On the Evaluation of Graph Centrality for Shape Matching

Graph centrality has been extensively applied in Social Net-work Analysis to model the interaction of actors and the information owinside a graph. In this paper, we investigate the usage of graph centrali-ties in the Shape Matching task. We create a graph-based representationof a shape and describe this graph by using di erent centrality measures.We build a Naive Bayes classi er whose input feature vector consists ofthe measurements obtained by the centralities and evaluate the di erentperformances for each centrality