Adobe Sensei Improves Selfies Using Artificial Intelligence
ADOBE’S NEW AI PLATFORM WILL TRANSFORM MOBILE PORTRAITURE AND YOUR SELFIE GAME
Today, Adobe revealed a sneak peak into one facet of its research technology powered by the mystical Adobe Sensei, its AI and machine learning platform, and it’s absolutely absurd. It demos a combination of perspective effect editing, automatic, software-only portrait masking, and portrait style transfer all within smartphone portraits, using deep learning and AI to turn ordinary smartphone portraits into “good ones.”
The one-minute video shows an average-looking man on an iPhone, as he turns a selfie into a portrait, adjusting distance and perspective with the slide of a finger. Then, in a tab called ‘segmentation’ he applies automatic portrait masking, letting him easily adjust depth of field in just one swipe.
Now this is where things get scary—in a part of the platform called ‘Search for styles,’ he’s able to comb through what appears to be a standard image search, and copy and paste the style of any photo he comes across to his own.
“In one click, you can edit it like a pro,” the video says.
I’m all for efficiency, but at what point do things become too easy? I would never argue that a platform such as this would cannibalize photography, but it surely raises questions of what defines plagiarism when it comes to digital imagery. Just imagine the edit of any photo you posted online could be easily taken and applied to any image by any person, with no way of protecting yourself.
That may very well be the future we’re looking at, but one things for sure, it’s that your selfie game won’t suffer.