Edge detection of piecewise smooth functions from undersampled Fourier data using variance signatures

Dennis Denker, Anne Gelb

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Detecting edges in piecewise smooth images is important in a variety of applications. In some cases, such as magnetic resonance imaging, data are collected as Fourier samples, and are often oversampled at low frequencies but more sparsely acquired in the high frequency range. This undersampling causes Fourier-based edge detection algorithms to yield false positives in smooth regions. The problem is compounded when the data are noisy. In this paper, we demonstrate that even when the given data are noisy and undersampled, the edge response behavior obtained using certain Fourier-based edge detection methods can be accurately characterized. In particular, the responses of different high order methods vary little, both in regions away from the discontinuities and at true edges. By contrast, the variance is large in cells neighboring each true edge. By locating these "double peaks" in variance, it is possible to determine edges from noisy and undersampled Fourier data. Alternative methods of generating and refining variance data are also explored. Specifically, a method based on regularized image reconstruction that leverages the sparsity of edges and image smoothness between these edges is effective at eliminating false positives.

Original languageEnglish (US)
Pages (from-to)A559-A592
JournalSIAM Journal on Scientific Computing
Volume39
Issue number2
DOIs
StatePublished - Jan 1 2017
Externally publishedYes

Fingerprint

Piecewise Smooth Functions
Edge Detection
Edge detection
Signature
Magnetic resonance
Image reconstruction
Refining
False Positive
Imaging techniques
High-order Methods
Magnetic Resonance Imaging
Image Reconstruction
Sparsity
Leverage
Low Frequency
Smoothness
Discontinuity
Vary
Alternatives
Cell

Keywords

  • Edge detection
  • Fourier data
  • L regularization
  • Undersampling
  • Variance

ASJC Scopus subject areas

  • Computational Mathematics
  • Applied Mathematics

Cite this

Edge detection of piecewise smooth functions from undersampled Fourier data using variance signatures. / Denker, Dennis; Gelb, Anne.

In: SIAM Journal on Scientific Computing, Vol. 39, No. 2, 01.01.2017, p. A559-A592.

Research output: Contribution to journalArticle

@article{752ae235fafd4984adffbe169f8f867c,
title = "Edge detection of piecewise smooth functions from undersampled Fourier data using variance signatures",
abstract = "Detecting edges in piecewise smooth images is important in a variety of applications. In some cases, such as magnetic resonance imaging, data are collected as Fourier samples, and are often oversampled at low frequencies but more sparsely acquired in the high frequency range. This undersampling causes Fourier-based edge detection algorithms to yield false positives in smooth regions. The problem is compounded when the data are noisy. In this paper, we demonstrate that even when the given data are noisy and undersampled, the edge response behavior obtained using certain Fourier-based edge detection methods can be accurately characterized. In particular, the responses of different high order methods vary little, both in regions away from the discontinuities and at true edges. By contrast, the variance is large in cells neighboring each true edge. By locating these {"}double peaks{"} in variance, it is possible to determine edges from noisy and undersampled Fourier data. Alternative methods of generating and refining variance data are also explored. Specifically, a method based on regularized image reconstruction that leverages the sparsity of edges and image smoothness between these edges is effective at eliminating false positives.",
keywords = "Edge detection, Fourier data, L regularization, Undersampling, Variance",
author = "Dennis Denker and Anne Gelb",
year = "2017",
month = "1",
day = "1",
doi = "10.1137/16M1068013",
language = "English (US)",
volume = "39",
pages = "A559--A592",
journal = "SIAM Journal of Scientific Computing",
issn = "1064-8275",
publisher = "Society for Industrial and Applied Mathematics Publications",
number = "2",

}

TY - JOUR

T1 - Edge detection of piecewise smooth functions from undersampled Fourier data using variance signatures

AU - Denker, Dennis

AU - Gelb, Anne

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Detecting edges in piecewise smooth images is important in a variety of applications. In some cases, such as magnetic resonance imaging, data are collected as Fourier samples, and are often oversampled at low frequencies but more sparsely acquired in the high frequency range. This undersampling causes Fourier-based edge detection algorithms to yield false positives in smooth regions. The problem is compounded when the data are noisy. In this paper, we demonstrate that even when the given data are noisy and undersampled, the edge response behavior obtained using certain Fourier-based edge detection methods can be accurately characterized. In particular, the responses of different high order methods vary little, both in regions away from the discontinuities and at true edges. By contrast, the variance is large in cells neighboring each true edge. By locating these "double peaks" in variance, it is possible to determine edges from noisy and undersampled Fourier data. Alternative methods of generating and refining variance data are also explored. Specifically, a method based on regularized image reconstruction that leverages the sparsity of edges and image smoothness between these edges is effective at eliminating false positives.

AB - Detecting edges in piecewise smooth images is important in a variety of applications. In some cases, such as magnetic resonance imaging, data are collected as Fourier samples, and are often oversampled at low frequencies but more sparsely acquired in the high frequency range. This undersampling causes Fourier-based edge detection algorithms to yield false positives in smooth regions. The problem is compounded when the data are noisy. In this paper, we demonstrate that even when the given data are noisy and undersampled, the edge response behavior obtained using certain Fourier-based edge detection methods can be accurately characterized. In particular, the responses of different high order methods vary little, both in regions away from the discontinuities and at true edges. By contrast, the variance is large in cells neighboring each true edge. By locating these "double peaks" in variance, it is possible to determine edges from noisy and undersampled Fourier data. Alternative methods of generating and refining variance data are also explored. Specifically, a method based on regularized image reconstruction that leverages the sparsity of edges and image smoothness between these edges is effective at eliminating false positives.

KW - Edge detection

KW - Fourier data

KW - L regularization

KW - Undersampling

KW - Variance

UR - http://www.scopus.com/inward/record.url?scp=85018323108&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85018323108&partnerID=8YFLogxK

U2 - 10.1137/16M1068013

DO - 10.1137/16M1068013

M3 - Article

AN - SCOPUS:85018323108

VL - 39

SP - A559-A592

JO - SIAM Journal of Scientific Computing

JF - SIAM Journal of Scientific Computing

SN - 1064-8275

IS - 2

ER -