Les travaux menés dans cette thèse concernent le développement d'un dispositif de vision permettant l'estimation de caractéristiques géométriques d'un feu de végétation en propagation. Ce dispositif est composé de plusieurs systèmes de stéréovision multimodaux générant des paires d'images stéréoscopiques à partir desquelles des points tridimensionnels sont calculés et les caractéristiques géométriques de feu tels que sa position, vitesse, hauteur, profondeur, inclinaison, surface et volume sont estimées. La première contribution importante de cette thèse est la détection de pixels de feu de végétation. Tous les algorithmes de détection de pixels de feu de la littérature ainsi que ceux développés dans le cadre de cette thèse ont été évalués sur une base de 500 images de feux de végétation acquises dans le domaine du visible et caractérisées en fonction des propriétés du feu dans l'image (couleur, fumée, luminosité). Cinq algorithmes de détection de pixels de feu de végétation basés sur la fusion de données issues d'images acquises dans le domaine du visible et du proche-infrarouge ont également été développés et évalués sur une autre base de données composée de 100 images multimodales caractérisées. La deuxième contribution importante de cette thèse concerne l'utilisation de méthodes de fusion d'images pour l'optimisation des points appariés entre les images multimodales stéréoscopiques.La troisième contribution importante de cette thèse est l'estimation des caractéristiques géométriques de feu à partir de points tridimensionnels obtenus depuis plusieurs paires d'images stéréoscopiques et recalés à l'aide de relevés GPS et d'inclinaison de tous les dispositifs de vision.Le dispositif d'estimation de caractéristiques géométriques à partir de systèmes de stéréovision a été évalué sur des objets rigides de dimensions connues et a permis d'obtenir les informations souhaitées avec une bonne précision. Les résultats des données obtenues pour des feux de végétation en propagation sont aussi présentés. / This thesis presents the geometrical characteristics measurement of spreading vegetation fires with multimodal stereovision systems. Image processing and 3D registration are used in order to obtain a three-dimensional modeling of the fire at each instant of image acquisition and then to compute fire front characteristics like its position, its rate of spread, its height, its width, its inclination, its surface and its volume. The first important contribution of this thesis is the fire pixel detection. A benchmark of fire pixel detection algorithms and of those that are developed in this thesis have been on a database of 500 vegetation fire images of the visible spectra which have been characterized according to the fire properties in the image (color, smoke, luminosity). Five fire pixel detection algorithms based on fusion of data from visible and near-infrared spectra images have also been developed and tested on another database of 100 multimodal images. The second important contribution of this thesis is about the use of images fusion for the optimization of the matching point’s number between the multimodal stereo images.The second important contribution of this thesis is the registration method of 3D fire points obtained with stereovision systems. It uses information collected from a housing containing a GPS and an IMU card which is positioned on each stereovision systems. With this registration, a method have been developed to extract the geometrical characteristics when the fire is spreading.The geometrical characteristics estimation device have been evaluated on a car of known dimensions and the results obtained confirm the good accuracy of the device. The results obtained from vegetation fires are also presented.

