Skip to main content
Seminar | Mathematics and Computer Science Division

Advanced Algorithms of X-ray Wavefront Sensing and Multi-contrast Imaging Methods for Synchrotron Facilities

LANS Seminar

Abstract: X-ray beam diagnostics, wavefront sensing, and at-wavelength metrology have become crucial and essential at synchrotron radiation and free-electron laser facilities worldwide. Specifically, the availability of advanced at-wavelength characterization tools and methods is critical to developing and fabricating the high-quality focusing optics required to achieve diffraction-limited focusing in many instruments deployed at modern X-ray light sources. The multi-contrast imaging methods are also actively investigated for both wavefront sensing and X-ray imaging of bio-samples, including absorption, quantitative phase, and dark-field information.

This presentation reports on the latest developments of the advanced algorithms of wavefront sensing and multi-contrast imaging methods at the Advanced Photon Source, including wavelet-transform-based speckle tracking (WXST), optimization-based coded-mask-based multi-contrast imaging (CMMI) method, and AI-accelerated real-time X-ray phase-contrast imaging method (Speckle-based Phase-contrast Imaging Neural Network, SPINNet). As a result of enhancements to both the hardware and software elements, these methods exhibit extraordinary performance in terms of computational efficiency, sensitivity, and resolution. In addition, the wavefront sensing and multi-contrast imaging method based on deep learning, SPINNet, has shown real-time measurement ability with superior quality compared with conventional methods, which promises applications in real-time beamline diagnostics and in-situ imaging of biomedical samples and battery materials both in 2D and 3D.