Research map / Graph learning / Knowledge and temporal graphs
End-to-End Graph-Constrained Vectorized Floorplan Generation with Panoptic Refinement
Jiachen Liu, Yuan Xue, Jose Duarte, Krishnendra Shekhawat, Zihan Zhou, Xiaolei Huang
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 |
|---|---|---|---|
| add hybrid sorting (ablation) | Base GPT-1 autoregressive model · RPLAN 6 rooms | Compatibility: 4.0/0.1 -> 3.6/0.1 | Improved |
| add hybrid sorting (ablation) | Base GPT-1 autoregressive model · RPLAN 6 rooms | Diversity: 27.8/0.5 -> 29.8/0.6 | Hurt |
| add GCN after hybrid sorting (ablation) | Hybrid-sorted model · RPLAN 6 rooms | Diversity: 29.8/0.6 -> 14.5/0.2 | Improved |
| add GCN after hybrid sorting (ablation) | Hybrid-sorted model · RPLAN 6 rooms | Compatibility: 3.6/0.1 -> 3.7/0.0 | No clear effect |
| add panoptic refinement after hybrid sorting (ablation) | Hybrid-sorted model · RPLAN 6 rooms | Diversity: 29.8/0.6 -> 14.0/0.3 | Improved |
| add panoptic refinement on top of GCN (ablation) | Hybrid sorting plus GCN · RPLAN 6 rooms | Compatibility: 3.7/0.0 -> 2.1/0.1 | Improved |
| add panoptic refinement on top of GCN (ablation) | Hybrid sorting plus GCN · RPLAN 6 rooms | Diversity: 14.5/0.2 -> 18.3/0.2 | Hurt |
| add geometric loss (ablation) | Full model without GL · RPLAN 6 rooms | Compatibility: 2.1/0.1 -> 1.9/0.0 | Improved |
| add geometric loss (ablation) | Full model without GL · RPLAN 6 rooms | Diversity: 18.3/0.2 -> 15.8/0.5 | Improved |
| use 5 refinement iterations instead of 1 (ablation) | PRN · RPLAN 6 rooms | Compatibility: 2.7/0.0 -> 1.9/0.0 | Improved |
| use 7 refinement iterations instead of 5 (ablation) | PRN · RPLAN 6 rooms | Compatibility: 1.9/0.0 -> 2.1/0.1 | No clear effect |
| our method vs HouseGAN++ at 8 rooms | Ours · RPLAN separate training (8 rooms) | Compatibility: 3.8/0.0 -> 4.6/0.0 | Hurt |
Closest papers by meaning
- FloorNet: A Unified Framework for Floorplan Reconstruction from 3D Scans (ECCV 2018 · similarity 0.90)
- Raster-To-Vector: Revisiting Floorplan Transformation (ICCV 2017 · similarity 0.90)
- VectorFloorSeg: Two-Stream Graph Attention Network for Vectorized Roughcast Floorplan Segmentation (CVPR 2023 · similarity 0.89)
- GSDiff: Synthesizing Vector Floorplans via Geometry-enhanced Structural Graph Generation (AAAI 2025 · similarity 0.89)
- HouseDiffusion: Vector Floorplan Generation via a Diffusion Model With Discrete and Continuous Denoising (CVPR 2023 · similarity 0.89)
- CAGE: Continuity-Aware edGE Network Unlocks Robust Floorplan Reconstruction (NeurIPS 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.
