Super-recognisers in Action: Evidence from Face-matching and Face Memory Tasks

Trust in automation and the accuracy of human-algorithm teams performing one-to-one face matching tasks

The human face is commonly used for identity verification. While this task was once exclusively performed by humans, technological advancements have seen automated facial recognition systems (AFRS) integrated into many identification scenarios. Although many state-of-the-art AFRS are exceptionally accurate, they often require human oversight or involvement, such that a human operator actions the final decision. Previously, we have shown that on average, humans assisted by a simulated AFRS (sAFRS) failed to reach the level of accuracy achieved by the same sAFRS alone, due to overturning the system's correct decisions and/or failing to correct sAFRS errors. The aim of the current study was to investigate whether participants' trust in automation was related to their performance on a one-to-one face matching task when assisted by a sAFRS. Participants (n = 160) completed a standard face matching task in two phases: an unassisted baseline phase, and an assisted phase where they were shown the identification decision (95% accurate) made by a sAFRS prior to submitting their own decision. While most participants improved with sAFRS assistance, those with greater relative trust in automation achieved larger gains in performance. However, the average aided performance of participants still failed to reach that of the sAFRS alone, regardless of trust status. Nonetheless, further analysis revealed a small sample of participants who achieved 100% accuracy when aided by the sAFRS. Our results speak to the importance of considering individual differences when selecting employees for roles requiring human-algorithm interaction, including identity verification tasks that incorporate facial recognition technologies.

Face Feature Change Detection Ability in Developmental Prosopagnosia and Super-Recognisers

In non-clinical populations, facial features (eyes, nose, mouth) may vary in their contribution to face identity perception. Changes to whole faces are easier to detect than changes to individual features, and eye changes are typically easier to detect than mouth changes, which in turn are easier to detect than nose changes. However, how this differs for people with face recognition difficulties (developmental prosopagnosia; DP) and for individuals with superior face recognition abilities (super-recognisers; SR) is not clear; although findings from previous studies have suggested differences, the nature of this difference is not understood. The aim of this study was to examine whether differences in the ability to detect feature changes in DPs and SRs were (a) quantitative, meaning that the pattern across feature changes remained the same but there was an overall upwards or downwards shift in performance, or (b) qualitative, meaning that the pattern across feature changes was different. Using a change detection task in which individual face features (eyes, nose, mouth) changed between sequentially presented faces, we found that while prosopagnosics showed a quantitative difference in performance with a downwards shift across all conditions, super-recognisers only showed qualitative differences: they were better able to detect when the face was the same and were marginally (but not non-significantly) worse at detecting when the eyes changed. Further, the only condition which distinguished between the three groups was the ability to identify when the same face was presented, with SRs being better than controls, and controls being better than DPs. Our findings suggest that, in feature-matching tasks, differences for DPs are due to them being overall worse at the task, while SRs use a qualitatively different strategy.

Face matching and self-insight: A Registered Report investigating individual differences in metacognitive sensitivity, efficiency, and bias

Deciding whether two face photographs depict the same person or not can be a challenging task, and there are substantial individual differences in face matching ability. Far less is known about differences in metacognitive ability; that is, how well people can estimate the quality of their own face matching judgements. The purpose of this Registered Report was to determine the relationship between first-order performance in a face matching task, and three metacognitive measures: metacognitive sensitivity (the information exploited by metacognition), metacognitive efficiency (the quality of metacognitive processing itself), and metacognitive bias (the overall tendency towards high or low confidence). Participants completed a 200-trial unfamiliar face matching task, providing a second-order (metacognitive) confidence rating after each first-order (cognitive) response. Cognitive performance and metacognitive sensitivity showed a substantial positive association (rs = .60), suggesting that they were based on overlapping information, with poorer performers having lower quality information available. Once these differences in the information available to metacognition were accounted for, the quality of metacognitive processing itself (metacognitive efficiency) did not vary systematically with cognitive performance (rs = .00). Moreover, poor performers were not less or more confident overall than good performers, as cognitive performance and metacognitive bias did not correlate significantly (rs = .11). These findings allow us to move beyond consideration of group-level insight and inform regarding individual differences in cognitive and metacognitive abilities.

Face identification proficiency test designed using item response theory

Measures of face-identification proficiency are essential to ensure accurate and consistent performance by professional forensic face examiners and others who perform face-identification tasks in applied scenarios. Current proficiency tests rely on static sets of stimulus items and so cannot be administered validly to the same individual multiple times. To create a proficiency test, a large number of items of "known" difficulty must be assembled. Multiple tests of equal difficulty can be constructed then using subsets of items. We introduce the Triad Identity Matching (TIM) test and evaluate it using item response theory (IRT). Participants view face-image "triads" (N = 225) (two images of one identity, one image of a different identity) and select the different identity. In Experiment 3, university students (N = 197) showed wide-ranging accuracy on the TIM test, and IRT modeling demonstrated that the TIM items span various difficulty levels. In Experiment 3, we used IRT-based item metrics to partition the test into subsets of specific difficulties. Simulations showed that subsets of the TIM items yielded reliable estimates of subject ability. In Experiments 3a and b, we found that the student-derived IRT model reliably evaluated the ability of non-student participants and that ability generalized across different test sessions. In Experiment 3c, we show that TIM test performance correlates with other common face-recognition tests. In summary, the TIM test provides a starting point for developing a framework that is flexible and calibrated to measure proficiency across various ability levels (e.g., professionals or populations with face-processing deficits).

