Research map / Label-efficient and robust learning / Domain adaptation and generalization
Domain Adaptation through Synthesis for Unsupervised Person Re-identification
Slawomir Bak, Peter Carr, Jean-Francois Lalonde
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 SyRI synthetic images to real training data (ablation) | joint re-id model · PRID2011 | rank-1: 7.0 -> 15.0 | Improved |
| add SyRI synthetic images to real training data (ablation) | joint re-id model · Market-1501 | rank-1: 44.7 -> 54.3 | Improved |
| add SyRI synthetic images to real training data (ablation) | joint re-id model · VIPeR | rank-1: 32.3 -> 36.4 | Improved |
| add SyRI synthetic images to real training data (ablation) | joint re-id model · iLIDS | rank-1: 51.0 -> 54.8 | Mixed |
| add plain CycleGAN translation without regularization (ablation) | R + S · iLIDS | rank-1: 54.8 -> 53.9 | Hurt |
| add identity mapping loss Lid to CycleGAN (ablation) | CycleGAN · PRID2011 | rank-1: 33.0 -> 40.0 | Improved |
| replace Lid with LRef (ablation) | CycleGAN + Lid · CUHK01 | rank-1: 54.0 -> 48.4 | Hurt |
| replace Lid with LRef (ablation) | CycleGAN + Lid · PRID2011 | rank-1: 40.0 -> 28.0 | Hurt |
| add foreground mask loss (ablation) | CycleGAN + LRef · PRID2011 | rank-1: 28.0 -> 43.0 | Improved |
| add foreground mask loss (ablation) | CycleGAN + LRef · Market-1501 | rank-1: 57.5 -> 65.7 | Improved |
| add illumination inference instead of random domain selection (ablation) | domain translation · PRID2011 | rank-1: 36.0 -> 43.0 | Improved |
| add illumination inference instead of random domain selection (ablation) | domain translation · VIPeR | rank-1: 38.9 -> 43.0 | Improved |
Closest papers by meaning
- A Novel Unsupervised Camera-Aware Domain Adaptation Framework for Person Re-Identification (ICCV 2019 · similarity 0.91)
- Camera-Driven Representation Learning for Unsupervised Domain Adaptive Person Re-identification (ICCV 2023 · similarity 0.89)
- Unsupervised Multi-Source Domain Adaptation for Person Re-Identification (CVPR 2021 · similarity 0.89)
- Alice Benchmarks: Connecting Real World Re-Identification with the Synthetic (ICLR 2024 · similarity 0.89)
- Image-Image Domain Adaptation With Preserved Self-Similarity and Domain-Dissimilarity for Person Re-Identification (CVPR 2018 · similarity 0.89)
- Diffusion-based Synthetic Data Generation for Visible-Infrared Person Re-Identification (AAAI 2025 · similarity 0.87)
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.
