Research map / Image generation / GANs
House-GAN++: Generative Adversarial Layout Refinement Network towards Intelligent Computational Agent for Professional Architects
Nelson Nauata, Sepidehsadat Hosseini, Kai-Hung Chang, Hang Chu, Chin-Yi Cheng, Yasutaka Furukawa
Check 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 iterative layout refinement with component-wise ground-truth conditioning and 50% heuristic constraint passing instead of House-GAN | House layout generation · RPLAN 8 rooms | FID: 66.4±1.7 -> 32.9±4.9 | Improved |
| Use iterative layout refinement with component-wise ground-truth conditioning and 50% heuristic constraint passing instead of House-GAN | House layout generation · RPLAN 8 rooms | GED: 5.3±0.0 -> 3.9±0.5 | Improved |
| Use iterative layout refinement with component-wise ground-truth conditioning and 50% heuristic constraint passing instead of House-GAN | House layout generation · RPLAN 8 rooms | realism: 0.0 -> 0.2 | Improved |
| Always provide all previous-layout constraints rather than provide them at a 50% chance (ablation) | House layout generation · RPLAN 8 rooms | FID: 32.9±4.9 -> 31.2±0.6 | Improved |
| Always provide all previous-layout constraints rather than provide them at a 50% chance (ablation) | House layout generation · RPLAN 8 rooms | GED: 3.9±0.5 -> 3.7±0.1 | Improved |
| Optimize a static room/door conditioning schedule for diversity (ablation) | House layout generation · RPLAN 8 rooms | FID: 32.9±4.9 -> 27.3±1.1 | Improved |
| Optimize a static room/door conditioning schedule for diversity (ablation) | House layout generation · RPLAN 8 rooms | GED: 3.9±0.5 -> 3.2±0.1 | Improved |
| Optimize a dynamic conditioning schedule for diversity instead of a static schedule (ablation) | House layout generation · RPLAN 8 rooms | FID: 27.3±1.1 -> 24.7±0.8 | Improved |
| Optimize a dynamic conditioning schedule for diversity instead of a static schedule (ablation) | House layout generation · RPLAN 8 rooms | GED: 3.2±0.1 -> 3.7±0.0 | Hurt |
| Optimize dynamic conditioning for compatibility rather than diversity (ablation) | House layout generation · RPLAN 8 rooms | GED: 3.7±0.0 -> 2.6±0.0 | Improved |
| Optimize dynamic conditioning for compatibility rather than diversity (ablation) | House layout generation · RPLAN 8 rooms | FID: 24.7±0.8 -> 35.7±1.3 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- House-GAN — House-GAN: Relational Generative Adversarial Networks for Graph-constrained House Layout Generation (ECCV 2020) · better on 10 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- House-GAN: Relational Generative Adversarial Networks for Graph-constrained House Layout Generation (ECCV 2020 · similarity 0.93)
- Building-GAN: Graph-Conditioned Architectural Volumetric Design Generation (ICCV 2021 · similarity 0.89)
- Graph Transformer GANs With Graph Masked Modeling for Architectural Layout Generation (TPAMI 2024 · similarity 0.88)
- MaskPLAN: Masked Generative Layout Planning from Partial Input (CVPR 2024 · similarity 0.86)
- GSDiff: Synthesizing Vector Floorplans via Geometry-enhanced Structural Graph Generation (AAAI 2025 · similarity 0.86)
- Constrained Layout Generation with Factor Graphs (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.