Face masks and fake masks: the effect of real and superimposed masks on face matching with super-recognisers, typical observers, and algorithms

Mask wearing has been required in various settings since the outbreak of COVID-19, and research has shown that identity judgements are difficult for faces wearing masks. To date, however, the majority of experiments on face identification with masked faces tested humans and computer algorithms using images with superimposed masks rather than images of people wearing real face coverings. In three experiments we test humans (control participants and super-recognisers) and algorithms with images showing different types of face coverings. In all experiments we tested matching concealed or unconcealed faces to an unconcealed reference image, and we found a consistent decrease in face matching accuracy with masked compared to unconcealed faces. In Experiment 1, typical human observers were most accurate at face matching with unconcealed images, and poorer for three different types of superimposed mask conditions. In Experiment 2, we tested both typical observers and super-recognisers with superimposed and real face masks, and found that performance was poorer for real compared to superimposed masks. The same pattern was observed in Experiment 3 with algorithms. Our results highlight the importance of testing both humans and algorithms with real face masks, as using only superimposed masks may underestimate their detrimental effect on face identification.

The face inversion effect does not provide a pure measure of holistic face processing

It is widely held that upright faces are processed more holistically than inverted faces and that this difference is reflected in the face inversion effect. It is not clear, however, how the inversion effect can best be measured, whether it is task specific, or even whether it specifically correlates with processing of upright faces. We examined these questions in a large sample (N = 420) who provided data on processing of upright and inverted stimuli in two different tasks with faces and one with objects. We find that the inversion effects are task dependent, and that they do not correlate better among face processing tasks than they do across face and object processing tasks. These findings were obtained regardless of whether inversion effects were measured by means of difference scores or regression. In comparison, only inversion effects based on regression predicted performance with upright faces in tasks other than those the inversion effects were derived from. Critically, however, inversion effects based on regression also predicted performance with inverted faces to a similar degree as they predicted performance with upright faces. Consequently, and contrary to what is commonly assumed, inversion effects do not seem to capture effects specific to holistic processing of upright faces. While the present findings do not bring us closer to an understanding of which changes in cognitive processing are induced by inversion, they do suggest that inversion effects do not reflect a unitary construct; an implicit assumption that seems to characterize much of the research regarding face processing.

Investigating people's metacognitive insight into their own face abilities

Within the domain of face processing, researchers have been interested in quantifying the relationship between objective (i.e., performance on laboratory tests of recognition and matching) and subjective measures of ability (typically, self-report questionnaires). Put simply, do people show high levels of metacognitive insight into their own abilities with faces? Although several studies have suggested that the association between these two types of measures may only be moderate, there remain several important issues that require consideration before this question can be sensibly investigated. First, specificity is needed regarding both objective and subjective measurements because both tend to span a wide range of potentially separable abilities. Second, experimental tasks appear to focus on different contexts to those tapped in self-report questionnaire items. Third, recent issues with statistical approaches and visualisation can result in numerical artefacts and misinterpretations. Finally, the sizes of population-level insights suggested by recent work provide only limited information regarding individuals within these populations, and so researchers aiming to identify people at the extremes of ability must be careful when drawing conclusions. Taken together, we argue that more attention to these issues is needed when attempting to investigate metacognitive insight within this domain.

A new approach to diagnosing and researching developmental prosopagnosia: Excluded cases are impaired too

Developmental prosopagnosia is characterized by severe, lifelong difficulties when recognizing facial identity. Unfortunately, the most common diagnostic assessment (Cambridge Face Memory Test) misses 50-65% of individuals who believe that they have this condition. This results in such excluded cases' absence from scientific knowledge, effect sizes of impairment potentially overestimated, treatment efficacy underrated, and may elicit in them a negative experience of research. To estimate their symptomology and group-level impairments in face processing, we recruited a large cohort who believes that they have prosopagnosia. Matching prior reports, 56% did not meet criteria on the Cambridge Face Memory Test. However, the severity of their prosopagnosia symptoms and holistic perception deficits were comparable to those who did meet criteria. Excluded cases also exhibited face perception and memory impairments that were roughly one standard deviation below neurotypical norms, indicating the presence of objective problems. As the prosopagnosia index correctly classified virtually every case, we propose it should be the primary method for providing a diagnosis, prior to subtype categorization. We present researchers with a plan on how they can analyze these excluded prosopagnosia cases in their future work without negatively impacting their traditional findings. We anticipate such inclusion will enhance scientific knowledge, more accurately estimate effect sizes of impairments and treatments, and identify commonalities and distinctions between these different forms of prosopagnosia. Owing to their atypicalities in visual perception, we recommend that the prosopagnosia index should be used to screen out potential prosopagnosia cases from broader vision research.

Data-driven studies in face identity processing rely on the quality of the tests and data sets

There is growing interest in how data-driven approaches can help understand individual differences in face identity processing (FIP). However, researchers employ various FIP tests interchangeably, and it is unclear whether these tests 1) measure the same underlying ability/ies and processes (e.g., confirmation of identity match or elimination of identity match) 2) are reliable, 3) provide consistent performance for individuals across tests online and in laboratory. Together these factors would influence the outcomes of data-driven analyses. Here, we asked 211 participants to perform eight tests frequently reported in the literature. We used Principal Component Analysis and Agglomerative Clustering to determine factors underpinning performance. Importantly, we examined the reliability of these tests, relationships between them, and quantified participant consistency across tests. Our findings show that participants' performance can be split into two factors (called here confirmation and elimination of an identity match) and that participants cluster according to whether they are strong on one of the factors or equally on both. We found that the reliability of these tests is at best moderate, the correlations between them are weak, and that the consistency in participant performance across tests and is low. Developing reliable and valid measures of FIP and consistently scrutinising existing ones will be key for drawing meaningful conclusions from data-driven studies.

