Anisotropic data with scale

Display a 3D image with anisotropic voxel spacing using the scale parameter so that the volume appears with correct proportions.

Microscopy data is often anisotropic: the voxel spacing differs across axes. Without scale, napari treats every voxel as a unit cube and the volume appears with incorrect proportions. Setting scale to the real voxel dimensions corrects this.

Toggle the visibility of each layer to compare the difference.

Tags: visualization-nD, layers

from skimage import data

import napari

# cells3d has voxel spacing approximately (0.29, 0.26, 0.26) in (z, y, x).
# We subsample z by 4 and x by 2 to simulate more strongly anisotropic data.
# After subsampling, the effective voxel spacing becomes (1.16, 0.26, 0.52).
# The ratio is approximately (4.5, 1, 2), which we pass to ``scale``.
cells = data.cells3d()
nuclei = cells[::4, 1, :, ::2]

scale = (4.5, 1, 2)

viewer = napari.Viewer(ndisplay=3)

viewer.add_image(
    nuclei,
    name='no scale',
    blending='additive',
    colormap='magenta',
)

viewer.add_image(
    nuclei,
    name='with scale',
    blending='additive',
    colormap='green',
    scale=scale,
)

viewer.layers['no scale'].bounding_box.line_color = 'magenta'
viewer.layers['no scale'].bounding_box.visible = True
viewer.layers['with scale'].bounding_box.line_color = 'green'
viewer.layers['with scale'].bounding_box.visible = True

viewer.camera.angles = (-45, 0, -60)
viewer.fit_to_view()

if __name__ == '__main__':
    napari.run()