Research map / Learning theory and algorithms / Combinatorial and search optimization
Globally Optimal Contrast Maximisation for Event-Based Motion Estimation
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 |
|---|---|---|---|
| Use globally optimal branch-and-bound contrast maximisation instead of the strongest locally optimised baseline | Rotational motion estimation · dynamic, boxes, poster | mean angular-velocity error φ (deg/s): 12.29 -> 10.09; 21.41 -> 17.97; 50.04 -> 46.34 | Helped |
| Use globally optimal branch-and-bound contrast maximisation instead of the strongest locally optimised baseline | Rotational motion estimation · star tracking | mean angular-velocity error φ (deg/s): 0.423 -> 0.234 | Helped |
| Use globally optimal branch-and-bound contrast maximisation instead of the strongest locally optimised baseline | Rotational motion estimation · dynamic, boxes, poster | mean parameter error ε: 14.55 -> 11.93; 21.95 -> 18.76; 49.49 -> 44.34 | Helped |
| Use globally optimal branch-and-bound contrast maximisation instead of the strongest locally optimised baseline | Rotational motion estimation · star tracking | mean parameter error ε: 0.294 -> 0.174 | Helped |
| Initialize the local solver from the previous time window rather than zero (ablation) | CMGD · dynamic, boxes, poster | mean angular-velocity error φ (deg/s): 20.07 -> 13.31; 31.47 -> 21.70; 54.92 -> 50.12 | Helped |
| Initialize the local solver from the previous time window rather than zero (ablation) | CMRW · dynamic, boxes, poster | mean angular-velocity error φ (deg/s): 18.59 -> 12.29; 32.23 -> 21.41; 58.71 -> 50.04 | Helped |
| Replace the contrast objective with the reward objective (ablation) | Local CM solver initialized at zero · dynamic, boxes, poster | mean angular-velocity error φ (deg/s): 20.07 -> 18.59; 31.47 -> 32.23; 54.92 -> 58.71 | No clear effect |
| Use the discrete event image and its bound rather than the continuous event image (ablation) | CMBnB · boxes, 10 ms subsequence | iterations to termination: 30k -> 12k | Helped |
| Use the discrete event image rather than the continuous event image (ablation) | Contrast maximisation · boxes subsequences | solution quality | No clear effect |
| Use global branch-and-bound rather than local conjugate-gradient optimisation | Contrast maximisation · dynamic, boxes, poster | runtime (seconds): 20.2 -> 163.2; 31.1 -> 278.3; 35.3 -> 320.6 | Hurt |
Closest papers by meaning
- Event Cameras, Contrast Maximization and Reward Functions: An Analysis (CVPR 2019 · similarity 0.90)
- A Unifying Contrast Maximization Framework for Event Cameras, With Applications to Motion, Depth, and Optical Flow Estimation (CVPR 2018 · similarity 0.90)
- EF-3DGS: Event-Aided Free-Trajectory 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.84)
- An N-Point Linear Solver for Line and Motion Estimation with Event Cameras (CVPR 2024 · similarity 0.84)
- EvEnhancer: Empowering Effectiveness, Efficiency and Generalizability for Continuous Space-Time Video Super-Resolution with Events (CVPR 2025 · similarity 0.84)
- Motion-adaptive Transformer for Event-based Image Deblurring (AAAI 2025 · similarity 0.83)
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.