Recognition for the Banknotes Grade Based on CPN

The counter propagation networks (CPN) is used to improve the accuracy of the grade of banknotes recognition. First, self-organizing map (SOM) is used to cluster banknotes data into regions based on the different feature of banknotes; second, the principal component analysis (PCA) is performed in each region to extract the main principal features of banknotes data; finally, the CPN is employed as the main classifier to identify the banknotes of different grades. The recognition effects of CPN and BP are compared in this paper. The results show that the reliability and speed of CPN are greatly better than that of the BP. The experimentation shows that the CPN can primly solve the recognition problem of the grade of banknotes.