Research map / 3D vision / Human pose estimation
PPT: Token-Pruned Pose Transformer for Monocular and Multi-View Human Pose Estimation
Haoyu Ma, Zhe Wang, Yifei Chen, Deying Kong, Liangjian Chen, Xingwei Liu, Xiangyi Yan, Hao Tang, Xiaohui Xie
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 token pruning (HTI, keep ratio 0.7) to TokenPose-S | TokenPose-S · COCO val | AP: 72.5 -> 72.2 | No clear effect |
| add token pruning to TokenPose-S | TokenPose-S · COCO val | GFLOPs: 2.2 -> 1.6 | Improved |
| add token pruning to TokenPose-B | TokenPose-B · COCO val | AP: 74.7 -> 74.4 | No clear effect |
| add token pruning to TokenPose-L/D6 | TokenPose-L/D6 · COCO val | AP: 75.4 -> 75.2 | No clear effect |
| add token pruning to TokenPose-S | TokenPose-S · MPII val | PCKh@0.5: 86.2 -> 87.3 | Improved |
| add token pruning to TokenPose-B | TokenPose-B · MPII val | PCKh@0.5: 89.7 -> 89.8 | No clear effect |
| add token pruning to TokenPose-S | TokenPose-S · COCO val | throughput: 651 -> 842 | Improved |
| add token pruning to TokenPose-S | TokenPose-S · COCO val | FPS: 120 -> 123 | No clear effect |
| lower keep ratio from 0.7 to 0.6 (ablation) | PPT-S · COCO val | AP: 72.2 -> 71.8 | Hurt |
| raise keep ratio from 0.7 to 0.8 (ablation) | PPT-S · COCO val | AP: 72.2 -> 72.0 | Hurt |
| lower keep ratio from 0.7 to 0.6 (ablation) | PPT-S · MPII val | PCKh@0.5: 87.3 -> 86.7 | Hurt |
| add human area fusion (token pruning) before cross-view transformers (ablation) | TokenPose+Transformers · Human3.6M 2D pose | JDR: 97.4 -> 97.9 | Improved |
Closest papers by meaning
- Hourglass Tokenizer for Efficient Transformer-Based 3D Human Pose Estimation (CVPR 2024 · similarity 0.88)
- Optimizing Human Pose Estimation Through Focused Human and Joint Regions (AAAI 2025 · similarity 0.87)
- SVTformer: Spatial-View-Temporal Transformer for Multi-View 3D Human Pose Estimation (AAAI 2025 · similarity 0.86)
- FusionFormer: A Concise Unified Feature Fusion Transformer for 3D Pose Estimation (AAAI 2024 · similarity 0.86)
- ExtPose: Robust and Coherent Pose Estimation by Extending ViTs (ICML 2025 · similarity 0.86)
- SDPose: Tokenized Pose Estimation via Circulation-Guide Self-Distillation (CVPR 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.
