Research map / Label-efficient and robust learning / Recognition and long-tail
Hierarchical Skeleton Meta-Prototype Contrastive Learning with Hard Skeleton Mining for Unsupervised Person Re-identification
Haocong Rao, Cyril Leung, Chunyan Miao
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 direct skeleton prototype contrastive learning (DPC) (ablation) | raw-skeleton baseline · KGBD | Rank-1: 20.5 -> 53.7 | Improved |
| add direct skeleton prototype contrastive learning (DPC) (ablation) | raw-skeleton baseline · KS20 | Rank-1: 17.0 -> 63.3 | Improved |
| replace DPC with meta-prototype contrastive learning (MPC) (ablation) | joint-level skeletons · KS20 | Rank-1: 63.3 -> 63.9 | Improved |
| add hierarchical skeleton representations (ablation) | MPC · KS20 | Rank-1: 63.9 -> 66.4 | Improved |
| add hierarchical skeleton representations (ablation) | MPC · BIWI-W | Rank-1: 22.4 -> 24.2 | Improved |
| add hard skeleton mining (HSM) (ablation) | MPC (joint-level) · IAS-B | Rank-1: 37.3 -> 42.8 | Improved |
| add HSM (ablation) | Hi + MPC · BIWI-S | Rank-1: 40.4 -> 47.5 | Improved |
| add HSM (ablation) | Hi + MPC · KS20 | Rank-1: 66.4 -> 69.6 | Improved |
| add HSM to limb-level only representation (ablation) | Hi-MPC · BIWI-W | Rank-1: 18.0 -> 17.5 | Hurt |
| combine limb, component and joint levels (MSMR) vs joint-only (ablation) | Hi-MPC with HSM · IAS-B | Rank-1: 41.7 -> 48.2 | Improved |
| increase meta-transformation heads from 1 to 16 (ablation) | Hi-MPC · KS20 | Rank-1: 67.2 -> 70.7 | Improved |
| increase meta-transformation heads from 1 to 16 (ablation) | Hi-MPC · BIWI-S | mAP: 15.3 -> 16.9 | No clear effect |
Closest papers by meaning
- Motif Guided Graph Transformers with Combinatorial Skeleton Prototype Learning for Skeleton-Based Person Re-Identification (AAAI 2025 · similarity 0.91)
- A Self-Supervised Gait Encoding Approach With Locality-Awareness for 3D Skeleton Based Person Re-Identification (TPAMI 2021 · similarity 0.88)
- TranSG: Transformer-Based Skeleton Graph Prototype Contrastive Learning With Structure-Trajectory Prompted Reconstruction for Person Re-Identification (CVPR 2023 · similarity 0.88)
- Skeletons Speak Louder than Text: A Motion-Aware Pretraining Paradigm for Video-Based Person Re-Identification (AAAI 2026 · similarity 0.87)
- Spindle Net: Person Re-Identification With Human Body Region Guided Feature Decomposition and Fusion (CVPR 2017 · similarity 0.86)
- From Poses to Identity: Training-Free Person Re-Identification via Feature Centralization (CVPR 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.
