Research map / Label-efficient and robust learning / Recognition and long-tail
Harmonious Attention Network for Person Re-Identification
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 |
|---|---|---|---|
| jointly learn soft and hard attention (ablation) | HA-CNN · Market-1501 SQ | Rank-1: 84.7 -> 91.2 | Helped |
| jointly learn soft and hard attention (ablation) | HA-CNN · DukeMTMC-ReID SQ | Rank-1: 72.4 -> 80.5 | Helped |
| add soft spatial attention (ablation) | HA-CNN · Market-1501 SQ | Rank-1: 84.7 -> 85.5 | Helped |
| add soft channel attention (ablation) | HA-CNN · Market-1501 SQ | Rank-1: 84.7 -> 86.8 | Helped |
| combine soft spatial and channel attention (ablation) | HA-CNN · Market-1501 SQ | Rank-1: 86.8 -> 88.5 | Helped |
| add hard regional attention to soft spatial and channel attention (ablation) | HA-CNN · Market-1501 SQ | Rank-1: 88.5 -> 91.2 | Helped |
| add cross-attention interaction learning (ablation) | HA-CNN · Market-1501 SQ | Rank-1: 86.6 -> 91.2 | Helped |
| add cross-attention interaction learning (ablation) | HA-CNN · Market-1501 SQ | mAP: 66.2 -> 75.7 | Helped |
| add cross-attention interaction learning (ablation) | HA-CNN · DukeMTMC-ReID SQ | Rank-1: 74.0 -> 80.5 | Helped |
| add cross-attention interaction learning (ablation) | HA-CNN · DukeMTMC-ReID SQ | mAP: 55.4 -> 63.8 | Helped |
| combine global and local features (ablation) | HA-CNN · Market-1501 SQ | Rank-1: 89.9 -> 91.2 | Helped |
| combine global and local features (ablation) | HA-CNN · DukeMTMC-ReID SQ | Rank-1: 78.9 -> 80.5 | Helped |
Closest papers by meaning
- Deep Neural Networks with Inexact Matching for Person Re-Identification (NeurIPS 2016 · similarity 0.91)
- Relation-Aware Global Attention for Person Re-Identification (CVPR 2020 · similarity 0.90)
- AXM-Net: Implicit Cross-Modal Feature Alignment for Person Re-identification (AAAI 2022 · similarity 0.90)
- Re-Identification With Consistent Attentive Siamese Networks (CVPR 2019 · similarity 0.90)
- Joint Learning of Single-Image and Cross-Image Representations for Person Re-Identification (CVPR 2016 · similarity 0.89)
- Interaction-And-Aggregation Network for Person Re-Identification (CVPR 2019 · similarity 0.89)
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.