The IMAGACT Visual Ontology. An Extendable Multilingual Infrastructure for the representation of lexical encoding of Action