Abstract
This chapter addresses automatic printed Arabic text recognition. Arabic text recognition has its own difficulties due to the cursive nature of the scripts, overlapping characters, large number of dots and diacritics, etc. In this chapter, we present a general framework for a printed Arabic text recognition system. We then discuss different phases of such a system, e.g., pre-processing, feature extraction, and classification. We present different reported techniques for each phase. In addition, different databases for printed Arabic text recognition are discussed here. We conclude this chapter by presenting several experimental results for hidden Markov model (HMM)-based printed Arabic text recognition.