Connect with us

Hi, what are you looking for?


Nvidia Research Creates Neuralangelo AI Model For 3D Reconstructions

The artificial intelligence model allows users to generate 3D structures from video with intricate levels of detail

Nvidia Research has created a new AI model for 3D reconstruction. Neuralangelo uses neural networks to convert 2D video clips into detailed 3D buildings, sculptures and other physical world objects. The AI model allows users to generate 3D structures with intricate levels of detail. The 3D objects can then be imported into design programs for further editing.

Neuralangelo can also translate complex textures like roof shingles and glass panes from 2D clips to 3D objects. With its high fidelity, the AI model’s reconstructions are easier for developers to quickly create viable virtual objects using smartphone video footage.

“The 3D reconstruction capabilities Neuralangelo offers will be a huge benefit to creators, helping them recreate the real world in the digital world,” says senior director of research and co-author of the paper Ming-Yu Liu. “This tool will eventually enable developers to import detailed objects — whether small statues or massive buildings — into virtual environments for video games or industrial digital twins.”

Neuralangelo Can Recreate Masterpieces

Nvidia Researchers showcased a demo of how Neuralangelo can recreate iconic objects such as Michelangelo’s David. The model is also capable of reconstructing building interiors which Nvidia showcased with a 3D model of the park at Nvidia’s Bay Area campus.

In the past, AI models have failed to reconstruct repetitive texture patterns and strong colour variations accurately. However, Neuralangelo is able to do so by adopting instant neural graphics primitives to capture fine details. The model uses 2D videos of an object taken from different angles to create a 3D object. First, the AI model makes a rough 3D representation. Then it optimises it and renders fine details.

Nvidia Research will present Neuralangelo along with nearly 30 other projects at the Conference on Computer Vision and Pattern Recognition (CVPR). The event will take place Jun 18th to 22nd in Vancouver. Nvidia also recently announced Nvidia ACE, a new custom AI model foundry service for games.

Written By

Jack Brassell is a freelance journalist and aspiring novelist. Jack is a self-proclaimed nerd with a lifelong passion for storytelling. As an author, Jack writes mostly horror and young adult fantasy. Also an avid gamer, she works as the lead news editor at Hardcore Droid. When she isn't writing or playing games, she can often be found binge-watching Parks & Rec or The Office, proudly considering herself to be a cross between Leslie Knope and Pam Beasley.

You May Also Like

Level Up

Eager to be at the metaverse frontier, but not sure how to get started? As exciting as the idea of a shared digital space...


New blockchain gaming platform based on Unreal Engine 5.


The record for the most expensive land sale in the metaverse has just been raised


Fashion brand teams up with proto-metaverse for two new eyewear options, the Helux and Hydra


Subscribe to the future