Abstract
In this work, we present VHS (Voxelized Homological Simplification), a C++ package whose purpose is to de-noise voxelized data and output a topologically accurate simplified shape. In contrast to previous work on voxelized homological simplification tools, our main goal is offering a better starting point for computing curve skeletons for shape analysis. This goal necessitates additional simplification beyond what other packages provide, although our approach extends and improves prior work on heuristic methods which compute approximate solutions for the homological simplification problem. Our tool is designed for and tested on voxelized plant roots, although it is potentially useful beyond this data set. While the homological simplification problem is NP-hard in general, our package is able to simplify almost all of the topological noise when used on data from plant root systems. Compared with existing simplification tools, our method strikes a better balance between topological simplicity and geometric accuracy, resulting in higher usability of the resulting skeletons. Our code is publicly available at https://github.com/davidletscher/VHS/.
| Original language | English |
|---|---|
| Article number | 102198 |
| Journal | Computational Geometry: Theory and Applications |
| Volume | 130 |
| DOIs | |
| State | Published - Jan 2026 |
Keywords
- Computational topology
- Persistent homology
- Shape simplification
- Skeletization
Fingerprint
Dive into the research topics of 'VHS: A package for homological simplification of voxelized plant root data for skeletonization'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver