基于动态特征的真伪笑容表达与识别

笑容是人类最普遍、最频繁的表情。人类进化出伪装笑容的能力,也拥有部分识别伪装的能力。在表情的表达与识别上,动态信息起着重要的作用。一方面,笑容表达的动态特征可能为区分真伪笑容提供重要的信息,所以我们拟借助近年发展的计算机视觉的特征提取技术,系统地量化分析真伪笑容的动态特征(时长、方向、速度、流畅性、运动对称性、不同部位同步性、头动模式等),考察笑容在不同伪装方式及不同情境下的区别与一致性,深入理解人类笑容表达的特点。另一方面,通过探索有效动态特征与正确识别率的关系,检验知觉-注意假说,了解真伪笑容的识别特点及研究识别机制。通过比较动态真伪笑容的表达特点与识别特点,进一步理解人类表情信号编码与解码之间的关系。     

Judgement of Authenticity of Smiles and Detection of Facial indexes]

The smile is one of the most often expressed emotions during social interactions. It can be authentic, that is, associated with a joyful emotional state in the person expressing it, but it can also be false, that is, deliberately produced in the absence of that emotional state in order to deceive one or more individuals (Ekman, 1993). Even though the fake smile very much resembles the authentic smile, it generally does not constitute the perfect simile. The fake smile more often has a certain degree of asymmetry than the authentic smile (Ekman, Hager, & Friesen, 1981) and it uses the cheek raiser action less often than with the authentic smile (Ekman, Friesen, & O'Sullivan, 1988; Frank, Ekman, & Friesen, 1993). This study looked at the knowledge that adults have of these differences as well as their perceptive ability to detect them. The visual stimuli presented to participants were prepared using the Facial Action Coding System (Ekman & Friesen, 1978). Results show that participants detected the differences between the two types of smile and that detection was better using smile asymmetry than with the cheek raiser action. Analysis of the use of response categories in the detection task indicated that participants underestimated the differences between smiles when they were different and that this tendency was more apparent with the cheek raiser detection method than for asymmetry detection. Participants also demonstrated a better knowledge of smile asymmetry than cheek raiser action. The knowledge gathered suggests that the ability of the receptor to judge smile authenticity is limited by perceptive factors. However, the mediation analyses that we conducted show the judging smile authenticity is not limited to simple perceptive detection of facial clues. Detecting facial clues is a necessary condition for correctly assessing smile authenticity, but it does not explain the variance in these assessments. We believe that this variance would be due more to the importance that participants give to facial clues. Finally, our results show that the capacity to detect differences between authentic and fake smiles is not easy to change. Participants who received modified information on changes of appearance linked to the two facial parameters were not more likely to detect the differences than participants who did not receive information.     

A smile biases the recognition of eye expressions: Configural projection from a salient mouth

A smile is visually highly salient and grabs attention automatically. We investigated how extrafoveally seen smiles influence the viewers' perception of non-happy eyes in a face. A smiling mouth appeared in composite faces with incongruent non-happy (fearful, neutral, etc.) eyes, thus producing blended expressions, or it appeared in intact faces with genuine expressions. Attention to the eye region was spatially cued while foveal vision of the mouth was blocked by gaze-contingent masking. Participants judged whether the eyes were happy or not. Results indicated that the smile biased the evaluation of the eye expression: The same non-happy eyes were more likely to be judged as happy and categorized more slowly as not happy in a face with a smiling mouth than in a face with a non-smiling mouth or with no mouth. This bias occurred when the mouth and the eyes appeared simultaneously and aligned, but also to some extent when they were misaligned and when the mouth appeared after the eyes. We conclude that the highly salient smile projects to other facial regions, thus influencing the perception of the eye expression. Projection serves spatial and temporal integration of face parts and changes.     

Processing valence and intensity of infant expressions: the roles of expertise and gender

Several studies have provided evidence of a women's better accuracy in interpreting emotional states. Despite this difference is generally ascribed to the primary role of female gender in the affective relation with the offspring, to date, little information is available regarding gender differences in the ability to interpret infant facial expressions. In the present study, we examined the roles of gender and expertise in interpreting infant expression in 34 men and women who differed in their experience with infants. Women showed a significantly higher level of decoding accuracy compared to men. Expertise positively affected facial expressions decoding among women only. Our results suggest that in judging emotional facial expressions of infants, there is an interaction of biological (i.e., gender) and cultural factors that is independent of a woman's socioeconomic status.     

