The increased concentration of greenhouse gases is associated with the global warming in the lower troposphere. For over twenty years, the space physics community has studied a hypothesis of global cooling in the thermosphere, attributable to greenhouse gases. While the global temperature increase in the lower troposphere has been relatively well established, the existence of global changes in the thermosphere is still under investigation. A central difficulty in reaching definite conclusions is the absence of data with sufficiently long temporal and sufficiently broad spatial coverage. Time series of data that cover several decades exist only in a few separated regions. The space physics community has struggled to combine the information contained in these data, and often contradictory conclusions have been reported based on the analyses relying on one or a few locations. To detect global changes in the ionosphere, we present a novel statistical methodology that uses all data, even those with incomplete temporal coverage. It is based on a new functional regression approach that can handle unevenly spaced, partially observed curves. While this research makes a solid contribution to the space physics community, our statistical methodology is very flexible and can be useful in other applied problems.

The increased concentration of greenhouse gases is associated with the global warming in the lower troposphere. For over twenty years, the space physics community has studied a hypothesis of global cooling in the thermosphere, attributable to greenhouse gases. While the global temperature increase in the lower troposphere has been relatively well established, the existence of global changes in the thermosphere is still under investigation. A central difficulty in reaching definite conclusions is the absence of data with sufficiently long temporal and sufficiently broad spatial coverage. Time series of data that cover several decades exist only in a few separated regions. The space physics community has struggled to combine the information contained in these data, and often contradictory conclusions have been reported based on the analyses relying on one or a few locations. To detect global changes in the ionosphere, we present a novel statistical methodology that uses all data, even those with incomplete temporal coverage. It is based on a new functional regression approach that can handle unevenly spaced, partially observed curves. While this research makes a solid contribution to the space physics community, our statistical methodology is very flexible and can be useful in other applied problems.

Philosophiae Doctor - PhD / This research investigated the application of recurrent neural networks (RNNs) for recognition of facial expressions based on facial action coding system (FACS). Support vector machines (SVMs) were used to validate the results obtained by RNNs. In this approach, instead of recognizing whole facial expressions, the focus was on the recognition of action units (AUs) that are defined in FACS. Recurrent neural networks are capable of gaining knowledge from temporal data while SVMs, which are time invariant, are known to be very good classifiers. Thus, the research consists of four important components: comparison of the use of image sequences against single static images, benchmarking feature selection and network optimization approaches, study of inter-AU correlations by implementing multiple output RNNs, and study of difference images as an approach for performance improvement. In the comparative studies, image sequences were classified using a combination of Gabor filters and RNNs, while single static images were classified using Gabor filters and SVMs. Sets of 11 FACS AUs were classified by both approaches, where a single RNN/SVM classifier was used for classifying each AU. Results indicated that classifying FACS AUs using image sequences yielded better results than using static images. The average recognition rate (RR) and false alarm rate (FAR) using image sequences was 82.75% and 7.61%, respectively, while the classification using single static images yielded a RR and FAR of 79.47% and 9.22%, respectively. The better performance by the use of image sequences can be at- tributed to RNNs ability, as stated above, to extract knowledge from time-series data. Subsequent research then investigated benchmarking dimensionality reduction, feature selection and network optimization techniques, in order to improve the performance provided by the use of image sequences. Results showed that an optimized network, using weight decay, gave best RR and FAR of 85.38% and 6.24%, respectively. The next study was of the inter-AU correlations existing in the Cohn-Kanade database and their effect on classification models. To accomplish this, a model was developed for the classification of a set of AUs by a single multiple output RNN. Results indicated that high inter-AU correlations do in fact aid classification models to gain more knowledge and, thus, perform better. However, this was limited to AUs that start and reach apex at almost the same time. This suggests the need for availability of a larger database of AUs, which could provide both individual and AU combinations for further investigation. The final part of this research investigated use of difference images to track the motion of image pixels. Difference images provide both noise and feature reduction, an aspect that was studied. Results showed that the use of difference image sequences provided the best results, with RR and FAR of 87.95% and 3.45%, respectively, which is shown to be significant when compared to use of normal image sequences classified using RNNs. In conclusion, the research demonstrates that use of RNNs for classification of image sequences is a new and improved paradigm for facial expression recognition.

