Research map / Image generation / Autoregressive and latent generation
Controllable Image Synthesis via SegVAE
Yen-Chi Cheng, Hsin-Ying Lee, Min Sun, Ming-Hsuan Yang
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 LSTM to learned prior and posterior (ablation) | SegVAE · HumanParsing | FID: 50.8830 -> 39.6496 | Improved |
| add LSTM to learned prior and posterior (ablation) | SegVAE · CelebAMask-HQ | FID: 34.5073 -> 28.8221 | Improved |
| add learned shape prior instead of fixed Gaussian (ablation) | SegVAE · HumanParsing | FID: 44.6217 -> 39.6496 | Improved |
| add learned shape prior instead of fixed Gaussian (ablation) | SegVAE · CelebAMask-HQ | FID: 33.8060 -> 28.8221 | Improved |
| add learned shape prior instead of fixed Gaussian (ablation) | SegVAE · HumanParsing | Diversity: .1625 -> .2072 | Improved |
| add LSTM to learned prior and posterior (ablation) | SegVAE · HumanParsing | Compatibility error: .7534 -> .6174 | Improved |
| add learned shape prior instead of fixed Gaussian (ablation) | SegVAE · CelebAMask-HQ | Compatibility error: .0839 -> .0754 | Improved |
| use Body->Clothes->Accessories order rather than Accessories first (ablation) | SegVAE · HumanParsing | FID: 40.2909 -> 39.6496 | Improved |
| use Face->Face features->Accessories order rather than Accessories first (ablation) | SegVAE · CelebAMask-HQ | FID: 32.0325 -> 28.8221 | Improved |
| use semantic category order instead of random order (ablation) | SegVAE · HumanParsing | FID: 52.6728 -> 39.6496 | Improved |
| swap Body and Clothes order (ablation) | SegVAE · HumanParsing | FID: 39.6496 -> 39.9008 | No clear effect |
| increase label-set length from 10 to 14 (ablation) | SegVAE · HumanParsing | FID: 39.2692 -> 44.2884 | Hurt |
Closest papers by meaning
- Semantic Probability Distribution Modeling for Diverse Semantic Image Synthesis (TPAMI 2022 · similarity 0.87)
- EditAR: Unified Conditional Generation with Autoregressive Models (CVPR 2025 · similarity 0.86)
- Unlocking Pre-trained Image Backbones for Semantic Image Synthesis (CVPR 2024 · similarity 0.86)
- SSMG: Spatial-Semantic Map Guided Diffusion Model for Free-Form Layout-to-Image Generation (AAAI 2024 · similarity 0.85)
- Stochastic Conditional Diffusion Models for Robust Semantic Image Synthesis (ICML 2024 · similarity 0.85)
- Scene Graph Disentanglement and Composition for Generalizable Complex Image Generation (NeurIPS 2024 · 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.
