Unsupervised deep learning for hippocampus segmentation in 7.0 Tesla MR images

Minjeong Kim, Guorong Wu, Dinggang Shen

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

13 Citations (Scopus)

Abstract

Recent emergence of 7.0T MR scanner sheds new light on the study of hippocampus by providing much higher image contrast and resolution. However, the new characteristics shown in 7.0T images, such as richer structural information and more severe intensity inhomogeneity, raise serious issues for the extraction of distinctive and robust features for accurately segmenting hippocampus in 7.0T images. On the other hand, the hand-crafted image features (such as Haar and SIFT), which were designed for 1.5T and 3.0T images, generally fail to be effective, because of the considerable image artifacts in 7.0T images. In this paper, we introduce the concept of unsupervised deep learning to learn the hierarchical feature representation directly from the pre-observed image patches in 7.0T images. Specifically, a two-layer stacked convolutional Independent Subspace Analysis (ISA) network is built to learn not only the intrinsic low-level features from image patches in the lower layer, but also the high-level features in the higher layer to describe the global image appearance based on the outputs from the lower layer. We have successfully integrated this deep learning scheme into a state-of-the-art multi-atlases based segmentation framework by replacing the previous hand-crafted image features by the hierarchical feature representations inferred from the two-layer ISA network. Promising hippocampus segmentation results were obtained on 20 7.0T images, demonstrating the enhanced discriminative power achieved by our deep learning method.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages1-8
Number of pages8
Volume8184 LNCS
ISBN (Print)9783319022666
DOIs
Publication statusPublished - 2013 Jan 1
Externally publishedYes
Event4th International Workshop on Machine Learning in Medical Imaging, MLMI 2013, Held in Conjunction with 16th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2013 - Nagoya, Japan
Duration: 2013 Sep 222013 Sep 22

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8184 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other4th International Workshop on Machine Learning in Medical Imaging, MLMI 2013, Held in Conjunction with 16th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2013
CountryJapan
CityNagoya
Period13/9/2213/9/22

Fingerprint

Hippocampus
Unsupervised Learning
Segmentation
Electric network analysis
Network Analysis
Deep learning
Patch
Subspace
Scale Invariant Feature Transform
Atlas
Scanner
Inhomogeneity

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Kim, M., Wu, G., & Shen, D. (2013). Unsupervised deep learning for hippocampus segmentation in 7.0 Tesla MR images. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8184 LNCS, pp. 1-8). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8184 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-02267-3_1

Unsupervised deep learning for hippocampus segmentation in 7.0 Tesla MR images. / Kim, Minjeong; Wu, Guorong; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8184 LNCS Springer Verlag, 2013. p. 1-8 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8184 LNCS).

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

Kim, M, Wu, G & Shen, D 2013, Unsupervised deep learning for hippocampus segmentation in 7.0 Tesla MR images. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 8184 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8184 LNCS, Springer Verlag, pp. 1-8, 4th International Workshop on Machine Learning in Medical Imaging, MLMI 2013, Held in Conjunction with 16th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2013, Nagoya, Japan, 13/9/22. https://doi.org/10.1007/978-3-319-02267-3_1
Kim M, Wu G, Shen D. Unsupervised deep learning for hippocampus segmentation in 7.0 Tesla MR images. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8184 LNCS. Springer Verlag. 2013. p. 1-8. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-02267-3_1
Kim, Minjeong ; Wu, Guorong ; Shen, Dinggang. / Unsupervised deep learning for hippocampus segmentation in 7.0 Tesla MR images. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8184 LNCS Springer Verlag, 2013. pp. 1-8 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{ce07f294657840f49318c1ead2e11858,
title = "Unsupervised deep learning for hippocampus segmentation in 7.0 Tesla MR images",
abstract = "Recent emergence of 7.0T MR scanner sheds new light on the study of hippocampus by providing much higher image contrast and resolution. However, the new characteristics shown in 7.0T images, such as richer structural information and more severe intensity inhomogeneity, raise serious issues for the extraction of distinctive and robust features for accurately segmenting hippocampus in 7.0T images. On the other hand, the hand-crafted image features (such as Haar and SIFT), which were designed for 1.5T and 3.0T images, generally fail to be effective, because of the considerable image artifacts in 7.0T images. In this paper, we introduce the concept of unsupervised deep learning to learn the hierarchical feature representation directly from the pre-observed image patches in 7.0T images. Specifically, a two-layer stacked convolutional Independent Subspace Analysis (ISA) network is built to learn not only the intrinsic low-level features from image patches in the lower layer, but also the high-level features in the higher layer to describe the global image appearance based on the outputs from the lower layer. We have successfully integrated this deep learning scheme into a state-of-the-art multi-atlases based segmentation framework by replacing the previous hand-crafted image features by the hierarchical feature representations inferred from the two-layer ISA network. Promising hippocampus segmentation results were obtained on 20 7.0T images, demonstrating the enhanced discriminative power achieved by our deep learning method.",
author = "Minjeong Kim and Guorong Wu and Dinggang Shen",
year = "2013",
month = "1",
day = "1",
doi = "10.1007/978-3-319-02267-3_1",
language = "English",
isbn = "9783319022666",
volume = "8184 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "1--8",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Unsupervised deep learning for hippocampus segmentation in 7.0 Tesla MR images

