rgfprg swamm06

Ontological Infrastructure for a Semantic Newspaper: 

Ontological Infrastructure for a Semantic Newspaper Roberto García1, Ferran Perdrix1,2, Rosa Gil1 1GRIHO – Human Computer Interaction Research Group Universitat de Lleida, Spain 2SEGRE Media Group, Spain

Contents: 

Contents Introduction Proposal Ontological framework Integration framework Conclusions Future Work

Contents: 

Contents Introduction Proposal Ontological framework Integration framework Conclusions Future Work

Introduction: 

Introduction Press and Media companies getting digital and Web Segre: newspaper, radio, television and web portal. Multiple kinds of media text, photo, video,… Heterogeneous sources agencies, journalists, partners, institutions,… Heterogeneity: difficult to integrate and manage.

Introduction: 

Introduction Related standards: International Press NewsCodes, subjects reference system, taxonomy NITF, news documents structure NewsML, model news as multimedia packages Multimedia MPEG-7, descriptive multimedia metadata TV-Anytime, multimedia taxonomies Common aspect: non formal semantics, XML-based

Introduction: 

Introduction Journalists News Agencies Legacy News+Media Receiver News+Photos Custom XML NITF, NewsCodes, NewsML,… Archivist User

Contents: 

Contents Introduction Proposal Ontological framework Integration framework Conclusions Future Work

Proposal: 

Proposal Semantic Metadata and Ontology facilitate management and integration. Related previous work: ELIN (Electronic Newspaper Initiative) NEPTUNO (Semantic Web Technologies for Digital Newspaper) NewMARS (Multimedia Advanced Redistribution Surveillance)

Proposal: 

Proposal Journalists News Agencies Legacy Receiver Semantic Repository User

Contents: 

Contents Introduction Proposal Ontological framework Integration framework Conclusions Future Work

Ontological Framework: 

Ontological Framework NewsML, NITF, NewsCodes, MPEG-7, TVAnytime XML  Semantic Web “XML Semantics Reuse Methodology”. ReDeFer implementation XSD2OWL: schema to ontology. XML2RDF: XML instance data to RDF instances. CS2OWL: classification scheme to ontology

Ontological Framework: 

Ontological Framework ReDeFer XSD2OWL Mappings:

Ontological Framework: 

Ontological Framework NewsCodes Subjects Ontology Subjects taxonomy NITF 3.3 Ontology Structure concepts (paragraph, subheadline,…) Metadata properties (copyright, authorship, issue date,…) NewsML 1.2 Ontology News multimedia structure (envelope, component, item,…) MPEG-7 Ontology Complete ontology (2372 classes and 975 properties) TVAnytime Ontologies Content and Format CSs

Ontological Framework: MPEG-7: 

Ontological Framework: MPEG-7 Validation, compare to other MPEG-7 Ontologies: Hunter02: not complete, RDF+DAML. Tsinaraki04: not complete, semantic part of MDS. Troncy03: not complete, from an ontology to MPEG-7.

Ontological Framework: MPEG-7: 

Ontological Framework: MPEG-7 Hunter02 MPEG-7 Ontology

Ontological Framework: MPEG-7: 

Ontological Framework: MPEG-7 MPEG-7 Ontology

Ontological Framework: MPEG-7: 

Ontological Framework: MPEG-7 Tsinaraki04 MPEG-7 Ontology <complexType name="AudioType"> <complexContent> <extension base= "mpeg7:MultimediaContentType"> <sequence> <element name="Audio" type="mpeg7:AudioSegmentType"/> </sequence> </extension> </complexContent> </complexType> Class (AudioType partial restriction(Audio cardinality(1)) MultimediaContentType) Class (AudioType partial restriction(Audio cardinality(1)) restriction(Audio allValuesFrom(AudioSegmentType))) MultimediaContentType)

Ontological Framework: Instances: 

Ontological Framework: Instances ReDeFer XML2RDF: XML tree  RDF graph. Deduce blank node types from XSD2OWL ontologies restrictions.

Ontological Framework: Instances: 

Ontological Framework: Instances XML2RDF example

Contents: 

Contents Introduction Proposal Ontological framework Integration framework Conclusions Future Work

Integration Framework: 

Integration Framework Load Ontological Framework

Integration Framework: 

Integration Framework NITF packaged in NewsML container IPTC’s NITF-to-NewsML Metadata Mapping Stylesheet <NewsML> <NewsItem> <NewsComponent> <DescriptiveMetadata> <SubjectCode> <Subject FormalName="04000000"/> </SubjectCode> </DescriptiveMetadata> <ContentItem> <DataContent> <nitf><body>…</body></nitf> </DataContent> </ContentItem> </NewsComponent> </NewsItem> </NewsML>

Integration Framework: 

Integration Framework NewsML multimedia items context and content-based MPEG-7 metadata XML2RDF: RDF for NewsML-NITF instances Bridge subjects to NewsCodes ontology RDF for MPEG-7 metadata

Integration Framework: 

Integration Framework

Integration Framework: 

Integration Framework

Contents: 

Contents Introduction Proposal Ontological framework Integration framework Conclusions Future Work

Conclusions: 

Conclusions

Conclusions: 

Conclusions Press and Media domain: heterogeneous and metadata intensive Semantic Web and Ontology facilitate management and integration Existing work NewsML, NITF, NewsCodes, MPEG-7, TVAnytime,…

Conclusions: 

Conclusions XSD2OWL: take profit from XML Schema hidden semantics We formalise them when building ontologies, but also implicitly when we make XML Schemas. XML2RDF: reuse existing XML metadata to add momentum to the Semantic Web

Contents: 

Contents Introduction Proposal Ontological framework Integration framework Conclusions Future Work

Future Work: 

Future Work Generate ontology for legacy system XML Map legacy ontology to NewsML-NITF ontologies Integrate automatic and assisted MPEG-7 metadata multimedia annotation Complete the integration framework

Future Work: 

Future Work User Interface: Rhizomik Media MPEG-7, TVAnytime, DC, Copyright Ontology… Rhizomer-based semantic portal Rhizomer

Thank you for your attention: 

Thank you for your attention More at: http://rhizomik.net …/redefer …/semanticnewspaper …/ontologies/mpeg7ontos Contact: roberto@rhizomik.net {fperdrix,rgil}@diei.udl.es