Demonstration of a family of X-ray dark-field retrieval approaches on a common set of samples.

There are various imaging setups capable of capturing dark-field images, each with its own capabilities and limitations. Across all these setups, the underlying physical mechanism responsible for dark-field contrast is generally defined similarly: small-angle X-ray scattering from sub-resolution structures, multiple refraction from large resolvable features and/or edge effects. Despite this conceptual consistency, there remains a lack of studies that directly compare different dark-field imaging setups using identical samples. We have recently developed a `family' of dark-field imaging techniques that, although they have different experimental setups, each extract dark-field contrast by directly resolving dark-field-associated local image blurring with a high-resolution camera. In this paper, we provide a qualitative comparison of the capabilities of this family of methods, covering both different experimental acquisition methods and image-retrieval algorithms for this subset of dark-field imaging techniques. We acquired X-ray imaging data from two test samples using synchrotron propagation-based, single-grid and speckle-based setups, and retrieved dark-field images from each dataset using an appropriate algorithm. The dark-field retrieval methods differed in their computational framework: single-shot algorithms were applied for the single-grid and speckle-based approaches, an additional multi-exposure method was used for speckle data, and a dual-energy method was applied to the propagation-based data. We find that all approaches successfully retrieve dark-field contrast in microstructure-dense regions, as expected. Differences arise at high-spatial-frequency sample features, such as edges, and some approaches exhibit artefacts. We attribute these differences to variations in experimental parameters (e.g. pixel size, sample-to-detector distance and X-ray energy) and to the algorithmic assumptions underlying each technique (e.g. the single-material assumption and treatment of phase effects). This study aims to guide users of dark-field imaging in selecting the most suitable technique for their imaging goals. To this end, we provide a summary table and highlight opportunities for future research into the sources of dark-field contrast across emerging methods.