Facial composite

PhotoFIT generation

Construction of the composite was originally performed by a trained artist, through drawing, sketching, or painting, in consultation with a witness or crime victim. In the 1970s techniques were devised for use by those less artistically skilled, employing interchangeable templates of separate facial features, such as "Photofit" in the UK, Smith & Wesson's "Identi-Kit" in the U.S. and PortraitPad.

In the last two decades, a number of computer based facial composite systems have been introduced; amongst the most widely used systems are SketchCop FACETTE Face Design System Software, "Identi-Kit 2000", FACES, E-FIT and PortraitPad. In the U.S. the FBI maintains that hand-drawing is its preferred method for constructing a facial composite. Many other police agencies, however, use software, since suitable artistic talent is often not available.

Evolutionary systems

Until quite recently, the facial composite systems used by international police forces were exclusively based on a construction methodology in which individual facial features (eyes, nose, mouth, eyebrows, etc.) are selected one at a time from a large database and then electronically 'overlaid' to make the composite image. Such systems are often referred to as feature-based since they essentially rely on the selection of individual features in isolation. However, after a long period of research and development work conducted largely within British Universities, systems based on a rather different principle are finding increasing use by police forces. These systems may be broadly described as holistic or global in that they primarily attempt to create a likeness to the suspect through an evolutionary mechanism in which a witness's response to groups of complete faces (not just features) converges towards an increasingly accurate image. Three such systems have come from academic beginnings, EFIT-V from the University of Kent; EvoFIT (http://www.EvoFIT.co.uk) from the University of Stirling, the University of Central Lancashire (UCLan) and the University of Winchester; and ID from the University of Cape Town, South Africa. GFE (https://sites.google.com/view/facevisionresearch) is an experimental evolutionary face compositing system using image gradient instead of luminance to represent faces, which seems to produce better quality composites.

Research

A general review of research into the evaluation of mechanical template techniques may be found in Davies and Valentine (2006). A review of research into more modern 'feature' and 'recognition' systems, and into methods for improving the effectiveness of composites, may be found in Frowd et al. (2008) and (2009).

The systems used in the UK have been subjected to a number of academic studies. These have typically shown that E-FIT and PRO-fit produce composites that are correctly named, either immediately or a few hours after construction, about 20% of the time (see Brace et al. (2000), Bruce et al. (2002), Davies et al. (2000) and Frowd et al. (2005)). When witnesses in these studies are required to wait two days before constructing a composite, which matches real use more closely, naming falls to a few percent at best (e.g. Frowd et al. [2005] and [2007]). The reason for the low level of naming from these systems appears to be that witnesses are unable to accurately construct the internal features of the face after long delays, the region that is important for recognition by another person later (Frowd et al. [2007]).

Evolutionary systems show a marked improvement in accuracy. In academic trials, research on a fairly-recent version of the EvoFIT system has shown correct naming levels of about 30% after a 2-day delay (see Frowd et al., 2010). Using more-recent construction techniques, the performance increased to 45% correct naming (Frowd et al., 2012). Using the very latest system, interview and enhancement techniques, naming of an EvoFIT composite is 74% correct (Frowd et al., 2013). Appropriately, the system does appear to behave more like a face recognition than a face recall system (Frowd et al., 2011) Accompanying the development of EvoFIT have been new approaches in the type of interview administered to eyewitnesses prior to face construction (e.g. Frowd et al., 2012). Similarly in extensive field use EFIT-V has shown a 40% naming rate over an 18-month period with 1000 interviews. The EvoFIT system has been similarly evaluated in formal police field-trials. These evaluations have reported a much higher naming rate for EvoFIT composites but, using the latest interview techniques, a suspect arrest rate of 60%. This latest police field trial has also indicated that an EvoFIT directly leads to the arrest of a suspect and then a conviction in 29% of cases. There have been many notable successes - for example, in this investigation, EvoFIT has directly led to the arrest of a serial rapist.

Usage

While the classic use of the facial composite is the citizen recognizing the face as an acquaintance, there are other ways where a facial composite can prove useful. The facial composite can contribute in law enforcement in a number of ways:

1. Identifying the suspect in a wanted poster.
2. Additional evidence against a suspect.
3. Assisting investigation in checking leads.
4. Warning vulnerable population against serial offenders.

Facial composites of various types have been used extensively in those television programs which aim to reconstruct major unsolved crimes with a view to gaining information from the members of the public, such as "America's Most Wanted" in the US and "Crimewatch" in the UK.

Notable cases

These notable cases had facial composites assist in identifying the perpetrator: