Estrelas anãs brancas são o último estágio evolutivo observável de mais de 95% das estrelas e também um resultado comum na evolução de estrelas binárias. O estudo de anãs brancas é, portanto, uma ferramenta poderosa na compreensão da evolução de estrelas simples e binárias, da função de massa inicial local, e da perda de massa após a sequência principal, levando-nos a uma melhor compreensão do histórico de formação e evolução estelar de diferentes populações. Para que esses estudos sejam possíveis, é necessária uma amostra grande e preferencialmente completa de anãs brancas, cobrindo todo o intervalo de parâmetros físicos. A maneira mais simples de obter isso é utilizando dados de grandes projetos de mapeamento. O Sloan Digital Sky Survey já permitiu o aumento do número de anãs brancas conhecidas em cinco vezes até o data release 10. Neste trabalho, estendemos a busca por anãs brancas aos novos objetos no data release 12, descobrindo 3 157 novas anãs brancas e 1 349 novas subanãs. Pela primeira vez, estendemos essa busca para log g < 6.5, correspondente a M < 0.3 M⊙. Anãs brancas abaixo desse limite de massa não podem formar-se em um tempo de Hubble; contudo, se a estrela é parte de um sistema de binárias próximas, a perda de massa pode ser tão intensa que a anã branca resultante tem massa inferior ao limite por evolução simples. Esses objetos são chamados anãs brancas de massa extremamente-baixa (ELMs, do inglês extremely-low mass white dwarfs) Elas têm Teff < 20 000 K e 5.0 . log g . 6.5 e espectros muito similares a estrelas A de sequência principal. Menos de cem são conhecidas, e a maioria dos objetos foi descoberta tendo em conta um critério de seleção tendencioso, que excluiu ELMs frias (Teff < 9 000 K) e com mais baixa massa (M . 0.15 M⊙), tornando difícil verificar modelos e compreender as propriedades das ELMs como classe. Nós identificamos milhares de objetos cujas propriedades físicas, temperatura efetiva e log g, estão no intervalo de produtos de evolução binária, como as ELMs. Nós os chamamos de sdAs, referindo-nos ao seu log g que as coloca abaixo da sequência principal e seu espectro dominado por hidrogênio. As sdAs parecem conter populações estelares sobrepostas, e encontramos que 7% são mais compatíveis com (pre-)ELMs do que com objetos de sequência principal. Obtivemos espectroscopia resolvida temporalmente para 26 objetos e pudemos confirmar que 15 estão em binárias. Um objeto faz parte de um sistema eclipsante, enquanto outro é pulsante — o oitavo membro da classe de ELMs pulsantes. Outros seis objetos também apresentaram pulsações em nossas observações, cinco desses estão na vizinhança da faixa de instabilidade das ELMs. Com estes resultados, aumentamos a população de ELMs por um fator de 20%, elevando a fração de ELMs de 4 para 20%, um valor que é consistente com as previsões de modelos evolucionários. / White dwarf stars are the final observable evolutionary state of over 95% of stars and also a common outcome in binary evolution. Therefore, studying white dwarfs is a powerful tool to understand both single and binary stellar evolution, local initial mass function, and post-main sequence mass loss, leading us to a better comprehension of the history of stellar formation and evolution of different stellar populations. In order to make this type of studies possible, a large and preferably complete sample of white dwarf stars, covering the whole range of physical parameters, is required. The simplest way to achieve that is to take advantage of data provided by large surveys. The Sloan Digital Sky Survey has already allowed the increase of the number of known white dwarf stars fivefold up to its data release 10. In this work, we extended the search for white dwarfs to the new objects in the data release 12, discovering 3 157 new white dwarfs and 1 349 new subdwarfs. For the first time, we have extended this search to log g < 6.5, corresponding to M < 0.3 M⊙. White dwarfs below this mass limit cannot be formed through single evolution within a Hubble time; however, if the star is part of a close binary system, the mass loss of the system may be so intense that the resulting white dwarf has mass below the single evolution limit. These objects are known as extremely-low mass white dwarfs (ELMs) They show Teff < 20 000 K and 5.0 . log g . 6.5 and spectra very similar to main sequence A stars. Less than a hundred of them are known, and most objects were discovered relying on biased selection criteria, that excluded cool (Teff < 9 000 K), lowermass (M . 0.15 M⊙) ELMs, making it difficult to validate the models and comprehend the properties of the ELMs as a class. We have identified thousands of objects whose physical properties, effective temperature and surface gravity, place them in the range of by-products of binary interaction such as the ELMs. We have called them sdAs, referring to their sub-main sequence log g and hydrogen dominated spectra. They seem to be composed of overlapping stellar populations, and we found that at least 7% are more likely ELMs or their precursors, the pre-ELMs, than main sequence stars. Obtaining time-resolved spectroscopy for 26 objects, we could confirm 15 to be in close binaries. One of them is also an eclipsing system, while another is a pulsator — the eighth member of the pulsating ELM class. Other six new pulsators were found as part of our follow-up, five of them in the vicinity of the ELM instability strip. With these results, we increase the population of ELMs by 20%, raising the fraction of cool ELMs from 4 to 20%, which is consistent with the predictions from the evolutionary models.

