Fuzzy Decomposition of Spatially Extended Objects

Modern database applications including computer-aided design, multimedia information systems, medical imaging, molecular biology, or geographical information systems impose new requirements on the effective and efficient management of spatial data. Particular problems arise from the need of high resolutions for large spatial objects. In this short paper, we sketch a new decompositioning approach based on clustering. We propose to describe a voxelized spatial object by a set of Gaussian distribution functions. Based on this decompositioning technique, we propose intersection queries which do not simply return a boolean value for each database object, but assign to each object a probability value indicating how likely an intersection is. The benefit of this approach compared to traditional approaches is that we do not any longer need an expensive refinement step for detecting whether objects intersect exactly on the fine-grained voxel sets.

• Type: Conference
• Year: 2005
• Where: SSDBM
• Authors: Hans-Peter Kriegel, Martin Pfeifle
• Keywords: Database, Large Spatial Objects, Spatial Object, SSDBM 2005, Voxelized Spatial Object