Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks

Lipeng Xie; Jayaram K. Udupa; Yubing Tong; Drew A. Torigian; Zihan Huang; Rachel M. Kogan; Jennifer Ben Nathan; David Wootton; Kokren Choy; Sanghun Sin; Mark E. Wagshul; Raanan Arens

doi:10.1117/12.2581974

Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks

Lipeng Xie, Jayaram K. Udupa, Yubing Tong, Drew A. Torigian, Zihan Huang, Rachel M. Kogan, Jennifer Ben Nathan, David Wootton, Kokren Choy, Sanghun Sin, Mark E. Wagshul, Raanan Arens

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

4 Scopus citations

Abstract

Upper airway segmentation in static and dynamic MRI is a prerequisite step for quantitative analysis in patients with disorders such as obstructive sleep apnea. Recently, some semi-Automatic methods have been proposed with high segmentation accuracy. However, the low efficiency of such methods makes it difficult to implement for the processing of large numbers of MRI datasets. Therefore, a fully automatic upper airway segmentation approach is needed. In this paper, we present a novel automatic upper airway segmentation approach based on convolutional neural networks. Firstly, we utilize the U-Net network as the basic model for learning the multi-scale feature from adjacent image slices and predicting the pixel-wise label in MRI. In particular, we train three networks with the same structure for segmenting the pharynx/larynx and nasal cavity separately in axial static 3D MRI and axial dynamic 2D MRI. The visualization and quantitative results demonstrate that our approach can be applied to various MRI acquisition protocols with high accuracy and stability.

Original language	English (US)
Title of host publication	Medical Imaging 2021
Subtitle of host publication	Biomedical Applications in Molecular, Structural, and Functional Imaging
Editors	Barjor S. Gimi, Andrzej Krol
Publisher	SPIE
ISBN (Electronic)	9781510640290
DOIs	https://doi.org/10.1117/12.2581974
State	Published - 2021
Event	Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging - Virtual, Online, United States Duration: Feb 15 2021 → Feb 19 2021

Publication series

Name	Progress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume	11600
ISSN (Print)	1605-7422

Conference

Conference	Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging
Country/Territory	United States
City	Virtual, Online
Period	2/15/21 → 2/19/21

Keywords

Upper airway
convolutional neural network
dynamic MRI
segmentation
sleep apnea
static MRI

ASJC Scopus subject areas

Electronic, Optical and Magnetic Materials
Biomaterials
Atomic and Molecular Physics, and Optics
Radiology Nuclear Medicine and imaging

Access to Document

10.1117/12.2581974

Cite this

Xie, L., Udupa, J. K., Tong, Y., Torigian, D. A., Huang, Z., Kogan, R. M., Nathan, J. B., Wootton, D., Choy, K., Sin, S., Wagshul, M. E., & Arens, R. (2021). Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks. In B. S. Gimi, & A. Krol (Eds.), Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging [116000J] (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 11600). SPIE. https://doi.org/10.1117/12.2581974

Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks. / Xie, Lipeng; Udupa, Jayaram K.; Tong, Yubing et al.
Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging. ed. / Barjor S. Gimi; Andrzej Krol. SPIE, 2021. 116000J (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 11600).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Xie, L, Udupa, JK, Tong, Y, Torigian, DA, Huang, Z, Kogan, RM, Nathan, JB, Wootton, D, Choy, K, Sin, S , Wagshul, ME & Arens, R 2021, Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks. in BS Gimi & A Krol (eds), Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging., 116000J, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 11600, SPIE, Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging, Virtual, Online, United States, 2/15/21. https://doi.org/10.1117/12.2581974

Xie L, Udupa JK, Tong Y, Torigian DA, Huang Z, Kogan RM et al. Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks. In Gimi BS, Krol A, editors, Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging. SPIE. 2021. 116000J. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE). doi: 10.1117/12.2581974

Xie, Lipeng ; Udupa, Jayaram K. ; Tong, Yubing et al. / Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks. Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging. editor / Barjor S. Gimi ; Andrzej Krol. SPIE, 2021. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE).

