Pavement texture has been measured predominantly by using two-dimensional (2D) profile methods. The 2D profile based mean profile depth (MPD) is still the well accepted texture index which has been found inadequate to characterize pavement texture especially when tire/pavement friction and noise are involved. There is a lack of standard 3D texture indices which show strong correlation with pavement friction and noise. There is a need to use 3D texture measurement for more comprehensive understanding of texture. The objectives of this thesis are to characterize pavement surfaces using 3D texture parameters based on 3D texture measurement and to explore the relationship between 3D texture parameters, pavement friction, and pavement noise. Field tests are conducted at various pavement sections for the measurements of texture, friction, and noise. The tested pavements include Interstate highway, MnROAD test facilities, airport runway, and municipal streets. The findings and contributions of this thesis are: • The pavement surface texture is measured in a 3D manner by using a line-laser scanner with both horizontal sample interval and vertical accuracy better than 0.05 mm. • A 3D texture analysis procedure with discrete wavelet transform (DWT) is proposed to separate macrotexture from microtexture and to define texture indices independently. • 3D parameters for macrotextures and microtexture are proposed and verified by field tests. • The relationship between 3D and 2D macrotexture indices [i.e. SMTD and MPD; Sq and root mean square roughness (RMSR)] are established, which is useful for the purposes of data comparison between 3D and 2D methods. • The relationship is investigated between 3D macrotexture parameters (SMTD and Sq) and pavement friction and noise. • It is found that texture distribution indices (i.e. Ssk and Sku) are significant contributors to pavement friction and noise. The new 3D texture analysis procedure and texture indices proposed in this thesis can be used to characterize various pavement textures (concrete pavement, asphalt pavement, and pavement contains recycled materials) in 3D manner, to compare 3D with 2D texture measurement/indices for quality control purposes, and to evaluate and predict pavement friction and noise. / February 2016

The use of virtual reality (VR) for visualization can revolutionize medical training by simulating real world medical training procedures through intuitive and engaging user interface. Existing virtual reality based visualization systems for human anatomy are based on 3D surface and volumetric models and simulative systems based on model libraries. The visual impact as well as facilitation for learning are inadequate in such systems. This thesis research is aimed at eliminating such inadequacies by developing a non-immersive virtual reality system framework for storage, access and navigation of real human cadaveric data. Based on this framework, a real time software system called virtual cadaver navigation system (VCNS) is developed, that can be used as an aid for teaching human anatomy. The hardware components of the system include, a mannequin, an examination probe similar to a medical ultrasound probe, and a personal computer. The examination probe is moved over the mannequin to obtain the virtual tomographic slice from the real cadaveric3-D volume data. A 3-D binary space partitioning tree structure is defined to organize the entire volumetric data, by subdividing it into small blocks of predefined size, called as bricks that are assigned a unique address for identification. As the examination probe is moved over the mannequin, the set of bricks intersecting the corresponding tomographic slice are determined by traversing the tree structure, and only, the selected bricks are accessed from the main memory and brought into the texture memory on the graphics accelerator card for visualization. The texture memory in the graphics card and the main memory are divided into slots of size, that is a multiple of the brick size, and a tagging scheme that relates the brick addresses, texture memory slots, and the main memory blocks is developed. Based on spatial, temporal and sequential locality of reference, only the currently required bricks as well as some of the neighboring bricks are loaded from the main memory into the texture memory, in order to maintain the highest frame rates required forreal time visualization. The above framework consisting of the data organization and the access mechanism are critical in terms of achieving the interactive frame rates required for real-time visualization. The input data to the system consists of non-segmented voxel data, and the data segmented and labelled based on tissue classification. The software system includes a labeling tool, in order to display the specific tissue information at the the location of the mouse cursor. This facility is useful in both teaching anatomy and self learning. Thus, the proposed VCNS system supports efficient navigation through the human body for learning anatomy and provides the knowledge of spatial locations and the interrelationship among the various organs of the body. A prototype software system has been developed, which is capable of achieving a throughput of 30 frames per second and has been tested with a 18-Gigabyte human cadaveric data obtained from the National Library of Medicine, on a personal computer with 64 Megabytes of texture memory and 512 Megabytes of main memory.

<p>A new approach for object replacement in 3D space is presented. Introducing a technique that replaces the older two dimensional (2D) based facial replacement method performed by compositing artist in motion picture productions and video commercial industry.</p><p>This method uses 4 digital video cameras filming an actor from 360 degrees, the cameras are placed with 90 degrees in between, the video footage acquired is then used to produce a 3D video texture consisting of video segments taken from different angles representing the object from 3D point of view.</p><p>The video texture is then applied to a 3D modelled head matching the geometry of the original object.</p><p>Offering the freedom of showing the object from any point of view from 3D space, which is not possible using the current two dimensional method where the actormust at all time face the camera.</p><p>The method is described in details with images showing every stage of the process.</p><p>Results are presented as still frames taken from the final video footage and as a video file demonstrating them.</p>

A new approach for object replacement in 3D space is presented. Introducing a technique that replaces the older two dimensional (2D) based facial replacement method performed by compositing artist in motion picture productions and video commercial industry. This method uses 4 digital video cameras filming an actor from 360 degrees, the cameras are placed with 90 degrees in between, the video footage acquired is then used to produce a 3D video texture consisting of video segments taken from different angles representing the object from 3D point of view. The video texture is then applied to a 3D modelled head matching the geometry of the original object. Offering the freedom of showing the object from any point of view from 3D space, which is not possible using the current two dimensional method where the actormust at all time face the camera. The method is described in details with images showing every stage of the process. Results are presented as still frames taken from the final video footage and as a video file demonstrating them.

