Research map / 3D vision / Human pose estimation
MetaPose: Fast 3D Pose From Multiple Views Without 3D Supervision
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 |
|---|---|---|---|
| Add neural Stage 2 refinement to the Stage 1 initialization (ablation) | MetaPose · Human3.6M, 4 cameras | PMPJPE (mm): 74 -> 32 | Helped |
| Add neural Stage 2 refinement to the Stage 1 initialization (ablation) | MetaPose · SkiPose-PTZ, 6 cameras | PMPJPE (mm): 81 -> 42 | Helped |
| Replace neural refinement with iterative gradient-based refinement (ablation) | MetaPose · Human3.6M, 4 cameras | PMPJPE (mm): 32 -> 43 | Hurt |
| Train Stage 2 with reprojection supervision instead of the probabilistic bundle-adjustment objective (ablation) | MetaPose · Human3.6M, 4 cameras | PMPJPE (mm): 39 -> 32 | Helped |
| Use single-view pose estimates for initialization instead of random Gaussian noise (ablation) | MetaPose · Human3.6M, 4 cameras | PMPJPE (mm): 36 -> 32 | Helped |
| Replace permutation-equivariant blocks with concatenated-input MLPs (ablation) | MetaPose · Human3.6M and SkiPose-PTZ | PMPJPE (mm): H36M, 4 cameras: 32 -> 30; SkiPose, 6 cameras: 42 -> 46 | Mixed |
| Add a personalized bone-length prior (ablation) | MetaPose (S1+S2) · Human3.6M, 2 cameras | PMPJPE (mm): 44 -> 37 | Helped |
| Add a teacher-student loss matching the iterative refiner (ablation) | MetaPose (S1+S2) · Human3.6M and SkiPose-PTZ | PMPJPE (mm): 32 -> 38; 50 -> 72 | Hurt |
| Use a single Gaussian instead of a Gaussian mixture (ablation) | MetaPose · Human3.6M, 2 cameras | PMPJPE (mm): 44 -> 48 | Hurt |
Closest papers by meaning
- Learning Monocular 3D Human Pose Estimation From Multi-View Images (CVPR 2018 · similarity 0.92)
- ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.88)
- MV-SSM: Multi-View State Space Modeling for 3D Human Pose Estimation (CVPR 2025 · similarity 0.88)
- SelfPose3d: Self-Supervised Multi-Person Multi-View 3d Pose Estimation (CVPR 2024 · similarity 0.88)
- PoseIRM: Enhance 3D Human Pose Estimation on Unseen Camera Settings via Invariant Risk Minimization (CVPR 2024 · similarity 0.88)
- COMPOSE: Hypergraph Cover Optimization for Multi-view 3D Human Pose Estimation (NeurIPS 2026 · 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.