Module PkgSurfaceMeshSkeletonizationRef

doc+code: mention that to get a better skeleton that is closer to the medial axis, the surface must be sufficiently well sampled so that the Voronoi poles lie on the media axis (see Amenta's paper). Propose the usage of the isotropic remeshing and see if we add a boolean to do it automatically in the api (correspondance would be broken)

code: implement the random sampling of surface using the work started by Alexandru during its gsoc to get a better approximation of poles

code: expose in polygon mesh processing the function to compute the voronoi pole of a close triangle mesh

code: expose in polygon mesh processing the function to remesh locally a triangle mesh with the angle and edge length parameters

code: avoid using EPEC for the triangulation