Research map / 3D vision / Depth and stereo
Event-Based Photometric Bundle Adjustment
Shuang Guo, Guillermo Gallego
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 |
|---|---|---|---|
| run EPBA refinement on CMax-ω rotations (quadratic loss) | CMax-ω front-end · ECRot synthetic playroom | ARE RMSE (deg): 3.223 -> 1.066 | Improved |
| run EPBA refinement on CMax-ω rotations (quadratic loss) | CMax-ω front-end · ECRot synthetic bicycle | ARE RMSE (deg): 1.69 -> 0.195 | Improved |
| replace quadratic loss with Huber loss (ablation) | EPBA with CMax-ω init · ECRot synthetic town | ARE RMSE (deg): 1.19 -> 0.188 | Improved |
| replace quadratic loss with Cauchy loss (ablation) | EPBA with CMax-ω init · ECRot synthetic town | ARE RMSE (deg): 1.19 -> 0.196 | Improved |
| use Huber loss instead of quadratic (ablation) | EPBA with CMax-ω init · ECRot synthetic playroom | squared photometric error (x10^6): 0.088 -> 0.102 | Hurt |
| use Huber or Cauchy loss instead of quadratic (ablation) | EPBA · ECD real sequences | ARE RMSE (deg): 2.951 vs 3.020 | No clear effect |
| use Huber loss instead of quadratic (ablation) | EPBA with CMax-ω init · ECD real boxes | squared photometric error (x10^5): 1.620 -> 1.699 | Hurt |
| add EPBA refinement to CMax-ω rotations | CMax-ω front-end · ECD real boxes | squared photometric error (x10^5): 3.736 -> 1.620 | Improved |
| use EPBA instead of EMBA for refining CMax-ω rotations | CMax-ω front-end · ECD real shapes | ARE RMSE (deg): 4.441 -> 3.020 | Improved |
| EPBA refinement of CMax-ω rotations vs CMax-SLAM | CMax-ω front-end · ECRot synthetic playroom | ARE RMSE (deg): 0.763 vs 1.066 | Hurt |
| assume wrong contrast threshold C=0.5 instead of true 0.2 (ablation) | EPBA (Quad) · bicycle | ARE RMSE (deg): 0.449 -> 0.747 | Hurt |
| assume contrast threshold C=0.1 instead of true 0.2 (ablation) | EPBA (Quad) · bicycle | ARE RMSE (deg): 0.449 -> 0.379 | No clear effect |
Closest papers by meaning
- Event-based Mosaicing Bundle Adjustment (ECCV 2024 · similarity 0.93)
- Towards Robust Event-guided Low-Light Image Enhancement: A Large-Scale Real-World Event-Image Dataset and Novel Approach (CVPR 2024 · similarity 0.86)
- EAG3R: Event-Augmented 3D Geometry Estimation for Dynamic and Extreme-Lighting Scenes (NeurIPS 2025 · similarity 0.85)
- EF-3DGS: Event-Aided Free-Trajectory 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.85)
- Event-3DGS: Event-based 3D Reconstruction Using 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.85)
- EventMamba: Enhancing Spatio-Temporal Locality with State Space Models for Event-Based Video Reconstruction (AAAI 2025 · 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.
