Graph grammar based analysis system of complex table form document

Abstract

Structure analysis of table form document is importantbecause printed documents and also electronical documentsonly provide geometrical layout and lexical information explicitly.To handle these documents automatically, logicalstructure information is necessary. In this paper, we firstpropose a general representation of table form documentbased on XML, which contains both structure and layoutinformation. Next, we present structure analysis systembased on graph grammar which represents document structureknowledge. As the relation between adjacent fields intable form documents become two dimensional, two dimensionalnotation is necessary to denote structural knowledge.Therefore, we adopt two dimensional graph grammar to denotethem. By using grammar notation, we can easily modifyand keep consistency of it, as the rules are relatively simple.Another advantage of using grammar notation is that,it can be used for generating documents only from logicalstructure. Experimental results have shown that the systemsuccessfully analyzed several kinds of table forms.