Online Arabic Handwritten Digits Recognition

This paper fills a void in the literature of online Arabic handwritten digits recognition as no systems are dedicated to this problem. The two main contributions of this paper are introducing a large online Arabic handwritten digits dataset and developing an efficient online Arabic handwritten digits recognition system. In the dataset, we collected 30,000 online Arabic digits from 300 writers. The developed system uses a combination of temporal and spatial features to recognize those digits. The system achieved 98.73% recognition rate. Comparison with a commercial product demonstrates the superiority of the proposed system.