Face averages and multiple images in a live matching task

We know from previous research that unfamiliar face matching (determining whether two simultaneously presented images show the same person or not) is very error-prone. A small number of studies in laboratory settings have shown that the use of multiple images or a face average, rather than a single image, can improve face matching performance. Here, we tested 1,999 participants using four-image arrays and face averages in two separate live matching tasks. Matching a single image to a live person resulted in numerous errors (79.9% accuracy across both experiments), and neither multiple images (82.4% accuracy) nor face averages (76.9% accuracy) improved performance. These results are important when considering possible alterations which could be made to photo-ID. Although multiple images and face averages have produced measurable improvements in performance in recent laboratory studies, they do not produce benefits in a real-world live face matching context.     

Unfamiliar face matching: Pairs out‐perform individuals and provide a route to training

Matching unfamiliar faces is known to be difficult. Here, we ask whether performance can be improved by asking viewers to work in pairs, a manipulation known to increase accuracy for low‐level visual discrimination tasks. Across four experiments we consistently find that face matching accuracy is higher for pairs of viewers than for individuals. This ‘pairs advantage’ is generally driven by adopting the response of the higher scoring partner. However, when the task becomes difficult, both partners' performance is improved by working in a pair. In two experiments, we find evidence that working in a pair can lead to subsequent improvements in individual performance, specifically for viewers whose accuracy is initially low. The pairs' technique therefore offers the opportunity for substantial improvements in face matching performance, along with an added training benefit.     

The face you recognize may not be the one you saw: memory conjunction errors in individuals with or without learning disability

Memory conjunction errors, that is, when a combination of two previously presented stimuli is erroneously recognized as previously having been seen, were investigated in a face recognition task with drawings and photographs in 23 individuals with learning disability, and 18 chronologically age-matched controls without learning disability. Compared to the controls, individuals with learning disability committed significantly more conjunction errors, feature errors (one old and one new component), but had lower correct recognition, when the results were adjusted for different guessing levels. A dual-processing approach gained more support than a binding approach. However, neither of the approaches could explain all of the results. The results of the learning disability group were only partly related to non-verbal intelligence.     

Identity specific adaptation with composite faces

A composite face, made from the top half of a celebrity face and the bottom half of an unfamiliar face, appears to be a single, “new” face (e.g., Young, Hellawell, & Hay, 1987). Composite faces were used within the face identity aftereffect (FIAE) paradigm, in which prolonged exposure to a face reduces sensitivity to it (adaptation). Adaptation occurred both with an intact face and with composites containing its upper half, but only when composites were explicitly recognized during the adaptation phase. Unrecognized composites produced no adaptation. These findings imply that the FIAE is a relatively high-level perceptual effect, given that identical stimuli either did or did not produce adaptation depending on whether or not they were recognized. They also suggest a perceptual locus for the “composite face effect”.     

Telling faces together: Learning new faces through exposure to multiple instances

We are usually able to recognize novel instances of familiar faces with little difficulty, yet recognition of unfamiliar faces can be dramatically impaired by natural within-person variability in appearance. In a card-sorting task for facial identity, different photos of the same unfamiliar face are often seen as different people. Here we report two card-sorting experiments in which we manipulate whether participants know the number of identities present. Without constraints, participants sort faces into many identities. However, when told the number of identities present, they are highly accurate. This minimal contextual information appears to support viewers in “telling faces together”. In Experiment 2 we show that exposure to within-person variability in the sorting task improves performance in a subsequent face-matching task. This appears to offer a fast route to learning generalizable representations of new faces.     

Face Recognition Depends on Specialized Mechanisms Tuned to View-Invariant Facial Features: Insights from Deep Neural Networks Optimized for Face or Object Recognition

Face recognition is a computationally challenging classification task. Deep convolutional neural networks (DCNNs) are brain-inspired algorithms that have recently reached human-level performance in face and object recognition. However, it is not clear to what extent DCNNs generate a human-like representation of face identity. We have recently revealed a subset of facial features that are used by humans for face recognition. This enables us now to ask whether DCNNs rely on the same facial information and whether this human-like representation depends on a system that is optimized for face identification. In the current study, we examined the representation of DCNNs of faces that differ in features that are critical or non-critical for human face recognition. Our findings show that DCNNs optimized for face identification are tuned to the same facial features used by humans for face recognition. Sensitivity to these features was highly correlated with performance of the DCNN on a benchmark face recognition task. Moreover, sensitivity to these features and a view-invariant face representation emerged at higher layers of a DCNN optimized for face recognition but not for object recognition. This finding parallels the division to a face and an object system in high-level visual cortex. Taken together, these findings validate human perceptual models of face recognition, enable us to use DCNNs to test predictions about human face and object recognition as well as contribute to the interpretability of DCNNs.     

