One of the major obstacles to pressure ulcer research is the difficulty in accurately measuring mechanical loading of specific anatomical sites. A human motion analysis system capable of automatically segmenting a patient's body into high-risk areas can greatly improve the ability of researchers and clinicians to understand how pressure ulcers develop in a hospital environment. This project has developed automated computational methods and algorithms to analyze pressure images acquired in a hospital setting. The algorithm achieved 99% overall accuracy for the classification of pressure images into three pose classes (left lateral, supine, and right lateral). An applied kinematic model estimated the overall pose of the patient. The algorithm accuracy depended on the body site, with the sacrum, left trochanter, and right trochanter achieving an accuracy of 87-93%. This project reliably segments pressure images into high-risk regions of interest.

De nombreux projets de numérisation sont actuellement menés en France et en Europe pour sauvegarder le contenu de dizaines de milliers de documents anciens. Les images de ces documents sont utilisées par les historiens pour identifier l’historique des livres. Cette thèse s’inscrit dans le projet Navidomass (ANR-06-MDCA-012) qui a pour but de valoriser le patrimoine écrit français de la Renaissance, en proposant d’identifier ses images pour les indexer. Dans le cadre de cette thèse, nous nous sommes particulièrement intéressés aux images graphiques. Ces documents,qui sont apparus avec le début de l’imprimerie, sont composées d’images complexes puisque composées de différentes couches d’informations (images de traits).Afin de répondre à ce problème, nous proposons un modèle ontologique d’analyse complexe d’images de documents anciens. Ce modèle permet d’intégrer dans une même base les connaissances propres aux historiens, et les connaissances extraites par des traitements d’images. De par la nature complexe de ces images, les méthodes habituelles d’analyse d’images et d’extraction automatique de connaissances sont inopérantes. Nous proposons donc une nouvelle approche d’analyse des images de documents anciens qui permet de les caractériser à partir de leurs spécificités. Cette approche commence par simplifier les images en les séparant en différentes couches d’informations (formes et traits). Puis, pour chaque couche, nous venons extraire des motifs utilisés pour décrire les images. Ainsi, les images sont caractérisées à l’aide de sacs de motifs fréquents, et de sacs de traits. Pour ces deux couches d’informations, nous venons également extraire des graphes de régions qui permettent d’extraire une connaissance structurelle des images. La complexification de ces deux descriptions est insérée dans la base de connaissances, pour permettre des requêtes complexes. Le but de cette base est de proposer à l’utilisateur de rechercher une image en indiquant soit un exemple d’images recherchées, soit des éléments caractéristiques des images. / In the general context of cultural heritage preservation campaigns, many digitization projects are being conducted in France and Europe to save the contents of thousands of ancient documents. Images of these documents are used by historians to identify the history of books. This thesis was led into the Navidomass project (ANR-06-MDCA-012) which aims at promoting the written heritage of the documents from the Renaissance, by proposing to identify its images. As part of this thesis, we are particularly interested in graphical images, and more specifically to dropcaps. These graphical images, which emerged with the beginning of printing, are complex images which can be seen as composed of different layers of information (images composed of strokes). To address this problem, we propose an ontological model of complex analysis of images of old documents. This model allows to integrate the knowledge specific to historians, and the knowledge extracted by image processing, into a single database. Due to the complex nature of these images, the usual methods of image analysis and automatic extraction of knowledge are inefficient. We therefore propose a new approach for analyzing images of old documents that can be characterized on their features basis. This approach begins by simplifying the images, separated in different layers of information (shapes and lines). Then, for each layer, we extract patterns used to describe the images. Thus, images are described with most common bags of patterns, and bags of stroke. For these two layers of information, we have also extracted graphs of regions that allow extracting a more structural knowledge of the images. A more complex description is then inserted into the knowledge base in order to allow complex queries. The purpose of this database is to offer the possiblity to make either query by example, or query by specific features of the images, to user.

