Cette thèse explore les dernières évolutions du débruitage d'images, et elle essaie de développer une vision synthétique des techniques utilisées jusqu'à présent. Elle aboutit à un nouvel algorithme de débruitage d'image évitant les artefacts et avec un meilleur PSNR que tous les algorithmes que nous avons pu évaluer. La première méthode que nous présentons est DA3D, un algorithme de débruitage fréquentiel avec guide, inspiré de DDID. La surprise de cet algorithme, c'est que le débruitage fréquentiel peut battre l'état de l'art sans produire artefacts. Cet algorithme produit des bons résultats non seulement en PSNR, mais aussi (et surtout) en qualité visuelle. DA3D marche particulièrement bien pour améliorer les textures des images et pour enlever les effets de staircasing.DA3D, guidé par un autre algorithme de débruitage améliore presque toujours le résultat de son guide. L'amélioration est particulièrement nette quand le guide est un algorithme à patchs, et alors on combine deux principes différents: auto-similarité suivi de seuillage fréquentiel. Le deuxième résultat présenté est une méthode universelle de débruitage multi-échelle, applicable à tout algorithme. Une analyse qualitative montre en effet que les algorithmes de débruitage à patchs éliminent surtout les hautes fréquences du bruit, à cause de la taille limitée des voisinages traités. Plutôt que d'agrandir ces voisinages nous décomposons l'image en une pyramide basée sur la transformée en cosinus discrète, avec une méthode de recomposition évitant le ringing. Cette méthode traite le bruit à basse fréquence, et améliore la qualité de l'image. Le troisième problème sérieux que nous abordons est l'évaluation des algorithmes de débruitage. Il est bien connu que le PSNR n'est pas un indice suffisant de qualité. Un artefact sur une zone lisse de l'image est bien plus visible qu'une altération en zone texturée. Nous proposons une nouvelle métrique basée sur un Smooth PSNR et un Texture PSNR, pour mesurer les résultats d'un algorithme sur ces deux types des régions. Il apparaît qu'un algorithme de débruitage, pour être considéré acceptable, doit avoir des bons résultats pour les deux métriques. Ces métriques sont finalement utilisées pour comparer les algorithmes de l'état de l'art avec notre algorithme final, qui combine les bénéfices du multi-échelle et du filtrage fréquentiel guidé. Les résultats étant très positifs, nous espérons que la thèse contribue à résoudre un vieux dilemme, pour lequel la méthode DDID avait apporté de précieuses indications : comment choisir entre le seuillage fréquentiel et les méthodes basées sur l'auto-similarité pour le débruitage d'images ? La réponse est qu'il ne faut pas choisir. Cette thèse termine avec quelques perspectives sur la faisabilité du débruitage "externe". Son principe est de débruiter un patch en utilisant une grande base de données externe de patches sans bruit. Un principe bayésien démontré par Levin et Nadler en 2011 implique que le meilleur résultat possible serait atteint avec cette méthode, à condition d'utiliser tous les patches observables. Nous donnons les arguments mathématiques prouvant que l'espace des patches peut être factorisé, ce qui permet de réduire la base de données de patches utilisés d'un facteur au moins 1000. / This thesis explores the last evolutions on image denoising, and attempts to set a new and more coherent background regarding the different techniques involved. In consequence, it also presents a new image denoising algorithm with minimal artifacts and the best PSNR performance known so far.A first result that is presented is DA3D, a frequency-based guided denoising algorithm inspired form DDID [Knaus-Zwicker 2013]. This demonstrates that, contrarily to what was thought, frequency-based denoising can beat state-of-the-art algorithms without presenting artifacts. This algorithm achieves good results not only in terms of PSNR, but also (and especially) with respect to visual quality. DA3D works particularly well on enhancing the textures of the images and removing staircasing effects.DA3D works on top of another denoising algorithm, that is used as a guide, and almost always improve its results. In this way, frequency-based denoising can be applied on top of patch-based denoising algorithms, resulting on a hybrid method that keeps the strengths of both. The second result presented is Multi-Scale Denoising, a framework that allows to apply any denoising algorithm on a multi-scale fashion. A qualitative analysis shows that current denoising algorithms behave better on high-frequency noise. This is due to the relatively small size of patches and search windows currently used. Instead of enlarging those patches, that can cause other sorts of problems, the work proposes to decompose the image on a pyramid, with the aid of the Discrete Cosine Transformation. A quantitative study is performed to recompose this pyramid in order to avoid the appearance of ringing artifacts. This method removes most of the low-frequency noise, and improves both PSNR and visual results for smooth and textured areas.A third main issue addressed in this thesis is the evaluation of denoising algorithms. Experiences indicate that PSNR is not always a good indicator of visual quality for denoising algorithms, since, for example, an artifact on a smooth area can be more noticeable than a subtle change in a texture. A new metric is proposed to improve on this matter. Instead of a single value, a ``Smooth PNSR'' and a ``Texture PSNR'' are presented, to measure the result of an algorithm for those two types of image regions. We claim that a denoising algorithm, in order to be considered acceptable, must at least perform well with respect to both metrics. Following this claim, an analysis of current algorithms is performed, and it is compared with the combined results of the Multi-Scale Framework and DA3D.We found that the optimal solution for image denoising is the application of a frequency shrinkage, applied to regular regions only, while a multiscale patch based method serves as guide. This seems to resolve a long standing question for which DDID gave the first clue: what is the respective role of frequency shrinkage and self-similarity based methods for image denoising? We describe an image denoising algorithm that seems to perform better in quality and PSNR than any other based on the right combination of both denoising principles. In addition, a study on the feasibility of external denoising is carried, where images are denoised by means of a big database of external noiseless patches. This follows a work of Levin and Nadler, in 2011, that claims that state-of-the-art results are achieved with this method if a large enough database is used. In the thesis it is shown that, with some observation, the space of all patches can be factorized, thereby reducing the number of patches needed in order to achieve this result. Finally, secondary results are presented. A brief study of how to apply denoising algorithms on real RAW images is performed. An improved, better performing version of the Non-Local Bayes algorithm is presented, together with a two-step version of DCT Denoising. The latter is interesting for its extreme simplicity and for its speed.

