You are here

Applications of Elastic Shape Analysis

Event Category:
Statistics and Data Science Seminar Series
Carlos Soto
University of Massachusetts Amherst

What is shape? In this talk I will give an overview of Elastic Shape Analysis (ESA) [Srivastava & Klassen, 2007], a framework for analyzing the shape of curves, and illustrate how it can be applied in statistical settings. The foundations of ESA are differential geometry in nature and hence, to no surprise, is an indirect consequence of the work of C.R. Rao. After the necessary and sufficient background I will demonstrate how to apply these tools in a statistical manner to accomplish tasks such as mean/median estimation, shape functional principal components, two-sample tests for populations of shapes, and hierarchical clustering of shapes.

Wednesday, March 6, 2024 - 1:30pm
LGRT 1685