Overview

Calculation of Persistent Homology for Pixel Data

PixHomology is an open-source software for image processing and analysis focused on persistent homology computation. It provides a set of tools and algorithms to explore the topological features of 2D images, enabling users to extract meaningful information about the underlying structures.

Features

• In its initial release, the software enables the computation of 0-dimensional homology on a 2D image.

• The construction of simplicial complexes involves establishing connections between each pixel and its 8 neighboring pixels.

Installation

PixHomology is available on PyPI and can be installed using pip. To install the package, run the following command:

Optional (create new environment):

conda create -n pixh python=3.11
conda activate pixh

Install PixHomology:

pip install pixhomology

Building from Source:

To build the source code, ensure you have CMake installed on your system. You can download CMake from cmake.org or install it using your package manager. To install PixHomology, follow these steps:

git clone https://github.com/riccardoc95/PixHomology.git
cd PixHomology
git submodule init
git submodule update
pip install .

For detailed installation instructions, please refer to the Installation Guide.

Usage

Here is a basic example of using PixHomology in Python:

import pixhomology as px
image = np.random.rand(10,10)
dgm = px.computePH(image)
...

For more examples and detailed usage instructions, check out the Tutorials and examples.

To test the performance of PixHomology follow the instructions from this repository.