SceneFoundry: Generating Interactive Infinite 3D Worlds

TL;DR

SceneFoundry is a novel framework that automatically generates large, interactive 3D environments with articulated furniture and diverse layouts, facilitating robotic training and embodied AI research.

Contribution

It introduces a language-guided diffusion approach with differentiable guidance for creating functionally realistic and navigable 3D worlds from natural language prompts.

Findings

Generates structurally valid and semantically coherent environments

Enables scalable robotic training in diverse scene types

Maintains physical usability with articulated objects and navigable spaces

Abstract

The ability to automatically generate large-scale, interactive, and physically realistic 3D environments is crucial for advancing robotic learning and embodied intelligence. However, existing generative approaches often fail to capture the functional complexity of real-world interiors, particularly those containing articulated objects with movable parts essential for manipulation and navigation. This paper presents SceneFoundry, a language-guided diffusion framework that generates apartment-scale 3D worlds with functionally articulated furniture and semantically diverse layouts for robotic training. From natural language prompts, an LLM module controls floor layout generation, while diffusion-based posterior sampling efficiently populates the scene with articulated assets from large-scale 3D repositories. To ensure physical usability, SceneFoundry employs differentiable guidance functions to regulate object quantity, prevent articulation collisions, and maintain sufficient walkable space for robotic navigation. Extensive experiments demonstrate that our framework generates structurally valid, semantically coherent, and functionally interactive environments across diverse scene types and conditions, enabling scalable embodied AI research. project page: https://anc891203.github.io/SceneFoundry-Demo/