Fourth International Conference Document Analysis and Recognition (ICDAR'97)
Automatic Knowledge Acquisition for Spatial Document Interpretation
Ulm, GERMANY
August 18-August 20
ISBN: 0-8186-7898-4
In this paper, a qualitative representation for the layout of structured documents as well as classes of documents is presented, which is established by the means of supervised learning from a labeled training set of documents. For this formal representation, an inference algorithm has been developed, adopted from error tolerant subgraph isomorphism, which assigns logic labels to layout objects of a test document.
Index Terms:
Spatial Representation, Machine Learning, Logic Labelling
Citation:
H. Walischewski, "Automatic Knowledge Acquisition for Spatial Document Interpretation," icdar, pp.243, Fourth International Conference Document Analysis and Recognition (ICDAR'97), 1997