3D Shape Matching Using Skeleton Graphs

We describe a method to match 3D polygonal geometry models using their internal skeleton graphs. These graphs carry information about the overall shape of the model. In order to make assumptions about the similarity of the 3D models, we search for the biggest common subgraph. As internal skeleton graphs normally do not consist of many edges, the problem of exponential time complexity due to NP completeness does not pose a problem. To measure similarity, we calculate a scalar match quality value based mainly on the edge ratios. In addition, we present an empirical study in order to evaluate the approach by trying to match a number of example models. It demonstrates that the skeleton based matching algorithm allows us to effectively compare two meshes based on their overall shape.

• Type: Conference
• Year: 2004
• Where: SIMVIS
• Authors: Angela Brennecke, Tobias Isenberg
• Keywords: Graphs Carry Information, Internal Skeleton Graphs, Overall Shape, SIMVIS 2004, SIMVIS 2007