The objective of the research presented in this thesis was to develop a concrete bridge deck topping mixture that resists the effects of differential shrinkage by decreasing shrinkage and increasing creep. . In addition, the amount of tensile creep that concrete experiences under long-term tensile stresses were quantified and compared to compressive creep values in order to gain a better understanding of how concrete behaves under tension. Test results show that the amount of tensile creep exceeded compressive creep by a factor of 2-5. Various shrinkage and creep models were compared against test data in order to quantify results and determine the best model to use for the mixes examined during this research project. Data analysis revealed that the AASHTO time dependent effects (shrinkage and creep) models outperformed the other models used in this research project. Other material property data including compressive strength, splitting tensile strength, Young's modulus of elasticity, and unrestrained shrinkage was also collected to compare against a common bridge deck topping mix to ensure that the mixes used in this research project are suitable for use in the field. A parametric study utilizing the Age Adjusted Effective Modulus (AAEM) method was performed which showed that the most important factor in reducing tensile stresses was to decrease the amount of shrinkage experienced by the concrete bridge deck topping mixture. Three concrete mixtures, one included saturated lightweight aggregates (SLWA), one including ground granulated blast furnace slag (GGBFS), and one incorporating both were tested. Preliminary results show that the inclusions of SLWA into a concrete mixture reduced shrinkage by 25% and overall tensile stress by 38%. / Master of Science

No description available.

Stylometry assumes that the essence of the individual style of an author can be captured using a number of quantitative criteria, such as the relative frequencies of noncontextual words (e.g., or, the, and, etc.). Several statistical methodologies have been developed for authorship analysis. Jockers et al. (2009) utilize Nearest Shrunken Centroid (NSC) classification, a promising classification methodology in DNA microarray analysis for authorship analysis of the Book of Mormon. Schaalje et al. (2010) develop an extended NSC classification to remedy the problem of a missing author. Dabney (2005) and Koppel et al. (2009) suggest other modifications of NSC. This paper develops a full Bayesian classifier and compares its performance to five versions of the NSC classifier using the Federalist Papers, the Book of Mormon text blocks, and the texts of seven other authors. The full Bayesian classifier was superior to all other methods.

