Hyperspectral (HS) imaging, which consists of acquiring a same scene in several hundreds of contiguous spectral bands (a three dimensional data cube), has opened a new range of relevant applications, such as target detection [MS02], classification [C.-03] and spectral unmixing [BDPD+12]. However, while HS sensors provide abundant spectral information, their spatial resolution is generally more limited. Thus, fusing the HS image with other highly resolved images of the same scene, such as multispectral (MS) or panchromatic (PAN) images is an interesting problem. The problem of fusing a high spectral and low spatial resolution image with an auxiliary image of higher spatial but lower spectral resolution, also known as multi-resolution image fusion, has been explored for many years [AMV+11]. From an application point of view, this problem is also important as motivated by recent national programs, e.g., the Japanese next-generation space-borne hyperspectral image suite (HISUI), which fuses co-registered MS and HS images acquired over the same scene under the same conditions [YI13]. Bayesian fusion allows for an intuitive interpretation of the fusion process via the posterior distribution. Since the fusion problem is usually ill-posed, the Bayesian methodology offers a convenient way to regularize the problem by defining appropriate prior distribution for the scene of interest. The aim of this thesis is to study new multi-band image fusion algorithms to enhance the resolution of hyperspectral image. In the first chapter, a hierarchical Bayesian framework is proposed for multi-band image fusion by incorporating forward model, statistical assumptions and Gaussian prior for the target image to be restored. To derive Bayesian estimators associated with the resulting posterior distribution, two algorithms based on Monte Carlo sampling and optimization strategy have been developed. In the second chapter, a sparse regularization using dictionaries learned from the observed images is introduced as an alternative of the naive Gaussian prior proposed in Chapter 1. instead of Gaussian prior is introduced to regularize the ill-posed problem. Identifying the supports jointly with the dictionaries circumvented the difficulty inherent to sparse coding. To minimize the target function, an alternate optimization algorithm has been designed, which accelerates the fusion process magnificently comparing with the simulation-based method. In the third chapter, by exploiting intrinsic properties of the blurring and downsampling matrices, a much more efficient fusion method is proposed thanks to a closed-form solution for the Sylvester matrix equation associated with maximizing the likelihood. The proposed solution can be embedded into an alternating direction method of multipliers or a block coordinate descent method to incorporate different priors or hyper-priors for the fusion problem, allowing for Bayesian estimators. In the last chapter, a joint multi-band image fusion and unmixing scheme is proposed by combining the well admitted linear spectral mixture model and the forward model. The joint fusion and unmixing problem is solved in an alternating optimization framework, mainly consisting of solving a Sylvester equation and projecting onto a simplex resulting from the non-negativity and sum-to-one constraints. The simulation results conducted on synthetic and semi-synthetic images illustrate the advantages of the developed Bayesian estimators, both qualitatively and quantitatively.

It has been claimed that action representations can be evoked by the image of a handled object (Tucker & Ellis, 1998). Contrary to this view, it may instead be the location of the object’s handle in visual space that generates a spatial code that in turn interacts with selection of response location. For example, an object with its handle extending into right visual space may bias attention to the right, resulting in a faster right- versus left-sided response (Cho & Proctor, 2010). In the current experiments I present evidence that under certain task conditions, images of objects evoke their corresponding action representations. When subjects engaged in laterality judgments to images of hands presented after or in conjunction with an image of a handled object, motor representations associated with that object were evoked. Although the location of the handle was irrelevant to the task, subjects were faster at responding when the depicted handle location and hand of response were aligned (i.e., right-handed key press to a right-handled frying pan) rather than misaligned. The effect of alignment remained constant across the response time distribution. When subjects made a crossed-hand response, the alignment effect was driven by a correspondence between the location of the object’s handle and the response hand, not the response location. These results contrast with what was found when observers responded to directional arrow cues in place of pictures of hands. With arrow cues, the observed alignment effect appeared to be driven by spatial correspondence between the location of the object’s body and the location of the response button. Moreover, in this case the alignment effect decreased across the response time distribution, in keeping with other cases of spatial compatibility effects (Proctor, Miles, & Baroni, 2011). I conclude that attention to an image of a hand can induce observers to activate motor affordances associated with pictured objects. / Graduate

