Efficient Gaussian Sampling for Solving Large-Scale Inverse Problems using MCMC - Archive ouverte HAL Access content directly
Journal Articles IEEE Transactions on Signal Processing Year : 2015

Efficient Gaussian Sampling for Solving Large-Scale Inverse Problems using MCMC

Abstract

The resolution of many large-scale inverse problems using MCMC methods requires a step of drawing samples from a high dimensional Gaussian distribution. Based on the reversible jump Markov chain framework, this paper proposes an efficient Gaussian sampling algorithm having a reduced computation cost and memory usage, as compared to classical methods based on Cholesky factorization. The main feature of the algorithm is to perform an approximate resolution of a linear system with a truncation level adjusted using a self-tuning adaptive scheme allowing to achieve the minimal computation cost. The connection between this algorithm and some existing strategies is discussed and its efficiency is illustrated on a linear inverse problem of image resolution enhancement.
Fichier principal
Vignette du fichier
Gilavert2014_RJPO_draft.pdf (824.27 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01059414 , version 1 (31-08-2014)

Licence

Attribution

Identifiers

Cite

Clément Gilavert, Saïd Moussaoui, Jérôme Idier. Efficient Gaussian Sampling for Solving Large-Scale Inverse Problems using MCMC. IEEE Transactions on Signal Processing, 2015, 63 (01), pp.70-80. ⟨10.1109/TSP.2014.2367457⟩. ⟨hal-01059414⟩
254 View
663 Download

Altmetric

Share

Gmail Facebook X LinkedIn More