Document Table Recognition by Graph Rewriting

Abstract

This paper proposes a bottom-up approach for identifying and recognizing tables within a document. This approach is based on the paradigm of graph rewriting. First, the document image is transformed into a layout graph whose nodes and edges respectively represent document entities and their interrelations. This graph is subsequently rewritten using a set of rules designed for and based on apriori document knowledge and general formatting conventions. The resulting graph provides both logical and layout views of the document content.