Research map / Label-efficient and robust learning / Domain adaptation and generalization
Joint Visual and Temporal Consistency for Unsupervised Domain Adaptive Person Re-Identification
Jianing Li, Shiliang 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 |
|---|---|---|---|
| add SAC (self-adaptive one-hot classification) to direct transfer (ablation) | direct transfer model · DukeMTMC -> Market1501 | rank-1 accuracy: 42.1 -> 64.5 | Improved |
| add SAC to direct transfer (ablation) | direct transfer model · Market1501 -> DukeMTMC | rank-1 accuracy: 31.8 -> 59.4 | Improved |
| replace visual-only clustering with MTC (visual plus temporal consistency) (ablation) | memory-bank baseline · DukeMTMC -> Market1501 | mAP: 46.6 -> 56.4 | Improved |
| replace visual-only clustering with MTC (ablation) | memory-bank baseline · Market1501 -> DukeMTMC | mAP: 43.6 -> 51.1 | Improved |
| combine SAC and MTC (JVTC) instead of the baseline (ablation) | memory-bank baseline · DukeMTMC -> Market1501 | mAP: 46.6 -> 61.1 | Improved |
| add SAC to MTC (ablation) | MTC · DukeMTMC -> Market1501 | mAP: 56.4 -> 61.1 | Improved |
| add SAC to MTC (ablation) | MTC · Market1501 -> DukeMTMC | rank-1 accuracy: 71.3 -> 75.0 | Improved |
| use joint visual-temporal similarity at test time (JVTC+) (ablation) | JVTC · DukeMTMC -> Market1501 | mAP: 61.1 -> 67.2 | Improved |
| use joint similarity at test time (JVTC+) (ablation) | JVTC · Market1501 -> DukeMTMC | mAP: 56.2 -> 66.5 | Improved |
| use joint similarity at test time (JVTC+) (ablation) | JVTC · MSMT17 (source DukeMTMC) | rank-1 accuracy: 45.4 -> 52.9 | Improved |
| use SAC alone instead of the visual-only baseline | memory-bank baseline · DukeMTMC -> Market1501 | rank-1 accuracy: 77.4 -> 64.5 | Hurt |
| add positive loss weights w1 and w2 for target losses (versus 0) (ablation) | JVTC · Market1501 and DukeMTMC-reID | ReID performance | Improved |
Closest papers by meaning
- Camera-Driven Representation Learning for Unsupervised Domain Adaptive Person Re-identification (ICCV 2023 · similarity 0.92)
- Reliability Exploration with Self-Ensemble Learning for Domain Adaptive Person Re-identification (AAAI 2022 · similarity 0.92)
- Unsupervised Person Re-Identification via Multi-Label Classification (CVPR 2020 · similarity 0.91)
- Domain Adaptive Attention Learning for Unsupervised Person Re-Identification (AAAI 2020 · similarity 0.91)
- A Novel Unsupervised Camera-Aware Domain Adaptation Framework for Person Re-Identification (ICCV 2019 · similarity 0.91)
- Delving into Probabilistic Uncertainty for Unsupervised Domain Adaptive Person Re-identification (AAAI 2022 · similarity 0.90)
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.
