Research map / Label-efficient and robust learning / Recognition and long-tail
Deep Learning for Person Re-Identification: A Survey and Outlook
Mang Ye, Jianbing Shen, Gaojie Lin, Tao Xiang, Ling Shao, Steven C. H. Hoi
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 deep learning with viewpoint information (VAL) instead of human-level matching | image-based Re-ID · Market-1501 | Rank-1: 93.5 -> 96.2 | Improved |
| move from RNN video model to spatial-temporal aggregation (GLTR) | video Re-ID · PRID-2011 | Rank-1: 70 -> 95.5 | Improved |
| move from RNN to ADFD | video Re-ID · iLIDS-VID | Rank-1: 58 -> 86.3 | Improved |
| move from IDEX to STC spatial-temporal aggregation | video Re-ID · MARS | Rank-1/mAP: 68.3%/49.3% -> 88.5%/82.3% | Improved |
| train on larger, harder dataset (MSMT17 vs Market-1501) with same method class | image-based Re-ID · MSMT17 vs Market-1501 | Rank-1: 96.2 -> 82.3 | Hurt |
| evaluate GLTR on large LS-VID instead of smaller video datasets | GLTR · LS-VID | Rank-1/mAP: 63.1%/44.43% | Hurt |
| add part-level features to global features | image-based Re-ID · Re-ID benchmarks | accuracy | Improved |
| add attention schemes | image-based Re-ID · Re-ID benchmarks | accuracy | Improved |
| combine identity loss with triplet loss (multi-loss) | image-based Re-ID · Re-ID benchmarks | accuracy | Improved |
Closest papers by meaning
- A Systematic Evaluation and Benchmark for Person Re-Identification: Features, Metrics, and Datasets (TPAMI 2018 · similarity 0.94)
- Multi-Scale Deep Learning Architectures for Person Re-Identification (ICCV 2017 · similarity 0.93)
- Resource Aware Person Re-Identification Across Multiple Resolutions (CVPR 2018 · similarity 0.92)
- Intra-Camera Supervised Person Re-Identification (IJCV 2021 · similarity 0.92)
- Unsupervised Pre-Training for Person Re-Identification (CVPR 2021 · similarity 0.91)
- Person Re-Identification by Camera Correlation Aware Feature Augmentation (TPAMI 2017 · similarity 0.91)
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.
