Image-based Meshing

This example generates different types of image-based meshes using the Stanford bunny as an example.

import mymesh
from mymesh import image, visualize
import matplotlib.pyplot as plt

threshold = 100
bunny_img = mymesh.demo_image('bunny')
voxelsize = (0.337891, 0.337891, 0.5) # mm

Voxel meshes represent every voxel (three dimensional pixel) with a box, and the voxel intensity value is stored as element data (ElemData['Image Data']) in the mesh. The full image can be converted to a voxel mesh, with an element for every voxel, or a threshold can be applied to extract an object from the image. The image can also be coarsened (or refined) with the scalefactor input to create a coarser mesh. For images other than the demo image, files (or directories of files, such as DICOMs) can be read with :func:mymesh.image.read, or 3D numpy arrays can be used directly (see also: Image-based Meshing)

bunny_voxel = image.VoxelMesh(bunny_img, voxelsize, scalefactor=0.5)
half_voxel = bunny_voxel.Clip(normal=[0,1,0])

bunny_voxel2 = image.VoxelMesh(bunny_img, voxelsize, threshold, scalefactor=0.2)

Calculating element centroids...Done

Surface meshes or tetrahedral volume meshes can also be extracted from the image by applying a threshold and interpolating to create a more ‘smooth’ surface (see also Contour).

bunny_surf = image.SurfaceMesh(bunny_img, voxelsize, threshold, scalefactor=0.5)
bunny_tet = image.TetMesh(bunny_img, voxelsize, threshold, scalefactor=0.5, interpolation='linear')
half_tet = bunny_tet.Clip(pt=[0,88,0], normal=[0,1,0])

Calculating element centroids...Done

fig, axes = plt.subplots(2, 2)
titles = ['Full Voxel', 'Thresholded Voxel', 'Surface Mesh', 'Tetrahedral Mesh']
meshes = [half_voxel, bunny_voxel2, bunny_surf, half_tet]
for m, ax, title in zip(meshes, axes.ravel(), titles):
    m.verbose=False
    if title == 'Full Voxel':
        subfig, subax = m.plot(scalars='Image Data', color='Greys_r',
                               show=False,return_fig=True, view='-x-z')
    elif title == 'Thresholded Voxel':
        subfig, subax = m.plot(show_edges=True,show=False,return_fig=True, view='-x-z')
    else:
        subfig, subax = m.plot(show=False,return_fig=True, view='-x-z')
    ax.imshow(subax.get_images()[0].get_array())
    ax.set_title(title)
    ax.set_axis_off()
    plt.close(subfig)

RFBOutputContext()
RFBOutputContext()
RFBOutputContext()
RFBOutputContext()