Research map / Label-efficient and robust learning / Recognition and long-tail
Human-in-the-Loop Vehicle ReID
Zepeng Li, Dongxiang Zhang, Yanyan Shen, Gang Chen
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 five rounds of IRIN feedback with perfect positive matches | IRIN on Veri776 · Veri776 | mAP: 81.6% to 95.2% | Improved |
| Use IRIN with oracle feedback instead of machine-only TransReID | Vehicle ReID · Veri776 | mAP: 80.8% to 95.2% | Improved |
| Use minimum-distance rather than maximum-distance positive selection (ablation) | IRIN · Veri776 | mAP: 94.5 -> 95.2; 92.1 -> 93.2; 88 -> 89.9 | Improved |
| Use active-learning candidate sampling instead of random sampling (ablation) | IRIN · Veri776 | mAP: 93.8 -> 95.2; 91.8 -> 93.2; 87.7 -> 89.9 | Improved |
| Use supervised contrastive loss instead of triplet loss (ablation) | IRIN · Veri776 | mAP: 92.3 -> 95.2; 90.6 -> 93.2; 85.9 -> 89.9 | Improved |
| Add a hard negative memory bank (ablation) | IRIN · Veri776 | mAP: 92.7 -> 95.2; 91.3 -> 93.2; 86.8 -> 89.9 | Improved |
| Increase ILOS distillation-loss weight from 0.1 to 0.2 | ILOS · Veri776 | mAP: 85.7 to 86.4 | Improved |
| Increase ILOS distillation-loss weight from 0.2 to 0.3 | ILOS · Veri776 | mAP: 86.4 to 86.3 | Hurt |
Closest papers by meaning
- Human-in-the-Loop Feature Selection (AAAI 2019 · similarity 0.82)
- Auto-ReID: Searching for a Part-Aware ConvNet for Person Re-Identification (ICCV 2019 · similarity 0.81)
- Interaction-And-Aggregation Network for Person Re-Identification (CVPR 2019 · similarity 0.81)
- Instruct-ReID: A Multi-purpose Person Re-identification Task with Instructions (CVPR 2024 · similarity 0.81)
- MSINet: Twins Contrastive Search of Multi-Scale Interaction for Object ReID (CVPR 2023 · similarity 0.81)
- Pose-Aware Multi-Level Feature Network for Human Object Interaction Detection (ICCV 2019 · similarity 0.81)
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.
