Solving robust regularization problems using iteratively re-weighted least squares

Khurrum Aftab Kiani, Tom Drummond

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

1 Citation (Scopus)

Abstract

Many computer vision problems are formulated as an objective function consisting of a sum of functions. In the case of ill-constrained problems, regularization terms are included in the objective function to reduce the ambiguity and noise in the solution. The most commonly used regularization terms are the L2 norm and the L1 norm. Since the last two decades, the class of regularized problems, especially the L1-regularized problems, has received much attention but still many regularized problems are either difficult to solve, or require complex optimization techniques. We propose a method based on an Iteratively Re-weighted Least Squares approach to minimize an objective function comprising a mixture of m-estimator regularization terms. In addition to the proof of convergence of the algorithm to the desired minimum, we show the applicability of the proposed algorithm by solving the problems of edge-preserved image denoising and image super-resolution. In both the cases, our experimental results show that the proposed algorithm gives superior results to the state-of-Art regularization methods.

Original languageEnglish
Title of host publication2017 IEEE Winter Conference on Applications of Computer Vision (WACV 2017)
Subtitle of host publicationSanta Rosa, California, USA, 24-31 March 2017
Place of PublicationPiscataway, NJ
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages483-492
Number of pages10
ISBN (Electronic)9781509048229
ISBN (Print)9781509048236
DOIs
Publication statusPublished - 11 May 2017
EventIEEE Winter Conference on Applications of Computer Vision 2017 - Santa Rosa, United States of America
Duration: 24 Mar 201731 Mar 2017
http://pamitc.org/wacv2017/
http://www.wikicfp.com/cfp/program?id=2993&s=WACV&f=Workshop%20on%20Applications%20of%20Computer%20Vision

Conference

ConferenceIEEE Winter Conference on Applications of Computer Vision 2017
Abbreviated titleWACV 2017
CountryUnited States of America
CitySanta Rosa
Period24/03/1731/03/17
Internet address

Cite this

Kiani, K. A., & Drummond, T. (2017). Solving robust regularization problems using iteratively re-weighted least squares. In 2017 IEEE Winter Conference on Applications of Computer Vision (WACV 2017): Santa Rosa, California, USA, 24-31 March 2017 (pp. 483-492). [7926643] Piscataway, NJ: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/WACV.2017.60
Kiani, Khurrum Aftab ; Drummond, Tom. / Solving robust regularization problems using iteratively re-weighted least squares. 2017 IEEE Winter Conference on Applications of Computer Vision (WACV 2017): Santa Rosa, California, USA, 24-31 March 2017. Piscataway, NJ : IEEE, Institute of Electrical and Electronics Engineers, 2017. pp. 483-492
@inproceedings{c08f78ce6f6e4695bc80746a5ab08044,
title = "Solving robust regularization problems using iteratively re-weighted least squares",
abstract = "Many computer vision problems are formulated as an objective function consisting of a sum of functions. In the case of ill-constrained problems, regularization terms are included in the objective function to reduce the ambiguity and noise in the solution. The most commonly used regularization terms are the L2 norm and the L1 norm. Since the last two decades, the class of regularized problems, especially the L1-regularized problems, has received much attention but still many regularized problems are either difficult to solve, or require complex optimization techniques. We propose a method based on an Iteratively Re-weighted Least Squares approach to minimize an objective function comprising a mixture of m-estimator regularization terms. In addition to the proof of convergence of the algorithm to the desired minimum, we show the applicability of the proposed algorithm by solving the problems of edge-preserved image denoising and image super-resolution. In both the cases, our experimental results show that the proposed algorithm gives superior results to the state-of-Art regularization methods.",
author = "Kiani, {Khurrum Aftab} and Tom Drummond",
year = "2017",
month = "5",
day = "11",
doi = "10.1109/WACV.2017.60",
language = "English",
isbn = "9781509048236",
pages = "483--492",
booktitle = "2017 IEEE Winter Conference on Applications of Computer Vision (WACV 2017)",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States of America",

}

Kiani, KA & Drummond, T 2017, Solving robust regularization problems using iteratively re-weighted least squares. in 2017 IEEE Winter Conference on Applications of Computer Vision (WACV 2017): Santa Rosa, California, USA, 24-31 March 2017., 7926643, IEEE, Institute of Electrical and Electronics Engineers, Piscataway, NJ, pp. 483-492, IEEE Winter Conference on Applications of Computer Vision 2017, Santa Rosa, United States of America, 24/03/17. https://doi.org/10.1109/WACV.2017.60

Solving robust regularization problems using iteratively re-weighted least squares. / Kiani, Khurrum Aftab; Drummond, Tom.

2017 IEEE Winter Conference on Applications of Computer Vision (WACV 2017): Santa Rosa, California, USA, 24-31 March 2017. Piscataway, NJ : IEEE, Institute of Electrical and Electronics Engineers, 2017. p. 483-492 7926643.

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

TY - GEN

T1 - Solving robust regularization problems using iteratively re-weighted least squares

AU - Kiani, Khurrum Aftab

AU - Drummond, Tom

PY - 2017/5/11

Y1 - 2017/5/11

N2 - Many computer vision problems are formulated as an objective function consisting of a sum of functions. In the case of ill-constrained problems, regularization terms are included in the objective function to reduce the ambiguity and noise in the solution. The most commonly used regularization terms are the L2 norm and the L1 norm. Since the last two decades, the class of regularized problems, especially the L1-regularized problems, has received much attention but still many regularized problems are either difficult to solve, or require complex optimization techniques. We propose a method based on an Iteratively Re-weighted Least Squares approach to minimize an objective function comprising a mixture of m-estimator regularization terms. In addition to the proof of convergence of the algorithm to the desired minimum, we show the applicability of the proposed algorithm by solving the problems of edge-preserved image denoising and image super-resolution. In both the cases, our experimental results show that the proposed algorithm gives superior results to the state-of-Art regularization methods.

AB - Many computer vision problems are formulated as an objective function consisting of a sum of functions. In the case of ill-constrained problems, regularization terms are included in the objective function to reduce the ambiguity and noise in the solution. The most commonly used regularization terms are the L2 norm and the L1 norm. Since the last two decades, the class of regularized problems, especially the L1-regularized problems, has received much attention but still many regularized problems are either difficult to solve, or require complex optimization techniques. We propose a method based on an Iteratively Re-weighted Least Squares approach to minimize an objective function comprising a mixture of m-estimator regularization terms. In addition to the proof of convergence of the algorithm to the desired minimum, we show the applicability of the proposed algorithm by solving the problems of edge-preserved image denoising and image super-resolution. In both the cases, our experimental results show that the proposed algorithm gives superior results to the state-of-Art regularization methods.

UR - http://www.scopus.com/inward/record.url?scp=85020216484&partnerID=8YFLogxK

U2 - 10.1109/WACV.2017.60

DO - 10.1109/WACV.2017.60

M3 - Conference Paper

AN - SCOPUS:85020216484

SN - 9781509048236

SP - 483

EP - 492

BT - 2017 IEEE Winter Conference on Applications of Computer Vision (WACV 2017)

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Piscataway, NJ

ER -

Kiani KA, Drummond T. Solving robust regularization problems using iteratively re-weighted least squares. In 2017 IEEE Winter Conference on Applications of Computer Vision (WACV 2017): Santa Rosa, California, USA, 24-31 March 2017. Piscataway, NJ: IEEE, Institute of Electrical and Electronics Engineers. 2017. p. 483-492. 7926643 https://doi.org/10.1109/WACV.2017.60