Ces travaux de recherche s’inscrivent dans la problématique scientifique de reconstruction du relief sous un couvert végétal à partir d’observations aéroportées. Le scanner laser aéroporté est une technique d’imagerie très prometteuse, notamment pour l’observation des zones forestières. Sa déclinaison "onde complète" consiste à émettre une impulsion laser et à enregistrer temporellement l’intégralité des échos de retour réfléchis par la scène. La forme des échos de retour fournit des informations sur l’épaisseur optique du couvert végétal. De nombreux systèmes commerciaux sont en exploitation, en particulier en topographie ou en bathymétrie. Mais ces systèmes ne sont pas dédiés à l’observation de la végétation. L’objectif de cette thèse est l’étude de l’intérêt de ces systèmes pour la construction de modèle numérique de terrain (MNT) sous couvert végétal. Elle est basée sur le développement d’outils de simulation du signal temporel incident au capteur lidar et de traitement des données. Dans un premier temps, le modèle physique de lidar onde complète, DELiS (n-Dimensional Estimation of Lidar Signals) a été développé. Il permet de simuler l’observation de scènes de végétation réalistes, tout en incluant la prise en compte de l’environnement extérieur (atmosphère, soleil) ainsi que des caractéristiques de la source et de la chaîne de détection (bruits de mesure). DELiS a été validé par confrontation à des résultats analytiques. Ensuite, DELIS a permis de comprendre et d’évaluer l’importance des diffusions multiples dans le couvert en fonction du champ de vue du lidar mais aussi de justifier l’utilisation d’acquisitions aéroportées petit champ pour simuler le signal d’un lidar spatial plus grand champ. Dans une deuxième étape, ses capacités de simulation ont été utilisées afin d’étudier l’intérêt du lidar onde complète pour la reconstruction d’un MNT sous couvert végétal. Dans ce but, nous avons développé et implémenté numériquement une méthode originale de traitement et de classification des données lidar onde complète permettant de séparer les échos lidar provenant du sol de ceux provenant de la végétation. Après classification des échos, nous avons reconstruit la géométrie du sol et des objets occultés par la végétation. Enfin, nous avons étudié comment combiner des données aéroportées acquises sous différents points de vue afin d’améliorer les reconstructions. Nos travaux montrent que le scanner laser aéroporté onde complète pourrait permettre d’obtenir en milieux forestier des reconstructions de la géométrie du terrain à des résolutions sub-métriques et avec une précision de l’ordre de 10 à 20 centimètres. La combinaison de visées multi-angulaire permet, par l’apport d’une quantité importante d’information supplémentaire, d’améliorer encore la reconstruction du MNT. Nous montrons cependant que les visées inclinées sont plus sensibles à la présence des troncs et branchages des arbres, éléments qui sont susceptibles d’introduire une erreur importante dans les processus de classification et de reconstruction. Pour cette raison, nous recommandons l’utilisation de la visée nadir pour la reconstruction mono-vue des modèles numériques de terrain, et nous proposons une méthode permettant de choisir de façon optimale les visées inclinées à ajouter pour l’observation détaillée d’une portion plus restreinte de la scène. / This research work regards the scientific challenge of reconstructing the ground and the object presents under a vegetation cover from airborne observations. Airborne laser scanning is a promising technology. Full-waveform devices are able to record the complete temporal return signal following the emission of a short laser pulse towards the ground. This offers a great potential for remote sensing of forested areas, since the laser pulse will travel through the vegetation. Many commercial systems are already operated for topography or bathymetry. Scientists have been using these systems for vegetation observation, even if they are not dedicated to this purpose. The objective of this thesis is to study the relevance of full-waveform lidars for the geometric reconstruction of digital terrain models (DTM) under vegetation. We also aim at developing simulation and data processing toolsthat will help design and optimize future sensors dedicated to vegetation observation. Our first task was the development of a new physical simulator for full-waveform lidar measurement. The DELiS model (n-Dimensional Estimation of Lidar Signals) is able tosimulate the observation of complex and realistic vegetation scenes while accounting for atmosphere and sun perturbations, and simulating the multiple scattering of the laser pulse in the canopy. We have also implemented a sensor model for simulation of the measurement, amplification and digitization noises. This operational simulation tool is a key asset for future physical studies as well as for designing and optimizing future sensors and data processing methods. After validating the DELiS model by confrontation with analytical results, we have used it for studying the interest of full-waveform lidar for digital terrain models reconstruction under vegetation. For this purpose, we have developed a full-waveform lidar data processing method for decomposition of the signals and classification of the lidar echoes into two classes : ’ground’ and ’vegetation’. We were then able to reconstruct ground geometry.Finally, we have led a study on the combination of multi-angular acquisitions for improvement of the reconstructions.Our work shows that airborne full-waveform lidar observations may allow ground reconstruction with sub-metric resolutions and a precision of 10 to 20 centimeters in forested areas. Combining multiple viewing angles provides additional data, and helps improving the precision of the reconstructions. Yet, we show that non-nadir viewing is much more sensitive to trunks and branches. These elements may be the cause of an additional error in the classification and reconstruction processes. For this reason, we recommend using nadir viewing for single-view ground reconstruction, and propose a method for optimally selecting non-nadir views for the detailed observation of restricted areas of interest.

