Research map / Label-efficient and robust learning / Domain adaptation and generalization
Deep Credible Metric Learning for Unsupervised Domain Adaptation Person Re-identification
Guangyi Chen, Yuhao Lu, Jiwen Lu, Jie Zhou
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 |
|---|---|---|---|
| add GAN-translated source images for pre-training (ablation) | UDA (reproduced) · Market-1501 to DukeMTMC-reID | mAP: 54.4 -> 60.4 | Improved |
| add GAN-translated source images for pre-training (ablation) | UDA (reproduced) · DukeMTMC-reID to Market-1501 | mAP: 56.5 -> 70.5 | Improved |
| add credible anchor mining and IMS loss to UDA (no GAN) (ablation) | UDA (reproduced) · Market-1501 to DukeMTMC-reID | mAP: 54.4 -> 60.2 | Improved |
| add credible anchor mining and IMS loss to UDA (no GAN) (ablation) | UDA (reproduced) · DukeMTMC-reID to Market-1501 | mAP: 56.5 -> 69.2 | Improved |
| add instance margin spreading loss (ablation) | UDA + GAN · Market-1501 to DukeMTMC-reID | mAP: 60.4 -> 62.2 | Improved |
| add instance margin spreading loss (ablation) | UDA + GAN · DukeMTMC-reID to Market-1501 | mAP: 70.5 -> 71.3 | Improved |
| add credible anchor mining strategy (KNN) (ablation) | UDA + GAN + IMSLoss · Market-1501 to DukeMTMC-reID | mAP: 62.2 -> 63.3 | Improved |
| add credible anchor mining strategy (KNN) (ablation) | UDA + GAN + IMSLoss · DukeMTMC-reID to Market-1501 | mAP: 71.3 -> 72.6 | Improved |
| use prototype similarity instead of KNN similarity for credibility (ablation) | DCML · Market-1501 to DukeMTMC-reID | mAP: 63.3 -> 63.5 | No clear effect |
| use prototype similarity instead of KNN similarity for credibility (ablation) | DCML · DukeMTMC-reID to Market-1501 | mAP: 72.6 -> 72.3 | No clear effect |
| use DCML instead of SSG++ | unsupervised domain adaptation re-ID · Market-1501 to DukeMTMC-reID | mAP: 60.3 -> 63.3 | Improved |
| use DCML instead of PAST | unsupervised domain adaptation re-ID · CUHK03 to DukeMTMC-reID | mAP: 51.8 -> 56.9 | Improved |
Closest papers by meaning
- A Novel Unsupervised Camera-Aware Domain Adaptation Framework for Person Re-Identification (ICCV 2019 · similarity 0.89)
- Unsupervised Multi-Source Domain Adaptation for Person Re-Identification (CVPR 2021 · similarity 0.89)
- Reliability Exploration with Self-Ensemble Learning for Domain Adaptive Person Re-identification (AAAI 2022 · similarity 0.88)
- Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised Domain Adaptation on Person Re-identification (ICLR 2020 · similarity 0.88)
- Self-Training With Progressive Augmentation for Unsupervised Cross-Domain Person Re-Identification (ICCV 2019 · similarity 0.88)
- Self-Similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-Identification (ICCV 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.
