No description available.

The purpose of this research was to enhance the methods towards the development of an efficient three dimensional face recognition system. More specifically, one of our aims was to investigate how the use of curvature of the diagonal profiles, extracted from 3D facial geometry models can help the neutral face recognition processes. Another aim was to use a gender classifier employed on 3D facial geometry in order to reduce the search space of the database on which facial recognition is performed. 3D facial geometry with facial expression possesses considerable challenges when it comes face recognition as identified by the communities involved in face recognition research. Thus, one aim of this study was to investigate the effects of the curvature-based method in face recognition under expression variations. Another aim was to develop techniques that can discriminate both expression-sensitive and expression-insensitive regions for ii face recognition based on non-neutral face geometry models. In the case of neutral face recognition, we developed a gender classification method using support vector machines based on the measurements of area and volume of selected regions of the face. This method reduced the search range of a database initially for a given image and hence reduces the computational time. Subsequently, in the characterisation of the face images, a minimum feature set of diagonal profiles, which we call T shape profiles, containing diacritic information were determined and extracted to characterise face models. We then used a method based on computing curvatures of selected facial regions to describe this feature set. In addition to the neutral face recognition, to solve the problem arising from data with facial expressions, initially, the curvature-based T shape profiles were employed and investigated for this purpose. For this purpose, the feature sets of the expression-invariant and expression-variant regions were determined respectively and described by geodesic distances and Euclidean distances. By using regression models the correlations between expressions and neutral feature sets were identified. This enabled us to discriminate expression-variant features and there was a gain in face recognition rate. The results of the study have indicated that our proposed curvature-based recognition, 3D gender classification of facial geometry and analysis of facial expressions, was capable of undertaking face recognition using a minimum set of features improving efficiency and computation.

No description available.

Cette thèse s'intéresse au recalage de données issues de capteurs 3D et multispectraux pour l'étude du patrimoine.Lorsque l'on étudie ce type d'objet, il y a souvent peu de points saillants naturels entre ces jeux de données complémentaires. Par ailleurs, l'utilisation de mires optiques est proscrite.Notre problème est donc de recaler des données multimodales sans points caractéristiques.Nous avons développé une méthode de recalage basé sur le suivi des systèmes d'acquisition en utilisant des techniques issues de la photogrammétrie.Des simulations nous ont permis d'évaluer la précision de la méthode dans trois configurations qui représentent des cas typiques dans l'étude d'objets du patrimoine.Ces simulations ont montré que l'on peut atteindre une précision du suivi de 0.020 mm spatialement et 0.100 mrad angulairement en utilisant quatre caméras 5 Mpx lorsque l'on numérise une zone de 400 mm x 700 mm.La précision finale du recalage repose sur le succès d'une série de calibrations optiques et géométriques, ainsi que sur leur stabilité pour la durée du processus d'acquisition.Plusieurs tests ont permis d'évaluer la précision du suivi et du recalage de plusieurs jeux de données indépendants; d'abord seulement 3D, puis 3D et multispecrales.Enfin, nous avons étendu notre méthode d'estimation de la réflectance à partir des données multispectrales lorsque celles-ci sont recalées sur un modèle 3D. / The concern and interest of this PhD thesis is the registration of featureless 3D and multispectral datasets describing cultural heritage objects.In this context, there are few natural salient features between the complementary datasets, and the use of targets is generally proscribed.We thus develop a technique based on the photogrammetric tracking of the acquisition systems in use.A series of simulations was performed to evaluate the accuracy of our method in three configurations chosen to represent a variety of cultural heritage objects.These simulations show that we can achieve a spatial tracking accuracy of 0.020 mm and an angular accuracy of 0.100 mrad using four 5 Mpx cameras when digitizing an area of 400 mm x 700 mm. The accuracy of the final registration relies on the success of a series of optical and geometrical calibrations and their stability for the duration of the full acquisition process.The accuracy of the tracking and registration was extensively tested in laboratory settings. We first evaluated the potential for multiview 3D registration. Then, the method was used for to project of multispectral images on 3D models.Finally, we used the registered data to improve the reflectance estimation from the multispectral datasets