Dans cette thèse sont présentées deux méthodes visant à obtenir automatiquement une description tridimensionnelle des mouvements d'une main étant donnée une séquence vidéo monoculaire de cette main. En utilisant l'information fournie par la vidéo, l'objectif est de déterminer l'ensemble des paramètres cinématiques nécessaires à la description de la configuration spatiale des différentes parties de la main. Cet ensemble de paramètres est composé des angles de chaque articulation ainsi que de la position et de l'orientation globale du poignet. Ce problème est un problème difficile. La main a de nombreux degrés de liberté et les auto-occultations sont omniprésentes, ce qui rend difficile l'estimation de la configuration des parties partiellement ou totalement cachées. Dans cette thèse sont proposées deux nouvelles méthodes qui améliorent par certains aspects l'état de l'art pour ce problème. Ces deux méthodes sont basées sur un modèle de la main dont la configuration spatiale est ajustée pour que sa projection dans l'image corresponde au mieux à l'image de main observée. Ce processus est guidé par une fonction de coût qui définit une mesure quantitative de la qualité de l'alignement de la projection du modèle avec l'image observée. La procédure d'ajustement du modèle est réalisée grâce à un raffinement itératif de type descente de gradient quasi-newton qui vise à minimiser cette fonction de coût.Les deux méthodes proposées diffèrent principalement par le choix du modèle et de la fonction du coût. La première méthode repose sur un modèle de la main composé d'ellipsoïdes et d'une fonction coût utilisant un modèle de la distribution statistique de la couleur la main et du fond de l'image.La seconde méthode repose sur un modèle triangulé de la surface de la main qui est texturé est ombragé. La fonction de coût mesure directement, pixel par pixel, la différence entre l'image observée et l'image synthétique obtenue par projection du modèle de la main dans l'image. Lors du calcul du gradient de la fonction de coût, une attention particulière a été portée aux termes dûs aux changements de visibilité de la surface au voisinage des auto-occultations, termes qui ont été négligés dans les méthodes préexistantes.Ces deux méthodes ne fonctionnement malheureusement pas en temps réel, ce qui rend leur utilisation pour l'instant impossible dans un contexte d'interaction homme-machine. L'amélioration de la performance des ordinateur combinée avec une amélioration de ces méthodes pourrait éventuellement permettre d'obtenir un résultat en temps réel. / In this thesis we propose two methods that allow to recover automatically a full description of the 3d motion of a hand given a monocular video sequence of this hand. Using the information provided by the video, our aimto is to determine the full set of kinematic parameters that are required to describe the pose of the skeleton of the hand. This set of parameters is composed of the angles associate to each joint/articulation and the global position and orientation of the wrist. This problem is extremely challenging. The hand as many degrees of freedom and auto-occlusion are ubiquitous, which makes difficult the estimation of occluded or partially ocluded hand parts.In this thesis, we introduce two novel methods of increasing complexity that improve to certain extend the state-of-the-art for monocular hand tracking problem. Both are model-based methods and are based on a hand model that is fitted to the image. This process is guided by an objective function that defines some image-based measure of the hand projection given the model parameters. The fitting process is achieved through an iterative refinement technique that is based on gradient-descent and aims a minimizing the objective function. The two methos differ mainly by the choice of the hand model and of the cost function.The first method relies on a hand model made of ellipsoids and a simple discrepancy measure based on global color distributions of the hand and the background. The second method uses a triangulated surface model with texture and shading and exploits a robust distance between the synthetic and observed image as discrepancy measure.While computing the gradient of the discrepancy measure, a particular attention is given to terms related to the changes of visibility of the surface near self occlusion boundaries that are neglected in existing formulations. Our hand tracking method is not real-time, which makes interactive applications not yet possible. Increase of computation power of computers and improvement of our method might make real-time attainable.

