Efficient retrieval of the regularized least-squares solution

An adaptive approach to the recovery of the least-squares (LS) solution employing regularization and conjugate gradient descent is developed. This is then implemented on image data corrupted by deterministic blur and additive Gaussian noise. The approach is useful where large data vectors are being...

Full description

Saved in:
Bibliographic Details
Published inOptical Engineering Vol. 37; no. 4; pp. 1283 - 1289
Main Author Sundaram, Ramakrishnan
Format Journal Article
LanguageEnglish
Published 01.04.1998
Subjects
Fourier domain formulations
Least-squares solution
Neural architectures
Restart procedures
Online AccessGet full text
ISSN0091-3286
1560-2303
DOI10.1117/1.601965

Cover

More Information
Summary:An adaptive approach to the recovery of the least-squares (LS) solution employing regularization and conjugate gradient descent is developed. This is then implemented on image data corrupted by deterministic blur and additive Gaussian noise. The approach is useful where large data vectors are being processed with the need to conserve space and obtain reliable estimates rapidly. ©
Bibliography:ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 23
ISSN:0091-3286
1560-2303
DOI:10.1117/1.601965