Segmentation of infant hippocampus using common feature representations learned for multimodal longitudinal data

Yanrong Guo, Guorong Wu, Pew Thian Yap, Valerie Jewells, Weili Lin, Dinggang Shen

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

2 Citations (Scopus)

Abstract

Aberrant development of the human brain during the first year after birth is known to cause critical implications in later stages of life. In particular, neuropsychiatric disorders, such as attention deficit hyperactivity disorder (ADHD), have been linked with abnormal early development of the hippocampus. Despite its known importance, studying the hippocampus in infant subjects is very challenging due to the significantly smaller brain size, dynamically varying image contrast, and large across-subject variation. In this paper, we present a novel method for effective hippocampus segmentation by using a multi- atlas approach that integrates the complementary multimodal information from longitudinal T1 and T2 MR images. In particular, considering the highly heterogeneous nature of the longitudinal data, we propose to learn their common feature representations by using hierarchical multi-set kernel canonical correlation analysis (CCA). Specifically, we will learn (1) within-time-point common features by projecting different modality features of each time point to its own modality-free common space, and (2) across-time-point common features by mapping all time-point-specific common features to a global common space for all time points. These final features are then employed in patch matching across different modalities and time points for hippocampus segmentation, via label propagation and fusion. Experimental results demonstrate the improved performance of our method over the state-of-the-art methods.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages63-71
Number of pages9
Volume9351
ISBN (Print)9783319245737
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015 - Munich, Germany
Duration: 2015 Oct 52015 Oct 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9351
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015
CountryGermany
CityMunich
Period15/10/515/10/9

Fingerprint

Hippocampus
Longitudinal Data
Brain
Segmentation
Modality
Labels
Fusion reactions
Disorder
Canonical Correlation Analysis
Atlas
Multiset
Patch
Fusion
Integrate
Propagation
kernel
Experimental Results
Demonstrate

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Guo, Y., Wu, G., Yap, P. T., Jewells, V., Lin, W., & Shen, D. (2015). Segmentation of infant hippocampus using common feature representations learned for multimodal longitudinal data. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9351, pp. 63-71). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9351). Springer Verlag. https://doi.org/10.1007/978-3-319-24574-4_8

Segmentation of infant hippocampus using common feature representations learned for multimodal longitudinal data. / Guo, Yanrong; Wu, Guorong; Yap, Pew Thian; Jewells, Valerie; Lin, Weili; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. p. 63-71 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9351).

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

Guo, Y, Wu, G, Yap, PT, Jewells, V, Lin, W & Shen, D 2015, Segmentation of infant hippocampus using common feature representations learned for multimodal longitudinal data. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 9351, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9351, Springer Verlag, pp. 63-71, 18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015, Munich, Germany, 15/10/5. https://doi.org/10.1007/978-3-319-24574-4_8
Guo Y, Wu G, Yap PT, Jewells V, Lin W, Shen D. Segmentation of infant hippocampus using common feature representations learned for multimodal longitudinal data. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351. Springer Verlag. 2015. p. 63-71. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-24574-4_8
Guo, Yanrong ; Wu, Guorong ; Yap, Pew Thian ; Jewells, Valerie ; Lin, Weili ; Shen, Dinggang. / Segmentation of infant hippocampus using common feature representations learned for multimodal longitudinal data. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. pp. 63-71 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{4af890a52efb41a092e54e6b6fe797dd,
title = "Segmentation of infant hippocampus using common feature representations learned for multimodal longitudinal data",
abstract = "Aberrant development of the human brain during the first year after birth is known to cause critical implications in later stages of life. In particular, neuropsychiatric disorders, such as attention deficit hyperactivity disorder (ADHD), have been linked with abnormal early development of the hippocampus. Despite its known importance, studying the hippocampus in infant subjects is very challenging due to the significantly smaller brain size, dynamically varying image contrast, and large across-subject variation. In this paper, we present a novel method for effective hippocampus segmentation by using a multi- atlas approach that integrates the complementary multimodal information from longitudinal T1 and T2 MR images. In particular, considering the highly heterogeneous nature of the longitudinal data, we propose to learn their common feature representations by using hierarchical multi-set kernel canonical correlation analysis (CCA). Specifically, we will learn (1) within-time-point common features by projecting different modality features of each time point to its own modality-free common space, and (2) across-time-point common features by mapping all time-point-specific common features to a global common space for all time points. These final features are then employed in patch matching across different modalities and time points for hippocampus segmentation, via label propagation and fusion. Experimental results demonstrate the improved performance of our method over the state-of-the-art methods.",
author = "Yanrong Guo and Guorong Wu and Yap, {Pew Thian} and Valerie Jewells and Weili Lin and Dinggang Shen",
year = "2015",
doi = "10.1007/978-3-319-24574-4_8",
language = "English",
isbn = "9783319245737",
volume = "9351",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "63--71",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Segmentation of infant hippocampus using common feature representations learned for multimodal longitudinal data

AU - Guo, Yanrong

AU - Wu, Guorong

AU - Yap, Pew Thian

AU - Jewells, Valerie

AU - Lin, Weili

AU - Shen, Dinggang

PY - 2015

Y1 - 2015

N2 - Aberrant development of the human brain during the first year after birth is known to cause critical implications in later stages of life. In particular, neuropsychiatric disorders, such as attention deficit hyperactivity disorder (ADHD), have been linked with abnormal early development of the hippocampus. Despite its known importance, studying the hippocampus in infant subjects is very challenging due to the significantly smaller brain size, dynamically varying image contrast, and large across-subject variation. In this paper, we present a novel method for effective hippocampus segmentation by using a multi- atlas approach that integrates the complementary multimodal information from longitudinal T1 and T2 MR images. In particular, considering the highly heterogeneous nature of the longitudinal data, we propose to learn their common feature representations by using hierarchical multi-set kernel canonical correlation analysis (CCA). Specifically, we will learn (1) within-time-point common features by projecting different modality features of each time point to its own modality-free common space, and (2) across-time-point common features by mapping all time-point-specific common features to a global common space for all time points. These final features are then employed in patch matching across different modalities and time points for hippocampus segmentation, via label propagation and fusion. Experimental results demonstrate the improved performance of our method over the state-of-the-art methods.

AB - Aberrant development of the human brain during the first year after birth is known to cause critical implications in later stages of life. In particular, neuropsychiatric disorders, such as attention deficit hyperactivity disorder (ADHD), have been linked with abnormal early development of the hippocampus. Despite its known importance, studying the hippocampus in infant subjects is very challenging due to the significantly smaller brain size, dynamically varying image contrast, and large across-subject variation. In this paper, we present a novel method for effective hippocampus segmentation by using a multi- atlas approach that integrates the complementary multimodal information from longitudinal T1 and T2 MR images. In particular, considering the highly heterogeneous nature of the longitudinal data, we propose to learn their common feature representations by using hierarchical multi-set kernel canonical correlation analysis (CCA). Specifically, we will learn (1) within-time-point common features by projecting different modality features of each time point to its own modality-free common space, and (2) across-time-point common features by mapping all time-point-specific common features to a global common space for all time points. These final features are then employed in patch matching across different modalities and time points for hippocampus segmentation, via label propagation and fusion. Experimental results demonstrate the improved performance of our method over the state-of-the-art methods.

UR - http://www.scopus.com/inward/record.url?scp=84951793320&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84951793320&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-24574-4_8

DO - 10.1007/978-3-319-24574-4_8

M3 - Conference contribution

SN - 9783319245737

VL - 9351

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

SP - 63

EP - 71

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

PB - Springer Verlag

ER -