Dundee team triumphs in global algorithm contest

The winning plaques

26 August 2014

A University of Dundee team has won an international competition to create software to help in the diagnosis of connective tissue diseases such as rheumatoid arthritis.

The six-strong team from the School of Computing’s Computer Vision and Image Processing group, triumphed over entries from 11 other institutions, including Cambridge and the Russian Academy of Science, to win the prestigious accolade.

Organisers challenged entrants to create an algorithm which can tell the difference between patterns of cells in images extracted using a technique called Indirect Immunofluorescence.

“When I heard about the competition I had a hunch we’d do well based on the skills and expertise we had in the group,” said Professor Stephen McKenna, Chair of Computer Vision.

After registering, Stephen and his team downloaded a dataset of thousands of images collected by the Queensland Medical Testing Laboratory in Australia.

They then worked on the problem over a period of 3 months, on top of their normal research commitments.

“What was nice about it was that it got people from different projects working together, and it clearly motivated us to solve the problem. That was great for everybody involved.”

PhD student Shazia Akbar travelled to Stockholm, where the International Conference on Pattern Recognition (ICPR) is being held, to accept the prize on behalf of the University.

The full team is as follows: PhD students Siyamalan Manivannan, Wenqi Li and Shazia Akbar, and School of Computing staff Dr Ruixuan (Roy) Wang, Dr Jianguo Zhang and Professor Stephen McKenna.

Stephen said there was no financial reward for winning but the team would publish papers from the work.

“We’re also considering what the next steps should be to take this work forward and how best to develop it further.”

The Universities of Salerno and Queensland - in conjunction with ICPR – organised the competition, which was entitled: “Performance Evaluation of Indirect Immunofluorescence Image Analysis Systems.”

There were two categories, both of which were won by the University of Dundee team.

Dr Janet Hughes, Dean of the School of Computing, said: “This is quite terrific news – a double win at such a prestigious conference is quite outstanding and deserves double congratulations.”

Read more at the College of Medicine, Dentistry & Nursing website.

Can't find what you're looking for? Contact Us