Jamie Burke

Thesis title: Image Analysis and Machine Learning for OCT Image Sequences in Non-Ocular Disease

Biography

Background

Studied mathematics as an undergraduate at the University of Edinburgh (2015-2019) before starting a PhD in September 2019 in the MRC Precision Medicine DTP researching image processing and machine learning with specific application to retinal and choroidal imaging using optical coherence tomography image data.

Jamie is now a third year student currently developing and applying novel image processing tools to ophthalmological image datasets in search for ocular imaging biomarkers of systemic disease.

Qualifications

Awarded a BSc(Hons) in Mathematics from the University of Edinburgh in July 2019.

Undergraduate teaching

I am lead tutor for the semester 1 advanced calculus course Advanced Methods of Applied Mathematics (MATH10086).

Postgraduate teaching

Since September 2020 I have been tutoring the semester 1 programming course Python Programming (MATH11199). Since January 2021 I have been tutoring the semester 2 course Machine Learning in Python (MATH11205).

Open to PhD supervision enquiries?

No

Research summary

Main research topics of interest include:

  • Image processing;
  • Optical coherence tomography;
  • Ophthalmic image analysis;
  • Systemic disease;
  • Machine learning;
  • Deep learning;
  • Bayesian statistics and modelling;
  • Network analysis;
  • Multimorbidity.

Current research interests

Current research involves the development of medical computer vision models for automated ophthalmic image analysis. I mainly work with optical coherence tomography (OCT) image data which visualises retinal and choroidal structures at the back of the eye. My main research focus is on developing image processing methodologies to extract meaningful information on these anatomical structures, with application to how these anatomies may change in response to treatment or development of systemic disease. From a technical perspective, I have built several image processing methodologies based on machine learning tools such as Gaussian process regression, clustering and quantisation, and artificial intelligence. These research problems were motivated by common computer vision tasks for OCT such as image segmentation, contrast enhancement and speckle denoising. From a clinical perspective, these tools are being applied to novel OCT datasets of cohorts from a diverse range of systemic disease, including chronic kidney disease, neurodegenerative disease and sepsis. Besides computer vision in ophthalmology, I have an avid interest in healthcare data analytics and the development of new analytical tools to better understand multimorbidity. These tools include include network analysis and probabilistic modelling.
This article was published on