GaunGAN, the new AI-powered real-time painting tool from Nvidia creates stunning sceneries from the most primitive brush strokes.
Environment artists and concept artists froze in horror because this little AI genius does what they do a billion times faster. You do not even need mind-blowing drawing skills to create stunning landscapes with this new app from Nvidia, all you need is some very basic understanding of shapes and composition — GauGAN will take care of the rest. The tool uses its AI-driven algorithm to create objects based on simple rough sketches and give you a much more refined version of what you paint with a regular brush.
NVIDIA Canvas uses a form of AI called generative adversarial networks, which consist of a generator and a discriminator. The generator is responsible for converting material maps into images. The discriminator has the knowledge about the features of various objects (e.g. that lakes contain reflections) and transfers this knowledge to the generator to make the imitation more accurate. GauGAN has been trained on an NVIDIA DGX system using over 5 million images.
Concept artist Jama Jurabaev explains that at the beginning of the creative process, it's you against a blank canvas and it can be hard to simply start painting because "you have an entire imaginary world you want to project on the screen" but all you see in front of you is a frustratingly empty white rectangle. Nvidia's new tool makes the first steps easier because with just a few strokes you no longer have a blank canvas, instead, you already see a beautiful piece coming together.
The "Styles" tab allows you to change the mood of the picture, alter the lighting and create either a dramatic effect or paint a peaceful picture. There are nine styles that modify the look and feel of a painting. You can also choose from 15 materials, including sky, grass, mountains, and others to tell the AI tool what you want to see in the final picture.
Jama Jurabaev emphasizes that this tool allows you to spend more time on the creative side of things rather than on the technical aspects of the process.