A hybrid simulation framework for computer simulation and modelling studies of grating-based x-ray phase-contrast images

摘要

Clinical studies performed using computer simulation are inexpensive, flexible methods that can be used to study aspects of a proposed imaging technique prior to a full clinical study. Typically, lesions are simulated into (experimental) data to assess the clinical potential of new methods or algorithms. In grating-based phase-contrast imaging (GB-PCI), full wave simulations are, however, computationally expensive due to the high periodicity of the gratings and therefore not practically applicable when large data sets are required. This work describes the development of a hybrid modelling platform that combines analytical and empirical input data for a more rapid simulation of GB-PCI images with little loss of accuracy. Instead of an explicit implementation of grating details, measured summary metrics (i.e. visibility, flux, noise power spectra, presampling modulation transfer function) are applied in order to generate transmission and differential phase images with large fields of view. Realistic transmission and differential phase images were obtained with good quantitative accuracy The different steps of the simulation framework, as well as the methods to measure the summary metrics, are discussed in detail such that the technique can be easily customized for a given system. The platform offers a fast, accurate alternative to full wave simulations when the focus switches from grating/system design and set up to the generation of GB-PCI images for an established system.