Thanks for the code. It is helpful for my work. Can it still work when template provided by you and my 3D object has very different size?

Hi, I am very interested in using meshmonk to crop, trim, and prune 3D facial images of my own participants. I noted that the DemoFaceMapping.m file could be directly modified to run on my own facial imaging data. However, I this is just a demo file. I think there should be some more formal way for me to use meshmonk rather than using DemoFaceMapping.m file for each of my sample image.

Following the README.md file of meshmonk, I run mex_windows_all.m script to mex meshmonk functions. However, this is the end of the file regarding how to use meshmonk from Matlab on Win. My question is what should I do after mexing? Will there be an .exe file created from the mexing? I just do not want to use DemoFaceMapping.m file for my mapping, and I wish to know the orthodox way of using meshmonk, after mexing.

Thank you for sharing the great work.

When I call compute_correspondences function using Matlab. I have got this error;

Unexpected Standard exception from MEX file.
What() is:Error: 'dimensionality' must match column count in data matrix

This occurs both demo and tutorial code, run on Mac OS.
Thank you in advance,

I run meshmonk on 64 bit Win 10 system.

When I run the demo file, I came across the Eigenvecdtor decomposre on Q failed issue, such issue has been reported by marvin521.

When I run the DemoFaceMapping.m file on two of my self-collected 3dMD facial surface mesh, I was able to view the mapping result as shown in the image below:

However, when I clicked the 'Save 3D' icon, I was returned error message:

Undefined function 'export' for input arguments of type 'shape3D'.

Error in viewer3D/constructToolbar>save_Callback (line 139)
                 export(obj.ShapeChildren{i});
 
Error while evaluating PushTool ClickedCallback.

Such issue happen when I tried the .m file on both of my sample.

Kind regards

Hello sir，
We have successfully complied meshmonk in windows with visual studio and put the compiled meshmonk.lib into the root folder of matlab ，but when we run the “UnderstandingNonRigidMappingWithMeshMonk.m” we came across the problem that the function of compute_correspondences could not be recognized.From my prospect，whether because of the misdiaplacement of the lib file(meshmonk.lib)induced such problem or any other posibilities which I don‘t know. Could you help me find out the reason or give
me more suggestions on solving such problem，thanks a lot！

Hi there,

I recently needed to reinstall Meshmonk again on my new M1 macbook.
I now get an issue in matlab when I want to run meshmonk, concerning the libmeshmonk.dylib architecture:

MException with properties:

identifier: 'MATLAB:mex:ErrInvalidMEXFile'
   message: 'Invalid MEX-file '/Users/pieter-janverhelst/Documents/MATLAB/ImgGENESoftware/myCompiledToolboxes/compute_correspondences.mexmaci64': dlopen(/Users/pieter-janverhelst/Documents/MATLAB/ImgGENESoftware/myCompiledToolboxes/compute_correspondences.mexmaci64, 0x0006): Library not loaded: libmeshmonk.dylib↵  Referenced from: /Users/pieter-janverhelst/Documents/MATLAB/ImgGENESoftware/myCompiledToolboxes/compute_correspondences.mexmaci64↵  Reason: tried: '/Applications/MATLAB_R2021b.app/bin/maci64/libmeshmonk.dylib' (no such file), '/Applications/MATLAB_R2021b.app/bin/maci64/./libmeshmonk.dylib' (no such file), '/Applications/MATLAB_R2021b.app/bin/maci64/../../sys/os/maci64/libmeshmonk.dylib' (no such file), '/Applications/MATLAB_R2021b.app/Contents/MacOS/libmeshmonk.dylib' (no such file), '/Applications/MATLAB_R2021b.app/Contents/MacOS/./libmeshmonk.dylib' (no such file), '/Applications/MATLAB_R2021b.app/Contents/MacOS/../../standalone/bin/maci64/libmeshmonk.dylib' (no such file), '/Applications/MATLAB_R2021b.app/Contents/MacOS/../../sys/os/maci64/libmeshmonk.dylib' (no such file), '/Applications/MATLAB_R2021b.app/Contents/MacOS/../../bin/maci64/libmeshmonk.dylib' (no such file), 'libmeshmonk.dylib' (no such file), '/usr/local/lib/libmeshmonk.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'x86_64')), '/usr/lib/libmeshmonk.dylib' (no such file), '/Users/pieter-janverhelst/Desktop/VECTRA/IMAGES/libmeshmonk.dylib' (no such file), '/usr/local/lib/libmeshmonk.dylib' (mach-o file, but is an incompatible architecture (have 'arm64', need 'x86_64')), '/usr/lib/libmeshmonk.dylib' (no such file)'
     cause: {}
     stack: [2×1 struct]
Correction: []

