Research map / Label-efficient and robust learning / Recognition and long-tail
Unsupervised Learning of Multi-Level Descriptors for Person Re-Identification
Yang Yang, Longyin Wen, Siwei Lyu, Stan Li
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 |
|---|---|---|---|
| use WLC coding instead of LLC | pixel-level descriptor with KISSME · VIPeR | Rank-1 matching rate (%): 12.5 -> 22.9 | Improved |
| use WLC coding instead of SC | patch-level descriptor · VIPeR | Rank-1 matching rate (%): 16.6 -> 22.3 | Improved |
| use WLC instead of LLC/SC for image-level descriptors | image-level descriptor · VIPeR | Rank-1 matching rate (%): 0.4 -> 6.6 | Improved |
| increase similarity constraint weight lambda1 from 0 to 0.001d^2 (ablation) | WLC · VIPeR | Rank-1 matching rate | Improved |
| partition image into horizontal stripes for image-level feature (ablation) | image-level descriptor · VIPeR | Rank-1 matching rate | Improved |
| use WLC instead of SC for speed | WLC · pixel-level extraction time | seconds: 0.12 -> 0.03 | Improved |
| increase number of bases for patch-level WLC from 70 to 100 (ablation) | WLC patch-level · VIPeR | Rank-1 matching rate (%): 22.3 -> 19.0 | Hurt |
| compare against GOG at Rank 5 | WLC multi-level fusion · VIPeR | Rank-5 matching rate: 79.7% vs 76.4% | Hurt |
Closest papers by meaning
- Unsupervised Person Re-Identification via Softened Similarity Learning (CVPR 2020 · similarity 0.87)
- Multi-Task Learning with Low Rank Attribute Embedding for Multi-Camera Person Re-Identification (TPAMI 2017 · similarity 0.86)
- Camera-Aware Proxies for Unsupervised Person Re-Identification (AAAI 2021 · similarity 0.85)
- Unsupervised Person Re-Identification by Soft Multilabel Learning (CVPR 2019 · similarity 0.85)
- Unsupervised Pre-Training for Person Re-Identification (CVPR 2021 · similarity 0.85)
- Unsupervised Person Re-Identification via Multi-Label Classification (CVPR 2020 · similarity 0.85)
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.
