PhD Studentship: Development of medical diagnostics based on Terahertz Imaging with Artificial Intelligence
University: University of Warwick
Department: Physics (host dept) and Computer Science
Location: Coventry, United Kingdom
We are seeking a PhD student to support the development of an effective medical diagnostic system based on terahertz (THz) imaging through the use of compressed sensing and machine learning techniques.
This is an exciting opportunity for a wide range of prospective PhD students interested in a multi-disciplinary training with leading terahertz imaging and machine learning/AI experts (see "Further details" at the end).
Project Goals:
As a PhD candidate, the student will be expected to be involved in some or all of the following components:
- Understanding the design and working of the THz imaging system
- Acquisition of patient data
- Analysis of patient data
- Development of a diagnostic system for various skin (or other medical) conditions through machine learning
- Analysis of patient specific parameters in data obtained from THz Imaging System
- Improvement in the design and data acquisition speed of THz system
Contact: If you are interested in this position, please contact Prof MacPherson directly on [Email Address Removed].
Funding: This is a fully funded-PhD studentship at standard UK Research Council rates, available with a flexible start date from October 2023 onwards for a 3.5 year period. The studentship covers university fees and a living stipend, and is available to UK and potentially international students. For more information please see http://go.warwick.ac.uk/PhysicsPG.
Eligibility
Bachelors or Masters in Physics (or related area) (equivalent to a UK 2.1 or better)
OR
Bachelors or Masters in Computer Science (or related area) (equivalent to a UK 2.1 or better)
Desired criterion:
- No medical background is needed
- Strong analytical skills
- Strong willingness to learn terahertz imaging system concepts and machine learning
- Demonstrable ability to work on multi-disciplinary projects spanning physics and computer science (strongly preferred)
- Experience of working on Terahertz imaging (optional) or familiarity with fundamentals of terahertz imaging (preferred)
- Experience in programming: Python (preferred) or Matlab (Required)
- Experience in experimental data analysis (required), machine learning (optional) and/or compressed sensing (optional)
- Experience with machine learning libraries such as scikit-learn, pytorch, or tensor-flow (preferred)
- Strong background in linear algebra, probability theory and calculus
- Strong academic writing and communication skills (required)
Further details:
Terahertz imaging:
Terahertz (THz) pulsed imaging is a new technique with high resolution (about 20 μm) and has only emerged recently as a potential new clinical tool for medical imaging. It is a totally non-destructive and non-ionising imaging modality as the average power of the pulse for producing the radiation is as small as 100 nW. The radiation produced is focused onto the sample of interest and then detected coherently. A point measurement is analogous to an ultrasound A-scan. Reflections off different layers are used to determine the structure at various depths. THz light is very sensitive to hydrogen bonds. By Fourier transforming the time-domain data, useful spectroscopic information can also be revealed. Potential applications range from security imaging to medical diagnosis, but the analysis and instrumentation needs to be tailored for each application investigated. THz light can be generated and detected using photoconductive antennas. The usable frequency range of a THz system depends on the photoconductive devices as well as other factors such as optical alignment.
Existing Work:
As recently reported by Sky International News (https://vimeo.com/757511937/c9a3e8cbd7), we have developed a handheld THz probe that is now being used in the first clinical trials of their kind at the University hospital in Coventry. The aim of this project is to take and analyse patient and volunteer data to develop classification algorithms for skin conditions. For example, it may be possible to use THz imaging to detect early stages of skin cancer, or the lateral extent of tumours that are beneath the skin’s surface, and not yet visible. Applicants do not need a medical background, more importantly, they need an interest in applying maths, programming to real world applications.
Supervisory Team:
The supervisory team consists of Prof. Emma MacPherson and Dr. Fayyaz Minhas who have significant experience in Terahertz imaging systems (MacPherson) including the development of clinical trial grade handheld THz probes and development of machine learning algorithms in biomedical informatics.
Prof MacPherson joined Warwick University in 2017 and received a Wolfson Merit award to support her research in the UK. The following links give more information about the Ultrafast THz Photonics group at Warwick and Prof MacPherson’s research group. This PhD project is part of the £8M EPSRC programme grant, Terabotics which spans Warwick, Leeds and Exeter universities.
Webpage: https://warwick.ac.uk/fac/sci/physics/staff/academic/emacpherson/
Dr. Fayyaz Minhas is an associate professor at Warwick Department of Computer Science and the Warwick Tissue Image Analysis Centre. He works on solving problems in biology and medicine using machine learning methods as well as the development of bespoke machine learning algorithms in the domains of biomedical informatics and computational pathology.
Webpage: https://warwick.ac.uk/fac/sci/dcs/people/fayyaz_minhas/
Continue with Facebook
