Proceedings of the Korea Database Society Conference (한국데이타베이스학회:학술대회논문집)
- 1997.10a
- /
- Pages.472-482
- /
- 1997
A Study of Efficient Access Method based upon the Spatial Locality of Multi-Dimensional Data
- Yoon, Seong-young (Multimedia/Graphics Lab. Yonsei University, Computer Science Dept.) ;
- Joo, In-hak (Multimedia/Graphics Lab. Yonsei University, Computer Science Dept) ;
- Choy, Yoon-chul (Multimedia/Graphics Lab. Yonsei University, Computer Science Dept.)
- Published : 1997.10.01
Abstract
Multi-dimensional data play a crucial role in various fields, as like computer graphics, geographical information system, and multimedia applications. Indexing method fur multi-dimensional data Is a very Important factor in overall system performance. What is proposed in this paper is a new dynamic access method for spatial objects called HL-CIF(Hierarchically Layered Caltech Intermediate Form) tree which requires small amount of storage space and facilitates efficient query processing. HL-CIF tree is a combination of hierarchical management of spatial objects and CIF tree in which spatial objects and sub-regions are associated with representative points. HL-CIF tree adopts "centroid" of spatial objects as the representative point. By reflecting objects′sizes and positions in its structure, HL-CIF tree guarantees the high spatial locality of objects grouped in a sub-region rendering query processing more efficient.