Research map / 3D vision / Depth and stereo
RPEFlow: Multimodal Fusion of RGB-PointCloud-Event for Joint Optical Flow and Scene Flow Estimation
Zhexiong Wan, Yuxin Mao, Jing Zhang, Yuchao Dai
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 |
|---|---|---|---|
| add event data to the framework (ablation) | RPEFlow · FlyingThings3D val | EPE2D: 2.067 -> 1.402 | Improved |
| add event feature to CamLiFlow decoder (ablation) | CamLiFlow · FlyingThings3D val | EPE2D: 2.20 -> 1.56 | Improved |
| add event feature to RAFT decoder (ablation) | RAFT · FlyingThings3D val | EPE2D: 3.12 -> 2.49 | Improved |
| replace concat fusion with attention fusion (no event) (ablation) | RPEFlow · FlyingThings3D val | EPE2D: 2.200 -> 2.133 | Improved |
| replace concat fusion with attention fusion (with event) (ablation) | RPEFlow · FlyingThings3D val | EPE2D: 1.519 -> 1.494 | Improved |
| add mutual information regularization (no event) (ablation) | RPEFlow · FlyingThings3D val | EPE2D: 2.133 -> 2.067 | Improved |
| add mutual information regularization (with event) (ablation) | RPEFlow · FlyingThings3D val | EPE2D: 1.494 -> 1.402 | Improved |
| add event into the concat fusion (ablation) | RPEFlow concat variant · FlyingThings3D val | EPE2D: 1.561 -> 1.519 | Improved |
| train 2D and 3D branches jointly instead of 2D only (ablation) | RPEFlow · FlyingThings3D val | EPE2D: 1.937 -> 1.402 | Improved |
| train 2D and 3D branches jointly instead of 3D only (ablation) | RPEFlow · FlyingThings3D val | EPE3D (Full): 0.078 -> 0.042 | Improved |
| add event to RGB+PC input with CamLiFlow, ACC1px | CamLiFlow · FlyingThings3D val | ACC1px: 84.6% -> 84.4% | No clear effect |
Later papers that compared with this method
1 later accepted paper reports RPEFlow as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- x^2-Fusion: Cross-Modality and Cross-Dimension Flow Estimation in Event Edge Space (CVPR 2026 · better on 9 of 9 shared results)
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.
- CamLiFlow — CamLiFlow: Bidirectional Camera-LiDAR Fusion for Joint Optical Flow and Scene Flow Estimation (CVPR 2022) · better on 6 of 6 shared results
- RAFT-3D — RAFT-3D: Scene Flow Using Rigid-Motion Embeddings (CVPR 2021) · better on 1 of 2 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
- Bring Event into RGB and LiDAR: Hierarchical Visual-Motion Fusion for Scene Flow (CVPR 2024 · similarity 0.91)
- Learning Optical Flow and Scene Flow With Bidirectional Camera-LiDAR Fusion (TPAMI 2023 · similarity 0.89)
- x^2-Fusion: Cross-Modality and Cross-Dimension Flow Estimation in Event Edge Space (CVPR 2026 · similarity 0.89)
- ARES: Unifying Asymmetric RGB-Event Stereo for Probabilistic Scene Flow Estimation (CVPR 2026 · similarity 0.89)
- Injecting Frame-Event Complementary Fusion into Diffusion for Optical Flow in Challenging Scenes (NeurIPS 2025 · similarity 0.86)
- Bridge Frame and Event: Common Spatiotemporal Fusion for High-Dynamic Scene Optical Flow (CVPR 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.