Volume rendering is one of the branches of scientific visualization. Its popularity has grown in the recent years, and due to the increase in the computation speed of the graphics hardware of the desktop systems, became more and more accessible. Visualizing volumetric datasets using volume rendering technique requires a large amount of trilinear interpolation operations that are computationally expensive. This situation used to restrict volume rendering methods to be used only in high-end graphics workstations or with special-purpose hardware. In this thesis, an application tool has been developed using hardware accelerated volume rendering techniques on commercial graphics processing devices. This implementation has been developed with a 3D texture based approach using bump mapping for building an illumination model with OpenGL API. The aim of this work is to propose visualization methods and tools for rendering medical image datasets at interactive rates. The methods and tool are validated and compared with a commercially available software.

The simulation of natural phenomena such as clouds, smoke, fire and water is one of the most important research areas in computer graphics. Clouds are an essential component of any outdoor virtual environment, they add an important element of visual detail without which the environment would feel unrealistic. This paper describes an approach for setting up a cloud simulation. Clouds in our system are modeled using cellular automaton. To accelerate the simulation and its visualization, we implement both  entirely on programmable floating-point graphics hardware. The main part of the algorithm is implemented in a fragment shader and therefore takes full advantage of the highly parallel structure. The algorithm can generate result at real-time or near real-time frame rates. We also simulate the interaction of clouds with light, including self-shadowing.

In recent years, texture synthesis has been a heated topic in computer graphics, and the development of advanced algorithms for generating high-quality 3D textures is an area of active research. A recently proposed model, Neural Cellular Automata, can synthesize realistic 2D texture images or videos. However, due to the complexity and non-differentiable nature of 3D rendering and the lack of definition of the neighborhood on 3D mesh objects, no one has extended the 2D Neural Cellular Automata to the 3D scenario. In this master’s thesis, we propose a novel method for modeling the neighborhood relationship on 3D mesh objects, drawing inspiration from a graph variant of the Neural Cellular Automata. We also design an end-to-end 3D texture synthesis pipeline, leveraging a differentiable renderer to enable the Graph Neural Cellular Automata to learn to synthesize desired 3D textures. Our method allows users to either give the text description of the target textures or present the target texture images as the objectives. We evaluate the effectiveness of our proposed method both qualitatively and quantitatively, comparing it with the state-of-the-art method to demonstrate that it achieves comparable or better results. Furthermore, we explore the homology between the graph variant of Neural Cellular Automata and the 2D model, examining whether our proposed model preserves critical properties of the 2D model such as zero-shot generalization and self-regeneration. Finally, we analyze the limitations and potential drawbacks of our proposed method and suggest directions for future research. In summary, this thesis proposes a novel approach to synthesizing high-quality 3D textures using the Graph Neural Cellular Automata model and a differentiable renderer. Our work provides a foundation for future research in this area, and we believe that our findings will contribute to the development of advanced algorithms for 3D texture synthesis. / Under de senaste åren har textursyntes varit ett hett ämne inom datorgrafik, och utvecklingen av avancerade algoritmer för att generera högkvalitativa 3D-texturer är ett aktivt forskningsområde. En nyligen föreslagen modell, Neural Cellular Automata, kan syntetisera realistiska 2D-texturbilder eller videor. Dock, på grund av komplexiteten och den icke-differentierbara naturen av 3D-rendering och bristen på definition av grannskapet på 3D-meshobjekt, har ingen utvidgat 2D Neural Cellular Automata till 3D-scenariot. I den här masteruppsatsen föreslår vi en ny metod för att modellera grannskapsrelationen på 3D-meshobjekt, inspirerade av en grafvariant av Neural Cellular Automata. Vi utformar också en ände-till-ände 3D-textursyntes pipeline, genom att utnyttja en differentierbar renderer för att möjliggöra för Graph Neural Cellular Automata att lära sig syntetisera önskade 3D-texturer. Vår metod tillåter användare att antingen ge textbeskrivningen av måltexturerna eller presentera måltexturbilderna som målen. Vi utvärderar effektiviteten av vår föreslagna metod både kvalitativt och kvantitativt, jämför den med den mest avancerade metoden för att visa att den uppnår jämförbara eller bättre resultat. Dessutom utforskar vi homologin mellan grafvarianten av Neural Cellular Automata och 2D-modellen, undersöker om vår föreslagna modell bevarar kritiska egenskaper hos 2D-modellen som zero-shot generalisering och självregenerering. Slutligen analyserar vi begränsningarna och eventuella nackdelar med vår föreslagna metod och föreslår riktningar för framtida forskning. Sammanfattningsvis föreslår denna avhandling en ny metod för att syntetisera högkvalitativa 3D-texturer med hjälp av Graph Neural Cellular Automata-modellen och en differentierbar renderer. Vårt arbete ger en grund för framtida forskning inom detta område, och vi tror att våra fynd kommer att bidra till utvecklingen av avancerade algoritmer för 3D-textursyntes.

This thesis deals with simulation and rendering of fluid based volumetric effects, especially effect of fire and smoke. Computations are accelerated on graphics card using modern graphics API with motivation to achieve realistic visual results as well as physically correct calculations. Implemented volumetric effects are distributed as dynamic library which allows addition of these effects to existing applications.