Ontology Based Multimedia Indexing

Ontology Based Multimedia Indexing

Mihaela Brut (Institut de Recherche en Informatique de Toulouse, France) and Florence Sedes (Institut de Recherche en Informatique de Toulouse, France)
Copyright: © 2010 |Pages: 19
DOI: 10.4018/978-1-61520-859-3.ch013

Abstract

The chapter goal is to provide responses to the following question: how the ontologies could be used in order to index and manage the multimedia collections? Alongside with reviewing the main standard formats, vocabularies and ontology categories developed especially for multimedia content description, the chapter emphasis the existing techniques for acquiring ontology-based indexing. Since a fully automatic such technique is not possible yet, the chapter also proposes a solution for indexing a multimedia collection by combining technologies from both semantic Web and multimedia indexation domains. The solution considers the management of multimedia metadata based on two correlated dictionaries: a metadata dictionary centralizes the multimedia metadata obtained through an automatic indexation process, while the visual concepts dictionary identifies the list of visual objects contained in multimedia documents and considered in the ontology-based annotation process. This approach facilitates as well the multimedia retrieval process.
Chapter Preview
Top

Vocabularies For Multimedia Annotation

In the recent years, multiple standards were developed for storing each type of multimedia content. Among these, the most popular are:

  • Video: MPEG-1, MPEG-2, MPEG-4, QuickTime, Sony DV, AVI, ASF, Real-Media, etc.

  • Audio: Raw PCM, WAV, MPEG-1, MP3, GSM, G.723, ADPCM, etc;

  • Image: JPEG, TIFF, BMP, GIF, etc.;

  • Multimedia presentations: SMIL, MHEG, SVG.

Alongside with the binary information that constitutes the content itself, some basically metadata (creation date, file size, title, author etc.) are also stored in these formats.

Complete Chapter List

Search this Book:
Reset