Non-local atlas-guided multi-channel forest learning for human brain labeling

Guangkai Ma, Yaozong Gao, Guorong Wu, Ligang Wu, Dinggang Shen

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

1 Citation (Scopus)

Abstract

Labeling MR brain images into anatomically meaningful regions is important in many quantitative brain researches. In many existing label fusion methods, appearance information is widely used. Meanwhile, recent progress in computer vision suggests that the context feature is very useful in identifying an object from a complex scene. In light of this, we propose a novel learning-based label fusion method by using both low-level appearance features (computed from the target image) and high-level context features (computed from warped atlases or tentative labeling maps of the target image). In particular, we employ a multi-channel random forest to learn the nonlinear relationship between these hybrid features and the target labels (i.e., corresponding to certain anatomical structures). Moreover, to accommodate the high inter-subject variations, we further extend our learning-based label fusion to a multi-atlas scenario, i.e., we train a random forest for each atlas and then obtain the final labeling result according to the consensus of all atlases. We have comprehensively evaluated our method on both LONI-LBPA40 and IXI datasets, and achieved the highest labeling accuracy, compared to the state-of-the-art methods in the literature.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages719-726
Number of pages8
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

Atlas
Labeling
Labels
Brain
Fusion
Fusion reactions
Random Forest
Target
Computer Vision
Computer vision
Scenarios
Human
Learning
Context

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Ma, G., Gao, Y., Wu, G., Wu, L., & Shen, D. (2015). Non-local atlas-guided multi-channel forest learning for human brain labeling. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9351, pp. 719-726). (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_86

Non-local atlas-guided multi-channel forest learning for human brain labeling. / Ma, Guangkai; Gao, Yaozong; Wu, Guorong; Wu, Ligang; 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. 719-726 (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

Ma, G, Gao, Y, Wu, G, Wu, L & Shen, D 2015, Non-local atlas-guided multi-channel forest learning for human brain labeling. 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. 719-726, 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_86
Ma G, Gao Y, Wu G, Wu L, Shen D. Non-local atlas-guided multi-channel forest learning for human brain labeling. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351. Springer Verlag. 2015. p. 719-726. (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_86
Ma, Guangkai ; Gao, Yaozong ; Wu, Guorong ; Wu, Ligang ; Shen, Dinggang. / Non-local atlas-guided multi-channel forest learning for human brain labeling. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. pp. 719-726 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{2bc3781db7fa47af80a884f3253d43d7,
title = "Non-local atlas-guided multi-channel forest learning for human brain labeling",
abstract = "Labeling MR brain images into anatomically meaningful regions is important in many quantitative brain researches. In many existing label fusion methods, appearance information is widely used. Meanwhile, recent progress in computer vision suggests that the context feature is very useful in identifying an object from a complex scene. In light of this, we propose a novel learning-based label fusion method by using both low-level appearance features (computed from the target image) and high-level context features (computed from warped atlases or tentative labeling maps of the target image). In particular, we employ a multi-channel random forest to learn the nonlinear relationship between these hybrid features and the target labels (i.e., corresponding to certain anatomical structures). Moreover, to accommodate the high inter-subject variations, we further extend our learning-based label fusion to a multi-atlas scenario, i.e., we train a random forest for each atlas and then obtain the final labeling result according to the consensus of all atlases. We have comprehensively evaluated our method on both LONI-LBPA40 and IXI datasets, and achieved the highest labeling accuracy, compared to the state-of-the-art methods in the literature.",
author = "Guangkai Ma and Yaozong Gao and Guorong Wu and Ligang Wu and Dinggang Shen",
year = "2015",
doi = "10.1007/978-3-319-24574-4_86",
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 = "719--726",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Non-local atlas-guided multi-channel forest learning for human brain labeling

AU - Ma, Guangkai

AU - Gao, Yaozong

AU - Wu, Guorong

AU - Wu, Ligang

AU - Shen, Dinggang

PY - 2015

Y1 - 2015

N2 - Labeling MR brain images into anatomically meaningful regions is important in many quantitative brain researches. In many existing label fusion methods, appearance information is widely used. Meanwhile, recent progress in computer vision suggests that the context feature is very useful in identifying an object from a complex scene. In light of this, we propose a novel learning-based label fusion method by using both low-level appearance features (computed from the target image) and high-level context features (computed from warped atlases or tentative labeling maps of the target image). In particular, we employ a multi-channel random forest to learn the nonlinear relationship between these hybrid features and the target labels (i.e., corresponding to certain anatomical structures). Moreover, to accommodate the high inter-subject variations, we further extend our learning-based label fusion to a multi-atlas scenario, i.e., we train a random forest for each atlas and then obtain the final labeling result according to the consensus of all atlases. We have comprehensively evaluated our method on both LONI-LBPA40 and IXI datasets, and achieved the highest labeling accuracy, compared to the state-of-the-art methods in the literature.

AB - Labeling MR brain images into anatomically meaningful regions is important in many quantitative brain researches. In many existing label fusion methods, appearance information is widely used. Meanwhile, recent progress in computer vision suggests that the context feature is very useful in identifying an object from a complex scene. In light of this, we propose a novel learning-based label fusion method by using both low-level appearance features (computed from the target image) and high-level context features (computed from warped atlases or tentative labeling maps of the target image). In particular, we employ a multi-channel random forest to learn the nonlinear relationship between these hybrid features and the target labels (i.e., corresponding to certain anatomical structures). Moreover, to accommodate the high inter-subject variations, we further extend our learning-based label fusion to a multi-atlas scenario, i.e., we train a random forest for each atlas and then obtain the final labeling result according to the consensus of all atlases. We have comprehensively evaluated our method on both LONI-LBPA40 and IXI datasets, and achieved the highest labeling accuracy, compared to the state-of-the-art methods in the literature.

UR - http://www.scopus.com/inward/record.url?scp=84951837197&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84951837197&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-24574-4_86

DO - 10.1007/978-3-319-24574-4_86

M3 - Conference contribution

AN - SCOPUS:84951837197

SN - 9783319245737

VL - 9351

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

SP - 719

EP - 726

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

PB - Springer Verlag

ER -