Research map / 3D vision / Human pose estimation
PoseAnchor: Robust Root Position Estimation for 3D Human Pose Estimation
Jun-Hee Kim, Jumin Han, Seong-Whan 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 |
|---|---|---|---|
| add direct root regression (Root+) to root-centered baseline (ablation) | MixSTE · Human3.6M | MPJPE: 40.9 -> 41.4 | Hurt |
| replace Root+ regression with zero-shot ITRR root estimation (ZS-anchor) (ablation) | MixSTE · Human3.6M | MRPE: 116.7 -> 100.5 | Improved |
| replace Root+ regression with zero-shot ITRR root estimation (ZS-anchor) (ablation) | MixSTE · Human3.6M | MPJPE: 41.4 -> 40.9 | No clear effect |
| add support-set-filtered training (PoseAnchor) to ZS-anchor (ablation) | MixSTE · Human3.6M | MPJPE: 40.9 -> 40.3 | Improved |
| add support-set-filtered training (PoseAnchor) to ZS-anchor (ablation) | MixSTE · Human3.6M | MRPE: 100.5 -> 96.7 | Improved |
| apply PoseAnchor instead of Root+ | Martinez et al. · Human3.6M | MRPE: 146.7 -> 133.7 | Improved |
| apply PoseAnchor instead of Root+ | VPose · Human3.6M | MRPE: 124.9 -> 112.0 | Improved |
| apply PoseAnchor instead of Root+ | GLA-GCN · Human3.6M | MRPE: 119.2 -> 106.4 | Improved |
| use PoseAnchor instead of Root+ under 2D noise sigma=10 (ablation) | MixSTE · Human3.6M GT 2D + noise | Abs-MPJPE: 524.7 -> 366.2 | Improved |
| use PoseAnchor instead of Root+ under 2D noise sigma=2 (ablation) | MixSTE · Human3.6M GT 2D + noise | Abs-MPJPE: 117.4 -> 82.5 | Improved |
| use PoseAnchor instead of Root+ without noise (ablation) | MixSTE · Human3.6M GT 2D + noise | Abs-MPJPE: 85.6 -> 82.2 | No clear effect |
| use ITRR support set to filter joints (ablation) | ZS-anchor · Human3.6M | 2D detection error (pixels): 79.9 vs 13.1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- MixSTE — MixSTE: Seq2seq Mixed Spatio-Temporal Encoder for 3D Human Pose Estimation in Video (CVPR 2022) · better on 2 of 2 shared results
- Ray3D — Ray3D: Ray-Based 3D Human Pose Estimation for Monocular Absolute 3D Localization (CVPR 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- A Simple yet Effective Baseline for 3D Human Pose Estimation (ICCV 2017 · similarity 0.88)
- PoseFormerV2: Exploring Frequency Domain for Efficient and Robust 3D Human Pose Estimation (CVPR 2023 · similarity 0.87)
- PoseAug: A Differentiable Pose Augmentation Framework for 3D Human Pose Estimation (CVPR 2021 · similarity 0.86)
- A Dual-Augmentor Framework for Domain Generalization in 3D Human Pose Estimation (CVPR 2024 · similarity 0.85)
- PoseIRM: Enhance 3D Human Pose Estimation on Unseen Camera Settings via Invariant Risk Minimization (CVPR 2024 · similarity 0.85)
- ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.85)
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.