This Ph.D. thesis tackles the issue of sill and video object categorization. The objective is to associate semantic labels to 2D objects present in natural images/videos. The principle of the proposed approach consists of exploiting categorized 3D model repositories in order to identify unknown 2D objects based on 2D/3D matching techniques. We propose here an object recognition framework, designed to work for real time applications. The similarity between classified 3D models and unknown 2D content is evaluated with the help of the 2D/3D description. A voting procedure is further employed in order to determine the most probable categories of the 2D object. A representative viewing angle selection strategy and a new contour based descriptor (so-called AH), are proposed. The experimental evaluation proved that, by employing the intelligent selection of views, the number of projections can be decreased significantly (up to 5 times) while obtaining similar performance. The results have also shown the superiority of AH with respect to other state of the art descriptors. An objective evaluation of the intra and inter class variability of the 3D model repositories involved in this work is also proposed, together with a comparative study of the retained indexing approaches . An interactive, scribble-based segmentation approach is also introduced. The proposed method is specifically designed to overcome compression artefacts such as those introduced by JPEG compression. We finally present an indexing/retrieval/classification Web platform, so-called Diana, which integrates the various methodologies employed in this thesis

Made available in DSpace on 2016-06-02T20:00:55Z (GMT). No. of bitstreams: 1 5728.pdf: 4886967 bytes, checksum: 9a66021bc6eec94037ca58204cb8fb84 (MD5) Previous issue date: 2013-08-23 / Financiadora de Estudos e Projetos / Stormwater flooding in urban centers today is the direct result of ground impermeabilization and the ineffective dimensioning of stormwater drainage systems. In order to control these events, innovative low impact stormwater drainage facilities are used in Brazil, which are efficient in controlling sharp peak flows in urban watersheds without, however, affecting the aesthetic appearance of the landscape in these environments. The purpose of this study is to model and estimate the volume of water infiltrated into the soil where a compensatory stormwater drainage system is inserted, which can contribute to groundwater recharge. Modeling was performed by means of a physical storage method and storage variation in soil using a tensiometer, and in digital form using the Hydrus 2D/3D software package. The monitored drainage system and the digitally designed system consist of a water retention basin, a ditch or filter, and an infiltration trench. Groundwater recharge was estimated based on soil analyses, geophysical investigations, flow measurements, and rainfall data, in addition to data collected during 454 days of continuous monitoring of the system by means of tensiometers installed on-site, rain gauges, and the construction of bimodal soil water retention curves using the filter paper method. The moisture content at different depths, the zero water flow plane, and the volumes moving in the soil matrix were calculated based on the identification of the soil water content and matric potential, the hydraulic conductivity in saturated conditions at the monitored depths, and the soil s characteristic curves. The results of the Soil Water Retention Curves (SWRC) and soil moisture content were negligible starting from a depth of 80cm. However, the groundwater recharge rate estimated by the tensiometer method and by the Hydrus digital model was about 45% of the total volume of rainfall. / Atualmente, eventuais problemas de enchentes e inundações, são causas diretas da impermeabilização de áreas nos centros urbanos e do ineficaz dimensionamento da drenagem pluvial nas cidades. No intuito de se controlar estes eventos, equipamentos de drenagem de baixo impacto são inovadoras no Brasil e eficientes no controle de vazões de pico acentuadas em bacias urbanas, sem sobretudo, renunciar à estética e à paisagem nestes ambientes. Neste trabalho objetivou-se modelar e estimar o volume de água infiltrado no solo, onde um sistema de drenagem pluvial compensatória está inserido, que é capaz de contribuir com a recarga freática. A modelagem foi realizada por meio do método de armazenamento e variação da armazenagem no solo com utilização de tensiômetro e de forma digital utilizando-se do software Hydrus 2D/3D. O sistema de drenagem monitorado e o diagramado digitalmente são constituídos de uma bacia de retenção de água, uma vala ou filtro gramado e uma trincheira de infiltração. Na tentativa de estimar a recarga freática realizaram-se, conjuntamente, análises de solo, investigações geofísicas, medidas de vazão, coleta de dados de chuva, além do monitoramento contínuo em 454 dias do sistema através da instalação de tensiômetros no local, pluviógrafos, além da construção de curvas bimodais de retenção de água nos solos por meio do método do papel filtro. A partir da identificação dos potenciais mátricos do solo, da condutividade hidráulica em condições saturadas para as profundidades monitoradas e da obtenção das curvas características do solo, foi possível calcular o teor de umidade em diferentes profundidades, identificar o plano de fluxo zero e calcular os volumes em movimento na matriz do solo. Os resultados obtidos pela verificação das Curvas de Retenção de Água no Solo (CRAs) e umidades de campo foram pouco expressivos a partir de 80cm de profundidade, entretanto, a recarga estimada obtida pelo método dos tensiômetros e pelo modelo digital Hydrus ficou em torno de um valor médio de 45% do volume total precipitado.

