$$\nu $$ν acceleration of statistical iterative methods for image restoration. Issue 5 (July 2016)

• Record Type: Journal Article
• Authors: Dell'Acqua, Pietro
• Is Part Of: Signal, image and video processing. Volume 10:Issue 5(2016)
• Page Start: 927
• Page End: 934
• Publication Date: 2016-07
• Subjects: Automatic acceleration -- $$\nu $$ν-method -- Lucy–Richardson method -- Image space reconstruction algorithm; Signal processing -- Digital techniques -- Periodicals; Image processing -- Digital techniques -- Periodicals; Digital video -- Periodicals; 621.3822
• DOI: 10.1007/s11760-015-0842-9
• Languages: English
• ISSNs: 1863-1703

Abstract

Abstract Image restoration problem is an important topic which appears in many different scientific areas. Several solving techniques are available, but generally in real applications, from which large-scale linear systems arise, the choice falls on iterative algorithms. In particular statistical methods (Lucy–Richardson method, image space reconstruction algorithm) have been extensively studied in the literature. Since they have low convergence rates, it is necessary to employ acceleration strategies. At this time, the most popular is the one introduced in 1997 by Biggs and Andrews, called automatic acceleration by the authors. In the present paper we describe $$\nu $$ ν acceleration, that is the translation of the idea behind $$\nu $$ ν -method, conceived for speeding up Landweber method, in the context of statistical methods. Computational results, which compare accelerated and classical methods, show the effectiveness of this strategy, which is able to get better performance than automatic acceleration.