Improving forensic perpetrator identification with Super-Recognizers

About a decade ago, Super-Recognizers (SRs) were first described as individuals with exceptional face identity processing abilities. Since then, various tests have been developed or adapted to assess individuals' abilities and identify SRs. The extant literature suggests that SRs may be beneficial in police tasks requiring individual identification. However, in reality, the performance of SRs has never been examined using authentic forensic material. This not only limits the external validity of test procedures used to identify SRs, but also claims concerning their deployment in policing. Here, we report the first-ever investigation of SRs' ability to identify perpetrators using authentic case material. We report the data of 73 SRs and 45 control participants. These include (a) performance on three challenging tests of face identity processing recommended by Ramon (2021) for SR identification; (b) performance for perpetrator identification using four CCTV sequences depicting five perpetrators and police line-ups created for criminal investigation purposes. Our findings demonstrate that the face identity processing tests used here are valid in measuring such abilities and identifying SRs. Moreover, SRs excel at perpetrator identification relative to control participants, with more correct perpetrator identifications, the better their performance across lab tests. These results provide external validity for the recently proposed diagnostic framework and its tests used for SR identification (Ramon, 2021). This study provides the first empirical evidence that SRs identified using these measures can be beneficial for forensic perpetrator identification. We discuss theoretical and practical implications for law enforcement, whose procedures can be improved via a human-centric approach centered around individuals with superior abilities.

Face matching and metacognition: investigating individual differences and a training intervention

Background

Although researchers have begun to consider metacognitive insight during face matching, little is known about the underlying mechanism. Here, I investigated whether objective ability, as well as self-assessed ability, were able to predict metacognitive performance, that is, the ability to differentiate correct and incorrect responses in terms of confidence. In addition, I considered whether a training intervention resulted in improvements to both face matching performance and metacognitive insight.

Methods

In this experiment (N = 220), participants completed a face matching task, with either a diagnostic feature training course or a control course presented at the halfway point. In addition, a second face matching task, as well as a self-report questionnaire regarding ability, were completed to provide measures of objective and self-assessed ability respectively.

Results

Higher self-assessed ability with faces, as well as higher objective ability with face matching, predicted better metacognitive performance, i.e., greater confidence in correct, in comparison with incorrect, responses. This pattern of results was evident both when objective ability was measured through performance on the same task used to measure metacognitive insight and when a different task was used. Finally, the training intervention failed to produce improvements in face matching performance and showed no evidence of altering metacognitive ability.

Discussion

The current work begins to address the mechanism underlying individual differences in metacognitive insight during face matching. Although support was provided for a competence-based account, where better face matchers showed greater performance on the task and were more successful in monitoring their performance, further work might focus on decoupling task performance and competence in order to more conclusively explain why some people are more insightful than others.

Idiosyncratic biases in the perception of medical images

Introduction

Radiologists routinely make life-altering decisions. Optimizing these decisions has been an important goal for many years and has prompted a great deal of research on the basic perceptual mechanisms that underlie radiologists' decisions. Previous studies have found that there are substantial individual differences in radiologists' diagnostic performance (e.g., sensitivity) due to experience, training, or search strategies. In addition to variations in sensitivity, however, another possibility is that radiologists might have perceptual biases-systematic misperceptions of visual stimuli. Although a great deal of research has investigated radiologist sensitivity, very little has explored the presence of perceptual biases or the individual differences in these.

Methods

Here, we test whether radiologists' have perceptual biases using controlled artificial and Generative Adversarial Networks-generated realistic medical images. In Experiment 1, observers adjusted the appearance of simulated tumors to match the previously shown targets. In Experiment 2, observers were shown with a mix of real and GAN-generated CT lesion images and they rated the realness of each image.

Results

We show that every tested individual radiologist was characterized by unique and systematic perceptual biases; these perceptual biases cannot be simply explained by attentional differences, and they can be observed in different imaging modalities and task settings, suggesting that idiosyncratic biases in medical image perception may widely exist.

Discussion

Characterizing and understanding these biases could be important for many practical settings such as training, pairing readers, and career selection for radiologists. These results may have consequential implications for many other fields as well, where individual observers are the linchpins for life-altering perceptual decisions.

Face Memory and Face Matching

Abstract. The Cambridge Face Memory Test Long (CFMT+) and the Glasgow Face Matching Test Short (GFMT-S) are frequently used tests in face recognition research. No test-retest results in conjunction with internal consistency, mean inter-item correlations (MICs), and pre-post mean differences have been reported. The internal consistency and the MICs provide insights into the homogeneity of items. In an online study ( N = 72), we investigated the test-retest reliability, Cronbach’s α, split-half reliability, MICs, and retest mean differences for the CFMT+ and the GFMT-S. The CFMT+ showed satisfactory reliability coefficients above .88, whereas the coefficients of the GFMT-S were mainly dissatisfactory and below .75. We argue that task characteristics like heterogeneous stimulus material might lower MICs, response behavior might enhance reliability, and practice effects might increase the means of the CFMT+ in repeated measurements. Therefore, an integrative evaluation of different psychometric parameters helps explaining variations of reliability in face recognition tests.

