Abstract

The iconic representation of video is film: sprocket holes, celluloid, a sharp blade and mylar tape; and these metaphors underpin the interaction paradigms of the most modern video editing software. Even 3D special effects are inserted as “layers” keyed by “mattes”.

In this talk I will discuss a number of more direct interactions with the moving image, dealing with 3D object insertion, object tracking, and new representations for video which were not possible without modern computation. I show how recent research in computer vision and computer graphics allows us to think of new paradigms in manipulating and representing video, ultimately editing in the image stream, not on it.