Made available in DSpace on 2016-12-23T13:56:04Z (GMT). No. of bitstreams: 1 DISSERTACAO FINAL.pdf: 1304058 bytes, checksum: 22df93a24605e7e83e388686a27fb38f (MD5) Previous issue date: 2009-05-29 / The tuberculosis continues to represent a severe worldwide public health issue, particularly in developing countries. Besides early diagnosis and therapy, one of the great challenges for its control is the scarce knowledge available about the transmission mechanisms and the related risks. The retrospective cohort described in this study aims to evaluate the risk of Mycobacterium tuberculosis infection, measured by tuberculin skin testing, in contacts of patients with cavitary and noncavitary pulmonary tuberculosis. Identification and screening of index cases between july/2003 and december/2007 was performed based on analysis of two databases: National information system of disease notification and TB notes. Only index cases, with age above 18 years old, living at Cariacica, Serra, Vila Velha and Vitoria cities were screened. Demographic, clinical and laboratory informations of the index cases and contact informations were captured from the medical charts. Bacteriological confirmation based on mycobacterial cultures was captured from TB notes database. among the 1662 index cases identified, 320 met the inclusion criteria: 154 (48,1%) with cavitary disease (C group) and 166 (51,9%) with non-cavitary disease (NC group) based on chest xray. among these 320 index cases, 1257 contacts were identified. most cases of C and NC groups denied previous contact with tuberculosis: 70,2% and 60,1%, respectively. index cases reported cough ranging between 0 and 980 days; mean of 93,5 and median of 60 days; 25 and 75 percentis were 30 and 90 days, respectively. period of cough, which represent time of disease, was higher in the c group (p=0,01, OR 1,82, IC 95%:2,8-13,5) when the threshold was 60 days. higher number of positive sputum smear patients was observed in the C group (p <0,00, OR=5,86, IC 95%: 2,8-13,5). among the 1257 enrolled contacts, 555 (44%) were contacts of the C group and 702 (56%) of the NC group patients. both groups were similar regarding gender, age and chest xray images. however, more reactors to tuberculin skin testing (PPD &#8805; 10 mm) were found among contacts of C group: 48% versus 40,6% (p= 0,009, OR1,35, IC 95% 1,07-1,7 ). Taking into account images observed in chest xray, M tuberculosis infection rate was 1,7 higher among contacts of cavitary in comparison with non-cavitary patients. After logistic regression analysis this association was statistically significant for two variables: period of cough &#8805; 30 days (p=0, 0007OR=3, 20, IC 95%: 1, 5-8.6,) and positivity of sputum smears (p=0, 0039 OR=2, 47, IC 95: 1, 24-4, 81). Positivity of sputum smears was also related to the presence of cavitary disease. Index cases included in C group had 5, 86 more chance of having positive sputum smear (IC 95%: 2, 8-13, 5). Using infection prevalence rate of 30% or more as the threshold for transmission analysis, the three following variables were related to infection transmission to contacts: period of cough, cavitary disease and positive sputum smear. It was concluded that: a period of cough &#8805;30 days was associated with TB infection; smear positivity was associated with a higher chance of infection among contacts. The association between cavitary disease and TB infection transmission to contacts was marginally statistically significant. / Um dos principais desafios para o controle