Options
Genty, Emilie
Nom
Genty, Emilie
Affiliation principale
Fonction
Collaborateur.trice scientifique
Email
emilie.genty@unine.ch
Identifiants
Résultat de la recherche
Voici les éléments 1 - 10 sur 12
- PublicationAccès libreASBAR: an Animal Skeleton-Based Action Recognition framework. Recognizing great ape behaviors in the wild using pose estimation with domain adaptation(2024)
; ; ; To date, the investigation and classification of animal behaviors have mostly relied on direct human observations or video recordings with posthoc analysis, which can be labor-intensive, time-consuming, and prone to human bias. Recent advances in machine learning for computer vision tasks, such as pose estimation and action recognition, thus have the potential to significantly improve and deepen our understanding of animal behavior. However, despite the increased availability of open-source toolboxes and large-scale datasets for animal pose estimation, their practical relevance for behavior recognition remains under-explored. In this paper, we propose an innovative framework, To date, the investigation and classification of animal behaviors have mostly relied on direct human observations or video recordings with posthoc analysis, which can be labor-intensive, time-consuming, and prone to human bias. Recent advances in machine learning for computer vision tasks, such as pose estimation and action recognition, thus have the potential to significantly improve and deepen our understanding of animal behavior. However, despite the increased availability of open-source toolboxes and large-scale datasets for animal pose estimation, their practical relevance for behavior recognition remains under-explored. In this paper, we propose an innovative framework, ASBAR, for Animal Skeleton-Based Action Recognition, which fully integrates animal pose estimation and behavior recognition. We demonstrate the use of this framework in a particularly challenging task: the classification of great ape natural behaviors in the wild. First, we built a robust pose estimator model leveraging OpenMonkeyChallenge, one of the largest available open-source primate pose datasets, through a benchmark analysis on several CNN models from DeepLabCut, integrated into our framework. Second, we extracted the great ape’s skeletal motion from the PanAf dataset, a large collection of in-the-wild videos of gorillas and chimpanzees annotated for natural behaviors, which we used to train and evaluate PoseConv3D from MMaction2, a second deep learning model fully integrated into our framework. We hereby classify behaviors into nine distinct categories and achieve a Top 1 accuracy of 74.98%, comparable to previous studies using video-based methods, while reducing the model’s input size by a factor of around 20. Additionally, we provide an open-source terminal-based GUI that integrates our full pipeline and release a set of 5,440 keypoint annotations to facilitate the replication of our results on other species and/or behaviors. All models, code, and data can be accessed at: https://github.com/MitchFuchs/asbar. - PublicationAccès libreInternational use of laughter in bonobos and chimpanzees?(Neuchâtel : Université de Neuchâtel, 2024)
; ; ; Caroline FrynsAlthough laughter is generally associated with humor and was long thought to be uniquely human, growing evidence highlights its existence in other species like primates or rats. Research on nonhuman primates’ laughter emphasizes its emotionality and context-specificity as it mostly occurs during play or when tickled, whereas human voluntary laughter presents such functional flexibility that no classification system seems to exhaust its meaning. Although laughter is identified as a tool to coordinate joint actions by easing transitions between topics and by managing social relationships in humans, these patterns have not yet been investigated in other primates. Focusing on dyadic social play in two species of great apes, bonobos (Pan paniscus) and chimpanzees (Pan troglodytes), we explore the interaction of coordination (transition between moments of an interaction) and social relationships (rank differences and social bond strength) in predicting laughter presence (1/0) and type (contact/non-contact). We show not only that laughter is used to reengage play in apes, but also that non-contact laughter seems to be a specialized coordination tool used to intentionally reengage play-activity in apes, almost absent at other moments of the interaction, be it during the Main Body or at other transition points. Rank differences and social bond strength did not however appear significant predictors neither of laughter presence nor of type. Although the same patterns were observed between species and although laughter quantity was not directly assessed, bonobos showed a higher propensity to laugh during the Main Body of the interaction and when reengaging play than chimpanzees. - PublicationAccès libreEvery product needs a process: unpacking joint commitment as a process across species(2022-7-25)
