This thesis presents a novel three dimensional (3D) image-analysis method for characterizing the physical characteristics of coarse particles in the field, and introduces new methodology for the total analysis of glacial till samples. The novel image analysis method, called the GID method, is capable of determining the size, shape and surface texture of each individual clast analysed. Images of particles are taken in the field and analysis is done in the laboratory. Therefore the GID method makes it feasible to analyse statistically representative large sample in short period; for poorly sorted sediments, such as till, one-tonne is required if the analysis includes cobble size. The capability of the GID method was demonstrated by studying coarse clasts (20-200 mm) from till. There is excellent agreement in the results of the size distribution obtained from the GID method and sieve analysis. The GID method results for size and shape parameters show high and very high repeatability. The particle angularity in the GID method has not been measured to acceptable level; the repeatability test shows some variability. The new methodology for total analysis of till applied the GID method at four different locations in Sweden. The total analysis included 3D size and shape distribution of coarse particles coupled to electrical resistivity, lithological distribution and magnetic susceptibility of the clasts. The results show clear difference in the till samples from the different sites. / <p>QC 20120828</p>

L’imagerie médicale joue un rôle de plus en plus important avec le développement de nombreuses techniques d’acquisition. Il faut principalement pouvoir restaurer (débruiter) les images et en faire une segmentation. Ainsi toute l’information qualitative et quantitative sera disponible pour affiner les diagnostics. Dans cette thèse nous proposons une contribution à cette analyse dans un contexte 3D. Nous étudions deux grands types de méthodes : les méthodes variationnelles et les méthodes par ondelettes. Nous commençons par présenter les modèles variationnels du second ordre, qui s’avèrent plus performants que la classique méthode du premier ordre de Rudin-Osher-Fatemi. Nous l’utilisons pour débruiter et segmenter après avoir donné un bref état de l’art des procédés d’acquisition des images en médecine. Nous introduisons ensuite la transformée en ondelettes et présentons des algorithmes basés sur cette méthode. Les résultats numériques montrent que ces méthodes sont performantes et compétitives. Le coeur de notre travail est de développer des rerésentations 3D qui sont bien adaptées à des données médicales complexes comme des images IRM sous échantillonnées, peu contrastées (cervelets de souris) ou des images IRM d’angiographie (cerveaux de souris). Chaque technique a ses avantages et ses inconvénients. Aussi nous proposons un modèle variationnel mixte second ordre / seuillage par ondelettes. Ce modèle se comporte particulièrement bien : le bruit est correctement éliminé et les contours et textures préservés. Pour finir, nous adaptons plusieurs méthodes de fermeture de contours (hystérésis et distance de chanfrein) dans un contexte 3D. Le mémoire se termine par une synthèses des résultats et une présentation de futures directions de recherche. / Medical procedures have become a critical application area that makes substantial use of image processing. Medical image processing tasks mainly deal with image restoration, image segmentation that bring out medical image details, measure quantitatively medical conditions etc. The diagnosis of a health problem is now highly dependent on the quality and the credibility of the image analysis. The practical contributions of this thesis can be considered in many directions for medical domain. This manuscript addresses a 3D image analysis with variational methods and wavelet transform in the context of medical image processing. We first survey the second-order variational minimization model, which was proved that better than the classical Rudin-Osher-Fatemi model. This method is considered in problems associated to image denoising, image segmentation, that makes a short state of the art on medical imaging processing techniques. Then we introduce the concept of wavelet transform and present some algorithms that also used in this domain. Experimental results show that these tools are very useful and competitive. The core of this research is the development of new 3D representations, which are well adapted to representing complicated medical data, and filament structures in 3D volumes: the cerebellum and mice vessels network. Each of these two based methods has advantages and disadvantages, we then propose a new modified model that combines these schemes in the rest of the thesis. In this situation we propose a new modified model that combines these schemes. With the new decomposition model, in the reconstructed image, noise can be removed successfully and contours, textures are well preserved. This leads to further improvements in denoising performance. Finally, the further part of the thesis is devoted to the description of contribution to extend some classical contour closing methods, namely hysteresis thresholding and contour closing based on chamfer distance transform, in the 3D context. The thesis concludes with a review of our main results and with a discussion of a few of many open problems and promising directions for further research and application.

