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 learning with multiple images leads to fast acquisition of familiarity for specific individuals

Matching unfamiliar faces is a difficult task. Here we ask whether it is possible to improve performance by providing multiple images to support matching. In two experiments we observe that accuracy improves as viewers are provided with additional images on which to base their match. This technique leads to fast learning of an individual, but the effect is identity-specific: Despite large improvements in viewers’ ability to match a particular person's face, these improvements do not generalize to other faces. Experiment 2 demonstrated that trial-by-trial feedback provided no additional benefits over the provision of multiple images. We discuss these results in terms of familiar and unfamiliar face processing and draw out some implications for training regimes.     

Identity From Variation: Representations of Faces Derived From Multiple Instances

Research in face recognition has tended to focus on discriminating between individuals, or “telling people apart.” It has recently become clear that it is also necessary to understand how images of the same person can vary, or “telling people together.” Learning a new face, and tracking its representation as it changes from unfamiliar to familiar, involves an abstraction of the variability in different images of that person's face. Here, we present an application of principal components analysis computed across different photos of the same person. We demonstrate that people vary in systematic ways, and that this variability is idiosyncratic—the dimensions of variability in one face do not generalize well to another. Learning a new face therefore entails learning how that face varies. We present evidence for this proposal and suggest that it provides an explanation for various effects in face recognition. We conclude by making a number of testable predictions derived from this framework.     

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.     

Not looking yourself: The cost of self‐selecting photographs for identity verification

Photo‐identification is based on the premise that photographs are representative of facial appearance. However, previous studies show that ratings of likeness vary across different photographs of the same face, suggesting that some images capture identity better than others. Two experiments were designed to examine the relationship between likeness judgments and face matching accuracy. In Experiment 1, we compared unfamiliar face matching accuracy for self‐selected and other‐selected high‐likeness images. Surprisingly, images selected by previously unfamiliar viewers – after very limited exposure to a target face – were more accurately matched than self‐selected images chosen by the target identity themselves. Results also revealed extremely low inter‐rater agreement in ratings of likeness across participants, suggesting that perceptions of image resemblance are inherently unstable. In Experiment 2, we test whether the cost of self‐selection can be explained by this general disagreement in likeness judgments between individual raters. We find that averaging across rankings by multiple raters produces image selections that provide superior identification accuracy. However, benefit of other‐selection persisted for single raters, suggesting that inaccurate representations of self interfere with our ability to judge which images faithfully represent our current appearance.     

Ensemble coding of facial identity is not refined by experience: Evidence from other-race and inverted faces

The ability to recognize identity despite within-person variability in appearance is likely a face-specific skill and shaped by experience. Ensemble coding – the automatic extraction of the average of a stimulus array – has been proposed as a mechanism underlying face learning (allowing one to recognize novel instances of a newly learned face). We investigated whether ensemble encoding, like face learning and recognition, is refined by experience by testing participants with upright own-race faces and two categories of faces with which they lacked experience: other-race faces (Experiment 1) and inverted faces (Experiment 2). Participants viewed four images of an unfamiliar identity and then were asked whether a test image of that same identity had been in the study array. Each test image was a matching exemplar (from the array), matching average (the average of the images in the array), non-matching exemplar (a novel image of the same identity), or non-matching average (an average of four different images of the same identity). Adults showed comparable ensemble coding for all three categories (i.e., reported that matching averages had been present more than non-matching averages), providing evidence that this early stage of face learning is not shaped by face-specific experience.     

Robust representations of individual faces in chimpanzees (<Emphasis Type="Italic">Pan troglodytes</Emphasis>) but not monkeys (<Emphasis Type="Italic">Macaca mulatta</Emphasis>)

Being able to recognize the faces of our friends and family members no matter where we see them represents a substantial challenge for the visual system because the retinal image of a face can be degraded by both changes in the person (age, expression, pose, hairstyle, etc.) and changes in the viewing conditions (direction and degree of illumination). Yet most of us are able to recognize familiar people effortlessly. A popular theory for how face recognition is achieved has argued that the brain stabilizes facial appearance by building average representations that enhance diagnostic features that reliably vary between people while diluting features that vary between instances of the same person. This explains why people find it easier to recognize average images of people, created by averaging multiple images of the same person together, than single instances (i.e. photographs). Although this theory is gathering momentum in the psychological and computer sciences, there is no evidence of whether this mechanism represents a unique specialization for individual recognition in humans. Here we tested two species, chimpanzees (Pan troglodytes) and rhesus monkeys (Macaca mulatta), to determine whether average images of different familiar individuals were easier to discriminate than photographs of familiar individuals. Using a two-alternative forced-choice, match-to-sample procedure, we report a behaviour response profile that suggests chimpanzees encode the faces of conspecifics differently than rhesus monkeys and in a manner similar to humans.     