Solar-type binaries with short orbital periods (P-close equivalent to 1-10. days; a less than or similar to 0.1. au) cannot form directly via fragmentation of molecular clouds or protostellar disks, yet their component masses are highly correlated, suggesting interaction during the pre-main-sequence (pre-MS) phase. Moreover, the close binary fraction of pre-MS stars is consistent with that of their MS counterparts in the field (F-close = 2.1%). Thus, we can infer that some migration mechanism operates during the early pre-MS phase (tau less than or similar to 5 Myr) that reshapes the primordial separation distribution. We test the feasibility of this hypothesis by carrying out a population synthesis calculation which accounts for two formation channels: Kozai-Lidov (KL) oscillations and dynamical instability in triple systems. Our models incorporate (1) more realistic initial conditions compared to previous studies, (2) octupole-level effects in the secular evolution, (3) tidal energy dissipation via weak-friction equilibrium tides at small eccentricities and via non-radial dynamical oscillations at large eccentricities, and (4) the larger tidal radius of a pre-MS primary. Given a 15% triple-star fraction, we simulate a close binary fraction from KL oscillations alone of F-close approximate to 0.4% after tau = 5. Myr, which increases to F-close 0.8% by tau = 5. Gyr. Dynamical ejections and disruptions of unstable coplanar triples in the disk produce solitary binaries with slightly longer periods P approximate to 10-100. days. The remaining approximate to 60% of close binaries with outer tertiaries, particularly those in compact coplanar configurations with log P-out (days) approximate to 2-5 (a(out) < 50 au), can be explained only with substantial extra energy dissipation due to interactions with primordial gas.

We image 104 newly identified low-mass (mostly M-dwarf) pre-main sequence (PMS) members of nearby young moving groups (YMGs) with Magellan Adaptive Optics (MagAO) and identify 27 stellar binaries with instantaneous projected separation as small as 40 mas. Fifteen were previously unknown. The total number of multiple systems in this sample including spectroscopic and visual binaries from the literature is 36, giving a raw stellar multiplicity rate of at least 35(-4)(+5)% for this population. In the separation range of roughly 1-300 au in which infrared AO imaging is most sensitive, the raw multiplicity rate is at least 24(-4)(+5)% for binaries resolved by the MagAO infrared camera (Clio). The M-star subsample of 87 stars yields a raw multiplicity of at least 30(-4)(+5)% over all separations, 21(-4)(+5)% for secondary companions resolved by Clio from 1 to 300 au (23(-4)(+5)% for all known binaries in this separation range). A combined analysis with binaries discovered by the Search for Associations Containing Young stars shows that stellar multiplicity fraction as a function of mass over the range of 0.2 to 1.2M(circle dot) appears to be linearly flat, in contrast to the field, where multiplicity increases with mass. After bias corrections are applied, the multiplicity of low-mass YMG members (0.2-0.6M(circle dot)) is in excess of the field. The overall multiplicity fraction is also consistent with being constant in age and across YMGs, which suggests that multiplicity rates for this mass range are largely set by 10 Myr without appreciable evolution thereafter.

Os condutores horizontais dos sistemas de coleta de esgotos sanitários, no Brasil e em quase todos os países do mundo, tem sido dimensionados considerando-se o escoamento em regime permanente, o que na realidade não ocorre. Sabendo-se que o escoamento no interior destes condutores acontece em regime não permanente, caracterizado pelo movimento de ondas, devido a aleatoriedade de uso dos aparelhos. Neste trabalho, apresentamos o estudo do escoamento não permanente, verificado, nos ramais e coletores dos sistemas de coleta de esgotos sanitários de edifícios residenciais. As equações diferenciais da continuidade e da quantidade de movimento, que regem este escoamento, são resolvidas pelo método das características. Apresentamos, também, os resultados e análise de várias simulações do programa de computador desenvolvido com base no método das características pelo professor Swaffield, da Universidade Heriot-Watt, Escócia. Os dados de entrada para estas simulações foram obtidos a partir de ensaios laboratoriais de algumas configurações de ambientes sanitários brasileiros, também apresentados neste trabalho. Os resultados destas simulações mostram que a consideração do escoamento em regime não permanente, possibilita a redução dos diâmetros dos ramais e coletores, e consequentemente, a redução dos custos destes sistemas. / Usually horizontal waste pipes of Building Drainage Systems, in Brazil and almost all of countries in the world, are dimensioned considering a steady flow, that does not exist. The random use of the sanitary appliances connected to the Drainage Systems of the building is the responsible for the unsteady flow. We present a study of unsteady partially filled pipeflow for drains and sewers. The differential equations of continuity and motion are solved via the method of characteristics. We present, also, the results and analysis of several simulations using a computational program developed by Prof. Swaffield-Heriot - Watt University, Scotland. The data entry were obtained through laboratorial tests which considered brazilian bathroom configurations, presented in this work. The results show us that the consideration of unsteady flow conduct to the reductions of horizontal waste pipes diameters, consequently to the reduction of the Building Drainage Systems costs.

