Nvidia Research announced Neuralangelo, an AI model that turns 2D video clips into 3D visual reconstructions. The platform, a reference to Michelangelo, can generate detailed and textured 3D structures from basic video clips. The model can render virtual 3D representations of buildings, sculptures, and other real-world objects or spaces, which can be imported into design apps for editing. - It requires at least several 2D video frames of the object at various angles.
- Neuralangelo has applications in areas like video game development, digital twin creation, and art.
- Nvidia says the AI model was adapted from Instant NeRF, its AI tool demoed last year that creates 3D objects from 2D photos in milliseconds.
- It could be used to capture video conferences in 3D, reconstruct scenes for 3D digital maps, and rapidly create places and avatars for the metaverse.
|
