09/3 Representation and data structures for meshless approaches

		Coarse, point-based rendering of a fracturing material (left), and detailed point-based rendering of the same material along with simulation particles (right).

In recent years, points have proved to be a very efficient rendering primitive in computer graphics, especially for very complex models where traditional mesh representations would produce excessively large numbers of polygons. Another advantage of point-based representations is that they completely lack topological information, therefore they can be stored and processed more efficiently than triangle meshes. Today, there are already very efficient and highly accurate techniques for rendering point-based surfaces. We believe that fusing point-based rendering methods with multiscale particle simulation will open up a new paradigm in soft-tissue surgery simulation.

Our goal is to investigate representations and data structures for meshless approaches to represent biological tissues, while keeping in mind aspects related to the physical simulation. One of the major challenges in surgical simulations is to reconstruct the surface of organs or other tissue during cutting operations. Point-based meshless representations do not require maintaining complex consistency constraints, therefore they lead to efficient surface reconstruction using a resampling and relaxation technique. Further investigation is needed in aspects such as multiresolution representations, data structures for fast and dynamic neighborhood searches, or tracking topological changes during virtual tissue cutting operations.