Research map / Label-efficient and robust learning / Recognition and long-tail
Cut out the Middleman: Revisiting Pose-based Gait Recognition
Yang Fu, Saihui Hou, Shibei Meng, Xuecai Hu, Chunshui Cao, Xu Liu, Yongzhen Huang
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 |
|---|---|---|---|
| use upstream pose-estimation heatmaps instead of heatmaps regenerated from keypoints (ablation) | GaitHeat · CASIA-B | mean rank-1 accuracy (%): 75.25 -> 88.17 | Improved |
| use upstream heatmaps instead of regenerated heatmaps (ablation) | GaitHeat · CCPG CL | rank-1 accuracy (%): 55.99 -> 73.59 | Improved |
| add Pose-Guided Heatmap Alignment (PGHA) (ablation) | GaitHeat · CASIA-B to CCPG, BG (cross-domain) | rank-1 accuracy (%): 14.59 -> 21.59 | Improved |
| add PGHA (ablation) | GaitHeat · CCPG to CCPG, CL | rank-1 accuracy (%): 70.63 -> 73.59 | Improved |
| add PGHA (ablation) | GaitHeat · CASIA-B in-domain | mean rank-1 accuracy (%): 89.61 -> 88.17 | Hurt |
| replace HRNet heatmaps with ViTPose heatmaps (GaitHeat++) | GaitHeat · SUSTech1K overall | rank-1 accuracy (%): 76.94 -> 90.00 | Improved |
| use heatmaps instead of skeletons as pose representation | pose-based gait recognition · SUSTech1K overall | rank-1 accuracy (%): 42.45 -> 76.94 | Improved |
| use heatmaps on cloth-changing cross-domain test (CASIA-B to CCPG) | GaitHeat · CASIA-B to CCPG, CL | rank-1 accuracy (%): 11.85 -> 5.63 | Hurt |
| heatmap-based pose method versus silhouette-based GaitGL on CASIA-B | GaitHeat++ · CASIA-B mean | rank-1 accuracy (%): 91.83 -> 95.94 | Improved |
Closest papers by meaning
- GPGait: Generalized Pose-based Gait Recognition (ICCV 2023 · similarity 0.93)
- SkeletonGait: Gait Recognition Using Skeleton Maps (AAAI 2024 · similarity 0.92)
- BiggerGait: Unlocking Gait Recognition with Layer-wise Representations from Large Vision Models (NeurIPS 2025 · similarity 0.88)
- Vocabulary-Guided Gait Recognition (NeurIPS 2025 · similarity 0.87)
- BigGait: Learning Gait Representation You Want by Large Vision Models (CVPR 2024 · similarity 0.87)
- Learning Visual Prompt for Gait Recognition (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.
