A Survey on Techniques Used in Arabic Handwritten Character Recognition

Arabic handwritten character recognition is one of the challenging tasks in the field of Natural Language Processing. Currently there are so many English recognition methods. But due to the diversity in Arabic character's position and shapes, there are only few Natural Language Processing metods available for Arabic Language. This research paper is an attempt to study various algorithms used for Arabic handwritten character recognition. Latin script OCR is a well-researched area. Arabic script OCR is an emerging area of intense research that follows few reasons: Firstly, around 200 million people in the world use Arabic as their first language. Secondly, around 1.6 billion people follow Islam wherein it is compulsory to recite the religious scripture which was revealed in Arabic. Thirdly, after achieving considerable success in Latin Text OCR, researchers have now focused on extending their prowess to the more challenging Arabic text OCR. Hand written character recognition using many methods like Support Vector Machine analysis algorithm, K Nearest algorithm, Bioinspired alogorithm like artificial neural networks and convolutional neural networks etc. [1,2,3] In this study, preprocessing, feature extraction and post processing methods are focused.