71 |
Parameter choice in Banach space regularization under variational inequalitiesHofmann, Bernd, Mathé, Peter 17 April 2012 (has links) (PDF)
The authors study parameter choice strategies for Tikhonov regularization of nonlinear ill-posed problems in Banach spaces. The effectiveness of any parameter choice for obtaining convergence rates depend on the interplay of the solution smoothness and the nonlinearity structure, and it can be expressed concisely in terms of variational inequalities. Such inequalities are link conditions between the penalty term, the norm misfit and the corresponding error measure. The parameter choices under consideration include an a priori choice, the discrepancy principle as well as the Lepskii principle. For the convenience of the reader the authors review in an appendix a few instances where the validity of a variational inequality can be established.
|
72 |
Reconnaissance des émotions par traitement d’images / Emotions recognition based on image processingGharsalli, Sonia 12 July 2016 (has links)
La reconnaissance des émotions est l'un des domaines scientifiques les plus complexes. Ces dernières années, de plus en plus d'applications tentent de l'automatiser. Ces applications innovantes concernent plusieurs domaines comme l'aide aux enfants autistes, les jeux vidéo, l'interaction homme-machine. Les émotions sont véhiculées par plusieurs canaux. Nous traitons dans notre recherche les expressions émotionnelles faciales en s'intéressant spécifiquement aux six émotions de base à savoir la joie, la colère, la peur, le dégoût, la tristesse et la surprise. Une étude comparative de deux méthodes de reconnaissance des émotions l'une basée sur les descripteurs géométriques et l'autre basée sur les descripteurs d'apparence est effectuée sur la base CK+, base d'émotions simulées, et la base FEEDTUM, base d'émotions spontanées. Différentes contraintes telles que le changement de résolution, le nombre limité d'images labélisées dans les bases d'émotions, la reconnaissance de nouveaux sujets non inclus dans la base d'apprentissage sont également prises en compte. Une évaluation de différents schémas de fusion est ensuite réalisée lorsque de nouveaux cas, non inclus dans l'ensemble d'apprentissage, sont considérés. Les résultats obtenus sont prometteurs pour les émotions simulées (ils dépassent 86%), mais restent insuffisant pour les émotions spontanées. Nous avons appliqué également une étude sur des zones locales du visage, ce qui nous a permis de développer des méthodes hybrides par zone. Ces dernières améliorent les taux de reconnaissance des émotions spontanées. Finalement, nous avons développé une méthode de sélection des descripteurs d'apparence basée sur le taux d'importance que nous avons comparée avec d'autres méthodes de sélection. La méthode de sélection proposée permet d'améliorer le taux de reconnaissance par rapport aux résultats obtenus par deux méthodes reprises de la littérature. / Emotion recognition is one of the most complex scientific domains. In the last few years, various emotion recognition systems are developed. These innovative applications are applied in different domains such as autistic children, video games, human-machine interaction… Different channels are used to express emotions. We focus on facial emotion recognition specially the six basic emotions namely happiness, anger, fear, disgust, sadness and surprise. A comparative study between geometric method and appearance method is performed on CK+ database as the posed emotion database, and FEEDTUM database as the spontaneous emotion database. We consider different constraints in this study such as different image resolutions, the low number of labelled images in learning step and new subjects. We evaluate afterward various fusion schemes on new subjects, not included in the training set. Good recognition rate is obtained for posed emotions (more than 86%), however it is still low for spontaneous emotions. Based on local feature study, we develop local features fusion methods. These ones increase spontaneous emotions recognition rates. A feature selection method is finally developed based on features importance scores. Compared with two methods, our developed approach increases the recognition rate.
|
73 |
Quadratic Inverse Problems and Sparsity Promoting RegularizationFlemming, Jens 16 January 2018 (has links) (PDF)
Ill-posed inverse problems with quadratic structure are introduced, studied and solved. As an example an inverse problem appearing in laser optics is solved numerically based on a new regularized inversion algorithm. In addition, the theory of sparsity promoting regularization is extended to situations in which sparsity cannot be expected and also to equations with non-injective operators.
