A New Character Segmentation Approach for Printed Arabic Text

Abstract

Optical character recognition has been the subject of considerable research activity for many years, segmentation is a very important step in the process of character recognition, and has a major impact on the quality of the recognition system. An error produced in this step can be very serious, since it affects the performance of the recognition systems carrying the character to be either misrecognized or in most cases rejected completely. In this paper we propose a segmentation approach for printed Arabic text for different segmentation stages; line segmentation, subword segmentation and character segmentation based on projection profile technique and pen thickness estimation, interesting results have been obtained.