Saving phase: Injectivity and stability for phase retrieval

Afonso Bandeira, Jameson Cahill, Dustin G. Mixon, Aaron A. Nelson

Research output: Contribution to journalArticle

Abstract

Recent advances in convex optimization have led to new strides in the phase retrieval problem over finite-dimensional vector spaces. However, certain fundamental questions remain: What sorts of measurement vectors uniquely determine every signal up to a global phase factor, and how many are needed to do so? Furthermore, which measurement ensembles yield stability? This paper presents several results that address each of these questions. We begin by characterizing injectivity, and we identify that the complement property is indeed a necessary condition in the complex case. We then pose a conjecture that 4M-4 generic measurement vectors are both necessary and sufficient for injectivity in M dimensions, and we prove this conjecture in the special cases where M=2,3. Next, we shift our attention to stability, both in the worst and average cases. Here, we characterize worst-case stability in the real case by introducing a numerical version of the complement property. This new property bears some resemblance to the restricted isometry property of compressed sensing and can be used to derive a sharp lower Lipschitz bound on the intensity measurement mapping. Localized frames are shown to lack this property (suggesting instability), whereas Gaussian random measurements are shown to satisfy this property with high probability. We conclude by presenting results that use a stochastic noise model in both the real and complex cases, and we leverage Cramer-Rao lower bounds to identify stability with stronger versions of the injectivity characterizations.

Original languageEnglish (US)
Pages (from-to)106-125
Number of pages20
JournalApplied and Computational Harmonic Analysis
Volume37
Issue number1
DOIs
StatePublished - 2014

Fingerprint

Phase Retrieval
Injectivity
Complement
Compressed sensing
Convex optimization
Cramer-Rao Lower Bound
Vector spaces
Compressed Sensing
Convex Optimization
Isometry
Leverage
Sort
Vector space
Lipschitz
Ensemble
Sufficient
Necessary Conditions
Necessary

Keywords

  • Bilipschitz function
  • Cramer-Rao lower bound
  • Phase retrieval
  • Quantum mechanics

ASJC Scopus subject areas

  • Applied Mathematics

Cite this

Saving phase : Injectivity and stability for phase retrieval. / Bandeira, Afonso; Cahill, Jameson; Mixon, Dustin G.; Nelson, Aaron A.

In: Applied and Computational Harmonic Analysis, Vol. 37, No. 1, 2014, p. 106-125.

Research output: Contribution to journalArticle

Bandeira, Afonso ; Cahill, Jameson ; Mixon, Dustin G. ; Nelson, Aaron A. / Saving phase : Injectivity and stability for phase retrieval. In: Applied and Computational Harmonic Analysis. 2014 ; Vol. 37, No. 1. pp. 106-125.
@article{783d6fcd223b47d984b69a0cdf02a4da,
title = "Saving phase: Injectivity and stability for phase retrieval",
abstract = "Recent advances in convex optimization have led to new strides in the phase retrieval problem over finite-dimensional vector spaces. However, certain fundamental questions remain: What sorts of measurement vectors uniquely determine every signal up to a global phase factor, and how many are needed to do so? Furthermore, which measurement ensembles yield stability? This paper presents several results that address each of these questions. We begin by characterizing injectivity, and we identify that the complement property is indeed a necessary condition in the complex case. We then pose a conjecture that 4M-4 generic measurement vectors are both necessary and sufficient for injectivity in M dimensions, and we prove this conjecture in the special cases where M=2,3. Next, we shift our attention to stability, both in the worst and average cases. Here, we characterize worst-case stability in the real case by introducing a numerical version of the complement property. This new property bears some resemblance to the restricted isometry property of compressed sensing and can be used to derive a sharp lower Lipschitz bound on the intensity measurement mapping. Localized frames are shown to lack this property (suggesting instability), whereas Gaussian random measurements are shown to satisfy this property with high probability. We conclude by presenting results that use a stochastic noise model in both the real and complex cases, and we leverage Cramer-Rao lower bounds to identify stability with stronger versions of the injectivity characterizations.",
keywords = "Bilipschitz function, Cramer-Rao lower bound, Phase retrieval, Quantum mechanics",
author = "Afonso Bandeira and Jameson Cahill and Mixon, {Dustin G.} and Nelson, {Aaron A.}",
year = "2014",
doi = "10.1016/j.acha.2013.10.002",
language = "English (US)",
volume = "37",
pages = "106--125",
journal = "Applied and Computational Harmonic Analysis",
issn = "1063-5203",
publisher = "Academic Press Inc.",
number = "1",

