Deconvolution by finite-size-source effects of x-ray phase-contrast images

Med Phys. 2011 Apr;38(4):1951-61. doi: 10.1118/1.3560889.

Abstract

Purpose: In the hard x-ray region, the cross sections for the phase shift of low-Z elements are about 1000 times larger than the absorption ones. As a consequence, phase contrast is detectable even when absorption contrast is minimal or absent. Therefore, phase-contrast imaging could become a valid alternative to absorption contrast without delivering high dose to tissue/human body parts.

Methods: To enhance the quality of phase-contrast images without increasing the dose, a possible approach could be the partial deconvolution of the finite source size effects by experimental phase-contrast images. The deconvolution procedure, the authors propose, employs the acquisition of two images on a suitable well-known test sample, one in contact and the other in phase-contrast conditions. Both acquired images are used along with a simulated phase-contrast image (obtained from the test sample in ideal conditions of pointlike source illumination) to correctly retrieve the experimental source distribution function. This information allows a generic experimental phase-contrast image, acquired in the same conditions, to be partially deconvolved by finite source size effects.

Results: The performed experimental tests indicate that deconvolved images are equivalent to those which would be obtained with a source 40% smaller than the actual size. In turn, this finding is equivalent to an increase of the "effective" lateral spatial coherence length. The corresponding quality improvement of the phase-contrast imaging is directly deducible by the presence of many Fresnel fringes, much better visible with respect to the original experimental phase-contrast images.

Conclusions: The use of a test standard sample, always possible in every experimental setup, to partially deconvolve the finite-size-source blurring effects shows that higher quality phase-contrast images could be readily available, making easier diagnoses and tissue/sample analyses. The method could give, in the future, the possibility to further lower the delivered dose to patients, organs, and tissues when compact room-sized and brilliant microfocus x-ray sources will be available for clinical applications in hospitals.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Absorption
  • Image Processing, Computer-Assisted / methods*
  • Radiography / methods*