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

LEARN: Learned Experts' Assessment-Based Reconstruction Network for Sparse-Data CT.

Hu Chen, Yi Zhang, Yunjin Chen

    IEEE Transactions on Medical Imaging
    |June 6, 2018
    PubMed
    Summary

    This study introduces a deep learning network (LEARN) for sparse-data computed tomography (CT) reconstruction. LEARN effectively learns regularization parameters, significantly improving image quality and speed compared to traditional methods.

    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

    Immunosuppressive therapies adversely affect blood biochemical parameters in patients with inflammatory bowel disease: a meta-analysis.

    The Journal of international medical research·2019
    Same author

    Facile synthesis of Ag-CuO/SBA-15 for aerobic epoxidation of olefins with high activity.

    Nanotechnology·2019
    Same author

    Reinforcement of Polylactic Acid for Fused Deposition Modeling Process with Nano Particles Treated Bamboo Powder.

    Polymers·2019
    Same author

    Comparison of deep learning and human observer performance for detection and characterization of simulated lesions.

    Journal of medical imaging (Bellingham, Wash.)·2019
    Same author

    Bending Flexibility of Moso Bamboo (<i>Phyllostachys Edulis</i>) with Functionally Graded Structure.

    Materials (Basel, Switzerland)·2019
    Same author

    CT Super-Resolution GAN Constrained by the Identical, Residual, and Cycle Learning Ensemble (GAN-CIRCLE).

    IEEE transactions on medical imaging·2019

    Area of Science:

    • Medical Imaging
    • Computational Imaging
    • Machine Learning

    Background:

    • Compressive sensing (CS) is vital for tomographic reconstruction with limited data, crucial for applications like few-view computed tomography (CT).
    • Adaptive parameter selection for regularization in iterative CS reconstruction remains a significant challenge.
    • Current methods often struggle with artifact reduction and feature preservation in sparse-data CT.

    Purpose of the Study:

    • To develop a data-driven deep learning network for adaptive regularization in sparse-data CT reconstruction.
    • To address the open problem of adaptive parameter selection in CS-based iterative reconstruction.
    • To demonstrate the superior performance of the proposed network in terms of image quality and computational efficiency.

    Main Methods:

    Related Experiment Videos

    • Unfolding a state-of-the-art iterative reconstruction scheme into a deep neural network architecture (LEARN).
    • Utilizing a data-driven approach for training the network, enabling it to learn regularization terms and parameters.
    • Employing the Mayo Clinic low-dose challenge dataset for experimental validation and comparison.

    Main Results:

    • The proposed LEARN network achieved superior performance in artifact reduction and feature preservation compared to existing state-of-the-art methods.
    • LEARN demonstrated significantly improved computational speed, reducing complexity by orders of magnitude.
    • Experimental results validated the network's effectiveness on a challenging low-dose CT dataset.

    Conclusions:

    • The learned experts' assessment-based reconstruction network (LEARN) effectively addresses adaptive regularization in sparse-data CT.
    • LEARN leverages application-specific knowledge learned from data for more favorable image recovery.
    • The proposed deep learning approach offers a promising, computationally efficient solution for sparse-data tomographic reconstruction.