On Reliability of Paper Currency Classifiers Using Neural Networks and PCA
Ali Ahmadi, Sigeru Omatu, Toshihisa Kosaka · IEEJ Transactions on Electronics Information and Systems · 2003
We propose an approach to increase the reliability of a neuro-classifier for paper currency recognition by using a principal component analysis (PCA) algorithm. The PCA is used to extract the main features of input data and reducing the data size. A learning vector quantization (LVQ) neural network is applied as the main classifier of the system. By defining a new algorithm for rating the reliability, we evaluate the reliability of the system for 1, 200 sample test data. The result shows that the average reliability measure is increased up to 99.6% when the number of PCA components as well as number of LVQ codebooks are taken properly.