The concept of portfolio optimization has been widely studied in the academy and implemented in the financial markets since its introduction by Markowitz 70 years ago. The problem of the mean-variance optimization framework caused by input uncertainty has been one of the foci in the previous research. In this study, several models (linear shrinkage and Black-Litterman) based on Bayesian approaches are studied to improve the estimation of inputs. Moreover, a new framework based on robust optimization is presented to mitigate the input uncertainty further.  An out-of-sample test is specially designed, and the results show that Bayesian models in this study can improve the optimization results in terms of higher Sharpe ratios (the quotient between portfolio returns and their risks). Both covariance matrix estimators based on the linear shrinkage method contain less error and provide better optimization results, i.e. higher Sharpe ratios. The Black-Litterman model with a proper choice of inputs can significantly improve the portfolio return. The new framework based on the combination of shrinkage estimators, Black-Litterman, and robust optimization presents a better way for portfolio optimization than the classical framework of mean-variance optimization.

Shotcrete is sprayed concrete applied pneumatically under high pressure and was invented in the beginning of the 1900's. This new technique decreased the construction time and since steel fibres were introduced in the shotcrete during the 1970's, shotcrete has been the primary support method for tunnels. Tunnels excavated with the drill and blast method creates a highly irregular rock surface which results in a shotcrete lining with varying thickness. The structural behaviour as well as the loads acting on the shotcrete lining depends on the interaction between the shotcrete, rock and rock bolts. There are several parameters influencing this interaction, e.g. bond strength, the stiffness of the rock and thickness of the shotcrete. All of these parameters are difficult to predict accurately which makes the structural design of the lining to a complex problem. This thesis present the first part of a research project with the long-term goal to improve the understanding of the structural behaviour of the shotcrete lining. To achieve this, numerical modelling have been used to study the build up of stresses and cracking of shotcrete when subjected to restrained loading caused by e.g. temperature differences and drying shrinkage. The response in the lining when subjected to a gravity load from a block has also been studied. The model is capable of describing the non-linear deformation behaviour of both plain and fibre reinforced shotcrete and uses presented in situ variations in thickness to more accurately account for the effects of expected variations in thickness. The thesis discuss and demonstrate the effect of important loads that acts on the shotcrete lining and how the irregular geometry of the rock surface in combination with the varying thickness of the shotcrete affect the development of stresses in the lining. It is also discussed how a full or partial bond failure affect the structural capacity of a shotcrete lining. / Sprutbetong är betong som appliceras pneumatiskt under högt tryckt, en metod utvecklad i början av 1900-talet. Kort därefter gjordes de första försöken att använda sprutbetong som bergförstärkning. Den här nya tekniken minskade produktionstiden och när stålfibrer introduceras under 1970-talet kunde det tunga arbetet med att placera armering minimeras. Sedan dess har sprutbetong blivit den preliminära förstärkningsmetoden, särskilt för tunnlar i hårt berg där tunna lager av sprutbetong ibland kan användas som den enda förstärkningsåtgärden.  Tunnlar byggs normalt genom metoden "borrning-sprängning" vilket leder till att bergytan där sprutbetongen appliceras få r en oregelbunden form. Under sprutning är det svårt att fastställa den exakta tjockleken och sprutbetongen har därmed en oregelbunden tjocklek. Beroende på in situ förhållanden kan oarmerad eller fiberarmerad sprutbetong i kombination med bergbultar användas för att förstärka berget. Det strukturella beteendet och lasterna som påverkar förstärkningen beror på interaktionen mellan sprutbetong, berg och bergbultar. Denna samverkan styrs av flera parametrar som t ex; vidhäftningshållfastheten, bergets styvhet och tjockleken hos sprutbetongen. Dessa parametrar är svåra att förutsäga vilket gör dimensionering av en sprutbetongförstärkningen till ett komplext problem. Den här uppsatsen presenterar den första delen av ett forskningsprojekt med det långsiktiga målet att öka förståelsen för det strukturella beteendet hos en sprutbetongförstärkning. För att uppnå detta har numerisk modellering använts för att studera spänningsuppbyggnaden och uppsprickningen av sprutbetong som utsätts för förhindrade rörelser orsakade av temperaturförändringar eller uttorkningskrympning. Sprutbetongens beteende när den utsätts för en blocklast har också studerats. En numeriskt modell för att analysera spänningarna i sprutbetong som tar hänsyn till tidsberoende materialegenskaper har använts. Modellen kan beskriva det icke-linjära deformationsbeteendet av oarmerad samt fiberarmerad sprutbetong och använder sig av presenterad fältdata för att beskriva de förväntade tjockleksvariationerna. Uppsatsen disskuterar och demonstrerar effekten av viktiga laster som verkar på sprutbetongförstärkningen och hur bergets oregelbundna yta i kombination med sprutbetongens varierande tjocklek påverkar spänningsuppbyggnaden i förstärkningen. Det diskuteras också hur ett fullständigt eller partiellt vidhäftningsbrott på verkar sprutbetongförstärkningens bärförmåga. / <p>QC 20170418</p>

