A Multimodal Biometric Recognition Algorithm Based on Second Generation Curvelet and 2D Gabor Filter

Abstract

Aiming at the problems of unimodal biometric recognition, such as easy interference and low recognition rate in practical application, a multimodal biometric recognition method is proposed in this paper. This method uses the second generation curvelet to extract face features and the 2D Gabor phase method to extract iris features. Then the two features are fused in the feature layer, and the fusion feature vectors are recognized by ELM classification. The experiments are carried out on the YaleB face dataset and CASIA-Iris-Lamp dataset. The experiment results show that the average accuracy of recognition of the algorithm can reach up to 99.74%, which is better than recognition methods of unimodal face and unimodal iris. The rate increases by 3.35%, which provides an effective model for multimodal biometric recognition.