音響信号の振幅差分特徴を用いたサポートベクターマシンによる疲弊紙幣の識別

抄録

The circulation bills include a large amount of fatigue bills which cause various types of problems such as the paper jam in automatic tellers due to overworked and exhausted bills. The high accuracy bill classification technique, which identifies the level of fatigue in the bills as well as distinguishes between the used and the new ones, is greatly required in order to prune these fatigued bills. The purpose of this paper is to propose a classification method of fatigue bills based on acoustic signals. The amplitude difference in the acoustic signal of bills is employed as characteristics of the bill classification. Further, a Support Vector Machine (SVM) is introduced as a classification method. We then perform the classification experiments using the acoustic signals from the practical bill conveyance device in order to show the effectiveness of the proposed method by the bill identification experimentation.