Classification of Three Fatigue Levels for Bills Using Acoustic Frequency Band Energy Patterns

Abstract

This paper proposes a method to classify three fatiguelevels of bills using acoustic frequency band energypatterns by a neural network. The proposed method dealswith an acoustic frequency band energy pattern from theacoustic signal which has been generated by the billpassing through a banking machine. The competitive neuralnetwork is used to classify a bill into one of threedifferent fatigue levels. We use two types of the LVQalgorithm for training of the competitive neural network.The experimental results show the effectiveness of theproposed method.