The common practice in sign language recognition is to first construct individual sign models, in terms of discrete state transitions, mostly represented using Hidden Markov Models, from manually isolated sign samples and then to use them to recognize signs in continuous sentences. In this thesis we use a continuous state space model, where the states are based on purely image-based features, without the use of special gloves. We also present an unsupervised approach to both extract and learn models for continuous basic units of signs, which we term as signemes, from continuous sentences. Given a set of sentences with a common sign, we can automatically learn the model for part of the sign,or signeme, that is least affected by coarticulation effects. We tested our idea using the publicly available Boston SignStreamDataset by building signeme models of 18 signs. We test the quality of the models by considering how well we can localize the sign in a new sentence. We also present the concept of smooth continuous curve based models formed using functional splines and curve registration. We illustrate this idea using 16 signs.

Tropical forests of the world are fast disappearing and there is a race to understand patterns of species distribution in space and time. Studying species distributions can provide better frameworks for conservation of these ecologically important patches of floral and faunal diversity. The island of Sri Lanka is a well known harbour of unique and highly threatened biodiversity. Tropical lowland forest is remnant in the south-west of the island now mainly existing in small patches. While most are small disturbed fragments, Sinharaja Forest Reserve represents one of the largest remaining patches of this important ecosystem. As a UNESCO World Heritage Site and a Man and Biosphere Reserve, it has a dual role as a conservation area and a historically important resource forest. While the distribution of vegetation diversity has been well documented, analyses of invertebrate species distributions are lacking. This thesis investigated a key arthropod group, ground dwelling ants, in relation to environmental gradients within the forest. Cumulative results demonstrate the high diversity of the forest patch. In an area representing less than half the reserve, over 173 ground dwelling ant species were found in distinct assemblages throughout the forest. Since the forest is located upon a series of parallel ridges, ant species distribution was first analysed in terms of this small elevation change. Species richness declined over a vertical incline from 430 m to 660 m, highlighting a possible small-scale, mountain mass effect. This section of the reserve is also characterised by a patch of once-logged forest (30 years previously). A study was undertaken to investigate whether there were residual effects of selective logging on the reserve. / Significant differences between species assemblages in once-logged forest and unlogged forest add to growing evidence that selectively logged forests continue to remain distinct from unlogged forest even after decades of regeneration. Ant distribution was then analysed for their relationship with habitat heterogeneity and tree species distribution. Long-term research on tree species in the SFR has demonstrated a close relationship to habitat complexity. Ant species appear to respond more to the structural heterogeneity of the vegetation than to actual topographic variation within the forest. From a conservation perspective, maintaining the integrity of this highly diverse forest is imperative. The impact of anthropogenic land uses surrounding the forest was investigated in terms of ant assemblages along the forest edges. Significant differences were found between assemblages within the edges bordered by different matrix types. Even relatively large forest remnants can be affected by the surrounding matrix land uses and encouraging the growth of structurally similar vegetation and maintaining low disturbance along the borders should attenuate the effect of the edge. Overall, the highly heterogeneous distribution of ant assemblages within the SFR demonstrates the potential for other small patches to be harbours of further species diversity. Future research should be undertaken to assess the diversity and distribution of ant species within this region and encourage the protection of this remnant diversity.

The estimation of the distribution of usual intake of nutrients is important for developing nutrition policies as well as for etiological research and educational purposes. In most nutrition surveys only a small number of repeated intake observations per individual are collected. Of main interest is the longterm usual intake which is defined as long-term daily average intake of a dietary component. However, dietary intake on a single day is a poor estimate of the individual’s long-term usual intake. Furthermore, the distribution of individual intake means is also a poor estimator of the distribution of usual intake since usually there is large within-individual compared to between-individual variability in the dietary intake data. Hence, the variance of the mean intakes is larger than the variance of the usual intake distribution. Essentially, the estimation of the distribution of long-term intake is equivalent to the estimation of a distribution of a random variable observed with measurement error. Some of the methods for estimating the distributions of usual dietary intake are reviewed in detail and applied to nutrient intake data in order to evaluate their properties. The results indicate that there are a number of robust methods which could be used to derive the distribution of long-term dietary intake. The methods share a common framework but differ in terms of complexity and assumptions about the properties of the dietary consumption data. Hence, the choice of the most appropriate method depends on the specific characteristics of the data, research purposes as well as availability of analytical tools and statistical expertise.

