Research map / 3D vision / Human pose estimation
Single-View Robot Pose and Joint Angle Estimation via Render & Compare
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 render-and-compare state refinement instead of DREAM keypoint-based pose estimation | Camera-to-robot pose estimation · Panda-ORB | ADD: 69.05 -> 80.54 | Helped |
| Use render-and-compare state refinement instead of DREAM keypoint-based pose estimation | Camera-to-robot pose estimation · Panda 3CAM-RS | ADD: 78.83 -> 76.90 | Hurt |
| Predict pose and joint angles without ground-truth joint angles | RoboPose · Panda DR | ADD: 92.70 -> 82.85 | Hurt |
| Predict pose and joint angles without ground-truth joint angles | RoboPose · Kuka DR | ADD: 89.62 -> 80.17 | Hurt |
| Estimate pose and joint angles directly rather than using CRAVES synt+real | CRAVES-lab · CRAVES-lab, all images | Translation error (cm), xyz: 10.1 -> 0.61 | Helped |
| Use the robot centroid rather than the root as pose-update reference point (ablation) | RoboPose · Panda-ORB | ADD: 75.02 -> 80.54 | Helped |
| Use the robot centroid rather than the middle part P4 as pose-update reference point (ablation) | RoboPose · Panda-ORB | ADD: 79.45 -> 80.54 | Helped |
| Use the robot centroid rather than hand part P7 as pose-update reference point (ablation) | RoboPose · Panda-ORB | ADD: 00.00 -> 80.54 | Helped |
| Randomly select an anchor among the 3 largest parts instead of using fixed anchor P4 (ablation) | RoboPose · Panda-ORB | ADD: 69.54 -> 71.36 | Helped |
| Train the refiner with multiple iterations rather than one iteration (ablation) | RoboPose · Panda-ORB, unknown joint angles, K_test=10 | ADD: 41.25 -> 70.77 | Helped |
| Run multiple refinement iterations at test time rather than a single iteration (ablation) | RoboPose · Panda-ORB video | ADD: 28.5 -> 80.7 | Helped |
| Increase training iterations from 3 to 5 (ablation) | RoboPose · Panda-ORB, unknown joint angles, K_test=10 | ADD: 70.39 -> 70.77 | Helped |
Closest papers by meaning
- RoboPEPP: Vision-Based Robot Pose and Joint Angle Estimation through Embedding Predictive Pre-Training (CVPR 2025 · similarity 0.89)
- CAP-Net: A Unified Network for 6D Pose and Size Estimation of Categorical Articulated Parts from a Single RGB-D Image (CVPR 2025 · similarity 0.86)
- R^2-Art: Category-Level Articulation Pose Estimation from Single RGB Image via Cascade Render Strategy (AAAI 2025 · similarity 0.86)
- 6D Object Pose Tracking in Internet Videos for Robotic Manipulation (ICLR 2025 · similarity 0.85)
- Vision Foundation Model Enables Generalizable Object Pose Estimation (NeurIPS 2024 · similarity 0.85)
- Pose6DAug: Physically Plausible Multi-View Object Swapping for Robot Data Augmentation (NeurIPS 2026 · 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.