Contents
Academic literature on the topic 'Acquisition automatique des données – Capture de mouvements'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the lists of relevant articles, books, theses, conference reports, and other scholarly sources on the topic 'Acquisition automatique des données – Capture de mouvements.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Dissertations / Theses on the topic "Acquisition automatique des données – Capture de mouvements"
Chaumeil, Anaïs. "Evaluation et développement de méthodes d'analyse du mouvement sans marqueurs à partir de vidéos." Electronic Thesis or Diss., Lyon 1, 2024. http://www.theses.fr/2024LYO10209.
Full textVideo-based markerless motion capture has benefitted, in the last few years, from the development of automatic point estimation methods, which are based on deep learning techniques. For motion analysis in biomechanics, these methods have numerous advantages, such as the possibility to analyse the movement without participant-worn equipment or outside of the laboratory. The goal of this thesis is thus to contribute to the evaluation and development of video-based markerless motion capture methods for applications in biomechanics. First, an existing video-based markerless motion capture method is evaluated for movements and kinematic parameters, rarely studied in the literature. Then, 2D keypoints estimated by automatic point estimation methods are characterized, and the influence of these characteristics on 3D point reconstruction is studied. Finally, a method using whole confidence heatmaps – which are obtained using automatic point estimation methods – to compute 3D kinematics is proposed and evaluated
Naert, Lucie. "Capture, annotation and synthesis of motions for the data-driven animation of sign language avatars." Thesis, Lorient, 2020. http://www.theses.fr/2020LORIS561.
Full textThis thesis deals with the capture, annotation, synthesis and evaluation of arm and hand motions for the animation of avatars communicating in Sign Languages (SL). Currently, the production and dissemination of SL messages often depend on video recordings which lack depth information and for which editing and analysis are complex issues. Signing avatars constitute a powerful alternative to video. They are generally animated using either procedural or data-driven techniques. Procedural animation often results in robotic and unrealistic motions, but any sign can be precisely produced. With data-driven animation, the avatar's motions are realistic but the variety of the signs that can be synthesized is limited and/or biased by the initial database. As we considered the acceptance of the avatar to be a prime issue, we selected the data-driven approach but, to address its main limitation, we propose to use annotated motions present in an SL Motion Capture database to synthesize novel SL signs and utterances absent from this initial database. To achieve this goal, our first contribution is the design, recording and perceptual evaluation of a French Sign Language (LSF) Motion Capture database composed of signs and utterances performed by deaf LSF teachers. Our second contribution is the development of automatic annotation techniques for different tracks based on the analysis of the kinematic properties of specific joints and existing machine learning algorithms. Our last contribution is the implementation of different motion synthesis techniques based on motion retrieval per phonological component and on the modular reconstruction of new SL content with the additional use of motion generation techniques such as inverse kinematics, parameterized to comply to the properties of real motions
Reverdy, Clément. "Annotation et synthèse basée données des expressions faciales de la Langue des Signes Française." Thesis, Lorient, 2019. http://www.theses.fr/2019LORIS550.
Full textFrench Sign Language (LSF) represents part of the identity and culture of the deaf community in France. One way to promote this language is to generate signed content through virtual characters called signing avatars. The system we propose is part of a more general project of gestural synthesis of LSF by concatenation that allows to generate new sentences from a corpus of annotated motion data captured via a marker-based motion capture device (MoCap) by editing existing data. In LSF, facial expressivity is particularly important since it is the vector of numerous information (e.g., affective, clausal or adjectival). This thesis aims to integrate the facial aspect of LSF into the concatenative synthesis system described above. Thus, a processing pipeline is proposed, from data capture via a MoCap device to facial animation of the avatar from these data and to automatic annotation of the corpus thus constituted. The first contribution of this thesis concerns the employed methodology and the representation by blendshapes both for the synthesis of facial animations and for automatic annotation. It enables the analysis/synthesis scheme to be processed at an abstract level, with homogeneous and meaningful descriptors. The second contribution concerns the development of an automatic annotation method based on the recognition of expressive facial expressions using machine learning techniques. The last contribution lies in the synthesis method, which is expressed as a rather classic optimization problem but in which we have included
Benchiheub, Mohamed-El-Fatah. "Contribution à l'analyse des mouvements 3D de la Langue des Signes Française (LSF) en Action et en Perception." Thesis, Université Paris-Saclay (ComUE), 2017. http://www.theses.fr/2017SACLS559/document.
