Design and evaluation of a machine learning-based multimodal biometric authentication system

This research describes a multimodal biometric identification system that uses machine learning and feature-level fusion. The importance of this study has to do with the use of a person’s face and palm print to identify themselves. The performance of a multimodal biometric identification system is enhanced by machine learning. Precision, recognition rate, equal error rate, and many more assessment criteria are used to assess performance. The accuracy of the proposed multimodal system is 90.18%, the false acceptance rate (FAR) is 3.44%, and the false recognition rate (FRR) is 2.96%. The multimodal system uses score-level fusion to obtain this outcome. It is shown that a multimodal system can achieve good accuracy with low FAR and FRR.