Matching unfamiliar faces is a challenging task and can be error prone, however the difficulty of the task is not widely known. Matching faces provides vital intelligence and evidence in police investigations and the court room.
The aim of this project is to investigate ways of improving accuracy in forensic face matching and face matching more widely in policing, looking at four strands to provide enhanced face matching expertise:
Short professional training courses
Trained forensic examiners
People who are naturally proficient at matching faces (super matchers)
Automated facial recognition technology
The project comprises of empirical evaluations of each of the above using police and forensic practitioners.
The project comprises of five studies:
An international survey of training practices in applied face matching
An evaluation of a professional two-day training course in applied face matching
A comparison of perceptual face-matching skills between untrained superior face matchers and trained forensic examiners
Combining face-matching decisions between humans and algorithms
Evaluating the operational accuracy of individual forensic face matching examiners and examiner teams