Edge detection from truncated Fourier data using spectral mollifiers

Douglas Cochran, Anne Gelb, Yang Wang

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Edge detection from a finite number of Fourier coefficients is challenging as it requires extracting local information from global data. The problem is exacerbated when the input data is noisy since accurate high frequency information is critical for detecting edges. The noise furthermore increases oscillations in the Fourier reconstruction of piecewise smooth functions, especially near the discontinuities. The edge detection method in Gelb and Tadmor (Appl Comput Harmon Anal 7:101-135, 1999, SIAM J Numer Anal 38(4):1389-1408, 2000) introduced the idea of "concentration kernels" as a way of converging to the singular support of a piecewise smooth function. The kernels used there, however, and subsequent modifications to reduce the impact of noise, were generally oscillatory, and as a result oscillations were always prevalent in the neighborhoods of the jump discontinuities. This paper revisits concentration kernels, but insists on uniform convergence to the "sharp peaks" of the function, that is, the edge detection method converges to zero away from the jumps without introducing new oscillations near them. We show that this is achievable via an admissible class of spectral mollifiers. Our method furthermore suppresses the oscillations caused by added noise.

Original languageEnglish (US)
Pages (from-to)737-762
Number of pages26
JournalAdvances in Computational Mathematics
Volume38
Issue number4
DOIs
StatePublished - 2013

Fingerprint

Edge Detection
Edge detection
Oscillation
Piecewise Smooth Functions
kernel
Discontinuity
Jump
Fourier coefficients
Uniform convergence
Converge
Zero

Keywords

  • Edge detection
  • Gibbs phenomenon
  • Poisson summation formula
  • Spectral mollifier
  • Truncated Fourier data

ASJC Scopus subject areas

  • Applied Mathematics
  • Computational Mathematics

Cite this

Edge detection from truncated Fourier data using spectral mollifiers. / Cochran, Douglas; Gelb, Anne; Wang, Yang.

In: Advances in Computational Mathematics, Vol. 38, No. 4, 2013, p. 737-762.

Research output: Contribution to journalArticle

Cochran, Douglas ; Gelb, Anne ; Wang, Yang. / Edge detection from truncated Fourier data using spectral mollifiers. In: Advances in Computational Mathematics. 2013 ; Vol. 38, No. 4. pp. 737-762.
@article{2f08f9fdee44443588ef1ecf164bdaf1,
title = "Edge detection from truncated Fourier data using spectral mollifiers",
abstract = "Edge detection from a finite number of Fourier coefficients is challenging as it requires extracting local information from global data. The problem is exacerbated when the input data is noisy since accurate high frequency information is critical for detecting edges. The noise furthermore increases oscillations in the Fourier reconstruction of piecewise smooth functions, especially near the discontinuities. The edge detection method in Gelb and Tadmor (Appl Comput Harmon Anal 7:101-135, 1999, SIAM J Numer Anal 38(4):1389-1408, 2000) introduced the idea of {"}concentration kernels{"} as a way of converging to the singular support of a piecewise smooth function. The kernels used there, however, and subsequent modifications to reduce the impact of noise, were generally oscillatory, and as a result oscillations were always prevalent in the neighborhoods of the jump discontinuities. This paper revisits concentration kernels, but insists on uniform convergence to the {"}sharp peaks{"} of the function, that is, the edge detection method converges to zero away from the jumps without introducing new oscillations near them. We show that this is achievable via an admissible class of spectral mollifiers. Our method furthermore suppresses the oscillations caused by added noise.",
keywords = "Edge detection, Gibbs phenomenon, Poisson summation formula, Spectral mollifier, Truncated Fourier data",
author = "Douglas Cochran and Anne Gelb and Yang Wang",
year = "2013",
doi = "10.1007/s10444-011-9258-4",
language = "English (US)",
volume = "38",
pages = "737--762",
journal = "Advances in Computational Mathematics",
issn = "1019-7168",
publisher = "Springer Netherlands",
number = "4",

}

TY - JOUR

T1 - Edge detection from truncated Fourier data using spectral mollifiers

AU - Cochran, Douglas

AU - Gelb, Anne

AU - Wang, Yang

PY - 2013

Y1 - 2013

N2 - Edge detection from a finite number of Fourier coefficients is challenging as it requires extracting local information from global data. The problem is exacerbated when the input data is noisy since accurate high frequency information is critical for detecting edges. The noise furthermore increases oscillations in the Fourier reconstruction of piecewise smooth functions, especially near the discontinuities. The edge detection method in Gelb and Tadmor (Appl Comput Harmon Anal 7:101-135, 1999, SIAM J Numer Anal 38(4):1389-1408, 2000) introduced the idea of "concentration kernels" as a way of converging to the singular support of a piecewise smooth function. The kernels used there, however, and subsequent modifications to reduce the impact of noise, were generally oscillatory, and as a result oscillations were always prevalent in the neighborhoods of the jump discontinuities. This paper revisits concentration kernels, but insists on uniform convergence to the "sharp peaks" of the function, that is, the edge detection method converges to zero away from the jumps without introducing new oscillations near them. We show that this is achievable via an admissible class of spectral mollifiers. Our method furthermore suppresses the oscillations caused by added noise.

AB - Edge detection from a finite number of Fourier coefficients is challenging as it requires extracting local information from global data. The problem is exacerbated when the input data is noisy since accurate high frequency information is critical for detecting edges. The noise furthermore increases oscillations in the Fourier reconstruction of piecewise smooth functions, especially near the discontinuities. The edge detection method in Gelb and Tadmor (Appl Comput Harmon Anal 7:101-135, 1999, SIAM J Numer Anal 38(4):1389-1408, 2000) introduced the idea of "concentration kernels" as a way of converging to the singular support of a piecewise smooth function. The kernels used there, however, and subsequent modifications to reduce the impact of noise, were generally oscillatory, and as a result oscillations were always prevalent in the neighborhoods of the jump discontinuities. This paper revisits concentration kernels, but insists on uniform convergence to the "sharp peaks" of the function, that is, the edge detection method converges to zero away from the jumps without introducing new oscillations near them. We show that this is achievable via an admissible class of spectral mollifiers. Our method furthermore suppresses the oscillations caused by added noise.

KW - Edge detection

KW - Gibbs phenomenon

KW - Poisson summation formula

KW - Spectral mollifier

KW - Truncated Fourier data

UR - http://www.scopus.com/inward/record.url?scp=84876424520&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84876424520&partnerID=8YFLogxK

U2 - 10.1007/s10444-011-9258-4

DO - 10.1007/s10444-011-9258-4

M3 - Article

AN - SCOPUS:84876424520

VL - 38

SP - 737

EP - 762

JO - Advances in Computational Mathematics

JF - Advances in Computational Mathematics

SN - 1019-7168

IS - 4

ER -