Noniterative Implementation of a Class of Iterative Signal Restoration Algorithms

D. O. Walsh, P. A. Uelaney, Michael W Marcellin

Research output: Contribution to journalArticle

Abstract

In this paper, we show that a class of iterative signal restoration algorithms, which includes as a special case the discrete Gerchberg-Papoulis algorithm, can always be implemented directly (i.e., noniteratively). In the exactly and overdetermined cases, the iterative algorithm always converges to a unique least squares solution. In the underdetermined case, it is shown that the iterative algorithm always converges to the sum of a unique minimum norm solution and a term dependent on initial conditions. For the purposes of early termination, it is shown that the output of the iterative algorithm at the rth iteration can be computed directly using a singular value decomposition-based algorithm. The computational requirements of various iterative and noniterative implementations are dicussed, and the effect of the relaxation parameter on the regularization capability of the iterative algorithm is investigated.

Original languageEnglish (US)
Pages (from-to)1320
Number of pages1
JournalIEEE Transactions on Signal Processing
Volume44
Issue number5
StatePublished - 1996

Fingerprint

Restoration
Singular value decomposition

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Signal Processing

Cite this

Noniterative Implementation of a Class of Iterative Signal Restoration Algorithms. / Walsh, D. O.; Uelaney, P. A.; Marcellin, Michael W.

In: IEEE Transactions on Signal Processing, Vol. 44, No. 5, 1996, p. 1320.

Research output: Contribution to journalArticle

@article{892a36b19cec446d8643fd6645318ad9,
title = "Noniterative Implementation of a Class of Iterative Signal Restoration Algorithms",
abstract = "In this paper, we show that a class of iterative signal restoration algorithms, which includes as a special case the discrete Gerchberg-Papoulis algorithm, can always be implemented directly (i.e., noniteratively). In the exactly and overdetermined cases, the iterative algorithm always converges to a unique least squares solution. In the underdetermined case, it is shown that the iterative algorithm always converges to the sum of a unique minimum norm solution and a term dependent on initial conditions. For the purposes of early termination, it is shown that the output of the iterative algorithm at the rth iteration can be computed directly using a singular value decomposition-based algorithm. The computational requirements of various iterative and noniterative implementations are dicussed, and the effect of the relaxation parameter on the regularization capability of the iterative algorithm is investigated.",
author = "Walsh, {D. O.} and Uelaney, {P. A.} and Marcellin, {Michael W}",
year = "1996",
language = "English (US)",
volume = "44",
pages = "1320",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - Noniterative Implementation of a Class of Iterative Signal Restoration Algorithms

AU - Walsh, D. O.

AU - Uelaney, P. A.

AU - Marcellin, Michael W

PY - 1996

Y1 - 1996

N2 - In this paper, we show that a class of iterative signal restoration algorithms, which includes as a special case the discrete Gerchberg-Papoulis algorithm, can always be implemented directly (i.e., noniteratively). In the exactly and overdetermined cases, the iterative algorithm always converges to a unique least squares solution. In the underdetermined case, it is shown that the iterative algorithm always converges to the sum of a unique minimum norm solution and a term dependent on initial conditions. For the purposes of early termination, it is shown that the output of the iterative algorithm at the rth iteration can be computed directly using a singular value decomposition-based algorithm. The computational requirements of various iterative and noniterative implementations are dicussed, and the effect of the relaxation parameter on the regularization capability of the iterative algorithm is investigated.

AB - In this paper, we show that a class of iterative signal restoration algorithms, which includes as a special case the discrete Gerchberg-Papoulis algorithm, can always be implemented directly (i.e., noniteratively). In the exactly and overdetermined cases, the iterative algorithm always converges to a unique least squares solution. In the underdetermined case, it is shown that the iterative algorithm always converges to the sum of a unique minimum norm solution and a term dependent on initial conditions. For the purposes of early termination, it is shown that the output of the iterative algorithm at the rth iteration can be computed directly using a singular value decomposition-based algorithm. The computational requirements of various iterative and noniterative implementations are dicussed, and the effect of the relaxation parameter on the regularization capability of the iterative algorithm is investigated.

UR - http://www.scopus.com/inward/record.url?scp=33747673572&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33747673572&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:33747673572

VL - 44

SP - 1320

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 5

ER -