O aumento da demanda por água potável em função do crescimento populacional, vem sendo objeto de estudo de diversos pesquisadores que buscam meios de otimizar o uso desse recurso. Desse modo, ao longo dos anos os equipamentos sanitários foram sendo aperfeiçoados para tornarem-se mais eficientes. Uma preocupação é o impacto nos ramais e coletores de sistemas prediais de esgotos sanitários, uma vez que os componentes economizadores ao demandarem menos água em sua operação, retornam vazões menores para a rede predial de esgoto sanitário e, consequentemente, geram ondas menores que impactam negativamente na autolimpeza dos condutos horizontais. Nesse sentido, o objetivo da pesquisa foi investigar, em laboratório, as vazões e profundidades da lâmina d´água do escoamento em ramais e coletores do sistema de esgotos sanitários de edifícios residenciais unifamiliares, tendo em vista a redução de diâmetros de 100 mm para 75 mm. A pesquisa foi realizada em laboratório vertical, onde foi montada uma configuração típica de um banheiro residencial, localizada no pavimento superior. Foram utilizadas bacias sanitárias com volume nominal de descarga de 6 L e 4,8 L, um chuveiro com vazão constante de 0,20 L/s e um lavatório com vazão constante de 0,15 L/s. As bacias sanitárias foram ensaiadas para calibração, segundo a NBR 15097 (ABNT, 2011) e ambas obtiveram resultados satisfatórios em todos os requisitos dessa norma. Foram variados os diâmetros dos ramais e coletores prediais de 100 mm para 75 mm, com e sem a contribuição da vazão em regime permanente de 0,35 L/s proveniente de um chuveiro e um lavatório e as declividades variadas em 0%, 1% e 2% nos ramais e 1%, 2% e 3% nos coletores de esgoto. Avaliou-se a influência dos parâmetros declividade, volume de descarga, presença ou não de vazão em regime permanente e diâmetro da tubulação, sobre a velocidade de escoamento da água. Concluiu-se que a redução dos diâmetros dos ramais e coletores de esgoto e a redução do volume de descarga de 6 L para 4,8 L contribuíram para o aumento da velocidade da onda sob as condições de estudo em todo o trecho ensaiado, ou seja, no trecho compreendido entre a bacia sanitária e o tubo de queda e ao longo de todo comprimento de 3 m do coletor de esgoto predial. Não foi estudada a influência da variação dos parâmetros sobre a velocidade de onda em trechos além do coletor de esgoto predial. Sob as condições de estudo, observou-se que a declividade não influencia, de maneira geral, no desempenho do sistema nos trechos ensaiados. A presença de uma vazão em regime permanente contribui para amortizar a energia da onda de descarga da bacia sanitária no ramal de descarga e consequentemente, reduz a velocidade de escoamento da água, o que pode contribuir para a piora do requisito de autolimpeza neste trecho da tubulação. No coletor de esgoto, não foi possível concluir que a presença de vazão em regime permanente tem impacto sobre a velocidade de escoamento da água. Por fim, conclui-se ser possível a redução dos diâmetros dos ramais e coletores do sistema predial de esgoto sanitário. / The increase in the demand for drinking water due to the population growth has been the object of study of several researchers who seek ways to optimize the use of this resource. In this way, over the years sanitary equipment has been improved to become more efficient. One concern is the impact on drains and main drains of building drainage system, since the economizing components, while using less water in their operation, return smaller flows to the sanitary sewage network and consequently lead to the formation of smaller waves that contribute to the self-cleaning of pipes. In this sense, the objective of the research was to investigate, in the laboratory, the flows and the depths of the flow in drains and main drains of residential building drainage system, in order to reduce diameters from 100 mm to 75 mm. The research was carried out in a vertical laboratory, where a typical configuration of a residential bathroom located on the second floor of a two story building was set up. A toilet with a nominal discharge volume of 6 L and 4,8L, a shower with a constant flow rate of 0.20 L/s and a lavatory with a constant flow rate of 0.15 L/s were used. The toilets were tested according to NBR15097 (ABNT, 2011) for calibration and both obtained satisfactory results in all the requirements of this standard. The diameters of the drains and main drains were varied from 100 mm to 75 mm with and without the contribution of the continuous flow rate of 0.35 L/s from a shower and lavatory and the slopes varying in 0%, 1% and 2% in the drains and 1%, 2% and 3% in the main drain. The influence of the parameters like slope, discharge volume, presence or not of continuous flow rate and pipe diameter, on the water flow velocity were evaluated. It was concluded that the reduction of the diameters of the drains and main drains and the reduction of the volume of discharge from 6 L to 4.8 L contribute to the increase of the velocity of the wave in the whole parts of the drain and main drain. It was not studied the influence of the variation of the parameters on the wave velocity on parts beyond the main drain. It was observed that the slope variation does not influence in general performance of the system. The presence of a continuous flow rate contributes to amortize the discharge wave energy from the toilet at the drains and consequently reduces the flow velocity of the water, which may contribute on deterioration of the self-cleaning requirements in this section of the pipeline. In the main drain, it was not possible to conclude that the presence of permanent flow has an impact on the flow velocity of the water.