Abstract Purpose: is to identify the challenges of multimodal transport service of ESLSE in the Europe trade routes operation from origin up to destination and to analyze the contribution of network partners to ESLSE solving these challenges.   Design\Data collection\Approach: This thesis relied on qualitative research design in order to identify the key challenges of multimodal transport services and to show how the network partners contribute to solving these challenges. Moreover, to achieve the objective of the study, semi structure interviews were used, and the responses from 14 experts were analyzed.   Theoretical framework: The theoretical base and concepts are used to determine the theoretical boundaries of the research. It provides an overview of the literature that exists in the challenges of multimodal transport services. The theoretical framework firstly encompassed the logistics performance and personnel, followed by the description of multimodal transport services, customs facilitation, ICT, infrastructure, network partner integration and port administration. Lastly, theoretical synthesis is developed that are identified from the theory to demonstrate the interrelationship between individual concepts.   Findings: The findings of this study revealed that multimodal transport service was impacted by a number of challenges even if the network partners contributed some solutions. The results of the study showed that the lack of skilled logistics personnel; poor ICT system; lack of integration between network partners; lack of effective infrastructure; inadequate and ineffective capacity of trucks; material theft; corruption; security risk; lack of prompt response in the operation between network partners; monopoly of the operation by ESLSE; and lack of quality of transported cargo are the main challenges for multimodal transport services. Managerial implication: For managers of network partners, it is recommended that they should be aware that the challenge of multimodal transport service is the result of poor ICT systems; lack of integration between network partners; lack of effective infrastructure; lack of skilled logistic personnel; inadequate and ineffective capacity of trucks; material theft; corruption; security risk; lack of quality of transported cargo; and lack of prompt response in the operation between the network partners Hence, they should take action to solve the problems by integrating with shipping agents and steering committees such as customers office, transport minister, and maritime authority. Moreover, the shipping agents managers can use the results of the study to enhance their contribution in order to solve the challenges of multimodal transport services by communicating with the ESLSE.   Limitations: The small sample size without adequately diverse geographical spread and sample of shipping agents and dry ports since the study is only limited to Europe trade routes and it did not take in to consideration other continents like Africa and Asia which could have given additional information on the topic. The sample of shipping agent and dry ports was limited with 3 out of 11 and 3 out of 7 respectively.   Originality\Value: This thesis is one of the first to analyze the challenges of multimodal transport in the case of Ethiopia to Europe trade operation by interviewing both the ESLSE and shipping agents.

Le cancer de la prostate est le cancer le plus fréquent chez l'homme en France et aux pays occidentaux. Il est la troisième cause de décès liés au cancer, étant responsable d'environ 10% des morts. La curiethérapie, une technique de radiothérapie, est liée à une meilleure qualité de vie après le traitement, par rapport aux autres méthodes de traitement. La curiethérapie de la prostate consiste à insérer des sources radioactives dans la prostate afin de délivrer une dose d'irradiation localisée à la tumeur tout en protégeant les tissus sains environnants. L'imagerie multimodale est utilisée afin d'améliorer la précision du traitement. Les images Tomodensitométriques préopératoires, appelées Computed Tomography (CT), peuvent être utilisées pour calculer une distribution personnalisée et plus précise de dose. Pendant l'intervention, le chirurgien utilise un système de guidage temps-réel par l'Ultrason Transrectale, Transrectal Ultrasound (TRUS), pour placer correctement les sources radioactives dans leurs positions souhaitées. Par conséquent, si les positions des sources sont déterminées sur l'image CT, elles doivent être transférées à l'image US. Cependant, un recalage US/CT direct et robuste est difficilement envisageable parce que les tissus mous, telle que la prostate, offrent peu de contraste en CT et en US. En revanche, l'Imagerie par Résonance Magnétique (IRM) fournit un meilleur contraste et peut, potentiellement, améliorer le traitement en améliorant la visualisation. Donc, ces trois modalités (IRM, CT et US) doivent être correctement alignées. Pour compenser les déformations de la prostate, due au changement de taille et forme