Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal.

Traumatic pelvic injuries associated with high-energy pelvic fractures are life-threatening injuries. Extensive bleeding is relatively common with pelvic fractures. However, bleeding is especially prevalent with high-energy fractures. Hemorrhage remains the major cause of death that occur within the first 24 hours after a traumatic pelvic injury. Emergent-life saving treatment is required for high-energy pelvic fractures associated with hemorrhage. A thorough understanding of potential sources of bleeding within a short period is essential for diagnosis and treatment planning. Computed Tomography (CT) images have been widely in use in identifying the potential sources of bleeding. A pelvic CT scan contains a large number of images. Analyzing each slice in a scan via simple visual inspection is very time consuming. Time is a crucial factor in emergency medicine. Therefore, a computer-assisted pelvic trauma decision-making system is advantageous for assisting physicians in fast and accurate decision making and treatment planning. The proposed project presents an automated system to detect and segment hemorrhage and combines it with the other extracted features from pelvic images and demographic data to provide recommendations to trauma caregivers for diagnosis and treatment. The first part of the project is to develop automated methods to detect arteries by incorporating bone information. This part of the project merges bone edges and segments bone using a seed growing technique. Later the segmented bone information is utilized along with the best template matching to locate arteries and extract gray level information of the located arteries in the pelvic region. The second part of the project focuses on locating the source of hemorrhage and its segmentation. The hemorrhage is segmented using a novel rule based hemorrhage segmentation approach. This approach segments hemorrhage through hemorrhage matching, rule optimization, and region growing. Later the position of hemorrhage in the image and the volume of the hemorrhage are determined to analyze hemorrhage severity. The third part of the project is to automatically classify the outcome using features extracted from the medical images and patient medical records and demographics. A multi-stage feature selection algorithm is used to select the predominant features among all the features. Finally, boosted logistic model tree is used to classify the outcome. The methods are tested on CT images of traumatic pelvic injury patients. The hemorrhage segmentation and classification results seem promising and demonstrate that the proposed method is not only capable of automatically segmenting hemorrhage and classifying outcome, but also has the potential to be used for clinical applications. Finally, the project is extended to abdominal trauma and a novel knowledge based heuristic technique is used to detect and segment spleen from the abdominal CT images. This technique is tested on a limited number of subjects and the results are promising.

Thèse numérisée par la Direction des bibliothèques de l'Université de Montréal.

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

Title: Vessel segmentation Author: Ján Dupej Department / Institute: Department of Software and Computer Science Education Supervisor of the master thesis: RNDr. Josef Pelikán, KSVI Abstract: In this thesis we researched some of the blood vessed segmentation and visualization techniques currently available for angiography on CT data. We then designed, implemented and tested a system that allows both semi-automatic and automatic vessel segmentation and visualization. For vessel segmantation and tracking we used a region-growing algorithm that we overhauled with several heuristics and combined with centerline detection. We then automated this algorithm by automatic seed generation. The visualization part is accomplished with an adaptation of the well-known straightened CPR method that we enhanced so that it visualizes the whole cross-section of the blood vessel, instead of just one line of it. Furthermore, we used the Bishop frame to maintain minimal twist of the curve-local coordinate system along the whole vessel. Keywords: vessel segmentation, medical data analysis, volume data

