Representation and visualization techniques of point-based surfaces

In this thesis the candidate deals with the problem of visualizing point-based models using the splatting techniques. The bulk of the problem is to visualize a continuous surface extracting information from unorganized points taken as samples of an unknown surface embedded in a 3D space. The initial unstructured set of points is at first enriched by local differential geometry properties computed considering samples from the neighborhood. Subsequentially in order to visualize the surface, the splatting technique is extended, taking into account the results of the pre-processing phase. The computation involved in the whole pipeline exploits the hardware acceleration of GPU cards.