EXPRESS: Distinct abilities associated with matching same identity faces vs. discriminating different faces: Evidence from individual differences in prosopagnosics and controls

Previous face matching studies provide evidence that matching same identity faces (match trials) and discriminating different face identities (non-match trials) rely on distinct processes. For example, instructional studies geared towards improving face matching in applied settings have often found selective improvements in match or non-match trials only. Additionally, a small study found that developmental prosopagnosics (DPs) have specific deficits in making match but not non-match judgments. In the current study, we sought to replicate this finding in DPs and examine how individual differences across DPs and controls in match vs. non-match performance relate to featural vs. holistic processing abilities. 43 DPs and 27 controls matched face images shown from similar front views or with varied lighting or viewpoint. Participants also performed tasks measuring featural (eyes/mouth) and holistic processing (part-whole task). We found that DPs showed worse overall matching performance than controls and that their relative match vs. non-match deficit depended on image variation condition, indicating that DPs do not consistently show match- or non-match-specific deficits. When examining the association between holistic and featural processing abilities and match vs. non-match trials in the entire group of DPs and controls, we found a very clear dissociation: Match trials significantly correlated with eye processing ability (r=.48) but not holistic processing (r=.11), whereas non-match trials significantly correlated with holistic processing (r=.32) but not eye processing (r=.03). This suggests that matching same identity faces relies more on eye processing while discriminating different faces relies more on holistic processing.

Forensic facial examiners versus super-recognizers: Evaluating behavior beyond accuracy

We evaluated the detailed, behavioral properties of face matching performance in two specialist groups: forensic facial examiners and super-recognizers. Both groups compare faces to determine identity with high accuracy and outperform the general population. Typically, facial examiners are highly trained; super-recognizers rely on natural ability. We found distinct behaviors between these two groups. Examiners used the full 7-point identity judgment scale (-3: "different"; +3: "same"). Super-recognizers' judgments clustered toward highly confident decisions. Examiners' judgments for same- and different-identities were symmetric across the scale midpoint (0); super-recognizers' judgments were not. Examiners showed higher identity judgment agreement than super-recognizers. Despite these qualitative differences, both groups showed insight into their own accuracy: more confident people and those who rated the task to be easier tended to be more accurate. Altogether, we show to better understand and interpret judgments according to the nature of someone's facial expertise, evaluations should assess more than accuracy.

Identifying criminals: No biasing effect of criminal context on recalled threat

To date, it is still unclear whether there is a systematic pattern in the errors made in eyewitness recall and whether certain features of a person are more likely to lead to false identification. Moreover, we also do not know the extent of systematic errors impacting identification of a person from their body rather than solely their face. To address this, based on the contextual model of eyewitness identification (CMEI; Osborne & Davies, 2014, Applied Cognitive Psychology, 28[3], 392-402), we hypothesized that having framed a target as a perpetrator of a violent crime, participants would recall that target person as appearing more like a stereotypical criminal (i.e., more threatening). In three separate experiments, participants were first presented with either no frame, a neutral frame, or a criminal frame (perpetrators of a violent crime) accompanying a target (either a face or body). Participants were then asked to identify the original target from a selection of people that varied in facial threat or body musculature. Contrary to our hypotheses, we found no evidence of bias. However, identification accuracy was highest for the most threatening target bodies high in musculature, as well as bodies paired with detailed neutral contextual information. Overall, these findings suggest that while no systematic bias exists in the recall of criminal bodies, the nature of the body itself and the context in which it is presented can significantly impact identification accuracy.

When two fields collide: Identifying "super-recognisers" for neuropsychological and forensic face recognition research

In the last decade, a novel individual differences approach has emerged across the face recognition literature. While the field has long been concerned with prosopagnosia (the inability to recognise facial identity), it has more recently become clear that there are vast differences in face recognition ability within the typical population. "Super-recognisers" are those individuals purported to reside at the very top of this spectrum. On one hand, these people are of interest to cognitive neuropsychologists who are motivated to explore the commonality of the face recognition continuum, whereas on the other hand, researchers from the forensic face matching field evaluate the implementation of super-recognisers into real-world police and security settings. These two rather different approaches have led to discrepancies in the definition of super-recognisers, and perhaps more fundamentally, the approach to identifying them, resulting in a lack of consistency that prohibits theoretical progress. Here, we review the protocols used in published work to identify super-recognisers, and propose a common definition and screening recommendations that can be adhered to across fields.

Visual search performance in 'CCTV' and mobile phone-like video footage

Finding an unfamiliar person in a crowd of others is an integral task for police officers, CCTV-operators, and security staff who may be looking for a suspect or missing person; however, research suggests that it is difficult and accuracy in such tasks is low. In two real-world visual-search experiments, we examined whether being provided with four images versus one image of an unfamiliar target person would help improve accuracy when searching for that person through video footage. In Experiment 1, videos were taken from above and at a distance to simulate CCTV, and images of the target showed their face and torso. In Experiment 2, videos were taken from approximately shoulder height, such as one would expect from body-camera or mobile phone recordings, and target images included only the face. Our findings suggest that having four images as exemplars leads to higher accuracy in the visual search tasks, but this only reached significance in Experiment 2. There also appears to be a conservative bias whereby participants are more likely to respond that the target is not in the video when presented with only one image as opposed to 4. These results point to there being an advantage for providing multiple images of targets for use in video visual-search.

Low prevalence match and mismatch detection in simultaneous face matching: Influence of face recognition ability and feature focus guidance

Simultaneous face matching to verify identity is key to security and policing. However, matching is error-prone, particularly when target-item prevalence is low. Two experiments examined whether superior face recognition ability and the use of internal or external facial feature guidance scales would reduce low prevalence effects. In Experiment 1, super-recognisers (n = 317) significantly outperformed typical-ability controls (n = 452), while internal feature guidance enhanced accuracy across all prevalence conditions. However, an unexpected effect in controls revealed higher accuracy in low prevalence conditions, probably because no low-match or low-mismatch prevalence information was provided. In Experiment 2, top-end-of-typical range ability participants (n = 841) were informed of their low prevalence condition and demonstrated the expected low-prevalence effects. Findings and implications are discussed.

