Kronecker Products in Image Restoration
J. Nagy and L. Perrone
A flexible preconditioning approach based on Kronecker product
and singular value decomposition (SVD)
approximations is presented. The approach can be used
with a variety of boundary conditions, depending on what is most
appropriate for the specific deblurring application. It is
shown that, regardless of the imposed boundary condition,
SVD approximations can be used effectively in filtering methods,
such as the truncated SVD,
as well as in designing preconditioners for iterative methods.