Modélisation du comportement dynamique des rotors d'hélicoptèresL'objectif de la thèse est l'étude et la modélisation du comportement dynamique et aérodynamique du rotor de l'hélicoptère en considérant à la fois les forces d'inertie et les forces aérodynamiques et en tenant compte des déformations élastiques des pales. L'algorithme de couplage proposé permet d'effectuer le calcul transitoire avec échange de données entre les solveurs fluide et structure à chaque pas de temps.La particularité de cette étude est l'utilisation du modèle aérodynamique de la ligne active, qui représente les forces de pale appliquées au fluide par des termes sources. Ces termes sources sont répartis dans les cellules de maillage à l'emplacement de la pale. Ainsi, la rotation, la torsion et le battement de la pale peuvent être représentés sans aucune déformation du maillage. Un avantage de la ligne active est que la simulation utilise un nombre réduit de nœuds, car des conditions aux limites «lois des parois» ne doivent pas être modélisées.Le cas d'un petit rotor expérimental d'hélicoptère est étudié en vol d'avancement. Les solveurs de fluide et de structure sont couplés pour calculer le comportement aérodynamique et dynamique du rotor. Pour ce faire, un algorithme de couplage faible en série décalé est appliqué. Les calculs itératifs sont contrôlés par un code spécialement conçu. Au début de chaque itération, le code calcule et répartit les termes sources dans le domaine fluide. A la fin du pas de temps, le code exécute le solveur de calcul de structure pour calculer un seul pas de temps. Ce solveur calcule le déplacement de la pale sous l'effet des forces aérodynamiques, élastiques et d'inertie et renvoi les résultats au solveur fluide. Les déplacements de la pale calculés servent de référence pour le solveur fluide au pas de temps suivant, pour distribuer les termes sources. Le calcul s'arrête lorsque le critère de convergence est vérifié.Afin de valider le cas simulé, des expérimentations sont réalisées en soufflerie. La puissance et la poussée aérodynamique du rotor sont mesurées. La Vélocimétrie par images de particules (PIV) est utilisée pour obtenir le champ de vitesse autour du rotor. Les mesures PIV à phase bloqué dans des plans azimutaux ont permis de reconstituer le champ d'écoulement 3D. La comparaison entre les résultats numériques et les expériences montre un bon accord et permet de valider la méthode de couplage proposée. / MODELING THE DYNAMIC BEHAVIOR OF HELICOPTER ROTORThe aim of the thesis is the investigation and modeling of dynamic and aerodynamic behavior of helicopter rotor considering both inertial and aerodynamic forces and taking into account the elastic deformation of the blades. The proposed coupling algorithm allows the transient calculations with data exchange between the fluid and structure solvers at each time step.The particularity of this research is the use of an actuator line aerodynamic model, which represents the blade forces applied to the fluid as source terms. These source terms are distributed in the grid cells where the blade is located. Thus the rotation, flapping and torsion of the blade can be represented without any grid deformation. An advantage of the actuator line is that the simulation uses a reduced number of nodes, because the “wall” boundary conditions do not need to be modeled.The case of small experimental helicopter rotor is studied in forward flight. The fluid and structure solvers are coupled to calculate aerodynamic and dynamic behavior of the rotor. For this purpose, a loosely coupling serial staggered algorithm is applied. The iterative calculations are controlled by specially developed code. At the beginning of each iteration, this code calculates and distributes the source terms in the fluid domain. At the end of the time step, the code runs the structural solver to execute a single time step. This solver calculates the blade displacement under aerodynamic, elastic and inertial forces, and the results are returned to the fluid solver. The calculated blade displacements serve as reference in the next fluid step to distribute the source terms. The calculation stops when the convergence criteria are met.In order to validate the simulated case, measurements are carried on in the wind tunnel. The power and aerodynamic thrust of the rotor are measured. Particle Image Velocimetry (PIV) is used to obtain the velocity field around the rotor. Phase locked measurement in azimuth planes enabled to reconstruct 3D flow field. The comparison between numerical results and experiments shows good agreement and permits to validate the proposed coupling method.

