Multi-class brain segmentation using atlas propagation and EM-based refinement

Christian Ledig, Robin Wolz, Paul Aljabar, Jyrki Lötjönen, Rolf Heckemann, Alex Hammers, Daniel Rueckert

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

21 Citations (Scopus)

Abstract

In recent years, multi-atlas segmentation has emerged as one of the most accurate techniques for the segmentation of brain magnetic resonance (MR) images, especially when combined with intensity-based refinement techniques such as graph-cut or expectation-maximization (EM) optimization. However, most of the work so far has focused on intensity-based refinement strategies for individual anatomical structures such as the hippocampus. In this work we extend a previously proposed framework for labeling whole brain scans by incorporating a global and stationary Markov random field that ensures the consistency of the neighbourhood relations between structures with an a-priori defined model. In particular we improve the segmentation result of a locally weighted multi-atlas fusion method for 41 different structures simultaneously by applying a subsequent EM optimization step. We evaluate the proposed approach on 30 manually annotated brain MR images and observe an improvement of label overlaps to a manual reference by up to 6%. We also achieved a considerably improved group separation when the proposed segmentation framework is applied to a volumetric analysis of 404 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort.
Original languageEnglish
Title of host publication2012 9th IEEE International Symposium on Biomedical Imaging (ISBI)
PublisherIEEE Institute of Electrical and Electronic Engineers
Pages896-899
ISBN (Electronic)978-1-4577-1858-8, 978-1-4577-1856-4
ISBN (Print)978-1-4577-1857-1
DOIs
Publication statusPublished - 2012
MoE publication typeA4 Article in a conference publication
Event9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012 - Barcelona, Spain
Duration: 2 May 20125 May 2012
Conference number: 9

Conference

Conference9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012
Abbreviated titleISBI 2012
CountrySpain
CityBarcelona
Period2/05/125/05/12

Fingerprint

Brain
Magnetic resonance
Volumetric analysis
Neuroimaging
Labeling
Labels
Fusion reactions

Keywords

  • brain atrophy
  • hippocampus
  • Alzheimers disease

Cite this

Ledig, C., Wolz, R., Aljabar, P., Lötjönen, J., Heckemann, R., Hammers, A., & Rueckert, D. (2012). Multi-class brain segmentation using atlas propagation and EM-based refinement. In 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI) (pp. 896-899). IEEE Institute of Electrical and Electronic Engineers . https://doi.org/10.1109/ISBI.2012.6235693
Ledig, Christian ; Wolz, Robin ; Aljabar, Paul ; Lötjönen, Jyrki ; Heckemann, Rolf ; Hammers, Alex ; Rueckert, Daniel. / Multi-class brain segmentation using atlas propagation and EM-based refinement. 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI). IEEE Institute of Electrical and Electronic Engineers , 2012. pp. 896-899
@inproceedings{9c10e09e776a45438d68ad2007b0e500,
title = "Multi-class brain segmentation using atlas propagation and EM-based refinement",
abstract = "In recent years, multi-atlas segmentation has emerged as one of the most accurate techniques for the segmentation of brain magnetic resonance (MR) images, especially when combined with intensity-based refinement techniques such as graph-cut or expectation-maximization (EM) optimization. However, most of the work so far has focused on intensity-based refinement strategies for individual anatomical structures such as the hippocampus. In this work we extend a previously proposed framework for labeling whole brain scans by incorporating a global and stationary Markov random field that ensures the consistency of the neighbourhood relations between structures with an a-priori defined model. In particular we improve the segmentation result of a locally weighted multi-atlas fusion method for 41 different structures simultaneously by applying a subsequent EM optimization step. We evaluate the proposed approach on 30 manually annotated brain MR images and observe an improvement of label overlaps to a manual reference by up to 6{\%}. We also achieved a considerably improved group separation when the proposed segmentation framework is applied to a volumetric analysis of 404 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort.",
keywords = "brain atrophy, hippocampus, Alzheimers disease",
author = "Christian Ledig and Robin Wolz and Paul Aljabar and Jyrki L{\"o}tj{\"o}nen and Rolf Heckemann and Alex Hammers and Daniel Rueckert",
note = "Project: 18493 Project : 71992",
year = "2012",
doi = "10.1109/ISBI.2012.6235693",
language = "English",
isbn = "978-1-4577-1857-1",
pages = "896--899",
booktitle = "2012 9th IEEE International Symposium on Biomedical Imaging (ISBI)",
publisher = "IEEE Institute of Electrical and Electronic Engineers",
address = "United States",

