Volume 48, pp. 310-328, 2018.

A conditional gradient method for primal-dual total variation-based image denoising

Abdeslem Hafid Bentbib, Abderrahman Bouhamidi, and Karim Kreit

Abstract

In this paper, we consider the problem of image denoising by total variation regularization. We combine the conditional gradient method with the total variation regularization in the dual formulation to derive a new method for denoising images. The convergence of this method is proved. Some numerical examples are given to illustrate the effectiveness of the proposed method.

Full Text (PDF) [3.1 MB]

Key words

ill-posed problem, regularization, total variation, conditional gradient, image restoration, image denoising, convex programming

AMS subject classifications

65F10, 65R30

Links to the cited ETNA articles

[6]Vol. 18 (2004), pp. 153-173 Daniela Calvetti, Bryan Lewis, Lothar Reichel, and Fiorella Sgallari: Tikhonov regularization with nonnegativity constraint

< Back