About this role
Funding Source: Internal
Sponsor/ Supporting Company: Department funding
Research Group: CiMAT
Stipend: A 3-year studentship with a standard EPSRC stipend (2025: £20,780; 2026: £21,805), in line with current PGR Stipend & Fee Rates
Eligibility: UK Citizen, EU and International applicants are not eligible for funding.
Start Date: 02/10/2026 (3 year funding period)
Project Overview
Non-clinical X-ray Computed Tomography (XCT) has evolved into a significant "big data" challenge, with a single scanner easily generating over 10TB of data annually. The sheer volume of this structured data creates substantial hurdles for storage, transmissibility, and long-term curation. This PhD project aims to address these challenges by researching and developing specialized lossless and lossy compression methods designed specifically for the spatial and sequential structures inherent in XCT projections and reconstructed volumes. The goal is to achieve a 60-80% reduction in data size without compromising the integrity of scientific information extracted from the scans.
The successful candidate will join the AC/DC research team to develop an open-source compression solution tailored for the XCT community. While generic compression tools exist, they often fail to fully exploit the specific redundancies found in 3D tomographic data. You will exploit your signal processing knowledge with statistical mathematical frameworks to maximise compression ratios. We will use predictor models which estimate projections or slices, storing only differences between the prediction and original data. Because errors are small and repetitive they can be efficiently compressed with classical predictors, and improved with data-driven models of XCT intensity statistics.
The project provides a unique opportunity to work at the intersection of mathematics, algorithms and high-end imaging, contributing to a solution that reduces the environmental and financial burden of large scale scientific data.
For informal enquiries, contact Dr Jay Warnett [email protected]
Essential and Desirable Criteria
Essential:
First degree (Batchelors and/or Masters) in Mathematics, Computer Science, or field with strong mathematical focus. An interest in signal processing and image representation Demonstratable experience in python for data handling and algorithm development
Desirable:
Knowledge and experience with imaging systems (X-ray CT, MRI or similar) Familiarity with image transforms, entropy coding or noise characterisation
Funding and Eligibility: Funded PhD. UK Student
Supervisors
Dr Jay Warnett (WMG)
Dr Randa Herzallah (Mathematics)