Generalised Approximate Message Passing for Non-I.I.D. Sparse Signals

Christian Schou Oxvig; Thomas Arildsen

Generalised Approximate Message Passing for Non-I.I.D. Sparse Signals

Christian Schou Oxvig, Thomas Arildsen

Research output: Contribution to conference without publisher/journal › Paper without publisher/journal › Research › peer-review

119 Downloads (Pure)

Abstract

Generalised approximate message passing (GAMP) is an approximate Bayesian estimation algorithm for signals observed through a linear transform with a possibly non-linear subsequent measurement model. By leveraging prior information about the observed signal, such as sparsity in a known dictionary, GAMP can for example reconstruct signals from under-determined measurements -- known as compressed sensing. In the sparse signal setting, most existing signal priors for GAMP assume the input signal to have i.i.d. entries. Here we present sparse signal priors for GAMP to estimate non-i.d.d. signals through a non-uniform weighting of the input prior, for example allowing GAMP to support model-based compressed sensing.

Original language	English
Publication date	22 Nov 2018
Number of pages	3
Publication status	Published - 22 Nov 2018
Event	international Traveling Workshop on Interactions between low-complexity data models and Sensing Techniques - Centre International de Rencontres Mathématiques, Marseille, France Duration: 21 Nov 2018 → 23 Nov 2018 Conference number: 4 https://sites.google.com/view/itwist18

Workshop

Workshop	international Traveling Workshop on Interactions between low-complexity data models and Sensing Techniques
Number	4
Location	Centre International de Rencontres Mathématiques
Country/Territory	France
City	Marseille
Period	21/11/2018 → 23/11/2018
Internet address	https://sites.google.com/view/itwist18

Keywords

compressed sensing
signal processing
estimation theory

Access to Document

wgamp-itwist18Accepted author manuscript, 633 KBLicence: CC BY 4.0

https://arxiv.org/abs/1812.00909Licence: CC BY 4.0

AUB Link

Search for the material in Aalborg University Library's search engine

Cite this

@conference{3c927b25be864633b81c177ae66e9b95,

title = "Generalised Approximate Message Passing for Non-I.I.D. Sparse Signals",

abstract = "Generalised approximate message passing (GAMP) is an approximate Bayesian estimation algorithm for signals observed through a linear transform with a possibly non-linear subsequent measurement model. By leveraging prior information about the observed signal, such as sparsity in a known dictionary, GAMP can for example reconstruct signals from under-determined measurements -- known as compressed sensing. In the sparse signal setting, most existing signal priors for GAMP assume the input signal to have i.i.d. entries. Here we present sparse signal priors for GAMP to estimate non-i.d.d. signals through a non-uniform weighting of the input prior, for example allowing GAMP to support model-based compressed sensing.",

keywords = "compressed sensing, signal processing, estimation theory",

author = "Oxvig, {Christian Schou} and Thomas Arildsen",

year = "2018",

month = nov,

day = "22",

language = "English",

note = "international Traveling Workshop on Interactions between low-complexity data models and Sensing Techniques ; Conference date: 21-11-2018 Through 23-11-2018",

url = "https://sites.google.com/view/itwist18",

}

Generalised Approximate Message Passing for Non-I.I.D. Sparse Signals. / Oxvig, Christian Schou; Arildsen, Thomas.
2018. Paper presented at international Traveling Workshop on Interactions between low-complexity data models and Sensing Techniques, Marseille, France.

Research output: Contribution to conference without publisher/journal › Paper without publisher/journal › Research › peer-review

TY - CONF

T1 - Generalised Approximate Message Passing for Non-I.I.D. Sparse Signals

AU - Oxvig, Christian Schou

AU - Arildsen, Thomas

N1 - Conference code: 4

PY - 2018/11/22

Y1 - 2018/11/22

N2 - Generalised approximate message passing (GAMP) is an approximate Bayesian estimation algorithm for signals observed through a linear transform with a possibly non-linear subsequent measurement model. By leveraging prior information about the observed signal, such as sparsity in a known dictionary, GAMP can for example reconstruct signals from under-determined measurements -- known as compressed sensing. In the sparse signal setting, most existing signal priors for GAMP assume the input signal to have i.i.d. entries. Here we present sparse signal priors for GAMP to estimate non-i.d.d. signals through a non-uniform weighting of the input prior, for example allowing GAMP to support model-based compressed sensing.

AB - Generalised approximate message passing (GAMP) is an approximate Bayesian estimation algorithm for signals observed through a linear transform with a possibly non-linear subsequent measurement model. By leveraging prior information about the observed signal, such as sparsity in a known dictionary, GAMP can for example reconstruct signals from under-determined measurements -- known as compressed sensing. In the sparse signal setting, most existing signal priors for GAMP assume the input signal to have i.i.d. entries. Here we present sparse signal priors for GAMP to estimate non-i.d.d. signals through a non-uniform weighting of the input prior, for example allowing GAMP to support model-based compressed sensing.

KW - compressed sensing

KW - signal processing

KW - estimation theory

UR - https://doi.org/10.5281/zenodo.1409655

UR - https://docs.google.com/spreadsheets/d/e/2PACX-1vSWjLALqTnovorQQmhKyGAoPOXHaESJWMowskOYz1vNdo1ZbgW0jV6giu9yVxpduAAB3Bl0HE-QbYq_/pubhtml?gid=1304672187&single=true&widget=false&headers=false&chrome=false

M3 - Paper without publisher/journal

T2 - international Traveling Workshop on Interactions between low-complexity data models and Sensing Techniques

Y2 - 21 November 2018 through 23 November 2018

ER -

Generalised Approximate Message Passing for Non-I.I.D. Sparse Signals

Abstract

Workshop

Keywords

Access to Document

AUB Link

Other files and links

Fingerprint

FastAFM: Enabling Fast Image Acquisition for Atomic Force Microscopy using Compressed Sensing