Interdisciplinary Institute in Artificial Intelligence

DESCRIPTION

X-ray phase contrast imaging (PCI) is the most promising advancement in X-ray imaging techniques since their invention by Röntgen, enabling an increase in sensitivity of more than three orders of magnitude and, consequently, the potential for significant reductions in imaging dose as well as greatly improved contrast in soft tissue without the use of contrast agents. To exploit the phase information, a computational step is required to reconstruct the phase—and possibly the attenuation—a process known as phase retrieval. Phase retrieval is a nonlinear ill-posed inverse problem. The development of deep learning methods in recent years has led to many advances in image and signal processing.

We have recently developed new nonlinear methods for phase retrieval based on deep learning and transfer learning (training on simulated data). These include end-to-end methods, methods based on Generative Adversarial Networks (GANs), and algorithm unrolling methods, in which parts of an iterative algorithm are replaced with neural networks and the iterations are learned, allowing the physics model of image formation to be taken into account while learning a regularization. These types of methods can reduce computation time by several orders of magnitude, while improving image quality and the robustness of reconstructions. These algorithms have yielded exceptional initial results when applied to test objects.

Based on these results in propagation-based PCI (PBI), the aim of this project is to extend the developed algorithms to tomographic and 4D imaging in PBI, as well as to related phase-contrast techniques. In this project, we will focus on high-frequency imaging (3D+t) using storage rings and X-ray Free Electron Lasers (XFELs), Modulation-Based Imaging (MoBI)—including dark-field imaging (DFI) (3D+scattering)—and phase retrieval using Spectral Photon Counting Imaging (SPCI) and Computed Tomography (SPCCT) (3D+spectrum). The developed algorithms will be made available through the PyPhase package. The developed algorithms will be applied to imaging of animal models of osteoarthritis and osteoarthrosis. Faster reconstructions, higher image quality, and more robust algorithms will improve diagnosis and analysis in both clinical and preclinical settings. This work requires strong expertise in inverse problems for the development of the algorithms and their components, as well as expertise in their applications to bring them into routine use. To meet the challenges posed by the research project, we have assembled an international team of imaging physicists, inverse problems specialists, algorithm specialists, and biologists.

ACTIVITIES

One Ph.D. student has been hired to work on the topic "Image reconstruction in modulation-based and spectral X-ray phase-contrast imaging using deep learning."

One postdoctoral fellow will be hired to work on the topic "3D and 4D phase retrieval in propagation-based X-ray phase-contrast imaging using deep learning."

CHAIR EVENTS

Langer M, Mom K, Sixou B, "Deep Learning for Phase Retrieval from Fresnel Diffraction Patterns," 11th Applied Inverse Problems Conference, Göttingen, Germany, invited talk, 2023.