database has to deal with complex data with complex
relationships that are distributed over multiple repositories with
different characteristics and image representations. We will build on
previous work in defining spatial description and query languages and
developing indexing structures. Data volumes make it important to integrate query processing
algorithms with hierarchical, multidimensional indexing structures and
study query optimization issues in this context.
Tools have been developed in recent years for data mining -- many for relational data and a few for spatial data. However, tools that can be used for mining image data are almost non-existent, though their importance has been widely recognized. Image data mining is a term whose definition is not always clear; at a minimum it combines aspects of computer vision, AI, and statistics. By the end of the project, we would like to make the case that something interesting, and even unexpected, has been discovered in a large image or spatial database.