Cross-cultural perceptions of facial resemblance between kin

Transforming faces to mimic natural kin: A comparison of different paradigms

The ability to detect phenotypic similarity or kinship in third-parties' faces is not perfect, but better than chance. Still, some humans are better than others at this task. Yet researchers in kinship detection have difficulties in building up large and diverse datasets of high-quality pictures of related persons. The current experiments tested a novel method for circumventing this difficulty by using morphing techniques in order to generate a wide array of stimuli derived from a limited number of individual pictures. Six experiments tested various stimuli (standard protocol, mirrored face, other-sex face, other-ethnicity face, other-expression face and antiface). Our benchmarks are the similarity or kinship scores achieved by participants when faced with pictures of real siblings. We show that all stimuli, except the antiface, elicit detection scores similar to those elicited by real pictures of actual siblings. In addition, by exploring different experiment parameters (simultaneous or sequential task, kinship or similarity task) and some individual characteristics, these experiments provide a better understanding of kinship detection in third parties. The validation of our new method will allow widening the range of available stimuli to the research community, and even to develop new ecologically relevant experimental protocols that are hardly or not feasible with veridical images.

Familiar faces as islands of expertise

Most people recognise and match pictures of familiar faces effortlessly, while struggling to match unfamiliar face images. This has led to the suggestion that true human expertise for faces applies only to familiar faces. This paper develops that idea to propose that we have isolated 'islands of expertise' surrounding each familiar face that allow us to perform better with faces that resemble those we already know. This idea is tested in three experiments. The first shows that familiarity with a person facilitates identification of their relatives. The second shows that people are better able to remember faces that resemble someone they already know. The third shows that while prompting participants to think about resemblance at study produces a large positive effect on subsequent recognition, there is still a significant effect if there is no such prompt. Face-space-R (Lewis, 2004) is used to illustrate a possible computational explanation of the processes involved.

Patch-Based Dual-Tree Complex Wavelet Transform for Kinship Recognition

Kinship recognition is a prominent research aiming to find if kinship relation exists between two different individuals. In general, child closely resembles his/her parents more than others based on facial similarities. These similarities are due to genetically inherited facial features that a child shares with his/her parents. Most existing researches in kinship recognition focus on full facial images to find these kinship similarities. This paper first presents kinship recognition for similar full facial images using proposed Global-based dual-tree complex wavelet transform (G-DTCWT). We then present novel patch-based kinship recognition methods based on dual-tree complex wavelet transform (DT-CWT): Local Patch-based DT-CWT (LP-DTCWT) and Selective Patch-Based DT-CWT (SP-DTCWT). LP-DTCWT extracts coefficients for smaller facial patches for kinship recognition. SP-DTCWT is an extension to LP-DTCWT and extracts coefficients only for representative patches with similarity scores above a normalized cumulative threshold. This threshold is computed by a novel patch selection process. These representative patches contribute more similarities in parent/child image pairs and improve kinship accuracy. Proposed methods are extensively evaluated on different publicly available kinship datasets to validate kinship accuracy. Experimental results showcase efficacy of proposed methods on all kinship datasets. SP-DTCWT achieves competitive accuracy to state-of-the-art methods. Mean kinship accuracy of SP-DTCWT is 95.85% on baseline KinFaceW-I and 95.30% on KinFaceW-II datasets. Further, SP-DTCWT achieves the state-of-the-art accuracy of 80.49% on the largest kinship dataset, Families In the Wild (FIW).

On the Complexity of Visual Judgement of Kinship

Discrimination of close relatives is a basic ability of humans, with demonstrated and important consequences in social and sexual behaviours. In this article, we investigate the visual judgement of kinship, that is the process of discriminating relatives based on visual cues and, in particular, on facial resemblance. Starting from triplets of face stimuli, we focus on a simple two-alternative forced choice protocol and we ask participants to evaluate kinship, similarity, or dissimilarity. Response times of the participants performing these visual judgements are recorded and further analysed. The analysis can also benefit from previous findings on the adopted face data set; in particular, results are compared with reference to an independently generated and statistically reliable similarity index, which is available for each possible considered pair of images. Our results confirm previous findings stating that kinship and similarity judgements are closely related and take longer, on average, than dissimilarity judgement. Moreover, they confirm that similarity and dissimilarity cannot be considered just as opposite concepts, and strongly support the existence of different pathways for similarity and dissimilarity judgements. Concerning kinship judgements, results confirm the assumption, inherent in previous models, of a close relationship between cues signalling for kinship and cues signalling for similarity but suggest the existence of a more complex process, where dissimilarity cues need to be explicitly included in order to model measured effects. Our results reinforce the idea that modulation mechanisms between similarity and dissimilarity measures could explain selective suppression or enhancement effects reported in previous works. A new framework is thus proposed hypothesising that kinship recognition is the result of a balanced evaluation of both similar or dissimilar pathways.

