The Knossos project: constructing multipoint-of-view and practice-adapted interfaces for indexing and retrieving information through iconic corpora

The Knossos project is a text-founded anthropocentric system where, by extensive use of shared intelligent skills, the user, in constant interaction with the machine, receives semantic assistance that enhances her/his descriptive competence and helps her/him to quickly set up powerful textual descriptions of image meaning(s) or retrieve by content image information using textual search techniques. Indexing and retrieving images are envisaged from a symbolic point of view. The Knossos approach appears therefore as an integrated approach concerning all types of images. The main scientific objective of the project is to show in which manner the concept of textuality may be used as semiotic middle offering a renovated space of modelling to image meaning. Indeed, the text may be taken as a central (perhaps the only) semiotic e-medium in a universe where the lexical paradigm is predominant. By doing this, the Knossos project takes place in the semantic Web debate. Semiotic experts who give more or less complete panel of relevant iconic categories of the corpus create iconic genders. The system may be used under an ontology creation mode presents the same functionalities but concerning different iconic corpora.