Fast one-dimensional wave-front propagation for x-ray differential phase-contrast imaging

Abstract

Abstract:Numerical wave-optical simulations of X-ray differential phase-contrast imaging using grating interferometry require the oversampling of gratings and object structures in the range of few micrometers. Consequently, fields of view of few millimeters already use large amounts of a computer's main memory to store the propagating wave front, limiting the scope of the investigations to only small-scale problems. In this study, we apply an approximation to the Fresnel-Kirchhoff diffraction theory to overcome these restrictions by dividing the two-dimensional wave front up into 1D lines, which are processed separately. The approach enables simulations with samples of clinically relevant dimensions by significantly reducing the memory footprint and the execution time and, thus, allows the qualitative comparison of different setup configurations. We analyze advantages as well as limitations and present the simulation of a virtual mammography phantom of several centimeters of size. Reiser, and F. Bamberg, "Assessment of grating-based X-ray phase-contrast CT for differentiation of invasive ductal carcinoma and ductal carcinoma in situ in an experimental ex vivo set-up." Eur. Radiol. 2, 381-387 (2012). 9. F. Pfeiffer, M. Bech, O. Bunk, P. Kraft, E. F. Eikenberry, C. Bronnimann, C. Grünzweig, and C. David, "Hard-xray dark-field imaging using a grating interferometer," Nat. Mater. 7, 134-137 (2008

In Silico Phase-Contrast X-Ray Imaging of Anthropomorphic Voxel-Based Phantoms

“…Here, accurate WP simulations require micrometer model sampling, in direct contradiction with the aforementioned scaling of voxelized models. Previous work on WP simulations has been limited by this sampling requirement [25], [26].…”

Simulations of multi-contrast x-ray imaging using near-field speckles

“…The numerical simulations of x-ray speckle tracking were carried out using an object-oriented framework written in Python with extensions in C++ [3,13]. The framework has been successfully employed for quantitative grating-based phase-contrast and dark-field imaging [3,14], and more recently for studies on the effects of the polychromaticity of the x-ray beam on the images retrieved using speckle tracking [12].…”

Big-data x-ray phase contrast imaging simulation challenges

“…For this work, it was decided that an incremental approach to building a complex system was most reasonable to help identify challenges early-on as there exists extensive studies on small-scale simulations. [1][2][3][4] One of the appealing features of XPCI is its greater sensitivity to density and material variations; however, greater sensitivity in a system implies that a greater level of caution is needed when designing algorithms for a BDXPCI simulation tool. Particular caution must be exercised when attempting to design algorithms that will operate on big data, as a small error can propagate through the algorithm-increasing as more information is processed by the method.…”

“…Unfortunately, the survey could not find any efforts in large-scale simulation as most efforts either restricted the simulation to very small fields-of-view, compressed information in object-space, short propagation lengths, or any combination thereof. [1][2][3][4] Work by Peterzol et. al.…”

“…al. 4 describes a fast XPCI simulation that greatly reduces computational resource requirements by decomposing a 2D plane wave into a set of 1D lines that can be efficiently processed in parallel. Wolf et.…”