المجلة الدولية للعلوم والتقنية

The Arabic text recognition (i.e.: word extraction, pattern recognition or segmentation) needs databases for training and validation (which are difficult to find such Arabic comprehensive databases), in addition to that, Arabic writing recognition involves different pre-processing, segmentation and recognition methods. To overcome these issues, we propose a different Arabic Optical Character Recognition "AOCR" approach, that is segmentation-free character recognition independent of a lexicon of words. A system for Arabic character recognition which is implemented using either Walsh Hadamard Transform (WHT) or Discrete Cosine Transform (DCT) is analysed. The system starts by scanning the document, then matching the archive of character images with the text image, and determines the positions of these characters in the text image. Three font types (Arial, Times New Roman and Simplified Arabic), and nine font sizes (8,10,12,14,16,20,24 and 28) are used for this analysis.