Research map / 3D vision / Human pose estimation
Camera Distance-Aware Top-Down Approach for 3D Multi-Person Pose Estimation From a Single RGB Image
Gyeongsik Moon, Ju Yong Chang, Kyoung Mu Lee
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| train RootNet and PoseNet separately instead of jointly with a shared backbone (ablation) | 3D multi-person pose pipeline · Human3.6M (Protocol 2) | MRPE: 138.2 -> 120.0 | Improved |
| train RootNet and PoseNet separately instead of jointly with a shared backbone (ablation) | 3D multi-person pose pipeline · Human3.6M (Protocol 2) | MPJPE: 116.7 -> 57.3 | Improved |
| train RootNet and PoseNet separately instead of jointly with a shared backbone (ablation) | 3D multi-person pose pipeline · Human3.6M (Protocol 2) | seconds per frame: 0.132 -> 0.141 | Hurt |
| use RootNet correction of k instead of k directly as depth (ablation) | DetectNet R-50 · MuPoTS-3D | AP25 root: 5.2 -> 28.5 | Improved |
| use RootNet correction of k instead of k directly as depth (ablation) | DetectNet R-50 · MuPoTS-3D | 3DPCK_abs: 9.6 -> 31.5 | Improved |
| use stronger DetectNet backbone X-101-32 instead of R-50 (ablation) | RootNet (Ours) · MuPoTS-3D | AP25 root: 28.5 -> 31.0 | Improved |
| use stronger DetectNet backbone X-101-32 instead of R-50 (ablation) | RootNet (Ours) · MuPoTS-3D | 3DPCK_abs: 31.5 -> 31.5 | No clear effect |
| use groundtruth box instead of detected box (ablation) | RootNet (Ours) · MuPoTS-3D | 3DPCK_abs: 31.5 -> 31.6 | No clear effect |
| use groundtruth root instead of RootNet (ablation) | DetectNet GT box · MuPoTS-3D | 3DPCK_abs: 31.6 -> 80.2 | Improved |
| use groundtruth root instead of RootNet (ablation) | DetectNet GT box · MuPoTS-3D | AUC_rel: 39.8 -> 39.8 | No clear effect |
| use RootNet instead of distance minimization baseline (ablation) | root localization · Human3.6M (Protocol 2) | MRPE: 267.8 -> 120.0 | Improved |
| use RootNet instead of distance minimization with RANSAC (ablation) | root localization · Human3.6M (Protocol 2) | MRPE: 213.1 -> 120.0 | Improved |
Closest papers by meaning
- Dual Networks Based 3D Multi-Person Pose Estimation From Monocular Video (TPAMI 2022 · similarity 0.90)
- Monocular 3D Multi-Person Pose Estimation by Integrating Top-Down and Bottom-Up Networks (CVPR 2021 · similarity 0.89)
- Graph-Based 3D Multi-Person Pose Estimation Using Multi-View Images (ICCV 2021 · similarity 0.88)
- Distribution-Aware Single-Stage Models for Multi-Person 3D Pose Estimation (CVPR 2022 · similarity 0.88)
- HMOR: Hierarchical Multi-Person Ordinal Relations for Monocular Multi-Person 3D Pose Estimation (ECCV 2020 · similarity 0.87)
- 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.
