Point-Based Multi-Resolution Rendering

For efficient processing, usually multi-resolution algorithms and data structures are employed: A model is represented at different levels-of-detail in order to adapt the representation accuracy to processing demands. A recent approach is the usage of point-based multi-resolution data structures: A geometric model is approximated by a hierarchy of sample points of different resolution. As being generally applicable and easy to use, such representations have recently gained some interest.

In my talk, I will discuss different data structures for point-based multi-resolution representations. Employing these data structures, I will describe different rendering applications such as real-time forward mapping, rendering of animated scenes, multi-resolution raytracing, volume rendering and sound rendering. Lastly, some directions for future work are discussed. The talk should demonstrate that point-based multi-resolution representations can solve rendering problems in various application areas that often have been hard to handle using former techniques.