Research map / Image generation / Autoregressive and latent generation
LayoutDiffusion: Controllable Diffusion Model for Layout-to-Image 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 LayoutDiffusion rather than the strongest listed GAN baseline | Layout-to-image generation · COCO-stuff 256×256 | FID: 29.10 -> 15.61 | Helped |
| Use LayoutDiffusion rather than the strongest listed GAN baseline | Layout-to-image generation · COCO-stuff 256×256 | CAS: 40.00 -> 47.74 | Helped |
| Use LayoutDiffusion rather than the strongest listed GAN baseline | Layout-to-image generation · VG 256×256 | FID: 28.06 -> 15.63 | Helped |
| Use LayoutDiffusion rather than the strongest listed GAN baseline | Layout-to-image generation · VG 256×256 | CAS: 34.48 -> 48.90 | Helped |
| Use LayoutDiffusion rather than LDM-8 (100 steps) | COCO-stuff 256×256 generation · COCO-stuff 256×256 | FID: 42.06 -> 31.68 | Helped |
| Use the small model configuration instead of LDM-8 | COCO-stuff 256×256 generation · COCO-stuff 256×256 | images / s: 0.457 -> 0.608 | Helped |
| Add LFM and OaCA (ablation) | LayoutDiffusion · COCO-stuff 128×128 | FID: 29.94 -> 16.57 | Helped |
| Add LFM and OaCA (ablation) | LayoutDiffusion · COCO-stuff 128×128 | CAS: 3.83 -> 43.60 | Helped |
| Add LFM and OaCA (ablation) | LayoutDiffusion · COCO-stuff 128×128 | DS: 0.70±0.08 -> 0.47±0.09 | Hurt |
| Add LFM (ablation) | LayoutDiffusion · COCO-stuff 128×128 | CAS: 30.86 -> 43.60 | Helped |
| Add OaCA (ablation) | LayoutDiffusion · COCO-stuff 128×128 | CAS: 40.67 -> 43.60 | Helped |
| Replace OaCA with ordinary cross attention (ablation) | LayoutDiffusion with LFM · COCO-stuff 128×128 | FID: 16.46 -> 16.57 | Hurt |
Closest papers by meaning
- HiCo: Hierarchical Controllable Diffusion Model for Layout-to-image Generation (NeurIPS 2024 · similarity 0.91)
- Build-A-Scene: Interactive 3D Layout Control for Diffusion-Based Image Generation (ICLR 2025 · similarity 0.90)
- Move Anything with Layered Scene Diffusion (CVPR 2024 · similarity 0.88)
- Visual Layout Composer: Image-Vector Dual Diffusion Model for Design Layout Generation (CVPR 2024 · similarity 0.88)
- InstanceAssemble: Layout-Aware Image Generation via Instance Assembling Attention (NeurIPS 2025 · similarity 0.88)
- Control and Realism: Best of Both Worlds in Layout-to-Image without Training (ICML 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.