marina@ai.toronto.edu (Marina Haloulos) (11/29/89)
Department of Computer Science, University of Toronto
(SF = Sandford Fleming Building, 10 King's College Road)
-------------------------------------------------------------
COLLOQUIUM
SF1105, at 11:00 a.m., Tuesday 5 December 1989
Professor Demetri Terzopoulos
University of Toronto
Visual Modelling: Deformable Models in Vision and Graphics
Computer vision and graphics are mutually converse disciplines --- the
former is concerned with the analysis of images, the latter with their
synthesis. They pose similar problems at the object modeling level. The
Visual Modeling research program takes a unified approach to vision and
graphics through the apparatus of computational physics. In this talk I
shall focus on physically-based, deformable models and associated force
field techniques for synthesizing and analyzing the shapes and motions of
nonrigid objects.
I will demonstrate the application of deformable models to vision problems
such as adaptive image sampling, image contour extraction, stereo and
motion estimation, visible-surface and 3D object reconstruction, and the
recovery of 3D shape and nonrigid motion of objects from dynamic stereo
imagery. The video presentation will also include computer graphics
animation of deformable models illustrating a variety of behaviors in
simulated physical worlds, along with a recent application: physically-
based modeling and interactive animation of the human face.