abstract: Energy return in footwear is associated with the damping behavior of midsole foams, which stems from the combination of cellular structure and polymeric material behavior. Recently, traditional ethyl vinyl acetate (EVA) foams have been replaced by BOOST(TM) foams, thereby reducing the energetic cost of running. These are bead foams made from expanded thermoplastic polyurethane (eTPU), which have a multi-scale structure consisting of fused porous beads, at the meso-scale, and thousands of small closed cells within the beads at the micro-scale. Existing predictive models coarsely describe the macroscopic behavior but do not take into account strain localizations and microstructural heterogeneities. Thus, enhancement in material performance and optimization requires a comprehensive understanding of the foam’s cellular structure at all length scales and its influence on mechanical response. This dissertation focused on characterization and deformation behavior of eTPU bead foams with a unique graded cell structure at the micro and meso-scale. The evolution of the foam structure during compression was studied using a combination of in situ lab scale and synchrotron x-ray tomography using a four-dimensional (4D, deformation + time) approach. A digital volume correlation (DVC) method was developed to elucidate the role of cell structure on local deformation mechanisms. The overall mechanical response was also studied ex situ to probe the effect of cell size distribution on the force-deflection behavior. The radial variation in porosity and ligament thickness profoundly influenced the global mechanical behavior. The correlation of changes in void size and shape helped in identifying potentially weak regions in the microstructure. Strain maps showed the initiation of failure in cell structure and it was found to be influenced by the heterogeneities around the immediate neighbors in a cluster of voids. Poisson’s ratio evaluated from DVC was related to the microstructure of the bead foams. The 4D approach taken here provided an in depth and mechanistic understanding of the material behavior, both at the bead and plate levels, that will be invaluable in designing the next generation of high-performance footwear. / Dissertation/Thesis / Doctoral Dissertation Materials Science and Engineering 2020

Our understanding of the complexity of the brain is limited by the data we can collect and analyze. Because of experimental limitations and a desire for greater detail, most investigations focus on just one aspect of the brain. For example, brain function can be studied at many levels of abstraction including, but not limited to, gene expression, protein interactions, anatomical regions, neuronal connectivity, synaptic plasticity, and the electrical activity of neurons. By focusing on each of these levels, neuroscience has built up a detailed picture of how the brain works, but each level is understood mostly in isolation from the others. It is likely that interaction between all these levels is just as important. Therefore, a key hypothesis is that functional units spanning multiple levels of biological organization exist in the brain. This project attempted to combine neuronal circuitry analysis with functional proteomics and anatomical regions of the brain to explore this hypothesis, and took an evolutionary view of the results obtained. During the process we had to solve a number of technical challenges as the tools to undertake this type of research did not exist. Two informatics challenges for this research were to develop ways to analyze neurobiological data, such as brain protein expression patterns, to extract useful information, and how to share and present this data in a way that is fast and easy for anyone to access. This project contributes towards a more wholistic understanding of the fruit fly brain in three ways. Firstly, a screen was conducted to record the expression of proteins in the brain of the fruit fly, Drosophila melanogaster. Protein expression patterns in the fruit fly brain were recorded from 535 protein trap lines using confocal microscopy. A total of 884 3D images were annotated and made available on an easy to use website database, BrainTrap, available at fruitfly.inf.ed.ac.uk/braintrap. The website allows 3D images of the protein expression to be viewed interactively in the web browser, and an ontology-based search tool allows users to search for protein expression patterns in specific areas of interest. Different expression patterns mapped to a common template can be viewed simultaneously in multiple colours. This data bridges the gap between anatomical and biomolecular levels of understanding. Secondly, protein trap expression patterns were used to investigate the properties of the fruit fly brain. Thousands of protein-protein interactions have been recorded by methods such as yeast two-hybrid, however many of these protein pairs do not express in the same regions of the fruit fly brain. Using 535 protein expression patterns it was possible to rule out 149 protein-protein interactions. Also, protein expression patterns registered