Research map / Label-efficient and robust learning / Recognition and long-tail
Point to Set Similarity Based Deep Feature Learning for Person Re-Identification
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the P2S triplet term to the P2P model (ablation) | part-based deep CNN · 3DPeS | Top1 matching rate: 61.97 -> 71.16 | Helped |
| add the P2S triplet term to the P2P model (ablation) | part-based deep CNN · CUHK01 | Top1 matching rate: 68.91 -> 77.34 | Helped |
| add the P2S triplet term to the P2P model (ablation) | part-based deep CNN · PRID2011 | Top1 matching rate: 62.24 -> 70.71 | Helped |
| add the P2S triplet term to the P2P model (ablation) | part-based deep CNN · Market1501 single-query | Top1 matching rate: 61.31 -> 70.72 | Helped |
| add the P2S triplet term to the P2P model (ablation) | part-based deep CNN · Market1501 multi-query | Top1 matching rate: 76.09 -> 85.78 | Helped |
| use adaptive symmetric-triplet weights µ=0.6, ν=0.4 instead of conventional weights µ=1.0, ν=0.0 (ablation) | P2S model · 3DPeS | Top1 matching rate: 64.38 -> 71.16 | Helped |
| use adaptive symmetric-triplet weights µ=0.6, ν=0.4 instead of conventional weights µ=1.0, ν=0.0 (ablation) | P2S model · CUHK01 | Top1 matching rate: 71.27 -> 77.34 | Helped |
| use adaptive symmetric-triplet weights µ=0.6, ν=0.4 instead of conventional weights µ=1.0, ν=0.0 (ablation) | P2S model · PRID2011 | Top1 matching rate: 65.32 -> 70.71 | Helped |
| use adaptive symmetric-triplet weights µ=0.6, ν=0.4 instead of conventional weights µ=1.0, ν=0.0 (ablation) | P2S model · Market1501 | Top1 matching rate: 63.82 -> 70.72 | Helped |
| set α=0.05 instead of α=0.1 (ablation) | P2S model · CUHK01 | Top1 matching rate: 77.3% -> 68.9% | Hurt |
| set α=0.15 instead of α=0.1 (ablation) | P2S model · CUHK01 | Top1 matching rate: 77.3% -> 72.5% | Hurt |
| set Cp=0.25 and Mp=0.25 instead of Cp=0.2 and Mp=0.3 (ablation) | P2S model · CUHK01 | Top1 matching rate: 77.3% -> 66.9% | Hurt |
Closest papers by meaning
- Learning a Discriminative Null Space for Person Re-Identification (CVPR 2016 · similarity 0.90)
- Semi-Supervised Bayesian Attribute Learning for Person Re-Identification (AAAI 2018 · similarity 0.89)
- Group Consistent Similarity Learning via Deep CRF for Person Re-Identification (CVPR 2018 · similarity 0.89)
- Eliminating Background-Bias for Robust Person Re-Identification (CVPR 2018 · similarity 0.88)
- Efficient and Deep Person Re-Identification Using Multi-Level Similarity (CVPR 2018 · similarity 0.88)
- Resource Aware Person Re-Identification Across Multiple Resolutions (CVPR 2018 · similarity 0.88)
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.