A Multimodal Biometric Recognition System Based on Fusion of Palmprint, Fingerprint and Face

This paper presents a multimodal biometric recognition system integrating palmprint, fingerprint and face based on score level fusion. The feature vectors are extracted independently from the pre-processed images of palmprint, fingerprint and face. The feature vectors of query images are then compared individually with the enrollment templates which are taken and stored during database preparation for each biometric trait respectively. The individual matching scores generated after matching of query images with database images are passed to the fusion module. Fusion module performs score normalization and fusion of normalized scores by weighted sum rule. Weights associated with each biometric trait for a specific user indicates the importance of corresponding biometric characteristic possessed by the user. These individual normalized scores along with their weights are finally combined into a total score by sum rule, which is passed to the decision module which declares the person as genuine or an imposter. The identity established by this system is more reliable than the identity established by individual biometric systems. Integrating multiple biometric traits improves recognition performance and reduces fraudulent access. The proposed multimodal biometric system overcomes the limitations of individual biometric systems and also meets the response time as well as the accuracy requirements.