ADM: a dynamic model for general multimedia storage and content-based retrieval

Content-based retrieval (CBR) has been considered effective for searching complex multimedia data from the Internet. Since multimedia has plentiful connotations, the user-dependent content definitions always are arbitrary but insufficient for general applications. In this paper, a new attribute-dispersed model (ADM) is proposed for general multimedia storage and content-based retrieval. According to the proposed model, different attributes of the multimedia object are firstly classified as independent ones and dependent ones, the later will be stored and managed separately, and new attributes or concepts can be easily expanded dynamically in a self-description and reference list (SDRL) scheme. It is expected that the proposed model can be a rational and practical solution for the efficient and effective storage and retrieval of general multimedia information