Study On Query Analysis For Multimedia Query Language UMQL

With the prevalence of various multimedia data, multimedia information retrieval has become one of the most important problems in the data management domain and acquired a series of achievements. A powerful multimedia query language is a very useful facility to specify users'multimedia query requirements, and therefore is one of the most essential components in multimedia information retrieval. In recent years, there have been many multimedia query language proposals, a majority of which are either designed for one particular medium (e.g. images) or specialized for a particular application (e.g., a digital library), consequently not competent for uniform multimedia information retrieval.A unified multimedia query language called UMQL has been proposed recently by our group, which is able to effectively support the content, structure, or spatio-temporal inner information of multimedia data. Firstly, multimedia information concerned by users are separated into three categories: feature, content, and spatio-temporal information, after an in-depth study on users'multimedia query requirements, and then proposed ten evaluation rules for effective unified multimedia query language. Secondly, an in-depth investigation of existing multimedia data models are made, where specially aims to UMQL fundamental model, i.e., a semi-structured organization data model. Finally, based on the evaluation rules proposed for unified multimedia query language, analysis and evaluation specialized for the ability of UMQL on multimedia description, consequently as the foundation for proposing an effective grammar analysis model of UMQL.Based on the language features of UMQL, regular expression is employed to specify lexical restrictive rules, Backus-Naur form is employed to specify syntactic restrictive rules, and logical algebra expression is employed to specify semantic restrictive rules, and then based on these grammar restrictive rules, an effective grammar analysis model for UMQL is proposed. Secondly, a UMQL grammar analyzer is designed and implemented, mainly including: a lexical analyzer, a syntactic analyzer, and a semantic analyzer, where the lexical analyzer is employed to translate the query sentences as the form of character sequences into token sequences, the syntactic analyzer is employed to generate a syntactic tree based on the token sequences and syntactic generation expressions, and the semantic analyzer is employed to check each language variable whether satisfying the semantic rules and then generate several UMQL semantic groups, so the query sentences as the form of character sequences are translated in to a system inner specification: a syntactic tree and several UMQL semantic groups, by the grammar analyzer. Finally, particular tests for the grammar analyzer are made, including: integrated test, error test and performance test, and then the concerned test reports are presented.