February 2007 – A review by Gary Wells and Lisa Hasels of Iowa State University published in Current Directions in Psychological Science has found that recent technological advances in facial composite systems have failed to improve identification and apprehension of criminal suspects.
The era of an artist’s skilful pencil sketches has been replaced by computer software offering witnesses a huge range of different facial features. The review highlights poor results from such facial composite systems. In one study, only 2.8 per cent of participants correctly identified a well-known celebrity created by other participants using face-composite software. In separate research, participants were unable to discriminate composites of their classmates from those of students at different schools.
The report suggests that these poor results do not reflect software deficiencies as such but more a discrepancy between how we remember faces and how composites are produced.
The authors comment:
“Numerous lines of evidence converge on the view that faces are generally processed, stored and retrieved at a holistic level rather than at the level of individual facial features.
“The psychological process of remembering faces may include more complex representations such as multidimensional similarity to other faces or relative sizes and distances of features and so on that are not readily retrieved by memory nor utilized by facial composite software.”
The report highlights encouraging results from early tests using the concept of holistic processing to develop face-composite systems using “whole-face” methods for recall. Generation of a random set of faces enables the witness to select the one most similar to their memory of the perpetrator. This “parent” image is subject to several mutations to produce a set of similar looking faces. The witness again chooses, continuing the process until unable to discriminate from the face in their memory.
The authors fit attempts to improve face-composite systems into the larger problem of the US criminal justice system. Analysis of the first 180 DNA exonerations involved mistaken eyewitness testimony in 75 per cent of cases. Poor composites may also encourage guilty suspects.
Gary Wells commented:
“Imagine the solace of the culprit who sees a composite of his face in the newspaper that looks nothing like his face.”
The report argues that increasing knowledge of memory processes involved in facial recognition will help improve the accuracy of facial composites and reduce the number of those falsely convicted.
The authors conclude:
“As the historical and natural home of the science of memory, psychological science has great promise for helping to solve an age-old problem.”