La géométrie numérique en temps réel est un domaîne de recherches émergent en informatique graphique.Pour pouvoir générer des images photo-réalistes de haute définition,beaucoup d'applications requièrent des méthodes souvent prohibitives financièrementet relativement lentes.Parmi ces applications, on peut citer la pré-visualisation d'architectures, la réalisation de films d'animation,la création de publicités ou d'effets spéciaux pour les films dits réalistes.Dans ces cas, il est souvent nécessaire d'utiliser conjointement beaucoup d'ordinateurs possédanteux-mêmes plusieurs unités graphiques (Graphics Processing Units - GPUs).Cependant, certaines applications dites temps-réel ne peuvent s'accomoder de telles techniques, car elles requièrentde pouvoir générer plus de 30 images par seconde pour offrir un confort d'utilisationet une intéraction avec des mondes virtuels 3D riches et réalistes.L'idée principale de cette thèse est d'utiliser la synthèse de géométrie,la géométrie numérique et l'analyse géométrique pourrépondre à des problèmes classiques en informatique graphique,telle que la génération de surfaces de subdivision, l'illumination globaleou encore l'anti-aliasing dans des contextes d'intéraction temps-réel.Nous présentons de nouveaux algorithmes adaptés aux architectures matérielles courantes pour atteindre ce but. / Eal-time geometry synthesis is an emerging topic in computer graphics.Today's interactive 3D applications have to face a variety of challengesto fulfill the consumer's request for more realism and high quality images.Often, visual effects and quality known from offline-rendered feature films or special effects in movie productions are the ultimate goal but hard to achieve in real time.This thesis offers real-time solutions by exploiting the Graphics Processing Unit (GPU)and efficient geometry processing.In particular, a variety of topics related to classical fields in computer graphics such assubdivision surfaces, global illumination and anti-aliasing are discussedand new approaches and techniques are presented.

Background. In computer graphics, it can be a time-consuming process to render photorealistic images. This rendering process, called “physically based rendering” uses complex algorithms to calculate the behavior of light. Fortunately, most renderers offer the possibility to alter the render-settings, allowing for a decrease in render time, but this usually comes at the cost of a lower quality image. Objectives. This study aims to identify what setting has the highest impact on the rendering process in the V-Ray renderer. It also examines if a perceived difference can be seen when reducing this setting. Methods. To achieve this, an experiment was done where 22 participants would indicate their preference for rendered images. The images were rendered in V-Ray with different settings, which affected their respective render time differently. Additionally, an objective image metric was used to analyze the images and try to form a correlation with the subjective results. Results. The results show that the anti-aliasing setting had the highest impact on render time as well as user preference. It was found that participants preferred images with at least 25% to 50% anti-aliasing depending on the scene. The objective results also coincided well enough with the subjective results that it could be used as a faster analytical tool to measure the quality of a computer-generated image. Prior knowledge of rendering was also taken into account but did not give conclusive results about user preferences. Conclusions. From the results it can be concluded that anti-aliasing is the most important setting for achieving good subjective image quality in V-Ray. Additionally, the use of an objective image assessment tool can drastically speed up the process for targeting a specific visual quality goal.

Voxel cone tracing (VCT) is a rendering method designed to approximate global illumination in a fast and efficient way. Global illumination means to render not only the direct lighting of a scene but also light from indirect sources, simulating how light in the real-world tend to bounce around and illuminate even the areas that are occluded from a direct light source. Rendering accurate global illumination in real-time has for a long time been a challenge in the field of computer graphics. This effect is most accurately simulated through expensive algorithms such as path tracing, where individual rays of light are traced from the pixels of a camera as they bounce around the scene and sample the environment. More efficient methods tend to rely on static image-based approaches, where global illumination is pre-rendered and baked into textures. VCT presents itself as a middle ground of the two, trading some of the accuracy for improved performance and the ability to work in a fully dynamic environment where objects and light sources may be moved around. VCT computes global illumination by volumetrically sampling a lower resolution voxel-based representation of the rendered scene by tracing cones.   The aim of this thesis is to further investigate the performance and image quality of VCT through an implementation and evaluation method. A VCT algorithm has been implemented that is capable of real-time global illumination in a dynamic setting, utilizing physically based rendering for improved image quality and a sparse 3D texture for efficient voxel storage. Performance was then measured in terms of rendering speed and memory usage. Image quality was evaluated through comparison with accurate path traced reference images. The results show that VCT is a promising rendering method for achieving a real-time approximation of global illumination, but that it also suffers from some issues regarding the image quality.

