Hi everyone,

I have a general question regarding KPM. Since kronecker product matrices have cartesian tiling, I was wondering if these could be made more efficient by implementing hexagonal sampling instead of cartesian within kronecker matrices. Is it possible to do that? I'm confused since kronecker matrices are separable and I just wanted to have a clear theoretical idea on how could I switch to hexagonal sampling or blue noise sampling to increase the efficiency of kronecker product matrices.
Your ideas would be warmly appreciated.