Spelling suggestions: "subject:"egmentation model"" "subject:"asegmentation model""
1 |
Power, competition and regulation : the case of the UK brewing sectorBobe, Jonathan Mark January 1999 (has links)
This thesis explores the role of unequal power relationships between business enterprises in the UK brewing sector and how these asymmetries shape the dynamic and direction of changei n patternso f geographicailn dustrialisation.P ower has,t o date,r emaineda largely neglected concept in economic relationships as considered in economic geography. A new model of geographical industrialisation is developed in this thesis that focuses on capital: capital relations, incorporates the dynamic nature of enterprises and the networks of relations within which they are embeddedt,h e asymmetryo f power relations within and betweene nterprisesa ndt he dynamicc hangesin markets tructured uring periodso f recession and restructuring. It further seeks to explore the relationship between stability and instability in the derivation of emerging patterns of geographical industrialisation. The model is based on the concept of circuits of power (Clegg, 1989) which has been successfully applied to economic geography over recent years (Taylor, 1995,1996; Taylor and Hallsworth, 1996,1999; Taylor eta!, 1995). In this model inequalities in power between enterprisese stablishesth e basesu pon which competition can take place and go on to create the context within which social relationships are established and can develop. However, as currently specified this approach neglects the collective agency of enterprises inherent in segmented economic sectors (Taylor and Thrift, 1982a, 1982b, 1983). By the incorporation of appropriate insights from the study of complexity, collective agency, the element of process within the circuits of power framework, can be more fully understood. In this way those processes that create instability and flux in enterprises, but which at the same time lead to periodic stabilisations, can be identified. The thesis is divided into four parts. Part I. makes explicit the limitations of current theories of geographical industrialisation (Chapters 1 and 2) and proposes a new model (Chapter 2), incorporating the concepts of circuits of power and complexity, that addresses these limitations. Part II of the thesis (Chapters 3,4 and 5) tests the model against historical trajectories of change in the UK brewing sector identifying six cycles of change since 1700. For each cycle, by applying the model, the processes that have instigated and promulgated change are made explicit. Distinct enterprise segmentations, associated with each period of relative stability during these cycles, are also identified. Part III of the thesis, through a questionnaire survey (Chapter 6) and a series of semistructured interviews (Chapter 7), uses the model to examine the state of the UK brewing sector at the present time. Chapter 6 identifies contemporary enterprise segments active within the sector and the differential action of pressures upon these segments. In doing so the path dependent trajectories ofchange ofenterprise segments, and the limitations imposed upon such trajectories, are made explicit. Chapter 7 considers, through the model, the day to day interactions of enterprise segments and how these interactions reinforce the negotiated inequalities inherent in asymmetrical power relations. Coping strategies adopted by enterprises during a period of instability are identified and the relationship between the market and interpersonal relationships are made explicit. It is concluded that the model proposed in this thesis provides for a more realistic interpretation of changing patterns of geographical industrialisation than previous models
|
2 |
Analyse multi-échelles et modélisation de la croissance foliaire chez Arabidopsis thaliana : mise au point et test d’un pipeline d’analyses permettant une analyse intégrée du développement de la cellule à la pousse entière / Multi-scale analysis and modeling of shoot growth in Arabidopsis thaliana – : development and testing of a pipeline of analysis methods enabling an integrative analysis of the development from cell to shoot scaleLievre, Maryline 15 December 2014 (has links)
Ce travail est basé sur le constat du manque de méthodes permettant l'analyse intégrée des processus contrôlant le développement végétatif d'Arabidopsis thaliana dans les études phénotypiques multi-échelles. Un phénotypage préliminaire de la croissance foliaire de 91 génotypes a permis de sélectionner 3 mutants et des variables d'intérêt pour une étude plus poussée du développement de la pousse. Un pipeline de méthodes d'analyses combinant techniques d'analyse d'images et modèles statistiques a été développé pour intégrer les mesures faites à l'échelle de la feuille et de la pousse. Des modèles multi-phasiques à changements de régime semi-markovien ont été estimés pour chaque génotype permettant une caractérisation plus pertinente des mutants. Ces modèles ont validé l'hypothèse selon laquelle le développement de la rosette peut être découpé en une suite de phases de développement, pouvant varier selon les génotypes. Ils ont aussi mis en évidence le rôle structurant de la variable «trichome abaxial», bien que les phases de développement ne puissent être entièrement expliquées par ce trait. Un 