Invariants of curves

Stochastic Analysis Seminar Series

Abstract:

The study of invariants of mathematical objects has a long history. A standard problem is for example the determination of features of a polynomial that stay invariant under linear coordinate transformations (modulo the determinant of the transformation). In the area of machine learning this topic has enjoyed a renaissance, since determining aspect of images which are invariant with respect to Euclidean transformations is often an important first step in image analysis.

The study of invariants of multidimensional curves is on the other hand not that well developed. Based on the signature, I will present a complete set of invariants to rotation (in 2d) and the general linear group (any dimension).

 

Location:
Speaker(s):

JOSCHA DIEHL

Start date:
Monday, March 3, 2014 - 14:15
End date:
Monday, January 20, 2014 - 15:15