Research map / Video / Video segmentation and tracking
Learning Efficient Meshflow and Optical Flow From Event Cameras
Xinglong Luo, Ao Luo, Kunming Luo, Zhengning Wang, Ping Tan, Bing Zeng, Shuaicheng Liu
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 ADM to EEMFlow | EEMFlow · HREM+ meshflow dt=1 | average EPE: 5.50 -> 5.05 | Improved |
| add ADM to EEMFlow | EEMFlow · HREM+ meshflow dt=4 | average EPE: 24.45 -> 22.43 | Improved |
| add ADM to TMA | TMA · DSEC | 1PE: 10.9 -> 10.0 | Improved |
| add ADM to BFlow | BFlow · DSEC | 1PE: 11.9 -> 10.2 | Improved |
| add ADM to EEMFlow+ | EEMFlow+ · DSEC | 1PE: 11.4 -> 10.3 | Improved |
| add self-corrector branch of CDC (ablation) | EEMFlow+ · DSEC | 1PE: 15.5 -> 14.1 | Improved |
| add self-correlation branch of CDC (ablation) | EEMFlow+ · DSEC | 1PE: 15.5 -> 12.8 | Improved |
| add both CDC branches (ablation) | EEMFlow+ · DSEC | 1PE: 15.5 -> 11.4 | Improved |
| use channel shuffle instead of standard convolutions in decoder (ablation) | EEMFlow · HREM+ dt=1 Fast | EPE: 9.45 -> 8.78 | Improved |
| use channel shuffle instead of depthwise convolutions (ablation) | EEMFlow · HREM+ dt=1 Fast | EPE: 11.79 -> 8.78 | Improved |
| use DFC instead of standard r=3 correlation (ablation) | EEMFlow · HREM+ dt=1 Slow | EPE: 2.58 -> 2.21 | Improved |
| use DFC instead of standard r=4 correlation (ablation) | EEMFlow · inference time | Time (ms): 27 -> 7 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- DPFlow — DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Framework (CVPR 2025) · better on 2 of 2 shared results
- HDRFlow — HDRFlow: Real-Time HDR Video Reconstruction with Large Motions (CVPR 2024) · better on 1 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
- Efficient Meshflow and Optical Flow Estimation from Event Cameras (CVPR 2024 · similarity 0.97)
- Learning Optical Flow from Event Camera with Rendered Dataset (ICCV 2023 · similarity 0.92)
- EDCFlow: Exploring Temporally Dense Difference Maps for Event-based Optical Flow Estimation (CVPR 2025 · similarity 0.90)
- Learning Normal Flow Directly From Events (ICCV 2025 · similarity 0.90)
- E-MoFlow: Learning Egomotion and Optical Flow from Event Data via Implicit Regularization (NeurIPS 2025 · similarity 0.87)
- EF-3DGS: Event-Aided Free-Trajectory 3D Gaussian Splatting (NeurIPS 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.
