Simulation, pattern recognition, modular neural networks, neuro-classifier

It is the purpose of the present paper to suggest an approach to utilization of mathematical models of a classification system for pattern recognition. Pattern recognition has a long history but has recently become much more widespread as the automated capture of signals and images has become cheaper. Very many of the applications of neural networks are to classification, and so are within the field of pattern recognition. This article describes a classification system for pattern recognition based on artificial neural networks with modular architecture. We use a three layer feedforward network model that is learned with the backpropagation algorithm for all experiments. Our experimental recognition objects were digits and their type fonts. We also propose outline further development on this topic in conclusion.