Acquisition of 3D model of real objects and scenes is indispensable and useful in many practical applications, such as digital archives, game and entertainment industries, engineering, advertisement. There are 2 main methods for 3D acquisition : laser-based reconstruction (active method) and image-based reconstruction from multiple images of the scene in different points of view (passive method). While laser-based reconstruction achieves high accuracy, it is complex, expensive and difficult to set up for large-scale outdoor reconstruction. Image-based, or multi-view stereo methods are more versatile, easier, faster and cheaper. By the time we begin this thesis, most multi-view methods could handle only low resolution images under controlled environment. This thesis targets multi-view stereo both both in large scale and high accuracy issues. We significantly improve some previous methods and combine them into a remarkably effective multi-view pipeline with GPU acceleration. From high-resolution images, we produce highly complete and accurate meshes that achieve best scores in many international recognized benchmarks. Aiming even larger scale, on one hand, we develop Divide and Conquer approaches in order to reconstruct many small parts of a big scene. On the other hand, to combine separate partial results, we create a new merging method, which can merge automatically and quickly hundreds of meshes. With all these components, we are successful to reconstruct highly accurate water-tight meshes for cities and historical monuments from large collections of high-resolution images (around 1600 images of 5 M Pixel images)

Two-dimensional images and three-dimensional volumes have become a staple ingredient of our artistic, cultural, and scientific appetite. Images capture and immortalize an instance such as natural scenes, through a photograph camera. Moreover, they can capture details inside biological subjects through the use of CT (computer tomography) scans, X-Rays, ultrasound, etc. Three-dimensional volumes of objects are also of high interest in medical imaging, engineering, and analyzing cultural heritage. They are produced using tomographic reconstruction, a technique that combine a large series of 2D scans captured from multiple views. Typically, penetrative radiation is used to obtain each 2D scan: X-Rays for CT scans, radio-frequency waves for MRI (magnetic resonance imaging), electron-positron annihilation for PET scans, etc. Unfortunately, their acquisition is influenced by noise caused by different factors. Noise in two-dimensional images could be caused by low-light illumination, electronic defects, low-dose of radiation, and a mispositioning tool or object. Noise in three-dimensional volumes also come from a variety of sources: the limited number of views, lack of captor sensitivity, high contrasts, the reconstruction algorithms, etc. The constraint that data acquisition be noiseless is unrealistic. It is desirable to reduce, or eliminate, noise at the earliest stage in the application. However, removing noise while preserving the sharp features of an image or volume object remains a challenging task. We propose a multi-scale method to smooth 2D images and 3D tomographic data while preserving features at a specified scale. Our algorithm is controlled using a single user parameter - the minimum scale of features to be preserved. Any variation that is smaller than the specified scale is treated as noise and smoothed, while discontinuities such as corners, edges and detail at a larger scale are preserved. We demonstrate that our smoothed data produces clean images and clean contour surfaces of volumes using standard surface-extraction algorithms. In addition to, we compare our results with results of previous approaches. Our method is inspired by anisotropic diffusion. We compute our diffusion tensors from the local continuous histograms of gradients around each pixel in image

Acquisition of 3D model of real objects and scenes is indispensable and useful in many practical applications, such as digital archives, game and entertainment industries, engineering, advertisement. There are 2 main methods for 3D acquisition : laser-based reconstruction (active method) and image-based reconstruction from multiple images of the scene in different points of view (passive method). While laser-based reconstruction achieves high accuracy, it is complex, expensive and difficult to set up for large-scale outdoor reconstruction. Image-based, or multi-view stereo methods are more versatile, easier, faster and cheaper. By the time we begin this thesis, most multi-view methods could handle only low resolution images under controlled environment. This thesis targets multi-view stereo both both in large scale and high accuracy issues. We significantly improve some previous methods and combine them into a remarkably effective multi-view pipeline with GPU acceleration. From high-resolution images, we produce highly complete and accurate meshes that achieve best scores in many international recognized benchmarks. Aiming even larger scale, on one hand, we develop Divide and Conquer approaches in order to reconstruct many small parts of a big scene. On the other hand, to combine separate partial results, we create a new merging method, which can merge automatically and quickly hundreds of meshes. With all these components, we are successful to reconstruct highly accurate water-tight meshes for cities and historical monuments from large collections of high-resolution images (around 1600 images of 5 M Pixel images)

