Remember back in 2015 when Google's DeepDream showed us how AI sees the world? Principally as stoner-peacock-feather-eyeball-cascade-from-Hellraiser? Well, thanks to TensorFlow -- Google's open-source machine learning tech -- we have something even more wonderful (and possibly weirder) to enjoy in 2017: Pix2Pix.
Inspired by Edges2Cats, Pix2Pix uses image-to-image tech to turn line art into a photo. It basically establishes a relationship between a given set of images and a drawing. Wielding crazy math (or magic), it gives a user the ability to draw whatever they want into a box, then process it into a photo-realistic image presented in another box. Edges2Cats creator Christopher Hesse actually created several flavors of these mind-bending modules and THEY ARE ALL WILD. So wild that, naturally, we had to try them on some GoComics stars. Apologies to all of our celebrated cartoonists. What I've done, I've done for science.
Charlie Brown of Peanuts
Jason from FoxTrot
Sarah of Sarah's Scribbles
Stephan Pastis as his Pearls Before Swine alter ego
Joey Alison Sayers of Joey Alison Sayers Comics
Big Nate's Nate Wright
Cathy run through edges2handbags, given her affection for them
Actually, Garfield looks better with Edges2Cats...
Webcomic Name creator Alex Norris already made one himself!
in case you ever wondered what blob and "oh no" speech bubble would look like if they were human pic.twitter.com/Qowz1leMXO— Alex Norris (@dorrismccomics) June 5, 2017
You can melt your mind with your own creations at Pix2Pix.