The Nvidia neural network turns photos into 3D models

The company has unveiled a new version of its Neural Radiance Fields (NeRF) artificial intelligence.

The tool is adapted to work with newer models of video cards, using which you can create 3D models even faster. Whereas the previous version of a neural network took several hours, now NeRF can do this in a fraction of a second.

The principle of artificial intelligence remains the same: the software is trained by examining pictures taken from different angles. The machine then studies the images, fills in the gaps and models the whole object.

You can use the tool  in different fields. For example, it can help create autopiloting systems. So, an autopilot will be able to navigate traffic at lightning speeds by estimating distances, dimensions, and shapes. Additionally, you can apply the neural network in architecture, construction and education.