This dissertation investigates the feasibility of creating real-time three-dimensional images, using data obtained from satellites. The aim is to enhance satellite imaging applications, by utilizing the normal 3-D visual perceptions of humans. A study is made of the different methods developed to create the illusion of seeing a three-dimensional object from essentially two-dimensional images. 3-D display devices based on the principles of human stereoscopic vision do exist. Other 3-D display techniques include holograms and volumetric displays. Satellite images are used in a wide range of applications, from urban planning, to earth surveillance, and even weather prediction. In the past, satellite imaging was the express domain of experts, trained in the analysis and interpretation of satellite images. However, in recent years, the acquisition and analysis of satellite images have been greatly facilitated by the growing number of commercial satellites in our skies, as well as readily available software packages. Satellite images are available in many types of image formats, and can represent a large variety of information about an area. The model developed for this dissertation (the ACSI-3D model) proposes a method for the conversion of satellite images to suitable input for a stereoscopic 3-D display device. The model covers the process from initial image acquisition to the final display. It consists of four basic phases: Image Acquisition, Stereopsis, Sequencing and Synchronization, and Display. The “Stereoscopic Image Pair Creator” prototype was developed to test parts of this model. / Ehlers, E.M., Prof.

The study of adolescents’ idols has an over 100-year tradition. The meta-analysis of Teigen, Normann, Bjorkheim and Helland (2000) showed that idols, which are commonly understood as role models, changed over the last century which is attributed to changes in the social context. The present paper argues that Social Identity Theory (Tajfel & Turner, 1979, 1986) offers an appropriate theoretical framework to conceptualize social context by hypothesising a functional relationship between idols and identity management strategies moderated by the status position of the adolescent’s group s/he belongs to. The hypothesised functional relationship was tested in two studies with white and black adolescent South Africans. The results of the two studies supported our assumptions that the functional relationship between idols and identity management strategies is indeed moderated by status position. The results also indicate that Social Identity Theory seems to be an appropriate theoretical framework when social context is particularly conceptualised as social change.

Most digital cameras capture images through Color Filter Arrays (CFA), and reconstruct the full color image from the CFA image. Each CFA pixel only captures one primary color component at each pixel location; the other primary components will be estimated using information from neighboring pixels. During the demosaicking algorithm, the unknown color components will be estimated at each pixel location. Most of the demosaicking algorithms use the RGB Bayer CFA pattern with Red, Green and Blue filters. Some other CFAs contain four color filters. The additional filter is a panchromatic/white filter, and it usually receives the full light spectrum. In this research, we studied and compared different four channel CFAs with panchromatic/white filter, and compared them with three channel CFAs. An appropriate demosaicking algorithm has been developed for each CFA. The most well-known three-channel CFA is Bayer. The Fujifilm X-Trans pattern has been studied in this work as another three-channel CFA with a different structure. Three different four-channel CFAs have been discussed in this research: RGBW-Kodak, RGBW-Bayer and RGBW- $5 \times 5$. The structure and the number of filters for each color are different for these CFAs. Since the Least-Square Luma-Chroma Demultiplexing method is a state of the art demosaicking method for the Bayer CFA, we designed the Least-Square method for RGBW CFAs. The effect of noise on different CFA patterns will be discussed for four channel CFAs. The Kodak database has been used to evaluate our non-adaptive and adaptive demosaicking methods as well as the optimized algorithms with the least square method. The captured values of white (panchromatic/clear) filters in RGBW CFAs have been estimated using red, green and blue filter values. Sets of optimized coefficients have been proposed to estimate the white filter values accurately. The results have been validated using the actual white values of a hyperspectral image dataset. A new denoising-demosaicking method for RGBW-Bayer CFA has been presented in this research. The algorithm has been tested on the Kodak dataset using the estimated value of white filters and a hyperspectral image dataset using the actual value of white filters, and the results have been compared. The results in both cases have been compared with the previous works on RGB-Bayer CFA, and it shows that the proposed algorithm using RGBW-Bayer CFA is working better than RGB-Bayer CFA in presence of noise.

This thesis proposes a novel chroma keying method based on stereo images, which can be applied to post-process the alpha matte generated by any existing matting approach. Given a pair of stereo images, a new matting Laplacian matrix is constructed based on the affinities between matching pixels and their neighbors from two frames. Based on the new matting Laplacian matrix, a new cost function is also formulated to estimate alpha values of the reference image through the propagation between stereo images.

