Research map / Label-efficient and robust learning / Recognition and long-tail
Person Re-identification with Deep Similarity-Guided Graph Neural Network
Yantao Shen, Hongsheng Li, Shuai Yi, Dapeng Chen, Xiaogang Wang
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 SGGNN to the base model (ablation) | siamese base model · Market-1501 | mAP (%): 76.4 -> 82.8 | Improved |
| add SGGNN to the base model (ablation) | siamese base model · CUHK03 | mAP (%): 88.9 -> 94.3 | Improved |
| add SGGNN to the base model (ablation) | siamese base model · DukeMTMC | mAP (%): 61.8 -> 68.2 | Improved |
| add similarity guidance (SG) to SGGNN (ablation) | SGGNN · Market-1501 | top-1 (%): 90.6 -> 92.3 | Improved |
| add similarity guidance (SG) to SGGNN (ablation) | SGGNN · DukeMTMC | mAP (%): 67.3 -> 68.2 | Improved |
| use SGGNN instead of k-reciprocal re-ranking (ablation) | base model · Market-1501 | mAP (%): 78.5 -> 82.8 | Improved |
| use SGGNN instead of random walk (ablation) | base model · DukeMTMC | mAP (%): 66.0 -> 68.2 | Improved |
| add SGGNN training for l2 feature distance (ablation) | base model · Market-1501 (l2 feature distance) | mAP (%): 74.6 -> 76.7 | Improved |
| add SGGNN training for l2 feature distance (ablation) | base model · CUHK03 (l2 feature distance) | top-1 (%): 91.0 -> 93.6 | Improved |
| reduce top-K from 100 to 50 (ablation) | SGGNN · DukeMTMC | mAP (%): 68.2 -> 66.6 | Hurt |
| increase top-K from 100 to 150 (ablation) | SGGNN · DukeMTMC | mAP (%): 68.2 -> 71.8 | Improved |
| change number of message passing steps K from 4 to 3 or 5 (ablation) | SGGNN · Market-1501 | mAP (%): 82.8 -> 82.0 | Hurt |
Closest papers by meaning
- Group Consistent Similarity Learning via Deep CRF for Person Re-Identification (CVPR 2018 · similarity 0.88)
- Learning Context Graph for Person Search (CVPR 2019 · similarity 0.88)
- Deep Constrained Dominant Sets for Person Re-Identification (ICCV 2019 · similarity 0.88)
- Learning Hybrid Relationships for Person Re-identification (AAAI 2021 · similarity 0.88)
- Consistent-Aware Deep Learning for Person Re-Identification in a Camera Network (CVPR 2017 · similarity 0.88)
- Cheb-GR: Rethinking K-nearest Neighbor Search in Re-ranking for Person Re-identification (CVPR 2025 · similarity 0.86)
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.