On étudie les valeurs spéciales de fonctions $L$ attachées aux formes modulaires, tordues par des caractères de Dirichlet de conducteur arbitraire. On construit des distributions à valeurs algébriques à partir de distributions à valeurs modulaires. On prouve sur ces dernières des résultats de congruence et d'admissibilité. On obtient alors les distributions scalaires en appliquant une forme linéaire, ce qui permet de montrer qu'elles vérifient elles aussi de bonnes propriétés. On établit ensuite un lien entre ces distributions scalaires et les valeurs spéciales qui nous intéressent, ce qui permet d'établir de nouvelles congruences entre elles.

<p>For random waves the free surface elevation can be described by a number of individual wave parameters. The main objective of this work has been to study the statistical properties of individual parameters in successive waves; the wave crest height, the wave height and the wave period.</p><p>In severe sea states the wave crest heights exhibit a nonlinear behavior, which must be reflected in the models. An existing marginal distribution that uses second order Stokes-type nonlinearity is transformed to a two-dimensional distribution by use of the two–dimensional Rayleigh distribution. This model only includes sum frequency effects. A two-dimensional distribution is also established by transforming a second order model including both sum and different frequency effects. Both models are based on the narrow-band assumption, and the effect of finite water depth is included. A parametric wave crest height distribution proposed by Forristall (2000) has been extended to two dimensions by transformation of the two-dimensional Weibull distribution. </p><p>Two successive wave heights are modeled by a Gaussian copula, which is referred to as the Nataf model. Results with two initial distributions for the transformation are presented, the Næss (1985) model and a two-parameter Weibull distribution, where the latter is in best agreement with data. The results are compared with existing models. The Nataf model has also been used for modeling three successive wave heights. Results show that the Nataf transformation of three successive wave heights can be approximated by a first order autoregression model. This means that the distribution of the wave height given the previous wave height is independent of the wave heights prior to the previous wave height. The simulation of successive wave heights can be done directly without simulating the time series of the complete surface elevation. </p><p>Successive wave periods are modeled with the Nataf transformation by using a two-parameter Weibull distribution and a generalized Gamma distribution as the initial distribution, where the latter is in best agreement with data. Results for the marginal and two-dimensional distributions are compared with existing models. In practical applications, it is often of interest to consider successive wave periods with corresponding wave heights exceeding a certain threshold. Results show that the distribution for successive wave periods when the corresponding wave heights exceed the root-mean-square value of the wave heights can be approximated by a multivariate Gaussian distribution. When comparing the results with data, a long time series is needed in order to obtain enough data cases. Results for three successive wave periods are also presented. </p><p>The models are compared with field data from the Draupner field and the Japan Sea, and with laboratory data from experiments at HR Wallingford. In addition, data from numerical simulations based on second order wave theory, including both sum and frequency effects, are included.</p>

Les hydrates de gaz sont des composés solides formés à partir de molécules de gaz emprisonnées dans des structures cristallines formées par des molécules d'eau reliées par liaisons hydrogène. Ils sont stables sous des conditions de haute pression et de basse température. Dans les conduites pétrolières, la formation d'hydrate de gaz peut être responsable du colmatage des conduites et du blocage des vannes. Pour éviter leur cristallisation, il existe plusieurs solutions : l'isolation ou le réchauffage de la conduite pétrolière ainsi que l'injection d'additifs cinétiques ou thermodynamiques. Une autre solution envisagée est l'utilisation d'additifs anti-agglomérants. Il s'agit d'agents tensio-actifs qui favorisent d'abord la formation d'une émulsion eau dans huile et ensuite limitent l'agglomération entre les cristaux une fois formés. De cette façon, la taille des particules d'hydrates serait limitée par la taille des gouttelettes d'eau dans