Performance Enhancement of Hopfield Neural Network as Associative Memory for Finger Print Images with Pseudoinverse

This paper is designed to analyse the performance of a Hopfield neural network for storage and recall of fingerprint images. The study implements a form of unsupervised learning. The paper first discusses the storage and recall via hebbian learning and the problem areas or the efficiency issues involved and then the performance enhancement via the pseudoinverse learning. Performance is measured with respect to storage capacity, recall of distorted or noisy patterns i.e association of a noisy version of a stored pattern to the original stored pattern for testing the accretive behaviour of the network and association of new or unstored patterns to some stored pattern.