Research map / 3D vision / Human pose estimation
PoseExaminer: Automated Testing of Out-of-Distribution Robustness in Human Pose and Shape Estimation
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 |
|---|---|---|---|
| Fine-tune SPIN on PoseExaminer-discovered failure modes | SPIN · 3DPW test | MPJPE: 94.11 -> 88.66 | Helped |
| Fine-tune SPIN on PoseExaminer-discovered failure modes | SPIN · cAIST | MPJPE: 108.09 -> 98.88 | Helped |
| Fine-tune SPIN on PoseExaminer-discovered failure modes | SPIN · cAIST-EXT | MPJPE: 133.65 -> 120.98 | Helped |
| Fine-tune HMR-EFT on PoseExaminer-discovered failure modes | HMR-EFT · cAIST | MPJPE: 98.55 -> 88.53 | Helped |
| Fine-tune PARE on PoseExaminer-discovered failure modes | PARE · 3DPW test | MPJPE: 81.81 -> 77.46 | Helped |
| Fine-tune PARE on PoseExaminer-discovered failure modes | PARE · cAIST | MPJPE: 99.15 -> 87.44 | Helped |
| Fine-tune PARE on PoseExaminer-discovered failure modes | PARE · cAIST-EXT | MPJPE: 117.45 -> 109.82 | Helped |
| Use easy rather than mixed PoseExaminer difficulty for fine-tuning (ablation) | PARE · 3DPW test | MPJPE: 81.81 -> 74.44 | Helped |
| Use easy rather than mixed PoseExaminer difficulty for fine-tuning (ablation) | PARE · cAIST | MPJPE: 99.15 -> 105.09 | Hurt |
| Use hard rather than mixed PoseExaminer difficulty for fine-tuning (ablation) | PARE · cAIST | MPJPE: 99.15 -> 86.92 | Helped |
| Replace failure-mode samples with random samples from the parameter space (ablation) | PARE · cAIST | MPJPE: 99.15 -> 95.10 | Helped |
| Increase failure-subspace evaluation samples from 200 to 5000 (ablation) | PARE · PARE failure subspaces | meanMPJPE: 150.06 -> 150.41 | No clear effect |
Closest papers by meaning
- $\text{Di}^2\text{Pose}$: Discrete Diffusion Model for Occluded 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.88)
- PoseIRM: Enhance 3D Human Pose Estimation on Unseen Camera Settings via Invariant Risk Minimization (CVPR 2024 · similarity 0.87)
- CUPS: Improving Human Pose-Shape Estimators with Conformalized Deep Uncertainty (ICML 2025 · similarity 0.87)
- HiPoser: 3D Human Pose Estimation with Hierarchical Shared Learning at Parts-Level Using Inertial Measurement Units (AAAI 2025 · similarity 0.87)
- TokenHMR: Advancing Human Mesh Recovery with a Tokenized Pose Representation (CVPR 2024 · similarity 0.87)
- GRPose: Learning Graph Relations for Human Image Generation with Pose Priors (AAAI 2025 · similarity 0.86)
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.