The Delaunay-Rips filtration is a lighter and faster alternative to the well-known Rips filtration for low-dimensional Euclidean point clouds. Despite these advantages, it has seldom been studied. In this paper, we aim to bridge this gap by providing a thorough theoretical and empirical analysis of this construction. From a theoretical perspective, we show how the persistence diagrams associated with the Delaunay-Rips filtration approximate those obtained with the Rips filtration. Additionally, we describe the instabilities of the Delaunay-Rips persistence diagrams when the input point cloud is perturbed. Finally, we introduce an algorithm that computes persistence diagrams of Delaunay-Rips filtrations in any dimension. We show that our method is faster and has a lower memory footprint than traditional approaches in low dimensions. Our C++ implementation, which comes with Python bindings, is available at https://github.com/MClemot/GeoPH.
翻译:Delaunay-Rips 过滤是用于低维欧几里得点云的著名 Rips 过滤的一种更轻量、更快速的替代方案。尽管具有这些优势,它却很少被研究。在本文中,我们旨在通过对此构造进行全面的理论和实证分析来弥补这一空白。从理论角度,我们展示了与 Delaunay-Rips 过滤相关联的持久性图如何逼近通过 Rips 过滤获得的持久性图。此外,我们描述了当输入点云受到扰动时 Delaunay-Rips 持久性图的不稳定性。最后,我们提出了一种可在任意维度计算 Delaunay-Rips 过滤持久性图的算法。我们证明,在低维情况下,我们的方法比传统方法速度更快且内存占用更低。我们提供了带有 Python 绑定的 C++ 实现,可在 https://github.com/MClemot/GeoPH 获取。