Hybrid Pre-log and Post-log Image Reconstruction for Computed Tomography

Guobao Wang, Jian Zhou, Zhou Yu, Wenli Wang, Jinyi Qi

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Tomographic image reconstruction for low-dose CT is increasingly challenging as dose continues to reduce in clinical applications. Pre-log domain methods and post-log domain methods have been proposed individually and each method has its own disadvantage. While having the potential to improve image quality for low-dose data by using an accurate imaging model, pre-log domain methods suffer slow convergence in practice due to the nonlinear transformation from the image to measurements. In contrast, post-log domain methods have fast convergence speed but the resulting image quality is suboptimal for low-count CT data because the log transformation is extremely unreliable for low-count measurements and undefined for negative values. This paper proposes a hybrid method that integrates the pre-log model and post-log model together to overcome the disadvantages of individual pre-log and post-log methods. We divide a set of CT data into high-count and low-count regions. The post-log weighted least squares model is used for measurements in the high-count region and the pre-log shifted Poisson model for measurements in the low-count region. The hybrid likelihood function can be optimized using an existing iterative algorithm. Computer simulations and phantom experiments show that the proposed hybrid method can achieve faster early convergence than the pre-log shifted Poisson likelihood method and better signal-to-noise performance than the post-log weighted least squares method.

Original languageEnglish (US)
JournalIEEE Transactions on Medical Imaging
DOIs
StateAccepted/In press - Sep 12 2017

Fingerprint

Computer-Assisted Image Processing
Image reconstruction
Tomography
Image quality
Least-Squares Analysis
Likelihood Functions
Imaging techniques
Computer simulation
Computer Simulation
Noise
Experiments

Keywords

  • Computational modeling
  • Computed tomography
  • Convergence
  • Data models
  • Image reconstruction
  • image reconstruction
  • iterative algorithm
  • low-dose CT
  • Noise measurement
  • noise model
  • shifted Poisson
  • weighted least squares

ASJC Scopus subject areas

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

Cite this

Hybrid Pre-log and Post-log Image Reconstruction for Computed Tomography. / Wang, Guobao; Zhou, Jian; Yu, Zhou; Wang, Wenli; Qi, Jinyi.

In: IEEE Transactions on Medical Imaging, 12.09.2017.

Research output: Contribution to journalArticle

@article{c129ad959de347b383939daca87109dc,
title = "Hybrid Pre-log and Post-log Image Reconstruction for Computed Tomography",
abstract = "Tomographic image reconstruction for low-dose CT is increasingly challenging as dose continues to reduce in clinical applications. Pre-log domain methods and post-log domain methods have been proposed individually and each method has its own disadvantage. While having the potential to improve image quality for low-dose data by using an accurate imaging model, pre-log domain methods suffer slow convergence in practice due to the nonlinear transformation from the image to measurements. In contrast, post-log domain methods have fast convergence speed but the resulting image quality is suboptimal for low-count CT data because the log transformation is extremely unreliable for low-count measurements and undefined for negative values. This paper proposes a hybrid method that integrates the pre-log model and post-log model together to overcome the disadvantages of individual pre-log and post-log methods. We divide a set of CT data into high-count and low-count regions. The post-log weighted least squares model is used for measurements in the high-count region and the pre-log shifted Poisson model for measurements in the low-count region. The hybrid likelihood function can be optimized using an existing iterative algorithm. Computer simulations and phantom experiments show that the proposed hybrid method can achieve faster early convergence than the pre-log shifted Poisson likelihood method and better signal-to-noise performance than the post-log weighted least squares method.",
keywords = "Computational modeling, Computed tomography, Convergence, Data models, Image reconstruction, image reconstruction, iterative algorithm, low-dose CT, Noise measurement, noise model, shifted Poisson, weighted least squares",
author = "Guobao Wang and Jian Zhou and Zhou Yu and Wenli Wang and Jinyi Qi",
year = "2017",
month = "9",
day = "12",
doi = "10.1109/TMI.2017.2751679",
language = "English (US)",
journal = "IEEE Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Hybrid Pre-log and Post-log Image Reconstruction for Computed Tomography

AU - Wang, Guobao

AU - Zhou, Jian

AU - Yu, Zhou

AU - Wang, Wenli

AU - Qi, Jinyi

PY - 2017/9/12

Y1 - 2017/9/12

N2 - Tomographic image reconstruction for low-dose CT is increasingly challenging as dose continues to reduce in clinical applications. Pre-log domain methods and post-log domain methods have been proposed individually and each method has its own disadvantage. While having the potential to improve image quality for low-dose data by using an accurate imaging model, pre-log domain methods suffer slow convergence in practice due to the nonlinear transformation from the image to measurements. In contrast, post-log domain methods have fast convergence speed but the resulting image quality is suboptimal for low-count CT data because the log transformation is extremely unreliable for low-count measurements and undefined for negative values. This paper proposes a hybrid method that integrates the pre-log model and post-log model together to overcome the disadvantages of individual pre-log and post-log methods. We divide a set of CT data into high-count and low-count regions. The post-log weighted least squares model is used for measurements in the high-count region and the pre-log shifted Poisson model for measurements in the low-count region. The hybrid likelihood function can be optimized using an existing iterative algorithm. Computer simulations and phantom experiments show that the proposed hybrid method can achieve faster early convergence than the pre-log shifted Poisson likelihood method and better signal-to-noise performance than the post-log weighted least squares method.

AB - Tomographic image reconstruction for low-dose CT is increasingly challenging as dose continues to reduce in clinical applications. Pre-log domain methods and post-log domain methods have been proposed individually and each method has its own disadvantage. While having the potential to improve image quality for low-dose data by using an accurate imaging model, pre-log domain methods suffer slow convergence in practice due to the nonlinear transformation from the image to measurements. In contrast, post-log domain methods have fast convergence speed but the resulting image quality is suboptimal for low-count CT data because the log transformation is extremely unreliable for low-count measurements and undefined for negative values. This paper proposes a hybrid method that integrates the pre-log model and post-log model together to overcome the disadvantages of individual pre-log and post-log methods. We divide a set of CT data into high-count and low-count regions. The post-log weighted least squares model is used for measurements in the high-count region and the pre-log shifted Poisson model for measurements in the low-count region. The hybrid likelihood function can be optimized using an existing iterative algorithm. Computer simulations and phantom experiments show that the proposed hybrid method can achieve faster early convergence than the pre-log shifted Poisson likelihood method and better signal-to-noise performance than the post-log weighted least squares method.

KW - Computational modeling

KW - Computed tomography

KW - Convergence

KW - Data models

KW - Image reconstruction

KW - image reconstruction

KW - iterative algorithm

KW - low-dose CT

KW - Noise measurement

KW - noise model

KW - shifted Poisson

KW - weighted least squares

UR - http://www.scopus.com/inward/record.url?scp=85030228993&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85030228993&partnerID=8YFLogxK

U2 - 10.1109/TMI.2017.2751679

DO - 10.1109/TMI.2017.2751679

M3 - Article

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

SN - 0278-0062

ER -