medialaxis3d

This package extends the scikit-image function medial_axis to the 3D case.

Install

pip install medialaxis3d

Dependencies

Automatically installed with pip:

• numpy
• scipy
• cython

Optional only for visualization

• napari

Documentation

WIP

Quickstart

Use it without returning the medial distance.

>>> import numpy as np
>>> import skimage as ski
>>> import medialaxis3d
>>> import napari

>>> rng = np.random.default_rng(1278)

>>> image = ski.data.binary_blobs(length = 128,
>>>                           blob_size_fraction = 0.2,
>>>                           n_dim = 3,
>>>                           volume_fraction = 0.6,
>>>                           rng = rng)

>>> skeleton = medialaxis3d.medial_axis_3d(image, 
>>>                                        return_distance = False, 
>>>                                        size = 8, 
>>>                                        rng = rng)

>>> viewer = napari.Viewer()
>>> viewer.add_image(image, 
>>>                  rendering = "attenuated_mip", 
>>>                  attenuation = 0.5, 
>>>                  scale = [1, 1, 1])
>>> viewer.add_image(skeleton, 
>>>                  interpolation3d = "nearest", 
>>>                  colormap = "magenta", 
>>>                  scale = [1, 1, 1])
>>> napari.run()

or use it to return the distance as well.

>>> import numpy as np
>>> import skimage as ski
>>> import medialaxis3d
>>> import napari

>>> rng = np.random.default_rng(1278)
>>> image = ski.data.binary_blobs(length = 128,
                              blob_size_fraction = 0.2,
                              n_dim = 3,
                              volume_fraction = 0.6,
                              rng = rng)

>>> skeleton, distance = medialaxis3d.medial_axis_3d(image, 
>>>                                                  return_distance = True, 
>>>                                                  size = 8, 
>>>                                                  rng = rng)

>>> viewer = napari.Viewer()
>>> viewer.add_image(image, 
>>>                  rendering = "attenuated_mip", 
>>>                  attenuation = 0.5, 
>>>                  scale = [1, 1, 1])
>>> viewer.add_image(skeleton*distance, 
>>>                  interpolation3d = "nearest", 
>>>                  colormap = "turbo", 
>>>                  scale = [1, 1, 1])
>>> napari.run()