2nd pipeline d'analyses combinant une méthode semi-automatique de segmentation d'images de l'épiderme foliaire et l'analyse des surfaces de cellules par un modèle de mélange de lois gamma à paramètres liés par une loi d'échelle a été développé. Ce modèle nous a permis d'estimer la loi du nombre de cycles d'endoréduplication. Nous avons mis en évidence que cette loi dépendait du rang de la feuille.Le cadre d'analyses multi-échelles développé et testé durant cette thèse devrait être assez générique pour être appliqué à d'autres espèces végétales dans diverses conditions environnementales. / This study is based on the observation of a lack of methods enabling the integrated analysis of the processes controlling the vegetative development in Arabidopsis thaliana during multi-scale phenotypic studies. A preliminary leaf growth phenotyping of 91 genotypes enabled to select 3 mutants and different variables of interest for a more in depth analysis of the shoot development.We developed a pipeline of analysis methods combining image analysis techniques and statistical models to integrate the measurements made at the leaf and shoot scales. Semi-Markov switching models were built for each genotype, allowing a more thorough characterization of the studied mutants. These models validated the hypothesis that the rosette can be structured into successive developmental phases that could change depending on the genotype. They also highlighted the structuring role of the ‘abaxial trichomes' variable, although the developmental phases cannot be explained entirely by this trait. We developed a second pipeline of analysis methods combining a semi-automatic method for segmenting leaf epidermis images, and the analysis of the obtained cell areas using a gamma mixture model whose parameters of gamma components are tied by a scaling rule. This model allowed us to estimate the mean number of endocycles. We highlighted that this mean number of endocycles was function of the leaf rank.The multi-scale pipeline of analysis methods that we developed and tested during this PhD should be sufficiently generic to be applied to other plant species in various environmental conditions.
|
3 |
Rekonstrukce pozadí z několika fotografií / Background Reconstruction from Several PhotographsMotáček, Vladimír January 2010 (has links)
This thesis concerns the background reconstruction from several photographs (so called depopulation scene efect). There are presented methods for obtaining the background from video and discussion of their use for photographs. The greatest emphasis is placed on the Gaussian mixture model and effort to improve this algorithm due to static image. The photographs should be taken with a tripod.
|
4 |
Unraveling Complexity: Panoptic Segmentation in Cellular and Space ImageryEmanuele Plebani (18403245) 03 June 2024 (has links)
<p dir="ltr">Advancements in machine learning, especially deep learning, have facilitated the creation of models capable of performing tasks previously thought impossible. This progress has opened new possibilities across diverse fields such as medical imaging and remote sensing. However, the performance of these models relies heavily on the availability of extensive labeled datasets.<br>Collecting large amounts of labeled data poses a significant financial burden, particularly in specialized fields like medical imaging and remote sensing, where annotation requires expert knowledge. To address this challenge, various methods have been developed to mitigate the necessity for labeled data or leverage information contained in unlabeled data. These encompass include self-supervised learning, few-shot learning, and semi-supervised learning. This dissertation centers on the application of semi-supervised learning in segmentation tasks.<br><br>We focus on panoptic segmentation, a task that combines semantic segmentation (assigning a class to each pixel) and instance segmentation (grouping pixels into different object instances). We choose two segmentation tasks in different domains: nerve segmentation in microscopic imaging and hyperspectral segmentation in satellite images from Mars.<br>Our study reveals that, while direct application of methods developed for natural images may yield low performance, targeted modifications or the development of robust models can provide satisfactory results, thereby unlocking new applications like machine-assisted annotation of new data.<br><br>This dissertation begins with a challenging panoptic segmentation problem in microscopic imaging, systematically exploring model architectures to improve generalization. Subsequently, it investigates how semi-supervised learning may mitigate the need for annotated data. It then moves to hyperspectral imaging, introducing a Hierarchical Bayesian model (HBM) to robustly classify single pixels. Key contributions of include developing a state-of-the-art U-Net model for nerve segmentation, improving the model's ability to segment different cellular structures, evaluating semi-supervised learning methods in the same setting, and proposing HBM for hyperspectral segmentation. <br>The dissertation also provides a dataset of labeled CRISM pixels and mineral detections, and a software toolbox implementing the full HBM pipeline, to facilitate the development of new models.</p>
|
Page generated in 0.2836 seconds