Fuelling prosperity and growth

Research that drives innovation

Knowledge exchange and the generation of new knowledge are not straight forward, automatic or even linear. Often these processes are unpredictable: the accumulated evidence points in one direction, but then things turn out quite differently, according to Vicki Bruce FBA, Professor of Psychology at Newcastle University.

‘Progress’, she says, ‘is like moving up a spring, so progress is onwards, but sometimes one appears to go backwards a bit. As you round the turn of the spring you ask questions in a different way or you are getting new technology that allows you to make an advance.’

Pioneering UK research on facial recognition has been heavily influential in the way that CCTV images and memories of faces are now used in criminal justice. When Bruce began her research in this area, psychologists were just starting to demonstrate that people could remember hundreds of pictures of faces in laboratory experiments. ‘And at the same time ... there were lots and lots of cases of appalling miscarriages of justice where witnesses had testified that innocent people had committed crimes. We had this extraordinary paradox: that people were good at remembering faces and very bad at remembering faces.’

Bruce worked on how physical features of the human face are recognised with physicists using laser scanning to build 3D surface images. The research was developed on the assumption that the brain recognised faces in three dimensions, but what the evidence proved was in fact that recognition was actually based on much simpler 2D patterns of lights and darks. This sensitivity to image features makes it possible to remember specific pictures very well, but can confound memory in more natural conditions. It was her metaphorical spring in operation – ‘how you can take these twists and turns and do research driven by one question then find something different’. The work in turn spurred study of how computers might recognise faces, and was then applied to the bases of resemblance between CCTV images and suspects, and how misleading these, too, can be.