Since the first multimedia database system, ORION was developed in 1987, the area and applications of Multimedia Databases have experienced tremendous growth. From the multimedia database viewpoint two main issues have to be considered:
1) Issues on Multimedia Database Modeling: A multimedia data model must deal with the issue of representing the multimedia objects, that is, designing the high and low-level abstractions of the media data to facilitate various These operations may include media objects selection, insertion, editing, indexing, browsing, querying, retrieval and communication. following issues: A good data model should fulfill the following issues:
I) it is necessary to model and store media components in the database, its storage mechanism will play an important factor affecting the performance of the multimedia systems.
ii) A representation should be provided for the logical media structure. It is essential to represent this structure explicitly both for querying and for representation.
iii) Semantic meaning should be modeled and linked to low-level characteristics and the structure of the media.
iv) The meta-data necessary for the operation of the system components needs to be determined and stored in the database.
v) it is necessary to rely on international standards, e.g., MPEG-7, in order to Guarantee interoperability for data sharing and data exchange.
2) Issues on Multimedia Indexing, Querying and Presentation: The key functionality in a multimedia database is how to access and how exchange multimedia information effectively. No matter what data model is used, data store mechanisms are employed, the most important thing is how to retrieve and communicate continuous and non-continuous media with a short real-time constraint. The key to retrieval process is similarity between two objects. The content of the object is analyzed and used to evaluate specified selection predicates, the process called content-based retrieval. In order to have an accurate representation of the multimedia objects in the database and the query object, different features have to be combined (e.g., texture, shape and color). The results are high-dimensional feature vectors (usually more than 1024 values). The efficiency of the similarity search in such spaces must be supported by the use of special multidimensional indexing structures and by dimension reduction methods, which however take care of the different nature of the feature components. Finally, similarity metrics have to re-visit to take the user perception into account.
In order to retrieve multimedia data from a database system, a query language must be provided. A multimedia query language must have abilities to handle complex spatial and temporal relationships. A powerful query language should have to deal with keywords, index on keywords and semantic contents of multimedia objects.
The key to efficient communication is to rely on standards for communicating meta-data and associated media data.
The issues considered above on data model, storage retrieval and query should not be limited to atomic-multimedia data like image, audio, or video.
