CS631 Paper Response Form

Paper Title: A Resolution Independent Video Language


Author(s): Jonathan Swartz and Brian C. Smith


Main Point(s):

Rivl is and extension of Tcl/Tk that provides data types for images (sampled in space) and video (sampled in time). Nice features are that Rivl abstracts from resolution (i.e. sampling rate of both images and video), is platform independent, supports a wide range of image and video formats and is easy to use. Rivl's lazy-evaluation interpreter dynamically optimizes execution by building execution graphs, propagating input and output specifications through these graphs and intersecting these specifications in order to determine which parts of the data actually have to be processed. These execution graphs are also used to determine optimal caching strategies, because they allow a limited glimpse into the future.

Possible Use(s):

Rivl makes image and video editing very comfortable because it relieves the user of issues like resolution and format. It also gives programmers the possibility to do rapid prototyping of new image and video processing algorithms.

Extensions:

When I've understood Brian correctly, Rivl doesn't include yet the compressed-domain processing stuff but I think that it would be very nice to have these primitives and some more complicated routines build out of them in Rivl. Apart from that, it would be fun to include more fancy processing algorithms into Rivl (image/video understanding, but also more effects, like the ones where the screen folds up into, say, a bird and flies out of the screen ect.)

CS631 home page