Currency recognition based on fusion two different principal component analysis
Cui Yan-li · Computer Engineering and Applications Journal · 2009
In order to extract the most useful features of currency,a novel method of currency recognition is proposed.First of all,two kinds of Principle Component Analysis (PCA) are used to respectively reduce the dimensionalities of the original image vector space.Then,rough set attribute optimization is introduced to optimize the parameters of feature vectors.At last,canonical correlation analysis is adopted to fuse these features.The results show that the performance of the proposed method can extract the more useful features,and the recognition results are better than those of using one kind of PCA.In addition,when the number of the training set is 20,the recognition rate is 98.78%.