Center for Imaging Science
Seminars/Colloquia/Invited Talks
Seminars
Francis Lazarus
Introduction to Computational Topology
| PLACE: | Clark 314
|
| EVENT: | CIS Seminar Series
|
| DATE: | May 3, 2005
|
| TIME: | 1:00 - 2:00
|
Presentation Slides
Abstract-
-
Computational topology deals with effective computations of
topological invariants over combinatorial structures. A topological
invariant of a (combinatorial) object is left fixed by a
transformation that do not alter its topology. A typical
transformation is a subdivision of a triangulation. Classical
invariants are algebraic, ranging from numbers (number of connected
components, Euler characteristic,...) to groups (the fundamental group
or the homology groups). When an invariant has multiple
representations, such as a set of generators for a group, its
computation can be coupled with an optimization problem such as: find
the minimal basis for the cycle space of a graph with respect to some
length definition.
Computational topology arises in many applications including
molecular analysis in bioinformatics, computer graphics (morphing,
texture mapping), shape and pattern recognition, etc...
In this talk I will introduce the main concepts of Computational
topology, the fundamental group and the homology groups, and present a
state of the art for low dimensional objects, including optimization
problems for homotopy of curves on a graph or on a surface.
Brief biography -
-
Francis Lazarus is a research scientist at CNRS (Centre National de la
Recherche Scientifique) in Poitiers, France.
His current research focus on computational aspects of topology, and
more specifically on optimization of curves on surfaces. His research
interests also include Computational Geometry and geometric problems
in Computer Graphics.
He received his doctorate in 1995 from the Department of Computer
Science of the University of Paris VII with a thesis on morphing
algorithms. He then spent 18 months as a Post-Doc in the IBM
T.J. Watson Research Center working on 3D geometry compression and
progressive transmission of polygonal models.
|