Anastasia graduated in 2024 from King’s College London with a BSc in Mathematics and then completed her Master of Philosophy in Data Intensive Science at the University of Cambridge. Her interest lies in deep learning, inverse problems, and image analysis.Outside of maths, she enjoys reading, hiking, shooting, and stargazing.

Project title:
AI for Parameter-Free Image Reconstruction

Supervisor(s):
Matthias Ehrhardt, Margaret Duff

Project description:
Neutron tomography is a non-destructive imaging technique that uses neutron beams to reveal the internal structure of materials and objects. However, access to neutron facilities is limited and costly, while measurements can be affected by low neutron counts, noise and experimental instabilities. These challenges can make it difficult to reconstruct clear and reliable images from the data collected. 

My research aims to make this reconstruction process more reliable and easier to use by automatically selecting some key settings, especially one known as the regularisation strength. This setting controls the balance between accurately representing the measured data and reducing noise while preserving important features in the image. Developing principled methods to select it automatically could improve image quality and consistency, reduce reliance on manual trial and error, and help researchers make better use of scarce neutron beam time.