L’hémophilie est une maladie génétique se traduisant par la déficience des facteurs VIII et IX de la coagulation et conduisant à une tendance hémorragique. L’intensité des traitements substitutifs en facteur VIII et IX est définie essentiellement sur le taux basal du facteur déficitaire et non pas sur la capacité propre à chaque patient à générer de la thrombine qui est l’enzyme clé dans la formation du caillot de fibrine. Le test de génération de thrombine pourrait être utilisé pour permettre une individualisation du traitement anti-hémophilique. En effet, le taux de facteur VIII ou IX nécessaire à la normalisation de la génération de thrombine est potentiellement variable d’un patient à l’autre pour une même sévérité d’hémophilie. On peut donc se demander quelle approche expérimentale permettrait de mettre en exergue le lien entre taux de facteur anti-hémophilique et la génération de thrombine. Est-il possible de modéliser mathématiquement la coagulation pour obtenir une relation, soit explicite, soit implicite, entre taux de facteurs et génération de thrombine ? Les modèles existants permettent-ils d'obtenir une telle relation ? Une vaste campagne expérimentale a donc été menée pour mettre en place une base de données qui a permis d’identifier les facteurs déterminants de la génération de thrombine et la relation entre génération de thrombine et taux de facteur anti-hémophilique, de définir leurs valeurs de références, ainsi que d’évaluer et de paramétrer de manière sujet-spécifique des modèles mathématiques de la coagulation. / Haemophilia is a genetic disease corresponding to the deficiency of coagulation factor VIII or IX and leading to a bleeding tendency. The current substitutive treatment is defined essentially by the basal level of deficient factor and not the individual capacity to generate thrombin, a key enzyme of the clot formation. The thrombin generation assay could help in the individualisation of the anti-haemophilia treatment. Indeed, the factor VIII or IX level needed to normalise the thrombin generation vary potentially from one patient to another for a same degree of severity. We can wonder which experimental approach could emphasise the relation between level of anti-haemophilic factor and thrombin generation. Is it possible to mathematically model coagulation to obtain a relation, either explicit, or implicit, between factor level and thrombin generation? Could existing models provide this relation? An extensive experimental campaign was carried out to build a database that has been used to identify the determinant coagulation factors of thrombin generation and the individual relation between thrombin generation and anti-haemophilic factor level, to define their reference values, and also to evaluate and parametrise subject-specifically mathematical models of the coagulation cascade

This thesis presents an all-inclusive framework on how the current formant tracking and audio (and/or visual)-to-articulatory inversion algorithms can be improved. The possible improvements are summarized as follows: The first part of the thesis investigates the problem of the formant frequency estimation when the number of formants to be estimated fixed or variable respectively. The fixed number of formant tracking method is based on the assumption that the number of formant frequencies is fixed along the speech utterance. The proposed algorithm is based on the combination of a dynamic programming algorithm and Kalman filtering/smoothing. In this method, the speech signal is divided into voiced and unvoiced segments, and the formant candidates are associated via dynamic programming algorithm for each voiced and unvoiced part separately. Individual adaptive Kalman filtering/smoothing is used to perform the formant frequency estimation. The performance of the proposed algorithm is compared with some algorithms given in the literature. The variable number of formant tracking method considers those formant frequencies which are visible in the spectrogram. Therefore, the number of formant frequencies is not fixed and they can change along the speech waveform. In that case, it is also necessary to estimate the number of formants to track. For this purpose, the proposed algorithm uses extra logic (formant track start/end decision unit). The measurement update of each individual formant trajectories is handled via Kalman filters. The performance of the proposed algorithm is illustrated by some examples The second part of this thesis is concerned with improving audiovisual to articulatory inversion performance. The related studies can be examined in two parts / Gaussian mixture model (GMM) regression based inversion and Jump Markov Linear System (JMLS) based inversion. GMM regression based inversion method involves modeling audio (and /or visual) and articulatory data as a joint Gaussian mixture model. The conditional expectation of this distribution gives the desired articulatory estimate. In this method, we examine the usefulness of the combination of various acoustic features and effectiveness of various types of fusion techniques in combination with audiovisual features. Also, we propose dynamic smoothing methods to smooth articulatory trajectories. The performance of the proposed algorithm is illustrated and compared with conventional algorithms. JMLS inversion involves tying the acoustic (and/or visual) spaces and articulatory space via multiple state space representations. In this way, the articulatory inversion problem is converted into the state estimation problem where the audiovisual data are considered as measurements and articulatory positions are state variables. The proposed inversion method first learns the parameter set of the state space model via an expectation maximization (EM) based algorithm and the state estimation is handled via interactive multiple model (IMM) filter/smoother.