La representación interactiva de escenarios naturales en tiempo real presenta problemas debido a la gran cantidad de información que debe ser procesada. Un bosque puede estar compuesto por decenas de miles de especies vegetales, las cuales a su vez están compuestas por decenas de miles de elementos (hojas, tallos y ramas). Una descripción precisa de una escena de este tipo implica el procesamiento de una cantidad de información intratable para su uso en aplicaciones en tiempo real. Durante los últimos 20 años se han publicado muchos trabajos de carácter científico para resolver este problema, aportando diferentes estrategias con particulares ventajas e inconvenientes. El objetivo principal de esta tesis es la presentación de un conjunto de técnicas para mejorar la eficiencia en la representación de escenas naturales en tiempo real. Para ello se han realizado aportaciones en diferentes ámbitos del área de los gráficos por computador. Por una parte, se ha propuesto un modelo de nivel de detalle multirresolución especialmente diseñado para la vegetación, basado en nivel de detalle de resolución variable, el cual permite la adaptación de la complejidad poligonal de especies vegetales en tiempo real. Esta propuesta se acompaña de una nueva aportación diseñada para el manejo de escenas masivamente pobladas, es decir, con miles de modelos multiresolución. Además se han propuesto nuevas técnicas para mejorar la visualización en el ámbito de la iluminación, tanto en cuanto a modelar los efectos de iluminación en árboles como a la representación eficiente de sombras visualmente realistas.

Shadows contribute significantly to the perceived realism of an image, and provide an important depth cue. Rendering high quality, antialiased shadows efficiently is a difficult problem. To antialias shadows, it is necessary to compute partial visibilities, but computing these visibilities using existing approaches is often too slow for interactive applications. Shadow maps are a widely used technique for real-time shadow rendering. One major drawback of shadow maps is aliasing, because the shadow map data cannot be filtered in the same way as colour textures. In this thesis, I present variance shadow maps (VSMs). Variance shadow maps use a linear representation of the depth distributions in the shadow map, which enables the use of standard linear texture filtering algorithms. Thus VSMs can address the problem of shadow aliasing using the same highly-tuned mechanisms that are available for colour images. Given the mean and variance of the depth distribution, Chebyshev's inequality provides an upper bound on the fraction of a shaded fragment that is occluded, and I show that this bound often provides a good approximation to the true partial occlusion. For more difficult cases, I show that warping the depth distribution can produce multiple bounds, some tighter than others. Based on this insight, I present layered variance shadow maps, a scalable generalization of variance shadow maps that partitions the depth distribution into multiple segments. This reduces or eliminates an artifact - "light bleeding" - that can appear when using the simpler version of variance shadow maps. Additionally, I demonstrate exponential variance shadow maps, which combine moments computed from two exponentially-warped depth distributions. Using this approach, high quality results are produced at a fraction of the storage cost of layered variance shadow maps. These algorithms are easy to implement on current graphics hardware and provide efficient, scalable solutions to the problem of shadow map aliasing.

Accelerated graphics cards, or Graphics Processing Units (GPUs), have become ubiquitous in recent years. On the right kinds of problems, GPUs greatly surpass CPUs in terms of raw performance. However, GPUs are currently used only for a narrow class of special-purpose applications; the raw processing power available in a typical desktop PC is unused most of the time. The goal of this work is to present an extension to JikesRVM that automatically executes suitable code on the GPU instead of the CPU. Both static and dynamic features are used to decide whether it is feasible and beneficial to off-load a piece of code on the GPU. Feasible code is discovered by an implementation of data dependence analysis. A cost model that balances the speedup available from the GPU against the cost of transferring input and output data between main memory and GPU memory has been deployed to determine if a feasible parallelization is indeed beneficial. The cost model is parameterized so that it can be applied to different hardware combinations. We also present ways to overcome several obstacles to parallelization inherent in the design of the Java bytecode language: unstructured control flow, the lack of multi-dimensional arrays, the precise exception semantics, and the proliferation of indirect references.

