Research map / 3D vision / Human pose estimation
Deep Reinforcement Learning for Active Human Pose 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 |
|---|---|---|---|
| use automatic stopping rather than a fixed view count | Pose-DRL-S · Maf; Ult; Pose; Maf + Ult; All | reconstruction error (mm/joint): 144.7 -> 130.3; 157.5 -> 135.4; 135.1 -> 135.3; 155.5 -> 134.2; 140.4 -> 135.0 | Helped |
| use automatic stopping rather than a fixed view count | Pose-DRL-M · Maf; Ult; Pose; Maf + Ult; All | reconstruction error (mm/joint): 114.8 -> 114.8; 118.0 -> 116.4; 106.7 -> 104.6; 117.6 -> 115.9; 112.8 -> 110.7 | Mixed |
| select views with Pose-DRL automatic policy instead of Max-Azim | Pose-DRL-S · Maf; Ult | reconstruction error (mm/joint): 156.3 -> 130.3; 171.4 -> 135.4 | Helped |
| select views with Pose-DRL automatic policy instead of Max-Azim | Pose-DRL-M · Maf; Ult | reconstruction error (mm/joint): 123.5 -> 114.8; 131.2 -> 116.4 | Helped |
| use only the base feature map as state information (ablation) | Pose-DRL-S · Maf; Ult; Pose | reconstruction error (mm/joint): 144.7 -> 153.5; 157.5 -> 166.9; 135.1 -> 134.4 | Mixed |
| propagate the fused pose estimate to the next active-view instead of resetting (ablation) | Pose-DRL-S · Maf; Ult; Pose | reconstruction error (mm/joint): 152.5 -> 144.7; 160.8 -> 157.5; 133.4 -> 135.1 | Mixed |
Closest papers by meaning
- Domes to Drones: Self-Supervised Active Triangulation for 3D Human Pose Reconstruction (NeurIPS 2019 · similarity 0.91)
- PoseAgent: Budget-Constrained 6D Object Pose Estimation via Reinforcement Learning (CVPR 2017 · similarity 0.90)
- Attention Mechanism Exploits Temporal Contexts: Real-Time 3D Human Pose Reconstruction (CVPR 2020 · similarity 0.90)
- Proactive Multi-Camera Collaboration for 3D Human Pose Estimation (ICLR 2023 · similarity 0.89)
- ActiveMoCap: Optimized Viewpoint Selection for Active Human Motion Capture (CVPR 2020 · similarity 0.89)
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · similarity 0.88)
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.