Processing of OLAP queries extended with image similarity predicates and spatial predicates over non-conventional data warehouses

Abstract

There is a new set of OLAP (on-line analytical processing) queries that can not currently be answered by conventional data warehousing (DWing) applications, since they require the management of both multimedia data, such as images, audio and videos, and geographic data, such as points, lines and polygons. For instance, suppose that a given medical team is interested in determining how many images are similar to a given breast cancer image to investigate the evolution of this disease in hospitals located up to 100 km away from São Carlos. This analysis can not be carried out by conventional DWing applications because it uses as a basis medical images of diseases, requires similarity comparisons over images, and encompasses the processing of a spatial predicate. This research project focuses on the processing of OLAP queries extended with image similarity predicates and spatial predicates over non-conventional data warehouses, specially image data warehouses and geographic data warehouses. Therefore, the project aims to investigate a gap in the literature of data warehouse related to the proposal of strategies to efficiently manipulate image and spatial data in the processing of OLAP queries. (AU)