Die Dissertation wurde am 26.04.2007 bei der Technischen Universität München eingereicht und durch die Fakultät für Informatik am 12.09.2007 angenommen. ii Copyright c ○ 2007 Bartosz von Rymon Lipiński. Alle Rechte vorbehalten.

www.elsevier.com/locate/cag We address the problem of an efficient image-space reconstruction of adaptively sampled scenes in the context of point-based and linebased graphics. The image-space reconstruction offers an advantageous time complexity compared to surface splatting techniques and, in fact, our improved GPU implementation performs significantly better than splatting implementations for large point-based models. We discuss the integration of elliptical Gaussian weights for enhanced image quality and generalize the image-space reconstruction to line segments. Furthermore, we present solutions for the efficient combination of points, lines, and polygons in a single image.

In this paper, we present a novel point-based digitally reconstructed radiography (PBDRR) method for large CT data sets. Three steps are mainly included in the proposed method, namely point sampling, point rendering and quantization. Finite samples are determined on the nature or user-specified transfer function by a Monte Carlo technique. These samples are projected onto an image plane and the intensity value of each pixel records the count of samples projected onto the corresponding pixel area. The transfer function is related with the X-ray photon energy and the CT value. From the view of Monte Carlo integration, sufficient samples make the estimated pixel intensities believable theoretically. Moreover, PBDRR has O(N 2) time complexity depending on the number of samples and is suitable for parallel processing and hardware acceleration. Some experimental results demonstrate the performance of the present method without redundant adjustments of the transfer function. 1

Abstract not found