Neural network application to container number recognition

Abstract

The authors describe an application of neural network technology to a practical problem: the recognition of container identification numbers. The general problem is described and technical difficulties are highlighted. A solution is proposed and a set of algorithms is implemented. Emphasis is placed here on the development of the character recognition algorithm, which uses a neural network model known as the neocognitron. The general recognition methodology is discussed and the experimental results are reported. It is observed that this neural network model can attain a high level of accuracy even though the input patterns are highly distorted.<>