On the effectiveness of persistent homology

Renata Turkeš, Guido Montúfar, Nina Otter

Abstract

Persistent homology (PH) is one of the most popular methods in Topological Data Analysis. Even though PH has been used in many different types of applications, the reasons behind its success remain elusive; in particular, it is not known for which classes of problems it is most effective, or to what extent it can detect geometric or topological features. The goal of this work is to identify some types of problems where PH performs well or even better than other methods in data analysis. We consider three fundamental shape analysis tasks: the detection of the number of holes, curvature and convexity from 2D and 3D point clouds sampled from shapes. Experiments demonstrate that PH is successful in these tasks, outperforming several baselines, including PointNet, an architecture inspired precisely by the properties of point clouds. In addition, we observe that PH remains effective for limited computational resources and limited training data, as well as out-of-distribution test data, including various data transformations and noise. For convexity detection, we provide a theoretical guarantee that PH is effective for this task in $\mathbb{R}^d$, and demonstrate the detection of a convexity measure on the FLAVIA data set of plant leaf images.

Introduction

While Persistent Homology (PH) is a cornerstone of Topological Data Analysis (TDA), the reasons behind its broad success have remained somewhat mysterious. In the world of data science, PH is often treated as a "black box" that provides useful features for machine learning. We know it works for everything from analyzing brain artery trees to classifying leaf morphology, but we lack a clear understanding of what it is actually seeing. Is it just capturing global topology, or is it sensing fine-grained geometry?

This paper aims to bridge this gap by investigating PH’s effectiveness in three fundamental shape analysis tasks: detecting the number of holes (topology), estimating the curvature of a surface (local geometry), and distinguishing between convex and concave shapes (global geometry). By comparing PH against standard machine learning baselines and specialized architectures like PointNet, the study seeks to define the "boundaries of effectiveness" for topological methods.

Methodology

The researchers utilize the standard PH pipeline—turning point clouds into simplicial complexes and tracking how topological features persist across different scales. Key methodological contributions include:

  • Tubular Filtrations: To tackle convexity, the authors introduce a novel "tubular filtration" based on the distance of points from a given line. If a shape is concave, this filtration identifies multiple disconnected components that eventually merge.
  • Theoretical Guarantees: The paper provides a theoretical proof (Theorem 1) demonstrating that a shape is convex if and only if every possible tubular filtration results in exactly one persistent component.
  • Signature Comparison: Various vectorization methods for topological data, including Persistence Images (PI) and Persistence Landscapes (PL), were tested using an SVM classifier.

Experimental Setup and Datasets

The study utilizes a mix of synthetic benchmarks and real-world data to evaluate performance:

  • Synthetic Holes: 2D and 3D point clouds sampled from shapes with 0, 1, 2, 4, or 9 holes.
  • Constant Curvature Disks: Point clouds sampled from surfaces with constant negative (hyperbolic), zero (Euclidean), and positive (spherical) curvature.
  • Convexity Benchmark: A balanced set of regular and random shapes generated via convex hulls and alpha shapes.
  • FLAVIA Dataset: A real-world dataset of 1,907 plant leaf images used to test the PH pipeline’s ability to predict convexity in biological data.
Example shapes used for the hole detection task, categorized by the number of holes.

Results and Discussion

Experiments revealed that PH is remarkably robust compared to deep learning baselines. While models like PointNet require large amounts of training data to learn shape features, PH achieves high accuracy with very few samples because the "knowledge" of shape is inherent to the mathematical framework.

Accuracy comparison of PH variants against ML and NN baselines across different data transformations.

Key findings include:

  • Hole Detection: PH achieved nearly 100% accuracy across all transformations (rotation, shear, stretch). In contrast, PointNet’s accuracy plummeted to ~20% under simple translations or rotations.
  • Curvature Estimation: PH (specifically using 0-dimensional lifespans) achieved a Mean Squared Error (MSE) of 0.06, vastly outperforming PointNet (MSE of 578.28). This suggests that "short bars" in persistence diagrams, often dismissed as noise, contain essential geometric information.
  • Out-of-Distribution Robustness: PH proved much more resilient than neural networks when faced with out-of-distribution data, such as testing on complex random shapes after training on simple regular polygons.

Conclusion

The paper concludes that PH should be viewed not just as a niche tool for topology, but as a robust, general-purpose framework for geometric feature extraction. Its inherent invariance to rotation and translation, combined with its data efficiency, makes it a powerful alternative or complement to deep learning in shape-related machine learning tasks.

Summary updated

Abstract

Persistent homology (PH) is one of the most popular methods in Topological Data Analysis. Even though PH has been used in many different types of applications, the reasons behind its success remain elusive; in particular, it is not known for which classes of problems it is most effective, or to what extent it can detect geometric or topological features. The goal of this work is to identify some types of problems where PH performs well or even better than other methods in data analysis. We consider three fundamental shape analysis tasks: the detection of the number of holes, curvature and convexity from 2D and 3D point clouds sampled from shapes. Experiments demonstrate that PH is successful in these tasks, outperforming several baselines, including PointNet, an architecture inspired precisely by the properties of point clouds. In addition, we observe that PH remains effective for limited computational resources and limited training data, as well as out-of-distribution test data, including various data transformations and noise. For convexity detection, we provide a theoretical guarantee that PH is effective for this task in $\mathbb{R}^d$, and demonstrate the detection of a convexity measure on the FLAVIA data set of plant leaf images. Due to the crucial role of shape classification in understanding mathematical and physical structures and objects, and in many applications, the findings of this work will provide some knowledge about the types of problems that are appropriate for PH, so that it can - to borrow the words from Wigner 1960 - ``remain valid in future research, and extend, to our pleasure", but to our lesser bafflement, to a variety of applications.
BibTeX
@inproceedings{turkes2022effectiveness,
  author = {Renata Turkeš and Guido Montúfar and Nina Otter},
  title = {{On the effectiveness of persistent homology}},
  booktitle = {Advances in Neural Information Processing Systems},
  year = {2022},
}
Figure