Optimizing content top N quiries

The AMIS project (started in summer 1997 and finished in 2002) was a research initiative sponsored by the Netherlands Organization for Scientific Research (NWO) organization. The project aimed to broaden and deepen the understanding of methods for indexing and searching in multimedia databases. The methods had to be both effective and efficient. In this project, several disciplines of computer science collaborated to achieve this one research goal of advancing insight into critical bottlenecks of multimedia system technology.

Five research groups throughout the Netherlands participated in the project, each accounting for a separate subproject. The Database group of the University of Twente was responsible for subproject 2. This subproject aimed to support query formulation and processing in a multimedia setting.

For several reasons techniques used in traditional database systems are inadequate to deal with multimedia data. This also holds for the natural major query class of top N content queries. Our research aimed at building an optimizer achitecture for our logical extended NF2 algebra, called Moa. In this architecture we wanted to implement top N optimizer facilities for content based retrieval. We were mainly intested in text and images. This top N research built on top of the architecture developed in the miRRor project.

More information can be found on the AMIS homepage.