TU‐B‐201B‐06: A Constrained Optimization Algorithm for CT Metal Artifact Reduction

X. Zhang, J. Wang, L. Xing

Research output: Contribution to journalArticle

Abstract

Purpose: The CT streak artifacts caused by metal implants limit various applications of CT imaging, such as target delineation and accurate dose calculations in radiation therapy. Both interpolation‐based one‐shot algorithms and iterative reconstruction algorithms have certain limitations in efficiently suppressing the artifacts. The purpose of this work is to develop an effective and robust iterative algorithm to minimize the metal artifacts and achieve clinically acceptable CT images. Methods and materials: A penalized‐weighted‐least‐squares method is first used to accurately identify the metal objects in image space. The binary metal image is then projected to sinogram to segment out metal corrupted projection data. A constrained‐optimization based iterative algorithm is then applied to the remaining data to reconstruct the CT image. The algorithm minimizes the total variation (TV) or an anisotropically penalized smoothness (PS) of the image, subject to a data fidelity constraint, with image non‐negativity enforced. The constrained minimization problem is solved through the alternation of projection onto convex sets (POCS) and steepest gradient descent of the objective function. The algorithm is evaluated using two numerical phantoms (shepp‐logan and QA, 350×350 (in pixels)) and experimental CatPhan®600 phantom. The experimental projection data was acquired using cone‐beam CT lose dose protocol (10mA/10ms), including 680 views for a full 360° rotation. Results: Both simulation and experimental studies showed that the constrained optimization algorithm has superior performance compared with analytical FDK algorithm. It significantly suppressed metal artifacts even in the presence of noise and system imperfections. The FWHM comparisons on CatPhan®600 suggested that image reconstructed by constrained optimization has a higher resolution than FDK reconstruction. Conclusion: The proposed constrained optimization algorithm can be used to significantly reduce metal artifacts and produce clinically acceptable image with relatively low dose. The proposed algorithm can also be generated to other incomplete/missing data problems in systems involving x‐ray source trajectories.

Original languageEnglish (US)
Number of pages1
JournalMedical Physics
Volume37
Issue number6
DOIs
StatePublished - 2010

Fingerprint

Artifacts
Metals
Noise
Radiotherapy
X-Rays

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging

Cite this

TU‐B‐201B‐06 : A Constrained Optimization Algorithm for CT Metal Artifact Reduction. / Zhang, X.; Wang, J.; Xing, L.

In: Medical Physics, Vol. 37, No. 6, 2010.

Research output: Contribution to journalArticle

@article{e6abd7c3c8164fe89728523435bf697b,
title = "TU‐B‐201B‐06: A Constrained Optimization Algorithm for CT Metal Artifact Reduction",
abstract = "Purpose: The CT streak artifacts caused by metal implants limit various applications of CT imaging, such as target delineation and accurate dose calculations in radiation therapy. Both interpolation‐based one‐shot algorithms and iterative reconstruction algorithms have certain limitations in efficiently suppressing the artifacts. The purpose of this work is to develop an effective and robust iterative algorithm to minimize the metal artifacts and achieve clinically acceptable CT images. Methods and materials: A penalized‐weighted‐least‐squares method is first used to accurately identify the metal objects in image space. The binary metal image is then projected to sinogram to segment out metal corrupted projection data. A constrained‐optimization based iterative algorithm is then applied to the remaining data to reconstruct the CT image. The algorithm minimizes the total variation (TV) or an anisotropically penalized smoothness (PS) of the image, subject to a data fidelity constraint, with image non‐negativity enforced. The constrained minimization problem is solved through the alternation of projection onto convex sets (POCS) and steepest gradient descent of the objective function. The algorithm is evaluated using two numerical phantoms (shepp‐logan and QA, 350×350 (in pixels)) and experimental CatPhan{\circledR}600 phantom. The experimental projection data was acquired using cone‐beam CT lose dose protocol (10mA/10ms), including 680 views for a full 360° rotation. Results: Both simulation and experimental studies showed that the constrained optimization algorithm has superior performance compared with analytical FDK algorithm. It significantly suppressed metal artifacts even in the presence of noise and system imperfections. The FWHM comparisons on CatPhan{\circledR}600 suggested that image reconstructed by constrained optimization has a higher resolution than FDK reconstruction. Conclusion: The proposed constrained optimization algorithm can be used to significantly reduce metal artifacts and produce clinically acceptable image with relatively low dose. The proposed algorithm can also be generated to other incomplete/missing data problems in systems involving x‐ray source trajectories.",
author = "X. Zhang and J. Wang and L. Xing",
year = "2010",
doi = "10.1118/1.3469202",
language = "English (US)",
volume = "37",
journal = "Medical Physics",
issn = "0094-2405",
publisher = "AAPM - American Association of Physicists in Medicine",
number = "6",

}

TY - JOUR

T1 - TU‐B‐201B‐06

T2 - A Constrained Optimization Algorithm for CT Metal Artifact Reduction

AU - Zhang, X.

