Research map / 3D vision / Human pose estimation
PoseFix: Model-Agnostic General Human Pose Refinement Network
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 |
|---|---|---|---|
| Train the refiner model-agnostically on realistically synthesized errors rather than fine-tuning it end-to-end with CPN | CPN pose refinement on MS COCO validation · MS COCO validation | AP: 70.1 (+0.4) -> 72.1 (+2.4) | Helped |
| Replace CPN RefineNet with model-agnostic PoseFix | CPN GlobalNet refinement · MS COCO validation | AP: 69.1 (+1.8) -> 71.5 (+4.2) | Helped |
| Apply PoseFix to Simple | Simple · PoseTrack 2018 validation | Total AP: 70.4 -> 75.0 | Helped |
| Use a coarse Gaussian-blob input rather than a fine input-pose representation (ablation) | PoseFix refinement pipeline · MS COCO validation | pose refinement accuracy | Helped |
| Estimate a fine-form output rather than a coarse Gaussian-heatmap output (ablation) | PoseFix refinement pipeline · MS COCO validation | pose refinement accuracy | Helped |
| Use coordinate-only supervision instead of the combined heatmap and coordinate losses (ablation) | C2F PoseFix · MS COCO validation | pose refinement accuracy | Hurt |
| Use heatmap-only supervision instead of the combined heatmap and coordinate losses (ablation) | C2F PoseFix · MS COCO validation | pose refinement accuracy | Hurt |
Closest papers by meaning
- PoseExaminer: Automated Testing of Out-of-Distribution Robustness in Human Pose and Shape Estimation (CVPR 2023 · similarity 0.89)
- Self-Correctable and Adaptable Inference for Generalizable Human Pose Estimation (CVPR 2023 · similarity 0.88)
- SynSP: Synergy of Smoothness and Precision in Pose Sequences Refinement (CVPR 2024 · similarity 0.88)
- Reconstructing Close Human Interaction with Appearance and Proxemics Reasoning (CVPR 2025 · similarity 0.87)
- PoseIRM: Enhance 3D Human Pose Estimation on Unseen Camera Settings via Invariant Risk Minimization (CVPR 2024 · similarity 0.85)
- GRPose: Learning Graph Relations for Human Image Generation with Pose Priors (AAAI 2025 · 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.