Research map / Learning theory and algorithms / Prediction and decision losses
Adaptive Annealing for Robust Geometric Estimation
Read the paperCheck 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 |
|---|---|---|---|
| Adapt the annealing step using Hessian positive-definiteness instead of a fixed scale decrement | GNC for robust 3D registration · 3DMatch instances with < 50% outliers | Global minimum reached (S%): Small γ: 89.3; Large γ: 80.4; GNCp: 91.1 | Helped |
| Use a small fixed annealing factor (γ = 1.4) rather than adaptive annealing (ablation) | GNC for robust 3D registration · 3DMatch instances with < 50% outliers | Global minimum reached (S%): 91.1 vs 89.3 | Hurt |
| Use a large fixed annealing factor (γ = 2.0) rather than adaptive annealing (ablation) | GNC for robust 3D registration · 3DMatch instances with < 50% outliers | Global minimum reached (S%): 91.1 vs 80.4 | Hurt |
| Replace exact Hessian evaluation with the piecewise polynomial Hessian approximation (ablation) | GNCp · Robust 3D registration | computational cost | Helped |
| Use stochastic Hessian approximation after the first 5 iterations (ablation) | GNCp · Robust 3D registration | efficiency | Helped |
| Perturb the current estimate by a large random rotation at each GNC stage and accept a lower-cost result (ablation) | GNCp · 3DMatch | failure instances: ≈12% to ∼9% | Helped |
Closest papers by meaning
- One Ring to Rule Them All: Certifiably Robust Geometric Perception with Outliers (NeurIPS 2020 · similarity 0.88)
- Convex Relaxation for Robust Vanishing Point Estimation in Manhattan World (CVPR 2025 · similarity 0.86)
- Scalable 3D Registration via Truncated Entry-wise Absolute Residuals (CVPR 2024 · similarity 0.85)
- Adaptive 3D Reconstruction via Diffusion Priors and Forward Curvature-Matching Likelihood Updates (NeurIPS 2025 · similarity 0.85)
- A robust inlier identification algorithm for point cloud registration via $\mathbf{\ell_0}$-minimization (NeurIPS 2024 · similarity 0.85)
- Revisiting Sampson Approximations for Geometric Estimation Problems (CVPR 2024 · similarity 0.84)
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.