Le travail présenté dans cette thèse s’inscrit dans le cadre de différents programmes de recherches sur la modélisation et la conception des machines synchrones à aimants permanents, pour des applications de transports terrestres. En effet, la tendance actuelle, que ce soit dans la traction ferroviaire, ou dans les véhicules électriques et/ou hybrides électriques, est d’utiliser de tels moteurs pour leurs grandes performances massiques et leur bon rendement. Cette tendance est également observable dans les grandes éoliennes à attaque directe. Toutefois, un inconvénient de ces machines est l’existence de pertes pouvant être importantes dans les aimants permanents. Ces pertes sont d’une part à l’origine d’une dégradation du rendement, mais elles peuvent aussi être à l’origine d’échauffements excessifs des aimants, avec des risques de désaimantation d’une part et des risques de décollement d’autre part. Dans ce contexte, l’objectif de notre travail de thèse a été d’établir de nouveaux modèles, plus précis, des pertes par courants de Foucault dans les aimants. Les modèles utilisés actuellement sont généralement des modèles bidimensionnels qui ignorent donc le fait que les courants de Foucault ont une répartition tridimensionnelle dans les aimants. Afin de valider le modèle développé, une maquette expérimentale a été mise en place. Dans un premier temps, le modèle de calcul des pertes par courants induits dans les pièces massives a été validé en utilisant une approche qui combine les résultats expérimentaux et ceux calculés analytiquement et numériquement. Ensuite différentes grandeurs globales et locales issues du modèle analytique ont été comparées aux éléments finis aussi bien en 2D qu'en 3D de même qu'aux mesures expérimentales. Les modèles et méthodes de calcul et de mesures proposés pourront être efficacement utilisés ultérieurement pour estimer les pertes par courants induits dans les aimants permanents de moteurs synchrone à aimants. / The work presented in this thesis is part of various research programs on the modeling and design of permanent magnet synchronous machines for land transport applications. Indeed, the current trend, whether in railway traction, or in electric and / or hybrid electric vehicles, is to use such engines for their high mass performance and good efficiency. This trend is also observable in large direct-attack wind turbines. However, one disadvantage of these machines is the existence of significant losses in the permanent magnets. These losses can cause a deterioration in efficiency, and also be the cause of excessive heating of the magnets, with risks of demagnetization and risks of. In this context, the aim of our thesis work was to establish new, more accurate models of eddy current losses in magnets. The models currently used are generally two-dimensional models which therefore ignore the three-dimensional distribution of eddy currents in the magnets. In order to validate the model developed, we set up an experimental test bench. In a first step, the calculation of induced current losses in massive pieces was validated using an approach that combines the experimental results with those calculated analytically and numerically. Then, different global and local quantities from the analytical model were compared to the finite elements in both 2D and 3D as well as experimental measurements. The models and methods of computation and measurements proposed can be effectively used later to estimate eddy current losses in the permanent magnets of synchronous magnet motors.