}

TY - JOUR

T1 - Saving phase

T2 - Injectivity and stability for phase retrieval

AU - Bandeira, Afonso

AU - Cahill, Jameson

AU - Mixon, Dustin G.

AU - Nelson, Aaron A.

PY - 2014

Y1 - 2014

N2 - Recent advances in convex optimization have led to new strides in the phase retrieval problem over finite-dimensional vector spaces. However, certain fundamental questions remain: What sorts of measurement vectors uniquely determine every signal up to a global phase factor, and how many are needed to do so? Furthermore, which measurement ensembles yield stability? This paper presents several results that address each of these questions. We begin by characterizing injectivity, and we identify that the complement property is indeed a necessary condition in the complex case. We then pose a conjecture that 4M-4 generic measurement vectors are both necessary and sufficient for injectivity in M dimensions, and we prove this conjecture in the special cases where M=2,3. Next, we shift our attention to stability, both in the worst and average cases. Here, we characterize worst-case stability in the real case by introducing a numerical version of the complement property. This new property bears some resemblance to the restricted isometry property of compressed sensing and can be used to derive a sharp lower Lipschitz bound on the intensity measurement mapping. Localized frames are shown to lack this property (suggesting instability), whereas Gaussian random measurements are shown to satisfy this property with high probability. We conclude by presenting results that use a stochastic noise model in both the real and complex cases, and we leverage Cramer-Rao lower bounds to identify stability with stronger versions of the injectivity characterizations.

AB - Recent advances in convex optimization have led to new strides in the phase retrieval problem over finite-dimensional vector spaces. However, certain fundamental questions remain: What sorts of measurement vectors uniquely determine every signal up to a global phase factor, and how many are needed to do so? Furthermore, which measurement ensembles yield stability? This paper presents several results that address each of these questions. We begin by characterizing injectivity, and we identify that the complement property is indeed a necessary condition in the complex case. We then pose a conjecture that 4M-4 generic measurement vectors are both necessary and sufficient for injectivity in M dimensions, and we prove this conjecture in the special cases where M=2,3. Next, we shift our attention to stability, both in the worst and average cases. Here, we characterize worst-case stability in the real case by introducing a numerical version of the complement property. This new property bears some resemblance to the restricted isometry property of compressed sensing and can be used to derive a sharp lower Lipschitz bound on the intensity measurement mapping. Localized frames are shown to lack this property (suggesting instability), whereas Gaussian random measurements are shown to satisfy this property with high probability. We conclude by presenting results that use a stochastic noise model in both the real and complex cases, and we leverage Cramer-Rao lower bounds to identify stability with stronger versions of the injectivity characterizations.

KW - Bilipschitz function

KW - Cramer-Rao lower bound

KW - Phase retrieval

KW - Quantum mechanics

UR - http://www.scopus.com/inward/record.url?scp=84900549836&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84900549836&partnerID=8YFLogxK

U2 - 10.1016/j.acha.2013.10.002

DO - 10.1016/j.acha.2013.10.002

M3 - Article

VL - 37

SP - 106

EP - 125

JO - Applied and Computational Harmonic Analysis

JF - Applied and Computational Harmonic Analysis

SN - 1063-5203

IS - 1

ER -