zorah/kernelmatching

Implementation of Efficient Deformable Shape Correspondence via Kernel Matching (https://arxiv.org/abs/1707.08991).

kernelmatching's Issues

geodesic_matlab_api.so is not a valid shared library

Should we build it? tried downloading the matlab library from here
but failed.
Do you remember how you built it?
Thanks a lot!

geodesic_matlab_api.so crush on build_adjacencies only when running on the tosca dataset

I'm able to run the kernelmatching example on the tosca example provided, but it crashes with the following coredump when running on two real samples of the faust dataset (~ 100K vertices).
Do you remember having problems running the example on that dataset? are there a need of modifications for that matter?
Thanks a lot for all your help zorah:)

/lib/x86_64-linux-gnu/libc.so.6+00257687 gsignal+00000199
[  1] 0x00007fb9318f2801                    /lib/x86_64-linux-gnu/libc.so.6+00264193 abort+00000321
[  2] 0x00007fb9318e239a                    /lib/x86_64-linux-gnu/libc.so.6+00197530
[  3] 0x00007fb9318e2412                    /lib/x86_64-linux-gnu/libc.so.6+00197650
[  4] 0x00007fb6ea9070e0 /home/dvir/PycharmProjects/gml/Dvir/FMunsup/misc/post_process_tools/KernelMatching-master/tools/geodesic_cpp/geodesic_matlab_api.so+00110816 _ZN8geodesic4Mesh17build_adjacenciesEv+00001008
[  5] 0x00007fb6ea90a905 /home/dvir/PycharmProjects/gml/Dvir/FMunsup/misc/post_process_tools/KernelMatching-master/tools/geodesic_cpp/geodesic_matlab_api.so+00125189 _ZN8geodesic4Mesh20initialize_mesh_dataIPdPlEEvjRT_jRT0_+00000997
[  6] 0x00007fb6ea8f69bc /home/dvir/PycharmProjects/gml/Dvir/FMunsup/misc/post_process_tools/KernelMatching-master/tools/geodesic_cpp/geodesic_matlab_api.so+00043452 new_mesh+00000460
[  7] 0x00007fb6ea6ea70b /tmp/tp1a42b427_f5e0_4b90_9973_4828b83ad0a7/geodesic_matlab_api_thunk_glnxa64.so+00001803 longlongvoidPtrlongvoidPtrvoidPtrvoidPtrThunk+00000145
[  8] 0x00007fb72227d223                            bin/glnxa64/libmwcli.so+00172579
[  9] 0x00007fb722295d28                            bin/glnxa64/libmwcli.so+00273704
[ 10] 0x00007fb7222962e8                            bin/glnxa64/libmwcli.so+00275176
[ 11] 0x00007fb91e2e0b02                   bin/glnxa64/libmwm_dispatcher.so+00543490
[ 12] 0x00007fb91e2bda40                   bin/glnxa64/libmwm_dispatcher.so+00399936 _ZN13Mfh_MATLAB_fn16dispatch_fh_implEMS_FviPP11mxArray_tagiS2_EiS2_iS2_+00000448
[ 13] 0x00007fb91e2bdc2e                   bin/glnxa64/libmwm_dispatcher.so+00400430 _ZN13Mfh_MATLAB_fn22dispatch_fh_with_reuseEiPP11mxArray_tagiS2_+00000030
[ 14] 0x00007fb91a801d4e                          bin/glnxa64/libmwm_lxe.so+13725006

Can you look the picture of my result and give some detail advice to me ?

I am so sorry to ask you again,but I am failed to get a good result using my data ,I have uploaded the picture of my result.Can you give me some detail advice on it.Looking forward to your response.Thank you very much.

remove outliers

Hi Zorah, can you give a brief introduction of how to use your code to run PMF in order to remove the outliers of my correspondence results? Thanks.

Are there any specific requirements for data sets?

Hello,I am very glad to find your works,and run your experiment about cats is successful, but when run my own data set is problematic - corresponding is disordered (a position of the original mesh is corresponding to multiple and discontinuous locations of the target mesh), so I want to ask about that if there are have any specific requirements for the data set in the experiment of cat? Looking forward to your reply. Thank you very much!!!!

remesh_and_fix algorithm

Hi,
In the example_run script you mention remesh_and_fix, which is indeed needed in some meshes:)

Can you please provide the script?
Thanks:)

Are there any specific requirements for data sets?

Hello,I am very glad to find your works,and run your experiment about cats is successful, but when run my own data set is problematic - corresponding is disordered (a position of the original mesh is corresponding to multiple and discontinuous locations of the target mesh), so I want to ask about that if there are have any specific requirements for the data set in the experiment of cat? Looking forward to your reply. Thank you very much!!!!

Are there any specific requirements for data sets?

Hello,I am very glad to find your works,and run your experiment about cats is successful, but when run my own data set is problematic - corresponding is disordered (a position of the original mesh is corresponding to multiple and discontinuous locations of the target mesh), so I want to ask about that if there are have any specific requirements for the data set in the experiment of cat? Looking forward to your reply. Thank you very much!!!!

What() is:invalid vector<T> subscript

I'm having the following error running the PMF example

Unexpected Standard exception from MEX file.What() is:invalid vector subscript..Error in example_run (line 154)X.desc = calc_shot(X.VERT',X.TRIV', 1:X.n, opts.shot_num_bins, opts.shot_radius*sqrt(X.area)/100, 3)';

with a mesh processed by the fixme algorithm (attached both the lores and the regular mesh)

Do you know what can be the problem?

Thanks a lot:)

Uploading lores_reg_mesh_fixed.zip…

Could you give your data of kids you uesd in this experiment?

Hello ,I am very sorry to bither you again and again.I think my data is suitabe for your first experiment about kids,but I do not konw how you deal with your data of kids and how it runs,especially do not find the input codes of the sparse collection of point-wise matches(as few as two).So could you give me the data of the kids you uesd in this experiment and the related implementation codes just on the kids.And if it is nessary to normalize the data before use it? My email is [email protected]. I will be forever grateful to you for extending my any help.

Are there any specific requirements for data sets?

Hello,I am very glad to find your works,and run your experiment about cats is successful, but when run my own data set is problematic - corresponding is disordered (a position of the original mesh is corresponding to multiple and discontinuous locations of the target mesh), so I want to ask about that if there are have any specific requirements for the data set in the experiment of cat? Looking forward to your reply. Thank you very much!!!!

Possible with Octave?

Hi,

Firstly thank you for this contribution. I am researching on a similar idea and was hoping to validate my method with your work by comparing the results. Is it possible that I can run your code on a Linux machine with Octave instead of Matlab? The Linux flavor I use is Ubuntu 18.04 with Octave 4.2.2. I was able to compile the MEX files without any fuss, the only problems I face is for using geodesic CPP. I see you are using CPP shared libraries with Matlab functions libisloaded, unloadlibrary, calllib, etc., which sadly isn't available in Octave. Do you know if there is a workaround for using the geodesics? Is it possible to use your exactgeodesics matlab mex? I was able compile that project sucessfully

Regards,
#0K