An Algorithm for Multimodal Biometric Recognition Based on Feature Level and the Second-Generation Curvelet Transform
Deng Wan-yu · Xi'an Jiaotong Daxue xuebao · 2009
A single sample biometric recognition approach is proposed based on the feature level and curvelet transform of the second-generation to improve the recognition rate of the single modal biometric system in application.Two kinds of biometric features are used.These are the palm-print feature and the face feature.All image samples are normalized and decomposed using the combination of curvelet wavelet transform.Then the normalized curvelet wavelet-transformed face and palm-print features are combined at the feature fusion level.The K-NN classifier is used to determine the final biometric classification,and then the recognition results are reported.The experimental results show that the proposed approach has better performance than the single modal solution:the best average recognition rate is improved to 92.40%,and the recognition rate is improved by 35.38% and 8.92% compared with single face feature and single palm-print feature respectively.