When intuition fails to align with data: A reply to Rossion (2013)

Holistic processing, a hallmark of face perception, is often measured in the so-called composite paradigm, in which participants are asked to match part of a stimulus while ignoring another part. In prior work, we recommended against the use of one version of the composite task we call the partial design, on the basis of confounds with response biases. Rossion wrote a lengthy piece that reviews the work that he has published using this design, raising a large number of criticisms, both about an alternative measure of holistic processing that we have used and advocated (which we call the complete design) and about our work in general. In this reply, we have limited our discussion to those issues that would be relevant to a researcher looking to decide which version of this composite paradigm to use, as we doubt a comprehensive reply would be of significant interest outside a very small circle.     

Face gender recognition in developmental prosopagnosia: Evidence for holistic processing and use of configural information

Face identification deficits in developmental prosopagnosics (DPs) have been thought to be due to general difficulties with processing configural face information and integrating configural and parts information into a coherent whole (holistic processing). Gender recognition provides a further opportunity to more fully examine this issue as this ability may be intact in DPs and it has been shown to depend on processing configural information and holistic processing in neurotypical individuals. In the present study we first determined that, indeed, gender discrimination performance was similar in DPs and controls. Second, we found that inversion and scrambling (which we propose measures holistic processing and sensitivity to configural information, respectively) produced comparable deficits in DPs and controls, suggesting that both groups use holistic processing and configural information to recognize gender. This indicates that holistic processing and using configural face information are not general impairments in DP and may be more specific to face identity.     

Why 8-year-olds cannot tell the difference between Steve Martin and Paul Newman: factors contributing to the slow development of sensitivity to the spacing of facial features

Children are nearly as sensitive as adults to some cues to facial identity (e.g., differences in the shape of internal features and the external contour), but children are much less sensitive to small differences in the spacing of facial features. To identify factors that contribute to this pattern, we compared 8-year-olds' sensitivity to spacing cues with that of adults under a variety of conditions. In the first two experiments, participants made same/different judgments about faces differing only in the spacing of facial features, with the variations being kept within natural limits. To measure the effect of attention, we reduced the salience of featural information by blurring faces and occluding features (Experiment 1). To measure the role of encoding speed and memory limitations, we presented pairs of faces simultaneously and for an unlimited time (Experiment 2). To determine whether participants' sensitivity would increase when spacing distortions were so extreme as to make the faces grotesque, we manipulated the spacing of features beyond normal limits and asked participants to rate each face on a "bizarreness" scale (Experiment 3). The results from the three experiments indicate that low salience, poor encoding efficiency, and limited memory can partially account for 8-year-olds' poor performance on face processing tasks that require sensitivity to the spacing of features, a kind of configural processing that underlies adults' expertise. However, even when the task is modified to compensate for these problems, children remain less sensitive than adults to the spacing of features.     

Incorporating multiple secondary targets into learning trials for individuals with autism spectrum disorder

The current study examined the outcome of presenting multiple secondary targets in learning trials for individuals with autism spectrum disorder. We compared conditions in which (a) a secondary target was presented in the antecedent and consequence of trials, (b) two secondary targets were presented in the consequence of trials, (c) one secondary target was presented in the consequence of each trial, and (d) no additional targets were presented trials. The participants acquired the majority of secondary targets. Presenting one or multiple secondary targets per trial, regardless of the location of these secondary targets, increased the efficiency of instruction in comparison to a condition with no secondary target.     

CAMERA: A system for fast and reliable acquisition of multiple ethological records

CAMERA is a system for collecting and editing ethological data. The hardware and software enable easy coding of complex behavioral interactions from video recordings. It is especially designed to improve accuracy, reliability, and training standards in coding behavior. The system also supports several basic analyses of collected data. CAMERA operates with both North American and European video signals.     

Face encoding is not categorical: Consistent evidence across multiple types of contingent aftereffects

ABSTRACT

Race-contingent aftereffects indicate that faces of different races are encoded via dissociable visual channels. Race-contingent aftereffects reflect perceptions of typicality, indicating a gradual transition of activity across channels as faces change from looking more typical of one race to the other We examine whether faces forming more discrete categories (sex: male/female; species: human/monkey) are encoded in a similar fashion, or whether they are instead encoded via more dichotomous categorical judgements. Curves representing the rate of change of aftereffect magnitude as stimuli changed from male to female produced shallow slopes, closely resembling ratings of typicality, but differing significantly from slopes for categorical judgements. For species, aftereffect slopes were significantly shallower than both ratings of typicality and categorical judgements. Overall, these results suggest that the encoding of facial properties such as these is not categorical, but instead involves a graded response as stimulus typicality varies. Aftereffect slopes were similar across the two experiments, raising the possibility of a common system that is recruited during contingent adaptation, regardless of the specific dimension involved or the categories used by the experimenter.     

