In this project artificial neural network has been called for its application as characters recognizing network. The network is made to learn as per the requirement by training them with some specific patterns that corresponds to the character. The number of input and output layer neurons is chosen. The training patterns and testing patterns are designed using matrices 0's and 1's. The weights in the network are adjusted using back propagation algorithm (delta rule) for training patterns and are checked for testing patterns. Then we train the network using those input patterns followed by testing the neural network with given training patterns

1. McCulloch, and Pitts, A logical calculus of the ideal imminent nervous activity. Bulletin of Mathematical Biophysics
2. Minsky, M. L. and papert S. 1969. Perceptions Cambridge, MA:MIT Press
3. Pitts, and McCulloch, W. W. 1947. How we Know universals. Bulletin of mathematical biophysics
4. Widrow,B and Angell. Reliable networks for computing and control aerospace Engineering.
5. Kohenen. T. 1984 Self-organisation and associative memory: series in information sciences.
6. Cohen. M. A. and Grossberg, Absolute stability of global pattern formation and parallel memory storage by competitive neural networks.
7. Hopfield,J. J Neural Networks and physical systems with emergent collective computational abilities.
8. F. J. Pieda, "generalization of back propagation to recurrent neural networks. "
9. B. K. Verma,"New methods of training the MLP,"
10. J. J. Mulawka ," Improving the training time of the back propagation algorithms,"
11. Dan W. Patterson, Artificial intelligence and expert system
12. G. L. Martin and J. A. Pittman,"recognition hand-printed letters and digits using back-propagation learning,"
13. J. P. Nadal,"Learning in feedforward layered networks: the algorithms
14. G. K. Pieda, "generalization of back propagation "

Neural Networks