Forensic Facial Composites

Frowd, Charlie orcid iconORCID: 0000-0002-5082-1259 (2021) Forensic Facial Composites. In: Methods, Measures, and Theories in Eyewitness Identification Tasks. Routledge, New York, pp. 34-64. ISBN 9781138612549

[thumbnail of Author Accepted Manuscript]
PDF (Author Accepted Manuscript) - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.


Official URL:


There are many types of evidence available after a crime has been committed. Some evidence is physical, including fingerprints, footprints and DNA (e.g., Bradbury & Feist, 2005). Forensic officers attempt to recover information of this type from a crime scene for analysis. The hope, of course, is that the outcome will lead to identification the person(s) responsible. Success usually depends on recidivism: an offender has been convicted of a previous offence and his or her identifying information is available to be searched on a database of fingerprints, footprints, DNA, etc.

A similar situation applies to CCTV footage (e.g., Davies & Thasen, 2000). The hope now is that the offender has been caught on camera and a clear image of the face can be extracted. The image can then be compared against a database of known offenders for potential matches. It can also be circulated within a police force, or in the media, with the aim that someone who is familiar with the person will report him or her to the authorities. This use of evidence is psychological in nature due to the need for a human to recognise the face; it is considered by Richard Kemp and colleagues in Chapter 1 of this volume. A different situation also involving human recognition occurs when police officers have only seen a photograph of a wanted person, and so the face is unfamiliar, as discussed by Kara Moore and James Lampinen in Chapter 3.

As part of collecting all available evidence, the police will interview victims or bystanders who were present when the crime took place. Specific interviewing techniques have been developed for observers who may be able to provide an account of the crime and those involved (e.g., Fisher, 1995). Evidence collected in this way can be particularly important in the absence of the aforementioned evidence, or when recovering useful evidence may take a long time such that further crime occurs.
One (or more) of these observers might also be able to construct a composite image of the offender’s face. This part of a person’s appearance is particularly important for identification—although gait, voice, characteristic motion and possibly other human characteristics may facilitate recognition to some extent (as mentioned later). As with CCTV evidence, an image is circulated for identification. Again, the aim is that someone who is familiar with the person will name the face to police, providing an investigation with a potential suspect (e.g., Ellis & Shepherd, 1992). Subsequent police-work will collect evidence to assess whether or not the named person is likely to be responsible for the crime. Sometimes, a facial composite is constructed of someone with whom the police would like to make contact in order to eliminate him or her from the investigation, or to locate a potential witness, but composites are usually created of an offender, specifically a perpetrator of serious crime.

In this chapter, my aim is to assess facial composites as a reliable method of identification. Face-production systems have changed greatly over the years, as have methods of their deployment. Early ‘feature’ systems in the 1970s and 1980s are not reliable, nor are their computerised descendants, but the newer ‘holistic’ systems (esp. EvoFIT) can now create identifiable faces, for the first time providing a reliable means to identify offenders using this forensic technique. I also look to the future to consider promising techniques that may allow composites to be made even more effective.

Repository Staff Only: item control page