WE‐A‐301‐02: Four‐Dimensional Cone Beam CT Reconstruction and Enhancement with a Temporal Non‐Local Means Method

X. Jia, Z. Tian, Y. Lou, S. Jiang

Research output: Contribution to journalArticle

Abstract

Purpose: Four‐dimensional Cone Beam Computer Tomography (4DCBCT) can provide respiratory phase resolved volumetric images in image guided radiation therapy. However, the undersampling of projections in each phase results in obvious streaking artifacts under the conventional FDK algorithm. In this work, we propose a novel iterative 4DCBCT reconstruction algorithm and an enhancement algorithm utilizing temporal nonlocal mean (TNLM) method. Methods: For reconstruction, we introduce a TNLM term in the optimization model. Such a term favors those 4DCBCT images such that any anatomical features at one spatial point in one phase can be found in a nearby spatial point in neighboring phases. As for the image enhancement, 4DCBCT images obtained from the FDK algorithm are enhancement by minimizing the TNLM function while requiring that the solution is close to the FDK results. A forward‐backward splitting algorithm and a Gauss‐Jacobi iteration method are employed to solve the problems. The algorithms are implemented on GPU to achieve a high computational speed. Our algorithms have been tested on a digital NCAT phantom and a clinical patient case. Results: Our reconstruction algorithm and the enhancement algorithm generate visually similar 4DCBCT images, but both better than FDK results. Quantitative evaluations indicate that, compared to the FDK reconstruction results, our reconstruction method improves contrast‐to‐noise‐ratio (CNR) by a factor of 2.56∼3.13 and our enhancement method increases CNR by 2.75∼3.33 times. The enhancement method also removes over 80% of the streak artifacts from the FDK reconstruction results. The total computation time is ∼460 sec and ∼610 sec on an NVIDIA Tesla C1060 GPU card for the reconstruction and the enhancement algorithm, respectively. Conclusions: Our new reconstruction and enhancement algorithms can effectively reduce the image artifacts in 4DCBCT. Comparing the two algorithms, the resulted image qualities are similar. The shorter computation time makes the enhancement algorithm more attractive.

Original languageEnglish (US)
Pages (from-to)3796
Number of pages1
JournalMedical Physics
Volume38
Issue number6
DOIs
StatePublished - 2011

Fingerprint

Cone-Beam Computed Tomography
Artifacts
Image-Guided Radiotherapy
Image Enhancement

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging

Cite this

WE‐A‐301‐02 : Four‐Dimensional Cone Beam CT Reconstruction and Enhancement with a Temporal Non‐Local Means Method. / Jia, X.; Tian, Z.; Lou, Y.; Jiang, S.

In: Medical Physics, Vol. 38, No. 6, 2011, p. 3796.

Research output: Contribution to journalArticle

@article{97b6a92081f14470a7552d425ba3f36f,
title = "WE‐A‐301‐02: Four‐Dimensional Cone Beam CT Reconstruction and Enhancement with a Temporal Non‐Local Means Method",
abstract = "Purpose: Four‐dimensional Cone Beam Computer Tomography (4DCBCT) can provide respiratory phase resolved volumetric images in image guided radiation therapy. However, the undersampling of projections in each phase results in obvious streaking artifacts under the conventional FDK algorithm. In this work, we propose a novel iterative 4DCBCT reconstruction algorithm and an enhancement algorithm utilizing temporal nonlocal mean (TNLM) method. Methods: For reconstruction, we introduce a TNLM term in the optimization model. Such a term favors those 4DCBCT images such that any anatomical features at one spatial point in one phase can be found in a nearby spatial point in neighboring phases. As for the image enhancement, 4DCBCT images obtained from the FDK algorithm are enhancement by minimizing the TNLM function while requiring that the solution is close to the FDK results. A forward‐backward splitting algorithm and a Gauss‐Jacobi iteration method are employed to solve the problems. The algorithms are implemented on GPU to achieve a high computational speed. Our algorithms have been tested on a digital NCAT phantom and a clinical patient case. Results: Our reconstruction algorithm and the enhancement algorithm generate visually similar 4DCBCT images, but both better than FDK results. Quantitative evaluations indicate that, compared to the FDK reconstruction results, our reconstruction method improves contrast‐to‐noise‐ratio (CNR) by a factor of 2.56∼3.13 and our enhancement method increases CNR by 2.75∼3.33 times. The enhancement method also removes over 80{\%} of the streak artifacts from the FDK reconstruction results. The total computation time is ∼460 sec and ∼610 sec on an NVIDIA Tesla C1060 GPU card for the reconstruction and the enhancement algorithm, respectively. Conclusions: Our new reconstruction and enhancement algorithms can effectively reduce the image artifacts in 4DCBCT. Comparing the two algorithms, the resulted image qualities are similar. The shorter computation time makes the enhancement algorithm more attractive.",
author = "X. Jia and Z. Tian and Y. Lou and S. Jiang",
year = "2011",
doi = "10.1118/1.3613286",
language = "English (US)",
volume = "38",
pages = "3796",
journal = "Medical Physics",
issn = "0094-2405",
publisher = "AAPM - American Association of Physicists in Medicine",
number = "6",

}

TY - JOUR

T1 - WE‐A‐301‐02

T2 - Four‐Dimensional Cone Beam CT Reconstruction and Enhancement with a Temporal Non‐Local Means Method

AU - Jia, X.

