To draw silhouettes, we have implemented several methods of extracting and displaying edge lines from polyhedral models, which we discuss in Section 5.1.1 and 5.1.2. These methods can be roughly broken down into two categories. The first assumes no prior knowledge or preprocessing of the model and heavily leverages commodity graphics hardware. The second set of methods use preprocessing of the model and are purely software algorithms. Both hardware and software algorithms clearly have a place. The set of hardware methods are useful because of ease of implementation. The software methods are advantageous due to their flexibility and lower computational complexity. All of the software methods assume that the models are either manifold or manifold with boundary.
We can also extract boundary curves, edges adjacent to only a single face, and creases, that is, the edge between two front facing polygons whose dihedral angle is above some threshold. The user is provided the option to draw these edges, dependent upon the model and intent of the image. The computation and drawing of creases is discussed in Section 5.1.3.