Finding an unfamiliar face in a line-up: Viewing multiple images of the target is beneficial on target-present trials but costly on target-absent trials

When viewing unfamiliar faces, photographs of the same person often are perceived as belonging to different people and photographs of different people as belonging to the same person. Identity matching of unfamiliar faces is especially challenging when the photographs are of a person whose ethnicity differs from that of the observer. In contrast, matching is trivial when viewing familiar faces, regardless of race. Viewing multiple images of an own-race target identity improves accuracy on a line-up task when the target is known to be present (Dowsett et al., 2016, Q J Exp Psychol, 69, 1), suggesting that exposure to within-person variability in appearance is key to face learning. Across three experiments, we show that viewing multiple images of a target identity also improves accuracy for other-race faces on target-present trials. However, viewing multiple images decreases accuracy (i.e., increases false alarms) on target-absent trials for both own- and other-race faces. We discuss the implications of our findings for models of face recognition and for forensic settings.     

Walk this way: approaching bodies can influence the processing of faces

A highly familiar type of movement occurs whenever a person walks towards you. In the present study, we investigated whether this type of motion has an effect on face processing. We took a range of different 3D head models and placed them on a single, identical 3D body model. The resulting figures were animated to approach the observer. In a first series of experiments, we used a sequential matching task to investigate how the motion of an approaching person affects immediate responses to faces. We compared observers’ responses following approach sequences to their performance with figures walking backwards (receding motion) or remaining still. Observers were significantly faster in responding to a target face that followed an approach sequence, compared to both receding and static primes. In a second series of experiments, we investigated long-term effects of motion using a delayed visual search paradigm. After studying moving or static avatars, observers searched for target faces in static arrays of varying set sizes. Again, observers were faster at responding to faces that had been learned in the context of an approach sequence. Together these results suggest that the context of a moving body influences face processing, and support the hypothesis that our visual system has mechanisms that aid the encoding of behaviourally-relevant and familiar dynamic events.     

Identifying facial emotions: valence specific effects and an exploration of the effects of viewer gender

The valence hypothesis suggests that the right hemisphere is specialised for negative emotions and the left hemisphere is specialised for positive emotions (Silberman & Weingartner, 1986). It is unclear to what extent valence-specific effects in facial emotion perception depend upon the gender of the perceiver. To explore this question 46 participants completed a free view lateralised emotion perception task which involved judging which of two faces expressed a particular emotion. Eye fixations of 24 of the participants were recorded using an eye tracker. A significant valence-specific laterality effect was obtained, with positive emotions more accurately identified when presented to the right of centre, and negative emotions more accurately identified when presented to the left of centre. The valence-specific laterality effect did not depend on the gender of the perceiver. Analysis of the eye tracking data showed that males made more fixations while recognising the emotions and that the left-eye was fixated substantially more than the right-eye during emotion perception. Finally, in a control condition where both faces were identical, but expressed a faint emotion, the participants were significantly more likely to select the right side when the emotion label was positive. This finding adds to evidence suggesting that valence effects in facial emotion perception are not only caused by the perception of the emotion but by other processes.     

The effect of inversion on 3- to 5-year-olds’ recognition of face and nonface visual objects

This study compared the effect of stimulus inversion on 3- to 5-year-olds’ recognition of faces and two nonface object categories matched with faces for a number of attributes: shoes (Experiment 1) and frontal images of cars (Experiments 2 and 3). The inversion effect was present for faces but not shoes at 3 years of age (Experiment 1). Analogous results were found for boys when faces were compared with frontal images of cars. For girls, stimulus inversion impaired recognition of both faces and cars at 3 to 4 years of age, becoming specific to faces only at 5 years of age (Experiments 2 and 3). Evidence demonstrates that the ability to extract the critical cues that lead to adults’ efficient face recognition is selectively tuned to faces during preschool years.     

Some effects of alcohol and eye movements on cross-race face learning

This study examines the impact of acute alcohol intoxication on visual scanning in cross-race face learning. The eye movements of a group of white British participants were recorded as they encoded a series of own-and different-race faces, under alcohol and placebo conditions. Intoxication reduced the rate and extent of visual scanning during face encoding, reorienting the focus of foveal attention away from the eyes and towards the nose. Differences in encoding eye movements also varied between own-and different-race face conditions as a function of alcohol. Fixations to both face types were less frequent and more lingering following intoxication, but in the placebo condition this was only the case for different-race faces. While reducing visual scanning, however, alcohol had no adverse effect on memory, only encoding restrictions associated with sober different-race face processing led to poorer recognition. These results support perceptual expertise accounts of own-race face processing, but suggest the adverse effects of alcohol on face learning published previously are not caused by foveal encoding restrictions. The implications of these findings for alcohol myopia theory are discussed.