Existing techniques in Arabic Characters Recognition (ACR)
Aid Ahmed Radaideh, Mohd Shafry Mohd Rahim · Journal of Informatics and Mathematical Sciences · 2016
Text or handwritten document writing recognition or understanding through characters whether online or off-line play an important role in many applications, e.g. recognizing text on bank cheques, editing old documents or translating online writing. The Arabic language characters recognition in Arabic text (hand written and printed text) is a challenging job, which is addressed by researchers in two different domains, i.e. online and off-line. The challenges are due to the cursive nature of the Arabic language. The approaches designed and adopted to achieve the target of Arabic characters recognition are also broadly classified into two, i.e. with segmentation and without segmentation. Machines can easily classify the characters if they are properly presented to it. The aim of this paper is to review critically state-of-the-art and to highlight the drawbacks in recent Arabic character recognition (ACR) for realization of the needs for robust online or off-line character recognition. This paper will properly enlighten the way for new researchers to improve or overcome the problems faced by previous studies and in addition provide a brief summary of ACR databases.