Machine learning in handwritten arabic text recognition

Automated recognition of handwritten text is one of the most interesting applications of machine learning. This chapter poses handwritten Arabic text recognition as a learning problem and provides an overview of the ML techniques that have been used to address this challenging task. The use of co-training for solving the problem of paucity of labeled training data and structural learning approaches to capture contextual information for feature enhancement have been presented. A system for recognition of Arabic PAWs using an ensemble of biased learners in a hierarchical framework and the use of techniques such as Artificial Neural Networks, Deep Belief Networks, and Hidden Markov models within this hierarchical framework to improve text recognition have been described. The chapter also describes some of the features that have been successfully used for handwritten Arabic text classification for completeness since the selection of discriminative features is critical to the success of the classification task.