Point-set registration methods using the EM algorithm and Softassign that exploits geometric and structural evidence.
The code contains rigid and non-rigid registration algorithms.
The parameters of the methods are especially set to be used with Chui's fish and Chinese character datasets.
Gerard Sanromà (2020). Smooth Point-set Registration using Neighboring Constraints (https://www.mathworks.com/matlabcentral/fileexchange/35179-smooth-point-set-registration-using-neighboring-constraints), MATLAB Central File Exchange. Retrieved .
the link to the Chui's fish dataset doesn't work anymore, is it possible to provide a working link?
you can monitor the status of the process at each iteration by checking the variables vgM (deformed point-set) and Sd (correspondence matrix), inside the main loop of function "match_chui" (or "match_chui_out" if you use the outliers' version).
Also, in case you use this software in your research, I kindly ask you to cite the paper with the same name in the Pattern Recognition Letters journal.
Hope that helps.
After using your non rigid demo, i have two questions:
Is there a function to visualize what is happening? And is it possible to use it for 3D data?