1. Introduction
There is widespread and increasing interest in converting paper-based document information to electronic format for computer storage, access and processing. The conversion is being routinely done for record management; automated document delivery, document achieving, journal distribution, and other applications. The field of document image understanding deals with a variety of documents such as business letters and faxes, forms, technical articles, postal mail pieces, and maps and drawings [1] [2]. As a result, transformation of paper documents to their electronic versions and subsequent document image understanding have become an important and a challenge application domain for computer and pattern recognition researchers [1] [2] [3].