AU - Wang, J.

AU - Xing, L.

PY - 2010

Y1 - 2010

N2 - Purpose: The CT streak artifacts caused by metal implants limit various applications of CT imaging, such as target delineation and accurate dose calculations in radiation therapy. Both interpolation‐based one‐shot algorithms and iterative reconstruction algorithms have certain limitations in efficiently suppressing the artifacts. The purpose of this work is to develop an effective and robust iterative algorithm to minimize the metal artifacts and achieve clinically acceptable CT images. Methods and materials: A penalized‐weighted‐least‐squares method is first used to accurately identify the metal objects in image space. The binary metal image is then projected to sinogram to segment out metal corrupted projection data. A constrained‐optimization based iterative algorithm is then applied to the remaining data to reconstruct the CT image. The algorithm minimizes the total variation (TV) or an anisotropically penalized smoothness (PS) of the image, subject to a data fidelity constraint, with image non‐negativity enforced. The constrained minimization problem is solved through the alternation of projection onto convex sets (POCS) and steepest gradient descent of the objective function. The algorithm is evaluated using two numerical phantoms (shepp‐logan and QA, 350×350 (in pixels)) and experimental CatPhan®600 phantom. The experimental projection data was acquired using cone‐beam CT lose dose protocol (10mA/10ms), including 680 views for a full 360° rotation. Results: Both simulation and experimental studies showed that the constrained optimization algorithm has superior performance compared with analytical FDK algorithm. It significantly suppressed metal artifacts even in the presence of noise and system imperfections. The FWHM comparisons on CatPhan®600 suggested that image reconstructed by constrained optimization has a higher resolution than FDK reconstruction. Conclusion: The proposed constrained optimization algorithm can be used to significantly reduce metal artifacts and produce clinically acceptable image with relatively low dose. The proposed algorithm can also be generated to other incomplete/missing data problems in systems involving x‐ray source trajectories.

AB - Purpose: The CT streak artifacts caused by metal implants limit various applications of CT imaging, such as target delineation and accurate dose calculations in radiation therapy. Both interpolation‐based one‐shot algorithms and iterative reconstruction algorithms have certain limitations in efficiently suppressing the artifacts. The purpose of this work is to develop an effective and robust iterative algorithm to minimize the metal artifacts and achieve clinically acceptable CT images. Methods and materials: A penalized‐weighted‐least‐squares method is first used to accurately identify the metal objects in image space. The binary metal image is then projected to sinogram to segment out metal corrupted projection data. A constrained‐optimization based iterative algorithm is then applied to the remaining data to reconstruct the CT image. The algorithm minimizes the total variation (TV) or an anisotropically penalized smoothness (PS) of the image, subject to a data fidelity constraint, with image non‐negativity enforced. The constrained minimization problem is solved through the alternation of projection onto convex sets (POCS) and steepest gradient descent of the objective function. The algorithm is evaluated using two numerical phantoms (shepp‐logan and QA, 350×350 (in pixels)) and experimental CatPhan®600 phantom. The experimental projection data was acquired using cone‐beam CT lose dose protocol (10mA/10ms), including 680 views for a full 360° rotation. Results: Both simulation and experimental studies showed that the constrained optimization algorithm has superior performance compared with analytical FDK algorithm. It significantly suppressed metal artifacts even in the presence of noise and system imperfections. The FWHM comparisons on CatPhan®600 suggested that image reconstructed by constrained optimization has a higher resolution than FDK reconstruction. Conclusion: The proposed constrained optimization algorithm can be used to significantly reduce metal artifacts and produce clinically acceptable image with relatively low dose. The proposed algorithm can also be generated to other incomplete/missing data problems in systems involving x‐ray source trajectories.

UR - http://www.scopus.com/inward/record.url?scp=85024797459&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85024797459&partnerID=8YFLogxK

U2 - 10.1118/1.3469202

DO - 10.1118/1.3469202

M3 - Article

AN - SCOPUS:85024797459

VL - 37

JO - Medical Physics

JF - Medical Physics

SN - 0094-2405

IS - 6

ER -