Birth Order Does Not Affect Ability to Detect Kin

Previous studies suggest that birth order affects kinship detection ability. Kaminski et al. (2010) argued that firstborns use contextual cues (e.g., maternal perinatal association) to assess kinship in their own family, leading to a disadvantage in assessing kinship from facial cues alone in strangers. In contrast, laterborns do not have the contextual cue of maternal perinatal association and hence rely more on facial cues, leading to an advantage in detecting kin from facial cues alone. However, Alvergne et al. (2010) found no evidence in support of such a birthorder effect. The current study aimed to replicate previous studies with better suited methods to determine the effect of birth order on kin recognition. 109 raters viewed 132 pairs of photographs of children (aged 3–17 years), and indicated whether each pair was related or unrelated. Half of the pairs were sibling pairs and half were unrelated child pairs that were age- and gender- matched to the related pairs. No image was shown more than once, related pairs were not known to be related to any other image in the study, and individuals from unrelated pairs were not known to be related to any other image. We used binomial logistic mixed effects modelling to predict kinship judgments from relatedness and birth order (with image pair and rater as random factors). Relatedness was the main factor driving kinship judgments; related child-pairs were more than twice as likely as unrelated pairs to be judged as kin. Kinship judgment accuracy was unaffected by rater birth order. These findings indicate that laterborns did not have an advantage in detecting child sibling pairs. Pre-registration, data, code, and preprint available at osf.io/h43ep.

Do mothers also "manipulate" grandparental care?

Paternity uncertainty has proven to be a robust ultimate hypothesis for predicting the higher investment in grandchildren observed among maternal grandparents compared to that of the paternal grandparents. Yet the proximate mechanisms for generating such preferred biases in grandparental investment remain unclear. Here we address two different questions for better understanding the proximate mechanisms leading to the observed bias in grandparental investments: (i) is there a larger emphasis on resemblance descriptions (between grandchildren and grandparent) among daughters than among sons, and (ii) do mothers really believe that their offspring more resemble their parents, that is, the children's grandparents, than fathers do? From questioning grandparents, we find that daughters more often and more intensely than sons express opinions about grandchild-grandparent resemblance. Moreover, daughters also seem to believe that their children more resemble their grandmother than sons do. The latter is, however, not the case for beliefs about children's resemblance to grandfathers. In sum, our results suggest that even in a population of Norwegians, strongly influenced by ideas concerning gender equality, there exist a sexual bias among parents in opinions and descriptions about grandchild-grandparent resemblance. This resemblance bias, which echoes that of mothers biasing resemblance descriptions of newborns to putative fathers, does not seem to represent a conscious manipulation. Yet it could be instrumental for influencing grandparental investments. We believe that a "manipulative mother hypothesis" might parsimoniously account for many of the results relating to biased alloparenting hitherto not entirely explained by "the paternity uncertainty hypothesis."

Prototype-Based Discriminative Feature Learning for Kinship Verification

In this paper, we propose a new prototype-based discriminative feature learning (PDFL) method for kinship verification. Unlike most previous kinship verification methods which employ low-level hand-crafted descriptors such as local binary pattern and Gabor features for face representation, this paper aims to learn discriminative mid-level features to better characterize the kin relation of face images for kinship verification. To achieve this, we construct a set of face samples with unlabeled kin relation from the labeled face in the wild dataset as the reference set. Then, each sample in the training face kinship dataset is represented as a mid-level feature vector, where each entry is the corresponding decision value from one support vector machine hyperplane. Subsequently, we formulate an optimization function by minimizing the intraclass samples (with a kin relation) and maximizing the neighboring interclass samples (without a kin relation) with the mid-level features. To better use multiple low-level features for mid-level feature learning, we further propose a multiview PDFL method to learn multiple mid-level features to improve the verification performance. Experimental results on four publicly available kinship datasets show the superior performance of the proposed methods over both the state-of-the-art kinship verification methods and human ability in our kinship verification task.

