An advanced neuro-classification of new and used bills using the spectral patterns is proposed. An acoustic spectral pattern is obtained from the output of the two-stage adaptive digital filters (ADFs) for time-series acoustic data. The acoustic spectral patterns are fed to a competitive neural network, and classified into some categories which show worn-out degrees of the bill. The proposed method is based on extension of an ADF, an individual adaptation (IA) algorithm, and a learning vector quantization (LVQ) algorithm. The experimental results show that the proposed method is useful to classify new and used bills.
Efficient systolic architectures for 1-D and 2-D DLMS adaptive digital filters
