Segmenting hippocampus from infant brains by sparse patch matching with deep-learned features

Yanrong Guo, Guorong Wu, Leah A. Commander, Stephanie Szary, Valerie Jewells, Weili Lin, Dinggang Shen

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

15 Citations (Scopus)

Abstract

Accurate segmentation of the hippocampus from infant MR brain images is a critical step for investigating early brain development. Unfortunately, the previous tools developed for adult hippocampus segmentation are not suitable for infant brain images acquired from the first year of life, which often have poor tissue contrast and variable structural patterns of early hippocampal development. From our point of view, the main problem is lack of discriminative and robust feature representations for distinguishing the hippocampus from the surrounding brain structures. Thus, instead of directly using the predefined features as popularly used in the conventional methods, we propose to learn the latent feature representations of infant MR brain images by unsupervised deep learning. Since deep learning paradigms can learn low-level features and then successfully build up more comprehensive high-level features in a layer-by-layer manner, such hierarchical feature representations can be more competitive for distinguishing the hippocampus from entire brain images. To this end, we apply Stacked Auto Encoder (SAE) to learn the deep feature representations from both T1- and T2-weighed MR images combining their complementary information, which is important for characterizing different development stages of infant brains after birth. Then, we present a sparse patch matching method for transferring hippocampus labels from multiple atlases to the new infant brain image, by using deep-learned feature representations to measure the inter-patch similarity. Experimental results on 2-week-old to 9-month-old infant brain images show the effectiveness of the proposed method, especially compared to the state-of-the-art counterpart 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
Pages308-315
Number of pages8
Volume8674 LNCS
EditionPART 2
ISBN (Print)9783319104690
DOIs
Publication statusPublished - 2014 Jan 1
Externally publishedYes
Event17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014 - Boston, MA, United States
Duration: 2014 Sep 142014 Sep 18

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume8674 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014
CountryUnited States
CityBoston, MA
Period14/9/1414/9/18

Fingerprint

Hippocampus
Patch
Brain
Segmentation
Unsupervised Learning
Atlas
Encoder
Labels
Paradigm
Entire
Tissue

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Guo, Y., Wu, G., Commander, L. A., Szary, S., Jewells, V., Lin, W., & Shen, D. (2014). Segmenting hippocampus from infant brains by sparse patch matching with deep-learned features. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 2 ed., Vol. 8674 LNCS, pp. 308-315). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8674 LNCS, No. PART 2). Springer Verlag. https://doi.org/10.1007/978-3-319-10470-6_39

Segmenting hippocampus from infant brains by sparse patch matching with deep-learned features. / Guo, Yanrong; Wu, Guorong; Commander, Leah A.; Szary, Stephanie; Jewells, Valerie; Lin, Weili; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8674 LNCS PART 2. ed. Springer Verlag, 2014. p. 308-315 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8674 LNCS, No. PART 2).

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

Guo, Y, Wu, G, Commander, LA, Szary, S, Jewells, V, Lin, W & Shen, D 2014, Segmenting hippocampus from infant brains by sparse patch matching with deep-learned features. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 2 edn, vol. 8674 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 8674 LNCS, Springer Verlag, pp. 308-315, 17th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2014, Boston, MA, United States, 14/9/14. https://doi.org/10.1007/978-3-319-10470-6_39
Guo Y, Wu G, Commander LA, Szary S, Jewells V, Lin W et al. Segmenting hippocampus from infant brains by sparse patch matching with deep-learned features. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 2 ed. Vol. 8674 LNCS. Springer Verlag. 2014. p. 308-315. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). https://doi.org/10.1007/978-3-319-10470-6_39
Guo, Yanrong ; Wu, Guorong ; Commander, Leah A. ; Szary, Stephanie ; Jewells, Valerie ; Lin, Weili ; Shen, Dinggang. / Segmenting hippocampus from infant brains by sparse patch matching with deep-learned features. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8674 LNCS PART 2. ed. Springer Verlag, 2014. pp. 308-315 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).
@inproceedings{27be7c0565744d60b87058546cfb72b6,
title = "Segmenting hippocampus from infant brains by sparse patch matching with deep-learned features",
abstract = "Accurate segmentation of the hippocampus from infant MR brain images is a critical step for investigating early brain development. Unfortunately, the previous tools developed for adult hippocampus segmentation are not suitable for infant brain images acquired from the first year of life, which often have poor tissue contrast and variable structural patterns of early hippocampal development. From our point of view, the main problem is lack of discriminative and robust feature representations for distinguishing the hippocampus from the surrounding brain structures. Thus, instead of directly using the predefined features as popularly used in the conventional methods, we propose to learn the latent feature representations of infant MR brain images by unsupervised deep learning. Since deep learning paradigms can learn low-level features and then successfully build up more comprehensive high-level features in a layer-by-layer manner, such hierarchical feature representations can be more competitive for distinguishing the hippocampus from entire brain images. To this end, we apply Stacked Auto Encoder (SAE) to learn the deep feature representations from both T1- and T2-weighed MR images combining their complementary information, which is important for characterizing different development stages of infant brains after birth. Then, we present a sparse patch matching method for transferring hippocampus labels from multiple atlases to the new infant brain image, by using deep-learned feature representations to measure the inter-patch similarity. Experimental results on 2-week-old to 9-month-old infant brain images show the effectiveness of the proposed method, especially compared to the state-of-the-art counterpart methods.",
author = "Yanrong Guo and Guorong Wu and Commander, {Leah A.} and Stephanie Szary and Valerie Jewells and Weili Lin and Dinggang Shen",
year = "2014",
month = "1",
day = "1",
doi = "10.1007/978-3-319-10470-6_39",
language = "English",
isbn = "9783319104690",
volume = "8674 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
number = "PART 2",
pages = "308--315",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 2",

