Separating edges from microstructure in X-ray dark-field imaging: evolving and devolving perspectives via the X-ray...

Area of Science:

• Physics
• Materials Science
• Medical Imaging

Background:

• X-ray dark-field (XDF) contrast arises from X-ray diffusion by sub-resolution structures (small-angle X-ray scattering) and resolvable sample edges.
• Existing experimental XDF techniques struggle to differentiate these two distinct contrast generation mechanisms.
• Speckle-based X-ray imaging (SBXI) analyzes changes in a reference speckle pattern's visibility to extract XDF.

Purpose of the Study:

• To present a novel algorithm for SBXI capable of separating microstructure-induced and edge-induced XDF contrast.
• To introduce a 'devolving' Fokker-Planck equation perspective for modeling XDF generation in paraxial X-ray imaging.
• To enable multimodal retrieval of attenuation, phase, and distinct XDF properties from experimental data.

Main Methods:

• Developed a variant of the multimodal intrinsic speckle-tracking (MIST) algorithm utilizing a 'devolving' Fokker-Planck equation.
• Compared single- and multiple-exposure multimodal retrieval algorithms based on both 'evolving' and 'devolving' Fokker-Planck perspectives.
• Applied the developed algorithms to experimental data from a phantom and an organic sample.

Main Results:

• The 'devolving' Fokker-Planck perspective successfully distinguished between XDF generated by unresolved microstructure and sharp edges.
• Demonstrated the capability to separate these two physically different XDF contrast mechanisms.
• Verified the algorithm's performance on both phantom and organic sample datasets.

Conclusions:

• The new algorithm provides complementary XDF images, separating sharp-edge scatter from microstructure-induced scatter.
• This separation is crucial for improving XDF computed tomography by reducing streaking artifacts.
• The method facilitates sample preparation by eliminating the need for embedding, as strong edge signals can be isolated.