Florentine coats of arms on the Web: experimenting retrieval based on text or image content

The paper describes two different ways of accessing a collection of Florentine coats of arms, which was digitised and made available on the Web. One way is a traditional approach: textual description and indexing. As heraldry uses a specific and complex terminology the textual approach seemed alright for experts but insufficient for laypersons. Therefore an additional retrieval mode was chosen which is based on the visual specification by users of the content of images representing the coat of arms. This access mode allows users who are not familiar with heraldry to access the collection.