Research map / 3D vision / Human pose estimation
The Devil Is in the Details: Delving Into Unbiased Data Processing for Human Pose Estimation
Junjie Huang, Zheng Zhu, Feng Guo, Guan Huang
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 |
|---|---|---|---|
| Apply UDP to HRNet-W32-256×192 | HRNet · COCO test-dev | AP: 73.5 -> 75.2 (+1.7) | Improved |
| Apply UDP to HRNet-W48-384×288 | HRNet · COCO test-dev | AP: 75.5 -> 76.5 (+1.0) | Improved |
| Apply UDP to HigherHRNet-W32-512×512 | HigherHRNet · COCO test-dev | AP: 64.1 -> 66.8 (+2.7) | Improved |
| Apply UDP to HigherHRNet-W32-512×512 | HigherHRNet · CrowdPose test | AP: 65.0 -> 66.1 (+1.1) | Improved |
| Add UCST without flipping testing (ablation) | HRNet-W32-256×192 · COCO val | AP: 74.5 -> 74.4 | No clear effect |
| Add flipping testing without UCST (ablation) | HRNet-W32-256×192 · COCO val | AP: 74.5 -> 73.3 | Hurt |
| Use UCST with flipping testing (ablation) | HRNet-W32-256×192 · COCO val | AP: 74.4 -> 75.7 | Improved |
| Replace UCST with the one-pixel SNOOP correction (ablation) | HRNet-W32-256×192 with flipping testing · COCO val | AP: 75.6 versus 75.7 | Mixed |
| Add extra compensation to SNOOP (ablation) | HRNet-W32-256×192 with flipping testing · COCO val | AP: 75.6 -> 75.8 | Improved |
| Add UKFT-CCRF to flipping testing without UCST (ablation) | HRNet-W32-256×192 · COCO val | AP: 74.5 versus 74.5 | No clear effect |
| Add UKFT-CCRF to flipping testing with UCST (ablation) | HRNet-W32-256×192 · COCO val | AP: 75.7 -> 76.8 | Improved |
| Use UKFT-CF instead of UKFT-CCRF with UCST (ablation) | HRNet-W32-256×192 · COCO val | AP: 76.8 versus 76.8 | No clear effect |
Closest papers by meaning
- Kinematic-Structure-Preserved Representation for Unsupervised 3D Human Pose Estimation (AAAI 2020 · similarity 0.87)
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · similarity 0.87)
- Distribution-Aware Coordinate Representation for Human Pose Estimation (CVPR 2020 · similarity 0.87)
- A Comparative Assessment of Accuracy in Video-Based Monocular Human Pose Estimation Frameworks (TPAMI 2026 · similarity 0.87)
- PoseTrans: A Simple yet Effective Pose Transformation Augmentation for Human Pose Estimation (ECCV 2022 · similarity 0.87)
- Video-Based Human Pose Regression via Decoupled Space-Time Aggregation (CVPR 2024 · 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.
