Comparison of cone beam artifacts reduction: Two pass algorithm vs TV-based CS algorithm

Shinkook Choi, Jongduk Baek

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)


In a cone beam computed tomography (CBCT), the severity of the cone beam artifacts is increased as the cone angle increases. To reduce the cone beam artifacts, several modified FDK algorithms and compressed sensing based iterative algorithms have been proposed. In this paper, we used two pass algorithm and Gradient-Projection-Barzilai-Borwein (GPBB) algorithm to reduce the cone beam artifacts, and compared their performance using structural similarity (SSIM) index. In two pass algorithm, it is assumed that the cone beam artifacts are mainly caused by extreme-density(ED) objects, and therefore the algorithm reproduces the cone beam artifacts(i.e., error image) produced by ED objects, and then subtract it from the original image. GPBB algorithm is a compressed sensing based iterative algorithm which minimizes an energy function for calculating the gradient projection with the step size determined by the Barzilai- Borwein formulation, therefore it can estimate missing data caused by the cone beam artifacts. To evaluate the performance of two algorithms, we used testing objects consisting of 7 ellipsoids separated along the z direction and cone beam artifacts were generated using 30 degree cone angle. Even though the FDK algorithm produced severe cone beam artifacts with a large cone angle, two pass algorithm reduced the cone beam artifacts with small residual errors caused by inaccuracy of ED objects. In contrast, GPBB algorithm completely removed the cone beam artifacts and restored the original shape of the objects.

Original languageEnglish
Title of host publicationMedical Imaging 2015
Subtitle of host publicationPhysics of Medical Imaging
EditorsChristoph Hoeschen, Despina Kontos, Christoph Hoeschen
ISBN (Electronic)9781628415025
Publication statusPublished - 2015
EventMedical Imaging 2015: Physics of Medical Imaging - Orlando, United States
Duration: 2015 Feb 222015 Feb 25

Publication series

NameProgress in Biomedical Optics and Imaging - Proceedings of SPIE
ISSN (Print)1605-7422


OtherMedical Imaging 2015: Physics of Medical Imaging
Country/TerritoryUnited States

Bibliographical note

Publisher Copyright:
© 2015 SPIE.

All Science Journal Classification (ASJC) codes

  • Electronic, Optical and Magnetic Materials
  • Biomaterials
  • Atomic and Molecular Physics, and Optics
  • Radiology Nuclear Medicine and imaging


Dive into the research topics of 'Comparison of cone beam artifacts reduction: Two pass algorithm vs TV-based CS algorithm'. Together they form a unique fingerprint.

Cite this