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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

A voxel-wise uncertainty-guided framework for glioma segmentation using spherical projection-based U-Net and localized refinement.

Medical physics·2026
Same author

PhysMorph: A biomechanical and image-guided deep learning framework for real-time multi-modal liver image registration.

Physics and imaging in radiation oncology·2026
Same author

An exploratory study on integrating radiomics with vision transformers for enhancing medical imaging classification accuracy.

Medical physics·2026
Same author

An Implicit Registration Framework Integrating Kolmogorov-Arnold Networks with Velocity Regularization for Image-Guided Radiation Therapy.

Bioengineering (Basel, Switzerland)·2025
Same author

Finite Element Method-Based Hybrid MRI/CBCT Generation to Improve Liver Stereotactic Body Radiation Therapy Targets Localization Accuracy.

IEEE transactions on radiation and plasma medical sciences·2025
Same author

A Radiogenomic Deep Ensemble Learning Model for Identifying Radionecrosis Following Brain Metastases (BM) Stereotactic Radiosurgery in Patients With Non-small Cell Lung Cancer BM.

Advances in radiation oncology·2025

Related Experiment Video

Updated: Sep 15, 2025

Four-Dimensional CT Analysis Using Sequential 3D-3D Registration
05:05

Four-Dimensional CT Analysis Using Sequential 3D-3D Registration

Published on: November 23, 2019

8.1K

Deep residual network-based projection interpolation and post-processing techniques for thoracic patient CBCT

Ke Lu1,2, Zeyu Zhang1,2, Lei Ren3

  • 1Medical Physics Graduate Program, Duke University, Durham, North Carolina, USA.

Medical Physics
|July 15, 2025
PubMed
Summary

Deep learning (DL) effectively interpolates cone-beam computed tomography (CBCT) projections, significantly reducing artifacts and patient dose. This novel DL workflow enhances image quality from sparsely sampled data.

Keywords:
CBCTdeep learningpost‐processingprojection interpolation

More Related Videos

Author Spotlight: Advancing 3D Modeling for Enhanced Diagnosis and Treatment of Pulmonary Nodules in Early-Stage Lung Cancer
07:53

Author Spotlight: Advancing 3D Modeling for Enhanced Diagnosis and Treatment of Pulmonary Nodules in Early-Stage Lung Cancer

Published on: October 13, 2023

1.6K
Three-Dimensional Phase Resolved Functional Lung Magnetic Resonance Imaging
10:44

Three-Dimensional Phase Resolved Functional Lung Magnetic Resonance Imaging

Published on: June 21, 2024

621

Related Experiment Videos

Last Updated: Sep 15, 2025

Four-Dimensional CT Analysis Using Sequential 3D-3D Registration
05:05

Four-Dimensional CT Analysis Using Sequential 3D-3D Registration

Published on: November 23, 2019

8.1K
Author Spotlight: Advancing 3D Modeling for Enhanced Diagnosis and Treatment of Pulmonary Nodules in Early-Stage Lung Cancer
07:53

Author Spotlight: Advancing 3D Modeling for Enhanced Diagnosis and Treatment of Pulmonary Nodules in Early-Stage Lung Cancer

Published on: October 13, 2023

1.6K
Three-Dimensional Phase Resolved Functional Lung Magnetic Resonance Imaging
10:44

Three-Dimensional Phase Resolved Functional Lung Magnetic Resonance Imaging

Published on: June 21, 2024

621

Area of Science:

  • Medical Imaging
  • Radiology
  • Computer Vision

Background:

  • Sparse sampling in cone-beam computed tomography (CBCT) reconstruction causes streak artifacts.
  • Conventional interpolation methods introduce blur and artifacts, limiting image quality.
  • Deep learning (DL) offers a potential solution to overcome these limitations in CBCT image reconstruction.

Purpose of the Study:

  • Develop a DL-based technique for interpolating sparsely sampled CBCT projections before reconstruction.
  • Implement DL for post-processing reconstructed CBCT images to enhance quality.
  • Reduce patient imaging dose through optimized DL interpolation and reconstruction.

Main Methods:

  • Linear interpolation of under-sampled projections followed by re-slicing.
  • Application of a deep residual U-Net (DRU) model to augment image quality of interpolated slices.
  • Reassembly of slices into densely-sampled projections for FDK reconstruction.
  • Secondary DRU model for post-processing the reconstructed CBCT volume.

Main Results:

  • Substantially improved peak-signal-to-noise-ratio (PSNR), structural-similarity-index-measure (SSIM), and root-mean-square-error (RMSE) compared to conventional methods.
  • DRU post-processing further enhanced image quality.
  • Achieved 86% reduction in patient imaging dose with a combined DL workflow using minimal projections (98 for half-fan, 49 for full-fan).

Conclusions:

  • The first demonstrated DL CBCT projection interpolation technique for real patient data.
  • DL interpolation and post-processing effectively reduce artifacts in CBCT images from under-sampled projections.
  • The proposed method significantly reduces patient imaging dose while maintaining image quality.