Research map / Graph learning / Graph self-supervised learning
Generative Layout Modeling Using Constraint Graphs
Wamiq Para, Paul Guerrero, Tom Kelly, Leonidas J. Guibas, Peter Wonka
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 |
|---|---|---|---|
| use free generation (no boundary input) instead of boundary-conditioned generation | constraint-graph layout model · RPLAN floor plans | FID: 27.27 -> 21.47 | Improved |
| condition on ground-truth exterior boundary | constraint-graph layout model · RPLAN floor plans | s_avg: 1.00 -> 1.03 | No clear effect |
| condition on ground-truth exterior boundary | constraint-graph layout model · LIFULL floor plans | s_avg: 1.00 -> 2.30 | Hurt |
| condition on ground-truth exterior boundary | constraint-graph layout model · LIFULL floor plans | FID: 26.15 -> 31.94 | Hurt |
| use proposed three-step constraint graph model instead of StyleGAN image generation | floor plan generation · RPLAN floor plans | s_avg: 19.67 -> 1.00 | Improved |
| use proposed model instead of StyleGAN | floor plan generation · LIFULL floor plans | s_avg: 16.27 -> 1.00 | Improved |
| use proposed free generation instead of RPLAN baseline (which gets ground-truth boundary) | floor plan generation · RPLAN floor plans | FID: 21.29 -> 21.47 | No clear effect |
| use proposed free generation instead of RPLAN baseline | floor plan generation · LIFULL floor plans | s_avg: 15.55 -> 1.00 | Improved |
| use proposed model instead of Graph2Plan (which gets ground-truth layout graph) | floor plan generation · RPLAN floor plans | s_avg: 8.46 -> 1.00 | Improved |
| use proposed model instead of StyleGAN for furniture layouts | furniture layout generation · Furniture layouts | s_avg: 9.94 -> 1.00 | Improved |
Closest papers by meaning
- LayoutFormer++: Conditional Graphic Layout Generation via Constraint Serialization and Decoding Space Restriction (CVPR 2023 · similarity 0.91)
- LayoutDM: Discrete Diffusion Model for Controllable Layout Generation (CVPR 2023 · similarity 0.90)
- Towards Aligned Layout Generation via Diffusion Model with Aesthetic Constraints (ICLR 2024 · similarity 0.89)
- MaskPLAN: Masked Generative Layout Planning from Partial Input (CVPR 2024 · similarity 0.87)
- Constrained Layout Generation with Factor Graphs (CVPR 2024 · similarity 0.86)
- Retrieval-Augmented Layout Transformer for Content-Aware Layout Generation (CVPR 2024 · similarity 0.85)
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.