entre les différentes acquisitions, un recalage non-rigide est nécessaire. Une méthode de recalage entièrement automatique est nécessaire, afin de faciliter son intégration au bloc opératoire. Nous proposons dans un premier temps un recalage IRM/CT basé sur la maximisation de l'information mutuelle en combinaison avec un champ de déformation paramétré par B-Splines. Nous proposons de contraindre le recalage sur des volumes d'intérêt (VOIs) afin d'améliorer la robustesse et le temps de calcul. L'approche proposée a été validée sur des jeux de données cliniques. Une évaluation quantitative a montré que l'erreur de recalage est égale à 1.15±0.20 mm; qui répond à la précision clinique souhaitée. Ensuite, nous proposons un deuxième recalage US/IRM, où nous utilisons une approche multi-résolution pour éviter les minima locaux et améliorer le temps de calcul. Un critère de similarité, qui met en corrélation l'intensité de l'image US avec l'intensité ainsi que le gradient de l'image IRM, a été utilisé afin de trouver la transformation qui aligne les deux images. Cette méthode a été validée sur un fantôme de prostate dans un premier temps pour évaluer sa faisabilité. Ensuite, elle a été validée sur des jeux de données cliniques en utilisant des critères qualitatives et quantitatives. La distance Hausdorff a montré que l'erreur de recalage est égale à 1.44±0.06 mm. L'approche proposée dans ce travail permet d'aller vers un protocole de curiethérapie guidée par l'imagerie multimodale qui puisse améliorer la précision globale de cette procédure. Malgré ces résultats plutôt encourageants, les travaux futurs impliqueront une évaluation plus approfondie sur plus de jeux de données afin d'évaluer la fiabilité et l'efficacité de cette méthode avant de l'intégrer au bloc opératoire. / Prostate cancer is the most common cancer in men in France and western countries. It is the third leading cause of death from cancer, being responsible for around 10% of deaths. Brachytherapy, a radiotherapy technique, is associated with a better health-related quality of life after the treatment, compared to other treatment techniques. Prostate brachytherapy involves the implantation of radioactive sources inside the prostate to deliver a localized radiation dose to the tumor while sparing the surrounding healthy tissues. Multi-modal imaging is used in order to improve the overall accuracy of the treatment. The pre-operative Computed Tomography (CT) images can be used to calculate a personalized and accurate dose distribution. During the intervention, the surgeon utilizes a real-time guiding system, Trasnrectal Ultrasound (TRUS), to accurately place the radioactive sources in their desired pre-planned positions. Therefore, if the positions of the sources were determined on CT, they need to be transferred to US. However, a robust and direct US/CT registration is hardly possible since they both provide low soft tissue contrast. Magnetic Resonance Imaging (MRI), on the other hand, has a superior contrast and can potentially improve the treatment planning and delivery by providing a better visualization. Thus, these three modalities (MRI, US and CT) need to be accurately registered. To compensate for prostate deformations, caused by changes in size and form between the different acquisitions, non-rigid registration is essential. Fully automatic registration methodology is necessary in order to facilitate its integration in a clinical workflow. At first, we propose a registration between pre-operative MR and CT images based on the maximization of the mutual information in combination with a deformation field parameterized by cubic B-Splines. We propose to constrain the registration to volumes of interest (VOIs) in order to improve the robustness and the computational efficiency. The proposed approach was validated on clinical patient datasets. Quantitative evaluation indicated that the overall registration error was of 1.15±0.20 mm; which satisfies the desired clinical accuracy. Then, we propose a second intra-operative US/MRI registration, where a multi-resolution approach is implemented to reduce the probability of local minima and improve the computational efficiency. A similarity measure, which correlates intensities of the US image with intensities and gradient magnitude of the MRI, is used to determine the transformation that aligns the two images. The proposed methodology was validated on a prostate phantom at first to assess its feasibility. Subsequently, the method was validated on clinical patient datasets and evaluated using qualitative and quantitative criteria, resulting in a registration error of 1.44±0.06 mm. The approach proposed in this work allows going towards a multimodal protocol for image-guided brachytherapy which can improve the overall accuracy of this procedure. Despite such encouraging results, future work will involve further evaluation on a larger number of datasets in order to assess the reliability and the efficiency of this methodology before integrating it in a clinical workflow.

