Finding needles in haystacks: Identity mismatch frequency and facial identity verification.

Abstract

Accurate person identification is central to all security, police, and judicial systems. A commonplace method to achieve this is to compare a photo-ID and the face of its purported owner. The critical aspect of this task is to spot cases in which these two instances of a face do not match. Studies of person identification show that these instances often go undetected when mismatches occur regularly in an experiment, but this differs from everyday operations in which identity mismatches are rare. The current study therefore examined whether infrequent identity mismatches are more likely to go undetected by observers. In Experiments 1 and 2, identity mismatches were detected equally under low (2%) and high (50%) mismatch prevalence. This pattern persisted when viewing conditions were optimized for person identification in Experiment 3, by using a card-sorting task in which all face identities could be viewed repeatedly, and also under increased task difficulty, by constraining viewing conditions temporally in Experiment 4. These results imply that the infrequent occurrence of identity mismatches in security settings such as passport control does not impair an observer's ability to detect these important events.

Human–Computer Interaction in Face Matching

“…To understand this within the context of passport control, the stimuli that were employed in this study portrayed considerable within‐person variability (see Jenkins, White, Van Montfort, & Burton, 2011; Megreya, Sandford, & Burton, 2013), and mismatches occurred infrequently (Bindemann, Avetisyan, & Blackwell, 2010; Papesh & Goldinger, 2014). Furthermore, the proportion of inconsistently labeled and unresolved trials was lower than the proportion of trials with consistent labels, given that the algorithms employed at passport control are projected to be highly accurate, and thus incorrect identifications should occur only rarely (FRONTEX, 2015b).…”

The Effect of Feedback on Face‐Matching Accuracy

“…For this purpose, observers were shown pairs of faces comprising photographs of the same person or two different people and match/mismatch decisions to these facial identities were required (as in, e.g., Bindemann et al, 2010;Burton et al, 2010;Megreya et al, 2011). Different photographs of the same person were provided on identity-match trials to eliminate simple picture-matching strategies (see, e.g., .…”

“…In this field, face-matching accuracy is typically illustrated by combining performance across blocks. Moreover, face-matching is frequently studied in bestcase scenarios, which are designed to produce the highest possible level of performance (see, e.g., Burton et al, 2010;Bindemann et al, 2010Bindemann et al, , 2012Megreya & Burton, 2006, or, conversely, under conditions that make this task particularly difficult (e.g., Bindemann & Sandford, 2011;Henderson et al, 2001;Kemp et al, 1997). Both of these approaches may be insensitive to the gradual decline in mismatch accuracy that was observed here, especially when performance data is also pooled across blocks.…”

The Effect of Image Pixelation on Unfamiliar‐Face Matching

“…Despite its applied usage, face matching is an error-prone task. Under seemingly optimized laboratory conditions, in which pairs of to-be-matched faces are depicted in the same lighting, expression and view, identification errors are typically made 10% to 30% of the time (see, e.g., Bindemann, Avetisyan, & Blackwell, 2010;Bindemann, Avetisyan, & Rakow, 2012;Burton, White & McNeill, 2010;Megreya, Bindemann, & Harvard, 2011;Megreya, White, & Burton, 2011). Accuracy declines even further under different task demands, for example, when a target has to be compared to two , five (Bindemann, Sandford, Gillatt, Avetisyan, & Megreya, 2012;Megreya, Bindemann, Havard, & Burton, 2013) or ten concurrent faces (e.g., Bruce et al, 1999;Megreya & Burton, 2006).…”

“…While face matching is often measured under highly controlled conditions, (e.g., Bindemann et al, 2010Burton et al, 2010;Megreya & Burton, 2006), many factors can make this task more difficult. Of these, poor image quality has been linked consistently to reduced performance in investigations of person identification from CCTV (e.g., Burton, Wilson, Cowan, & Bruce, 1999;Henderson et al, 2001;Liu, Seetzen, Burton, & Chaudhuri, 2003;Lee et al, 2009).…”