The code needs vec2mat from the communication toolbox. Since I don't have this toolbox, I used the following script: function [ mat, npadded ] = vec2mat( vec, N ) v = vec(:); nelem = size(v,1); nrows = ceil(nelem/N); npadded = nrows*N - nelem; newv = [v; zeros(npadded,1)]; mat=reshape(newv,N, nrows)'; end . The program worked but created unexpected complex values and the patch function could not run, for the example given in the readme. For the same example but replacing the translation by [1 1 1] gave me an interpolation result which does not look plausible and it appears that the created variables have a lot of NaN values (I had a warning that the matrix is singular to working precision, so that could explain it). Another reason could be my own implementation of vec2mat given above.
If you want to contribute with another review, please follow these instructions.
Please consider to cut/paste/edit the raw JSON data attached to this paper.