As a core problem in video analysis, action recognition is of great significance for many higher-level tasks, both in research and industrial applications. With more and more video data being produced and shared daily, effective automatic action recognition methods are needed. Although, many deep-learning methods have been proposed to solve the problem, recent research reveals that single-stream, RGB-based networks are always outperformed by two-stream networks using both RGB and optical flow as inputs. This dependence on optical flow, which indicates a deficiency in learning motion, is present not only in 2D networks but also in 3D networks. This is somewhat surprising since 3D networks are explicitly designed for spatio-temporal learning. In this thesis, we assume that this deficiency is caused by difficulties associated with learning from videos exhibiting strong temporal variations, such as sudden motion, occlusions, acceleration, or deceleration. Temporal variations occur commonly in real-world videos and force a neural network to account for them, but often are not useful for recognizing actions at coarse granularity. We propose a Dynamic Equilibrium Module (DEM) for spatio-temporal learning through adaptive Eulerian motion manipulation. The proposed module can be inserted into existing networks with separate spatial and temporal convolutions, like the R(2+1)D model, to effectively handle temporal video variations and learn more robust spatio-temporal features. We demonstrate performance gains due to the use of DEM in the R(2+1)D model on miniKinetics, UCF-101, and HMDB-51 datasets.

No description available.

No description available.

No description available.

Découvrir les environs tout en marchant à l'aide d'un appareil mobile est désormais possible en associant les nouvelles technologies telles que la réalité augmentée (RA) et les services basés sur la localisation. De nombreuses applications mobiles ont été conçues à cet effet récemment. Les environs d'un utilisateur en mouvement sont présentés sur l'écran de son appareil mobile en utilisant les données de géo-localisation et elle peut interagir avec son environnement grâce à une couche d'information numérique superposée sur la scène capturée. Toutefois, la quantité d'informations géo-référencées produites et accessibles augmentant chaque jour, il y a un besoin croissant d'une approche générique qui vise à présenter l'information de manière structurée et adaptée aux différents domaines d'application et à la diversité des profils des utilisateurs.Notre travail se concentre sur la définition d'un modèle de données pour les objets 3D destiné à être intégré dans des applications de RA mobiles. Le modèle de données que nous proposons donne accès à l'information spatio-temporelle et thématique disponible sur les objets du monde réel. Alors que la réalité augmentée semble être bien adaptée pour la recherche d'informations à partir de la localisation, la plupart des approches basées sur la RA se concentre sur des scénarios spécifiques à un domaine d'application. D'autre part, le nuage des données ouvertes et liées (LOD Cloud en anglais) est un ensemble en constante extension de données structurées et publiées sur le Web. Il contient donc de grandes quantités d'informations à références spatio-temporelles qui peuvent être exploitées par les applications mobiles et présentées en utilisant la RA pour favoriser l'interaction. Dans notre recherche, ce nuage du Web des Données sert de source principale d'information au cours de l'expérience de découverte des environs à l'aide de la RA proposée à l'utilisateur.À cette fin, nous publions sur le Web les objets 3D correspondant aux objets du monde réel en utilisant notre modèle de données, ce qui permet de les lier avec le nuage du Web des Données. De fait, un objet du monde réel peut être représenté selon trois dimensions d'information : thématique (qui décrit les rôles ou fonctions de l'objet), spatiale (à travers sa géométrie 3D), et temporelle (par la période qui correspond à son existence). De plus, chaque changement dans la vie d'un objet du monde réel est représenté dans le modèle par des informations sémantiques. En conséquence, sur son appareil mobile, l'utilisateur visualise une vue de RA construite avec des modèles 3D légers et transparents. Il peut interagir avec ces objets de RA afin d'accéder aux informations relatives aux objets du monde réel correspondants qui se trouvent aux alentours et vers lesquels il pointe son téléphone mobile. Ces objets de RA permettent à l'utilisateur de construire mentalement la relation référentielle entre les objets virtuels et réels, tandis que le développeur de l'application mobile peut créer des expériences basées sur des thématiques différentes disponibles à travers sur le nuage du Web des Données (architecture, histoire, gastronomie, loisirs, etc.).En nous appuyant sur cette base de connaissances conceptualisée, nous proposons une architecture appelée ARCAMA-3D (Réalité Augmentée et 3D pour les applications mobiles sensibles au contexte), dont les modules permettent aux concepteurs et développeurs de créer leurs propres applications de RA pour différents domaines en étant en mesure d'étendre le modèle de données, de lier les modèles 3D avec d'autres ensembles de données ouvertes et liées disponibles dans le nuage, et d'alimenter ces applications de découverte à partir de sources de données spécifiques. Notre proposition est illustrée sur le cas d'étude que constitue le Monastère Royal de Brou, en France, à travers un scénario de cas d'utilisation pour un utilisateur visitant ce monument historique. / Discovering the surroundings while walking using a mobile device is now possible by coupling new technologies such as Augmented Reality (AR) and location-based services. Many mobile applications have been designed for that purpose recently. The surroundings of the user are presented on the screen of her mobile device using the location data and she can interact with her environment through a digital information layer superimposed on the captured scene. However, since the amount of geo-referenced information increases every day, there is a growing need for a generic approach that aims at presenting information in a structured manner and adapted to different application domains and users profiles.Our work focuses on the definition of a data model for 3D objects to be used in mobile AR applications. The data model we propose prioritizes access to available spatiotemporal and thematic information about real-world objects. While Augmented Reality appears to be well-suited for searching location-based information, most of the AR approaches focus on domain specific scenarios and we observe that there is no generic data model dedicated to information search and discovery that could be re-used in various AR applications. On the other hand, Linked Open Data (LOD) cloud is an ever-growing set of structured and interlinked datasets published on the Web. It contains vast amounts of spatiotemporal information that can be exploited by location-based mobile applications and presented using AR for fostering interaction. In our research, the LOD cloud serves as a basis for information retrieval during the AR experience of the user.For this purpose, we first publish 3D objects that correspond to real-world objects (buildings, monuments, etc.) on the Web by using our data model and interlink them with data sources on the LOD cloud. Then, using our data model, any real-world object can be represented by three informational dimensions: themes (describing the roles or functions of this object), space (through its 3D geometry), and time (the period linked to its existence). Each change in the life of a real-world object is represented in the model by semantic information following the LOD publishing principles. This allows a binding between the content of our data model and the LOD cloud. As a consequence, on her mobile device, the user visualizes an AR view built with light and transparent 3D models. She can interact with these AR objects in order to access to information related to the corresponding real-world objects of her surroundings she is pointing at. These AR objects allow the user to mentally construct the referential relationships between virtual and real-world objects, while the mobile application developer can create experiences based on different concepts found on the cloud (thematic and temporal concepts, etc.). This way, the LOD cloud, as a growing and updated structured source of semantic data, becomes the main source of information and facilitates knowledge discovery in AR applications.Using this conceptualized knowledge base, we propose our architecture, called ARCAMA-3D (Augmented Reality for Context Aware Mobile Applications with 3D), whose modules allow application designers and developers to create their own AR applications for different domains by being able to extend the data model, bind 3D models with other data sources on the LOD cloud, cover a selected part of LOD and feed their application only with these specific data sources. We develop our ideas working on the case study of the Royal Monastery of Brou, in France, and implement a use case scenario for a user visiting the monastery.

