Problem: Error in the tutorial of "3D reconstruction of mesh model" using matching cube

[swjn2017USTC]

Hello, Thank you so much for developing such a useful tool. I have a question. I try to run your "3D Reconstruction" tutorial link using the Drosophila Embryo data ("E7-9h_cellbin_tdr_v2.h5ad"). It raises no errors until I try to make a 3D reconstruction of mesh model using the code below:
embryo_mesh, _, _ = st.tdr.construct_surface(pc=embryo_pc, key_added="tissue", alpha=0.6, cs_method="marching_cube")
Here is the error log:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[10], line 1
----> 1 embryo_mesh, _, _ = st.tdr.construct_surface(pc=embryo_pc, key_added="tissue", alpha=0.6, cs_method="marching_cube")
      2 # st.pl.three_d_plot(model=st.tdr.collect_models([embryo_mesh, embryo_pc]), key="tissue", model_style=["surface", "points"], jupyter="static", cpo=cpo)
      3 # st.tdr.save_model(model=embryo_mesh, filename="embryo_mesh_model.vtk")

File ~/anaconda3/envs/Spateo/lib/python3.9/site-packages/spateo/tdr/models/models_individual/mesh.py:235, in construct_surface(pc, key_added, label, color, alpha, uniform_pc, uniform_pc_alpha, cs_method, cs_args, nsub, nclus, smooth, scale_distance, scale_factor)
    232 uniform_surfs = []
    233 for sub_surf in csurf.split_bodies():
    234     # Repair the surface mesh where it was extracted and subtle holes along complex parts of the mesh
--> 235     sub_fix_surf = fix_mesh(mesh=sub_surf.extract_surface())
    237     # Get a uniformly meshed surface using voronoi clustering.
    238     sub_uniform_surf = uniform_mesh(mesh=sub_fix_surf, nsub=nsub, nclus=nclus)

File ~/anaconda3/envs/Spateo/lib/python3.9/site-packages/spateo/tdr/models/models_individual/mesh_utils.py:137, in fix_mesh(mesh)
    132 except ImportError:
    133     raise ImportError(
    134         "You need to install the package `pymeshfix`. \nInstall pymeshfix via `pip install pymeshfix`"
    135     )
--> 137 meshfix = mf.MeshFix(mesh)
    138 meshfix.repair(verbose=False)
    139 fixed_mesh = meshfix.mesh.triangulate().clean()

File ~/anaconda3/envs/Spateo/lib/python3.9/site-packages/pymeshfix/meshfix.py:81, in MeshFix.__init__(self, *args)
     78             tri_mesh = mesh.triangulate()
     79             faces = tri_mesh.faces
---> 81         self.f = np.ascontiguousarray(faces.reshape(-1, 4)[:, 1:])
     83 else:
     84     raise TypeError("Invalid input. Please load a surface mesh or face and vertex arrays")

ValueError: cannot reshape array of size 15942 into shape (4)

Thank you！
Best,

[YifanLu2000]

Hi @swjn2017USTC, I’m not entirely sure about the reason, but you might try adjusting the parameters of the marching cube, for example mc_scale_factor and levelset

pc_mesh, _, _ = st.tdr.construct_surface(
    pc= pc, 
    key_added="tissue", 
    label='tissue_mesh', 
    color=palette[celltype],
    alpha=1, 
    cs_method="marching_cube", 
    cs_args={"mc_scale_factor": 1.2, 'dist_sample_num': 1000, 'levelset': 0.1}, 
    smooth=2000, 
    scale_factor=1,
)

These parameters mainly control the resolution and threshold of the generated surface, which can help produce a smoother or more detailed mesh. Note that typically mc_scale_factor > , levelset > 0 and levelset < 1.