against a common template brain were used to produce new anatomical breakdowns of the fruit fly brain. Clustering techniques were able to naturally segment brain regions based only on the protein expression data. This is just one example of how, by combining proteomics with anatomy, we were able to learn more about both levels of understanding. Results are analysed further in combination with networks such as genetic homology networks, and connectivity networks. We show how the wealth of biological and neuroscience data now available in public databases can be combined with the Brain- Trap data to reveal similarities between areas of the fruit fly and mammalian brain. The BrainTrap data also informs us on the process of evolution and we show that genes found in fruit fly, yeast and mouse are more likely to be generally expressed throughout the brain, whereas genes found only in fruit fly and mouse, but not yeast, are more likely to have a specific expression pattern in the fruit fly brain. Thus, by combining data from multiple sources we can gain further insight into the complexity of the brain. Neural connectivity data is also analyzed and a new technique for enhanced motifs is developed for the combined analysis of connectivity data with other information such as neuron type data and potentially protein expression data. Thirdly, I investigated techniques for imaging the protein trap lines at higher resolution using electron microscopy (EM) and developed new informatics techniques for the automated analysis of neural connectivity data collected from serial section transmission electron microscopy (ssTEM). Measurement of the connectivity between neurons requires high resolution imaging techniques, such as electron microscopy, and images produced by this method are currently annotated manually to produce very detailed maps of cell morphology and connectivity. This is an extremely time consuming process and the volume of tissue and number of neurons that can be reconstructed is severely limited by the annotation step. I developed a set of computer vision algorithms to improve the alignment between consecutive images, and to perform partial annotation automatically by detecting membrane, synapses and mitochondria present in the images. Accuracy of the automatic annotation was evaluated on a small dataset and 96% of membrane could be identified at the cost of 13% false positives. This research demonstrates that informatics technology can help us to automatically analyze biological images and bring together genetic, anatomical, and connectivity data in a meaningful way. This combination of multiple data sources reveals more detail about each individual level of understanding, and gives us a more wholistic view of the fruit fly brain.

[pt] Na indústria do petróleo, problemas como dano mecânico causam redução da porosidade e permeabilidade de uma formação rochosa, reduzindo a produtividade e injetividade de poços de sistemas de produção de óleo e gás. Na perfuração do poço há alteração do estado de tensões no seu entorno, causando uma deformação na rocha que pode induzir a uma perda significativa da permeabilidade. Nesta dissertação foi realizado um estudo da influência do dano mecânico na porosidade de rochas do tipo arenito. Para isso, utilizou-se a técnica não-destrutiva de microtomografia de raios-x, que permite a visualização da estrutura interna de materiais, acoplada a uma célula desenvolvida para aplicação, in situ, de tensão hidrostática. Uma amostra de arenito como 8 mm de diâmetro foi tomografada em 3 condições: sem carregamento, após a aplicação de tensão hidrostática de 3300 psi e após o descarregamento. A célula permitiu que as variações de carga fossem realizadas sem retirar a amostra do tomógrafo, permitindo uma comparação quantitativa entre as imagens 3D. Nas 3 condições foram obtidos dados como porosidade total, variação da área porosa em cada camada, volume e forma dos poros. / [en] In the oil industry, problems such as mechanical damage reduce the porosity and permeability of a rock formation, reducing the productivity and injectivity of wells in oil and gas production systems. During the well drilling there is a change in the state of the stress in its surroundings, causing a deformation in the rock that can induce a significant loss of permeability. In this dissertation, it was carried out a study regarding the influence of mechanical damage on the porosity of sandstone rocks. In order to do this, the non-destructive technique of x-ray microtomography was used, which allows the visualization of the materials internal structure, coupled to a cell developed for in situ application of hydrostatic stress. A sandstone sample of 8 mm in diameter was scanned under 3 conditions: without load, after application of 3300 psi hydrostatic stress and after unloading. The cell allowed the load variations to be performed without removing the sample from the tomograph, allowing a quantitative comparison between the 3D images in the 3 conditions. Data such as total porosity, variation of the porous area in each layer, volume and shape of the pores were obtained.

