Research map / Label-efficient and robust learning / Recognition and long-tail
AANet: Attribute Attention Network for Person Re-Identifications
Check 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 |
|---|---|---|---|
| Add the part task to the global ID task (ablation) | AANet-50 · DukeMTMC-reID | mAP: 62.92 -> 66.35 | Helped |
| Add the part task to the global ID task (ablation) | AANet-50 · DukeMTMC-reID | Rank-1 accuracy: 80.18 -> 82.93 | Helped |
| Add attribute classification to global ID and part tasks (ablation) | AANet-50 · DukeMTMC-reID | mAP: 66.35 -> 67.28 | Helped |
| Add attribute classification to global ID and part tasks (ablation) | AANet-50 · DukeMTMC-reID | Rank-1 accuracy: 82.93 -> 83.29 | Helped |
| Use uncertainty learning to weight the three tasks (ablation) | AANet-50 with global ID, part, and attribute tasks · DukeMTMC-reID | mAP: 67.28 -> 70.47 | Helped |
| Use uncertainty learning to weight the three tasks (ablation) | AANet-50 with global ID, part, and attribute tasks · DukeMTMC-reID | Rank-1 accuracy: 83.29 -> 85.44 | Helped |
| Add the attribute attention map task (ablation) | AANet-50 with uncertainty-weighted global ID, part, and attribute tasks · DukeMTMC-reID | mAP: 70.47 -> 72.56 | Helped |
| Add the attribute attention map task (ablation) | AANet-50 with uncertainty-weighted global ID, part, and attribute tasks · DukeMTMC-reID | Rank-1 accuracy: 85.44 -> 86.42 | Helped |
| Integrate attribute, part, and global-ID learning | AANet-50 versus PCB · DukeMTMC-reID | mAP: 69.2 -> 72.56 | Helped |
| Integrate attribute, part, and global-ID learning | AANet-50 versus PCB · Market1501 | Rank-1 accuracy: 93.8 -> 93.89 | Helped |
| Use AANet rather than SPReID at the same backbone depth | ResNet-152 person re-ID · Market1501 | mAP: 83.36 -> 83.41 | Helped |
| Use AANet rather than SPReID with re-ranking | ResNet-152 person re-ID · Market1501 | mAP: 90.96 -> 92.38 | Helped |
Closest papers by meaning
- Relation-Aware Global Attention for Person Re-Identification (CVPR 2020 · similarity 0.88)
- Harmonious Attention Network for Person Re-Identification (CVPR 2018 · similarity 0.88)
- Multi-Level Factorisation Net for Person Re-Identification (CVPR 2018 · similarity 0.88)
- Relation Network for Person Re-Identification (AAAI 2020 · similarity 0.87)
- Attention-Aware Compositional Network for Person Re-Identification (CVPR 2018 · similarity 0.87)
- AD-Cluster: Augmented Discriminative Clustering for Domain Adaptive Person Re-Identification (CVPR 2020 · similarity 0.87)
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.