<title>Generalization of Strang's preconditioner with applications to iterative deconvolution</title>
Lists
Only you see this
Abstract
In this paper, we proposed a method to generalize Strang's circulant preconditioner for arbitrary n-by-n matrices A<SUB>n</SUB>. The [n/2]th column of our circulant preconditioner S<SUB>n</SUB> is equal to the [n/2]th column of the given matrix A<SUB>n</SUB>. Thus if A<SUB>n</SUB> is a square Toeplitz matrix, then S<SUB>n</SUB> is just the Strang circulant preconditioner. When S<SUB>n</SUB> is not Hermitian, our circulant preconditioner can be defined as (S<SUP>*</SUP><SUB>n</SUB>S<SUB>n</SUB>)<SUP>1/2</SUP>. This construction is similar to the forward-backward projection method used in constructing preconditioners for tomographic inversion problems in medical imaging. Comparisons of our preconditioner S<SUB>n</SUB> with other circulant-based preconditioners are carried out for some 1D Toeplitz least squares problems: minb - Ax<SUB>2</SUB>. Preliminary numerical results show that S<SUB>n</SUB> performs quite well. Test results are also reported for a 2D deconvolution problem arising in ground-based atmospheric imaging.
BibTeX
@inproceedings{chan1994lt,
author = {Raymond H. Chan and Michael K. Ng and Robert J. Plemmons},
title = {{\<title\>Generalization of Strang's preconditioner with applications to iterative deconvolution\</title\>}},
booktitle = {Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE},
year = {1994},
doi = {10.1117/12.190864},
}Loading related content...