Document Type

Presentation Abstract

Presentation Date

10-15-2018

Abstract

[Download attached PDF for the complete abstract with photos placed before and after the abstract text.]

Applications of geometric data analysis often involve producing collections of subspaces, such as illumination spaces for digital imagery. For a given collection of subspaces, a natural task is to find the mean of the collection. A robust suite of algorithms has been developed to generate mean representatives for a collection of subspaces of fixed dimension, or equivalently, a collection of points on a particular Grassmann manifold. These representatives include the ag mean, the normal mean, the projection mean, and the Karcher mean. In this talk, we catalogue the types of means and present comparative heuristics for the suite of mean representatives. We respond to, and at times, challenge, the conclusions of a recent paper outlining various means built via tangent-bundle maps on the Grassmann manifold.

Additional Details

Monday, October 15, 2018 at 3:00 p.m. in Math 103
Refreshments at 4:00 p.m. in Math Lounge 109

Download

Share

COinS