PhiP-G: Physics-Guided Text-to-3D Compositional Scene Generation
This addresses the challenge of generating physically realistic compositional 3D scenes for applications in gaming, simulation, and virtual reality, representing a novel integration of methods rather than a paradigm shift.
The paper tackles the problem of generating physically plausible 3D scenes from text descriptions by proposing PhiP-G, a framework that integrates scene graph generation, targeted asset creation, and physics-guided layout planning. The result is state-of-the-art performance in CLIP scores, competitive quality on T³Bench, and a 24x efficiency improvement.
Text-to-3D asset generation has achieved significant optimization under the supervision of 2D diffusion priors. However, when dealing with compositional scenes, existing methods encounter several challenges: 1). failure to ensure that composite scene layouts comply with physical laws; 2). difficulty in accurately capturing the assets and relationships described in complex scene descriptions; 3). limited autonomous asset generation capabilities among layout approaches leveraging large language models (LLMs). To avoid these compromises, we propose a novel framework for compositional scene generation, PhiP-G, which seamlessly integrates generation techniques with layout guidance based on a world model. Leveraging LLM-based agents, PhiP-G analyzes the complex scene description to generate a scene graph, and integrating a multimodal 2D generation agent and a 3D Gaussian generation method for targeted assets creation. For the stage of layout, PhiP-G employs a physical pool with adhesion capabilities and a visual supervision agent, forming a world model for layout prediction and planning. Extensive experiments demonstrate that PhiP-G significantly enhances the generation quality and physical rationality of the compositional scenes. Notably, PhiP-G attains state-of-the-art (SOTA) performance in CLIP scores, achieves parity with the leading methods in generation quality as measured by the T$^3$Bench, and improves efficiency by 24x.