Universität Bielefeld - Graduiertenkolleg Aufgabenorientierte
Kommunikation
The Rise and Fall of Generic Object Recognition
Department of Computer Science and
Center for Cognitive Science
Rutgers University
Tuesday, May 23rd, 2000, 2.15 p.m., Hörsaal
Generic object recognition enjoyed the spotlight in the 1970's, sporting
prototypical models composed of parameterized, 3-D volumetric parts.
These systems were ambitious, but met with little practical success:
they were slow, cubmersome, and often restricted to contrived scenes.
Since then, research in object recognition has steadily moved away from
generic object recognition and closer towards exemplar-based
recognition, in which the exact geometry or appearance of an object
instance is known. In this talk, I will discuss this trend and the
effect it is having on the computer vision community. Next, I will
review a number of generic object recognition systems that, over the
years, have resisted this trend. Finally, I will discuss why the time is
right for generic object recognition to make a comeback.
Anke Weinberger, 2000-04-19