ArchivIA Università degli Studi di Catania

ArchivIA - Archivio istituzionale dell'Universita' di Catania >
Tesi >
Tesi di dottorato >
Area 09 - Ingegneria industriale e dell'informazione >

Please use this identifier to cite or link to this item:

Issue Date: 3-May-2011
Authors: Pino, Carmelo
Title: A Semantic-Based and Adaptive Architecture for Automatic Multimedia Retrieval Composition
Abstract: In this thesis, we present a MMR (Multimedia Retrieval) system that is multi-domain and task independent. The use of MMR systems for different domains and tasks poses several limitations, mainly related to poor flexibility and adaptability to the different domains and user requirements. Therefore, it is desirable to have systems capable to automatically adapt to the domain they deal with. Another issue with MMR systems regards the relevance of the results that, as is known, depends strongly on the specific task. In order to address that, a retrieval system must be able to adapt automatically to a specific context and domain respecting the constraints imposed by the user and by the notion of relevance that one would like to apply. This, of course, implies that the system must understand what are the low-level features that fit a given purpose and which sequence of steps may produce satisfying results. In this thesis, a semantic-based system that uses ontologies for describing not only the application domain but also the algorithm's steps to be performed for respecting user's and domain requirements is proposed. More in detail, an ontological model, defined by the user, allows the system to adapt the retrieval mechanism to the application domain. Specifically, a user, by his/her requests, can choose from a set of predefined OWL ontology models and can add constraints/concepts to these models in order to formulate accurate and specific requests. According to the instances generated for each request, the system generates the interfaces (GUI) for the retrieval system specific for the domain (e.g. music, video, images) by a procedure self-guided by the defined ontology. Some examples of ontology-based GUI auto-composition and their related performance will be shown. Finally, the proposed architecture is scalable and flexible because of 1) a mechanism that allows an easy definition of new multimedia processing algorithms and 2) the use of semantic web-oriented technologies.
Appears in Collections:Area 09 - Ingegneria industriale e dell'informazione

Files in This Item:

File Description SizeFormatVisibility
Thesis_Carmelo_Pino.pdf2,53 MBAdobe PDFView/Open

Items in ArchivIA are protected by copyright, with all rights reserved, unless otherwise indicated.

Share this record




Stumble it!



  Browser supportati Firefox 3+, Internet Explorer 7+, Google Chrome, Safari

ICT Support, development & maintenance are provided by the AePIC team @ CILEA. Powered on DSpace Software.