Research on the Neural Network with RBF for Currency Recognition

B. Sun and F. Takeda (Japan)


Neural network, RBF, Currency, Recognition


In this paper, a new type of neural network (NN) is proposed to recognize the Chinese currency. In our previous researches, the recognition system with the NN using sigmoid function as unit function had been applied successfully for many kinds of worldwide currency. In this research, a new kind of currency, Chinese currency is recognized as an object, and radial basis function (RBF) is used in the units of hidden layer and output layer of the NN. Since the RBF is a localized function, the NN proposed in this paper not only can recognize the classes having been learned correctly, but also can effectively reject the unknown currency. The simulation on PC shows that the NN system is effective for classification of known data and rejection of unknown data.

Important Links:

Go Back