Automating Artl@s – extracting data from exhibition catalogues

Catalogues, which have been published for centuries, are an extremely precious resource for scholars. Using the Artl@s database as an example, where exhibition catalogues are transformed into a georeferenced database, we question the possibility of an (almost) automatic transformation of pdfs into semantically annotated data. To do so, we present and analyse the graphic organisation of exhibition catalogues, before exploring a possible modeling into TEI (involving possible enhancement of the guidelines).