Identification of visual paternity cues in humans

Understanding how individuals identify their relatives has implications for the evolution of social behaviour. Kinship cues might be based on familiarity, but in the face of paternity uncertainty and costly paternal investment, other mechanisms such as phenotypic matching may have evolved. In humans, paternal recognition of offspring and subsequent discriminative paternal investment have been linked to father-offspring facial phenotypic similarities. However, the extent to which paternity detection is impaired by environmentally induced facial information is unclear. We used 27 portraits of fathers and their adult sons to quantify the level of paternity detection according to experimental treatments that manipulate the location, type and quantity of visible facial information. We found that (i) the lower part of the face, that changes most with development, does not contain paternity cues, (ii) paternity can be detected even if relational information within the face is disrupted and (iii) the signal depends on the presence of specific information rather than their number. Taken together, the results support the view that environmental effects have little influence on the detection of paternity using facial similarities. This suggests that the cognitive dispositions enabling the facial detection of kinship relationships ignore genetic irrelevant facial information.

Neighborhood repulsed metric learning for kinship verification

Kinship verification from facial images is an interesting and challenging problem in computer vision, and there are very limited attempts on tackle this problem in the literature. In this paper, we propose a new neighborhood repulsed metric learning (NRML) method for kinship verification. Motivated by the fact that interclass samples (without a kinship relation) with higher similarity usually lie in a neighborhood and are more easily misclassified than those with lower similarity, we aim to learn a distance metric under which the intraclass samples (with a kinship relation) are pulled as close as possible and interclass samples lying in a neighborhood are repulsed and pushed away as far as possible, simultaneously, such that more discriminative information can be exploited for verification. To make better use of multiple feature descriptors to extract complementary information, we further propose a multiview NRML (MNRML) method to seek a common distance metric to perform multiple feature fusion to improve the kinship verification performance. Experimental results are presented to demonstrate the efficacy of our proposed methods. Finally, we also test human ability in kinship verification from facial images and our experimental results show that our methods are comparable to that of human observers.

The Evolution of Altruistic Preferences: Mothers versus Fathers

What can evolutionary biology tell us about male-female differences in preferences concerning family matters? Might mothers be more solicitous toward offspring than fathers, for example? The economics literature has documented gender differences-children benefit more from money put in the hands of mothers rather than fathers, for example-and these differences are thought to be partly due to preferences. Yet for good reason family economics is mostly concerned with how prices and incomes affect behavior against a backdrop of exogenous preferences. Evolutionary biology complements this approach by treating preferences as the outcome of natural selection. We mine the well-developed biological literature to make a prima facie case for evolutionary roots of parental preferences. We consider the most rudimentary of traits-sex differences in gamete size and internal fertilization-and explain how they have been thought to generate male-female differences in altruism toward children and other preferences related to family behavior. The evolutionary approach to the family illuminates connections between issues typically thought distinct in family economics, such as parental care and marriage markets.

Does facial resemblance enhance cooperation?

Facial self-resemblance has been proposed to serve as a kinship cue that facilitates cooperation between kin. In the present study, facial resemblance was manipulated by morphing stimulus faces with the participants' own faces or control faces (resulting in self-resemblant or other-resemblant composite faces). A norming study showed that the perceived degree of kinship was higher for the participants and the self-resemblant composite faces than for actual first-degree relatives. Effects of facial self-resemblance on trust and cooperation were tested in a paradigm that has proven to be sensitive to facial trustworthiness, facial likability, and facial expression. First, participants played a cooperation game in which the composite faces were shown. Then, likability ratings were assessed. In a source memory test, participants were required to identify old and new faces, and were asked to remember whether the faces belonged to cooperators or cheaters in the cooperation game. Old-new recognition was enhanced for self-resemblant faces in comparison to other-resemblant faces. However, facial self-resemblance had no effects on the degree of cooperation in the cooperation game, on the emotional evaluation of the faces as reflected in the likability judgments, and on the expectation that a face belonged to a cooperator rather than to a cheater. Therefore, the present results are clearly inconsistent with the assumption of an evolved kin recognition module built into the human face recognition system.

Perceptual Factors Affecting the Ability to Assess Facial Resemblance between Parents and Newborns in Humans

People undeniably pay attention to faces, and facial resemblance may act as a kinship cue. However, previous studies have shown that the ability to detect kinship through facial resemblance is limited, and it has been suggested that this may be due to several types of perceptual factors. To further understand the processes that underpin kinship judgment, it is important to investigate which perceptual factors predict the probability of parent–child pairs being detected as related. To this end, we performed two experiments. In the first, we evaluated the ability of human observers to match newborns with one of their parents. In the second, we explored three perceptual factors that may have influenced kinship detection (gender discrimination, facial attractiveness, and perceptual similarity). Results showed that the participants were able to match newborns with one of their parents, even though the task was perceived as difficult. Moreover, our study goes further than previous findings, showing that the perceptual factors investigated may significantly contribute to kinship detection.