Personalized Model And Retrieval On Multimedia

As rapid development of computer networks and multimedia information, the user's search requirements are constantly changing, they are no longer satisfied with the text retrieval. Users concern much about the multimedia information retrieval. Most of the current models focus on multi-media personalized search terms. However, the semantic information has not been taken into account and thus existing techniques cannot satisfy the requirements of personalized multimedia information retrieval. Therefore, the design of multimedia semantic information, taking into account the needs of individual users of the retrieval methods, is of great significance.Based on a generalized multimedia database model (G3M), we propose a multimedia personalized model (G3M-UP) using ontology. As MPEG-7 is one of the most general multimedia description standards and G3M is a general multimedia database model based on MPEG-7, they are used for G3M-UP. In G3M-UP, users' personalized description is carried out with semantics of multimedia & ontology, the semantics of multimedia and users' personalized description. According to users' query history and personalized description, users' mixed feedback information is implemented to update users' personalized description.By comprehensive analysis of the advantages of storage for G3M, we conclude that our proposed G3M storage approach extends the SQL + XML hybrid storage solution. The approach utilizes the characteristics of domain knowledge, description of the preferences of individual users and user feedback information G3M-UP model.The searching schema is proposed with variety of multimedia searching and character of G3M-UP. Based on G3M-UP, building and update of personalized description is complemented. As personalized description is utilized into query searching schema, users' personalized searching requirement is satisfied.