Fourier transform based iterative method for x-ray differential phase-contrast computed tomography

TL;DR

This paper introduces a Fourier transform based iterative method for x-ray phase-contrast tomography that improves image quality with incomplete or noisy data, especially within regions of interest, by directly reconstructing the refractive index from differential phase shift data.

Contribution

It presents a novel iterative reconstruction technique leveraging Fourier transforms for phase-contrast tomography, enhancing robustness to data incompleteness and noise, and enabling interior imaging with prior knowledge.

Findings

Effective in reconstructing images from incomplete data

Superior noise robustness compared to analytic methods

Successful biological tissue imaging demonstrations

Abstract

Biological soft tissues encountered in clinical and pre-clinical imaging mainly consist of light element atoms, and their composition is nearly uniform with little density variation. Thus, x-ray attenuation imaging suffers from low image contrast resolution. By contrast, x-ray phase shift of soft tissues is about a thousand times greater than x-ray absorption over the diagnostic energy range, thereby a significantly higher sensitivity can be achieved in terms of phase shift. In this paper, we propose a novel Fourier transform based iterative method to perform x-ray tomographic imaging of the refractive index directly from differential phase shift data. This approach offers distinct advantages in cases of incomplete and noisy data than analytic reconstruction, and especially suitable for phase-contrast interior tomography by incorporating prior knowledge in a region of interest (ROI). Biological experiments demonstrate the merits of the proposed approach.