A random block-coordinate primal-dual proximal algorithm with application to 3D mesh denoising

Audrey Repetti; Emilie Chouzenoux; Jean-Christophe Pesquet

doi:10.1109/ICASSP.2015.7178634

A random block-coordinate primal-dual proximal algorithm with application to 3D mesh denoising

Audrey Repetti, Emilie Chouzenoux, Jean-Christophe Pesquet

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

2 Citations (Scopus)

Abstract

Primal-dual proximal optimization methods have recently gained much interest for dealing with very large-scale data sets encoutered in many application fields such as machine learning, computer vision and inverse problems [1-3]. In this work, we propose a novel random block-coordinate version of such algorithms allowing us to solve a wide array of convex variational problems. One of the main advantages of the proposed algorithm is its ability to solve composite problems involving large-size matrices without requiring any inversion. In addition, the almost sure convergence to an optimal solution to the problem is guaranteed. We illustrate the good performance of our method on a mesh denoising application.

Original language	English
Title of host publication	2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Publisher	IEEE
Pages	3561-3565
Number of pages	5
ISBN (Electronic)	9781467369978
DOIs	https://doi.org/10.1109/ICASSP.2015.7178634
Publication status	Published - 6 Aug 2015
Event	40th IEEE International Conference on Acoustics, Speech and Signal Processing 2015 - Brisbane, Australia Duration: 19 Apr 2015 → 24 Apr 2015

Conference

Conference	40th IEEE International Conference on Acoustics, Speech and Signal Processing 2015
Abbreviated title	ICASSP 2015
Country	Australia
City	Brisbane
Period	19/04/15 → 24/04/15

Access to Document

10.1109/ICASSP.2015.7178634

Fingerprint Dive into the research topics of 'A random block-coordinate primal-dual proximal algorithm with application to 3D mesh denoising'. Together they form a unique fingerprint.

Cite this

@inproceedings{6185244039944a3c8a151d5f1c2a2b31,

title = "A random block-coordinate primal-dual proximal algorithm with application to 3D mesh denoising",

abstract = "Primal-dual proximal optimization methods have recently gained much interest for dealing with very large-scale data sets encoutered in many application fields such as machine learning, computer vision and inverse problems [1-3]. In this work, we propose a novel random block-coordinate version of such algorithms allowing us to solve a wide array of convex variational problems. One of the main advantages of the proposed algorithm is its ability to solve composite problems involving large-size matrices without requiring any inversion. In addition, the almost sure convergence to an optimal solution to the problem is guaranteed. We illustrate the good performance of our method on a mesh denoising application.",

author = "Audrey Repetti and Emilie Chouzenoux and Jean-Christophe Pesquet",

year = "2015",

month = aug,

day = "6",

doi = "10.1109/ICASSP.2015.7178634",

language = "English",

pages = "3561--3565",

booktitle = "2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)",

publisher = "IEEE",

address = "United States",

note = "40th IEEE International Conference on Acoustics, Speech and Signal Processing 2015, ICASSP 2015 ; Conference date: 19-04-2015 Through 24-04-2015",

}

Repetti, A, Chouzenoux, E & Pesquet, J-C 2015, A random block-coordinate primal-dual proximal algorithm with application to 3D mesh denoising. in 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, pp. 3561-3565, 40th IEEE International Conference on Acoustics, Speech and Signal Processing 2015, Brisbane, Australia, 19/04/15. https://doi.org/10.1109/ICASSP.2015.7178634

TY - GEN

T1 - A random block-coordinate primal-dual proximal algorithm with application to 3D mesh denoising

AU - Repetti, Audrey

AU - Chouzenoux, Emilie

AU - Pesquet, Jean-Christophe

PY - 2015/8/6

Y1 - 2015/8/6

N2 - Primal-dual proximal optimization methods have recently gained much interest for dealing with very large-scale data sets encoutered in many application fields such as machine learning, computer vision and inverse problems [1-3]. In this work, we propose a novel random block-coordinate version of such algorithms allowing us to solve a wide array of convex variational problems. One of the main advantages of the proposed algorithm is its ability to solve composite problems involving large-size matrices without requiring any inversion. In addition, the almost sure convergence to an optimal solution to the problem is guaranteed. We illustrate the good performance of our method on a mesh denoising application.

AB - Primal-dual proximal optimization methods have recently gained much interest for dealing with very large-scale data sets encoutered in many application fields such as machine learning, computer vision and inverse problems [1-3]. In this work, we propose a novel random block-coordinate version of such algorithms allowing us to solve a wide array of convex variational problems. One of the main advantages of the proposed algorithm is its ability to solve composite problems involving large-size matrices without requiring any inversion. In addition, the almost sure convergence to an optimal solution to the problem is guaranteed. We illustrate the good performance of our method on a mesh denoising application.

U2 - 10.1109/ICASSP.2015.7178634

DO - 10.1109/ICASSP.2015.7178634

M3 - Conference contribution

SP - 3561

EP - 3565

BT - 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

PB - IEEE

T2 - 40th IEEE International Conference on Acoustics, Speech and Signal Processing 2015

Y2 - 19 April 2015 through 24 April 2015

ER -