Google has launched “Mini” stickers for iOS and Android, which use machine learning to craft personalized emoji from your photo. More precisely, the feature uses a combination of machine learning, neural networks and artist illustrations to conjure up the best representation of you, taking into account various characteristics like your skin tone, hair color and style, eye color, face shape and facial hair. Just access Mini from within Gboard and start the creation process by taking a selfie. It will then automatically create your avatar and generate packs of stickers you can use.
Formerly known as “Project Puppetron” (see previous re: the underlying tech), Adobe’s new Characterizer will soon be part of Character Animator CC:
You will be able to bring original art into Character Animator, record a series of sounds and facial expressions, and Characterizer will generate a new unique character… In a matter of seconds, you have a completely unique puppet that’s ready for your performance, regardless of your previous animation experience.
Here’s a quick tour from my buddy Dave:
Meanwhile After Effects is adding some fun new chops to its puppeting toolset:
“Teaching Google Photoshop” has long been my working mantra here—i.e. getting computers to see like artists & wield their tools. In a similar vein, researchers from Adobe & MIT have teamed up on “An AI for CGI“—tech that automatically separates image elements in discreet regions for augmentation. I can’t wait to give it a try:
Photoshop’s venerable (oh God, how can it already be venerable?!) Content-Aware Phil is growing up nicely, as seen in this new sneak peek from my friend Meredith. I look forward to seeing whether the increased power/finesse more than offsets the apparent jump in complexity—as I expect it will.