About the Project
The University of Dundee has been awarded a £7-million grant from the National Institute of Health Research’s Global Health programme to establish a major new Scotland-India (Dundee - Chennai) clinical partnership to combat diabetes. The new project will see Dundee’s world-leading expertise in the use of genomics linked to medical records to deliver improved care in diabetes `twinned’ with the large patient data set (covering over 400,000 Indian diabetic patients) collected by Dr Mohan’s Diabetes Specialities Centre, the largest clinical network of diabetes care in India.
The project has 10 PhD studentships to allocate, some of which on the development of software tools for the analysis of very large volumes of images from portable and conventional (clinic-based) fundus cameras, for exploring retinal biomarkers in conjunction with extensive, cross-linked data sets available to the project in Dundee and in Chennai.
The successful candidate will join a large and interdisciplinary project team.
He/she will join a computer vision/machine learning team of two RAs in the Dundee VAMPIRE group, part of the Computer Vision and Image Processing group (CVIP) co-ordinated by Professors Emanuele Trucco and Stephen McKenna.
The project VAMPIRE/CVIP team, in turn, will work in close collaboration with the investigators, RAs and PhD students in the Division of Molecular and Clinical Medicine, the Health Informatics Centre, and within the Dundee iCIGHT interdisciplinary initiative (informatics, Clinical Imaging, Genomics and Healthcare Technologies).
In addition, the research will be embedded within the activities of VAMPIRE (Vessel Assessment and Measurement for Images of the Retina), a research initiative started in 2007 by the Universities of Edinburgh and Dundee. VAMPIRE has been growing a suite of software tools supporting research on retinal biomarkers for a range of systemic conditions, but also clinical assessment (lesion detection and quantitative assessment) and rich support for diagnosis. Present and recent VAMPIRE collaborators include Harvard Medical School, Tufts Medical School, UCLA Jules Stein Eye Institute, and ASTAR Singapore.
RESEARCH
• You will develop, and contribute to develop,
o algorithms and software for retinal image analysis, including their systematic validation on data provided by the project and public data sets, for conventional and portable fundus cameras; this will benefit from, and build on, the substantial experience and software available in the VAMPIRE team (vampire.computing.dundee.ac.uk).
o an investigation of the associations between quantitative retinal properties and diabetes-related clinical outcomes (kidney complications, diabetic retinopathy, others) based on machine learning and deep learning, in collaboration with the clinical statistics team, and harmonize findings from machine learning and conventional statistics.
• You will be trained in the use of the VAMPIRE package for semi-automatic measurement of large quantities of fundus-camera retinal images, and participate in a program of quantitation of retinal vascular properties in large image sets (tens and hundreds of thousands).
• You will be prepared to co-supervise VAMPIRE and CVIP MSc and final-year CompSci students choosing projects related to the program.
• You will collaborate closely with the RAs and other PhDs in the Dundee team, collectively taking responsibility for co-organizing regular, technical project meeting in Dundee, as well as collaborate with the India team with modalities to be defined, including regular skype meetings and email contacts.
• You will be prepared to take part in project meetings in Dundee and in Chennai, as needed, prepare any necessary presentation material, and contribute to organization as appropriate.
• You will write, and contribute to write, high-quality technical papers for conferences and journals.
• You will have opportunities to contribute to Computer Science taught modules, typically as paid lab tutor.
• Where appropriate, you will contribute to public engagements activities, e.g. Science Festivals, Café Science and others.
REQUIREMENTS
We are looking for candidates with:
• a graduate degree in a numerate discipline with a very good to excellent score;
• experience of image processing, computer vision and machine learning via taught courses and projects,
• ideally medical image analysis and specifically retinal image analysis;
• excellent skills in programming (e.g. MATLAB, C++) under Windows and/or UNIX;
• a strong commitment to carry out interdisciplinary research;
• an open, collaborative and collegiate attitude to working in a group
• excellent communication skills, written and oral.
• Experience of retinal image analysis and computer vision, especially if supported by publications, will be a definite advantage,
• as will exposure to or participation in clinical studies.
• You will be ready to move to Dundee, Scotland, for the full duration of your PhD.
RESTRICTIONS
Please notice that NIHR funding rules impose that ****ONLY INDIAN NATIONALS CAN APPLY FOR THESE POSITIONS****