da tuberculose (TB) é ainda o pouco conhecimento disponível sobre os mecanismos intrínsecos de sua transmissão e os graus de risco associados a eles. Este estudo, uma coorte retrospectiva, teve por objetivo avaliar o risco de infecção por Mycobacterium tuberculosis, medido pela prova tuberculínica, em contatos de pacientes com tuberculose pulmonar cavitária e não cavitária. A identificação e a seleção dos casos índices foram realizadas por meio da análise de dois bancos de dados: Sistema de Informação de Agravos de Notificação (SINAN) e TB notes. O período abrangido foi de julho de 2003 a dezembro de 2007. Foram selecionados indivíduos residentes em quatro municípios da Região Metropolitana de Vitória (Cariacica, Serra, Vila Velha e Vitória). De todos os casos índices com idade superior a 18 anos notificados no SINAN, obteve-se os prontuários médicos para análise dos dados demográficos, clínicos e laboratoriais e as informações dos respectivos contatos. A confirmação bacteriológica por cultura dos pacientes foi obtida no banco de dados TB-notes. Dos 1662 casos índices identificados, 320 preencheram os critérios de inclusão do estudo 154 (48,1%) com doença cavitária (grupo C) e 166 (51,9%) sem doença cavitária (grupo NC) à radiografia do tórax. Os 320 casos índices geraram 1.257 contatos. A maioria dos casos do grupo C e NC não tinha história epidemiológica de contato prévio com a doença 70,2% e 60,1%, respectivamente. O tempo de tosse variou de 0-980 dias, com média de 93,5 dias (± 130 dias), e mediana, de 60 dias. Os percentis 25 e 75 foram, respectivamente, de 30 e 90 dias. O tempo de tosse, que, na prática, retrata o tempo de doença, foi maior nos pacientes do grupo C (p=0,01, RC 1,82, IC 95%: 1,01-3,04), quando o ponto de corte foi de 60 dias. Houve uma maior concentração de pacientes com baciloscopia positiva no grupo C, o que representa uma diferença estatisticamente significativa (p <0,00, RC=5,86, IC 95%: 2,8-13,5). Dos 1.257 contatos arrolados, 555 (44%) eram contatos do grupo C, e 702 (56%) do grupo NC. Os dois grupos de contatos foram semelhantes em relação ao gênero, idade e resultado da radiografia do tórax. Houve, entretanto, diferença na prevalência de positividade da prova tuberculínica (&#8805;10mm), que foi maior nos contatos do grupo C: 48% versus 40,6% do grupo NC (p= 0,009, RC 1,35, IC95% 1,07-1,7). Os casos índices com doença cavitária infectaram 1,7 vezes mais seus contatos do que os casos índices sem cavidade à radiografia do tórax (IC 95%: 0.94-3.08, p=0, 061). Após análise de regressão logística essa associação mostrou-se estatisticamente significativa para as variáveis: tempo de tosse superior a 30 dias (RC=3,20, IC 95%: 1,5-8.6, p=0, 0007) e positividade da baciloscopia do escarro (RC=2,47, IC 95: 1,24-4,81 p=0, 0039). A positividade da baciloscopia do escarro também esteve associada à presença de doença cavitária. Os casos índices incluídos nessa categoria possuíam 5,86 mais chance de apresentarem exame direto do escarro positivo (IC95%: 2,8-13,5). Essas três variáveis (tempo de tosse, doença cavitária e baciloscopia positiva do escarro) estiveram relacionadas à transmissão da infecção aos contatos, quando se utilizou, como ponto de corte para análise da transmissão, a presença de infecção igual ou superior a 30% nos contatos dos respectivos casos índices. Concluímos que o tempo de tosse superior a 30 dias e a baciloscopia positiva estão fortemente associados à transmissão por M.tuberculosis. Esta associação teve significância limítrofe para doença cavitária.

