Comparison among digital fundamental groups and its applications (Q2482186)

A digital image is a discrete object consisting of a strictly positive integer \(n\), a subset \(X\) of \(\mathbb{Z}^n\) (the set of \(n\)-tuples of integers) and an adjacency relation \(k\) on \(X\). Various notions coming from algebraic topology, as homotopy, relative homotopy, fundamental group and covering space, were adapted to this digital setting. These ideas are essential to the pre-preprocessing operations in pattern recognition, which reduce a digital image into a \(k\)-homotopically equivalent frame. There exist three kinds of digital \(k\)-fundamental groups of a digital image \((X,k)\), the first one due to Kong and Malgouyres, the second one due to Boxer and Han, and the third one due to Ayala, Domínguez, Francés and Quintero. Since each of these digital \(k\)-fundamental groups has an intrinsic feature of its own, its usages depend on the situation. This paper, after an introduction to make it self-contained, compares the first two notions of fundamental groups and is intended to show the strong merit of the second one with respect to the first one in relation to the classification of digital images.