While there is plenty of empirical evidence which demonstrates how good people are at recognizing familiar faces, we are now starting to realize just how poor people are at unfamiliar face recognition.
National security officials, officers in the criminal justice system and retail staff frequently rely on face recognition to establish and authenticate the identity of an individual. At UK Border Control, officials work to ensure that only those passengers whose passport photo matches their face are allowed to enter the country. In the criminal justice system, police officers often utilise CCTV images as a means of identifying the perpetrator of a crime. In addition, cashiers in retail stores must examine face-photo ID cards in order to prohibit the illegal sale of age-restricted goods. Each of these occupations relies on the ability to detect correctly whether or not the face of an unfamiliar person matches a face photo on an ID card or an image still.
Our reliance on face recognition for identity verification may stem from the fact that in some instances we show a high level of expertise in this area. For example, we are able to recognise familiar faces across a large range of highly variable photos, apparently effortlessly. However, in a striking contrast, recent research has shown that we are surprisingly poor at recognising new instances of an unfamiliar person. This distinction has major implications for applied professions in which accurate unfamiliar face recognition is vital.
As stated above, our reliance on face recognition in applied domains may stem from the assumption that we have a degree of expertise with faces in general. It is likely that this assumption has arisen from our remarkable ability to recognise novel and highly variable instances of familiar people (i.e. family, friends, colleagues, celebrities). However, recent research has shown that this ability does not generalise to the identification of similar instances of unfamiliar people.
For example, Burton et al (2010) developed the Glasgow Face Matching Test (GFMT; see Figure 1 below), a psychometric test of unfamiliar face recognition ability, in which participants are required to decide whether a pair of face photos depicts two instances of the same person (taken seconds apart using different cameras) or two different people. This type of 1-1 matching task is the experimental analogue of the paired comparisons made by passport officers (face-passport photo) and cashiers (face-photo ID) on a daily basis. Despite the GFMT using high quality front-facing images and the insistence that accuracy be prioritised over speed, error rates of between 15% and 20% are common among samples of university student participants on this task.
In contrast to the 1-1 matching tested by the GFMT, Bruce et al (1999) had previously developed a task which modelled a police line-up scenario which involved a series of 1-10 matching arrays. As seen in Figure 3, participants were presented with a single high quality front-face image still of a ‘suspect’ (taken from high quality video footage), below which was presented an array of 10 face photos. Participants were required to decide whether or not there was a novel instance of the suspect in the array (i.e. was he present or absent in the line-up). Error rates on this task were reported as being at 30% despite the array photos being taken on the same day as the video footage, in the same pose and in optimal lighting conditions (note that when Henderson et al (2001) used typical - low quality - CCTV video images, error rates on a similar task rose to 75%). It should be noted that recent work has demonstrated that live face-to-face photo matching studies (Megreya & Burton, 2008; Davis & Valentine, 2009) report equally poor performance.
The findings from these studies establish a critical distinction between the ease with which we recognise familiar faces and the difficulties that we encounter when attempting to recognise a novel instance of an unfamiliar face, even when the viewing conditions are favourable. It cannot be overstated that performance levels on these tasks produce concerning levels of error. In daily life it only requires a single face recognition error by professionals to sell age restricted goods to a minor, to allow an undesirable individual through passport control or to convict an innocent suspect.
This text is an excerpt from a Robertson & Burton (2015) review article published in KEESING: The Journal of Documents and Identity.