; ; ; - PublicationAccès libreHow 2- and 4-year-old children coordinate social interactions with peers(2022-7-25)
; ; ; - PublicationAccès libreEvidence of joint commitment in great apes' natural joint actions.(2021-12-01T00:00:00Z)
; ; ; ; - PublicationAccès libreAssessing joint commitment as a process in great apes.(2021-08-20T00:00:00Z)
; ; ; ; ; - PublicationAccès libreBonobos engage in joint commitment(2020-12-18)
; ; ; ; - PublicationAccès libreHow apes get into and out of joint actions: Shared intentionality as an interactional achievement(2020)
; ; ; Compared to other animals, humans appear to have a special motivation to share experiences and mental states with others (Clark, 2006; Grice, 1975), which enables them to enter a condition of ‘we’ or shared intentionality (Tomasello & Carpenter, 2005). Shared intentionality has been suggested to be an evolutionary response to unique problems faced in complex joint action coordination (Levinson, 2006; Tomasello, Carpenter, Call, Behne, & Moll, 2005) and to be unique to humans (Tomasello, 2014). The theoretical and empirical bases for this claim, however, present several issues and inconsistencies. Here, we suggest that shared intentionality can be approached as an interactional achievement, and that by studying how our closest relatives, the great apes, coordinate joint action with conspecifics, we might demonstrate some correlate abilities of shared intentionality, such as the appreciation of joint commitment. We provide seven examples from bonobo joint activities to illustrate our framework. - PublicationAccès libreSocial play as joint action: A framework to study the evolution of shared intentionality as an interactional achievement.(2017-12-01T00:00:00Z)
; ; ; Social play has a complex, cooperative nature that requires substantial coordination. This has led researchers to use social games to study cognitive abilities like shared intentionality, the skill and motivation to share goals and intentions with others during joint action. We expand this proposal by considering play as a joint action and examining how shared intentionality is achieved during human joint action. We describe how humans get into, conduct, and get out of joint actions together in an orderly way, thereby constructing the state of "togetherness" characteristic of shared intentionality. These processes play out as three main phases, the opening (where participants are ratified and joint commitments are established), the main body (where progress, ongoing commitments, and possible role reversals are coordinated), and the closing (where the intention to terminate the action is coordinated and where participants take leave of each other). We use this process in humans as a framework for examining how various animal species get into, maintain, and get out of play bouts. This comparative approach constitutes an alternative measure of those species' possession of shared intentionality. Using this framework, we review the play literature on human children and different social species of mammals and birds in search of behavioral markers of shared intentionality in the coordination of play bouts. We discuss how our approach could shed light on the evolution of the special human motivation to cooperate and share psychological states with others. - PublicationAccès libreComplex patterns of signalling to convey different social goals of sex in bonobos, Pan paniscus(2015)
; ; Sexual behaviour in bonobos (Pan paniscus) functions beyond mere reproduction to mediate social interactions and relationships. In this study, we assessed the signalling behaviour in relation to four social goals of sex in this species: appeasement after conflict, tension reduction, social bonding and reproduction. Overall, sexual behaviour was strongly decoupled from its ancestral reproductive function with habitual use in the social domain, which was accompanied by a corresponding complexity in communication behaviour. We found that signalling behaviour varied systematically depending on the initiator's goals and gender. Although all gestures and vocalisations were part of the species-typical communication repertoire, they were often combined and produced flexibly. Generally, gestures and multi-modal combinations were more flexibly used to communicate a goal than vocalisations. There was no clear relation between signalling behaviour and success of sexual initiations, suggesting that communication was primarily used to indicate the signaller's intention, and not to influence a recipient's willingness to interact sexually. We discuss these findings in light of the larger question of what may have caused, in humans, the evolutionary transition from primate-like communication to language.