Nvidia is no stranger to developing ground-breaking technologies, such as correcting blurry photographs or utilizing AI to create portraits of individuals. They’ve now astonished us once again, this time with a tool that converts your sketches into gorgeous pieces of art.
GauGAN is the name of the company’s latest tool, a mix of the generative adversarial networks (GAN) acronym and the last name of French painter Paul Gauguin. This new technology, according to Nvidia, may assist everyone, from architects and urban planners to landscape designers and game developers, in constructing virtual environments.

Nvidia notes, “GauGAN lets users create their segmentation maps and alter the image. And labeling each section with names such as sand, sky, sea, or snow.”

ACCORDING TO THE COMPANY, the AI was trained on a million photographs: “Draw in a pond, and adjacent things like trees and rocks would emerge as reflections in the water.”
Image Courtesy: Nvidia
Because the technology has taught with genuine photos, it understands and can produce reflections in natural bodies of water.

Bryan Catanzaro, vice president of applied deep learning research, characterizes it as “like a coloring book image that shows where a tree is, where the sun is, where the sky is.”

GauGAN enables the user to alter the produced pictures to match a particular painter’s style and adjust the scene’s daylight.

“This technique isn’t simply copying and pasting textures or sewing together fragments of other photos,” Bryan Catanzaro explains. “It’s synthesising new visuals, similar to how an artist would sketch anything,” says the researcher.

In June, this new approach will be described in a research paper given at the Computer Vision. Those attending this week’s GPU Technology Conference may even check out GauGAN at the Nvidia booth!
Please let us know your opinions in the comments section below.