HAL Portal UBS (Université Bretagne Sud)
Not a member yet
14201 research outputs found
Sort by
Development of decision support tools by model order reduction for active endovascular navigation
No full textInternational audienceEndovascular therapies consist in treating vascular pathologies mini-invasively by inserting long tools towards the area to treat. However, some trajectories are so-called complex (e.g. Supra-Aortic Trunks (SATs)). In order to facilitate the access to complex targets by catheterization, an active guidewire made of Shape Memory Alloy has been developed. Our study focuses on the navigation of this device and associated catheters towards neurovascular targets through the left carotid artery. In a previous study, a finite element model was developed to simulate the navigation of the active guidewire and catheters from the aortic arch to the hooking of the left carotid artery of patient-specific aortas. However, numerical simulations are time-consuming and cannot be used directly in the clinic routine to provide navigation assistance. We present in this study the development of numerical charts aiming to provide a real-time computation, based on high-fidelity FE simulations, of: 1. the behaviour of the active guidewire; 2. the navigation of the active guidewire and associated catheters within a given anatomy for specific guidewire and navigation parameters. These charts are developed using the HOPGD method and demonstrate their ability to provide an accurate real-time response from a limited number of preliminary high-fidelity computations
Cannabis et grossesse : une diversité de pratiques de repérage parmi des sages-femmes libérales françaises
No full textInternational audienceThis article focuses on the practices of French midwives in identifying the risk associated with cannabis during pregnancy. The study is based on a content analysis of semistructured interviews conducted with 37 independent midwives. The professionals interviewed systematically address the issue of cannabis use with pregnant patients. Their practices are not very formalized and are very diverse. Suggestions are made to improve the effectiveness of this screening.Cet article s’intéresse aux pratiques des sages-femmes françaises en matière d’identification du risque lié à la consommation de cannabis pendant la grossesse. L’étude est basée sur une analyse de contenu d’entretiens semi-directifs réalisés auprès de 37 sages-femmes libérales. Les professionnelles interrogées abordent de manière systématique la question de la consommation de cannabis avec leurs patientes enceintes selon des pratiques peu formalisées et très diversifiées. Des suggestions sont formulées pour améliorer l’efficacité de ce repérage
Predicting Museum Visitors Intention through Nonverbal Cues : the Potential of Facial Expressions
No full textInternational audienceUnderstanding museum visitor engagement is crucial for optimizing communication strategies and exhibition design. Museums need to anticipate visit intentions and potential visitors. This study explores facial expression analysis as a nonverbal predictor of museum visit intentions, focusing on joy intensity in response to promotional exhibition visuals. Using a mixed-effects logistic regression model on 61 participants, we demonstrate that joy expression significantly predicts visit intention, particularly among undecided visitors. The results highlight variations across age, gender, and cultural background, suggesting tailored marketing strategies. By integrating facial emotion recognition into predictive models, this research contributes to anticipating cultural engagement and emphasizing the potential of nonverbal cues in enhancing audience segmentation and communication effectiveness in cultural institutions. These findings provide new avenues for museum marketing and visitor experience managemen
MONSTER: Monash Scalable Time Series Evaluation Repository
No full textInternational audienceWe introduce Monster-the MONash Scalable Time Series Evaluation Repository-a collection of large datasets for time series classification and associated set of classification tasks that jointly define a new time series classification benchmark. The field of time series classification has benefitted from common benchmarks set by the UCR and UEA time series classification repositories. However, the datasets in these benchmarks are small, with median training set sizes of 217 and 255 examples, respectively. In consequence they favour a narrow subspace of models that are optimised to achieve low classification error on a wide variety of smaller datasets, that is, models that minimise variance, and give little weight to computational issues such as scalability. Our hope is to diversify the field by introducing benchmarks using larger datasets. We believe that there is enormous potential for new progress in the field by engaging with the theoretical and practical challenges of learning effectively from larger quantities of data.45 pages; 38 figure
Effort Minimization and the Built Environment: Public Health Implications
No full textInternational audiencePromoting physical activity represents a major public health opportunity due to its significant impact on physical and mental health. Despite ongoing efforts, public health interventions often struggle to achieve sustainable behavioral changes. Instead of explicitly or implicitly attributing such failures to a lack of individual motivation, it is essential to consider the characteristics of contemporary environments that promote physical inactivity. We propose an explanatory framework that integrates the theory of effort minimization in physical activity with the postulates of the ecological model of physical activity behavior. According to theory of effort minimization in physical activity, humans have an innate tendency to avoid physical effort, making it difficult to adopt an active lifestyle in environments where opportunities to minimize effort are pervasive. Complementarily, the ecological model emphasizes the key role of built environment in providing behavior settings—those social and physical situations that can promote and sometimes demand certain actions and discourage or prohibit others. Building from the theory of effort minimization in physical activity, we suggest that redesigning the built environment so that being active is the default behavioral option, while ensuring that it elicits positive affective responses, could be a decisive strategy. Such an approach could not only increase physical activity levels across the population but also help to reduce gender differentials and sociospatial inequalities in participation
Quelques rappels sur l'exercice du pouvoir disciplinaire à l'encontre d'un agent excédant sous couvert d'humour les limites admises
No full textInternational audienceObservations sous TA Marseille, 4 juillet 2025, no 240105
Faire l’histoire de la santé et de la médecine aujourd’hui : objets, enjeux et méthodes
No full textInternational audienc
Towards a Digital Virtual Twin for Heterogeneous Thermomechanical Experiments
No full textInternational audienceABSTRACT The combination of digital image correlation with infrared thermography (DIC‐IRT) in heterogeneous thermomechanical tests enables the simultaneous measurement of temperature and deformation fields and can be used to identify thermomechanical constitutive parameters using inverse identification methods. However, two major challenges remain. The first is the quantification of experimental uncertainties on identification results. The second is the design of optimal test configurations that allow a single heterogeneous DIC‐IRT experiment to be reliably exploited for the inverse identification of unknown constitutive parameters. To address the first challenge, this paper presents the first developments towards a digital virtual twin (DVT) able to simulate heterogeneous thermomechanical tests integrating DIC‐IRT. The proposed methodology relies on virtual image deformation, which mimics kinematic field measurements with DIC. To include thermal field measurements, a numerical measurement chain is added to simulate the acquisition of infrared images and the spatial synchronization with kinematic data. Based on this approach, a metrological study has enabled the characterization of parameters required to reproduce the experimental uncertainties in the measurement chain. The DVT was then used to simulate a thermomechanical test, which consists of a thick steel sample subjected to heterogeneous thermal strain and temperature fields. The test is designed to identify the thermal expansion coefficient and thermophysical properties from a single experiment using finite element model updating (FEMU). By processing virtual DIC‐IRT data with FEMU, the confidence intervals on the identified parameters were predicted based on random and systematic errors in DIC‐IRT measurements. Finally, the main limitation hampering the use of the proposed DVT for optimizing the design of heterogeneous thermomechanical tests is discussed, along with potential strategies to overcome it
ITERATED RANDOM WALKS IN RANDOM SCENERY (PAPAPA)
No full textVersion acceptée pour publication à SPAWe establish a limit theorem for a new model of 3-dimensional random walk in an inhomogeneous lattice with random orientations. This model can be seen as a 3dimensional version of the Matheron and de Marsily model [12]. This new model leads us naturally to the study of iterated random walk in random scenery, which is a new process that can be described as a random walk in random scenery evolving in a second random scenery. We use the french acronym PAPAPA for this new process and answer a question about its stochastic behaviour asked about twenty years ago by Stéphane Le Borgne.</div