Face processing in police service: the relationship between laboratory-based assessment of face processing abilities and performance in a real-world identity matching task

In the present study, we investigated whether police officers’ performance in searching for unfamiliar faces in a video-based real-world task is predicted by laboratory-based face processing tests that are typically used to assess individual differences in face processing abilities. Specifically, perceptual performance in the field was operationalized via the identification of target individuals in self-made close-circuit television (CCTV) video tapes. Police officers’ abilities in the laboratory were measured by the Cambridge Face Memory Test long form (CFMT+). We hypothesized that the CFMT+ predicts individual differences in the CCTV task performance. A total of N = 186 police officers of the Rhineland-Palatinate State Police participated in the study (i.e., N = 139 novice and advanced cadets with either 3 months, 15 months or 24 months of pre-service experience; N = 47 experienced police officers with three years of pre-service experience and at least two years of full-service experience, who participated in the assessment center of the special police forces, specifically the surveillance and technical unit). Results revealed that the CFMT+ explained variance in the CCTV task. In sample 1, CFMT+ scores predicted hits, but not false alarms. In contrast, in sample 2, CFMT+ scores were correlated with both hits and false alarms. From a theoretical perspective, we discuss factors that might explain CCTV task performance. From a practical perspective, we recommend that personnel selection processes investigating individual differences of police officers’ face processing abilities should comprise of two steps. At first, laboratory-based tests of face processing abilities should be applied. Subsequently, to validate laboratory-based individual differences in face processing abilities, we recommend that work samples such as CCTV tasks from the field should be added.

Super-Recognizers - a novel diagnostic framework, 70 cases, and guidelines for future work

When you hear the word Super-Recognizer, you may think of comic-book-hero-esque agents searching the underground to find people who went missing decades ago. Compared to this fantasy, the reality seems somewhat less exciting. Super-Recognizers (SRs) were initially reported a decade ago as a collateral while developing tests for developmental prosopagnosia. Today, the topic of SRs sparks interest from groups seeking to enhance scientific knowledge, public safety, or their monetary gain. With no immediate consequences of erroneous SR identification, there has been no pressure to establish a clear SR definition. This promotes heterogenous empirical evidence and the proliferation of unsupported claims in the media. Not only is this status quo unfortunate, it stands in opposition to the potential of special populations - both for science and application. SRs are a special population with imminent real-world value that can advance our understanding of brain functioning. To exploit their potential, I propose a needed formal framework for SR diagnosis, and introduce 70 cases identified based hereupon. These cases represent the core of a growing SR cohort, studied in my lab in the course of a long-term, multi-methodological research agenda involving academic and government collaborators. Finally, I provide recommendations for those interested in SR work, and highlight current caveats and future challenges.

Developmental prosopagnosics and super recognizers rely on the same facial features used by individuals with normal face recognition abilities for face identification

Face recognition depends on the ability of the face processing system to extract facial features that define the identity of a face. In a recent study we discovered that altering a subset of facial features changed the identity of the face, indicating that they are critical for face identification. Changing another set of features did not change the identity of a face, indicating that they are not critical for face identification. In the current study, we assessed whether developmental prosopagnosics (DPs) and super recognizers (SRs) also rely more heavily on these critical features than non-critical features for face identification. To that end, we presented to DPs and SRs faces in which either the critical or the non-critical features were manipulated. In Study 1, we presented SRs with a famous face recognition task. We found that overall SRs recognized famous faces that differ in either critical or non-critical features better than controls. Similar to controls, changes in critical features had a larger effect on SRs' face recognition than changes in non-critical features. In Study 2, we presented an identity matching task to DPs and SRs. Similar to controls, DPs and SRs perceived faces that differed in critical features as more different than faces that differed in non-critical features. Taken together, our results indicate that SRs and DPs use the same critical features for face identification as normal individuals. These findings emphasize the fundamental role of this subset of features for face identification.

The part-whole effect in super-recognisers and typical-range-ability controls

Face recognition skills are distributed on a continuum, with developmental prosopagnosics and super-recognisers at the bottom and top ends, respectively. Holistic processing propensity is associated with face recognition ability and may be impaired in some developmental prosopagnosics and enhanced in some super-recognisers. Across two experiments we compared holistic processing of 75 super-recognisers and 89 typical-range ability controls using The Part-Whole Effect (PWE) paradigm. A subgroup of super-recognisers demonstrated enhanced PWEs in the nose region, suggesting they integrate the nose into the holistic face percept more effectively than controls. Focussed processing of the nose region, an optimal viewing position to extract the holistic properties of faces, has previously been associated with superior face recognition, and this may partly explain the superiority of some super-recognisers. However, a few super-recognisers generated extreme nose region performance patterns in an opposite direction across both experiments, suggesting their superiority is driven by alternative mechanisms. These results support proposals that super-recognition is associated with heterogeneous underlying processes.

Does Cross-Race Contact Improve Cross-Race Face Perception? A Meta-Analysis of the Cross-Race Deficit and Contact

Contact with racial outgroups is thought to reduce the cross-race recognition deficit (CRD), the tendency for people to recognize same-race (i.e., ingroup) faces more accurately than cross-race (i.e., outgroup) faces. In 2001, Meissner and Brigham conducted a meta-analysis in which they examined this question and found a meta-analytic effect of r = -.13. We conduct a new meta-analysis based on 20 years of additional data to update the estimate of this relationship and examine theoretical and methodological moderators of the effect. We find a meta-analytic effect of r = -.15. In line with theoretical predictions, we find some evidence that the magnitude of this relationship is stronger when contact occurs during childhood rather than adulthood. We find no evidence that the relationship differs for measures of holistic/configural processing compared with normal processing. Finally, we find that the magnitude of the relationship depends on the operationalization of contact and that it is strongest when contact is manipulated. We consider recommendations for further research on this topic.

