Thesis supervisor: Gábor Valasek
Location of studies (in Hungarian): EÖTVÖS LORÁND TUDOMÁNYEGYETEM Abbreviation of location of studies: ELTE
Description of the research topic:
In the field of computer graphics and data visualization, we assume the presence of a continuous input signal most of the time. However, in practice, the input consists of discretely sampled data. This necessitates a reconstruction step, which approximates the original continuous signal from the discrete samples. If we assume that the original signal does not have a dedicated direction, it can be shown that using alternative regular lattices (BCC, FCC) for reconstruction has the mathematical potential to result in higher reconstruction quality even with fewer samples.
Comparison of different lattices or different reconstruction methods and filters on the same lattice can be achieved with the help of test functions. A chosen test function is discretized at various resolutions, then the quality of reconstructions is measured. In volume visualization, the de-facto standard test function is the Marschner-Lobb signal, which has been used in numerous publications for decades. The signal is well-suited for visual and qualitative comparisons.
One of the main goals of our research is to thoroughly examine the applicability of the Marschner-Lobb signal as a base for comparison of reconstruction methods, and to check whether the widely used discretization resolutions are theoretically adequate. Especially in cases when more than one regular lattice type is involved in the comparisons. Further goals are to seek the frontiers of known algorithms, and to find new novel methods
Required language skills: magyar Recommended language skills (in Hungarian): magyar, angol