Vascular smooth muscle cell (VSMC) accumulation is a hallmark of atherosclerosis and vascular injury. However, fundamental aspects of proliferation and the phenotypic changes within individual VSMCs, which underlie vascular disease remain unresolved. In particular, it is not known if all VSMCs proliferate and display plasticity, or whether individual cells can switch to multiple phenotypes. To assess whether proliferation and plasticity in disease is a general characteristic of VSMCs or a feature of a subset of cells, multi-colour lineage labelling is used to demonstrate that VSMCs in injury-induced neointimal lesions and in atherosclerotic plaques are oligo-clonal, derived from few expanding cells, within mice. Lineage tracing also revealed that the progeny of individual VSMCs contribute to both alpha Smooth muscle actin (aSma)-positive fibrous cap and Mac-3-expressing macrophage-like plaque core cells. Co-staining for phenotypic markers further identified a double-positive aSma+ Mac3+ cell population, which is specific to VSMC-derived plaque cells. In contrast, VSMC-derived cells generating the neointima after vascular injury generally retained expression of VSMC markers and upregulation of Mac3 was less pronounced. Monochromatic regions in atherosclerotic plaques and injury-induced neointima did not contain VSMC-derived cells expressing a different fluorescent reporter protein, suggesting that proliferation-independent VSMC migration does not make a major contribution to VSMC accumulation in vascular disease. Similarly, VSMC proliferation was examined in an Angiotensin II perfusion model of aortic aneurysm in mice, oligo-clonal proliferation was observed in remodelling regions of the vasculature, however phenotypic changes were observed in a large proportion of VSMCs, suggesting that the majority of VSMCs have some potential to modulate their phenotype. To understand the mechanisms behind the inherent VSMC heterogeneity and observed functionality, the single cell transcriptomic techniques Smart-seq2 and the Chromium 10X system were optimized for use on VSMCs. The work within this thesis suggests that extensive proliferation of a low proportion of highly plastic VSMCs results in the observed VSMC accumulation after injury, and the atherosclerotic and aortic aneurysm models of cardiovascular disease.

Modern day power systems present an open access environment, inspiring participation from small scale and large power suppliers. With multiple players in the system driven by the market, proper monitoring and control of system becomes a major concern. This transformation is accompanied by dynamic consumption patterns and rising power demands. The expanding network encompassing EHV/AC network, HVDC and FACTS devices, along with increased penetration of renewable sources, viz. solar and wind energy at medium and low voltage levels, adds to the problem. Independent System Operators (ISO) are entrusted with ensuring smooth operation, and employing proper preventive measures to eliminate a possible cascade tripping leading to a partial or large-scale blackout. To aid the operator in the process of ensuring secure operation of the grid, there are many tools that provide required information and guidance. Power flow tracing is one such tool that aids the operator in congestion management, transmission pricing, transaction evaluation, loss allocation and reactive power optimization. In this thesis, a novel active power flow tracing approach is proposed that takes into account, the real-time operating conditions and network topology. It provides the decomposition of active power flow in a line into respective components injected by various generators in the system. It also provides the contribution of the generators to various loads in the system. The approach is simple and computationally fast, making it an ideal tool to aid preventive control decisions. Based on the proposed active power flow tracing, a congestion management approach is developed. The approach indicates the least number of generators that need to be coordinated for generation rescheduling, so as to alleviate overloading in affected transmission lines and transformers. The approach also takes into consideration the operating constraints on the system, while computing the optimal rescheduling amongst selected generators using LP technique. The thesis also presents a real power loss allocation approach based on the proposed power flow tracing. Loss allocation is an important part of tariff design as the cost associated with losses amounts to a sizable fraction of total revenue collected from the loads. The approach provides information as to how losses are distributed among loads and how much each generator is providing for the loss share of each load. The approaches developed in the thesis are illustrated on a sample 10-bus equivalent system, IEEE 30-bus, and IEEE 39-bus systems. Results for typical case studies are presented for practical systems of 72-bus equivalent and 203-bus equivalent of Indian Southern grid.

O núcleo dorsal da rafe (DR) envia densas projeções para o prosencéfalo e está implicado em várias funções complexas. Muitos estudos indicam que a atividade neuronal do DR é controlada por neurônios do córtex pré-frontal (PFC). Devido à escassez de informações detalhadas sobre as conexões entre o PFC e o DR, objetivamos mapear sistematicamente as projeções do PFC para o DR. O traçador neuronal toxina colérica subunidade b (CTb) foi injetado nas diferentes partes do DR. Com técnica imunoistoquímica contra a CTb, células CTb-ir foram encontradas em maior escala no córtex polar frontal e na parede medial do PFC. No PFC lateral, o córtex da insula agranular dorsal, apresentou maior densidade de neurônios CTb-ir. No PFC orbital, o córtex orbital lateral e dorso-lateral eferentam principalmente a parte central do DR. Injeções situadas na parte caudal do DR resultaram em marcação atenuada. Os resultados indicam que o PFC envia robustas projeções para o DR de forma diferenciada. Através dessas projeções o PFC pode exercer um controle \"top down\" sobre os neurônios do DR. / The dorsal raphe nucleus (DR) sends massive projections to the forebrain and is implicated in a variety of complex functions. Several studies indicate that neuronal activity in the DR is controlled by the prefrontal cortex (PFC). Since there is no detailed information about the projections from the PFC to the DR, the goal of the present study was to map these projections systematically. The neuronal tracer cholera toxin, subunit b (CTb) was injected at different levels of the DR. By immunohistochemical methods, CTb-ir cells were found highly enriched in the frontal polar cortex and in the medial wall of the PFC. In the lateral PFC, the dorsal lateral insular cortex presented the highest density of CTb-ir neurons. In the orbital PFC, the lateral and dorso-lateral orbital cortex project principally to the central segment of the DR. Injections in the caudal segment of the DR resulted in reduced CTb staining. Our results indicate that the PFC sends massive topographically organized projections to the DR, which may exert a \"top down\" control over neurons in the DR.

