Towards efficient horizontal multimedia database fragmentation using semantic-based predicates implication

Abstract

Partitioning techniques are traditionally used in distributed system design to reduce accesses to irrelevant information by grouping data frequently accessed together in specific fragments. Here, we address the primary horizontal fragmentation of textually annotated multimedia data. In this study, we discuss the issue of identifying semantic implications between textual-based multimedia predicates, as a crucial phase in the efficient partitioning of multimedia data. Our proposal integrates knowledge bases as a framework for assessing the semantic relatedness between predicate values and operators. We developed a prototype implementing the various aspects of multimedia semantic predicates implications. Experimental results show that the proposed method is polynomial in the number of user predicates as well as the sizes of the knowledge bases being employed. Realworld multimedia fragmentation tests are ongoing.