Clara graduated from Durham university in 2021 with a BSc in Mathematics and Philosophy. She then completed a master’s in mathematical modelling at Sorbonne university in 2023. Within the masters, Clara specialised in optimisation and worked with EDF for five months on the shape optimisation of overflow spillways on water dams. During her studies, Clara has covered a lot of pure mathematics and is now very interested in the implementation of this theory in real-world contexts.

In her free time, Clara is a keen musician – playing the piano and violin, and enjoys dance and spending time outdoors.

Project title:
Multi-modality image reconstruction for Ptychography and XRF imaging

Supervisor(s):
Matthias Ehrhardt, Paul Quinn (Ada Lovelace Centre)

Project description:
Multi-modality imaging is the practice of creating multiple images of a single sample while using different techniques. Common applications can be found in medical imaging with PET-CT or PET-MRI scans, or in collection of satellite data. For this project, we study multi-modality imaging done at the UK’s national synchrotron facility Diamond Light source. The imaging techniques we have chosen to focus on are Ptychography and X-ray Fluorescence (XRF) which, when combined give complementary information on the sample’s elemental composition.

While using different modalities allows us to gather more information on the sample, constraints in the acquisition mean that we often get a high resolution Ptychography image and a low resolution XRF. This project aims to explore how the high resolution Ptychography can be used to increase the resolution of the XRF. Two main approaches are considered here, based on the following observation. Ptychography and XRF measure physical quantities that are closely related, and both based on the sample’s elemental composition. Therefore, the images should show the same structure and present edges in the same places. Thus, a first approach for the super-resolution of the XRF image is to match its edges to the Ptychography image during reconstruction. A different and more Physics-based approach is to consider the Physical process captured in both images and formulate the problem as an inverse problem, where the unknown is the sample’s elemental concentration maps and both images are used as data in the reconstruction.

This project is in collaboration with Diamond Light source, the Ada Lovelace Center and The Science and Technology Facilities Council (STFC).