Research map / Label-efficient and robust learning / Recognition and long-tail
Recover and Identify: A Generative Dual Model for Cross-Resolution Person Re-Identification
Yu-Jhe Li, Yun-Chun Chen, Yen-Yu Lin, Xiaofei Du, Yu-Chiang Frank 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 image-level adversarial loss (ablation) | CRGAN · MLR-CUHK03 (ablation) | Rank-1 (%): 82.1 -> 79.8 without it | Improved |
| add image-level adversarial loss (ablation) | CRGAN · MLR-CUHK03 (ablation) | LPIPS: 0.07 -> 0.17 without it | Improved |
| add feature-level adversarial loss (ablation) | CRGAN · MLR-CUHK03 (ablation) | Rank-1 (%): 82.1 -> 67.6 without it | Improved |
| add HR reconstruction loss (ablation) | CRGAN · MLR-CUHK03 (ablation) | Rank-1 (%): 82.1 -> 66.7 without it | Improved |
| add classification loss (ablation) | CRGAN · MLR-CUHK03 (ablation) | Rank-1 (%): 82.1 -> 1.7 without it | Improved |
| add classification loss (ablation) | CRGAN · MLR-CUHK03 (ablation) | PSNR: 20.2 -> 21.4 without it | Hurt |
| use joint representation [f,g] instead of f only (ablation) | CRGAN · MLR-CUHK03 | Rank-1 (%): 77.6 -> 82.1 | Improved |
| use joint representation [f,g] instead of g only (ablation) | CRGAN · MLR-CUHK03 | Rank-1 (%): 79.7 -> 82.1 | Improved |
| use joint representation [f,g] instead of g only (ablation) | CRGAN · CAVIAR | Rank-1 (%): 38.9 -> 42.8 | Improved |
| use f only rather than g only (ablation) | CRGAN · CAVIAR | Rank-1 (%): 38.9 -> 41.5 | Improved |
| test on unseen down-sampling rate r=8 instead of seen rates | CRGAN · MLR-CUHK03 | Rank-1 (%): 82.1 -> 78.6 | Hurt |
| use CRGAN instead of CSR-GAN on unseen resolution | CRGAN vs CSR-GAN · MLR-CUHK03 (unseen r=8) | Rank-1 (%): 62.1 -> 78.6 | Improved |
Closest papers by meaning
- Learning Resolution-Invariant Deep Representations for Person Re-Identification (AAAI 2019 · similarity 0.95)
- Deep Low-Resolution Person Re-Identification (AAAI 2018 · similarity 0.92)
- Joint Discriminative and Generative Learning for Person Re-Identification (CVPR 2019 · similarity 0.91)
- Pose-Normalized Image Generation for Person Re-identification (ECCV 2018 · similarity 0.90)
- Instance-Guided Context Rendering for Cross-Domain Person Re-Identification (ICCV 2019 · similarity 0.90)
- Person Re-Identification by Camera Correlation Aware Feature Augmentation (TPAMI 2017 · similarity 0.90)
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.