l'émulsion. Cette méthode a été utilisée dans les travaux de Camargo (2001) à l'IFP et puis de Fidel-Dufour (2004) à l'ENSM de Saint-Etienne. Cette thèse est une étude consacrée à la caractérisation du couplage entre la cristallisation des hydrates et la rhéologie des écoulements pétroliers diphasiques (émulsion eau dans huile) laminaires ou turbulents dont l'objectif final est la production d'un modèle de cristallisation en écoulement. Elle s'appuie sur les mesures réalisées à l'aide de la sonde FBRM (Focused Beam Reflectance Measurement) qui permet des mesures en longueurs de cordes (CLD) in-situ lors de la formation d'hydrates en systèmes dispersés. La formation des hydrates de gaz en écoulement est étudiée de manière expérimentale sur deux dispositifs : une boucle de circulation Archimède située à l'ENSM-SE et une boucle de circulation Lyre à l'IFP Lyon. Les résultats obtenus avec les deux dispositifs sont comparés. La plus grande partie de ce travail a porté sur l'interprétation des mesures de longueurs de corde de la FBRM. Une série d'algorithmes permettant de générer des agrégats aléatoires ont été élaborés, suivis du calcul de leurs CLD. Ces CLD sont comparées avec les CLD obtenues expérimentalement permettant ainsi le suivi de l'agglomération pendant la cristallisation en écoulement.

This work focuses on the emission processes of metal-organic clusters MmMen, (M is the organic molecule and Me the metal atom) ejected from self assembled monolayers (SAMs) of alkanethiols on gold after keV ion bombardment. These aggregates are often observed upon energetic ion bombardment of strongly bound molecules like SAMs. The explanation of this effect remains elusive, especially for large clusters as those observed in our study. The emission of these clusters is investigated using ToF-SIMS under 15 keV Ga+ bombardment. In particular, we have measured the energy distributions (KEDs), which are informative of the physical processes of sputtering. We have probed both the influence of the intermolecular forces and the adsorbate-metal bonding on the cluster ion emission. Importantly, our KEDs revealed that a significant fraction of MmMen clusters is formed via the metastable decay of larger aggregates in the acceleration section of the spectrometer. This is the experimental evidence that another cluster formation channel has to be considered in addition to the recombination mechanisms proposed by other groups. In parallel to these experiments, we have used classical molecular dynamics (MD) simulations to model an overlayer of octanethiols on gold. A realistic potential has been used including long-range forces between the hydrocarbon chains of the alkanethiols. Our key finding concerns the emission of large clusters which were not observed under sub-keV projectile impact. Statistically, they are predominantly formed in high yield events, where many fragments and (supra)molecular species are ejected. From the microscopic viewpoint, these events mostly stem from the confinement of the projectile and recoil atom energies in a finite nanovolume of the surface. As a result of the high local energy density, molecular aggregates desorb from an overheated liquid-like region surrounding the impact point. In summary, from a combined experimental and computational study we have shown that analytical models involving linear collision cascades and recombination processes are insufficient to describe metal-thiolate cluster emission from SAMs under keV ion bombardment. The detailed MD investigation have allowed us to obtain a general picture of the emission of these aggregates in which the mechanisms at play are reminiscent of those high yields events (megaevents) with non linear effects used usually to account for large (bio)molecule desorption.

This thesis presents the results of a study on diamonds from the Artemisia kimberlite, Nunavut, Canada. This study integrates the isotopic, chemical and physical characteristics of the diamonds in an attempt to provide insight into the nature of the micro-/macro-diamond genetic relationship, and its possible implication pertaining to the use of size frequency distributions in modern kimberlite exploration programs. Despite geochemical commonalities of the Artemisia micro- and macro-diamonds, distinct signatures are observed, particularly in regards to enriched and depleted carbon isotopic compositions, the abundance of nitrogen and hydrogen impurities and different resorption histories. This implies that, despite an observed lognormal size frequency distribution for Artemisia diamonds, micro- and macro-diamonds do not represent a single population at this locality.

