2D Topological and Direction Relations in the World of Minimum Bounding Circles
IDEAS '99 Proceedings of the 1999 International Symposium on Database Engineering & Applications
| Hi-index | 0.00 |
For efficient retrieval of 2D objects by shape, we propose three index structures on features that are extracted from the objects' minimum bounding circles (MBC). Besides traditional applications (e.g., CAD/CAM and Trademark registry), new multimedia applications such as structured video, animation, and MPEG-4 standard require the storage and management of well-defined objects. A sample query type in these applications is to find objects that are similar to or exactly match a given object. Due to large size of these multimedia databases and the complexity of the 2D object comparison algorithms, appropriate indexing techniques in order to invoke the comparison algorithms for only a subset of the stored objects is crucial.In this paper, by using MBC of an object, we identify six common features for that object such as the center coordinate and the radius of its MBC. The other four features are based on the vertices of the object that are on its MBC, termed touch points. A major observation is that those features are unique per object and can be utilized to filter out non-similar candidates. By utilizing these features in three alternative index structures, we guarantee no false drops for our similarity searches while reducing the number of false hits. Finally, we propose variations to known comparison algorithms, again based on our identified features, in order to eliminate false hits. To evaluate our techniques, we conducted a simulation study on a database of 2D objects. The results show the superiority of our techniques as compared to a naive indexing which was based on the number of vertices of the objects (at least 40% improvement in I/O cost). We also identify one of the indexing structures as the superior one, independent of the size of the database and the number of vertices of the objects.