Tumblelog by Soup.io
Newer posts are loading.
You are at the newest post.
Click here to check if anything new just came in.

The holy grail is to give Photoshop computer vision. The app should simply select "objects" the way users see, like a "beach ball" or a "tree" or a "head," not as "blob of color one," "blob of color two." Then the user should be able to do what she pleases to the object, with the software filling in the details like what might've been behind that object — something that's available in a nascent form in CS6. Content vision also means the software should know when you're working on a family photo and when you're working on a logo, adjusting color grading techniques accordingly. It means unifying many of Photoshop's features — which, once again, its architecture is uniquely suited to do.

Jeff wants to teach Photoshop to be more user-aware as well — what he's really working toward is artificial intelligence. It's a surprising idea coming from Adobe, because it seems like something that will be solved by a cutting-edge internet giant like Google, not the developers of a 23-year-old app originally meant to display photos on black-and-white Macintosh screens. But that’s where Photoshop is today.

Reposted byjaerksebastiankomandorJohann

Don't be the product, buy the product!