Hi, thank you so much for sharing your unwrapping algorithm!!

I would like to adapt your pipeline to unwrapping segmented malpighian tubule cells in Drosophila. Step0 of u-Unwrap3D only considers isotropic data, how do you recommend dealing with anisotropic data.

For example, from Step0: Mean curvature measurement from binary segmentation it does not allow updating for anisotropic data only one input voxel size since its isotropic?.

We map the curvature values to a colorscheme for display. To set the scale, we use the voxel size to convert to metric units of $\mu m^{-1}$
"""

we generate colors from the mean curvature

voxel_size = 0.104 #um
surf_H_colors = vol_colors.get_colors(surf_H/voxel_size, # 0.104 is the voxel resolution -> this converts to um^-1
colormap=cm.Spectral_r,
vmin=-1.,
vmax=1.) # colormap H with lower and upper limit of -1, 1 um^-1.

set the vertex colors to the computed mean curvature color

img_binary_surf_mesh.visual.vertex_colors = np.uint8(255*surf_H_colors[...,:3])

save the mesh for viewing in an external program such as meshlab which offers much better rendering capabilities

tmp = img_binary_surf_mesh.export(os.path.join(savefolder, 'curvature_binary_mesh_'+basefname+'.obj')) # tmp is used to prevent printing to screen.