Multi-stage FCM-based intensity inhomogeneity correction for MR brain image segmentation

László Szilágyi, Sándor M. Szilágyi, László Dávid, Z. Benyó

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)

Abstract

Intensity inhomogeneity or intensity non-uniformity (INU) is an undesired phenomenon that represents the main obstacle for MR image segmentation and registration methods. Various techniques have been proposed to eliminate or compensate the INU, most of which are embedded into clustering algorithms, and they generally have difficulties when INU reaches high amplitudes. This paper proposes a multiple stage fuzzy c-means (FCM) based algorithm for the estimation and compensation of INU, by modeling it as a slowly varying additive or multiplicative noise, supported by a pre-filtering technique for Gaussian and impulse noise elimination. The slowly varying behavior of the bias or gain field is assured by a smoothing filter that performs a context dependent averaging, based on a morphological criterion. The experiments using 2-D synthetic phantoms and real MR images show, that the proposed method provides accurate segmentation. The resulting segmentation and fuzzy membership values can serve as excellent support for 3-D registration and segmentation techniques.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages527-536
Number of pages10
Volume5164 LNCS
EditionPART 2
DOIs
Publication statusPublished - 2008
Event18th International Conference on Artificial Neural Networks, ICANN 2008 - Prague, Czech Republic
Duration: Sep 3 2008Sep 6 2008

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume5164 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other18th International Conference on Artificial Neural Networks, ICANN 2008
CountryCzech Republic
CityPrague
Period9/3/089/6/08

Fingerprint

Gaussian noise (electronic)
Impulse noise
Fuzzy C-means
Image registration
Image segmentation
Inhomogeneity
Clustering algorithms
Image Segmentation
Non-uniformity
Noise
Brain
Segmentation
Cluster Analysis
Experiments
Fuzzy Membership
Impulse Noise
Multiplicative Noise
Additive Noise
Gaussian Noise
Image Registration

Keywords

  • Context dependent filter
  • Fuzzy c-means algorithm
  • Image segmentation
  • Intensity inhomogeneity
  • Magnetic resonance imaging
  • Morphological operations

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Szilágyi, L., Szilágyi, S. M., Dávid, L., & Benyó, Z. (2008). Multi-stage FCM-based intensity inhomogeneity correction for MR brain image segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 2 ed., Vol. 5164 LNCS, pp. 527-536). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5164 LNCS, No. PART 2). https://doi.org/10.1007/978-3-540-87559-8_55

Multi-stage FCM-based intensity inhomogeneity correction for MR brain image segmentation. / Szilágyi, László; Szilágyi, Sándor M.; Dávid, László; Benyó, Z.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5164 LNCS PART 2. ed. 2008. p. 527-536 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5164 LNCS, No. PART 2).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Szilágyi, L, Szilágyi, SM, Dávid, L & Benyó, Z 2008, Multi-stage FCM-based intensity inhomogeneity correction for MR brain image segmentation. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 2 edn, vol. 5164 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 5164 LNCS, pp. 527-536, 18th International Conference on Artificial Neural Networks, ICANN 2008, Prague, Czech Republic, 9/3/08. https://doi.org/10.1007/978-3-540-87559-8_55
Szilágyi L, Szilágyi SM, Dávid L, Benyó Z. Multi-stage FCM-based intensity inhomogeneity correction for MR brain image segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 2 ed. Vol. 5164 LNCS. 2008. p. 527-536. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). https://doi.org/10.1007/978-3-540-87559-8_55
Szilágyi, László ; Szilágyi, Sándor M. ; Dávid, László ; Benyó, Z. / Multi-stage FCM-based intensity inhomogeneity correction for MR brain image segmentation. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5164 LNCS PART 2. ed. 2008. pp. 527-536 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).
@inproceedings{553bf35da2924fb683c006b3b50ce280,
title = "Multi-stage FCM-based intensity inhomogeneity correction for MR brain image segmentation",
abstract = "Intensity inhomogeneity or intensity non-uniformity (INU) is an undesired phenomenon that represents the main obstacle for MR image segmentation and registration methods. Various techniques have been proposed to eliminate or compensate the INU, most of which are embedded into clustering algorithms, and they generally have difficulties when INU reaches high amplitudes. This paper proposes a multiple stage fuzzy c-means (FCM) based algorithm for the estimation and compensation of INU, by modeling it as a slowly varying additive or multiplicative noise, supported by a pre-filtering technique for Gaussian and impulse noise elimination. The slowly varying behavior of the bias or gain field is assured by a smoothing filter that performs a context dependent averaging, based on a morphological criterion. The experiments using 2-D synthetic phantoms and real MR images show, that the proposed method provides accurate segmentation. The resulting segmentation and fuzzy membership values can serve as excellent support for 3-D registration and segmentation techniques.",
keywords = "Context dependent filter, Fuzzy c-means algorithm, Image segmentation, Intensity inhomogeneity, Magnetic resonance imaging, Morphological operations",
author = "L{\'a}szl{\'o} Szil{\'a}gyi and Szil{\'a}gyi, {S{\'a}ndor M.} and L{\'a}szl{\'o} D{\'a}vid and Z. Beny{\'o}",
year = "2008",
doi = "10.1007/978-3-540-87559-8_55",
language = "English",
isbn = "3540875581",
volume = "5164 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 2",
pages = "527--536",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 2",