Cancer is a complex disease and presents one of the greatest challenges in modern medicine. Despite remarkable advances in treatment of several cancer types, relapse and resistance to therapy remain recurring outcomes in patients, which underscores a need for personalized treatment approaches. These complications have been related to the high genetic diversity observed within tumors, termed intratumor heterogeneity (ITH). While specific mutational profiles have been associated with the development of heterogeneous tumors, the relationship between ITH and phenotype could unveil features that undergo selection and convey fitness. Features presented in the transcriptome, as markers of heterogeneity, might therefore be valuable biomarkers. In this project, these features are explored by assuming a linear relationship between genetic ITH measures and gene expression data from The Cancer Genome Atlas samples. By first reducing the number of variables among the transcriptome to the differentially expressed genes between low and high ITH samples, the association between specific gene expression profiles and ITH is sought with a linear model. By using two different methods for estimating ITH, called Expands and PhyloWGS, the association was modeled with each method. Interestingly, the model based on Expands captured the elevated expression of a chaperone gene DNAJC18 as being consistently associated with lower ITH in four cancer types. On the other hand, models based on PhyloWGS presented lower predictive power. These results demonstrate that the transcriptome can be used to predict genetic ITH, although this depends on the method used for characterizing ITH. / Cancer är en komplex sjukdom och en av de största utmaningarna i dagens medicin. Trots stora framsteg i behandlingen av flera cancerformer är återfall och terapiresistens återkommande problem vilket talar starkt för behov av individualiserad behandling. Dessa komplikationer har relaterats till den höga genetiska variabiliteten som observeras inom tumörer, även kallad intratumoral heterogenitet (ITH). Undersökning av relationen mellan ITH och fenotypisk data kan ta fram markörer som är involverade i cancerutvecklingen som bidragare till heterogenitet. Genom att modellera associationen mellan transcriptomen och ITH kan man även hitta kliniskt relevanta biomarkörer. I detta projekt undersöks relationen mellan genutryck och ITH genom att applicera linjär regression. Genom att först reducera antalet variabler i transkriptomen till de diferentiellt utryckta gener, används linjära modellen för att ta fram specifika gener vars utryck kan relateras till ändringar i ITH uppskattad för The Cancer Genome Atlas prover. ITH uppskattas med två algoritmiska metoder, kallade Expands och PhyloWGS. Resultaten visade att förhöjd uttryck an genen DNAJC18 är associerad med lägre ITH uppskattad med Expands bland fyra cancer typer. Trots detta visade inte genutryck och ITH uppskattat med PhyloWGS lika starkt linjärt samband.

Concrete slabs-on-ground are commonly used in many types of industrial floors, warehouses, highways, parking lots and buildings. Cracks and deflection of slabs are undesired events caused by differential shrinkage, which limits the service life of the slabs. Non-linear behavior of cracks and deflections, interaction of concrete and reinforcement increase the complexity in predicting the occurrence and positioning of cracks. The Eurocode 2 provides a reference for theoretical approximation for design of concrete structures. This thesis intent to investigate the crack behavior of slabs-on-ground subjected to gradient shrinkage using nonlinear finite element analysis, as implemented in the software package Atena 2D. The first part of the thesis is focused on suitable modeling techniques for predicting cracks in concrete slabs-on-ground due to gradient shrinkage. The second part is directed towards parametric studies, performed to explore the significance of varying thickness, length, concrete strength class, bond types, reinforcement content and friction coefficient. The results obtained with the Atena 2D was validated using the design software WIN-statik for calculating the maximum crack width in the context of obtaining realistic results. Finally, the WSP guide recommended parameters were tested as inputs to the model. A slab-on-ground was modeled in Atena 2D considering these as statically indeterminate structures, where both slab and grade were included and the convergence analysis performed under plane stress conditions enabling prediction of the maximum crack widths for increasing applied shrinkage loads. Parametric studies demonstrate the dependency of the slab length, showing that a smaller length reduces the crack width, since such a slab is less constrained by the sub-base. To avoid cracks in the slabs their relative thickness should not be increased above a certain thickness, instead the reinforcement content should be increased. The numerical simulation shows that different concrete strength classes give similar cracks widths. Sand as sub-base provides less crack widths for interface materials EPS, sand and gravel. Although, dry sand as interface material gives similar crack widths as EPS, it is the best to use EPS that is also used to retard the moisture diffusion from the sub-base. The numerical model developed was validated for the recommended values given by the WSP guide, which gives less crack widths and deflections. The numerical model gives less crack widths compared to the Eurocode 2, which considers only the statistically determinant problems overestimating the crack widths. The presented examples demonstrate that the developed model can accurately predict crack formation, crack behavior and vertical deflection in concrete slabs-on-ground subjected to gradient shrinkage loads.

