Research map / 3D vision / 3D generation
SceneTeller: Language-to-3D Scene Generation
Basak Melis Ocal, Maxim Tatarchenko, Sezer Karaoglu, Theo Gevers
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 |
|---|---|---|---|
| select in-context exemplars by retrieval instead of randomly (ablation) | SceneTeller layout generation (R+GPT) · 3D-FRONT layouts | mIoU (%): 29.90 -> 44.20 | Improved |
| select in-context exemplars by retrieval instead of randomly (ablation) | SceneTeller layout generation (R+GPT) · 3D-FRONT layouts | OOB rate (%): 76.11 -> 54.65 | Improved |
| add GPT paraphrasing to rule-based prompts with retrieval exemplars (ablation) | SceneTeller layout generation · 3D-FRONT layouts | OOB rate (%): 64.87 -> 54.65 | Improved |
| add GPT paraphrasing to rule-based prompts with retrieval exemplars (ablation) | SceneTeller layout generation · 3D-FRONT layouts | mIoU (%): 45.50 -> 44.20 | No clear effect |
| add negative exemplars (pos+neg) in place of all retrieved exemplars (ablation) | SceneTeller layout generation · 3D-FRONT layouts | mIoU (%): 31.10 -> 41.90 vs random; 45.50 retrieval | Mixed |
| use GPT-4 in-context layout generation instead of human drawing | layout generation · 3D-FRONT layouts | mIoU (%): 28.70 -> 44.60 | Improved |
| use GPT-4 in-context layout generation instead of human drawing | layout generation · 3D-FRONT layouts | OOB rate (%): 12.09 -> 56.78 | Hurt |
| generate layout first then assemble and stylize with 3DGS versus baseline scene generators | SceneTeller · ScanNet FID | FID: 238.75 -> 167.27 | Improved |
Later papers that compared with this method
2 later accepted papers report SceneTeller as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Scenethesis: A Language and Vision Agentic Framework for 3D Scene Generation (ICLR 2026 · better on 4 of 4 shared results)
- Learning 3D Object Spatial Relationships from Pre-trained 2D Diffusion Models (ICCV 2025 · better on 4 of 4 shared results)
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.
- LucidDreamer — LucidDreamer: Towards High-Fidelity Text-to-3D Generation via Interval Score Matching (CVPR 2024) · better on 6 of 6 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
- InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior (ICLR 2024 · similarity 0.90)
- SceneWeaver: All-in-One 3D Scene Synthesis with an Extensible and Self-Reflective Agent (NeurIPS 2025 · similarity 0.88)
- Hierarchically-Structured Open-Vocabulary Indoor Scene Synthesis with Pre-trained Large Language Model (AAAI 2025 · similarity 0.88)
- Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning (NeurIPS 2025 · similarity 0.88)
- SceneCraft: Layout-Guided 3D Scene Generation (NeurIPS 2024 · similarity 0.88)
- ArtiScene: Language-Driven Artistic 3D Scene Generation Through Image Intermediary (CVPR 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.
