GBFPUM - A MATLAB Package for Partition of Unity Based Signal Interpolation and Approximation on Graphs

Abstract

This is a manual for the software package GBFPUM, a MATLAB toolbox for the generation of a partition of unity on graphs and their application as an interpolation and approximation tool for graph signals. GBFPUM combines local kernel approximation based on graph basis functions (GBFs) with a partition of unity method (PUM) in order to obtain a low-cost global interpolation or classification scheme for large graphs. We give a detailed description of all the routines implemented in the MATLAB package and show how the code can be used to interpolate graph signals in concrete examples.

De Rossi A., Cavoretto R., Erb W. (2022) "GBFPUM - A MATLAB Package for Partition of Unity Based Signal Interpolation and Approximation on Graphs " Dolomites Research Notes on Approximation, 15(2), 25-34. DOI: 10.14658/PUPJ-DRNA-2022-2-3  
Year of Publication
2022
Journal
Dolomites Research Notes on Approximation
Volume
15
Issue Number
2
Start Page
25
Last Page
34
Date Published
10/2022
ISSN Number
2035-6803
Serial Article Number
3
DOI
10.14658/PUPJ-DRNA-2022-2-3
Issue
Section
SpecialIssue