Psychophysical profiles in super-recognizers

Facial identity matching ability varies widely, ranging from prosopagnosic individuals (who exhibit profound impairments in face cognition/processing) to so-called super-recognizers (SRs), possessing exceptional capacities. Yet, despite the often consequential nature of face matching decisions—such as identity verification in security critical settings—ability assessments tendentially rely on simple performance metrics on a handful of heterogeneously related subprocesses, or in some cases only a single measured subprocess. Unfortunately, methodologies of this ilk leave contributions of stimulus information to observed variations in ability largely un(der)specified. Moreover, they are inadequate for addressing the qualitative or quantitative nature of differences between SRs’ abilities and those of the general population. Here, therefore, we sought to investigate individual differences—among SRs identified using a novel conservative diagnostic framework, and neurotypical controls—by systematically varying retinal availability, bandwidth, and orientation of faces’ spatial frequency content in two face matching experiments. Psychophysical evaluations of these parameters’ contributions to ability reveal that SRs more consistently exploit the same spatial frequency information, rather than suggesting qualitatively different profiles between control observers and SRs. These findings stress the importance of optimizing procedures for SR identification, for example by including measures quantifying the consistency of individuals’ behavior.

The Oxford Face Matching Test: A non-biased test of the full range of individual differences in face perception

Tests of face processing are typically designed to identify individuals performing outside of the typical range; either prosopagnosic individuals who exhibit poor face processing ability, or super recognisers, who have superior face processing abilities. Here we describe the development of the Oxford Face Matching Test (OFMT), designed to identify individual differences in face processing across the full range of performance, from prosopagnosia, through the range of typical performance, to super recognisers. Such a test requires items of varying difficulty, but establishing difficulty is problematic when particular populations (e.g., prosopagnosics, individuals with autism spectrum disorder) may use atypical strategies to process faces. If item difficulty is calibrated on neurotypical individuals, then the test may be poorly calibrated for atypical groups, and vice versa. To obtain items of varying difficulty, we used facial recognition algorithms to obtain face pair similarity ratings that are not biased towards specific populations. These face pairs were used as stimuli in the OFMT, and participants were required to judge whether the face images depicted the same individual or different individuals. Across five studies the OFMT was shown to be sensitive to individual differences in the typical population, and in groups of both prosopagnosic individuals and super recognisers. The test-retest reliability of the task was at least equivalent to the Cambridge Face Memory Test and the Glasgow Face Matching Test. Furthermore, results reveal, at least at the group level, that both face perception and face memory are poor in those with prosopagnosia, and are good in super recognisers.

Performance of typical and superior face recognizers on a novel interactive face matching procedure

Unfamiliar simultaneous face matching is error prone. Reducing incorrect identification decisions will positively benefit forensic and security contexts. The absence of view-independent information in static images likely contributes to the difficulty of unfamiliar face matching. We tested whether a novel interactive viewing procedure that provides the user with 3D structural information as they rotate a facial image to different orientations would improve face matching accuracy. We tested the performance of 'typical' (Experiment 1) and 'superior' (Experiment 2) face recognizers, comparing their performance using high-quality (Experiment 3) and pixelated (Experiment 4) Facebook profile images. In each trial, participants responded whether two images featured the same person with one of these images being either a static face, a video providing orientation information, or an interactive image. Taken together, the results show that fluid orientation information and interactivity prompt shifts in criterion and support matching performance. Because typical and superior face recognizers both benefited from the structural information provided by the novel viewing procedures, our results point to qualitatively similar reliance on pictorial encoding in these groups. This also suggests that interactive viewing tools can be valuable in assisting face matching in high-performing practitioner groups.

The effect of face masks and sunglasses on identity and expression recognition with super-recognizers and typical observers

Face masks present a new challenge to face identification (here matching) and emotion recognition in Western cultures. Here, we present the results of three experiments that test the effect of masks, and also the effect of sunglasses (an occlusion that individuals tend to have more experienced with) on (i) familiar face matching, (ii) unfamiliar face matching and (iii) emotion categorization. Occlusion reduced accuracy in all three tasks, with most errors in the mask condition; however, there was little difference in performance for faces in masks compared with faces in sunglasses. Super-recognizers, people who are highly skilled at matching unconcealed faces, were impaired by occlusion, but at the group level, performed with higher accuracy than controls on all tasks. Results inform psychology theory with implications for everyday interactions, security and policing in a mask-wearing society.

Surgical face masks impair human face matching performance for familiar and unfamiliar faces

In response to the COVID-19 pandemic, many governments around the world now recommend, or require, that their citizens cover the lower half of their face in public. Consequently, many people now wear surgical face masks in public. We investigated whether surgical face masks affected the performance of human observers, and a state-of-the-art face recognition system, on tasks of perceptual face matching. Participants judged whether two simultaneously presented face photographs showed the same person or two different people. We superimposed images of surgical masks over the faces, creating three different mask conditions: control (no masks), mixed (one face wearing a mask), and masked (both faces wearing masks). We found that surgical face masks have a large detrimental effect on human face matching performance, and that the degree of impairment is the same regardless of whether one or both faces in each pair are masked. Surprisingly, this impairment is similar in size for both familiar and unfamiliar faces. When matching masked faces, human observers are biased to reject unfamiliar faces as "mismatches" and to accept familiar faces as "matches". Finally, the face recognition system showed very high classification accuracy for control and masked stimuli, even though it had not been trained to recognise masked faces. However, accuracy fell markedly when one face was masked and the other was not. Our findings demonstrate that surgical face masks impair the ability of humans, and naïve face recognition systems, to perform perceptual face matching tasks. Identification decisions for masked faces should be treated with caution.

