PhD project

Deep Learning for Computational Imaging in Spectral Computed Tomography

On this page
Application deadline

31 January 2023

Image reconstruction is an inverse problem that deals with the estimation of an image from measurement data obtained from physics-based acquisition processes. As instance, modern techniques used in oncology such as X-ray Computed tomography (CT) require fast and accurate image reconstruction algorithms to deal with incomplete data to speed-up the acquisition, noise and artefacts.

Spectral CT is a particular data acquisition modality able to provide different spectral channels information used to separate individual materials and to provide concentration maps of contrast agents/basis materials. Exploiting X-ray measurements acquired at multiple energies, spectral CT has the ability to recover the concentration maps of the constituents of the tissues in a quantitative manner. Recent developments in energy selective photon counting detectors have boosted the research in this area by utilizing photon-counting detectors which count the photon received at different energies intervals or by using different multiple X-ray sources at different energies.

This project exploits a new approach based on robust deep learning for material decomposition and model-based large-scale optimization that allows for direct estimation of material concentration from the measurements enabling quantitative estimation and reducing the X-ray dose. Deep learning methods, that use for training past datasets of successfully reconstructed images together with the measurement that produced them, have been shown to produce some impressive results in image reconstruction. However, the problem with such data-driven imaging is currently the reliability of the results and robustness to model acquisition mismatch, i.e. different scanners used in training and testing which hinders the use for clinical assessment. Indeed, even small deviations in the data can result in large differences in the outcome which has devastating implications for many applications. Furthermore, the collection of training ground truth clinical data is sometimes impossible to obtain under indirect acquisition of the specimen through medical scanners.

In this project you will answer these questions by developing new algorithms, based on deep learning and optimization, and analysis techniques towards robust computational image reconstruction for spectral CT. The student is expected to further advance current developments of supervised (or self-supervised) model-based deep learning with convolutional network and deep generative prior. The expected project’s outcomes will involve writing research papers with the developed methods together with the creation of the software to reproduce the results and clinical validation.

The methods which will be developed will be:

  1. Develop diffusion models and generative adversarial networks for self-supervised deep learning for spectral CT reconstruction.
  2. Design image reconstruction for X-ray CT and spectral photon counting detectors
  3. Critically analyse the results of the developed methods on simulated and real spectral CT data.

How to apply

  1. Email Dr Alessandro Perelli to:
    1. Send a copy of your CV
    2. Discuss your potential application and any practicalities (e.g. suitable start date).
  2. After discussion with Dr Alessandro Perelli, formal applications can be made via our direct application system.
Apply for the Doctor of Philosophy (PhD) degree in Biomedical Engineering

Supervisors

Principal supervisor

Funding

PhD funding

The Chinese Scholarship Council provides opportunities for Chinese Students to undertake a PhD programme in any research field at the School of Life Sciences and the School of Science and Engineering. Successful applicants will receive support to enter the China Scholarship Council (CSC) competition scheme.

Funding eligibility: China