Comparison of feed-forward neural net algorithms in application to character recognition

• Published in: 2001 IEEE Tencon - 2001 IEEE Region 10 Conference: Electrical and Electronic Technology Vol. 1; pp. 165 - 169 vol.1
• Main Author: Kamruzzaman, J.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 2001
• Subjects: Backpropagation algorithms; Character recognition; Data preprocessing; Feature extraction; Gravity; Multi-layer neural network; Neural networks; Pattern recognition; Smoothing methods; System testing
• ISBN: 0780371011; 9780780371019
• DOI: 10.1109/TENCON.2001.949573

In a neural network based character recognition system it is important to choose a training algorithm with high generalization ability. In this paper, we apply three different multilayer feedforward training algorithms namely, backpropagation, double backpropagation and weight smoothing algorithm in a neural network based invariant character recognition model. The model consists of a preprocessor and a classifier. The preprocessor extracts geometrical features of the input character and passes the feature values through a rapid transform block which performs a cyclic shift invariant transform on its input. The classifier is a neural network classifier. Simulation results with 26 English capital letters show that the recognition system achieves best performance with significantly high recognition rate when trained with weight smoothing learning algorithm.