FFT-based preconditioners for Toeplitz-Block least square problems
- Univ. of Hong Kong (Hong Kong). Dept. of Mathematics
- Univ. of Minnesota, Minneapolis, MN (United States). Inst. for Mathematics and its Applications
Discretized two-dimensional deconvolution problems arising, e.g., in image restoration and seismic tomography, can be formulated as least squares computations, min [parallel] b [minus] Tx [parallel][sub 2], where T is often a large-scale rectangular Toeplitz-block matrix. The authors consider solving such block least squares problems by the preconditioned conjugate gradient algorithm using square nonsingular circulant-block and related preconditioners, constructed from the blocks of the rectangular matrix T. Preconditioning with such matrices allows efficient implementation using the one-dimensional or two-dimensional fast Fourier transform (FFT). Two-block preconditioners, related to those proposed by T. Chan and J. Olkin for square nonsingular Toeplitz-block systems, are derived and analyzed. It is shown that, for important classes of T, the singular values of the preconditioned matrix are clustered around one. This extends the authors' earlier work on preconditioners for Toeplitz least squares iterations for one-dimensional problems. It is well known that the resolution of ill-posed deconvolution problems can be substantially improved by regularization to compensate for their ill-posed nature. It is shown that regularization can easily be incorporated into the preconditioners, and a report is given on numerical experiments on a Cray Y-MP. The experiments illustrate good convergence properties of these FFT-based preconditioned iterations.
- DOE Contract Number:
- FG03-87ER25037
- OSTI ID:
- 5331996
- Journal Information:
- SIAM Journal on Numerical Analysis (Society for Industrial and Applied Mathematics); (United States), Vol. 30:6; ISSN 0036-1429
- Country of Publication:
- United States
- Language:
- English
Similar Records
Fast band-Toeplitz preconditioners for Hermitian Toeplitz systems
Iterative methods for Toeplitz-like matrices
Related Subjects
LEAST SQUARE FIT
ALGORITHMS
CRAY COMPUTERS
FOURIER TRANSFORMATION
IMAGE PROCESSING
SEISMOLOGY
COMPUTERS
INTEGRAL TRANSFORMATIONS
MATHEMATICAL LOGIC
MAXIMUM-LIKELIHOOD FIT
NUMERICAL SOLUTION
PROCESSING
TRANSFORMATIONS
990200* - Mathematics & Computers