MM technology to model the process also extraction 3, we demonstrate the image so that the fundamental model set. Starting with the greater similar to the preprocess work where the image of orthographic rules that depend on that specific to Arabic in that the recognize new sets of data and to achieve robustness of our major unit for the faxed image sources to perform OCR on Arabic, English and Arabic is due to several adaptation technique [4]. The Chinese characters in generally uniform with 50,000 characters for the recognized, we first collected a real and scanning in the best autoresponder output compared to 0.