Research map / Image generation / Editing and style transfer
Unpaired Person Image Generation With Semantic Parsing Transformation
Sijie Song, Wei Zhang, Jiaying Liu, Zongming Guo, Tao Mei
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 semantic parsing transformation (TS-Pred vs Baseline) (ablation) | unpaired person generation · DeepFashion | IS: 3.140 -> 3.201 | Improved |
| add semantic parsing transformation (TS-Pred vs Baseline) (ablation) | unpaired person generation · Market-1501 | IS: 2.776 -> 3.462 | Improved |
| add semantic parsing transformation (TS-Pred vs Baseline) (ablation) | unpaired person generation · DeepFashion | SSIM: 0.698 -> 0.724 | Improved |
| add semantic parsing transformation (TS-Pred vs Baseline) (ablation) | unpaired person generation · Market-1501 | SSIM: 0.157 -> 0.180 | Improved |
| use end-to-end training instead of two-stage with predicted maps (ablation) | semantic+appearance networks · DeepFashion | IS: 3.201 -> 3.441 | Improved |
| use end-to-end training instead of two-stage with predicted maps (ablation) | semantic+appearance networks · Market-1501 | IS: 3.462 -> 3.499 | Improved |
| use end-to-end training instead of two-stage with predicted maps (ablation) | semantic+appearance networks · Market-1501 | mask-IS: 3.546 -> 3.680 | Improved |
| use end-to-end training instead of two-stage with predicted maps (ablation) | semantic+appearance networks · DeepFashion | SSIM: 0.724 -> 0.736 | Improved |
| E2E training versus ground-truth semantic maps (TS-GT) (ablation) | semantic+appearance networks · DeepFashion | IS: 3.350 -> 3.441 | No clear effect |
| E2E training versus ground-truth semantic maps (TS-GT) (ablation) | semantic+appearance networks · Market-1501 | IS: 3.472 -> 3.499 | Improved |
| E2E training versus ground-truth semantic maps (TS-GT) (ablation) | semantic+appearance networks · Market-1501 | SSIM: 0.200 -> 0.203 | Improved |
| use semantic-aware style loss instead of mask-style or patch-style loss (ablation) | TS-GT · DeepFashion | visual quality | Improved |
Closest papers by meaning
- Unsupervised Person Image Generation With Semantic Parsing Transformation (CVPR 2019 · similarity 0.99)
- Learning Semantic Person Image Generation by Region-Adaptive Normalization (CVPR 2021 · similarity 0.89)
- Structure-aware Person Image Generation with Pose Decomposition and Semantic Correlation (AAAI 2021 · similarity 0.89)
- Disentangled Pose and Appearance Guidance for Multi-Pose Generation (CVPR 2025 · similarity 0.87)
- SCULPT: Shape-Conditioned Unpaired Learning of Pose-dependent Clothed and Textured Human Meshes (CVPR 2024 · similarity 0.86)
- Coarse-to-Fine Latent Diffusion for Pose-Guided Person Image Synthesis (CVPR 2024 · 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.
