Research map / Image generation / Autoregressive and latent generation
Diverse Semantic Image Synthesis via Probability Distribution Modeling
Check 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 INADE instead of GroupDNet | semantic image synthesis · Cityscapes | FID: 41.12 -> 38.04 | Helped |
| Use INADE instead of SPADE | semantic image synthesis · Cityscapes | mIoU: 61.38 -> 61.02 | Hurt |
| Use INADE instead of SEAN | semantic image synthesis · ADE20K | FID: 28.11 -> 29.60 | Hurt |
| Use INADE instead of BicycleGAN | semantic image synthesis · ADE20K | LPIPS: 0.443 -> 0.400 | Hurt |
| Use INADE instead of SEAN | semantic image synthesis · CelebAMask-HQ | FID: 24.30 -> 22.55 | Helped |
| Use INADE instead of GroupDNet | semantic image synthesis · CelebAMask-HQ | mIoU: 76.13 -> 74.08 | Hurt |
| Use INADE instead of SEAN | semantic image synthesis · DeepFashion | FID: 7.37 -> 9.96 | Hurt |
| Use INADE instead of GroupDNet | multimodal synthesis · DeepFashion2 | mISD: 0.0500 -> 0.0580 | Helped |
| Add prior noise remapping during training (ablation) | INADE · DeepFashion | FID: 12.09 -> 9.97 | Helped |
| Add prior noise remapping during training (ablation) | INADE · DeepFashion | LPIPS: 0.184 -> 0.225 | Helped |
| Use unified rather than independent noise sampling across layers (ablation) | INADE · DeepFashion | FID: 248.33 -> 9.97 | Helped |
| Use unified rather than independent noise sampling across layers (ablation) | INADE · DeepFashion | LPIPS: 0.624 -> 0.225 | Hurt |
Closest papers by meaning
- PLACE: Adaptive Layout-Semantic Fusion for Semantic Image Synthesis (CVPR 2024 · similarity 0.90)
- Stochastic Conditional Diffusion Models for Robust Semantic Image Synthesis (ICML 2024 · similarity 0.89)
- Unlocking Pre-trained Image Backbones for Semantic Image Synthesis (CVPR 2024 · similarity 0.87)
- Noise Diffusion for Enhancing Semantic Faithfulness in Text-to-Image Synthesis (CVPR 2025 · similarity 0.87)
- SynthRGB-T: Language-Vision Guided Image Translation for Diversity Synthesis (CVPR 2026 · similarity 0.86)
- SSMG: Spatial-Semantic Map Guided Diffusion Model for Free-Form Layout-to-Image Generation (AAAI 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.