Les travaux effectués dans cette thèse industrielle concernent l'amélioration du calculateur d'itinéraire de Cityway, société spécialisée dans les technologies de l’information appliquées à la mobilité.Nous avons d'abord établi un état de l'art exhaustif, accompagné d'une mise en perspective de l'existant Cityway avec celui-ci. Cela nous a permis d'aider l'entreprise à prendre du recul sur son produit et de justifier les axes de recherche choisis pour nos travaux.Nous nous sommes ensuite intéressés à l'aspect multicritère du problème. En effet, le calculateur, basé sur l'algorithme de Dijkstra, permet de trouver des trajets minimisant une somme pondérée de critères. Nous avons développé un algorithme multilabel permettant de conserver et étendre plusieurs labels au même nœud. Malgré une légère augmentation des temps de calculs, des résultats satisfaisants ont été obtenus dans une application bicritère de ce nouvel algorithme.Nous avons également travaillé sur la génération et la sélection de trajets alternatifs. La génération s'appuie sur les algorithmes monolabel ou multilabel. La sélection s'appuie quant à elle sur la définition d'une distance entre les solutions et des méthodes de regroupement.Enfin, nous nous sommes intéressés à l'optimisation du calcul du critère lexicographique de durée minimale dans le cas bicritère. Pour qu'un trajet soit intéressant, il faut qu'il soit optimal sur les critères usuels, mais aussi qu'il dure le moins longtemps possible. L'utilisation de certaines propriétés sur ce critère permet de réduire des temps de calcul initialement trop longs. / The work carried out in this industrial PhD aims at improving the route planner of Cityway, a company specialized in information technologies applied to mobility. We first established an exhaustive state of the art, and compared it to the existing Cityway product. This allowed us to help the company take a step back from its urgent needs, and justify the research guidelines chosen for our work.We then looked at the multi-criteria aspect of the problem. Indeed, the trip planner, based on the Dijkstra algorithm, makes it possible to find paths minimizing a weighted sum of criteria. We have developed a multilabel algorithm to maintain and extend multiple labels at the same node. Despite a slight increase in computation time, satisfactory results were obtained in a bicriteria application of this new algorithm.We also worked on the generation and selection of alternative routes. The generation algorithm relies on the existing monolabel or newly developed multilabel algorithms. The selection algorithm is based on the definition of a distance between trips and adaptations of existing clustering algorithms to this specific case.Finally, we were interested in what we called the lexicographic criterion. For a trip to be interesting, it must be optimal on the usual criterion of earliest arrival, and, for trips arriving at the same time, on the latest departure criterion. The use of certain properties on this criterion makes it possible to reduce computation times on the bicriteria case.

Submitted by Claudia Rocha (claudia.rocha@udesc.br) on 2018-03-15T16:54:10Z No. of bitstreams: 1 PGCA16MA213.pdf: 1442712 bytes, checksum: 89506b3035f7acc4dd2a3d494c07e5e8 (MD5) / Made available in DSpace on 2018-03-15T16:54:10Z (GMT). No. of bitstreams: 1 PGCA16MA213.pdf: 1442712 bytes, checksum: 89506b3035f7acc4dd2a3d494c07e5e8 (MD5) Previous issue date: 2016-10-27 / The surgical procedure mastectomy is the most appropriate treatment in cases of tumors of the mammary glands in female dogs. In addition to the drug combinations that seek multimodal anesthesia, the use of regional anesthesia techniques can also be used for this procedure. Blockade of the transversus abdominis plane (TAP) is a regional anesthetic technique designed to desensitize the branches that innervate the abdominal muscles, abdominal subcutaneous tissue and parietal peritoneum, with the deposition of the local anesthetic in the fascia between the internal oblique and transversus abdominis identified by ultrasound. The objective of this study was to evaluate the analgesic effectiveness of the TAP block, associated with intercostal blocks in trans and postoperative of animals submitted at radical unilateral mastectomy, compared to analgesia with continuous infusion of morphine, lidocaine, and ketamine (MLK), and testing sensitivity to rescue analgesics three different methods for assessing acute pain in dogs: Visual Analogue Scale (VAS), Glasgow Composite pain Scale and Melbourne Scale. The project was carried out from April to July of 2016, with animals from the clinical routine Veterinary Hospital HCV-CAV / UDESC. Three groups were used with 6 dogs each: TAP Group, submitted to the transversus abdominis block associated with intercostal blocks, performed with bupivacaine without vasoconstrictor; Group MLK subjected to intravenous infusion of morphine, lidocaine and ketamine (MLK) during surgeries; and control group, with animals that did not receive any locoregional anesthesia or analgesics infusion. The application of analgesic rescues with fentanyl was higher in the control group during the trans-operative assessment and did not differ between TAP and MLK groups. In the postoperative assessments, there were no differences between the groups in relation to the request redemptions of morphine, however, Glasgow and EVA were more sensitive. The TAP block associated with intercostal blocks can replace the infusion of MLK in bitches submitted to