Search papers, labs, and topics across Lattice.
NeoWorld-Pro introduces a novel approach to converting monocular images into interactive 3D environments by reformulating scene reconstruction as procedural programming. This framework utilizes the capabilities of multi-layered language models (MLLMs) to generate executable programs from single RGB images, which define object geometry and physical properties. The integration of a physics-in-the-loop mechanism allows for iterative refinement of these programs, resulting in superior performance in tasks like stable stacking and manipulation compared to existing methods.
Transforming a single RGB image into a fully interactive 3D scene could redefine how we create simulation assets for Embodied AI.
The advancement of Embodied AI necessitates high-quality simulation assets that faithfully mirror the real world. However, transforming raw visual observations into simulation-ready scenes remains challenging due to the lack of physical grounding and scene-level interactivity in current image-to-URDF methods. We propose NeoWorld-Pro, a framework that reformulates monocular scene reconstruction as procedural programming for interactive 3D environments. Leveraging the zero-shot reasoning and code synthesis capabilities of MLLMs, NeoWorld-Pro converts a single RGB image into executable programs specifying object geometry, articulation, and physical properties. A physics-in-the-loop mechanism then iteratively refines the generated programs by validating their execution in a physics engine, enforcing physically plausible articulations, valid object compositions and interactions, and accurate spatial relationships. Experiments show that NeoWorld-Pro outperforms open-loop and prior monocular reconstruction methods, while enabling complex downstream tasks such as stable stacking and fine-grained manipulation.
