Handwritten Arabic word recognition: A review of common approaches

• Published in: 2008 International Conference on Computer and Communication Engineering pp. 801 - 805
• Main Authors: Assma, O.H., Khalifa, O.O., Hassan, A.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-05-2008
• Subjects: Character recognition; Dictionaries; Handwriting recognition; Hidden Markov models; Image databases; Image recognition; Image segmentation; Shape; Testing; Text recognition
• ISBN: 9781424416912; 1424416914
• DOI: 10.1109/ICCCE.2008.4580716

Automated methods for the recognition of Arabic script are at an early stage compared to their equivalent for the recognition of Latin and Chinese. In this paper, different approaches used for handwritten Arabic word recognition were reviewed. An introduction to the Arabic script is given, followed by a description of algorithms for the process involved: segmentation, feature extraction, classification, and recognition. However, an automatic recognition of text on scanned images has enabled many applications such as words spotting in large volumes of documents, automatic sorting of postal mail, and convenient editing of previously printed documents. The domain of handwriting in the Arabic script presents unique technical challenges and has been addressed more recently than other domains. Many different methods have been proposed and applied to various types of images, and a variety of approaches have been already developed, tested and returned good results. Yet the database used was too small compared to the huge size of Arabic texts, which gives way to other approaches to be developed based on a bigger database. Since handwriting recognition is such a large subject, there is plenty of scope for the work in this field. Finally, a comparison to show pros and cons of the approaches reviewed is conducted.