Case Study: Algorithms for forensic imaging

Television and newspaper reports of serious crime are frequently accompanied by E-Fit images of the suspects the police wish to interview. Improvements in the accuracy and composition of photo-fit and E-Fit systems, would assist witnesses in recognition and the police in suspect identification. The Industrial Mathematics KTN awarded a CASE studentship at Kent University's Forensic Imaging Group to help engineer a novel facial composite system.

The research aims to place evolutionary, multidimensional search strategies on a rigorous mathematical base, with research into compact, parametric representations of the human face based on linear and non-linear models which will ultimately lead to better E-Fit tools. The work encourages hands-on industrial and commercial collaboration, by the student, through work with the industrial partner VisionMetric Ltd.

Chris Solomon, Technical Director of VisionMetric Ltd. said that "the grant review was very thorough" and that "the KTN was very constructive, and improved the project".