Mona Lisa

Original Image

This page is an example for the paper A Framework for Media Adaptation Using the Web and the Semantic Web, by Sébastien Laborie and Antoine Zimmermann, accepted for publication in SMAP 2007. It contains a picture in PNG format (http://www.inrialpes.fr/exmo/people/zimmer/smap2007/monalisa.png) with an alt attribute having value "Mona Lisa". The image format and resolution cannot be displayed by certain devices. For instance, the Samsung U600 (a recently released mobile phone) can display JPEG pictures but is not able to display PNG format.

This picture is described in two different RDF/XML files: d1.rdf and d2.rdf. The first description uses an existing ontology found on the Web (Exif vocabulary workspace). The second description uses the vocabulary defined in ontology O1.owl illustrated in the paper in Figure 3.

Alternative Images

Using the alt attribute from the original image as a query in Google Image Search, here are the results for small images:

Using this method, the small picture on the left is a candidate replacement for adapting the original picture. However, considering the RDF/XML description d3.rdf based on ontology O2.owl, a semantic approach would discard this candidate.

Moreover, another possible candidate is the picture on the left. Besides, with the RDF/XML descriptions d4.rdf and d5.rdf based on O1.owl and O2.owl, respectively, it would be selected as a good replacement candidate with a semantic approach.

Ontology Alignment

Since the descriptions are based on two different ontologies, it is necessary to match corresponding concepts or properties. We propose such an alignment between O1.owl and O2.owl, available at alignment.rdf. It is written in the Alignment Format defined in An API for Ontology Alignment. This helps finding the most similar descriptions independently of the terminology used to describe them.

monalisa.html: last modified 2007/08/21 10:12:02 by Antoine Zimmermann.