In the context: Nvidia’s GauGAN technology has already shown what it can do by turning simple sketches into photorealistic images. We’ve since seen it used in Nvidia Canvas, but it looks like the GPU giant is aiming for higher targets with its AI, launching a new version capable of turning words into images.
Nvidia first showed its GauGAN technology in 2019, but it was only recently that we saw it featured in a product available to the general public. Called Canvas, this program can be very fun to use, allowing users to create stunning photo-like images with simple sketches.
Several months have passed since the announcement of Canvas, but work on GauGAN continues, and now it is coming to version 2.0. The technology has become even more impressive since now able to turn words to photorealistic images, providing a result similar to what we get when using the paint to image function.
As shown in the video above, write something in the text box and an image will be generated immediately based on your words. Add an adjective or replace a noun in a phrase and the image changes accordingly.
For additional personalization, users can combine text and image drawing functions. By using written words to create bases and drawings to detail an image, you can change the shape, size, and texture of any object in an image.
To achieve these results, Nvidia’s GauGAN 2 text-to-image feature uses a generative adversarial AI network model that “combines segmentation mapping, drawing, and text-to-image generation.” This model has been trained using 10 million landscape images, so it should be well prepared for whatever you throw at it (or not).
You can try using Nvidia’s interactive AI demo for GauGAN 2 using a web browser. To play with it, you’ll need to agree to Nvidia’s terms first (check the box at the bottom of the page).