Word level script identification in bilingual documents through discriminating features

Abstract

India is a multi-lingual and multi-script country where a line of a bilingual document page may contain text words in regional language and numerals in English. For Optical Character Recognition (OCR) of such a document page, it is necessary to identify different script forms before running an individual OCR of the scripts. In this paper, we examine the use of discriminating features (aspect ratio, strokes, eccentricity, etc,) as a tool for determining the script at word level in three bilingual documents representing Kannada, Tamil and Devnagari containing English numerals, based on the observation that every text has the distinct visual appearance. The knearest neighbour algorithm is used to classify the new word images. The proposed algorithm is tested on 2500 sample words with various font styles and sizes. The results obtained are quite encouraging. © 2007 IEEE.