Heart disease is the major cause of death in the developed world. Due to its fast, portable, low-cost and harmless way of imaging the heart, echocardiography has become the most frequent tool for diagnosis of cardiac function in clinical routine. However, visual assessment of heart function from echocardiography is challenging, highly operatordependant and is subject to intra- and inter observer errors. Therefore, development of automated methods for echocardiography analysis is important towards accurate assessment of cardiac function. In this thesis we develop new ways to model echocardiography data using Bayesian machine learning methods and concern three problems: (i) wall motion analysis in 2D stress echocardiography, (ii) segmentation of the myocardium in 3D echocardiography, and (iii) standard views extraction from 3D echocardiography. Firstly, we propose and compare four discriminative methods for feature extraction and wall motion classification of 2D stress echocardiography (images of the heart taken at rest and after exercise or pharmalogical stress). The four methods are based on (i) Support Vector Machines, (ii) Relevance Vector Machines, (iii) Lasso algorithm and Regularised Least Squares, (iv) Elastic Net regularisation and Regularised Least Squares. Although all the methods are shown to have superior performance to the state-of-the-art, one conclusion is that good segmentation of the myocardium in echocardiography is key for accurate assessment of cardiac wall motion. We investigate the application of one of the most promising current machine learning techniques, called Decision Random Forests, to segment the myocardium from 3D echocardiograms. We demonstrate that more reliable and ultrasound specific descriptors are needed in order to achieve the best results. Specifically, we introduce two sets of new features to improve the segmentation results: (i) LoCo and GloCo features with a local and a global shape constraint on coupled endoand epicardial boundaries, and (ii) FA features, which use the Feature Asymmetry measure to highlight step-like edges in echocardiographic images. We also reinforce the traditional features such as Haar and Rectangular features by aligning 3D echocardiograms. For that we develop a new registration technique, which is based on aligning centre lines of the left ventricles. We show that with alignment performance is boosted by approximately 15%. Finally, a novel approach to detect planes in 3D images using regression voting is proposed. To the best of our knowledge we are the first to use a one-step regression approach for the task of plane detection in 3D images. We investigate the application to standard views extraction from 3D echocardiography to facilitate efficient clinical inspection of cardiac abnormalities and diseases. We further develop a new method, called the Class- Specific Regression Forest, where class label information is incorporating into the training phase to reinforce the learning from semantically relevant to the problem classes. During testing the votes from irrelevant classes are excluded from voting to maximise the confidence of output predictors. We demonstrate that the Class-Specific Regression Random Forest outperforms the classic Regression Random Forest and produces results comparable to the manual annotations.

Cette thèse porte sur la modélisation morphologique de matériaux cimentaires, et sur l'analyse de leurs propriétés linéaires élastiques. Dans cet objectif, des images 3D, obtenues par micro-tomographie, de matériaux cimentaires (mortier et béton) sont étudiées. Dans un premier temps, l'image de mortier est segmentée afin d'obtenir une image de microstructure réelle pour des calculs en élasticité linéaire. L'image de béton est utilisée, après traitement, pour la détermination des caractéristiques morphologiques du matériau. Un modèle aléatoire de béton est ensuite développé et validé par des données morphologiques. Ce modèle comporte trois phases qui correspondent à la matrice, les granulats et les pores. La phase des granulats est modélisée par implantation sans recouvrement de polyèdres de Poisson. Pour cela, un algorithme de génération vectorielle de polyèdres de Poisson est mis en place et validé par des mesures morphologiques. Enfin, les propriétés linéaires élastiques effectives de la microstructure de mortier et de microstructures simulées sont déterminées par méthode FFT (Fast-Fourier Transform), pour différents contrastes entre le module de Young des granulats et de la matrice. Cette étude des propriétés effectives est complétée par une analyse locale des champs dans la matrice, afin de déterminer l'arrangement spatial entre les zones de concentration de contraintes dans la matrice, et les différentes phases de la microstructure (granulats et pores). Une caractérisation statistique des champs est de plus réalisée, avec notamment le calcul du Volume Élémentaire Représentatif (VER). Une comparaison des propriétés élastiques effectives et locales obtenues d'une part sur une microstructure simulée contenant des polyèdres et d'autre part sur une microstructure contenant des sphères est de plus effectuée. / The goal of this thesis is to develop morphological models of cementitious materials and use these models to study their local and effective response. To this aim, 3D images of cementitious materials (mortar and concrete), obtained by micro-tomography, are studied. First, the mortar image is segmented in order to obtain an image of a real microstructure, to be used for linear elasticity computations. The image of concrete is used, after being processed, to determine various morphological characteristics of the material. A random model of concrete is then developed and validated by means of morphological data. This model is made up of three phases, corresponding to the matrix, aggregates and voids. The aggregates phase is modelled by implantation of Poisson polyhedra without overlap. For this purpose, an algorithm suited to the vector generation of Poisson polyhedra is introduced and validated with morphological measurements. Finally, the effective linear elastic properties of the mortar and other simulated microstructures are estimated with the FFT (Fast-Fourier Transform) method, for various contrasts between the aggregates and matrix' Young moduli. To complete this work, focused on effective properties, an analysis of the local elastic response in the matrix phase is undertaken, in order to determine the spatial arrangement between stress concentration zones in the matrix and the phases of the microstructure (aggregates and voids). Moreover, a statistical fields characterization, in the matrix, is achieved, including the determination of the Representative Volume Element (RVE) size. Furthermore, a comparison between effective and local elastic properties obtained from microstructures containing polyhedra and spheres is carried out.

