A Method of Measuring Shape Similarity between multi-scale objects

Similarity measure is a key issue in evaluation of map generalization, object matching and object recognition. The measures of similarity include shape similarity, location similarity and semantic content similarity (Frank & Ester, 2006). Among these similarity measures, the shape similarity measure is very important because of the easy collecting of the necessary parameters and the well matching of human intuition (Zhang et al., 2002). As a matter of fact, most existing shape similarity measures (e.g. Arkin et al., 1991; Basri et al., 1998; Veltkamp & Latecki, 2006) are developed based on the concept of dissimilarity by measuring the distance or cost between objects. In practice, it is not as easy as a real similarity measure to distinguish how similar two objects are.