}

Ledig, C, Wolz, R, Aljabar, P, Lötjönen, J, Heckemann, R, Hammers, A & Rueckert, D 2012, Multi-class brain segmentation using atlas propagation and EM-based refinement. in 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI). IEEE Institute of Electrical and Electronic Engineers , pp. 896-899, 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012, Barcelona, Spain, 2/05/12. https://doi.org/10.1109/ISBI.2012.6235693

Multi-class brain segmentation using atlas propagation and EM-based refinement. / Ledig, Christian; Wolz, Robin; Aljabar, Paul; Lötjönen, Jyrki; Heckemann, Rolf; Hammers, Alex; Rueckert, Daniel.

2012 9th IEEE International Symposium on Biomedical Imaging (ISBI). IEEE Institute of Electrical and Electronic Engineers , 2012. p. 896-899.

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

TY - GEN

T1 - Multi-class brain segmentation using atlas propagation and EM-based refinement

AU - Ledig, Christian

AU - Wolz, Robin

AU - Aljabar, Paul

AU - Lötjönen, Jyrki

AU - Heckemann, Rolf

AU - Hammers, Alex

AU - Rueckert, Daniel

N1 - Project: 18493 Project : 71992

PY - 2012

Y1 - 2012

N2 - In recent years, multi-atlas segmentation has emerged as one of the most accurate techniques for the segmentation of brain magnetic resonance (MR) images, especially when combined with intensity-based refinement techniques such as graph-cut or expectation-maximization (EM) optimization. However, most of the work so far has focused on intensity-based refinement strategies for individual anatomical structures such as the hippocampus. In this work we extend a previously proposed framework for labeling whole brain scans by incorporating a global and stationary Markov random field that ensures the consistency of the neighbourhood relations between structures with an a-priori defined model. In particular we improve the segmentation result of a locally weighted multi-atlas fusion method for 41 different structures simultaneously by applying a subsequent EM optimization step. We evaluate the proposed approach on 30 manually annotated brain MR images and observe an improvement of label overlaps to a manual reference by up to 6%. We also achieved a considerably improved group separation when the proposed segmentation framework is applied to a volumetric analysis of 404 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort.

AB - In recent years, multi-atlas segmentation has emerged as one of the most accurate techniques for the segmentation of brain magnetic resonance (MR) images, especially when combined with intensity-based refinement techniques such as graph-cut or expectation-maximization (EM) optimization. However, most of the work so far has focused on intensity-based refinement strategies for individual anatomical structures such as the hippocampus. In this work we extend a previously proposed framework for labeling whole brain scans by incorporating a global and stationary Markov random field that ensures the consistency of the neighbourhood relations between structures with an a-priori defined model. In particular we improve the segmentation result of a locally weighted multi-atlas fusion method for 41 different structures simultaneously by applying a subsequent EM optimization step. We evaluate the proposed approach on 30 manually annotated brain MR images and observe an improvement of label overlaps to a manual reference by up to 6%. We also achieved a considerably improved group separation when the proposed segmentation framework is applied to a volumetric analysis of 404 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort.

KW - brain atrophy

KW - hippocampus

KW - Alzheimers disease

U2 - 10.1109/ISBI.2012.6235693

DO - 10.1109/ISBI.2012.6235693

M3 - Conference article in proceedings

SN - 978-1-4577-1857-1

SP - 896

EP - 899

BT - 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI)

PB - IEEE Institute of Electrical and Electronic Engineers

ER -

Ledig C, Wolz R, Aljabar P, Lötjönen J, Heckemann R, Hammers A et al. Multi-class brain segmentation using atlas propagation and EM-based refinement. In 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI). IEEE Institute of Electrical and Electronic Engineers . 2012. p. 896-899 https://doi.org/10.1109/ISBI.2012.6235693