This paper may not exactly answer your question, but there should be another one from the same project that gets you close:

http://www.cs.vu.nl/~guus/papers/Akker11a.pdf

I've been looking through several ontologies in the cultural heritage domain, and they should suffice. In particular, I've found the painting ontology of MOLTO project, which extends the model CIDOC-CRM; ICONCLASS, a classification system that has already an implmentetation in SKOS, and a direct implementation of the CIDOC-CRM model in OWL DL 1.0 called Erlangen-CRM.

Anyway, I've cross-read the paper you've sent me and I think that, at first, it is closely related to the subject that we want to deal with. Thank you very much for the reference.

The Europeana Data Model (EDM) might work for you too. We are working to publish Smithsonian data and would like to connect with people working on this domain. Check us out at isi.edu/integration/karma, and feel free to contact me at [email protected].

Pedro

Do you see this link?

http://www.ou.edu/ouphil/faculty/irvin/ArtworksObjectsStructures.pdf

http://onlinelibrary.wiley.com/doi/10.1111/j.1540-6245.2011.01496.x/full

@Ana, I'm afraid that both papers you've linked are too phylosophical for the kind of project we are developing. I mean, we need an ontology to model and build a data set regarding the Linked Data initiative. Anyway, thank you very much :)

It's highly unlikely that you will find a ready made list of category terms, let alone an ontology. Roman Ingarden's "Ontology of the Literary Work of Art" and "Cognition of the Literary Work of Art" does touch on paintings, but in terms of ontological structure rather than simple datasets. Simplistic adoption of such a dataset without considering the implications of using it may also land you in trouble.

Always remember: “Classifications that appear natural, eloquent, and homogeneous within a given human context appear forced and heterogeneous outside of that context.”

Bowker, G. C. and S. L. Star (1999). Sorting Things Out: Classification and its consequences. Cambridge, MIT Press.

@John I agree with you up to a extent. Maybe I've reduced the expression way too much by saying that I'm looking for a data set. What I'm looking for is an ontology to express the kind of compositions that a painting might exhibit (e.g.: portraits, landscapes, ...) and what defines it.

Of course, I'm taking into account that the adoption of any model may bias the view that my system offer of the reality, but, in my opinion, this simplification comes handy when it comes to classifying the resources and helping the users to express their information needs. I completely agree with the quote you've marked: one of the consequences of taking decisions at modeling time when developing knowledge based information systems is this dependence of the context.

Thank you very much, John :)

Thanks for that clarification Carlos. In that case you may well find Ingarden very useful. If he is not sufficiently specific on artworks, his approach to literature, and the borderline cases in particular might help. I used his analysis of scientific works in my thesis, which was looking at developing a categorisation scheme.

Lamp, J.W. (2011) Information categorisation: an emergent approach University of Melbourne

http://repository.unimelb.edu.au/10187/10102

Thank you very much, I'll have a look to Ingarden and to your thesis :)

I am not sure that this can help you - but maybe will show some ideas:

ONTOLOGY-BASED ANNOTATION OF PAINTINGS WITH ARTISTIC CONCEPTS MARCHENKO YELIZAVETA

(http://scholarbank.nus.edu/termsofuse;jsessionid=A03E20CE11293E0028A3ED7E92940D0A

@Carlos, have you managed to find the ontology you were looking for? I am currently searching for exacly the same thing, and apparently experiencing the same problem as you several years ago - no resources that could be easily accessible (or maybe 'findable'). 

@Szymon, In the end, we used a little adapted version of the painting ontology of MOLTO project, as it was expressive enough for our purposes. 

Recalling from the previous answers, it extends the model CIDOC-CRM; ICONCLASS, a classification system that has already an implmentetation in SKOS, and a direct implementation of the CIDOC-CRM model in OWL DL 1.0 called Erlangen-CRM; but I'm afraid I cannot recall further details right now (I would have to check it in the project proposal document). 

@Carlos, thank you for you answer. How did you get the access to the MOLTO project ontology? Is is available for download or did you access it via some kind of API?

I got the reference from "Multilingual access to cultural heritage content

on the Semantic Web", http://aclweb.org/anthology/W/W13/W13-2715.pdf

The URL is still working: 

http://spraakdata.gu.se/svedd/painting-ontology/painting.owl

Hope you find it useful :) 

Conference Paper Quantitative Analysis of Art Market Using Ontologies, Named ...

This paper contains an ontology for lots sold at the art market, perhaps it might be something interesting for you.