Extreme Makeover: Computer Science Edition
zunavision.stanford.edu
zunavision.stanford.edu
It's pretty neat that the technology not only handles pans, zooms, and rotation, but also shadows and people walking in front of the 'screen'. It seems like a much more advanced version of the first down line for football on tv. I wonder in what other interesting ways it will be used when this can be done in real time.
http://query.nytimes.com/gst/fullpage.html?res=9C06E7DB173AF...
I've seen this ability to replace a moving rectangle in a video with other content in 3D software before. With high definition video, I imagine this process is more accurate than ever.
I think a bigger question is how to get someone to pay you to add fake ads to your YouTube submissions, and how to track that somebody is actually doing this, and for how many seconds. That could get very interesting.
Also, related:
Related, the other side: how to detect that a video WAS modified:
http://news.cnet.com/Are-fake-videos-next/2100-1008_3-611344...
They already have Adwords, and they acquired that huge banner ad company, doubleclick. They also acquired an in-game ad company (which is probably the most similar to this concept, in terms of technology).
A missing component is them paying the creator of the video...
If everything can be modified in the background in near realtime, how can we be assured any video is real? We probably cant already, but its just sort of an interesting realization.
I never knew we were even close to being able to do stuff like this with video. These techniques are actually able to use depth-estimation algorithms to something useful. Perhaps we'll soon be able to make convincing 3D models based on video footage.