An evaluation of post-production facial composite enhancement techniques
| Published date | 09 November 2015 |
| Date | 09 November 2015 |
| Pages | 307-318 |
| DOI | https://doi.org/10.1108/JFP-08-2015-0042 |
| Author | Josh P Davis,Stacie Simmons,Lucy Sulley,Chris Solomon,Stuart Gibson |
| Subject Matter | Health & social care,Criminology & forensic psychology,Forensic practice |
An evaluation of post-production facial
composite enhancement techniques
Josh P. Davis, Stacie Simmons, Lucy Sulley, Chris Solomon and Stuart Gibson
Dr Josh P. Davis is Senior
Lecturer at the Department of
Psychology, Social Work and
Counselling, University of
Greenwich, London, UK.
Stacie Simmons and Lucy Sulley,
both are based at the
Department of Psychology,
Social Work and Counselling,
University of Greenwich,
London, UK.
Dr Chris Solomon is Reader at the
School of Physical Sciences,
University of Kent, Canterbury, UK.
Dr Stuart Gibson is based at the
School of Physical Sciences,
University of Kent, Canterbury, UK.
Abstract
Purpose –The purpose of this paper is to describe four experiments evaluating post-production
enhancement techniques with facial composites mainly created using the EFIT-V holistic system.
Design/methodology/approach –Experiments1-4 were conducted in two stages.In Stage 1, constructors
created between one and four individual composites of unfamiliar targets. These were merged to create
morphs. Additionally in Experiment 3, composites were vertically stretched. In Stage 2, participants familiar
with the targets named or provided target-similarity ratings to the images.
Findings –In Experiments 1-3, correct naming rates were significantly higher to between-witness 4-morphs,
within-witness 4-morphs and vertically stretched composites than to individual composites. In Experiment 4,
there was a positive relationship between composite-target similarity ratings and between-witness
morph-size (2-, 4-, 8-, 16-morphs).
Practical implications –The likelihood of a facial composite being recognised can be improved by
morphing and vertical stretch.
Originality/value –This paper improves knowledge of the theoretical underpinnings of these facial
composite post-production enhancement techniques. This should encourage acceptance by the criminal
justice system, and lead to better detection outcomes.
Keywords Face recognition, Memory, E-FIT, EFIT-V, Eyewitness, Facial composite, Morphing,
Vertical stretch
Paper type Research paper
Introduction
In a police investigation, an eyewitness may from memory construct a facial composite of the
suspect, with the aim that someone familiar with that suspect recognises that image. Traditional
mechanical (e.g. Identikit, Photofit) and computerised (e.g. E-FIT, FACES, PRO-fit) feature-based
systems required witnesses to describe the offender’s face, and to assemble individual facial
features. This is challenging as humans mainly process faces holistically as Gestalts (Tanaka and
Farah, 1993). Consequently, these composites are often poor likenesses (for reviews see Davies
and Valentine, 2007; Frowd, 2015; see Fodarella et al., 2015, for descriptions of composite
systems). With holistic systems (e.g. EFIT-V, EvoFIT, ID), witnesses select computer-generated
realistic whole face images from a series of arrays until the final composite matches their memory
of the suspect. Accessing recognition, rather than recall, holistic composites are often better
likenesses than feature-based composites (Davis et al., 2010; Frowd et al., 2007b), enhancing
police suspect identification rates (Frowd et al., 2010; Solomon et al., 2012).
Regardless of system, composite quality depends on other factors including the witness’
memory, the event (e.g. offender viewing conditions) and offender characteristics. For instance,
distinctive faces promote composite naming (e.g. Frowd et al., 2005). Post-production
techniques such as displaying multiple composites (Frowd et al., 2006), dynamic animated
Received 4 August 2015
Revised 4 August 2015
Accepted 4 August 2015
The authors would like to thank Josi
Cakebread, Amy Skelton, Sarah
Thorniley, Helen Little, Sarah Poland,
Natalie Baker, Sheena Belfon, and
Lala Jammeh for their assistance in
collecting data for this project. The
research was partly funded by two
internal University of Greenwich
grants to the first author. Part of this
research was presented at the
American Psychology-Law Society
Annual Meeting, 4th International
Congress of Psychology and Law,
Miami, Florida, USA, March 2011.
Chris Solomon and Stuart Gibson
are faculty members of the
University of Kent and directors of
VisionMetric Ltd. VisionMetric Ltd
markets the EFITV and E-FIT facial
composite systems. Solomon and
Gibson’s contribution to this work
was to facilitate the data collection
that took place at the University of
Kent, development of the software
and provision of software support.
Data collection, analyses and
interpretation were performed by
Davis, Simmons, and Sulley.
DOI 10.1108/JFP-08-2015-0042 VOL. 17 NO. 4 2015, pp. 307-318, © Emerald Group Publishing Limited, ISSN 2050-8794
j
JOURNAL OF FORENSIC PRACTICE
j
PAG E 30 7
To continue reading
Request your trial