Le travail de thèse exposé dans ce manuscrit s'intéresse à la simulation d'un mouvement de manipulation d'objet, plus particulièrement à la saisie suivie de la rotation d'une sphère selon un axe fixe. Le but ici est, à partir d'une analyse de mouvements réels, d'être capable de proposer un algorithme de simulation reproduisant des mouvements comparables aux données expérimentales, en fournissant en entrée le moins d'informations possibles et en essayant d'introduire de la variabilité dans le mouvement simulé. 12 sujets volontaires ont participé à l'expérimentation. Nous leur avons demandé de saisir et de tourner une sphère de 60mm de diamètre avec des amplitudes allant de 45° à 360°, dans les deux sens. L'analyse de ces données expérimentales, complétée par des simulations de l'effet des limites articulaires sur le mouvement nous ont permis d'étudier plusieurs hypothèses sur le contrôle de mouvement telles que l'hypothèse de confort final (« end-state comfort ») et le principe de travail minimum. Une des conclusions majeures est que l'anticipation posturale au moment de saisie semble être expliquée davantage par le confort en fin de mouvement qu'au moment de saisie. Basé sur ces observations, nous proposons un algorithme de simulation avec pour nouveautés la prise en compte de l'espace de mouvements admissibles par les limites articulaires ainsi que l'introduction de la variabilité au choix de la posture de saisie. Les premiers résultats semblent être en accord en grande partie avec les observations expérimentales donnant une base de travail pour aller vers des outils de simulation se rapprochant de plus en plus vers un comportement "humain" / The thesis work presented in this manuscript focuses on the simulation of an handling motion, more specifically on the grasp followed by the rotation of a sphere along a fixed single axis. The aim here is, from the analysis of actual motions, to be able to propose an simulation algorithm reproducing motions comparable to experimental data, with the less input as possible and trying to introduce some variability into the simulated motion. 12 volunteers participated to the experiment. Subjects were asked to grasp and turn a sphere of 60mm of diameter. Amplitudes of rotations were ranged from 45° to 360°, in both directions. Experimental data analysis, completed with some simulations of the effect of joint limits on motion, allowed us to investigate several motion control hypothesis as the end-state comfort hypothesis or the minimum work principle. One of the main conclusions is that postural anticipation when grasping seems to be more explained by the comfort at the end of the motion than when grasping. Based on these observations, we proposed a simulation algorithm being original by the way of how it takes into account possible motions allowed by joint limits and by the introduction of variability into the simulated grasp posture. The first results seem to follow most part of the experimental observations giving a strong basis to go towards simulation tools that will come closer to a “human” behavior

La biométrie est une alternative qui se base sur l'identification des personnes à partir de leurs caractéristiques physiques (empreinte digitale, forme de la main, empreinte palmaire) et/ou comportementales (voix, signature dynamique). La biométrie tend à réaliser deux buts importants dans notre vie courante. Le premier but est de réaliser la sécurité en éliminant le doute sur l'identité d'une personne et le second but est de faciliter l'identification des individus. En effet, cette méthode d'identification est de plus en plus préférée par rapport aux méthodes traditionnelles impliquant les mots de passe et les badges. Les travaux de recherche de cette thèse s'inscrivent dans le cadre de la reconnaissance de personnes à l'aide de la biométrie de la main. L'objectif principal est de concevoir un système biométrique multimodal basé sur la fusion de la forme de la main et de l'empreinte palmaire.La première partie de cette thèse propose un nouveau système uni-modal de vérification de la forme de la main. En effet, ce système est basé d'une part, sur la détection du meilleur ensemble des points-clés localisés sur le contour de la main pour adopter la description SIFT (Scale Invariant Feature Transform). D'autre part, un raffinement de correspondance, basé région et apparence de la main est proposé, afin de raffiner autant que possible les points-clés faussement matchés.Tandis que la deuxième partie consiste à proposer un nouveau système d'identification palmaire. En effet, la méthode de représentation parcimonieuse est adoptée afin de décrire le trait biométrique de l'empreinte palmaire. Elle est basée sur l'extraction de descripteurs SIFT de chacun des points-clés détectés. Notre troisième partie concerne la proposition de différentes méthodes de fusion multi-types de la multi modalité, comprenant la fusion multi-représentation, la fusion multi-biométrique et la fusion multi-instance. En effet, la fusion multi-représentation est basée sur la combinaison