The present research is conducted to investigate long term (more than two years) free and confined drying shrinkage magnitude and behaviour of self-compacting concrete (SCC) and compare with normal concrete (NC). For all SCCs mixes, Portland cement was replaced with 0-60% of fly ash (FA), fine and coarse aggregates were kept constant at 890 kg/m3 and 780 kg/m3, respectively. Two different water binder ratios of 0.44 and 0.33 were examined for both SCCs and NCs. Fresh properties of SCCs such as filling ability, passing ability, viscosity and resistance to segregation and hardened properties such as compressive and flexural strengths, water absorption and density of SCCs and NCs were also determined. Experimental results of free drying shrinkage obtained from this study together with collected comprehensive database from different sources available in the literature were compared to five existing models, namely the ACI 209R-92 model, BSEN-92 model, ACI 209R-92 (Huo) model, B3 model, and GL2000 model. To assess the quality of predictive models, the influence of various parameters (compressive strength, cement content, water content and relative humidity) on the drying shrinkage strain are studied. An artificial neural network models (ANNM) for prediction of drying shrinkage strains of SCC was developed using the same data used in the existing models. Two ANNM sets namely ANNM1 and ANNM2 with different numbers of hidden layer neurones were constructed. Comparison between the results given by the ANNM1 model and the results obtained by the five existing predicted models were presented. The results showed that, using up to 60% of FA as cement replacement can produce SCC with a compressive strength as high as 30 MPa and low drying shrinkage strain. SCCs long-term drying shrinkage from 356 to 1000 days was higher than NCs. Concrete filled elliptical tubes (CFET) with self-compacting concrete containing FA up to 60% are recommended for use in construction in order to prevent confined drying strain. ACI 209R-92 model provided a better prediction of drying shrinkage compared with the other four models. However, a very high predictability with high accuracy was achieved with the ANNM1 model with a mean of 1.004. Moreover, by using ANNM models, it is easy to insert any of factors effecting drying shrinkage to the input parameters to predict drying shrinkage strain of SCC. / Ministry of Higher Education, Libya

Yes / As a viable option to upcycle construction and demolition waste (CDW) into value-added materials, geopolymer technology is emerging. Most studies investigate CDWs in a separated form or in combination with mainstream pozzolanic/cementitious materials focusing only on fundamental properties of geopolymer pastes, not considering to scale such materials to the level of their application in the forms of structural mortars/concretes or to characterize long-term performance/durability. This study investigated the development and characterization of ambient-cured mortars with mixed CDW-based geopolymer binders and untreated fine recycled concrete aggregates (FRCA). Mixture of CDW-based roof tile (RT), red clay brick (RCB), hollow brick (HB), concrete (C), and glass (G) was used as the precursor, while ground granulated blast furnace slag (S) was used in some mixtures to partly replace CDW precursors. Compressive strength, durability-related parameters including drying shrinkage, water absorption, and efflorescence, microstructure and materials sustainability were evaluated. Results showed that 28 d compressive strength results above 30 and 50 MPa is achievable with the entirely CDW-based and slag-substituted mortars, which were found improvable to have entirely CDW-based structural concretes. Drying shrinkage of the mortars is slightly higher than that of conventional cementitious/geopolymeric systems although it can be minimized significantly through mixture optimization. Water absorption values remain comparable with the literature. CDW-based geopolymer mortars outperform Portland cement mortars in terms of CO2 emission and energy requirement. Our findings show that via utilizing CDW-based constituents in mixed form as precursor and waste aggregates, it is possible to develop greener construction materials with acceptable strength and long-term performance. / This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 894100. The authors also wish to acknowledge the support of the Scientific and Technical Research Council (TUBITAK) of Turkey provided under project: 117M447.