A Neural Network for Real-World Postal Address Recognition

Abstract

In this paper, we present a description of an implemented system for the recognition of printed and handwritten postal addresses, based on Artificial Neural Networks (ANNs). Two classification methods were compared for the task of character and address recognition. We compared two neural network techniques, measuring recognition rate and accuracy. The C programming language, a SUN workstation, and the SP2 Supercomputer were used for the experiments. The system has been successfully tested on real world printed and handwritten postal addresses. Some experimental results are presented in this paper. 1. Introduction ANNs have been successfully used in such areas as pattern recognition [1], medical applications [2], fingerprint analysis [3] and signature verification [4] to name just a few. This paper attempts to take one step forward in producing another successful practical application: Recognition of handwritten postal addresses. As outlined in [5] and more recently in [6], there has b...