Research map / Label-efficient and robust learning / Recognition and long-tail
Easy Identification From Better Constraints: Multi-Shot Person Re-Identification From Reference Constraints
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 reference constraints with moderate positive mining and a kernelized metric | Person re-identification · CAVIAR | Rank-1 accuracy: 53.54 -> 73.36 | Helped |
| Use viewpoint-alignment reference constraints with a kernelized metric | Person re-identification · PRID 2011 | Rank-1 accuracy: 51.00 -> 70.90 | Helped |
| Use class-based discriminative-space reference constraints with a linear metric | Person re-identification · iLIDS-VID | Rank-1 accuracy: 30.66 -> 44.67 | Helped |
| Use class-based discriminative-space reference constraints with a kernelized metric | Person re-identification · Market-1501, Sing-Q | Rank-1 accuracy: 59.47 -> 63.20 | Helped |
| Use reference constraints with viewpoint alignment and a linear metric | Person re-identification · Market-1501, Multi-Q | Rank-1 accuracy: 57.70 -> 72.40 | Helped |
| Switch from a linear metric to a kernelized metric (ablation) | Reference-constraint methods · CAVIAR | Rank-1 accuracy: TC-L 65.25 -> TC-K 66.80; TM-L 70.90 -> TM-K 73.36; TA-L 68.73 -> TA-K 61.02 | Mixed |
| Use viewpoint-alignment references instead of class-based references (ablation) | Reference-constraint metric learning · PRID 2011 | Rank-1 accuracy: TC-L 70.10; TM-L 64.80; TA-L 70.40; TC-K 66.90; TM-K 65.10; TA-K 70.90 | Mixed |
| Use viewpoint-alignment references instead of class-based references (ablation) | Reference-constraint metric learning · iLIDS-VID | Rank-1 accuracy: TC-L 44.67; TM-L 38.67; TA-L 42.67; TC-K 37.33; TM-K 39.33; TA-K 42.00 | Mixed |
Closest papers by meaning
- Online Joint Multi-Metric Adaptation From Frequent Sharing-Subset Mining for Person Re-Identification (CVPR 2020 · similarity 0.87)
- Learning a Discriminative Null Space for Person Re-Identification (CVPR 2016 · similarity 0.87)
- Improving Person Re-Identification via Pose-Aware Multi-Shot Matching (CVPR 2016 · similarity 0.86)
- Similarity Learning With Spatial Constraints for Person Re-Identification (CVPR 2016 · similarity 0.86)
- Top-Push Video-Based Person Re-Identification (CVPR 2016 · similarity 0.86)
- One-Shot Metric Learning for Person Re-Identification (CVPR 2017 · 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.