Hybrid Binarization Approach for Typewritten Gurmukhi Script Documents
In this digital world having motto “Save Paper Save Trees”, most of the documentation work is being transformed into digital format. This paper presents a hybrid approach for the binarization of Typewritten Gurmukhi script documents. These typewritten documents suffer from various degradations such as ageing, faded characters and noise. In typewritten documents, retaining of characters is very important for character segmentation. Binarization is the preprocessing phase of the Optical Character Recognition system. We have presented a hybrid approach for image restoration by using the blend of local and global threshold method . We have also compared results of our technique with other binarization techniques: Sauvola, Otsu, Bernsen and H- DIBCO’10. Our approach combines the noise removal and image restoration goals into a single framework, thus producing high quality document images from degraded typewritten documents.