Research map / Image generation / Autoregressive and latent generation
Layout generation
Generate graphic and design layouts with transformers and diffusion models.
39 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2019–2026) · an approach in Autoregressive and latent generation, keyword cluster “motion · layout · object”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
18 papers in 2023–24 and 12 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 3.8% to 1.3%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2019 · Image Generation From Layout (CVPR 2019 · 221 citations)
- 2020 · BachGAN: High-Resolution Image Synthesis From Salient Object Layout (CVPR 2020 · 40 citations)
- 2021 · Variational Transformer Networks for Layout Generation (CVPR 2021 · 124 citations)
- 2022 · Coarse-to-Fine Generative Modeling for Graphic Layouts (AAAI 2022 · 41 citations)
- 2023 · LayoutDiffusion: Controllable Diffusion Model for Layout-to-Image Generation (CVPR 2023 · 172 citations)
- 2024 · Retrieval-Augmented Layout Transformer for Content-Aware Layout Generation (CVPR 2024 · 21 citations)
- 2025 · Graphic Design with Large Multimodal Model (AAAI 2025 · 10 citations)
- 2026 · OcclusionFormer: Arranging Z-Order for Layout-Grounded Image Generation (ICML 2026 · 0 citations)
Earliest papers here
- Image Generation From Layout (CVPR 2019 · 221 citations)
- LayoutGAN: Generating Graphic Layouts with Wireframe Discriminators (ICLR 2019 · 31 citations)
- BachGAN: High-Resolution Image Synthesis From Salient Object Layout (CVPR 2020 · 40 citations)
What has been tried
31 papers here report 347 measured changes (a component added, removed or swapped, compared with a baseline): helped 73%, no clear effect 3%, hurt 22%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add retrieved-layout feature fL and input-image feature fI alongside fC | PKU test · FID | 4.67 -> 3.45 | Helped | Retrieval-Augmented Layout Transformer for Content-Aware Layout Generation (CVPR 2024 · 21 citations) |
| Use the saliency map mask instead of a full mask for the attention constraint | Web-design test set · FID | 34.49 -> 31.52 | Helped | Desigen: A Pipeline for Controllable Design Template Generation (CVPR 2024 · 14 citations) |
| Remove conditioning on previously predicted attributes | RPLAN test, boundary-only input · fidimg | 4.182 -> 23.103 | Hurt | MaskPLAN: Masked Generative Layout Planning from Partial Input (CVPR 2024 · 14 citations) |
| Add the image diffusion domain to vector-only VLC | Crello · composed image FID | 5.74 -> 5.83 | Hurt | Visual Layout Composer: Image-Vector Dual Diffusion Model for Design Layout Generation (CVPR 2024 · 12 citations) |
| Use Qwen1.5-7B instead of Qwen1.5-0.5B | Crello · S_GI | 6.35 -> 6.90 | Helped | Graphic Design with Large Multimodal Model (AAAI 2025 · 10 citations) |
| Add UGC to the HiCo-base configuration with GBB and mask FuseNet | HiCo-7K ablation · FID | 15.26 -> 16.85 | Hurt | HiCo: Hierarchical Controllable Diffusion Model for Layout-to-image Generation (NeurIPS 2024 · 5 citations) |
| use layered design composition with rendered previous layers as context | Crello-v4 design composition · LLaVA-OV design and layout | 7.84 -> 8.08 | Helped | From Elements to Design: A Layered Approach for Automatic Graphic Design Composition (CVPR 2025 · 3 citations) |
| Use pixel-space wireframe input instead of object-space sequence input | PubLaynet · F1 | 29.52 -> 59.82 | Helped | Spot the Error: Non-autoregressive Graphic Layout Generation with Wireframe Locator (AAAI 2024 · 3 citations) |
| Add amodal mask supervision to CreatiLayout | OverLayBench-Simple · O-mIoU (%) | 32.52 -> 37.69; +15.90% | Helped | OverLayBench: A Benchmark for Layout-to-Image Generation with Dense Overlaps (NeurIPS 2025 · 1 citation) |
| Add aesthetic constraints to LACE | PubLayNet unconditional · Alignment | 0.238 -> 0.141 | Helped | Towards Aligned Layout Generation via Diffusion Model with Aesthetic Constraints (ICLR 2024 · 1 citation) |
| Train the adversary jointly instead of keeping it fixed | Zelda game levels · % Unique | 0.22 -> 0.995 | Helped | GenCO: Generating Diverse Designs with Combinatorial Constraints (ICML 2024 · 0 citations) |
| add Plücker manifold parameterization | PartNet Mobility, unconditional generation · MMD | 0.0270 -> 0.0264 | Helped | MIDGArD: Modular Interpretable Diffusion over Graphs for Articulated Designs (NeurIPS 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 6 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- COCO · 2 of 6 papers
- LayoutSAM · 2 of 6 papers
Baselines
- FlexDM · 2 of 6 papers
- GLIGEN · 2 of 6 papers
- InstanceDiffusion · 2 of 6 papers
- MIGC · 2 of 6 papers
- HiCo · 2 of 6 papers
- 3DIS · 2 of 6 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 30 papers; from OpenAlex and the papers' affiliations).
- Microsoft Research Asia · 3 papers
- Microsoft · 3 papers
- South China University of Technology · 2 papers
- ByteDance · 2 papers
- CyberAgent · 2 papers
- Zhejiang University · 2 papers
- Xi’an Jiaotong University · 2 papers
- Adobe · 2 papers
- University of California, San Diego · 2 papers
Use it
Download BibTeX of 36 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Compositional scene generation (105 papers)
- Diffusion planning and policies (58 papers)
Elsewhere in the topic
- Classical generative image synthesis (147 papers)
- High-resolution diffusion synthesis (140 papers)
- Retrieval and matching with generative features (140 papers)
- Generative latent space tooling (129 papers)
- Generalizable fake image detection (119 papers)
- Latent diffusion with learned tokenizers (114 papers)
- Synthetic training data from generators (111 papers)
- Next-scale and token autoregressive images (90 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).