Date: Mon, 25 Nov 1996 20:26:44 GMT Server: Apache/1.1.1 Content-type: text/html Content-length: 3793 Last-modified: Fri, 05 Apr 1996 18:20:05 GMT John R. Kender



next up previous contents
Next: Shape from Darkness Up: Faculty Research Guide Previous: Gail E. Kaiser

John R. Kender

Research Area: Computer Vision, Robotic Navigation, Artificial Intelligence

My main research interests center around the higher levels of computer visual perception of objects and space. These higher levels involve surfaces, volumes, place locations, topological relationships, and verbal descriptions. This is the mental landscape in which vision meets mainstream AI, and where knowing the "where" of things is more important than knowing what the "what" of things are.

This research most recently has had four components. The most classical component is the development of shape from darkness, a (non-human) method to determining object surface shape from the shadows that the object casts, including shadows that it casts on itself. Less classical are two researches into navigation and description. The first of these uses topological relationships of objects to describe a navigational path without using metric quantities: this is qualitative direction giving, without the use of compass or odometer, much as people do on the back of an envelope. The second explores how the location of a single object in a cluttered enviroment can be described purely in spatial terms: medical radiologists do this for a living. The final investigation seeks to develop a Cheshire mouse: an indicating and selecting input device that requires no instrumentation other than the images of the moving and gesturing hand. Like Carrol's Cheshire cat, all that remains of the standard mouse is the smile (on the hand of the user).

We review these four investigations in order, from most classical in terms of vision research, to most multidisciplinary. We state what the objective of each is, summarize current progress, and sketch one or two areas of further research that would have the most scientific and engineering payoff.





Sabah S. al-Binali
Fri Sep 22 16:39:42 EDT 1995