Nvidia’s Latest AI Software Turns Doodles Into Realistic Landscapes

Nvidia has demonstrated the latest technology, which will make artists more productive with the help of the AI. At the company's Graphics Technology Conference (GTC) 2019, the company has unveiled a mind-blowing new image creator. The software called GauGAN uses generative adversarial networks (GANs) that allow users to transform rough sketches into detailed and quite photorealistic landscapes.

Generative adversarial network (GAN) is an AI model which described as a “smart paintbrush”. Nvidia software can support people to make impressive images with some basic outlines of scenes. The GauGAN is able to generate landscapes instantly, and it’s surprisingly intuitive.

Nvidia Official Demo: Changing Sketches into Photorealistic Masterpieces

The deep learning model has trained GauGAN on million images then GauGAN can synthesize thousands of objects and their relation to other objects in the real world. Swapping a segment label from “grass” to “snow” makes the entire image change to a winter scene. The GauGAN will produce results unique to the input. In comparison with other similar tools in the past, it seems to be the most polished demonstration concept.

GauGAN gives an auto-reflection scene artwork.

Nvidia does not have any plans to release the software commercially. The company claims that they are working to improve this. In the future, Nvidia expects the useful tool would allow everyone from artists to architects to game developers the ability to create high-quality virtual worlds in just seconds.