Deep Bv: A Fully Automated System for Brain Ventricle Localization and Segmentation In 3D Ultrasound Images of Embryonic Mice

Ziming Qiu, Jack Langerman, Nitin Nair, Orlando Aristizabal, Jonathan Mamou, Daniel H. Turnbull, Jeffrey Ketterling, Yao Wang

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

Abstract

Volumetric analysis of brain ventricle (BV) structure is a key tool in the study of central nervous system development in embryonic mice. High-frequency ultrasound (HFU) is the only non-invasive, real-time modality available for rapid volumetric imaging of embryos in utero. However, manual segmentation of the BV from HFU volumes is tedious, time-consuming, and requires specialized expertise. In this paper, we propose a novel deep learning based BV segmentation system for whole-body HFU images of mouse embryos. Our fully automated system consists of two modules: localization and segmentation. It first applies a volumetric convolutional neural network on a 3D sliding window over the entire volume to identify a 3D bounding box containing the entire BV. It then employs a fully convolutional network to segment the detected bounding box into BV and background. The system achieves a Dice Similarity Coefficient (DSC) of 0.8956 for BV segmentation on an unseen 111 HFU volume test set surpassing the previous state-of-the-art method (DSC of 0.7119) by a margin of 25%.

Original languageEnglish (US)
Title of host publication2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538659168
DOIs
StatePublished - Jan 16 2019
Event2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Philadelphia, United States
Duration: Dec 1 2018 → …

Publication series

Name2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings

Conference

Conference2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018
CountryUnited States
CityPhiladelphia
Period12/1/18 → …

Fingerprint

Brain
Ultrasonics
Volumetric analysis
Embryonic Structures
Neurology
Embryonic Development
Central Nervous System
Learning
Neural networks
Imaging techniques

ASJC Scopus subject areas

  • Signal Processing
  • Health Informatics

Cite this

Qiu, Z., Langerman, J., Nair, N., Aristizabal, O., Mamou, J., Turnbull, D. H., ... Wang, Y. (2019). Deep Bv: A Fully Automated System for Brain Ventricle Localization and Segmentation In 3D Ultrasound Images of Embryonic Mice. In 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings [8615610] (2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SPMB.2018.8615610

Deep Bv : A Fully Automated System for Brain Ventricle Localization and Segmentation In 3D Ultrasound Images of Embryonic Mice. / Qiu, Ziming; Langerman, Jack; Nair, Nitin; Aristizabal, Orlando; Mamou, Jonathan; Turnbull, Daniel H.; Ketterling, Jeffrey; Wang, Yao.

2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. 8615610 (2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings).

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

Qiu, Z, Langerman, J, Nair, N, Aristizabal, O, Mamou, J, Turnbull, DH, Ketterling, J & Wang, Y 2019, Deep Bv: A Fully Automated System for Brain Ventricle Localization and Segmentation In 3D Ultrasound Images of Embryonic Mice. in 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings., 8615610, 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018, Philadelphia, United States, 12/1/18. https://doi.org/10.1109/SPMB.2018.8615610
Qiu Z, Langerman J, Nair N, Aristizabal O, Mamou J, Turnbull DH et al. Deep Bv: A Fully Automated System for Brain Ventricle Localization and Segmentation In 3D Ultrasound Images of Embryonic Mice. In 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. 8615610. (2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings). https://doi.org/10.1109/SPMB.2018.8615610
Qiu, Ziming ; Langerman, Jack ; Nair, Nitin ; Aristizabal, Orlando ; Mamou, Jonathan ; Turnbull, Daniel H. ; Ketterling, Jeffrey ; Wang, Yao. / Deep Bv : A Fully Automated System for Brain Ventricle Localization and Segmentation In 3D Ultrasound Images of Embryonic Mice. 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. (2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings).
@inproceedings{ce5615c2c8e7457397f9601fad82cca7,
title = "Deep Bv: A Fully Automated System for Brain Ventricle Localization and Segmentation In 3D Ultrasound Images of Embryonic Mice",
abstract = "Volumetric analysis of brain ventricle (BV) structure is a key tool in the study of central nervous system development in embryonic mice. High-frequency ultrasound (HFU) is the only non-invasive, real-time modality available for rapid volumetric imaging of embryos in utero. However, manual segmentation of the BV from HFU volumes is tedious, time-consuming, and requires specialized expertise. In this paper, we propose a novel deep learning based BV segmentation system for whole-body HFU images of mouse embryos. Our fully automated system consists of two modules: localization and segmentation. It first applies a volumetric convolutional neural network on a 3D sliding window over the entire volume to identify a 3D bounding box containing the entire BV. It then employs a fully convolutional network to segment the detected bounding box into BV and background. The system achieves a Dice Similarity Coefficient (DSC) of 0.8956 for BV segmentation on an unseen 111 HFU volume test set surpassing the previous state-of-the-art method (DSC of 0.7119) by a margin of 25{\%}.",
author = "Ziming Qiu and Jack Langerman and Nitin Nair and Orlando Aristizabal and Jonathan Mamou and Turnbull, {Daniel H.} and Jeffrey Ketterling and Yao Wang",
year = "2019",
month = "1",
day = "16",
doi = "10.1109/SPMB.2018.8615610",
language = "English (US)",
series = "2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
booktitle = "2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings",

