Research map / 3D vision / Human pose estimation
Learning Monocular 3D Human Pose Estimation From Multi-View Images
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 |
|---|---|---|---|
| Add multi-view consistency to training on labeled S1 (ablation) | Ours, normalized pose distance · Human3.6M, known rotations | NMPJPE (mm): 83.4 -> 78.2 | Helped |
| Add multi-view consistency to training on labeled S1+S5 (ablation) | Ours, normalized pose distance · Human3.6M, known rotations | NMPJPE (mm): 76.1 -> 70.8 | Helped |
| Use normalized pose distance instead of MSE (ablation) | Ours, weakly supervised · Human3.6M, S1 only, known rotations | NMPJPE (mm): 88.8 -> 78.2 | Helped |
| Use normalized pose distance instead of MSE (ablation) | Ours, weakly supervised · Human3.6M, S1+S5, known rotations | NMPJPE (mm): 74.7 -> 70.8 | Helped |
| Add the regularization term that keeps unlabeled predictions near initial estimates (ablation) | Ours, multi-view plus supervised loss · Human3.6M | validation NMPJPE | Helped |
| Omit the regularization term (ablation) | Ours, multi-view plus supervised loss · Human3.6M training | training loss convergence | Hurt |
| Estimate camera rotations instead of using known rotations (ablation) | Ours, weakly supervised · Human3.6M, S1 labeled | NMPJPE improvement: 7.4 to 3.4 mm | Hurt |
| Add multi-view consistency using unlabeled S2-S8 data (ablation) | Ours, normalized pose distance · MPII-3DHP, supervised S1 | NMPJPE (mm): 124.9 -> 119.8 | Helped |
| Add multi-view consistency on unlabeled views during training (ablation) | Ours, normalized pose distance · MPII-3DHP, novel viewpoints on S8 | NMPJPE (mm): 125.4 -> 108.7 | Helped |
| Update the initial reference model during training instead of keeping it fixed (ablation) | Ours, regularizer · Experiments reported in the paper | prediction accuracy | No clear effect |
Closest papers by meaning
- CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild (CVPR 2021 · similarity 0.92)
- MetaPose: Fast 3D Pose From Multiple Views Without 3D Supervision (CVPR 2022 · similarity 0.92)
- Weakly-Supervised 3D Human Pose Learning via Multi-View Images in the Wild (CVPR 2020 · similarity 0.90)
- Multiview-Consistent Semi-Supervised Learning for 3D Human Pose Estimation (CVPR 2020 · similarity 0.90)
- MV-SSM: Multi-View State Space Modeling for 3D Human Pose Estimation (CVPR 2025 · similarity 0.89)
- PoseIRM: Enhance 3D Human Pose Estimation on Unseen Camera Settings via Invariant Risk Minimization (CVPR 2024 · similarity 0.89)
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.