Script independent text pre-processing and segmentation for OCR

Abstract

Optical Character Recognition (OCR) systems have been numerously developed for the recognition of printed script in many languages. Multiple approaches to pre-processing and segmentation exist for various scripts where OCR accuracy mainly depends on the text pre-processing and segmentation algorithm being used for the document. When the document is scanned it can be put in any arbitrary angle which would appear in the image as skew angle. Our experimental results proposed in the paper assures the superior algorithm for correction of skew angle of the text document. Projection Profile based methods used makes segmentation easy to separate the text in document image into lines, words and characters independent of the Language in the Text.