![]() | Only 14 pages are availabe for public view |
Abstract > Semantic content-based qlle/)’ing and retrieval of multimedia data are the core filllctions of many ” modern applications sllch as content-based video-on-demand for home-lise, video databases for i • J cinematic production and the evolving Internet 2 applications. ’~In this thesis, existing models of multimedia semantic content-based querying are studied classified and criticized. It is showlI that they suffer from domain dependellcy, media type dependency alldlor lack of methodology. Also, a standard unified model does not exist. A new enhanced model, which is domain independent and less complex. was proposed. I Furthermore, a standard methodology was followed ill the design of this model. A general .; conceptual-level model, that covers the concepts mentioned in all the reported models, was firstly ’. proposed using the standard Entity Relationship model. Then, the model was mapped into a relational design. The relational model was used to make use of existing sophisticated RDBMSs. Also, SQL was IIsed as the que/)’ language for its popularity and simplicity. The model was .? validated and the relational design was verified . . •An architecture jar the system, which includes different system-components and modules, was •. proposed. In this architecture, the content-data was stored in a separate relational database to :J support hybrid media sources. A prototype was implemented to demonstrate the model. Some case studies of different que!)’ types supported and not supported by previous models were applied and their results \vere disClissed. |