Research map / Label-efficient and robust learning / Recognition and long-tail
SVDNet for Pedestrian Retrieval
Yifan Sun, Liang Zheng, Weijian Deng, Shengjin Wang
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 |
|---|---|---|---|
| add SVDNet restraint-relaxation training | CaffeNet FC6 · Market-1501 | rank-1 (%): 55.3 -> 80.5 | Improved |
| add SVDNet restraint-relaxation training | CaffeNet FC6 · Market-1501 | mAP (%): 30.4 -> 55.9 | Improved |
| add SVDNet restraint-relaxation training | ResNet-50 Pool5 · Market-1501 | rank-1 (%): 73.8 -> 82.3 | Improved |
| add SVDNet restraint-relaxation training | CaffeNet FC6 · CUHK03 | rank-1 (%): 38.6 -> 68.5 | Improved |
| add SVDNet restraint-relaxation training | CaffeNet FC6 · DukeMTMC-reID | rank-1 (%): 46.9 -> 67.6 | Improved |
| add SVDNet restraint-relaxation training | ResNet-50 Pool5 · CUHK03 | rank-1 (%): 66.2 -> 81.8 | Improved |
| add SVDNet restraint-relaxation training | ResNet-50 Pool5 · DukeMTMC-reID | rank-1 (%): 65.5 -> 76.7 | Improved |
| replace W with US (SVD) at Step 1, no further training | CaffeNet FC7 Step-0 model · Market-1501 | rank-1 (%): 63.6 -> 63.6 | No clear effect |
| replace W with U instead of US at Step 1 (ablation) | CaffeNet FC7 Step-0 model · Market-1501 | rank-1 (%): 63.6 -> 61.7 | Hurt |
| replace W with QD instead of US at Step 1 (ablation) | CaffeNet FC7 Step-0 model · Market-1501 | rank-1 (%): 63.6 -> 61.6 | Hurt |
| use US (keep singular values) instead of U in RRI (ablation) | SVDNet CaffeNet FC6 · Market-1501 | rank-1 (%): 76.2 -> 80.5 | Improved |
| use US instead of UV^T in RRI (ablation) | SVDNet CaffeNet FC6 · Market-1501 | rank-1 (%): 57.4 -> 80.5 | Improved |
Closest papers by meaning
- Person Search via A Mask-guided Two-stream CNN Model (ECCV 2018 · similarity 0.87)
- Combined Depth Space Based Architecture Search for Person Re-Identification (CVPR 2021 · similarity 0.87)
- Joint Detection and Identification Feature Learning for Person Search (CVPR 2017 · similarity 0.87)
- Backbone Cannot Be Trained at Once: Rolling Back to Pre-Trained Network for Person Re-Identification (AAAI 2019 · similarity 0.86)
- Auto-ReID: Searching for a Part-Aware ConvNet for Person Re-Identification (ICCV 2019 · similarity 0.86)
- A Multi-Task Deep Network for Person Re-Identification (AAAI 2017 · 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.