}

TY - GEN

T1 - Multi-stage FCM-based intensity inhomogeneity correction for MR brain image segmentation

AU - Szilágyi, László

AU - Szilágyi, Sándor M.

AU - Dávid, László

AU - Benyó, Z.

PY - 2008

Y1 - 2008

N2 - Intensity inhomogeneity or intensity non-uniformity (INU) is an undesired phenomenon that represents the main obstacle for MR image segmentation and registration methods. Various techniques have been proposed to eliminate or compensate the INU, most of which are embedded into clustering algorithms, and they generally have difficulties when INU reaches high amplitudes. This paper proposes a multiple stage fuzzy c-means (FCM) based algorithm for the estimation and compensation of INU, by modeling it as a slowly varying additive or multiplicative noise, supported by a pre-filtering technique for Gaussian and impulse noise elimination. The slowly varying behavior of the bias or gain field is assured by a smoothing filter that performs a context dependent averaging, based on a morphological criterion. The experiments using 2-D synthetic phantoms and real MR images show, that the proposed method provides accurate segmentation. The resulting segmentation and fuzzy membership values can serve as excellent support for 3-D registration and segmentation techniques.

AB - Intensity inhomogeneity or intensity non-uniformity (INU) is an undesired phenomenon that represents the main obstacle for MR image segmentation and registration methods. Various techniques have been proposed to eliminate or compensate the INU, most of which are embedded into clustering algorithms, and they generally have difficulties when INU reaches high amplitudes. This paper proposes a multiple stage fuzzy c-means (FCM) based algorithm for the estimation and compensation of INU, by modeling it as a slowly varying additive or multiplicative noise, supported by a pre-filtering technique for Gaussian and impulse noise elimination. The slowly varying behavior of the bias or gain field is assured by a smoothing filter that performs a context dependent averaging, based on a morphological criterion. The experiments using 2-D synthetic phantoms and real MR images show, that the proposed method provides accurate segmentation. The resulting segmentation and fuzzy membership values can serve as excellent support for 3-D registration and segmentation techniques.

KW - Context dependent filter

KW - Fuzzy c-means algorithm

KW - Image segmentation

KW - Intensity inhomogeneity

KW - Magnetic resonance imaging

KW - Morphological operations

UR - http://www.scopus.com/inward/record.url?scp=51849157762&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=51849157762&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-87559-8_55

DO - 10.1007/978-3-540-87559-8_55

M3 - Conference contribution

AN - SCOPUS:51849157762

SN - 3540875581

SN - 9783540875581

VL - 5164 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 527

EP - 536

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -