Titanium alloys are strong candidates for the aerospace industry and biomaterial applications because of their low density, high strength-to-weight ratio and very high strength even at temperatures up to 600°C. Like many other engineering alloys, titanium alloys are prone to strong preferred crystallographic orientation development during thermomechanical processing. Part of the titanium processing route is to heat treat the material above the β transus for the purpose of homogenization and associated phase transformation. This heat treatment dramatically affected the microstructure and texture evolution. Theoretically, such heat treatment should result in a nearly random texture if all variants during α→β→α phase transformation are active. In reality, significant textures are observed after such a heat treatment process. The present project aims at developing a detailed understanding of the root cause for this relatively strong texture by means of EBSD and in-situ neutron diffraction studies. The effect of β grain growth on variant selection during β to α phase transformation has been investigated by using two variants of Ti-6Al-4V with and without 0.4 wt% yttrium addition. The aim of adding yttrium was to control β grain growth above the β transus by pinning grain boundaries with yttria. Both materials were first thermomechanically processed to generate a similar starting microstructures and crystallographic textures. Subsequently, both materials were solution heat treated above the β transus followed by slow cooling to promote growth of the α lath structure from grain boundary α. Detailed EBSD and in situ neutron diffraction analysis were carried out to study microstructure and texture evolution. The variant selection calculation suggests that more variant selection occurred in convectional material with a large grain size compared to material with yttrium addition. In situ measurements showed that β texture strengthened significantly above the β transus with increasing β grain size. There was no significant variant selection during α→β transformation; variant selection noticeably increased during β→α transformation with increasing β grain size. Additional interrupted cooling experiments followed by EBSD analysis showed early nucleation of α variants with a 'butterfly morphology' from β grain boundaries that have a pair of β grain with a common <110> pole. These observations suggest reduced nucleation energies for α formation in such circumstances allowing extensive growth of these α variants into unoccupied β grains making it a dominant variant. The influence of rolling temperatures (i.e. at 800 ºC and 950 ºC) to produce different starting texture, on texture evolution and variant selection during α→β→α transformation was also investigated. Laboratory X-ray, EBSD and in-situ neutron diffraction texture analyses were carried out. Even though the transformation texture is stronger at 800 ºC, the degree of variant selection is stronger in materials rolled at950 ºC compared to material rolled at 800 ºC. Here, the enhanced variant selectionfor the material rolled at 950 ºC was related to the different β texture. It is suggested that the combination of a particular β texture components promote variant nucleation that can increase the likelihood of having β grain pairs with a common <110> pole.

Image segmentation is an important step in image processing. A traditional way how to segment an image is a texture-based segmentation that uses texture features to describe image texture. In this work, Local Binary Patterns (LBP) are used for image texture representation. Texture feature is a histogram of occurences of LBP codes in a small image window. The work also aims to comparison of results of various modifications of Local Binary Patterns and their usability in the image segmentation which is done by unsupervised clustering of texture features. The Fuzzy C-Means algorithm is finally used for the clustering in this work.