The separation of natural and anthropogenically caused climatic changes is an important task of contemporary climate research. For this purpose, a detailed knowledge of the natural variability of the climate during warm stages is a necessary prerequisite. Beside model simulations and historical documents, this knowledge is mostly derived from analyses of so-called climatic proxy data like tree rings or sediment as well as ice cores. In order to be able to appropriately interpret such sources of palaeoclimatic information, suitable approaches of statistical modelling as well as methods of time series analysis are necessary, which are applicable to short, noisy, and non-stationary uni- and multivariate data sets. Correlations between different climatic proxy data within one or more climatological archives contain significant information about the climatic change on longer time scales. Based on an appropriate statistical decomposition of such multivariate time series, one may estimate dimensions in terms of the number of significant, linear independent components of the considered data set. In the presented work, a corresponding approach is introduced, critically discussed, and extended with respect to the analysis of palaeoclimatic time series. Temporal variations of the resulting measures allow to derive information about climatic changes. For an example of trace element abundances and grain-size distributions obtained near the Cape Roberts (Eastern Antarctica), it is shown that the variability of the dimensions of the investigated data sets clearly correlates with the Oligocene/Miocene transition about 24 million years before present as well as regional deglaciation events. Grain-size distributions in sediments give information about the predominance of different transportation as well as deposition mechanisms. Finite mixture models may be used to approximate the corresponding distribution functions appropriately. In order to give a complete description of the statistical uncertainty of the parameter estimates in such models, the concept of asymptotic uncertainty distributions is introduced. The relationship with the mutual component overlap as well as with the information missing due to grouping and truncation of the measured data is discussed for a particular geological example. An analysis of a sequence of grain-size distributions obtained in Lake Baikal reveals that there are certain problems accompanying the application of finite mixture models, which cause an extended climatological interpretation of the results to fail. As an appropriate alternative, a linear principal component analysis is used to decompose the data set into suitable fractions whose temporal variability correlates well with the variations of the average solar insolation on millenial to multi-millenial time scales. The abundance of coarse-grained material is obviously related to the annual snow cover, whereas a significant fraction of fine-grained sediments is likely transported from the Taklamakan desert via dust storms in the spring season. / Die Separation natürlicher und anthropogen verursachter Klimaänderungen ist eine bedeutende Aufgabe der heutigen Klimaforschung. Hierzu ist eine detaillierte Kenntnis der natürlichen Klimavariabilität während Warmzeiten unerlässlich. Neben Modellsimulationen und historischen Aufzeichnungen spielt hierfür die Analyse von sogenannten Klima-Stellvertreterdaten eine besondere Rolle, die anhand von Archiven wie Baumringen oder Sediment- und Eisbohrkernen erhoben werden. Um solche Quellen paläoklimatischer Informationen vernünftig interpretieren zu können, werden geeignete statistische Modellierungsansätze sowie Methoden der Zeitreihenanalyse benötigt, die insbesondere auf kurze, verrauschte und instationäre uni- und multivariate Datensätze anwendbar sind. Korrelationen zwischen verschiedenen Stellvertreterdaten eines oder mehrerer klimatologischer Archive enthalten wesentliche Informationen über den Klimawandel auf großen Zeitskalen. Auf der Basis einer geeigneten Zerlegung solcher multivariater Zeitreihen lassen sich Dimensionen schätzen als die Zahl der signifikanten, linear unabhängigen Komponenten des Datensatzes. Ein entsprechender Ansatz wird in