Rapportens namn: Vilka är egentligen våra kunder?- en kvalitativ studie om segmentering i B2B företag Frågeställning: Hur utarbetas och genomförs segmentering av företag verksamma på B2B marknaden? Syfte: Syftet med arbetet är att undersöka hur företag utarbetar och genomför segmentering på B2B marknaden. Vidare undersöka vilka faktorer som är viktiga att ta i beaktning för företag vid utförandet av segmentering på B2B marknaden.Vi ämnar även att presentera både teoretiska och praktiska förslag på hur B2B företag bör arbeta med segmenteringsstrategier. Metod: Uppsatsen är en kvalitativ fallstudie som antagit en abduktiv forskningsansats. Datainsamlingen har skett vi semistrukturerade djupintervjuer med sju olika B2B företag. Resultat och slutsatser: Resultatet visar hur företag idag utarbetar och genomför segmentering. Vidare vilka kunskaper och vilken förståelse de besitter utifrån de segmenteringsvariabler som framkommit i studiens teoretiska referensram. Slutsatser kunde dras kring att företag i dagsläget inte i många fall arbetar aktivt och iterativt med att segmentera sin marknad och har inte följt en specifik och tydlig process vid utförandet. De besitter kunskaper som hade kommit till användning vid en vidare utveckling av de segment de arbetar utifrån idag. Teoretiskt och praktiskt bidrag: Uppsatsen presenterar en reviderad segmenteringsmodell utifrån den teoretiska referensramen i kombination med den insamlade empirin. Behov och potential framkommer som kompletterande faktorer att ta i beaktning. Det praktiska bidraget ger förslag på hur företag bör utarbeta och genomföra segmentering. Vidare presenteras förslag till segmenteringsstrategi till uppsatsens uppdragsgivare. / Name of report: - a qualitative study of segmenting the B2B market Research question: Who are actually our customers?- a qualitative study concerning segmentation in B2B companies Purpose: The purpose of this thesis is to examine how companies active in the B2B market prepare and implement segmentation. Furthermore to examine which different factors to take into consideration when segmenting the B2B market. We also intend to provide theoretical as well as practical implications of how B2B companies should work with segmentation. Method: This thesis is of qualitative character and follows an abductive research approach. The empirical data of this study have been gathered through semi-structured interviews with seven different B2B companies. Results and conclusions: The results shows how companies work with segmentation today, furthermore which knowledge they possess concerning the presented segmentation variables from the study’s theoretical framework. Conlusions have been made that in many cases companies do not actively and ongoing work with segmenting their market. Furthermore they have not followed an specific and clear process when segmenting. Companies do however possess knowledge which can come to use in order to develop the segment for which they work with today. Theoretical and practical contribution: The thesis presents a revised segmentation model based on the theoretical framework in combination with the collected empirical results. Customers needs and potential emerged as complementary factors to take into account. The practical contribution provides suggestions on how companies should develop and implement segmentation. Furthermore, proposals for segmentation strategy are presented to the principal of this thesis.

La topologie algébrique, bien que domaine abstrait des mathématiques, apporte de nouveaux concepts pour le traitement d'images. En effet, ces tâches sont complexes et restent limitées par différents facteurs tels que la nécessité d’utiliser un paramétrage, l'influence de l'arrière-plan ou la superposition d'objets. Nous proposons ici des méthodes dérivées de la topologie algébrique qui diffèrent des méthodes classiques de traitement d'images par l’intégration d’informations locales vers des échelles globales grâce à des invariants topologiques. Une première méthode de segmentation d'images a été développée en ajoutant aux caractéristiques statistiques classiques d’autres de nature topologique calculées par homologie persistante. Une autre méthode basée sur des complexes topologiques a été développée dans le but de segmenter les objets dans des images 2D et 3D. Cette méthode segmente des objets dans des images multidimensionnelles et fournit une réponse à certains problèmes habituels en restant robuste vis à vis du bruit et de la variabilité de l'arrière-plan. Son application aux images de grande taille peut se faire en utilisant des superpixels. Nous avons également montré que l'homologie relative détecte le mouvement d’objets dans une séquence d'images qui apparaissent et disparaissent du début à la fin. Enfin, nous posons les bases d’un ensemble de méthodes d'analyse d'images basé sur la théorie des faisceaux qui permet de fusionner des données locales en un ensemble cohérent. De plus, nous proposons une seconde approche qui permet de comprendre et d'interpréter la structure d’une image en utilisant les invariants fournis par la cohomologie des faisceaux. / Algebraic topology, which is often appears as an abstract domain of mathematics, can bring new concepts in the execution of the image processing tasks. Indeed, these tasks might be complex and limited by different factors such as the need of prior parameters, the influence of the background, the superposition of objects. In this thesis, we propose methods derived from algebraic topology that differ from classical image processing methods by integrating local information at global scales through topological invariants. A first method of image segmentation was developed by adding topological characteristics calculated through persistent homology to classical statistical characteristics. Another method based on topological complexes built from pixels was developed with the purpose to segment objects in 2D and 3D images. This method allows to segment objects in multidimensional images but also to provide an answer to known issues in object segmentation remaining robust regarding the noise and the variability of the background. Our method can be extended to large scale images by using the superpixels concept. We also showed that the relative version of homology can be used effectively to detect the movement of objects in image sequences. This method can detect and follow objects that appear and disappear in a video sequence from the beginning to the end of the sequence. Finally, we lay the foundations of a set of methods of image analysis based on sheaf theory that allows the merging of local data into a coherent whole. Moreover, we propose a second approach that allows to understand and interpret scale analysis and localization by using the sheaves cohomology.