Les propriétés des matériaux cellulaires dépendent de leur architecture et des défauts de coulée. L'architecture se réfère à la forme et la distribution de la phase solide. Les défauts correspondent à la présence et aux distributions des cavités et d'intermétalliques dans la phase solide du fait de la procédure de fabrication. Deux types de matériaux produits de différentes façons sont étudiés dans cette thèse. D'une part, deux mousses ERG de tailles de pores différentes ont été choisies pour étudier l'effet de la présence des intermétalliques sur la plasticité et l'endommagement. Des tests de micro-traction et des expériences de nanoindentation ont été réalisés sur des éprouvettes extraites de la mousse pour déterminer leur comportement micro-élastoplastique de la phase solide. D'autre part, deux structures ayant la même forme et le même motif répétitif, mais différentes épaisseurs d'entretoises et de nœuds ont été produites par fusion sélective par laser pour étudier aussi la plasticité et l'endommagement. Ce travail de thèse visait à développer une procédure de modélisation par éléments finis générique basée sur les images 3D pour prendre en compte l'effet de la porosité locale et la présence des intermétalliques dans le comportement. Les états initiaux des échantillons ont été numérisés en utilisant des méthodes de tomographie "locale" et "stitching" à haute résolution. Les géométries 3D maillées, la porosité locale et les propriétés élastiques-plastiques de chaque élément ont été directement renseignées à partir des images 3D à haute résolution. Les procédures de déformation et de rupture des échantillons ont été illustrées en effectuant des expériences in-situ/ex-situ couplées à une numérisation tomographique à basse résolution. Des modèles éléments finis conformes à l'image 3D ont été développés pour la simulation des essais de traction/compression et montrent que la prise en compte des hétérogénéités locales de microstructure permet de prédire plus finement le comportement mécanique des structures cellulaires, en particulier dans la rupture. L'étude visait également à déterminer la conductivité thermique d'une mousse ERG hautement poreuse en utilisant des calculs par éléments finis basés sur l'image. Les résultats ont été vérifiés en comparant avec la conductivité thermique mesurée à partir des expériences de plaques chauffées. / The properties of cellular materials depend on their architecture and casting defects. The architecture refers to shape and distribution of the solid phase. Defects correspond to the presence and distribution of cavities or intermetallic particles in the solid phase due to the fabrication procedure. Two types of materials produced by different fabricating routes are studied in this manuscript. On the one hand, two ERG foams with different cell sizes were chosen to study the effect of the presence of intermetallic particles on the plasticity and damage. Micro-tensile tests and nanoindentation experiment were also performed on the struts extracted from the foam to determine their micro elastoplastic behaviour. On the other hand, two structures with the same shape and repetitive pattern but different struts and nodes thicknesses were produced by selective laser melting manufacturing route to study the effect of porosity on plasticity and damage. This PhD-work aimed at developing a generic image-based finite element procedure to take into account the effect of the local porosity and the presence of intermetallic particles into the finite element simulations of the cellular materials. The initial state of the samples was pictured by performing high resolution "local" tomography and "stitching" methods. The 3D geometries were meshed and the local porosity and elastic-plastic properties of each element were directly informed according to high-resolution 3D images. The deformation and fracture procedures of the samples were pictured by performing in-situ/ex-situ experiments coupled with low-resolution tomography scanning. 3D image-based finite element models were developed for the simulation of the tension/compression tests. The microstructurally informed FE models better capture the mechanical behaviour of the cellular structures, especially for the prediction of the fracture. The study also aimed at determining the thermal conductivity of a highly porous ERG foam using image-based finite element calculations. The results were verified by comparing with the measured thermal conductivity from guarded hot plates experiments.