Generic Automatic Recognition System for Handwritten Arabic Script Documents

This proposal addresses the feasibility of developing a generic system framework for automatically recognizing handwritten text for non-Arabic languages using Arabic-style script such as Urdu or Pashto. Despite recent progress in automatic handwritten Arabic recognition, little attention has been paid to non-Arabic languages using Arabic-style scripts. Nonetheless, automatic recognition of non-Arabic languages using Arabic-style scripts such as Urdu and Pashto has great potential in both military and commercial applications. Hence, the goal of this proposal is to develop a generic system framework for automatically recognizing handwritten text for non-Arabic languages using Arabic-style script. In particular, we focus on extending the optical Arabic text recognition capability to the Urdu language in which computing standards are established. The generic Arabic-style script recognition system will be based on the Hidden Markov Model approach that is suitable for cursive and context-sensitive Urdu scripts. Various feature-extraction algorithms will be implemented based on the proposed generic handwritten Arabic-style script recognition framework. We shall evaluate the basic performance of the developed system using a handwritten Urdu database collected during Phase I.