KITTI after full supervision (Full+Self) | PointPWC-Net · KITTI | EPE3D: 0.0694 -> 0.0430 | Improved |
| PointPWC-Net versus HPLFlowNet (fully supervised) | PointPWC-Net · FlyingThings3D | EPE3D: 0.0804 -> 0.0588 | Improved |
| PointPWC-Net versus HPLFlowNet on KITTI without fine-tuning | PointPWC-Net · KITTI | EPE3D: 0.1169 -> 0.0694 | Improved |
| self-supervised PointPWC-Net versus CPD(non-rigid) on KITTI | PointPWC-Net · KITTI | EPE3D: 0.4144 -> 0.2549 | Improved |
| supervised training (Full) versus self-supervised training | PointPWC-Net · FlyingThings3D | EPE3D: 0.1213 -> 0.0588 | Improved |
| PointPWC-Net runtime versus HPLFlowNet | PointPWC-Net · FlyingThings3D | runtime (ms): 98.4 -> 117.4 | No clear effect |
Later papers that compared with this method
11 later accepted papers report PointPWC-Net as a baseline in their results tables; 10 of them report a better number on most of the results they share with it.
- GenFlow3D: Generative Scene Flow Estimation and Prediction on Point Cloud Sequences (ICCV 2025 · better on 2 of 2 shared results)
- FastPCI: Motion-Structure Guided Fast Point Cloud Frame Interpolation (ECCV 2024 · better on 1 of 1 shared result)
- RMS-FlowNet++: Efficient and Robust Multi-scale Scene Flow Estimation for Large-Scale Point Clouds (IJCV 2024 · better on 1 of 1 shared result)
- IHNet: Iterative Hierarchical Network Guided by High-Resolution Estimated Information for Scene Flow Estimation (ICCV 2023 · better on 1 of 1 shared result)
- DELFlow: Dense Efficient Learning of Scene Flow for Large-Scale Point Clouds (ICCV 2023 · better on 1 of 1 shared result)
- FH-Net: A Fast Hierarchical Network for Scene Flow Estimation on Real-World Point Clouds (ECCV 2022 · better on 3 of 4 shared results)
- Multiway Non-Rigid Point Cloud Registration via Learned Functional Map Synchronization (TPAMI 2022 · better on 1 of 1 shared result)
- Learning Scene Dynamics from Point Cloud Sequences (IJCV 2022 · better on 1 of 1 shared result)
3 more not listed.
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- FlowNet3D — FlowNet3D: Learning Scene Flow in 3D Point Clouds (CVPR 2019) · better on 1 of 1 shared result
- HPLFlowNet — HPLFlowNet: Hierarchical Permutohedral Lattice FlowNet for Scene Flow Estimation on Large-Scale Point Clouds (CVPR 2019) · better on 0 of 1 shared result
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
- Deep Scene Flow Learning: From 2D Images to 3D Point Clouds (TPAMI 2023 · similarity 0.90)
- PointFlowNet: Learning Representations for Rigid Motion Estimation From Point Clouds (CVPR 2019 · similarity 0.90)
- RCP: Recurrent Closest Point for Point Cloud (CVPR 2022 · similarity 0.90)
- DELFlow: Dense Efficient Learning of Scene Flow for Large-Scale Point Clouds (ICCV 2023 · similarity 0.90)
- SCOOP: Self-Supervised Correspondence and Optimization-Based Scene Flow (CVPR 2023 · similarity 0.89)
- LinNet: Linear Network for Efficient Point Cloud Representation Learning (NeurIPS 2024 · similarity 0.88)
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.
