Pixel-wise bias approximation and correction for convolutional neural network noise reduction in CT

Nathan R. Huber; Hao Gong; Thomas Huber; David Campeau; Scott Hsieh; Shuai Leng; Lifeng Yu; Cynthia McCollough

doi:10.1117/12.2612703

Pixel-wise bias approximation and correction for convolutional neural network noise reduction in CT

Nathan R. Huber, Hao Gong, Thomas Huber, David Campeau, Scott Hsieh, Shuai Leng, Lifeng Yu, Cynthia McCollough

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

Abstract

This study introduces a framework to approximate the bias inflicted by CNN noise reduction of CT exams. First, CNN noise reduction was used to approximate the noise-free image and noise-only image of a CT scan. The noise and signal were then recombined with spatial decoupling to simulate an ensemble of 100 images. CNN noise reduction was applied to the simulated ensemble and pixel-wise bias calculated. This bias approximation technique was validated within natural images and phantoms. The technique was then tested on ten whole-body-low-dose CT (WBLD-CT) patient exams. Bias correction led to improved contrast of lung and bone structures.

Original language	English (US)
Title of host publication	Medical Imaging 2022
Subtitle of host publication	Physics of Medical Imaging
Editors	Wei Zhao, Lifeng Yu
Publisher	SPIE
ISBN (Electronic)	9781510649378
DOIs	https://doi.org/10.1117/12.2612703
State	Published - 2022
Event	Medical Imaging 2022: Physics of Medical Imaging - Virtual, Online Duration: Mar 21 2022 → Mar 27 2022

Publication series

Name	Progress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume	12031
ISSN (Print)	1605-7422

Conference

Conference	Medical Imaging 2022: Physics of Medical Imaging
City	Virtual, Online
Period	3/21/22 → 3/27/22

Keywords

Bias
Bootstrap Approximation
CNN Denoising
Uncertainty
Whole-Body-Low-Dose CT

ASJC Scopus subject areas

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

Access to Document

10.1117/12.2612703

Cite this

Huber, N. R., Gong, H., Huber, T., Campeau, D., Hsieh, S., Leng, S., Yu, L., & McCollough, C. (2022). Pixel-wise bias approximation and correction for convolutional neural network noise reduction in CT. In W. Zhao, & L. Yu (Eds.), Medical Imaging 2022: Physics of Medical Imaging Article 120311C (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 12031). SPIE. https://doi.org/10.1117/12.2612703

Pixel-wise bias approximation and correction for convolutional neural network noise reduction in CT. / Huber, Nathan R.; Gong, Hao; Huber, Thomas et al.
Medical Imaging 2022: Physics of Medical Imaging. ed. / Wei Zhao; Lifeng Yu. SPIE, 2022. 120311C (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 12031).

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

Huber, NR, Gong, H, Huber, T, Campeau, D, Hsieh, S , Leng, S , Yu, L & McCollough, C 2022, Pixel-wise bias approximation and correction for convolutional neural network noise reduction in CT. in W Zhao & L Yu (eds), Medical Imaging 2022: Physics of Medical Imaging., 120311C, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 12031, SPIE, Medical Imaging 2022: Physics of Medical Imaging, Virtual, Online, 3/21/22. https://doi.org/10.1117/12.2612703

@inproceedings{a3b523c8a9064d53bcaedd5692690654,

title = "Pixel-wise bias approximation and correction for convolutional neural network noise reduction in CT",

abstract = "This study introduces a framework to approximate the bias inflicted by CNN noise reduction of CT exams. First, CNN noise reduction was used to approximate the noise-free image and noise-only image of a CT scan. The noise and signal were then recombined with spatial decoupling to simulate an ensemble of 100 images. CNN noise reduction was applied to the simulated ensemble and pixel-wise bias calculated. This bias approximation technique was validated within natural images and phantoms. The technique was then tested on ten whole-body-low-dose CT (WBLD-CT) patient exams. Bias correction led to improved contrast of lung and bone structures.",

keywords = "Bias, Bootstrap Approximation, CNN Denoising, Uncertainty, Whole-Body-Low-Dose CT",

author = "Huber, {Nathan R.} and Hao Gong and Thomas Huber and David Campeau and Scott Hsieh and Shuai Leng and Lifeng Yu and Cynthia McCollough",

note = "Publisher Copyright: {\textcopyright} 2022 SPIE.; Medical Imaging 2022: Physics of Medical Imaging ; Conference date: 21-03-2022 Through 27-03-2022",

year = "2022",

doi = "10.1117/12.2612703",

language = "English (US)",

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

publisher = "SPIE",

editor = "Wei Zhao and Lifeng Yu",

booktitle = "Medical Imaging 2022",

}

TY - GEN

T1 - Pixel-wise bias approximation and correction for convolutional neural network noise reduction in CT

AU - Huber, Nathan R.

AU - Gong, Hao

AU - Huber, Thomas

AU - Campeau, David

AU - Hsieh, Scott

AU - Leng, Shuai

AU - Yu, Lifeng

AU - McCollough, Cynthia

PY - 2022

Y1 - 2022

N2 - This study introduces a framework to approximate the bias inflicted by CNN noise reduction of CT exams. First, CNN noise reduction was used to approximate the noise-free image and noise-only image of a CT scan. The noise and signal were then recombined with spatial decoupling to simulate an ensemble of 100 images. CNN noise reduction was applied to the simulated ensemble and pixel-wise bias calculated. This bias approximation technique was validated within natural images and phantoms. The technique was then tested on ten whole-body-low-dose CT (WBLD-CT) patient exams. Bias correction led to improved contrast of lung and bone structures.

AB - This study introduces a framework to approximate the bias inflicted by CNN noise reduction of CT exams. First, CNN noise reduction was used to approximate the noise-free image and noise-only image of a CT scan. The noise and signal were then recombined with spatial decoupling to simulate an ensemble of 100 images. CNN noise reduction was applied to the simulated ensemble and pixel-wise bias calculated. This bias approximation technique was validated within natural images and phantoms. The technique was then tested on ten whole-body-low-dose CT (WBLD-CT) patient exams. Bias correction led to improved contrast of lung and bone structures.

KW - Bias

KW - Bootstrap Approximation

KW - CNN Denoising

KW - Uncertainty

KW - Whole-Body-Low-Dose CT

UR - http://www.scopus.com/inward/record.url?scp=85131197047&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85131197047&partnerID=8YFLogxK

U2 - 10.1117/12.2612703

DO - 10.1117/12.2612703

M3 - Conference contribution

AN - SCOPUS:85131197047

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

BT - Medical Imaging 2022

A2 - Zhao, Wei

A2 - Yu, Lifeng

PB - SPIE

T2 - Medical Imaging 2022: Physics of Medical Imaging

Y2 - 21 March 2022 through 27 March 2022

ER -

Pixel-wise bias approximation and correction for convolutional neural network noise reduction in CT

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this