Friday, February 24, 2017
3:00 PM - 4:00 PM
VC 2306
Event Type
Speaker or Lecture: Featured
Contact
Agustin, Marc
618-650-3681
Sponsor
Marcus Agustin
Link
https://ems.siue.edu/MasterCalendar/EventDetails.aspx?EventDetailId=120128
The Department of Mathematics and Statistics would like to
invite faculty and students to a special colloquium talk on Friday, February
24th at 3pm. The speaker is Dr. Chalani Prematilake from Texas Tech University
and her talk is titled Comparing the Performances of Different
Parameterizations used for Dimension Reduction in Approximating Shapes of
Planar Contours. Please see below for more information about the colloquium
talk.
Mathematics and Statistics Colloquium
SIUE features: Dr. Chalani Prematilake - Department of Mathematics and Statistics, Texas Tech
University
Comparing the Performances of Different Parameterizations
used for Dimension Reduction in Approximating Shapes of Planar Contours
Abstract:
As a result of rapid advancements of technology, high
dimensional data can be easily found almost everywhere. One type of such data
is the shape of a contour, where a contour may be viewed as the 2D outline of
the image of an object. This type of data arises in medical imaging as well as
in computer vision and can be modeled as data on a manifold. It can thus be
studied using statistical shape analysis.
Practically speaking, each observed contour, while
theoretically infinite dimensional, must be discretized for computations. As
such, the coordinates for each contour are obtained at k sampling times,
resulting in the contour being represented as a k-dimensional complex vector.
While choosing large values of k will result in closer approximations to the
original contour, this will also result in higher computational costs in the
subsequent analysis. The goal of this study is to determine reasonable values
for k so as to keep the computational cost low, while maintaining accuracy. To
do this, we consider three parameterizations for selecting sample points and
determine lower bounds for k for obtaining a desired level of approximation
error using two different criteria.
Because this process is computationally inefficient to
perform on a large scale, we then develop models for predicting the lower
bounds for k based on simple characteristics of the contours. Current procedure
develops simple linear regression models to predict the lower bounds for k
using three characteristics of the original contour, length of the contour, total
absolute curvature, and the number of times the curvature changes its sign.
Study discusses procedures for model validation as well as the effectiveness
added to the models by smoothing.
Key Words: Statistical shape analysis, Statistics in
manifolds, planar contours