Fusion of dorsal palm vein and palm print modalities for higher security applications

Biometrics is extensively used for person identification for various security applications. However existing systems using single biometric modality do not provide required level of accuracy and robustness. Combining various biometric modalities has proved to be more efficient in terms of enhancing accuracy. Verification using palm print modality has definitely achieved a great success as a unique and reliable biometric characteristic. Still the demand of robust and highly accurate biometric system can not be fulfilled using only palm print modality. The combined use of palm print and palm vein modality can substantially increase the robustness, anti-spoof capability and accuracy of the system. This paper presents a multimodal biometric identification system based on the fusion of Palm print and Palm vein modalities of the human hand. Here in the first module, the hand image is first preprocessed and its palm print features are extracted using wavelet decomposition technique. Palm vein features are extracted using matched filter technique. For every modality, separate matcher is used for recognition. Decisions obtained by both the matchers are logically ANDed together to recognize the person. Module 1 gives 96% accuracy with zero FAR. In the second module, a rough-to-fine hierarchical feature matching is done for efficient hand recognition. Module 2 gives 97.25% accuracy with very low FAR.