Variational Deep Learning for Low-Dose Computed Tomography

Erich Kobler, Matthew Muckley, Baiyu Chen, Florian Knoll, Kerstin Hammernik, Thomas Pock, Daniel Sodickson, Ricardo Otazo

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

Abstract

In this work, we propose a learning-based variational network (VN) approach for reconstruction of low-dose 3D computed tomography data. We focus on two methods to decrease the radiation dose: (1) x-ray tube current reduction, which reduces the signal-to-noise ratio, and (2) x-ray beam interruption, which undersamples data and results in images with aliasing artifacts. While the learned VN denoises the current-reduced images in the first case, it reconstructs the undersampled data in the second case. Different VNs for denoising and reconstruction are trained on a single clinical 3D abdominal data set. The VNs are compared against state-of-the-art model-based denoising and sparse reconstruction techniques on a different clinical abdominal 3D data set with 4-fold dose reduction. Our results suggest that the proposed VNs enable higher radiation dose reductions and/or increase the image quality for a given dose.

LanguageEnglish
Title of host publication2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers
Pages6687-6691
Number of pages5
Volume2018-April
ISBN (Print)9781538646588
DOIs
StatusPublished - 10 Sep 2018
Event2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Calgary, Canada
Duration: 15 Apr 201820 Apr 2018

Conference

Conference2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018
CountryCanada
CityCalgary
Period15/04/1820/04/18

Fingerprint

Tomography
Dosimetry
X rays
Image quality
Signal to noise ratio
Deep learning

Keywords

  • Compressed sensing
  • Computed tomography
  • Machine learning
  • Medical imaging
  • Variational networks

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Kobler, E., Muckley, M., Chen, B., Knoll, F., Hammernik, K., Pock, T., ... Otazo, R. (2018). Variational Deep Learning for Low-Dose Computed Tomography. In 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings (Vol. 2018-April, pp. 6687-6691). [8462312] Institute of Electrical and Electronics Engineers. DOI: 10.1109/ICASSP.2018.8462312

Variational Deep Learning for Low-Dose Computed Tomography. / Kobler, Erich; Muckley, Matthew; Chen, Baiyu; Knoll, Florian; Hammernik, Kerstin; Pock, Thomas; Sodickson, Daniel; Otazo, Ricardo.

2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April Institute of Electrical and Electronics Engineers, 2018. p. 6687-6691 8462312.

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

Kobler, E, Muckley, M, Chen, B, Knoll, F, Hammernik, K, Pock, T, Sodickson, D & Otazo, R 2018, Variational Deep Learning for Low-Dose Computed Tomography. in 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. vol. 2018-April, 8462312, Institute of Electrical and Electronics Engineers, pp. 6687-6691, 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018, Calgary, Canada, 15/04/18. DOI: 10.1109/ICASSP.2018.8462312
Kobler E, Muckley M, Chen B, Knoll F, Hammernik K, Pock T et al. Variational Deep Learning for Low-Dose Computed Tomography. In 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April. Institute of Electrical and Electronics Engineers. 2018. p. 6687-6691. 8462312. Available from, DOI: 10.1109/ICASSP.2018.8462312
Kobler, Erich ; Muckley, Matthew ; Chen, Baiyu ; Knoll, Florian ; Hammernik, Kerstin ; Pock, Thomas ; Sodickson, Daniel ; Otazo, Ricardo. / Variational Deep Learning for Low-Dose Computed Tomography. 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings. Vol. 2018-April Institute of Electrical and Electronics Engineers, 2018. pp. 6687-6691
@inproceedings{cd2d3a251a914f7982f09a913b78c5cd,
title = "Variational Deep Learning for Low-Dose Computed Tomography",
abstract = "In this work, we propose a learning-based variational network (VN) approach for reconstruction of low-dose 3D computed tomography data. We focus on two methods to decrease the radiation dose: (1) x-ray tube current reduction, which reduces the signal-to-noise ratio, and (2) x-ray beam interruption, which undersamples data and results in images with aliasing artifacts. While the learned VN denoises the current-reduced images in the first case, it reconstructs the undersampled data in the second case. Different VNs for denoising and reconstruction are trained on a single clinical 3D abdominal data set. The VNs are compared against state-of-the-art model-based denoising and sparse reconstruction techniques on a different clinical abdominal 3D data set with 4-fold dose reduction. Our results suggest that the proposed VNs enable higher radiation dose reductions and/or increase the image quality for a given dose.",
keywords = "Compressed sensing, Computed tomography, Machine learning, Medical imaging, Variational networks",
author = "Erich Kobler and Matthew Muckley and Baiyu Chen and Florian Knoll and Kerstin Hammernik and Thomas Pock and Daniel Sodickson and Ricardo Otazo",
year = "2018",
month = "9",
day = "10",
doi = "10.1109/ICASSP.2018.8462312",
language = "English",
isbn = "9781538646588",
volume = "2018-April",
pages = "6687--6691",
booktitle = "2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers",
address = "United States",

}

TY - GEN

T1 - Variational Deep Learning for Low-Dose Computed Tomography

AU - Kobler,Erich

AU - Muckley,Matthew

AU - Chen,Baiyu

AU - Knoll,Florian

AU - Hammernik,Kerstin

AU - Pock,Thomas

AU - Sodickson,Daniel

AU - Otazo,Ricardo

PY - 2018/9/10

Y1 - 2018/9/10

N2 - In this work, we propose a learning-based variational network (VN) approach for reconstruction of low-dose 3D computed tomography data. We focus on two methods to decrease the radiation dose: (1) x-ray tube current reduction, which reduces the signal-to-noise ratio, and (2) x-ray beam interruption, which undersamples data and results in images with aliasing artifacts. While the learned VN denoises the current-reduced images in the first case, it reconstructs the undersampled data in the second case. Different VNs for denoising and reconstruction are trained on a single clinical 3D abdominal data set. The VNs are compared against state-of-the-art model-based denoising and sparse reconstruction techniques on a different clinical abdominal 3D data set with 4-fold dose reduction. Our results suggest that the proposed VNs enable higher radiation dose reductions and/or increase the image quality for a given dose.

AB - In this work, we propose a learning-based variational network (VN) approach for reconstruction of low-dose 3D computed tomography data. We focus on two methods to decrease the radiation dose: (1) x-ray tube current reduction, which reduces the signal-to-noise ratio, and (2) x-ray beam interruption, which undersamples data and results in images with aliasing artifacts. While the learned VN denoises the current-reduced images in the first case, it reconstructs the undersampled data in the second case. Different VNs for denoising and reconstruction are trained on a single clinical 3D abdominal data set. The VNs are compared against state-of-the-art model-based denoising and sparse reconstruction techniques on a different clinical abdominal 3D data set with 4-fold dose reduction. Our results suggest that the proposed VNs enable higher radiation dose reductions and/or increase the image quality for a given dose.

KW - Compressed sensing

KW - Computed tomography

KW - Machine learning

KW - Medical imaging

KW - Variational networks

UR - http://www.scopus.com/inward/record.url?scp=85054201849&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2018.8462312

DO - 10.1109/ICASSP.2018.8462312

M3 - Conference contribution

SN - 9781538646588

VL - 2018-April

SP - 6687

EP - 6691

BT - 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings

PB - Institute of Electrical and Electronics Engineers

ER -