Sign In

Communications of the ACM

ACM TechNews

Beyond Deep Fakes


The new method was used to transfer the facial expressions of comedian John Oliver to those of a cartoon character.

Researchers at Carnegie Mellon University have developed a new method of automatically transforming the content of one video into the style of another.

Credit: Carnegie Mellon University

Carnegie Mellon University researchers have developed a new method of automatically transforming the content of one video into the style of another.

Since the data-driven method (called Recycle-GAN) does not require human intervention, it can quickly transform large amounts of video, making it useful for movie production.

In addition, the new system can be used to colorize black-and-white films or create content for virtual reality experiences, as well as potentially being used to create "deep fakes," videos in which a person's image is inserted without permission.

The technique relies on generative adversarial networks that help computers understand how to apply the style of one image to another.

From Carnegie Mellon University
View Full Article

 

Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account