Abstract Character  recognition  has  served  as  one  of  the  prin cipal  proving  grounds  for  neural  network  methods  an d  has  emerged as one of the most successful applications  of this technology. In this paper, a new network is designed to recognize a  set  of  handwritten  arabic  characters.  This  new  netw ork  consists  of  two  stages. The  first  is  to  recognize  the  main  shape  of  the  character,  and  the  second  stage  is  for  dots  recogni tion. Also,  the  characteristics,  structure,  and  the training algorithm for the  network are presented. 

References

[1] Amin A., Off Line Arabic Character Recognition: A Survey, in Proceedings of Document Analysis and Recognition (ICDAR'9 ), Germany, vol. 2, pp. 596 599, 1997.

[2] Amin A., Recognition of Hand printed Characters Based on Structural Description and Inductive Logic Programming, Pattern Recognition Letters, vol. 24, no. 16, pp. 3187 3196, 2003.

[3] Burges C., A Tutorial on Support Vector Machines for Pattern Recognition, Computer Journal of Data Mining and Knowledge Discovery , vol. 2, no. 2, pp. 121 167, 1998.

[4] El Wakil S. and Shoukry A., On line Recognition of Handwritten Isolated Arabic Characters, Official Journal of the Pattern Recognition Society , vol. 22, no. 2, pp 97 105, 1989.

[5] Fukushima K., Neocognitron: A Self Organizing Neural Network Model for a Mechanism of Pattern Recognition Unaffected by a Shift in Position, Computer Journal of Biological Cybernetics, vol. 36, no. 4, pp. 193 202, 1980. Layer Units Connections with previous layer Free parameters Input 256 First hidden layer 384 9984 534 Second hidden layer 192 19392 792 Output 68 13124 13124 Summation 900 42500 14450 Handwriting Arabic Character Recognition LeNet Using Neural Network 309

[6] Hashimoto W., Quadratic Forms in Natural Images, Computer Journal of Network Computation in Neural Systems , vol. 14, no. 4, pp. 765 788, 2003.

[7] Kharma N. and Ward R., A Novel Invariant Mapping Applied to Hand written Arabic Character Recognition, Official Journal of the Pattern Recognition Society , vol. 34, no. 11, pp. 2115 2120, 2001.

[8] Kherallah M., Haddad A., and Alimi A., On Line Handwritten Digit Recognition Based on Trajectory and Velocity Modeling, Pattern Recognition Letters , vol. 29, no. 5, pp. 580 594, 2007.

[9] LeCun Y., Boser B., Denker J., Henderson D., Howard R., Hubbard W., and Jackel L., Handwritten Diget Recognition with a Back- Propagation Network , Morgan Kaufman, USA, pp. 396 404, 1990.

[10] LeCun Y., Bottou L., Bengio Y., and Haffner P., Gradient Based Learning Applied to Document Recognition, in Proceedings of the IEEE, USA, vol. 86, no. 11, pp. 2278 2324, 1998.

[11] Meuller R., Mika S., Ratsch G., Tsuda K., and Sch olkopf B., An Introduction to Kernel Based Learning Algorithms, IEEE Computer Journal of Transactions on Neural Networks , vol. 12, no. 2, pp. 181 202, 2001.

[12] Naoum R., Al Rawi H., and Jabir A., Supervised Neocognitron for Handwritten Arabic Characters Recognition, Basrah Science, vol. 18, no. 2, pp. 109 126, 2000.

[13] Saleh, S. An On-Line Automatic Arabic Document Reader , MSc Thesis, University of Basrah, Iraq, 1998.

[14] Yang M., Graham L., Colin H., and Swe M., Segmentation and Recognition of Phonetic Features in Handwritten Pitman Shorthand, Computer Journal of Pattern Recognition , vol. 41, no. 4, pp. 1280 1294, 2007.

[15] Yuhas B. and Ansari N., Neural Networks in Telecommunications , Kluwer Academic Publishers, UK, 1994.

[16] Abdulla S., Al Nassiri A., and Abdul Salam R., Off Line Arabic Handwritten Word Segmentation Using Rotational Invariant Segments Features, in International Arab Journal of Information Technology, vol. 5, no. 2, pp. 200 208, 2008. Rashad Al-Jawfi is lecturer in fractal geometry and neural networks. He holds a PhD degree in mathematics and computer science from the University of Baghdad. He is a member of IUM. 307 The International Arab Journal of Inf ormation Technology, Vol. 6, No. 3, July 2009