Researchers from University of Washington have developed a technique to drastically enhance video clips using high-res photos taken of the same scene.
The technique can be used for several image enhancements including; fixing exposure, bringing in lost details, enhancing colors and depth, removing camera shakes, adding image data or completely eliminating objects from a scene.

Loading Video
(Note: You need Flash 9+ to watch these videos)
It is all very amazing.
(via Motionographer)


Fascinating stuff this. I love the step in their process that they call "spacetime fusion" (I'd give you the timestamp but you don't show that).

Writer, Project Manager and Interaction Designer
follow me on Facebook, Twitter, Flickr, or visit my website
Jeton, they should
Jonathan, Now I do (showing timestamps - just added it). And space-time fusion is cool.
wow - this is amazing stuff! its a very clever piece of joined up thinking - well explained too.
Wow! These stuff definitely has to be in every video editing software out there...

all your effects with stabilization, reflections and changing pictures already existed. never heard of monet, syntheyes or ???
the new stuff is the thing about images + video. the rest already exists and is better than your implementation

alex, yes those effects all exists, but they work in an entirely different way. here things like the masks on the photos are all automatically adapted. and please tell me one software that creates the stylization effects in the last scenes in the same quality, all do it by doing it per frame and just adapting their seed. also the ease of object removal is unmatched.



Baekdal.com is (nearly) everywhere. Check out our social profiles:
Jeton
visit my website
That is really impressive.
I think the movie industry needs to talk with these researchers.