Sparse signal approximation via nonseparable regularization

Ivan Selesnick, Masoud Farshchian

Research output: Contribution to journalArticle

Abstract

The calculation of a sparse approximate solution to a linear system of equations is often performed using either L1-norm regularization and convex optimization or nonconvex regularization and nonconvex optimization. Combining these principles, this paper describes a type of nonconvex regularization that maintains the convexity of the objective function, thereby allowing the calculation of a sparse approximate solution via convex optimization. The preservation of convexity is viable in the proposed approach because it uses a regularizer that is nonseparable. The proposed method is motivated and demonstrated by the calculation of sparse signal approximation using tight frames. Examples of denoising demonstrate improvement relative to L1 norm regularization.

Original languageEnglish (US)
Article number7857076
Pages (from-to)2561-2575
Number of pages15
JournalIEEE Transactions on Signal Processing
Volume65
Issue number10
DOIs
StatePublished - May 15 2017

Fingerprint

Convex optimization
Linear systems

Keywords

  • convex function
  • denoising
  • optimization
  • sparse approximation
  • Sparse signal model

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Sparse signal approximation via nonseparable regularization. / Selesnick, Ivan; Farshchian, Masoud.

In: IEEE Transactions on Signal Processing, Vol. 65, No. 10, 7857076, 15.05.2017, p. 2561-2575.

Research output: Contribution to journalArticle

@article{9cd3b5b7c9c34d16ad57e63295628fb5,
title = "Sparse signal approximation via nonseparable regularization",
abstract = "The calculation of a sparse approximate solution to a linear system of equations is often performed using either L1-norm regularization and convex optimization or nonconvex regularization and nonconvex optimization. Combining these principles, this paper describes a type of nonconvex regularization that maintains the convexity of the objective function, thereby allowing the calculation of a sparse approximate solution via convex optimization. The preservation of convexity is viable in the proposed approach because it uses a regularizer that is nonseparable. The proposed method is motivated and demonstrated by the calculation of sparse signal approximation using tight frames. Examples of denoising demonstrate improvement relative to L1 norm regularization.",
keywords = "convex function, denoising, optimization, sparse approximation, Sparse signal model",
author = "Ivan Selesnick and Masoud Farshchian",
year = "2017",
month = "5",
day = "15",
doi = "10.1109/TSP.2017.2669904",
language = "English (US)",
volume = "65",
pages = "2561--2575",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "10",

}

TY - JOUR

T1 - Sparse signal approximation via nonseparable regularization

AU - Selesnick, Ivan

AU - Farshchian, Masoud

PY - 2017/5/15

Y1 - 2017/5/15

N2 - The calculation of a sparse approximate solution to a linear system of equations is often performed using either L1-norm regularization and convex optimization or nonconvex regularization and nonconvex optimization. Combining these principles, this paper describes a type of nonconvex regularization that maintains the convexity of the objective function, thereby allowing the calculation of a sparse approximate solution via convex optimization. The preservation of convexity is viable in the proposed approach because it uses a regularizer that is nonseparable. The proposed method is motivated and demonstrated by the calculation of sparse signal approximation using tight frames. Examples of denoising demonstrate improvement relative to L1 norm regularization.

AB - The calculation of a sparse approximate solution to a linear system of equations is often performed using either L1-norm regularization and convex optimization or nonconvex regularization and nonconvex optimization. Combining these principles, this paper describes a type of nonconvex regularization that maintains the convexity of the objective function, thereby allowing the calculation of a sparse approximate solution via convex optimization. The preservation of convexity is viable in the proposed approach because it uses a regularizer that is nonseparable. The proposed method is motivated and demonstrated by the calculation of sparse signal approximation using tight frames. Examples of denoising demonstrate improvement relative to L1 norm regularization.

KW - convex function

KW - denoising

KW - optimization

KW - sparse approximation

KW - Sparse signal model

UR - http://www.scopus.com/inward/record.url?scp=85017578146&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85017578146&partnerID=8YFLogxK

U2 - 10.1109/TSP.2017.2669904

DO - 10.1109/TSP.2017.2669904

M3 - Article

AN - SCOPUS:85017578146

VL - 65

SP - 2561

EP - 2575

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 10

M1 - 7857076

ER -