Multifunctional Biometric Authentication with Face Net or Gaussian Mixture Model for Face and Voice Recognition

Abstract

Advancements in information technology have made information security a crucial aspect of the field. Authentication plays a key role in maintaining security, requiring users to be identified through biometrics that analyze specific physiological and behavioral traits. Reliable personal recognition systems are essential for verifying the identity of individuals accessing various services, ensuring that only authorized users can utilize these services. This case study focuses on enhanced accuracy in multisensory biometric identification, namely voice and face recognition, which effectively reduces the equal mistake rate. The suggested solution uses a Gaussian mixture model for voice recognition, a FaceNet model for facial identification, and score-level fusion to determine user identity. The findings show that this new strategy has the lowest equal error rate when compared to existing methodologies.