Research map / 3D vision / 3D generation
MajutsuCity: Language-driven Aesthetic-adaptive City Generation with Controllable 3D Assets and Layouts
Zilong Huang, Jun He, Xiaobin Huang, Ziyi Xiong, Yang Luo, Junyan Ye, Weijia Li, Yiping Chen, Ting Han
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 |
|---|---|---|---|
| replace CLIP text encoder with LongCLIP for layout generation (ablation) | Layout diffusion (SD v2.1) · City layout generation | FID: 28.0 -> 22.7 | Improved |
| add fine-grained spatial text descriptions instead of short prompts (ablation) | Layout diffusion (SD v2.1) · City layout generation | FID: 35.7 -> 28.0 | Improved |
| add LongCLIP and spatial text together (ablation) | Layout diffusion (SD v2.1) · City layout generation | KID: 0.025 -> 0.013 | Improved |
| add LongCLIP to spatial-text model (ablation) | Layout diffusion (SD v2.1) · City layout generation | IS: 3.07 -> 3.14 | No clear effect |
| text-to-layout diffusion with spatial text vs CityCraft | MajutsuCity · City layout generation | FID: 28.4 -> 22.7 | Improved |
| explicit mesh assets generated per instance vs neural representations (CityDreamer) | MajutsuCity · City scene generation (AQS, GPT) | Structural and View Consistency: 4.20 -> 8.56 | Improved |
| generate stylized assets instead of retrieving from a fixed library (CityCraft) | MajutsuCity · City scene generation (AQS, GPT) | Scene Richness and Complexity: 6.11 -> 8.33 | Improved |
| GaussianCity (3DGS) vs mesh-based methods on SVC under AQS versus RDR | Baselines · City scene generation | SVC: AQS GPT 6.73 vs 6.17, 6.00; RDR GPT 23.56 vs 25.31, 24.44 | Mixed |
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.
- CityDreamer — CityDreamer: Compositional Generative Model of Unbounded 3D Cities (CVPR 2024) · better on 2 of 2 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
- Yo'City: Personalized and Boundless 3D Realistic City Scene Generation via Self-Critic Expansion (CVPR 2026 · similarity 0.89)
- CityDreamer: Compositional Generative Model of Unbounded 3D Cities (CVPR 2024 · similarity 0.88)
- ArtiScene: Language-Driven Artistic 3D Scene Generation Through Image Intermediary (CVPR 2025 · similarity 0.86)
- SceneCraft: Layout-Guided 3D Scene Generation (NeurIPS 2024 · similarity 0.86)
- Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning (NeurIPS 2025 · similarity 0.85)
- InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior (ICLR 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.