AU - Kim, Minjeong

AU - Wu, Guorong

AU - Shen, Dinggang

PY - 2013/1/1

Y1 - 2013/1/1

N2 - Recent emergence of 7.0T MR scanner sheds new light on the study of hippocampus by providing much higher image contrast and resolution. However, the new characteristics shown in 7.0T images, such as richer structural information and more severe intensity inhomogeneity, raise serious issues for the extraction of distinctive and robust features for accurately segmenting hippocampus in 7.0T images. On the other hand, the hand-crafted image features (such as Haar and SIFT), which were designed for 1.5T and 3.0T images, generally fail to be effective, because of the considerable image artifacts in 7.0T images. In this paper, we introduce the concept of unsupervised deep learning to learn the hierarchical feature representation directly from the pre-observed image patches in 7.0T images. Specifically, a two-layer stacked convolutional Independent Subspace Analysis (ISA) network is built to learn not only the intrinsic low-level features from image patches in the lower layer, but also the high-level features in the higher layer to describe the global image appearance based on the outputs from the lower layer. We have successfully integrated this deep learning scheme into a state-of-the-art multi-atlases based segmentation framework by replacing the previous hand-crafted image features by the hierarchical feature representations inferred from the two-layer ISA network. Promising hippocampus segmentation results were obtained on 20 7.0T images, demonstrating the enhanced discriminative power achieved by our deep learning method.

AB - Recent emergence of 7.0T MR scanner sheds new light on the study of hippocampus by providing much higher image contrast and resolution. However, the new characteristics shown in 7.0T images, such as richer structural information and more severe intensity inhomogeneity, raise serious issues for the extraction of distinctive and robust features for accurately segmenting hippocampus in 7.0T images. On the other hand, the hand-crafted image features (such as Haar and SIFT), which were designed for 1.5T and 3.0T images, generally fail to be effective, because of the considerable image artifacts in 7.0T images. In this paper, we introduce the concept of unsupervised deep learning to learn the hierarchical feature representation directly from the pre-observed image patches in 7.0T images. Specifically, a two-layer stacked convolutional Independent Subspace Analysis (ISA) network is built to learn not only the intrinsic low-level features from image patches in the lower layer, but also the high-level features in the higher layer to describe the global image appearance based on the outputs from the lower layer. We have successfully integrated this deep learning scheme into a state-of-the-art multi-atlases based segmentation framework by replacing the previous hand-crafted image features by the hierarchical feature representations inferred from the two-layer ISA network. Promising hippocampus segmentation results were obtained on 20 7.0T images, demonstrating the enhanced discriminative power achieved by our deep learning method.

UR - http://www.scopus.com/inward/record.url?scp=84886731548&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84886731548&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-02267-3_1

DO - 10.1007/978-3-319-02267-3_1

M3 - Conference contribution

AN - SCOPUS:84886731548

SN - 9783319022666

VL - 8184 LNCS

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

SP - 1

EP - 8

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

PB - Springer Verlag

ER -