Identifying cobordisms using kernel persistence
Yossi Bokor Bleile, Lisbeth Fajstrup, Teresa Heiss, Anne Marie Svane, Søren Strandskov Sørensen
TL;DR
The paper addresses identifying tunnels, modeled as open cobordisms, between disjoint subcomplexes in a filtered cell complex, with motivation from material science to detect atomic tunnels in point clouds. It introduces a homological construction via a map $Φ$ from $Ker(iota^A_*) ⊕ Ker(iota^B_*)$ to $Ker(iota^{A∪B}_*)$ and treats $Cok(Φ)$ as a persistence module to capture birth and death events of cobordisms along the filtration. Two main algorithmic contributions are presented: (i) kernel persistence to compute birth times in the relevant kernel modules, and (ii) a pairing procedure that combines these to identify birth–death pairs in $Cok(Φ)$ and to produce explicit cobordism representatives. The framework enables systematic detection and characterization of tunnel-like voids in atomic structures from point clouds, with extensions proposed to higher-order cobordisms, a Voronoi-dual formulation, and handling of practical issues such as slice thickness and periodic boundaries.
Abstract
Motivated by applications in chemistry, we give a homlogical definition of tunnels, or more generally cobordisms, connecting disjoint parts of a cell complex. For a filtered complex, this defines a persistence module. We give a method for identifying birth and death times using kernel persistence and a matrix reduction algorithm for pairing birth and death times.
