Demixing sines and spikes

Robust spectral super-resolution in the presence of outliers

Carlos Fernandez-Granda, Gongguo Tang, Xiaodong Wang, Le Zheng

Research output: Contribution to journalArticle

Abstract

We consider the problem of super-resolving the line spectrum of a multisinusoidal signal from a finite number of samples, some of which may be completely corrupted. Measurements of this form can be modeled as an additive mixture of a sinusoidal and a sparse component. We propose to demix the two components and super-resolve the spectrum of the multisinusoidal signal by solving a convex program. Our main theoretical result is that-up to logarithmic factors-this approach is guaranteed to be successful with high probability for a number of spectral lines that is linear in the number of measurements, even if a constant fraction of the data are outliers. The result holds under the assumption that the phases of the sinusoidal and sparse components are random and the line spectrum satisfies a minimum-separation condition. We show that the method can be implemented via semi-definite programming, and explain how to adapt it in the presence of dense perturbations as well as exploring its connection to atomic-norm denoising. In addition, we propose a fast greedy demixing method that provides good empirical results when coupled with a local non-convex-optimization step.

Original languageEnglish (US)
Pages (from-to)105-168
Number of pages64
JournalInformation and Inference
Volume7
Issue number1
DOIs
StatePublished - Jan 1 2018

Fingerprint

Super-resolution
Spike
Outlier
Line
Convex Program
Local Optimization
Nonconvex Optimization
Semidefinite Programming
Denoising
Resolve
Logarithmic
Perturbation
Norm

Keywords

  • Atomic norm
  • Continuous dictionary
  • Convex optimization
  • Greedy methods
  • Line spectra estimation
  • Outliers
  • Semi-definite programming
  • Sparse recovery
  • Super-resolution

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Analysis
  • Applied Mathematics
  • Statistics and Probability
  • Numerical Analysis

Cite this

Demixing sines and spikes : Robust spectral super-resolution in the presence of outliers. / Fernandez-Granda, Carlos; Tang, Gongguo; Wang, Xiaodong; Zheng, Le.

In: Information and Inference, Vol. 7, No. 1, 01.01.2018, p. 105-168.

Research output: Contribution to journalArticle

@article{afc4c94669f84413abb51ac9155cbb55,
title = "Demixing sines and spikes: Robust spectral super-resolution in the presence of outliers",
abstract = "We consider the problem of super-resolving the line spectrum of a multisinusoidal signal from a finite number of samples, some of which may be completely corrupted. Measurements of this form can be modeled as an additive mixture of a sinusoidal and a sparse component. We propose to demix the two components and super-resolve the spectrum of the multisinusoidal signal by solving a convex program. Our main theoretical result is that-up to logarithmic factors-this approach is guaranteed to be successful with high probability for a number of spectral lines that is linear in the number of measurements, even if a constant fraction of the data are outliers. The result holds under the assumption that the phases of the sinusoidal and sparse components are random and the line spectrum satisfies a minimum-separation condition. We show that the method can be implemented via semi-definite programming, and explain how to adapt it in the presence of dense perturbations as well as exploring its connection to atomic-norm denoising. In addition, we propose a fast greedy demixing method that provides good empirical results when coupled with a local non-convex-optimization step.",
keywords = "Atomic norm, Continuous dictionary, Convex optimization, Greedy methods, Line spectra estimation, Outliers, Semi-definite programming, Sparse recovery, Super-resolution",
author = "Carlos Fernandez-Granda and Gongguo Tang and Xiaodong Wang and Le Zheng",
year = "2018",
month = "1",
day = "1",
doi = "10.1093/imaiai/iax005",
language = "English (US)",
volume = "7",
pages = "105--168",
journal = "Information and Inference",
issn = "2049-8772",
number = "1",

}

TY - JOUR

T1 - Demixing sines and spikes

T2 - Robust spectral super-resolution in the presence of outliers

AU - Fernandez-Granda, Carlos

AU - Tang, Gongguo

AU - Wang, Xiaodong

AU - Zheng, Le

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We consider the problem of super-resolving the line spectrum of a multisinusoidal signal from a finite number of samples, some of which may be completely corrupted. Measurements of this form can be modeled as an additive mixture of a sinusoidal and a sparse component. We propose to demix the two components and super-resolve the spectrum of the multisinusoidal signal by solving a convex program. Our main theoretical result is that-up to logarithmic factors-this approach is guaranteed to be successful with high probability for a number of spectral lines that is linear in the number of measurements, even if a constant fraction of the data are outliers. The result holds under the assumption that the phases of the sinusoidal and sparse components are random and the line spectrum satisfies a minimum-separation condition. We show that the method can be implemented via semi-definite programming, and explain how to adapt it in the presence of dense perturbations as well as exploring its connection to atomic-norm denoising. In addition, we propose a fast greedy demixing method that provides good empirical results when coupled with a local non-convex-optimization step.

AB - We consider the problem of super-resolving the line spectrum of a multisinusoidal signal from a finite number of samples, some of which may be completely corrupted. Measurements of this form can be modeled as an additive mixture of a sinusoidal and a sparse component. We propose to demix the two components and super-resolve the spectrum of the multisinusoidal signal by solving a convex program. Our main theoretical result is that-up to logarithmic factors-this approach is guaranteed to be successful with high probability for a number of spectral lines that is linear in the number of measurements, even if a constant fraction of the data are outliers. The result holds under the assumption that the phases of the sinusoidal and sparse components are random and the line spectrum satisfies a minimum-separation condition. We show that the method can be implemented via semi-definite programming, and explain how to adapt it in the presence of dense perturbations as well as exploring its connection to atomic-norm denoising. In addition, we propose a fast greedy demixing method that provides good empirical results when coupled with a local non-convex-optimization step.

KW - Atomic norm

KW - Continuous dictionary

KW - Convex optimization

KW - Greedy methods

KW - Line spectra estimation

KW - Outliers

KW - Semi-definite programming

KW - Sparse recovery

KW - Super-resolution

UR - http://www.scopus.com/inward/record.url?scp=85071674056&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071674056&partnerID=8YFLogxK

U2 - 10.1093/imaiai/iax005

DO - 10.1093/imaiai/iax005

M3 - Article

VL - 7

SP - 105

EP - 168

JO - Information and Inference

JF - Information and Inference

SN - 2049-8772

IS - 1

ER -