多维数据索引结构回顾

Review of Multi-Dimensional Index Structures

一、引言最初,多维数据库主要用于计算机图形学、地理信息系统等.现在,多维数据库的应用扩展到医学图像处理、计算机视觉和多媒体数据库等领域.尤其是在多媒体数据库中,对多媒体对象的描述更加复杂,例如:对图像来说,常用颜色直方图、主色调、Tamura纹理等特征描述图像;对文本来说,常用矢量空间模型来描述文档;对视频来说,常用颜色、纹理、形状和运动等特征来描述视频镜头.但是,不管采用哪一种描述方法,都需要用多维空间中的点、线段或区域等来表示这些多媒体对象.

Similar to conventional relation databases, fast search and query operations in multi-dimensional databases require efficient index structures. The traditional index structures can't be applied to the multi-dimensional databases because of the special characteristics. Many multi-dimensional index structures have been proposed. In this paper, we classify these index structures by their intrinsic features , and introduce some representative index structures in detail.