Research map / Label-efficient and robust learning / Domain adaptation and generalization
Unsupervised Domain Adaptation in the Dissimilarity Space for Person Re-identification
Djebril Mekhazni, Amran Bhuiyan, George Ekladious, Eric Granger
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 between-class dissimilarity MMD loss to supervised loss (ablation) | D-MMD · Duke to Market | rank-1 accuracy: 36.1 -> 51.8 | Improved |
| add within-class dissimilarity MMD loss to supervised loss (ablation) | D-MMD · Duke to Market | rank-1 accuracy: 36.1 -> 45.8 | Improved |
| add within-class loss to between-class loss (ablation) | D-MMD · Duke to Market | rank-1 accuracy: 51.8 -> 66.6 | Improved |
| add feature-based MMD loss to both dissimilarity losses (ablation) | D-MMD · Duke to Market | rank-1 accuracy: 66.6 -> 70.6 | Improved |
| add feature-based MMD loss to both dissimilarity losses (ablation) | D-MMD · Duke to Market | mAP: 45.4 -> 48.8 | Improved |
| add between-class loss to within-class loss (ablation) | D-MMD · Market to Duke | rank-1 accuracy: 30.3 -> 60.5 | Improved |
| add feature-based MMD loss to both dissimilarity losses (ablation) | D-MMD · Market to Duke | rank-1 accuracy: 60.5 -> 63.5 | Improved |
| add feature-based MMD loss to both dissimilarity losses (ablation) | D-MMD · Market to Duke | mAP: 42.9 -> 46.0 | Improved |
| use between-class instead of within-class term alone (ablation) | D-MMD · Market to Duke | rank-1 accuracy: 30.3 -> 45.6 | Improved |
| use D-MMD instead of ECN | D-MMD · Duke to Market | rank-1 accuracy: 75.6 -> 70.6 | Hurt |
| use D-MMD instead of ECN | D-MMD · Duke to Market | mAP: 43.0 -> 48.8 | Improved |
| use D-MMD instead of BUC | D-MMD · Market to Duke | rank-1 accuracy: 47.4 -> 63.5 | Improved |
Closest papers by meaning
- Unsupervised Domain Adaptation with Background Shift Mitigating for Person Re-Identification (IJCV 2021 · similarity 0.95)
- Exert Diversity and Mitigate Bias: Domain Generalizable Person Re-identification with a Comprehensive Benchmark (IJCV 2024 · similarity 0.91)
- A Novel Unsupervised Camera-Aware Domain Adaptation Framework for Person Re-Identification (ICCV 2019 · similarity 0.90)
- Unsupervised Multi-Source Domain Adaptation for Person Re-Identification (CVPR 2021 · similarity 0.90)
- Joint Disentangling and Adaptation for Cross-Domain Person Re-Identification (ECCV 2020 · similarity 0.89)
- Towards Discriminative Representation Learning for Unsupervised Person Re-Identification (ICCV 2021 · 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.
