Research map / Label-efficient and robust learning / Recognition and long-tail
Foreground-Aware Pyramid Reconstruction for Alignment-Free Occluded Person Re-Identification
Lingxiao He, Yinggang Wang, Wu Liu, He Zhao, Zhenan Sun, Jiashi Feng
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 foreground-aware pyramid reconstruction matching | Baseline ReID model · Occluded REID | Rank-1: 42.12 -> 78.30 | Improved |
| add foreground-aware pyramid reconstruction matching | Baseline ReID model · Partial REID | Rank-1: 53.33 -> 81.00 | Improved |
| add foreground-aware pyramid reconstruction matching | Baseline ReID model · Partial iLIDS | Rank-1: 52.94 -> 68.08 | Improved |
| add foreground-aware pyramid reconstruction matching | Baseline ReID model · Market1501 | Rank-1: 94.06 -> 95.42 | Improved |
| add foreground-aware pyramid reconstruction matching | Baseline ReID model · CUHK03 | Rank-1: 73.57 -> 76.08 | Improved |
| add foreground-aware pyramid reconstruction matching | Baseline ReID model · DukeMTMC | Rank-1: 87.30 -> 88.64 | Improved |
| use FPR instead of DSR reconstruction matching | DSR · Occluded REID | Rank-1: 72.80 -> 78.30 | Improved |
| use FPR instead of DSR reconstruction matching | DSR · Partial REID | Rank-1: 73.67 -> 81.00 | Improved |
| use FPR instead of DSR reconstruction matching | DSR · Market1501 | mAP: 85.78 -> 86.58 | Improved |
| use FPR instead of DSR reconstruction matching | DSR · DukeMTMC | mAP: 77.07 -> 78.42 | Improved |
| set foreground loss weight alpha to 0.02 (swept 0.01 to 0.04) (ablation) | FPR · Occluded datasets (Rank-1 and mAP) | Rank-1 / mAP | Improved |
| set foreground label threshold tau near 0.35 (swept 0 to 1) (ablation) | FPR · Occluded datasets (Rank-1 and mAP) | Rank-1 / mAP | Improved |
Closest papers by meaning
- Occluded Person Re-Identification With Single-Scale Global Representations (ICCV 2021 · similarity 0.92)
- Learning To Know Where To See: A Visibility-Aware Approach for Occluded Person Re-Identification (ICCV 2021 · similarity 0.92)
- Adversarially Occluded Samples for Person Re-Identification (CVPR 2018 · similarity 0.91)
- Dynamic Feature Pruning and Consolidation for Occluded Person Re-identification (AAAI 2024 · similarity 0.91)
- From Poses to Identity: Training-Free Person Re-Identification via Feature Centralization (CVPR 2025 · similarity 0.88)
- Occluded Person Re-identification via Saliency-Guided Patch Transfer (AAAI 2024 · similarity 0.86)
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.