The humanities have for a long time been a collection of fields that have not gained from the advancements in computational power, as predicted by Moore´s law.  Fields like medicine, biology, physics, chemistry, geology and economics have all developed quantitative tools that take advantage of the exponential increase of processing power over time.  Recent advances in computerized pattern recognition, in combination with a rapid digitization of historical document collections around the world, is about to change this. The first part of this dissertation focuses on constructing a full system for finding handwritten words in historical manuscripts. A novel segmentation algorithm is presented, capable of finding and separating text lines in pre-modern manuscripts.  Text recognition is performed by translating the image data of the text lines into sequences of numbers, called features. Commonly used features are analysed and evaluated on manuscript sources from the Uppsala University library Carolina Rediviva and the US Library of Congress.  Decoding the text in the vast number of photographed manuscripts from our libraries makes computational linguistics and social network analysis directly applicable to historical sources. Hence, text recognition is considered a key technology for the future of computerized research methods in the humanities. The second part of this thesis addresses digital palaeography, using a computers superior capacity for endlessly performing measurements on ink stroke shapes. Objective criteria of character shapes only partly catches what a palaeographer use for assessing similarity. The palaeographer often gets a feel for the scribe's style.  This is, however, hard to quantify.  A method for identifying the scribal hands of a pre-modern copy of the revelations of saint Bridget of Sweden, using semi-supervised learning, is presented.  Methods for production year estimation are presented and evaluated on a collection with close to 11000 medieval charters.  The production dates are estimated using a Gaussian process, where the uncertainty is inferred together with the most likely production year. In summary, this dissertation presents several novel methods related to image analysis and machine learning. In combination with recent advances of the field, they enable efficient computational analysis of very large collections of historical documents. / q2b

Modern cars are often equipped with sensors like radar, infrared cameras and stereo cameras that collect information about its surroundings. By using a stereo camera, it is possible to receive information about the distance to points in front of the car. This information can be used to estimate the height of the predicted path of the car. An application which does this is the stereo based Road surface preview (RSP) algorithm. By using the output from the RSP algorithm it is possible to use active suspension control, which controls the vertical movement of the wheels relative to the chassis. This application primarily makes the driving experience more comfortable, but also extends the durability of the vehicle. The idea behind this Master’s thesis is to create an evaluation tool for the RSP algorithm, which can be used at arbitrary roads.  The thesis describes the proposed evaluation tool, where focus has been to make an accurate comparison of camera data received from the RSP algorithm and laser data used as ground truth in this thesis. Since the tool shall be used at the company proposing this thesis, focus has also been on making the tool user friendly. The report discusses the proposed methods, possible sources to errors and improvements. The evaluation tool considered in this thesis shows good results for the available test data, which made it possible to include an investigation of a possible improvement of the RSP algorithm.

[pt] A partir dos anos 70 constatou-se a formação de uma massa aderida às paredes do canal radicular, após a instrumentação, que foi denominada smear layer. Esta camada é composta de material orgânico e inorgânico, tecido pulpar e bactérias, e contribui para obliterar os túbulos dentinários. A remoção desta camada melhora a adaptação entre o material obturador e as paredes dentinárias, elimina bactérias, limpa e alarga os túbulos. Essa remoção é realizada por soluções quelantes, mas a substância e o tempo ideais permanecem desconhecidos. Os quelantes atuam removendo Cálcio da microestrutura dentinária. No presente estudo, foi analisado o poder quelante de MTAD, HEBP e EDTA sobre a dentina coronária, avaliando a fração de área tubular aberta (AreaP) e a massa de Cálcio quelado (MNC (Ca)). Três hipóteses foram propostas e testadas: Não existe diferença entre os quelantes quanto a AreaP (H1) e quanto a MNC(Ca) (H2); existe correlação entre AreaP e MNC(Ca) (H3). Através de Microscopia Óptica Co-localizada (MCL) foram capturadas imagens de vários campos de cada amostra, em diferentes tempos de ataque ácido. Uma seqüência de processamento e análise de imagens foi utilizada para medir AreaP versus tempo, para cada quelante. Esta análise foi complementada qualitativamente com Microscopia Eletrônica de Varredura (MEV). As medidas de MNC(Ca) foram realizadas por Espectroscopia de Absorção Atômica (AA) sobre as soluções quelantes utilizadas no experimento de MCL. H1 foi negada, com MTAD sendo o mais potente e HEBP o menos potente dos quelantes. Os resultados de MEV corroboram os resultados de MCL. Os resultados de AA apresentaram grande dispersão, mas indicam que H2 é falsa. H3 foi rejeitada. / [en] In the 70's it was discovered that after instrumentation of the root canal there was the formation of a smear layer, attached to the walls of the canal. This layer is composed of organic and inorganic material, pulp tissue and bacteria, and contributes to the obliteration of the dentine tubules. The removal of this layer improves the adaptation of the filling material to the dentine walls, eliminates bacteria, cleans and enlarges the tubules. Smear layer removal is achieved through the use of chelators but the ideal substance and application time remain unknown. Chelators act through the removal of Calcium from the dentine microstructure. In the present study, the chelating power of MTAD, HEBP and EDTA on coronal dentine was analyzed through the measurement of the area fraction of open tubules (AreaP) and of the chelated Calcium mass (MNC(Ca)). Three hypotheses were proposed and tested: There is no difference between chelators regarding AreaP (H1) and regarding MNC(Ca) (H2); there is correlation between AreaP and MNC(Ca) (H3). Images of several fields of each sample, for different chelation times, were acquired through Co-Site Optical Microscopy (CSOM). An image processing and analysis sequence was employed to measure AreaP versus time, for each chelator. This analysis was qualitatively complemented by Scanning Electron Microscopy (SEM). Measurements of MNC(Ca) were obtained by Atomic Absorption Spectroscopy (AA) from the chelator solutions used during the CSOM experiments. H1 was denied, with MTAD found to be the most powerful chelator while HEBP was the weakest. SEM results corroborate these findings. AA results showed wide dispersion but indicate that H2 is false. H3 was rejected.