Les mers côtières jouent un rôle essentiel pour les sociétés humaines. Mais la concentration et la diversité des activités qui s’y développent exercent une pression croissante sur cet espace et les milieux associés en générant des interactions parfois conflictuelles entre activités. La compréhension de ces interactions constitue un enjeu en termes de recherche et pour la société civile. Une méthodologie visant à décrire la distribution spatio-temporelle de différentes activités en mer côtière est donc conçue et mise en œuvre en rade de Brest. La collecte de données spatiales, temporelles, quantitatives et qualitatives combine l’analyse de bases de données spatio-temporelles comme celles issues du Système d’Identification Automatique (AIS), et le dépouillement d’entretiens semi-directifs menés auprès de personnes-ressources. À partir des données hétérogènes collectées, une information structurée dans une base de données spatio-temporelle (BDST) est produite. Dans un premier temps, son exploitation cartographique par un Système d’Information Géographique (SIG) permet la réalisation d’instantanés au pas de temps quotidien sur l’ensemble d’une année. La qualité de l’information temporelle et quantitative puis la nature et la source de l’information spatiale sont renseignées. Dans un second temps, la BDST est mobilisée pour identifier, spatialiser et quantifier les conflits d’usages potentiels et les interactions spatio-temporelles potentielles négatives entre les activités considérées en rade de Brest. / Coastal seas play an essential role for human societies who develop many and diverse activities. These space and resource consuming activities induce an increase pressure on the environment and sometimes generate conflicting interactions among various activities. Understanding these interactions remains a challenge for research and civil society. A methodology is proposed to describe the spatial and temporal distribution of several activities in coastal sea. An application is developed in the bay of Brest (Brittany, France). Spatial, temporal, quantitative and qualitative data acquisition combines both analysis of spatio-temporal databases such as automatic identification system (AIS) databases, and results from semi-structured interviews with key-informants. The heterogeneous data collected are stored in a spatio-temporal database (STDB). First, the STDB is used with a Geographic Information System (GIS) to produce temporal snapshots of daily human activities patterns within a year. The quality of temporal and quantitative information and the nature and source of spatial information are also provided. Secondly, the STBD enables to identify, quantify and map potential uses conflicts and potential negative spatial interactions between activities in the bay of Brest.

Le domaine de la modélisation du trafic routier vise à comprendre son évolution. Dans les dernières années, plusieurs modèles du trafic ont été proposés dans l’objectif de géolocaliser les embouteillages au sein du trafic, détecter des motifs dans le trafic routier, estimer l’état du trafic etc. La plupart des modèles proposés considèrent le trafic routier en termes de ses constituants ou comme une entité agrégée en fonction de l’échelle choisie et expliquent l’évolution du trafic quantitativement en tenant compte des relations entre les variables de trafic comme le flot, la densité et la vitesse. Ces modèles décrivent le trafic en utilisant des données très précises acquises par différents capteurs. La précision des données rend son calcul coûteux en termes de ressources requises. Une des solutions à ce problème est la représentation qualitative du trafic routier qui réduit le nombre de ressources de traitement nécessaires. Puisque le trafic routier est un phénomène spatio-temporel, les modèles proposés pour représenter ce type de phénomène pourraient être appliqués dans le cas du trafic routier. Les modèles spatio-temporels, proposés par la communauté de l’Analyse Spatio-Temporelle, ont comme objectif la représentation d’un phénomène tant du point de vue qualitatif que quantitatif. Certains de ces modèles proposent une discrétisation des phénomènes modélisés en considérant un phénomène comme constitué d’entités. Appliquée au trafic routier, cette notion permet d’identifier différentes entités, comme les véhicules, les piétons, les bâtiments etc., qui le constituent. Ces entités influent sur l’évolution du trafic. Les modèles spatio-temporels qualitatifs définissent l’effet des différentes entités les unes sur les autres en terme de relations spatiales. L’évolution spatio-temporelle du phénomène modélisé est représenté par la variation temporelle de ces relations. La prise en compte des entités du trafic et des relations spatiales formalise une structure qui peut être représentée en utilisant un graphe, où les nœuds modélisent des entités et les arcs des relations spatiales. Par conséquent, l’évolution du trafic, modélisée via ce graphe, devient l’évolution du graphe et peut être représenté en terme de la variation de la structure du graphe ainsi que celle des attributs de ses nœuds et de ses arcs. Dans cette thèse, nous proposons une modélisation du trafic routier de ce type basée sur la théorie des graphes. Une des applications à la modélisation du trafic routier est la détection des motifs pertinents au sein du trafic. Dans les modèles du trafic existants, les motifs détectés sont statistiques et sont représentés en utilisant des caractéristiques numériques. Le modèle que nous pro posons dans cette thèse met en avant la structure représentant le trafic routier et peut donc être utilisé pour définir des motifs structurels du trafic qui prennent en compte des différentes entités du trafic et leurs relations. Ces motifs structurels sont sous-jacents à une modélisation sous forme de graphe dynamique. Dans cette thèse, nous proposons un algorithme pour détecter ces motifs structurels du trafic dans le graphe spatio-temporel représentant le trafic routier. Ce problème est formalisé comme celui de l’isomorphisme de sous-graphe pour des graphes dynamiques. L’algorithme proposé est évalué en fonction desdifférents paramètres de graphes. / For past several decades, researchers have been interested in understanding traffic evolution, hence, have proposed various traffic models to identify bottleneck locations where traffic congestion occurs, to detect traffic patterns, to predict traffic states etc. Most of the existing models consider traffic as many-particle system, describe it using different scales of representation and explain its evolution quantitatively by deducing relations between traffic variables like flow, density and speed. Such models are mainly focused on computing precise information about traffic using acquired traffic data. However, computation of such precise information requires more processing resources. A way to remedy this problem is to consider traffic evolution in qualitative terms which reduces the required number of processing resources. Since traffic is spatio-temporal in nature, the models which deal with spatio-temporal phenomenon can be applied in case of traffic. Such models represent spatio-temporal phenomenon from qualitative as well as quantitative standpoints. Depending on the intended application, some models are able to differentiate between various entities taking part in the phenomenon, which proves useful in case of traffic since different objects like vehicles, buildings, pedestrians, bicycles etc., directly affecting traffic evolution, can be included in traffic models. Qualitative spatio-temporal models consider the effects of different entities on each other in terms of spatial relations between them and spatio-temporal evolution of the modeled phenomenon is described in terms of variation in such relations over time. Considering different traffic constituents and spatial relations between them leads to the formation of a structure which can be abstracted using graph, whose nodes represent individual constituents and edges represent the corresponding spatial relations. As a result, the evolution of traffic, represented using graph, is described in terms of evolution of the graph itself, i. e. change in graph structure and attributes of nodes and edges, with time. In this thesis, we propose such a graph model to represent traffic. As mentioned above, one of the applications of existing traffic models is in detecting traffic patterns. However, since such models consider traffic quantitatively, in terms of acquired traffic data, the patterns detected using such models are statistical (a term employed by Pattern Recognition researchers) in the sense that they are represented using numerical description. Since graph-based traffic model proposed in this thesis represents the structure of traffic, it can be employed to redefine the meaning of traffic patterns from statistical to structural (also a term from Pattern Recognition community). Structural traffic patterns include different traffic constituents and their inter-links and are represented using time-varying graphs. An algorithm to detect a given structural traffic pattern in the spatio-temporal graph representing traffic is proposed in this thesis. It formalizes this problem as subgraph isomorphism for time-varying graphs. In the end, the performance of the algorithm is tested using various graph parameters.