Compressive phase retrieval via generalized approximate message passing

Philip Schniter, Sundeep Rangan

Research output: Contribution to journalArticle

Abstract

In phase retrieval, the goal is to recover a signal x ε CN from the magnitudes of linear measurements Ax εM. While recent theory has established that M\approx 4N intensity measurements are necessary and sufficient to recover generic x, there is great interest in reducing the number of measurements through the exploitation of sparse x, which is known as compressive phase retrieval. In this work, we detail a novel, probabilistic approach to compressive phase retrieval based on the generalized approximate message passing (GAMP) algorithm. We then present a numerical study of the proposed PR-GAMP algorithm, demonstrating its excellent phase-transition behavior, robustness to noise, and runtime. Our experiments suggest that approximately M ≥ 2K\log2(N/K) intensity measurements suffice to recover K-sparse Bernoulli-Gaussian signals for A with i.i.d Gaussian entries and K\ll N. Meanwhile, when recovering a 6678-sparse 65536-pixel grayscale image from 32768 randomly masked and blurred Fourier intensity measurements at 30 dB measurement SNR, PR-GAMP achieved an output SNR of no less than 28 dB in all of 100 random trials, with a median runtime of only 7.3 seconds. Compared to the recently proposed CPRL, sparse-Fienup, and GESPAR algorithms, our experiments suggest that PR-GAMP has a superior phase transition and orders-of-magnitude faster runtimes as the sparsity and problem dimensions increase.

Original languageEnglish (US)
Article number6998861
Pages (from-to)1043-1055
Number of pages13
JournalIEEE Transactions on Signal Processing
Volume63
Issue number4
DOIs
StatePublished - Feb 15 2015

Fingerprint

Message passing
Phase transitions
Pixels
Experiments

Keywords

  • Belief propagation
  • compressed sensing
  • estimation
  • phase retrieval

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Signal Processing

Cite this

Compressive phase retrieval via generalized approximate message passing. / Schniter, Philip; Rangan, Sundeep.

In: IEEE Transactions on Signal Processing, Vol. 63, No. 4, 6998861, 15.02.2015, p. 1043-1055.

Research output: Contribution to journalArticle

@article{e5020843a7554eb19e018ce58b999292,
title = "Compressive phase retrieval via generalized approximate message passing",
abstract = "In phase retrieval, the goal is to recover a signal x ε CN from the magnitudes of linear measurements Ax εM. While recent theory has established that M\approx 4N intensity measurements are necessary and sufficient to recover generic x, there is great interest in reducing the number of measurements through the exploitation of sparse x, which is known as compressive phase retrieval. In this work, we detail a novel, probabilistic approach to compressive phase retrieval based on the generalized approximate message passing (GAMP) algorithm. We then present a numerical study of the proposed PR-GAMP algorithm, demonstrating its excellent phase-transition behavior, robustness to noise, and runtime. Our experiments suggest that approximately M ≥ 2K\log2(N/K) intensity measurements suffice to recover K-sparse Bernoulli-Gaussian signals for A with i.i.d Gaussian entries and K\ll N. Meanwhile, when recovering a 6678-sparse 65536-pixel grayscale image from 32768 randomly masked and blurred Fourier intensity measurements at 30 dB measurement SNR, PR-GAMP achieved an output SNR of no less than 28 dB in all of 100 random trials, with a median runtime of only 7.3 seconds. Compared to the recently proposed CPRL, sparse-Fienup, and GESPAR algorithms, our experiments suggest that PR-GAMP has a superior phase transition and orders-of-magnitude faster runtimes as the sparsity and problem dimensions increase.",
keywords = "Belief propagation, compressed sensing, estimation, phase retrieval",
author = "Philip Schniter and Sundeep Rangan",
year = "2015",
month = "2",
day = "15",
doi = "10.1109/TSP.2014.2386294",
language = "English (US)",
volume = "63",
pages = "1043--1055",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "4",

}

TY - JOUR

T1 - Compressive phase retrieval via generalized approximate message passing

AU - Schniter, Philip

AU - Rangan, Sundeep

PY - 2015/2/15

Y1 - 2015/2/15

N2 - In phase retrieval, the goal is to recover a signal x ε CN from the magnitudes of linear measurements Ax εM. While recent theory has established that M\approx 4N intensity measurements are necessary and sufficient to recover generic x, there is great interest in reducing the number of measurements through the exploitation of sparse x, which is known as compressive phase retrieval. In this work, we detail a novel, probabilistic approach to compressive phase retrieval based on the generalized approximate message passing (GAMP) algorithm. We then present a numerical study of the proposed PR-GAMP algorithm, demonstrating its excellent phase-transition behavior, robustness to noise, and runtime. Our experiments suggest that approximately M ≥ 2K\log2(N/K) intensity measurements suffice to recover K-sparse Bernoulli-Gaussian signals for A with i.i.d Gaussian entries and K\ll N. Meanwhile, when recovering a 6678-sparse 65536-pixel grayscale image from 32768 randomly masked and blurred Fourier intensity measurements at 30 dB measurement SNR, PR-GAMP achieved an output SNR of no less than 28 dB in all of 100 random trials, with a median runtime of only 7.3 seconds. Compared to the recently proposed CPRL, sparse-Fienup, and GESPAR algorithms, our experiments suggest that PR-GAMP has a superior phase transition and orders-of-magnitude faster runtimes as the sparsity and problem dimensions increase.

AB - In phase retrieval, the goal is to recover a signal x ε CN from the magnitudes of linear measurements Ax εM. While recent theory has established that M\approx 4N intensity measurements are necessary and sufficient to recover generic x, there is great interest in reducing the number of measurements through the exploitation of sparse x, which is known as compressive phase retrieval. In this work, we detail a novel, probabilistic approach to compressive phase retrieval based on the generalized approximate message passing (GAMP) algorithm. We then present a numerical study of the proposed PR-GAMP algorithm, demonstrating its excellent phase-transition behavior, robustness to noise, and runtime. Our experiments suggest that approximately M ≥ 2K\log2(N/K) intensity measurements suffice to recover K-sparse Bernoulli-Gaussian signals for A with i.i.d Gaussian entries and K\ll N. Meanwhile, when recovering a 6678-sparse 65536-pixel grayscale image from 32768 randomly masked and blurred Fourier intensity measurements at 30 dB measurement SNR, PR-GAMP achieved an output SNR of no less than 28 dB in all of 100 random trials, with a median runtime of only 7.3 seconds. Compared to the recently proposed CPRL, sparse-Fienup, and GESPAR algorithms, our experiments suggest that PR-GAMP has a superior phase transition and orders-of-magnitude faster runtimes as the sparsity and problem dimensions increase.

KW - Belief propagation

KW - compressed sensing

KW - estimation

KW - phase retrieval

UR - http://www.scopus.com/inward/record.url?scp=84922369056&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84922369056&partnerID=8YFLogxK

U2 - 10.1109/TSP.2014.2386294

DO - 10.1109/TSP.2014.2386294

M3 - Article

VL - 63

SP - 1043

EP - 1055

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 4

M1 - 6998861

ER -