Path Similarity Skeleton Graph Matching for 3D Objects

Simone Schäfer, University of Koblenz-Landau

Date: 14/04/2011
Time: 14:15-15:45
Room: AE-A 101

Matching of salient features on an object's shape is a key component in various applications, such as image retrieval or object classification. Bai and Latecki proposed a promising skeleton based matching algorithm for 2D shapes based on the geodesic paths between skeleton end points. In this presentation, the proposed algorithm for matching 2D shapes will be introduced. Strengths and weaknesses of the algorithm are highlighted. Furthermore, extending the algorithm into the third dimension is examined. Ideas are presented that have been developed for improving the algorithm's performance in 3D images in general and pre- and postoperative blood vessel images in particular.