The UKBB has a repository of over 100,000 clinical eye images that hold a wealth of information about the natural history of eye disease. To analyse these images individually would take trained professionals several years to complete at significant cost. In addition, medical images often contain a lot of detail and subtlety making automated computational analysis very difficult. Crowdsourcing is an emerging concept that has attracted significant attention in recent years. Put simply, it is the process of outsourcing numerous tasks to many untrained individuals. By asking tens of thousands of people their opinion, crowdsourcing has the potential for dramatic financial benefits in large scale image analysis and can provide rapid analysis of pattern-recognition tasks that may otherwise be very challenging to solve. We have developed an interactive online training module and webpage and we plan to promote public participation to assist in the classification of ophthalmic medical images. The crowdsourcing research group aims to develop the technique pioneered in the development of the Oxford English Dictionary, genealogy and other branches of science including astronomy, for possible future use in very large scale datasets such as UK Biobank.

Research Group Team

  • Dr Danny Mitry – Edinburgh University/Moorfields Eye Hospital, London
  • Dr Tunde Peto - Queens University Belfast
  • Prof James Morgan – Cardiff University
  • Prof Paul Foster – Moorfields & UCL Institute of Ophthalmology
  • Prof Manuel Trucco – University of Dundee
  • Dr Tom McGillivray - University of Edinburgh