Some very smart folks in the University of Washington’s Computer Science department are cooking up some pretty amazing procedures that could well revolutionize the process of creating visual effects.
First up is a process for interactively extracting a matte for an image – what’s so incredible about this is that the object being extracted is not shot on a greenscreen, the background does not need to be static, and the camera does not need to be locked down. Say what? See this link for an abstract and a video.
Next, even more incredible, is a process for enhancing video by using a few high-resolution static images of the same scene. Enhance how, you may ask? How about giving the entire video more resolution (even though the camera is moving and the still image is from just one angle), increasing dynamic range, and changing or completely removing objects from a scene – all using an almost automatic process. Read about it and watch the video by clicking here
It’s research and raw code at this point, but the results provide a window to where visual effects may well be headed. Fascinating stuff.