Research map / 3D vision / Human pose estimation
PoseAgent: Budget-Constrained 6D Object Pose Estimation via Reinforcement Learning
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 PoseAgent with τmax=6 rather than Krull et al. | Pose estimation, N=210 · Total | percent correct poses: 60.06 -> 62.18 | Helped |
| Use PoseAgent with τmax=7 rather than Krull et al. | Pose estimation, N=420 · Total | percent correct poses: 68.03 -> 70.62 | Helped |
| Use PoseAgent with τmax=6 rather than Krull et al. | Pose estimation, N=210 · Samurai 2 | percent correct poses: 60.30 -> 54.82 | Hurt |
| Use PoseAgent with τmax=7 rather than Krull et al. | Pose estimation, N=420 · Samurai 2 | percent correct poses: 66.45 -> 61.13 | Hurt |
| Use PoseAgent rather than Krull et al. | Pose estimation, N=210 · Cat 2 | percent correct poses: 63.05 -> 71.52 | Helped |
| Use PoseAgent rather than Krull et al. | Pose estimation, N=420 · Toolbox 2 | percent correct poses: 64.79 -> 71.60 | Helped |
| Use dynamic policy-based refinement rather than random refinement | PoseAgent, N=210 · Total | percent correct poses: 38.47 -> 62.18 | Helped |
| Use dynamic policy-based refinement rather than best-energy-only refinement | PoseAgent, N=420 · Total | percent correct poses: 50.21 -> 70.62 | Helped |
| Set τmax to 3 instead of 6 (ablation) | PoseAgent, N=210 · Total | percent correct poses: 62.18 -> 58.94 | Hurt |
| Set τmax to 5 instead of 6 (ablation) | PoseAgent, N=210 · Total | percent correct poses: 62.18 -> 61.47 | Hurt |
| Set τmax to 7 instead of 6 (ablation) | PoseAgent, N=210 · Total | percent correct poses: 62.18 -> 60.88 | Hurt |
| Set τmax to 3 instead of 7 (ablation) | PoseAgent, N=420 · Total | percent correct poses: 70.62 -> 67.37 | Hurt |
Closest papers by meaning
- PFRL: Pose-Free Reinforcement Learning for 6D Pose Estimation (CVPR 2020 · similarity 0.92)
- Deep Reinforcement Learning for Active Human Pose Estimation (AAAI 2020 · similarity 0.90)
- Knowledge Distillation for 6D Pose Estimation by Aligning Distributions of Local Predictions (CVPR 2023 · similarity 0.88)
- FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (CVPR 2024 · similarity 0.88)
- 6D Object Pose Tracking in Internet Videos for Robotic Manipulation (ICLR 2025 · similarity 0.85)
- Pos3R: 6D Pose Estimation for Unseen Objects Made Easy (CVPR 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.