La fiabilité est l'un des principaux objectifs de tout fabricant de moteurs électriques. Au cours des 20 dernières années, l'apparition de l'électronique de puissance dans la partie alimentation es moteurs électriques de traction a donné la possibilité de contrôle de couple à toute vitesse de rotation. En utilisant ce type d'alimentation pour les moteurs asynchrones, ceux-ci sont devenus une alternative à coût réduit et très fiable aux moteurs synchrones utilisés auparavant dans la traction. L'inconvénient de l'électronique de puissance, c'est que des fronts dV/dt sont appliqués aux moteurs. Ces fronts excitent les couplages parasites dans les moteurs, de sorte que les courants vagabonds peuvent circuler dans le moteur, et passer par les roulements. Cela amène à une augmentation du taux de défaillance. L'objectif de ce travail est de développer un modèle analytique, utilisable à partir de la phase de conception du moteur, qui permet de prévoir le type et la quantité des courants qui sont susceptibles de se produire.

La mise en correspondance d'images multimodales est devenue un axe de recherche essentiel en traitement d'images quels que soient les domaines d'applications. Dans le cadre médical, de nombreuses modalités d'imagerie permettent d'explorer l'organisme mais la grande diversité des phénomènes physiques mis en jeu implique que les organes seront analysés selon des points de vue très différents, spécifiques à chacune des modalités. La fusion d'images multimodales permet de palier cette difficulté en exploitant la complémentarité des différentes modalités pour extraire des données nouvelles par la juxtaposition des données initiales et faisant ainsi de la mise en correspondance un élément d'analyse incontournable. La présente étude est orientée sur la recalage et la mise en correspondance de données issues de systèmes d'imageries projective et tomographique d'angiographie cérébrale. L'angiographie par rayons X (2D) est la référence actuelle pour l'étude des vaisseaux intracrâniens mais elle souffre d'un manque de flexibilité et d'absence d'informations tridimensionnelles. Le recalage et la mise en correspondance de cette modalité avec l'Imagerie par Résonance Magnétique permet, d'une part de compléter les données de l'image radiologique et, d'autre part, de valider les innovations technologiques apportées aux dernières générations d'imageurs. La méthode proposée se base sur un recalage automatique et rigide entre les deux modalités. L'originalité de l'approche réside en l'absence de marqueurs ou référentiels externes, basant ainsi le recalage uniquement sur les caractéristiques anatomiques individuelles. Les retombées médicales de ce travail sont multiples tant dans le domaine diagnostique que thérapeutique. A titres d'exemples, une information fine et précise de la vascularisation peut se révéler être décisive dans un cadre neurochirurgical et la mise en correspondance d'images acquises à des instants différents permet un suivi plus objectif d'une pathologie.

The goal of our studies was to test the effect of segmentation, cueing, and 2D/3D presentations to foster complex animation rocessing. The material was an upright mechanical piano system. We used an eye tracking system which provides information about learners' attention direction during the animation processing. We analyzed the effect of the format presentations and the eye movements during learning. Based on animation and multimedia research background, four experiments were conducted. In the first experiment the effect of the presentation of simplified external representations on learning from complex animation was investigated. Experiment two and three aimed at studying the cognitive processes involved in learning to complex mechanism system with new cueing techniques with spatial-temporal colored tokens. In the fourth experiment, 2D and 3D presentation of the same animated content were compared. Results of these experiments showed that (1) the use of a dual format presentation is better for developing a dynamic mental model from the animation than a single format, (2) the signaling strategies using cued tokens of dual format can guide efficiently learner's building of mental model and can enhance learner's comprehension of complex system, (3) a sequential format presentation followed by an animation format presentation helps the learner to understand the key stages of a dynamic process and to create a high quality mental model, (4) 3D animation presentation is better than 2D animation presentation to direct attention on relevant component of the animation. For depth processing, comprehension with 3D animation presentation is better than 2D animation format. Eye tracking measures provided insights into how people understood and learned complex systems. The analysis of eye tracking also contributed to the understanding of the subject's perceptual processing during learning. Overall, results provide a significant contribution in the field of learning with complex animation. Recommendations in the ergonomics area for the design of animations are proposed.