}

TY - GEN

T1 - Segmenting hippocampus from infant brains by sparse patch matching with deep-learned features

AU - Guo, Yanrong

AU - Wu, Guorong

AU - Commander, Leah A.

AU - Szary, Stephanie

AU - Jewells, Valerie

AU - Lin, Weili

AU - Shen, Dinggang

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Accurate segmentation of the hippocampus from infant MR brain images is a critical step for investigating early brain development. Unfortunately, the previous tools developed for adult hippocampus segmentation are not suitable for infant brain images acquired from the first year of life, which often have poor tissue contrast and variable structural patterns of early hippocampal development. From our point of view, the main problem is lack of discriminative and robust feature representations for distinguishing the hippocampus from the surrounding brain structures. Thus, instead of directly using the predefined features as popularly used in the conventional methods, we propose to learn the latent feature representations of infant MR brain images by unsupervised deep learning. Since deep learning paradigms can learn low-level features and then successfully build up more comprehensive high-level features in a layer-by-layer manner, such hierarchical feature representations can be more competitive for distinguishing the hippocampus from entire brain images. To this end, we apply Stacked Auto Encoder (SAE) to learn the deep feature representations from both T1- and T2-weighed MR images combining their complementary information, which is important for characterizing different development stages of infant brains after birth. Then, we present a sparse patch matching method for transferring hippocampus labels from multiple atlases to the new infant brain image, by using deep-learned feature representations to measure the inter-patch similarity. Experimental results on 2-week-old to 9-month-old infant brain images show the effectiveness of the proposed method, especially compared to the state-of-the-art counterpart methods.

AB - Accurate segmentation of the hippocampus from infant MR brain images is a critical step for investigating early brain development. Unfortunately, the previous tools developed for adult hippocampus segmentation are not suitable for infant brain images acquired from the first year of life, which often have poor tissue contrast and variable structural patterns of early hippocampal development. From our point of view, the main problem is lack of discriminative and robust feature representations for distinguishing the hippocampus from the surrounding brain structures. Thus, instead of directly using the predefined features as popularly used in the conventional methods, we propose to learn the latent feature representations of infant MR brain images by unsupervised deep learning. Since deep learning paradigms can learn low-level features and then successfully build up more comprehensive high-level features in a layer-by-layer manner, such hierarchical feature representations can be more competitive for distinguishing the hippocampus from entire brain images. To this end, we apply Stacked Auto Encoder (SAE) to learn the deep feature representations from both T1- and T2-weighed MR images combining their complementary information, which is important for characterizing different development stages of infant brains after birth. Then, we present a sparse patch matching method for transferring hippocampus labels from multiple atlases to the new infant brain image, by using deep-learned feature representations to measure the inter-patch similarity. Experimental results on 2-week-old to 9-month-old infant brain images show the effectiveness of the proposed method, especially compared to the state-of-the-art counterpart methods.

UR - http://www.scopus.com/inward/record.url?scp=84906980652&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84906980652&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-10470-6_39

DO - 10.1007/978-3-319-10470-6_39

M3 - Conference contribution

AN - SCOPUS:84906980652

SN - 9783319104690

VL - 8674 LNCS

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

SP - 308

EP - 315

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

PB - Springer Verlag

ER -