@inproceedings{2fe434afa8334aa9a5e1099a041c593e,

title = "Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks",

abstract = "Upper airway segmentation in static and dynamic MRI is a prerequisite step for quantitative analysis in patients with disorders such as obstructive sleep apnea. Recently, some semi-Automatic methods have been proposed with high segmentation accuracy. However, the low efficiency of such methods makes it difficult to implement for the processing of large numbers of MRI datasets. Therefore, a fully automatic upper airway segmentation approach is needed. In this paper, we present a novel automatic upper airway segmentation approach based on convolutional neural networks. Firstly, we utilize the U-Net network as the basic model for learning the multi-scale feature from adjacent image slices and predicting the pixel-wise label in MRI. In particular, we train three networks with the same structure for segmenting the pharynx/larynx and nasal cavity separately in axial static 3D MRI and axial dynamic 2D MRI. The visualization and quantitative results demonstrate that our approach can be applied to various MRI acquisition protocols with high accuracy and stability. ",

keywords = "Upper airway, convolutional neural network, dynamic MRI, segmentation, sleep apnea, static MRI",

author = "Lipeng Xie and Udupa, {Jayaram K.} and Yubing Tong and Torigian, {Drew A.} and Zihan Huang and Kogan, {Rachel M.} and Nathan, {Jennifer Ben} and David Wootton and Kokren Choy and Sanghun Sin and Wagshul, {Mark E.} and Raanan Arens",

note = "Funding Information: This work is supported by an NIH grant 1R01HL130468-A1. Publisher Copyright: {\textcopyright} COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.; Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging ; Conference date: 15-02-2021 Through 19-02-2021",

year = "2021",

doi = "10.1117/12.2581974",

language = "English (US)",

series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

publisher = "SPIE",

editor = "Gimi, {Barjor S.} and Andrzej Krol",

booktitle = "Medical Imaging 2021",

}

TY - GEN

T1 - Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks

AU - Xie, Lipeng

AU - Udupa, Jayaram K.

AU - Tong, Yubing

AU - Torigian, Drew A.

AU - Huang, Zihan

AU - Kogan, Rachel M.

AU - Nathan, Jennifer Ben

AU - Wootton, David

AU - Choy, Kokren

AU - Sin, Sanghun

AU - Wagshul, Mark E.

AU - Arens, Raanan

N1 - Funding Information: This work is supported by an NIH grant 1R01HL130468-A1. Publisher Copyright: © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

PY - 2021

Y1 - 2021

N2 - Upper airway segmentation in static and dynamic MRI is a prerequisite step for quantitative analysis in patients with disorders such as obstructive sleep apnea. Recently, some semi-Automatic methods have been proposed with high segmentation accuracy. However, the low efficiency of such methods makes it difficult to implement for the processing of large numbers of MRI datasets. Therefore, a fully automatic upper airway segmentation approach is needed. In this paper, we present a novel automatic upper airway segmentation approach based on convolutional neural networks. Firstly, we utilize the U-Net network as the basic model for learning the multi-scale feature from adjacent image slices and predicting the pixel-wise label in MRI. In particular, we train three networks with the same structure for segmenting the pharynx/larynx and nasal cavity separately in axial static 3D MRI and axial dynamic 2D MRI. The visualization and quantitative results demonstrate that our approach can be applied to various MRI acquisition protocols with high accuracy and stability.

AB - Upper airway segmentation in static and dynamic MRI is a prerequisite step for quantitative analysis in patients with disorders such as obstructive sleep apnea. Recently, some semi-Automatic methods have been proposed with high segmentation accuracy. However, the low efficiency of such methods makes it difficult to implement for the processing of large numbers of MRI datasets. Therefore, a fully automatic upper airway segmentation approach is needed. In this paper, we present a novel automatic upper airway segmentation approach based on convolutional neural networks. Firstly, we utilize the U-Net network as the basic model for learning the multi-scale feature from adjacent image slices and predicting the pixel-wise label in MRI. In particular, we train three networks with the same structure for segmenting the pharynx/larynx and nasal cavity separately in axial static 3D MRI and axial dynamic 2D MRI. The visualization and quantitative results demonstrate that our approach can be applied to various MRI acquisition protocols with high accuracy and stability.

KW - Upper airway

KW - convolutional neural network

KW - dynamic MRI

KW - segmentation

KW - sleep apnea

KW - static MRI

UR - http://www.scopus.com/inward/record.url?scp=85103212505&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85103212505&partnerID=8YFLogxK

U2 - 10.1117/12.2581974

DO - 10.1117/12.2581974

M3 - Conference contribution

AN - SCOPUS:85103212505

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2021

A2 - Gimi, Barjor S.

A2 - Krol, Andrzej

PB - SPIE

T2 - Medical Imaging 2021: Biomedical Applications in Molecular, Structural, and Functional Imaging

Y2 - 15 February 2021 through 19 February 2021

ER -

Automatic upper airway segmentation in static and dynamic MRI via deep convolutional neural networks

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this