Capturing the Essence of Shape of Polygonal Meshes
Author(s): Tobias Isenberg.
PhD Thesis: University of Magdeburg, Germany,
2004.
[BibTeX] 
Abstract:
Geometric models are the basis of computer graphics. Due to the growing computing
power and hardware support more and increasingly complex models are created. In
order to efficiently store, evaluate, manipulate, and match these, it is necessary to capture and extract the essence of shape. In particular for polygonal models, a concept is developed that allows adaptation of the notion of importance to the specific application without changing the extraction algorithm itself. For the use with this concept, a variety of criteria are described and conceived that capture what commonly is considered to be important in a polygonal mesh. In addition, an algorithm is developed within the concept that allows extraction of external skeletons as one aspect of essence of shape. Moreover, it is demonstrated that the concept also covers different notions of shape such as silhouettes. Since these have different application-specific requirements, a different algorithm is presented for fulfilling them. Finally, the methods that were presented are discussed with respect to potential application areas and a number of examples are shown.