Authors:
Mustafa Sert
1
and
Buyurman Baykal
2
Affiliations:
1
Baskent University, Turkey
;
2
Middle East Technical University, Turkey
Keyword(s):
Audio databases, audio modeling, content-based retrieval,MPEG-7.
Related
Ontology
Subjects/Areas/Topics:
Image and Video Databases
;
Multimedia
;
Multimedia Indexing and Retrieval
;
Multimedia Signal Processing
;
Telecommunications
Abstract:
The modeling of multimedia databases for multimedia information systems is a complicated task. The designer has to model the structure and the dynamic behavior of multimedia objects, as well as the interactions between them. In this paper, we present a data model for audio database applications in the context of MPEG-7. The model is based on the object-oriented paradigm and as well as low-level and high-level signal features, which are standardized within the MPEG-7 framework, thus enabling interoperability of data resources. The model consists of two parts: a structural model, which provides a structural view of raw audio data, and an interpretation model, which allows semantic labels to be associated with audio data. We make use of an object-oriented approach to capture the audio events and objects in our model. Compared to similar models, particular attention is paid to integration issues of the model with commercial database management systems. Temporal relations between audio ob
jects and events are also considered in this study.
(More)