unilateral mastectomy / O procedimento cirúrgico de mastectomia é o tratamento mais indicado nos casos de tumores de glândulas mamárias em cadelas. Além das associações medicamentosas que buscam a anestesia multimodal, o emprego de técnicas de anestesia locorregional também podem ser utilizadas para esse procedimento. O bloqueio do plano transverso abdominal (TAP), é uma técnica de anestesia regional projetada para dessensibilizar os ramos que inervam os músculos abdominais, tecido subcutâneo abdominal e peritônio parietal, por meio da deposição do anestésico local na fáscia entre o músculo oblíquo interno e transverso abdominal identificados através da ecografia. O objetivo deste trabalho, foi avaliar a eficiência analgésica do bloqueio TAP, associado aos bloqueios intercostais, no trans e pós-operatórios de animais submetidos à mastectomia unilateral radical, comparativamente à analgesia com infusão continua de morfina, lidocaína e cetamina (MLK), testando a sensibilidade ao resgate analgésicos de três diferentes métodos de avaliação de dor aguda em cães: Escala Visual Analógica (EVA), Escala Composta de Dor de Glasgow e Escala de Melbourne. O projeto foi desenvolvido no período de Abril a Julho do ano de 2016, com animais provenientes da rotina clínica do Hospital de Clínicas Veterinárias HCV-CAV/UDESC. Foram utilizados três grupos com 6 cadelas cada: Grupo TAP, submetido ao bloqueio transverso abdominal associado a bloqueios intercostais, realizados com bupivacaína sem vasoconstritor; Grupo MLK, submetido a infusão intravenosa de morfina, lidocaína e cetamina (MLK) durante o transoperatório; e Grupo controle, com animais que não receberam nenhum tipo de anestesia locorregional ou infusão de analgésicos. O requerimento de resgastes analgésico com fentanil, foi maior no grupo Controle durante a avaliação trans-operatória, e não diferiu entre os grupos TAP e MLK. Nas avaliações de pós-operatório, não houveram diferenças entre os grupos em relação ao requerimento de resgates com morfina, no entanto, Glasgow e a EVA mostraram-se mais sensíveis. O bloqueio TAP associado aos bloqueios intercostais, causa analgesia similar à infusão de MLK em cadelas submetidas a mastectomia unilateral

Made available in DSpace on 2014-06-11T19:29:12Z (GMT). No. of bitstreams: 0 Previous issue date: 2009-10-23Bitstream added on 2014-06-13T18:38:58Z : No. of bitstreams: 1 hiratsuka_a_me_ilha.pdf: 1245218 bytes, checksum: e01d077a2a1eac561a669b606897fe0a (MD5) / Este trabalho apresenta uma análise de impactos ambientais e econômico, entre dois cenários para o transporte de açúcar da Usina Pioneiros Bioenergia S/A, até o Terminal Exportador de Açúcar Guarujá. O Cenário 1 é composto somente pelo modo de transporte rodoviário, enquanto o Cenário 2 é constituído pelo transporte multimodal, ou seja, a combinação do modo hidroviário e o rodoviário. Para proceder a análise dos dois Cenários, consideraram-se parâmetros econômico e ambientais. O parâmetro econômico é o valor do frete por tonelada de açúcar transportado. Para o parâmetro ambiental, foi utilizado um modelo de índices de destruição ambiental, causados pela emissão de poluentes na queima de combustível. Através dos resultados dos parâmetros obtidos, pôde-se comprovar que a utilização de transporte multimodal, principalmente para grandes quantidades de carga a longas distâncias, possibilita ganhos econômicos e também a redução de impactos ambientais / This study presents the analysis of the environmental and economic impacts in the transportation of sugar, it compares two scenarios of sugar transportation from Usina Pioneiros Bioenergia S/A to Guarujá Export Sugar Terminal. In scenario 1 the sugar is transported using only road transportation, while in the scenario 2 multimodal transport is used, where road and waterway transportation are used. Economic and environmental parameters were taken into account in the analysis of both situations. The economic parameter is the value of freight per ton of transported sugar. For the environmental parameter the environmental destruction index was used, it quantifies the impact of fuel emissions. The results obtained show that, especially for large quantities of cargo over long distances, the use of multimodal transport brings economic gains and also reduces negative environmental impact

Este trabalho trata do problema de otimização do resultado para uma empresa (maximização da receita) que presta o serviço de transporte de carga conteinerizada por via marítima num cenário onde a demanda dos clientes supera a oferta de capacidade disponível. No trabalho são descritas as características do problema da escolha do mix de carga em um horizonte de planejamento típico de várias semanas (tipicamente de 6 a 8) com detalhamento diário, ou seja, caracterizando um problema de múltiplos períodos. Também são consideradas as restrições de peso e de volume dos navios utilizados e disponibilidade de contêineres. A proposta é maximizar o retorno gerado pelo transporte através da escolha do conjunto de clientes que apresentem a melhor rentabilidade total dentre um universo definido em um processo mensal de levantamento de demanda. Contribuem para o aumento da complexidade da modelagem do problema a necessidade de utilização de variáveis inteiras em função da escolha de cada carga de cada cliente, considerando também vários navios em múltiplas rotas e programação do reposicionamento de contêineres vazios em conjunto com os cheios. Para a resolução deste problema foi desenvolvida uma heurística para a solução do