orientationpy 0.3.1

Package for greyscale orientation analysis on 2D and 3D images

OrientationPy is the pythonic successor to the well-loved OrientationJ Fiji Plugin (github) from the BIG (Biomedical Imaging Group) in EPFL. It is a library that takes in 2D images or 3D volumes and computes the orientation of the greylevels. A key step is the computation of image gradients, for which we offer a number of different techniques.

Please visit the OrientationPy online docs.

Highlights

Please see the Example Gallery for some small examples of the use of the code, allowing the computation of orientations maps like this:

Other noteworthy python projects

Other orientation or structure tensor software that we’re aware of in python:

• skimage.feature.structure_tensor: Sobel filter for gradient

• structure-tensor: Only using finite differences for gradient

• pynamix.measure.orientation_map: FFT-based approach

• PyTextureAnalysis: 2D orientation package with overlay rendering

Acknowledgements

• The initial prototype was coded by Alexandru Vasile during a 6-week visit to the EPFL Center for Imaging, supported by Mariana Verezhak from PSI

• We used some of this data for testing: https://zenodo.org/record/5483719#.Yyra4vFBy2A

• We initialised this repository from a cookie-cutter: https://github.com/Nekroze/cookiecutter-pypackage

History

• 0.3.0.0 (2023-08-07) Closed form solution for eigenvalue decomposition

• 0.2.0.4 (2023-01-20) Anisotropy of voxel sizes implemented

• 0.2.0.3 (2023-01-13) Moved code to /src/orientationpy so it gets packaged?

• 0.2.0.2 (2023-01-13) Moved code to /src so it gets packaged?

• 0.2.0.1 (2023-01-13) Moved to pyproject.toml

• 0.2.0.0 (2023-01-13) First version on pip (non-working due to requirements.txt)