Development of technique for recognition of old newspapers of Gurumukhi script

Abstract

Pattern recognition emphasizes on recognition of data patterns based on certain features. Document Analysis and Recognition (DAR) is a branch of pattern recognition in which information typed or handwritten on a paper is extracted automatically. Over the period of time, papers are the eminent source of dissemination of information. With the advent of technology in computer science, researchers thought of paperless work culture. Nowadays, with inexpensive high quality scanning devices in combination of high speed computer systems and OCR packages, recognition of these kinds of documents is possible for most of the users. One more important source of information is newspapers. Newspapers bring a huge collection of news from around the whole world. Newspapers serve us a very primary source of information by providing data about historical and current events. newlineThis paramount source of information should be zealously preserved because newspapers deteriorate with time and the information printed on the papers can lose. Preservation of information is feasible in generating digital images or microfilms from the newspaper pages but they do not serve the purpose of full text search. Full text search is feasible through optical character recognition (OCR) on the newspaper text. Currently, there is no such research work present that has dealt with recognition of a newspaper article text printed in Gurumukhi script. This work would be the first attempt in this direction. newline