Identity Authentication Algorithm for Railway Real-Name Ticket Checking System Based on the Facial Features of Passengers
Chang Liu · 2013
In order to efficiently complete the checking process for real-name train tickets,an identity authentication algorithm based on the facial Gabor features of passengers was proposed.Firstly,the two dimensional Gabor filters were used to extract the Gabor features of face images,and the features were transformed and reconstructed.Secondly,the intra-personal differences and the extra-personal differences of the reconstructed feature matrix were calculated to acquire the initial samples.The dimension reduction of the initial samples was conducted with the principal component analysis(PCA),and then the support vector machine(SVM) classifiers were established.Finally,the identity authentication of passengers was completed by SVM classifiers.The experiments demonstrate that the proposed algorithm achieves 94.14% correct identity authentication accuracy on the F-Gnet,and has the robustness to the changes of illumination and expression.