Develop A Prototype for an Arabic Handwriting Recognition system that has
Reliable Accuracy (> 90%)
- Large Vocabulary (> 500k words)
- Real Time Performance (including on mobile devices)
Build a Hidden Markov Model (HMM) based system to provide solutions for most of the difficulties inherent in recognizing Arabic script along with new preprocessing technique that will be applied on the delayed strokes to match the structure of the HMM model.