Research map / Label-efficient and robust learning / Recognition and long-tail
Occlude Them All: Occlusion-Aware Attention Network for Occluded Person Re-ID
Peixian Chen, Wenfeng Liu, Pingyang Dai, Jianzhuang Liu, Qixiang Ye, Mingliang Xu, Qi’an Chen, Rongrong Ji
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 occlusion augmentation (ablation) | OAMN baseline (Lide+Ltri) · Occluded-Duke | Rank-1: 36.9 -> 51.2 | Improved |
| add grouped triplet loss Ltri-g (ablation) | OAMN with OA · Occluded-Duke | Rank-1: 51.2 -> 55.5 | Improved |
| add attention-guided mask loss Lmask (ablation) | OAMN with OA + Ltri-g · Occluded-Duke | Rank-1: 55.5 -> 57.7 | Improved |
| add occlusion unification at test time (ablation) | OAMN with OA + Ltri-g · Occluded-Duke | Rank-1: 55.5 -> 60.3 | Improved |
| add both Lmask and OU (ablation) | OAMN with OA + Ltri-g · Occluded-Duke | Rank-1: 55.5 -> 62.6 | Improved |
| use grouped triplet loss instead of vanilla triplet loss (ablation) | OAMN · Occluded-Duke | Rank-1: 55.2 -> 62.6 | Improved |
| use grouped triplet loss instead of vanilla triplet loss (ablation) | OAMN · Partial-REID | Rank-1: 76.7 -> 86.0 | Improved |
| augment occlusion at all four locations instead of top/bottom only (ablation) | OAMN · Occluded-Duke | Rank-1: 56.5 -> 62.6 | Improved |
| augment both 1/2 and 1/4 occlusion areas instead of 1/4 only (ablation) | OAMN · Occluded-Duke | Rank-1: 52.8 -> 62.6 | Improved |
| attach mask module at layer-3 instead of layer-1 (ablation) | OAMN · Occluded-Duke | Rank-1: 57.5 -> 62.6 | Improved |
| attach mask module at layer-3 instead of layer-4 (ablation) | OAMN · Partial-REID | Rank-1: 76.3 -> 86.0 | Improved |
| apply OAMN on top of its baseline for holistic ReID | OAMN · Market-1501 | Rank-1: 85.7 -> 93.2 | Improved |
Closest papers by meaning
- Feature Erasing and Diffusion Network for Occluded Person Re-Identification (CVPR 2022 · similarity 0.93)
- Learning To Know Where To See: A Visibility-Aware Approach for Occluded Person Re-Identification (ICCV 2021 · similarity 0.92)
- Attention Disturbance and Dual-Path Constraint Network for Occluded Person Re-identification (AAAI 2024 · similarity 0.92)
- Occluded Person Re-Identification With Single-Scale Global Representations (ICCV 2021 · similarity 0.92)
- Dynamic Feature Pruning and Consolidation for Occluded Person Re-identification (AAAI 2024 · similarity 0.90)
- Occluded Person Re-identification via Saliency-Guided Patch Transfer (AAAI 2024 · similarity 0.87)
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.
