You might be interested in "deep fakes" deep learning video editing software. It uses a library of 1 person's face as training, and overlays that person's face onto a person in a given video, mimicking mouth movement and expressions

@mcpaccard "generated" isn't it just slightly mixing up existing photos?

@piggo @mcpaccard I think the video accompanying the paper gives a better overview of what they are doing.


@mcpaccard : Si ça c'est pas un signe qu'on est déjà dans la Matrice :troll:

The technology is great, but unfortunately I can only think of bad applications coming out of this.
If anyone has a good idea, I'd be glad to read it ! :)

@mcpaccard Why don't you tell the relevant news? We can have now literally infinite cat pictures!

«Uncurated set of images produced by our style-based
generator (config F ) with the LSUN C AT dataset at 256²»
@mcpaccard (just kidding, it concerns me a lot. thank you for the info!)
Sign in to participate in the conversation

A Mastodon instance for and by people who make things!