zmesh 1.13.0

Multilabel marching cubes and simplification of volumetric data.

zmesh: Multi-Label Marching Cubes & Mesh Simplification

from zmesh import Mesher

labels = ... # some dense volumetric labeled image
mesher = Mesher( (4,4,40) ) # anisotropy of image

# initial marching cubes pass
# close controls whether meshes touching
# the image boundary are left open or closed
mesher.mesh(labels, close=False)

# If you don't mind shuffling the vertices and faces
# from older versions of zmesh, this option unlocks
# some performance optimizations.
mesher.mesh(labels, preserve_order=False) 

meshes = []
for obj_id in mesher.ids():
  meshes.append(
    mesher.get(
      obj_id, 
      normals=False, # whether to calculate normals or not

      # tries to reduce triangles by this factor
      # 0 disables simplification
      reduction_factor=100, 

      # Max tolerable error in physical distance
      # note: if max_error is not set, the max error
      # will be set equivalent to one voxel along the 
      # smallest dimension.
      max_error=8,
      # whether meshes should be centered in the voxel
      # on (0,0,0) [False] or (0.5,0.5,0.5) [True]
      voxel_centered=False, 
    )
  )
  mesher.erase(obj_id) # delete high res mesh

mesher.clear() # clear memory retained by mesher

mesh = meshes[0]
mesh = mesher.simplify(
  mesh, 
  # same as reduction_factor in get
  reduction_factor=100, 
  # same as max_error in get
  max_error=40, 
  compute_normals=False, # whether to also compute face normals
) # apply simplifier to a pre-existing mesh

# use an fqmr derived non-topology preserving algorithm
# this useful particularly for multi-resolution meshes
# where visual appearence is more important than connectivity
# This function has many parameters, see help(zmesh.simplify_fqmr)
mesh = zmesh.simplify_fqmr(
  mesh, 
  triangle_count=(mesh.faces.shape[0] // 10),
)

# compute normals on a pre-existing mesh
mesh = zmesh.compute_normals(mesh) 

# run face based connected components
ccls = zmesh.face_connected_components(mesh)
# run vertex based connected components
ccls = zmesh.vertex_connected_components(mesh)

# remove small components based on vertices or faces
mesh = zmesh.dust(mesh, threshold=100, metric="vertices")
# remove components bigger than the threshold using invert
mesh = zmesh.dust(mesh, threshold=100, metric="vertices", invert=True)
# retain only the largest k connected components
mesh = zmesh.largest_k(mesh, k=1, metric="vertices")
# retain only the smallest k connected components
mesh = zmesh.largest_k(mesh, k=1, metric="vertices", invert=True)

mesh.vertices
mesh.faces 
mesh.normals
mesh.triangles() # compute triangles from vertices and faces

# Extremely common obj format
with open('iconic_doge.obj', 'wb') as f:
  f.write(mesh.to_obj())

# Common binary format
with open('iconic_doge.ply', 'wb') as f:
  f.write(mesh.to_ply())

# Neuroglancer Precomputed format
with open('10001001:0', 'wb') as f:
  f.write(mesh.to_precomputed())

Note: mesher.get_mesh has been deprecated in favor of mesher.get which fixed a long standing bug where you needed to transpose your data in order to get a mesh in the correct orientation.

Installation

If binaries are not available for your system, ensure you have a C++ compiler installed.

pip install zmesh

Performance Tuning & Notes

• The mesher will consume about double memory in 64 bit mode if the size of the object exceeds <1023, 1023, 511> on the x, y, or z axes. This is due to a limitation of the 32-bit format.
• The mesher is ambidextrous, it can handle C or Fortran order arrays.
• The maximum vertex range supported .simplify after converting to voxel space is 2²⁰ (appx. 1M) due to the packed 64-bit vertex format.

Related Projects

• zi_lib - zmesh makes heavy use of Aleks' C++ library.
• Igneous - Visualization of connectomics data using cloud computing.

Example Performance

This was a limited experiment conducted on a Macbook Pro M3 comparing Zmesh 1.12.0 and scikit-image==0.26.0. The volume is a 512^3 uint32 segmentation of a mouse visual cortex containing 2523 shapes of various sizes including parts of dendrites, a nucleus, and a glia.

Note that this is not really an apples-to-apples comparison because scikit-image is specialized for continuous value images like CT scans not segmentation, and so the resulting meshes are very different.

This is mesher.mesh(image).

Marching Cubes Data	ZMESH Time (s)	ZMESH MVx/sec	SKIMAGE Time (s)	SKIMAGE MVx/sec	N
Black	0.891	451.35	NOT HANDLED	—	3
Filled	0.961	418.12	NOT HANDLED	—	3
connectomics.npy	4.107	97.89	9.861	40.77	3
random	6.950	12.81	40.509	2.20	1

The meshes can then be extracted (mesher.get):

Simplification Factor	Max Error	Labels per Second	N
0	N/A	478.2	3
100	40	14.7	3

Credits

Thanks to Aleks Zlateski for creating and sharing this beautiful mesher.

Later changes by Will Silversmith, Nico Kemnitz, and Jingpeng Wu.

Thank you to Sven Forstmann, Kristof S., Brénainn Woodsend, and others for pyfqmr which we have adapted here for non-topology preserving simplification and fast OBJ reading. See https://github.com/Kramer84/pyfqmr-Fast-Quadric-Mesh-Reduction/

References

1. W. Lorensen and H. Cline. "Marching Cubes: A High Resolution 3D Surface Construction Algorithm". pp 163-169. Computer Graphics, Volume 21, Number 4, July 1987. (link)
2. M. Garland and P. Heckbert. "Surface simplification using quadric error metrics". SIGGRAPH '97: Proceedings of the 24th annual conference on Computer graphics and interactive techniques. Pages 209–216. August 1997. doi: 10.1145/258734.258849 (link)
3. H. Hoppe. "New Quadric Metric for Simplifying Meshes with Appearance Attributes". IEEE Visualization 1999 Conference. pp. 59-66. doi: 10.1109/VISUAL.1999.809869 (link)