Between 1999 and 2009 the Swedish Armed Forces underwent an immense reformation. This thesis aims to contribute to the discussion on small state foreign policy behaviour within the wider field of international relations by explaining why Sweden decided to reform their national defence from an invasion-based defence to a mission-based defence. Using a neoclassical realist approach and a process tracing methodology, two hypotheses are tested to analyze a predicted causal relationship between systemic imperatives, the search for influence, and the decision to reform the Armed Forces. The first hypothesis is that when a state’s relative power rises it will initially seek, but ultimately attempt to gain, more influence abroad. The second hypothesis is that the search for influence is the driving force behind the reformation of the Armed Forces. The analysis corroborates both hypotheses, although the former with a relative amount of uncertainty, and argues that both system and intervening unit-level variables matter. The main conclusion is that the favourable situation in the international system has permitted an increased search for influence. This search for influence has in turn been the driving force behind the reformation. Since the first hypothesis has a relative amount of uncertainty connected to it, the certainty of the conclusions in this thesis can be discussed, and must therefore be subject to further empirical research.

When the Tsunami occurred on Boxing Day 2004 it caused a major stress upon the governmental agencies in not only the affected areas but also in states that had a large number of citizens in the affected area. The governmental agencies were not prepared for the demands from the citizenry in these countries to deal with rescue and crisis management in areas situated halfway around the globe. This thesis main purpose is to research what the causal mechanisms of Sweden’s and Finland’s crisis management were through the method of process tracing. Taking its theoretical stance in Allison’s &amp; Zelikow’s three models of decision making, the thesis compares the course of events of the first two critical days in Sweden and Finland to establish what mistakes were made during this critical period. The main findings of the essay show that none of the theories is on its own capable of explaining the chain of events, but a combination of the organisational model and the governmental politics model do prove the necessity of trained crisis personnel as decision-makers early on in crisis management as well as the need for an organisational structure with a linear and clear chain of command.

Les cellules souches assurent le développement des tissus, leur renouvellement ainsique leur réparation suite à des blessures. L’une des questions clés du domaine de labiologie des cellules souches est l’identification des différents types cellulaires qu’unecellule souche peut donner. Les cellules souches peuvent être multipotentes, c’est-à-direcapables de donner naissance à plusieurs types cellulaires différents, ou unipotentes,c’est-à-dire qu’elles ne peuvent alors se différencier qu’en un seul type cellulaire. Lesexpériences de traçage cellulaire sont réalisées quotidiennement en biologie dudéveloppement et en biologie des cellules souches afin d’évaluer le devenir des cellulessouches in vivo. Cependant, il n’existe à ce jour aucune méthode rigoureuse permettantd’interpréter les résultats d’expériences de traçage cellulaire de manière non ambigüe etde déterminer la multipotence ou l’unipotence des cellules souches avec grande précisionet de manière statistiquement fiable. Nous avons développé de nouvelles méthodes afind’évaluer avec une très grande précision le caractère unipotent ou multipotent descellules souches du sein et de la prostate. Ces nouvelles découvertes démontrent demanière non ambigüe que la prostate provient de cellules souches multipotentes, alorsque seules des cellules souches unipotentes contribuent au développement et auremodelage de la glande mammaire au stade adulte. D’autre part, nous montrons que cesont des cellules souches multipotentes qui sont responsables des phases précoces dudéveloppement embryonnaire de la glande mammaire, et que ces cellules deviennentunipotentes peu avant la naissance. Nous avons étudié les mécanismes régulant le passagede l’état multipotent à l’état unipotent et démontrons que le facteur de transcription p63joue un rôle crucial dans la restriction du potentiel de différenciation des cellules souchesmammaires embryonnaires. Enfin, nous montrons que les cellules souches mammairesadultes, normalement unipotentes, peuvent redevenir multipotentes en conditionsphysiopathologiques telles que l’ablation spécifique d’une lignée cellulaire mammaire ouau cours de l’initiation tumorale. Nous essayons donc de comprendre de manière généraleles mécanismes impliqués dans le passage de l’état unipotent à l’état multipotent descellules souches mammaires adultes, et d’élucider les similarités existant entre lesdifférentes conditions dans lesquelles des cellules souches mammaires multipotentessont observées. / Doctorat en Sciences biomédicales et pharmaceutiques (Médecine) / info:eu-repo/semantics/nonPublished

