linkedtv / ontologies

Repository for LinkedTV ontologies
3 stars 1 forks source link

Add a new type for Chapters #8

Open lyndonnixon opened 10 years ago

lyndonnixon commented 10 years ago

WP1/Exmaralda should tag extracted video chapters as news items (RBB) or art objects (TKK). cf. https://github.com/linkedtv/wp1/issues/4#issuecomment-50581638

This tag needs to be propagated in the LinkedTV RDF either through a subclassing of LinkedTV Chapters or an additional property-value on the Chapter instance.

lyndonnixon commented 10 years ago

I'd propose creating subclasses of linkedtv:Chapter, maybe in a separate namespace simply to keep this out of the "domain generic" LinkedTV Ontology.This makes it clear other scenarios could define their own specific classes and properties in other namespaces, but using the www.linkedtv.eu domain still to indicate that these (at least) are part of the "official" LinkedTV ecosystem.

e.g.

linkedtvrbb: http://www.linkedtv.eu/ontology/rbb# linkedtvavro: http://www.linkedtv.eu/ontology/avro#

linkedtvrbb:NewsItemChapter rdf:subClassOf linkedtv:Chapter linkedtvavro:ArtObjectChapter rdf:subClassOf linkedtv:Chapter

@rtroncy what do you think

rtroncy commented 10 years ago

I'm, in principle, against redefining the wheel. A (seed) program has normally already a set of global metadata, and we have worked on some genre taxonomies for programs. Hence, at the program level, one should already know whether this is a news program, a documentary, a soap, etc.

At the chapter level, EBU has translated its taxonomy in OWL/RDF, largely re-using the Ontology for Media Resources and the Media Fragments URI work. Hence:

lyndonnixon commented 10 years ago

OK great I didnt think to look for news items in other ontologies. I'm fine with those changes. So what's the process for agreed changes to the core ontology and publishing a new version at the namespace? A changelog will be useful too. Then e.g. we can immediately announce to the relevant partners - in this case @jblom - the new properties/classes so that they can use them in their tools.

e-apostolidis commented 10 years ago

I just informed Jose-Luis about the new format of the EXMARaLDA files that are created by the updated WP1 REST Service. This format differs from the previous one to the part related to the detected chapters of the video (given that chapter segmentation was among the selected analysis components). The new format contains information about the type of the detected chapters (annotated as "Art Object" or "News Item") and the type of the extracted keyframes for the videos of the documentary scenario (annotated as "Bumper" or "NoBumper").

So, the TV2RDF mechanism has to be updated accordingly, in order to parse the new EXMARaLDA files and retrieve this additional information.