A multimedia presentation is a synchronized, sequential or concurrent, and possibly interactive, delivery of streams of multimedia data such as video, audio, text and still images using audio-video output devices. Presently, multimedia presentations, created by multimedia authoring tools, are used in many applications such as computer-aided training, computer-aided learning, online books, etc..
To deliver a multimedia presentation effectively, multimedia presentation systems need to have powerful presentation languages for authoring a semantically coherent presentation, database support for storing and querying presentations, and efficient evaluation of queries.
In this thesis, we propose a spatio-temporal object-oriented data model for multimedia data. The data model can deal with not only the natural hierarchical organization of multimedia data, but also contents associated with the data. We then define a database query language for the retrieval of segments and the construction of presentation streams. For this purpose, we extend the Object Query Language (OQL) and name the resulting language Presentation Stream Construction Language (PSCL).
Presentations can be modeled as directed acyclic graphs (DAG), which represent the playout order of streams in a presentation naturally.
Using presentation graphs as the basic structures, we define query languages, namely GCalculus, GCalculus/S and GVISUAL, for querying presentation graphs and their content information. GCalculus and GCalculus/S are 2nd order calculus-based languages, and GVISUAL is a user-friendly graphical language. GCalculus/S does not have the universal quantifier of predicate calculus and the universal path quantifier (which GCalculus has) and provides a formal basis for GVISUAL.
We provide "bottom-up" evaluations of GCalculus/S queries, using an algebraic language, called O-Algebra. We extend O-Algebra with three new operators X$\sp{Next},\ {\rm X}\sp{Connect}$ and X$\sp{Until}$ which correspond to temporal logic operators, Next, Connected and Until of GCalculus/S. To evaluate the resulting O-Algebra expressions, we define the node code system which assigns unique codes to each node of a presentation graph, and evaluates the newly defined operators X$\sp{Next},\ {\rm X}\sp{Connect}$ and X$\sp{Until}$ efficiently by eliminating graph traversals of presentation graphs.
We also provide efficient evaluations of GVISUAL queries by transforming GVISUAL to O-Algebra expressions directly without the intermediate stage of GCalculus/S, i.e., the two transformations, one from GVISUAL to GCalculus/S, and another from GCalculus/S to O-algebra, are replaced by a direct transformation from GVISUAL to O-Algebra.
To summarize, in this thesis, we integrate presentations into database systems. To achieve the integration, we provide a data model for multimedia data, query languages for the retrieval of presentations based on content and query processing techniques.
Recommendations
Query Processing Techniques for Multimedia Presentations
A multimedia presentation is a synchronized, and possibly interactive, delivery of multimedia data to users. We expect that, in the future, multimedia presentations will be stored into and queried from multimedia databases. In an earlier work, we have ...
A Multimedia Query Specification Language
IW-MMDBMS '95: Proceedings of the International Workshop on Multi-Media Database Management SystemsAbstract: Typically, multimedia applications involve thousands of hours of video, images, audio, text and graphics that need to be stored, retrieved and manipulated in a large multimedia database. There is therefore an important need for novel ...