Reducing emotional reasoning: An experimental manipulation in individuals with fear of spiders

We investigated the effects of smiling on perceptions of positive, neutral and negative verbal statements. Participants viewed computer-generated movies of female characters who made angry, disgusted, happy or neutral statements and then showed either one of two temporal forms of smile (slow vs. fast onset) or a neutral expression. Smiles significantly increased the perceived positivity of the message by making negative statements appear less negative and neutral statements appear more positive. However, these smiles led the character to be seen as less genuine than when she showed a neutral expression. Disgust + smile messages led to higher judged happiness than did anger + smile messages, suggesting that smiles were seen as reflecting humour when combined with disgust statements, but as masking negative affect when combined with anger statements. These findings provide insights into the ways that smiles moderate the impact of verbal statements.     

Decoders' processing of emotional facial expression — a top-down or bottom-up mechanism?

To date little evidence is available as to how emotional facial expression is decoded, specifically whether a bottom-up (data-driven) or a top-down (schema-driven) approach is more appropriate in explaining the decoding of emotions from facial expression. A study is reported (conducted with N = 20 subjects each in Germany and Italy), in which decoders judged emotions from photographs of facial expressions. Stimuli represented a selection of photographs depicting both single muscular movements (action units) in an otherwise neutral face, and combinations of such action units. Results indicate that the meaning of action units changes often with context; only a few single action units transmit specific emotional meaning, which they retain when presented in context. The results are replicated to a large degree across decoder samples in both nations, implying fundamental mechanisms of emotion decoding.     

When does the brain distinguish between genuine and ambiguous smiles? An ERP study

Event-related brain potentials (ERPs) were recorded to assess the processing time course of ambiguous facial expressions with a smiling mouth but neutral, fearful, or angry eyes, in comparison with genuinely happy faces (a smile and happy eyes) and non-happy faces (neutral, fearful, or angry mouth and eyes). Participants judged whether the faces looked truly happy or not. Electroencephalographic recordings were made from 64 scalp electrodes to generate ERPs. The neural activation patterns showed early P200 sensitivity (differences between negative and positive or neutral expressions) and EPN sensitivity (differences between positive and neutral expressions) to emotional valence. In contrast, sensitivity to ambiguity (differences between genuine and ambiguous expressions) emerged only in later LPP components. Discrimination of emotional vs. neutral affect occurs between 180 and 430 ms from stimulus onset, whereas the detection and resolution of ambiguity takes place between 470 and 720 ms. In addition, while blended expressions involving a smile with angry eyes can be identified as not happy in the P200 (175–240 ms) component, smiles with fearful or neutral eyes produce the same ERP pattern as genuinely happy faces, thus revealing poor discrimination.     

Brighten up: Smiles facilitate perceptual judgment of facial lightness

The metaphoric expression ‘bright smile’ may reflect the actual judgment of facial lightness under varying emotional expressions. The present research examined whether people in fact judge smiling faces as perceptually brighter than frowning faces. Four studies demonstrated that participants believed smiling faces were brighter compared to frowning faces in a binary choice task and in an absolute judgment task. The results suggest that emotional expressions (i.e., smiles and frowns) can bias judgments of facial brightness in ways consistent with the metaphor. Among other implications, such results suggest that stereotypes about darker-skinned individuals may be attenuated by smiles.     

Smiles as signals of lower status in football players and fashion models: Evidence that smiles are associated with lower dominance and lower prestige

Across four studies, the current paper demonstrates that smiles are associated with lower social status. Moreover, the association between smiles and lower status appears in the psychology of observers and generalizes across two forms of status: prestige and dominance. In the first study, faces of fashion models representing less prestigious apparel brands were found to be more similar to a canonical smile display than the faces of models representing more prestigious apparel brands.?In a second study, after being experimentally primed with either high or low prestige fashion narratives, participants in the low prestige condition were more likely to perceive smiles in a series of photographs depicting smiling and non-smiling faces. A third study of football player photographs revealed that the faces of less dominant (smaller) football players were more similar to the canonical smile display than the faces of their physically larger counterparts. Using the same football player photographs, a fourth study found that smiling was a more reliable indicator of perceived status-relevant personality traits than perceptions of the football players' physical sizes inferred from the photographs.     

