Generalized Subspace Pursuit and an application to sparse Poisson denoising - Usages et représentations territoriales en contexte amazonien traditionnel Access content directly
Conference Papers Year : 2014

Generalized Subspace Pursuit and an application to sparse Poisson denoising

Abstract

We present a generalization of Subspace Pursuit, which seeks the k-sparse vector that minimizes a generic cost function. We introduce the Restricted Diagonal Property, which much like RIP in the classical setting, enables to control the convergence of Generalized Subspace Pursuit (GSP). To tackle the problem of Poisson denoising, we propose to use GSP together with the Moreau-Yosida approximation of the Poisson likelihood. Experiments were conducted on synthetic, exact sparse and natural images corrupted by Poisson noise. We study the influence of the different parameters and show that our approach performs better than Subspace Pursuit or l1-relaxed methods and compares favorably to state-of-art methods.
Fichier principal
Vignette du fichier
GSP.pdf (524.77 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01071760 , version 1 (06-10-2014)

Identifiers

  • HAL Id : hal-01071760 , version 1

Cite

François-Xavier Dupé, Sandrine Anthoine. Generalized Subspace Pursuit and an application to sparse Poisson denoising. International Conference on Image Processing (ICIP) 2014, IEEE conference on,, Oct 2014, Paris, France. ⟨hal-01071760⟩
429 View
606 Download

Share

Gmail Facebook X LinkedIn More