Seems to be a compatibility issue between the X86_64 design and the new arm64 architecture. Is this a known issue and are there any solutions? Thanks in advance!

Matlab 2019b

when running the BatchProcessingWithMeshmonk.m with own obj face I get this error.
K>> [usemtl,mtl,maps] = parseMTL(filename,usePrefix);
Unrecognized function or variable 'parseMTL'.

The obj file is called 007.obj and mtl file is called 007.mtl

Hello,

I was following the installation instructions for OSX and got to the last part of copying the relevant header files into /usr/local/include/src.

I observed while mexing that 2 of the files in /usr/local/include/src couldn't find this particular file - nanoflann.hpp. This is because nanoflann.hpp is in a subfolder of it's own called vendor. Copying it into src did the trick.

If you are able to reproduce the problem and confirm the issue, I'll be happy to send in a pull request. Thanks!

Hi, I have difficulty in compiling example file. when I try to compile the example using make example compilation fails.

meshmonk/example.cpp:62: undefined reference to 'read_obj_files'
meshmonk/example.cpp:213: undefined reference to 'pyramid_registration'

but when I check the meshmonk.cpp file and meshmonk.hpp file, these functions were already defined.

I'm using cpp_project branch

I use a long term version of Ubuntu and have started matlab with these options.
LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.30:/usr/local/lib/libmeshmonk.so:/usr/local/lib/libOpenMeshCore.so:/usr/local/lib/libOpenMeshTools.so matlab

I have gcc 11. How can I solve this problem?

Hello. When I try to use shape mappers on my on obj files I get the following error.
Could you please help me to recognzie the issue?

eigenvector decomposer on Q failed!
Q : -nan(ind) -nan(ind) -nan(ind) -nan(ind)
-nan(ind) -nan(ind) -nan(ind) -nan(ind)
-nan(ind) -nan(ind) -nan(ind) -nan(ind)
-nan(ind) -nan(ind) -nan(ind) -nan(ind)
...
Thank You!

I have compiled the mex files and libmeshmonk.so.
When I run a demo I get this error.
Invalid MEX-file '/home/knutjb/meshmonks/meshmonk/matlab/compute_correspondences.mexa64': /usr/local/lib64/libmeshmonk.so: undefined symbol:
_ZTIN8OpenMesh12BasePropertyE

Error in test_compute_correspondences (line 38)
compute_correspondences(floatingFeatures, targetFeatures,..

Hi, thanks for great work.

When I use meshmonk, I can get aligned pointcloud and mesh information in obj.FloatingShape, but I don't know how to find texture information.

Could you teach me how to find texture, or do I have to mannually match texture?

Thanks

To avoid version mismatches with Eigen, Nanaflann & OpenMesh we need to include the correct versions of these libs in a vendors directory.

This will allow to always build the software without any dependency.

Error log

src/NeighbourFinder.hpp:102:19: error: no matching constructor for
initialization of 'nanoflann::KDTreeEigenMatrixAdaptor<Matrix<float, -1,
6, 0, -1, 6> >'
...= new nanoflann::KDTreeEigenMatrixAdaptor(_inSourcePoints,
^ ~~~~~~~~~~~~~~~~~
src/CorrespondenceFilter.cpp:34:22: note: in instantiation of member function
'registration::NeighbourFinder<Eigen::Matrix<float, -1, 6, 0, -1, 6>
>::set_source_points' requested here
_neighbourFinder.set_source_points(_inTargetFeatures);
^
/usr/local/include/nanoflann.hpp:1318:3: note: candidate constructor not viable:
no known conversion from 'const Eigen::Matrix<float, -1, 6, 0, -1, 6>' to
'const int' for 1st argument
KDTreeEigenMatrixAdaptor(const int dimensionality, cons...
^
/usr/local/include/nanoflann.hpp:1329:3: note: candidate constructor not viable:
requires 1 argument, but 2 were provided
KDTreeEigenMatrixAdaptor(const self_t&);
^
After short investigating, here is the found line in NeighbourFinder.hpp
_kdTree = new nanoflann::KDTreeEigenMatrixAdaptor<VecMatType(_inSourcePoints, _leafSize);

