Research map / Label-efficient and robust learning / Domain adaptation and generalization
Generalizing Person Re-Identification by Camera-Aware Invariance Learning and Cross-Domain Mixup
Chuanchen Luo, Chunfeng Song, Zhaoxiang Zhang
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 |
|---|---|---|---|
| replace camera-agnostic neighborhood loss with intra-camera loss (ablation) | Ls + Lag · Duke→Market | rank-1 (%): 60.9 -> 70.6 | Improved |
| replace camera-agnostic neighborhood loss with intra-camera loss (ablation) | Ls + Lag · Market→Duke | rank-1 (%): 49.8 -> 70.0 | Improved |
| add inter-camera neighborhood loss Linter (ablation) | Ls + Lintra · Duke→Market | rank-1 (%): 70.6 -> 81.2 | Improved |
| add inter-camera neighborhood loss Linter (ablation) | Ls + Lintra · Market→Duke | rank-1 (%): 70.0 -> 76.2 | Improved |
| add cross-domain mixup loss Lm in place of source classification (ablation) | Ls + Lintra · Duke→Market | rank-1 (%): 70.6 -> 76.8 | Improved |
| add cross-domain mixup loss Lm in place of source classification (ablation) | Ls + Lintra · Market→Duke | rank-1 (%): 70.0 -> 73.7 | Improved |
| add cross-domain mixup loss Lm in place of source classification (ablation) | Ls + Lintra + Linter · Duke→Market | rank-1 (%): 81.2 -> 88.1 | Improved |
| add cross-domain mixup loss Lm in place of source classification (ablation) | Ls + Lintra + Linter · Market→Duke | rank-1 (%): 76.2 -> 79.5 | Improved |
| raise scaling factor s from 10 to 14 (ablation) | Ours · Duke→Market | rank-1 (%): 88.1 -> 74.4 | Hurt |
| lower scaling factor s from 10 to 6 (ablation) | Ours · Duke→Market | rank-1 (%): 88.1 -> 80.3 | Hurt |
| vary Beta parameter alpha of mixup (ablation) | Ours · Duke→Market | rank-1 (%): 86.7 -> 88.2 | No clear effect |
Closest papers by meaning
- Learning to Adapt Across Dual Discrepancy for Cross-Domain Person Re-Identification (TPAMI 2022 · similarity 0.94)
- Learning to Adapt Invariance in Memory for Person Re-identification (TPAMI 2020 · similarity 0.93)
- A Novel Unsupervised Camera-Aware Domain Adaptation Framework for Person Re-Identification (ICCV 2019 · similarity 0.92)
- Invariance Matters: Exemplar Memory for Domain Adaptive Person Re-Identification (CVPR 2019 · similarity 0.92)
- Joint Disentangling and Adaptation for Cross-Domain Person Re-Identification (ECCV 2020 · similarity 0.91)
- Generalizing A Person Retrieval Model Hetero- and Homogeneously (ECCV 2018 · similarity 0.91)
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.