[pt] Neste trabalho, propõe-se uma metodologia para a caracterização de sínter de minério de ferro através de microscopia digital e análise de imagens. O sínter é um material multifásico, com macro e microestrutura heterogêneas. Em geral, as principais fases são hematita, magnetita, ferritos e silicatos, além de poros. Empregando um microscópio óptico automatizado, imagens individuais em alto aumento, cobrindo toda a área da seção transversal das amostras, foram obtidas. Da mesma forma e cobrindo a mesma área, foram obtidas imagens de mosaico em baixo aumento. Os mosaicos fornecem uma visão qualitativa poderosa da amostra inteira, e uma avaliação quantitativa das fases principais, embora com resolução limitada. Uma comparação quantitativa das frações de fase entre o mosaico e as imagens em alto aumento foi realizada através de uma rotina automática de processamento e análise de imagens, também desenvolvida neste trabalho. Enquanto as fases mais finas, como os silicatos, só foram bem discriminadas na melhor resolução óptica empregada, fases preponderantes, como a hematita, foram identificadas em todas as resoluções. Para hematita, magnetita e ferritos, o maior erro relativo na fração de área, entre as imagens obtidas com as lentes de 5 e 20X, foi de 12 %, em 3 amostras distintas. Os resultados para os silicatos foram menos exatos com um erro relativo até 44 %. Uma comparação com resultados de difração de raios-x, usando o método de Rietveld, também foi realizada. Estes resultados indicam que a microscopia digital fornece um método flexível de caracterização destes materiais, permitindo combinar informação global qualitativa e semi- quantitativa, com informação local quantitativa. / [en] In the present work, a method for the characterization of iron ore sinter, based on digital microscopy and image analysis is proposed. Iron ore sinters are multi-phase materials, with heterogeneous macro and micro structure. In general, the main phases are: hematite, magnetite, ferrites and silicates, besides pores. Employing an automated optical microscope, individual high magnification images were obtained, covering the full cross section of the samples. Likewise, and covering the same area, low magnification mosaic images were obtained. Mosaics provide a powerful qualitative view of the whole sample, and a quantitative evaluation of the main phases, albeit with limited resolution. A quantitative comparison of phase fractions between the mosaic and the high magnification images was performed through an automatic image processing and analysis routine, also developed in the present work. While finer phases, such as silicates, could only be discriminated at the best employed optical resolution, the main phases, such as hematite, were identified at all resolutions. For hematite, magnetite, and ferrites, the largest relative error in area fraction was 12%, when comparing images acquired with the 5X and 20X objective lenses, for three different samples. Results for silicates were less accurate, with relative errors up to 44%. A comparison with x-ray diffraction results, employing the Rietveld method, was also performed. These results indicate that digital microscopy provides a flexible method for the characterization of these materials, allowing the combination of global qualitative information with local quantitative data.

