A Novel Recognition Method for Nigeria Paper Currency Based on Support Vector Machine & Genetic Algorithms

This paper presents a novel recognition method for paper currency based on support vector machine(SVM) and genetic algorithm. Firstly the scaned original paper currency images are pre-processed by applying hough transform to obtain data sets of regulated samples, which are further divided into two collections with one as train sets and the other as test sets. The feature vectors are then extracted from the regulated images for the learning or verifying of SVM with radical basis function (RBF)[1]. Genetic algorithm is applied to optimize the parameters -c and -g of the classification model for better classifying performance. The proposed model has been used for the recognition of Nigeria paper currencies and shows promising results.