All of the below projects are eligible for the Vice Chancellors Scholarship. For more information about this award, and other Postgraduate funding, please see the University Scholarship finder.
Please note that this projects is in competition with a number of other projects across the School of Physical Sciences for three funded positions. There will be an internal competition and the quality of the applicants will play a role in the final decision of the School.
PhD in Few Shot Learning for the Identification of Rare Pathologies in Medical Images
Supervisor: Dr Stuart Gibson and Prof Adrian Podoleanu
A PhD position is available in the field of Applied Artificial Intelligence and Medical Imaging. This project is in competition with other projects offered by the School of Physical Sciences for one of a number of Vice Chancellor’s PhD Studentships.
Artificial intelligence (AI) has become prevalent in many areas including autonomous (self-driving) vehicles, recommender systems and healthcare. In particular deep neural networks have revolutionised the field of computer vision, allowing accurate and fast identification of objects within images for problems comprising thousands of object classes. An estimated 90 percent of all healthcare data originates from medical images. Access to such large datasets has allowed, for example, DeepMind to train an AI based image interpretation system that is capable of automatically detecting a variety of eye diseases. The work employed optical coherence tomography (OCT), a technology developed intensively in the School of Physical Sciences.
Conversely, this project will investigate few-shot learning methods for medical image interpretation in cases were large datasets are not readily available. We anticipate that this work will be particularly useful for the identification of rare diseases. The project will draw upon the growing expertise in applied AI within the School of Physical Sciences, developed through previous and current projects (e.g. detection of diabetic macular edema in retinal scans, classification and quantification of chemical species and facial identification using brainwaves). Where possible the project will make use of existing links with our local NHS Trust and the new Kent and Medway Medical School.