Research map / Label-efficient and robust learning / Recognition and long-tail
Eliminating Background-Bias for Robust 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 |
|---|---|---|---|
| Add person-region guided pooling (ablation) | Person re-identification model · Market-1501 | Top-1 accuracy: 75.6 -> 80.5 | Helped |
| Add random-background data augmentation | Proposed person re-identification model · CUHK03 | Top-1 accuracy: 91.7 -> 92.5 | Helped |
| Add random-background data augmentation | Proposed person re-identification model · CUHK01 | Top-1 accuracy: 80.7 -> 82.5 | Helped |
| Add random-background data augmentation | Proposed person re-identification model · VIPeR | Top-1 accuracy: 50.6 -> 51.9 | Helped |
| Add random-background data augmentation | Proposed person re-identification model · 3DPeS | Top-1 accuracy: 64.1 -> 65.6 | Helped |
| Add random-background data augmentation | Proposed person re-identification model · Market-1501 | Top-1 accuracy: 80.5 -> 81.2 | Helped |
| Add the third part-specific branch to the guided-pooling network (ablation) | Two-branch upper+lower model · Market-1501 | Top-1 accuracy: 80.0 -> 80.5 | Helped |
| Use online rather than offline background augmentation (ablation) | Main network · Market-1501 | Top-1 accuracy: 76.1 -> 79.2 | Helped |
| Set online background replacement probability to 0.5 rather than 0.25 (ablation) | Main network · Market-1501 | Top-1 accuracy: 79.2 -> 79.5 | No clear effect |
| Set online background replacement probability to 0.75 rather than 0.5 (ablation) | Main network · Market-1501 | Top-1 accuracy: 79.5 -> 78.8 | Hurt |
| Train with random-background augmentation (ablation) | DGD · CUHK03 random-background test set | Top-1 accuracy: 87.2 -> 64.0 | Hurt |
| Use the proposed method instead of DGD | Person re-identification model · CUHK03 random-background test set | Top-1 accuracy: 64.0 -> 91.1 | Helped |
Closest papers by meaning
- Deep Neural Networks with Inexact Matching for Person Re-Identification (NeurIPS 2016 · similarity 0.90)
- Resource Aware Person Re-Identification Across Multiple Resolutions (CVPR 2018 · similarity 0.89)
- A Multi-Task Deep Network for Person Re-Identification (AAAI 2017 · similarity 0.88)
- Point to Set Similarity Based Deep Feature Learning for Person Re-Identification (CVPR 2017 · similarity 0.88)
- Human Semantic Parsing for Person Re-Identification (CVPR 2018 · similarity 0.88)
- Exploring the Camera Bias of Person Re-identification (ICLR 2025 · similarity 0.88)
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.