Full textNowadays, Sign Language (SL) is still little described, particularly for what concerns the movement of articulators. Research on SL has focused on understanding and modeling linguistic properties. Few investigations have been carried out to understand the kinematics and dynamics of the movement itself and what it brings to understand the LS SL generated by models. This thesis deals with the analysis of movement in the French Sign Language LSF with a main focus on its production as well as its understanding by deaf people.Better understanding the movement in SL requires the creation of new resources for the scientific community studying SL. In this framework, we have created and annotated a corpus of 3D motion data from the upper body and face, using a motion capture system. The processing of this corpus made it possible to specify the kinematics of the movement in SL during the signs and the transitions.The first contribution of this thesis was to quantify to what extent certain classical laws, known in motor control, remained valid during the movements of SL, in order to know if the knowledge acquired in motor control could be exploited in SL.Finding relevant information of the movement that is crucial for understanding SL represented the second part of this thesis. We were basically interested to know which aspects of the movement of SL production models should be replicated as a priority. In this approach, we have examined to what extent deaf individuals, whether signers or not, were able to understand SL according to the amount of information available to them
Badreddine, Wafa. "Communication Protocols in Wireless Body Area Networks (WBAN)." Electronic Thesis or Diss., Sorbonne université, 2018. http://www.theses.fr/2018SORUS214.
Full textThe rapid advances in sensors and ultra-low power wireless communication has enabled a new generation of wireless sensor networks: Wireless Body Area Networks (WBAN). WBAN is a recent challenging area. There are several concerns in this area ranging from energy efficient communication to designing delay efficient protocols that support nodes dynamic induced by human body mobility. In WBAN tiny devices are deployed in/on or around a human body, are able to detect and collect the physiological phenomena of the human body (such as: EEG, ECG, SpO2, etc.), and transmit this information to a collector point (i.e Sink) that will process it, take decisions, alert or record. WBAN differs from typical large-scale wireless sensor networks WSN in many aspects: Network size is limited to a dozen of nodes, in-network mobility follows the body movements and the wireless channel has its specificities. Links have a very short range and a quality that varies with the wearer's posture. The transmission power is kept low to improve devices autonomy and reduce wearers electromagnetic exposition. Consequently, the effects of body absorption, reflections and interference cannot be neglected and it is difficult to maintain a direct link (one-hop) between the Sink and all WBAN nodes. Thus, multi-hop communication represents a viable alternative. In this work we investigate energy-efficient multi-hop communication protocols in WBAN. Our work is part of SMART-BAN Self-organizing Mobility Aware, Reliable and Timely Body Area Networks project. In order to evaluate our communication protocols described in the sequel in a specific WBAN scenario, we implemented them under the Omnet++ simulator that we enriched with the Mixim project and a realistic human body mobility and channel model issued from a recent research on biomedical and health informatics. We are interested in WBAN where sensors are placed on the body. We focus on two communication primitives: broadcast and converge-cast. For the broadcasting problem in WBAN, we analyze several broadcast strategies inspired from the area of DTN then we propose two novel broadcast strategies MBP: Mixed Broadcast Protocol and Optimized Flooding: -MBP (Mixed Broadcast Protocol): We proposed this strategy as a mix between the dissemination-based and knowledge-based approaches. -OptFlood (Optimized Flooding): This strategy takes into account the strengths and weaknesses of the basic strategy Flooding. Optimized Flooding is a revised version of Flooding whose purpose is to keep the good end-to-end delay given by Flooding while lowering energy consumption with the simplest way and the minimum cost. Additionally, we performed investigations of independent interest related to the ability of all the studied strategies to ensure the FIFO order consistency property (i.e. packets are received in the order of their sending) when stressed with various transmission rates. These investigations open new and challenging research directions. With no exception, the existing flat broadcast strategies register a dramatic drop of performances when the transmission rate is superior to 11Kb/s. There, we propose the first network-MAC layer broadcast protocol, CLBP, designed for multi-hop communication and resilient to human body postures and mobility. Our protocol is optimized to exploit the human body mobility by carefully choosing the most reliable communication paths in each studied posture. Moreover, our protocol includes a slot assignment mechanism that reduces the energy consumption, collisions, idle listening and overhearing. Additionally, CLBP includes a synchronization scheme that helps nodes to resynchronize with the Sink on the fly. Our protocol outperforms existing flat broadcast strategies in terms of percentage of covered nodes, energy consumption and correct reception of FIFO-ordered packets and maintains its good performances up to 190Kb/s transmission rates. [...]
Gisselbrecht, Thibault. "Algorithmes de bandits pour la collecte d’informations en temps réel dans les réseaux sociaux." Electronic Thesis or Diss., Paris 6, 2017. http://www.theses.fr/2017PA066655.
