RoomPlanner: Explicit Layout Planner for Easier LLM-Driven 3D Room Generation

📄 Paper | 🧠 Code

RoomPlanner is the first fully automatic 3D room generation framework for creating realistic indoor scenes from only short text descriptions. Without any manual layout design or panoramic image guidance, it generates explicit layout criteria for rational spatial placement through hierarchical language-driven agent planning.

🖼️ Visual Results

Example 1
A diverse set of 3D scenes generated by RoomPlanner, demonstrating its ability to create both common and uncommon room types.
Example 2
Compared to previous methods, ie.,visual-guided method Pano2Room (Pu, Zhao, and Lian 2024) and rule-based method DreamScene (Li et al. 2024a), our approach effectively generates indoor scenes characterized by (a) more realism, (b) smoother mesh structures, and (c) support for a diverse array of editions, including operations such as rotation, translation, importing/deleting 3D assets, and style variations.
Example 3
Text-to-layout mapping with spatial reasoning visualization.
Example 4
Object placement refinement guided by hierarchical planner.
Example 5
Scene generated with minimal text prompt ("modern living room").
Example 6
Demonstration of spatial feasibility and realism in layout.

🎥 Demo Videos