Generalized Samplets

Abstract

Recently, the concept of samplets has been introduced by transferring the construction of Tausch-White wavelets to scattered data. A multiresolution analysis tailored to discrete data which directly enables data compression, feature detection and adaptivity was obtained. This was done by using finite linear combinations of dirac distributions.

In this project we take a frame theoretic viewpoint. We e.g. investigate generalized finite samplets by combining functionals, and extending the approach to redundant and infinite dimensional settings.

This project is clsoely related to the SNSF Starting grant: "Multiresolution methods for unstructured data", see https://data.snf.ch/grants/grant/211684

Mathematics

Team

Cooperation Partner