Acceleration of the PDHGM on Partially Strongly Convex Functions

Tuomo Valkonen, Thomas Pock

Research output: Contribution to journalArticleResearchpeer-review

Abstract

We propose several variants of the primal–dual method due to Chambolle and Pock. Without requiring full strong convexity of the objective functions, our methods are accelerated on subspaces with strong convexity. This yields mixed rates, O(1 / N2) with respect to initialisation and O(1 / N) with respect to the dual sequence, and the residual part of the primal sequence. We demonstrate the efficacy of the proposed methods on image processing problems lacking strong convexity, such as total generalised variation denoising and total variation deblurring.

Original languageEnglish
Pages (from-to)394-414
Number of pages21
JournalJournal of Mathematical Imaging and Vision
Volume59
Issue number3
DOIs
Publication statusPublished - 1 Nov 2017

Fingerprint

convexity
Convex function
Convexity
Image processing
Primal-dual Method
Deblurring
Total Variation
Denoising
Initialization
image processing
Efficacy
Image Processing
Objective function
Subspace
Demonstrate

Keywords

  • Accelerated
  • Primal–dual
  • Subspace
  • Total generalised variation

ASJC Scopus subject areas

  • Statistics and Probability
  • Modelling and Simulation
  • Condensed Matter Physics
  • Computer Vision and Pattern Recognition
  • Geometry and Topology
  • Applied Mathematics

Cite this

Acceleration of the PDHGM on Partially Strongly Convex Functions. / Valkonen, Tuomo; Pock, Thomas.

In: Journal of Mathematical Imaging and Vision, Vol. 59, No. 3, 01.11.2017, p. 394-414.

Research output: Contribution to journalArticleResearchpeer-review

@article{43a3601da0bf4e3c981aa728beea9d33,
title = "Acceleration of the PDHGM on Partially Strongly Convex Functions",
abstract = "We propose several variants of the primal–dual method due to Chambolle and Pock. Without requiring full strong convexity of the objective functions, our methods are accelerated on subspaces with strong convexity. This yields mixed rates, O(1 / N2) with respect to initialisation and O(1 / N) with respect to the dual sequence, and the residual part of the primal sequence. We demonstrate the efficacy of the proposed methods on image processing problems lacking strong convexity, such as total generalised variation denoising and total variation deblurring.",
keywords = "Accelerated, Primal–dual, Subspace, Total generalised variation",
author = "Tuomo Valkonen and Thomas Pock",
year = "2017",
month = "11",
day = "1",
doi = "10.1007/s10851-016-0692-2",
language = "English",
volume = "59",
pages = "394--414",
journal = "Journal of Mathematical Imaging and Vision",
issn = "0924-9907",
publisher = "Springer Science+Business Media B.V",
number = "3",

}

TY - JOUR

T1 - Acceleration of the PDHGM on Partially Strongly Convex Functions

AU - Valkonen, Tuomo

AU - Pock, Thomas

PY - 2017/11/1

Y1 - 2017/11/1

N2 - We propose several variants of the primal–dual method due to Chambolle and Pock. Without requiring full strong convexity of the objective functions, our methods are accelerated on subspaces with strong convexity. This yields mixed rates, O(1 / N2) with respect to initialisation and O(1 / N) with respect to the dual sequence, and the residual part of the primal sequence. We demonstrate the efficacy of the proposed methods on image processing problems lacking strong convexity, such as total generalised variation denoising and total variation deblurring.

AB - We propose several variants of the primal–dual method due to Chambolle and Pock. Without requiring full strong convexity of the objective functions, our methods are accelerated on subspaces with strong convexity. This yields mixed rates, O(1 / N2) with respect to initialisation and O(1 / N) with respect to the dual sequence, and the residual part of the primal sequence. We demonstrate the efficacy of the proposed methods on image processing problems lacking strong convexity, such as total generalised variation denoising and total variation deblurring.

KW - Accelerated

KW - Primal–dual

KW - Subspace

KW - Total generalised variation

UR - http://www.scopus.com/inward/record.url?scp=85006085564&partnerID=8YFLogxK

U2 - 10.1007/s10851-016-0692-2

DO - 10.1007/s10851-016-0692-2

M3 - Article

VL - 59

SP - 394

EP - 414

JO - Journal of Mathematical Imaging and Vision

JF - Journal of Mathematical Imaging and Vision

SN - 0924-9907

IS - 3

ER -