Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/68369
Title: Alternating direction method of multipliers for a class of nonconvex and nonsmooth problems with applications to background/foreground extraction
Authors: Yang, L 
Pong, TK 
Chen, XJ 
Keywords: Nonsmooth and nonconex optimization
Alternating dirtection method of multipliers
Dual step-size
Background/foreground extraction
Issue Date: 2017
Publisher: Society for Industrial and Applied Mathematics
Source: SIAM journal on imaging sciences, 2017, v. 10, no. 1, p. 74-110 How to cite?
Journal: SIAM journal on imaging sciences 
Abstract: In this paper, we study a general optimization model, which covers a large class of existing models for many applications in imaging sciences. To solve the resulting possibly nonconvex, nonsmooth and non-Lipschitz optimization problem, we adapt the alternating direction method of multipliers (ADMM) with a general dual step-size to solve a reformulation that contains three blocks of variables, and analyze its convergence. We show that for any dual step-size less than the golden ratio, there exists a computable threshold such that if the penalty parameter is chosen above such a threshold and the sequence thus generated by our ADMM is bounded, then the cluster point of the sequence gives a stationary point of the nonconvex optimization problem. We achieve this via apotential function specifically constructed for our ADMM. Moreover, we establish the global conver-gence of the whole sequence if, in addition, this special pot ential function is a Kurdyka-Lojasiewicz function. Furthermore, we present a simple strategy for initializing the algorithm to guarantee bound-edness of the sequence. Finally, we perform numerical experiments comparing our ADMM with the proximal alternating linearized minimization (PALM) proposed in [5] on the background/foreground extraction problem with real data. The numerical results show that our ADMM with a nontrivial dual step-size is efficient.
URI: http://hdl.handle.net/10397/68369
ISSN: 1936-4954
DOI: 10.1137/15M1027528
Appears in Collections:Journal/Magazine Article

Access
View full-text via PolyU eLinks SFX Query
Show full item record

Page view(s)

37
Checked on Aug 21, 2017

Google ScholarTM

Check

Altmetric



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.