NVIDIA’s AI Tool Transforms MS Paint-Like Scribbles Into Realistic Landscapes2 min read


Video screenshot via NVIDIA

Landscape paintings that take weeks, or even months, to complete are now a thing of the past.

If your illustration expertise involves sketching inverted W’s to create birds in the sky, NVIDIA’s latest AI technology might be something worth coveting. The tech firm, known for its over-the-top machine learning projects, has unveiled an AI tool that instantly converts doodles into photorealistic landscapes.

The tool is called ‘GauGAN’, named after French post-Impressionist artist Paul Gauguin. The AI was fed with about a million photos to grasp how to recreate and fill in areas of landscapes.

Side-by-side samples reveal that all you need to do input rough brushstrokes, to which ‘GauGAN’ will generate elements of landscapes in real time, essentially turning any MS Painter and stickman artist into an actual pro.

“A novice painter might set brush to canvas aiming to create a stunning sunset landscape—craggy, snow-covered peaks reflected in a glassy lake—only to end up with something that looks more like a multi-colored inkblot,” the company describes in a blog post.

“But a deep learning model developed by NVIDIA Research can do just the opposite: it turns rough doodles into photorealistic masterpieces with breathtaking ease.”

To bring your vision to reality, the ‘GauGAN’ interface is accompanied with options that recreate natural features like skies, grass, trees, rocks, sand, flowers, mountains, and clouds.

“It’s like a coloring book picture that describes where a tree is, where the sun is, where the sky is,” details Bryan Catanzaro, Vice President of Applied Deep Learning Research at NVIDIA. “And then the neural network is able to fill in all of the detail and texture, and the reflections, shadows and colors, based on what it has learned about real images.”

At present, the tool isn’t available publicly, but NVIDIA hopes technology like ‘GauGAN’ will benefit anyone from architects to casual artist in the future.


Video screenshot via NVIDIA


Video screenshot via NVIDIA


Video screenshot via NVIDIA


Video screenshot via NVIDIA

[via PetaPixel, videos and screenshots via NVIDIA]