de descripteurs SIFT et les caractéristiques géométriques de la main au niveau des scores, pour la vérification de la forme de la main. La fusion multi-biométrique est basée sur la combinaison des deux modalités biométriques à savoir la forme de la main et l'empreinte palmaire, au niveau des caractéristiques et de la décision. Par contre, la fusion multi-instance est basée sur la combinaison des empreintes palmaires droite et gauche, au niveau du rang.Ces différentes méthodes de fusion ont prouvé leur efficacité en obtenant de meilleurs taux de reconnaissance, qui sont compétitifs par rapport à d'autres approches multimodales de la biométrie de la main. / Biometry is a technology which is based on the personal identification using their physical features (fingerprint, hand geometry, palmprint) and/or behavioral features (voice, dynamic signature). Biometry aims to achieve two important goals in our current life. The first one is to ensure security by eliminating doubt regarding the identity of a person and the second one is to facilitate the identification of individuals. Indeed, this method of identification is increasingly preferred over traditional methods including passwords and badges. The research works of this thesis talk about the personal recognition using hand biometrics. The main objective is to design a multimodal biometric system based on the fusion of hand shape and palmprint modalities.Our first part is to propose a new unimodal biometric system for hand shape verification. In fact, this system is based firstly, on the detection of the best set of keypoints located on the contour of the hand for further SIFT (Scale Invariant Feature Transform) description. On the other hand, a matching refinement based hand region and appearance is proposed in order to refine as much as possible false matched keypoints.Our second part consists in the proposition of a new palmprint identification system. In fact, the sparse representation method is adopted in order to describe the palmprint biometric trait. It is based on the extraction on SIFT descriptors for each detected keypoint.Our third part concerns the proposition of multi-type fusion methods for multimodality, including the multi-representation fusion, the multi-biometric fusion and the multi-instance fusion. Indeed, the multi-representation fusion method is based on the combination of SIFT descriptors and geometrical features of the hand, at score level. The multi-biometric fusion method is based on the fusion of hand shape and palmprint modalities, at feature and decision levels. On the other hand, the multi-instance fusion method is based on the combination of left and right palmprints, at rank level.These different methods of fusion have proven their effectiveness by achieving encouraging recognition rates that are competitive to other popular multimodal hand biometric approaches.

La main est l’organe de préhension chez l’homme qui lui permet de manipuler des objets de tailles et de formes diverses et variées. Elle lui permet également d’effectuer des tâches, parfois d’une grande complexité, avec force ou avec une grande finesse. Il s’agit d’un outil d’une importance cruciale au quotidien, tant sur le plan domestique que sur le plan professionnel, et la perte de fonctionnalité de la main peut rapidement devenir un handicap pour certaines personnes. Bien qu’il existe encore peu d’études sur la capture et l’analyse des mouvements de la main dans la littérature scientifique, de nombreux domaines comme la médecine, l’ergonomie, le sport, la robotique, la réalité virtuelle ou les jeux vidéo s’y intéressent de plus en plus, afin de permettre aux personnes d’exploiter au mieux les fonctionnalités de la main tout en la préservant, ou de créer des interfaces homme-machine afin que celui-ci puisse interagir avec des robots ou en environnement virtuel. L’objectif de cette thèse a été de développer une méthode d’analyse cinématique in vivo et personnalisée afin de contribuer à une meilleure compréhension des mouvements de la main de l’homme. Une première partie du travail consiste à mettre en place un protocole de capture du mouvement de la main sur des sujets masculins et féminins de différentes tranches d’âges, allant de 20 à 50 ans. Les captures de mouvement ont été effectuées à l’aide d’un système optoélectronique et de 56 marqueurs passifs collés sur la peau de la main du sujet. Deux types de mouvements ont été capturés : les mouvements fonctionnels en flexion-extension et en abduction-adduction de la main, et les mouvements de prise d’objets cylindriques et sphériques. Les captures de mouvement sont ensuite labellisées afin d’identifier les différents marqueurs et de pouvoir en extraire les trajectoires. La deuxième partie du travail consiste à développer une méthode d’analyse cinématique des mouvements externes de la main à partir des trajectoires des marqueurs et de la valider à l’aide d’un modèle in silico. En plus d’estimer les paramètres cinématiques avec précision dans le cas où les trajectoires des marqueurs sont