I keep getting the following error when running the tutorial "UnderstandingNonRigidMappingWithMeshMonk.m" on windows. Compiled everything as instructed.

The first few operations using the Shape3D and Viewer class work. However, the error occurs when running the cell "Step 1 - compute correspondences".

"
Undefined function 'compute_correspondences' for input arguments of type 'single'.

Error in UnderstandingNonRigidMappingWithMeshMonk (line 139)
compute_correspondences(floatingFeatures, targetFeatures,...
"

A similar error occurs later on:
"
Undefined function 'compute_nonrigid_transformation' for input arguments of type 'uint32'.

Error in UnderstandingNonRigidMappingWithMeshMonk (line 219)
compute_nonrigid_transformation(floatingFeatures, correspondingFeatures,...
"

Hi, thanks for the great work.

When I ran the script 'DemoFaceMapping.m', matlab gives me the errors:

Are you ready? Y/N [Y]: Y
Started...
eigenvector decomposer on Q failed!
Q : nan nan nan nan
nan nan nan nan
nan nan nan nan
nan nan nan nan

xxx
shape3D/get.TriHandle (line 76)
out = triangulation(obj.Faces,obj.Vertices);% triangulation is a matlab construct that helps in a lot of computations

BTW, I am using OSX, and I have installed all the packages following the instructions except that my eigen version is 3.3.4

especially the paper for nonrigid registration method. think you!

Hi
I was successfully able to follow the tutorial and use the demo data to run "UnderstandingNonRigidMappingWithMeshMonk.m". Thanks for providing that.

I would like to know how to determine the accuracy of the registration process? Is there an evaluation metric, of how well the target is registered onto the template as well as how accurate are the face vertices?

I have the data with 3D facial scans and I ran through the above matlab script. Visually, the registration did not look as accurate as it was with your demo data. But I would like to compare to accuracy (quantify with an evaluation metric) of registration of my data compared to the demo data. Also, I do not have any manual landmark indications, just the 3D facial scan data - obj, mtl files.

Thanks

Hi,
I ran the demo successfully, but the colors randomly generated of floating and target are often close during visualization, so is there a way to specify the color for single object.
Thanks.

The relevant code is as follows:

%% Visualize
s = shape3D;
s.Vertices = correspondingFeatures(:,1:3);
st = shape3D;
st.Vertices = targetFeatures(:,1:3);
v = viewer(s);
viewer(st,v);

Hello,

I am trying to perform rigid and non-rigid transformations on the FLAME template using my own target 3D scanned faces. Both the target and the template have been landmarked with 32 points. The template and the target look as follows -

Template object

Target 3D object

As far as I know there are no NAN values in either the template or the object file. But when I run the BatchProcessingWithMeshMonk.m file, I get the following error -

eigenvector decomposer on Q failed!
Q : nan nan nan nan
nan nan nan nan
nan nan nan nan
nan nan nan nan

I isolated the line which is causing this error. It starts with the map() function in the above file and I traced it back to the rigid transformation switch case..

Further I observed that the program failed on update() in compute_rigid_registration function in meshmonk.cpp which in turn pointed me to the actually print statement of the error message I'm receiving in RigidTransformer.cpp - https://github.com/TheWebMonks/meshmonk/blob/master/src/RigidTransformer.cpp#L82

I'm kinda stumped as to how I can debug my particular scenario and check why the EigenVector decomposer is not working. Could you please give me some pointers?

Thank you,
Niraj Pandkar