Maybe better define an RDF vocabulary for that instead, so that individual DIVs and IMGs can be correctly annotated in HTML. ;)