A hierarchical variational Bayesian approximation approach in acoustic imaging

Ning Chu 1, 2 Ali Mohammad-Djafari 2 Nicolas Gac 2 José Picheral 1
1 E3S - Supélec Sciences des Systèmes [Gif-sur-Yvette]
2 L2S - Laboratoire des signaux et systèmes
Abstract : Acoustic imaging is a powerful technique for acoustic source localization and power reconstruction from limited noisy measurements at microphone sensors. But it inevitably con-fronts a very ill-posed inverse problem which causes unexpected solution uncertainty. Recently, the Bayesian inference methods using sparse priors have been effectively investigated. In this paper, we propose to use a hierarchical variational Bayesian approximation for robust acoustic imaging. And we explore the Student-t priors with heavy tails to enforce source sparsity, and to model non-Gaussian noise respectively. Compared to conventional methods, the proposed approach can achieve the higher spatial resolution and wider dynamic range of source powers for real data from automo-bile wind tunnel.
Keywords : Student-t prior Variational Bayesian Approximation Non-Gaussian noise Acoustic imaging
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [12 references]  Display  Hide  Download
https://hal-supelec.archives-ouvertes.fr/hal-01103784
Contributor : Alexandra Siebert <>
Submitted on : Thursday, January 15, 2015 - 1:46:00 PM
Last modification on : Sunday, September 16, 2018 - 10:06:06 PM
Document(s) archivé(s) le : Thursday, April 16, 2015 - 10:40:51 AM

File

Maxent2014-CHU.pdf
Files produced by the author(s)

Identifiers

Collections

SUPELEC | SUP_LSS | SUP_SIGNAUX | SUP_SSE | SUP_E3S | L2S

Citation

Ning Chu, Ali Mohammad-Djafari, Nicolas Gac, José Picheral. A hierarchical variational Bayesian approximation approach in acoustic imaging. 34th International Workshop on Bayesian Inference and Maximun Entropy Methods in Science and Engineering (MaxEnt'14), Sep 2014, Amboise, France. pp.572 - 579, ⟨10.1063/1.4906024⟩. ⟨hal-01103784⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

440

Files downloads

162

Contact

support.ccsd.cnrs.fr
hal.support@ccsd.cnrs.fr
Logo CCSD