Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Regularization parameter selection for penalized-likelihood list-mode image reconstruction in PET.

Mengxi Zhang1, Jian Zhou2, Xiaofeng Niu2

  • 1Department of Biomedical Engineering, University of California, Davis, CA, United States of America.

Physics in Medicine and Biology
|April 13, 2017
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Unsupervised deep learning framework for data-driven gating in positron emission tomography.

Medical physics·2023
Same author

Deep Learning Based Joint PET Image Reconstruction and Motion Estimation.

IEEE transactions on medical imaging·2021
Same author

Artificial Intelligence in PET: An Industry Perspective.

PET clinics·2021
Same author

Anatomically aided PET image reconstruction using deep neural networks.

Medical physics·2021
Same author

Deep learning-based image quality improvement of <sup>18</sup>F-fluorodeoxyglucose positron emission tomography: a retrospective observational study.

EJNMMI physics·2021
Same author

Generative adversarial network based regularized image reconstruction for PET.

Physics in medicine and biology·2020

We developed a new cross-validation log-likelihood (CVLL) method to optimize regularization parameters for penalized likelihood (PL) reconstruction in positron emission tomography (PET) imaging. This approach enhances image quality without needing ground truth data.

Area of Science:

  • Medical Imaging
  • Nuclear Medicine
  • Image Reconstruction

Background:

  • Penalized likelihood (PL) reconstruction offers improved image quality in positron emission tomography (PET) compared to traditional ordered-subsets expectation-maximization (OSEM).
  • Selecting optimal regularization parameters for PL is difficult due to the absence of ground truth and diverse penalty functions.

Purpose of the Study:

  • To introduce a novel cross-validation log-likelihood (CVLL) function for automated regularization parameter selection in PL reconstruction.
  • To demonstrate the CVLL method's applicability to list-mode PET data without requiring true image information.

Main Methods:

  • Developed a cross-validation log-likelihood (CVLL) function for regularization parameter selection.
  • Performed statistical analysis on the mean and variance of the CVLL.

Related Experiment Videos

  • Validated the method using both simulated and real patient PET data.

    • Main Results:

    • The CVLL function provides an unbiased estimate of the log-likelihood function from noise-free data.
    • The predicted variance of CVLL aids in verifying statistical significance between parameter choices.
    • Optimally selected parameters resulted in visually superior PET image quality.

    Conclusions:

    • The proposed CVLL method effectively determines optimal regularization parameters for PL reconstruction in PET.
    • This technique enhances PET image quality and is practical for clinical application.
    • The CVLL method offers a robust solution for parameter selection challenges in PET image reconstruction.