Research map / 3D vision / 3D generation
SPREAD: Spatial-Physical REasoning via geometry Aware Diffusion
Minzhang Li, Kuixiang Shao, Xuebing Li, Yuyang Jiao, Yinuo Bai, Hengan Zhou, Sixian Shen, Jiayuan Gu, Jingyi Yu
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 geometry-aware perceiver module (ablation) | SPREAD · ProcTHOR | Colmesh: 0.241 -> 0.225 | Improved |
| add geometry-aware perceiver module (ablation) | SPREAD · ProcTHOR | Stability: 0.934 -> 0.938 | Improved |
| add multi-guidance framework (collision, gravity, relation) (ablation) | SPREAD · ProcTHOR | Colmesh: 0.225 -> 0.121 | Improved |
| add multi-guidance framework (ablation) | SPREAD · ProcTHOR | GRecall: 0.965 -> 0.979 | Improved |
| add multi-guidance framework (ablation) | SPREAD · ProcTHOR | Isaac Sim stability: 0.938 -> 0.950 | Improved |
| add collision guidance (ablation) | SPREAD · ProcTHOR | interpenetration | Improved |
| add gravity guidance (ablation) | SPREAD · ProcTHOR | floating objects | Improved |
| SPREAD vs baselines on 3D-FRONT FID | SPREAD · 3D-FRONT | FID: 68.0 (ATISS) vs 64.7 | No clear effect |
| SPREAD vs InstructScene on 3D-FRONT FID | SPREAD · 3D-FRONT | FID: 61.3 -> 64.7 | Hurt |
| SPREAD vs InstructScene on mesh collisions | SPREAD · 3D-FRONT Bedroom | Colmesh: 0.285 -> 0.097 | Improved |
| SPREAD vs InstructScene on Isaac stability | SPREAD · ProcTHOR | Isaac Sim stability: 0.876 -> 0.950 | Improved |
| SPREAD vs baselines on inference time | SPREAD · Inference time per scene | seconds: 10.25 (DiffuScene) -> 14.72 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- InstructScene — InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior (ICLR 2024) · better on 8 of 9 shared results
- DiffuScene — DiffuScene: Denoising Diffusion Models for Generative Indoor Scene Synthesis (CVPR 2024) · better on 5 of 6 shared results
- ATISS — ATISS: Autoregressive Transformers for Indoor Scene Synthesis (NeurIPS 2021) · 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
- MetaSpatial: Reinforcing 3D Spatial Reasoning in VLMs for the Metaverse (ICLR 2026 · similarity 0.88)
- PhyScene: Physically Interactable 3D Scene Synthesis for Embodied AI (CVPR 2024 · similarity 0.87)
- MesaTask: Towards Task-Driven Tabletop Scene Generation via 3D Spatial Reasoning (NeurIPS 2025 · similarity 0.86)
- SpatialStack: Layered Geometry-Language Fusion for 3D VLM Spatial Reasoning (CVPR 2026 · similarity 0.86)
- SceneWeaver: All-in-One 3D Scene Synthesis with an Extensible and Self-Reflective Agent (NeurIPS 2025 · similarity 0.86)
- DiffusionSfM: Predicting Structure and Motion via Ray Origin and Endpoint Diffusion (CVPR 2025 · similarity 0.86)
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.
