A segmentation scheme of arabic words with harakat

Today, Arabic is one of the big challenges in Optical Character Recognition (OCR) to support a digital communication. There are many research on arabic OCR, either printed or handwritten input. However, the research on arabic OCR with harakat is still little due to the high degree of difficulty in segmentation techniques. In this paper, we propose a new segmentation scheme of the arabic character with harakat. The segmentation process is one of significant step in arabic OCR. The performance of proposed character segmentation is 97.55% and the performance of harakat separation is 88.78%. It means that the proposed algorithm has been running well.