Research map / Label-efficient and robust learning / Recognition and long-tail
Person Re-identification in Identity Regression Space
Hanxiao Wang, Xiatian Zhu, Shaogang Gong, Tao Xiang
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 OneHot identity coding instead of random coding (ablation) | IRS (LOMO) · Market-1501 (single-query) | Rank-1 (%): 47.4 -> 57.7 | Improved |
| use OneHot identity coding instead of random coding (ablation) | IRS (LOMO) · CUHK03 (manual) | Rank-1 (%): 51.7 -> 61.6 | Improved |
| use OneHot coding instead of FDA coding (ablation) | IRS (LOMO) · Market-1501 (single-query) | Rank-1 (%): 55.6 -> 57.7 | Improved |
| use OneHot coding instead of FDA coding (ablation) | IRS (LOMO) · VIPeR and CUHK01 | Rank-1 (%): 45.1 -> 45.1 | No clear effect |
| fuse WHOS, LOMO and CNN features by score-level fusion instead of single LOMO (ablation) | IRS · VIPeR | Rank-1 (%): 45.1 -> 54.6 | Improved |
| fuse WHOS, LOMO and CNN features instead of CNN alone (ablation) | IRS · CUHK01 (486/485 split) | Rank-1 (%): 68.6 -> 80.8 | Improved |
| use CUHK03-trained CNN feature without fine-tuning instead of LOMO (ablation) | IRS · VIPeR | Rank-1 (%): 45.1 -> 33.1 | Hurt |
| use CNN feature instead of LOMO (ablation) | IRS · CUHK03 (detected) | Rank-1 (%): 53.4 -> 80.3 | Improved |
| use CNN feature instead of LOMO (ablation) | IRS · Market-1501 (single-query) | Rank-1 (%): 57.7 -> 72.7 | Improved |
| add IRS regression with 2 FC layers on top of DGD CNN features (ablation) | DGD · CUHK03 (manual) | Rank-1 (%): 73.7 -> 76.6 | Improved |
| use 3 FC layers instead of 2 in deep learning regression (ablation) | IRS (DLR) · CUHK03 (manual) | Rank-1 (%): 76.6 -> 74.2 | Hurt |
| use ridge regression instead of deep learning regression (ablation) | IRS · CUHK03 (manual) | Rank-1 (%): 76.6 -> 81.5 | Improved |
Closest papers by meaning
- Joint Disentangling and Adaptation for Cross-Domain Person Re-Identification (ECCV 2020 · similarity 0.87)
- Implicit Sample Extension for Unsupervised Person Re-Identification (CVPR 2022 · similarity 0.87)
- Learning a Discriminative Null Space for Person Re-Identification (CVPR 2016 · similarity 0.86)
- Unsupervised Pre-Training for Person Re-Identification (CVPR 2021 · similarity 0.86)
- Self-Training With Progressive Augmentation for Unsupervised Cross-Domain Person Re-Identification (ICCV 2019 · similarity 0.86)
- Self-Similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-Identification (ICCV 2019 · 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.