Fluorescence is a common effect in nature, it re-emits light by absorbing photons, caus- ing a wavelength shift from a shorter wavelength to a longer one. In recent years, there is an increased interest in including fluorescence in physically-based rendering. Fluorescence behavior is properly represented as a re-radiation matrix: for a given input wavelength, this matrix indicates how much energy is re-emitted at all other wavelengths. However, such a 2D representation has a significant memory footprint, especially when a scene con- tains a high number of fluorescent objects or fluorescent textures. This thesis proposes using Gaussian Mixture Domain to model re-radiation, which allows us to significantly reduce the memory footprint. Instead of storing the full matrix, we work with a set of Gaussian parameters that also allow direct importance sampling. When accuracy is a concern, one can still use the re-radiation matrix data and just benefit from impor- tance sampling provided by the Gaussian Mixture. Our method is useful when numerous fluorescent materials are present in a scene, particularly for textures with fluorescent components. 1

Ray tracing is widely used technique for realistic rendering of computer scenes. Its major drawback is time needed to compute the image, therefore it's usually parallelized. This thesis describes parallelization and ray tracing in general. It explains the possibility of how can be ray tracing parallelized as well as it defines the problems which may occur during the process. The result is parallel rendering application which uses selected ray tracing software and measurement of how successful this application is.

Goal of this diploma thesis is to get knowledge about history and evolution of computer graphic in area of realistic image synthesis, get knowledge about rendering system mental ray and about writing shader for it and write several shader. Create manual about writing  shaders for mental ray. Get knowledge about program Maya and create a tutorial hot to get the shader into it.

This work envolves basic options of human visual sense, describes matter of depth perception. We analyze binocular vision of humans, its limits, which must be considered, and followed when trying to project stereoscopic pictures. Then, we mention the most common methods of stereoscopic projection, reader is made acquainted with their options, advantages, and difficulties. The biggest part of this work is dedicated to the optimalizations themselves. They are focused on the most effective rendering of 3D scenes using stereoscopic technology. The effort is to make the best stereoscopic engine as possible and to inspect and improve displayed scene, so that the stereoeffect would be the best. Trivisio ARvision-3D HMD was used for testing purposes.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第18615号 / 情博第539号 / 新制||情||96(附属図書館) / 31515 / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 松山 隆司, 教授 美濃 導彦, 准教授 中澤 篤志, 講師 延原 章平 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第19186号 / 工博第4063号 / 新制||工||1627(附属図書館) / 32178 / 京都大学大学院工学研究科電気工学専攻 / (主査)教授 小山田 耕二, 教授 小林 哲生, 教授 中村 裕一 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

[pt] Um dos principais desafios em algoritmos de visualização volumétrica é calcular a integral volumétrica de maneira eficiente, mantendo uma precisão mínima adequada. Geralmente, métodos de integração numérica utilizam passos de tamanho constante, não incluindo nenhuma estratégia de controle numérico. Como uma possível solução, métodos numéricos adaptativos podem ser utilizados, pois conseguem adaptar o tamanho do passo de integração dada uma tolerância de erro pré-definida. Em CPU, os algoritmos adaptativos de integração numérica são, normalmente, implementados recursivamente. Já em GPU, é desejável eliminar implementações recursivas. O presente trabalho propõe um algoritmo adaptativo e iterativo para a avaliação da integral volumétrica em malhas regulares, apresentando soluções para manter o controle do passo da integral interna e externa. Os resultados do trabalho buscaram comparar a precisão e eficiência do método proposto com o modelo de integração com passo de tamanho constante, utilizando a soma de Riemann. Verificou-se que o algoritmo proposto gerou resultados precisos, com desempenho competitivo. As comparações foram feitas em CPU e GPU. / [en] One of the main challenges in volume rendering algorithms is how to compute the Volume Rendering Integral accurately, while maintaining good performance. Commonly, numerical methods use equidistant samples to approximate the integral and do not include any error estimation strategy to control accuracy. As a solution, adaptive numerical methods can be used, because they can adapt the step size of the integration according to an estimated numerical error. On CPU, adaptive integration algorithms are usually implemented recursively. On GPU, however, it is desirable to eliminate recursive algorithms. In this work, an adaptive and iterative integration strategy is presented to evaluate the volume rendering integral for regular volumes, maintaining the control of the step size for both internal and external integrals. A set of computational experiments were made comparing both accuracy and efficiency against the Riemann summation with uniform step size. The proposed algorithm generates accurate results, with competitive performance. The comparisons were made using both CPU and GPU implementations.