Research map / 3D vision / 3D generation
Text-to-3D scene generation
Creates 3D scenes from text via latent diffusion, direct training and compositional factoring.
58 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2022–2026) · an approach in 3D generation, keyword cluster “3d generation · text · consistent”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
22 papers in 2023–24 and 35 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 4.9% to 4.0%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2022 · Zero-Shot Text-Guided Object Generation With Dream Fields (CVPR 2022 · 381 citations)
- 2023 · SceneScape: Text-Driven Consistent Scene Generation (NeurIPS 2023 · 24 citations)
- 2024 · One-2-3-45++: Fast Single Image to 3D Objects with Consistent Multi-View Generation and 3D Diffusion (CVPR 2024 · 141 citations)
- 2025 · Structured 3D Latents for Scalable and Versatile 3D Generation (CVPR 2025 · 247 citations)
- 2026 · Free-Form Scene Editor: Enabling Multi-Round Object Manipulation Like in a 3D Engine (AAAI 2026 · 1 citation)
Earliest papers here
- Zero-Shot Text-Guided Object Generation With Dream Fields (CVPR 2022 · 381 citations)
- SceneScape: Text-Driven Consistent Scene Generation (NeurIPS 2023 · 24 citations)
- Text-To-4D Dynamic Scene Generation (ICML 2023 · 23 citations)
What has been tried
35 papers here report 347 measured changes (a component added, removed or swapped, compared with a baseline): helped 74%, no clear effect 4%, hurt 19%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use SLAT resolution 64 with 8 channels rather than resolution 32 with 16 channels | 3D asset reconstruction · PSNR | 31.64 -> 32.74 | Helped | Structured 3D Latents for Scalable and Versatile 3D Generation (CVPR 2025 · 247 citations) |
| Replace consistent multi-view generation with Zero123 XL | GSO · CLIP similarity | 71.9 vs. 73.4 | Hurt | One-2-3-45++: Fast Single Image to 3D Objects with Consistent Multi-View Generation and 3D Diffusion (CVPR 2024 · 141 citations) |
| increase the interval length δT | text-to-3D generation · detail | Hurt | LucidDreamer: Towards High-Fidelity Text-to-3D Generation via Interval Score Matching (CVPR 2024 · 118 citations) | |
| add 3D-aware SDS and image guidance | Text-to-4D ablation prompts · CLIP Score | 29.3 -> 35.0 | Helped | 4D-fy: Text-to-4D Generation Using Hybrid Score Distillation Sampling (CVPR 2024 · 55 citations) |
| Add style consistency loss | Synthetic Object Dataset (15 scenes; scenes requiring style consistency) · CLIP_sim | 0.312* -> 0.319* | Helped | FocalDreamer: Text-Driven 3D Editing via Focal-Fusion Assembly (AAAI 2024 · 38 citations) |
| Use 24 rather than 12 decoder layers | GSO novel views · LPIPS | 0.1266 -> 0.1258 | Helped | Instant3D: Fast Text-to-3D with Sparse-view Generation and Large Reconstruction Model (ICLR 2024 · 32 citations) |
| add perspective-view guidance to panoramic-only guidance | 10 indoor test prompts · CLIP-AP | 59.7 -> 97.6 | Helped | Towards Text-guided 3D Scene Composition (CVPR 2024 · 25 citations) |
| Add structural guidance from the coarse 3D prior | A head of the Terracotta Army · Geometric fidelity | Helped | Sherpa3D: Boosting High-Fidelity Text-to-3D Generation via Coarse 3D Prior (CVPR 2024 · 22 citations) | |
| Increase the number of segmentation samples from 1 to 10 | Part segmentation, automatic · mAP50 | 45.2 -> 59.3 | Helped | PartGen: Part-level 3D Generation and Reconstruction with Multi-view Diffusion Models (CVPR 2025 · 18 citations) |
| Alternate local foreground editing with global full-image editing instead of using local editing only | Text-driven editing · Background preservation and foreground-background harmony | Helped | Customize your NeRF: Adaptive Source Driven 3D Scene Editing via Local-Global Iterative Training (CVPR 2024 · 15 citations) | |
| add SDS++ refinement | T3Bench Single-Object-with-Surroundings · BRISQUE | 37.1 -> 32.3 | Helped | Director3D: Real-world Camera Trajectory and 3D Scene Generation from Text (NeurIPS 2024 · 10 citations) |
| Add automatic alignment and cleaning | Rotated SRN Cars · FID | 46.77 -> 8.69 | Helped | DIRECT-3D: Learning Direct Text-to-3D Generation on Massive Noisy 3D Data (CVPR 2024 · 9 citations) |
What reviewers will expect
Datasets and baselines named in the 17 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- ABO · 3 of 17 papers
- Objaverse · 3 of 17 papers
- Toys4K · 2 of 17 papers
- Objaverse-XL · 2 of 17 papers
- Google Scanned Objects · 2 of 17 papers
- 3D-FRONT · 2 of 17 papers
- 3D-FUTURE · 2 of 17 papers
- PartObjaverse-Tiny · 2 of 17 papers
Baselines
- LucidDreamer · 4 of 17 papers
- GaussianDreamer · 3 of 17 papers
- Shap-E · 2 of 17 papers
- DreamFusion · 2 of 17 papers
- HiFA · 2 of 17 papers
- LGM · 2 of 17 papers
- HoloPart · 2 of 17 papers
- MVDream · 2 of 17 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 57 papers; from OpenAlex and the papers' affiliations).
- Peking University · 4 papers
- Zhejiang University · 4 papers
- Huawei · 3 papers
- Sun Yat-sen University · 3 papers
- The Chinese University of Hong Kong · 3 papers
- Tsinghua University · 3 papers
- Meta · 3 papers
- MBZUAI · 2 papers
- Peng Cheng Laboratory · 2 papers
- Xi’an Jiaotong University · 2 papers
Use it
Download BibTeX of 25 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Diffusion-based novel view synthesis (156 papers)
- Score distillation for text-to-3D (106 papers)
- Feed-forward 3D asset generation and editing (116 papers)
Elsewhere in the topic
- Learning 3D reconstruction from images (196 papers)
- Semantic scene completion (153 papers)
- Generative indoor scene synthesis (131 papers)
- Hand-object interaction generation (116 papers)
- Open-vocabulary 3D scene understanding (95 papers)
- 2D-to-3D pretrained representations (94 papers)
- Monocular 3D object reconstruction (94 papers)
- Scene-aware human motion synthesis (82 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).