Author: A.W.M. Smeulders, L. Hardman, G. Schreiber, J.M. Geusebroek
Reference: ACM Workshop on Multimedia Information Retrieval, New York, 2002
Link: http://dare.uva.nl/... (pdf)
Abstract:
We discuss access to e-documents from three different perspectives beyond the plain keyword web-search of the entire document. The first one is the situation-depending delivery of multimedia documents adapting the preferred form (picture, text, speech) to the available information capacity or need exemplified by documents from the annotated media database of the Rijksmuseum. It goes beyond Quality of Service methods which insist on delivering information in the same form even if that is no longer effective. Secondly, we discuss the use of ontologies to provide access across diverse library categorizations as part of the W3C semantic web. The system translates codes in the one catalogue system into a set of codes in another expanding the potential access to digital heritage knowledge across all library systems in the ontology, such as AAT, Worldnet and IconClass. Thirdly, we discuss access to the pictorial contents of paintings by computer vision techniques, here showing in examples of Pieter de Hoogh and Johannes Vermeer which one of the two consistently painting photometric realistic in addition to adhering to the geometric realism as they both did. It is concluded access is the key issue in digital cultural heritage - be it access by situational delivery of e-document of cultural heritage, be it access to diverse knowledge systems, or be it access to the pictorial content of the picture.
No comments:
Post a Comment