Research map / Point clouds / Point cloud classification
Self-Point-Flow: Self-Supervised Scene Flow Estimation From Point Clouds With Optimal Transport and Random Walk
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 cycle-consistency regularization | Self-Point-Flow with FlowNet3D · FT3D_s | EPE: 0.1208 -> 0.1009 | Helped |
| Add cycle-consistency regularization | Self-Point-Flow with FlowNet3D · KITTI_s | EPE: 0.1271 -> 0.1120 | Helped |
| Add KITTI_v training data | Self-Point-Flow trained on KITTI_r · KITTI_t | EPE: 0.1029 -> 0.0895 | Helped |
| Add color and surface-normal descriptors to matching (ablation) | Self-Point-Flow pseudo-label generation · KITTI_o | AS: 10.19 -> 24.36 | Helped |
| Add one-to-one global matching constraints (ablation) | Coordinate-only greedy point matching · KITTI_o | AS: 1.85 -> 10.19 | Helped |
| Use pre-warped points and hard correspondences rather than original-point hard matching (ablation) | Self-Point-Flow pseudo-label generation · KITTI_o | AS: 12.52 -> 24.36 | Helped |
| Use hard rather than soft correspondences on pre-warped points (ablation) | Self-Point-Flow pseudo-label generation · KITTI_o | AS: 13.16 -> 24.36 | Helped |
| Add KNN-based naive smoothing (ablation) | Self-Point-Flow pseudo-label generation · KITTI_o | AS: 24.36 -> 27.53 | Helped |
| Add undirected-graph random-walk smoothing (ablation) | Self-Point-Flow pseudo-label generation · KITTI_o | AS: 24.36 -> 40.88 | Helped |
| Add directed-graph random-walk label generation for unlabeled points (ablation) | Self-Point-Flow with undirected-graph smoothing · KITTI_o | AS: 40.88 -> 41.74 | Helped |
| Increase random-walk iterations from 1 to convergence (ablation) | Self-Point-Flow pseudo-label refinement · KITTI_o | AS: 37.29 -> 41.74 | Helped |
Closest papers by meaning
- SCOOP: Self-Supervised Correspondence and Optimization-Based Scene Flow (CVPR 2023 · similarity 0.93)
- Self-Supervised Robust Scene Flow Estimation via the Alignment of Probability Density Functions (AAAI 2022 · similarity 0.93)
- RigidFlow: Self-Supervised Scene Flow Learning on Point Clouds by Local Rigidity Prior (CVPR 2022 · similarity 0.92)
- ZeroFlow: Scalable Scene Flow via Distillation (ICLR 2024 · similarity 0.88)
- 3DSFLabelling: Boosting 3D Scene Flow Estimation by Pseudo Auto-labelling (CVPR 2024 · similarity 0.87)
- Sonata: Self-Supervised Learning of Reliable Point Representations (CVPR 2025 · similarity 0.87)
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.