UNSW Face Test: A screening tool for super-recognizers

We present a new test-the UNSW Face Test (www.unswfacetest.com)-that has been specifically designed to screen for super-recognizers in large online cohorts and is available free for scientific use. Super-recognizers are people that demonstrate sustained performance in the very top percentiles in tests of face identification ability. Because they represent a small proportion of the population, screening large online cohorts is an important step in their initial recruitment, before confirmatory testing via standardized measures and more detailed cognitive testing. We provide normative data on the UNSW Face Test from 3 cohorts tested via the internet (combined n = 23,902) and 2 cohorts tested in our lab (combined n = 182). The UNSW Face Test: (i) captures both identification memory and perceptual matching, as confirmed by correlations with existing tests of these abilities; (ii) captures face-specific perceptual and memorial abilities, as confirmed by non-significant correlations with non-face object processing tasks; (iii) enables researchers to apply stricter selection criteria than other available tests, which boosts the average accuracy of the individuals selected in subsequent testing. Together, these properties make the test uniquely suited to screening for super-recognizers in large online cohorts.

Differences between and within individuals, and subprocesses of face cognition: implications for theory, research and personnel selection

Recent investigations of individual differences have demonstrated striking variability in performance both within the same subprocess in face cognition (e.g. face perception), but also between two different subprocesses (i.e. face perception versus face recognition) that are assessed using different tasks (face matching versus face memory). Such differences between and within individuals between and within laboratory tests raise practical challenges. This applies in particular to the development of screening tests for the selection of personnel in real-world settings where faces are routinely processed, such as at passport control. The aim of this study, therefore, was to examine the performance profiles of individuals within and across two different subprocesses of face cognition: face perception and face recognition. To this end, 146 individuals completed four different tests of face matching-one novel tool for assessing proficiency in face perception, as well as three established measures-and two benchmark tests of face memory probing face recognition. In addition to correlational analyses, we further scrutinized individual performance profiles of the highest and lowest performing observers identified per test, as well as across all tests. Overall, a number of correlations emerged between tests. However, there was limited evidence at the individual level to suggest that high proficiency in one test generalized to other tests measuring the same subprocess, as well as those that measured a different subprocess. Beyond emphasizing the need to honour inter-individual differences through careful multivariate assessment in the laboratory, our findings have real-world implications: combinations of tests that most accurately map the task(s) and processes of interest are required for personnel selection.

Body Language Influences on Facial Identification at Passport Control: An Exploration in Virtual Reality

Person identification at airports requires the matching of a passport photograph to its bearer. One aim of this process is to find identity impostors, who use valid identity documents of similar-looking people to avoid detection. In psychology, this process has been studied extensively with static pairs of face photographs that require identity match (same person shown) versus mismatch (two different people) decisions. However, this approach provides a limited proxy for studying how other factors, such as nonverbal behaviour, affect this task. The current study investigated the influence of body language on facial identity matching within a virtual reality airport environment, by manipulating activity levels of person avatars queueing at passport control. In a series of six experiments, detection of identity mismatches was unaffected when observers were not instructed to utilise body language. By contrast, under explicit instruction to look out for unusual body language, these cues enhanced detection of mismatches but also increased false classification of matches. This effect was driven by increased activity levels rather than body language that simply differed from the behaviour of the majority of passengers. The implications and limitations of these findings are discussed.

Human-algorithm teaming in face recognition: How algorithm outcomes cognitively bias human decision-making

In face recognition applications, humans often team with algorithms, reviewing algorithm results to make an identity decision. However, few studies have explicitly measured how algorithms influence human face matching performance. One study that did examine this interaction found a concerning deterioration of human accuracy in the presence of algorithm errors. We conducted an experiment to examine how prior face identity decisions influence subsequent human judgements about face similarity. 376 volunteers were asked to rate the similarity of face pairs along a scale. Volunteers performing the task were told that they were reviewing identity decisions made by different sources, either a computer or human, or were told to make their own judgement without prior information. Replicating past results, we found that prior identity decisions, presented as labels, influenced volunteers' own identity judgements. We extend these results as follows. First, we show that the influence of identity decision labels was independent of indicated decision source (human or computer) despite volunteers' greater distrust of human identification ability. Second, applying a signal detection theory framework, we show that prior identity decision labels did not reduce volunteers' attention to the face pair. Discrimination performance was the same with and without the labels. Instead, prior identity decision labels altered volunteers' internal criterion used to judge a face pair as "matching" or "non-matching". This shifted volunteers' face pair similarity judgements by a full step along the response scale. Our work shows how human face matching is affected by prior identity decision labels and we discuss how this may limit the total accuracy of human-algorithm teams performing face matching tasks.

Hyper-realistic Face Masks in a Live Passport-Checking Task

Hyper-realistic face masks have been used as disguises in at least one border crossing and in numerous criminal cases. Experimental tests using these masks have shown that viewers accept them as real faces under a range of conditions. Here, we tested mask detection in a live identity verification task. Fifty-four visitors at the London Science Museum viewed a mask wearer at close range (2 m) as part of a mock passport check. They then answered a series of questions designed to assess mask detection, while the masked traveller was still in view. In the identity matching task, 8% of viewers accepted the mask as matching a real photo of someone else, and 82% accepted the match between masked person and masked photo. When asked if there was any reason to detain the traveller, only 13% of viewers mentioned a mask. A further 11% picked disguise from a list of suggested reasons. Even after reading about mask-related fraud, 10% of viewers judged that the traveller was not wearing a mask. Overall, mask detection was poor and was not predicted by unfamiliar face matching performance. We conclude that hyper-realistic face masks could go undetected during live identity checks.

