Abstract
A set of image-space techniques for visualizing silhouettes of 3D models is presented. We assume that very little information is available about the data and rely only on simple counting arguments. This also allows us to automatically handle surface boundaries. While based on previous work, the presented technique is generally simpler and provides some extra convenience compared to the few existing methods which work under similar conditions. We also introduce a method for visualizing all silhouettes of a model in a single view which in many cases allows faster and better understanding of underlying geometric structure.