Dans le cadre de cette thèse, nous nous intéressons à la sélection des attributs en s'appuyant sur la théorie des graphes dans les différents contextes d'apprentissage non supervisé, semi-supervisé et supervisé. En particulier, nous nous intéressons aux scores de classement d'attributs basés sur des contraintes must-link et cannot-link. En effet, ces contraintes sont faciles à obtenir dans le cadre des applications réelles. Elles nécessitent juste de formuler pour deux données si elles se ressemblent et donc doivent être regroupées ensemble ou non, sans requérir d'informations détaillées sur les classes à retrouver. Les scores de contraintes ont montré de bonnes performances pour la sélection semi-supervisée des attributs. Cependant, ils sont fortement dépendants du sous-ensemble de contraintes disponibles. Nous proposons alors un score qui utilise à la fois l'ensemble des contraintes disponibles et les propriétés locales des données non contraintes. Des expériences réalisées sur des bases de données artificielles et réelles montrent que ce nouveau score est moins dépendant de l'ensemble de contraintes disponibles que les scores existants tout en atteignant des performances de classification similaires. La sélection semi-supervisée d'attributs a également été appliquée avec succès à la classification de textures couleur. En effet, parmi les nombreux attributs de texture pouvant être extraits des images couleur, il est nécessaire de sélectionner les plus pertinents afin d'améliorer la qualité de classification. / Within the framework of this thesis, we are interested in feature selection methods based on graph theory in different unsupervised, semi-supervised and supervised learning contexts. We are particularly interested in the feature ranking scores based on must-link et cannot-link constraints. Indeed, these constraints are easy to be obtained on real applications. They just require to formalize for two data samples if they are similar and then must be grouped together or not, without detailed information on the classes to be found. Constraint scores have shown good performances for semi-supervised feature selection. However, these scores strongly depend on the given must-link and cannot-link subsets built by the user. We propose then a new semi-supervised constraint scores that uses both pairwise constraints and local properties of the unconstrained data. Experiments on artificial and real databases show that this new score is less sensitive to the given constraints than the previous scores while providing similar performances. Semi supervised feature selection was also successfully applied to the color texture classification. Indeed, among many texture features which can be extracted from the color images, it is necessary to select the most relevant ones to improve the quality of classification.

À l’heure actuelle, l’Internet est devenu une des sources d’information les plus importantes dans des nombreux domaines, dont le domaine de la santé. Les images médicales portent des informations importantes pour le diagnostic, l’enseignement et la recherche, des informations parfois difficiles à décrire textuellement. Avec le développement des technologies d’acquisition, les images médicales sont de plus en plus nombreuses dans les bases d’images et dans les documents en ligne. Vu l’importance des images médicales pour le diagnostic, l’enseignement et même l’éducation civique, des méthodes et des outils efficaces d’annotation, d’indexation et de recherche des images médicales sont nécessaires. Dans ce contexte, nous proposons une architecture pour l’annotation des images médicales incluses dans des documents de santé en ligne. Notre système extrait des informations médicales spécifiques (i.e. modalité médicale, région anatomique) à partir de plusieurs sources, et combine ces informations pour annoter les images. Cette annotation est nécessaire pour pouvoir retrouver les images à l’intérieur des documents sur le web. Dans nos expérimentations, nous avons implémenté deux approches (chacune basée sur une source d’information) : une première orientée vers le contenu des images, et une deuxième orientée sur le contexte des images (régions textuelles associées aux images). / Today, Internet has become a major source of information in many areas, including health. Medical images are carrying crucial information for diagnostic, teaching and research, moste of the time this information being very difficult to describe using only text. In this context, we are proposing architecture for the annotation of medical images included in online health documents. Our system extracts specific medical information (modality, anatomical regions ...) from several sources and combines this information to annotate the images. This annotation is necessary to be able to search for images inside documents. We implemented two distinct approaches (each based on a different type of information) : one oriented towards the image content, and a second oriented on the image context (image-related textual regions).

The purpose of this thesis is to identify and discuss various popular images of the Mormon women of Utah between 1852 and 1890, the period during which the Latter-day Saints openly practiced plural marriage. The phrase "in the toils" refers to the basic image present in the minds of many Americans--that the women of the church were an oppressed, unhappy, enslaved group of individuals. This image, expressed in different ways, is found in many published writings of the period examined. After demonstrating the presence of this "in the toils" image, this study then attempts to analyze and evaluate its significance. Certainly this negative image had a bearing on anti-Mormon sentiment in general and opposition to the practice of polygamy. Moreover, the image seemed to generate more from nineteenth century values toward women and the family than it did from observable realities of the Mormon woman's condition. Several more realistic and historically valid images of the nineteenth century Mormon woman are suggested here. Another aspect of this subject is the manner in which the Mormon women themselves responded vii to their misrepresentation among non-Mormons. ' They identified themselves with an image of moving "onward for Zion," many of them, at least publicly, wholeheartedly endorsing their church and its teaching concerning plurality of wives. This paper has been based on numerous types of primary sources published between 1852 and 1890, including periodical articles, novels, reformers' tracts, travel accounts, newspapers, and public documents. The Mormon perspective has been studied through sermons, public testimonials of Mormon women, the Woman's Exponent, and a number of secondary sources.

No description available.