A Method Combining Syntax Analysis and Correction Rules to Re-construct the Re-flowable Document

Abstract

To improve the shortcomings of fault-tolerance ability in the re-construction method for re-flowable document structure, a new method combining left-corner method and correction rules is proposed, where the xml schema is applied to construct a syntax tree of typesetting rules of document components, and left-corner method is applied to analyze the logical components of the document supervised by the syntax tree. In the analysis process, the correction rules are used to correct the possible errors existed in document component and eventually get the most likely document structure. The results show that the algorithm can effectively improve the fault tolerance in the document structure reconstruction and the accuracy of document structure recognition, which forms the foundation for document understanding and format checking.