High-speed processing of X-ray wavefront marking data with the Unified Modulated Pattern Analysis (UMPA) model

TL;DR

This paper introduces a new, high-speed implementation of the UMPA model for wavefront-marking X-ray imaging, enabling rapid processing of large datasets and extended field of view, with discussion on artifacts and new features.

Contribution

The paper presents a novel, faster UMPA implementation that significantly improves processing speed and field of view in wavefront-marking X-ray imaging.

Findings

Achieved rapid processing of large datasets

Extended the field of view in imaging

Discussed artifacts and introduced new features

Abstract

Wavefront-marking X-ray imaging techniques use e.g., sandpaper or a grating to generate intensity fluctuations, and analyze their distortion by the sample in order to retrieve attenuation, phase-contrast, and dark-field information. Phase contrast yields an improved visibility of soft-tissue specimens, while dark-field reveals small-angle scatter from sub-resolution structures. Both have found many biomedical and engineering applications. The previously developed Unified Modulated Pattern Analysis (UMPA) model extracts these modalities from wavefront-marking data. We here present a new UMPA implementation, capable of rapidly processing large datasets and featuring capabilities to greatly extend the field of view. We also discuss possible artifacts and additional new features.