signal processing

Topology-based Filtering of Graph Signals via Persistent Homology

arXiv:2408.14109

summary

The paper introduces a topology-based filter for signals defined on graph vertices and mesh faces that removes features with low topological persistence, offering a controllable alternative to traditional graph-spectral filters.

Abstract

We study topology-based filtering of vertex-defined signals on graphs and their two-dimensional analogues. Unlike graph-spectral filters, the proposed approach distinguishes features by topological persistence rather than by spatial wavelength or periodicity. We consider graphs with faces embedded in surfaces, a class that includes discrete models of images and meshes. We prove that, in general, exact simultaneous removal of low-persistence features in dimensions and is impossible. This motivates a relaxed formulation, for which we introduce the Low Persistence Filter (LPF). The LPF removes finite-persistence features below a prescribed threshold while controlling the resulting perturbation of the signal. We illustrate the method on one-dimensional signals, two-dimensional images, and signals on triangular meshes. A Python implementation is publicly available.

codes available at https://github.com/mvlier/topapprox

Topics & keywords

#graph signal processing#persistent homology#topological data analysis#filter design#mesh processinglow persistence filterLPFℓ∞ perturbationtopological persistencegraph spectral
Topology-based Filtering of Graph Signals via Persistent Homology · wovepaper