The annealing behaviour, including studies of recrystallisation kinetics and development of crystallographic texture, of two low carbon steels after different cold rolling reductions have been investigated using Optical Microscopy (OM), Electron Back-Scatter Diffraction (EBSD) and Transmission Electron Microscopy (TEM). The primary recrystallisation behaviour of 20, 50, 70 and 90% cold rolled Interstitial Free (IF) and High Strength Low Alloy (HSLA) steels was studied. The HSLA was initially processed to give a volume fraction of about 0.2 of fine pearlite colonies, which acted as mechanically hard particles. The presence of such particles on the HSLA steel significantly reduced the temperature needed for recrystallisation by enhancing the recrystallisation and acting as nucleation sites by the Particle Stimulated Nucleation (PSN) mechanism. The inhomogeneous deformation and the local orientation changes introduced in the neighbourhood of the carbide particles (i.e. the particle deformation zone) were observed using TEM and selected area electron diffraction. The JMAK model was used to analyse the recrystallisation kinetics of the two steels. The experimental data plotted according to the JMAK model could be represented by straight lines with a JMAK exponent n falling in the range from 1.4 to 2.0. The development of crystallographic textures after cold rolling reductions and subsequent recrystallisation has been investigated. The texture development is shown to be largely dependent on the rolling reduction. With increasing rolling reduction, the annealing texture show gradual intensification of α- and γ-fibre components. However, the intensity of both fibres is weaker in the HSLA steel. Despite the dominance of PSN in HSLA steel, the recrystallisation textures were similar to that of the rolling textures with weaker intensity of both fibres. To help clarify the reasons for that, in-situ EBSD experiments of recrystallising HSLA steel deformed to 50% and 70% have been carried out. It shows that the formation of the nucleus seems to occur within the deformation zones in regions away from the particle surface leading to recrystallisation textures similar to that of the rolling textures (i.e. both α- and γ-fibre exist). The validity of this assumption has been confirmed by the use of Monte Carlomodelling. This model was used to simulate, in general way, and study the phenomenon of PSN during the recrystallisation process. The simulation shows the discontinuous evolution of the subgrains in the deformation zone to form recrystallisation nuclei around the particle. It shows also the subsequent growth of these nuclei to consume the matrix region around the particle. The simulation results are shown to match with the experimentally observed features of the recrystallisation phenomena in low carbon steel containing coarse cementite particles.

Dans le cadre de cette thèse, nous nous intéressons à la sélection des attributs en s'appuyant sur la théorie des graphes dans les différents contextes d'apprentissage non supervisé, semi-supervisé et supervisé. En particulier, nous nous intéressons aux scores de classement d'attributs basés sur des contraintes must-link et cannot-link. En effet, ces contraintes sont faciles à obtenir dans le cadre des applications réelles. Elles nécessitent juste de formuler pour deux données si elles se ressemblent et donc doivent être regroupées ensemble ou non, sans requérir d'informations détaillées sur les classes à retrouver. Les scores de contraintes ont montré de bonnes performances pour la sélection semi-supervisée des attributs. Cependant, ils sont fortement dépendants du sous-ensemble de contraintes disponibles. Nous proposons alors un score qui utilise à la fois l'ensemble des contraintes disponibles et les propriétés locales des données non contraintes. Des expériences réalisées sur des bases de données artificielles et réelles montrent que ce nouveau score est moins dépendant de l'ensemble de contraintes disponibles que les scores existants tout en atteignant des performances de classification similaires. La sélection semi-supervisée d'attributs a également été appliquée avec succès à la classification de textures couleur. En effet, parmi les nombreux attributs de texture pouvant être extraits des images couleur, il est nécessaire de sélectionner les plus pertinents afin d'améliorer la qualité de classification. / Within the framework of this thesis, we are interested in feature selection methods based on graph theory in different unsupervised, semi-supervised and supervised learning contexts. We are particularly interested in the feature ranking scores based on must-link et cannot-link constraints. Indeed, these constraints are easy to be obtained on real applications. They just require to formalize for two data samples if they are similar and then must be grouped together or not, without detailed information on the classes to be found. Constraint scores have shown good performances for semi-supervised feature selection. However, these scores strongly depend on the given must-link and cannot-link subsets built by the user. We propose then a new semi-supervised constraint scores that uses both pairwise constraints and local properties of the unconstrained data. Experiments on artificial and real databases show that this new score is less sensitive to the given constraints than the previous scores while providing similar performances. Semi supervised feature selection was also successfully applied to the color texture classification. Indeed, among many texture features which can be extracted from the color images, it is necessary to select the most relevant ones to improve the quality of classification.

