Scatter kernel estimation with an edge-spread function method for cone-beam computed tomography imaging

Heng Li, Radhe Mohan, X. Ronald Zhu

Research output: Contribution to journalArticle


The clinical applications of kilovoltage x-ray cone-beam computed tomography (CBCT) have been compromised by the limited quality of CBCT images, which typically is due to a substantial scatter component in the projection data. In this paper, we describe an experimental method of deriving the scatter kernel of a CBCT imaging system. The estimated scatter kernel can be used to remove the scatter component from the CBCT projection images, thus improving the quality of the reconstructed image. The scattered radiation was approximated as depth-dependent, pencil-beam kernels, which were derived using an edge-spread function (ESF) method. The ESF geometry was achieved with a half-beam block created by a 3 mm thick lead sheet placed on a stack of slab solid-water phantoms. Measurements for ten water-equivalent thicknesses (WET) ranging from 0 cm to 41 cm were taken with (half-blocked) and without (unblocked) the lead sheet, and corresponding pencil-beam scatter kernels or point-spread functions (PSFs) were then derived without assuming any empirical trial function. The derived scatter kernels were verified with phantom studies. Scatter correction was then incorporated into the reconstruction process to improve image quality. For a 32 cm diameter cylinder phantom, the flatness of the reconstructed image was improved from 22% to 5%. When the method was applied to CBCT images for patients undergoing image-guided therapy of the pelvis and lung, the variation in selected regions of interest (ROIs) was reduced from >300 HU to <100 HU. We conclude that the scatter reduction technique utilizing the scatter kernel effectively suppresses the artifact caused by scatter in CBCT.

Original languageEnglish (US)
Pages (from-to)6729-6748
Number of pages20
JournalPhysics in medicine and biology
Issue number23
StatePublished - Dec 7 2008
Externally publishedYes

ASJC Scopus subject areas

  • Radiological and Ultrasound Technology
  • Radiology Nuclear Medicine and imaging

Fingerprint Dive into the research topics of 'Scatter kernel estimation with an edge-spread function method for cone-beam computed tomography imaging'. Together they form a unique fingerprint.

  • Cite this