AU - Tian, Z.

AU - Lou, Y.

AU - Jiang, S.

PY - 2011

Y1 - 2011

N2 - Purpose: Four‐dimensional Cone Beam Computer Tomography (4DCBCT) can provide respiratory phase resolved volumetric images in image guided radiation therapy. However, the undersampling of projections in each phase results in obvious streaking artifacts under the conventional FDK algorithm. In this work, we propose a novel iterative 4DCBCT reconstruction algorithm and an enhancement algorithm utilizing temporal nonlocal mean (TNLM) method. Methods: For reconstruction, we introduce a TNLM term in the optimization model. Such a term favors those 4DCBCT images such that any anatomical features at one spatial point in one phase can be found in a nearby spatial point in neighboring phases. As for the image enhancement, 4DCBCT images obtained from the FDK algorithm are enhancement by minimizing the TNLM function while requiring that the solution is close to the FDK results. A forward‐backward splitting algorithm and a Gauss‐Jacobi iteration method are employed to solve the problems. The algorithms are implemented on GPU to achieve a high computational speed. Our algorithms have been tested on a digital NCAT phantom and a clinical patient case. Results: Our reconstruction algorithm and the enhancement algorithm generate visually similar 4DCBCT images, but both better than FDK results. Quantitative evaluations indicate that, compared to the FDK reconstruction results, our reconstruction method improves contrast‐to‐noise‐ratio (CNR) by a factor of 2.56∼3.13 and our enhancement method increases CNR by 2.75∼3.33 times. The enhancement method also removes over 80% of the streak artifacts from the FDK reconstruction results. The total computation time is ∼460 sec and ∼610 sec on an NVIDIA Tesla C1060 GPU card for the reconstruction and the enhancement algorithm, respectively. Conclusions: Our new reconstruction and enhancement algorithms can effectively reduce the image artifacts in 4DCBCT. Comparing the two algorithms, the resulted image qualities are similar. The shorter computation time makes the enhancement algorithm more attractive.

AB - Purpose: Four‐dimensional Cone Beam Computer Tomography (4DCBCT) can provide respiratory phase resolved volumetric images in image guided radiation therapy. However, the undersampling of projections in each phase results in obvious streaking artifacts under the conventional FDK algorithm. In this work, we propose a novel iterative 4DCBCT reconstruction algorithm and an enhancement algorithm utilizing temporal nonlocal mean (TNLM) method. Methods: For reconstruction, we introduce a TNLM term in the optimization model. Such a term favors those 4DCBCT images such that any anatomical features at one spatial point in one phase can be found in a nearby spatial point in neighboring phases. As for the image enhancement, 4DCBCT images obtained from the FDK algorithm are enhancement by minimizing the TNLM function while requiring that the solution is close to the FDK results. A forward‐backward splitting algorithm and a Gauss‐Jacobi iteration method are employed to solve the problems. The algorithms are implemented on GPU to achieve a high computational speed. Our algorithms have been tested on a digital NCAT phantom and a clinical patient case. Results: Our reconstruction algorithm and the enhancement algorithm generate visually similar 4DCBCT images, but both better than FDK results. Quantitative evaluations indicate that, compared to the FDK reconstruction results, our reconstruction method improves contrast‐to‐noise‐ratio (CNR) by a factor of 2.56∼3.13 and our enhancement method increases CNR by 2.75∼3.33 times. The enhancement method also removes over 80% of the streak artifacts from the FDK reconstruction results. The total computation time is ∼460 sec and ∼610 sec on an NVIDIA Tesla C1060 GPU card for the reconstruction and the enhancement algorithm, respectively. Conclusions: Our new reconstruction and enhancement algorithms can effectively reduce the image artifacts in 4DCBCT. Comparing the two algorithms, the resulted image qualities are similar. The shorter computation time makes the enhancement algorithm more attractive.

UR - http://www.scopus.com/inward/record.url?scp=85024778662&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85024778662&partnerID=8YFLogxK

U2 - 10.1118/1.3613286

DO - 10.1118/1.3613286

M3 - Article

AN - SCOPUS:85024778662

VL - 38

SP - 3796

JO - Medical Physics

JF - Medical Physics

SN - 0094-2405

IS - 6

ER -