An evaluation of US systems for facial composite production

Abstract

Witness and victims of serious crime are normally requested to construct a facial composite of a suspect's face. While modern systems for constructing composites have been evaluated extensively in the UK, this is not the case in the US. In the current work, two popular computerized systems in the US, FACES and Identikit 2000, were evaluated against a 'reference' system, PRO-fit, where performance is established. In Experiment 1, witnesses constructed a composite with both PRO-fit and FACES using a realistic procedure. The resulting composites were very poorly named, but the PRO-fit emerged best in 'cued' naming and two supplementary measures: composite sorting and likeness ratings. In Experiment 2, PRO-fit was compared with Identikit 2000, a sketch-like feature system. Spontaneous naming was again very poor, but both cued naming and sorting suggested that the systems were similar. The results support previous findings that modern systems do not produce identifiable composites.

Configural and featural information in facial-composite images

“…In fact, some composites are of such poor quality that they are rarely if ever correctly named (e.g. Frowd et al, 2005aFrowd et al, , 2007c), a situation, as mentioned above, which would lead to low experimental power and ineffective analyses for the current study. We overcame this issue by selecting celebrity images which people had correctly named from previous projects (Frowd et al, 2004(Frowd et al, , 2005b(Frowd et al, , 2007b.…”

“…Frowd et al, 2007c), and so it is sensible to consider experimental power and effect size. The aim was to detect a medium effect size, to be of practical significance, and this was achieved by estimating the number of composites required per condition.…”

The benefit of context for facial-composite construction

“…We should acknowledge that changes made to PRO-fit for these studies necessitated using the software differently to how it was originally intended. However, developers of systems using isolated-feature selection (e.g., the FACES system that is popular in the US, Frowd et al, 2007b) should be aware that their software could be usefully improved based on these findings.…”

“…The composites produced in this study were unlikely to be of good quality, since the constructors constructed the images themselves rather than with a trained composite-system operator, and so a sensitive measure of composite quality was used (Frowd et al, 2007b): naming participants were shown composites from both context conditions and selected an identity for each from a list of written names corresponding to the identities. This so-called constrained-naming task aimed to facilitate performance.…”

“…Frowd et al, 2005aFrowd et al, , 2007bFrowd et al, , 2016. Research has demonstrated that such a delay negatively affects both face recall and recognition (e.g.…”