Autodesk has recently introduced Project Bernini, an exciting research effort focused on developing generative AI for the Design and Make industries.
This innovative project offers the first experimental generative AI model capable of rapidly generating superior 3D shapes from various inputs, including single 2D images, multiple images showing different views of an object, point clouds, voxels, and text.
The initial Bernini model is specifically geared toward professional geometric workflows, allowing it to generate multiple functional variations of a 3D shape based on a given input.
By emphasizing accuracy and productivity in 3D modeling, Autodesk aims to enhance the creative workflow for architects, engineers, manufacturers, and filmmakers.
While the technology is still in its experimental phase, Autodesk has not yet disclosed specific integration plans with its existing products
New multi-modal AI model Generates 3D models from text, images, point clouds or voxels
Project Bernini, a remarkable AI technology for 3D model generation, stands out due to its versatility in accepting various input types.
Unlike conventional methods that generate 3D meshes solely from text prompts or infer them from images or video, Bernini operates seamlessly with text, images, voxels, and point clouds.
Developed by Autodesk Research’s AI Lab, this model builds upon the foundation of Make-A-Shape and was trained on an extensive dataset of “10 million publicly available shapes,” encompassing both CAD models and organic forms.
Is Project Bernini optimized for creating simple hard surface object?
In the video demonstration, Project Bernini showcases its ability to generate a variety of hard-surface objects, including a chair, spatula, and water pitcher. Autodesk’s initial focus lies in creating functional 3D structures, as these models are intended for real-world applications.
Currently, the AI model excels at shape exploration during the early stages of design, although it produces relatively low-detail 3D forms.
However, as it is trained on larger, high-quality professional datasets—such as buildings, car designs, and game characters—it is expected to become increasingly valuable.
Which Autodesk applications will Project Bernini be used in?
Project Bernini remains in a “strictly experimental” phase and is not yet accessible for public use. Autodesk has not disclosed a specific timeline for its public availability or the integration plans with its existing products and services.
While the video showcases 3ds Max as an existing application, it appears to be used for editing a model exported from Project Bernini rather than directly for the generative AI process itself