Segmentation is one of the first steps in human visual system which helps us see the world around us. Humans pre-attentively segment scenes into regions of unique textures in around 10-20 ms. In this thesis, we address the problem of segmentation by grouping dense pixel-wise descriptors. Our work is based on the fact that human vision has a feed forward and a feed backward loop, where low level feature are used to refine high level features in forward feed, and higher level feature information is used to refine the low level features in backward feed. Most vision algorithms are based on a feed-forward loop, where low-level features are used to construct and refine high level features, but they don’t have the feed back loop. We have introduced ”Shape-Tailored Local Descriptors”, where we use the high level feature information (region approximation) to update low level features i.e. the descriptor, and the low level feature information of the descriptor is used to update the segmentation regions. Our ”Shape-Tailored Local Descriptor” are dense local descriptors which are tailored to an arbitrarily shaped region, aggregating data only within the region of interest. Since the segmentation, i.e., the regions, are not known a-priori, we propose a joint problem for Shape-Tailored Local Descriptors and Segmentation (regions). Furthermore, since natural scenes consist of multiple objects, which may have different visual textures at different scales, we propose to use a multi-scale approach to segmentation. We have used a set of discrete scales, and a continuum of scales in our experiments, both resulted in state-of-the-art performance. Lastly we have looked into the nature of the features selected, we tried handcrafted color and gradient channels and we have also introduced an algorithm to incorporate learning optimal descriptors in segmentation approaches. In the final part of this thesis we have introduced techniques for unsupervised learning of descriptors for segmentation. This eliminates the problem of deep learning methods where we need huge amounts of training data to train the networks. The optimum descriptors are learned, without any training data, on the go during segmentation.

Extension is a three-sectional, one-movement composition for orchestra exploring various permutations of a single motivic unit. The central priority has been to present this motive in a variety of textural situations with a harmonic accompaniment evolving from a macrotonal to a microtonal setting. Some of the devices utilized to realize this priority are mixed-instrument timbral combinations, tone clusters, multiphonics for brass and woodwinds, multiple stops for strings, and superimpositions of multiphonics. Extension is unique in two areas. First, the evolutionary progression from a macrotonal to microtonal harmonic texture is made possible by expanding the priorities of instrumental performing. Second, the use of multiphonics for full orchestra is unique to this work.

https://dc.etsu.edu/engr-1110-oer/1013/thumbnail.jpg

https://dc.etsu.edu/entc-2170-oer/1013/thumbnail.jpg

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Cette thèse introduit une méthode programmable de conception d'arrangements 2D pour les textures faites de petits éléments gémoétriques. Ces textures sont omniprésentes dans de nombreux domaines, par exemple l'illustration assistée par ordinateur. Notre approche cible les artistes-techniciens qui concevront chaque arrangement en écrivant un script. Ces scripts sont basés sur trois types d'opérateurs dédiés à la conception rapide d'arrangements stationnaires : les partitions qui décrivent l'organisation globale de l'arrangement, les associations qui controllent l'organisation locale des éléments et les combinaisons qui assemblent différents arrangements. Nous montrons que ce simple ensemble d'opérateurs est suffisant pour atteindre une variété d'arrangements bien plus grande que les méthodes précédentes. De plus les modifications du script mènent à des changements prédictibles dans l'arrangement généré, ce qui permet facilement de concevoir des structures complexes de façon itérative. Enfin, notre ensemble d'opérateurs est extensible et peut être adapté à différents besoins selon les applications. En particulier, nous amenons une contribution supplémentaire à cette nouvelle approche : nous introduisons une méthode aux performances interactives pour synthétiser des arrangements infinis de types réguliers et semi-réguliers. Cette méthode est badée sur la carte planaire pavée, une nouvelle structure de données que nous introduisons et qui contient peu mais suffisamment d'information pour représenter ces arrangements infinis. Nous montrons également comment étendre notre système pour concevoir des arrangements incluant des variations spatiales en utilisant des champs de contrôle. Nous appliquons cette extension au domaine de la cartographie, dans le cadre d'une collaboration avec l'Institut National de l'Information Géographique et Forestière (IGN-France). Le but de ce projet est de synthétiser automatiquement les motifs de hachurage représentant les zones rocheuses des montagnes dans les cartes topographiques. / This thesis introduces a programmable method to design 2D arrangements for element textures: Textures made of small geometric elements. These textures are ubiquitous in numerous applications of computer-aided illustration. Our approach targets technical artists who will design an arrangement by writing a script. Such scripts are based on three types of operators dedicated to the fast creation of stationary arrangements: Partitioning operators for defining the broad-scale organization of the arrangement, mapping operators for controlling the local organization of elements, and merging operators for mixing different arrangements. We show that this simple set of operators is sufficient to reach a much broader variety of arrangements than previous methods. Editing the script leads to predictable changes in the synthesized arrangement, which allows an easy iterative design of complex structures. Finally, our operator set is extensible and can be adapted to application-dependent needs. In particular, we introduce an additional contribution to this main framework: We present a method that demonstrates interactive performances at synthesizing (infinite) regular and semi-regular arrangements using the Tiled Planar Map, a novel data structure that contains few yet sufficient information for representing these arrangements. Finally, we show how to extend our system for designing spatially-varying textures using control fields. In particular, we present a practical application to cartography in which we collaborated with the French National Mapping Agency (IGNFrance) to automatically synthesize hatchings in rocky mountain areas of topographic maps.

