Research map / Label-efficient and robust learning / Recognition and long-tail
Population-Based Evolutionary Gaming for Unsupervised Person Re-identification
Yunpeng Zhai, Peixi Peng, Mengxi Jia, Shiyong Li, Wei‐Qiang Chen, Xuesong Gao, Yonghong Tian
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 |
|---|---|---|---|
| train eight models jointly with shared pseudo-labels instead of one model (ablation) | single model baseline · Market-1501 | mAP: 69.6 -> 76.4 | Improved |
| add population mutual learning (ablation) | Multi-model · Market-1501 | mAP: 76.4 -> 78.5 | Improved |
| add population mutual learning (ablation) | Multi-model · DukeMTMC-reID | mAP: 63.4 -> 66.1 | Improved |
| add selection of 3 of 8 networks (ablation) | Multi-model + PML · Market-1501 | mAP: 78.5 -> 79.2 | Improved |
| add reproduction and mutation (ablation) | Multi-model + PML + Sel. · Market-1501 | mAP: 79.2 -> 84.3 | Improved |
| add reproduction and mutation (ablation) | Multi-model + PML + Sel. · DukeMTMC-reID | mAP: 66.9 -> 71.9 | Improved |
| add multiple architectures on top of single-architecture PEG (ablation) | Single architecture + PEG · Market-1501 | mAP: 80.1 -> 84.3 | Improved |
| add temporal average model to mutual learning (ablation) | PML · Market-1501 (3 networks) | mAP: 73.3 -> 79.2 | Improved |
| add mutual identity loss (ablation) | PML · Market-1501 (3 networks) | mAP: 77.2 -> 79.2 | Improved |
| add mutual triplet loss (ablation) | PML · Market-1501 (3 networks) | mAP: 77.0 -> 79.2 | Improved |
| use all models for pseudo-labels with ClusterContrast without selection (ablation) | ClusterContrast · Market-1501 | mAP: 83.3 -> 79.4 | Hurt |
| add selection before multi-model CCL training (ablation) | CCL-Multi · Market-1501 | mAP: 79.4 -> 84.3 | Improved |
Closest papers by meaning
- Self-Reinforcing Prototype Evolution with Dual-Knowledge Cooperation for Semi-Supervised Lifelong Person Re-Identification (ICCV 2025 · similarity 0.85)
- Deep Reinforcement Active Learning for Human-in-the-Loop Person Re-Identification (ICCV 2019 · similarity 0.85)
- ABD-Net: Attentive but Diverse Person Re-Identification (ICCV 2019 · similarity 0.85)
- Self-Similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-Identification (ICCV 2019 · similarity 0.85)
- Lifelong Person Re-Identification via Adaptive Knowledge Accumulation (CVPR 2021 · similarity 0.84)
- Mixture of Submodules for Domain Adaptive Person Search (CVPR 2025 · similarity 0.83)
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.
