Multi-atlas and multi-modal hippocampus segmentation for infant MR brain images by propagating anatomical labels on hypergraph

Pei Dong, Yanrong Guo, Dinggang Shen, Guorong Wu

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

4 Citations (Scopus)

Abstract

Accurate segmentation of hippocampus from infant magnetic resonance (MR) images is very important in the study of early brain development and neurological disorder. Recently, multi-atlas patch-based label fusion methods have shown a great success in segmenting anatomical structures from medical images. However, the dramatic appearance change from birth to 1-year-old and the poor image contrast make the existing label fusion methods less competitive to handle infant brain images. To alleviate these difficulties, we propose a novel multi-atlas and multi-modal label fusion method, which can unanimously label for all voxels by propagating the anatomical labels on a hypergraph. Specifically, we consider not only all voxels within the target image but also voxels across the atlas images as the vertexes in the hypergraph. Each hyperedge encodes a high-order correlation, among a set of vertexes, in different perspectives which incorporate (1) feature affinity within the multi-modal feature space, (2) spatial coherence within target image, and (3) population heuristics from multiple atlases. In addition, our label fusion method further allows those reliable voxels to supervise the label estimation on other difficult-to-label voxels, based on the established hyperedges, until all the target image voxels reach the unanimous labeling result. We evaluate our proposed label fusion method in segmenting hippocampus from T1 and T2 weighted MR images acquired from at 2-week-old, 3-month-old, 6-month-old, 9-month-old, and 12-month-old. Our segmentation results achieves improvement of labeling accuracy over the conventional state-of-the-art label fusion methods, which shows a great potential to facilitate the early infant brain studies.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages188-196
Number of pages9
Volume9467
ISBN (Print)9783319281933
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event1st International Workshop on Patch-Based Techniques in Medical Imaging, Patch-MI 2015 - Munich, Germany
Duration: 2015 Oct 92015 Oct 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9467
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other1st International Workshop on Patch-Based Techniques in Medical Imaging, Patch-MI 2015
CountryGermany
CityMunich
Period15/10/915/10/9

Fingerprint

Hippocampus
Magnetic Resonance
Atlas
Voxel
Magnetic resonance
Hypergraph
Labels
Brain
Fusion
Segmentation
Fusion reactions
Magnetic Resonance Image
Target
Labeling
Medical Image
Feature Space
Affine transformation
Patch
Disorder
Heuristics

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Dong, P., Guo, Y., Shen, D., & Wu, G. (2015). Multi-atlas and multi-modal hippocampus segmentation for infant MR brain images by propagating anatomical labels on hypergraph. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9467, pp. 188-196). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9467). Springer Verlag. https://doi.org/10.1007/978-3-319-28194-0_23

Multi-atlas and multi-modal hippocampus segmentation for infant MR brain images by propagating anatomical labels on hypergraph. / Dong, Pei; Guo, Yanrong; Shen, Dinggang; Wu, Guorong.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9467 Springer Verlag, 2015. p. 188-196 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9467).

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

Dong, P, Guo, Y, Shen, D & Wu, G 2015, Multi-atlas and multi-modal hippocampus segmentation for infant MR brain images by propagating anatomical labels on hypergraph. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 9467, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9467, Springer Verlag, pp. 188-196, 1st International Workshop on Patch-Based Techniques in Medical Imaging, Patch-MI 2015, Munich, Germany, 15/10/9. https://doi.org/10.1007/978-3-319-28194-0_23
Dong P, Guo Y, Shen D, Wu G. Multi-atlas and multi-modal hippocampus segmentation for infant MR brain images by propagating anatomical labels on hypergraph. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9467. Springer Verlag. 2015. p. 188-196. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-28194-0_23
Dong, Pei ; Guo, Yanrong ; Shen, Dinggang ; Wu, Guorong. / Multi-atlas and multi-modal hippocampus segmentation for infant MR brain images by propagating anatomical labels on hypergraph. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9467 Springer Verlag, 2015. pp. 188-196 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{5bec91261e9b48d58944e96ee2834ed0,
title = "Multi-atlas and multi-modal hippocampus segmentation for infant MR brain images by propagating anatomical labels on hypergraph",
abstract = "Accurate segmentation of hippocampus from infant magnetic resonance (MR) images is very important in the study of early brain development and neurological disorder. Recently, multi-atlas patch-based label fusion methods have shown a great success in segmenting anatomical structures from medical images. However, the dramatic appearance change from birth to 1-year-old and the poor image contrast make the existing label fusion methods less competitive to handle infant brain images. To alleviate these difficulties, we propose a novel multi-atlas and multi-modal label fusion method, which can unanimously label for all voxels by propagating the anatomical labels on a hypergraph. Specifically, we consider not only all voxels within the target image but also voxels across the atlas images as the vertexes in the hypergraph. Each hyperedge encodes a high-order correlation, among a set of vertexes, in different perspectives which incorporate (1) feature affinity within the multi-modal feature space, (2) spatial coherence within target image, and (3) population heuristics from multiple atlases. In addition, our label fusion method further allows those reliable voxels to supervise the label estimation on other difficult-to-label voxels, based on the established hyperedges, until all the target image voxels reach the unanimous labeling result. We evaluate our proposed label fusion method in segmenting hippocampus from T1 and T2 weighted MR images acquired from at 2-week-old, 3-month-old, 6-month-old, 9-month-old, and 12-month-old. Our segmentation results achieves improvement of labeling accuracy over the conventional state-of-the-art label fusion methods, which shows a great potential to facilitate the early infant brain studies.",
author = "Pei Dong and Yanrong Guo and Dinggang Shen and Guorong Wu",
year = "2015",
doi = "10.1007/978-3-319-28194-0_23",
language = "English",
isbn = "9783319281933",
volume = "9467",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "188--196",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Multi-atlas and multi-modal hippocampus segmentation for infant MR brain images by propagating anatomical labels on hypergraph