The thermal conductivity of various porous thermal barrier coating (TBC) systems used in elevated temperature for industrial applications has been evaluated using a proposed six-phase model. These porous TBC systems rely on microstructural properties and yield different types of porosity. These microstructural properties can influence the thermal conductivity of TBC systems. The purpose of this thesis is to assess the thermal conductivity of TBC systems based on microstructural attributes, particularly the effect of different types of porosity. Thus, the first component of this thesis investigates the microstructural characterization of various TBC systems using image analysis (IA) technique. In this technique, scanning electron microscopy (SEM) and light optical microscopy (LOM) micrographs were used to measure the porosity level of different TBC materials. The volumetric fraction of porosity along with orientation, shape, and morphology have a considerable impact on the total thermal conductivity of TBCs. The second component of this thesis evaluates the thermal conductivity of these porous TBC systems by taking into account the effect of the heat treatment process. The IA results reveal that as long as the porosity content increases, the thermal conductivity decreases for all of the TBC materials studied in this thesis. Further, while the content of microcracks and non-flat porosity play a crucial role in reducing the thermal conductivity of TBC materials, the other types of porosity (open randomly oriented, penny-shaped, and interlamellar) exert less impact on the thermal conductivity of TBCs. Comparing the results of the proposed six-phase model to experimental values and finite element analysis (FEA) values showed a relatively good agreement. The proposed six-phase model can predict the thermal conductivity of porous microstructure of TBC systems close to real measured values; therefore, the proposed six-phase model may be utilized to fabricate the porous microstructure of TBCs.

[pt] A produção do ferro primário, intermediário na produção do aço, é realizada majoritariamente pela rota tradicional do alto-forno. No entanto, tecnologias alternativas estão aumentando suas participações no mercado. Dentre elas, a tecnologia emergente brasileira de autorredução Tecnored é bastante promissora e é objeto de estudo desta dissertação. A tecnologia Tecnored utiliza briquetes autorredutores de minério de ferro em fornos de cuba para obter metal líquido como produto final. Os aglomerados autorredutores, curados a frio, são produzidos a partir de uma mistura de fluxantes, ligantes, finos de minério de ferro e/ou resíduos ferrosos, e um agente redutor carbonoso. A caracterização do insumo para a produção de ferro-primário se faz necessária, de modo a conhecer a distribuição de poros para avaliar a integridade estrutural e a mecânica dos fluidos durante a redução, e quantificar e avaliar a distribuição do ferro metálico presente no aglomerado. Esta dissertação objetiva desenvolver uma metodologia para quantificação de ferro metálico nos briquetes autorredutores de minério de ferro, por meio de rotinas automáticas de processamento de imagens capturadas em microscópio ótico de luz refletida (MLR), e mapear os poros e o ferro metálico ao longo das seções retiradas para análise, também de forma automática. Para a pesquisa realizada, foram utilizados dois tipos de redutores – Carvão Mineral (CM) e Coque Verde de Petróleo (CVP), e diferentes condições de tempo e espessura de briquete. O processo experimental apresentou alguns desafios, desde o corte realizado na preparação das seções polidas para observação em MRL, até a identificação e mapeamento das fases presentes no briquete. Para a validação da metodologia, os resultados da quantificação do ferro metálico realizada por análise de imagens foram comparados com a técnica tradicional de análise química. Os mapas de porosidade e ferro metálico permitiram uma avaliação qualitativa das variáveis tempo e espessura do briquete, para cada tipo de redutor. / [en] The production of primary iron, intermediate step in steelmaking chain, is mostly done by the traditional route of blast furnace. However, alternative technologies are increasing their market share. Among the technologies, Tecnored is a very promising Brazilian ironmaking process and the goal of this master dissertation is the study of it iron-bearing burden. The Tecnored technology utilizes self-reduced iron ore briquettes in a modulate shaft furnace to obtain hot metal. The self-reduced agglomerates, cold bonded, are produced from a mixture of flux, binder, fines of iron ore, residues containing iron and carbonaceous material as a reducing agent. The burden characterization in ironmaking process is essential to study the pores distribution and evaluate the structural integrity and mechanics of fluids during the reduction process, and quantify and analyze the metallic iron distribution along the briquette’s volume. This dissertation main goal is to develop a new methodology of quantification of the metallic iron in selfreduced iron ore briquettes, through automatic routines of image processing, captured in bright field of a reflected light optical microscopy (MRL), and mapping pores and metallic iron automatically along the cross-sections analyzed. The tested briquettes utilized two reducing agents – coal fines (CM) and coke of petroleum (CVP), and it was submitted in two different reduction times and two different briquettes’ thickness. The experimental process presented some challenges, from the cross section cut to be polished and observed in MRL, to the identification and mapping of the phases in the briquettes. To validate the methodology, the results of the quantification of metallic iron through image analysis were compared to the results of the traditional technique of chemical analysis. The porosity and metallic iron maps provided a qualitative evaluation of reduction time and the effect of briquettes’ thickness, for each type of cabonaceus reducing agents.

