A Style-Based Generator Architecture for Generative Adversarial Networks from Nvidia.

A Style-Based Generator Architecture for Generative Adversarial Networks.
"Networks Learn to generate entirely new images that mimic the appearance of real photos, With the limitation on control over the generated images."
The new generator which can automatically learn to separate different aspects of the images without any human supervision.All the images in Video were produced using a generator and the photographs are not of some real people.
It's unbelievable to see the clarity of the images and i see upcoming potential usage of this technology. Though Google does started to edit real images from AI this is something incredible the way generators edit, change and give options to choose the strength of characteristics to create an image.
Please see the video until the end to see how generators were used in animals, Automobiles, and types of furniture.
To learn more please do visit Nvidia research site.

Leave a comment

Make sure you enter all the required information, indicated by an asterisk (*). HTML code is not allowed.