Title: Geometric and invariant aspects of complex functional data
Authors: Karthik Bharath - University of Nottingham (United Kingdom) [presenting]
Abstract: High-resolution imaging has led to increasing availability of functional data with rich geometric information: range and/or domain might be manifolds with symmetries; or, set of functions themselves might have a nonlinear structure. We will consider some specific examples (e.g. closed 2D curves representing tumour shapes) of such data, discuss their common features, and present some of the challenges involved in developing statistical models and methods that are compatible with the underlying geometry and invariances.