[pt] Com os métodos de instrumentação usados atualmente na terapia endodôntica ocorre a formação de uma camada residual composta de matéria orgânica e inorgânica que freqüentemente pode reter bactérias e seus subprodutos. Os métodos disponíveis para a remoção do smear layer incluem o ataque químico, técnicas ultra-sônicas e a laser, sendo que nenhum desses métodos se mostra totalmente eficaz, não existindo um consenso universal a respeito. A ação do EDTA, do EDTAC e do ácido cítrico sobre a dentina radicular foi analisada neste estudo por dois métodos. Primeiramente, foi realizada uma análise da microdureza radicular antes e depois da aplicação dos quelantes. Em seguida, utilizou-se microscopia de força atômica (AFM) para mapear e caracterizar as alterações microestruturais ocorridas na dentina durante o processo de quelação dos íons de cálcio. No caso da microscopia, realizou-se experimentos em que a ação quelante foi observada em tempo real, in situ, através do uso de um porta-amostras especial - uma célula de líquido - que permitiu observar a superfície das amostras durante o condicionamento ácido. Seqüências de imagens foram capturadas durante o processo e uma rotina de processamento e análise digital de imagens (PADI) foi criada para avaliar qualitativa e quantitativamente o efeito das substâncias quelantes. O desenvolvimento desta metodologia de microscopia in situ e análise digital é uma das principais contribuições da presente dissertação. / [en] Instrumentation methods currently used in endodontic therapy create a residual layer composed of organic and inorganic material that frequently retains bacteria and their subproducts. The available methods to remove this smear layer include chemical etching, ultrasonic and laser techniques but none of these is completely efficient and there is no universal consensus regarding this subject. In the present study the effect of EDTA, EDTAC and citric acid upon radicular dentin was analyzed by two methods. Dentin microhardness was measured before and after etching. Atomic force microscopy was used to map and characterize dentinary microstructural changes throughout the chelation process. Real time, in situ, observations of the process were performed through the use of a special sample holder - a liquid cell - that allowed the observation of the samples surfaces during etching. Image sequences were acquired and an image processing and analysis procedure was developed to evaluate, both qualitatively and quantitatively, the effect of the 3 chelators. The development of this methodology employing in situ microscopy and digital image analysis is one of the main contributions of the present dissertation.

The world of industrial manufacturing has changed a lot during the past decades. It has gone from a labour-intensive process of manual control of machines to a fully connected and automated process. The next big leap in industrial manufacturing is known as industry 4.0 or smart manufacturing. With industry 4.0 comes increased integration between IT systems and the factory floor. This change has proven challenging to implement into existing factories many with the intended lifespan of several decades. One of the single most important parameters to measure is the operating hours of each machine. This information can help companies better utilize their resources and save huge amounts of money.  The goal is to develop a solution which can track the operating hours of the machines using image analysis and the signal lights already mounted on the machines. Using methods commonly used for traffic light recognition in autonomous cars, a system with an accuracy of over 99% during the specified conditions, has been developed. It is believed that if more diverse video data becomes available a system, with high reliability that generalizes well, could be developed using similar methodology. / Industriell tillverkning har förändrats mycket under de senaste decennierna. Det har gått från en process som krävt mycket manuellt arbete till en process som är nästan helt uppkopplad och automatiserad. Nästa stora steg inom industriell tillverkning går under benämningen industri 4.0 eller smart tillverkning. Med industri 4.0 kommer en ökad integration mellan IT-system och fabriksgolvet. Denna förändring har visat sig vara särskilt svår att implementera i redan existerande fabriker som kan ha en förväntad livstid på flera årtionden. En av de viktigaste parametrarna att mäta inom industriell tillverkning är varje maskins operativa timmar. Denna information kan hjälpa företag att bättre utnyttja tillgängliga resurser och därigenom spara stora summor pengar. Målet är att utveckla en lösning som, med hjälp av bildanalys och de signalljus som maskinerna kommer utrustade med, kan mäta maskinernas operativa timmar. Med hjälp av metoder som vanligen används för trafikljusigenkänning i autonoma fordon har ett system med en träffsäkerhet på över 99% under de förutsättningar som presenteras i rapporten utvecklats. Om mer video med större variation blir tillgänglig är det mycket troligt att det går att utveckla ett system som har hög pålitlighet i de flesta produktionsmiljöer.