modelo matemático que analisa a rentabilidade relativa de cada carga segundo critérios de ocupação (volume e peso) dos navios. A partir daí, as cargas mais rentáveis são alocadas e são verificadas as restrições de capacidade do navio utilizado e a disponibilidade de contêineres vazios em um processo interativo, onde são analisadas as opções de reposicionamento dos contêineres vazios até a obtenção da solução ou recusa da carga. A heurística proposta permite considerar diferentes critérios de rentabilidade dos clientes e cargas, tendo sido aplicada a um problema real para a sua validação. Os resultados obtidos mostram uma oportunidade de melhoria no processo atual tanto no aspecto de aumento da lucratividade do negócio, objetivo principal do trabalho, como em outras questões como a programação antecipada do reposicionamento dos contêineres vazios e a flexibilização do espaço alocado para esta movimentação. / This work presents a problem of profit optimization (revenue maximization) in container shipment company by maritime modal in a scenario where demand surplus the available capacity. In this work is described the characteristics of cargo mix problem in a multiple planning period of several weeks (typically from 6 to 8) with diary scheduling, which means that it turns to a multiple period problem. It is also considered the weight and volume restrictions of the boats and availability of containers. The proposal is to maximize the return generated by the transportation service by choosing the set of clients that generates the bigger profitability in a major group defined in a monthly process of demand evaluation. The complexity of the problem is enhanced by the utilization of integer variables, because of the need of choosing each client and its cargo, also considering that it may be more than one ship in multiple routes and the empty containers transportation along with the cargo. For the solution of the problem, it was developed a heuristic for the solution of the mathematical model that analyses the profitability of each client according to occupation (volume and weight) of the ships. From this moment on, the most profitable cargos are allocated and the restriction of capacity of the ships and availability of empty containers are in a interactive process until the solution is found or the client discharged. This method was tested in a real problem of a maritime transportation company, comparing the various profitability criteria of the clients and cargos. The obtained results show an opportunity of actual process improvement concerning the profitability of the business, main objective of this work, as in the anticipation of the empty containers repositioning and flexibilization of the space allocated to this operation.

The purpose with this thesis is to analyze how written text and images interact in a traditional and digital learning component in chemistry. • How does text and images interact in a traditional and digital teaching material for grades 4-6 in the field of water characteristics and circuits? • What are the multimodal similarities and differences between the two teaching tools? To answer the questions, a quality multimodal text analysis has been used together with a model that focuses on the text's overall structure and the interaction between the text's parts. The result showed that there was interaction between text and images in all the teaching materials. The properties and circuits of the water are reinforced through pictures and accompanying text in both study units. The difference was in the traditional teaching medium, where there were more pictures and shorter text, while in the digital learning medium there were long information-proof texts.

Literacy is an invaluable asset to have, and has allowed for communication, documentation and the spreading of ideas since the beginning of the written language. With technological advancements, and new possibilities to communicate, it is important to question the degree to which people’s abilities to utilise these new methods have developed in relation to these emerging technologies. The purpose of this bachelor’s thesis is to analyse the state of students’ at Dalarna University mulitimodal literacy, as well as their experience of multimodality in their education. This has led to the two main research questions: What is the state of the students at Dalarna University multimodal literacy? And: How have the students at Dalarna University experienced multimodality in education? The paper is based on a mixed-method study that incorporates both a quantitative and qualitative aspect to it. The main thrust of the research paper is, however, based on a quantitative study that was conducted online and emailed to students via their program coordinators. The scope of the research is in audio-visual modes, i.e. audio, video and images, while textual literacy is presumed and serves as an inspiration to the study. The purpose of the study is to analyse the state of the students’ multimodal literacy and their experience of multimodality in education. The study revealed that the students at Dalarna University have most skill in image editing, while not being very literate in audio or video editing. The students seem to have had mediocre experience creating meaning through multimodality both in private use and in their respective educational institutions. The study also reveals that students prefer learning by means of video (rather than text or audio), yet are not able to create meaning (communicate) through it.