Full textIn this thesis, we study the problem of real time data capture on social media. Due to the different limitations imposed by those media, but also to the very large amount of information, it is not possible to collect all the data produced by social networks such as Twitter. Therefore, to be able to gather enough relevant information related to a predefined need, it is necessary to focus on a subset of the information sources. In this work, we focus on user-centered data capture and consider each account of a social network as a source that can be listened to at each iteration of a data capture process, in order to collect the corresponding produced contents. This process, whose aim is to maximize the quality of the information gathered, is constrained at each time step by the number of users that can be monitored simultaneously. The problem of selecting a subset of accounts to listen to over time is a sequential decision problem under constraints, which we formalize as a bandit problem with multiple selections. Therefore, we propose several bandit models to identify the most relevant users in real time. First, we study of the case of the so-called stochastic bandit, in which each user corresponds to a stationary distribution. Then, we introduce two contextual banditmodels, one stationary and the other non stationary, in which the utility of each user can be estimated more efficiently by assuming some underlying structure in the reward space. In particular, the first approach introduces the notion of profile, which corresponds to the average behavior of each user. On the other hand, the second approach takes into account the activity of a user at a given instant in order to predict his future behavior. Finally, we are interested in models that are able to take into account complex temporal dependencies between users, with the use of a latent space within which the information transits from one iteration to the other. Moreover, each of the proposed approaches is validated on both artificial and real datasets
Li, Jingting. "Facial Micro-Expression Analysis." Thesis, CentraleSupélec, 2019. http://www.theses.fr/2019CSUP0007.
Full textThe Micro-expressions (MEs) are very important nonverbal communication clues. However, due to their local and short nature, spotting them is challenging. In this thesis, we address this problem by using a dedicated local and temporal pattern (LTP) of facial movement. This pattern has a specific shape (S-pattern) when ME are displayed. Thus, by using a classical classification algorithm (SVM), MEs are distinguished from other facial movements. We also propose a global final fusion analysis on the whole face to improve the distinction between ME (local) and head (global) movements. However, the learning of S-patterns is limited by the small number of ME databases and the low volume of ME samples. Hammerstein models (HMs) are known to be a good approximation of muscle movements. By approximating each S-pattern with a HM, we can both filter outliers and generate new similar S-patterns. By this way, we perform a data augmentation for S-pattern training dataset and improve the ability to differentiate MEs from other facial movements. In the first ME spotting challenge of MEGC2019, we took part in the building of the new result evaluation method. In addition, we applied our method to spotting ME in long videos and provided the baseline result for the challenge. The spotting results, performed on CASME I and CASME II, SAMM and CAS(ME)2, show that our proposed LTP outperforms the most popular spotting method in terms of F1-score. Adding the fusion process and data augmentation improve even more the spotting performance
Ben, Messaoud Rim. "Towards efficient mobile crowdsensing assignment and uploading schemes." Thesis, Paris Est, 2017. http://www.theses.fr/2017PESC1031/document.
Full textThe ubiquity of sensors-equipped mobile devices has enabled people to contribute data via crowdsensing systems. This emergent paradigm comes with various applications. However, new challenges arise given users involvement in data collection process. In this context, we introduce collaborative sensing schemes which tackle four main questions: How to assign sensing tasks to maximize data quality with energy-awareness? How to minimize the processing time of sensing tasks? How to motivate users to dedicate part of their resources to the crowdsensing process ? and How to protect participants privacy and not impact data utility when reporting collected sensory data ? First, we focus on the fact that smart devices are energy-constrained and develop task assignment methods that aim to maximize sensor data quality while minimizing the overall energy consumption of the data harvesting process. The resulting contribution materialized as a Quality and Energy-aware Mobile Sensing Scheme (QEMSS) defines first data quality metrics then models and solves the corresponding optimization problem using a Tabu-Search based heuristic. Moreover, we assess the fairness of the resulted scheduling by introducing F-QEMSS variant. Through extensive simulations, we show that both solutions have achieved competitive data quality levels when compared to concurrent methods especially in situations where the process is facing low dense sensing areas and resources shortcomings. As a second contribution, we propose to distribute the assignment process among participants to minimize the average sensing time and processing overload com- pared to a fully centralized approach. Thus, we suggest to designate some participants to carry extra sensing tasks and delegate them to appropriate neighbors. The new assign- ment is based on predicting users local mobility and sensing preferences. Accordingly, we develop two new greedy-based assignment schemes, one only Mobility-aware (MATA) and the other one accounting for both preferences and mobility (P-MATA), and evaluate their performances. Both MATA and P-MATA consider a voluntary sensing process and show that accounting for users preferences minimize the sensing time. Having showing that, our third contribution in this thesis is conceived as an Incentives-based variant, IP-MATA+. IP-MATA+ incorporates rewards in the users choice model and proves their positive impact on enhancing their commitment especially when the dedicated budget is shared function of contributed data quality. Finally, our fourth and last contribution addresses the seizing of users privacy concerns within crowdsensing systems. More specifically, we study the minimization of the incurred privacy leakage in data uploading phase while accounting for the possible quality regression. That is, we assess simultaneously the two competing goals of ensuring queriers required data utility and protecting participants’ sensitive information. Thus, we introduce a trust entity to the crowdsensing traditional system. This entity runs a general privacy-preserving mechanism to release a distorted version of sensed data that responds to a privacy-utility trade-off. The proposed mechanism, called PRUM, is evaluated on three sensing datasets, different adversary models and two main data uploading scenarios. Results show that a limited distortion on collected data may ensure privacy while maintaining about 98% of the required utility level.The four contributions of this thesis tackle competing issues in crowdsensing which paves the way at facilitating its real implementation and aims at broader deployment