Options
Zuberbühler, Klaus
Nom
Zuberbühler, Klaus
Affiliation principale
Fonction
Professeur ordinaire
Email
klaus.zuberbuehler@unine.ch
Identifiants
Résultat de la recherche
Voici les éléments 1 - 10 sur 194
- PublicationAccès libreFrom Forest to Zoo: Domain Adaptation in Animal Behavior Recognition for Great Apes with ChimpBehave(2024-06-17)
; ; ; ; This paper addresses the significant challenge of recognizing behaviors in non-human primates, specifically focusing on chimpanzees. Automated behavior recognition is crucial for both conservation efforts and the advancement of behavioral research. However, it is significantly hindered by the labor-intensive process of manual video annotation. Despite the availability of large-scale animal behavior datasets, the effective application of machine learning models across varied environmental settings poses a critical challenge, primarily due to the variability in data collection contexts and the specificity of annotations. In this paper, we introduce ChimpBehave, a novel dataset featuring over 2 hours of video (approximately 193,000 video frames) of zoo-housed chimpanzees, meticulously annotated with bounding boxes and behavior labels for action recognition. ChimpBehave uniquely aligns its behavior classes with existing datasets, allowing for the study of domain adaptation and cross-dataset generalization methods between different visual settings. Furthermore, we benchmark our dataset using a state-of- theart CNN-based action recognition model, providing the first baseline results for both within and cross-dataset settings. The dataset, models, and code can be accessed at: https://github.com/MitchFuchs/ChimpBehave - 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 libreJuvenile vervet monkeys rely on others when responding to danger(2023)
; ; AbstractPrimate alarm calls are mainly hardwired but individuals need to adapt their calling behaviours according to the situation. Such learning necessitates recognising locally relevant dangers and may take place via their own experience or by observing others. To investigate monkeys alarm calling behaviour, we carried out a field experiment in which we exposed juvenile vervet monkeys to unfamiliar raptor models in the presence of audiences that differed in experience and reliability. We used audience age as a proxy for experience and relatedness as a proxy for reliability, while quantifying audience reactions to the models. We found a negative correlation between alarm call production and callers’ age. Adults never alarm called, compared to juveniles. We found no overall effect of audience composition and size, with juveniles calling more when with siblings than mothers or unrelated individuals. Finally, concerning audience reactions to the models, we observed juveniles remained silent with vigilant mothers and only alarm called with ignoring mothers, whereas we observed the opposite for siblings: juveniles remained silent with ignoring siblings and called with vigilant siblings. Despite the small sample size, juvenile vervet monkeys, confronted with unfamiliar and potentially dangerous raptors, seem to rely on others to decide whether to alarm call, demonstrating that the choice of the model may play an important key role in the ontogeny of primate alarm call behaviour. - PublicationAccès libreCall combinations and compositional processing in wild chimpanzees(2023)
; - 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 libreA new method to determine the diet of pygmy hippopotamus in Taï National Park, Côte d’Ivoire(2021)
; AbstractDiet determination of endangered species is an essential element in defining successful conservation strategies and optimising captive breeding programmes. In this study, we developed a new diet identification system, derived from standard faecal analysis, to determine the diet of an elusive and endangered herbivore, the pygmy hippopotamus (Choeropsis liberiensis). We collected faecal samples from 10 free‐ranging individuals covering a combined home range area of about 50 km2 in Taï National Park, Côte d’Ivoire. In subsequent laboratory analyses, we extracted a large number of leaf epidermis fragments from spatially separated faecal samples and compared them with a reference plant database. Using Multiple Correspondence Analysis (MCA) of epidermis fragments combined with direct visual inspection, we identified the most frequently consumed plant species, which revealed that pygmy hippopotami qualified as intermediate feeders. Their diet was based on at least seven species of monocotyledonae, dicotyledonae and fern groups, with a preference for a small number of other plant species. We evaluate the merit of our method and discuss our findings for developing effective conservation and captive breeding strategies in an endangered species with a wild population of less than 2500 adult individuals. - PublicationAccès libreBonobos engage in joint commitment(2020-12-18)
; ; ; ;