Cette thèse s'inscrit dans le contexte de la caractérisation du cancer de la prostate à partir de données d'imagerie multimodale. L'objectif est d'identifier et de caractériser les tumeurs prostatiques à partir d'observation in-vivo, incluant imagerie par résonance magnétique multiparamétrique (IRMm), tomodensitométrie (TDM) et tomographie par émission de positons (TEP), avec une référence biologique sous forme de lames histologiques fournies par l'analyse anatomopathologique de la prostate après prostatectomie radicale. Nous proposons dans un premier temps deux méthodes de recalage afin de réaliser la mise en correspondance des données multimodales dans un référentiel spatial commun défini par l'IRM. Le premier algorithme s'attache à l'alignement des images TDM/TEP et IRM, en combinant information de contour et probabilité de présence de la prostate. Le second a pour objectif de recaler les images histologiques et IRM. En suivant le protocole de Stanford, la pièce de prostatectomie est découpée plus finement fournissant ainsi plus de lames histologiques qu'en routine clinique. Leur correspondance avec les coupes IRM est alors estimée à l'aide de l'information à priori de la découpe et de points saillants (SURF) extraits dans les deux modalités. Cette initialisation sert de base à un recalage affine puis non-rigide basé sur l'information mutuelle et des cartes de distance obtenues à partir des structures intraprostatiques. Dans un second temps, des caractéristiques structurelles (Haar, Gabor, etc) et fonctionnelles (Ktrans, Kep, SUV, TLG, etc) sont extraites en chaque voxel de la prostate à partir des images IRM et TEP. À ces vecteurs de caractéristiques sont associés les labels biologiques correspondant obtenus lors de l'analyse anatomopathologique. Parmi ces labels, nous disposons d'un score d'agressivité appelé score de Gleason et de scores immunohistochimiques quantifiant certains aspects biologiques des tissus, tels que l'hypoxie et la prolifération cellulaire. Finalement, ces couples (vecteurs de caractéristiques/informations biologiques) servent de données d'apprentissage à l’entraînement de classifieurs RF et SVM, utilisés par la suite pour caractériser les tumeurs à partir de nouvelles observations in-vivo. Dans ces travaux, nous réalisons une étude de faisabilité sur neuf patients. / This thesis concerns the prostate cancer characterization based on multimodal imaging data. The purpose is to identify and characterize the tumors using in-vivo observations including mMRI and PET/CT, with a biological reference obtained from anatomopathological analysis of radical prostatectomy specimen providing histological slices. Firstly, we propose two registration methods to match the multimodal images in the the spatial reference defined by MRI. The first algorithm aims at aligning PET/CT images with MRI by combining contours information and presence probability of the prostate. The objective of the second is to register the histological slices with the MRI. Based on the Stanford protocol, a thinner cutting of the radical prostatectomy specimen is done providing more slices compared to clinical routine. The correspondance between histological and MRI slices is then estimated using a combination of the prior information of the slicing and salient points (SURF) extracted in both modalities. This initialization step allows for an affine and non-rigid registration based on mutual information and intraprostatic structures distance map. Secondly, structural (Haar, Garbor, etc) and functional (Ktrans, Kep, SUV, TLG, etc) descriptors are extracted for each prostate voxel over MRI and PET images. Corresponding biological labels obtained from the anatomopathological analysis are associated to the features vectors. The biological labels are composed by the Gleason score providing an information of aggressiveness and immunohistochemistry grades providing a quantification of biological process such as hypoxia and cell growth. Finally, these pairs (features vectors/biological information) are used as training data to build RF and SVM classifiers to characterize tumors from new in-vivo observations. In this work, we perform a feasibility study with nine patients.