Research map / 3D vision / 3D generation
Roam2Room: A Unified Floorplan-to-Furnished Framework for Controllable Indoor Scene Generation
Wenbo Li, Zipeng Qin, Xiaoliang Ju, Rongyao Fang, Hongsheng LI
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 |
|---|---|---|---|
| represent floorplans as a K-D tree instead of the canonical representation (ablation) | LLM floorplan generator · Floorplan generation (500 samples) | floorplans with disconnected rooms: 16.2% -> 12.3% | Improved |
| use K-D tree representation (canonical gives overlaps and holes) (ablation) | LLM floorplan generator · Floorplan generation (500 samples) | overlapping room polygons: 9.2% overlap and 2.0% holes under canonical | Improved |
| fine-tune Qwen3-4B on 300K real floorplans vs ProcTHOR-10K generation | floorplan generation · Floorplan generation (500 samples) | graph validity: 72.6% -> 87.7% | Improved |
| fine-tune on real floorplans vs ProcTHOR-10K | floorplan generation · Floorplan generation (500 samples) | typed graph diversity: 6.2 -> 22.4 | Improved |
| add top-down view roaming (ablation) | layout generation pipeline · Room layout generation (ablation) | volume density: 0.22 -> 0.35 | Improved |
| add ego-centric view roaming (ablation) | layout generation pipeline · Room layout generation (ablation) | volume density: 0.20 -> 0.35 | Improved |
| add VLM-based recursive refiner (ablation) | layout generation pipeline · Room layout generation (ablation) | collision ratio: 0.20 -> 0.05 | Improved |
| add VLM-based recursive refiner (ablation) | layout generation pipeline · Room layout generation (ablation) | footprint object density: 3.02 -> 4.16 FOD; CR 0.20 -> 0.05 | Mixed |
| add manipulable object placement (ablation) | layout generation pipeline · Room layout generation (ablation) | footprint object density: 1.82 -> 4.16 | Improved |
| add manipulable object placement (ablation) | layout generation pipeline · Room layout generation (ablation) | collision ratio: 0.05 -> 0.05 | No clear effect |
| use full pipeline instead of LayoutVLM | layout generation · Room layout generation (4 room types, 25 each) | collision ratio: 0.20 -> 0.05 | Improved |
| use full pipeline instead of LayoutGPT | layout generation · Room layout generation (4 room types, 25 each) | collision ratio: 0.05 -> 0.05 | No clear effect |
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.
- Holodeck — Holodeck: Language Guided Generation of 3D Embodied AI Environments (CVPR 2024) · better on 3 of 3 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
- OptiScene: LLM-driven Indoor Scene Layout Generation via Scaled Human-aligned Data Synthesis and Multi-Stage Preference Optimization (NeurIPS 2025 · similarity 0.91)
- Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning (NeurIPS 2025 · similarity 0.89)
- DeBaRA: Denoising-Based 3D Room Arrangement Generation (NeurIPS 2024 · similarity 0.89)
- InternScenes: A Large-scale Simulatable Indoor Scene Dataset with Realistic Layouts (NeurIPS 2025 · similarity 0.88)
- SceneCraft: Layout-Guided 3D Scene Generation (NeurIPS 2024 · similarity 0.88)
- CASAGPT: Cuboid Arrangement and Scene Assembly for Interior Design (CVPR 2025 · similarity 0.88)
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.