Un nouveau dispositif tactile capable d'enregistrer et de reproduire des scènes tactiles à l'aide de forces tangentielles a été conçu et réalisé sur la base des limites et des exigences du système somatosensoriel humain. Les micro-déformations tangentielles du doigt coulissant sur la texture sont mesurées par le transducteur dans une bande passante de 500Hz et sont reproduites par le déplacement d'une plaque de verre sous l'action contrôlée d'un moteur électrodynamique critiquement amorti. Dans le but d'évaluer la pertinence des signaux sensoriels pour la reproduction d'une scène tactile, les grandeurs physiques qui influent sur la perception tactile humaine ont été étudiées. En utilisant une méthode d'escalier, il a été démontré que des ondelettes avec différentes combinaisons d'amplitude et de durée sont ressenties égales en intensité. Ces résultats suggèrent qu'il existe des quantités physiques communes - des invariants - pour ces signaux auques le cerveau est sensible, ce qui pourrait se rapporter à une constance perceptuelle dans l'exploration d'aspérités. En analysant les forces de frottement créées par un doigt glissant sur une surface dotées de points braille à des vitesses différents, il a été constaté que, bien que la réponse mécanique instantanée varie dans son ensemble, l'intégrale des forces tangentielles locales au cours d'une période de déformation courte reste constante. Ces enregistrements ont ensuite été classés selon leur vitesse et utilisés comme stimuli dans une tâche de comparaison. Les participants devaient glisser leur index sur une plateforme de verre qui vibrait afin de reproduire les points brailles de hauteurs différentes enregistrés à la même. / A new tactile device able to record and reproduce tactile scenes using tangential forces was designed and realized based on the limits and requirements of our somatosensory system. The tangential micro-deformations of the finger sliding on a textured surface can be measured with 500 Hz-bandwidth and reproduced by vibrating a glass plate under the controlled action of a critically damped electrodynamic actuator. In an effort to identify what sensory cues are relevant to tactile sensations for the reproduction of a scene, the physical quantities that influence tactile perception were studied. Using a staircase method, it was demonstrated that tactile wavelets with different combinations of amplitude and duration could be felt perceptually equal in intensity. These results suggested that there are common physical quantities – invariants – for these signals that the brain is sensitive to, which could relate to a perceptual constancy in asperity exploration. By analyzing the friction forces of a finger exploring braille dots with different pressures and velocities profiles, we found that although the mechanical response at a highly localized stimulus varies as a whole, the integral of the local tangential forces during a short deformation period remained constant. These recordings were then categorized by velocity and used as stimuli in a comparison task in which participants explored virtual dots of different heights at different speeds. While sliding on a glass platform which vibrated tangentially to reproduce a braille dot recorded at the same exploration speed, they were asked to report which of the two stimuli was stronger (or ‘higher’), a task that they could successfull.

Nous présentons une étude expérimentale de la dynamique d'une sphère près d'une paroi texturée dans un fluide visqueux. Le déplacement de la sphère est mesuré par interférométrie laser à haute fréquence. La texture consiste en un réseau de micro-piliers, dont on fait varier les paramètres géométriques (hauteur et concentration) sur une large gamme, afin d'étudier leur influence sur la dynamique d'approche, de collision et de rebond de la sphère. A petit nombre de Reynolds, la vitesse d'approce de la sphère est augmentée comparativement au cas d'une paroi lisse, en raison de la modification de la force de lubrification par la texture. Cette modification est décrite par plusieurs modèles hydrodynamiques. Une longueur effective de glissement au-dessus du réseau de piliers est mesurée. A nombre de Stokes modéré, l'inertie de la sphère n'est pas négligeable, et elle peut se coller ou rebondir sur les piliers. La dynamique de collage présente différents régimes qui sont modélisés en tenant compte de la modification de traînée par la texture. Le nombre de Stokes critique pour la transition de rebond est modifié par la texture. La dynamique des micro-rebonds est caractérisée. Enfin, la dynamique de contact sphère-piliers est décrite par une loi de Hertz modifiée prenant en compte la géométrie de la texture. / We investigate experimentally the dynamics of a sphere near a textured wall in a viscous fluid. High frequency laser interferometry is used to measure small displacements of the sphere. The texture is comprised of arrays of micro-pillars, whose geometrical parameters are systematically varied in order to study their influence of the dynamics of approach, collision and rebound of the sphere. Under creeping flow conditions, the sphere velocity near the textured wall is increased compared to the case of a smooth wall, as a result of the modification of the lubrication force due to the texture. We describe this modification using several hydrodynamic models. An effective slip length over the pillar array is measured. At moderate Stokes number the sphere inertia is non-negligible and the sphere can either stick or bounce off the wall. The sticking dynamics exhibit several regimes which can be described by taking into account the modification of the lubrication force by the texture. The micro-pillars modify the critical Stokes number for the bouncing transition. The micro-rebounds are characterized. Finally, the contact dynamics between the sphere and the pillars are described by a Hertz law modified by the texture geometry.