Skip to Main content Skip to Navigation
Journal articles

Bayesian selection for the l2-Potts model regularization parameter: 1D piecewise constant signal denoising

Abstract : Piecewise constant denoising can be solved either by deterministic optimization approaches, based on the Potts model, or by stochastic Bayesian procedures. The former lead to low computational time but require the selection of a regularization parameter, whose value significantly impacts the achieved solution, and whose automated selection remains an involved and challenging problem. Conversely, fully Bayesian formalisms encapsulate the regularization parameter selection into hierarchical models, at the price of high computational costs. This contribution proposes an operational strategy that combines hierarchical Bayesian and Potts model formulations, with the double aim of automatically tuning the regularization parameter and of maintaining computational effciency. The proposed procedure relies on formally connecting a Bayesian framework to a l2-Potts functional. Behaviors and performance for the proposed piecewise constant denoising and regularization parameter tuning techniques are studied qualitatively and assessed quantitatively, and shown to compare favorably against those of a fully Bayesian hierarchical procedure, both in accuracy and in computational load.
Complete list of metadatas

Cited literature [52 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01887965
Contributor : Open Archive Toulouse Archive Ouverte (oatao) <>
Submitted on : Monday, October 8, 2018 - 11:00:22 AM
Last modification on : Thursday, September 24, 2020 - 2:02:06 PM
Long-term archiving on: : Wednesday, January 9, 2019 - 1:03:28 PM

File

frecon_19068..pdf
Files produced by the author(s)

Identifiers

Citation

Jordan Frecon, Nelly Pustelnik, Nicolas Dobigeon, Herwig Wendt, Patrice Abry. Bayesian selection for the l2-Potts model regularization parameter: 1D piecewise constant signal denoising. IEEE Transactions on Signal Processing, Institute of Electrical and Electronics Engineers, 2017, 65 (19), pp.5215-5224. ⟨10.1109/TSP.2017.2715000⟩. ⟨hal-01887965⟩

Share

Metrics

Record views

259

Files downloads

595