Intelligent Parameter Tuning in Optimization-based Iterative CT Reconstruction via Deep Reinforcement Learning

Chenyang Shen, Yesenia Gonzalez, Liyuan Chen, Steve B. Jiang, Xun Jia

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

A number of image-processing problems can be formulated as optimization problems. The objective function typically contains several terms specifically designed for different purposes. Parameters in front of these terms are used to control the relative importance among them. It is of critical importance to adjust these parameters, as quality of the solution depends on their values. Tuning parameters is a relatively straightforward task for a human, as one can intuitively determine the direction of parameter adjustment based on the solution quality. Yet manual parameter tuning is not only tedious in many cases, but becomes impractical when a number of parameters exist in a problem. Aiming at solving this problem, this paper proposes an approach that employs deep reinforcement learning to train a system that can automatically adjust parameters in a humanlike manner. We demonstrate our idea in an example problem of optimization-based iterative CT reconstruction with a pixelwise total-variation regularization term. We set up a Parameter- Tuning Policy Network (PTPN), which maps a CT image patch to an output that specifies the direction and amplitude by which the parameter at the patch center is adjusted. We train the PTPN via an end-to-end reinforcement learning procedure.We demonstrate that under the guidance of the trained PTPN, reconstructed CT images attain quality similar or better than those reconstructed with manually tuned parameters.

Original languageEnglish (US)
JournalIEEE Transactions on Medical Imaging
DOIs
StateAccepted/In press - Apr 4 2018

Fingerprint

Reinforcement learning
Tuning
Learning
Social Adjustment
Electronic guidance systems
Image quality
Image processing
Reinforcement (Psychology)
Direction compound

Keywords

  • Computed tomography
  • Image quality
  • Image reconstruction
  • Image reconstruction -iterative methods
  • Inverse methods
  • Machine learning
  • Machine learning
  • Optimization
  • Radio frequency
  • Tuning
  • x-ray imaging and computed tomography

ASJC Scopus subject areas

  • Software
  • Radiological and Ultrasound Technology
  • Computer Science Applications
  • Electrical and Electronic Engineering

Cite this

Intelligent Parameter Tuning in Optimization-based Iterative CT Reconstruction via Deep Reinforcement Learning. / Shen, Chenyang; Gonzalez, Yesenia; Chen, Liyuan; Jiang, Steve B.; Jia, Xun.

In: IEEE Transactions on Medical Imaging, 04.04.2018.

Research output: Contribution to journalArticle

@article{068c5891825d4fda84318e88193147ce,
title = "Intelligent Parameter Tuning in Optimization-based Iterative CT Reconstruction via Deep Reinforcement Learning",
abstract = "A number of image-processing problems can be formulated as optimization problems. The objective function typically contains several terms specifically designed for different purposes. Parameters in front of these terms are used to control the relative importance among them. It is of critical importance to adjust these parameters, as quality of the solution depends on their values. Tuning parameters is a relatively straightforward task for a human, as one can intuitively determine the direction of parameter adjustment based on the solution quality. Yet manual parameter tuning is not only tedious in many cases, but becomes impractical when a number of parameters exist in a problem. Aiming at solving this problem, this paper proposes an approach that employs deep reinforcement learning to train a system that can automatically adjust parameters in a humanlike manner. We demonstrate our idea in an example problem of optimization-based iterative CT reconstruction with a pixelwise total-variation regularization term. We set up a Parameter- Tuning Policy Network (PTPN), which maps a CT image patch to an output that specifies the direction and amplitude by which the parameter at the patch center is adjusted. We train the PTPN via an end-to-end reinforcement learning procedure.We demonstrate that under the guidance of the trained PTPN, reconstructed CT images attain quality similar or better than those reconstructed with manually tuned parameters.",
keywords = "Computed tomography, Image quality, Image reconstruction, Image reconstruction -iterative methods, Inverse methods, Machine learning, Machine learning, Optimization, Radio frequency, Tuning, x-ray imaging and computed tomography",
author = "Chenyang Shen and Yesenia Gonzalez and Liyuan Chen and Jiang, {Steve B.} and Xun Jia",
year = "2018",
month = "4",
day = "4",
doi = "10.1109/TMI.2018.2823679",
language = "English (US)",
journal = "IEEE Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Intelligent Parameter Tuning in Optimization-based Iterative CT Reconstruction via Deep Reinforcement Learning

AU - Shen, Chenyang

AU - Gonzalez, Yesenia

AU - Chen, Liyuan

AU - Jiang, Steve B.

AU - Jia, Xun

PY - 2018/4/4

Y1 - 2018/4/4

N2 - A number of image-processing problems can be formulated as optimization problems. The objective function typically contains several terms specifically designed for different purposes. Parameters in front of these terms are used to control the relative importance among them. It is of critical importance to adjust these parameters, as quality of the solution depends on their values. Tuning parameters is a relatively straightforward task for a human, as one can intuitively determine the direction of parameter adjustment based on the solution quality. Yet manual parameter tuning is not only tedious in many cases, but becomes impractical when a number of parameters exist in a problem. Aiming at solving this problem, this paper proposes an approach that employs deep reinforcement learning to train a system that can automatically adjust parameters in a humanlike manner. We demonstrate our idea in an example problem of optimization-based iterative CT reconstruction with a pixelwise total-variation regularization term. We set up a Parameter- Tuning Policy Network (PTPN), which maps a CT image patch to an output that specifies the direction and amplitude by which the parameter at the patch center is adjusted. We train the PTPN via an end-to-end reinforcement learning procedure.We demonstrate that under the guidance of the trained PTPN, reconstructed CT images attain quality similar or better than those reconstructed with manually tuned parameters.

AB - A number of image-processing problems can be formulated as optimization problems. The objective function typically contains several terms specifically designed for different purposes. Parameters in front of these terms are used to control the relative importance among them. It is of critical importance to adjust these parameters, as quality of the solution depends on their values. Tuning parameters is a relatively straightforward task for a human, as one can intuitively determine the direction of parameter adjustment based on the solution quality. Yet manual parameter tuning is not only tedious in many cases, but becomes impractical when a number of parameters exist in a problem. Aiming at solving this problem, this paper proposes an approach that employs deep reinforcement learning to train a system that can automatically adjust parameters in a humanlike manner. We demonstrate our idea in an example problem of optimization-based iterative CT reconstruction with a pixelwise total-variation regularization term. We set up a Parameter- Tuning Policy Network (PTPN), which maps a CT image patch to an output that specifies the direction and amplitude by which the parameter at the patch center is adjusted. We train the PTPN via an end-to-end reinforcement learning procedure.We demonstrate that under the guidance of the trained PTPN, reconstructed CT images attain quality similar or better than those reconstructed with manually tuned parameters.

KW - Computed tomography

KW - Image quality

KW - Image reconstruction

KW - Image reconstruction -iterative methods

KW - Inverse methods

KW - Machine learning

KW - Machine learning

KW - Optimization

KW - Radio frequency

KW - Tuning

KW - x-ray imaging and computed tomography

UR - http://www.scopus.com/inward/record.url?scp=85045201602&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85045201602&partnerID=8YFLogxK

U2 - 10.1109/TMI.2018.2823679

DO - 10.1109/TMI.2018.2823679

M3 - Article

C2 - 29870371

AN - SCOPUS:85045201602

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

SN - 0278-0062

ER -