/nlreg1d

Code and data associated with nonlinear registration paper

Primary LanguageHTMLMIT LicenseMIT

nlreg1d

Nonlinear registration for univariate one-dimensional data.


This repository contains code and data associated with the paper:

Pataky TC, Robinson MA, Vanrenterghem J, Donnelly CJ (2022) Simultaneously assessing amplitude and temporal effects in biomechanical trajectories using nonlinear registration and statistical nonparametric mapping. Journal of Biomechanics 136: 111049.


⚠️ This repository contains primarily wrapper functions to key functionality in fdasrsf and scikit-fda and spm1d; little-to-no new functionality is introduced. The primary purpose of this repository is not to introduce new functionality, but instead to assemble existing, relatively complex functionality into an easy-to-use interface. Users are encouraged to use the original packages: fdasrsf, scikit-fda and spm1d.


Dependencies

Dependencies include:

  • python 3.8
  • numpy 1.22
  • scipy 1.8
  • matplotlib 3.5
  • spm1d = 0.4
  • fdasrsf 2.3
  • scikit-fda 0.7

Additional version details for all dependencies are provided in this repository's env.yml file.


Installation

To install all dependencies for this repository, we strongly suggest creating an Anaconda environment using the env.yml file found in this repository, according to the following instructions.


First download the code then navigate to the repository. For example:

cd /Users/USERNAME/Downloads/nlreg1d/

Next create the environment:

conda env create -f env.yml

Activate the environment:

conda activate nlreg1d

Last, add the parent directory for nlreg1d to the Python path. One way to do this is inside a Python script:

>>> import sys
>>> dir0 = '/Users/USERNAME/Downloads/nlreg1d-main'  # adjust this to the location on your computer
>>> if dir0 not in sys.path:
>>>     sys.path.insert(0, dir0)
>>>
>>> import nlreg1d


You should now be able to run all scripts / notebooks in IPython / Jupyter.

<br>

To launch IPython:

	ipython --pylab

To launch Jupyter:

	jupyter notebook


<br>

When finished, don't forget to deactivate the environment:

	conda deactivate

<br>

## Get started

Refer to documentation in the `Notebooks` folder of this repository for usage instructions and examples.