Face detection dissociates from face identification

We describe three experiments in which viewers complete face detection tasks as well as standard measures of unfamiliar face identification. In the first two studies, participants viewed pareidolic images of objects (Experiment 1) or cloud scenes (Experiment 2), and their propensity to see faces in these scenes was measured. In neither case is performance significantly associated with identification, as measured by the Cambridge Face Memory or Glasgow Face Matching Tests. In Experiment 3 we showed participants real faces in cluttered scenes. Viewers’ ability to detect these faces is unrelated to their identification performance. We conclude that face detection dissociates from face identification.     

Recognizing disguised faces

Across three experiments, we evaluated the effects of “disguises” on observers’ face identification performance using naturalistic images in which individuals posed with a variety of wigs and eyeglasses. Experiment 1 tested recognition memory performance with and without disguises and showed that any changes in these facial attributes hindered performance, replicating previous findings. More interestingly, Experiment 1 revealed that a change in hairstyle or the removal of eyeglasses had more impact on performance than did the addition of eyeglasses. In Experiment 2, disguised and undisguised faces were presented upright or inverted to examine whether the performance decrements seen in Experiment 1 were attributable to disruption of processing strategies for faces. Despite showing an overall effect of disguise on performance, there was no interaction between inversion and disguise. This suggests that disguises are not directly affecting processing strategies, but instead disguises are likely encoded as part of the overall representation. Similarly, in Experiment 3, the composite face paradigm was used to examine whether the impact of disguises is attributable to disruption of holistic processing for faces. Again, we found an overall effect of disguise on performance, but here we also observed that the presence of disguises interacted with participants’ tendency to form holistic face representations.     

The sibling familiarity effect: Is within-person facial variability shared across siblings?

Humans are experts at familiar face recognition, but poor at unfamiliar face recognition. Familiarity is created when a face is encountered across varied conditions, but the way in which a person’s appearance varies is identity-specific, so familiarity with one identity does not benefit recognition of other individuals. However, the faces of biological siblings share structural similarities, so we explored whether the benefits of familiarity are shared across siblings. Results show that familiarity with one half of a sibling pair improves kin detection (experiment 1), and that unfamiliar face matching is more accurate when targets are the siblings of familiar versus unfamiliar individuals (experiment 2). PCA applied to facial images of celebrities and their siblings demonstrates that faces are generally better reconstructed in the principal components of a same-sex sibling than those of an unrelated individual. When we encounter the unfamiliar sibling of someone we already know, our pre-existing representation of their familiar relation may usefully inform processing of the unfamiliar face. This can benefit both kin detection and identity processing, but the benefits are constrained by the degree to which facial variability is shared.     

Newborns' face recognition is based on spatial frequencies below 0.5 cycles per degree

A critical question in Cognitive Science concerns how knowledge of specific domains emerges during development. Here we examined how limitations of the visual system during the first days of life may shape subsequent development of face processing abilities. By manipulating the bands of spatial frequencies of face images, we investigated what is the nature of the visual information that newborn infants rely on to perform face recognition. Newborns were able to extract from a face the visual information lying from 0 to 1 cpd (Experiment 1), but only a narrower 0-0.5 cpd spatial frequency range was successful to accomplish face recognition (Experiment 2). These results provide the first empirical support of a low spatial frequency advantage in individual face recognition at birth and suggest that early in life low-level, non-specific perceptual constraints affect the development of the face processing system.     

Flexible voices: Identity perception from variable vocal signals

Human voices are extremely variable: The same person can sound very different depending on whether they are speaking, laughing, shouting or whispering. In order to successfully recognise someone from their voice, a listener needs to be able to generalize across these different vocal signals (‘telling people together’). However, in most studies of voice-identity processing to date, the substantial within-person variability has been eliminated through the use of highly controlled stimuli, thus focussing on how we tell people apart. We argue that this obscures our understanding of voice-identity processing by controlling away an essential feature of vocal stimuli that may include diagnostic information. In this paper, we propose that we need to extend the focus of voice-identity research to account for both “telling people together” as well as “telling people apart.” That is, we must account for whether, and to what extent, listeners can overcome within-person variability to obtain a stable percept of person identity from vocal cues. To do this, our theoretical and methodological frameworks need to be adjusted to explicitly include the study of within-person variability.

