A Novel Multimodal Biometric System using Face and FKP

Abstract

Single modality based biometric authentication systems are suffered from some security issues such as inter-class variation, spoof attacks and so on. A multimodal biometrics system overcomes these security issues by fusion of two or more modalities of an individual. In this paper, a novel multimodal biometric system is proposed that integrates face and Finger Knuckle Print (FKP) at the feature level. An image transformation algorithm is applied on the face and FKP images at sensor level. Scale Invariant Feature Transform (SIFT) algorithm is used to extract the feature vector of the transformed images. These transformed feature vectors are used for further processing. KNN and SVM classifiers are used for classification at matching level. An image transformation technique secures the original biometric modalities of a person. The experimental work has been performed on public dataset with the help of MATLAB 2017b and result shows that proposed multimodal system outperform as compared with other.