Cette thèse pour but de développer des méthodes de segmentation pour des scènes fortement structurées (ex. bâtiments et environnements urbains) ou faiblement structurées (ex. paysages ou objets naturels). En particulier, les images de bâtiments peuvent être décrites en termes d'une grammaire de formes, et une dérivation de cette grammaire peut être inférée pour obtenir une segmentation d'une image. Cependant, il est difficile et long d'écrire de telles grammaires. Pour répondre à ce problème, nous avons développé une nouvelle méthode qui permet d'apprendre automatiquement une grammaire à partir d'un ensemble d'images et de leur segmentation associée. Des expériences montrent que des grammaires ainsi apprises permettent une inférence plus rapide et produisent de meilleures segmentations. Nous avons également étudié une méthode basée sur les auto-contextes pour segmenter des scènes fortement structurées et notamment des images de bâtiments. De manière surprenante, même sans connaissance spécifique sur le type de scène particulier observé, nous obtenons des gains significatifs en qualité de segmentation sur plusieurs jeux de données. Enfin, nous avons développé une technique basée sur les réseaux de neurones convolutifs (CNN) pour segmenter des images de scènes faiblement structurées. Un filtrage adaptatif est effectué à l'intérieur même du réseau pour permettre des dépendances entre zones d'images distantes. Des expériences sur plusieurs jeux de données à grande échelle montrent là aussi un gain important sur la qualité de segmentation / The aim of this thesis is to develop techniques for segmenting strongly-structuredscenes (e.g. building images) and weakly-structured scenes (e.g. natural images). Buildingimages can naturally be expressed in terms of grammars and inference is performed usinggrammars to obtain the optimal segmentation. However, it is difficult and time consum-ing to write such grammars. To alleviate this problem, a novel method to automaticallylearn grammars from a given training set of image and ground-truth segmentation pairs isdeveloped. Experiments suggested that such learned grammars help in better and fasterinference. Next, the effect of using grammars for strongly structured scenes is explored.To this end, a very simple technique based on Auto-Context is used to segment buildingimages. Surprisingly, even with out using any domain specific knowledge, we observedsignificant improvements in terms of performance on several benchmark datasets. Lastly,a novel technique based on convolutional neural networks is developed to segment imageswithout any high-level structure. Image-adaptive filtering is performed within a CNN ar-chitecture to facilitate long-range connections. Experiments on different large scale bench-marks show significant improvements in terms of performance

Esta pesquisa é resultado do reconhecimento, no mercado hoteleiro, de algumas vantagens em se conhecer o conjunto de necessidades específicas de um determinado tipo de hóspede. O turismo de negócios é visto como um importante segmento do mercado turístico e como forte gerador da demanda hoteleira em algumas regiões. A proposta é, a partir da apresentação de características do mercado turístico e do setor hoteleiro, discutir a percepção do turista de negócios quanto aos atributos oferecidos pelos hotéis que justificam a preferência por dado empreendimento para sua estadia. A pesquisa bibliográfica, aliada a um estudo exploratório na cidade de São Paulo, permite algumas conjecturas a respeito do comportamento do turista de negócios no que diz respeito à hospedagem. Em virtude do dinamismo desta demanda, espera-se incentivar com esta pesquisa a continuidade, por profissionais e acadêmicos, dos estudos a respeito do mercado hoteleiro para o turismo de negócios. / business tourist behavior concerning hosting. Due to the dynamism of hotel demand, is expected that professionals and academics be stimulated to prospect and study the hotel market for business tourism. This research is the outcome of the recognition, in the hotel market, of the advantages of knowing the needs of a certain guest profile. Business tourism is an important segment of the tourism market and generates a great demand in some specific regions. The purpose of the research is, from the overview of the characteristics of tourism market and hotels segment, discuss the business tourist perspective of the attributes offered by hotels that justify their preference for a specific project. Research bibliography combined with an exploration in the city of São Paulo, allowed some hypothesis regarding

The task of object detection and pose estimation has widely been done using template matching techniques. However, these algorithms are sensitive to outliers and occlusions, and have high latency due to their iterative nature. Recent research in computer vision and deep learning has shown great improvements in the robustness of these algorithms. However, one of the major drawbacks of these algorithms is that they are specific to the objects. Moreover, the estimation of pose depends significantly on their RGB image features. As these algorithms are trained on meticulously labeled large datasets for object's ground truth pose, it is difficult to re-train these for real-world applications. To overcome this problem, we propose a two-stage pipeline of convolutional neural networks which uses RGB images to localize objects in 2D space and depth images to estimate a 6DoF pose. Thus the pose estimation network learns only the geometric features of the object and is not biased by its color features. We evaluate the performance of this framework on LINEMOD dataset, which is widely used to benchmark object pose estimation frameworks. We found the results to be comparable with the state of the art algorithms using RGB-D images. Secondly, to show the transferability of the proposed pipeline, we implement this on ATLAS robot for a pick and place experiment. As the distribution of images in LINEMOD dataset and the images captured by the MultiSense sensor on ATLAS are different, we generate a synthetic dataset out of very few real-world images captured from the MultiSense sensor. We use this dataset to train just the object detection networks used in the ATLAS Robot experiment.