parfaites, l’évaluation des méthodes fonctionnelles montre que les mouvements de la main sont assimilables à des mouvements plans, circulaires ou sphériques en fonction de l’articulation étudiée. La construction des repères fonctionnels à partir des paramètres cinématiques pour chaque segment de la main permet ensuite de décrire les mouvements de rotation des articulations à l’aide des angles de Cardan. Les courbes des angles de Cardan obtenues à partir des trajectoires des marqueurs du modèle in silico permettent de valider la méthode de décomposition développée pour l’ensemble des articulations, à l’exception des articulations trapézo-métacarpienne (TMC) et métacarpo-phalangienne (MCP1) du pouce qui sont plus difficiles à analyser. La dernière partie de ce travail de thèse consiste à analyser les mouvements fonctionnels et les mouvements de prise d’objets à partir des captures expérimentales. Les courbes des angles de Cardan obtenues à partir des mouvements fonctionnels correspondent globalement à celles obtenues dans la littérature, sauf dans le cas de l’articulation TMC. De plus, ces courbes montrent que les rotations articulaires ne s’effectuent pas uniquement autour d’un axe, mais autour d’un axe de rotation dominant et d’un ou deux axes de rotation secondaires. Toutefois, les courbes des angles de Cardan des rotations autour des axes secondaires ne correspondent pas toujours à celles présentées dans la littérature. Bien que peu de mouvements de prise d’objets aient pu être analysés, certaines corrélations intéressantes ont pu être trouvées entre les postures de la main et la géométrie des objets manipulés, notamment au niveau des articulations MCP et IPD. / The human hand is a prehensile organ which allows people to handle objects with various sizes and shapes. It is wonderful tool that can be used to perform different simple or complex tasks with strength or great dexterity. It is also a crucial tool in the daily life, both at home and in the workspace, and loss of hand functionality may quickly become disabling for some people. There are few studies in the literature. However, motion capture and kinematic analysis of the hand is becoming more and more of an interest in different areas such as medicine, ergonomics, sport, robotics, virtual reality and video games. Results from these studies have improved knowledge about skills of the hand and how to preserve them. The studies have also improved interactions between people and computers in order to command robots or to progress in virtual reality. The aim of the thesis was to develop methods for an in vivo and subject-specific kinematic analysis in order to contribute to the improvement of knowledge about the human hand motion. A first part of this thesis was to develop a protocol for the motion capture of the hand for male and female subjects aged from 20 to 50 years old. The motion capture was performed using an optoelectronic system with passive markers glued on the skin of the hand. Two sorts of movements were captured. Firstly, functional movements like flexion-extension and abduction-adduction. Secondly, prehensile movements of cylindrical and spherical objects. Then, markers on the motion captures were identified in order to extract their trajectories. The second part of this thesis consisted in the development of a method for the kinematic analysis of external hand movements from the marker trajectories. Validation of this method was achieved using a model of the hand developed in silico. Since no noise was added to the marker trajectories in the silico model; kinematic parameters were estimated with precision. Moreover, assessment of the functional methods showed that the hand motions can be approximated by a plane, a circular arc or a spherical cap depending on the joint studied. After constructing the functional coordinate systems for each segment of the hand using the joint kinematic parameters, it was possible to decompose any joint rotation into three Cardan angles. This decomposition method was validated using the marker trajectories of the hand model, except for the trapeziometacarpal (TMC) and the metacarpophalangeal (MCP1) joints of the thumb which are more difficult to study. The last part of this thesis consisted in the analysis of the functional and the prehensile movements from the motion captures. The curves of the Cardan angles obtained from the functional movements are similar to those presented in the literature for all of the joints, except for the TMC joint. It was also noticed that the joint rotations do not occur around only one axis, but around one dominant axis and one or two secondary axes. However, some differences between the curves of the Cardan angles around the secondary axes obtained in this thesis and those presented in the literature were noticed for some joints. Despite only few prehensile grasps were analyzed, some interesting correlations were also found between the hand shape and the objects grasped, more particularly at the metacarpophalangeal (MCP) and the distal interphalangeal (DIP) joints.