Changes the dtype of voxel_res from uint32 to double

seung-lab / zmesh

Marching Cubes & Mesh Simplification on multi-label 3D images.

GNU General Public License v3.0

58 stars 8 forks source link

Changes the dtype of voxel_res from uint32 to double #19

Closed normanrz closed 3 years ago

normanrz commented 3 years ago

I was trying to mesh a dataset that has a voxel size of 11.23 x 11.23 x 28. I was seeing gaps between chunks of the mesh and an offset to the underlying data. After a bit of debugging, I noticed that the decimal places are cut off when casting the voxel resolution to uint32. However, the meshing code itself doesn't seem to require unsigned ints for the simplification where the voxel size is used. With this PR, I propose to change the dtype from uint32 to double.

normanrz commented 3 years ago

Sorry, I just noticed the other PR. Let me know if you want to close this one instead of the other.

william-silversmith commented 3 years ago

Sorry this took so long. I'm going to release a version of this library with the non-integer resolutions shortly.