Research map / Label-efficient and robust learning / Recognition and long-tail
Robust Person Re-Identification by Modelling Feature Uncertainty
Tianyuan Yu, Da Li, Yongxin Yang, Timothy M. Hospedales, Tao Xiang
Read the paperCheck your idea against 137,217 papers
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 |
|---|---|---|---|
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · Market-1501, 20% random label noise | mAP: 45.36 -> 53.40 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · CUHK01, 10% random label noise | mAP: 52.02 -> 63.36 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · CUHK03, 50% random label noise | mAP: 6.44 -> 10.61 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · Market-1501, 10% patterned label noise | mAP: 25.87 -> 27.04 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · CUHK01, 10% patterned label noise | mAP: 42.87 -> 52.72 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · CUHK03, 10% patterned label noise | mAP: 10.10 -> 10.87 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · Market-1501, closed-world, no added noise | mAP: 67.66 -> 70.82 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · CUHK01, closed-world, no added noise | mAP: 60.70 -> 70.70 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · DukeMTMC-ReID, closed-world, no added noise | mAP: 54.17 -> 55.98 | Improved |
| model each feature as a Gaussian with learned variance (DistributionNet) | ResNet50 ReID model · Market-1501 open-world set verification, FTR 0.1% | TTR: 43.03 -> 55.76 | Improved |
| use CleanNet with 10% clean reference set and removal of 20% suspected noisy samples instead of baseline | ResNet-Baseline · Market-1501, 20% random label noise | mAP: 45.36 -> 44.22 | Hurt |
| add Bootstrap hard label refinement | ResNet-Baseline · Market-1501, closed-world, no added noise | mAP: 67.66 -> 69.09 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CleanNet — CleanNet: Transfer Learning for Scalable Image Classifier Training With Label Noise (CVPR 2018) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Adversarially Occluded Samples for Person Re-Identification (CVPR 2018 · similarity 0.90)
- Exploring the Camera Bias of Person Re-identification (ICLR 2025 · similarity 0.89)
- Pose Transferrable Person Re-Identification (CVPR 2018 · similarity 0.89)
- A Multi-Task Deep Network for Person Re-Identification (AAAI 2017 · similarity 0.88)
- Benchmarks for Corruption Invariant Person Re-identification (NeurIPS 2021 · similarity 0.88)
- Learning Disentangled Representation for Robust Person Re-identification (NeurIPS 2019 · 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.
