PyCoShREM

Python port of the Matlab Toolbox Complex Shearlet-Based Ridge and Edge Measurement by Rafael Reisenhofer.

Easy Installation

You should be able to directly install PyCoShREM via pip with

$ sudo pip install pycoshrem

Manual Installation

If you are not able to install via pip (see: setuptools), pull this repository and make sure the neccessary dependecies are installed.

Dependencies

The implementation relies heavily on numpy and scipy, make sure these are installed.

The thinning-algorithm used is from scikit-kimage.

To install the packages system-wide, run

$ sudo pip install numpy scipy scikit-image pillow

To install for a single user only, run

$ pip install --user numpy scipy scikit-image pillow

Since these packages use C-Extensions, on some Distributions (e.g. Debian) you will also need the python-dev packages, notably Python.h-Headers installed.

Example

Several example scripts are included examples folder within the repository, which performs edge and ridge detection on the different kind of image data.

Documentation

Documentation is almost non-existing at the moment, but more will be added in the future.

All stand-alone documentation can be found in the docs folder. Sphinx with the napoleon extension is used to create the documentation from Google-Style docstrings.

Build Documentation

Make sure Sphinx is installed

$ sudo pip install Sphinx

To build the documentation using sphinx change to the docs folder and use the provided Makefile, e.g.:

$ cd docs
$ make html

This will create a html-Documentation in docs/_build/html/.

Tests

Test coverage is very low at the moment. Tests are provided via the Python unittest module and can be run from the projects root directory via

$ python -m unittest

Acknowledgements

Most code in this package is directly ported from the Complex Shearlet-Based Ridge and Edge Measure Toolbox (CoShREM). However some parts of CoShREM were taken from other projects:

The method

• coshrem.shearlet.yapuls

is originally taken from Yet Another Wavelet Toolbox (YAWTb).

The methods

• coshrem.shearlet.shear and
• coshrem.shearlet.padarray

are originally taken from ShearLab 3D.

The flame front data images

• K05_CH_full.txt
• K05_OH_full.txt

and the mock image

• B0.png

used in tests and examples were kindly provided by Johannes Kiefer.

License