Bahraini Paper Currency Recognition

Abstract

This paper presents a new technique to recognize paper currency. Bahrain paper currency has been considered as a special case, although the system can be trained for other currencies too. The recognition of paper currency has been successfully attempted to be recognized from both sides. It uses two classifiers, the weighted Euclidean distance using suitable weights and the Neural Network. The proposed technique is based on extracting some specific features of the paper currency. In addition to finding and extracting features, the technique also includes various preprocessing steps. Various factors like the image size, edge detection, the Euler number and the correlation coefficient play important role in the recognition process. The system, using weighted Euclidean distance, yields results up to 96.4% of accuracy of recognition. Whereas, the two level of feed forward back propagation neural network classifier, which has many conditions that effect on the accuracy rate, can recognize up to 95.5% of accuracy in best case of the recognition.