Ai Dreams Forum
Software & Hardware => Graphics and Video Software => Topic started by: Art on November 30, 2014, 01:20:11 am
-
This is the forefront of some amazing work with facial animation software and a Kinect device.
Stay with the video for a short time and see all that it can do. Very cool technology mapping!
Faceshift: Markerless Motion Capture (http://www.youtube.com/watch?v=24qUFDdZAG8#ws)
-
That's clever. Looks like movie tech is coming into our homes at last.
I had toyed with getting a Kinect to make motion capture animations.
-
Using a kinect is cheating in a way, it says 'markerless' but its using a depth cam... Im trying to get this to work with just an ordinary web cam, but a kinect definitely makes the job easier.
-
I'm with ranch. While this tech is amazing and relatively cheap, it is using so many unnatural sensors and means nothing to our field of research. We can tell depth, distance, faces and everything by just looking at a horrible VGA recording. A true achievement in AI and robotics research would be teaching a program to do this, thus making it more human.
Maybe I like that idea because I could test it at home with my webcam and not having to give $200 just to have a day of fun (as the case with my PS3, which I sold within a week of use lol).
-
cheers.
Even though the kinect is excellent, im definitely getting mine soon just for kicks, If someone made a kinect that worked off ordinary video, then it would be able to restore any video to 3d, as long as it wasnt too lossy - id have to give up eventually, sometimes the information just isnt there.
-
Yes you are right. Sometimes the info isn't there. If I weren't on the phone I'd make up a longer post. For now I'll just say that human visual perception is much more complicated. I'll use this chance to mention the concept of imagination, or maybe visual/auditory estimation ? Can we explain it like that?
By merely seeing someone's face you can imagine (or make an estimate of) it in all three dimensions.