Abstract:
In the paper issues of table (logical) layout analysis are discussed. The task of structuring table information presented on unstructured documents and addressed to human comprehension to structured representation is particularly considered. The table transformation (conversion) system from semistructured representation to relation in database is proposed. The system provides a semi-automatic recovering dimensions (domains) used in a table. Proposed transformation is focused on tables originally generated from databases.
Keywords:document analysis and recognition, information extraction from tables, table analysis and processing, table conversion.