Research map / Image generation / Autoregressive and latent generation
A Variational U-Net for Conditional Appearance and Shape Generation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the variational U-Net instead of pix2pix | Conditional image reconstruction · Market-1501 | SSIM: 0.166 -> 0.353 | Helped |
| Use the variational U-Net instead of pix2pix | Conditional image reconstruction · DeepFashion | SSIM: 0.646 -> 0.786 | Helped |
| Use the variational U-Net instead of pix2pix | Conditional image reconstruction · Market-1501 | IS: 2.289 -> 3.214 | Helped |
| Use the variational U-Net instead of pix2pix | Conditional image reconstruction · DeepFashion | IS: 2.640 -> 3.087 | Helped |
| Add appearance representation z (ablation) | Variational U-Net · Image generation ablation | IS: 2.211 -> 3.087 | Helped |
| Add appearance representation z (ablation) | Variational U-Net · Image transfer ablation | IS: 2.211 -> 3.504 | Helped |
| Add KL regularization (ablation) | Variational U-Net · Reconstruction ablation | IS: 3.168 -> 3.087 | Hurt |
| Add KL regularization (ablation) | Variational U-Net · Transfer ablation | IS: 3.594 -> 3.504 | Hurt |
| Add KL regularization (ablation) | Appearance transfer in the variational U-Net · Appearance transfer | Shape-invariance of appearance representation | Helped |
Closest papers by meaning
- Visual Object Networks: Image Generation with Disentangled 3D Representations (NeurIPS 2018 · similarity 0.87)
- Michelangelo: Conditional 3D Shape Generation based on Shape-Image-Text Aligned Latent Representation (NeurIPS 2023 · similarity 0.87)
- Disentangled Pose and Appearance Guidance for Multi-Pose Generation (CVPR 2025 · similarity 0.86)
- Generating Furry Cars: Disentangling Object Shape and Appearance across Multiple Domains (ICLR 2021 · similarity 0.86)
- SCULPT: Shape-Conditioned Unpaired Learning of Pose-dependent Clothed and Textured Human Meshes (CVPR 2024 · similarity 0.84)
- A Simple Approach to Unifying Diffusion-based Conditional Generation (ICLR 2025 · similarity 0.84)
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.