Research map / Point clouds / Point cloud classification
RaLiFlow: Scene Flow Estimation with 4D Radar and LiDAR Point Clouds
Jingyun Fu, Zhiyu Xiang, Na Zhao
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 |
|---|---|---|---|
| fuse radar and LiDAR with RaLiFlow instead of LiDAR-only DeFlow | RaLiFlow · VoD val, LiDAR flow | 3-way EPE: 0.0691 -> 0.0556 | Improved |
| fuse radar and LiDAR with RaLiFlow instead of radar-only CMFlow | RaLiFlow · VoD val, radar flow | 3-way EPE: 0.1180 -> 0.0588 | Improved |
| use DBCF cross-attention instead of concatenation (ablation) | RaLiFlow · VoD val, LiDAR flow | 3-way EPE: 0.0603 -> 0.0556 | Improved |
| add Gaussian radar-dynamics weights to DBCF (ablation) | RaLiFlow · VoD val, LiDAR flow | 3-way EPE: 0.0562 -> 0.0556 | Improved |
| add Gaussian radar-dynamics weights to DBCF (ablation) | RaLiFlow · VoD val, LiDAR flow | FS EPE: 0.0269 -> 0.0269 | No clear effect |
| add instance-level flow consistency loss (ablation) | RaLiFlow · VoD val, LiDAR flow | 3D EPE: 0.0272 -> 0.0252 | Improved |
| add instance-level flow consistency loss (ablation) | RaLiFlow · VoD val, radar flow | 3D EPE: 0.0401 -> 0.0366 | Improved |
| mask unreliable radar points in radar flow loss (ablation) | RaLiFlow · VoD val, radar flow | FD EPE: 0.1320 -> 0.1298 | Improved |
| add radar to LiDAR input (fusion vs LiDAR only) (ablation) | RaLiFlow-dagger · VoD val, LiDAR flow | 3-way EPE: 0.0673 -> 0.0562 | Improved |
| add LiDAR to radar input (fusion vs radar only) (ablation) | RaLiFlow-dagger · VoD val, radar flow | 3-way EPE: 0.0861 -> 0.0596 | Improved |
Closest papers by meaning
- Weakly Supervised Cross-Modal Learning for 4D Radar Scene Flow Estimation (ICML 2026 · similarity 0.90)
- Hidden Gems: 4D Radar Scene Flow Learning Using Cross-Modal Supervision (CVPR 2023 · similarity 0.89)
- HVA-Fusion:Hierarchical Velocity-Aware 4D Radar-LiDAR Fusion for Robust 3D Object Detection (ECCV 2026 · similarity 0.89)
- ICP-Flow: LiDAR Scene Flow Estimation with ICP (CVPR 2024 · similarity 0.86)
- HGSFusion: Radar-Camera Fusion with Hybrid Generation and Synchronization for 3D Object Detection (AAAI 2025 · similarity 0.86)
- Motion Cues from Image-based Point Tracking for LiDAR Scene Flow Estimation (NeurIPS 2026 · similarity 0.85)
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.