La reconstruction 3D multi-vue atteint désormais un niveau de maturité industrielle : des utilisateurs non-experts peuvent produire des modèles 3D large-échelle de qualité à l'aide de logiciels commerciaux. Ces reconstructions utilisent des capteurs haut de gamme comme des LIDAR ou des appareils photos de type DSLR, montés sur un trépied et déplacés autour de la scène. Ces protocoles d'acquisition sont mal adaptés à l’inspection d’infrastructures de grande taille, à géométrie complexe. Avec l'évolution rapide des capacités des micro-drones, il devient envisageable de leur confier ce type de tâche. Un tel choix modifie les données d’acquisition : on passe d’un ensemble restreint de photos de qualité, soigneusement acquises par l’opérateur, à une séquence d'images à cadence vidéo, sujette à des variations de qualité image dues, par exemple, au bougé et au défocus.Les données vidéo posent problème aux logiciels de photogrammétrie du fait de la combinatoire élevée engendrée par le grand nombre d’images. Nous proposons d’exploiter l’intégralité des images en deux étapes. Au cours de la première, la reconstruction 3D est obtenue en sous-échantillonnant temporellement la séquence, lors de la seconde, la restitution haute résolution de texture est obtenue en exploitant l'ensemble des images. L'intérêt de la texture est de permettre de visualiser des détails fins du modèle numérisé qui ont été perdus dans le bruit géométrique de la reconstruction. Cette augmentation de qualité se fait via des techniques de Super Résolution (SR).Pour atteindre cet objectif nous avons conçu et réalisé une chaîne algorithmique prenant, en entrée, la séquence vidéo acquise et fournissant, en sortie, un modèle 3D de la scène avec une texture sur-résolue. Cette chaîne est construite autour d’un algorithme de reconstruction 3D multi-vues de l’état de l’art pour la partie géométrique.Une contribution centrale de notre chaîne est la méthode de recalage employée afin d’atteindre la précision sub-pixellique requise pour la SR. Contrairement aux données classiquement utilisées en SR, nos prises de vues sont affectées par un mouvement 3D, face à une scène à géométrie 3D, ce qui entraîne des mouvements image complexes. La précision intrinsèque des méthodes de reconstruction 3D est insuffisante pour effectuer un recalage purement géométrique, ainsi nous appliquons un raffinement supplémentaire par flot optique. Le résultat de cette méthode de restitution de texture SR est d'abord comparée qualitativement à une approche concurrente de l’état de l’art.Ces appréciations qualitatives sont renforcées par une évaluation quantitative de qualité image. Nous avons à cet effet élaboré un protocole d’évaluation quantitatif de techniques de SR appliquées sur des surfaces 3D. Il est fondé sur l'utilisation de mires fractales binaires, initialement proposées par S. Landeau. Nous avons étendu ces idées au contexte de SR sur des surfaces courbes. Cette méthode est employée ici pour valider les choix de notre méthode de SR, mais elle s'applique à l'évaluation de toute texturation de modèle 3D.Enfin, les surfaces spéculaires présentes dans les scènes induisent des artefacts au niveau des résultats de SR en raison de la perte de photoconsistence des pixels au travers des images à fusionner. Pour traiter ce problème nous avons proposé deux méthodes correctives permettant de recaler photométriquement nos images et restaurer la photoconsistence. La première méthode est basée sur une modélisation des phénomènes d’illumination dans un cas d'usage particulier, la seconde repose sur une égalisation photométrique locale. Les deux méthodes testées sur des données polluées par une illumination variable s'avèrent effectivement capables d'éliminer les artefacts. / Multi-view 3D reconstruction techniques have reached industrial level maturity : non-expert users are now able to use commercial software to produce quality, large scale, 3D models. These reconstructions use top of the line sensors such as LIDAR or DSLR cameras, mounted on tripods and moved around the scene. Such protocols are not designed to efficiently inspect large infrastructures with complex geometry. As the capabilities of micro-drones progress at a fast rate, it is becoming possible to delegate such tasks to them. This choice induces changes in the acquired data : rather than a set of carefully acquired images, micro-drones will produce a video sequence with varying image quality, due to such flaws as motion blur and defocus. Processing video data is challenging for photogrammetry software, due to the high combinatorial cost induced by the large number of images. We use the full image sequence in two steps. Firstly, a 3D reconstruction is obtained using a temporal sub-sampling of the data, then a high resolution texture is built from the full sequence. Texture allows the inspector to visualize small details that may be lost in the noise of the geometric reconstruction. We apply Super Resolution techniques to achieve texture quality augmentation. To reach this goal we developed an algorithmic pipeline that processes the video input and outputs a 3D model of the scene with super resolved texture. This pipeline uses a state of the art 3D reconstruction software for the geometric reconstruction step. The main contribution of this pipeline is the image registration method used to achieve the sub-pixel accuracy required for Super Resolution. Unlike the data on which Super Resolution is generally applied, our viewpoints are subject to relative 3D motion and are facing a scene with 3D geometry, which makes the motion field all the more complex. The intrinsic precision of current 3D reconstruction algorithms is insufficient to perform a purely geometric registration. Instead we refine the geometric registration with an optical flow algorithm. This approach is qualitatively to a competing state of the art method. qualitative comparisons are reinforced by a quantitative evaluation of the resulting image quality. For this we developed a quantitative evaluation protocol of Super Resolution techniques applied to 3D surfaces. This method is based on the Binary Fractal Targets proposed by S. Landeau. We extended these ideas to the context of curved surfaces. This method has been used to validate our choice of Super Resolution algorithm. Finally, specularities present on the scene surfaces induce artefacts in our Super Resolution results, due to the loss of photoconsistency among the set of images to be fused. To address this problem we propose two corrective methods designed to achieve photometric registration of our images and restore photoconsistency. The first method is based on a model of the illumination phenomena, valid in a specific setting, the second relies on local photometric equalization among the images. When tested on data polluted by varying illumination, both methods were able to eliminate these artefacts.

Dénommée multiparamétrique par adjonction de séquences fonctionnelle aux conventionnelles, L’IRM de prostate a montré ses performances pour la détection du cancer de prostate par un score radiologique visuel, subjectif. D’autres applications sont en cours d’investigations comme la stratification, le planning thérapeutique ou encore le suivi oncologique.La première partie s’attache à décrire, élaborer et appliquer une méthodologie de recalage non rigide en 3D entre l’histologie du spécimen de prostatectomie totale et les différentes séquences de l’IRMmp. Après avoir capturé une déformation et un changement de volume de la prostate entre les états in vivo et ex vivo par IRM, la méthode de recalage multimodalité appliquée à une population de prostatectomie totale précédée d’une IRM démontre une sous-estimation du volume de cancer par l’IRM, sujette à une stratification. Les implications se trouvent dans la détection, la stratification et le planning thérapeutique. La deuxième partie propose une analyse de texture des différentes séquences et cartographies quantitatives en diffusion et perfusion pour la détection et la stratification du cancer. Cette approche multiparamétrique de « Score d’Entropie » est testée dans une population pilote au moment des biopsies et présente des performances diagnostique pour sélectionner les lésions à biopsier. Ce score d’entropie participe de la stratification du cancer en corrélant positivement avec le score de Gleason et la longueur de cancer biopsique.La troisième partie explore le rôle de l’IRM dans le suivi d’une thérapie émergente, dite focale, du cancer. Il s’agit d’un travail de recalage non-rigide longitudinal sur une cohorte de patients traités par thérapie focale en vue de compenser les déformation focalement induites. Il apparaît que ce type de recalage peut permettre un suivi objectif des résultats d’ablation et potentiellement élaborer une cible biopsique et radiologique dans le suivi oncologique. / Conventional prostate MRI, enhanced by diffusion and perfusion sequences, and then named multiparametric, showed high performances for detection of prostate cancer using visual scoring. Indications in stratification, prognosis, treatment planning and follow up are currently under investigations.First part of this work attached itself to describe, elaborate and use a non-rigid image fusion method in 3D between gold standard histology of radical prostatectomy and MRI. Investigations captured the significant differences in shape and volume of in vivo and ex vivo prostate using MRI. The developed multimodality fusion method was applied to a cohort of patients who underwent MRI prior surgery. Results showed a stratified underestimation of cancer volume by MRI. Clinical output resides in detection, stratification and surgical planning.The second part proposed some texture analysis of sequences and quantitative maps. As a multiparametric approach, the Entropy Score is applied in a pilot cohort at time of biopsy and showed some potential usefulness to select MRI targets without compromising detection of significant cancer. By positively correlating with the Gleason Score and the maximal core length of cancer, Entropy Score participates of stratification of cancer.The third part explored application of image registration in the longitudinal follow up of an emergent therapy, said focal (FT). As a conservative approach, FT induces very local deformation of the gland which appears to be appropriately modelled by non-rigid registration, then opening possibilities to guide further control biopsy and radiologic assessment.

Les aciers renforcés par dispersion d’oxydes (ODS) sont envisagés comme matériaux de gainage pour les réacteurs à neutrons rapides au sodium. Ils présentent de très bonnes propriétés mécaniques, notamment en fluage, du fait de la présence des nano-renforts ainsi qu’une résistance au gonflement élevée de par leur matrice ferrito-martensitique. La gamme de fabrication des ODS ferritiques est complexe et génère une forte texture morphologique et cristallographique. Les microstructures doivent donc être recristallisées mais cette recristallisation est très souvent hétérogène et difficile à maitriser. Cette étude vise à améliorer la compréhension des mécanismes de recristallisation qui dépend fortement du taux de renforts incorporé et de la déformation appliquée. Dans ce but, l’élaboration de nuances modèles avec différents taux de renforts a été réalisée et des caractérisations multi-échelles (EBSD, DNPA, DSC) ont été effectuées. Il est apparu que la nature chimique des nano-oxydes influence fortement la stabilité des microstructures et leur capacité à restaurer après consolidation. Ces travaux ont également permis de mettre en évidence l’étroite relation entre le chemin de déformation à froid, la distribution de l’énergie stockée et la microstructure après recristallisation. Lorsque le traitement thermomécanique est optimal, des microstructures recristallisées homogènes et peu texturées peuvent être obtenues. Il est aussi apparu que la diminution du taux de renforts ne permet pas de faciliter clairement la recristallisation des ODS. Le développement d’un modèle numérique Monte Carlo pour simuler la croissance de grains et la recristallisation en présence de particules a permis d’aider à la compréhension des mécanismes. De plus, une analyse des propriétés en traction des différents états métallurgiques a montré que la recristallisation améliorait nettement la ductilité à chaud sans diminuer significativement les résistances mécaniques. Les résultats obtenus sur matériaux modèles ont permis de mieux comprendre les mécanismes mis en jeu lors de la recristallisation des ODS et d’analyser les gammes de fabrication actuelles des tubes ODS. Diverses optimisations possibles sont discutées à la lumière des résultats obtenus. / Oxide dispersion strengthened steels are considered as cladding materials for Sodium Fast Reactor. Due to the nano-precipitates incorporated within the microstructure, their mechanical properties are very high, especially under creep loading. Moreover, the ferritic-martensitic lattice gives ODS steels a remarkable swelling resistance under radiation. The manufacturing route of ferritic ODS steels is complex and induces a strong morphologic and crystallographic texture. The microstructure needs to be recrystallized but this recrystallization is often heterogeneous and difficult to control. In this framework, this investigation aims at achieving a better understanding of the recrystallization mechanisms in ODS steels that strongly rely both on the nano-oxydes concentration and the applied deformation. Thereby, several model grades with various concentrations of oxydes were elaborated and characterized at different scales (EBSD, SANS, DSC). It appears that the chemical composition of the nano-oxydes strongly influence the microstructure stability and its ability to recover after consolidation. Throughout this study, the strong relationship between the cold-deformation path, the stored energy distribution and the recrystallized microstructure has been highlighted. When the thermo-mechanical treatment is optimal, homogeneous recrystallized microstructures weakly textured can be achieved. Experiments also showed that decreasing the nano-oxydes concentration does not clearly favor recrystallization. Monte Carlo modelling was performed to simulate grain growth and recrystallization in the presence of second phase particles and the simulations helped to identify the mechanisms that occur experimentally. Furthermore, the tensile properties of different metallurgical states were investigated and it came out that the recrystallization notably increases the ductility at high temperature without significantly affecting the yield strength. The results obtained on model materials allowed to achieve a better understanding of the mechanisms implied in the recrystallization process of ODS steels. Several possible optimizations to the current manufacturing route are discussed from the conclusions of this study.

Les pratiques d’agriculture et de production alimentaire évoluent de façon profonde afin de proposer une offre alimentaire permettant de répondre aux enjeux de la croissance démographique, et de satisfaire des critères de santé et de durabilité. Cependant, le succès des produits alimentaires qui découlent de ces mutations repose avant tout sur leur appréciation sensorielle par le consommateur. En particulier, les perceptions de texture jouent un rôle majeur dans les choix et préférences des consommateurs.Ces travaux de thèse s’insèrent dans la problématique générale de la compréhension des mécanismes de perception de texture qui découlent des interactions mécaniques entre la langue, l’aliment et le palais durant le processus oral. L’objectif de la thèse réside dans le développement d’une méthode nouvelle et originale, basée sur les ultrasons quantitatifs, pour l’évaluation non-destructive, non-invasive et en temps réel des propriétés mécaniques du système langue-aliment-palais.Un montage bio-mimétique a été développé au laboratoire afin de reproduire en conditions contrôlées, in vitro, le comportement du système langue-palais durant la compression d’un aliment. Ce système a permis de progressivement prendre en compte la complexité de l’environnement physiologique, en intégrant notamment les caractéristiques liées à la rugosité, à la lubrification et à la déformabilité de la langue. Un ensemble de gels modèles d’agar et de gélatine a été développé, permettant d’obtenir une gamme de produits alimentaires avec des caractéristiques physiques et sensorielles contrastées. Un dispositif expérimental à ultrasons quantitatifs a ensuite été conçu afin de mesurer la réponse échographique du système langue-aliment-palais au cours de compressions uni-axiales imposées par le système biomimétique sur les aliments modèles. Des techniques de traitement du signal et des outils de simulation de propagation d’ondes ont permis de développer des indicateurs quantitatifs fournissant des informations sur les mécanismes physiques observés aux interfaces du système langue-aliment-palais. Deux paramètres ultrasonores ont alors été identifiés : i) la réflectivité de l’interface entre la langue et l’aliment, qui permet de caractériser le contact intime entre l’aliment et les aspérités de la surface de la langue, et ii) le temps de vol des ondes ultrasonores dans l’aliment, qui traduit l’état de déformation subi par ce dernier. Ces propriétés issues des signaux ultrasonores ont été confrontées à des données d’analyse sensorielle évaluées grâce à un panel de juges entraînés sur les aliments modèles sélectionnés. Les attributs sensoriels « humide » et « glissant » ont par exemple pu être mis en relation avec les mesures de réflectivité ultrasonore à l’interface langue-aliment, elles-mêmes liées à l’état de lubrification de l’interface.Ces travaux de thèse ont ainsi permis de montrer le potentiel des méthodes ultrasonores pour mieux comprendre et caractériser, en conditions in vitro contrôlées, des phénomènes physiques qui peuvent jouer un rôle important dans les mécanismes de perception de texture de l’aliment. Ils permettront d’ouvrir la voie aux perspectives de transfert de la technologie en conditions in vivo, directement sur le consommateur. / Agriculture and agri-food industry practices evolve in a deep way to supply a new food offer allowing to address population growth issues, and to meet health and sustainability criteria. However, the success of commercial products emerging from these transformations highly depends on sensory appreciation by the consumer. In particular, texture perceptions have been shown to play a major role in the choices and preferences of consumers.This thesis work fits within the general issue of understanding the mechanisms of texture perception resulting from the mechanical interactions between tongue, food and palate during oral processing. The aim of this thesis lies in the development of a new and original method, based on quantitative ultrasounds, for the non-destructive, non-invasive and real time evaluation of the mechanical properties of the tongue-food-palate system.A bio-mimicking system was developed in the laboratory to reproduce in controlled in vitro conditions, the behavior of tongue-palate system during the compression of a food sample. This system allowed to consider gradually the complexity of the physiological environment, by integrating characteristics related to tongue roughness, lubrication and deformability. A set of model gels of agar and gelatin was developed to obtain a range of food products with contrasting physical and sensory characteristics. Then, a quantitative ultrasound device was designed to measure the ultrasound pulse-echo response of the tongue-food-palate system throughout a uniaxial compression achieved by the bio-mimicking apparatus on the products. Signal processing methods and wave propagation simulation tools were used to develop quantitative indicators related to physical mechanisms occurring in tongue-food system interfaces. Two ultrasonic parameters were then identified: i) the reflectivity of tongue-food interface, which makes it possible to characterize the intimate contact between food and the asperities of tongue surface, and ii) the time of flight of ultrasonic waves through the food sample, which is related to the level of deformation undergone by the sample. Finally, the properties stemming from ultrasonic signals were compared with sensory data estimated on the set of food samples, thanks to a panel of trained judges. “Moist” and “slippery” sensory attributes were for instance shown to correlate with ultrasound reflectivity measurements, which are themselves related to lubrication conditions at tongue-food interface.This thesis work allowed to show the potential of ultrasonic methods to better understand and characterize, in controlled in vitro conditions, physical phenomena which may play an important role in the mechanisms of food texture perception. This work must pave the way for future investigations on the transfer of the technology to in vivo conditions, directly on the consumer.

<p> A new parameter-free texture feature-based seeded region growing algorithm is proposed in this dissertation for automated segmentation of organs in abdominal MR images. This algorithm requires that a user only mouse clicks twice to identify the upper left and lower right corners of a rectangular region of interest (ROI). With this given ROI, a seed point is automatically selected based on homogeneity criteria. Intensity as well as four texture features: 20 cooccurrence texture features, Gabor texture feature, and both 20 and 3D semivariogram texture features are extracted from the image and a seeded region growing algorithm is performed on these feature spaces. A threshold is then obtained by taking a lower value just before the one which results in an ' explosion '. An optional Snake post-processing tool is also provided to obtain better organ delineation. The comparative results of the texture features and intensity are reported using both normal digital images and abdominal MR images acquired from ten patients. Comparisons of Before and After Snake are also presented. Generally, Gabor texture feature is found to perform the best among all features . The experimental results of the proposed approach show that it is fast and accurate when combined with Gabor texture feature or intensity feature and should prove a boon to production radiological batch processing. </p> / Thesis / Doctor of Philosophy (PhD)

No description available.

Thèse numérisée par la Division de la gestion de documents et des archives de l'Université de Montréal.

Texture analysis is an active area of research and plays an important role in computer vision applications. Texture, along with color and shape, contains important features of an image. Texture analysis allows to characterize regions inside an image by using descriptors. These descriptors are applied in the study of texture classification, in which the goal is to identify features that characterize a particular texture and assign a label to an image based on these features. Because of the importance of texture analysis in computer vision, researchers are continually devising and developing new descriptors, with the aim to improve the discriminative power of texture features of an image. A difficult task in texture analysis is to compare these descriptors and verify which are the most suitable for each type of image. The lack of a good review and comparison of descriptors cause that some applications do not use the most appropriate descriptor for a specific type of texture. Therefore, in this dissertation it was developed a research and collaboration platform for the analysis and comparison of texture descriptors and texture datasets. The platform aims to support the researchers in the area of texture analysis, specifically in texture classification. The platform was useful to perform an extensive comparison of texture descriptors and various texture datasets. Using the platform, in some datasets the results produced were better than those previously found in the literature. The results indicate that the classification accuracy varies according to the descriptor and classifier employed. By varying the parameters of texture descriptors it was possible to get different, yet better, classification accuracies. / A análise de textura é uma área ativa de pesquisa que desempenha um papel importante em aplicações de visão computacional. A textura, juntamente com a cor e a forma, contém características importantes de uma imagem. A análise de textura permite caracterizar regiões dentro de uma imagem usando descritores. Esses descritores são aplicados no estudo de classificação de texturas, no qual o objetivo é identificar características que distingam uma determinada textura e atribuir um rótulo a uma imagem baseada nessas características. Devido à importância da análise de textura na visão computacional, os pesquisadores estão continuamente criando e desenvolvendo novos descritores, com o objetivo de melhorar o poder discriminativo dessas características em uma imagem. Uma tarefa difícil na análise de textura é comparar esses descritores e verificar quais são os mais adequados para cada tipo de tipo de imagem. A falta de uma boa revisão e comparação de descritores de textura pode fazer com que algumas aplicações não utilizem o descritor mais adequado para um tipo específico de textura. Portanto, nesta dissertação foi desenvolvida uma plataforma de pesquisa e colaboração para a análise e comparação de descritores de textura e conjuntos de dados de textura. A plataforma visa apoiar os pesquisadores na área de análise de textura, especificamente na classificação de texturas. A plataforma foi útil para realizar uma comparação extensiva de descritores de textura e vários conjuntos de dados de textura. Com essa plataforma, em alguns conjuntos de dados os resultados encontrados foram melhores que aqueles encontrados anteriormente na literatura. Os resultados indicam que a acurácia de classificação muda segundo o descritor e o classificador usado. Mudando os valores dos parâmetros dos descritores de textura foi possível obter acurácias diferentes e até melhores.