The AI model transforms 2D video clips into detailed 3D structures for export and use in design applications across art, game dev, robotics, and digital twins.
NVIDIA has just released the source code for Neuralangelo, a popular AI model from NVIDIA Research for generating 3D reconstructions. Neuralangelo turns 2D video clips into detailed 3D structures — generating lifelike virtual replicas of buildings, sculptures, and other real-world objects.
The NVIDIA blog “Digital Renaissance: NVIDIA Neuralangelo Research Reconstructs 3D Scenes” highlights the new AI model, comparing it to Michelangelo's ability to create lifelike visions from marble blocks. Neuralangelo generates 3D structures with intricate details and textures that can then be imported into design applications for use in art, video game development, robotics, and industrial digital twins.
Watch: Digital Renaissance: Neuralangelo by NVIDIA Research Reconstructs 3D Scenes from 2D Video Clips:
Neuralangelo is one of nearly 30 projects presented last June by NVIDIA Research at the Conference on Computer Vision and Pattern Recognition (CVPR) in Vancouver. The papers spanned topics including pose estimation, 3D reconstruction, and video generation.
Source: NVIDIA