Cascade-based Multimodal Biometric Recognition System with Fingerprint and Face

In recent years, biometrics has played a vital role in protecting users' privacy and enabling secure authentication. Multimodal security, comprised of identity cards with attached unique passwords, is used to authenticate the genuine or imposter person; however, it is not a perfect security framework. Biometric traits are unique to every individual and hence proved to be very secure. This paper proposes a hybrid approach by combining cascaded and fusion-based multimodal biometric framework using fingerprint and face traits. The fingerprint and face features are extracted using the minutiae feature extraction algorithm and principal component analysis (PCA) algorithm. The hybrid approach is applied to a self -built database of around 450 fingerprints and 450 face images. The performance of the proposed hybrid system is determined using False Acceptance Rate (FAR), False Rejection Rate (FRR), Equal Error Rate (EER), and Accuracy evaluation parameters. The unimodal system of fingerprint and face at Level I and Level II yields an accuracy of 100% at a threshold value of 0.42 and 0.54, respectively. The multimodal system at Level III delivers an accuracy rate of 93.70%, 99.20%, 97.85%, and 99.50% at a threshold value of 0.75, 0.60, 0.30, and 0.60 in sum level fusion, product level fusion, min level fusion, and max level fusion schemes, respectively.