Eventual linear convergence rate of an exchange algorithm for superresolution - Archive ouverte HAL Access content directly
Conference Papers Year : 2019

Eventual linear convergence rate of an exchange algorithm for superresolution

Abstract

We study an iterative discretization algorithm for solving optimization problems regularized by the total variation norm. Its design relies on ideas from the Frank-Wolfe algorithm, as well as from semi-infinite programming. For smooth regularity terms, we prove an eventual linear convergence rate guarantee.
Fichier principal
Vignette du fichier
Spars_manuscript.pdf (155.78 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-02374440 , version 1 (03-01-2021)

Identifiers

  • HAL Id : hal-02374440 , version 1

Cite

Frédéric de Gournay, Pierre Weiss, Axel Flinth. Eventual linear convergence rate of an exchange algorithm for superresolution. SPARS, Jul 2019, TOULOUSE, France. ⟨hal-02374440⟩
12 View
19 Download

Share

Gmail Facebook Twitter LinkedIn More