In this contribution we consider different hypertext structures one can encounter in documents. Methods for automatically finding those hypertext structures in paper documents [1] and in photographs with captions are presented [2]. Main focus is on finding relations between the content of a figure and associated text. The structures derived form the basis for convenient Internet access to the documents.
[1] M. Worring and A.W.M. Smeulders : From linear to non-linear reading: a case study to provide Internet access to paper documents. ICDAR97, Ulm.
[2] R. Shrihari : Automatic indexing and content-based retrieval of captioned images. IEEE Computer 28-9, 1995