ktch: model-based morphometrics in Python

ktch is written for researchers and engineers who need to quantify morphological properties, whatever the field.

It is a Python package for model-based morphometrics: the quantitative study of morphological properties and diversity through explicit models. It covers theoretical morphology models, harmonic descriptors, and landmark methods. Although each approach assumes a different model and applies to objects of different dimensions, topologies, and structures, they all share one underlying idea. ktch implements that idea as a scikit-learn compatible API. Thus the methods compose into pipelines, and a study can quantify different aspects of morphological properties within one workflow.

ktch provides the functionality specific to morphometrics and leaves the rest to the Python data analysis ecosystem. An analysis can therefore keep using the tools it already relies on for preprocessing, model selection, and visualization.

The elliptic Fourier analysis tutorial shows what this looks like in practice, including the scikit-learn estimator API (e.g., fit_transform).

Installation

Get started by installing ktch. ktch is available from PyPI and conda-forge.

Tutorials

Step-by-step guides for learning ktch through hands-on examples, from landmark methods to harmonic analysis and coiling models.

How-to guides

Task-oriented guides for common operations, such as reading morphometric file formats and visualizing results.

Explanation

Conceptual explanations of morphometric methods and theory, including what each method assumes about the target morphological properties.

API reference

Complete API documentation for all classes, functions, and modules.

Getting help

Bug reports, feature requests, and questions are welcome via GitHub Issues.