Skin cancers are very common in humans. The two most frequent epithelial skin cancers are the basal cell carcinoma (BCC) and the squamous cell carcinoma (SCC). For the vast majority of cancers, the cell at the origin of tumour initiation is still unknown and assumptions concerning their origin rely mainly on morphological and immunohistochemical studies. Recently, adult stem cells (SCs) have been suggested to be at the origin of tumour initiation based on their long term self-renewing capacities. According to these, two important questions arise; do epithelial skin cancers arise from mutations in a specific cell lineage of the epidermis? And are the stem cells more competent to initiate tumors than committed cells?<p>BCCs result from aberrant activation of HH signaling and several mouse models carrying mutations in HH signaling genes are capable to form tumors resembling to human BCCs. <p>To identify the cell lineage at the origin of BCC and to investigate the role of stem cells in tumor initiation, we followed a genetic approach where we conditionally expressed SmoM2 oncogene (a constitutively active Smoothened mutant) in distinct skin epidermal compartments including SCs. Targeting basal epidermis cells, showed that only SmoM2-clones in the inter follicular epidermis (IFE) and the infundibulum can progress into BCC, whereas SmoM2 expression in Bulge SCs or in matrix transit amplifying progenitor cells never leads to BCC formation. Progressively after SmoM2 expression, tumor-initiating cells lose their normal differentiation to adopt a hair placode-like shape and markers, demonstrating that biochemical and morphological tumour features can be misleading in extrapolating their cellular origin.<p><p>The molecular changes occurring in tumor initiating cells and the mechanisms regulating the early steps of cancer development are poorly characterized for the majority of tumors. To address these questions in BCC, we took advantage of our ability to isolate SmoM2 expressing cells at different stages of tumor initiation and progression. Transcriptional profiling of SmoM2-basal IFE cells isolated one week (normal histology) and 4 weeks (dysplastic lesion), suggests that adult IFE cells undergo a reprogramming into embryonic hair follicle (EHFP) like fate. In addition, we showed that Wnt/β-catenin signaling is essential for BCC initiating cell reprogramming into EHFP like fate and for tumor initiation in a cell autonomous manner. Finally, we show that EHFP reprogramming occurs also in human BCCs in addition to the presence of a similar canonical Wnt activation signature to the one revealed in the SmoM2-BCC mouse model.<p><p>SCC is the second most frequent skin cancers after BCC and mutations in p53 and Ras genes has been suggested to be potentially the primary events in this tumour. SCCs present signs of squamous differentiation, suggesting that SCCs may originate from the inter follicular epidermis (IFE). To identify the cell lineage at the origin of SCC and the role of the hair follicle SCs in tumor initiation, we use a genetic tools driving oncogenic KRas (KRasG12D) expression at physiological levels in different epidermal compartments. <p>Targeting KRasG12D expression in bulge SCs and their progeny or in IFE results in benign tumor development with no sign of malignant transformation. In contrast, KRasG12D expression in HF Transit amplifying (TA) matrix cells do not promotes any macroscopic tumors or microscopic defects in the epidermis. Interestingly, papillomas arising from the IFE express follicular markers such as CD34 and K17, indicating that the expression of HF markers by tumor cells does not necessarily reflect their cellular origin. Using a combination of deletion of both p53 alleles together with KRasG12D expression, we showed that bulge SCs and/or their progeny but not HF matrix TA cells, promote SCC formation, suggesting that additional genetic hits such as p53 are required to promote full-blown invasive skin SCC. <p><p>In summary, our work demonstrated the non-follicular origin of BCC resulting from Smo mutation, as well as the implication of the IFE progenitors in tumor initiation. We also revealed the progressive reprogramming of BCC initiating cells towards an EHFP-like fate and the key role of Wnt/β-catenin pathway in this process. In contrast, we showed the competence of several epidermal lineages to initiate benign tumors upon expression of KRasG12D oncogene at physiological levels. We also demonstrated that lineage -specific markers expression within tumor cells does not necessarily reflect their cellular origin. Finally, we demonstrated the requirement of additional hits, such as P53 loss, to promote malignant progression in the context of oncogenic Ras.<p> / Doctorat en Sciences biomédicales et pharmaceutiques / info:eu-repo/semantics/nonPublished