The value of genuine and polite smiles

Humans show remarkable ability to adapt their social behavior to suit the changing requirements of their interactions. An interaction partner's social cues, particularly facial expressions, likely play an important role in motivating and reinforcing this behavioral adaptation. Over three studies, we test a key aspect of this idea. Specifically, we ask how the reinforcement value of facial expressions compares to that of nonsocial feedback and to what degree two frequently occurring expressions (genuine and polite smiles) differ in reinforcement value. Our findings show that social feedback is preferred over nonsocial feedback and that genuine smiles are preferred over polite smiles. Based on a logistic model of our data, we show that both monetary and social values of stimuli contribute significantly to participants' decisions. Indeed, participants were willing to sacrifice the chance of a monetary reward to receive a genuine smile and produced inflated estimates of the value of genuinely smiling faces. These findings suggest that genuine smiles, and potentially other social cues, may be useful social reinforcers and therefore important in the control of social behavior on a moment-to-moment basis during interaction.     

Social anxiety and interpretation of ambiguous smiles

This study investigated whether social anxiety facilitates the discrimination between genuine and ambiguous smiles. Socially anxious (N=20) and nonanxious (N=20) participants categorized as “happy” or “not happy” faces with either (1) a truly happy expression (i.e., happy eyes and a smile), (2) truly nonhappy expressions (e.g., angry eyes and an angry mouth), or (3) blended expressions with a smiling mouth and nonhappy (e.g., angry, sad, etc.) eyes. Results indicated that, relative to nonanxious participants, those high in social anxiety were more likely to judge as “not happy” any blended expression with nonhappy eyes, and they were faster in judging as “not happy” the blended expressions with angry, fearful, or disgusted eyes (but not those with sad, surprised, or neutral eyes). These results suggest, respectively, that social anxiety inhibits a benign interpretation of all the ambiguous expressions with a smile, and speeds up the detection of threatening eyes in such expressions. Importantly, no differences appeared as a function of social anxiety for truly happy or nonhappy faces. This rules out a response-bias explanation, and also reveals that social anxiety does not affect sensitivity in the recognition of prototypical expressions.     

The eyes have it: making positive expressions more positive and negative expressions more negative

Facial expressions frequently involve multiple individual facial actions. How do facial actions combine to create emotionally meaningful expressions? Infants produce positive and negative facial expressions at a range of intensities. It may be that a given facial action can index the intensity of both positive (smiles) and negative (cry-face) expressions. Objective, automated measurements of facial action intensity were paired with continuous ratings of emotional valence to investigate this possibility. Degree of eye constriction (the Duchenne marker) and mouth opening were each uniquely associated with smile intensity and, independently, with cry-face intensity. In addition, degree of eye constriction and mouth opening were each unique predictors of emotion valence ratings. Eye constriction and mouth opening index the intensity of both positive and negative infant facial expressions, suggesting parsimony in the early communication of emotion.     

Duchenne smile,emotional experience,and autonomic reactivity: a test of the facial feedback hypothesis

This study examined the modulatory function of Duchenne and non-Duchenne smiles on subjective and autonomic components of emotion. Participants were asked to hold a pencil in their mouth to either facilitate or inhibit smiles and were not instructed to contract specific muscles. Five conditions--namely lips pressing, low-level non-Duchenne smiling, high-level non-Duchenne smiling, Duchenne smiling, and control--were produced while participants watched videoclips that were evocative of positive or negative affect. Participants who displayed Duchenne smiles reported more positive experience when pleasant scenes and humorous cartoons were presented. Furthermore, they tended to exhibit different patterns of autonomic arousal when viewing positive scenes. These results support thefacial feedback hypothesis and suggest that facial feedback has more powerful effects when facial configurations represent valid analogs of basic emotional expressions.     

