Progress in persistence for shape analysis

Abstract

Persistent topology mitigates the excessive freedom of topological equivalence by studying not just a topological space but a filtration of it. This makes it a very effective class of shape descriptors, with an impressive potential for applications in the image context, in particular when it comes to images of natural origin. Research in this field is lively and follows various threads. The talk will sample some recent results without any attempt to completeness.