A visual effects expert is carrying out a technical experiment using a machine learning algorithm, transforming animated characters on social media into plausible real people … but a nightmare!
You ever thought of meet from life i Pixar or Disney characters, the nice Russell from Up or Flash from The Incredibles? As a provocation and experiment, a technician from visual effects Californian named Nathan Shipley has started sharing graphic proofs with Twitter algorithms machine learning: for the uninitiated, in simple terms these are computer procedures, governed by an Artificial Intelligence that learns the more the more it performs similar tasks, able to modify images by following and applying rules and logic. In this case, it “humanizes” characters in stylized or cartoon graphics.
Today’s reverse toonification experiments with art from @Pixar for Incredibles 2, Up, & Coco.
More Pixar in thread! ⬇️ pic.twitter.com/uI2v4Tn96m
— Nathan Shipley (@CitizenPlain) October 15, 2020
The result is truly disturbing and we honestly hope it doesn’t go beyond these test nightmare, but it is an example of where CGI can be pushed if integrated with these algorithms: for a psychological reason, the apps that transform us into cartoons do not disturb us, but the reverse process, the flooding of the imagination into our real world , hits below the belt. It’s another face of software development that has led to apps that simulate aging on our photos and – in the worst case – deepfakes.
The encoder and frameworks that allow all this travel under the name Pixel2Style2Pixel, with sources available on Github.
More in thread: pic.twitter.com/se4XM5S4J7
— Nathan Shipley (@CitizenPlain) October 13, 2020