Beyond smile dynamics: mimicry and beliefs in judgments of smiles

The judgment that a smile is based on "true," usually positive, feelings affects social interaction. However, the processes underlying the interpretation of a smile as being more or less genuine are not well understood. The aim of the present research was to test predictions of the Simulation of Smiles Model (SIMS) proposed by Niedenthal, Mermillod, Maringer, and Hess (2010). In addition to the perceptual features that can guide the judgment of a smile as genuine, the model identifies the conditions that the judgments rely on: (a) the embodiment of the facial expression and its corresponding state, and (b) beliefs about the situations in which genuine smiles are most often expressed. Results of two studies are consistent with the model in that they confirm the hypotheses that facial mimicry provides feedback that is used to judge the meaning of a smile, and that beliefs about the situations in which a smile occurs guides such judgments when mimicry is inhibited.     

Many moral buttons or just one? Evidence from emotional facial expressions

We investigated whether moral violations involving harm selectively elicit anger, whereas purity violations selectively elicit disgust, as predicted by the Moral Foundations Theory (MFT). We analysed participants’ spontaneous facial expressions as they listened to scenarios depicting moral violations of harm and purity. As predicted by MFT, anger reactions were elicited more frequently by harmful than by impure actions. However, violations of purity elicited more smiling reactions and expressions of anger than of disgust. This effect was found both in a classic set of scenarios and in a new set in which the different kinds of violations were matched on weirdness. Overall, these findings are at odds with predictions derived from MFT and provide support for “monist” accounts that posit harm at the basis of all moral violations. However, we found that smiles were differentially linked to purity violations, which leaves open the possibility of distinct moral modules.     

Smiling virtual agent in social context

A smile may communicate different communicative intentions depending on subtle characteristics of the facial expression. In this article, we propose an algorithm to determine the morphological and dynamic characteristics of virtual agent's smiles of amusement, politeness, and embarrassment. The algorithm has been defined based on a virtual agent's smiles corpus constructed by users and analyzed with a decision tree classification technique. An evaluation, in different contexts, of the resulting smiles has enabled us to validate the proposed algorithm.     

Influence of emotional facial expressions on 3-5-year-olds’ face recognition

Three experiments examined 3- and 5-year-olds’ recognition of faces in constant and varied emotional expressions. Children were asked to identify repeatedly presented target faces, distinguishing them from distractor faces, during an immediate recognition test and during delayed assessments after 10 min and one week. Emotional facial expression remained neutral (Experiment 1) or varied between immediate and delayed tests: from neutral to smile and anger (Experiment 2), from smile to neutral and anger (Experiment 3, condition 1), or from anger to neutral and smile (Experiment 3, condition 2). In all experiments, immediate face recognition was not influenced by emotional expression for either age group. Delayed face recognition was most accurate for faces in identical emotional expression. For 5-year-olds, delayed face recognition (with varied emotional expression) was not influenced by which emotional expression had been displayed during the immediate recognition test. Among 3-year-olds, accuracy decreased when facial expressions varied from neutral to smile and anger but was constant when facial expressions varied from anger or smile to neutral, smile or anger. Three-year-olds’ recognition was facilitated when faces initially displayed smile or anger expressions, but this was not the case for 5-year-olds. Results thus indicate a developmental progression in face identity recognition with varied emotional expressions between ages 3 and 5.     

Perceiving emotions: Cueing social categorization processes and attentional control through facial expressions

Individuals spontaneously categorise other people on the basis of their gender, ethnicity and age. But what about the emotions they express? In two studies we tested the hypothesis that facial expressions are similar to other social categories in that they can function as contextual cues to control attention. In Experiment 1 we associated expressions of anger and happiness with specific proportions of congruent/incongruent flanker trials. We also created consistent and inconsistent category members within each of these two general contexts. The results demonstrated that participants exhibited a larger congruency effect when presented with faces in the emotional group associated with a high proportion of congruent trials. Notably, this effect transferred to inconsistent members of the group. In Experiment 2 we replicated the effects with faces depicting true and false smiles. Together these findings provide consistent evidence that individuals spontaneously utilise emotions to categorise others and that such categories determine the allocation of attentional control.     

Smiling virtual agent in social context

A smile may communicate different communicative intentions depending on subtle characteristics of the facial expression. In this article, we propose an algorithm to determine the morphological and dynamic characteristics of virtual agent’s smiles of amusement, politeness, and embarrassment. The algorithm has been defined based on a virtual agent’s smiles corpus constructed by users and analyzed with a decision tree classification technique. An evaluation, in different contexts, of the resulting smiles has enabled us to validate the proposed algorithm.