Research map / Learning theory and algorithms / Prediction and decision losses
Object Pose Estimation With Statistical Guarantees: Conformal Keypoint Detection and Geometric Uncertainty Propagation
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 |
|---|---|---|---|
| Increase ε from 0.1 to 0.4 for gt-ball | RANSAG average pose · LineMOD Occlusion | 2D projection success rate: 67.33 -> 70.71 | Helped |
| Increase ε from 0.1 to 0.4 for gt-ellipse | RANSAG average pose · LineMOD Occlusion | 2D projection success rate: 70.61 -> 71.39 | Helped |
| Increase ε from 0.1 to 0.4 for frcnn-ball | RANSAG average pose · LineMOD Occlusion | 2D projection success rate: 56.93 -> 61.58 | Helped |
| Increase ε from 0.1 to 0.4 for frcnn-ellipse | RANSAG average pose · LineMOD Occlusion | 2D projection success rate: 57.73 -> 61.25 | Helped |
| Use ellipse rather than ball keypoint prediction sets with groundtruth boxes at ε=0.1 | Conformalized heatmap and RANSAG · LineMOD Occlusion | 2D projection success rate: 67.33 -> 70.61 | Helped |
| Use ellipse rather than ball keypoint prediction sets with Faster RCNN boxes at ε=0.1 | Conformalized heatmap and RANSAG · LineMOD Occlusion | 2D projection success rate: 56.93 -> 57.73 | Helped |
| Use groundtruth rather than Faster RCNN boxes with ball sets at ε=0.1 | Conformalized heatmap and RANSAG · LineMOD Occlusion | 2D projection success rate: 56.93 -> 67.33 | Helped |
| Replace the maximum keypoint nonconformity aggregation with a sum and robust Geman-McClure cost (ablation) | Conformal keypoint detection · ICP keypoint prediction sets | prediction set | Mixed |
| Apply a monotonically increasing robust cost to each keypoint score while keeping maximum aggregation (ablation) | Conformal keypoint detection · ICP keypoint prediction sets | prediction set | No clear effect |
| Use first-order rather than second-order semidefinite relaxation (ablation) | Worst-case pose error bounds · LineMOD Occlusion | rotation error bound tightness: 100° | Hurt |
Closest papers by meaning
- Instance-Adaptive and Geometric-Aware Keypoint Learning for Category-Level 6D Object Pose Estimation (CVPR 2024 · similarity 0.89)
- 3D-Aware Hypothesis & Verification for Generalizable Relative Object Pose Estimation (ICLR 2024 · similarity 0.88)
- ProbPose: A Probabilistic Approach to 2D Human Pose Estimation (CVPR 2025 · similarity 0.88)
- UA-Pose: Uncertainty-Aware 6D Object Pose Estimation and Online Object Completion with Partial References (CVPR 2025 · similarity 0.87)
- KeyPose: Category-Level 6D Object Pose Estimation with Self-Adaptive Keypoints (AAAI 2025 · similarity 0.87)
- Structure-Aware Correspondence Learning for Relative Pose Estimation (CVPR 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.