The Intelligent Space Generation Team is dedicated to enhancing AI's capabilities in three-dimensional understanding and hallucination, utilizing AI to create virtual spaces that rival the real physical world. The laboratory's research encompasses multi-view generation, native 3D generation, scene layout generation, animation generation and so on, with applications spanning industrial manufacturing, AR/VR and embodied artificial intelligence.
The laboratory boasts profound technical expertise in 3D-related domains. Addressing the challenges of difficult digital asset acquisition and high production costs, the team has developed the Craftsman 3D Asset Generation Model, significantly improving the quality and efficiency of 3D model generation. This model is one of the pioneering open-source generative models to extend diffusion models to 3D data on an international scale.
Currently, the team is focused on enhancing the model's ability to generate complex geometries and textures, while advancing the development of interactive 3D space generation. This enables the creation of 3D objects and scene assets characterized by intricate geometric details, high-definition textures, logically arranged spatial layouts, and scientifically accurate motion sequences.