AU - Dong, Pei

AU - Guo, Yanrong

AU - Shen, Dinggang

AU - Wu, Guorong

PY - 2015

Y1 - 2015

N2 - Accurate segmentation of hippocampus from infant magnetic resonance (MR) images is very important in the study of early brain development and neurological disorder. Recently, multi-atlas patch-based label fusion methods have shown a great success in segmenting anatomical structures from medical images. However, the dramatic appearance change from birth to 1-year-old and the poor image contrast make the existing label fusion methods less competitive to handle infant brain images. To alleviate these difficulties, we propose a novel multi-atlas and multi-modal label fusion method, which can unanimously label for all voxels by propagating the anatomical labels on a hypergraph. Specifically, we consider not only all voxels within the target image but also voxels across the atlas images as the vertexes in the hypergraph. Each hyperedge encodes a high-order correlation, among a set of vertexes, in different perspectives which incorporate (1) feature affinity within the multi-modal feature space, (2) spatial coherence within target image, and (3) population heuristics from multiple atlases. In addition, our label fusion method further allows those reliable voxels to supervise the label estimation on other difficult-to-label voxels, based on the established hyperedges, until all the target image voxels reach the unanimous labeling result. We evaluate our proposed label fusion method in segmenting hippocampus from T1 and T2 weighted MR images acquired from at 2-week-old, 3-month-old, 6-month-old, 9-month-old, and 12-month-old. Our segmentation results achieves improvement of labeling accuracy over the conventional state-of-the-art label fusion methods, which shows a great potential to facilitate the early infant brain studies.

AB - Accurate segmentation of hippocampus from infant magnetic resonance (MR) images is very important in the study of early brain development and neurological disorder. Recently, multi-atlas patch-based label fusion methods have shown a great success in segmenting anatomical structures from medical images. However, the dramatic appearance change from birth to 1-year-old and the poor image contrast make the existing label fusion methods less competitive to handle infant brain images. To alleviate these difficulties, we propose a novel multi-atlas and multi-modal label fusion method, which can unanimously label for all voxels by propagating the anatomical labels on a hypergraph. Specifically, we consider not only all voxels within the target image but also voxels across the atlas images as the vertexes in the hypergraph. Each hyperedge encodes a high-order correlation, among a set of vertexes, in different perspectives which incorporate (1) feature affinity within the multi-modal feature space, (2) spatial coherence within target image, and (3) population heuristics from multiple atlases. In addition, our label fusion method further allows those reliable voxels to supervise the label estimation on other difficult-to-label voxels, based on the established hyperedges, until all the target image voxels reach the unanimous labeling result. We evaluate our proposed label fusion method in segmenting hippocampus from T1 and T2 weighted MR images acquired from at 2-week-old, 3-month-old, 6-month-old, 9-month-old, and 12-month-old. Our segmentation results achieves improvement of labeling accuracy over the conventional state-of-the-art label fusion methods, which shows a great potential to facilitate the early infant brain studies.

UR - http://www.scopus.com/inward/record.url?scp=84955243899&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84955243899&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-28194-0_23

DO - 10.1007/978-3-319-28194-0_23

M3 - Conference contribution

AN - SCOPUS:84955243899

SN - 9783319281933

VL - 9467

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

SP - 188

EP - 196

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

PB - Springer Verlag

ER -