     

7～10岁自闭症谱系障碍儿童对情绪面孔的觉察与加工

为了考察自闭症谱系障碍(autism spectrum disorder,ASD)儿童对不同情绪面孔的觉察和加工情况,设计两个眼动实验任务,要求14名7～10岁ASD儿童和20名同年龄正常儿童观看图片。实验一采用将情绪面孔嵌入风景图片中引起语义不一致的刺激;实验二采用含有情绪面孔的无意义背景乱序图片刺激。结果发现:(1)ASD儿童对不同情绪面孔的觉察时间都显著长于正常儿童;(2)与正常儿童一样,ASD儿童表现出对恐惧面孔的注意偏向;(3)实验二中,ASD儿童对不同情绪面孔内部特征区的注意分配与正常儿童不同;正常儿童能注意最能展示该类情绪特征信息的区域,如恐惧的眼睛、愉快的嘴巴,而ASD儿童对三类情绪面孔特征区的注意分配方式相似;(4)两个实验条件下,ASD者对不同情绪面孔的觉察、加工模式与正常儿童相似。     

Event-related potentials reveal the development of stable face representations from natural variability

Natural variability between instances of unfamiliar faces can make it difficult to reconcile two images as the same person. Yet for familiar faces, effortless recognition occurs even with considerable variability between images. To explore how stable face representations develop, we employed incidental learning in the form of a face sorting task. In each trial, multiple images of two facial identities were sorted into two corresponding piles. Following the sort, participants showed evidence of having learnt the faces performing more accurately on a matching task with seen than with unseen identities. Furthermore, ventral temporal event-related potentials were more negative in the N250 time range for previously seen than for previously unseen identities. These effects appear to demonstrate some degree of abstraction, rather than simple picture learning, as the neurophysiological and behavioural effects were observed with novel images of the previously seen identities. The results provide evidence of the development of facial representations, allowing a window onto natural mechanisms of face learning.     

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.     

Face processing in adolescents with autistic disorder: the inversion and composite effects

Two experiments with upright and inverted face and object images were carried out to investigate whether face processing in autism is more feature-based than in individuals with typical development. Participants were 17 high-ability adolescents with autistic disorder (16-24 years), 24 typically developing children (9-10 years) and 16 adults (18-33 years). In Experiment 1, a normal inversion effect was found for the adolescents with autism in a standard face recognition paradigm with reduced memory demands, except for a subgroup with low social intelligence who were not better in recognizing upright relative to inverted photographs of faces. In Experiment 2, the group with autism did not show the composite effect like the adult group did: they recognized face halves as well in aligned composite faces as in non-aligned composite faces. The results on the inversion task suggest that most adolescents with autism form a normal configuration-based face representation, but the absence of the composite effect indicates that they are less prone to use the contextual information of the face in a visual-search task.     

Tolerance for distorted faces: Challenges to a configural processing account of familiar face recognition

Face recognition is widely held to rely on ‘configural processing’, an analysis of spatial relations between facial features. We present three experiments in which viewers were shown distorted faces, and asked to resize these to their correct shape. Based on configural theories appealing to metric distances between features, we reason that this should be an easier task for familiar than unfamiliar faces (whose subtle arrangements of features are unknown). In fact, participants were inaccurate at this task, making between 8% and 13% errors across experiments. Importantly, we observed no advantage for familiar faces: in one experiment participants were more accurate with unfamiliars, and in two experiments there was no difference. These findings were not due to general task difficulty – participants were able to resize blocks of colour to target shapes (squares) more accurately. We also found an advantage of familiarity for resizing other stimuli (brand logos). If configural processing does underlie face recognition, these results place constraints on the definition of ‘configural’. Alternatively, familiar face recognition might rely on more complex criteria – based on tolerance to within-person variation rather than highly specific measurement.     

Individual differences in face and voice matching abilities: The relationship between accuracy and consistency

Deciding whether two different face photographs or voice samples are from the same person represent fundamental challenges within applied settings. To date, most research has focussed on average performance in these tests, failing to consider individual differences and within‐person consistency in responses. Here, participants completed the same face (Experiment 1) or voice matching test (Experiment 2) on two separate occasions, allowing comparison of overall accuracy across the two timepoints as well as consistency in trial‐level responses. In both experiments, participants were highly consistent in their performances. In addition, we demonstrated a large association between consistency and accuracy, with the most accurate participants also tending to be the most consistent. This is an important result for applied settings in which organisational groups of super‐matchers are deployed in real‐world contexts. Being able to reliably identify these high performers based upon only a single test informs regarding recruitment for law enforcement agencies worldwide.     

On the costs of accessible attitudes: detecting that the attitude object has changed

The present research examined whether individuals with more accessible attitudes have more difficulty detecting that the attitude object has changed. While being repeatedly exposed to photographs of undergraduates, participants either rehearsed their attitudes toward each photo or performed a control task. They then saw these original photos and computer-generated morphs representing varying degrees of change in an original. Participants in the attitude rehearsal condition required more time to correctly identify morphs that were similar to the original as "different" (Experiment 1) and made more errors in response to such morphs (Experiment 2). Experiment 3 revealed that participants with accessible attitudes perceived relatively less change; they were less likely to view a morph as a photo of a novel person and more likely to view it as a different photo of a person seen before. The costs and benefits of accessible attitudes are discussed.