Normative data for two challenging tests of face matching under ecological conditions

Background

Unfamiliar face processing is an ability that varies considerably between individuals. Numerous studies have aimed to identify its underlying determinants using controlled experimental procedures. While such tests can isolate variables that influence face processing, they usually involve somewhat unrealistic situations and optimized face images as stimulus material. As a consequence, the extent to which the performance observed under laboratory settings is informative for predicting real-life proficiency remains unclear.

Results

We present normative data for two ecologically valid but underused tests of face matching: the Yearbook Test (YBT) and the Facial Identity Card Sorting Test (FICST). The YBT (n = 252) measures identity matching across substantial age-related changes in facial appearance, while the FICST (n = 218) assesses the ability to process unfamiliar facial identity despite superficial image variations. To determine the predictive value of both tests, a subsample of our cohort (n = 181) also completed a commonly used test of face recognition and two tests of face perception (the long form of the Cambridge Face Memory Test (CFMT+), the Expertise in Facial Comparison Test (EFCT) and the Person Identification Challenge Test (PICT)).

Conclusions

Focusing on the top performers identified independently per test, we made two important observations: 1) YBT and FICST performance can predict CFMT+ scores and vice versa; and 2) EFCT and PICT scores neither reliably predict superior performance in ecologically meaningful and challenging tests of face matching, nor in the most commonly used test of face recognition. These findings emphasize the necessity for using challenging and ecologically relevant, and thus highly sensitive, tasks of unfamiliar face processing to identify high-performing individuals in the normal population.

The wisdom of the crowd: A case of post- to ante-mortem face matching by police super-recognisers

This case report describes novel methodology used to identify a 43-year-old post-mortem photo of a drowned male recovered from a London river in the 1970s. Embedded in an array of foils, police super-recognisers (n=25) possessing superior simultaneous face matching ability, and police controls (n=139) provided confidence ratings as to the similarity of the post-mortem photo to an ante-mortem photo of a man who went missing at about the same time. Indicative of a match, compared to controls, super-recognisers provided higher ratings to the target than the foils. Effects were enhanced when drawing on the combined wisdom of super-recogniser crowds, but not control crowds. These findings supported additional case evidence allowing the coroner to rule that the deceased male and missing male were likely one and the same person. A description of how similar super-recogniser wisdom of the crowd procedures could be applied to other visual image identification cases when no other method is feasible is provided.

A grey area: how does image hue affect unfamiliar face matching?

The role of image colour in face identification has received little attention in research despite the importance of identifying people from photographs in identity documents (IDs). Here, in two experiments, we investigated whether colour congruency of two photographs, shown side by side, affects face-matching accuracy. Participants were presented with two images from the Models Face Matching Test (experiment 1) and a newly devised matching task incorporating female faces (experiment 2) and asked to decide whether they show the same person or two different people. The photographs were either both in colour, both in grayscale, or mixed (one in grayscale and one in colour). Participants were more likely to accept a pair of images as a "match", i.e. same person, in the mixed condition, regardless of whether the identity of the pair was the same or not. This demonstrates a clear shift in bias between "congruent" colour conditions and the mixed trials. In addition, there was a small decline in accuracy in the mixed condition, relative to when the images were presented in colour. Our study provides the first evidence that the hue of document photographs matters for face-matching performance. This finding has important implications for the design and regulation of photographic ID worldwide.

Facial Identification at a Virtual Reality Airport

Person identification at airports requires the comparison of a passport photograph with its bearer. In psychology, this process is typically studied with static pairs of face photographs that require identity-match (same person shown) versus mismatch (two different people) decisions, but this approach provides a limited proxy for studying how environment and social interaction factors affect this task. In this study, we explore the feasibility of virtual reality (VR) as a solution to this problem, by examining the identity matching of avatars in a VR airport. We show that facial photographs of real people can be rendered into VR avatars in a manner that preserves image and identity information (Experiments 1 to 3). We then show that identity matching of avatar pairs reflects similar cognitive processes to the matching of face photographs (Experiments 4 and 5). This pattern holds when avatar matching is assessed in a VR airport (Experiments 6 and 7). These findings demonstrate the feasibility of VR as a new method for investigating face matching in complex environments.

Super-recognizers: From the lab to the world and back again

The recent discovery of individuals with superior face processing ability has sparked considerable interest amongst cognitive scientists and practitioners alike. These ‘Super‐recognizers’ (SRs) offer clues to the underlying processes responsible for high levels of face processing ability. It has been claimed that they can help make societies safer and fairer by improving accuracy of facial identity processing in real‐world tasks, for example when identifying suspects from Closed Circuit Television or performing security‐critical identity verification tasks. Here, we argue that the current understanding of superior face processing does not justify widespread interest in SR deployment: There are relatively few studies of SRs and no evidence that high accuracy on laboratory‐based tests translates directly to operational deployment. Using simulated data, we show that modest accuracy benefits can be expected from deploying SRs on the basis of ideally calibrated laboratory tests. Attaining more substantial benefits will require greater levels of communication and collaboration between psychologists and practitioners. We propose that translational and reverse‐translational approaches to knowledge development are critical to advance current understanding and to enable optimal deployment of SRs in society. Finally, we outline knowledge gaps that this approach can help address.