|
74 |
Contribution à la résolution de problèmes inverses sous contraintes et application de méthodes de conception robuste pour le dimensionnement de pièces mécaniques de turboréacteurs en phase avant-projets. / Contribution to solving inverse problems under constraints and application of robust design methods for the design of mechanical parts of preliminary design stageBiret, Maëva 18 November 2016 (has links)
L'objectif de ce travail est de proposer une nouvelle démarche pour améliorer et accélérer les études de dimensionnement des pièces de turboréacteurs en avant-projets. Il s'agit de fournir une méthodologie complète pour la conception robuste sous contraintes. Cette méthodologie consiste en trois étapes : la réduction de la dimension et la méta-modélisation, la conception robuste sous contraintes puis la résolution de problèmes inverses sous contraintes. Ce sont les trois principaux sujets abordés dans cette thèse. La réduction de la dimension est un pré-traitement indispensable à toute étude. Son but est de ne conserver, pour une sortie choisie du système, que les entrées influentes. Ceci permet de réduire la taille du domaine d'étude afin de faciliter la compréhension du système et diminuer les temps de calculs des études. Les méthodes de méta-modélisations contribuent également à ces deux objectifs. L'idée est de remplacer le code de calculs coûteux par un modèle rapide à évaluer et qui représente bien la relation entre la sortie étudiée et les entrées du système. La conception robuste sous contraintes est une optimisation bi-objectifs où les différentes sources d'incertitudes du système sont prises en compte. Il s'agit, dans un premier temps, de recenser et modéliser les incertitudes puis de choisir une méthode de propagation de ces incertitudes dans le code de calculs. Ceci permet d'estimer les moments (moyenne et écart-type) de la loi de la sortie d'intérêt. L'optimisation de ces moments constitue les deux objectifs de la conception robuste. En dernier lieu, il s'agit de choisir la méthode d'optimisation multi-objectifs qui sera utilisée pour obtenir l'optimum robuste sous contraintes. La partie innovante de cette thèse porte sur le développement de méthodes pour la résolution de problèmes inverses mal posés. Ce sont des problèmes pour lesquels il peut y avoir une infinité de solutions constituant des ensembles non convexes et même disjoints. L'inversion a été considérée ici comme un complément à l'optimisation robuste dans laquelle l'optimum obtenu ne satisfaisait pas une des contraintes. Les méthodes d'inversion permettent alors de résoudre ce problème en trouvant plusieurs combinaisons des entrées qui satisfont la contrainte sous la condition de rester proche de l'optimum robuste. Le but est d'atteindre une valeur cible de la contrainte non satisfaite tout en respectant les autres contraintes du système auxquelles on ajoute la condition de proximité à l'optimum. Appliquée au dimensionnement d'un compresseur HP en avants-projets, cette méthodologie s'inscrit dans l'amélioration et l'accélération des études marquées par de nombreux rebouclages chronophages en termes de ressources informatiques et humaines. / The aim of this PhD dissertation is to propose a new approach to improve and accelerate preliminary design studies for turbofan engine components. This approach consists in a comprehensive methodology for robust design under constraints, following three stages : dimension reduction and metamodeling, robust design under constraints and finally inverse problem solving under constraints. These are the three main subjects of this PhD dissertation. Dimension reduction is an essential pre-processing for any study. Its aim is to keep only inputs with large effects on a selected output. This selection reduces the size of the domain on which is performed the study which reduces its computational cost and eases the (qualitative) understanding of the system of interest. Metamodeling also contributes to these two objectives by replacing the time-consuming computer code by a faster metamodel which approximates adequately the relationship between system inputs and the studied output. Robust design under constraints is a bi-objectives optimization where different uncertainty sources are included. First, uncertainties must be collected and modeled. Then a propagation method of uncertainties in the computation code must be chosen in order to estimate moments (mean and standard deviation) of output distribution. Optimization of these moments are the two robust design objectives. Finally, a multi-objectives optimization method has to be chosen to find a robust optimum under constraints. The development of methods to solve ill-posed inverse problems is the innovative part of this PhD dissertation. These problems can have infinitely many solutions constituting non convex or even disjoint sets. Inversion is considered here as a complement to robust design in the case where the obtained optimum doesn't satisfy one of the constraints. Inverse methods then enable to solve this problem by finding several input datasets which satisfy all the constraints and a condition of proximity to the optimum. The aim is to reach a target value of the unsatisfied constraint while respecting other system constraints and the optimum proximity condition. Applied to preliminary design of high pressure compressor, this methodology contributes to the improvement and acceleration of studies currently characterized by a numerous of loopbacks which are expensive in terms of cpu-time and human resources.
|
75 |
On the Autoconvolution Equation and Total Variation ConstraintsFleischer, G., Gorenflo, R., Hofmann, B. 30 October 1998 (has links)
This paper is concerned with the numerical analysis of the autoconvolution equation
$x*x=y$ restricted to the interval [0,1]. We present a discrete constrained least
squares approach and prove its convergence in $L^p(0,1),1<p<\infinite$ , where
the regularization is based on a prescribed bound for the total variation of admissible
solutions. This approach includes the case of non-smooth solutions possessing jumps.