der vorliegenden Arbeit vorgestellt, kritisch diskutiert und im Hinblick auf die Analyse von paläoklimatischen Zeitreihen weiterentwickelt. Zeitliche Variationen der entsprechenden Maße erlauben Rückschlüsse auf klimatische Veränderungen. Am Beispiel von Elementhäufigkeiten und Korngrößenverteilungen des Cape-Roberts-Gebietes in der Ostantarktis wird gezeigt, dass die Variabilität der Dimension der untersuchten Datensätze klar mit dem Übergang vom Oligozän zum Miozän vor etwa 24 Millionen Jahren sowie regionalen Abschmelzereignissen korreliert. Korngrößenverteilungen in Sedimenten erlauben Rückschlüsse auf die Dominanz verschiedenen Transport- und Ablagerungsmechanismen. Mit Hilfe von Finite-Mixture-Modellen lassen sich gemessene Verteilungsfunktionen geeignet approximieren. Um die statistische Unsicherheit der Parameterschätzung in solchen Modellen umfassend zu beschreiben, wird das Konzept der asymptotischen Unsicherheitsverteilungen eingeführt. Der Zusammenhang mit dem Überlapp der einzelnen Komponenten und aufgrund des Abschneidens und Binnens der gemessenen Daten verloren gehenden Informationen wird anhand eines geologischen Beispiels diskutiert. Die Analyse einer Sequenz von Korngrößenverteilungen aus dem Baikalsee zeigt, dass bei der Anwendung von Finite-Mixture-Modellen bestimmte Probleme auftreten, die eine umfassende klimatische Interpretation der Ergebnisse verhindern. Stattdessen wird eine lineare Hauptkomponentenanalyse verwendet, um den Datensatz in geeignete Fraktionen zu zerlegen, deren zeitliche Variabilität stark mit den Schwankungen der mittleren Sonneneinstrahlung auf der Zeitskala von Jahrtausenden bis Jahrzehntausenden korreliert. Die Häufigkeit von grobkörnigem Material hängt offenbar mit der jährlichen Schneebedeckung zusammen, während feinkörniges Material möglicherweise zu einem bestimmten Anteil durch Frühjahrsstürme aus der Taklamakan-Wüste herantransportiert wird.

In den letzten Jahren ist eine Vielzahl neuer organischer Polymere und niedermolekularer Verbindungen synthetisiert worden, die sich als aktive Komponente für Elektrolumineszenz-Bauelemente und andere elektronische Anwendungen eignen. Trotz der großen technologischen Bedeutung und des erheblichen Fortschrittes, der bei der Herstellung solcher Materialien erzielt worden ist, sind grundlegende physikalische Eigenschaften dieser Materialklassen noch nicht ausreichend erforscht. Insbesondere das Auftreten lokalisierter Zustände innerhalb der Bandlücke hat besondere Bedeutung für ihre elektronischen Eigenschaften. Sowohl die Präsenz dieser flachen traps (Fallen, Löcher) als auch der Einfluß der Herstellungsbedingungen auf die tiefen und flachen lokalisierten Zustände wurden bisher nicht systematisch untersucht.<br /> Thermische Techniken sind wichtige Methoden, um lokalisierte Niveaus in organischen und anorganischen Materialien zu erforschen. Themisch-Stimulierte Lumineszenz (TSL), Thermisch-Stimulierte Ströme (TSC) und Thermisch-Stimulierte Depolarisierte Ströme (TSDC) ermöglichen die Untersuchung flacher und tiefer traps; in Verbindung mit DiElektrischer Spektroskopie (DES) können außerdem Polarisations- und Depolarisationseffekte studiert werden.<br /> Mit Hilfe numerischer Simulationen haben wir die kinetischen Gleichungen erster und zweiter Ordnung untersucht, die sich durch schwaches bzw. starkes Wieder-Fangen beschreiben lassen. In diesen Gleichungen haben wir Gaussian-, exponentielle und quasi-kontinuierliche Verteilungen von lokalisierten Zustände berücksichtigt. Durch Veränderung der beiden wichtigsten Parameter (Tiefe der traps E und Häufigkeit) konnte die Form der thermischen Maxima untersucht werden. Auch die die Gaussian-Verteilung bestimmenden Faktoren wurden verändert. <br /> Diese theoretischen Ergebnisse wurden auf die experimentellen Glow-Kurven angewandt. Dünne Filme aus polymeren und niedermolekularen Verbindungen (Polyphenylquinoxaline, Trisphenylquinoxaline und Oxadiazole), die wegen ihrer technologischen Bedeutung ausgewählt wurden, zeigen komplexes thermisches Verhalten.