Jay Rose Posted December 1, 2016 Report Share Posted December 1, 2016 A group at MIT has come up with a way to generate very real-looking motion videos, with people and objects doing what they'd normally do, based on a single still photo of the scene! Article at http://web.mit.edu/vondrick/tinyvideo/ has before-and-after shots, along with a discussion of the algorithm and machine learning involved. The process is in its infancy and generating only short, small-format clips. But we've all seen how quickly these things can be developed, once you throw money and programmers at them. For those who say, "well, what about the story? We'll always need to have writers and actors and directors..." I respond - not really tongue-in-cheek - what about reality tv? There's a heck of a lot of programming being developed where the story is tacked on in post. And there's a lot of money to be made when this process is developed well enough to fill cable tv slots. Not to mention all its other implications... Quote Link to comment Share on other sites More sharing options...
Dutch Posted December 1, 2016 Report Share Posted December 1, 2016 Scary, and no doubt we're heading in that direction. Quote Link to comment Share on other sites More sharing options...
Philip Perkins Posted December 1, 2016 Report Share Posted December 1, 2016 yeah we're doomed. Now can I get back to buying more gear? Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.