A new lecture support based on on-line arabic handwriting recognition

In this paper we have developed a prototype of engineering education using interactive learning based on PDAs and online handwriting recognition. This lecture support system concerns the fourth students and consists of handwritten Arabic letters and digits learning. The student writes one Arabic letter on the surface of digital tablet or on the screen of tablet PC. An evaluation will be done on the handwritten letter correctness. The recognition system was based on the use of neural networks developed in a fuzzy concept. The extraction of the handwritten trajectory parameters is based on ldquoBeta-elliptical modelrdquo. To test the performance of our system, we developed an Arabic database ldquoADABrdquo which contains 100 000 letters, 30 000 digits and 500 words. The recognition rate obtained on Arabic letter set is about 82.82 %.