<br /> Insbesondere hoch geordnete Filme eines amphiphil substituierten 2-(p-nitrophenyl)-5-(p-undecylamidophenyl)-1,3,4-oxadiazols (NADPO) zeichnen sich durch komplexe TSL-Diagramme aus. Im Bereich von Em = 4 meV wurde eine Region flacher traps gefunden. Zwei weitere TSL-Maxima treten bei Tm = 221.5 K bzw. Tm = 254.2 K auf. Sie besitzen Aktivierungsenergien von Em= 0.63 eV bzw. Em = 0.66 eV, ihre Frequenzfaktoren betragen s = 2.4x1012 s-1 bzw. s = 1.85x1011 s-1, sie zeigen Breiten der Verteilung von s = 0.045 eV bzw. s = 0.088 eV. <br /> Des weiteren zeigt diese Arbeit, daß die Form der Glow-Kurven stark von der Anregungstemperatur und vom thermischen Kreislauf beeinflußt wird. / New polymers and low molecular compounds, suitable for organic light emitting devices and organic electronic applications, have been synthesised in this years in order to obtain electron transport characteristics compatible with requirements for applications in real plastic devices. However, despite of the technological importance and of the relevant progress in devices manufacture, fundamental physical properties of such class of materials are still not enough studied. In particular extensive presence of distributions of localised states inside the band gap has a deep impact on their electronic properties. Such presence of shallow traps as well as the influence of the sample preparation conditions on deep and shallow localised states have not been, until now, systematically explored.<br /> The thermal techniques are powerful tools in order to study localised levels in inorganic and organic materials. Thermally stimulated luminescence (TSL), thermally stimulated currents (TSC) and thermally stimulated depolarisation currents (TSDC) allow to deeply look to shallow and deep trap levels as well as they permit to study, in synergy with dielectric spectroscopy (DES), polarisation and depolarisation effects. <br /> We studied, by means of numerical simulations, the first and the second order kinetic equations characterised by negligible and strong re-trapping respectively. We included in the equations Gaussian, exponential and quasi-continuous distributions of localised states. The shapes of the theoretical peaks have been investigated by means of systematic variation of the two main parameters of the equations, i. e. the energy trap depth E and the frequency factor a and of the parameters regulating the distributions, in particular for a Gaussian distribution the distribution width s and the integration limits. <br /> The theoretical findings have been applied to experimental glow curves. Thin films of polymers and low molecular compounds. Polyphenylquinoxalines, trisphenylquinoxalines and oxadiazoles, studied because of their technological relevance, show complex thermograms, having several levels of localised states and depolarisation peaks. <br /> In particular well ordered films of an amphiphilic substituted 2-(p-nitrophenyl)-5-(p-undecylamidophenyl)-1,3,4-oxadiazole (NADPO) are characterised by rich TSL thermograms. A wide region of shallow traps, localised at Em = 4 meV, has been successfully fit by means of a first order kinetic equation having a Gaussian distribution of localised states. <br /> Two further peaks, having a different origin, have been characterised. The peaks at Tm = 221.5 K and Tm = 254.2 have activation energy of Em= 0.63 eV and Em = 0.66 eV, frequency factor s = 2.4x1012 s-1 and s = 1.85x1011 s-1, distribution width s = 0.045 eV and s = 0.088 eV respectively. <br /> Increasing the number of thermal cycle, a peak, probably connected with structural defects, appears at Tm = 197.7 K. The numerical analysis of this peak was performed by means of a first order equation containing a Gaussian distribution of traps. The activation energy of the trap level is centred at Em = 0.55 eV. The distribution is perfectly symmetric with a quite small width s = 0.028 eV. The frequency factor is s = 1.15 x 1012 s-1, resulting of the same order of magnitude of its neighbour peak at Tm = 221.5 K, having both, probably, the same origin.<br /> Furthermore the work demonstrates that the shape of the glow curves is strongly influenced by the excitation temperature and by the thermal cycles. For that reason Gaussian distributions of localised states can be confused with exponential distributions if the previous thermal history of the samples is not adequately considered.