Les images de synthèse sont présentes à travers un grand nombre d'applications tel que les jeux vidéo, le cinéma, l'architecture, la publicité, l'art, la réalité virtuelle, la visualisation scientifique, l'ingénierie en éclairage, etc. En conséquence, la demande en photoréalisme et techniques de rendu rapide ne cesse d'augmenter. Le rendu réaliste d'une scène virtuelle nécessite l'estimation de son illumination globale grâce à une simulation du transport de lumière, un processus coûteux en temps de calcul dont la vitesse de convergence diminue généralement lorsque la complexité de la scène augmente. En particulier, une forte illumination indirecte combinée à de nombreuses occlusions constitue une caractéristique globale de la scène que les techniques existantes ont du mal à gérer. Cette thèse s'intéresse à ce problème à travers l'application de techniques d'analyse de formes pour le rendu 3D.Notre principal outil est un squelette curviligne du vide de la scène, représenté par un graphe contenant des informations sur la topologie et la géométrie de la scène. Ce squelette nous permet de proposer de nouvelles méthodes pour améliorer des techniques de rendu temps réel et non temps réel. Concernant le rendu temps réel, nous utilisons les informations géométriques du squelette afin d'approximer le rendu des ombres projetés par un grand nombre de points virtuels de lumière représentant l'illumination indirecte de la scène 3D.Pour ce qui est du rendu non temps réel, nos travaux se concentrent sur des algorithmes basés sur l'échantillonnage de chemins, constituant actuellement le principal paradigme en rendu physiquement plausible. Notre squelette mène au développement de nouvelles stratégies d'échantillonnage de chemins, guidés par des caractéristiques topologiques et géométriques. Nous adressons également ce problème à l'aide d'un second outil d'analyse de formes: la fonction d'ouverture du vide de la scène, décrivant l'épaisseur locale du vide en chacun de ses points. Nos contributions offrent une amélioration des méthodes existantes and indiquent clairement que l'analyse de formes offre de nombreuses opportunités pour le développement de nouvelles techniques de rendu 3D / Nowadays, computer generated images can be found everywhere, through a wide range of applications such as video games, cinema, architecture, publicity, artistic design, virtual reality, scientific visualization, lighting engineering, etc. Consequently, the need for visual realism and fast rendering is increasingly growing. Realistic rendering involves the estimation of global illumination through light transport simulation, a time consuming process for which the convergence rate generally decreases as the complexity of the input virtual 3D scene increases. In particular, occlusions and strong indirect illumination are global features of the scene that are difficult to handle efficiently with existing techniques. This thesis addresses this problem through the application of discrete shape analysis to rendering. Our main tool is a curvilinear skeleton of the empty space of the scene, a sparse graph containing important geometric and topological information about the structure of the scene. By taking advantage of this skeleton, we propose new methods to improve both real-time and off-line rendering methods. Concerning real-time rendering, we exploit geometric information carried by the skeleton for the approximation of shadows casted by a large set of virtual point lights representing the indirect illumination of the 3D scene. Regarding off-line rendering, our works focus on algorithms based on path sampling, that constitute the main paradigm of state-of-the-art methods addressing physically based rendering. Our skeleton leads to new efficient path sampling strategies guided by topological and geometric features. Addressing the same problem, we also propose a sampling strategy based on a second tool from discrete shape analysis: the opening function of the empty space of the scene, describing the local thickness of that space at each point. Our contributions demonstrate improvements over existing approaches and clearly indicate that discrete shape analysis offers many opportunities for the development of new rendering techniques