Article View: csu.cs
Article #13CSU CS BMAC Seminar on 10/14/96
From: barbara mclean
Date: Thu, 10 Oct 1996 00:00
Date: Thu, 10 Oct 1996 00:00
51 lines
2838 bytes
2838 bytes
DEPARTMENT OF COMPUTER SCIENCE
COLORADO STATE UNIVERSITY
BARNEY'S
MONDAY AFTERNOON CLUB SEMINAR
Speaker: Bruce Draper
Computer Science Department
Colorado State University
Date: October 14, 1996
Time: 3:30 - 4:30 p.m.
Place: 100 USC
Refreshments: following the talk
Knowledge-Directed Vision, Past and Future
The knowledge-directed approach to image interpretation,
popular in the 1980's, sought to identify objects in
unconstrained two-dimensional images and to determine the
three-dimensional relationships between these objects and
the camera. Among the issues faced by these systems were
variations among instances of a class; differences in
attributes among object class definitions (e.g., shape,
color, texture, size, context, substructures); and handling
uncertainty. Knowledge-based systems tried to overcome
these problems by applying large amounts of object- and
domain-specific knowledge to interpretation problems.
In this talk we argue that knowledge-directed vision
systems failed for two reasons. The first is that the
low- and mid-level vision procedures that were relied
upon to perform the basic tasks of vision were too
immature at the time to support the lofty interpretation
goals of these systems. This problem, we conjecture,
has been largely solved by recent advances. The other
impediment was that the control problem for computer
vision was never properly addressed as an independent
problem. This talk will conclude with a brief review of
several new projects which, although still in the early
stages of development, address the complex control issues
that continue to obstruct the development of knowledge-
based and model-based vision systems.
Note: this talk is based on the paper "Knowledge-Directed
Vision: Control, Learning and Integration" by B. Draper,
A. Hanson and E. Riseman, to appear in this November's
issue of the Proceedings of the IEEE.
Technical Details
Message-ID:
<53jk0m$fk@vivaldi.cs.colostate.edu>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 10244d,d57fe679260d54b6,start X-Google-Attributes: gid10244d,public X-Google-Thread: 15fcd9,d57fe679260d54b6,start X-Google-Attributes: gid15fcd9,public X-Google-Thread: 148df6,d57fe679260d54b6,start X-Google-Attributes: gid148df6,public X-Google-Thread: 12897c,d57fe679260d54b6,start X-Google-Attributes: gid12897c,public From: mclean@vivaldi.cs.colostate.edu (barbara mclean) Subject: CSU CS BMAC Seminar on 10/14/96 Date: 1996/10/10 Message-ID: <53jk0m$fk@vivaldi.cs.colostate.edu> X-Deja-AN: 188788746 distribution: co organization: Colorado State University -- Computer Science Department newsgroups: csu.cs,csu.cs.grad,csu.general,csu.lance