(Sunday, 7/29/2018) 4:00 PM - 6:00 PM
Room: Davidson Ballroom B
Purpose: To develop a fast, flexible, and accurate wave propagation model to serve as a forward model for X-ray phase-contrast imaging (XPCI) in cone-beam geometry. This will enable computed tomography (CT) phase reconstructions using iterative techniques and performance optimization of XPCI systems through simulations.
Methods: The parallel-beam beam propagation method (BPM) is a wave propagation model often used imaging that incorporates multiple scattering. We expanded it to include the Fresnel scaling theorem, which incorporates cone-beam effects into wave propagation models. We call this new approach cone-beam-enabled beam propagation method (cbeBPM).To demonstrate the validity of cbeBPM, simulations were performed for special cases that have exact solutions. Mie theory provides an exact analytic solution for electromagnetic plane waves scattering from a single sphere or cylinder, making those excellent systems with which to compare the accuracy and computation time of different models. Furthermore, cbeBPM was used to demonstrate the importance of allowing multiple scattering events and diffraction within the object.
Results: Various systems resulted in orders-of-magnitude faster computation and <1% error compared to known solutions. For the case of 3keV photons illuminating an 8µm sphere, the accuracy of cbeBPM compared to the exact solution is 0.03%, while cbeBPM’s computation time is more than 10,000 times faster. For thick and multi-scattering samples, cbeBPM preserves information that is lost in a projection cone-beam approach. For instance, a system of two 4 μm-diameter water spheres separated axially by 1cm and illuminated with 30keV X-rays produces errors if only one scattering event is allowed.
Conclusion: Compared to other existing forward models for X-ray imaging, only cbeBPM allows multiple scattering events, thick samples, and cone-beam geometry. The efficiency of this approach lends flexibility to iterative phase-contrast CT reconstruction methods. This is crucially important for the development of next-generation XPCI-CT systems and their eventual clinical translation.
Funding Support, Disclosures, and Conflict of Interest: This work is supported by a Mayo/ASU team science award.
Keywords
Phase Contrast, Simulation, Multiple Scattering
Taxonomy
Contact Email
