Research map / Graph learning / Graph self-supervised learning
TranSG: Transformer-Based Skeleton Graph Prototype Contrastive Learning With Structure-Trajectory Prompted Reconstruction for Person Re-Identification
Haocong Rao, 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 graph-based full-relation learning with SGT to the raw-joint baseline (ablation) | TranSG · BIWI-S | mAP: 9.3 -> 19.0 | Improved |
| Add graph-based full-relation learning with SGT to the raw-joint baseline (ablation) | TranSG · BIWI-S | R1: 24.8 -> 42.4 | Improved |
| Replace direct supervised learning with graph prototype contrastive learning (ablation) | SGT · BIWI-S | mAP: 19.0 -> 26.7 | Improved |
| Replace direct supervised learning with graph prototype contrastive learning (ablation) | SGT · BIWI-S | R1: 42.4 -> 66.6 | Improved |
| Add structure-trajectory prompted reconstruction (ablation) | SGT + GPC · BIWI-S | mAP: 26.7 -> 30.1 | Improved |
| Add structure-trajectory prompted reconstruction (ablation) | SGT + GPC · BIWI-S | R1: 66.6 -> 68.7 | Improved |
| Use naive prototype contrastive learning on original sequences instead of the raw-joint baseline (ablation) | TranSG ablation · BIWI-S | mAP: 9.3 -> 11.3 | Improved |
| Use naive prototype contrastive learning on original sequences instead of the raw-joint baseline (ablation) | TranSG ablation · BIWI-S | R1: 24.8 -> 38.1 | Improved |
| Use TranSG on unlabeled skeletons rather than SPC-MGR | Unsupervised person re-identification · BIWI-S | mAP: 16.0 -> 14.2 | Hurt |
| Use TranSG on unlabeled skeletons rather than SPC-MGR | Unsupervised person re-identification · BIWI-S | R1: 34.1 -> 42.2 | Improved |
Later papers that compared with this method
1 later accepted paper reports TranSG as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Motif Guided Graph Transformers with Combinatorial Skeleton Prototype Learning for Skeleton-Based Person Re-Identification (AAAI 2025 · better on 12 of 12 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Hierarchical Skeleton Meta-Prototype Contrastive Learning with Hard Skeleton Mining for Unsupervised Person Re-identification (IJCV 2023 · similarity 0.88)
- Motif Guided Graph Transformers with Combinatorial Skeleton Prototype Learning for Skeleton-Based Person Re-Identification (AAAI 2025 · similarity 0.88)
- Person Re-Identification by Unsupervised $$\ell _1$$ Graph Learning (ECCV 2016 · similarity 0.85)
- Learning Robust Event-Guided Representations for Person Re-Identification (IJCV 2026 · similarity 0.84)
- FeatER: An Efficient Network for Human Reconstruction via Feature Map-Based TransformER (CVPR 2023 · similarity 0.83)
- Graph Correspondence Transfer for Person Re-Identification (AAAI 2018 · similarity 0.83)
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.