Image registration is widely used in different areas nowadays. Usually, the efficiency, accuracy, and robustness in the registration process are concerned in applications. This thesis studies these issues by presenting an efficient intensity-based mono-modality rigid 2D-3D image registration method and constructing a novel mathematical model for intensity-based multi-modality rigid image registration. For mono-modality image registration, an algorithm is developed using RapidMind Multi-core Development Platform (RapidMind) to exploit the highly parallel multi-core architecture of graphics processing units (GPUs). A parallel ray casting algorithm is used to generate the digitally reconstructed radiographs (DRRs) to efficiently reduce the complexity of DRR construction. The optimization problem in the registration process is solved by the Gauss-Newton method. To fully exploit the multi-core parallelism, almost the entire registration process is implemented in parallel by RapidMind on GPUs. The implementation of the major computation steps is discussed. Numerical results are presented to demonstrate the efficiency of the new method. For multi-modality image registration, a new model for computing mutual information functions is devised in order to remove the artifacts in the functions and in turn smooth the functions so that optimization methods can converge to the optimal solutions accurately and efficiently. With the motivation originating from the objective to harmonize the discrepancy between the image presentation and the mutual information definition in previous models, the new model computes the mutual information function using both the continuous image function representation and the mutual information definition for continuous random variables. Its implementation and complexity are discussed and compared with other models. The mutual information computed using the new model appears quite smooth compared with the functions computed by others. Numerical experiments demonstrate the accuracy and efficiency of optimization methods in the case that the new model is used. Furthermore, the robustness of the new model is also verified.

With the plateauing of processor frequencies and increase in energy consumption in computing, application developers are seeking new sources of performance acceleration. Heterogeneous platforms with multiple processor architectures offer one possible avenue to address these challenges. However, modern heterogeneous programming models tend to be either so low-level as to severely hinder programmer productivity, or so high-level as to limit optimization opportunities. The novel systems presented in this thesis strike a better balance between abstraction and transparency, enabling programmers to be productive and produce high-performance applications on heterogeneous platforms. This thesis starts by summarizing the strengths, weaknesses, and features of existing heterogeneous programming models. It then introduces and evaluates four novel heterogeneous programming models and runtime systems: JCUDA, CnC-CUDA, DyGR, and HadoopCL. We'll conclude by positioning the key contributions of each piece in this thesis relative to the state-of-the-art, and outline possible directions for future work.

This thesis treats topics within the area of object recognition. A real-time view matching method has been developed to compute the transformation between two different images of the same scene. This method uses a color based region detector called MSCR and affine transformations of these regions to create affine-invariant patches that are used as input to the SIFT algorithm. A parallel method to compute the SIFT descriptor has been created with relaxed constraints so that the descriptor size and the number of histogram bins can be adjusted. Additionally, a matching step to deduce correspondences and a parallel RANSAC method have been created to estimate the undergone transformation between these descriptors. To achieve real-time performance, the implementation has been targeted to use the parallel nature of the GPU with CUDA as the programming language. Focus has been put on the architecture of the GPU to find the best way to parallelize the different processing steps. CUDA has also been combined with OpenGL to be able to use the hardware accelerated anisotropic sampling method for affine transformations of regions. Parts of the implementation can also be used individually from either Matlab or by using the provided C++ library directly. The method was also evaluated in terms of accuracy and speed. It was shown that our algorithm has similar or better accuracy at finding correspondences than SIFT when the 3D geometry changes are large but we get a slightly worse result on images with flat surfaces.

In today’s world of high tech manufacturing and computer-aided design simulations of models is at theheart of the whole manufacturing process. Trying to represent and study the variables of real worldmodels using simulation computer programs can turn out to be a very expensive and time consumingtask. On the other hand advancements in modern multi-core CPUs and general purpose GPUs promiseremarkable computational power. Properly utilizing this computational power can provide reduced simulation time. To this end modernmodeling environments provide different optimization and parallelization options to take advantage ofthe available computational power. Some of these parallelization approaches are based onautomatically extracting parallelism with the help of a compiler. Another approach is to provide themodel programmers with the necessary language constructs to express any potential parallelism intheir models. This second approach is taken in this thesis work. The OpenModelica modeling and simulation environment for the Modelica language has beenextended with new language constructs for explicitly stating parallelism in algorithms. This slightlyextended algorithmic subset of Modelica is called ParModelica. The new extensions allow modelswritten in ParModelica to be translated to optimized OpenCL code which can take advantage of thecomputational power of available Multi-core CPUs and general purpose GPUs.