}

TY - GEN

T1 - Deep Bv

T2 - A Fully Automated System for Brain Ventricle Localization and Segmentation In 3D Ultrasound Images of Embryonic Mice

AU - Qiu, Ziming

AU - Langerman, Jack

AU - Nair, Nitin

AU - Aristizabal, Orlando

AU - Mamou, Jonathan

AU - Turnbull, Daniel H.

AU - Ketterling, Jeffrey

AU - Wang, Yao

PY - 2019/1/16

Y1 - 2019/1/16

N2 - Volumetric analysis of brain ventricle (BV) structure is a key tool in the study of central nervous system development in embryonic mice. High-frequency ultrasound (HFU) is the only non-invasive, real-time modality available for rapid volumetric imaging of embryos in utero. However, manual segmentation of the BV from HFU volumes is tedious, time-consuming, and requires specialized expertise. In this paper, we propose a novel deep learning based BV segmentation system for whole-body HFU images of mouse embryos. Our fully automated system consists of two modules: localization and segmentation. It first applies a volumetric convolutional neural network on a 3D sliding window over the entire volume to identify a 3D bounding box containing the entire BV. It then employs a fully convolutional network to segment the detected bounding box into BV and background. The system achieves a Dice Similarity Coefficient (DSC) of 0.8956 for BV segmentation on an unseen 111 HFU volume test set surpassing the previous state-of-the-art method (DSC of 0.7119) by a margin of 25%.

AB - Volumetric analysis of brain ventricle (BV) structure is a key tool in the study of central nervous system development in embryonic mice. High-frequency ultrasound (HFU) is the only non-invasive, real-time modality available for rapid volumetric imaging of embryos in utero. However, manual segmentation of the BV from HFU volumes is tedious, time-consuming, and requires specialized expertise. In this paper, we propose a novel deep learning based BV segmentation system for whole-body HFU images of mouse embryos. Our fully automated system consists of two modules: localization and segmentation. It first applies a volumetric convolutional neural network on a 3D sliding window over the entire volume to identify a 3D bounding box containing the entire BV. It then employs a fully convolutional network to segment the detected bounding box into BV and background. The system achieves a Dice Similarity Coefficient (DSC) of 0.8956 for BV segmentation on an unseen 111 HFU volume test set surpassing the previous state-of-the-art method (DSC of 0.7119) by a margin of 25%.

UR - http://www.scopus.com/inward/record.url?scp=85062082672&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062082672&partnerID=8YFLogxK

U2 - 10.1109/SPMB.2018.8615610

DO - 10.1109/SPMB.2018.8615610

M3 - Conference contribution

AN - SCOPUS:85062082672

T3 - 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings

BT - 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -