Dans beaucoup de pays du monde, on observe une importante augmentation du nombre de personnes âgées vivant seules. Depuis quelques années, un nombre significatif de projets de recherche sur l’assistance aux personnes âgées ont vu le jour. La plupart de ces projets utilisent plusieurs modalités (vidéo, son, détection de chute, etc.) pour surveiller l'activité de la personne et lui permettre de communiquer naturellement avec sa maison "intelligente", et, en cas de danger, lui venir en aide au plus vite. Ce travail a été réalisé dans le cadre du projet ANR VERSO de recherche industrielle, Sweet-Home. Les objectifs du projet sont de proposer un système domotique permettant une interaction naturelle (par commande vocale et tactile) avec la maison, et procurant plus de sécurité à l'habitant par la détection des situations de détresse. Dans ce cadre, l'objectif de ce travail est de proposer des solutions pour la reconnaissance des sons de la vie courante dans un contexte réaliste. La reconnaissance du son fonctionnera en amont d'un système de Reconnaissance Automatique de la Parole. Les performances de celui-ci dépendent donc de la fiabilité de la séparation entre la parole et les autres sons. Par ailleurs, une bonne reconnaissance de certains sons, complétée par d'autres sources informations (détection de présence, détection de chute, etc.) permettrait de bien suivre les activités de la personne et de détecter ainsi les situations de danger. Dans un premier temps, nous nous sommes intéressés aux méthodes en provenance de la Reconnaissance et Vérification du Locuteur. Dans cet esprit, nous avons testé des méthodes basées sur GMM et SVM. Nous avons, en particulier, testé le noyau SVM-GSL (SVM GMM Supervector Linear Kernel) utilisé pour la classification de séquences. SVM-GSL est une combinaison de SVM et GMM et consiste à transformer une séquence de vecteurs de longueur arbitraire en un seul vecteur de très grande taille, appelé Super Vecteur, et utilisé en entrée d'un SVM. Les expérimentations ont été menées en utilisant une base de données créée localement (18 classes de sons, plus de 1000 enregistrements), puis le corpus du projet Sweet-Home, en intégrant notre système dans un système plus complet incluant la détection multi-canaux du son et la reconnaissance de la parole. Ces premières expérimentations ont toutes été réalisées en utilisant un seul type de coefficients acoustiques, les MFCC. Par la suite, nous nous sommes penchés sur l'étude d'autres familles de coefficients en vue d'en évaluer l'utilisabilité en reconnaissance des sons de l'environnement. Notre motivation fut de trouver des représentations plus simples et/ou plus efficaces que les MFCC. En utilisant 15 familles différentes de coefficients, nous avons également expérimenté deux approches pour transformer une séquence de vecteurs en un seul vecteur, à utiliser avec un SVM linéaire. Dans le première approche, on calcule un nombre fixe de coefficients statistiques qui remplaceront toute la séquence de vecteurs. La seconde approche (une des contributions de ce travail) utilise une méthode de discrétisation pour trouver, pour chaque caractéristique d'un vecteur acoustique, les meilleurs points de découpage permettant d'associer une classe donnée à un ou plusieurs intervalles de valeurs. La probabilité de la séquence est estimée par rapport à chaque intervalle. Les probabilités obtenues ainsi sont utilisées pour construire un seul vecteur qui remplacera la séquence de vecteurs acoustiques. Les résultats obtenus montrent que certaines familles de coefficients sont effectivement plus adaptées pour reconnaître certaines classes de sons. En effet, pour la plupart des classes, les meilleurs taux de reconnaissance ont été observés avec une ou plusieurs familles de coefficients différentes des MFCC. Certaines familles sont, de surcroît, moins complexes et comptent une seule caractéristique par fenêtre d'analyse contre 16 caractéristiques pour les MFCC / In many countries around the world, the number of elderly people living alone has been increasing. In the last few years, a significant number of research projects on elderly people monitoring have been launched. Most of them make use of several modalities such as video streams, sound, fall detection and so on, in order to monitor the activities of an elderly person, to supply them with a natural way to communicate with their “smart-home”, and to render assistance in case of an emergency. This work is part of the Industrial Research ANR VERSO project, Sweet-Home. The goals of the project are to propose a domotic system that enables a natural interaction (using touch and voice command) between an elderly person and their house and to provide them a higher safety level through the detection of distress situations. Thus, the goal of this work is to come up with solutions for sound recognition of daily life in a realistic context. Sound recognition will run prior to an Automatic Speech Recognition system. Therefore, the speech recognition’s performances rely on the reliability of the speech/non-speech separation. Furthermore, a good recognition of a few kinds of sounds, complemented by other sources of information (presence detection, fall detection, etc.) could allow for a better monitoring of the person's activities that leads to a better detection of dangerous situations. We first had been interested in methods from the Speaker Recognition and Verification field. As part of this, we have experimented methods based on GMM and SVM. We had particularly tested a Sequence Discriminant SVM kernel called SVM-GSL (SVM GMM Super Vector Linear Kernel). SVM-GSL is a combination of GMM and SVM whose basic idea is to map a sequence of vectors of an arbitrary length into one high dimensional vector called a Super Vector and used as an input of an SVM. Experiments had been carried out using a locally created sound database (containing 18 sound classes for over 1000 records), then using the Sweet-Home project's corpus. Our daily sounds recognition system was integrated into a more complete system that also performs a multi-channel sound detection and speech recognition. These first experiments had all been performed using one kind of acoustical coefficients, MFCC coefficients. Thereafter, we focused on the study of other families of acoustical coefficients. The aim of this study was to assess the usability of other acoustical coefficients for environmental sounds recognition. Our motivation was to find a few representations that are simpler and/or more effective than the MFCC coefficients. Using 15 different acoustical coefficients families, we have also experimented two approaches to map a sequence of vectors into one vector, usable with a linear SVM. The first approach consists of computing a set of a fixed number of statistical coefficients and use them instead of the whole sequence. The second one, which is one of the novel contributions of this work, makes use of a discretization method to find, for each feature within an acoustical vector, the best cut points that associates a given class with one or many intervals of values. The likelihood of the sequence is estimated for each interval. The obtained likelihood values are used to build one single vector that replaces the sequence of acoustical vectors. The obtained results show that a few families of coefficients are actually more appropriate to the recognition of some sound classes. For most sound classes, we noticed that the best recognition performances were obtained with one or many families other than MFCC. Moreover, a number of these families are less complex than MFCC. They are actually a one-feature per frame acoustical families, whereas MFCC coefficients contain 16 features per frame

In this thesis a clustering of the Stockholm county housing market has been performed using different clustering methods. Data has been derived and different geographical constraints have been used. DeSO areas (Demographic statistical areas), developed by SCB, have been used to divide the housing market in to smaller regions for which the derived variables have been calculated. Hierarchical clustering methods, SKATER and Gaussian mixture models have been applied. Methods using different kinds of geographical constraints have also been applied in an attempt to create more geographically contiguous clusters. The different methods are then compared with respect to performance and stability. The best performing method is the Gaussian mixture model EII, also known as the K-means algorithm. The most stable method when applied to bootstrapped samples is the ClustGeo-method. / I denna uppsats har en klustring av Stockholms läns bostadsmarknad genomförts med olika klustringsmetoder. Data har bearbetats och olika geografiska begränsningar har använts. DeSO (Demografiska Statistiska Områden), som utvecklats av SCB, har använts för att dela in bostadsmarknaden i mindre regioner för vilka områdesattribut har beräknats. Hierarkiska klustringsmetoder, SKATER och Gaussian mixture models har tillämpats. Metoder som använder olika typer av geografiska begränsningar har också tillämpats i ett försök att skapa mer geografiskt sammanhängande kluster. De olika metoderna jämförs sedan med avseende på kvalitet och stabilitet. Den bästa metoden, med avseende på kvalitet, är en Gaussian mixture model kallad EII, även känd som K-means. Den mest stabila metoden är ClustGeo-metoden.

Master of Science / Department of Statistics / Weixin Yao / Mixtures of factor analyzers have been popularly used to cluster the high dimensional data. However, the traditional estimation method is based on the normality assumptions of random terms and thus is sensitive to outliers. In this article, we introduce a robust estimation procedure of mixtures of factor analyzers using the trimmed likelihood estimator (TLE). We use a simulation study and a real data application to demonstrate the robustness of the trimmed estimation procedure and compare it with the traditional normality based maximum likelihood estimate.

Doctor of Philosophy / Department of Statistics / Weixin Yao and Kun Chen / Ordinary least-squares (OLS) estimators for a linear model are very sensitive to unusual values in the design space or outliers among y values. Even one single atypical value may have a large effect on the parameter estimates. In this proposal, we first review and describe some available and popular robust techniques, including some recent developed ones, and compare them in terms of breakdown point and efficiency. In addition, we also use a simulation study and a real data application to compare the performance of existing robust methods under different scenarios. Finite mixture models are widely applied in a variety of random phenomena. However, inference of mixture models is a challenging work when the outliers exist in the data. The traditional maximum likelihood estimator (MLE) is sensitive to outliers. In this proposal, we propose a Robust Mixture via Mean shift penalization (RMM) in mixture models and Robust Mixture Regression via Mean shift penalization (RMRM) in mixture regression, to achieve simultaneous outlier detection and parameter estimation. A mean shift parameter is added to the mixture models, and penalized by a nonconvex penalty function. With this model setting, we develop an iterative thresholding embedded EM algorithm to maximize the penalized objective function. Comparing with other existing robust methods, the proposed methods show outstanding performance in both identifying outliers and estimating the parameters.

Doctor of Philosophy / Department of Statistics / Kun Chen and Weixin Yao / This proposal contains two projects that are related to robust mixture models. In the robust project, we propose a new robust mixture of regression models (Bai et al., 2012). The existing methods for tting mixture regression models assume a normal distribution for error and then estimate the regression param- eters by the maximum likelihood estimate (MLE). In this project, we demonstrate that the MLE, like the least squares estimate, is sensitive to outliers and heavy-tailed error distributions. We propose a robust estimation procedure and an EM-type algorithm to estimate the mixture regression models. Using a Monte Carlo simulation study, we demonstrate that the proposed new estimation method is robust and works much better than the MLE when there are outliers or the error distribution has heavy tails. In addition, the proposed robust method works comparably to the MLE when there are no outliers and the error is normal. In the second project, we propose a new robust mixture of linear mixed-effects models. The traditional mixture model with multiple linear mixed effects, assuming Gaussian distribution for random and error parts, is sensitive to outliers. We will propose a mixture of multiple linear mixed t-distributions to robustify the estimation procedure. An EM algorithm is provided to and the MLE under the assumption of t- distributions for error terms and random mixed effects. Furthermore, we propose to adaptively choose the degrees of freedom for the t-distribution using profile likelihood. In the simulation study, we demonstrate that our proposed model works comparably to the traditional estimation method when there are no outliers and the errors and random mixed effects are normally distributed, but works much better if there are outliers or the distributions of the errors and random mixed effects have heavy tails.

The peaks over threshold (POT) method provides an attractive framework for estimating the risk of extreme events such as severe storms or large insurance claims. However, the conventional POT procedure, where the threshold excesses are modelled by a generalized Pareto distribution, suffers from small samples and subjective threshold selection. In recent years, two alternative approaches have been proposed in the form of mixture models that estimate the threshold and a folding procedure that generates larger tail samples. In this paper the empirical performances of the conventional POT procedure, the folding procedure and a mixture model are compared by modelling data sets on fire insurance claims and hurricane damage costs. The results show that the folding procedure gives smaller standard errors of the parameter estimates and in some cases more stable quantile estimates than the conventional POT procedure. The mixture model estimates are dependent on the starting values in the numerical maximum likelihood estimation, and are therefore difficult to compare with those from the other procedures. The conclusion is that none of the procedures is overall better than the others but that there are situations where one method may be preferred.

Positron emission tomography (PET) is a non-invasive imaging technique which utilizes positron-emitting radiopharmaceuticals (PERs) to characterize biological processes in tissues of interest. A PET scanner is usually composed of multiple scintillation crystal detectors placed in a ring so as to capture coincident photons from a position annihilation. These detectors require a crystal lookup table (CLUT) to map the detector response to the crystal of interaction. These CLUTs must be accurate, lest events get mapped to the wrong crystal of interaction degrading the final image quality. This work describes an automated algorithm, for CLUT generation, focused around Gaussian Mixture Models (GMM) with Thin Plate Splines (TPS). The algorithm was tested with flood image data collected from 16 detectors. The method maintained at least 99.8% accuracy across all tests. This method is considerably faster than manual techniques and can be adapted to different detector configurations. / February 2016

Master of Science / Department of Statistics / Weixin Yao / In this report, we introduce the minimum Hellinger distance (MHD) estimation method and review its history. We examine the use of Hellinger distance to obtain a new efficient and robust estimator for a class of semiparametric mixture models where one component has known distribution while the other component and the mixing proportion are unknown. Such semiparametric mixture models have been used in biology and the sequential clustering algorithm. Our new estimate is based on the MHD, which has been shown to have good efficiency and robustness properties. We use simulation studies to illustrate the finite sample performance of the proposed estimate and compare it to some other existing approaches. Our empirical studies demonstrate that the proposed minimum Hellinger distance estimator (MHDE) works at least as well as some existing estimators for most of the examples considered and outperforms the existing estimators when the data are under contamination. A real data set application is also provided to illustrate the effectiveness of our proposed methodology.

This thesis is mostly about supervised visual recognition problems. Based on a general definition of categories, the contents are divided into two parts: one which models categories and one which is not category based. We are interested in data driven solutions for both kinds of problems. In the category-free part, we study novelty detection in temporal and spatial domains as a category-free recognition problem. Using data driven models, we demonstrate that based on a few reference exemplars, our methods are able to detect novelties in ego-motions of people, and changes in the static environments surrounding them. In the category level part, we study object recognition. We consider both object category classification and localization, and propose scalable data driven approaches for both problems. A mixture of parametric classifiers, initialized with a sophisticated clustering of the training data, is demonstrated to adapt to the data better than various baselines such as the same model initialized with less subtly designed procedures. A nonparametric large margin classifier is introduced and demonstrated to have a multitude of advantages in comparison to its competitors: better training and testing time costs, the ability to make use of indefinite/invariant and deformable similarity measures, and adaptive complexity are the main features of the proposed model. We also propose a rather realistic model of recognition problems, which quantifies the interplay between representations, classifiers, and recognition performances. Based on data-describing measures which are aggregates of pairwise similarities of the training data, our model characterizes and describes the distributions of training exemplars. The measures are shown to capture many aspects of the difficulty of categorization problems and correlate significantly to the observed recognition performances. Utilizing these measures, the model predicts the performance of particular classifiers on distributions similar to the training data. These predictions, when compared to the test performance of the classifiers on the test sets, are reasonably accurate. We discuss various aspects of visual recognition problems: what is the interplay between representations and classification tasks, how can different models better adapt to the training data, etc. We describe and analyze the aforementioned methods that are designed to tackle different visual recognition problems, but share one common characteristic: being data driven. / <p>QC 20140604</p>

La courbe S-N est le moyen le plus courant d'analyse et de prédiction de la durée de vie d'un matériau, d'un composant ou d'une structure. Cependant, les modèles standards, qu'ils soient basés sur la théorie de la rupture ou sur des modèles probabilistes n'ajustent pas la courbe dans la totalité sans information sur la microstructure du matériau. Or, cette information provient d'analyses fractographiques souvent coûteuses et rarement disponibles dans le cadre d'une production industrielle. D'un autre côté, les modèles statistiques ne proposent pas d'interprétation matériau et ne peuvent pas être utilisées pour réaliser des prévisions. Les résultats d'un test de fatigue sont par ailleurs très dispersés, plus particulièrement pour les fortes durées de vie, lieu d'apparition d'un phénomène de bi-modalité. Ces constats sont la raison de la proposition d'un nouveau modèle probabiliste. Celui-ci est composé d'un modèle de mélange spécifique, prenant en compte l'approche apportée par la mécanique de la rupture sans nécessiter de d'information supplémentaire sur la microstructure du matériau. Il utilise le fait que la fatigue peut être vue comme la somme d'un amorçage de fissure suivi de sa propagation. Les paramètres du modèle sont estimés à l'aide d'un algorithme EM, où la phase de maximisation combine une méthode d'optimisation de Newton-Raphson et une intégration de type Monte-Carlo. Le modèle "amorçage-propagation" offre une représentation parcimonieuse des courbes $S-N$ dont les paramètres peuvent être facilement interprétés par des ingénieurs matériau. Ce modèle a été testé à l'aide de simulations et appliqué à des données réelles (données sur l'Inconel 718). Ceci nous a permis de mettre en évidence le bon ajustement du modèle à nos données, et ce, pour toutes les déformations disponibles. / S-N curve is the main tool to analyze and predict fatigue lifetime of a material, component or structure. But, standard models based on mechanic of rupture theory or standard probabilistic models for analyzing S-N curves could not fit S-N curve on the whole range of cycles without microstructure information. This information is obtained from costly fractography investigation rarely available in the framework of industrial production. On the other hand, statistical models for fatigue lifetime do not need microstructure information but they could not be used to service life predictions because they have no material interpretation. Moreover, fatigue test results are widely scattered, especially for High Cycle Fatigue region where split $S-N$ curves appear. This is the motivation to propose a new probabilistic model. This model is a specific mixture model based on a fracture mechanic approach, and does not require microstructure information. It makes use of the fact that the fatigue lifetime can be regarded as the sum of the crack initiation and propagation lifes. The model parameters are estimated with an EM algorithm for which the maximisation step combines Newton-Raphson optimisation method and Monte Carlo integrations. The resulting model provides a parsimonious representation of S-N curves with parameters easily interpreted by mechanic or material engineers. This model has been applied to simulated and real fatigue test data sets. These numerical experiments highlight its ability to produce a good fit of the S-N curves on the whole range of cycles.