Research map / 3D vision / Human pose estimation
Weakly-Supervised 3D Human Pose Learning via Multi-View Images in the Wild
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 3DHP and MannequinChallenge multi-view data to H36M training (ablation) | Weakly-supervised 3D pose estimator · Human3.6M | MPJPE: 69.1 -> 67.4 | Helped |
| Add MannequinChallenge multi-view data to 3DHP multi-view training (ablation) | Weakly-supervised 3D pose estimator · Human3.6M | 3D pose error (mm): 106.2 -> 93.6 | Helped |
| Use known camera extrinsics instead of estimating them from predictions (ablation) | Weakly-supervised model with 2D-annotated multi-view data · Human3.6M | 3D pose error (mm): 59.3 -> 57.2 | Helped |
| Remove 2D pose annotations from multi-view training data (ablation) | Weakly-supervised model using known camera extrinsics · Human3.6M | 3D pose error (mm): 57.2 -> 62.3 | Hurt |
| Remove 2D pose annotations from multi-view training data (ablation) | Weakly-supervised model estimating camera extrinsics · Human3.6M | 3D pose error (mm): 59.3 -> 69.1 | Hurt |
| Add multi-view consistency to MPII-only 2D supervision (ablation) | 2D pose estimator · Human3.6M | 2D pose error (px): 8.9 -> 8.4 | Helped |
| Add multi-view consistency with known camera extrinsics to MPII-only 2D supervision (ablation) | 2D pose estimator · Human3.6M | 2D pose error (px): 8.9 -> 8.3 | Helped |
| Add additional weak supervision from 3DHP and MannequinChallenge (ablation) | Fully supervised model · Human3.6M | MPJPE: 55.5mm vs 56.1mm | Hurt |
| Add 3DHP and MannequinChallenge multi-view data (ablation) | Semi-supervised model · Human3.6M | MPJPE: 62.8mm -> 59.7mm | Helped |
| Add 3DHP and MannequinChallenge multi-view data (ablation) | Weakly-supervised model · Human3.6M | MPJPE: 69.1mm -> 67.4mm | Helped |
| Use MannequinChallenge data as the only multi-view training source (ablation) | Weakly-supervised 3D pose estimator · Training convergence | convergence | Hurt |
Closest papers by meaning
- Deductive Learning for Weakly-Supervised 3D Human Pose Estimation via Uncalibrated Cameras (AAAI 2021 · similarity 0.92)
- CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild (CVPR 2021 · similarity 0.92)
- Multiview-Consistent Semi-Supervised Learning for 3D Human Pose Estimation (CVPR 2020 · similarity 0.92)
- Self-Supervised 3D Human Pose Estimation via Part Guided Novel Image Synthesis (CVPR 2020 · similarity 0.91)
- SelfPose3d: Self-Supervised Multi-Person Multi-View 3d Pose Estimation (CVPR 2024 · similarity 0.87)
- Unsupervised Learning of Category-Level 3D Pose from Object-Centric Videos (CVPR 2024 · similarity 0.87)
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.