Improving the character recognition efficiency of feed forward BP neural network

TL;DR

This paper enhances feed-forward back-propagation neural networks for character recognition by experimenting with multiple hidden layers and a modified momentum term, achieving higher accuracy with improved training efficiency.

Contribution

It introduces a modified additional momentum term and evaluates the impact of varying hidden layer counts on recognition accuracy.

Findings

Recognition accuracy increased with more hidden layers.

Modified momentum term improved training speed and accuracy.

Optimal hidden layer configuration balances accuracy and training time.

Abstract

This work is focused on improving the character recognition capability of feed-forward back-propagation neural network by using one, two and three hidden layers and the modified additional momentum term. 182 English letters were collected for this work and the equivalent binary matrix form of these characters was applied to the neural network as training patterns. While the network was getting trained, the connection weights were modified at each epoch of learning. For each training sample, the error surface was examined for minima by computing the gradient descent. We started the experiment by using one hidden layer and the number of hidden layers was increased up to three and it has been observed that accuracy of the network was increased with low mean square error but at the cost of training time. The recognition accuracy was improved further when modified additional momentum term was used.