Research map / 3D vision / Human pose estimation
PoseTrans: A Simple yet Effective Pose Transformation Augmentation for Human Pose Estimation
Wentao Jiang, Sheng Jin, Wentao Liu, Chen Qian, Ping Luo, Si Liu
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 PoseTrans augmentation | HRNet-W32 (256x192) · MS-COCO val | AP: 74.4 -> 75.5 | Improved |
| add PoseTrans augmentation | SBL-ResNet-50 · MS-COCO val | AP: 70.4 -> 72.3 | Improved |
| add PoseTrans augmentation | HigherHRNet-W32 (bottom-up) · MS-COCO val | AP: 67.1 -> 68.4 | Improved |
| add PoseTrans augmentation | HRNet-W32 (256x192) · PoseTrack2018 val | Total AP: 83.0 -> 84.9 | Improved |
| add PoseTrans augmentation | HRNet-W32 (256x192) · MS-COCO val | AP_BAL: 65.4 -> 67.9 | Improved |
| add PoseTrans on top of DarkPose | HRNet-W32 + Dark · MS-COCO val | AP: 75.6 -> 76.0 | Improved |
| use PoseTrans instead of Cutout, GridMask, Photometric Distortion, InstaBoost or ASDA | HRNet-W32 · MS-COCO val | AP: 74.7 -> 75.5 | Improved |
| add PTM instead of over-sampling (ablation) | HRNet-W32 · MS-COCO val | AP: 74.3 -> 75.5 | Improved |
| add PTM instead of re-weighting (ablation) | HRNet-W32 · MS-COCO val | AP_BAL: 65.8 -> 67.9 | Improved |
| add PCM rarest-pose selection (ablation) | PoseTrans · MS-COCO val | AP_BAL: 66.1 -> 67.9 | Improved |
| increase candidate pool from T=3 to T=5 (ablation) | PoseTrans · MS-COCO val | AP: 75.2 -> 75.5 | Improved |
| add pose discriminator (ablation) | PoseTrans · MS-COCO val | AP: 75.0 -> 75.5 | Improved |
Closest papers by meaning
- PoseAug: A Differentiable Pose Augmentation Framework for 3D Human Pose Estimation (CVPR 2021 · similarity 0.90)
- PoseTrack: A Benchmark for Human Pose Estimation and Tracking (CVPR 2018 · similarity 0.89)
- PoseFix: Model-Agnostic General Human Pose Refinement Network (CVPR 2019 · similarity 0.89)
- FlexPose: Pose Distribution Adaptation with Limited Guidance (AAAI 2025 · similarity 0.88)
- A Dual-Augmentor Framework for Domain Generalization in 3D Human Pose Estimation (CVPR 2024 · similarity 0.87)
- PoseIRM: Enhance 3D Human Pose Estimation on Unseen Camera Settings via Invariant Risk Minimization (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.
