Experience AI in action from NVIDIA Research

AI and deep learning is serious business at NVIDIA, but that doesn’t mean you can’t have a ton of fun putting it to work. Researchers at NVIDIA challenge themselves each day to answer the “what ifs” that push deep learning architectures and algorithms to richer practical applications. See some of that work in these fun, intriguing, artful and surprising projects.

Sign up for notifications when new apps are added and get the latest NVIDIA Research news.

NVIDIA Vid2Vid Cameo

NVIDIA Vid2Vid Cameo

Roll out of bed, fire up your laptop, and look picture-perfect for every video call. With Vid2Vid Cameo, one of the deep learning models behind the NVIDIA Maxine SDK for video conferencing, users can submit a polished, work-appropriate 2D photo or cartoon avatar prior to a call and the AI model will use GANs to synthesize realistic talking head videos based on the still image. In this demonstration, the AI maps facial movements to capture real-time motion and changing viewpoints.

View Research Paper > | Watch Video > | Resources >

NVIDIA AI Playground GameGAN


Forty years since PAC-MAN first hit arcades in Japan, the retro classic has been reimagined, courtesy of artificial intelligence (AI). Trained on 50,000 episodes of the game, GameGAN, a powerful new AI model created by NVIDIA Research, can generate a fully functional version of PAC-MAN—this time without an underlying game engine. This means that even without understanding a game’s fundamental rules, AI can recreate the game with convincing results. It's the first neural network model that mimics a computer game engine by harnessing generative adversarial networks, or GANs.

PAC-MAN™&©BANDAI NAMCO Entertainment Inc

View Research Paper > | Watch Video > | Resources >

NVIDIA AI Playground Ganimal


We’ve all passed a Chihuahua on the street that’s the size of a guinea pig with the attitude of a German Shepherd. With GANimal, you can bring your pet’s alter ego to life by projecting their expression and pose onto other animals.

Once you input an image into the GANimal app, the image translation network unleashes your pet’s true self by projecting their unique characteristics onto everything from a lynx to a Saint Bernard.

View Research Paper > | Watch Video > | Resources >



GauGAN, named after post-Impressionist painter Paul Gauguin, creates photorealistic images from segmentation maps, which are labeled sketches that depict the layout of a scene.

Artists can use paintbrush and paint bucket tools to design their own landscapes with labels like river, rock and cloud. A style transfer algorithm allows creators to apply filters — changing a daytime scene to sunset, or a photorealistic image to a painting. Users can even upload their own filters to layer onto their masterpieces, or upload custom segmentation maps and landscape images as a foundation for their artwork

View Research Paper > | Read Blog > | Resources >

Image Inpainting

NVIDIA Image Inpainting

Step right up and see deep learning inference in action on your very own portraits or landscapes. Our researchers developed state-of-the-art image reconstruction that fills in missing parts of an image with new pixels that are generated from the trained model, independent from what’s missing in the photo. Give it a shot with a landscape or portrait. Erase at will — get rid of that photobomber, or your ex, and then see what happens when new pixels are painted into the holes.

View Research Paper > | Read Story > | Resources >


Explore the NVIDIA AI Art Gallery

Join Our Team

Come solve the greatest challenges of our time.