Moreover, an adaption to the Sobolev space $H^1(0,1)$ and some remarks on monotone
functions are added. The paper is completed by a numerical case study concerning
the determination of non-monotone smooth and non-smooth functions x from the autoconvolution
equation with noisy data y.
|
76 |
Realization of source conditions for linear ill-posed problems by conditional stabilityHofmann, Bernd, Yamamoto, Masahiro 19 May 2008 (has links)
We prove some sufficient conditions for obtaining convergence rates in regularization
of linear ill-posed problems in a Hilbert space setting and show that these
conditions are directly related with the conditional stability in several concrete inverse
problems for partial differential equations.
|
77 |
Optimal rates for Lavrentiev regularization with adjoint source conditionsPlato, Robert, Mathé, Peter, Hofmann, Bernd January 2016 (has links)
There are various ways to regularize ill-posed operator equations in Hilbert space. If the underlying operator is accretive then Lavrentiev regularization (singular perturbation) is an immediate choice. The corresponding convergence rates for the regularization error depend on the given smoothness assumptions, and for general accretive operators these may be both with respect to the operator or its adjoint. Previous analysis revealed different convergence rates, and their optimality was unclear, specifically for adjoint source conditions. Based on the fundamental study by T. Kato, Fractional powers of dissipative operators. J. Math. Soc. Japan, 13(3):247--274, 1961, we establish power type convergence rates for this case. By measuring the optimality of such rates in terms on limit orders we exhibit optimality properties of the convergence rates, for general accretive operators under direct and adjoint source conditions, but also for the subclass of nonnegative selfadjoint operators.
|
78 |
Regularization of an autoconvolution problem occurring in measurements of ultra-short laser pulsesGerth, Daniel 26 September 2011 (has links)
Introducing a new method for measureing ultra-short laser pulses, the research group "Solid State Light Sources" of the Max Born Institute for Nonlinear Optics and Short Pulse Spectroscopy, Berlin, encountered a new type of autoconvolution problem. The so called SD-SPIDER method aims for the reconstruction of the real valued phase of a complex valued laser pulse from noisy measurements. The measurements are also complex valued and additionally influenced by a device-related kernel function. Although the autoconvolution equation has been examined intensively in the context of inverse problems, results for complex valued functions occurring as solutions and right-hand sides of the autoconvolution equation and for nontrivial kernels were missing. The thesis is a first step to bridge this gap. In the first chapter, the physical background is explained and especially the autoconvolution effect is pointed out. From this, the mathematical model is derived, leading to the final autoconvolution equation. Analytical results are given in the second chapter. It follows the numerical treatment of the problem in chapter three. A regularization approach is presented and tested with artificial data. In particular, a new parameter choice rule making use of a specific property of the SD-SPIDER method is proposed and numerically verified. / Bei der Entwicklung einer neuen Methode zur Messung ultra-kurzer Laserpulse stieß die Forschungsgruppe "Festkörper-Lichtquellen" des Max-Born-Institutes für Nichtlineare Optik und Kurzzeitspektroskopie, Berlin, auf ein neuartiges Selbstfaltungsproblem. Die so genannte SD-SPIDER-Methode dient der Rekonstruktion der reellen Phase eines komplexwertigen Laserpulses mit Hilfe fehlerbehafteter Messungen. Die Messwerte sind ebenfalls komplexwertig und zusätzlich beeinflusst von einer durch das Messprinzip erzeugten Kernfunktion. Obwohl Selbstfaltungsgleichungen intensiv im Kontext Inverser Probleme untersucht wurden, fehlen Resultate für komplexwertige Lösungen und rechte Seiten ebenso wie für nichttriviale Kernfunktionen. Die Diplomarbeit stellt einen ersten Schritt dar, diese Lücke zu schließen. Im ersten Kapitel wird der physikalische Hintergrund erläutert und insbesondere der Selbstfaltungseffekt erklärt. Davon ausgehend wird das mathematische Modell aufgestellt. Kapitel zwei befasst sich mit der Analysis der Gleichung. Es folgt die numerische Behandlung des Problems in Kapitel drei. Eine Regularisierungsmethode wird vorgestellt und an künstlichen Daten getestet. Insbesondere wird eine neue Regel zur Wahl des Regularisierungsparameters vorgeschlagen und numerisch bestätigt, welche auf einer speziellen Eigenschaft des SD-SPIDER Verfahrens beruht.
|
79 |
Parameter choice in Banach space regularization under variational inequalitiesHofmann, Bernd, Mathé, Peter January 2012 (has links)
The authors study parameter choice strategies for Tikhonov regularization of nonlinear ill-posed problems in Banach spaces. The effectiveness of any parameter choice for obtaining convergence rates depend on the interplay of the solution smoothness and the nonlinearity structure, and it can be expressed concisely in terms of variational inequalities. Such inequalities are link conditions between the penalty term, the norm misfit and the corresponding error measure. The parameter choices under consideration include an a priori choice, the discrepancy principle as well as the Lepskii principle. For the convenience of the reader the authors review in an appendix a few instances where the validity of a variational inequality can be established.
|
80 |
Quadratic Inverse Problems and Sparsity Promoting Regularization: Two subjects, some links between them, and an application in laser opticsFlemming, Jens 11 January 2018 (has links)
Ill-posed inverse problems with quadratic structure are introduced, studied and solved. As an example an inverse problem appearing in laser optics is solved numerically based on a new regularized inversion algorithm. In addition, the theory of sparsity promoting regularization is extended to situations in which sparsity cannot be expected and also to equations with non-injective operators.
|
Page generated in 0.0561 seconds