Research map / 3D vision / Human pose estimation
Monocular 3D Multi-Person Pose Estimation by Integrating Top-Down and Bottom-Up Networks
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 multi-person pose estimation to the top-down branch (ablation) | Top-down network · MuPoTS-3D | PCK_abs: 42.8 -> 45.7 | Helped |
| add top-down combined heatmaps to the bottom-up input (ablation) | Bottom-up network · MuPoTS-3D | PCK_abs: 40.2 -> 41.5 | Helped |
| replace hard integration with the integration network (ablation) | Top-down and bottom-up branches · MuPoTS-3D | PCK_abs: 46.2 -> 47.4 | Helped |
| replace confidence-weighted linear fusion with the integration network (ablation) | Top-down and bottom-up branches · MuPoTS-3D | PCK_abs: 46.7 -> 47.4 | Helped |
| add pose matching (ablation) | Top-down and bottom-up integration · MuPoTS-3D | PCK_abs: 45.3 -> 47.4 | Helped |
| add reprojection loss alongside the multi-perspective loss and discriminator (ablation) | Semi-supervised training · MuPoTS-3D | AUC_rel: 49.9 -> 50.6 | Helped |
| add the discriminator to reprojection-loss training (ablation) | Semi-supervised training · MuPoTS-3D | AUC_rel: 43.4 -> 49.9 | Helped |
| use multi-perspective loss instead of reprojection loss (ablation) | Semi-supervised training · MuPoTS-3D | AUC_rel: 43.4 -> 32.2 | Hurt |
| use the integrated top-down/bottom-up method | Monocular 3D multi-person pose estimation · JTA | F1 (t=0.4m): 50.82 -> 57.22 | Helped |
| use the integrated top-down/bottom-up method | Monocular 3D multi-person pose estimation · MuPoTS-3D | PCK_abs: 45.7 -> 48.0 | Helped |
| use the method on the occlusion subset rather than the full test set | Our method · 3DPW | PA-MPJPE: 62.9 -> 75.6 | Mixed |
Closest papers by meaning
- Graph and Temporal Convolutional Networks for 3D Multi-person Pose Estimation in Monocular Videos (AAAI 2021 · similarity 0.91)
- Monocular 3D Pose and Shape Estimation of Multiple People in Natural Scenes - The Importance of Multiple Scene Constraints (CVPR 2018 · similarity 0.91)
- Combining Detection and Tracking for Human Pose Estimation in Videos (CVPR 2020 · similarity 0.90)
- Simple Pose: Rethinking and Improving a Bottom-up Approach for Multi-Person Pose Estimation (AAAI 2020 · similarity 0.90)
- CoMotion: Concurrent Multi-person 3D Motion (ICLR 2025 · similarity 0.88)
- SelfPose3d: Self-Supervised Multi-Person Multi-View 3d Pose Estimation (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.