Electronic publishing has become an irreversible trend in publishing industry. An important step in electronic publishing is to convert a scanned copy or a digital image of a document into editable texts, which is very important to the people with visual impairment, who access the information by means of screen-readers. Existing methods for document parsing and paragraph reconstruction only work on PostScript documents with a simple and standard layout. These techniques are severely limited when applied to documents with complex layout and irregular typesetting. To address this issue, we propose a rectangle segmenting approach to make paragraph reconstruction for Postscript documents with complex layout.
展开▼
